Handbook of Developmental Cognitive Neuroscience, 2nd edition (Developmental Cognitive Neuroscience)

HANDBOOK OF DEVELOPMENTAL COGNITIVE NEUROSCIENCE Second Edition

Developmental Cognitive Neuroscience Neurodevelopmental Disorders, Helen Tager-Flusberg, ed. (1999) Handbook of Developmental Cognitive Neuroscience, Charles A. Nelson and Monica Luciana, eds. (2001) Modeling Neural Development, Arjen van Ooyen, ed. (2003) Handbook of Developmental Cognitive Neuroscience, second edition, Charles A. Nelson and Monica Luciana, eds. (2008)

HANDBOOK OF DEVELOPMENTAL COGNITIVE NEUROSCIENCE Second Edition

Edited by Charles A. Nelson and Monica Luciana

A BRADFORD BOOK THE MIT PRESS CAMBRIDGE, MASSACHUSETTS LONDON, ENGLAND

© 2008 Massachusetts Institute of Technology All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher. MIT Press books may be purchased at special quantity discounts for business or sales promotional use. For information, please e-mail [email protected] or write to Special Sales Department, The MIT Press, 55 Hayward Street, Cambridge, MA 02142. This book was set in Baskerville by SNP Best-set Typesetter Ltd., Hong Kong and was printed and bound in the United States of America. Library of Congress Cataloging-in-Publication Data Handbook of developmental cognitive neuroscience / edited by Charles A. Nelson and Monica Luciana.—2nd ed. p. ; cm.—(Developmental cognitive neuroscience) Includes bibliographical references and index. ISBN 978-0-262-14104-8 (hardcover : alk. paper) 1. Developmental neurobiology—Handbooks, manuals, etc. 2. Cognitive neuroscience—Handbooks, manuals, etc. I. Nelson, Charles A. (Charles Alexander) II. Luciana, Monica. III. Series. [DNLM: 1. Nervous System—growth & development. 2. Central Nervous System Diseases—physiopathology. 3. Cognition—physiology. 4. Human Development. 5. Perception—physiology. WL 102 H23535 2008] QP363.5.H365 2008 612.8'2—dc22 2008007886 10 9 8 7 6 5 4 3 2 1

CONTENTS

Preface to the Second Edition

I.

xi

FUNDAMENTALS OF DEVELOPMENTAL NEUROBIOLOGY 1

A. GENERAL PRINCIPLES

3

1.

The Formation of Axons and Dendrites by Developing Neurons Paul Letourneau

2.

Imaging Developmental Changes in Gray and White Matter in the Human Brain 23 Elizabeth D. O’Hare and Elizabeth R. Sowell

3.

Gyrification and Development of the Human Brain Tonya White and Claus C. Hilgetag

4.

Adult Neurogenesis in the Hippocampus 51 Yevgenia Kozorovitskiy and Elizabeth Gould

5.

The LHPA System and Neurobehavioral Development 63 Amanda R. Tarullo, Karina Quevedo, and Megan R. Gunnar

6.

The Effects of Monoamines on the Developing Nervous System Gregg D. Stanwood and Pat Levitt

5

39

83

v

B. STRUCTURAL FOUNDATIONS OF SENSATION, PERCEPTION, AND COGNITION 95 7.

Mechanisms of Auditory Reorganization during Development: From Sounds to Words 97 Richard N. Aslin, Meghan A. Clayards, and Neil P. Bardhan

8.

Brain Correlates of Language Processing during the First Years of Life 117 Angela D. Friederici

9.

Brain-Behavior Relationships in Early Visual Development Bogdan F. Iliescu and James L. Dannemiller

127

10. Motor Systems Development 147 Rosa M. Angulo-Barroso and Chad W. Tiernan 11. Neurodevelopment of Social Cognition 161 Melissa D. Bauman and David G. Amaral 12. Pre- and Postnatal Morphological Development of the Human Hippocampal Formation 187 László Seress and Hajnalka Ábrahám 13. Structural Development of the Human Prefrontal Cortex Ivica Kostovic´, Milosˇ Judasˇ, and Zdravko Petanjek

213

14. White Matter Maturation and Cognitive Development during Childhood 237 Torkel Klingberg

II. METHODOLOGICAL PARADIGMS

245

15. Electrophysiological Methods in Studying Infant Cognitive Development 247 Gergely Csibra, Elena Kushnerenko, and Tobias Grossmann 16. Eye Tracking Studies of Normative and Atypical Development Canan Karatekin

263

17. Diffusion Tensor Imaging 301 Jeffrey R. Wozniak, Bryon A. Mueller, and Kelvin O. Lim 18. Functional MRI Methods in Developmental Cognitive Neuroscience Kathleen M. Thomas and Angela Tseng

311

19. Mechanisms of Language Acquisition: Imaging and Behavioral Evidence 325 Jacques Mehler, Marina Nespor, Judit Gervain, Ansgar Endress, and Mohinish Shukla

vi

contents

20. Magnetic Resonance Spectroscopy of Developing Brain Balasrinivasa Rao Sajja and Ponnada A. Narayana

337

21. The Integration of Neuroimaging and Molecular Genetics in the Study of Developmental Cognitive Neuroscience 351 Essi Viding, Douglas E. Williamson, Erika E. Forbes, and Ahmad R. Hariri 22. Neural Network Models of Cognitive Development 367 Yuko Munakata, Jennifer Merva Stedron, Christopher H. Chatham, and Maria Kharitonova

III. NEURAL PLASTICITY IN DEVELOPMENT

383

23. Early Brain Injury, Plasticity, and Behavior 385 Bryan Kolb, Wendy Comeau, and Robbin Gibb 24. Developmental Plasticity and Reorganization of Function Following Early Diffuse Brain Injury 399 Linda Ewing-Cobbs, Mary R. Prasad, and Khader M. Hasan 25. Plasticity of the Visual System 415 Daphne Maurer, Terri L. Lewis, and Catherine J. Mondloch 26. Cross-Modal Plasticity in Development: The Case of Deafness Teresa V. Mitchell 27. Plasticity of Speech (Animal Model) Teresa A. Nick

IV. COGNITION

439

453

465

28. The Development and Integration of the Dorsal and Ventral Visual Pathways in Object Processing 467 Mark H. Johnson, Denis Mareschal, and Gergely Csibra 29. Attention in Young Infants: A Developmental Psychophysiological Perspective 479 John E. Richards 30. Nonhuman Primate Models of Memory Development Jocelyne Bachevalier

499

31. Neurocognitive Mechanisms for the Development of Face Processing 509 Michelle de Haan 32. The Development of Visuospatial Processing Joan Stiles, Brianna Paul, and Wendy Ark

521

contents

vii

33. Mechanisms of Change: A Cognitive Neuroscience Approach to Declarative Memory Development 541 Jenny Richmond and Charles A. Nelson 34. The Development of Executive Function in Childhood Philip David Zelazo, Stephanie M. Carlson, and Amanda Kesek

553

35. The Development of Prefrontal Cortex Functions in Adolescence: Theoretical Models and a Possible Dissociation of Dorsal versus Ventral Subregions 575 Elizabeth A. Olson and Monica Luciana 36. Cognition and Aging: Typical Development 591 Jonas Persson and Patricia A. Reuter-Lorenz 37. Cognition and Aging-Dementia 607 Mischa de Rover, Sharon Morein-Zamir, Andrew D. Blackwell, and Barbara J. Sahakian

V. NEURODEVELOPMENTAL ASPECTS OF CLINICAL DISORDERS 621 38. The Role of Nutrition in Cognitive Development 623 Anita J. Fuglestad, Raghavendra Rao, and Michael K. Georgieff 39. Fetal Alcohol Syndrome 643 Sarah N. Mattson, Susanna L. Fryer, Christie L. McGee, and Edward P. Riley 40. Impact of Prenatal Cocaine Exposure on the Developing Nervous System 653 Eric M. Langlois and Linda C. Mayes 41. Neurocognitive Models of Early-Treated Phenylketonuria: Insights from Meta-analysis and New Molecular Genetic Findings 677 Marilyn Welsh, Kathryn DeRoche, and David Gilliam 42. Research into Williams Syndrome: The State of the Art Annette Karmiloff-Smith

691

43. Neurocognitive Development in Autism 701 Mikle South, Sally Ozonoff, and Robert T. Schultz 44. Tics and Compulsions: Disturbances of Self-Regulatory Control in the Development of Habitual Behaviors 717 Rachel Marsh, James F. Leckman, Michael H. Bloch, Yanki Yazgan, and Bradley S. Peterson 45. Developmental Dyslexia 739 Guinevere F. Eden and D. Lynn Flowers

viii

contents

46. The Development and Cognitive Neuroscience of Anxiety Daniel S. Pine and Christopher S. Monk

755

47. Developmental Neuropsychology of Unipolar Depressions Ian M. Goodyer and Zoë Kyte

771

VI. EMOTION/COGNITION INTERACTIONS

785

48. Toward a Neurobiology of Attachment 787 Myron A. Hofer and Regina M. Sullivan 49. Sleep, Cognition, and Emotion: A Developmental View Oskar G. Jenni and Ronald E. Dahl

807

50. Neural Systems, Gaze Following, and the Development of Joint Attention 819 Peter Mundy and Amy Van Hecke 51. The Biology of Temperament: An Integrative Approach 839 Nathan A. Fox, Heather A. Henderson, Koraly Pérez-Edgar, and Lauren K. White 52. The Developing Adolescent Brain: Molecular Mechanisms Underlying Nicotine Vulnerability 855 Charles F. Landry, Terri L. Schochet, and Ann E. Kelley 53. Environmental Influences on Brain-Behavioral Development: Evidence from Child Abuse and Neglect 869 Jessica E. Shackman, Alison B. Wismer Fries, and Seth D. Pollak 54. Neurocognitive Development of Performance Monitoring and Decision Making 883 Eveline A. Crone and Maurits W. van der Molen

Contributors Index

897

901

contents

ix

Preface to the Second Edition The first edition of this Handbook appeared in 2001, at a time when the field of developmental cognitive neuroscience had only recently taken root. The volume contained fortyone chapters distributed over eight topical areas, including overviews of the fundamentals of developmental neurobiology, a surveying of methodological paradigms, neural plasticity and its expression during development and in the context of disease, sensory and motor system development, language development, cognition (broadly construed), neurodevelopmental aspects of clinical disorders, and emotion-cognition interactions. Seven years have now passed since the first edition was published and the field of developmental cognitive neuroscience has expanded enormously. To illustrate how the field has grown, we recently conducted Medline searches spanning the interval from 1902 to 2007 using the following three search parameters combined: brain, development, and cognition. From 1902 to 2001, there were 972 articles that represented this intersection of topics. From 2002 to 2007 alone, there were 988 articles. Thus, developmental cognitive neuroscience, following the pattern of its parent discipline, cognitive neuroscience, is growing at an exponential rate, with evidence of massive proliferation over the past five years. Many accomplishments within the field have resulted from the application of new methods to developmental samples. This proliferation of activity is also evident through other, more concrete, indices of change, including (a) an exponential increase in the number of developmental papers published in the Journal of Cognitive Neuroscience, (b) the appearance of special issues on this topic in a number of other journals, including Developmental Review, Child Development, Human Development, Neuropsychologia, and Developmental Psychology, (c) authored and edited volumes by a number of senior investigators (e.g., Mark Johnson, Michelle de Haan), and finally (d) the ease with which we were able to expand this volume. We have expanded to fifty-four chapters from the original forty-one. More importantly, we now present a number of areas that in our view represent new inroads made possible by advances in both developmental and cognitive neuroscience. First, there is a greater emphasis on affective and social neuroscience. This offshoot of cognitive neuroscience has firmly taken root in the adult literature and is gradually trickling down to the developmental literature. Second, we have placed a greater emphasis on clinical disorders. We have done so primarily because such work is inherently translational in nature, and translational research is currently receiving a great deal of attention by many working at the interface of brain and behavior. Finally, one chapter exclusively, and several to a lesser degree, discuss the breakthroughs being made in imaging genomics. In our mind the intersection of brain, behavior, and genetics represents an exciting new area of inquiry that will gain considerable traction in coming years, due in large part to advances being made in genetics/genomics and in neuroimaging. We are pleased to bring you this second edition and trust that it will serve as a resource for all those interested in the development of brain-behavior relations in the context of both typical and atypical development. Charles A. Nelson Monica Luciana

xi

I FUNDAMENTALS OF DEVELOPMENTAL NEUROBIOLOGY

A. GENERAL PRINCIPLES

1

The Formation of Axons and Dendrites by Developing Neurons PAUL LETOURNEAU

Introduction The neuronal circuitry that underlies human behavior and other neural functions develops over a prolonged period lasting from the second fetal month through adolescent years. These circuits arise from the extensive development of elaborate neuronal processes, as neurons express intrinsic morphogenetic behaviors, while interacting with other cells and molecules of the developing nervous system. First, immature neurons migrate from their birthplaces to the sites where they are organized into layers, nuclei, and ganglia of neuronal perikarya. Next, immature neurons sprout axons and dendrites that elongate, sometimes for many centimeters, to make synaptic connections with target neurons or other cells. This chapter describes intrinsic mechanisms of morphogenesis of axons and dendrites and the extrinsic environmental features that regulate where and when axons and dendrites grow to create neural circuits. The ability to extend neuronal processes, or neurites, is intrinsic to neurons. This is demonstrated when immature neurons, such as from prenatal hippocampus, are placed into tissue culture. Within a few hours the neurons sprout processes that elongate onto the substrate, each tipped by an adherent motile structure called a growth cone. These neurites mature to become axons and dendrites and form synapses in vitro. These events in a neutral in vitro environment show that the neuronal phenotype defines the intrinsic behaviors that produce neuronal shape. The most significant cellular components in neuronal morphogenesis are the protein polymers of the neuronal cytoskeleton. In the next section the neuronal cytoskeleton and the intrinsic mechanisms of neurite formation and elongation will be discussed. In the following three sections the regulation of axonal and dendritic growth by extrinsic molecules will be discussed.

The dynamic neuronal cytoskeleton Neuronal morphogenesis depends on the organization and dynamic properties of two cytoskeletal polymers, microtubules and actin filaments (Dent and Gertler, 2003; Luo, 2002). These cytoskeletal polymers are present in all cell types, although specific mechanisms determine cytoskeletal functions in neurons.

Microtubules Provide Support and a Means of Transport Microtubules are hollow cylinders 25 nm in diameter that extend through the cytoplasm of neuronal perikarya, axons, and dendrites (figures 1.1, 1.2). The wall of a microtubule consists of subunits of highly conserved proteins, alpha tubulin and beta tubulin. Microtubules have no defined length, and single neuronal microtubules can exceed 100 μm (Letourneau, 1982). Microtubules are rigid and resist compression to support the elaborate extensions of axons and dendrites. Microtubules are also the “rails” along which organelles are transported via linkage to the motor proteins, kinesins, and dynein (Hirokawa and Takemura, 2004). These two functions, providing structural support and being rails for intracellular transport, are the functions of neuronal microtubules. Formation of Microtubules in Cells Tubulin subunits polymerize by endwise addition to form microtubules. Because of inherent asymmetry of the tubulin protein, microtubules are polarized with a distinct molecular face at each end. Tubulin subunits are added more rapidly at one end, called the plus (+) end, while the less likely end for growth is called the minus (−) end (figure 1.1 and plate 1). Microtubules in neurons are formed in the centrosomal region near the nucleus and extend throughout the perikaryon with their minus ends anchored at the centrosome. The plus ends of cytoplasmic microtubules undergo bouts of growing and shrinking called dynamic instability, in which a microtubule end may undergo rapid disassembly, either completely or partially, which is followed by “rescue” and renewed growth (Tanaka, Ho, and Kirschner, 1995). Regulation of Microtubule Organization by MAPs In neurons, microtubule organization is regulated by a group of proteins called MAPs (microtubule-associated proteins). MAPs bind to microtubules and regulate all aspects of their organization, including assembly and disassembly, stability, and binding to neurofilaments, actin filaments, and other microtubules (Dehmelt and Halpain, 2004; Gordon-Weeks, 2000). Motor proteins, such as kinesin, bind to microtubules and move cargo toward microtubule plus ends, while dynein motors move cargo toward microtubule minus ends. The protein katanin binds microtubules and

5

Figure 1.1 Actin filaments and microtubules are polarized polymers. Actin filaments are polarized polymers for which the addition of ATP-actin is more likely at the barbed end than the pointed end. After hydrolysis of ATP-actin to ADP-actin, subunits dissociate at the pointed end. Microtubules are also polarized structures with

GTP-tubulin dimers adding to the plus or growing end and GDPtubulin dimers dissociating from the minus end. Microtubules also exhibit posttranslational modifications (detyrosination shown here) that correlate with the age and stability of the polymer. (From Dent and Gertler, 2003.) (See plate 1.)

severs them, promoting reorganization of microtubules and remodeling of neuronal shape (Baas and Buster, 2004). Some maps, such as MAP2, are localized in dendrites, while other MAPs, such as tau and MAP1B, are localized in axons. Several features distinguish microtubules in axons and dendrites. Unlike most cell types, the minus ends of microtubules in axons and dendrites are not anchored to the centrosome; rather, microtubules lie entirely within these processes. Microtubules are formed at the centrosome and then transported into axons or dendrites. Nearly all axonal microtubules have their plus ends oriented toward the terminal, while microtubules in dendrites have mixed polarity, some with plus ends and some with minus ends oriented toward dendritic termini. Many axonal and dendritic microtubules are highly stable as a result of enzymatic modifications of the tubulin protein and from binding of certain MAPs. Although microtubules must always be present to support neurites, it is uncertain how microtubules and tubulin subunits are advanced as neurites grow (Baas and Buster, 2004). Dynein motor molecules can slide short microtubules along,

depending on microtubule length and connections with other structures. Long microtubules in axons are stationary, although their plus ends undergo considerable dynamic instability of growth and shrinkage. Possibly, tubulin subunits or short microtubules are transported distally via dynein motors and then disassembled to release tubulin for addition to longer, stable microtubules. This dynamic assembly of tubulin onto existing microtubules is a critical event in the morphogenesis of axons and dendrites (Tanaka and Kirschner, 1995).

6

Actin Filaments in Neurons Actin filaments are the other important cytoskeletal components in neuronal morphogenesis (Dent and Gertler, 2003; Luo, 2002). In mature neurons, actin filaments form a cortical meshwork beneath the plasma membrane that organizes ion channels, vesicles, membrane proteins, and neurotransmitter receptors at nodes of Ranvier and at synapses. However, at the ends of growing axons and dendrites, elaborate networks of actin filaments are the organizing component that drives the searching behaviors that are necessary for navigation of

fundamentals of developmental neurobiology

Figure 1.2 The distribution of microtubules and actin filaments in developing neurons and in axonal growth cones. Microtubules (green) are densely packed with the neuronal cell bodies (S) and are bundled in the axons and branches. Actin filaments are arrayed in filament networks and bundles in the peripheral domains (P) of the

growth cones and along the shafts of the axons, where small areas of actin filament dynamics may give rise to collateral branches (B). In a growth cone, the microtubules from the central bundle of the central domain (C) splay apart, and individual microtubules extend into the P domain and into filopodia (arrows). (See plate 2.)

growth cones to their synaptic targets (figure 1.2 and plate 2; Letourneau, 1979, 1983; Yamada, Spooner, and Wessells, 1971).

Like microtubule polymerization, actin filaments polymerize by endwise addition of subunits. Also, like microtubules, the inherent asymmetry of the actin subunit leads to polarity of actin filaments, in which the “barbed” end is favored for polymerization and the “pointed” end is where actin subunits are lost from filaments. Again, like microtubules, neurons contain many proteins, whose function is to regulate the polymerization, stability, and interactions of actin filaments.

Organization of Actin in Cells Actin filaments are polymers of the conserved globular protein actin (figure 1.1). Actin filaments with a diameter of about 6–7 nm are individually not stiff, but bundles of actin filaments have stiffness. Unlike the cortical networks in mature neurons, actin filament arrays in growth cones are extensive, especially at the motile leading margin, where a dynamic actin filament network fills flattened projections, called lamellipodia, and bundles of actin filaments fill the cores of transient, fingerlike projections, called filopodia (figure 1.2; Letourneau, 1983).

Regulation of Actin Filament Organization by ABPs Actin-binding proteins (ABPs) have numerous functions (Dent and Gertler, 2003; Pollard and Borisy, 2003). One class of ABPs binds actin subunits, regulating

letourneau: formation of axons and dendrites by developing neurons

7

their availability for polymerization; other ABPs cross-link actin filaments into meshworks and bundles. ABPs that bind the barbed and pointed ends of actin filaments regulate the addition and loss of actin subunits to filaments. Several ABPs bind actin filaments and sever them, promoting the remodeling of actin filament arrays. In growth cones, actin filament barbed ends face the leading cell margin, where the addition of actin subunits is promoted by several ABPs. Myosins are motor molecules that bind and move cargoes along actin filaments. There are more than 10 myosins, which share common features of their motor activity, but which differ in the direction that they move cargoes along filaments and in cargoes that are moved (Brown and Bridgman, 2004). Myosins in growth cones interact with actin filaments and generate forces to move actin filaments, vesicles, or other cargoes and to exert tensions on cytoskeletal components and associated structures (Rochlin et al., 1995). Myosin II in growth cones is particularly important in

generating developing critical to developing

Figure 1.3 The interwoven network of signaling molecules that link guidance receptors with cytoskeletal dynamics underlying growth cone motility. Membrane receptors for extracellular guidance cues may function either alone or in a complex to activate cytoplasmic adaptors and mediators. The Rho family of GTPases may be pivotal links between guidance signals and actin-associated

proteins, which are responsible for regulating the assembly and disassembly of actin filaments. Similar types of molecules are represented by symbols of similar color and shape. Lines depict activation pathways that have been demonstrated experimentally in different systems. (From Song and Poo, 2001.)

8

forces to move components and reshape axons and dendrites. In summary, ABPs are regulating the behaviors of growth cones of axons and dendrites.

Regulation of Microtubule and Actin Organization and Dynamics by Cytoplasmic Signaling Pathways As noted previously, the organization of microtubules and actin filaments is regulated by MAPs and ABPs. The dynamic changes in cytoskeletal organization that drive neuronal morphogenesis reflect the activities of MAPs and ABPs. Certainly, levels of these proteins are regulated by gene transcription and protein synthesis, but in an immediate fashion, MAPs and ABPs are regulated by intracellular signaling and cytoplasmic second-messenger pathways. Cytoskeletal organization can be rapidly changed by fluctuations in levels of small molecules such as Ca++ ions, cAMP, cGMP, and phosphoinositides that bind MAPs and

fundamentals of developmental neurobiology

ABPs and regulate them allosterically (Dent and Gertler, 2003; Song and Poo, 2001; figure 1.3). The addition to and removal of phosphate groups from MAPs and ABPs by protein kinases and phosphates also rapidly regulate their activities. These molecules and pathways are, in turn, regulated by events at the plasma membrane, where adhesive proteins, growth factors, and other ligands bind membrane receptor proteins to trigger events that locally and temporally modulate the levels and activities of these regulatory molecules. Thus cytoplasmic signaling activities that cascade from ligand-receptor interactions at the plasma membrane rapidly and locally regulate cytoskeletal organization during neuronal morphogenesis (Dent and Gertler, 2003; Gallo and Letourneau, 2004). The Rho family of small guanosine triphosphatase (GTPase) proteins, in particular RhoA, Rac1, and Cdc42, are important regulatory proteins that relay signaling from the cell surface intracellularly to the cytoskeleton (Jaffe and Hall, 2005; figure 1.3). Rho GTPases bind to and regulate MAPs and ABPs or their upstream regulators, such as protein kinases and phosphatases. A critical feature of GTPases is that their activity is rapidly switched on or off, depending on whether they are bound to the nucleotides GTP (on) or guanosine diphosphate (GDP) (off). A rich variety of guanine nucleotide exchange factor proteins (GEFs) selectively activate GTPases by exchanging GDP for GTP; GTPase-activating proteins (GAPs) stimulate hydrolysis of GTP to inactive GTPases; and GDP dissociation inhibitors (GDIs) inhibit activation of GTPases by GEFs. These GEFs, GAPs, and GDIs are regulated by cell surface ligand-receptor interactions. Thus by regulating GTPases these membrane events regulate cytoskeletal proteins. Activation of RhoA, Rac1, or Cdc42 has distinct effects on actin filament organization (Jaffe and Hall, 2005). Rac1GTP activates several ABPs to stimulate actin polymerization and formation of lamellipodia, while Cdc42-GTP also stimulates actin polymerization and formation of filopodia. RhoA-GTP activates the kinase ROCK, which phosphorylates several substrates to suppress actin polymerization and activates the motor protein myosin II, increasing mechanical tensions and rearrangements of actin filaments. If RhoA levels are highly elevated, strong contractile forces in the growth cone can cause collapse of microtubule arrays and significant neurite retraction. All three Rho GTPases are present in the growth cone and contribute to growth cone motility. Microtubule organization and polymerization are also regulated by Rho GTPases, although the mechanisms are less well understood than for actin filaments. Microtubule-Actin Interactions Are Important Two particular interactions of actin filaments, which we will describe, are particularly important in neurite elongation and growth cone migration. As mentioned earlier,

microtubules maintain the shapes of axons and dendrites and resist compressive forces that would collapse or withdraw these processes. Proteins that mediate interactions between microtubule plus ends and actin filaments are particularly significant, because these proteins may be important in the initiation of neurites from a spherical perikaryon or in directing the advance of a growth cone (Rodriquez et al., 2003). These microtubule-actin interactions link the microtubule functions of structural support and organelle transport to the dynamic cortical actin filaments and associated membrane receptors that detect extrinsic signals and regulate the cytoskeletal activities that shape the developing neuron (figure 1.4). A Stage 1.1

Stage 1.2

Stage 1.3

Stage 2

B MAPs stabilize microtubules MAPs might promote microtubule bundling MAPs might promote microtubule/actin interactions

MAPs Actin Microtubules

Figure 1.4 A model for cytoskeletal reorganization during neurite initiation. (A) Shortly after plating, cultured hippocampal neurons extend a uniform lamellipodium that surrounds the cell soma (stage 1.1). Preceding the initial neurite outgrowth, the lamellipodium becomes segmented at one or more sites (stage 1.2). Then the lamellipodium migrates away from the cell soma to form a growth cone, concurrent with microtubule advance into the initiation site and formation of an ordered microtubule array (stage 1.3). The newly formed protrusion then elongates, and microtubules become tightly packed into parallel arrays inside the nascent neurite (stage 2). Actin-microtubule interactions are present in lamellipodia at all stages. Panel B depicts a more detailed view of the proposed cytoskeletal organization in stage 1.3. Interestingly, microtubules preferentially grow along actin bundles in filopodia, suggesting that a physical link between the structures exists. Multifunctional MAPs like MAP1B, MAP2, or plakins are candidates to act as such links. (From Dehmelt and Halpain, 2004.)

letourneau: formation of axons and dendrites by developing neurons

9

Actin Filaments and Adhesive Contacts In addition to interactions with microtubules, another key function of actin filaments involves the adhesive interactions of cells that are mediated by membrane receptor proteins that form noncovalent bonds between cells or between cells and extracellular matrices (ECM). The major adhesion receptors are the cadherins and the adhesion proteins of the immunoglobulin-like superfamily, which mediate cell-cell adhesions, and the integrin proteins, which mediate cell adhesion to ECM. As cell-cell contacts are initiated by intercellular binding, receptors cluster within the plasma membrane to form discrete adhesive contacts. By way of transmembrane linkage these clustered adhesion receptors create docking sites for signaling enzymes, kinases, GEFs, GAPs, and a number of ABPs that link actin filaments to the adhesive sites and induce actin polymerization. Thus adhesive sites are loci from which regulatory signals emanate and where actin filament organization and anchorage are regulated (Zamir and Geiger, 2001).

A mechanism for neurite initiation and growth In this section, neuritogenesis, neurite elongation, and growth cone migration by neurons will be described, emphasizing the dynamic cytoskeleton of actin filaments and microtubules. When developing neurons are placed in culture, the neurons settle on the substrate, and extend and withdraw cylindrical filopodia and flattened lamellipodia, like waves lapping on a beach (see figure 1.4). This motility is driven by actin filament polymerization, which pushes the cell margin outward, while simultaneously myosin II, located behind the cell margin, pulls newly formed filaments backward in a retrograde flow. The rearward transported filaments are severed and depolymerized, and if the protrusion and retrograde flow are equal, these activities produce no net change. Initially, microtubules remain in a loose network around the nucleus, and any microtubules that enter the protrusions are swept back with the retrograde flow of actin. However, eventually a filopodium or lamellipodium thickens and moves away from the cell body, tethered by a cylindrical nascent neurite. The critical step that distinguishes neurite formation from the initial protrusive activity occurs when microtubules and associated organelles enter and remain within a filopodial or lamellipodial protrusion and the protrusive motility moves forward ahead of the microtubules and organelles (Da Silva and Dotti, 2002; figure 1.4). Several activities may prompt neurite initiation. An increased expression of MAPs, such as MAP2, tau, and MAP1B, may stabilize microtubules, enhancing their resistance to the myosin-based retrograde forces pulling actin back from the leading margin (Dehmelt and Halpain, 2004). At sites where protrusions make firm adhesive contacts with the substrate, actin filaments become anchored to the adhesive apparatus,

10

and retrograde flow stops, creating space into which microtubules can advance. In addition, cytoplasmic signals generated at the adhesive sites may promote microtubule transport and polymerization. Finally, actin filaments linked to adhesive sites can interact with myosin II motors and pull microtubules and organelles toward the adhesive sites in opposition to the retrograde flow of untethered actin filaments (Suter and Forscher, 2000). The significance of these outwardly directed forces in neurite initiation is illustrated by findings that neurites can be pulled out from a neuron by attaching an adhesive bead to a neuronal surface and then pulling the bead and attached elongating neurite away from the nerve cell body (Fass and Odde, 2003). Organization of Growth Cones and Growth Cone Migration A typical neurite has a central bundle of microtubules with associated organelles and a motile terminal expansion, the growth cone (Gordon-Weeks, 2000; figures 1.2, 1.5). At the growth cone’s leading margin, called the P-domain (peripheral), vigorous actin polymerization pushes the cell margin forward, balanced by the myosinpowered rearward sliding of untethered actin filaments. Only when the leading edge forms transient adhesive contacts that link to actin filaments does the retrograde flow attenuate. At the base of a growth cone, microtubulebased motor proteins move microtubules and organelles from the neurite into the central growth cone, comprising the C-domain (central). From the C-domain individual microtubules extend into the P-domain, sliding forward powered by molecular motors and elongating by adding subunits to microtubule plus ends. Retrograde flow pulls most of these microtubules back into the C-domain (Schaefer, Kabir, and Forscher, 2002). Importantly, some microtubules advance into filopodia or lamellipodia, stabilized at adhesive sites (figure 1.2; Letourneau, 1979; Suter and Forscher, 2000). If these microtubules persist and are followed by other microtubules and organelles, the Cdomain advances and the neurite extends. To complete the cycle of growth cone movement, actin filaments and membrane components that are not stabilized by adhesions or associations with microtubules are recycled at the back of the growth cone by the myosin II–powered retrograde flow and by disassembly of actin filaments and endocytosis of plasma membrane. Thus neurite elongation proceeds by three activities (Dent and Gertler, 2003; figure 1.5): (1) the advance, expansion, and adhesion of the leading margin of the growth cone, driven by actin polymerization; (2) the advance of microtubules via polymerization, transport, and linkage to actin and adhesive sites (Letourneau, 1979); and (3) the advance of organelles via microtubule-based transport. The coordination of actin-driven membrane expansion, formation of adhesive contacts, and myosin II–powered exertion of

fundamentals of developmental neurobiology

Figure 1.5 Stages of axon and branch growth. Three stages of axon outgrowth have been termed protrusion, engorgement, and consolidation (Goldberg and Burmeister, 1986). Protrusion occurs with the rapid extension of filopodia and thin lamellar protrusions, often between filopodia. These extensions are primarily composed of bundled and meshlike F-actin networks. Engorgement occurs when microtubules invade protrusions bringing membranous

vesicles and organelles (mitochondria, endoplasmic reticulum). Consolidation occurs when the majority of F-actin depolymerizes in the neck of the growth cone, allowing the membrane to shrink around the bundle of microtubules, forming a cylindrical axon shaft. This process also occurs during the formation of collateral branches off the growth cone or axon shaft. (From Dent and Gertler, 2003.)

tension on these adhesive sites generates a force that pulls the growth cone forward. Thus neurite elongation involves “push” from the advance of microtubules and “pull” from myosin II–powered tension generated at adhesive sites at the growth cone margin (Lamoureux, Buxbaum, and

Heidemann, 1989; Letourneau, 1981; Letourneau et al., 1987). Experimental studies show that the “push” of microtubule advance is necessary for neurite elongation, while the “pull” of actin-based motility in growth cones is neither necessary nor sufficient for neurite elongation. However,

letourneau: formation of axons and dendrites by developing neurons

11

growth cone “pull” accelerates neurite elongation and, as described later, is necessary for growth cone navigation. Growth Cone Turning Growth cone navigation to synaptic targets occurs by the selective turning, advance, or retreat of a growth cone in response to guidance cues that a growth cone encounters within developing tissues. As described previously, a neurite elongates by the advance of microtubules and organelles from the growth cone Cdomain into the P-domain. In a neutral in vitro environment, this elongation may occur first to one side and then to the other, keeping the growth cone on a straight path. In a complex in vivo environment, however, there are local differences in adhesive surfaces, extrinsic factors, or other ligands that interact with growth cone receptors to generate local differences in the activities of Rho GTPases, protein kinases, protein phosphatases, or second messengers, such as Ca++ or cyclic nucleotides (Gomez and Zheng, 2006; Guan and Rao, 2003; Song and Poo, 2001). On

Figure 1.6 Summary of reorganization of actin filaments and microtubules in the peripheral domain of growth cones that is involved in turning toward an attractant and turning away from a repellent guidance molecule. An attractant promotes actin

12

the one hand, if these local variations in regulatory cues are sufficiently strong or persistent, they produce local differences in actin-based motility and microtubule advance that cause growth cone turning. This outcome might occur because the P-domain expands faster on one side as a result of locally enhanced actin polymerization, reduced retrograde actin flow, or linkage of actin filaments to adhesive sites (figure 1.6, upper panels). Localized signals might directly promote microtubule polymerization or stabilization, so that microtubules preferentially advance to one side of the P-domain (Challacombe et al., 1997; Dickson, 2002; Tanaka and Kirschner, 1995). On the other hand (see figure 1.6, lower panels), if local differences in signals triggered by extrinsic cues reduce actin-mediated protrusion on one side of a growth cone or if myosin II– powered retrograde flow of actin filaments increases on one side of a growth cone, microtubule advance on that side will be reduced, and the growth cone will turn toward the other side.

polymerization, adhesion, and microtubule advance, while a repellent inhibits actin polymerization and advance of microtubules. (Adapted from Dickson, 2002.)

fundamentals of developmental neurobiology

Mechanisms of Branching Branches of neurites, axons, or dendrites are formed in two ways: by a growth cone splitting or by a new branch sprouting from the neurite shaft behind a growth cone. In either case, the acquisition of stable microtubules is key to forming a branch (figures 1.2, 1.5). In a growth cone, part of the P-domain and associated C-domain may separate from the whole and establish an independent growth cone and a new branch of the parent neurite. This result may occur when a growth cone “pulls” in two directions (figure 1.2). Branch formation along a neurite is initiated by localized protrusion of filopodia or lamellipodia (figure 1.2; Gallo and Letourneau, 1998). This mechanism is particularly prevalent in the branching morphogenesis of dendrites. This localized actin-based motility may occur until microtubules enter an actin-filled nascent branch by transport or by polymerization of microtubules from the main neurite (Gallo and Letourneau, 1999). Microtubule ends in the main neurite may become linked to actin filaments of the protrusion and be pulled into the branch. The microtubule-severing protein katanin may promote branch formation by severing microtubules in the neurite shaft to create microtubule ends that can be moved into a nascent branch (Baas and Buster, 2004). Once stable microtubules are established, the advance of microtubules and organelles into the branch sustains its growth. The Differentiation of Axons and Dendrites; Polarization of Neuronal Form A hippocampal neuron in vitro initially sprouts several similar neurites that extend slowly. After 18–24 hours one neurite expands its growth cone and elongates significantly faster than the others. This neurite becomes the axon, and it accumulates proteins typical of axons, such as the MAPs tau and MAP1B, and GAP43, a protein involved in actin motility (Mandell and Banker, 1996). Several molecules and pathways may be critical to axonal specification, including PI3 kinase, the Par complex, and small Rho GTPases (Arimura and Kaibuchi, 2005; Wiggin et al., 2005). These molecules concentrate at the tips of newly specified axons and are implicated in regulating key activities, such as actin filament organization, microtubule polymerization or stability, and transport and addition of plasma membrane components. It is unclear whether axonal specification always begins with the same upstream event, such as concentration of PI3 kinase activity in a neurite tip, or whether concentration of any of the previously mentioned molecules or signals is sufficient to specify axonal character. In vitro manipulations, such as focally pulling on a neurite or presenting adhesive proteins to one neurite will induce a neurite to become the axon. Thus extrinsic signals can influence the intrinsic mechanism of axonal specification, perhaps by locally activating PI3 kinase or other components of the mechanism. After one neurite becomes the axon, the other neurites become

dendrites. Less is known about the mechanisms of dendrite specification. Acquisition of microtubules with mixed polarities may be important, as well as localization of cytoskeletal, membrane, and signaling components that regulate dendritic characteristics.

Regulation of neuronal morphogenesis in vivo The previous section focused on the intrinsic mechanisms of neurite initiation and elongation, growth cone migration and turning, neurite branching, and the specification of axons. This section will discuss the roles of extrinsic molecules and signaling events in regulating neuronal morphogenesis in the developing human brain. The neutral environment of a tissue culture dish facilitates understanding these intrinsic mechanisms. However, the in vivo environment is never neutral, and spatial and temporal patterns of distribution of axonal guidance cues in the environment of the developing brain shape these intrinsic morphogenetic mechanisms to generate neural circuits (Tessier-Lavigne and Goodman, 1996). Neuronal Migration Immature neurons arise from proliferation of neural precursors in the ventricular zone of the developing brain. From their birth immature neurons become polarized by asymmetry in local cues, including the adhesive protein laminin in the underlying extracellular matrix (ECM) of the ventricular layer, as well as growth factors, morphogens, and guidance molecules, such as sonic hedgehog and netrin, produced by the surrounding neuroepithelial cells. These newly born neurons migrate out of the ventricular zone of the telencephalon to establish the cortical plate in a wave of migration between 6 and 18 gestational weeks (Ramakers, 2005). Migrating neurons retain their initial polarization and encounter additional cues as they migrate upward. Neural migration stops at the outer marginal zone, where reelin, produced by CajalRetzius cells of the marginal zone, triggers neurons to cease expressing integrin adhesion receptors. Younger neurons migrate past older neurons to reach the marginal zone, so the upper layer II contains the youngest neurons, while the oldest neurons inhabit the lowest layer VI. Neuronal Polarization and the Initial Growth of Axons and Dendrites Neurons sprout axons soon after ceasing migration, as early as the seventh week in the cortex. In a neutral tissue culture environment, it is a random decision as to which neurite sprouted from a neuron becomes the axon, but cortical neurons in vivo always sprout their axon in the same direction that the axon will grow. In the model organism, Caenorhabditis elegans, a diffusible molecule netrin produced by ventrally located cells causes localized activity of PI3 kinase in young neurons, which then sprout

letourneau: formation of axons and dendrites by developing neurons

13

their axon toward the netrin source (Adler et al., 2006). PI3 kinase is involved in axonal specification of mammalian neurons, and thus localization of PI3 kinase in response to a local cue may both specify axonal identity and regulate actin motility to control the direction of axonal initiation. Other factors are implicated in regulating the initial direction of cortical axonal growth. Immature cortical pyramidal neurons first extend an axon toward the ventricle, followed by an apical dendrite, which grows toward the pial surface. Unexpectedly, these opposite directions of axonal versus dendritic growth are regulated by the same extracellular molecule, semaphorin 3A (Sema3A), produced by cells near the pial surface and released to create an extracellular gradient (Whitford et al., 2002). Axons are repelled by Sema3A, while the subsequently formed apical dendrites of these neurons are attracted by Sema3A. The difference in directions of these processes lies not in a local difference in expression of membrane receptors for Sema3A, but rather in a local difference in distribution of signaling proteins that modulate levels of the cyclic nucleotide cGMP. The combination of Sema3A signaling and high cGMP levels in the apical dendrite promotes actin polymerization and dendritic growth, while Sema3A signaling in the axon combined with low cGMP activates the GTPase RhoA, which depresses actin dynamics and activated myosin II contractility, so the axonal growth cone migrates away from the Sema3A source. Thus the opposite responses of axons and dendrites to

Sema3A are due to an asymmetric distribution of cytoplasmic signaling components in dendrites versus axons.

Figure 1.7 Summary of the action of guidance cues that are involved in growth cone navigation. Short-range cues on surfaces that growth cones come into contact with act to promote or inhibit growth cone adhesion and migration. Long-range cues are diffus-

ible molecules released from intermediate or synaptic targets that attract or repel migrating growth cones. Growth cones integrate information coming simultaneously from multiple cues during navigation. (From Tessier-LaVigne and Goodman, 1996.)

14

Axonal Guidance Once sprouted from neuronal perikarya, axons follow stereotypical routes to their targets. This pathfinding occurs by growth cone navigation; that is, a growth cone detects and responds to physical and chemical features in its environment (figure 1.7). The protrusion of filopodia and lamellipodia from the growth cone of a 1μm-diameter axon allows exploration of an expanded search area 25 μm or more across. When filopodial and lamellipodial protrusion is suppressed, axons grow, but they do not navigate accurately, because without filopodial and lamellipodial protrusions a growth cone’s search area is too small to localize guidance cues. If the path of a growth cone to its target is long, the path is divided into several segments, each ending at an intermediate target to which the growth cone navigates. Often these intermediate targets represent a choice point at which a growth cone turns or changes direction as it enters the next segment of its journey. Pathways for growth cone navigation contain molecules that promote adhesion and growth cone migration. Molecules that repress adhesion or actin dynamics are expressed adjacent to these pathways, acting like “guard rails” to keep growth cones migrating on the proper path. Several proteins have been identified as negative guidance cues, including slit proteins, Sema3A, and several

fundamentals of developmental neurobiology

ephrinA’s. Each negative cue is detected by a different specific receptor with specific signaling mechanisms, although common features of these mechanisms include disruption of growth cone adhesions, suppression of actin polymerization, and activation of RhoA to stimulate myosin II–mediated contraction, leading to growth cone collapse and sometimes retraction of entire axonal branches or segments (Guan and Rao, 2003). Some molecules simply mark a path as positive or negative without providing directional information, while other molecules are soluble, are released by navigation targets, and are distributed in gradients that provide directional information to growth cones. At any instant a growth cone is detecting several guidance molecules, so growth cone migration depends on integrating the intracellular signals simultaneously triggered from multiple receptors. The following section describes specific features of growth cone guidance in the developing CNS. Most of the molecular information about growth cone guidance comes from studies of model vertebrate systems, but the timing of the events in human brain development is included (Ramakers, 2005).

Navigation of Corticofugal Axons As stated earlier, cortical neurons sprout their axons away from the pial sur-

face in response to a gradient of the repellent cue Sema3A. At eight weeks the earliest corticofugal axons reach their first target, the intermediate zone, attracted by Sema3C, expressed in the subventricular zone (figure 1.8 and plate 3; figure 1.9 and plate 4). The intermediate zone is rich in extracellular matrix and contains laminin, an adhesive protein that binds growth cone integrin receptors to form adhesive contacts that promote actin polymerization and give growth cones traction to migrate. The intermediate zone is the first choice point for corticofugal axons, as they encounter the repellent Sema3A, expressed by the underlying ventricular zone. Growth cones of corticothalamic and corticospinal axons turn laterally to exit the dorsal telencephalon through the internal capsule, while growth cones of corticocortical axons turn medially. The molecules or cells that mediate this first decision are unknown. The internal capsule contains the attractant netrin-1, which along with laminin promotes growth through the internal capsule. As these axons traverse the internal capsule, they are prevented from moving medially by expression of the repellent cues, slit-1 and slit-2, in the ganglionic eminence (Bagri et al., 2002). At the telecephalic-diencephalic boundary these corticofugal axons reach another choice point and split into two groups. Corticothalamic axons turn toward the thalamus, while corticospinal axons continue caudally, avoiding slit proteins expressed in ventromedial diencephalon.

Figure 1.8 The trajectory of growing thalamocortical and corticothalamic fibers involves multiple steps and both attractive and repulsive guidance cues. The expression of guidance molecules is related to each of these steps: Slit is a repellent that steers thalamic axons emerging from the diencephalon and in the ventral telencephalon. Ephrin-A5 is involved in sorting thalamocortical axons in the ventral telencephalon. Netrin-1 is an attractive factor for

both populations of fibers in the internal capsule. Semaphorins 3A and 3C steer cortical fibers to penetrate the intermediate zone and then turn. EphA4 in the thalamus and ephrin-A5 in the cortex are involved in the establishment of topographic connections. Th, thalamus; Hyp, hypothalamus; IC, internal capsule; GE, ganglionic eminence; Ncx, neocortex. (From Uziel et al., 2006.) (See plate 3.)

Growth cone navigation along major pathways during cerebral cortical development

letourneau: formation of axons and dendrites by developing neurons

15

Figure 1.9 Schematic diagrams of coronal sections through the developing forebrain showing the trajectory of corticospinal (red), corticothalamic (blue), and thalamocortical (purple) axons in relation to regions that express slit-1 (blue) and slit-2 (yellow) at selected

levels. Regions depicted in green express both slit-1 and slit-2. CGE, caudal ganglionic eminence; H, hippocampus; ic, internal capsule; NCx, neocortex. (From Bagri et al., 2002.) (See plate 4.)

Growth cones of corticothalamic axons may turn in response to attractants released from their thalamic target, or they may recognize early thalamocortical axons and grow along them to reach the thalamus. Axons express several adhesion molecules, including L1 and N-cadherin, that bind homophilically to the same molecules on growth cones to form adhesive contacts that promote growth cone migration. Growth cone migration along previously extended axons is a major means of axonal growth in many tracts, and it is common that the first axons that establish a path become “pioneer fibers” that are followed by subsequent growth cones. Corticospinal axons continue toward the hindbrain until they reach the decussation area at 10 gestational weeks in humans (Ramakers, 2005; Ten Donkelaar et al., 2004). Although corticospinal axons were previously repelled from the midline by slit proteins and other negative cues, these repellents are not expressed in the decussation area, and corticospinal axons now respond to attractants such as netrin, to cross the midline, completing the decussation by week 17. Renewed expression of midline repellents, slit and ephrin3B, caudal to the decussation, prevents corticospinal axons from recrossing as they grow down the spinal cord. The lumbrosacral area is reached by 29 weeks, but growth cones do not enter spinal cord gray matter for several weeks. Innervation of target areas of gray matter by corticospinal axons occurs in an interesting manner. Corticospinal axons

initially extend beyond their target areas. Eventually, target cells release attractants and express adhesive ligands that specifically activate local regions along the afferent axons. Activation of Rho GTPases and ABPs induces localized actin-based protrusive activity from the axonal shafts, followed by collateral branches that grow into the targets. The axonal segments that extend beyond the innervated target are then eliminated via retraction involving myosin II. This exuberant growth of axons followed by retraction of mistargeted axonal segments is a common feature in the development of many cortical circuits (Innocenti and Price, 2005). The corticocortical fibers that form the corpus callosum make several guidance decisions after their first decision to turn medially in the intermediate zone (Richards, 2002). The molecules that guide these decisions are unknown, although the callosal path may be “pioneered” by axons from the cingulated cortex, creating an axonal path that is followed by neocortical axons to the midline. The growth cones of corticocortical axons are attracted by netrin-1, produced by midline cells, and channeled to cross the midline by repulsion from slit proteins expressed by cells above and below the developing corpus callosum. In humans the corpus callosum begins forming by 11–12 weeks and is well developed by 18–20 weeks. After navigating dorsally and into the contralateral hemisphere, the axons reach the cortical subplate where they extend and branch, remaining for several

16

fundamentals of developmental neurobiology

Figure 1.10 Mechanisms and molecules controlling retinotopic mapping in chicks and rodents. The names and/or distributions of molecules known, or potentially able, to control the dominant mechanisms at each stage are listed. The gradients represent the consensus distribution for a combination of related molecules (i.e.,

ephrin-A’s), which are not listed individually owing to distinctions in the individual members expressed and the precise distributions between species. Molecules other than those listed are likely to participate. (From McLaughlin and O’Leary, 2005.) (See plate 5.)

weeks before sprouting collateral branches at 28 weeks into their appropriate final target regions of the cortex.

signaling activated by guidance cues, and temporal and spatial differences in the expression of guidance cues and their receptors by developing tissues and neuronal populations. An interesting recent finding is that growth cone responses to guidance cues may depend on bursts of local protein synthesis of receptors or signaling components within a growth cone. For example, Sema3A rapidly stimulates synthesis of the GTPase RhoA from mRNA within growth cones (Wu et al., 2005). RhoA activity is necessary for Sema3A induction of growth cone collapse. Some growth cones cross the ventral spinal cord and only then synthesize and express EphA receptors that mediate a repulsive response to midline ephrins, preventing recrossing the midline (Brittis, Lu, and Flanagan, 2002). Much remains to be learned about how growth cones detect guidance cues and integrate complex signals to navigate to their intermediate and final targets.

Navigation of Thalamocortical Growth Cones Thalamocortical afferent axons begin their navigation by growing ventrally until they are stopped by repulsion from slit proteins expressed by the underlying hypothalamus (Lopez-Bendito and Molnar, 2003; Uziel et al., 2006; figure 1.8). Then the growth cones turn laterally, being attracted by netrin-1 expressed by cells in the internal capsule. The growth cones turn dorsally and migrate toward the cortex within the internal capsule, keeping lateral in response to slit proteins expressed by the ganglionic eminence (Bagri et al., 2002). Within the internal capsule, thalamocortical axons meet corticofugal fibers, which they follow toward their cortical targets. Thalamocortical axons penetrate the cortical subplate between 9 and 18 weeks in developing humans. By 24 weeks they fill the upper subplate and extend branches exploring for their correct cortical targets. Thalamocortical axons finally enter the cortex between 26 and 28 weeks, prior to the entry of callosal axons. The preceding paragraphs have described how axons navigate to their targets by detecting and responding to guidance molecules that regulate growth cone motility. It may seem that the relatively limited numbers of guidance molecules, laminins, ephrins, semaphorins, netrins, slits, and immunoglobulin-like adhesion molecules are too few to account for the complexity of neural circuitry (Yu and Bargmann, 2001). However, this diversity of axonal pathways arises from cell-type–specific differences in expression of receptors for guidance cues, in downstream cytoplasmic

Patterning Axonal Distribution within Targets Once a group of axons reach their synaptic target, they become organized into patterns that represent physiologically relevant topography or sensory parameters. The distribution of retinal ganglion cells’ axons in their midbrain target (optic tectum or superior colliculus) is a model system in understanding this process (McLaughlin and O’Leary, 2005). Gradients in the distribution of ephrins and their Eph receptors on cells across the optic tectum (or colliculus) and the incoming retinal axons and growth cones are key features that determine the topography of retinal inputs to the tectum (figure 1.10 and plate 5). Ephrin-A2 and -A5 are expressed in an increasing gradient from the anterior to posterior

letourneau: formation of axons and dendrites by developing neurons

17

tectum. EphA receptors bind ephrin-A ligands and trigger decreased Rac1 and Cdc42 activities and increased RhoA activity, stimulating growth cone repulsion. Growth cones of temporal retinal axons express high levels of EphA receptors, so they stop and innervate the anterior tectum, while nasal retinal growth cones, expressing lower levels of EphA receptors, extend to the posterior tectum, because they are less repelled by the ephrin-A gradient. The distances that retinal growth cones migrate along the anterior-posterior tectal gradient of increasing ephrin-A expression are determined by the relative levels of EphA expression by growth cones. Retinal mapping along the medial-lateral tectal axis involves gradients in the distribution of ephrin-Bs and their EphB receptors among retinal axons and tectal cells. The identification of signaling activity from the cytoplasmic domain of ephrin-B ligands indicates that both ephrins and EphB receptors can activate cytoplasmic signaling to regulate axonal targeting along the mediallateral tectal axis. This mechanism for topographic mapping of connections by gradients of cell surface ligands and receptors was proposed by Roger Sperry (1963) as the chemoaffinity hypothesis. The discovery of gradients in expression of ephrins-A2 and -A5 confirmed Sperry’s hypothesis. It has become clear that the initial distributions of axons, as regulated by these gradients, is not final, and that subsequent remodeling of axons due to further cellular interactions and physiological activities is necessary to create more precise neural circuits. The patterning of inputs to a target depends on activities distributed along the afferent axons, in addition to the growth cones. Local signaling by guidance cues or other physiological events along axonal shafts can rapidly regulate activities of RhoA or Rac1 and Cdc42 to regulate actin dynamics and myosin II activity to induce retraction or addition of collateral or terminal branches along developing axonal shafts (Gallo and Letourneau, 1998). The accessibility and simple anatomy of the retinotectal projection have allowed much progress in understanding the patterning of developing neural circuits. The discovery of gradients in the distributions of ephrin-A and EphA receptors in the neocortex and thalamus, respectively, indicates that gradients of interacting ephrins and their receptors have similar roles in regulating axonal guidance and patterning of thalamocortical connections to their targets in the primary sensory regions of the cerebral cortex (Uziel et al., 2006; figure 1.8). Similar mechanisms may operate in patterning the development of circuits in other regions (Flanagan, 2006). In addition to the development of the correct distribution of axons within a target, axons must recognize the target neurons with which they make synapses. Several cell surface and extracellular molecules are expressed in a lamina-specific manner in the developing cortex, including

18

cadherins, Eph receptors, ephrin ligands, proteoglycans, and neurotrophins (Lopez-Bendito and Molnar, 2003). These molecular differences may provide cues for thalamocortical and corticocortical axons to terminate in the correct layer.

Development of dendrites The dendritic arborization of a neuron contains the synaptic inputs to the neuron and is where synaptic inputs are integrated before the initiation of action potentials. Thus dendritic arbors are critical to the processing of neural information for behavior and other neural activities. Like the formation of axons, dendrite formation is intrinsic to the neuronal phenotype. In fact, different neuronal types in a neutral tissue culture environment will form dendritic arbors that are reminiscent of their characteristic in vivo morphologies. As described earlier, the same basic mechanisms of actin filament and microtubule dynamics operate to drive the formation of dendrites, although dendrites are more numerous, shorter, and more elaborately branched than axons, due to expression of dendritic-specific cytoskeletal, membrane, and signaling proteins. Generally, a neuron initiates dendrites after it is actively engaged in axonal elongation. This lag may be several days, and may be due to both environmental factors and intrinsic factors, such as changes in expression of specific cytoskeletal proteins. The sites of dendrite initiation from a neuron may be determined by previous cell interactions; for example, the apical dendrites of cerebral cortical neurons are formed from the leading process with which immature neurons had migrated from the ventricular lining of the cortex. As described previously, the apical dendrites of cortical neurons are oriented by an attractive response to Sema3A, produced at the pial surface. Other extrinsic proteins produced by neighboring cells or afferent axons promote the formation of dendrites, including osteogenic protein-1 (BMP7) and neurotrophins BDNF and NT-3 (Whitford et al., 2002). Thus intrinsic regulation of cytoskeletal and membrane components combined with availability of extrinsic factors, such as osteogenic protein-1 and neurotrophins, orchestrates the initiation and elongation of branched dendritic arbors. However, as described in the following paragraphs, the formation of dendrites is a prolonged activity, and the final shaping of dendritic arbors depends heavily on afferent inputs and interactions with axon terminals (Van Aelst and Cline, 2004). Visualization of the morphogenesis of individual dendrites in developing brains of living frogs and zebra fish has revealed rapidly changing addition and loss of small branches and arbors as dendrites interact with afferent axons. Filopodia transiently extend from dendritic shafts and termini, and if contacts are made with axonal growth cones, the dendritic filopodium may be stabilized, and nascent synapses may form. However, many of these contacts and

fundamentals of developmental neurobiology

synapses are brief, and the terminal axonal and dendritic branches may be retracted. Synaptic activity is a factor in dendritic morphogenesis, and activation of NMDA receptors at nascent synapses may regulate Rho GTPases to modulate actin filament dynamics that underlie the extension and retraction of dendritic filopodia (Van Aelst and Cline, 2004). The roles of these synapses in regulating dendritic growth may also change as the synapses mature. Postsynaptic activation at early synapses may stimulate formation of more dendritic filopodia and elaboration of dendritic branches, while signaling at more mature synapses may generate stop-growing signals to stabilize dendritic arbors. New excitatory synapses contain NDMA receptors only, and AMPA receptors are added later. Addition of AMPA receptors to synapses may be required for retention of synapses and stabilization of dendritic arborizations. The final shaping of axonal terminals is also dependent on interactions with dendrites and postsynaptic contacts. Retrograde synaptic interactions may signal growth cones to reduce their dynamic activity, stop, and transform to a presynaptic ending. Motor axons growing on muscle fibers of mice that lack the Achreceptor-aggregating protein, agrin, or the agrin receptor component, MUSK, extend abnormally long distances across muscle surfaces, implicating MUSK and agrin in an axonal “stop signal.” The neuromuscular junction contains a laminin isoform, S-laminin, that inhibits axonal growth. Nitric oxide, which is released by dendrites in response to synaptic activity, may be a retrograde signal that stops axonal growth in synaptic regions. Dendritogenesis in the Prenatal and Postnatal Human Brain Neurons begin to form dendrites soon after they initiate axon formation, although dendrites are initially short and slow growing. Apical dendrites are present on cortical pyramidal neurons by 12–13 weeks’ gestation. However, once innervating axons arrive in the cortical plate at 26–28 weeks, dendrite formation accelerates as a result of synaptic contacts, electrical stimulation by axons, and the release of neurotrophins and other factors from axons. In humans, most dendritic growth occurs postnatally in conjunction with synaptogenesis and the increased physiological experience and activity of postnatal life. Dendrite formation in the developing human brain has been examined most thoroughly in the visual cortex and prefrontal cortex (Ramakers, 2005). In the visual cortex most dendritic branches develop prenatally, and postnatal growth involves dendritic lengthening by terminal growth of branches as synapses are added. Total dendrite length of pyramidal neurons in the visual cortex increases rapidly in the first few postnatal months, increasing two- or threefold and reaching the adult levels by 1–2 years. In the prefrontal cortex, synaptogenesis and dendritic growth proceed more slowly than in the visual cortex. During the first postnatal year the

length of dendrites increases 5- to 10-fold by branching and elongating, while after the first year most growth occurs by elongation of branches. By two years of age the total dendritic length per pyramidal neuron is only half the adult level. Yet, at age two the average dendrite length per neuron in the prefrontal cortex is longer than dendritic length in the visual cortex, consistent with the greater dendritic and synaptic complexity in the more integrative cortical regions, compared to unimodal primary cortical regions. These measures of dendrite elaboration in the developing human brain are mostly based on anatomical studies involving Golgi staining of fixed neurons. These data are static and fail to account for the dynamic activities of dendritic elongation, branching, and retraction that are revealed from realtime visualization of dendrite growth and synaptogenesis in living embryos, as mentioned earlier. Much remains to be learned about how axonal and dendritic shapes are sculpted over a period of years, as the result of interactions between genetically defined mechanisms of neuronal growth and a dynamic flux of intercellular molecular signaling, synaptogenesis, and the unpredictable physiological activity of postnatal experience.

Summary Neural circuits arise by a morphogenetic process in which axons and dendrites are formed according to intrinsic neuronal mechanisms that respond to extrinsic regulatory interactions with molecules, cells, and features of the developing organism. The driving force for axonal and dendritic growth is the advance of microtubules and associated organelles, while the actin-based motility of growth cones at the ends of elongating processes allows exploration of local tissue environments for molecular guidance cues. Binding of guidance cues to their receptors on growth cones triggers cytoplasmic signaling that regulates actin filament organization, mechanical forces, and microtubule advance to locally direct growth cone migration, turning, and branching. Axonal growth cones reach their synaptic targets by navigating to a series of intermediate targets, guided by positive and negative responses to surface-bound and soluble molecular cues. Axonal projections within a target are initially patterned according to gradients in the expression of molecules, such as ephrins and Eph rceptors, on axons and target cells. Synaptogenesis, other cellular interactions, and physiological activities adjust and refine axonal growth and branching within a target to achieve more accurate axonal topography. Formation of dendrites begins before afferent axons arrive and involves interactions of intrinsic and extrinsic mechanisms that regulate the orientation and rates of dendritic growth. Dendritic growth accelerates when axons arrive and initiate synaptogenesis. The final shaping of arborizations of dendrites and axons depends on mutual interactions, and

letourneau: formation of axons and dendrites by developing neurons

19

physiological activity has a major role in this final phase of the formation of neural circuits. In the developing human brain, axonal navigation to targets begins in the first trimester and continues throughout the second and into the third trimester. Dendrite growth begins in the second trimester, accelerates in the third trimester, and continues most vigorously through the first 2–3 years and then for years afterward, as dendrites and axonal terminal arbors are sculpted and refined by experience. acknowledgments

The author thanks the members of his laboratory who have been dedicated and enthusiastic in research on axonal growth and guidance for 30 years. The author’s research has been supported by the National Institutes of Health, the National Science Foundation, and the Minnesota Medical Foundation. REFERENCES

Adler, C. E., R. D. Fetter, and C. I. Bargmann, 2006. Unc-6/ netrin induces neuronal asymmetry and defines the site of axon formation. Nature Neurosci. 9:511–518. Arimura, N., and K. Kaibuchi, 2005. Key regulators in neuronal polarity. Neuron 48:881–884. Baas, P. W., and D. W. Buster, 2004. Slow axonal transport and the genesis of neuronal morphology. J. Neurobiol. 58:3– 17. Bagri, A., O. Marin, A. S. Plump, J. Mak, S. J. Pleasure, J. L. R. Rubenstein, and M. Tessier-Lavigne, 2002. Slit proteins prevent midline crossing and determine the dorsoventral position of major axonal pathways in the mammalian forebrain. Neuron 33:233–248. Brittis, P. A., Q. Lu, and J. G. Flanagan, 2002. Axonal protein synthesis provides a mechanism for localized regulation at an intermediate target. Cell 110:223–235. Brown, M. E., and P. C. Bridgman, 2004. Myosin function in nervous and sensory systems. J. Neurobiol. 58:118–130. Challacombe, J. F., D. M. Snow, and P. C. Letourneau, 1997. Dynamic microtubule ends are required for growth cone turning to avoid an inhibitory guidance cue. J. Neurosci. 17:3085– 3095. Da Silva, J. S., and C. G. Dotti, 2002. Breaking the neuronal sphere: Regulation of the actin cytoskeleton in neuritogenesis. Nature Rev. Neurosci. 3:694–704. Dehmelt, L., and S. Halpain, 2004. Actin and microtubules in neurite initiation: Are MAPs the missing link? J. Neurobiol. 58:18–33. Dent, E. W., and F. B. Gertler, 2003. Cytoskeletal dynamics and transport in growth cone motility and axon guidance. Neuron 40:209–227. Dickson, B. J., 2002. Molecular mechanisms of axonal guidance. Science 298:1959–1964. Fass, J. N., and D. J. Odde, 2003. Tensile force–dependent neurite elicitation via anti-β1 integrin antibody-coated magnetic beads. Biophys. J. 85:623–636. Flanagan, J. G., 2006. Neural map specification by gradients. Curr. Opin. Neurobiol. 16:59–66. Gallo, G., and P. C. Letourneau, 1998. Localized sources of neurotrophins initiate axon collateral sprouting. J. Neurosci. 18:5403–5414.

20

Gallo, G., and P. C. Letourneau, 1999. Different contributions of microtubule dynamics and transport to the growth of axons and collateral sprouts. J. Neurosci. 19:3860–3873. Gallo, G., and P. C. Letourneau, 2004. Regulation of growth cone actin filaments by guidance cues. J. Neurobiol. 58:92–102. Goldberg, D. J., and D. W. Burmeister, 1986. Stages in axon formation: Observations of growth of aplysia axons in culture using video-enhanced contrast-differential interference contrast microscopy. J. Cell Biol. 103:1921–1931. Gomez, T. M., and J. Q. Zheng, 2006. The molecular basis for calcium-dependent axon pathfinding. Nature Rev. Neurosci. 7:115– 125. Gordon-Weeks, P. R., 2000. Neuronal Growth Cones. Cambridge, UK: Cambridge University Press. Guan, K.-L., and Y. Rao, 2003. Signaling mechanisms mediating neuronal responses to guidance cues. Nature Rev. Neurosci. 4:941– 956. Hirokawa, N., and R. Takemura, 2004. Molecular motors in neuronal development, intracellular transport and diseases. Curr. Opin. Neurobiol. 14:564–573. Innocenti, G. M., and D. J. Price, 2005. Exuberance in the development of cortical networks. Nature Rev. Neurosci. 6:955–965. Jaffe, A. B., and A. Hall, 2005. Rho GTPases: Biochemistry and biology. Annu. Rev. Cell Dev. Biol. 21:247–269. Lamoureux, P., R. E. Buxbaum, and S. R. Heidemann, 1989. Direct evidence that growth cones pull. Nature 340:159–162. Letourneau, P. C., 1979. Cell-substratum adhesion of neurite growth cones, and its role in neurite elongation. Exp. Cell Res. 124:127–138. Letourneau, P. C., 1981. Immunocytochemical evidence for colocalization in neurite growth cones of actin and myosin and their relationship to cell-substratum adhesions. Dev. Biol. 85:113– 122. Letourneau, P. C., 1982. Analysis of microtubule number and length in cytoskeletons of cultured chick sensory neurons. J. Neurosci. 2:806–814. Letourneau, P. C., 1983. Differences in the organization of actin in the growth cones compared with the neurites of cultured neurons from chick embryos. J. Cell Biol. 97:963–973. Letourneau, P. C., T. A. Shattuck, and A. H. Ressler, 1987. “Pull” and “push” in neurite elongation: Observations on the effects of different concentrations of cytochalasin B and taxol. Cell Motil. Cytoskeleton 8:193–209. Lopez-Bendito, G., and Z. Molnar, 2003. Thalamocortical development: How are we going to get there? Nature Rev. Neurosci. 4:276–289. Luo, L., 2002. Actin cytoskeleton regulation in neuronal morphogenesis and structural plasticity. Annu. Rev. Cell Dev. Biol. 18:601– 635. Mandell, J. W., and G. A. Banker, 1996. Microtubule-associated proteins, phosphorylation gradients and the establishment of neuronal polarity. Perspect. Dev. Neurobiol. 4:125–135. McLaughlin, T., and D. D. M. O’Leary, 2005. Molecular gradients and development of retinotopic maps. Annu. Rev. Neurosci. 28:327–355. Pollard, T. D., and G. G. Borisy, 2003. Cellular motility driven by assembly and disassembly of actin filaments. Cell 112:453– 465. Ramakers, G. J. A., 2005. Neuronal network formation in human cerebral cortex. Prog. Brain Res. 147:3–14. Richards, L. J., 2002. Axonal pathfinding mechanisms at the cortical midline and the development of the corpus callosum. Braz. J. Med. Biol. Res. 35:1431–1439.

fundamentals of developmental neurobiology

Rochlin, M. W., K. Itoh, R. S. Adelstein, and P. C. Bridgman, 1995. Localization of myosin II A and B isoforms in cultured neurons. J. Cell Sci. 108:3661–3670. Rodriguez, O. C., A. W. Schaefer, C. A. Mandato, P. Forscher, W. M. Bement, and C. M. Waterman-Storer, 2003. Conserved microtubule-actin interactions in cell movement and morphogenesis. Nature Cell Biol. 5:599–609. Schaefer, A. W., N. Kabir, and P. Forscher, 2002. Filopodia and actin arcs guide the assembly and transport of two populations of microtubules with unique dynamic parameters in neuronal growth cones. J. Cell Biol. 158:139–152. Song, H.-J., and M.-m. Poo, 2001. The cell biology of axonal navigation. Nature Cell Biol. 3:E81–E88. Sperry, R. W., 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. USA 50:703– 710. Suter, D. M., and P. Forscher, 2000. Substrate-cytoskeletal coupling as a mechanism for regulation of growth cone motility and guidance. J. Neurobiol. 44:97–113. Tanaka, E., T. Ho, and M. W. Kirschner, 1995. The role of microtubule dynamics in growth cone motility and axonal growth. J. Cell Biol. 128:139–155. Tanaka, E., and M. W. Kirschner, 1995. The role of microtubules in growth cone turning at substrate boundaries. J. Cell Biol. 128:127–137. Ten Donkelaar, H. J., M. Lammens, P. Wesseling, A. Hori, A. Keyser, and J. Rotteveel, 2004. Development and malforma-

tions of the human pyramidal tract. J. Neurol. 251:1429– 1442. Tessier-Lavigne, M., and C. S. Goodman, 1996. The molecular biology of axon guidance. Science 274:1123–1133. Uziel, D., P. Garcez, R. Lent, C. Peuckert, R. Niehage, F. Weth, and J. Bolz, 2006. Connecting the thalamus and cortex: The role of ephrins. Anat. Rec. A 288A–142. Van Aelst, L., and H. T. Cline, 2004. Rho GTPases and activity-dependent dendrite development. Curr. Opin. Neurobiol. 14:297–304. Whitford, K. L., P. Dijkhuizen, F. Polleux, and A. Ghosh, 2002. Molecular control of cortical dendrite development. Annu. Rev. Neurosci. 25:127–149. Wiggin, G. R., J. P. Fawcett, and T. Pawson, 2005. Polarity proteins in axon specification and synaptogenesis. Dev. Cell 8:803–816. Wu, K. Y., U. Hengst, L. J. Cox, E. Z. Macosko, A. Jeromin, E. R. Urquhart, and S. R. Jaffrey, 2005. Local translation of RhoA regulates growth cone collapse. Nature 436:1020– 1024. Yamada, K. M., B. S. Spooner, and N. K. Wessells, 1971. Ultrastructure and function of growth cones and axons of cultured nerve cells. J. Cell Biol. 49:614–635. Yu, T. W., and C. I. Bargmann, 2001. Dynamic regulation of axon guidance. Nature Neurosci. 4:1169–1176. Zamir, E., and B. Geiger, 2001. Molecular complexity and dynamics of cell-matrix adhesions. J. Cell Sci. 114:3583–3590.

letourneau: formation of axons and dendrites by developing neurons

21

2

Imaging Developmental Changes in Gray and White Matter in the Human Brain ELIZABETH D. O’HARE AND ELIZABETH R. SOWELL

Introduction Human brain maturation is a dynamic and complex process that extends well into adulthood. Understanding developmental changes in brain structure is of fundamental importance in the field of developmental cognitive neuroscience, as it may help elucidate the specific neurobiological changes underlying the maturation of a variety of cognitive processes. Much progress has been made in characterizing structural brain development during the past few decades, largely because of the availability of noninvasive imaging tools such as magnetic resonance imaging (MRI). Because of its relative safety, MRI is ideal for use in developmental populations, where studying individuals at multiple time points is necessary to characterize longitudinal maturational changes in brain structure and function. In vivo neuroimaging methods also have the distinct advantage of allowing for investigations into the relationship between brain structure and brain function, an issue of critical importance in developmental cognitive neuroscience. The main focus of this chapter will be on neuroimaging studies of normative brain maturation that have been performed with a variety of brain-mapping techniques. These studies have allowed for the mapping of structural changes throughout the brain and have advanced our understanding of the timing and localization of alterations in gray and white matter that occur throughout development. We start by reviewing the postmortem literature, and then turn our attention to the in vivo literature. We will briefly review the earliest quantitative imaging studies of brain development that used volumetric methods. Although these studies are highly intuitive and critical for assessing global changes in brain morphology, sophisticated mapping techniques, such as voxel-based morphometry (VBM) and cortical pattern matching (CPM), provide certain advantages. Specifically, they allow for visualization of changes occurring at the cortical surface and throughout the brain that are unbiased by the observable sulcal cortical boundaries necessary for making the anatomical delineations required by volumetric studies. Particular emphasis will be given to cortical gray

matter and peripheral white matter differentiation across development, as recent advancements in our laboratory have allowed for the measurement of cortical thickness in millimeters across the entire brain surface with submillimeter accuracy. We conclude with a discussion of potential cognitive correlates of brain structural maturation.

Postmortem studies: Synaptic modification and myelination Prior to the advent of neuroimaging tools, researchers were limited to inferring structural brain changes from postmortem data. Both approaches have distinct advantages and disadvantages. A major concern with postmortem studies is their generalizability, due to the questionable normalcy of participants studied after death and the lack of samples from younger age ranges. In contrast, despite the advantages associated with performing in vivo MRI imaging in developmental populations described earlier, these techniques measure changes in MR signal values that are only indirectly linked to cellular changes in the brain. Nevertheless, while we cannot directly measure structural changes at the cellular level with MRI, the spatial and temporal patterns of maturational change observed in recent imaging studies reflect patterns that were observed postmortem, demonstrating the validity and compatibility of these methods. Brain development can be characterized as a dynamic process of progressive and regressive changes, which are influenced by both complex genetic programs and experience-dependent plasticity. At birth the human brain contains on the order of 100 billion neurons (Kandel, Schwartz, and Jessell, 2000). As the newborn brain grows in size and complexity, these neurons undergo dendritic branching and arborization, synaptogenesis, myelination, and ultimately synaptic pruning. Huttenlocher’s pioneering work in the early 1980s charted the time courses of synaptogenesis and synaptic pruning. His series of histological studies demonstrated that synaptic density is high at birth and continues to increase throughout the first year of postnatal life, reaching its maximum between 12 and 18 months postnatal. Synaptic density then shows a marked decrease

23

between late childhood and early adulthood, presumably owing to the effects of experience-associated synaptic pruning, ultimately decreasing to about 60 percent of the maximum value attained shortly after birth (Huttenlocher, 1979; Huttenlocher et al., 1982). Interestingly, this pattern of synaptic formation and elimination shows regional variation. In primary visual cortex, synaptogenesis occurs more rapidly after birth and reaches a maximum at about 4 months postnatal. Synaptic pruning then begins and continues to about 4 years of age, at which point synaptic density in this region reaches adult levels. In contrast, peak synaptic density in the prefrontal cortex is not attained until 3–4 years of age. Synaptic pruning in this region is offset relative to that of primary visual cortex. Furthermore, prefrontal pruning lasts longer, with the most substantial decline occurring in middle and late adolescence (Huttenlocher and Dabholkar, 1997). In addition to the developmental processes of synaptogenesis and synaptic pruning, the neurons of the developing human brain are being myelinated. Early work by Yakovlev and Lecours demonstrated that myelination begins late in the second trimester of fetal development and extends into the second and third decades of life (Yakovlev and Lecours, 1967). Studies examining the patterns of myelination in the developing brain show that this process also follows a specific spatial and temporal pattern (Brody et al., 1987). Myelination is thought to progress from inferior to superior brain regions and from posterior to anterior regions. That is, the cerebellum myelinates prior to the cerebral hemispheres, and the occipital lobes prior to the frontal lobes. Thus myelination of the brain regions responsible for higher cognitive functioning, the dorsal frontal lobes, is still occurring throughout adolescence (Yakovlev and Lecours, 1967; Brody et al., 1987), a maturational period marked by the finetuning of cognitive control and executive function (Cohen et al., 2000; Bunge et al., 2002). The work of Kaes in 1907 was among the first to illustrate this complex relationship between age and the myelination of cortical regions known to subserve distinct cognitive functions. These postmortem studies of cortical width demonstrated that the primary cortices in which the myelination process is completed earlier show little age-related change. In contrast, the association cortices of the frontal and parietal lobes that are characterized by an extended period of myelination show pronounced age-related change (Albert and Knoefel, 1994). Composite maps of cortical thickness in individuals ranging from 3 months to 97 years of age demonstrated the progressive spread of intracortical myelination into frontal and parietal cortices during the first four decades of life. This proliferation of myelination into the cortex appears to result in gray matter “thinning.” Thus Kaes’s work suggests that loss of cortical thickness with age is due, in part, to an increased proliferation of myelin into the

24

cortical ribbon, and not just to the effects of synaptic pruning and cell loss. Further evidence for an extended period of myelination in certain brain regions was demonstrated by the work of Benes and colleagues who observed a 95 percent increase in the extent of myelination relative to brain weight between the first and second decades of life within the superior medullary lamina of the parahippocampal gyrus (Benes et al., 1994). The authors describe the connectivity of this region and indicate that some of the axonal myelination occurring here could be on axons originating in the cingulate gyrus, a region known for playing a critical role in the regulation of cognitive control (Carter, Botrinick, and Cohen, 1999). The work of Benes and colleagues provides additional, although indirect, evidence for a possible relationship between improved functionality and increased myelination. As discussed in the sections that follow, a major advantage of in vivo studies is the ability to relate changes in cortical structure with changes in cognition and behavior. Taken together, the postmortem studies that we have described illustrate that the brain is undergoing progressive and regressive age-related changes throughout development. Concomitant reductions in synaptic density and increases in axonal myelination are the hallmarks of experience-based neural plasticity and are consistent with the principle of selective specialization (Edelman, 1993). Postulated to be the basis of the formation of cognitive networks that underlie higher cognitive processes (Post and Weiss, 1997; Fuster, 2002), this process involves the initial overproduction of neurons and synaptic connections during infancy and early childhood followed by activity-dependent fine-tuning of neural activity via synaptic pruning that continues well into adolescence. This then leads to efficient networks of neuronal connections that are in turn continuously changing with experience (Post and Weiss, 1997; Kandel, Schwartz, and Jessell, 2000). Despite the characterization of the processes of synaptic pruning and axonal myelination that are occurring throughout development, the relationship between postmortem findings and those from in vivo studies remains unclear (Sowell, Thompson, and Toga, 2004). Histological studies yield only fragmentary evidence for brain maturational changes in late childhood and adolescence, because of the underrepresentation of postmortem data from this age range. Furthermore, questions remain as to whether synaptic pruning or axonal myelination contributes most to the gross morphological changes observed in neuroimaging studies. Giedd and colleagues have suggested that reductions in synaptic density are unlikely to account for the large volume decreases in gray matter observed throughout development and that the balance between the decreasing number of neurons and the increasing size of glial cells attributable to myelination is primarily responsible for determining the overall size of

fundamentals of developmental neurobiology

brain structures (Giedd, Snell, et al., 1996). Further, the early work of Kaes described previously illustrates the proliferation of myelin into the cortical ribbon during childhood and adolescence, resulting in what appears to be cortical “thinning” during that time period. The relative contributions of myelination and synaptic pruning to the global changes in brain size or cortical gray and white matter distributions observed across development may in part be elucidated by findings from brain-mapping studies that are the focus of the remainder of this chapter.

Anatomically based parcellation methods: Volumetric studies Some of the earliest quantitative brain imaging studies in children and adolescents used volumetric parcellation methods. In this approach, the brain is subdivided into separate anatomical regions via either stereotaxic coordinates (Jernigan, Archibald, et al., 1991; Reiss et al., 1996), manual definition of regions of interest (Giedd, Vaituzis, et al., 1996; Sowell, Trauner, et al., 2002), or automated lobar region definition (Giedd et al., 1999). Tissue segmentation and volumetric measurements are then used to estimate the volumes of gray matter, white matter, and cerebrospinal fluid (CSF) within each lobe or region of interest. Given postmortem findings of regional differences in the temporal and spatial patterns of synaptic pruning and myelination across the brain, developmentally related changes in gray and white matter volumes would be expected to show similar regional differences. In the first quantitative structural MRI study in normal children, Jernigan and Tallal (1990) reported that children (ages 8–10) had significantly more cortical gray matter than young adults, despite the fact that the young adults had larger total brain volumes. A second study by Jernigan and colleagues extended this finding with the observation that the timing of gray matter loss had different trajectories depending on the brain region. The earliest gray matter loss occurred in the deep motor nuclei around early childhood and then later in the association cortices of the parietal and frontal lobes during early adolescence (Jernigan, Trauner, et al., 1991). Although these studies did not measure synaptic density directly, this was the first in vivo morphological evidence to support the postmortem findings of Huttenlocher (Huttenlocher, 1979; Huttenlocher et al., 1982) and of Yakovlev and Lecours (1967), regarding the regional and temporal patterns of cellular maturation. Since the initial work of Jernigan and colleagues, cortical gray matter volume decreases during development have been reported by several other groups (Pfefferbaum et al., 1994; Reiss et al., 1996). As with the histological studies described previously, volumetric studies also suggest regional differences in the extent of gray matter loss during childhood

and adolescence. Taken together, these studies demonstrate highly significant decreases in gray matter along with concomitant increases in white matter in the dorsal association cortices of the parietal and frontal lobes, with increasing age. In contrast, the ventral aspects of the temporal lobes change less dramatically between childhood and adolescence (Giedd et al., 1999; Sowell, Trauner, et al., 2002). Interestingly, these structural changes appear to be related to the changing cognitive capacities of children and adolescents. A significant positive correlation between frontal gray matter volume and performance on a verbal learning task was observed in a study investigating brain-behavior relationships (Sowell, Delis, et al., 2001). It is important to note that some studies have observed nonlinear age effects on gray matter volume in various cortical regions (Giedd et al., 1999; Gogtay et al., 2004). These studies have reported an initial increase in gray matter density that peaks between ages 10 and 12, depending on gender, in frontal and parietal lobes. Gray matter density then declines during the adolescent and postadolescent periods. It may be that this nonlinearity in gray matter volume change is due to methodological differences, as those studies reporting initial increases in gray matter have been longitudinal in design and have had more power to detect individual growth patterns, even in the presence of large between-individual variation (Giedd et al., 1999). Furthermore, there is evidence to suggest that the observation of nonlinear age effects on gray matter is contingent upon the age range examined. Reports of initial gray matter increase have come from studies where the youngest subjects studied were 4 years of age (Giedd et al., 1999; Gogtay et al., 2004), in contrast to those reporting progressive gray matter loss (Jernigan, Archibald, et al., 1991; Sowell, Trauner, et al., 2002), where the youngest subjects were 7 years old. Despite differences in findings between research groups, all seem to agree that regionally specific patterns of gray matter loss occur during late childhood and adolescence. Furthermore, this pattern appears consistent with what would be expected given the results from postmortem studies and given the known pattern of cognitive developmental changes that occur during adolescence.

Whole-brain mapping methods: Voxel-based morphometry (VBM) Although volumetric studies provide further evidence for continued gray matter loss and white matter gain throughout child and adolescent development, these studies are unable to precisely localize maturational changes in the brain, given that only gross lobar structures, thus far, have been reliably identified and manually defined. To address the need for more accurate localization of maturational changes, methods were developed that allowed for the

o’hare and sowell: imaging developmental changes

25

assessment of structural changes in brain tissues on a voxelby-voxel basis. Modifying methods initially used to analyze functional imaging data, our laboratory used voxel-based morphometry methods (VBM) (Ashburner and Friston, 2000) to localize age-related gray matter density reductions between childhood and adolescence (Sowell, Thompson, et al., 1999b). The main advantage of VBM over volumetric parcellation methods is that it allows for the automated measurement of developmental changes in gray or white matter throughout the entire brain via spatial normalization of volumes into a standard space and the scaling of images so that each voxel coordinate is anatomically comparable across subjects. Results from these analyses revealed that the frontal and parietal gray matter volume reductions observed in the volumetric studies of brain maturation resulted mostly from gray matter reductions in diffuse dorsal regions of these association cortices (Sowell, Thompson, et al., 1999a). In a similar study, Paus and colleagues used VBM to assess white matter changes in children and adolescents and observed significant age-related increases in white matter density within the left and right internal capsule and the posterior portion of the left arcuate fasciculus (Paus et al., 1999). As these white matter pathways connect regions known to be important for speech and motor functions, this finding suggests a relationship between increased myelination and development of cognitive functions (Paus et al., 1999). A third study elucidated the full spectrum of structural brain maturation by using VBM to examine the pattern of maturation between adolescence and adulthood. As described earlier, between childhood and adolescence, cortical changes were distributed in frontal and parietal regions (Sowell, Thompson, et al., 1999a). In contrast, however, the pattern of cortical changes between adolescents and adults was localized to large regions of the dorsal, mesial, and orbitofrontal cortex, with relatively little gray matter loss in the parietal lobes (Sowell, Thompson, et al., 1999b) (see figure 2.1 and plate 6). This postadolescent gray matter loss in frontal regions makes sense when one considers that the cognitive functions typically ascribed to the frontal lobes, known collectively as executive functions (EF), show protracted courses of development that appear to parallel the lengthy course of structural development characteristic of this region (Fuster, 1997; Diamond, 2000).

Figure 2.1 VBM reveals the full spectrum of gray matter density loss across childhood, adolescence, and adulthood. The top panel shows the child minus adolescent statistical map for negative age effects, the bottom panel shows the same maps for adolescent minus adult. Areas in color represent clusters of gray matter density reduction observed between these age groups. These maps are threedimensional renderings of statistical maps shown inside the transparent cortical surface rendering of one representative subject’s brain. Color coding is applied to each cluster based on its location within the representative brains. Clusters are shown in the frontal lobes (purple), parietal lobes (red), occipital lobes (yellow), temporal lobes (blue), and subcortical regions (green) (Sowell, Thompson, et al., 1999a, 1999b). (Reproduced with permission from Sowell, Thompson, and Toga, 2004.) (See plate 6.)

Cortical mapping methods: Gray matter density Although the VBM methodological approach has advantages over volumetric studies, VBM methods are not without shortcomings. These methods rely on automated image registration techniques to normalize brain volumes across subjects. This is problematic because considerable variability in sulcal patterns exists across individuals and across cortical

26

regions. Studies have demonstrated significant cortical sulcal variability in children (Sowell, Thompson, et al., 2002), adults (Narr et al., 2001), and between males and females (Luders et al., 2004). Thus, when brain volumes are normalized without taking sulcal variability into account, cortical anatomical regions are likely not well matched across subjects.

fundamentals of developmental neurobiology

Cortical pattern matching (CPM) methods were developed to address the problem of intersubject cortical surface variability and to improve spatial normalization. Because CPM allows for cortical anatomy to be matched across subjects, these methods can account for interindividual differences in cortical sulcal patterns. With CPM, sulcal landmarks, manually defined in each individual, are used as anchors to drive fluid-warping algorithms (Thompson et al., 2004). These methods allow for the comparison of cortical features of interest at anatomically matched points across all subjects, and statistical analyses can then be used to help localize the effects of age. The studies to be discussed here quantify local gray matter using a measure termed gray matter density. This measures the proportion of segmented gray matter in a small region of a fixed radius (15 mm) around each matched cortical point (Thompson, Mega, et al., 2001; Sowell, Thompson, et al., 2002) (see figure 2.2 and plate 7). Our colleagues and we have used CPM techniques to measure developmental changes in gray matter density in several different populations. In the first of these, we observed a pattern of gray matter density change that was expected given the results from the volumetric studies in the same subjects described previously. Reporting on a sample of 35 individuals between the ages of 7 and 30 years, we observed gray matter density reductions most prominent in the parietal cortices during childhood, followed by a dramatic acceleration in gray matter density loss in the dorsal frontal lobes during the postadolescent years. Interestingly, this loss in gray matter density was inversely correlated with brain growth; that is, regions that showed gray matter density reductions were also expanding, as shown in the composite maps in figure 2.3 and plate 8 (Sowell, Thompson, et al., 2001). A reduction in the number of cortical synapses could result in the observation of reduced gray matter density. However, these results show local brain growth in the same regions where gray matter density reduction is occurring (rather than brain shrinkage). An increase in the amount of myelin could also result in a reduction in the amount of brain tissue that has a gray matter appearance, as suggested in the postmortem work by Kaes described earlier, and given that nonmyelinated peripheral axonal and dendritic fibers do not have normal white matter signal values on T1-weighted MRI (Barkovich et al., 1988). That is, nonmyelinated axonal fibers in the peripheral cortices would not stain for myelin in postmortem studies, and would thus appear more like gray matter at a gross level in MRI. Furthermore, this tissue would have an MR signal value more similar to that of gray matter. If the loss of gray matter observed in the volumetric, VBM, and CPM studies described earlier was caused only by regressive changes such as synaptic pruning or cell loss, we would not see local brain growth during the same time frame in the same cortical regions. Rather, taken together, these observations suggest that the apparent thinning of

cortex (loss of gray matter density) could also result from increased myelination. In summary, although the precise relationship between brain growth and gray matter density loss remains unclear, the in vivo findings of brain growth spatially and temporally concomitant with cortical thinning highlight the combinatorial nature of regressive (synaptic pruning) and progressive (myelination) changes in gray and white matter structure in the developing human brain. As mentioned previously, in contrast to the pronounced postadolescent decrease in gray matter density characteristic of the dorsal frontal and parietal lobes, small increases in gray matter density between childhood and young adulthood have now been observed in bilateral posterior perisylvian regions in three independent samples of normally developing individuals (Sowell, Thompson, et al., 2002; Sowell et al., 2003; Sowell, Thompson, Leonard, et al., 2004). In a longitudinal study of normally developing children, we observed gray matter thickness increases in the left inferior frontal sulcus (i.e., Broca’s area) (Sowell, Thompson, Leonard, et al., 2004). This unique pattern of gray matter thickening specific to primary language regions leads to the speculation that cortical thickening may be specifically related to gains in language processing. This notion is further supported by studies showing the time between childhood and adolescence as a period of intense learning and modification of language functioning (Sakai, 2005). Later, during the adolescent and postadolescent period, characterized by gains in executive functioning (Luciana and Nelson, 1998; Fuster, 2002; Rosso et al., 2004), changes in gray matter consist of cortical thinning in the frontal regions typically associated with these executive functions. Unfortunately, little work has been done to evaluate relationships between changing brain morphology and changing cognitive functions, but some recent work from our laboratory addressing this issue is discussed later. More recently, CPM studies have used regression analyses to create plots of linear and nonlinear effects of age on gray matter density. In a sample of 176 normal individuals ranging in age from 7 to 87 years studied cross-sectionally, we reported a significant nonlinear decline in gray matter density with age. This decline was most rapid between ages 7 and 60, and was localized to the dorsal frontal and parietal association cortices, on both the lateral and medial surfaces of the hemispheres (Sowell et al., 2003). In contrast, gray matter density gain occurred with increasing age in the left posterior region (perisylvian cortex) until about age 30 (see figure 2.4 and plate 9). The cortices that are known from histological studies to myelinate earliest, namely, the primary visual and auditory cortices and limbic regions, showed a more linear pattern of aging than either the posterior temporal language regions or the associative cortices of the parietal and frontal lobes (Sowell et al., 2003). In addition to evaluating local changes in gray matter over the brain

o’hare and sowell: imaging developmental changes

27

28

fundamentals of developmental neurobiology

Figure 2.2 Cortical pattern matching methods and gray matter density. Top left: Three representative brain image data sets with the original MRI, tissue-segmented images, and surface renderings with sulcal contours shown in pink. Top right: Surface rendering of one representative subject with cutout showing tissue-segmented coronal slice and axial slice superimposed within the surface. Sulcal lines are shown where they would lie on the surface in the cutout region. Note the sample spheres over the right hemisphere inferior frontal sulcus (lower sphere) and on the middle region of the precentral sulcus (upper sphere) that illustrate varying degrees of gray matter density. In the blown-up panel, note that the upper sphere has a higher gray matter density than the lower sphere as it contains only blue pixels (gray matter) within the brain. The lower sphere also contains green pixels (white matter) that would lower the gray matter proportion within it. In actual analyses, gray matter proportion is measured within 15-mm spheres centered across every point over the cortical surface. Bottom: Sulcal anatomical delineations are defined according to color. These are the contours drawn on each individual’s surface rendering according to a reliable, written protocol (see also figure 2.5). (Reproduced with permission from Sowell, Thompson, and Toga, 2004.) (See plate 7.)

Figure 2.3 Reductions in gray matter density are occurring in the same locations as brain growth. Composite statistical maps (top) showing the correspondence in age effects for changes in brain growth (defined here as distance from center, or DFC) and changes in gray matter in the child-to-adolescent contrast (A). Shown in green is the Pearson’s R map of all positive correlation coefficients for DFC, and in blue is the probability map of all regions of significant gray matter loss (surface point significance threshold P = 0.05). In red are regions of overlap in the gray and DFC statistical maps. A similar composite map for the adolescent-to-adult age effects is also shown (B). Note the highly spatially consistent relationship between brain growth and reduction in gray matter density. The shapes of the regions of greatest age-related change for the two maps (gray matter and DFC) are nearly identical in many frontal regions in the adolescent-to-adult contrast. Very few regions of gray matter density reduction fall outside regions of increases in DFC. Shown in images in the lower part of this figure (left, right, and top

views) are the difference between Pearson’s correlation coefficients for the age effects for gray matter density and the age effects for DFC between childhood and adolescence (C) and between adolescence and adulthood (D). The color bar represents corresponding Z scores ranging from −5 to +5 for the difference between correlation coefficients for DFC and gray matter. Highlighted in red are regions of significant negative correlation between DFC and gray matter density (P = 0.05), showing that the relationship between regions of greatest gray matter density reduction are statistically the same as the regions with the greatest brain growth, particularly in the adolescent-to-adulthood years. Highlighted in white are the regions where the difference between correlation coefficients for the gray matter and DFC maps is positive, indicating that the change with age is in the same direction for both variables (i.e., increased DFC change goes with increased gray matter density change). (Reproduced with permission from Sowell, Thompson, et al., 2001.) (See plate 8.)

surface, we also plotted the total volumes of gray matter, white matter, and CSF across the age range. While gray matter volume declined continuously with increasing age, white matter volume first increased, reaching its peak around age 50, and then declined. Despite this comprehensive characterization of the effects of age on both local and global measures of brain structure, cross-sectional studies cannot control for interindividual variance in brain maturation. Longitudinal studies are ideal for this purpose because they limit this variance and increase external validity and generalizability. In one of the first longitudinal mapping studies of normative human brain development, Gogtay and colleagues (2004) reported on a sample of 13 individuals between the ages of 4 and 21 years. A total of 52 images from the 13

subjects were analyzed with approximately 2 years between imaging sessions. Animation of the trajectory of gray matter loss over time revealed a shifting pattern of gray matter change that first appeared in dorsal parietal and primary sensorimotor regions between the ages of 4 and 8 years. Gray matter loss then spread laterally and caudally in temporal cortices, and finally extended anteriorly into dorsal frontal cortex (Gogtay et al., 2004). Similar to our findings (Sowell et al., 2003), the authors note that maturation of lower-order visual and somatosensory cortices occurs before that of the higher-order association cortices. Taken together, these CPM studies of gray matter density change suggest that loss of gray matter is most prominent in the association cortices of the frontal and parietal lobes. While gray matter loss in these regions shows a protracted

o’hare and sowell: imaging developmental changes

29

Figure 2.4 Plots of the relationship between age and gray matter density reveal different trajectories of gray matter changes for different brain regions. Shown is a surface rendering of a human brain (left hemisphere; left is anterior, right is posterior) with scatter plots for gray matter density at various points over the brain surface.

The graphs are laid over the brain approximately where the measurements were taken. The axes for every graph are identical, with gray matter density plotted on the x-axis and age (in years) plotted on the y-axis (Sowell et al., 2003). (See plate 9.)

course that extends into the postadolescent years, gray matter loss has a different trajectory in primary sensorimotor cortices, where it begins and ends earlier. Notably, these patterns of gray matter loss are complemented by brain growth, probably because of increased myelination in the same anatomical regions. In contrast to this pattern, the posterior temporal lobes show subtle increases in gray matter with age up until approximately age 30 before a subsequent decline (Sowell et al., 2003). Although the precise nature of the cellular changes underlying these in vivo observations remains unclear, the similarity of these patterns of results to those of postmortem histological studies portends their validity.

three-dimensional Eikonal fire equation (Sapiro, 2001), which automatically determines cortical thickness throughout brain volumes with submillimeter accuracy (see figure 2.5 and plate 10). A recent study by our group combined a longitudinal design with gray matter thickness measurements to map changes in cortical gray matter across development. We measured gray matter thickness in millimeters in a sample of 45 normally developing individuals between the ages of 5 and 11 years. Each subject was studied twice with two years between scans. The group average cortical thickness maps were remarkably similar to those described by von Economo in his postmortem sample (von Economo, 1929). Both maps revealed that cortex was thickest in the most dorsal aspects of the frontal and parietal lobes (approximately 4–5 mm), and thinnest in primary visual cortex along the banks of the calcarine sulcus in the medial aspect of the occipital lobe (less than 2 mm) (see figure 2.6 and plate 11). In addition to the striking correspondence of this in vivo data to that of Von Economo’s postmortem data, our data revealed statistically significant cortical thinning of approximately 0.15–0.30 mm per year, most prominently in right dorsal frontal and bilateral parietal regions. Significant

Cortical mapping methods: Gray matter thickness Studies of gray matter density changes are difficult to interpret because they must be reported as a percentage change between one group and another (Sowell, Thompson, Leonard, et al., 2004). Methodological advances now allow for the measurement of gray matter thickness in millimeters (Fischl and Dale, 2000; Jones, Buchbinder, and Aharon, 2000; Miller et al., 2000; Kabani et al., 2001). Our laboratory has recently developed similar methods by using the

30

fundamentals of developmental neurobiology

Figure 2.5 Cortical pattern matching methods and gray matter thickness. The skull-stripped, 3 D, gray-scale image volume is shown in the upper left for one representative subject. Surface renderings (upper right) are automatically rendered for each subject using the signal value that best differentiates cortical surface sulcal CSF from cortical gray matter. Thirty-five sulcal landmarks on the lateral and medial surfaces are identified and manually traced. After sulcal patterns are demarcated, surface renderings are flattened to a 2 D planar format. In the bottom left the flattened sulcally delineated surface renderings are shown for four individual subjects. Note the crosshairs in each map: while at slightly different locations in the image, they represent the same sulcal anatomy in each subject (i.e., homologous surface points). A complex deformation, or warping transform, is then applied that aligns the sulcal anatomy of each subject with an average sulcal pattern derived for

the group. Given that the deformation maps associate cortical locations with the same relation to the primary sulcal pattern across subjects (i.e., the crosshairs in all for subjects illustrated here), a local measurement of cortical thickness can be made in each subject and averaged across equivalent cortical locations in all subjects. This is illustrated in the bottom right panel. Cortical thickness, defined as the 3 D distance (in mm) between the inner gray matter/ white matter border and the closest point on the outer CSF/gray matter boundary, is calculated using the Eikonal fire equation (illustrated in more detail in figure 2.6). Using these methods, the average thickness value within a 15-mm sphere can be calculated and averaged across subjects to estimate cortical thickness within groups of individuals. On the bottom right is a group average map of cortical thickness. (See plate 10.)

increases in cortical thickness were observed in canonical language regions of the temporal and frontal lobes (Wernicke’s and Broca’s areas, respectively). This cortical thickening was on the order of approximately 0.10–0.15 mm per year (see figure 2.7 and plate 12). The correspondence of the changes in gray matter thickness to previously described gray matter density changes is striking. These results highlighted that measurable cortical changes were occurring

within individuals over a relatively brief time period during development, and they validate previous cross-sectional findings. Furthermore, they allowed for the quantitative assessment of the magnitude of change in identifiable units (i.e., millimeters). Finally, these results confirmed cortical thickening in posterior perisylvian regions and for the first time revealed cortical thickening in the left inferior frontal language region.

o’hare and sowell: imaging developmental changes

31

Figure 2.6 Cortical thickness maps: (A) original T1-weighted image for one representative subject, (B) tissue segmented image, (C) gray matter thickness image where thickness is progressively coded in millimeters from inner to outer layers of cortex using the 3 D Eikonal fire equation. Note the images were resampled to a voxel size of 0.33 mm cubed, so the thickness measures are at a submillimeter level of precision, according to the color bar on the right (mm). Figures A through C are sliced at the same level in all three image volumes from the same subject. Shown in (D) is an in

32

vivo average cortical thickness map created from these 45 subjects at the first scan. The brain surface is color coded according to the color bar where thickness is shown in millimeters. Our average thickness map can be compared to an adapted version of the 1929 cortical thickness map of von Economo (von Economo, 1929) (E). Color coding has been applied over his original stippling pattern, respecting the boundaries of his original work, to highlight the similarities between the two maps. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See plate 11.)

fundamentals of developmental neurobiology

Figure 2.7 These maps show the statistical significance of annualized change in cortical thickness measurements. Color coding represents t values at each cortical surface point according to the color bar at the near right (ranging from t = −3.00 to t = 3.0). Significant values are overlaid in shades of red (significant thickness

decreases, TD) and white (significant thickness increases, TI), according to the color bar at the far right. Arrows point to three regions of significant increases in gray matter thickness. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See plate 12.)

As noted, these findings represent the third observation of specific age-related increases in gray matter in the primary language cortices, drawn from three independent samples (Sowell, Thompson, et al., 2002; Sowell et al., 2003; Sowell, Thompson, Leonard, et al., 2004). In addition, it appears that these regions have a more protracted course of development than any other cortical region (Sowell et al., 2003). That the structural maturation pattern is markedly different for brain regions critical for language processing is perhaps not surprising, given the complexity and protracted nature of the language-learning process. Future work will combine structural and functional MRI methods to characterize the specific relationship between these observed increases in gray matter thickness in primary language cortices and the functional maturation of language processes.

matter thickness and performance on a test of general verbal intellectual functioning, the vocabulary subtest of the Wechsler Intelligence Scale for children (Wechsler, 1991). Greater gray matter thinning was correlated with better performance on this test in diffuse areas of the left hemisphere (Sowell, Thompson, Leonard, et al., 2004) (see figure 2.8 and plate 13). Although the only regions to survive correction for multiple comparisons were the left lateral dorsal frontal and the left lateral parietal, these findings were consistent with expectations, given that the task represents relatively global verbal intellectual abilities and that the significant correlations were observed in the language-dominant left hemisphere. More recently, we examined more specific cognitive functions—phonological processing and motor speed and dexterity—and their relationships to changes in cortical thickness in these same children (Lu et al., 2007). We expected that gray matter thickening in the left inferior frontal gyrus, the only region showing increases in gray matter thickness in the left hemisphere, would be associated with developmental changes in phonological processing. In order to establish the specificity of the relation between gray matter thickening in the left inferior frontal gyrus and phonological processing, we predicted that thickness change in this region would correlate with improving phonological skills but not with nonlanguage measures such as motor dexterity and strength. In turn, we expected improved hand motor skills to correlate with gray matter thickness values in the hand motor region, but not the left inferior frontal cortex. As predicted, increased gray matter thickness in left inferior frontal regions was

Cognitive correlates of developmental changes in gray matter What is the relationship between the changes in gray and white matter distributions that we have described and the changing cognitive capacities of children and adolescents? Unfortunately, little is known on this topic, but a handful of studies have begun to address this important question. This section contains a brief discussion of these studies, which have largely focused on correlating IQ measures with structural changes in the developing brain. In the same sample of 45 normally developing children studied longitudinally for cortical thickness change (described earlier), we evaluated relationships between changes in gray

o’hare and sowell: imaging developmental changes

33

Figure 2.8 Brain-behavior maps showing the relationship between vocabulary scores and cortical thickness. These maps show the p value for negative correlations between change in cortical thickness (time 2 value − time 1 value) and change in vocabulary scores (time 2 score − time 1 score). Regions in color represent

negative p values, that is, regions where greater thinning was associated with greater vocabulary improvement. Regions in white were not significant. No positive correlations reached significance. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See plate 13.)

significantly correlated with improving phonological processing, but not with improving motor processing. In contrast, motor processing improvement, but not phonological processing improvement, was significantly correlated with decreases in gray matter thickness in the hand region of left primary motor cortex (see figure 2.9 and plate 14). This double dissociation illustrates a specific correspondence between gray matter thickness change and cognitive maturation in the left inferior frontal cortex, a brain region known to be important for language processing. Evidence suggests that relationships between intelligence and gray matter structure may show regional differences that vary with age. Studies examining the correlation between gray matter volume or density and IQ found a significant positive correlation in children in the anterior cingulate (Wilke et al., 2003). In adolescents, gray matter density was positively correlated with IQ most prominently in the orbitofrontal cortex (Frangou, Chitins, and Williams, 2004). And finally, in adults, Haier and colleagues (2004) used VBM to assess gray matter and found a significant positive correlation between gray matter density and IQ in the lateral prefrontal cortex (Haier et al., 2004). As described previously, we found significant correlation between frontal (and parietal) gray matter thinning and vocabulary, which is likely reflective of more general verbal IQ and thus is quite similar to the other findings of frontal lobe–IQ correlations (Thompson, Cannon, et al., 2001; Toga and Thompson, 2005). While different measures and methods were used, common to all studies were significant relationships between frontal lobe structure and general intellectual functioning.

In a recent study, Shaw and colleagues used a longitudinal design to examine the relationship between gray matter thickness and intellectual ability in 307 normally developing children and adolescents. When examining all subjects, the authors noted modest positive correlations between IQ and cortical thickness in most of the frontal, parietal, and occipital cortex, and modest negative correlations between these variables in the anterior temporal cortex (Shaw et al., 2006). Apparently surprised by the relatively modest correlations between cortical thickness and IQ with such a large sample, the authors decided to further investigate their data by splitting subjects into average, high, and superior IQ groups. In these analyses, they found a significant interaction between IQ group and age in the prefrontal cortex, suggesting that the relationship between frontal cortical thickness and IQ varied as a function of IQ level. A shift was observed from negative correlations between IQ and frontal lobe cortical thickness in the younger children, to a strong positive correlation in later childhood through adulthood. That is, frontal cortex was thinner in the superior IQ children in the earlier years, and thicker in the superior IQ children in the older years. This result occurred in large part because the superior-intelligence group showed a marked increase in cortical thickness in the medial prefrontal cortex that peaked at age 11. In contrast, the average-intelligence group showed a decline in cortical thickness throughout the ages of 7–19. The high-intelligence group showed a trajectory intermediate to those of the other two intelligence groups. Growth curves demonstrated that the superior-intelligence group had the most rapid rate of cortical thinning, while the high-

34

fundamentals of developmental neurobiology

A

B

Figure 2.9 Pearson’s R correlations between changes in gray matter thickness and phonological processing (panel A) and between thickness and motor processing (panel B). Regions in white represent positive correlations with a threshold of p < 0.05. Regions in red represent negative correlations with a threshold of p < 0.05. Increases in gray matter thickness in left inferior

frontal regions significantly correlate with improvement in phonological processing, but not with motor processing, while decreases in gray matter thickness in primary motor cortex significantly correlate with improvement in motor processing, but not with phonological processing. (Reproduced with permission of Oxford University Press from Lu et al., 2007.) (See plate 14.)

and average-intelligence groups had slower rates. These findings suggest that instead of intelligence correlating with total gray matter volume or gray matter thickness across the age span, these correlations depend upon the age range studied. While considerable work is yet to be done in connecting structural brain changes to changes in cognitive functioning, the work to date leaves little doubt that measurable changes in gray (and white) matter tissues are linked to changes in cognitive abilities. The work of Lu and colleagues shows that

the direction of the relationship (i.e., positive or negative), depends on the cognitive function and brain regions evaluated, and the work of Shaw and colleagues demonstrates that the time at which these brain-behavior relationship evaluations are made is also critical. Generally, regions that show cortical thinning with development tend to show negative correlations with cognitive functions subserved by those regions. That is, improved functioning is associated with decreasing cortical thickness in regions that thin with age (which cover a wide area of most of the dorsal and

o’hare and sowell: imaging developmental changes

35

ventral frontal and dorsal parietal cortices), and in regions that thicken with development, improved functioning is associated with increasing cortical thickness.

Summary, conclusions, and future directions Postmortem studies have demonstrated that brain maturation is characterized by a combination of regressive and progressive changes in cortical structure. Changes in the relative proportions of gray and white matter observed with MRI are a hallmark of human brain development. The most prominent finding appears to be decreases in cortical gray matter along with concomitant increases in cortical white matter across the dorsal aspects of the higher-order associative cortices in the frontal and parietal lobes. This dynamic process may continue well into adulthood and appears to be related to functional increases in cognitive capacities associated with these regions. While we cannot determine the cellular etiology of gray and white matter changes with MRI, it is becoming more apparent that synaptic pruning and increased myelination contribute to cortical thinning during development. Again, during aging, cortical thinning continues, though it is likely more degenerative in nature, and not due to the same progressive changes that occur during development. As we discussed, while most developmental structural MRI studies have focused on gray matter, increased myelination into the inner layers of cortex likely results in observations of “cortical thinning” on MRI. That is, tissue that has gray matter signal on MRI in young subjects may actually be unmyelinated axonal and dendritic fibers. This issue may be largely semantic in nature, but does suggest that we not interpret findings of “gray matter” changes on MRI to strictly refer to changes in the cortical neuropil (i.e., cell bodies and their synaptic processes). New and converging findings from several independent samples suggest that the cortex may actually continue to increase in thickness throughout adolescence and young adulthood in some brain regions, namely, those associated with primary expressive and receptive language functioning. It is possible that there are unique aspects of language functioning that require extended stages of plasticity into adulthood, where decreased plasticity and increased efficiency may be occurring simultaneously in regions of the brain that show cortical thinning. It is not at all clear from the postmortem literature what kinds of cellular processes could lead to the thickening of cortex, but some recent controversial work suggests adult neurogenesis in animal models (discussed by Kozorovitskiy and Gould in chapter 4 of this volume; Gould et al., 1999). While we are keenly interested in the cellular etiology of structural changes observed with MRI, the only definitive way to evaluate this would be to combine MRI with postmortem material in the same individuals. That is, we could measure cortical thickness using MRI

36

within numerous individuals, and then correlate measures of cortical thickness in the same cortical locations with postmortem material from the same individuals. Unfortunately, this is not likely to be highly fruitful given the paucity of postmortem material in the child-to-adolescent age range. Animal models could also provide potential answers to these interesting and important questions. Relationships between changes in brain structure and changes in cognitive function have been unfortunately sparse. The few studies that have been conducted suggest that changing brain structure is related to changing cognitive function. The direction and pattern of these relationships varies depending on the typical developmental pattern of brain region subserving the cognitive function being measured, the age range of the subjects studied, and perhaps the overall intellectual capacity of the cohort under investigation. Future work should combine the quantification of brain structural changes with detailed analyses of changing cognitive abilities as assessed with functional MRI in an effort to provide an even greater understanding of the brain changes subserving cognitive development. acknowledgments

Support was provided by the National Institutes of Health (NIMH K01 MH01733 and NIDA R21 DA015878 and R01 DA017830 awarded to ERS, and NIAAA F31AA16039 awarded to EDO’H). Further funding was provided by NIH/NCRR resource grant P41 RR013642 and NIH Roadmap for Medical Research Grant U54 RR021813. REFERENCES

Albert, M. L., and J. E. Knoefel, 1994. Clinical Neurology of Aging. New York. Oxford Univerity Press. Ashburner, J., and K. J. Friston, 2000. Voxel-based morphometry—The methods. NeuroImage 11:805–821. Barkovich, A. J., B. O. Kjos, D. E. Jackson, Jr., and D. Norman, 1988. Normal maturation of the neonatal and infant brain: MR imaging at 1.5 T. Radiology 166:173–180. Benes, F. M., M. Turtle, Y. Khan, and P. Farol, 1994. Myelination of a key relay zone in the hippocampal formation occurs in the human brain during childhood: adolescence, and adulthood. Arch. Gen. Psychiatry 51:477–484. Brody, B. A., H. C. Kinney, A. S. Kloman, and F. H. Gilles, 1987. Sequence of central nervous system myelination in human infancy. I. An autopsy study of myelination. J. Neuropathol. Exp. Neurol. 46:283–301. Bunge, S. A., N. M. Dudukovic, M. E. Thomason, C. J. Vaidya, and J. D. Gabrieli, 2002. Immature frontal lobe contributions to cognitive control in children: Evidence from fMRI. Neuron 33:301–311. Carter, C. S., M. M. Botvinick, and J. D. Cohen, 1999. The contribution of the anterior cingulate cortex to executive processes in cognition. Rev. Neurosci. 10:49–57. Cohen, J. D., M. Botvinick, and C. S. Carter, 2000. Anterior cingulate and prefrontal cortex: Who’s in control? Nature Neurosci. 3:421–423.

fundamentals of developmental neurobiology

Diamond, A., 2000. Close interrelation of motor development and cognitive development and of the cerebellum and prefrontal cortex. Child Dev. 71:44–56. Edelman, G. M., 1993. Neural Darwinism: Selection and reentrant signaling in higher brain function. Neuron 10(2):115– 125. Fischl, B., and A. M. Dale, 2000. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc. Natl. Acad. Sci. USA 97:11050–11055. Frangou, S., X. Chitins, and S. C. Williams, 2004. Mapping IQ and gray matter density in healthy young people. NeuroImage 23:800–805. Fuster, J. M., 1997. The prefrontal cortex: Anatomy, physiology, and neuropsychology of the frontal lobe. New York: LippincottRaven. Fuster, J. M., 2002. Frontal lobe and cognitive development. J. Neurocytol. 31:373–385. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, T. Paus, A. C. Evans, and J. L. Rapoport, 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2:861– 863. Giedd, J. N., J. W. Snell, N. Lange, J. C. Rajapakse, B. J. Casey, P. L. Kozuch, A. C. Vaituzis, Y. C. Vauss, S. D. Hamburger, D. Kaysen, and J. L. Rapoport, 1996. Quantitative magnetic resonance imaging of human brain development: Ages 4–18. Cerebral Cortex 6:551–560. Giedd, J. N., A. C. Vaituzis, S. D. Hamburger, N. Lange, J. C. Rajapakse, D. Kaysen, Y. C. Vauss, and J. L. Rapoport, 1996. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: Ages 4–18 years. J. Comp. Neurol. 366:223–230. Gogtay, N., J. Giedd, L. Lusk, K. M. Hayashi, D. K. Greenstein, A. C. Vaituzis, T. F. Nugent, D. Herman, L. S. Clasen, A. W. Toga, J. L. Rapoport, and P. M. Thompson, 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101:8174– 8179. Gould, E., A. J. Reeves, M. S. Graziano, and C. G. Gross, 1999. Neurogenesis in the neocortex of adult primates. Science 286: 548–552. Haier, R. J., R. E. Jung, R. A. Yeo, K. Head, and M. T. Alkire, 2004. Structural brain variation and general intelligence. NeuroImage 23:425–433. Huttenlocher, P. R., 1979. Synaptic density in human frontal cortex—Developmental changes and effects of aging. Brain Res. 163:195–205. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Huttenlocher, P. R., C. De Courten, L. J. Garey, and H. van der Loos, 1982. Synaptic development in human cerebral cortex. Int. J. Neurol. 17:144–154. Jernigan, T. L., S. L. Archibald, M. T. Berhow, E. R. Sowell, D. S. Foster, and J. R. Hesselink, 1991. Cerebral structure on MRI: Part I: Localization of age-related changes. Biol. Psychiatry 29:55–67. Jernigan, T. L., and P. Tallal, 1990. Late childhood changes in brain morphology observable with MRI. Dev. Med. Child Neurol. 32:379–385. Jernigan, T. L., D. A. Trauner, J. R. Hesselink, and P. A. Tallal, 1991. Maturation of human cerebrum observed in vivo during adolescence. Brain 114:2037–2049.

Jones, S. E., B. R. Buchbinder, and I. Aharon, 2000. Threedimensional mapping of cortical thickness using Laplace’s equation. Hum. Brain Mapp. 11:12–32. Kabani, N., G. Le Goualher, D. MacDonald, and A. C. Evans, 2001. Measurement of cortical thickness using an automated 3– D algorithm: A validation study. NeuroImage 13:375–380. Kandel, E. R., J. H. Schwartz, T. M. Jessell, 2000. Principles of Neural Science. New York: McGraw-Hill. Lu, L. H., C. M. Leonard, P. M. Thompson, E. Kan, J. Jolley, S. E. Welcome, A. W. Toga, and E. R. Sowell, 2007. Normal developmental changes in inferior frontal gray matter are associated with improvement in phonological processing: A longitudinal MRI analysis. Cerebral Cortex 17:1092–1099. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36:273–293. Luders, E., K. L. Narr, P. M. Thompson, D. E. Rex, L. Jancke, H. Steinmetz, and A. W. Toga, 2004. Gender differences in cortical complexity. Nature Neurosci. 7:799–800. Miller, M. I., A. B. Massie, J. T. Ratnanather, K. N. Botteron, and J. G. Csernansky, 2000. Bayesian construction of geometrically based cortical thickness metrics. NeuroImage 12:676–687. Narr, K. L., P. M. Thompson, T. Sharma, J. Moussai, C. Zoumalan, J. Rayman, and A. W. Toga, 2001. Three-dimensional mapping of gyral shape and cortical surface asymmetries in schizophrenia: Gender effects. Am. J. Psychiatry 158:244– 255. Paus, T., A. Zijdenbos, K. Worsley, D. L. Collins, J. Blumenthal, J. N. Giedd, J. L. Rapoport, and A. C. Evans, 1999. Structural maturation of neural pathways in children and adolescents: In vivo study. Science 283:1908–1911. Pfefferbaum, A., D. H. Mathalon, E. V. Sullivan, J. M. Rawles, R. B. Zipursky, and K. O. Lim, 1994. A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch. Neurol. 51:874–887. Post, R. M., and S. R. Weiss, 1997. Emergent properties of neural systems: How focal molecular neurobiological alterations can affect behavior. Dev. Psychopathol. 9:907–929. Reiss, A. L., M. T. Abrams, H. S. Singer, J. L. Ross, and M. B. Denckla, 1996. Brain development, gender and IQ in children: A volumetric imaging study. Brain 119:1763–1774. Rosso, I. M., A. D. Young, L. A. Femia, and D. A. Yurgelun-Todd, 2004. Cognitive and emotional components of frontal lobe functioning in childhood and adolescence. Ann. NY Acad. Sci. 1021:355–362. Sakai, K. L. 2005. Language acquisition and brain development. Science 310:815–819. Sapiro, G. 2001. Geometric Partial Differential Equations and Image Analysis. Cambridge, UK: Cambridge University Press. Shaw, P., D. Greenstein, J. Lerch, L. Clasen, R. Lenroot, N. Gogtay, A. Evans, J. Rapoport, and J. Giedd, 2006. Intellectual ability and cortical development in children and adolescents. Nature 440:676–679. Sowell, E. R., D. Delis, J. Stiles, and T. L. Jernigan, 2001. Improved memory functioning and frontal lobe maturation between childhood and adolescence: A structural MRI study. J. Int. Neuropsychol. Soc. 7:312–322. Sowell, E. R., B. S. Peterson, P. M. Thompson, S. E. Welcome, A. L. Henkenius, and A. W. Toga, 2003. Mapping cortical change across the human life span. Nature Neurosci. 6:309– 315. Sowell, E. R., P. M. Thompson, C. J. Holmes, R. Batth, T. L. Jernigan, and A. W. Toga, 1999a. Localizing age-related

o’hare and sowell: imaging developmental changes

37

changes in brain structure between childhood and adolescence using statistical parametric mapping. NeuroImage 9:587–597. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 1999b. In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nature Neurosci. 2:859–861. Sowell, E. R., P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, and A. W. Toga, 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24(38):8223–8231. Sowell, E. R., P. M. Thompson, D. Rex, D. Kornsand, K. D. Tessner, T. L. Jernigan, and A. W. Toga, 2002. Mapping sulcal pattern asymmetry and local cortical surface gray matter distribution in vivo: Maturation in perisylvian cortices. Cerebral Cortex 12:17–26. Sowell, E. R., P. M. Thompson, K. D. Tessner, and A. W. Toga, 2001. Mapping continued brain growth and gray matter density reduction in dorsal frontal cortex: Inverse relationships during postadolescent brain maturation. J. Neurosci. 21:8819– 8829. Sowell, E. R., P. M. Thompson, and A. W. Toga, 2004. Mapping changes in the human cortex throughout the span of life. The Neuroscientist 10:372–392. Sowell, E. R., D. A. Trauner, A. Gamst, and T. L. Jernigan, 2002. Development of cortical and subcortical brain structures in childhood and adolescence: A structural MRI study. Dev. Med. Child. Neurol. 44:4–16. Thompson, P. M., T. D. Cannon, K. L. Narr, T. van Erp, V. P. Poutanen, M. Huttunen, J. Lonnqvist, C. G.

38

Standertskjold-Nordenstam, J. Kaprio, M. Khaledy, R. Dail, C. I. Zoumalan, and A. W. Toga, 2001. Genetic influences on brain structure. Nature Neurosci. 4:1253–1258. Thompson, P. M., K. M. Hayashi, E. R. Sowell, N. Gogtay, J. N. Giedd, J. L. Rapoport, G. I. de Zubicaray, A. L. Janke, S. E. Rose, J. Semple, D. M. Doddrell, Y. Wang, T. G. van Erp, T. D. Cannon, and A. W. Toga, 2004. Mapping cortical change in Alzheimer’s disease, brain development, and schizophrenia. NeuroImage 23 Suppl 1:S2–18. Thompson, P. M., M. S. Mega, C. Vidal, J. L. Rapoport, and A. W. Toga, 2001. Detecting disease-specific patterns of brain structure using cortical pattern matching and a population-based probabilistic brain atlas. In M. Insana and R. Leahy, eds., IEEE Conference on Information Processing in Medical Imaging (IPMI), Lecture Notes in Computer Science (LNCS), 488–501. New York: Springer-Verlag. Toga, A. W., and P. M. Thompson, 2005. Genetics of brain structure and intelligence. Annu. Rev. Neurosci. 28:1–23. von Economo, C. V., 1929. The Cytoarchitectonics of the Human Cerebral Cortex. London: Oxford Medical Publications. Wechsler, D., 1991. Manual for the Wechsler Intelligence Scale for Children, 3rd ed. San Antonio, TX: Psychological Corporation. Wilke, M., J. H. Sohn, A. W. Byars, and S. K. Holland, 2003. Bright spots: Correlations of gray matter volume with IQ in a normal pediatric population. NeuroImage 20:202–215. Yakovlev, P. I., and A. R. Lecours, 1967. The myelogenetic cycles of regional maturation of the brain. In A. Minkowski (ed.), Regional Development of the Brain in Early Life. Oxford, UK: 3–70. Blackwell Scientific.

fundamentals of developmental neurobiology

3

Gyrification and Development of the Human Brain TONYA WHITE AND CLAUS C. HILGETAG

Introduction The cerebral cortex of the human brain has a readily identified characteristic pattern of grooves and folds. Little is known about the mechanisms behind the emergence of these grooves and folds, known as gyri and sulci. The formation of the characteristic gyri and sulci mainly occurs before birth in a process known as gyrification. Many researchers who study the brain utilize gyri as landmarks to define specific anatomic or functional brain patterns, with little thought given to how they develop. However, the theories surrounding gyrification are fascinating, and, as biological form is closely linked to function, they are potentially relevant for understanding the development of brain function and its localization. It is the primary goal of this chapter to describe the stages and mechanisms of gyrification of the human brain. However, these processes are nested and dependent upon the other major neurodevelopmental processes, which will also be outlined here. The major components of gyrification take place primarily during the third trimester of fetal life, or between approximately 26 and 40 weeks’ gestational age.

Brain development prior to gyrification Formation of the Neural Tube (Neurulation) The mating of egg and sperm is followed by rapid stem cell growth and division. It is not until the second week of uterine life that the dividing cells begin to differentiate into one of three different layers. These layers, known as the endoderm, mesoderm, and ectoderm continue to undergo rapid cell division and differentiate into the different organs of the body. It is the differentiation of the ectodermal layer that forms both the skin and the central nervous system. The neural plate is the first structure of the CNS to develop within the ectodermal layer (Kandel, Schwartz, and Jessell, 2000; Pomeroy and Kim, 2000). A groove is created in the midline of the neural plate, forming perhaps the first short-lived sulcus in the brain. This neural groove rapidly closes to form a neural tube. The caudal region of the neural tube will later become the spinal cord, and the cortical and subcortical structures will spring forth from the rostral region. The cells within the ectoderm that are not involved

with the formation of the neural tube will differentiate into the epidermal layer of the skin. Finally, the neural crest cells between the neural tube and the ectodermal wall differentiate into the peripheral nervous system. During the fifth week of fetal life, the cerebral vesicles begin to take form. The walls surrounding the cerebral vesicles become thick and form the lamina terminalis by the seventh week of fetal life (Destrieux, Velut, and Kakou, 1998). The rostral region of the neural tube undergoes exuberant growth involving an overabundance of neural and glial cells. Clusters of neurons along the midline of the ventral region of the neural tube differentiate further into the thalamus, basal ganglia, hypothalamus, and brain stem (Pomeroy and Kim, 2000). The dorsal region of the lamina terminalis, known as the lamina reunions, folds to become the commissural plate (Rakic and Yakovlev, 1968). The classic concept of a “fate map” has been invoked to describe the differentiation of embryonic precursor structures, which are frequently separated by early limiting sulci, into specific adult brain structures, such as the brain stem or the cerebral lobes (His, 1874; Swanson, 2003). In general, a remarkable variety of morphogenetic and particularly folding processes occur even at this early stage of brain development. These processes are likely driven by differential growth rates of the different embryonic territories. Neuronal Migration Neurons migrate out from the ventral regions of the neural tube in a specific pattern. This mechanism, which has been described as the radial unit hypothesis (Rakic 1988, 1995, 2000), involves neurons formed via mitosis at the ventricular zone migrating along radial glial guide cells in the outer layers of the brain. This outer layer becomes the cortical rim of gray matter (GM) in the adult brain. Before six weeks of fetal life, the neural progenitor cells located in the ventricular zone begin symmetric cell division, with each stem cell producing two identical stem cells with each mitotic cycle (Rakic, 1988). Thus this period results in an exponential growth in neuronal progenitor cells. Then, at approximately six weeks gestational age, the progenitor cells gradually initiate asymmetric division, with one daughter cell remaining as an undifferentiated stem cell, while the other daughter cell matures into a neuron that migrates outward to the cortex.

39

The migration forms an inside-out pattern, with later generations passing through the previously developed cells before reaching their ultimate migratory position in the gray matter of the cortex (Sidman and Rakic, 1973). The cortical GM rim consists of six layers of cells that have migrated in this inside-out pattern. During these two phases of symmetric (before six weeks) and asymmetric (six weeks to 12 weeks) cell division, small perturbations can influence the thickness or the surface area of the cortex. In turn, these events can influence gyrification. Before six weeks of age, one additional mitotic cycle would potentially double the number of neural progenitor cells. This point can be illustrated in the story of the father who told his son at the beginning of the month, “I will either give you $20 now, or I will give you a penny now, and double the amount each day for the next month (that is, 1 cent, 2 cents, 4 cents, 16 cents, and so on).” If the son were to choose the penny, he would have more than 5.3 million dollars for a 30-day month or 10.6 million for a month with 31 days. The surface area of the brain has a close association with the number of radial units formed by symmetrical division along the ventricular zone (Rakic 1988, 1995). A larger number of radial units equates with a larger number of lined projections to the cortical plate and thus a greater surface area of the brain (Rakic, 1974; Sidman and Rakic, 1973). Since each round of mitosis results in an exponential increase in the number of progenitor cells, small changes affecting the duration of symmetric growth will have a dramatic impact on surface area (Rakic, 1995). This developmental principle has been called “late equals large,” as neurons migrating into late-developing brain structures undergo a longer period of symmetrical division, resulting in a larger size of these structures. The principle has been verified for the developmental time table and corresponding size of brain structures in a large number of different species (Finlay and Darlington, 1995; Striedter, 2005). When the embryos of monkeys are irradiated during the symmetric phase of progenitor cell division, there is a decrease in total surface area of the brain. However, when radiation is applied after six weeks, during the phase of asymmetric cell division, it results in a deletion of cortical cell layers and, in turn, a decrement in cortical thickness (Rakic, 1995), as well as disrupting the development of gyrification (Stewart et al., 1975). Thus the thickness of the sixlayered cortex is influenced by the asymmetric period of cell division. It has been found recently that the migration of neurons into the cortex is not as straightforward as initially thought; for example, several neuronal subpopulations show different migratory patterns (Nadarajah et al., 2003). In particular, the mode of radial migration is mostly followed by pyramidal cells, while different types of cortical interneurons pursue a tangential migratory path into their target layers

40

(Kriegstein and Noctor, 2004). The effect of the tangential mode of migration on cortical morphology is likely small or has only a local influence on the development of the cortical layer. The vast majority of cortical neurons are pyramidal cells that migrate along the predominant radial path. These pyramidal cells form connections with interneurons that are located within the vicinity of their migratory path. Another factor affecting brain morphology is cell death. This process can lead to the elimination of up to 50 percent of the initially formed neurons. However, cell death is a regressive event that occurs early in development (Cowan et al., 1984; Levitt, 2003) and thus may not be directly involved in cortical gyrification, which is a later process. Nonetheless, there are potential indirect consequences of cell death for gyrification and brain morphology that are as yet poorly understood. Development of Connectivity between Cortical and Subcortical Structures As the radial glial cells migrate toward the cortical plate, they meet with afferent cell projections from the thalamus (Rakic, 1988). These projections migrate primarily into layer IV of the cortex (Wise and Jones, 1978). The connection of these nerve cells allows for communication between the subcortical and cortical structures. Such a direct relation between cortical GM and the thalamus would presume a volumetric relation between the two structures. Studies have demonstrated high correlations between volumes of the thalamus and cortical GM, even when controlling for intracranial volume (White, Andreasen, and Nopoulos, 2002). This relation between the thalamus and cortical gray matter meshes well with the present neurodevelopmental and neuroanatomic understanding of thalamic/cortical GM connectivity. Studies that demonstrate volume decreases in both the thalamus and cortical regions (i.e., reductions in volume of both the prefrontal cortex and mediodorsal nucleus of the thalamus) may reflect aberrant patterns of connectivity between the two regions. (White, Andreasen, and Nopoulos, 2002). As the neurons migrate into the cortical plate, they extend apical and basilar dendrites (Juraska and Fifkova, 1979). The apical dendrites subsequently extend additional branches, with increasing complexity that continues during postnatal development (Conel, 1939). Chemical signals guide these developing dendrites toward their ultimate location, where synapses are formed (Sperry, 1963). While these connections form in the absence of the release of neuronal neurotransmitter, their maintenance, once formed, is activity dependent (Verhage, Maia, and Plomp, 2000). Connections are strengthened through activity, and those connections that see little traffic are eventually pruned (Luo and O’Leary, 2005). Thus a certain level of growth and pruning allows for inherent plasticity during neurodevelopment and explains

fundamentals of developmental neurobiology

why children with large sections of the brain removed may meet with minimal consequences as compared with the same resection in an adult patient. The development of the cerebral cortex requires an orchestration between the processes of neuronal migration, subcortical and cortical connectivity, interneuron development, and gyrification, all of which are temporally overlapping processes (Darlington, Dunlop, and Finlay, 1999). The formation of cortical and thalamocortical connectivity is not fully complete before gyrification, and a complex interplay exists between these processes that is as yet not completely understood.

The phylogeny of gyrification The elephant brain is four times larger, and the brain of the sperm whale is five to six times larger, than the human brain. Although these mammals certainly have a considerable level of intelligence when viewed within an environmentally adaptive context, larger absolute brain size does not necessarily translate into greater intelligence. A larger animal will generally possess a larger brain, owing to the larger surface area of the body and need for a greater number of neurons to cover the sensory and motor domains of the body. One can account for this relationship by normalizing brain mass by the weight of the animal. Brain weight scales to the ¾ power of body weight across both primates and nonprimates (Hofman, 1982). In addition, humans, porpoises, and dolphins have a disproportionately large brain compared to their body mass (Allman, 1998). Since humans, porpoises, and dolphins have the greatest amount of cortical folding, the relationship between body weight and surface area is considerably greater in these mammals. Finally, since the brain is a costly organ with respect to energy metabolism, there is evidence that one limitation in the enlargement of the brain is related to the balance between brain and body energy demands (Armstrong, 1982). The most conspicuous feature of the human brain is the disproportionately large size of its cerebral cortex, which is due to an extension of the developmental period of this latedeveloping structure (Finlay and Darlington, 1995). The result is a related increase in the surface area of the brain. Greater surface area equates to a larger amount of cortical gray matter and thus greater computational power. The phylogenetic increase in the surface area of the human brain has far exceeded growth in the cortical thickness (Welker, 1990). For example, in humans the surface area of the brain is 1,700 times larger than in shrews, yet the thickness of the cortex is only six times greater (Hofman, 1989). Compared to macaque monkeys, the surface area of the human brain is approximately ten times greater, whereas the human cortex is only twice as thick (Rakic, 1995). These comparisons indicate that during evolution the cortex

expanded laterally rather than vertically (Chenn and Walsh, 2002), resulting in a convoluted human cortical sheet that is about three times as large as the inner surface of the skull (Hilgetag and Barbas 2005, 2006; Richman et al., 1975; Toro and Burnod, 2005; Van Essen, 1997; Welker, 1990). Theoretically, the number of neurons in the cortex could also be increased by adding to cortical thickness rather than cortical surface. In this way, tripling cortical thickness, from about 5 to 15 mm, should allow a smooth human cortex with seemingly only a minor increase in brain volume. However, modeling studies (Murre and Sturdy, 1995; Ruppin, Schwartz, and Yeshurun, 1993) have demonstrated that this idea is ill-fated. Given the formidable degree of connectivity among cortical neurons (each forming, on average, a thousand or more connections with other neurons [Braitenberg and Schüz, 1998], the volume of wire grows exponentially with the number of neurons. Thus the extra projections required to link neurons in the additional cortical layers would lead to very uneconomical wiring in the thickened cortex, since connections within the cortex would need to take detours around the additional wiring volume (Chklovskii, Schikorski, and Stevens, 2002). These theoretical studies support the idea that the segregation of brain tissue into components of cell bodies within the GM and connections within the white matter (WM), in concert with the volume-saving folding of the cortical sheet, reflects an optimal wiring and volume arrangement for the very dense connectivity found in the primate cerebral cortex (Murre and Sturdy, 1995; Ruppin, Schwartz, and Yeshurun, 1993). Therefore, the computational prowess of the human brain appears linked to the overproportional size of its cerebral cortex and the intricate architecture of the cerebral cortex. This architecture includes the segregation of WM and GM and limiting the thickness of the gray matter sheet, which is enacted through a lateral increase of the cerebral cortex during evolution. At what point, however, did cortical folding become necessary? It appears that gyrification is strongly related to absolute increases in brain size. Prosimian and primate brains of up to 10 cm3 volume are generally smooth, while for larger volumes there is a close correlation between the degree of gyrification and absolute brain size (figure 10.1 in Striedter, 2005; Zilles et al., 1989). This critical threshold in brain size is likely related to constraints on other body systems and developmental events, such as the maximum birth size of the embryonic skull and its relation to the size of the pelvis and the birth process in humans (Striedter, 2005).

Theories of the ontogeny of gyrification What mechanisms produced the convoluted human cortex? Was it merely additional symmetric cell division with

white and hilgetag: gyrification and development of the human brain

41

subsequent squashing of the excess gray matter into the given space? Or, alternatively, were there additional neurodevelopmental changes that allowed not only for the increase in surface area, but also a mechanism to allow the newfound gray matter to be compacted within the given volume in a characteristic pattern? One need only study several brains within a species to notice the dramatic similarity in the folding patterns. Indeed, studies of monozygotic twins have demonstrated a strong genetic contribution to the development of brain topology (Bartley et al., 1993; Bartley, Jones, and Weinberger, 1997; Lohmann et al., 1999; White et al., 2002). In addition, specific genes have been found that are related to disorders affecting gyrification (Piao et al., 2004). Finally, as we explore the mechanistic theories of gyrification, is it possible that the folding patterns of the brain not only provide for increased surface area of the cortex, but also that gyrification enhances the efficiency of neuronal processing within the brain? During the third trimester of fetal life, the brain develops from a lissencephalic, or smooth, structure to a brain with convolutions that are characteristic of an adult brain (Armstrong et al., 1995; Naidich, Grant, and Altman, 1994; Retzius, 1891; Welker, 1990). Zilles and colleagues (1988) described a “gyrification index” (GI) that was applied to measure the developmental trajectory of gyrification in humans (Armstrong et al., 1995). The GI was calculated on coronal slices of the brain by calculating the ratio between the outline of the cortical surface and a lissencephalic outline of the brain (i.e., excluding the traces into the sulci). From this work on postmortem samples, they found that the gyrification index increases dramatically during the third trimester of fetal life, then remains relatively constant throughout development (Armstrong et al., 1995; Dareste, 1862). Since the brain nearly triples its volume from birth to adulthood, the process of gyrification continues to develop, maintaining this constant ratio. There is a sexual dimorphism in gyrification, with the female cerebral cortex more strongly convoluted than the male cortex (Luders et al., 2004). This finding implies that the volume difference between female and male brains is partly offset by a more efficient packing of the cortical sheet in the brains of females. The exact mechanisms underlying the gyrification of the brain are as yet unknown, although two overarching theories have emerged. These include the theory of gyrogenesis, or the theory that the gyri form as a result of active growth to specific regions of the brain (Le Gros Clark, 1945) and mechanical theories of gyrification based on the physical self-organization of the brain (Hilgetag and Barbas, 2005, 2006; Richman et al., 1975; Toro and Burnod, 2005; Van Essen, 1997; Welker, 1990). The mechanical theories can be further subdivided into those that involve differential growth patterns within the cortex (Richman

42

et al., 1975) and axonal-tension-based theories (Van Essen, 1997). A stimulating mechanical hypothesis was proposed by David Van Essen in 1997. This concept, known as the axonal-tension-based morphogenesis of the cerebral cortex, postulated that neuronal connectivity during early neurodevelopment is involved in producing fiber tension that draws interconnected regions closer together. If tension produced by the neuronal connections is involved in the mechanisms of gyrification, then changes in the patterns of the gyri and sulci are an expected outcome of brain changes that alter the connectivity between brain regions. Indeed, local and remote changes of gyrification have been observed after experimental white matter lesions in the developing primate brain (Goldman-Rakic, 1980; Goldman-Rakic and Rakic, 1984). The axonal-tension concept has also been supported by recent experimental findings in the primate brain as well as modeling studies (Hilgetag and Barbas, 2006; Toro and Burnod, 2005), and these studies support a link between brain surface morphology and regional neuronal connectivity within a developmental framework. Implications of the Mechanical Folding Hypothesis Combining the age-related differences in the morphology of the cerebral cortex with changes in neural connectivity is intriguing. The age-related decrease in synaptic and dendritic arborization may result in decreasing the tensile forces that are involved in the morphogenesis of the cerebral cortex. Histological studies of the neuronal pathways have found that the neural fibers on average tend to traverse horizontal to the surface in the sulci, whereas fiber pathways course on average more tangential in the gyri (Welker, 1990). From a mechanical perspective, a release of tension along the line of average tensile would result in widening of the sulci and greater curvature of the gyri. These changes have been found in a group of healthy adults (Magnotta et al., 1999). It is a well-known neuroscience concept that brain regions that wire together, fire together, that is, cooperate in brain function (Hilgetag et al., 2000). Thus another advantage of the tension-based morphogensis hypothesis is that regions that have strong interconnections and form functional circuits are maintained in relatively close proximity. This arrangement contributes to greater compactness of the brain and decreases the conduction time of neuronal signals along axonal fibers, thus enhancing overall efficiency (Van Essen, 1997). However, it needs to be kept in mind that cortical folding due to the balancing of axonal tension is played out as a global tug of war. Thus individual projections may deviate from the general shortening and straightening of fibers during development, so that some projections may remain relatively long and connect distant regions of the brain (Hilgetag and Barbas, 2006). This

fundamentals of developmental neurobiology

process would account for the long tail in the distance distribution of cortical projections (Kaiser and Hilgetag, 2006). While an admixture of long projections to mostly short cortical connectivity increases the global wiring length of the brain, it also contributes to a reduction of neural processing steps by providing long-distance shortcuts across the system (Kaiser and Hilgetag, 2006). Thus the overall efficiency of brain structure and function is determined by the simultaneous adaptation to multiple, partly opposing constraints. Generally, aspects of physical self-organization may make a substantial contribution to many morphogenetic processes. For instance, stretching and compression forces produced during the folding of the cortex influence the relative laminar thickness of cortical layers, resulting in a larger ratio of thickness of upper to deep layers in sulcal regions compared to gyral regions (Hilgetag and Barbas, 2006). Moreover, neuronal migration itself may be affected by the mechanical forces produced during cortical folding, as there is potential overlap in the developmental timetables of migration and folding. For example, the radial migration of late-maturing neurons through the already existing deep layers in budding gyri may be affected by lateral compression from the ongoing folding process, resulting in additional mechanical resistance to the neurons’ migration, and potentially in an additional accumulation of cells in the deep layers. In agreement with this idea, there are more pyramidal neurons and glia in the deep layers of gyri than in sulci or straight portions of cortex (Hilgetag and Barbas, 2006, unpublished observations); by contrast, there is no such clear trend in the data from monkey prefrontal cortex for different types of inhibitory interneurons (unpublished observations), which are predominantly migrating tangentially. Heritability of Gyrification Numerous studies have capitalized on nature’s clones, monozygotic (MZ) or identical twins, studied with and without dizygotic (fraternal twins) to study the interplay between genetic and environmental influences in neurodevelopment. Since parents and close friends are able to tell one identical twin from another, they are not completely identical (Machin, 1996). Considering the complexity of neurodevelopment, the role of stochastic processes (i.e., there are not enough genes to program all the specific connections of all neurons and their synapses), and the observation that some processes (i.e., pruning) are influenced by environmental factors, it should not be surprising that certain aspects of brain development are different in identical twins. Interestingly, the surface morphology of the brain has much greater variability than midline, volume, or subcortical regions of the brain (Bartley, Jones, and Weinberger, 1997; White et al., 2002). Studies evaluating the gyral patterns in monozygotic twins demonstrate that the deeper and developmentally earlier

sulci of the brain are more highly correlated than the tertiary sulci (Lohmann, von Cramon, and Steinmetz, 1999). In addition, there are regional differences between more conserved and more variable parts of the cortical landscape (Thompson et al., 2001). Nongenetic factors tend to have a greater influence on the tertiary sulci, which develop mainly after birth. Measures of gyral and sulcal curvature of the brain are significantly less correlated in twins than is the thickness of the cortex (White, Andreasen, and Nopoulos, 2002), supporting the phylogenetic differences between cortical thickness and surface area. It is possible that the greater nonshared environmental influences that are present for postnatal twins, coupled with the pronounced cortical plasticity inherent in early development, modulate the changes in cortical surface morphology. The strongest evidence for surface pattern dissimilarities between monozygotic twins came from a twin study by Bartley and colleagues (1997) that included both monozygotic and dizygotic twins. This study, which utilized a crosscorrelation algorithm to compare structural brain images, demonstrated that the majority of the morphologic variance between the brain surface morphology of MZ twins was a result of random environmental effects, while brain size, however, appeared to be strongly determined by genetic factors. Steinmetz and colleagues (1995) also demonstrated that MZ twins discordant for handedness exhibited differing degrees of asymmetry of the planum temporale. Epigenetic factors are probable explanations, not only for asymmetries of the planum temporale, but also for other surface measures of the brain, which show significantly lower correlations than the volumetric brain measures. It needs to be stressed, however, that genetic factors and physical self-organization are not mutually exclusive processes, but interact at all stages of development. For instance, genetic factors may underlie the timing of development of different cortical layers and areas, and mechanical factors may come into play as the cortex grows and cortical regions are interconnected by axonal fibers, resulting in the selforganization of migrating neurons, cortical layers, and cortical convolutions. Thus a dynamic interplay of genetics and simple physical principles underlies the formation of cortical convolutions as well as cortical architecture. Postnatal Changes in Brain Morphology and Gyrification During the first year of life, neurodevelopment changes take place at a rapid pace (Huttenlocher, 1979, 1990; Huttenlocher and Dabholkar, 1997; Huttenlocher and de Courten, 1987; Huttenlocher et al., 1982; Reiss et al., 1996; Yakovlev and Lecours, 1967). Synaptic density peaks by approximately the fourth postnatal month in the striate cortex (Huttenlocher and de Courten, 1987) and by approximately 12 postnatal months in the prefrontal cortex (Huttenlocher, 1979). Brain weight peaks at approximately

white and hilgetag: gyrification and development of the human brain

43

ten years of age, and MRI studies have demonstrated only a minor increase in mass after four to five years of age (Pfefferbaum et al., 1994; Reiss et al., 1996). Studies of gyrification during postnatal development have differed, depending on whether the sample was from postmortem brains or subjects who received magnetic resonance (MR) scans. Armstrong and colleagues (1995) demonstrated that postmortem brains maintained a constant gyrification index from birth through late adulthood. Alternatively, Magnotta and colleagues (1999) showed gradual changes in brain curvature, with the gyri developing greater curvature and the sulci developing less curvature, or becoming more broadened with increasing age. Since Armstrong and colleagues did not specifically study differences between the gyri and the sulci, and Magnotta and colleagues did not measure the gyrification index, it is possible that the opposite differences between the sulcal and gyral findings on MR, which were opposite in direction, canceled each other out, thus maintaining a constant gyrification index. Mixed results have resulted from evaluating age-related changes in total cell counts of the cortex (for review see Peters et al., 1998). Since the original studies of Brody (1955), who found that as many as half of the neurons in regions of the frontal and temporal lobes are lost, there have been conflicting reports of the extent of neuronal loss in the cortex (Anderson et al., 1983; Cragg, 1975; Devaney and Johnson, 1980; Haug, 1986; Haug et al., 1984; Henderson, Tomlinson, and Gibson, 1980; Shefer, 1973). The earlier studies that reported dramatic cell death apparently did not account for methodological techniques that resulted in a greater shrinkage of younger brains. This age-dependent shrinkage resulted in the appearance of a greater compaction of neurons within the brain (Haug, 1986; Haug et al., 1984). More recent studies that utilized techniques to reduce or control for brain shrinkage have not replicated the early findings, although the debate continues (Leuba and Kraftsik, 1994; Peters et al., 1998; Terry, De Teresa, and Hansen, 1987). The current consensus is that cortical neurons are generally preserved during adolescence and adulthood with at most a 10 percent reduction in neuronal numbers (Peters et al., 1998). Since this loss is mainly in older adults, little cell death appears to occur during adolescence, at least in the absence of the use of illicit substances. Even in the absence of considerable neuronal cell loss, developmental differences in brain structure and function have been described during adolescence and adult life (Huttenlocher, 1979; Huttenlocher et al., 1982; Giedd et al., 1996; Sowell et al., 2003, 2004; Sowell, Thompson, et al., 2002; Sowell, Trauner, et al., 2002). Huttenlocher and colleagues (Huttenlocher, 1979; Huttenlocher et al., 1982) have reported substantial decreases in synaptic density in the middle frontal gyrus during adolescence and early adulthood, this decrease being without considerable

44

neuronal loss. This decrease in synaptic density is associated with decreased plasticity and is likely related to the pruning of connections that have less functional efficiency in the developing neuropil (Easter et al., 1985). Histological studies have also demonstrated an age-related decrease in cortical thickness associated with the changes taking place in the neuropil (Jacobs et al., 1997). The thinning of the cortical GM during late childhood and adolescent development involves primarily modulations within the neuropil, namely, synaptic pruning and dendritic arborization. The age-related findings identified histologically have also been shown using magnetic resonance imaging (MRI) techniques (Giedd et al., 1997, 1996; Sowell et al., 2003, 2004; Sowell, Thompson, et al., 2002; Sowell, Trauner, et al., 2002; Thompson and Nelson, 2001). Cross-sectional MRI studies of development have consistently demonstrated a decrease in gray matter volume starting in late childhood or early adolescence and progressing into late adulthood (Caviness et al., 1996; De Bellis et al., 2001; Giedd, 2004; Giedd et al., 1996; Gogtay et al., 2004; Jernigan and Tallal, 1990; Jernigan et al., 1991; Lenroot and Giedd, 2006; Pfefferbaum et al., 1994; Reiss et al., 1996; Sowell et al., 2003; Sowell et al., 2004; Sowell, Trauner, et al., 2002). The age of peak GM volume varies between studies and ranges from 4 years (Pfefferbaum et al., 1994) to early adolescence (Giedd et al., 1996), and the cerebellum appears to be one of the later-developing brain structures. There appear to be age-related differences in volume loss, such that between childhood and adolescence there is greater gray matter loss in the parietal lobes (Sowell et al., 1999), whereas comparing adolescents to adults, the gray matter loss occurs more in the frontal and subcortical regions (Giedd et al., 1996; Sowell et al., 1999). MRI studies have demonstrated reductions in the thickness of the cortex that correspond with the decreases in gray matter volume (Sowell, Thompson, et al., 2002).

Clinical conditions associated with abnormalities in gyrification Using the analogy of an archaeologist piecing together specific clues about past events that are based on current finds, so too is it conceivable to piece together the timing of specific clinical aberrations of development based on current data. For example, based on what is known regarding neurodevelopment and gyrification—e.g., the recent reports of disruption of neuronal migration in rats exposed to long durations of ultrasound (Ang et al., 2006)—both changes in the thickness of the cortex and alterations in gyrification would be expected. Clinical conditions can be broken down into primary disorders of gyrification (those that cause gross disruptions in the gyrification that are readily identified on imaging or

fundamentals of developmental neurobiology

postmortem samples) and secondary alterations in gyrification (those that may influence gyrification through downstream processes). These distinctions define an arbitrary line, as the difference between a primary and secondary disorder of gyrification may only relate to the timing or the extent of the specific neurodevelopmental insult. Primary Disorders of Gyrification Lissencephaly (stemming from the Greek word lissos, meaning smooth, and enkephalos, meaning brain) is an umbrella term that describes a number of rare developmental malformations characterized by an absence or reduction of the gyri and sulci. Children with lissencephaly have severe developmental delays and mental retardation and are typically diagnosed within the first six months of life (Ross and Walsh, 2001). Malformations that are closely related to lissencephaly include pachygyria (pachy stemming from the Greek word pachys, meaning “thick”), a malformation in which the gyri are thicker and the sulci are less deep, and polymicrogyria, which results in a cortex with multiple small gyri and shallow sulci. Evidence shows that individuals with lissencephaly and related disorders have an arrest of normal neuronal migration during the third to fourth month of uterine life (Bielschowsky, 1923; Stewart, Richman, and Caviness Jr., 1975). This is evidenced by histological findings of heterotopic clumps of neurons in subcortical brain regions (Bielschowsky, 1923; Stewart, Richman, and Caviness Jr., 1975). As would be expected from the disruption of neuronal migration, the cortex is reduced from six to four layers, although the actual thickness of the cortex is increased (Bielschowsky, 1923; Josephy, 1944; Stewart, Richman, and Caviness Jr., 1975). This increase in thickness is related to the disruption in the migration and the subsequent spreading of the neurons within the cortex. Since the surface area of lissencephalic brains is reduced substantially and the gyri and sulci fail to form, neuronal migration plays an integral role in gyrification. It is possible that the dilution of neuronal fibers within the thicker cortex, coupled with the increased failure of tangential neuronal fibers to adequately produce neural networks, results in a lack of tension within the cortical surface. This then disrupts the formation of sulci and gryi during brain growth. Interestingly, it is postulated that when the disruption in the neuronal migration takes place earlier (before 18 weeks), there is a greater likelihood that the brain will show fewer gyri and sulci (i.e., lissencephaly), whereas between 18 and 24 weeks the gross morphological defects of gyrification will be less pronounced (pachygyria). Radial and horizontal migration of the deeper layers, prior to an insult to the CNS, may produce enough tension within these deeper layers to result in the development of some, albeit aberrant, gyrification.

Neuropsychiatric Disorders Associated with Altered Gyrification With the emerging evidence supporting a tension-based hypothesis for cortical morphogenesis (Hilgetag and Barbas, 2005, 2006; Van Essen, 1997), there is a direct translation between changes in neuronal connectivity and gyral and sulcal morphology (White et al., 2003; Zilles et al., 1988). Whereas the gyrification index described by Zilles and colleagues (1988) remains constant from birth through adulthood (Armstrong et al., 1995), it is well equipped to determine alterations that occurred during uterine life. Thus intrauterine events that alter gyrification would be evidenced in an alteration in the GI, and thus would support neurodevelopmental hypotheses for specific disorders (Kulynych et al., 1997; Vogeley et al., 2000). But because the gyrification index does not differentiate between sulcal and gyral changes, subtle changes in gyral and sulcal morphology that counteract each other would go undetected during normal aging and development using the GI (Kulynych et al., 1997). Thus Magnotta and colleagues (1999) developed an image analysis algorithm that is able to independently differentiate between sulcal and gyral changes in cortical surface morphology. Most studies of neuropsychiatric disorders utilize the GI to measure patient control differences. Studies in schizophrenia have been mixed, with several studies demonstrating a global decrease in the GI of the left cortex (Kulynych et al., 1997; Sallet et al., 2003) and the right cortex (Sallet et al., 2003; Vogeley et al., 2000). Alternatively, studies of gyrification in schizophrenia have also shown no patient/ control differences (Highley et al., 2003) and, interestingly, an increased GI in the right temporal lobe (Harris, Yates, et al., 2004). Given the variability of the findings of these studies and the clinical heterogeneity that is well known in schizophrenia, a full consensus has not been achieved regarding the specificity of neurodevelopmental abnormalities in gyrification in schizophrenia. Utilizing techniques developed by Magnotta and colleagues (1999) that are able to differentiate alterations in the gyri and sulci independently, White and colleagues (2003) found opposing changes in the gyri and sulci in the frontal and temporal regions in a group of children and adolescents with schizophrenia. The sulci developed greater curvature (i.e., became more broad), whereas the gyri became more peaked, demonstrating an increased curvature. Based on the histological differences between the sulci and the gyri (Welker, 1990), it was postulated that these changes were consistent with aberrant pruning under the tension-based morphogenesis hypothesis of gyrification. Several studies have emerged evaluating those at an increased risk for developing schizophrenia. Harris, Whalley, and colleagues (2004) found an increase in GI in the right prefrontal cortex of patients with schizophrenia, whereas Jou and colleagues (2005) found a decrease in the left frontal

white and hilgetag: gyrification and development of the human brain

45

lobe. Finally, individuals with velocardiofacial syndrome (VCFS), who have a nearly 30 times greater risk of developing schizophrenia than the general population, showed a significant decrease in the GI in both the frontal and parietal lobes (Schaer et al., 2006). Autism, which has been shown to have an altered developmental trajectory in brain volume (Courchesne, Redcay, and Kennedy, 2004; Piven et al., 1996), has only one study evaluating gyrification. In a cross-sectional study of brain gyrification in patients with autism, Hardan and colleagues (2004) found an increased GI in the frontal brain region. One hypothesis that may explain an increased GI in patients with autism, in addition to the reported decrease in GI in individuals with dyslexia (Casanova et al., 2004), involves abnormalities in the minicolumns, or the white matter fibers that course through the gyri (Casanova et al., 2002, 2006). Alterations in brain size or gyrification may alter the volume of white matter passing through these gyral columns, resulting in altered communication between distant brain regions and affecting processing speed and the integration of neuronal signals. Interestingly, we found a narrowing of the gyral alterations in our young patients with schizophrenia. Finally, an increase in GI has been demonstrated in patients with Williams syndrome (Schmitt et al., 2002). Although the increases were global, they were more pronounced in the left frontal, right parietal, and right occipital lobes. An increased GI may reflect a regionally diminished rate of pruning of synaptic and neuronal connections. The direct relationship between alterations in gyrification and subsequent behavioral changes for each of these disorders is an evolving area of research. However, a better understanding of the processes associated with gyrification of the brain may help us achieve a better understanding of factors associated with the pathology within the onset and course of these disorders.

Conclusions The development of the central nervous system and formation of the gyri and sulci in the brain involve an orchestration of highly complex processes. These processes begin in utero and progress throughout the life span. During the first two weeks of fetal life, there is a differentiation of cells within the ectoderm into cell lines that are destined to become the cells of the central nervous system. These cells divide and multiply rapidly, forming first the neural plate, then the neural tube. Cells that line the ventricular zone change from symmetric division to asymmetric division at approximately six weeks of gestational age. This time point is important, since alterations during the period of symmetric cell division are more likely to influence the surface area of the brain, and thus changes in the gyrification patterns.

46

Gyrification of the deepest, primary sulci begins near the end of the second trimester, although the brain has a predominant lissencephalic character at the beginning of the third trimester. The major changes in gyrification of the human brain take place during the third trimester with the necessary scaffolding likely being constructed prior to this time (Rakic, 2004). Although the mechanisms underlying gyrification are unknown, an interesting theory proposes that regions of greater neural connectivity apply greater tension, such that as the brain grows, areas of greater connectivity tend to be drawn closer together (Van Essen, 1997). This hypothesis would account for gyri being functional units with more efficient neural communication. Importantly, there are significant temporal overlaps between different developmental events, such as neuronal migration, formation of connections, and gyrification (Darlington, Dunlop, and Finlay, 1999), so that the processes of neuronal migration, formation of connections, and gyrification likely interact and affect each other. The most visible changes in brain structures occur during uterine life and the first few early years following birth. However, more subtle brain changes can be seen throughout childhood and adolescence, even into early to middle adulthood. Structural brain changes during adolescence and young adulthood include a decrease in gray matter, density changes of cells within the neuropil, and an increase in myelination. The brain surface morphology continues to change as well during adolescence and adult life, with the gyri becoming more steep and the sulci developing a broader appearance. These latter findings are associated with an increase of cerebrospinal fluid bathing the outer layer of the brain. With the continued evolution of electrophysiological and neuroimaging tools that are able to increase our knowledge of the complex processes involved with neurodevelopment, the upcoming years will continue to exponentially add to our understanding of brain development and gyrification. At the moment, however, many questions remain about the mechanisms and implications of the process of gyrification in the human brain. For example, what are the exact developmental mechanisms of gyrification, and which aberrations from typical development lead to the observable differences in brain shape in different patient groups? Answers to these questions will depend on a greater understanding of developmental events. (For example, when do specific cortical areas and their interconnections form?) Another important open question concerns the relationship between the amount, distribution, and development of white matter in the brain and the global and local degree of gyrification. Morphometric studies, as reviewed here, have demonstrated systematic differences in WM as well as gyrification between normal subjects and patients with a variety of neuropathologic conditions. While the axonal-tension hypothesis sug-

fundamentals of developmental neurobiology

gests that the two variables are linked through brain development, it is not yet clear how their relationship can be formalized in the context of given morphometric data. For instance, does a relative regional decrease in WM always lead to a regional increase or a regional decrease of gyrification, or can such relations only be addressed at the level of the whole brain? The recent advances in diffusion tensor imaging applied to studies of gyrification may help to address this question. Finally, processes of gyrification lead to systematic morphologic differences in different parts of the cortical landscape, affecting overall thickness and laminar morphology, as well as morphology at the cellular level (Hilgetag and Barbas, 2005, 2006; Richman et al., 1975; Toro and Burnod, 2005; Van Essen, 1997; Welker, 1990). Are there resulting functional differences between gyri and sulci as well? Such ideas may be tested by electrophysiology or high magnetic field functional imaging of cortical areas that extend across different cortical terrains at high resolution. These methodologies are likely to further our understanding of the relationship between brain structure and function in the developing brain. acknowledgments

The writing of this chapter was made possible by NIMH grant (MH068540). Correspondence concerning this chapter should be directed to Tonya White, M.D., at the Division of Child and Adolescent Psychiatry F256/2B, University of Minnesota, 2450 Riverside Avenue, Minneapolis, MN 55401, USA. We would like to acknowledge Dr. Canan Karatekin for her careful review of the manuscript and also thank Ms. Roxana Voitcu for her helpful comments. REFERENCES

Allman, J., 1998. Evolving Brains. New York: Scientific American Library. Anderson, J. M., B. M. Hubbard, G. R. Coghill, and W. Slidders, 1983. The effect of advanced old age on the neurone content of the cerebral cortex: Observations with an automatic image analyser point counting method. J. Neurol. Sci. 58:235–246. Ang, E. S., Jr., V. Gluncic, A. Duque, M. E. Schafer, and P. Rakic, 2006. Prenatal exposure to ultrasound waves impacts neuronal migration in mice. Proc. Natl. Acad. Sci. USA 103: 12903–12910. Armstrong, E., 1982. A look at relative brain size in mammals. Neurosci. Lett. 34:101–104. Armstrong, E., A. Schleicher, H. Omran, M. Curtis, and K. Zilles, 1995. The ontogeny of human gyrification. Cerebral Cortex 5:56–63. Bartley, A. J., D. W. Jones, E. F. Torrey, J. R. Zigun, and D. R. Weinberger, 1993. Sylvian fissure asymmetries in monozygotic twins: A test of laterality in schizophrenia. Biol. Psychiatry 34:853–863. Bartley, A. J., D. W. Jones, and D. R. Weinberger, 1997. Genetic variability of human brain size and cortical gyral patterns. Brain 120:257–269.

Bielschowsky, M., 1923. Ueber die oberflaechengestaltung des grosshirnmantels bei mikrogyrie und bei normaler entwicklung. J. Psychol. Neurol. (Lpz.) 30:29–76. Braitenberg, V., and A. Schüz, 1998. Cortex: Statistics and Geometry of Neuronal Connectivity, 2nd ed. Berlin: Springer. Brody, H. D., 1955. Organization of the cerebral cortex. J. Comp. Neurol. 1023:511–556. Casanova, M. F., J. Araque, J. Giedd, and J. M. Rumsey, 2004. Reduced brain size and gyrification in the brains of dyslexic patients. J. Child Neurol. 19:275–281. Casanova, M. F., D. P. Buxhoeveden, A. E. Switala, and E. Roy, 2002. Minicolumnar pathology in autism. Neurology 58:428– 432. Casanova, M. F., I. A. van Kooten, A. E. Switala, et al., 2006. Minicolumnar abnormalities in autism. Acta Neuropathol. (Berl.) 112:287–303. Caviness, V. S., Jr., D. N. Kennedy, C. Richelme, J. Rademacher, and P. A. Filipek, 1996. The human brain age 7–11 years: A volumetric analysis based on magnetic resonance images. Cerebral Cortex 6:726–736. Chenn, A., and C. A. Walsh, 2002. Regulation of cerebral cortical size by control of cell cycle exit in neural precursors. Science 297:365–369. Chklovskii, D. B., T. Schikorski, and C. F. Stevens, 2002. Wiring optimization in cortical circuits. Neuron 34:341–347. Conel, J. L., 1939. The Postnatal Development of the Human Cerebral Cortex. Cambridge, MA: Harvard University Press. Courchesne, E., E. Redcay, and D. P. Kennedy, 2004. The autistic brain: Birth through adulthood. Curr. Opin. Neurol. 17:489– 496. Cowan, W. M., J. W. Fawcett, D. D. O’Leary, and B. B. Stanfield, 1984. Regressive events in neurogenesis. Science 225: 1258–1265. Cragg, B. G., 1975. The density of synapses and neurons in normal, mentally defective aging human brains. Brain 98:81– 90. Dareste, M. C., 1862. Sur les rapports de la masse encéphalique avec le développement de l’intelligence. Bull. Soc. Anthropol. 3:26– 54. Darlington, R. B., S. A. Dunlop, and B. L. Finlay, 1999. Neural development in metatherian and eutherian mammals: Variation and constraint. J. Comp. Neurol. 411:359–368. De Bellis, M. D., M. S. Keshavan, S. R. Beers, et al., 2001. Sex differences in brain maturation during childhood and adolescence. Cerebral Cortex 11:552–557. Destrieux, C., S. Velut, and M. Kakou, 1998. [Development of the corpus callosum (CC)]. Neurochirurgie 44:11–16. Devaney, K. O., and H. A. Johnson, 1980. Neuron loss in the aging visual cortex of man. J. Gerontol. 35:836–841. Easter, S. S., D. Purves, P. Rakic, and N. C. Spitzer, 1985. The changing view of neural specificity. Science 230:507–511. Finlay, B. L., and R. B. Darlington, 1995. Linked regularities in the development and evolution of mammalian brains. Science 268:1578–1584. Giedd, J. N., 2004. Structural magnetic resonance imaging of the adolescent brain. Ann. NY Acad. Sci. 1021:77–85. Giedd, J. N., F. X. Castellanos, J. C. Rajapakse, A. C. Vaituzis, and J. L. Rapoport, 1997. Sexual dimorphism of the developing human brain. Prog. Neuropsychopharmacol. Biol. Psychiatry 21:1185– 1201. Giedd, J. N., J. W. Snell, N. Lange, et al., 1996. Quantitative magnetic resonance imaging of human brain development: Ages 4–18. Cerebral Cortex 6:551–560.

white and hilgetag: gyrification and development of the human brain

47

Gogtay, N., J. N. Giedd, L. Lusk, et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101:8174–8179. Goldman-Rakic, P. S., 1980. Morphological consequences of prenatal injury to the primate brain. Prog. Brain Res. 53:1–19. Goldman-Rakic, P. S., and P. Rakic, 1984. Experimental modification of gyral patterns. In N. Geschwind and A. Galaburda (eds.), Cerebral Dominance, 179–192. Cambridge, MA: Harvard University Press. Hardan, A. Y., R. J. Jou, M. S. Keshavan, R. Varma, and N. J. Minshew, 2004. Increased frontal cortical folding in autism: A preliminary MRI study. Psychiatry Res. 131:263–268. Harris, J. M., H. Whalley, S. Yates, P. Miller, E. C. Johnstone, and S. M. Lawrie, 2004. Abnormal cortical folding in high-risk individuals: A predictor of the development of schizophrenia? Biol. Psychiatry 56:182–189. Harris, J. M., S. Yates, P. Miller, J. J. Best, E. C. Johnstone, and S. M. Lawrie, 2004. Gyrification in first-episode schizophrenia: A morphometric study. Biol. Psychiatry 55:141–147. Haug, H., 1986. History of neuromorphometry. J. Neurosci. Methods 18:1–17. Haug, H., S. Kuhl, E. Mecke, N. L. Sass, and K. Wasner, 1984. The significance of morphometric procedures in the investigation of age changes in cytoarchitectonic structures of human brain. J. Hirnforsch. 25:353–374. Henderson, G., B. E. Tomlinson, and P. H. Gibson, 1980. Cell counts in human cerebral cortex in normal adults throughout life using an image analysing computer. J. Neurol. Sci. 46:113– 136. Highley, J. R., L. E. DeLisi, N. Roberts, et al., 2003. Sexdependent effects of schizophrenia: An MRI study of gyral folding, and cortical and white matter volume. Psychiatry Res. 124:11–23. Hilgetag, C. C., and H. Barbas, 2005. Developmental mechanics of the primate cerebral cortex. Anat. Embryol. (Berl.) 210:411– 417. Hilgetag, C. C., and H. Barbas, 2006. Role of mechanical factors in the morphology of the primate cerebral cortex. PLoS Comput. Biol. 2:e22. Hilgetag, C. C., G. A. Burns, M. A. O’Neill, J. W. Scannell, and M. P. Young, 2000. Anatomical connectivity defines the organization of clusters of cortical areas in the macaque monkey and the cat. Philos. Trans. R. Soc. London B. Biol. Sci. 355:91–110. His, W., 1874. Unsere Körperform und das physiologische problem ihrer Entstehung. Leipzig: F. C. W. Vogel. Hofman, M. A., 1982. Encephalization in mammals in relation to the size of the cerebral cortex. Brain Behav. Evol. 20:84–96. Hofman, M. A., 1989. On the evolution and geometry of the brain in mammals. Prog. Neurobiol. 32:137–158. Huttenlocher, P. R., 1979. Synaptic density in human frontal cortex—Developmental changes and effects of aging. Brain Res. 163:195–205. Huttenlocher, P. R., 1990. Morphometric study of human cerebral cortex development. Neuropsychologia 28:517–527. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Huttenlocher, P. R., and C. de Courten, 1987. The development of synapses in striate cortex of man. Hum. Neurobiol. 6:1–9. Huttenlocher, P. R., C. De Courten, L. J. Garey, and H. van der Loos, 1982. Synaptic development in human cerebral cortex. Int. J. Neurol. 17:144–154.

48

Jacobs, B., L. Driscoll, and M. Schall, 1997. Life-span dendritic and spine changes in areas 10 and 18 of human cortex: A quantitative Golgi study. J. Comp. Neurol. 386:661–680. Jernigan, T. L., and P. Tallal, 1990. Late childhood changes in brain morphology observable with MRI. Dev. Med. Child Neurol. 32:379–385. Jernigan, T. L., D. A. Trauner, J. R. Hesselink, and P. A. Tallal, 1991. Maturation of human cerebrum observed in vivo during adolescence. Brain 114:2037–2049. Josephy, H., 1944. Congenital agyria and a defect of the corpus callosum. J. Neuropathol. Exp. Neurol. 3. Jou, R. J., A. Y. Hardan, and M. S. Keshavan, 2005. Reduced cortical folding in individuals at high risk for schizophrenia: A pilot study. Schizophr. Res. 75:309–313. Juraska, J. M., and E. Fifkova, 1979. A Golgi study of the early postnatal development of the visual cortex of the hooded rat. J. Comp. Neurol. 183:247–256. Kaiser, M., and C. C. Hilgetag, 2006. Nonoptimal component placement, but short processing paths, due to long-distance projections in neural systems. PLoS Comput. Biol. 2:e95. Kandel, E. R., J. H. Schwartz, and T. M. Jessell, 2000. Principles of Neural Science, 4th ed. New York: McGraw-Hill Health Professions Division. Kriegstein, A. R., and S. C. Noctor, 2004. Patterns of neuronal migration in the embryonic cortex. Trends Neurosci. 27:392– 399. Kulynych, J. J., L. F. Luevano, D. W. Jones, and D. R. Weinberger, 1997. Cortical abnormality in schizophrenia: An in vivo application of the gyrification index. Biol. Psychiatry 41:995– 999. Le Gros Clark, W. E., 1945. Deformation patterns on the cerebral cortex. In W. E. LeGross Clark and P. B. Medawar (eds.), Essays on Growth and Form, 1–23. Oxford, UK: Oxford University Press. Lenroot, R. K., and J. N. Giedd, 2006. Brain development in children and adolescents: Insights from anatomical magnetic resonance imaging. Neurosci. Biobehav. Rev. 30:718–729. Leuba, G., and R. Kraftsik, 1994. Changes in volume, surface estimate, three-dimensional shape and total number of neurons of the human primary visual cortex from midgestation until old age. Anat. Embryol. (Berl.) 190:351–366. Levitt, P., 2003. Structural and functional maturation of the developing primate brain. J. Pediatr. 143:S35–45. Lohmann, G., D. Y. von Cramon, and H. Steinmetz, 1999. Sulcal variability of twins. Cerebral Cortex 9:754–763. Luders, E., K. L. Narr, P. M. Thompson, et al., 2004. Gender differences in cortical complexity. Nature Neurosci. 7:799–800. Luo, L., and D. D. O’Leary, 2005. Axon retraction and degeneration in development and disease. Annu. Rev. Neurosci. 28:127– 156. Machin, G. A., 1996. Some causes of genotypic and phenotypic discordance in monozygotic twin pairs. Am. J. Med. Genet. 61:216–228. Magnotta, V. A., N. C. Andreasen, S. K. Schultz, et al., 1999. Quantitative in vivo measurement of gyrification in the human brain: Changes associated with aging. Cerebral Cortex 9:151– 160. Murre, J. M., and D. P. Sturdy, 1995. The connectivity of the brain: Multilevel quantitative analysis. Biol. Cybern. 73:529– 545. Nadarajah, B., P. Alifragis, R. Wong, and J. Parnavelas, 2003. Neuronal migration in the developing cerebral cortex: Observations based on real-time imaging. Cerebral Cortex 13:607–611.

fundamentals of developmental neurobiology

Naidich, T. P., J. L. Grant, N. Altman, et al., 1994. The developing cerebral surface. Preliminary report on the patterns of sulcal and gyral maturation—Anatomy, ultrasound, and magnetic resonance imaging. Neuroimaging. Clin. N. Am. 4:201–240. Peters, A., J. H. Morrison, D. L. Rosene, and B. T. Hyman, 1998. Feature article: Are neurons lost from the primate cerebral cortex during normal aging? Cerebral Cortex 8:295–300. Pfefferbaum, A., D. H. Mathalon, E. V. Sullivan, J. M. Rawles, R. B. Zipursky, and K. O. Lim, 1994. A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch. Neurol. 51:874–887. Piao, X., R. S. Hill, A. Bodell, et al., 2004. G protein–coupled receptor-dependent development of human frontal cortex. Science 303:2033–2036. Piven, J., S. Arndt, J. Bailey, and N. Andreasen, 1996. Regional brain enlargement in autism: A magnetic resonance imaging study. J. Am. Acad. Child Adolesc. Psychiatry 35:530–536. Pomeroy, S. L., and J. Y. Kim, 2000. Biology and pathobiology of neuronal development. Ment. Retard. Dev. Disabil. Res. Rev. 6: 41–46. Rakic, P., 1974. Neurons in rhesus monkey visual cortex: Systematic relation between time of origin and eventual disposition. Science 183:425–427. Rakic, P., 1988. Specification of cerebral cortical areas. Science 241:170–176. Rakic, P., 1995. A small step for the cell, a giant leap for mankind: A hypothesis of neocortical expansion during evolution. Trends Neurosci. 18:383–388. Rakic, P., 2000. Radial unit hypothesis of neocortical expansion. Novartis Found. Symp. 228:30–42. Rakic, P., 2004. Neuroscience: Genetic control of cortical convolutions. Science 303:1983–1984. Rakic, P., and P. I. Yakovlev, 1968. Development of the corpus callosum and cavum septi in man. J. Comp. Neurol. 132:45– 72. Reiss, A. L., M. T. Abrams, H. S. Singer, J. L. Ross, and M. B. Denckla, 1996. Brain development, gender and IQ in children: A volumetric imaging study. Brain 119:1763–1774. Retzius, A., 1891. Ueber den Bau der Oberflächenschicht der Grosshirnrinde beim Menschen und bei den Säugethieren. Verh. Biol. Ver. 3:90–103. Richman, D. P., R. M. Stewart, J. W. Hutchinson, and V. S. Caviness, Jr., 1975. Mechanical model of brain convolutional development. Science 189:18–21. Ross, M. E., and C. A. Walsh, 2001. Human brain malformations and their lessons for neuronal migration. Annu. Rev. Neurosci. 24: 1041–1070. Ruppin, E., E. L. Schwartz, and Y. Yeshurun, 1993. Examining the volume efficiency of the cortical architecture in a multiprocessor network model. Biol. Cybern. 70:89–94. Sallet, P. C., H. Elkis, T. M. Alves, et al., 2003. Reduced cortical folding in schizophrenia: An MRI morphometric study. Am. J. Psychiatry 160:1606–1613. Schaer, M., J. E. Schmitt, B. Glaser, F. Lazeyras, J. Delavelle, and S. Eliez, 2006. Abnormal patterns of cortical gyrification in velo-cardio-facial syndrome (deletion 22q11.2): An MRI study. Psychiatry Res. 146:1–11. Schmitt, J. E., K. Watts, S. Eliez, U. Bellugi, A. M. Galaburda, and A. L. Reiss, 2002. Increased gyrification in Williams syndrome: Evidence using 3D MRI methods. Dev. Med. Child Neurol. 44:292–295. Shefer, V. F., 1973. Absolute number of neurons and thickness of the cerebral cortex during aging, senile and vascular dementia,

and Pick’s and Alzheimer’s diseases. Neurosci. Behav. Physiol. 6:319–324. Sidman, R. L., and P. Rakic, 1973. Neuronal migration, with special reference to developing human brain: A review. Brain Res. 62:1–35. Sowell, E. R., B. S. Peterson, P. M. Thompson, S. E. Welcome, A. L. Henkenius, and A. W. Toga, 2003. Mapping cortical change across the human life span. Nature Neurosci. 6:309–315. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 1999. In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nature Neurosci. 2:859–861. Sowell, E. R., P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, and A. W. Toga, 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24:8223–8231. Sowell, E. R., P. M. Thompson, D. Rex, et al., 2002. Mapping sulcal pattern asymmetry and local cortical surface gray matter distribution in vivo: Maturation in perisylvian cortices. Cerebral Cortex 12:17–26. Sowell, E. R., D. A. Trauner, A. Gamst, and T. L. Jernigan, 2002. Development of cortical and subcortical brain structures in childhood and adolescence: A structural MRI study. Dev. Med. Child Neurol. 44:4–16. Sperry, R. W., 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. USA 50:703– 710. Steinmetz, H., A. Herzog, G. Schlaug, Y. Huang, and L. Jancke, 1995. Brain (A) symmetry in monozygotic twins. Cerebral Cortex 5:296–300. Stewart, R. M., D. P. Richman, and V. S. Caviness, Jr., 1975. Lissencephaly and pachygyria: An architectonic and topographical analysis. Acta. Neuropathol. (Berl.) 31:1–12. Striedter, G., 2005. Principles of Brain Evolution. Sunderland, MA: Sinauer Associates. Swanson, L., 2003. Brain Architecture. Oxford, UK: Oxford University Press. Terry, R. D., R. DeTeresa, and L. A. Hansen, 1987. Neocortical cell counts in normal human adult aging. Ann. Neurol. 21:530– 539. Thompson, P. M., T. D. Cannon, K. L. Narr, et al., 2001. Genetic influences on brain structure. Nature Neurosci. 4:1253–1258. Thompson, R. A., and C. A. Nelson, 2001. Developmental science and the media: Early brain development. Am. Psychol. 56:5–15. Toro, R., and Y. Burnod, 2005. A morphogenetic model for the development of cortical convolutions. Cerebral Cortex 15:1900– 1913. Van Essen, D. C., 1997. A tension-based theory of morphogenesis and compact wiring in the central nervous system. Nature 385:313–318. Verhage, M., A. S. Maia, J. J. Plomp, et al., 2000. Synaptic assembly of the brain in the absence of neurotransmitter secretion. Science 287:864–869. Vogeley, K., T. Schneider-Axmann, U. Pfeiffer, et al., 2000. Disturbed gyrification of the prefrontal region in male schizophrenic patients: A morphometric postmortem study. Am. J. Psychiatry 157:34–39. Welker, W., 1990. Why does cerebral cortex fissure and fold? In E. G. Jones and A. Peters (eds.), Cerebral Cortex, vol. 8B, 3–136. New York: Plenum Press. White, T., N. C. Andreasen, and P. Nopoulos, 2002. Brain volumes and surface morphology in monozygotic twins. Cerebral Cortex 12:486–493.

white and hilgetag: gyrification and development of the human brain

49

White, T., N. C. Andreasen, P. Nopoulos, and V. Magnotta, 2003. Gyrification abnormalities in childhood- and adolescentonset schizophrenia. Biol. Psychiatry 54:418–426. Wise, S. P., and E. G. Jones, 1978. Developmental studies of thalamocortical and commissural connections in the rat somatic sensory cortex. J. Comp. Neurol. 178:187–208. Yakovlev, P. I., and A. R. Lecours, 1967. The myelogenetic cycles of regional maturation of the brain. In A. Minkowski (ed.),

50

Regional Development of the Brain in Early Life, 3–70. Oxford, UK: Blackwell. Zilles, K., E. Armstrong, K. H. Moser, A. Schleicher, and H. Stephan, 1989. Gyrification in the cerebral cortex of primates. Brain Behav. Evol. 34:143–150. Zilles, K., E. Armstrong, A. Schleicher, and H. J. Kretschmann, 1988. The human pattern of gyrification in the cerebral cortex. Anat. Embryol. (Berl.) 179:173–179.

fundamentals of developmental neurobiology

4

Adult Neurogenesis in the Hippocampus YEVGENIA KOZOROVITSKIY AND ELIZABETH GOULD

Introduction Most neurons in the adult mammalian brain are produced during the embryonic period, but a substantial number of granule cells are born in the cerebellum, olfactory bulb, and hippocampus during the postnatal period (Altman and Das, 1966; Altman, 1969). In the olfactory bulb and the dentate gyrus of the hippocampus, neurogenesis continues throughout adulthood at relatively high levels (Altman and Das, 1965). Adult neurogenesis under normal conditions has also been reported in multiple other brain regions, including neocortex, striatum, amygdala, hypothalamus, and substantia nigra (Altman and Das, 1965; Kaplan, 1981; Huang, De Vries, and Bittman, 1998; Gould, Reeves, et al., 1999, 2001; Zhao et al., 2003; Bernier et al., 2002; Fowler et al., 2002; Fowler, Johnson, and Wang, 2005; Dayer et al., 2005; Xu et al., 2005; Bedard, Gravel, and Parent, 2006; Luzzati et al., 2006; Runyan, Weickert, and Saunders, 2006), but these findings remain controversial. In contrast, the idea that neurogenesis persists in the adult hippocampus has gained widespread acceptance in the neuroscience community, although the role of adult-generated neurons in the function of neural systems and in behavior is still a matter of debate. Adult neurogenesis in the hippocampus has been demonstrated in the brains of all vertebrate species investigated thus far, including humans (Altman and Das, 1965; Barnea and Nottebohm, 1994; Polenov and Chetverukhin, 1993; PerezCanellas and Garcia-Verdugo, 1996; Gould, Reeves, et al., 1998, 1999; Eriksson et al., 1998; Zupanc, 1999), suggesting that it is a highly conserved process. Quantitative estimates indicate that approximately 10,000 new cells are added to the young adult rat dentate gyrus every day (Cameron and McKay, 2001); a substantial but apparently smaller amount of neurogenesis occurs in the adult primate hippocampus (Eriksson et al., 1998; Gould, Reeves, et al., 1998, 1999). The discovery of adult neurogenesis has already altered the way the adult brain is viewed in terms of its potential for structural change. Studying adult neurogenesis also opens the possibility of therapeutic application in cases of developmental abnormalities, neurodegenerative disease, or traumatic brain injury. To further these goals, the role that adult-generated neurons play in normal functional circuitry must be elucidated, and the factors that regulate the produc-

tion, differentiation, survival, and integration of newly generated neurons need to be identified. In this chapter, we consider some of the key experimental findings that led to the discovery of adult neurogenesis, discussing the techniques used to quantify neurogenesis in the adult brain. We also examine several factors and conditions that regulate neurogenesis in the hippocampus and consider the possible role of adult-generated cells in hippocampal function and behavior. Adult neurogenesis is an important example of a developmental neural process that persists throughout life, with postnatal experience substantially modifying both brain and behavior.

The discovery of neurogenesis in the adult mammalian brain The study of adult neurogenesis depends on methods that can selectively label newly generated neurons. Early histologists relied on methods which, despite excellent spatial resolution, were not adequately cell-type specific and thus severely limited the investigation of developmental or adult neurogenesis. In 1965, Altman used 3H-thymidine autoradiography and suggested the possibility that neurogenesis occurs in the olfactory bulb and dentate gyrus of adult rats (Altman and Das, 1965). This technique takes advantage of the fact that cells undergoing DNA synthesis incorporate thymidine. When injected into an animal, 3H-thymidine is taken up by cells in the DNA synthetic phase of the cell cycle, radioactively tagging the dividing cell and all its subsequent progeny (although the signal becomes diluted with each subsequent division). Stable and specific, this technique proved useful for mapping mitotically active cell populations at the single-cell level. While it was clear from the work of Altman that new cells are added to the adult rodent olfactory bulb and hippocampus, the identity and fate of these cells remained in question, because light microscopic examination was not sufficient to unequivocally identify them as neurons. Kaplan and Hinds partially resolved this issue by using electron microscopy to demonstrate, at the ultrastructural level, that adultgenerated neurons in the dentate gyrus of the rat receive synaptic input on their cell bodies and dendrites (Kaplan and Hinds, 1977). However, since some glial cells receive

51

synaptic input (Oppenheim, Chu-Wang, and Maderdrut, 1978), immunohistochemical advances that enable staining for cell-type-specific antigens and combined thymidine labeling/retrograde axon tracing were required to generate definitive evidence in support of adult neurogenesis in the dentate gyrus. These studies showed that 3H-thymidine-labeled cells expressed markers of mature neuronal phenotype, such as neuron-specific enolase (Cameron et al., 1993; Okano, Pfaff, and Gibbs, 1993). In addition, 3H-thymidine-labeled cells were shown to incorporate retrograde tracers injected into the CA3 region (Stanfield and Trice, 1988), suggesting that they direct axons to the major target site of developmentally generated granule cells. Altogether, these studies provided convincing evidence that the adult mammalian dentate gyrus produces cells with neuronal characteristics, which are incorporated into the hippocampal circuitry. Although the existence of adult neurogenesis in the brains of rodents had been firmly established, early discoveries did not succeed in encouraging the in-depth study of this phenomenon. The observation that the adult brains of less complex vertebrates, such as fish and amphibians, undergo a remarkable degree of regeneration, whereas the adult mammalian brain does not, seemed incongruous with the concept of persistent neurogenesis in the adult mammalian brain. Brain development in mammals was thought to be a temporally defined process, completed long before adulthood. The lack of interest in adult neurogenesis within the neuroscience community may have been partially attributable to the inability to find adult neurogenesis in primates. For example, 3H-thymidine-labeled cells in the dentate gyrus of adult macaques were few in number and lacked the morphological characteristics of neurons; it was therefore concluded that they were most likely to be glia (Rakic, 1985; Eckenhoff and Rakic, 1988). As a result, adult neurogenesis was discounted as being specific to only certain vertebrate species and therefore of relatively little importance. Methodological Advances Lead to the Discovery of Hippocampal Neurogenesis in Adult Primates The development of 5-bromo-2′-deoxyuridine (BrdU) labeling was a major breakthrough in the study of adult neurogenesis. This method provides a sensitive means for assessing the number of adult-generated cells. BrdU is an analogue of thymidine, which is incorporated into DNA and can be immunohistochemically detected using specific antibodies. This technique has several advantages over 3H-thymidine autoradiography: (1) BrdU-labeled cells throughout an entire brain section (40 μm thick or even thicker) can easily be visualized, while 3H-thymidine autoradiography only allows for the detection of labeled cells in the top 1–3 μm of a tissue section; (2) the sensitivity of BrdU labeling can be amplified, since it is an immunocytochemical method; and (3) the technique is highly compatible with the immunolabeling of

52

cell-specific antigens used to determine the phenotype of cells. These advantages of the BrdU labeling method allowed researchers to reinvestigate the issue of adult neurogenesis in primates, establishing its occurrence in adult monkeys and humans (Gould, Reeves, et al., 1998, 1999; Eriksson et al., 1998; Kornack and Rakic, 1999). BrdU labeling combined with stereological techniques allowed for the quantification of neurogenesis in the adult brain. These studies have revealed that several thousand new granule neurons are generated daily in the dentate gyrus of adult rats (Tanapat et al., 1999; Cameron and McKay, 2001). The majority of these adult-generated cells express immature and mature neuronal markers, such as beta III–tubulin (TUJ1) and neuronal nuclei (NeuN). Since the advent of BrdU labeling, additional methods for tagging new neurons have been developed. One promising method involves infecting the new neurons with a retrovirus that drives expression of a fluorescent protein (van Praag et al., 2002; Zhao et al., 2006). Retroviral tools have enabled a detailed morphological characterization of dendrites and axons from new neurons, as well as electrophysiological recordings from the new neurons.

Development of the dentate gyrus During embryonic development, granule neurons arise from progenitor cells located in a discrete part of the neuroepithelium in the wall of the lateral ventricle, migrate across the incipient hippocampus and come to reside in the granule cell layer (Altman and Bayer, 1990). Some progenitor cells remain in the hilus and the subgranular zone of the dentate gyrus without undergoing final cell division. These progenitors continue to divide, giving rise to daughter cells that differentiate into granule neurons in the adult brain (figure 4.1). In the rat, most dentate gyrus neurons are produced during the first two postnatal weeks; by the end of the second week, the granule cell layer has been formed, and cell proliferation and migration decrease significantly. In contrast, the granule cell layer in the macaque forms during the prenatal period (Nowakowski and Rakic, 1981). Yet new granule neurons continue to be produced in the adult dentate gyrus of both rodents and primates from progenitors located in the hilus and subgranular zone. The addition of these new neurons to the dentate gyrus in adulthood reflects a turnover of the adult-generated neuron population, because available evidence suggests that many of the new cells die within weeks of their birth (Cameron et al., 1993).

Hormones regulate adult neurogenesis Steroid hormones play a well-known role in the organization and activation of certain behaviors. Numerous studies

fundamentals of developmental neurobiology

Total number of BrdU-labeled cells in the dentate gyrus

4000

Figure 4.1 Granule cell neurogenesis in the adult dentate gyrus of the hippocampus. Precursor cells residing in the hilus and the subgranular zone of the dentate gyrus divide throughout adult life, giving rise to daughter cells that migrate into the granule cell layer and differentiate into granule neurons.

suggest that these hormones affect behavior by altering the structural development of the brain. Over the past few decades, studies have demonstrated that steroid hormones also alter the structure of the adult brain by influencing dendrites, synapses, and cell survival (Sapolsky, Krey, and MeEwen, 1985; Woolley et al., 1990). More recently, adrenal and ovarian steroids were observed to alter the proliferation of granule cell precursors in the adult brain (Cameron and Gould, 1994; Tanapat et al., 1999). Some evidence also supports the involvement of thyroid hormones in the modulation of adult hippocampal neurogenesis (Desouza et al., 2005; Ambrogini et al., 2005), but this chapter concentrates on the effects of adrenal and ovarian steroids, which have been studied in depth. Adrenal Steroids Several lines of evidence point to an inverse relationship between adrenal steroid levels and the proliferation of granule cell precursors. First, a negative correlation between levels of circulating adrenal steroids and hippocampal granule neuron production is observed across the life span. In the rat, during the first two postnatal weeks known as the stress hyporesponsive period, low levels of circulating adrenal steroids coincide with maximal granule neuron production (Schlessinger et al., 1975; Sapolsky and Meaney, 1986). During adulthood, when levels of circulating adrenal steroids are higher, the rate of granule neuron production decreases. With aging, the production of new cells is further diminished in aged rats (Kuhn et al., 1996) and macaques (Gould, Reeves, et al., 1999), coincident with increases in the levels of circulating glucocorticoids (Sapolsky and Altmann, 1991; Sapolsky, 1992). Experimental manipulations of glucocorticoids have confirmed this negative relationship between the levels of adrenal steroids and granule neuron production. Increases in glucocorticoids during the stress hyporesponsive period diminish the rate of granule cell production in the dentate gyrus (Gould et al., 1991). Removal of adrenal steroids by means of

*

3000

2000

1000

0 Ovx

Ovx + E2

Ovx + E2 + P4

Figure 4.2 Stereological estimates of the number of BrdU-labeled cells in the dentate gyrus of adult ovariectomized (Ovx) rats that were treated with vehicle estrogen, estrogen, and progesterone. Treatment with progesterone 48 hours following the last two daily injections with estradiol rapidly reverses an estrogen-induced increase in cell proliferation. Error bars represent SEM, p < 0.05 relative to controls.

adrenalectomy in adult animals leads to an increase in cell proliferation and neuronal production, whereas treatment with the adrenal glucocorticoid corticosterone results in a decrease in these measures (Gould et al., 1992). Removal of circulating glucocorticoids in aged rats returns cell proliferation in the dentate gyrus to the levels observed in adrenalectomized young adults (Cameron and McKay, 1999). Collectively, these observations illustrate that adrenal steroids are potent mediators of adult neurogenesis and suggest that stressful psychological and physical experiences may act to suppress adult neurogenesis. Ovarian Steroids In contrast to the suppressive action of adrenal steroids on adult neurogenesis, the ovarian steroid hormone estrogen has been shown to stimulate the production of new granule neurons in the dentate gyrus (Tanapat et al., 1999; Tanapat, Hastings, and Gould, 2005; Ormerod et al., 2003). In the rat, a natural fluctuation in cell proliferation is observed across the estrous cycle—the production of new granule cells is greatest during proestrus, the stage of maximal estrogen levels. The removal of estrogen by means of ovariectomy results in a decrease in the proliferation of granule cell precursors, while replenishing estrogen levels rapidly reverses this effect (figure 4.2). In addition to increasing the production of new cells, estrogen affects the survival of adult-generated cells in the dentate gyrus (Tanapat et al., 1999). These effects account for a sex difference in adult neurogenesis, such that female rats produce more new granule cells than males (Tanapat et al., 1999). This effect is temporary, since under standard laboratory conditions,

kozorovitskiy and gould: adult neurogenesis in the hippocampus

53

Number of BrdU-labeled cells

many of these new cells degenerate over time. Estrogeninduced increases in the pool of immature granule neurons may still exert an impact on hippocampal function, as will be examined further later on.

Experience regulates adult neurogenesis Because the majority of neurons in the dentate gyrus are produced postnatally, this brain region has the potential to undergo functionally significant experience-dependent structural changes during postnatal development and adulthood. The remainder of this chapter focuses on several experiences currently known to regulate adult neurogenesis in the hippocampus and the implications of these findings for the function of the hippocampus, as well as behavior. Stress Numerous studies have demonstrated that stressful experiences alter adult neurogenesis in the dentate gyrus by decreasing cell proliferation (see Mirescu and Gould, 2006, for review). During the stress hyporesponsive period, most stressors that normally activate the hypothalamic-pituitaryadrenal (HPA) axis fail to elevate circulating glucocorticoids. However, exposure to the odors of natural predators is sufficient to increase levels of circulating glucocorticoids in male rat pups (Tanapat, Galea, and Gould, 1998). Given that adrenal steroids are known to suppress the production of new neurons, it is not surprising that exposure to the odors of an unfamiliar adult male rat also suppresses the proliferation of granule cell precursors in male rat pups (Tanapat, Galea, and Gould, 1998). Stressful experiences during the early postnatal period can have a very longlasting impact on hippocampal neurogenesis. For example, daily maternal separation is associated with reduced levels of neurogenesis in the offspring when they grow up (figure 4.3), long after the basal and stress levels of the main rat stress hormone corticosterone return to normal (Mirescu, Peters, and Gould, 2004). Stress suppresses cell proliferation in the dentate gyrus of adult animals as well. In adult male rats, exposure to trimethylthiazoline, the main component of fox feces, results in an activation of the HPA axis accompanied by an inhibition of the proliferation of granule cell precursors (Mirescu, Peters, and Gould, 2004; Galea, Tanapat, and Gould, 2006). The effect of stress on neurogenesis is observed across many mammalian species, using a variety of stress-inducing paradigms. For example, adult marmosets demonstrate significant decreases in the proliferation of granule cell precursors when exposed to the social stress of a resident intruder paradigm (Gould et al., 1998); adult tree shrews show decreased neurogenesis after exposure to subordination stress (Gould et al., 1997); and a similar picture is observed in adult rats following sleep deprivation (Guzman-Marin et al., 2003; Roman et al., 2005; Mirescu et al., 2006; van der Borght

54

Cage control Large platform Small platform

8000

6000

*

4000

* 2000

0 2 hr

1 wk

3 wk

Survival following BrdU injection

Figure 4.3 Reduced cell proliferation and adult neurogenesis after prolonged sleep deprivation. Rats subjected to 72 hours of sleep deprivation (small platform, large platform controls, cage controls) received a single injection of BrdU and were perfused 2 hours, 1 week, or 3 weeks thereafter. Sleep-deprived rats had lower numbers of BrdU-labeled cells in the subgranular zone/granule cell layer of the dentate gyrus. Error bars represent SEM, *p < 0.05 relative to LP and CC controls, ♦ p < 0.05 relative to LP controls.

et al., 2006). The impact of stress on adult neurogenesis can be additive over time. Chronic exposure to subordination stress, which does not evoke adaptation of the HPA axis, results in continual suppression of cell proliferation and a decrease in the volume of the granule cell layer (Fuchs et al., 1997). Collectively, these studies indicate that stress inhibits the proliferation of granule cell precursors during development, as well as in adulthood, in a variety of mammals. The factors that underlie the effect of stress on cell proliferation are not fully understood; however, glucocorticoid-induced reduction in cell proliferation through an NMDA receptor– dependent mechanism is likely to play a role (Cameron and Gould, 1994; Gould et al., 1997). While the stress-induced rise in glucocorticoid levels downregulates these measures, a stress hormone–independent rebound in cell proliferation has been reported following stressor cessation (figure 4.4) (Mirescu et al., 2006). The relationship between stressful experiences and adult neurogenesis reductions is not always simple. For example, wheel running increases the levels of circulating glucocorticoids in adult rodents (Droste et al., 2003), while at the same time enhancing neurogenesis by affecting cell proliferation (van Praag, Kempermann, and Gage, 1999; Stranahan, Khalil, and Gould, 2006). Since running is a positive stressor, in the sense that rats seek access to a running wheel, one possibility is that positive and negative stressors affect adult neurogenesis differently. In addition, the effect of running on the number of adult-generated neurons is mediated by social context—after a brief period of running, socially housed rats have increased levels of neurogenesis, while singly housed animals show a decrement in this measure

fundamentals of developmental neurobiology

*

Cage control Small platform

10000 8000 6000 4000 2000

10000 BrdU-labeled cells in the dentate gyrus

Number of BrdU-labeled cells

12000

Control Runner

*

8000 6000 4000 2000 0 3d

6d

12 d

24 d

48 d

0 6 hr

1 wk

2 wk

Recovery period

Figure 4.4 Lasting effects of prolonged sleep deprivation on cell proliferation and adult neurogenesis. Rats subjected to 72 hours of sleep deprivation received a single injection of BrdU 6 hours, 1 week, or 2 weeks later, and were perfused in 2 weeks. Adult neurogenesis was enhanced after 1 week of recovery from sleep deprivation. Error bars represent SEM, p < 0.05 relative to controls.

Figure 4.5 Social context mediates the influence of short-term running on adult neurogenesis. Rats received daily injection of BrdU and were sacrificed on day 12. The number of BrdU-labeled cells was greater in group-housed runners compared to controls, but lower in single-housed runners compared to controls. Error bars represent SEM, p < 0.05 relative to controls.

(figure 4.5); rats living in isolation eventually manifest the running-related increase in adult neurogenesis, but this effect requires long-term exercise (figure 4.6) (Stranahan, Khalil, and Gould, 2006). Stressors appear to regulate adult neurogenesis in distinct ways that may depend on the valence of the stressor and the social context of the experience. Learning Studies have reported changes in the number of new hippocampal granule neurons in adult animals living in conditions associated with enhanced learning opportunities.

Figure 4.6 A longer duration of physical activity is required to enhance cell proliferation in socially isolated rats. Separate cohorts of individually housed rats ran for 3–48 days before being injected once with BrdU. The number of BrdU-labeled cells in the dentate gyrus was increased only after 48 days of running. Error bars represent SEM, p < 0.05 relative to controls.

One example is the striking relationship between experience and hippocampal neurogenesis in black-capped chickadees (Barnea and Nottebohm, 1994). New neurons persist for longer periods of time during parts of the year when these birds engage in seed storage and retrieval, behaviors that are likely to involve spatial navigation learning and, thus, the hippocampal region. Additionally, black-capped chickadees that live in the wild retain more new hippocampal neurons than those that live in captivity. Others have shown that mice living in an enriched laboratory environment maintain more new hippocampal granule neurons than those living in standard laboratory control cages. In rodents, the influence of environmental complexity on the number of new granule neurons appears to be maintained throughout the life span, affecting juvenile, young adult, and aged rodents (Kempermann, Kuhn, and Gage, 1997; 1998). While many variables such as stress, social interaction, nutrition, and activity levels differ between animals living in the wild and in a laboratory setting, these findings present the possibility that increased learning opportunities alter the survival of new neurons. It is possible that the positive impact of social dominance on adult neurogenesis in the dentate gyrus (figure 4.7) (Kozorovitskiy and Gould, 2004) might also relate to the differences in the opportunities for social learning that are available to dominant and subordinate animals living in a seminaturalistic environment. Multiple studies have addressed the link between learning and adult neurogenesis in rodents, and the overall results have been mixed (see Leuner, Gould, and Shors, 2006, for review). We and others have found that the types of learning that require the hippocampus tend to increase the number of new neurons in the dentate gyrus of adult rats (Leuner et al., 2004; Olariu et al., 2005). In animals living in standard laboratory conditions, many adult-born cells degenerate over

kozorovitskiy and gould: adult neurogenesis in the hippocampus

55

BrdU+ cells in the DG

7500

Con Dom Sub

2004). In contrast, training on tasks that do not require the hippocampus does not affect the number of new neurons. It is unclear what specific aspects of hippocampal-dependent tasks are necessary and sufficient for the enhanced survival of adult-generated cells. Electrophysiological studies have shown that both hippocampal-dependent and hippocampalindependent tasks activate hippocampal neurons (Weisz, Clark, and Thompson, 1984). One possible difference between the two sets of tasks is their difficulty. Several traditionally used learning tasks that require the hippocampus are more difficult to acquire than those that do not rely on this brain region. Although attempts have been made to determine whether task difficulty and hippocampal dependence are critical for the influence on adult-generated cells (Leuner et al., in press), these issues remain unresolved. In addition to studies finding enhanced adult neurogenesis with certain types of learning, other reports have either failed to find such increases or instead have found decreases in the numbers of new neurons (see Leuner, Gould, and Shors, 2006, for review). Some of these discrepancies can be attributed to methodological variations, such as differences in schedules of BrdU injection and survival times. A recent study has verified that the learning-induced enhancement of neurogenesis in the dentate gyrus alters only those cells that are produced within a specific time period prior to the learning (Epp et al., 2006). Thus studies that have utilized different paradigms, BrdU doses, or time points could easily miss such an effect. It is additionally possible that other types of learning tasks, particularly those that are stressful, might diminish the number of new neurons in the dentate gyrus, obscuring the enhancing effect of learning on this brain measure. Clearly, the effects of learning on adult neurogen-

*

5000

*

2500

0 Cage

Enriched

Unenriched

Figure 4.7 Status in the social dominance hierarchy influences adult neurogenesis. Dominant rats (Dom) had more BrdU-labeled cells in the dentate gyrus compared with subordinate animals (Sub) and cage controls (Con), 2 weeks after a single BrdU injection. This difference was maintained in animals whether or not they had access to an enriched environment during the survival time after BrdU injection. Error bars represent SEM, p < 0.05 relative to controls.

time (Cameron and McKay, 2001; Dayer et al., 2005). Training on hippocampal-dependent tasks during the time period when many newly generated granule neurons die significantly enhances the rate of their survival, although not their production, in adult rats. This increase in survival persists for months (figure 4.8), long after the time that task performance becomes independent of the hippocampus (Leuner et al., 6000

6000

*

5000

4000 3000 2000 1000

4000 3000 2000 1000

0 Trace paired

* 3000 2000

0 Trace paired

Unpaired

1 day

Figure 4.8 Trace memory formation persistently enhances the survival of newly born cells in the dentate gyrus. Total numbers of BrdU-labeled cells in the dentate gyrus of animals 1, 30, or 60 days after trace conditioning or unpaired training. Trace conditioning

56

4000

1000

0 Unpaired

5000

* BrdU-labeled cells

BrdU-labeled cells

BrdU-labeled cells

5000

6000

30 days

Unpaired

Trace paired

60 days

increased the number of BrdU-labeled cells when compared with exposure to unpaired stimuli at all survival times. Error bars represent SEM, p < 0.05 relative to controls.

fundamentals of developmental neurobiology

esis in the hippocampus are complex and likely dependent on the age of the neurons and the particular learning task examined. Although the exact manner in which learning alters adult neurogenesis remains unresolved, the new neurons are likely altered by learning. Indeed, recent studies suggesting that new neurons are activated by exploration and learning experiences (Ramirez-Amaya et al., 2006; Snyder and Cameron, 2006), confirm this hypothesis.

Functional significance of adult neurogenesis The functional significance of neurons generated in the dentate gyrus of adult animals remains unknown. Given their incorporation into the hippocampus, which is important for learning and memory, and the effect of learning on the number and activation of these new cells, it seems reasonable to consider a potential role for these new neurons in learning. Altman was the first to propose that neurons generated during the postnatal developmental period might be important for forming associations (Altman and Das, 1965, 1967; Altman et al., 1973), suggesting that the granule “microneurons” are morphologically well suited for the development of learning processes. Subsequently, Nottebohm suggested that new hippocampal neurons may be a cellular substrate for learning in the adult (Nottebohm, 1989). Electrophysiological studies of adult-born fluorescently labeled neurons indicate that the new cells eventually develop passive membrane properties, action potentials, and synaptic inputs similar to the surrounding granule neurons (van Praag et al., 2002; Zhao et al., 2006), although in their response to the neurotransmitter GABA they seem to resemble immature granule cells generated during early development (Overstreet et al., 2005). It has been suggested that new neurons positively contribute to the amount of synaptic plasticity shown by the hippocampus in adult animals (Snyder, Kee, and Wojtowicz, 2001), which may imply a functional utility for adult neurogenesis. A continually rejuvenating population of new neurons seems well suited for the proposed transient role of the hippocampus in information storage (Squire, 1992; Squire and Zola, 1998). Aside from the data showing that learning alters new neuron number and activation, additional lines of evidence suggest that new neurons might be involved in learning. First, several studies report a positive correlation between the number of new neurons and learning performance (see Leuner, Gould, and Shors, 2006, for review, as well as for consideration of the studies that do not reach similar conclusions). Second, multiple studies attempting to interfere with adult neurogenesis have found impairments in certain types of learning and memory tasks.

Parallel Changes in Adult Neurogenesis and Learning Since new granule cells require time to differentiate and become integrated into circuitry, changes in cell proliferation are not likely to result in immediate functional consequences, although adult-generated cells may be capable of exerting an impact on hippocampal function prior to complete maturation. In addition, it is important to note that acute changes in cell proliferation and neuronal survival may not be of sufficient magnitude to produce an observable functional impact. As a result, conditions under which cell proliferation and survival are chronically enhanced or diminished are of particular interest, because they are most likely to elucidate the functional consequences of changes in adult neuron production and survival. In general, chronic increases in the factors that negatively regulate adult granule neuron production are associated with poor performance on hippocampal-dependent tasks, whereas chronic increases in the factors that positively regulate adult granule neuron production are associated with enhanced cognitive function. Several studies have demonstrated that chronic stress results in an impairment of hippocampaldependent learning (Luine et al., 1994, Bodnoff et al., 1995; Luine et al., 1996; Krugers et al., 1997). These stress-induced impairments are not permanent; the performance of animals tested on a spatial task long after the termination of stress is similar to unstressed controls (Luine et al., 1994, 1996). This observation is consistent with a possible role for adultgenerated cells in hippocampal function: the deficit may only last as long as neurogenesis is impaired. It should be noted that previous work has reported that brief stress enhances hippocampal-dependent learning (Shors, Weiss, and Thompson, 1992). However, these behavioral changes were observed shortly after stress and may likely involve other cellular mechanisms, such as changes in synaptic plasticity. Several studies have demonstrated that estrogen has a positive effect on the acquisition of hippocampal-dependent tasks. Although estrogen treatment has generally been found to enhance learning, some studies report decreases in performance during times of high levels of circulating estrogen. As in the case of acute stress studies, the time frame examined in many studies is likely to be too early to involve changes in neuron production. Chronic treatment with estrogen for more than four days results in enhanced performance on a hippocampal-dependent task (Luine et al., 1998). This finding is consistent with the observation that adult-generated cells do not extend axons prior to four days after their production (Hastings and Gould, 1999). It is likely that sufficient time for adultgenerated cells to extend axons is required for estrogen-induced increases in cell production to have a functional effect. Consistent with studies that have demonstrated a positive effect of estrogen on the performance of hippocampaldependent tasks, a sex difference, favoring females, in spatial

kozorovitskiy and gould: adult neurogenesis in the hippocampus

57

navigation learning in rats has been reported (Perrot-Sinal et al., 1996). Previous studies of sex differences in spatial navigation learning in rodents on this task have yielded conflicting data (Bucci, Chiba, and Gallagher, 1995; Galea, Kavaliers, and Ossenkopp, 1996). However, the results of the former study demonstrate that although male rats initially perform better than female rats on the task, females learn better than males once the animals have been acclimated to the testing apparatus. Thus a sex difference in reaction to the novelty of this task prevents females from performing well. After this performance confound is removed, females learn certain aspects of this task better than males. Similarly, another study demonstrated a sex difference favoring females in hippocampal-dependent learning as well. This study reported that females learn trace eyeblink conditioning faster than males (Wood and Shors, 1998). Taken together with the observation that females produce more new granule neurons than males (Tanapat et al., 1999), these data indicate an additional positive relationship between new granule neurons and certain types of learning. Studies have also demonstrated a positive correlation between the number of new granule neurons and performance on hippocampal-dependent learning tasks following enriched environment living (Kempermann, Kuhn, and Gage, 1997, 1998). However, it should be noted that many of the factors and conditions known to alter the number of new neurons, either by affecting the proliferation of precursor cells or by altering the survival of new neurons, also affect other measures in the hippocampus—for example, synaptogenesis, dendritic architecture, and dendritic spines. Finally, the hippocampus is not the sole brain region affected by experiential and hormonal treatments that affect adult neurogenesis. For example, the amygdala and prefrontal cortex have been identified as brain regions sensitive to experience and hormones. Thus any behavioral changes observed after experiences that impact adult neurogenesis cannot be attributed solely to the alterations in the number of new neurons. Blockade of Neurogenesis and Learning Several studies have attempted to determine the function of adult neurogenesis in the hippocampus by inhibiting the process and examining the behavioral consequences. Collectively, these experiments suggest that adult-generated neurons are important for certain types of learning and memory. However, the specific conclusions of these studies differ. Some have shown that decreasing adult neurogenesis impairs trace eyeblink conditioning, trace fear conditioning, context fear conditioning, spatial navigation memory, and object recognition, while others have observed no effect on context fear conditioning and spatial navigation learning (see Leuner, Gould, and Shors, 2006, for review). One reason for these discrepancies involves the different methods used to

58

reduce adult neurogenesis, as well as the different paradigms employed to assess learning. Some of the techniques that have been used to study the connection between learning and adult neurogenesis are treatment with the antimitotic drug MAM, global or focal irradiation, and transgenic models with impairments in cell proliferation induced in adulthood (Shors et al., 2002; Achanta, Fuss, and Martinez, 2006; Fuller et al., 2006). Methods for decreasing adult neurogenesis work on different timescales and vary in their effectiveness. The possibility that they affect other aspects of brain function also remains an important potential confound in interpretation of these studies. In addition to assessing whether learning and memory are affected by ablation of adult neurogenesis, studies have examined another class of behaviors involving the hippocampus related to anxiety regulation. Santarelli and colleagues (2003) have shown that focal irradiation of the hippocampus inhibits adult neurogenesis and prevents the anxiolytic action of antidepressants in an animal model of chronic stress. Although these findings are also open to the same criticisms made for similar studies of learning and memory, they suggest the possibility that structural plasticity in the hippocampus is also involved in the therapeutic response of antidepressants, at least with regard to anxiety. Thus the definitive answer to the question of whether adult neurogenesis is necessary for hippocampal function in learning and/or anxiety regulation awaits further investigation, including the development of new methodologies for selectively depleting the hippocampus of new neurons without causing unrelated changes in other cell populations.

Conclusions It is evident that new neurons are generated in the hippocampus of adult mammals, including humans. The number of new granule neurons produced in adulthood and the variety of species in which this process has been reported suggest that new neurons are important for the function of this brain region. Studies carried out over the past several years have demonstrated that the production of new granule cells in the hippocampus can be modulated by hormones. The production of new neurons can be inhibited by adrenal steroids and stimulated by ovarian steroids, both of which act by altering cell proliferation. Furthermore, experiences, such as stress and learning, control the production of new neurons, by affecting either cell proliferation or cell survival. Collectively, these observations suggest that newly generated granule cells may provide an important cellular substrate by which hormones and experience alter hippocampal function. Cellular phenomena such as neurogenesis, axon extension, dendritic development, synaptogenesis, and cell death are traditionally viewed as developmental processes. The

fundamentals of developmental neurobiology

continual addition and replacement of new neurons that integrate into hippocampal circuitry during adulthood indicates that these developmental processes continue throughout life in certain brain regions. The extent to which adult neurogenesis and other forms of structural plasticity contribute to brain function under normal and pathological conditions remains to be elucidated.

REFERENCES Achanta, P., M. Fuss, and J. L. Martinez, Jr., 2006. Ionizing radiation given three months prior to training impairs the formation of trace fear memories. Soci. Neurosci. Abstr. 667:23. Altman, J., 1969. Autoradiographic and histological studies of postnatal neurogenesis. IV. Cell proliferation and migration in the anterior forebrain, with special reference to persisting neurogenesis in the olfactory bulb. J. Comp. Neurol. 137: 433–457. Altman, J., and S. A. Bayer, 1990. Migration and distribution of two populations of hippocampal granule cell precursors during the perinatal and postnatal periods. J. Comp. Neurol. 301: 365–381. Altman, J., R. L. Brunner, and S. A. Bayer, 1973. The hippocampus and behavioral maturation. Behav. Biol. 8:557–596. Altman, J., and G. D. Das, 1965. Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol. 124:319–335. Altman, J., and G. D. Das, 1966. Autoradiographic and histological studies of postnatal neurogenesis. I. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions. J. Comp. Neurol. 126:337–389. Altman, J., and G. D. Das, 1967. Postnatal neurogenesis in the guinea-pig. Nature 214:1098–1101. Ambrogini, P., R. Cuppini, P. Ferri, C. Mancini, S. Ciaroni, A. Voci, E. Gerdoni, and G. Gallo, 2005. Thyroid hormones affect neurogenesis in the dentate gyrus of adult rat. Neuroendocrinology 81:244–253. Barnea, A., and F. Nottebohm, 1994. Seasonal recruitment of hippocampal neurons in adult free-ranging black-capped chickadees. Proc. Natl. Acad. Sci. USA 91:11217–11221. Bedard, A., C. Gravel, and A. Parent, 2006. Chemical characterization of newly generated neurons in the striatum of adult primates. Exp. Brain Res. 170:501–512. Bernier, P. J., A. Bedard, J. Vinet, M. Levesque, and A. Parent, 2002. Newly generated neurons in the amygdala and adjoining cortex of adult primates. Proc. Natl. Acad. Sci. USA 99: 11464–11469. Bodnoff, S. R., A. G. Humphreys, J. C. Lehman, D. M. Diamond, G. M. Rose, and M. J. Meaney, 1995. Enduring effects of chronic corticosterone treatment on spatial learning, synaptic plasticity, and hippocampal neuropathology in young and midaged rats. J. Neurosci. 15:61–69. Bucci, D. J., A. A. Chiba, and M. Gallagher, 1995. Spatial learning in male and female Long-Evans rats. Behav. Neurosci. 109:180–183. Cameron, H. A., and E. Gould, 1994. Adult neurogenesis is regulated by adrenal steroids in the dentate gyrus. Neuroscience 61:203–209.

Cameron, H. A., and R. D. Mckay, 1999. Restoring production of hippocampal neurons in old age. Nature Neurosci. 2:894–897. Cameron, H. A., and R. D. Mckay, 2001. Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. J. Comp. Neurol. 435:406–417. Cameron, H. A., C. S. Woolley, B. S. McEwen, and E. Gould, 1993. Differentiation of newly born neurons and glia in the dentate gyrus of the adult rat. Neuroscience 56:337–344. Dayer, A. G., K. M. Cleaver, T. Abouantoun, and H. A. Cameron, 2005. New GABAergic interneurons in the adult neocortex and striatum are generated from different precursors. J. Cell Biol. 168:415–427. Desouza, L. A., U. Ladiwala, S. M. Daniel, S. Agashe, R. A. Vaidya, and V. A. Vaidya, 2005. Thyroid hormone regulates hippocampal neurogenesis in the adult rat brain. Mol. Cell Neurosci. 29:414–426. Droste, S. K., A. Gesing, S. Ulbricht, M. B. Muller, A. C. Linthorst, and J. M. Reul, 2003. Effects of long-term voluntary exercise on the mouse hypothalamic-pituitary-adrenocortical axis. Endocrinology 144:3012–3023. Eckenhoff, M. F., and P. Rakic, 1988. Nature and fate of proliferative cells in the hippocampal dentate gyrus during the life span of the rhesus monkey. J. Neurosci. 8:2729–2747. Epp, J. R., M. D. Spritzer, and L. A. Galea, 2006. Hippocampusdependent learning promotes survival of new cells in the dentate gyrus during the axon-extension period of cell maturation. Soci. for Neurosci. Abstr. 266:8. Eriksson, P. S., E. Perfilieva, T. Bjork-Eriksson, A. M. Alborn, C. Nordborg, D. A. Peterson, and F. H. Gage, 1998. Neurogenesis in the adult human hippocampus. Nature Med. 4: 1313–1317. Fowler, C. D., F. Johnson, and F. Wang, 2005. Estrogen regulation of cell proliferation and distribution of estrogen receptoralpha in the brains of adult female prairie and meadow voles. J. Comp. Neurol. 489:166–179. Fowler, C. D., Y. Liu, C. Ouimet, and Z. Wang, 2002. The effects of social environment on adult neurogenesis in the female prairie vole. J. Neurobiol. 51:115–128. Fuchs, E, G. Flugge, B. S. McEwen, P. Tanapat, and E. Gould, 1997. Chronic subordination stress inhibits neurogenesis and decreases the volume of the granule cell layer. Society for Neuroscience 23:317. Fuller, C. L., C. A. Rabinak, R. J. Lichtenwalner, A. J. Velander, C. F. Burant, S. Maren, and M. J. Parent, 2006. Targeted genetic ablation of neural progenitors in the adult mouse hippocampus. Society for Neuroscience 464:11. Galea, L. A., M. Kavaliers, and K. P. Ossenkopp, 1996. Sexually dimorphic spatial learning in meadow voles Microtus pennsylvanicus and deer mice Peromyscus maniculatus. J. Exp. Biol. 199:195–200. Galea, L. A., P. Tanapat, and E. Gould, 2006. Exposure to predator odor suppresses cell proliferation in the dentate gyrus of adult rats via a cholinergic mechanism. Soc. Neurosci. Abstr. 22. Gould, E., H. A. Cameron, D. C. Daniels, C. S. Woolley, and B. S. McEwen, 1992. Adrenal hormones suppress cell division in the adult rat dentate gyrus. J. Neurosci. 12:3642–3650. Gould, E., B. S. McEwen, P. Tanapat, L. A. M. Galea, and E. Fuchs, 1997. Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J. Neurosci. 17:2492–2498. Gould, E., A. J. Reeves, M. Fallah, P. Tanapat, C. C. Gross, and E. Fuchs, 1999. Hippocampal neurogenesis in adult Old World primates. Proc. Natl. Acad. Sci. USA 96:5263–5267.

kozorovitskiy and gould: adult neurogenesis in the hippocampus

59

Gould, E., P. Tanapat, N. B. Hastings, and T. J. Shors, 1999. Neurogenesis in adulthood: A possible role in learning. Trends Cogn. Sci. 3:186–191. Gould, E., P. Tanapat, B. S. McEwen, G. Flugge, and E. Fuchs, 1998. Proliferation of granule cell precursors in the dentate gyrus of adult monkeys is diminished by stress. Proc. Natl. Acad. Sci. USA 95:3168–3171. Gould, E., C. S. Woolley, H. A. Cameron, D. C. Daniels, and B. S. McEwen, 1991. Adrenal steroids regulate postnatal development of the rat dentate gyrus. II. Effects of glucocorticoids and mineralocorticoids on cell birth. J. Comp. Neurol. 313: 486–493. Gould, E., N. Vail, M. Wagers, and C. G. Gross, 2001. Adultgenerated hippocampal and neocortical neurons in macaques have a transient existence. Proc. Natl. Acad. Sci. USA 98: 10910–10917. Guzman-Marin, R., N. Suntsova, D. R. Stewart, H. Gong, R. Szymusiak, and D. McGinty, 2003. Sleep deprivation reduces proliferation of cells in the dentate gyrus of the hippocampus in rats. J. Physiol. 549:563–571. Hastings, N. B., and E. Gould, 1999. Rapid extension of axons into the CA3 region by adult-generated granule cells. J. Comp. Neurol. 413:146–154. Huang, L. G. J. DeVries, and E. L. Bittman, 1998. Photoperiod regulates neuronal bromodeoxyuridine labeling in the brain of a seasonally breeding mammal. J. Neurobiol. 36:410–420. Kaplan, M. S., 1981. Neurogenesis in the 3-month-old rat visual cortex. J. Comp. Neurol. 195:323–338. Kaplan, M. S., and J. W. Hinds, 1977. Neurogenesis in the adult rat: Electron microscopic analysis of light radioautographs. Science 197:1092–1094. Kempermann, G., H. G. Kuhn, and F. H. Gage, 1997. More hippocampal neurons in adult mice living in an enriched environment. Nature 386:493–495. Kempermann, G., H. G. Kuhn, and F. H. Gage, 1998. Experienceinduced neurogenesis in the senescent dentate gyrus. J. Neurosci. 18:3206–3212. Kornack, D. R., and P. Rakic, 1999. Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Proc. Natl. Acad. Sci. USA 96:5768–5773. Kozorovitskiy, Y., and E. Gould, 2004. Dominance hierarchy influences adult neurogenesis in the dentate gyrus. J. Neurosci. 24:6755–6759. Krugers, H. J., B. R. Douma, G. Andringa, B. Bohus, J. Korf, and P. G. Luiten, 1997. Exposure to chronic psychosocial stress and corticosterone in the rat: Effects on spatial discrimination learning and hippocampal protein kinase C gamma immunoreactivity. Hippocampus 7:427–436. Kuhn, H. G., H. Dickinson-Anson, and F. H. Gage, 1996. Neurogenesis in the dentate gyrus of the adult rat: Age-related decrease of neuronal progenitor proliferation. J. Neurosci. 16:2027–2033. Leuner, B., E. Gould, and T. J. Shors, 2006. Is there a link between adult neurogenesis and learning? Hippocampus 16:216 –224. Leuner, B., S. Mendolia-Loffredo, Y. Kozorovitskiy, D. Samburg, E. Gould, and T. Shors, 2004. Learning enhances the survival of new neurons beyond the time when the hippocampus is required for memory. J. Neurosci. 24:7477–7481. Leuner, B., J. Waddell, E. Gould, and T. J. Shors, in press. Temporal discontiguity is neither necessary nor sufficient for learning-induced effects on adult neurogenesis. J. Neurosci. 26: 13437–13442.

60

Luine, V., C. Martinez, M. Villegas, A. M. Magarinos, and B. S. McEwen, 1996. Restraint stress reversibly enhances spatial memory performance. Physiol. Behav. 59:27–32. Luine, V., M. Villegas, C. Martinez, and B. S. McEwen, 1994. Repeated stress causes reversible impairments of spatial memory performance. Brain Res. 639:167–170. Luine, V. N., S. T. Richards, V. Y. Wu, and K. D. Beck, 1998. Estradiol enhances learning and memory in a spatial memory task and effects levels of monoaminergic neurotransmitters. Horm. Behav. 34:149–162. Luzzati, F., S. De Marchis, A. Fasolo, and P. Peretto, 2006. Neurogenesis in the caudate nucleus of the adult rabbit. J. Neurosci. 26:609–621. Mirescu, C., and E. Gould, 2006. Stress and adult neurogenesis. Hippocampus 16:233–238. Mirescu, C., J. D. Peters, and E. Gould, 2004. Early life experience alters response of adult neurogenesis to stress. Nature Neurosci. 7:841–846. Mirescu, C., J. D. Peters, L. Noiman, and E. Gould, 2006. Sleep deprivation inhibits adult neurogenesis in the hippocampus by elevating glucocorticoids. Proc. Natl. Acad. Sci. USA 103:19170– 19175. Nottebohm, F., 1989. Hormonal regulation of synapses and cell number in the adult canary brain and its relevance to theories of long-term memory storage. In Neural Control of Reproductive Function, ed. J. Lakoski, J. R. Perez-Polo, and D. Rossin, 583– 601. New York: Liss. Nowakowski, R. S., and P. Rakic, 1981. The site of origin and route and rate of migration of neurons to the hippocampal region of the rhesus monkey. J. Comp. Neurol. 196:129–154. Okano, H. J., D. W. Pfaff and R. B. Gibbs, 1993. RB and Cdc2 expression in brain: Correlations with 3H-thymidine incorporation and neurogenesis. J. Neurosci. 13:2930–2938. Olariu, A., K. M. Cleaver, L. E. Shore, M. D. Brewer, and H. A. Cameron, 2005. A natural form of learning can increase and decrease the survival of new neurons in the dentate gyrus. Hippocampus 15:750–762. Oppenheim, R. W., I. W. Chu-Wang, and J. L. Maderdrut, 1978. Cell death of motoneurons in the chick embryo spinal cord. III. The differentiation of motoneurons prior to their induced degeneration following limb-bud removal. J. Comp. Neurol. 177: 87–111. Ormerod, B. K., T. T. Lee, and L. A. Galea, 2003. Estradiol initially enhances but subsequently suppresses (via adrenal steroids) granule cell proliferation in the dentate gyrus of adult female rats. J. Neurobiol. 55:247–260. Overstreet, W. L., D. A. Bromberg, A. L. Bensen, and G. L. Westbrook, 2005. GABAergic signaling to newborn neurons in dentate gyrus. J. Neurophysiol. 94:4528–4532. Perez-Canellas, M. M., and J. M. Garcia-Verdugo, 1996. Adult neurogenesis in the telencephalon of a lizard: A [3H]thymidine autoradiographic and bromodeoxyuridine immunocytochemical study. Brain Res. Dev. Brain Res. 93:49–61. Perrot-Sinal, T. S., M. A. Kostenuik, K. P. Ossenkopp, and M. Kavaliers, 1996. Sex differences in performance in the Morris water maze and the effects of initial nonstationary hidden platform training. Behav. Neurosci. 110:1309–1320. Polenov, A. L., and V. K. Chetverukhin, 1993. Ultrastructural radioautographic analysis of neurogenesis in the hypothalamus of the adult frog, Rana temporaria, with special reference to physiological regeneration of the preoptic nucleus. II. Types of neuronal cells produced. Cell Tissue Res. 271:351– 362.

fundamentals of developmental neurobiology

Rakic, P., 1985. Limits of neurogenesis in primates. Science 227:1054–1056. Ramirez-Amaya, V., D. F. Marrone, F. H. Gage, and C. A. Barnes, 2006. Integration of new neurons into functional neural networks. J. Neurosci. 26:12237–12241. Roman, V., K. Van der Borght, S. A. Leemburg, E. A., Van der Zee, and P. Meerlo, 2005. Sleep restriction by forced activity reduces hippocampal cell proliferation. Brain Res. 1065: 53–59. Runyan, C. A., C. S. Weickert, and R. C. Saunders, 2006. Adult neurogenesis and immediate early gene response to working memory stimulation in the primate prefrontal cortex. Society for Neuroscience 318.10. Santarelli, L., M. Saxe, C. Gross, A. Surget, F. Battaglia, S. Dulawa, N. Weisstaub, J. Lee, R. Duman, O. Arancio, C. Belzung, and R. Hen, 2003. Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science 301:805–809. Sapolsky, R. M., 1992. Do glucocorticoid concentrations rise with age in the rat? Neurobiol. Aging 13:171–174. Sapolsky, R. M., and J. Altmann, 1991. Incidence of hypercortisolism and dexamethasone resistance increases with age among wild baboons. Biol. Psychiatry 30:1008–1016. Sapolsky, R. M., L. C. Krey, and B. S. McEwen, 1985. Prolonged glucocorticoid exposure reduces hippocampal neuron number: Implications for aging. J. Neurosci. 5:1222–1227. Sapolsky, R. M., and M. J. Meaney, 1986. Maturation of the adrenocortical stress response: Neuroendocrine control mechanisms and the stress hyporesponsive period. Brain Res. 396: 64–76. Schlessinger, A. R., W. M. Cowan, and D. I. Gottlieb, 1975. An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. Comp. Neurol. 159:149–175. Shors, T., D. Townsend, M. Zhao, Y. Kozorovitskiy, and E. Gould, 2002. Neurogenesis may relate to some but not all types of hippocampal-dependent learning. Hippocampus 12:578–584. Shors, T. J., C. Weiss, and R. F. Thompson, 1992. Stress-induced facilitation of classical conditioning. Science 257:537–539. Snyder, J., and H. A. Cameron, 2006. Immediate early gene activation of adult-born neurons by the water maze. Society for Neuroscience 318:12. Snyder, J., N. Kee, and J. M. Wojtowicz, 2001. Effects of adult neurogenesis on synaptic plasticity in the rat dentate gyrus. J. Neurophysiol. 85:2423–2431. Squire, L. R., 1992. Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychol. Rev. 99:195–231. Squire, L. R., and S. M. Zola, 1998. Episodic memory, semantic memory, and amnesia. Hippocampus 8:205–211. Stanfield, B. B., and J. E. Trice, 1988. Evidence that granule cells generated in the dentate gyrus of adult rats extend axonal projections. Exp. Brain Res. 72:399–406.

Stranahan, A. M., D. Khalil, and E. Gould, 2006. Social isolation delays the positive effects of running on adult neurogenesis. Nature Neurosci. 9:526–533. Tanapat, P., L. A. Galea, and E. Gould, 1998. Stress inhibits the proliferation of granule cell precursors in the developing dentate gyrus. Int. J. Dev. Neurosci. 16:235–239. Tanapat, P., N. B. Hastings, and E. Gould, 2005. Ovarian steroids influence cell proliferation in the dentate gyrus of the adult female rat in a dose- and time-dependent manner. J. Comp. Neurol. 481:252–265. Tanapat, P., N. B. Hastings, A. J. Reeves, and E. Gould, 1999. Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci. 19:5792–5801. van der Borght, K., F. Ferrari, K. Klauke, V. Roman, R. Havekes, A. Sgoifo, E. A. van der Zee, and P. Meerlo, 2006. Hippocampal cell proliferation across the day: Increase by running wheel activity, but no effect of sleep and wakefulness. Behav. Brain Res. 167:36–41. van Praag, H., G. Kempermann, and F. H. Gage, 1999. Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nature Neurosci. 2:266–270. van Praag, H., A. F. Schinder, B. R. Christie, N. Toni, T. D. Palmer, and F. H. Gage, 2002. Functional neurogenesis in the adult hippocampus. Nature 415:1030–1034. Weisz, D. J., G. A. Clark, and R. F. Thompson, 1984. Increased responsivity of dentate granule cells during nictitating membrane response conditioning in rabbit. Behav. Brain Res. 12: 145–154. Wood, G. E., and T. J. Shors, 1998. Stress facilitates classical conditioning in males, but impairs classical conditioning in females through activational effects of ovarian hormones. Proc. Natl. Acad. Sci. USA 95:4066–4071. Woolley, C. S., E. Gould, M. Frankfurt, and B. S. McEwen, 1990. Naturally occurring fluctuation in dendritic spine density on adult hippocampal pyramidal neurons. J. Neurosci. 10: 4035–4039. Xu, Y., N. Tamamaki, T. Noda, K. Kimura, Y. Itokazu, N. Matsumoto, M. Dezawa, and C. Ide, 2005. Neurogenesis in the ependymal layer of the adult rat 3rd ventricle. Exp. Neurol. 192:251–264. Zhao, M., S. Momma, K. Delfani, M. Carlen, R. M. Cassidy, C. B. Johansson, H. Brismar, O. Shupliakov, J. Frisen, and A. M. Janson, 2003. Evidence for neurogenesis in the adult mammalian substantia nigra. Proc. Natl. Acad. Sci. USA 100: 7925–7930. Zhao, C., E. M. Teng, R. G. Summers, Jr., G. L. Ming, and F. H. Gage, 2006. Distinct morphological stages of dentate granule neuron maturation in the adult mouse hippocampus. J. Neurosci. 26:3–11. Zupanc, G. K., 1999. Neurogenesis, cell death and regeneration in the adult gymnotiform brain. J. Exp. Biol. 202 (Part 10): 1435–1446.

kozorovitskiy and gould: adult neurogenesis in the hippocampus

61

5

The LHPA System and Neurobehavioral Development AMANDA R. TARULLO, KARINA QUEVEDO, AND MEGAN R. GUNNAR

It has long been known that stress has significant impacts on the developing brain (Levine, Alpert, and Lewis, 1957; for review see Levine, 2005). Stressors in the form of abuse, neglect, parental loss, and poverty increase the risk of emotional disorders and poor academic outcomes (for review see Cicchetti and Cohen, 2006). As we will describe, research on animals and adults supports the hypothesis that activity of the limbic-hypothalamic-pituitary-adrenocortical (LHPA) system plays a role in mediating the impact of adverse experiences on the developing brain. Based on this research, studies of children are increasingly employing measures of cortisol, the major hormonal product of the LHPA system in humans, both to understand individual differences in stress vulnerability and resilience and to identify pathways through which adverse experiences affect brain development. Despite the ease of measuring cortisol, which can now be done using noninvasive sampling of saliva, interpretation of cortisol-behavior findings is often fraught with ambiguity. For instance, sometimes both positive and negative associations are obtained for the same target behaviors. In this chapter, we describe the anatomy and physiology of the LHPA system, mechanisms through which this system may impact brain development, developmental changes in animals and children in the reactivity of this system to stressors, and the critical role that psychosocial processes play in regulating its activity during infancy and childhood. Finally, we provide a very brief introduction to work on genes that may be important in stress vulnerability and resilience. While the information in this chapter will not reduce the complexity of interpreting cortisol-brainbehavior relations in studies of human development, it should help researchers new to this area understand why these relations are complex and often dependent on context and age.

LHPA anatomy and physiology The LHPA system is one of the two primary systems regulating mammalian stress responses (Stratakis and Chrousos, 1995; see table 5.1). The other mammalian stress system, the sympathetic adrenomedullary (SAM) system, will not be covered extensively in this chapter because of space limita-

tions. The LHPA stress response pathway begins with the secretion of corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP) from the medial parvocellular region of the paraventricular nucleus (mpPVN) of the hypothalamus, and culminates in increased adrenocortical production of stress hormones called glucocorticoids (cortisol in humans and primates, corticosterone in rodents). As depicted in figure 5.1, the initiation of this cascade is modulated by an intricate network of both excitatory and inhibitory limbic inputs to the mpPVN (Gunnar and Davis, 2003; Herman and Cullinan, 1997). The excitatory inputs include the central amygdala and lateral bed nucleus of the stria terminalis (BNST), while the inhibitory inputs include the medial BNST, prefrontal cortex, preoptic area, and ventral subiculum (Herman et al., 2002, 2004; Shekhar et al., 2005). For systemic stressors—that is, context-independent physiological stressors to the body such as hypoxia or infection—the mpPVN is activated by means of signals relayed from the brain stem, often through limbic circuits, and the response would occur even if the animal were unconscious. Processive stressors, in contrast, such as being chased by a predator or separated from a caregiver, require interpretation (processing) by the animal: assessing environmental threat and emotional significance by comparing the current situation to past experience (Gunnar and Vazquez, 2006; Herman and Cullinan, 1997). Such psychological processing, occurring at the level of the limbic system and prefrontal cortex, influences the signals sent to the mpPVN. Once released by the mpPVN, CRH and AVP travel to the pituitary, where CRH binds to the anterior pituitary CRH receptor 1 (CRH r1), stimulating synthesis of the proopiomelanocortin (POMC) molecule. In turn, adrenocorticotropic hormone (ACTH) is derived from the POMC molecule and released in pulses from the anterior pituitary. AVP potentiates this process. ACTH then enters the bloodstream and binds to receptors on the cortex of the adrenal gland, stimulating the synthesis and release of cortisol. This cascade is a rather slow process, such that peak cortisol levels in the bloodstream occur approximately 20–30 minutes after activation of the mpPVN. Cortisol binds to intracellular receptors throughout the brain and periphery, influencing gene transcription and thus having long-lasting effects on

63

64

fundamentals of developmental neurobiology

Cortex of the adrenal gland (adrenal cortex)

Anterior pituitary (AP)

Hippocampus

Amygdala

Regions Associated with LHPA Stress Response Hypothalamus (mp PVN) Neurobiological Effects Stimulates the release of ACTH by modulating the actions of CRH 1. Stimulates production of propiomelanocortin (POMC) in the anterior pituitary 2. Stimulates the release of norepinephrine via the locus coeruleus 3. Enhances neuronal excitability 4. Mediates mechanisms of synaptic plasticity involved in memory formation and learning

Some impacts via retrograde passage from pituitary to CNS; see work by de Wied (e.g., de Wied and Jolles, 1982) Glucocorticoids: Basal levels cortisol in humans 1. Regulation of and other primates, metabolism and corticosterone in energy utilization rodents 2. Enhances synaptic plasticity underlying learning Chronic stress levels 1. Inhibits growth and reproduction 2. Contains immune and inflammatory responses 3. Dendritic atrophy and apoptosis

1. Adenocorticotropic hormone (ACTH) (derived from POMC)

Corticotropinreleasing hormone (CRH)

Secreted Agent Arginine Vasopressin (AVP)—just in PVN

Sense of increased vigor followed, if elevations prolonged, by increased negative emotionality

1. Augments arousal, alertness and readiness to respond 2. Increases startle 3. Elevations in amygdala trigger fear and anxiety 4. Increase is related to avoidance and decreased social behavior

Stress-Related Behavioral Symptoms

Acute: increases release Chronic: increases or decreases release

Acute or chronic: increases or sensitizes its release

Acute or chronic: increases its release

Effects of Stress

Mineralocorticoid (MR) Higher affinity for cortisol, binds at basal levels in the central nervous system Outside the brain, MR binds aldosterone Glucocorticoid (GR) Lower affinity for cortisol, binds at high concentrations

Maintains electrical activity in neurons, blood pressure, HPA rhythm Facilitates cerebral glucose availability, synaptic plasticity, and SAM system’s immediate stress response

Induces the termination of the HPA activation Reduces cerebral glucose availability, thus increasing risk of neuronal death Impairs synaptic plasticity and memory formation

Hippocampus Frontal cortex (in primates)

Pituitary Hypothalamus Hippocampus Amygdala Medial frontal cortex and other limbic regions

Stimulates the synthesis and release of glucocorticoids

Cortex of the adrenal gland ACTH-R

1. Evokes anxiolitic and antidepressive responses 2. May promote recovery from stress and adaptation 3. Increases vasodilatation and diminishes blood pressure

Subcortical areas Lateral septal nuclei Choroid plexus Olfactory bulb Amygdala Hippocampus

CRH-2R: lower affinity for CRH. Binds preferentially with urocortins (II and III)

1. Mediates changes in ACTH release 2. Mediates the fast fight-flight response 3. Mediates defensive responses to processive stressors 4. Anxiogenic and depressogenic effects

Effects of Eeceptor Activations Increases sensitivity of AP to CRH

Neocortical areas Cerebellum Hippocampus Pituitary Hypothalamus Amygdala

Predominant Expression of Relevant Receptors Pituitary

CRH-1R: higher affinity for CRH

Receptors of Interest V1b

Table 5.1 Regions involved in the LHPA system: their hormonal products, neurobiological and behavioral effects, response to stress, and associated receptors

NE Locus coeruleus

Anterior cingulate

Spinal Cord

Brain stem Amygdala

Hippocampus

ACh NE

Hypothalamus Adrenal Orbital PFC

BNST CRH

POA

Anterior pituitary

ACTH

Medulla

EPI

Cortex

GC

Ventral subiculum Cortico-Limbic Level

Hypothalamic–Brain Stem Level

Neural-to-Adrenal Level

Figure 5.1 Schematic of activation pathways for the LHPA system. For processive stressors, activation of the LHPA cascade at the level of the medial parvocellular region of the paraventricular nucleus (mpPVN) of the hypothalamus depends on cortico-limbiclevel excitatory and inhibitory inputs. The anterior cingulate (ACC), orbital frontal cortex (OFC), amygdala, bed nucleus of the stria terminalis (BNST), hippocampus, and preoptic area (POA) all have direct inputs to the hypothalamus. The ACC, OFC, and amygdala are also reciprocally interconnected. The amygdala and ventral subiculum also have pathways to the hypothalamus by way of the BNST. The hippocampus and the amygdala are connected to the locus coeruleus (LC), which releases norepinephrine to brain areas involved in alerting. For systemic stressors, the mpPVN is activated by the brain stem, either directly or by way of limbic circuits. Once activated in response to a systemic or processive stressor,

the mpPVN produces corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP), which travel through the hypophysial portal system to the anterior pituitary gland, stimulating the production and release of adrenocorticotropic hormone (ACTH). ACTH stimulates cells in the adrenal cortex to produce glucocorticoids (cortisol in humans). Cortico-limbic regions including the hippocampus and ACC maintain feedback control of the PVN. The hypothalamus is also a key region in the sympathetic adrenomedullary system. The lateral hypothalamus activates nuclei in the brain stem, including the parabrachial nuclei, that regulate release of sympathetic hormones (norepinephrine, NE, and epinephrine, EPI) and parasympathetic hormones (acetylcholine, ACH), with pathways traveling from the spinal cord to preganglionic nuclei or to target organs (e.g., the adrenal medulla). (Adapted from Gunnar and Davis, 2003.)

physiology and behavior (de Kloet, 1991; Sapolsky, Romero, and Munck, 2000). Modulation and eventual termination of the LHPA stress response is achieved by means of multiple negative feedback loops, with varying time courses and mechanisms. Within minutes, cortisol binds to receptors in the anterior pituitary, hypothalamus, and hippocampus, inhibiting the LHPA axis at multiple levels by means of this signaling (see Dallman et al., 1992). Over the course of hours to days, cortisol also down-regulates CRH gene expression in the hypothalamus and POMC gene expression in the anterior pituitary, thereby exerting a prolonged suppressive effect on ACTH secretion (Gunnar and Vazquez, 2006). The LHPA system does not lie dormant between stress responses. Rather, the stress response is superimposed on an LHPA circadian rhythm. Basal cortisol levels peak about 30 minutes after waking and gradually decline across the day to reach their nadir in the late evening, near the onset of sleep (Daly and Evans, 1974; Kwak et al., 1993). In humans, this circadian rhythm is evident by 6 weeks of age (Larson et al., 1998), and the decline across the day becomes more

stable by 4–6 years of age as children give up daytime naps and develop a more adultlike sleep schedule (Watamura et al., 2004). Thus developmental researchers are frequently interested in assessing LHPA rhythmicity and reactivity as a means of studying the influence of stressful experience on the developing brain, as well as individual differences in stress vulnerability and resilience. Over the past several decades of research, a number of methodological considerations have come to light.

LHPA measurement in developmental research Interrogating the LHPA axis in the context of human development presents a host of challenges. Because of ethical and practical considerations, human development researchers are generally limited to measuring cortisol levels and must rely on animal models to draw tentative inferences about what is occurring at the levels of the pituitary and hypothalamus (Gunnar and Vazquez, 2006). Ethical concerns also restrict the paradigms available to human development researchers for assessing LHPA reactivity. Pharmacologic

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

65

probes are rarely used with children, and often ethically approved psychosocial stressors do not reliably activate the LHPA axis. For infants, separation from a caregiver generally elevates cortisol levels, although elevations in response to brief separations are not always observed among infants over 12 months of age (e.g., Spangler and Schieche, 1998). For older children and adults, manipulations that involve a threat to the social self (such as critical social evaluations by authority figures) sometimes are effective in provoking a stress response (Dickerson and Kemeny, 2004). However, the inconsistency of these mild manipulations in raising cortisol levels in children necessitates a heavy reliance on animal models to inform our understanding of cortisol reactivity to more severe stressors. Collecting salivary cortisol samples is a noninvasive, relatively well defined procedure. In contrast, interpreting the significance of the cortisol values obtained from those samples is a complicated, ambiguous undertaking requiring consideration of many variables that affect the LHPA system, such as collection context and time of day. Therefore, at the design phase of a study, careful attention must be paid to deciding when, where, and how frequently to sample cortisol (Gunnar and Talge, 2007). Studies measuring cortisol reactivity in children have often compared a cortisol sample collected upon arrival at the laboratory to samples collected following a psychosocial stressor. However, this approach is misleading because the first sample is not a typical baseline measure: It reflects an LHPA response to coming to the laboratory. For reasons that remain mysterious, infants and preschoolers often show cortisol levels that are suppressed upon arrival at the laboratory compared to samples collected at home at the same time of day (Goldberg et al., 2003; Gunnar et al., 1989; Larson, Gunnar, and Hertsgaard, 1991; Legendre and Trudel, 1996; Lundberg, Westermark, and Rasch, 1993). Several studies have found that older children (ages 9 and up) had higher cortisol levels upon arrival at the laboratory than time-matched home samples (Gunnar et al., in press; Tottenham et al., 2001), indicating that there may be developmental differences in response to the laboratory collection context between early and later childhood. Because of the multifaceted dynamic processes affecting cortisol levels, no one sample can prudently be considered “baseline.” Researchers commonly use one or both of the following methods to address the baseline problem. First, they collect home samples at the same time of day as the laboratory visit, which is equivalent to animal researchers’ definition of baseline levels as those observed when the animal has been left undisturbed in the home cage. Second, when children are assessed in the laboratory, researchers allow for an extended period of acclimation to the laboratory context prior to collecting baseline samples (Gunnar and Talge, 2007; van Goozen et al., 2007).

66

Because of the LHPA diurnal rhythm, time of day is another key design consideration, with the best time to sample varying depending on the question of interest. Latent state-trait modeling has demonstrated that the maximum trait component occurs shortly after waking, at the diurnal peak, making this a good sampling time for researchers interested in stable individual differences in basal cortisol levels (Kirschbaum et al., 1990; Shirtcliff et al., 2005). Conversely, researchers interested in stress reactivity may choose to sample in the late afternoon or evening, when state components are higher and the system is more responsive to stressors (Dallman et al., 1992; Kirschbaum et al., 1990). Variability in cortisol levels from day to day presents another interpretive challenge. Sampling on a single day is unlikely to be sufficient to detect stable individual differences. Aggregating time-matched samples across three or more days provides a more reliable measure (Gunnar and Talge, 2007). However, the day-to-day variation itself may be informative, particularly in identifying abnormal LHPA function and risk for psychopathology. In a recent study by Halligan and colleagues (2004), children whose mothers were clinically depressed during their first year of life, at 13 exhibited higher and more variable cortisol levels soon after awakening than did children of nondepressed mothers, even after controlling for maternal depression postinfancy. Goodyer and colleagues (2000) noted that among adolescents at high risk for depression, obtaining one or more abnormally high morning cortisol levels across four days of sampling predicted onset of depression in the ensuing year. Given these findings, the use of statistical methods such as hierarchical linear modeling to isolate variability may yield a richer picture of LHPA function than can be obtained by simply aggregating sample values. Thus, while collection of salivary cortisol samples is quite straightforward, meaningful assessment of LHPA function in children is a far more problematical and ambiguous undertaking than it might at first appear. Ethical and practical issues constrain the techniques available to human development researchers to probe the LHPA axis. Research design and the interpretation of cortisol results must take into account the time of day, collection context, developmental differences in LHPA reactivity, and variability in cortisol from day to day. On the positive side, consideration of all these variables may yield a richer, more refined picture of the effects of early adverse experience on cortisol rhythm and reactivity. However, human development researchers owe a substantial debt to the adult and animal literatures for informing our understanding of the LHPA system. While interpretation of adult and animal studies is also challenging, there are somewhat less severe constraints on researchers’ ability to probe the axis. The next section provides an overview of the influence of cortisol and CRH on the brain,

fundamentals of developmental neurobiology

relying heavily on the human adult literature and on animal models.

Influence of cortisol and CRH on the brain Cortisol targets tissues throughout the brain and periphery, making adjustments in the allocation of the body’s resources in response to stressors in the environment that threaten homeostasis. McEwen (1998) referred to this maintenance of stability through change as allostasis. Cortisol readily permeates the blood-brain barrier, and its activity influences multiple brain regions including (but not limited to) the hippocampus, amygdala, hypothalamus, prefrontal cortex, and orbital and medial cortical regions. Cortisol operates on tissues through well-understood genomic mechanisms and relatively little understood nongenomic mechanisms (de Kloet, 1991). Cortisol mediates genomic effects by binding to its receptors, after which the hormone-receptor complex is translocated to the cell nucleus, where cortisol interacts with glucocorticoid receptive elements (GREs) on the genome and modulates gene expression. One reason for interest in the LHPA system among developmental researchers is that cortisol is a gene transcription factor that plays complex roles in modulating genes involved in brain development as well as brain function (Gunnar and Vazquez, 2006). Nongenomic effects of cortisol will be briefly described later; here, we discuss the better-understood genomic effects that occur when cortisol binds to its receptors. There are two receptors that bind cortisol in the brain, mineralocorticoid receptors (MR) and glucocorticoid receptors (GR). These receptors mediate different types of effects (de Kloet, 1991). It may seem odd that MRs mediate cortisol effects in the brain, because outside the brain, these receptors bind aldosterone, a hormone involved in salt-water balance. However, the enzyme 11 beta hydroxysteroid dehydrogenase, which protects MRs from cortisol in the periphery, is not sufficiently present in the brain. As a result, MRs bind cortisol in the central nervous system. An inverted U function characterizes the relationship between cortisol and physical and behavioral health, such that moderate levels of cortisol are most adaptive, while both chronically low and chronically high levels yield differential but similarly deleterious effects. Two characteristics of MR and GR account for this paradox: cortisol’s differential affinity for these two types of receptors and the different types of effects they generate when bound and activated (Sapolsky, 1997). Cortisol has more than 10 times higher affinity for binding with MR than with GR (de Kloet, Oitzl, and Joels, 1993). Accordingly, at typical basal levels, 80–90 percent of MRs are bound with cortisol while very few GRs in the brain are bound (de Kloet, 1991). Thus, in this basal range, MRmediated cortisol effects predominate. These effects include maintaining a steady electrical current in the brain so that

neurons will be able to respond to neurotransmitters, maintaining the LHPA circadian rhythm, and facilitating cerebral glucose availability (Bradbury, Akana, and Dallman, 1994; see for review Gunnar and Vazquez, 2006). These MR-mediated effects are classified as permissive because they support the ability to immediately respond to stressors by means of the other stress system (i.e., the fast-acting sympathetic adrenomedullary system; Ingle, 1952; Sapolsky, Romero, and Munck, 2000). So, while stress-related increases in cortisol are much too slow to be helpful in responding to imminent physical threats, basal levels do play a crucial role in facilitating rapid responding to stressors (e.g., fight/flight responses; Sapolsky, Romero, and Munck, 2000). As cortisol levels increase in response to a stressor or at the morning peak of the circadian rhythm, cortisol molecules in the brain will occupy the remaining MRs and begin to occupy GRs. In the brain, GR-mediated effects tend to be suppressive, in that they counteract the responses of other stress-reactive systems (e.g., turning off stress-induced immune system responses and opposing impacts of stressinduced catecholamine actions on neural systems). GRmediated activity in the hippocampus, hypothalamus, and pituitary also serves to contain stress-induced activations of the LHPA axis in a process termed negative feedback. The suppressive effects of these stress-induced cortisol elevations are thought to serve the function of restoring homeostasis in the aftermath of a challenge (Sapolsky, Romero, and Munck, 2000). Many GR-mediated effects counteract MR-mediated effects. For example, MRs increase cerebral glucose availability, while GRs reduce cerebral glucose transport (Sapolsky, Romero, and Munck, 2000). Similarly, while GRs inhibit hippocampal neurons, impairing synaptic plasticity and memory formation, MRs boost synaptic plasticity and facilitate memory formation by lowering the refractory period of hippocampal neurons (for review, see Gunnar and Quevedo, 2007). When MRs and GRs act in opposition to one another, the effect of cortisol depends on the ratio of MR to GR occupation (de Kloet, 1991). If cortisol levels are chronically elevated, this ratio will tilt toward GR-mediated effects, with deleterious consequences including dendritic atrophy, cell death, and impaired learning and memory. In addition to suppressive effects, GRs also mediate preparatory effects, that is, long-term changes in gene expression that influence the LHPA response to future stressors (Sapolsky, Romero, and Munck, 2000). For example, chronically high levels of GR occupation can lower the amygdala’s threshold for responding to threatening stimuli and activating the mpPVN, with the result that LHPA responses may become more frequent and prolonged (Rosen and Schulkin, 1998). However, preparatory impacts of cortisol on CRHproduction in the mpPVN may oppose those in the amygdala, resulting in blunted cortisol responses to mpPVN stimulation (Rosen and Schulkin, 1998). These opposing

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

67

effects make it very difficult to predict, a priori, whether exposure to chronic or frequent high levels of cortisol will increase or decrease cortisol reactions to subsequent stressors (Friese et al., 2005). Additionally, chronically high cortisol levels, acting through GRs, can put the body’s physical health at risk because of their immunosuppressive effects (inhibiting cytokine transcription). However, chronically low cortisol levels can also be deleterious to physical and behavioral health because not enough MRs will be occupied to maintain the body in a state of preparedness to cope with the stressors it will inevitably encounter (de Kloet, 1991). The relation of both very high and very low cortisol activity to poor physical and mental health reflects the inverted Ushaped relation between this neuroendocrine system and healthy adaptation. The effects of LHPA activity also depend upon the location of the MR or GR receptors within the brain. Differences among species have been identified in the pattern of distribution of MR and GR receptors. In rats, MRs are mainly confined to the hippocampus and lateral septum. GRs, while also plentiful in the hippocampus and lateral septum, are more broadly expressed, appearing in the mpPVN, central amygdala, and nucleus tractus solitarius, among other locations (Reul and de Kloet, 1985; Sanchez, 2006). Evidence from nonhuman primates indicates relatively fewer GRs in the hippocampus compared to rodents (Sanchez et al., 2000). In primates, high concentrations of MRs and GRs have also been observed in the prefrontal cortex and other cortical regions, implicating cortisol in cognitive and emotional regulatory function (López, Akil, and Watson, 1999; Patel et al., 2000; Sanchez, 2006; Sanchez et al., 2000). Effects of cortisol produced by impacts on gene transcription take many minutes to hours (Sapolsky, Romero, and Munck, 2000). Recent evidence indicates that cortisol can have impacts on behavior and neural activity that occur too quickly to operate through genomic mechanisms and that are observed even in the presence of drugs that block MR and GR binding (see review by Makara and Haller, 2001). Nongenomic, rapid impacts of cortisol are not well understood, but they hold promise for a much better appreciation of the role of cortisol in adaptation. Notably, the picture that is emerging for these nongenomic effects includes evidence that they involve specific binding sites in the synaptic membrane, are involved in modulating neurotransmitter interactions with their receptors, and may have particularly marked effects on neurons that are active in response to the stressor event. These effects do not appear to be specific; rather, they seem to enhance activity that is already occurring. As a consequence, rapid nongenomic effects may exacerbate individual differences in adaptive responses to threat. In other cases, rapid nongenomic effects have been found to be the opposite of more slowly emerging genomic effects

68

(Makara and Haller, 2001). In sum, the more that is learned about rapid nongenomic and slower genomic impacts of cortisol (and corticosterone), the more apparent it becomes that the HPA system has multiple and often contradictory roles in adaptation, roles that depend on timing, context, and the specific neural system being affected. The complexity and context-dependent findings of cortisol-behavior studies mirror the complexity and subtlety of the neurophysiology of this system. Although studies with children must rely on measures of cortisol, many of the effects attributed to the LHPA system may have more to do with its releasing hormone, corticotropin-releasing hormone (CRH), than with cortisol itself. CRH and its associated receptors and ligands have a prominent role in the organization of behavioral, autonomic, and neuroendocrine responses to processive stressors. CRH is produced not only by the hypothalamus, but also in the central nucleus of the amygdala and other fear- and anxietyrelated regions, and the CRH involved in behavioral impacts is likely produced by these extrahypothalamic sites (Heinrichs and Koob, 2004; Rosen and Schulkin, 1998). Similarly, CRH receptors are located in regions involved in appraisal and processing of psychological threat (e.g., cingulate cortex, orbital/medial prefrontal cortex, amygdala, bed nucleus of the stria terminalis, and locus coeruleus; Bale and Vale, 2004). As with cortisol, CRH has two primary receptors (1 and 2) that have distinct distributions in the brain and appear to have opposing effects (Dautzenberg and Hauger, 2002; de Kloet, 2004; Vermetten and Bremner, 2002). CRH-1 receptors are believed to mediate many of the fear- and anxiety-like effects of CRH (M. Davis, 1997; M. Davis et al., 1993; LeDoux and Phelps, 2000), while CRH-2 receptors are believed to mediate many of the vegetative effects observed in chronic stress (e.g., stress-induced suppression of eating; Vermetten and Bremner, 2002). Notably, impacts of altered amygdala-produced CRH can be observed in the absence of robust activation of the LHPA axis or elevations in glucocorticoids by the adrenal (Makino et al., 1999). To summarize, the LHPA system affects the brain through multiple, often contradictory mechanisms to promote allostasis, the maintenance of stability through change. Slow, long-lasting genomic effects of cortisol in the brain are mediated by two receptors, MR and GR, with opposing influences and differential affinities for cortisol, such that an inverted U-shaped function characterizes the relationship between cortisol levels and healthy adaptation. In an emerging research area, evidence is accumulating for faster-acting nongenomic mechanisms of cortisol activity as well. In the context of processive stressors, CRH also appears to have widespread impact on the brain. This multilayered system of checks and balances is not fully mature at birth, yet cortisol and CRH have critical roles in brain development.

fundamentals of developmental neurobiology

We turn now to consider how these hormones shape the developing brain.

Influence of cortisol and CRH during brain development Prenatally and postnatally, cortisol is required for neurogenesis, synaptogenesis, gliogenesis, neural apoptosis, myelination, and development of neurotransmitter systems (Duman, Malberg, and Thome, 1999; for review see Gunnar and Vazquez, 2006). As noted, cortisol is a gene transcription factor, which allows even small variations in cortisol levels to have prolonged and widespread influences on neural development. These influences depend, of course, on the timing of cortisol variations in relation to periods of brain development. Several lines of evidence document the effects of elevated cortisol on the developing brain. First, in human fetuses at risk for premature birth, the medically prescribed administration of a synthetic form of cortisol called dexamethasone (DEX) creates a natural experiment to examine the effects of elevated cortisol on the pre- and perinatal brain. Also, 11β-HSD in the placenta helps to buffer the fetus from natural maternal cortisol, but DEX is unaffected by this placental barrier (Seckl, Cleasby, and Nyirenda, 2000). Consequently, DEX, which is administered to accelerate fetal lung maturation, has the significant side effect of flooding the fetal brain with an active and long-acting form of cortisol. In human infants, it is difficult to disentangle the effects of DEX exposure from the effects of being born prematurely. Animal models are helpful in isolating these effects and have demonstrated long-term consequences of prenatal exposure to excess glucocorticoids. In fetal rats, DEXinduced elevations inhibit neurogenesis, gliogenesis, cell division, and myelination. Fetal DEX exposure has been linked to permanent blunting of norepinephrine expression in the hippocampus and cerebral cortex; elevated serotonin expression in the hypothalamus, hippocampus, and brain stem; and elevated LHPA activity with impaired negative feedback (see for reviews Matthews, 2000; Whitelaw and Thoresen, 2000). Animal studies also suggest a link between prenatal glucocorticoids and long-term emotional and behavioral functioning. Prenatal DEX exposure is associated in adulthood both with impaired coping in adverse situations in adulthood and with elevated CRH in the central nucleus of the amygdala, a region instrumental in fear and anxiety (Welberg et al., 2001). Seckl and Meaney (2004) suggest that the long-term consequences of excess glucocorticoid exposure reflect prenatal programming of the HPA system, including permanent effects on GR gene expression. Cortisol also affects myelination in the developing brain because GRs are expressed in oligodendrocytes, the glial cells that manufacture myelin sheaths in the CNS (Cintra et al., 1994; Huang et al., 2001). Huang and associates (2001) examined the effect of repeated prenatal administration of

exogenous cortisol to sheep on the corpus callosum, a major white matter tract critical to cognitive and attentional processes. The corpus callosum was selected as the focus of this study because it was hypothesized to be particularly vulnerable to the effects of repeated exogenous cortisol administration as a result of its immaturity at the time of cortisol administration and its prolonged period of myelination. Indeed, they found that myelination was delayed in this region, suggesting another potential side effect of high levels of prenatal cortisol. Therapeutic administration of synthetic glucocorticoids to human adults to treat brain tumors also results in white matter abnormalities, specifically a widespread decrease in extracellular fluid concentrations in white matter (Minamikawa et al., 2004). The hippocampus appears to be particularly vulnerable to chronically elevated cortisol. Its structure and function are sensitive to environmental influence throughout its extended period of postnatal development, and in the dentate gyrus of the hippocampus, neurogenesis continues even in adulthood (Gould and Tanapat, 1999). Exposure to chronic stress has been associated with decreased dendritic arborization in the CA3 layer of the hippocampus (Watanabe et al., 1992), a prolonged suppressive effect on neurogenesis in the dentate gyrus (Fuchs, Uno, and Flugge, 1995; Gould and Tanapat, 1999), and impaired performance on hippocampally mediated learning and memory tasks (see for review Gould and Tanapat, 1999). Across the life span, granule neuron proliferation in the dentate gyrus is inversely related to cortisol levels (Sapolsky and Meaney, 1986; Schlessinger, Cowan, and Gottlieb, 1975). In rats, granule neuron proliferation peaks in the postnatal stress hyporesponsive period (SHRP), when corticosterone levels are low, and slows down when basal corticosterone levels rise near the end of the SHRP (Schlessinger, Cowan, and Gottlieb, 1975). Elevating corticosterone during the SHRP by exogenous administration or by exposure to an intense stressor (odor of a predator) decreases granule neuron proliferation in the developing dentate gyrus (Gould et al., 1991; Tanapat, Galea, and Gould, 1998). Most granule neuron precursors do not have MRs or GRs (Cameron, Wooley, and Gould, 1993), so cortisol impedes granule neuron proliferation indirectly through an NMDA-receptor-mediated pathway. Specifically, elevated cortisol (or corticosterone) levels stimulate the hippocampus to release glutamate, which inhibits granule neuron proliferation in the dentate gyrus (Gould and Tanapat, 1999). The deleterious effects on the hippocampus of chronic exposure to cortisol may be partially mediated by alterations in the expression of neurotrophic factors. Neurotrophic factors are proteins that facilitate neurogenesis and synaptogenesis. They also preserve existing neurons by preventing apoptosis. In the dentate gyrus and other hippocampal regions of rats subjected to chronic stress (immobilization),

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

69

Smith and colleagues (1995a, 1995b) found greatly decreased brain-derived neurotrophic factor (BDNF) mRNA, paired with increased neurotrophin-3 (NT-3) mRNA. The elevation in NT-3 expression was mediated by corticosterone. The mechanism for BDNF down-regulation was less clear, as a decrement in BDNF expression also was observed in the dentate gyrus of adrenalectomized rats exposed to chronic stress. The authors suggest that corticosterone likely plays some role in the suppression of BDNF, but other elements of the stress response must also be involved (Smith et al., 1995a, 1995b). While this study was conducted with adult rats, it has implications for the effects of glucocorticoids on the dentate gyrus and other hippocampal regions during their extended period of postnatal development. By altering expression of these neurotrophic factors, chronically elevated cortisol (in primates) and corticosterone (in rodents) could disrupt neurogenesis, synaptogenesis, and selective neuronal survival, perhaps inducing long-term effects on learning and memory. Because glucocorticoid effects can vary depending on species and developmental status, however, it will be important to replicate these findings in nonhuman primate infants whose postnatal neural development is more comparable to human development. Notably, there is already evidence that exposing the fetal rhesus monkey to dexamethasone impairs the development of the hippocampus, although to our knowledge this has not been studied in relation to alterations in BDNF activity as a potential mechanism in this process (Uno et al., 1990). In addition to these cortisol effects, CRH also affects the developing brain. The mechanisms through which CRH influences development depend on the timing of exposure, likely related to maturational changes in neurotransmitter production as well as receptor density and distribution. In the rodent hippocampus, the number of CRH-expressing neurons and CRH-1 receptors peaks on postnatal days 11– 18 and then declines drastically to adult levels (Chen et al., 2001). This abundance of CRH neurons is believed to enhance synaptic transmission, thus contributing to postnatal processes of learning and memory. Excessive CRH during this period of particular hippocampal excitability could make rodents prone to pathological outcomes. Similar windows of vulnerability may be present in the development of the human nervous system, and processes of priming involving CRH-1 receptors may be related to the development of human stress-related disorders. Experiments with rodents illustrate how CRH mediates mechanisms of neural plasticity that provoke long-term changes in limbic pathways involved in emotional disorders. For example, activation of CRH receptors facilitated longterm potentiation in the mouse hippocampus and enhanced context-dependent fear conditioning in rats subjected to acute stress (Blank et al., 2002). As further evidence of CRHinduced priming effects, the enhanced learning that rodents

70

typically show under stress did not occur in the context of pharmacological inhibition of hippocampal CRH receptors (Blank et al., 2002). A recent study demonstrated that CRH induces long-lasting cellular changes in the amygdala, which increase the rodent’s stress vulnerability (Rainnie et al., 2004). A potent CRH receptor agonist was chronically infused into the basolateral amygdala. The anxiety-like syndrome these rodents developed was correlated with cellular mechanisms of neural plasticity known to mediate long-term learning. The repeated activation of the basolateral amygdala resulted in long-term changes in the sensitivity of its neurons, including reduction of spontaneous inhibitory synaptic potentials and reduced expression of the inhibitory neurotransmitter GABA (Rainnie et al., 2004). This downregulation of inhibitory mechanisms resulted in chronic amygdalar hyperexcitability. These studies implicate cortisol and CRH in a variety of crucial neurodevelopmental processes and demonstrate that distortions in levels of these hormones can have persistent adverse effects in shaping the developing brain. For the developmentalist, such findings immediately raise the issue of the neurodevelopmental consequences of early stressful experiences. In the animal literature, research on this issue has a long history (reviewed in Levine, 2005). Here we review current findings from the animal literature on effects of early experience, to lay the groundwork for discussing the findings on this topic from studies of human development.

Early experience effects in animal models Studies in both rats and nonhuman primates indicate that early experiences, particularly those involving variations in parental care, can have long-term effects on reactivity and regulation of the LHPA system (reviewed in Sanchez, 2006). In nature, variations in parental care occur on a continuum. Experimental manipulations have yielded evidence that the developing LHPA system is sensitive not only to gross deviations in care such as peer rearing (e.g., Champoux et al., 1989), but also to species-typical variations in parenting quality. For instance, Francis and colleagues (1999) crossfostered the offspring of mothers who were at the high and low ends of the continuum of licking/grooming and archback nursing, key indicators of parenting quality in the rat. These variations within the normal range of maternal behavior influenced stress reactivity in the foster offspring. Many of these effects appear to involve regulation of MR and GR development by means of alterations in GR gene expression. This has been convincingly demonstrated in studies of infant rats. In rats, hippocampal GR expression increases throughout development (Suchecki, Rosenfeld, and Levine, 1993; M. Schmidt et al., 2003). Quality of parental care permanently and radically affects the density of hippocampal GR expression, thereby shaping LHPA reactivity in adulthood

fundamentals of developmental neurobiology

(Meaney and Szyf, 2005). Highly nurturant parental care increases hippocampal GR density and leads to more efficient feedback regulation of the LHPA axis, while extended parental separation and low nurturant parental care result in fewer hippocampal GRs and prolonged LHPA reactivity to processive stressors (for review see Gunnar and Vazquez, 2006). In elegant work, Meaney and colleagues (see Meaney and Szyf, 2005) traced these effects to the role of parental care in regulating methylation of the GR gene. Although long-term impacts of variations in parental care and parental separation have been observed in nonhuman primates, it is not clear whether these are mediated in the same way as in rodents (see Sanchez, Ladd, and Plotsky, 2001). For example, in nonhuman primates, postnatal levels of GR expression do not appear to vary developmentally, and therefore do not appear sensitive to social experience. In rats, MR density is highest during the early postnatal period, peaking in the hippocampus around postnatal day 10, and is sensitive to early social experience (Vazquez et al., 1998, 1993). One species of New World monkey shows a developmental peak in MR expression coinciding with weaning. Pryce and colleagues (2005) posit that the weaning phase could therefore represent a sensitive period for social experience to shape MR expression in primates, with potential long-term consequences for LHPA basal levels in the wake of early social adversity. It is not yet known whether humans share this MR-sensitive period. It would be informative to replicate the study with Old World monkeys, who are more closely evolutionarily related to humans. These findings exemplify the ubiquitous challenge in translational research of determining whether observed phenomena are species-specific or more broadly applicable (Sanchez, 2006). At the same time, they illustrate the rich potential of preclinical models to explicate the mechanisms of social influence on the LHPA system. Notably, some of the effects of early experience on brain development may operate by means of CRH rather than through modulations of glucocorticoids. During the stress hyporesponsive period in the rodent, perturbations that do not elevate corticosterone do increase CRH activity in the brain (Smith et al., 1997). Disturbances in parental care also produce persistent increases in CRH activity that, even in the absence of elevated corticosterone, would be capable of the deleterious impacts on hippocampal development noted to co-occur with disturbances in parental care (Baram et al., 1997, 2001; Brunson et al., 2001). Furthermore, although long-term changes in basal cortisol levels have been difficult to observe in nonhuman primate studies of disturbances in parental care, there is evidence of chronic increases in cerebral spinal levels of CRH in rhesus monkeys who had been reared by mothers foraging under unpredictable conditions compared to those whose mothers enjoyed predictable foraging conditions (e.g., Coplan et al., 1996). Finally, early

disturbances in parental care shift the balance in CRH receptors toward those that mediate fearful or anxious behaviors and heightened reactivity of the LHPA and sympathetic adrenomedullary (SAM) system (see review, Sanchez, Ladd, and Plotzky, 2001). Thus early rearing experiences in rodents and monkeys may produce long-term impacts on brain and behavior through multiple LHPArelated mechanisms. These animal models have helped to inform hypotheses about LHPA regulation in the complex realm of human social experience.

Social experience and LHPA regulation in human development The LHPA system is immature at birth, and LHPA basal function and reactivity continue to evolve throughout childhood (for review see Gunnar and Donzella, 2001). During this extended period of development, social experiences contribute in important ways to shaping these brain circuits. The sensitivity of the developing LHPA axis to social experience is both an asset and a liability. On the positive side, sensitive, responsive caregiving buffers the LHPA system during the first few years of life, preventing or mitigating cortisol elevations. However, social experience is also one of the major sources of stress that challenges the immature LHPA axis. In the absence of sensitive, responsive care, young children faced with psychosocial stressors such as maternal separation, child maltreatment, or social deprivation are highly vulnerable to cortisol elevations and longterm disturbances in LHPA regulation. Human newborns have a highly reactive LHPA system, exhibiting marked elevations in cortisol and ACTH in response to stressors such as well-baby physical examinations or childhood inoculations (for review see Gunnar, 1992). Over the course of the first year of life, LHPA reactivity to mild stressors declines precipitously, despite the fact that behavioral distress in response to these same stressors continues to be observed (Gunnar, Broderson, Krueger, et al., 1996; Larson et al., 1998; Lewis and Ramsay, 1995). By 12 months of age, it is difficult to observe mean increases in cortisol to mild stressors such as childhood inoculation injections, and this resistance to producing elevations in cortisol at least at the mean or group level persists through the toddler and preschool years (de Haan et al., 1998; Gunnar et al., 1997; Nachmias et al., 1996). This period of blunted LHPA reactivity in young humans appears to be functionally equivalent to the stress hyporesponsive period (SHRP) that occurs in rat pups from postnatal days 4 to 14, which is hypothesized to shield the developing brain from potentially damaging elevations in corticosterone. In rats and humans alike, the SHRP is mediated by sensitive, responsive caregiving. For rat pups, if maternal caregiving is sufficiently disorganized, marked increases in

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

71

corticosterone are observed even during the SHRP (Suchecki, Rosenfeld, and Levine, 1993). In humans, sensitive, responsive caregiving by the attachment figure is believed to result in formation of a secure attachment bond (Sroufe, 1983), and securely attached toddlers do not exhibit cortisol elevations to mild stressors while in the presence of their attachment figure (Ahnert et al., 2004; Gunnar, Broderson, Nachmias, et al., 1996; Nachmias et al., 1996; Spangler and Grossman, 1993; Spangler and Schieche, 1998). In contrast, toddlers with a disorganized/disoriented attachment status may be particularly stress vulnerable. Unlike securely attached toddlers, disorganized/disoriented toddlers show high cortisol levels in response to repeated maternal separations (Hertsgaard et al., 1995; Spangler and Grossman, 1993). Disorganized/disordered attachment behavior is a pattern of behavior often associated with maltreatment in which toddlers are ambivalent about whether to approach or avoid the primary caregiver and are unable to use their caregivers as a coping resource in stressful situations (Main and Solomon, 1990; van Ijzendoorn et al., 1999). Studies with children have shown that sensitive, responsive alternate caregivers, such as child care providers and babysitters, are also able to buffer infants and toddlers from cortisol elevations to mild stressors even when the mother is absent (Dettling et al., 2000; Gunnar et al., 1992). However, in a laboratory manipulation, maternal separation did result in significant increases in cortisol for 9-month-olds when the babysitter was instructed to be distant and perfunctory (Gunnar et al., 1992). These findings reveal the major drawback of the immature LHPA system’s dependence on social regulation: When deprived of a sensitive, responsive caregiver, toddlers become highly vulnerable to activation of the LHPA axis with significant, and sometimes large, increases in cortisol to even mild stressors (for review see Gunnar and Donzella, 2001). Findings from animal research indicate long-term consequences of chronic cortisol elevations in the context of disrupted care. These findings raise the specter of increased risk for psychopathology in genetically vulnerable children deprived of sensitive, response care during the toddler years. However, prospective research on the functional consequences of disrupted care for the developing LHPA axis has yet to be conducted in humans. Short- and long-term functional consequences would likely depend on the frequency and chronicity of LHPA elevations. Young children chronically subjected to inadequate or disrupted care tend to show dysregulation of basal LHPA function. Specifically, children exposed to deprived rearing environments (such as orphanages or neglectful homes) have blunted early morning cortisol levels and do not show the typical decline in cortisol levels over the course of the day (Bruce et al., under review; Carlson and Earls, 1997; Fisher et al., 2000; Kroupina, Gunnar, and Johnson, 1997). This flattening of the diurnal rhythm is also observed in chroni-

72

cally stressed and neglected rhesus monkey infants (Boyce et al., 1995; McCormack et al., 2003; Sanchez, 2006). These blunted or low early-morning levels may reflect changes in the LHPA axis and in stress-related neural activity in limbic and cortical regions that will, over time, result in increases in basal cortisol levels. For example, toddlers recently adopted from an orphanage or other institution and those studied soon after placement in foster care after being removed from maltreating parents often exhibit extremely low or blunted early-morning cortisol levels (Bruce et al., 2000; Dozier et al., 2006; Gunnar et al., 2006). However, studies of such children many years after adoption from orphanages or rescue from abusive parental care have often yielded evidence of elevated basal cortisol levels for at least some of the children. For example, elevated basal cortisol levels have been reported for physically or sexually abused children who have internalizing disorders when studied months or years after rescue (Carrion et al., 2002; Cicchetti and Rogosch, 2001a, 2001b; De Bellis et al., 1994, 1999). In two studies conducted at day camps, depressed maltreated children showed a rise in cortisol levels across the day instead of the expected decline (Hart, Gunnar, and Cicchetti, 1996; Kaufman, 1991). Finally, orphanageadopted children who were the most severely affected by their preadoption experiences as evidenced by severe growth delays at adoption had elevated early morning cortisol levels when studied an average of 6 years postadoption (Kertes et al., in press). Studies of adults maltreated as young children also tend to confirm the long-term consequences of maltreatment on activity of the LHPA system. However, the precise effects observed depend on the methodology employed—that is, pharmacological probes that activate specific levels of the axis versus processive stressors that rely on cortico-limbic circuits to activate the axis. In addition, particularly for processive stressors, whether hyper- or hypoactivity of the LHPA axis is observed depends on whether or not the adult is suffering concurrently from a clinical affective disorder and on the nature of that disorder (posttraumatic stress disorder or depression, or both). Among adults, PTSD is associated, anomalously, with low basal cortisol levels and blunted cortisol responses to many psychosocial stressors (Yehuda et al., 2001), while depression is associated with elevated basal cortisol levels, a blunted diurnal rhythm in cortisol, and hyperreactivity of the LHPA system to stressors (McEwen, 2005). Childhood maltreatment effects do not precisely mirror the effects observed for adults with these disorders who do not have a history of maltreatment during childhood, although conclusions remain tentative because not all studies of adults with childhood maltreatment histories have been careful to employ a nonmaltreated affective disorder comparison group (see for discussion, Heim, Plotsky, and Nemeroff, 2004). On the whole, however, as in the

fundamentals of developmental neurobiology

studies of children, a history of maltreatment increases the likelihood of abnormalities in the cortisol daily rhythm, atypically low basal ACTH levels (suggestive of down-regulation in response to chronic CRH drive on the pituitary), and problems in regulating ACTH and cortisol responses to psychosocial threat (see Heim et al., 2000, 2001; reviewed in Tarullo and Gunnar, 2006). These data tend to support the hypothesis that chronic stress during development sensitizes brain circuits to perceived environmental threat, influencing the frequency, amplitude, and duration of future LHPA stress responses and resulting in increased stress vulnerability and elevated risk of internalizing disorders (Heim and Nemeroff, 2001). Even for typically developing children not exposed to the extreme adversity of maltreatment or deprivation, social experience remains one of the most significant sources of stress. Many two- to four-year-old children show a rise in cortisol levels over the course of the day when attending full-day child care, but not on the days they are at home (Dettling, Gunnar, and Donzella, 1999, 2000; Tout et al., 1998; Watamura et al., 2003, 2004). For children of this age, peer interactions are becoming increasingly salient, but they are still in the process of acquiring the social skills to negotiate those interactions. Long hours in the challenging peer environment of child care may strain the emerging regulatory capacities of the LHPA system (reviewed in Gunnar and Donzella, 2001; Gunnar and Quevedo, 2007). Indeed, the magnitude of cortisol increases over the course of the childcare day is greatest for the children who are least skilled at negotiating social interactions: those rejected by peers and those rated as less socially competent and less capable of emotion regulation (Dettling et al., 1999, 2000; Gunnar et al., 1997, 2003). As further evidence that the developing LHPA axis is under strong social regulation by adults, children who receive sensitive, responsive, individualized care from child-care providers show little or no evidence of daycare cortisol elevations, even when their behavior is such that it would provoke negative or hostile responses from peers (Dettling et al., 2000). While we do not yet understand the significance for later development of the small but frequent cortisol elevations observed in less sensitive and responsive day-care settings, this important question is currently under active investigation. A key consideration in this research area will be the possible interaction between variations in parental care and variations in child care. Another challenge will be determining whether these cortisol elevations specifically affect development over and above the impact of being in a lower quality child-care setting. LHPA sensitivity to child care is age specific, such that, as a group, children aged 5 and older do not show these increases in cortisol across the day in school or other group care settings (for review see Gunnar and Quevedo, 2007). This developmental change may reflect the improved social skills of older

children which facilitate successful peer interactions, or improved LHPA self-regulatory capacities, or some combination of these factors. While increases in cortisol across the child-care day have been observed in a number of studies, it is often challenging to find stressors that reliably increase cortisol in laboratorybased studies. As noted, the problem of identifying ethical and effective laboratory stressors emerges around 12 to 18 months, at around the period when an adult with whom the child has a secure relationship history or even a sensitive, responsive unfamiliar adult can buffer reactivity of this neuroendocrine system. Researchers have employed a variety of stressors in studies with preschool and school-aged children, but with little success in provoking a mean or average increase in cortisol. While some of these putative stressors, such as exposure to challenging cognitive tasks, would not be effective with adults either (see review, Dickerson and Kemeny, 2004), others, such as the threat of speaking publicly about one’s most embarrassing moment, fit criteria for effective stressors among adult subjects (L. Schmidt et al., 1999). This group difference between children and adults in LHPA reactivity to laboratory tasks raises the question of what accounts for the transition from child to adult reactivity patterns. Several researchers have suggested that with puberty the LHPA system becomes more responsive, and that this heightened responsiveness may help explain the rise in prevalence of affective disorders around midpuberty (Spear, 2000; Walker, Walder, and Reynolds, 2001). When one takes a longer developmental perspective, including infancy, early and middle childhood, and adolescence, an alternative but not inconsistent hypothesis is that as children move into adolescence, the period of relative stress hyporesponsivity of the LHPA system in humans slowly draws to a close. The association of puberty with heightened LHPA responsivity may reflect maturational changes in the biology of the LHPA axis, experience-driven alterations in how the adolescent perceives and processes threatening situations, or an interaction between biological maturation and experience. There is increasing evidence that reactivity of the LHPA system and its relation to social experience change during the transition to adolescence, likely in connection with pubertal changes. Basal cortisol levels, particularly morning levels, increase from ages 6 to 17 (Kiess et al., 1995; Legro et al., 2003; Netherton et al., 2004; Shirtcliff et al., in press). Two studies have found that the timing of this increase in morning levels is linked to pubertal status, occurring around Tanner stage 3 (Halligan et al., 2004; Netherton et al., 2004). Gonadal steroids influence the LHPA system, and animal models indicate that estrogen stimulates LHPA activity (Netherton et al., 2004). Thus it is not surprising that sex differences in LHPA functioning also emerge around Tanner stage 3. In one study, girls who had reached midpuberty had

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

73

higher cortisol morning levels than either midpubertal boys or prepubescent girls and boys (Netherton et al., 2004). Several studies have also noted that cortisol reactivity to laboratory stressors increases with age and pubertal status (Klimes-Dougan et al., 2001; Walker, Walder, and Reynolds, 2001; Wewerka et al., 2007). Notably, in the one study that employed the Trier Social Stress Test (the stressor shown to be most capable of elevating cortisol in studies of adults, see Dickerson and Kemeny, 2004), 9-, 11- and prepubertal 13-year-olds failed to exhibit increases in cortisol to this stressor task, although they all reported being stressed or anxious during testing. In contrast, both pubertal 13-yearolds and 15-year-olds showed significant elevations in the range typically noted in studies of adults (Wewerka et al., 2007). These changes in the LHPA axis in human children parallel changes seen in rodents at the end of the relative stress hyporesponsive period (Vazquez, 1998). At this point, it is not clear whether these changes reflect alterations in the axis and its regulation by cortico-limbic circuits or alterations in psychosocial regulation of the axis, that is, a decrease in the capacity of parental presence and availability to buffer its activation. Both psychological and physiological changes associated with the transition to adolescence may be involved. Furthermore, it is unlikely that these changes involve only the LHPA system. Other systems involved in reactivity and regulation of stress may also undergo developmental changes around the same period of time (see Gunnar and Quevedo, 2007). For example, there are well-known changes in sleep around the pubertal transition that may also increase vulnerability to stressors by reducing the child’s ability to use sleep as a stress regulator (Dahl and Lewin, 2002). Thus, at the group level, we have some understanding of the relation between social experience and LHPA development. The LHPA system is under strong social regulation in the first years of life, a fact which can be considered a double-edged sword. Sensitive, responsive care buffers LHPA reactivity, but inadequate care leaves young children vulnerable to LHPA dysregulation. Elevated levels of cortisol and CRH can have long-term detrimental effects on developing neural circuits, including increased stress vulnerability. However, at the individual level, LHPA outcomes of exposure to early social adversity are heterogeneous and difficult to predict. In the final section of the chapter, we explore individual factors that may interact with experience in shaping LHPA development.

Individual differences in LHPA function: Temperament and genes In considering individual variability, it can be useful to employ the developmental psychopathology concepts of multifinality—that is, multiple outcomes from the same starting point—and equifinality, multiple paths to the same

74

end point (see Cicchetti and Tucker, 1994, for further discussion of these concepts). For instance, in different individuals, exposure to chronic, severe early social adversity may lead to LHPA hyperreactivity, LHPA hyporeactivity, or apparently normal LHPA function (multifinality). Conversely, a postinstitutionalized child several years after adoption may have LHPA function that closely resembles that of a child who has experienced low social adversity throughout life, but these two children may have arrived at that outcome by very different pathways (equifinality). Thus, while early social experience exerts a profound influence on the developmental path of the LHPA system, that influence is filtered through the lenses of an individual’s genetic predispositions and current neurodevelopmental status. The developmental implications of social adversity also depend in part on the broader context of concurrent and subsequent social experiences. To explore individual differences in LHPA development, we will likely need to learn a great deal more about how genetic predispositions and social experience interact at different points in development. A starting point in this analysis may be to consider interactions between stable individual differences in temperament and children’s social experience in predicting differences in basal and stress responses of the LHPA axis at different points in development. Here we use behavioral inhibition as an example. Behavioral inhibition has received extensive research attention because the 5–10 percent of children who fall in this temperamental category are at increased risk of developing internalizing disorders (Kagan, Reznick, and Snidman, 1987). Behavioral inhibition—characterized by acute shyness, avoidance of novelty, fearfulness, and vigilance— demonstrates moderate temporal stability and a significant heritable component. As further evidence of the biological basis of behavioral inhibition, extremely inhibited children exhibit elevated heart rate, low vagal tone, and greater relative right frontal EEG activation (Fox et al., 2001; Kagan, Reznick, and Snidman, 1988). This physiological profile is consistent with elevated amygdala activity: The amygdala activates the sympathetic nervous system (raising heart rate), inhibits the parasympathetic nervous system (lowering vagal tone), and sends projections to the right frontal cortex. It has been hypothesized that in inhibited children, LHPA hyperactivity increases CRH activity in the amygdala, which would lower the threshold for perception of threat and result in more frequent, prolonged LHPA stress responses, paving the way to development of an anxiety disorder (Rosen and Schulkin, 1998). The body of evidence on this topic indicates that behavioral inhibition interacts with social context in predicting activity of the LHPA axis. For instance, there are no differences in cortisol levels for behaviorally inhibited and noninhibited children in the presence of a sensitive, responsive child-care provider, but if the child-care provider is not

fundamentals of developmental neurobiology

sensitive and responsive, more temperamentally inhibited children do exhibit higher cortisol levels (Dettling et al., 2000). Similarly, while exposure to a frightening toy does not provoke cortisol increases in securely attached infants or in bold infants regardless of attachment status, the combination of insecure attachment and behavioral inhibition is associated with an elevation in cortisol (Nachmias et al., 1996). Generally, main effects of temperament are not observed in the proximate LHPA response to novel stressors, such as laboratory manipulations, entrance into a new playgroup, or adjusting to a new school year (reviewed in Gunnar and Vazquez, 2006). In these situations, a rise in cortisol may be considered an adaptive response to provide resources to cope with the challenge. However, early in a new school year, inhibited temperament predicts higher home levels of cortisol (Bruce et al., 2002; E. Davis et al., 1999; de Haan et al., 1998), suggesting difficulty in terminating the LHPA response following removal from the novel context. Inhibited children also have elevated basal cortisol levels, especially levels obtained early in the morning (Buss et al., 2003; Kagan, Reznick, and Snidman, 1987; L. Schmidt and Fox, 1998). Inhibited children may be less able to regulate LHPA responses to familiar stressors (for review see Gunnar and Vazquez, 2006). This pattern would be consistent with studies of adults conducted by van Eck and colleagues (1996a, 1996b), in which highly anxious adults were more likely to show a cortisol response to familiar daily hassles, but did not differ from other adults in response to a novel laboratory stressor. The interaction of behavioral inhibition with social context in predicting LHPA activity cautions against expecting that genetic predispositions will exert main effects on activity of the LHPA system. Rather, we might expect that individuals with particular genetic polymorphisms may be at risk for altered LHPA activity under certain conditions and given certain rearing histories. The question is, Which genetic polymorphisms may be relevant? Here we will review several that are receiving attention in research on stress and stress vulnerability: genes involved in regulation of brain-derived neurotrophic factor (BDNF), those involved in regulation of serotonin (i.e., the serotonin transporter gene), and several genes that operate within the LHPA system. We do not intend this discussion to be exhaustive, but merely to point to the type of gene research that will likely alter the nature of developmental research on stress and stress vulnerability in the near future. BDNF is a protein involved in neural plasticity and produced in high levels in the hippocampus and cortex. Neural activity, such as that involved in the encoding of new information, increases gene transcription of BDNF (Hartmann and Lessmann, 2001). BDNF increases neuronal excitability and synaptic transmission, and thus plays a critical role in facilitating neural plasticity (Alder et al., 2003). Chronic

exposure to corticosterone decreases BDNF expression in the rodent hippocampus, and its reduced expression may contribute to dendritic atrophy (Smith et al., 1995a, 1995b). Thus glucocorticoid-induced decreases in hippocampal BDNF have been related to stress-induced memory deficits. Additionally, BDNF down-regulation has been linked to the pathogenesis of depression, and some antidepressants ameliorate stress-induced reductions of hippocampal BDNF (Duman, 2002; Shirayama et al., 2002). In humans, a common single nucleotide polymorphism (SNP) in the BDNF gene has been identified that results in a valine (Val) to methionine (Met) substitution. The Met allele appears to reduce the efficiency of BDNF regulation (less trafficking), with impacts on hippocampal and cortical morphology (Cheng et al., 2005). Theoretically, the lowfrequency Val/Met and Met/Met genotypes should be associated with greater vulnerability to chronic stress than the predominant Val/Val genotype. This hypothesis is consistent with evidence that bipolar disorders, substance abuse problems, and mood disorders—all disorders that are increased in probability by histories of stress during development—appear to be associated with this BDNF polymorphism (Cheng et al., 2005; Green et al., 2006; Parsian et al., 2004; Rybakowski et al., 2006; Tsai et al., 2006). To our knowledge, there have been no studies of the effects of glucocorticoid infusions in human adults with the Val/Met genotype versus the more efficient BDNF allele. There have been studies using mice with homo- and heterozygous knockouts of the BDNF gene. On the whole, there is little evidence that BDNF knockouts are more stress reactive (e.g., Chourbaji et al., 2004). In contrast, there is evidence that BDNF knockout mice that also lack the serotonin transporter gene are hyperfearful and exhibit exaggerated increases in corticosterone to processive stressors (RenPatterson et al., 2005). While it is difficult to extrapolate from knockout mice studies to human development, this latter study does suggest that both the serotonin system and neurotrophic factors may be important in regulating the development of stress reactivity and regulation. Indeed, there is increasing evidence that genetic polymorphisms affecting the efficiency of serotonin regulation do influence reactivity and regulation of the LHPA system and moderate the developmental impact of adverse experiences. In both human and nonhuman primates, a functional polymorphism has been identified in the length of the serotonin transporter (5HTT) gene (Bennett et al., 2002). Its short version is associated with decreased serotonergic function (Holden, 2003). In nonhuman primates, parental deprivation in the form of peer rearing as compared to mother rearing has been shown to reduce cerebrospinal levels of serotonin. However, these effects were expressed only in those peer-reared animals with at least one short copy of the serotonin transporter gene (Bennett et al., 2002). Likewise,

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

75

relative to mother-reared monkeys, peer-reared monkeys exhibit greater increases in LHPA reactivity to the stress of separation and single-case housing for several days. Again, this effect is exacerbated in monkeys carrying at least one short serotonin transporter allele (Barr et al., 2004). Importantly, cortisol and corticosterone are known to affect serotonin regulation, and there is growing evidence that the effects of elevated glucocorticoids operate, at least in part, through promoting synthesis of the serotonin transporter (Tafet, Toister-Achitiv, and Shinitzky, 2001). These findings may partially account for a report by Caspi and colleagues (2003) that the increase in risk of depression among individuals who were maltreated as children was moderated by the serotonin transporter polymorphism. Risk of depression by age 25 increased among those who were maltreated as children, but the increase in risk depended on whether the individual carried one or two short copies of the serotonin transporter gene. As noted, children who have been severely maltreated may exhibit chronic increases in cortisol, which in turn may influence serotonin regulation and expression of the serotonin transporter (Cicchetti and Toth, 2005). These effects may be amplified in children with one or two short copies of the serotonin transporter gene, significantly increasing their risk for developing affective disorders (Caspi et al., 2003). As noted earlier, in knockout mice models, impairment in both serotonin and BDNF regulation interacted to enhance stress reactivity and fearful behavior (RenPatterson et al., 2005). This statement may also be true for human development; in at least one study, depressive symptoms were most evident in children who carried both the MET allele of the BDNF polymorphism and the short copy of the serotonin transporter gene. However, the impact of these two genes in combination increased depressive symptoms in maltreated but not in nonmaltreated children (Kaufman et al., 2006). Unfortunately, none of the studies examining gene-by-experience interactions for maltreated children included measures of LHPA activity, so we can only speculate that activity of this system may have helped to mediate the gene-by-experience impacts on the development of depression. Another class of genetic polymorphisms that may play a role in individual differences in stress reactivity and the impact of early stressful life events involves variations in genes directly involved in activity of the LHPA system. Several polymorphisms have been described for the GR gene, including the BclI and the ER22/23EK. The BclI polymorphism has been associated with hypersensitivity to glucocorticoids and consists of a single nucleotide change (C to G) in the GR gene, the effects of which are not yet fully understood (van Rossum et al., 2006). The ER22/23EK polymorphism is related to glucocorticoid resistance and consists of two linked nucleotide changes resulting in GR receptors that have less affinity for cortisol. Recent research

76

has found that homozygous carriers of both these polymorphisms had an increased risk of developing a major depressive episode, but the ER22/23EK carriers had a faster clinical response to antidepressants (van Rossum et al., 2006). Several polymorphisms in the CRH gene have also been identified, and at least one recent study associated variations in the region of the CRH gene with panic disorder in adults and behavioral inhibition in children (Smoller et al., 2005). These are only a few of the genetic polymorphisms that are currently being explored in LHPA-axisrelevant genes. Results to date, however, suggest that many of these polymorphisms will affect the development and regulation of the LHPA system, and thus children’s vulnerability to stressors. As repeatedly noted in this section on individual differences, neither temperament nor genetic variations appear to have deterministic effects on development. In many cases, their effects are moderated by the care and experiences children have during childhood. Some of these genetic differences may affect reactivity and regulation of the LHPA system, while others may moderate the impact of LHPA reactivity on other neural systems. Regardless, their impact on the brain likely depends on whether the child experiences significant stressors during periods of rapid brain development. Introducing genetics and candidate gene analyses into the study of stress and development may help us understand the development of individual differences in vulnerability to stress, but it is not likely to reduce the importance of research on childhood experiences. Rather, it may provide new avenues through which to comprehend the role of experiences in shaping the development of individual differences in stress vulnerability and resilience.

Conclusions It is well established that stress has considerable influence on the developing brain. Stressful life events including abuse, neglect, and parental loss are associated with emotional disorders and lower academic achievement. In the last several decades, researchers studying child development have increasingly turned to measuring cortisol as a means of examining the potential role of the LHPA system in mediating these impacts and in helping to explain individual differences in stress vulnerability and resilience. Notwithstanding the ease of introducing measures of salivary cortisol into studies in child development, the complexity of the neurophysiology and developmental psychobiology of the LHPA system belies simple interpretation of cortisol findings. Nevertheless, based on the accumulation of both animal and human studies, there is ample reason to expect that this neuroendocrine system plays important roles in brain development and functioning and in mediating the impact of adverse experiences on the developing brain. In addition,

fundamentals of developmental neurobiology

adverse care during development appears to shape reactivity and regulation of this system in humans in ways that complement those observed in more detailed and controlled animal studies. Individual differences abound, both in vulnerability to adversity and in reactivity of the LHPA system. Genetic differences among individuals may impact LHPA activity and interact with effects of glucocorticoid- and CRH-induced effects on brain targets. Increased understanding of this genetic variability will likely help explain why some children are more stress resilient than others. However, it is unlikely that the growing infusion of gene analyses into our studies of stress reactivity, regulation, and impacts will relegate the role of experience to the background. Rather, the inclusion of molecular genetic information may provide us with a better understanding of the processes through which experiences impact stress vulnerability and resilience during development. acknowledgments

Preparation of this manuscript was supported by a National Science Foundation predoctoral fellowship to the first author, a National Institute of Mental Health predoctoral fellowship (T32 MH15755) to the second author, and a National Institute of Mental Health Senior Scientist Award (K05 MH66208) to the third author. Correspondence regarding this article should be addressed to Megan R. Gunnar, University of Minnesota Institute of Child Development, 51 E. River Rd, Minneapolis, MN 55455. E-mail: [email protected].

REFERENCES Ahnert, L., M. Gunnar, M. Lamb, and M. Barthel, 2004. Transition to child care: Associations with infant-mother attachment, infant negative emotion and cortisol elevations. Child Dev. 75:639–650. Alder, J., S. Thakker-Varia, D. A. Bangasser, M. Kuroiwa, M. R. Plummer, T. J. Shors, et al., 2003. Brain-derived neurotrophic factor–induced gene expression reveals novel actions of VGF in hippocampal synaptic plasticity. J. Neurosci. 23(34): 10800–10808. Bale, T. L., and W. W. Vale, 2004. CRF and CRF receptors: Role in stress responsivity and other behaviors. Annu. Rev. Pharmacol. Toxicol. 44:525–527. Baram, T. Z., D. T. Chalmers, C. Chen, Y. Koutsoukos, and E. B. De Souza, 1997. The CRF1 receptor mediates the excitatory actions of corticotropin releasing factor (CRF) in the developing rat brain: In vivo evidence using a novel, selective, non-peptide CRF receptor antagonist. Brain. Res. 770(1–2):89–95. Baram, T. Z., S. Yi, S. Avishai-Eliner, and L. Schultz, 2001. Developmental neurobiology of the stress response: Multilevel regulation of corticotropin-releasing hormone function. Mol. Psychiatry 6(6):647–656. Barr, C. S., T. K. Newman, C., Shannon, C. Parker, R. L. Dvoskin, M. L. Becker, et al., 2004. Rearing condition and rh5-HTTLPR interact to influence limbic-hypothalamicpituitary-adrenal axis response to stress in infant macaques. Biol. Psychiatry 55:733–738.

Bennett, A. J., K. P. Lesch, A. Heils, J. C. Long, J. G. Lorenz, and S. E. Shoal, 2002. Early experience and serotonin transporter gene variation interact to influence primate CNS function. Mol. Psychiatry 7:118–122. Blank, T., I. Nijholt, K. Eckart, and J. Spiess, 2002. Priming of long-term potentiation in mouse hippocampus by corticotropinreleasing factor and acute stress: Implications for hippocampusdependent learning. J. Neurosci. 22:3788–3794. Boyce, W. T., M. Champoux, S. J. Suomi, and M. R. Gunnar, 1995. Salivary cortisol in nursery reared rhesus monkeys: Reactivity to peer interactions and altered circadian activity. Dev. Psychobiol. 28:257–267. Bradbury, M. J., S. F. Akana, and M. F. Dallman, 1994. Roles of type I and II corticosteroid receptors in regulation of basal activity in the hypothalamopituitary-adrenal axis during the diurnal trough and the peak: Evidence for a nonadditive effect of combined receptor occupation. Endocrinology 134(3): 1286–1296. Bruce, J., E. P. Davis, and M. Gunnar, 2002. Individual differences in children’s cortisol response to the beginning of a new school year. Psychoneuroendocrinology 27:635–650. Bruce, J., P. A. Fisher, K. C. Pears, and S. Levine, under review. Morning cortisol levels in preschool-aged foster children: Differential effects of maltreatment type. Bruce, J., M. Kroupina, S. Parker, M. Gunnar, 2000. The relationships between cortisol patterns, growth retardation, and developmental delays in postinstitutionalized children. Presented at International Conference on Infant Studies, Brighton, England. Brunson, K. L., M. Eghbal-Ahmadi, R. Bender, Y. Chen, and T. Z. Baram, 2001. Long-term, progressive hippocampal cell loss and dysfunction induced by early-life administration of corticotropin-releasing hormone reproduce the effects of early-life stress. Proc. Natl. Acad. Sci. USA 98(15):8856–8861. Buss, K. A., J. R. M. Schumacher, I. Dolski, N. H. Kalin, H. H. Goldsmith, and R. J. Davidson, 2003. Right frontal brain activity, cortisol, and withdrawal behavior in 6-month-old infants. Behav. Neurosci. 117:11–20. Cameron, H., C. Wooley, and E. Gould, 1993. Adrenal steroid receptor immunoreactivity in cells born in the adult rat dentate gyrus. Brain Res. 611:342–346. Carlson, M., and F. Earls, 1997. Psychological and neuroendocrinological sequelae of early social deprivation in institutionalized children in Romania. Ann. NY Acad. Sci. 807:419–428. Carrion, V., C. Weems, R. Ray, B. Glaser, D. Hessl, and A. Reiss, 2002. Diurnal salivary cortisol in pediatric posttraumatic stress disorder. Biol. Psychiatry 51:575–582. Caspi, A., K. Sugden, T. Moffit, A. Taylor, I. W. Craig, H. Harrington, et al., 2003. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301:386–389. Champoux, M., L. Coe, S. Schanberg, C. Kuhn, and S. Suomi, 1989. Hormonal effects of early rearing conditions in the infant rhesus monkey. Am. J. Primatol. 19:111–117. Chen, Y., R. A. Bender, M. Frotscher, and T. Z. Baram, 2001. Novel and transient populations of corticotropin-releasing hormone-expressing neurons in developing hippocampus suggest unique functional roles: A quantitative spatiotemporal analysis. J. Neurosci. 21:7171–7181. Cheng, C.-Y., C.-J. Hong, Y. W. Yu, T.-J. Chen, H.-C. Wu, and S.-J. Tsai, 2005. Brain-derived neurotrophic factor (val66met) genetic polymorphism is associated with substance abuse in males. Mol. Brain Res. 140:86–90.

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

77

Chourbaji, S., R. Hellweg, D. Brandis, B. Zorner, C. Zacher, U. E. Lang, et al., 2004. Mice with reduced brain-derived neurotrophic factor expression show decreased choline acetyltransferase activity, but regular brain monoamine levels and unaltered emotional behavior. Mol. Brain Res. 121:28–36. Cicchetti, D., and D. Cohen, eds., 2006. Developmental Psychopathology: Risk, Disorder, and Adaptation, 2nd ed., vol. 3. Hoboken, NJ: Wiley. Cicchetti, D., and F. Rogosch, 2001a. Diverse patterns of neuroendocrine activity in maltreated children. Dev. Psychopathol. 13:677–720. Cicchetti, D., and F. Rogosch, 2001b. The impact of child maltreatment and psychopathology on neuroendocrine functioning. Dev. Psychopathol. 13:783–804. Cicchetti, D., and S. L. Toth, 2005. Child maltreatment. Annu. Rev. Clin. Psychol. 1(1):409–438. Cicchetti, D., and D. Tucker, 1994. Development and selfregulatory structures of the mind. Dev. Psychopathol. 6:533–549. Cintra, A., M. Zoli, J. Rosen, J. Agnati, S. Okret, A. Wikstrom, J. Gustaffsson, and K. Fuxe, 1994. Mapping and computer assisted morphometry and microdensitometry of glucocorticoid receptor immunoreactive neurons and glial cells in the rat central nervous system. Neuroscience 62:843–897. Coplan, J. D., M. W. Andrews, L. A. Rosenblum, M. J. Owens, S. Friedman, J. M. Gorman, et al., 1996. Persistent elevations of cerebrospinal fluid concentrations of corticotropin-releasing factor in adult nonhuman primates exposed to early-life stressors: Implications for the pathophysiology of mood and anxiety disorders. Proc. Natl. Acad. Sci. USA 93:1619–1623. Dahl, R. E., and D. S. Lewin, 2002. Pathways to adolescent health: Sleep regulation and behavior. J. Adolesc. Health 31(Suppl 6):175–184. Dallman, M., S. Akana, K. Scribner, M. Bradbury, C. Walker, A. Strack, and C. Cascio, 1992. Stress, feedback and facilitation in the hypothalamo-pituitary-adrenal axis. J. Neuroendocrinol. 4:517–526. Daly, J. R., and J. I. Evans, 1974. Daily rhythms of steroid and associated pituitary hormones in man and their relationship to sleep. In Advances in Steroid Biochemistry and Pharmacology, ed. M. H. Biggs and G. A. Christie, 61–109. New York: Academic Press. Dautzenberg, F. M., and R. L. Hauger, 2002. The CRF peptide family and their receptors: Yet more partners discovered. Trends Pharmacol. Sci. 23:71. Davis, E., B. Donzella, W. K. Krueger, and M. Gunnar, 1999. The start of a new school year: Individual differences in salivary cortisol response in relation to child temperament. Dev. Psychobiol. 35:188–196. Davis, M., 1997. Neurobiology of fear responses: The role of the amygdala. J. Neuropsychiatry Clin. Neurosci. 9:382–402. Davis, M., W. A. Falls, S. Campeau, and M. Kim, 1993. Fearpotentiated startle: A neural and pharmacological analysis. Behav. Brain Res. 58:175–198. De Bellis, M., A. Baum, B. Birmaher, M. Keshavan, C. Eccard, A. Boring, F. Jenkins, and N. Ryan, 1999. Developmental traumatology, part 1: Biological stress systems. Biol. Psychiatry 45:1259–1270. De Bellis, M., G. Chrousos, L. Dorn, L. Burke, K. Helmers, M. Kling, P. Trickett, and F. Putman, 1994. Hypothalamicpituitary-adrenal axis dysregulation in sexually abused girls. J. Clin. Endocrinol. Metab. 78:249–255. de Haan, M., M. Gunnar, K. Tout, J. Hart, and K. Stansbury, 1998. Familiar and novel context yield different associations

78

between cortisol and behavior among 2-year-olds. Dev. Psychobiol. 31:93–101. de Kloet, E. R., 1991. Brain corticosteroid receptor balance and homeostatic control. Front. Neuroendocrinol. 12:95–164. de Kloet, E. R., 2004. Hormones and the stressed brain. Ann. NY Acad. Sci. 1018:1–5. de Kloet, E. R., M. Oitzl, and M. Joels, 1993. Functional implications of brain corticosteroid receptor diversity. Cell. Mol. Neurobiol. 13:433–455. de Wied, D., and J. Jolles, 1982. Neuropeptides derived from pro-opiocortin: Behavioral, physiological, and neurochemical effects. Physiol. Rev. 62:977–1059. Dettling, A., M. Gunnar, and B. Donzella, 1999. Cortisol levels of young children in full-day childcare centers: Relations with age and temperament. Psychoneuroendocrinology 24:505–518. Dettling, A., S. W. Parker, S. Lane, A. Sebanc, and M. R. Gunnar, 2000. Quality of care and temperament determine whether cortisol levels rise over the day for children in full-day childcare. Psychoneuroendocrinology 25:819–836. Dickerson, S. S., and M. E. Kemeny, 2004. Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research. Psychol. Bull. 130:355–391. Dozier, M., E. Peloso, M. K. Gordon, M. Manni, M. R. Gunnar, K. C. Stovall-McClough, et al., 2006. Foster children’s diurnal production of cortisol: An exploratory study. Child Maltreatment 11:189–197. Duman, R., 2002. Pathophysiology of depression: The concept of synaptic plasticity. Eur. Psychiatry 17(Suppl 3):306s–310s. Duman, R. S., J. Malberg, and J. Thome, 1999. Neural plasticity to stress and antidepressant treatment. Biol. Psychiatry 46:1181– 1191. Fisher, P. A., M. R. Gunnar, P. Chamberlain, and J. B. Reid, 2000. Preventive intervention for maltreated preschool children: Impact on children’s behavior, neuroendocrine activity, and foster parent functioning. J. Am. Acad. Child Adolesc. Psychiatry 39:1356–1364. Fox, N. A., H. A. Henderson, K. H. Rubin, S. D. Calkins, and L. A. Schmidt, 2001. Continuity and discontinuity of behavioral inhibition and exuberance: Psychophysiological and behavioral influences across the first four years of life. Child Dev. 72:1–21. Francis, D., J. Diorio, D. Liu, and M. Meaney, 1999. Nongenomic transmission across generations of maternal behavior and stress responses in the rat. Science 286:1155–1158. Friese, E., J. Hesse, J. Hellhammer, and D. Hellhammer, 2005. A new view on hypocortisolism. Psychoneuroendocrinology 30: 1010–1016. Fuchs, E., H. Uno, and G. Flugge, 1995. Chronic psychosocial stress induces morphological alterations in hippocampal pyramidal neurons of the tree shrew. Brain Res. 673:275–282. Goldberg, S., R. Levitan, E. Leung, M. Masellis, V. Basile, C. Nemeroff, et al., 2003. Cortisol concentrations in 12- to 18month-old infants: Stability over time, location, and stressor. Biol. Psychiatry 54:719–726. Goodyer, I. M., J. Herbert, A. Tamplin, and P. Altham, 2000. First-episode major depression in adolescents: Affective, cognitive and endocrine characteristics of risk status and predictors of onset. Br. J. Psychiatry 176:142–149. Gould, E., and P. Tanapat, 1999. Stress and hippocampal neurogenesis. Biol. Psychiatry 46:1472–1479. Gould, E., C. Wooley, H. Cameron, D. Daniels, and B. McEwen, 1991. Adrenal steroids regulate postnatal development of the rat dentate gyrus: II. Effects of glucocorticoids and mineralocorticoids on cell birth. J. Comp. Neurol. 313:486–493.

fundamentals of developmental neurobiology

Green, E. K., R. Raybould, S. MacGregor, S. Hyde, A. H. Young, M. C. O’Donovan, et al., 2006. Genetic variation of brain-derived neurotrophic factor (BDNF) in bipolar disorder: Case-control study of over 3000 individuals from the UK. Br. J. Psychiatry 188:21–25. Gunnar, M., 1992. Reactivity of the hypothalamic-pituitaryadrenocortical system to stressors in normal infants and children. Pediatrics 80:491–497. Gunnar, M., L. Brodersen, K. Krueger, and J. Rigatuso, 1996. Dampening of adrenocortical responses during infancy: Normative changes and individual differences. Child Dev. 67:877–889. Gunnar, M., L. Brodersen, M. Nachmias, K. Buss, and R. Rigatuso, 1996. Stress reactivity and attachment security. Dev. Psychobiol. 29:10–36. Gunnar, M., and E. Davis, 2003. Stress and emotion in early childhood. In Handbook of Psychology: Developmental Psychology, vol. 6, ed. R. Lerner, M. A. Easterbrooks, and J. Mistry, 113–134. Hoboken, NJ: Wiley. Gunnar, M., and B. Donzella, 2001. Social regulation of the LHPA axis in early human development. Psychoneuroendocrinology 27:199–220. Gunnar, M., P. Fisher, and the Early Experience, Stress, and Prevention Network, 2006. Bringing basic research on early experience and stress neurobiology to bear on preventive interventions for neglected and maltreated children. Dev. Psychopathol. 18:651–677. Gunnar, M., M. Larson, L. Hertsgaard, M. Harris, and L. Brodersen, 1992. The stressfulness of separation among 9month-old infants: Effects of social context variables and infant temperament. Child Dev. 63:290–303. Gunnar, M., S. Mangelsdorf, M. Larson, and L. Hertsgaard, 1989. Attachment, temperament and adrenocortical activity in infancy: A study of psychoendocrine regulation. Dev. Psychol. 25:355–363. Gunnar, M., and K. Quevedo, 2007. Developmental biology of stress. Annu. Rev. Psychol. 58:145–174. Gunnar, M. R., and N. M. Talge, 2007. Neuroendocrine measures in developmental research. In Developmental Psychophysiology: Theory, Systems, and Methods, ed. L. A. Schmidt and S. Segalowitz, 343–366. New York: Cambridge University Press. Gunnar, M., K. Tout, M. de Haan, S. Pierce, and K. Stansbury, 1997. Temperament, social competence, and adrenocortical activity in preschoolers. Dev. Psychobiol. 31:65–85. Gunnar, M., K. Tout, M. A. Donzella, and M. van Dulmen, 2003. Temperament, peer relationships, and cortisol activity in preschoolers. Dev. Psychbiol. 43:346–358. Gunnar, M., and D. Vazquez, 2006. Stress neurobiology and developmental psychopathology. In Developmental Psychopathology, 2nd ed., vol. 2, ed. D. Cicchetti and D. Cohen, 533–577. New York: Wiley. Gunnar, M. R., S. Wewerka, K. Frenn, J. D. Long, and C. Griggs, in press. Development changes in HPA activity over the transition to adolescence: Normative changes and associations with pubertal stage. Dev. Psychopathol. Halligan, S., J. Herbert, I. M. Goodyer, and L. Murray, 2004. Exposure to postnatal depression predicts elevated cortisol in adolescent offspring. Biol. Psychiatry 55:376–381. Hart, J., M. Gunnar, and D. Cicchetti, 1996. Altered neuroendocrine activity in maltreated children related to symptoms of depression. Dev. Psychopathol. 8:201–214. Hartmann M. H. R., and V. Lessmann, 2001. Synaptic secretion of BDNF after high-frequency stimulation of glutamatergic synapses. EMBO J. 20:5887–5897.

Heim, C., and C. Nemeroff, 2001. The role of childhood trauma in the neurobiology of mood and anxiety disorders: Preclinical and clinical studies. Biol. Psychiatry 49:1023–1039. Heim, C., D. Newport, R. Bonsall, A. Miller, and C. Nemeroff, 2001. Altered pituitary adrenal axis responses to provocative challenge tests in adult survivors of childhood abuse. Am. J. Psychiatry 158:575–581. Heim, C., D. J. Newport, S. Heit, Y. P. Graham, M. B. R. Wilcox, A. H. Miller, and C. Nemeroff, 2000. Pituitaryadrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. JAMA 284:592–597. Heim, C., P. Plotsky, and C. B. Nemeroff, 2004. The importance of studying the contributions of early adverse experiences to the neurobiological findings in depression. Neuropsychopharmacology 29:641–648. Heinrichs, S. C., and G. F. Koob, 2004. Corticotropin-releasing factor in brain: A role in activation, arousal, and affect regulation. J. Pharmacol. Exp. Ther. 311:427–440. Herman, J. P., and W. E. Cullinan, 1997. Neurocircuitry of stress: Central control of the hypothalamo pituitary adrenocortical axis. Trends Neurosci. 20:78–84. Herman, J. P, N. K. Muller, and H. Figueiredo, 2004. Role of GABA and glutamate circuitry in hypothalamo-pituitary adrenocortical stress integration. Ann. NY Acad. Sci. 1018:35–45. Herman, J. P., J. G. Tasker, D. R. Ziegler, and W. E. Cullinan, 2002. Local circuit regulation of paraventricular nucleus stress integration glutamate-GABA connections. Pharmacol. Biochem. Behav. 71:457–468. Hertsgaard, L., M. R. Gunnar, M. Erickson, and M. Nachmias, 1995. Adrenocortical responses to the strange situation in infants with disorganized/disoriented attachment relationships. Child Dev. 66:1100–1106. Holden, C., 2003. Getting the short end of the allele. Science, 301(5631):291–293. Huang, W., C. Harper, S. Evans, J. Newnham, and S. Dunlop, 2001. Repeated prenatal corticosteroid administration delays myelination of the corpus callosum in fetal sheep. Int. J. Dev. Neurosci. 19:415–425. Ingle, D., 1952. The role of the adrenal cortex in homeostasis. J. Endocrinol. 8:23–37. Kagan, J., J. S. Reznick, and N. Snidman, 1987. The physiology and psychology of behavioral inhibition in children. Child Dev. 58:1459–1473. Kagan, J., J. S. Reznick, and N. Snidman, 1988. Biological bases of childhood shyness. Science 240:167–171. Kaufman, J., 1991. Depressive disorders in maltreated children. J. Am. Acad. Child Adolesc. Psychiatry 30:257–265. Kaufman, J., B. Z. Yang, H. Douglas-Palumberi, D. Grasso, D. Lipschitz, S. Houshyar, et al., 2006. Brain-derived neurotrophic factor–5HTTLPR gene interactions and environmental modifiers of depression in children. Biol. Psychiatry 59:673–680. Kertes, D. A., M. Gunnar, N. J. Madsen, and J. Long, in press. Early deprivation and home basal cortisol levels: A study of internationally adopted children. Dev. Psychopathol. Kiess, W., R. A. Meidert, K. Dressensorfer, U. Schriever, A. Kessler, A. Konig, H. P. Schwarz, and C. J. Strasburger, 1995. Salivary cortisol levels throughout childhood and adolescence: Relation with age, pubertal stage, and weight. Pediatr. Res. 37:502–506. Kirschbaum, C., R. Steyer, M. Eid, U. Patalla, P. Schwenkmezger, and D. Hellhammer, 1990. Cortisol and behavior. 2. Application of a latent state-trait model to salivary cortisol. Psychoneuroendocrinology 15:297–307.

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

79

Klimes-Dougan, B., P. D. Hastings, D. A. Granger, B. A. Usher, and C. Zahn-Waxler, 2001. Adrenocortical activity in at-risk and normally developing adolescents: Individual differences in salivary cortisol basal levels, diurnal variation, and responses to social challenges. Dev. Psychol. 37:695–719. Kroupina, M., M. R. Gunnar, and D. E. Johnson, 1997. Report on salivary cortisol levels in a Russian baby home. Institute of Child Development, University of Minnesota, Minneapolis. Kwak, S., M. Morano, E. Young, S. Watson, and H. Akil, 1993. Diurnal CRH mRNA in the hypothalamus: Decreased expression in the evening is not dependent on endogenous glucocorticoids. Neuroendocrinology 57:96–105. Larson, M. C., M. R. Gunnar, and L. Hertsgaard, 1991. The effects of morning naps, car trips, and maternal separation on adrenocortical activity in human infants. Child Dev. 62: 362–372. Larson, M., B. P. White, A. Cochran, B. Donzella, and M. Gunnar, 1998. Dampening of the cortisol response to handling at 3-months in human infants and its relation to sleep, circadian cortisol activity, and behavioral distress. Dev. Psychobiol. 33: 327–337. LeDoux, J. E., and E. A. Phelps, 2000. Emotional networks in the brain. In Handbook of Emotions, 2nd ed., ed. M. Lewis and J. Haviland-Jones, 157–172. New York: Guilford. Legendre, A., and M. Trudel, 1996. Cortisol and behavioral responses of young children coping with a group of unfamiliar peers. Merrill-Palmer Q. 42:554–577. Legro, R. S., H. M. Lin, L. M. Demers, and T. Lloyd, 2003. Urinary free cortisol increases in adolescent Caucasian females during perimenarche. J. Clin. Endocrinol. Metab. 88:215–219. Levine, S., 2005. Stress: An historical perspective. In Handbook on Stress, Immunology and Behavior, ed. T. Steckler, N. Kalin, and J. M. H. M. Reul, 3–23. Amsterdam: Elsevier Science. Levine, S., M. Alpert, and G. Lewis, 1957. Infantile experience and the maturation of the pituitary adrenal axis. Science 126: 1347. Lewis, M., and D. Ramsay, 1995. Developmental change in infants’ responses to stress. Child Dev. 66:657–670. López, J. F., H. Akil, and S. J. Watson, 1999. Role of biological and psychological factors in early development and their impact on adult life: Neural circuits mediating stress. Biol. Psychiatry. 46:1461–1471. Lundberg, U., O. Westermark, and B. Rasch, 1993. Cardiovascular and neuroendocrine activity in preschool children: Comparison between day care and home levels. Scand. J. Psychol. 34:774–779. Main, M., and J. Solomon, 1990. Procedures for identifying infants as disorganized/disoriented during the Ainsworth strange situation. In Attachment in the Preschool Years: Theory, Research, and Intervention, ed. M. T. Greenberg, D. Cicchetti, and M. E. Cummings, 121–160. Chicago: University of Chicago Press. Makara, G. B., and J. Haller, 2001. Non-genomic effects of glucocorticoids in the neural system: Evidence, mechanisms and implications. Prog. Neurobiol. 65:367–390. Makino, S., T. Shibasaki, N. Yamauchi, T. Nishioka, T. Mimoto, I. Wakabayashi, et al., 1999. Psychological stress increased corticotropin-releasing hormone MRNA and content in the central nucleus of the amygdala but not in the hypothalamic paraventricular nucleus in the rat. Brain Res. 850:136–143. Matthews, S. G., 2000. Antenatal glucocorticoids and programming of the developing CNS. Pediatr. Res. 47:291–300. McCormack, K., A. Grand, J. LaPrairie, R. Fulks, A. Graff, D. Maestripieri, P. M. Plotsky, and M. M. Sanchez, 2003.

80

Behavioral and neuroendocrine outcomes of infant maltreatment in rhesus monkeys: The first four months. Soc. Neurosci. (Abs.). 641:14. McEwen, B., 1998. Stress, adaptation, and disease: Allostasis and allostatic load. Ann. NY Acad. Sci. 840:33–44. McEwen, B., 2005. Glucocorticoids, depression, and mood disorders: Structural remodeling in the brain. Metabolism 54(5 [Suppl 1]):20–23. Meaney, M. J., and M. Szyf, 2005. Maternal care as a model for experience-dependent chromatin plasticity? Trends Neurosci. 28: 456–463. Minamikawa, S., K. Kono, K. Nakayama, H. Yokote, T. Tashiro, A. Nishio, et al., 2004. Glucocorticoid treatment of brain tumor patients: Changes of apparent diffusion coefficient values measured by MR diffusion imaging. Neuroradiology 46:805– 811. Nachmias, M., M. Gunnar, S. Mangelsdorf, R. Parritz, and K. Buss, 1996. Behavioral inhibition and stress reactivity: Moderating role of attachment security. Child Dev. 67:508–522. Netherton, C. M., I. M. Goodyer, A. Tamplin, and J. Hertbert, 2004. Salivary cortisol and dehydroepiandrosterone in relation to puberty and gender. Psychoneuroendocrinology 29:125–140. Parsian, A., R. Sinha, B. Racette, J. H. Zhao, and J. S. Perlmutter, 2004. Association of a variation in the promoter region of the brain-derived neurotrophic factor gene with familial Parkinson’s disease. Parkinsonism Rel. Disord. 10:213–219. Patel, P. D., J. F. Lopez, D. M. Lyons, S. Burke, M. Wallace, and A. F. Schatzberg, 2000. Glucocorticoid and mineralocorticoid receptor mRNA expression in squirrel monkey brain. J. Psychiatr. Res. 34:383–392. Pryce, C. R., J. Feldon, E. Fuchs, I. Knuesel, T. Oertle, C. Sengstag, M. Spengler, E. Weber, A. Weston, and A. JongenRelo, 2005. Postnatal ontogeny of hippocampal expression of the mineralocorticoid and glucocorticoid receptors in the common marmoset monkey. Eur. J. Neurosci. 21:1521–1535. Rainnie, D. G., R. Bergeron, T. J. Sajdyk, M. Patil, D. R. Gehlert, and A. Shekhar, 2004. Corticotrophin releasing factor–induced synaptic plasticity in the amygdala translates stress into emotional disorders. J. Neurosci. 24:3471–3479. Ren-Patterson, R. F., L. W. Cochran, A. Holmes, S. Sherrill, S. J. Huang, T. Tolliver, et al., 2005. Loss of brain-derived neurotrophic factor gene allele exacerbates brain monoamine deficiencies and increases stress abnormalities of serotonin transporter knockout mice. J. Neurosci. Res. 15:756–771. Reul, J. M., and E. R. de Kloet, 1985. Two receptor systems for corticosterone receptors in rat brain: Microdistribution and differential occupation. Endocrinology 117:2505–2511. Rosen, J. B., and J. Schulkin, 1998. From normal fear to pathological anxiety. Psychol. Rev. 105:325–350. Rybakowski, J. K., A. Borkowska, M. Skibinska, A. Szczepankiewicz, P. Kapelski, A. Leszczynska-Rodziewicz, et al., 2006. Prefrontal cognition in schizophrenia and bipolar illness in relation to VAL66MET polymorphism of the brain-derived neurotrophic factor gene. Psychiatry Clin. Neurosci. 60(1):70–76. Sanchez, M., 2006. The impact of early adverse care on HPA axis development: Nonhuman primate models. Horm. Behav. 50:623 –631. Sanchez, M. M., C. O. Ladd, and P. M. Plotsky, 2001. Early adverse experience as a developmental risk factor for later psychopathology: Evidence from rodent and primate models. Dev. Psychopathol. 13:419–449. Sanchez, M. M., L. J., Young, P. M. Plotsky, and T. R. Insel, 2000. Distribution of corticosteroid receptors in the rhesus brain:

fundamentals of developmental neurobiology

Relative absence of glucocorticoid receptors in the hippocampal formation. J. Neurosci. 20:4657–4668. Sapolsky, R., 1997. McEwen-induced modulation of endocrine history: A partial review. Stress 2:1–12. Sapolsky, R., and M. Meaney, 1986. Maturation of the adrenocortical stress response: Neuroendocrine control mechanisms and the stress hyporesponsive period. Brain Res. 396:64– 76. Sapolsky, R., L. Romero, and A. Munck, 2000. How do glucocorticoids influence stress responses? Integrating permissive, stimulatory, and preparative actions. Endocr. Rev. 21:55–89. Schlessinger, A., W. Cowan, and D. Gottlieb, 1975. An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. Comp. Neurol. 159:149–175. Schmidt, L., and N. Fox, 1998. Fear-potentiated startle responses in temperamentally different human infants. Dev. Psychobiol. 32:113–120. Schmidt, L. A., N. A. Fox, E. M. Sternberg, P. W. Gold, C. C. Smith, and J. Schulkin, 1999. Adrenocortical reactivity and social competence in seven-year-olds. Personality and Individual Differences 26:977–985. Schmidt, M., L. Enthoven, M. van der Mark, S. Levine, E. R. de Kloet, and M. S. Oitzl, 2003. The postnatal development of the hypothalamic-pituitary-adrenal axis in the mouse. Int. J. Dev. Neurosci. 21:125–132. Seckl, J. R., M. Cleasby, and M. J. Nyirenda, 2000. Glucocorticoids, 11-beta-hydroxysteroid dehydrogenase, and fetal programming. Kidney Int. 57:1412–1417. Seckl, J. R., and M. J. Meaney, 2004. Glucocorticoid programming. Ann. NY Acad. Sci. 1032:63–84. Shekhar, A., W. Truitt, D. Rainnie, and T. Sajdyk, 2005. Role of stress, corticotrophin releasing factor (CRF) and amygdala plasticity in chronic anxiety. Stress 8:209–219. Shirayama, Y., A. C. Chen, S. Nakagawa, D. S. Russell, and R. S. Duman, 2002. Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J. Neurosci. 22:3251–3261. Shirtcliff, E., D. Granger, A. Booth, and D. Johnson, 2005. Low salivary cortisol levels and externalizing behavior problems in youth. Dev. Psychopathol. 17:167–184. Smith, M. A., S. Y. Kim, H. J. Van Oers, and S. Levine, 1997. Maternal deprivation and stress induce immediate early genes in the infant rat brain. Endocrinology 138(11):4622–4628. Smith, M. A., S. Makino, R. Kvetnansky, and R. M. Post, 1995a. Effects of stress on neurotrophic factor expression in the rat brain. In Stress: Basic Mechanisms and Clinical Implications, ed. G. Chrousos, R. McCarty, K. Pacak, G. Cizza, E. Sternberg, et al., 234–239. New York Academy of Sciences. Smith, M., S. Makino, R. Kvetnansky, and R. Post, 1995b. Stress and glucocorticoids affect the expression of brain derived neurotropic factors and neurotrophin-3 mRNA in the hippocampus. J. Neurosci. 15:1768–1777. Smoller, J. W., L. H. Yamaki, J. A. Fagerness, J. Biederman, S. Racette, N. M. Laird, et al., 2005. The corticotropin-releasing hormone gene and behavioral inhibition in children at risk for panic disorder. Biol. Psychiatry 57:1485–1492. Spangler, G., and K. E. Grossmann, 1993. Biobehavioral organization in securely and insecurely attached infants. Child Dev. 64:1439–1450. Spangler, G., and M. Schieche, 1998. Emotional and adrenocortical responses of infants to the strange situation: The differential

function of emotional expression. Int. J. Behav. Dev. 22:681– 706. Spear, L. P., 2000. The adolescent brain and age-related behavioral manifestations. Neurosci. Biobehav. Rev. 24:417– 463. Sroufe, L. A., 1983. Infant-caregiver attachment and patterns of adaptation in preschool: The roots of maladaptation and competence. In Development and Policy Concerning Children with Special Needs, Minnesota Symposia on Child Psychology, vol. 16, ed. M. Perlmutter, 41–83. Hillsdale, NJ: Erlbaum Associates. Stratakis, C. A., and G. P. Chrousos, 1995. Neuroendocrinology and pathophysiology of the stress system. In Stress: Basic Mechanisms and Clinical Implications, ed. G. P. Chrousos, R. McCarty, K. Pacak, G. Cizza, E. Sternberg, et al. New York Academy of Sciences. Suchecki, D., P. Rosenfeld, and S. Levine, 1993. Maternal regulation of the hypothalamic-pituitary-adrenal axis in the rat: The roles of feeding and stroking. Dev. Brain. Res. 75:185– 192. Tafet, G. E., M. Toister-Achituv, and M. Shinitzky, 2001. Enhancement of serotonin uptake by cortisol: A possible link between stress and depression. Cogn. Affective Behav. Neurosci. 1:96–104. Tanapat, P., L. Galea, and E. Gould, 1998. Stress inhibits the proliferation of granule cell precursors in the developing dentate gyrus. Int. J. Dev. Neurosci. 16:235–239. Tarullo, A., and M. Gunnar, 2006. Child maltreatment and the developing HPA axis. Horm. Behav. 50:632–639. Tottenham, N., S. W. Parker, and C. Liu, 2001. Individual differences in cardiac and HPA reactivity to a psychological stressor. Paper presented at the Society for Research in Child Development, Minneapolis, MN, April. Tout, K., M. de Haan, E. Kipp-Campbell, and M. Gunnar, 1998. Social behavior correlates of adrenocortical activity in daycare: Gender differences and time-of-day effects. Child Dev. 69:1247 –1262. Tsai, S.-J., C.-J. Hong, H.-C. Liu, T.-Y. Liu, and Y.-J. Liou, 2006. The brain-derived neurotrophic factor gene as a possible susceptibility candidate for Alzheimer’s disease in a Chinese population. Dement. Geriatr. Cogn. Disord. 21:139–143. Uno, H., L. Lohmiller, C. Thieme, J. W. Kemnitz, M. J. Engle, E. B. Roecker, et al., 1990. Brain damage induced by prenatal exposure to dexamethasone in fetal rhesus macaques. I. Hippocampus. Dev. Brain Res. 52:157–167. van Eck, M., H. Berkhof, N. Nicolson, and J. Sulon, 1996a. The effects of perceived stress, traits, mood states, and stressful daily events on salivary cortisol. Psychosom. Med. 58:447– 458. van Eck, M., N. Nicolson, H. Berkhof, and J. Sulon, 1996b. Individual differences in cortisol responses to a laboratory speech task and their relationships to responses to stressful daily events. Biol. Psychol. 43:69–84. van Goozen, S, W. Matthys, P. Cohen-Kettenis, J. Buittelaar, and H. van Engeland, 2000. Hypothalamic-pituitary-adrenal axis and autonomic nervous system activity in disruptive children and matched controls. J. Am. Acad. Child Adolesc. Psychiatry 39:1438–1445. van Ijzendoorn, H. W., C. Schuengel, and M. J. BakersmansKranenburg, 1999. Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants, and sequelae. Dev. Psychopathol. 11:225–249. van Rossum, E. F., E. B. Binder, M. Majer, J. W. Koper, M. Ising, S. Modell, et al., 2006. Polymorphisms of the glucocorticoid

tarullo, quevedo, and gunnar: lhpa system and neurobehavioral development

81

receptor gene and major depression. Biol. Psychiatry. 59:681 –688. Vazquez, D. M., 1998. Stress and the developing limbic-hypothalamic-pituitary-adrenal axis. Psychoneuroendocrinology 23:663–700. Vazquez, D. M., J. F. López, M. I. Morano, S. P. Kwak, S. J. Watson, and H. Akil, 1998. Alpha, Beta and Gamma mineralocorticoid receptor mRNA splice variants: Differential expression and rapid regulation in the developing hippocampus. Endocrinology 139:3165–3177. Vazquez, D. M., I. M. Morano, J. F. Lopez, S. J. Watson, and H. Akil, 1993. Short-term adrenalectomy increases glucocorticoid and mineralocorticoid receptor mRNA in selective areas of the developing hippocampus. Mol. Cell Neurosci. 4:455–471. Vermetten, E., and J. Bremner, 2002. Circuits and systems in stress. I. Preclinical studies. Depress. Anxiety 15:126–147. Walker, E. F., D. J. Walder, and F. Reynolds, 2001. Developmental changes in cortisol secretion in normal and at risk youth. Dev. Psychopathol. 13:721–732. Watamura, S., B. Donzella, J. Alwin, and M. Gunnar, 2003. Morning to afternoon increases in cortisol concentrations for infants and toddlers at child care: Age differences and behavioral correlates. Child Dev. 74:1006–1020.

82

Watamura, S., B. Donzella, D. Kertes, and M. Gunnar, 2004. Developmental changes in baseline cortisol activity in early childhood: Relations with napping and effortful control. Dev. Psychobiol. 45:125–133. Watanabe, Y., E. Gould, and B. McEwen, 1992. Stress induces atrophy of apical dendrites of hippocampal CA3 pyramidal neurons. Brain Res. 588:341–345. Welberg, L., J. Seckl, and M. Holmes, 2001. Prenatal glucocorticoid programming of brain corticosteroid receptors and corticotrophin-releasing hormone: Possible implications for behaviour. Neuroscience 104:71–79. Wewerka, S., K. Frenn, and M. R. Gunnar, March 2007. Developmental Changes to the Trier Social Stress Test in 9-, 11-, 13and 15-Year-Olds: Associations with Puberty. Paper presented at Society for Research in Child Development, Boston, MA. Whitelaw, A., and M. Thoresen, 2000. Antenatal steroids and the developing brain. Arch. Dis. Child. Fetal Neonatal Ed. 83:154– 157. Yehuda, R., S. L. Halligan, and R. Grossman, 2001. Childhood trauma and risk for PTSD: Relationship to intergenerational effects of trauma, parental PTSD, and cortisol excretion. Dev. Psychopathol. 13:733–753.

fundamentals of developmental neurobiology

6

The Effects of Monoamines on the Developing Nervous System GREGG D. STANWOOD AND PAT LEVITT

Introduction The development of the brain relies on the spatial and temporal regulation of cell-cell interactions that are controlled by contact-mediated and diffusible effector substances. Both the availability of molecular signals and the ability of developing cells to respond to those signals, by expression of specific receptors, are essential to ontogenesis. Transcription factors, cell adhesion and guidance molecules, and neurotrophic factors all have been established as mediators of tissue patterning, histotypic organization, and circuit formation. It is clear, however, that many molecules exhibit pleiotropic activities, serving as regulators of distinct cellular functions at different times in development and adulthood. Neurotransmitters and neuromodulators are now recognized as exhibiting multiple activities, performing very different roles in cellular communication in the mature brain and during development (Lauder, 1993; Levitt et al., 1997; Herlenius and Lagercrantz, 2004). Here, we focus on several features of brain development to understand the potential impact that an altered neurochemical environment may have on histogenesis. In this chapter we describe research findings that implicate monoamine systems in the regulation of neural development. These neuromodulators are particularly susceptible to modifications by exposure to drugs of abuse and psychotherapeutics during pre- and postnatal development.

Fundamentals of cerebral cortical development The cerebral cortex mediates higher cognitive functions and is responsible for the integration of complex sensory, motor, and homeostatic information. Defects in cortical development, therefore, can have a profound impact on mature brain functions. It has been suggested that developmental anomalies in cortical development underlie certain types of psychopathology, such as schizophrenia (Weinberger, 1995; D. Lewis and Levitt, 2002), and forms of mental retardation and autism (Charman, 1999; Levitt, Eagleson, and Powell, 2004). The molecular and cellular bases that tie developmental defects to cortical dysfunction in these disorders remain unknown, but we know that influences on cell-cell interactions that mediate specific developmental events are

likely targets. This also seems to be true for nongenetic alterations in development, such as prenatal exposure to toxicants, stress, or drugs of abuse (Trask and Kosofsky, 2000; Stanwood and Levitt, 2004; Andersen, 2005). The Basics of Brain Development Temporally overlapping events, grouped in five major categories, contribute to the formation of all brain structures from the neural tube. Progenitor cells of the germinal matrix, situated along the forerunner of the ventricular system, give rise to all neurons and macroglia in a well-controlled proliferative process. Recent evidence suggests that specialized radial neuroepithelial cells serve two purposes: precursor cells for both neurons and glia (Campbell and Gotz, 2002; Gotz and Huttner, 2005), and a radial scaffold to guide the directed migration of postmitotic neurons from their place of origin to their final resting position (Noctor et al., 2001; Kriegstein, 2005). Neuronal differentiation involves the expression of specific gene products that, together with the appearance of polarized structural features (axons and dendrites), contribute to the remarkable phenotypic diversity of the nervous system. Glial cells differentiate early to form specialized migration guides, radial glia, and later to form astrocytes and oligodendrocytes. Progenitor cells, neurons, and glia also undergo naturally occurring cell death, a complex process that appears to be a normal developmental mechanism to establish appropriate quantitative relationships between projection and target neurons and between neurons and glia. Synaptogenesis is a temporally extended developmental event, beginning prenatally and lasting through adolescence in the central nervous system (CNS) of all mammals, which is critical in the formation of synapses between specific target populations of neurons. These developmental milestones are described schematically in figure 6.1. Cerebral Cortical Histogenesis The neocortex is a sixlayered structure that exhibits very similar features across its tangential extent. Thus all functional areas have repeated laminar and columnar organization that is assembled during development in well-defined temporal and spatial patterns. Very similar temporally ordered patterns of cortical development occur in all mammalian species. In the rat, total gestation is approximately 21 days, with the peak of

83

Figure 6.1 Schematic representation of developmental milestones in the forebrain, depicting proliferation and migration, aggregation, circuit formation, and synaptogenesis. The approximate times of each event in the human brain are labeled on the left, and corresponding days in the rodent brain are on the right. Neurons and glia are produced from proliferative zones, and a variety of physical and chemical cues contribute to their migration to specific locations in the brain. Transient structures, such as the subplate in the cerebral cortex, provide cues that assist in the gen-

eration of topographic axon projections. Some elements of cell differentiation and synaptogenesis occur over protracted periods of time and peak after birth. Abbreviations: E, embryonic day; P, postnatal day; W, age in gestational weeks; M, age in months after birth; PZ, proliferative zone; MZ, marginal zone; Bf, basal forebrain; Th, thalamus; SP, subplate; CP, cortical plate; CC, corpus callosum; TC, thalamocortical projections; CTX, cerebral cortex. (Figure is modeled on Rakic, 1995.) (See plate 15.)

cortical neuronogenesis occurring at about embryonic day (E) 16. In the mouse, total gestation is about 19 days, and E15 corresponds to the peak of neuron proliferation in the cerebral cortex. In humans, neuron production, migration, and differentiation occur prenatally in the cerebral cortex, beginning by the end of the second month of gestation and peaking prior to midgestation. By the third trimester, basic organization and a minimum of connections are already formed, and myelination begins (figure 6.1 and plate 15; Levitt, 2003). The first neurons produced are actually not located in the forerunner of the cortex, the cortical plate, but rather form a structure called the preplate, which eventually is split into a subplate and a supraplate in the marginal zone by the first neurons destined for the cortical plate. The process of splitting the preplate is an important first step in establishing

appropriate migration patterns of neurons. In the mutant mouse reeler, in which expression of the reelin gene is defective, this process is abnormal and the cortex is disorganized. Subplate neurons are transient cells that serve as temporary targets for subcortical axons from the thalamus. These neurons aid in the guidance of axons to correct target regions of the cerebral cortex, and in the formation of precise sensory maps within specific layers of the cerebral cortex (Kanold et al., 2003). Neurons arising from the proliferative zone along the ventricular surface of the dorsal telencephalon (dorsal pallium) attain a bipolar morphology and migrate along radial glia to reach the cortical plate, with neurons born early residing deep and those born subsequently more superficial. This so-called inside-out settling pattern is a hallmark of cerebral cortical formation and allows one to define the

84

fundamentals of developmental neurobiology

time of origin of neurons residing in specific layers of the cortex. Migrating cells exhibit polarized features, with the trailing process of migrating cells generally forming the axon, which can be recognized as neurons migrate. These axons grow into the intermediate zone, the forerunner of cortical white matter, which resides between the subplate and the germinal matrix during development. Following the production of neurons destined for layers 6–4, a new, superficial zone of progenitor cells appears, termed the subventricular zone (SVZ). These progenitor cells maintain a position away from the luminal surface and produce the neurons of the superficial cortical layers and most of the macroglial cells (astrocytes and oligodendrocytes). The SVZ is maintained in adults in a restricted region along the rostro-ventral end of the lateral ventricle and can produce new neurons and glia in the mature forebrain. Recent studies have demonstrated that not all neurons of the cerebral cortex arise from the dorsal telencephalon. Rather, it appears that most of the GABA interneurons of the cerebral cortex are generated in the ganglionic eminence of the ventral telencephalon and migrate tangentially to reach all areas of the cerebral cortex (Anderson et al., 1997; Nadarajah and Parnavelas, 2002; Levitt, Eagleson, and Powell, 2004). These interneurons comprise 12–18 percent of all neurons in the cerebral cortex and exhibit many phenotypic properties of neurons as they migrate, including the formation of axons and dendrites and the synthesis of neurotransmitters. Neurotransmitters themselves can influence neuronal migration, and it is therefore possible that unique regulatory interactions between amino-acid-transmitter– synthesizing neurons occur during this migratory process (Behar et al., 1999). Neuronal polarity reflects the asymmetry of information processing and the molecular and structural differences in the accumulation of proteins and organelles. As noted previously, polarity is seen first when a neuron becomes postmitotic, with dendrite- and axon-specific proteins expressed during migration. Dendrites grow slowly over many weeks and months, whereas axons grow several orders of magnitude more rapidly, reaching target areas in some instances prior to neurons reaching their final resting position. Complex molecular signaling is responsible for the regulation of axonal growth. Diffusible and membrane proteins that comprise different families, including cadherins, semaphorins, the receptor tyrosine kinase Ephs and their membrane ligands, the ephrins, members of the Ig superfamily, and netrins are responsible both for chemoattraction and chemorepulsion. These signals can act locally, or at a distance to direct extension of axons along specific pathways to reach their proper targets. In the cerebral cortex, transcription factors control the specification of different functional areas that initially facilitate the correct targeting of axons from the dorsal thalamus prenatally, before extensive den-

dritic development of projection neurons. Downstream effectors of axon guidance are thus distributed in unique patterns to produce a basic blueprint of long projections. This process appears to be independent of specific activity in early developing circuits. In contrast, activity-dependent processes appear to be essential for mediating synaptogenesis. Whereas axon targeting occurs rapidly, quantitative analysis of synaptogenesis in the cortex indicates that only a small fraction of adult synapses are present by birth. In all mammalian species, the process of synapse formation and pruning occurs over a relatively long time period, beyond puberty. For example, in primates synapse number peaks early postnatally and through adolescence, plateaus for several years, and ultimately is reduced by almost 40 percent through a normal process of retraction and remodeling (Caviness et al., 1997; Bourgeois, Goldman-Rakic, and Rakic, 1999; Levitt, 2003). Recent studies indicate that regulation of neuronal excitability can have dramatic effects on synapse number (Burrone, O’Byrne, and Murthy, 2002) by controlling the expression of intracellular signaling molecules and transcription factors, such as Mef2, which in turn influence gene transcription (Flavell et al., 2006; Shalizi et al., 2006). Moreover, other molecules, some of which heretofore were considered irrelevant to neuronal function and development, and which can be modulated by physiological activity, including major histocompatibility complex class I receptors and complement proteins, have important roles in regulating synaptogenesis (Boulanger and Shatz, 2004; Bjartmar et al., 2006). The control of dendritic growth is complex, with neurotransmitters, neurotrophins, tyrosine kinases, small GTPases, and afferent-driven physiological activity among the key regulators (McAllister, 2000). In the visual system, neurotrophins are transported in a retrograde manner by developing thalamic neurons from the cortex and increase the complexity of dendritic branching. Ephrins, Eph receptors, semaphorins, and small GTPase signaling proteins also have profound influences on axonal and dendritic development. In vitro studies have shown that neurotransmitters, in the absence of synapses, also can modulate cell migration and dendritic growth. For example, the amino acids glutamate and GABA have complex effects on cell survival and neurite growth. Similarly, monoamines can serve as negative or positive regulators, depending upon which subtypes of receptor proteins are activated (see next section).

Developmental neuropharmacology of brain monoamines Dopamine (DA) In order to place the developmental roles of monoamines in a proper context, it is necessary to understand their normal ontogeny, pharmacology, and functions. Dopamine (DA) has been implicated in a variety of functions in the mature CNS, including motor control,

stanwood and levitt: effects of monoamines

85

cognition, and endocrine, emotional, and cardiovascular regulation. Abnormalities in brain DA systems are thought to contribute to several major neurological and psychiatric disorders including schizophrenia, Parkinson’s disease, attention-deficit hyperactivity disorder (ADHD), and drug addiction. Many exogenously administered drugs, including psychostimulants, act directly on DA systems and can produce long-lasting alterations in endogenous dopaminergic and nondopaminergic functions. The catecholamine DA is synthesized from the amino acid tyrosine through the actions of tyrosine hydroxylase (TH), the rate-limiting enzyme in the process, and followed by aromatic amino acid decarboxylase. The rate at which DA is synthesized is controlled by several mechanisms, including end-product inhibition, changes in the number or structure of TH molecules, and changes in the availability of necessary cofactors for tyrosine hydroxylation. Because of these regulatory processes, neurons usually are able to match the rate of DA synthesis to the rate of DA utilization, thereby avoiding either the buildup or the depletion of the transmitter. As is the case for all neurotransmitters, DA release occurs in response to an influx of calcium into the nerve terminal, which is triggered by the arrival of an action potential. Like DA synthesis, several processes regulate DA release. For example, DA can act back on the terminal from which it was released to inhibit subsequent release. Such influences represent negative feedback loops and act to maintain the rate of DA utilization within relatively narrow limits. DA release can also be potentiated or attenuated by both local and distal actions of other neurotransmitters. Dopamine induces a wide range of cellular and biochemical effects in neurons by way of its interactions with specific receptor proteins. These effects include relatively rapid (seconds) modulation of biochemical events in the target cell, resulting in changes in the responsiveness of the cell to other neuronal inputs, as well as more gradual (minutes-hours) alterations in gene expression. DA receptors belong to a large superfamily of neurotransmitter and hormone receptors that are characterized by an extracellular N-terminus, intracellular C-terminus, seven transmembrane domains, and coupling to specific effector functions through guanine nucleotide binding proteins (G proteins). DA receptors are classified into two subfamilies according to pharmacological profiles and sequence homology: the D1-like receptor subtypes (D1, D5) and the D2-like receptor subtypes (D2, D3, D4) (Sibley and Monsma, 1992). The neurotransmitter actions of released DA and other monoamines are typically terminated by transport back into presynaptic terminals by plasma transporter proteins that form high-affinity uptake sites. Once taken up, monoamines can be metabolized in the nerve terminal by monoamine oxidase (MAO) or further sequestered into storage vesicles by vesicular transporters for later reuse. In human prefrontal

86

cortex, nonneuronal catechol-o-methyl transferase (COMT) is primarily responsible for inactivation, because the DA transporter is not present in these synapses (Tunbridge, Harrison, and Weinberger, 2006). There are several major dopaminergic pathways. Axons of DA-containing cells in the substantia nigra (SN) pars compacta form the nigrostriatal tract, which provides the dopaminergic innervation of the caudate and putamen, or striatum. The striatum is the major output component of the basal ganglia, a group of nuclei involved in motor and cognitive functions. Degeneration of nigrostriatal DA neurons is the primary pathology in Parkinson’s disease. The mesolimbic and mesocortical DA systems largely arise from the ventral tegmental area (VTA), which lies medially to the SN in the midbrain. The mesolimbic DA system innervates ventrally located subcortical regions such as the nucleus accumbens, olfactory tubercle, and amygdala. The nucleus accumbens is thought to be a site of interface between limbic and motor systems and is a critical substrate for the development and regulation of goal-directed behaviors. The mesocortical system provides dopaminergic afferents to the medial prefrontal (mPFC) and anterior cingulate (ACC) cortices. These regions have been implicated in cognitive, emotional, and attentional processes (Goldman-Rakic, 1998; Elston, 2003; Dalley, Cardinal, and Robbins, 2004), and disruption of mesocortical DA neurotransmission has been associated with disease states including schizophrenia, ADHD, and depression. Furthermore, polysynaptic circuits through glutamatergic pathways produce anatomical substrates by which mesocortical DA dysregulation can secondarily alter dopaminergic activity in nigrostriatal and mesoaccumbens neurons, a process that may contribute to pathophysiology. The rate-limiting enzyme in DA synthesis, TH, is first apparent at E12–13 in the rat midbrain, and is present by E14 in the rabbit. Axons of dopaminergic cells reach the cortex a few days later. Limbic cortical regions, such as the ACC and mPFC, receive the densest dopaminergic innervation. This input is thus already present in the cortex even while more superficial cortical layers (II–IV) are beginning to form, consistent with a morphogenic role of DA. In the monkey, DA neurons of the SN/VTA are produced by E30 of a 165-day gestation period (Levitt and Rakic, 1982). In humans, midbrain DA neurons appear in the second month of gestation (Olson and Seiger, 1972). The mechanisms responsible for the proper guidance of dopaminergic afferents to the cortex and the morphogenic properties of these afferents on cortical neurons are not well understood, but netrins and ephrins have been implicated. Transcripts for the D1, D2, and D3 receptors can be detected in the striatum and cortex by E14 in the rat. In addition, D1 and D2 receptor proteins are measurable

fundamentals of developmental neurobiology

prenatally, and they increase throughout prenatal and early postnatal development to reach adult levels of expression between P14 and P21 in rodents. DA receptors functionally couple to G proteins very soon after their appearance. DA also is likely to have early biological activity in the primate brain. In the monkey, for example, DA receptors appear in target regions of DA input by the first quarter of gestation. In the human fetus, D1-like and D2-like receptor binding sites have been detected at gestational week six. Therefore, in all species examined, DA receptors are present very early in prenatal development, consistent with a role for DA in regulating neuronal differentiation and circuit formation. In vitro studies have supported a role for DA as both a promoter and an inhibitor of neurite growth. The actions of DA on outgrowth are modified by the complement of receptors that are activated, and as a function of the neuronal cell type being modulated. In cortical neurons, for example, selective D1 receptor activation decreases neurite outgrowth in a dose-dependent manner, whereas D2 receptor activation increases outgrowth. In striatal neurons, however, these effects are reversed, with D1 receptor activation serving to promote neuronal differentiation and process outgrowth. DA signaling also appears to be involved in prenatal neurogenesis itself within the neuroepithelial precursors of the striatum and cerebral cortex, by way of influences on cell cycle length (Ohtani et al., 2003; Zhang et al., 2005). The phenotypic differentiation of inhibitory GABAergic interneurons may also be modulated by dopaminergic stimulation. Studies from our laboratory and others investigating the effects of prenatal cocaine exposure suggest that modification of DA D1 receptor signaling during a sensitive period of prenatal development induces permanent effects on circuit formation and function (discussed in a later section of this chapter). Recent data also suggest that transient overexpression of the D2 receptor in the developing striatum can cause lifelong changes in the activity of D1 receptor systems in the prefrontal cortex (Kellendonk et al., 2006). Finally, DAdependent processes also alter postnatal development of brain circuits, especially during the periods of synaptic maturation and refinement. Serotonin (5-HT) The indoleamine 5-HT is found in cardiovascular tissue, blood cells, and the nervous system. In the CNS, the cell bodies for serotonergic neurons reside in the midline raphe nuclei in the brain stem. In addition, 5-HT serves as a precursor for melatonin production in the pineal gland. It is formed by the hydroxylation and decarboxylation of tryptophan by the actions of tryptophan hydroxylase and aromatic amino acid decarboxylase, respectively. Similarly to the catecholamines, 5-HT synthesis is controlled by several regulatory mechanisms, including end-product inhibition, changes in the availability of tryptophan hydroxylase, and the activation of autoreceptors.

Receptor proteins for 5-HT are diverse and numerous, as more than 15 different 5-HT receptors have been cloned. These are grouped into seven families (5-HT1–5-HT7). Moreover, alternative mRNA splicing and mRNA editing create at least 20 additional 5-HT receptors with different binding affinities and physiological functions. Similar to the other monoamines, most 5-HT receptors are G-protein– coupled receptors; a notable exception is the ligand-gated ion channel 5-HT3 receptor. The indoleamine 5-HT is removed from the synapse by the actions of a high affinity 5-HT transporter, followed by intracellular metabolism by MAO. Important roles for 5-HT have been implicated in a wide variety of behaviors and conditions including anxiety, appetite, aggression, schizophrenia, migraine, sexual behavior, and drug abuse. The indoleamine 5-HT is one of the earliest developing and most widely distributed neurotransmitter systems in the mammalian brain (Whitaker-Azmitia, 2001; Gaspar, Cases, and Maroteaux, 2003; Luo, Persico, and Lauder, 2003). Serotonergic neurons are first evident by E12 in the rat midbrain (Lauder and Bloom, 1974), by the end of the first month of gestation in the monkey (Levitt and Rakic, 1982), and by the fifth week in humans (Olson and Seiger, 1972). Serotonergic axons reach their forebrain targets prenatally in the rodent (E16–17), but terminal arborization and peak 5-HT levels occur during the second and third postnatal weeks, during synaptogenesis and circuit refinement. The influence of 5-HT occurs at specific sensitive periods. For example, removal of 5-HT during very early fetal development in rats can cause a permanent reduction in the number of cortical neurons in the adult brain. Also, 5-HT-releasing fibers influence Cajal-Retzius cells within the marginal zone of the cerebral cortex. Later in prenatal development, 5-HT receptor activation modifies specific aspects of dendritic development in differentiating neurons. Neurite outgrowth by dorsal thalamic neurons is modulated in a hierarchical fashion by different 5-HT receptors (Persico, Di Pino, and Levitt, 2006). Furthermore, 5-HT plays a negative feedback role on its own neurons. Unfortunately, there are large gaps in our current understanding of where and when distinct 5-HT receptor subtypes are expressed during prenatal brain development. Recent studies from our laboratory on the expression patterns in mice indicate highly regulated patterns of 5-HT receptor expression in the cortex, basal forebrain, and dorsal thalamus (Bonnin et al., 2006). A particularly striking example of a role for 5-HT in circuit formation has been demonstrated in mice where the gene encoding MAO-A has been disrupted (Cases et al., 1996, 1998). Loss of MAO-A results in increases in 5-HT and NE levels, the former being directly responsible for the failure of the development of barrellike structures related to topographic vibrissae representation in the somatosensory cortex. This pattern of cytoarchitectonic organization,

stanwood and levitt: effects of monoamines

87

unique to rodent somatosensory cortex, normally forms during the first postnatal week of life. Barrel formation in the MAO-A null mice can be restored with early intervention, using a drug to inhibit 5-HT synthesis during the critical period for barrel formation. Pharmacological inhibition of MAO-A activity in wild type mice or total absence of 5HT also leads to a loss of barrel formation, demonstrating the remarkable sensitivity of somatosensory pattern formation to 5-HT. Although barrels are absent in the somatosensory cortex in the MAO-A mutant line, thalamic and brain stem barrellike patterns are still evident. Cortical barrel formation is dependent upon activity arising from peripheral whisker barrels, transmitted through several synapses in the brain stem and thalamus. So why do non-monoaminergic afferents arising from regions with seemingly normal barrel formation produce abnormal barrel formation at the next synapse, in the somatosensory cortex? During normal development, the ventrobasal thalamic projections to somatosensory cortex express transiently the presynaptic 5-HT transporter, the vesicular monoamine transporter, and accumulate 5-HT during the critical period in the first postnatal week. The modulation of cortical barrel formation by 5-HT is shown further in the study in which genetic deletion of presynaptic 5-HT1B receptors, in the context of excessive 5HT produced by MAO-A disruption, results in normal barrel development (Salichon et al., 2001). Collectively, these studies have provided the strongest evidence to date that alterations in monoamine levels during development can lead to aberrant neuronal projection patterns and target organization. These data also have led to the hypothesis that exposure to selective 5-HT reuptake inhibitors (SSRIs) during thalamocortical synaptic formation may have deleterious consequences. Recent studies in rodents have suggested that this may be the case (Xu, Sari, and Zhou, 2004). In addition, 5-HT plays a role in the establishment of adult anxiety behavior through a mechanism in which activity need only be disrupted during the early postnatal period (Gross et al., 2002). Mutation of the 5-HT1A receptor in mice causes increases in anxiety-related behavior. This defect can be rescued by expression of the receptor in specific regions of the forebrain using a conditional transactivation genetic system. This conditional knockout strategy was used to show that, whereas repression of receptor expression in the adult is ineffective, repression of receptor expression until three weeks of age is sufficient to produce adult mice with increased anxiety-related behavior. Forebrain 5-HT1A receptor expression during the early postnatal period, but not in the adult, is thus necessary for the expression of normal anxiety responses. Norepinephrine (ne) Norepinephrine is involved in attention, anxiety, arousal, and learning and memory. The cell bodies of NE neurons are concentrated in the brain

88

stem, particularly in the locus coeruleus. From this structure, five major NE tracts originate that innervate the whole brain. DA serves as a metabolic intermediate within NE neurons, with conversion of DA to NE occurring through the action of the enzyme dopamine β-hydroxylase. There are at least ten identified NE receptors, which are segregated into two families (α and β). NE neurons in the pons (locus coeruleus) and brain stem first appear at relatively early stages in the CNS, approximately E13 in the rat, E30 in the monkey (Levitt and Rakic, 1982), and by 5–6 weeks in humans (Olson and Seiger, 1972). Despite the fact that many types of psychotropic medications have actions on the noradrenergic system, very little research has been conducted on the developmental roles of NE during pre- and postnatal development. One developmental function that has been described for NE during embryogenesis is on the development of Cajal-Retzius cells in the cerebral cortex (Naqui et al., 1999). These cells are the first neurons to be born in the cortex and are instrumental in neuronal migration and laminar formation (figure 6.1). Furthermore, the α2a NE receptor (as well as other monoamine receptors) is expressed by migrating neurons in the intermediate zone in close association with radial glia (Lidow and Wang, 1995). Thus NE may be involved in regulating the generation, migration, and maturation of cerebral cortical cells, but directed studies are needed in this area. Postnatally, the best characterized effect of NE has been in studies examining the role of neuromodulators in criticalperiod plasticity in sensory systems (Manunta and Edeline, 2004).

An example: Alterations in brain development and function due to prenatal cocaine exposure Given the modulatory influence of monoamines on specific aspects of neural development, it is understandable that prenatal exposure to drugs that affect these systems, such as drugs of abuse and psychotherapeutics, can have pronounced effects on the development of the cerebral cortex. Because of space limitations, our discussion will concentrate specifically on the effects of in utero cocaine, and we direct the reader to other sources for discussions of other psychopharmacological insults (Slotkin, 1998; Trask and Kosofsky, 2000; Thadani, 2002; Carlezon and Konradi, 2004; Olney et al., 2004; Andersen, 2005). The primary pharmacological sites of action of cocaine and other psychostimulants in the brain are the high-affinity transporters for DA, 5-HT, and NE. Cocaine binds to these transport proteins and blocks the reuptake of the neurotransmitters, thus prolonging their time in the extracellular space. This action permits the monoamine to bind to its receptor proteins for more sustained periods, resulting in excessive activation of these receptors, particularly those located

fundamentals of developmental neurobiology

extrasynaptically. Cocaine is a drug of abuse in adolescents and adults, produces a host of neuroadaptations in the brain of the user that are associated with addiction (Hyman and Malenka, 2001), and can potently modulate monoaminergic systems during prenatal development if the drug is used during pregnancy. Clinical reports of the impact of prenatal cocaine exposure have been diverse, as some suggest gross physical malformations, others observe specific deficits in cognitive and emotional development, and yet others indicate no effects. The variable outcomes are at least in part the result of important covariates such as the timing and amount of cocaine use during pregnancy, polydrug use, and the quality of pre- and postnatal care (Singer et al., 2004; Mayes, 2003). Severe physical abnormalities in cocaine-exposed infants typically occur following very prolonged and high levels of cocaine intake by the mother. However, there are measurable deficits in cognition and emotional regulation in children exposed to relatively low doses of cocaine in the womb. These disturbances are due primarily to alterations in attention, language development, and regulation of arousal, and are likely caused by improper functioning of limbic cortices and monoamine systems. Prospective longitudinal studies of cocaine-exposed infants and school-age children (Mayes et al., 1995; Richardson, Conroy, and Day, 1996; Singer et al., 2002; Arendt et al., 2004) have produced the least confounded and most compelling reports. For example, Singer and colleagues have assessed a prospective cohort where the relationship between prenatal cocaine exposure and developmental and cognitive outcomes has been assessed in 218 cocaine-exposed and 197 unexposed infants. The rate of developmental delay was nearly doubled (13.7% versus 7.1%, respectively). Importantly, this still ongoing study has controlled for prenatal exposure to other drugs, gestational age at birth, weight at birth, and caregiver characteristics. Disturbances reminiscent of children diagnosed with ADHD are consistently found in children exposed to cocaine prenatally (Leech et al., 1999; Mayes, 2002; B. Lewis et al., 2004; M. Lewis et al., 2004; Singer et al., 2004). These changes include deficits in recognition memory, task persistence, distractibility, and stress responsiveness, and may be worsening as these children continue to age. Interestingly, increased risk of ADHD also has been documented in the offspring of women who smoke or consume alcohol heavily during their pregnancy. Reports also have suggested that prenatal cocaine exposure delays language development during infancy and early childhood. There are numerous inconsistencies in the basic science literature regarding the biological effects of prenatal cocaine exposure, likely stemming from differences in dosing, route of administration, time of exposure, age of assessment, and species (see Stanwood and Levitt, 2001, for review of this

issue). Many models have utilized subcutaneous or intraperitoneal injections of high doses of cocaine (20–80 mg/kg). However, cocaine has complicated and diverse pharmacological effects on biological targets in the brain and the periphery of both the mother and fetus, and can exhibit pronounced teratological effects at high doses. We have suggested that a low-dose model, using a route of administration that closely mimics the pharmacokinetic response in human cocaine abusers, provides a reproducible system that facilitates the assessment of specific effects on the organization of cocaine’s targets in the CNS, the biogenic amine systems. A low-dose (2–4 mg/kg), intravenous rabbit model that resulted in highly reproducible and selective brain defects was thus initiated by a group of investigators. There are specific, dose-dependent, and permanent effects on behavioral functioning and on the structure and function of cortices receiving a rich DA innervation (Levitt et al., 1997; Harvey, 2004; Stanwood and Levitt, 2004). The effects are produced either by exposure for the majority of pregnancy or, perhaps more interestingly, during a short sensitive period during the human equivalent of the second trimester (Stanwood, Washington, and Levitt, 2001). The anatomical defects include aberrant growth of dendrites of cortical projection and interneurons, suggesting disruption of local circuitry, and behavioral abnormalities that involve Pavlovian learning and stereotypic motor behavior (figure 6.2 and plate 16). Selective deficits in aspects of cognition and/or attention also have been noted (Thompson, Levitt, and Stanwood, 2005). Possibly the most remarkable demonstration of the impact of prenatal cocaine exposure, and perhaps the point of origin for the structural and behavioral disturbances, is the striking reduction in coupling of the DA D1 receptor to its G protein (Jones et al., 2000; Zhen et al., 2001), which is initiated prenatally and sustained into adulthood (figure 6.2). This effect appears to be specific for D1-Gs coupling, because D2 and muscarinic cholinergic receptor coupling to Gi/Go proteins is normal. We have viewed the D1 receptor signaling defect as a manifestation of a cellular strategy, distinct from that in adults, for adapting to disrupted balance of DA signaling during development. Strikingly, the D1 receptor knockout mouse exhibits similar structural alterations in cortical development to rabbits in which prenatal cocaine reduces D1-Gs coupling (Stanwood, Parlaman, and Levitt, 2005). The neuroadaptive changes that occur in response to repeated exposure to cocaine vary with the maturational state during which the exposure occurs. The offspring of rabbits treated with cocaine show near complete loss of an amphetamine-induced stereotyped behavior (head bobbing) when tested as young adults (Stanwood and Levitt, 2003). In contrast, the mothers of these offspring, receiving cocaine at the same dose and duration, and with the same period of withdrawal, exhibit profound sensitization to the behavioral

stanwood and levitt: effects of monoamines

89

Figure 6.2 Schematic representation of cellular effects of in utero cocaine exposure in the rabbit. Rabbits exposed to low doses of cocaine intravenously exhibit alterations in the structure and function of neurons in the anterior cingulate cortex. The apical dendrites of pyramidal neurons (red) exhibit an undulating trajectory. The number of interneurons in which immunoreactivity for GABA is detectable (black) is increased. There also is an increase in parvalbumin (green) immunostaining in the dendrites of a subset of

these neurons. D1 dopamine receptor coupling to Gs protein (blue) is reduced, whereas coupling of the D2 receptor to Gi (orange) is unaffected. These changes likely influence the balance of excitatory and inhibitory influences in the cingulate cortex and produce the aberrant behavioral phenotypes exhibited by these rabbits, including psychostimulant responsiveness and attentional deficits (see text for more detail). (Adapted from Levitt et al., 1997). (See plate 16.)

effects of amphetamine (Stanwood and Levitt, 2003). Our most recent data suggest a similar double dissociation in the reinforcing effects of cocaine following previous in utero or adult exposure to the drug (B. Thompson, G. Stanwood, and P. Levitt, unpublished observations). Two groups have utilized intravenous administration of cocaine in rats (Mactutus, Herman, and Booze, 1994; Mactutus, 1999; Morrow, Elsworth, and Roth, 2002, 2003). Although the studies have emphasized the analysis of distinct biological measurements and markers, direct parallels between models are beginning to be established. Furthermore, models of higher dose cocaine administration in rodents (Kosofsky, 1998; Rocha, Mead, and Kosofsky, 2002; Melnick and Dow-Edwards, 2003) and nonhuman primates (Lidow, 1998; Chelonis, Gillam, and Paule, 2003) have revealed additional disruptions of CNS development by cocaine, including potential effects on neuronal migration, differentiation, and cell survival. We propose that the loss of dopamine D1 receptor signaling may represent a conserved feature of prenatal cocaine exposure and quite possibly the primary cellular mechanism used to compensate for excessive DA release and receptor stimulation during neuronal differentiation. Given the primarily extrasynaptic location of the D1 receptor, as compared to the intact primarily synaptically located D2 receptor, and the activation of the D1 receptor by phasic, but not tonic, DA release, we predict that the basal functioning of DA-modulated circuits would be minimally disrupted.

Environmental or pharmacological challenges, which would engage both receptors, however, would be influenced substantially, with the outcome likely to produce atypical DA modulation of circuits. Furthermore, the long-lasting decrease in stimulant-induced behavior in the rabbit model suggests that brain pharmacology is permanently altered and “typical” responses to drugs of abuse or psychotherapeutics cannot be assumed to occur in cocaine-exposed children. The data further suggest that some children exposed prenatally to cocaine will respond poorly to mild psychostimulants such as methylphenidate, a disturbing prediction given the high incidence of ADHD in prenatal cocaineexposed children. An alternative strategy might include treatment aimed at restoring the “balance” of D1 and D2 receptor signaling, which could be more efficacious (see next section). Further interactions between DA and the other monoamines are likely to play a role. Cocaine also blocks the reuptake of NE and 5-HT, and, as described earlier, these neuromodulators also influence excitatory and inhibitory activity within the brain. Although no evidence for changes in 5-HT or NE innervation or activity has been identified in the rabbit model, deficits in serotonergic functioning certainly have been described in higher dose models of prenatal cocaine exposure. Intriguingly, the ACC and mPFC, the cortical regions in which we have observed in utero cocaine-induced changes in our rabbit model, contribute to the neural control of

90

fundamentals of developmental neurobiology

attention and are sites of brain dysfunction in ADHD. Children exposed in utero to cocaine perform poorly on tasks that depend on the proper functioning of the mPFC and ACC, showing deficits similar to those of children diagnosed with ADHD. Reduced signaling through the DA D1 receptor appears to underlie the structural and functional deficits observed in prenatal cocaine-exposed rabbits, and thus it is intriguing to speculate that similar processes may be occurring in cocaine-exposed children. The development of a therapeutic strategy to restore proper D1 receptor coupling in these children might therefore be efficacious in the treatment of attentional disturbances. Reduced D1 receptor coupling could have other severe consequences throughout their lifetimes for children exposed prenatally to cocaine. D1 receptor activation is an important substrate for reward pathways in the brain. Although not yet studied in the human population directly, one prediction is that drug-exposed children would likely experience defects in endogenous reward systems that could lead to the occurrence of anhedonia and/or depression. In fact, reports

from animal models suggest that the efficacy of reinforcers is reduced following prenatal cocaine exposure.

Figure 6.3 D1- and D2-like DA receptors produce distinct, and sometimes opposite, neuromodulatory effects on cell signaling, electrophysiological responses, and behavior. They are often coexpressed within specific functional brain circuits, and sometimes even by the same neurons. A reduction in D1 receptor signaling could shift the balance of activity to D2 receptor subtypes. Normal physiological responsiveness may not necessarily depend on the

levels of receptor (above a certain minimum), but rather a balance between opposing activities. Thus, in considering strategies for restoring normal DA responsiveness, blockade of D2 receptors may have beneficial effects on the net outflow of DA-dependent neural circuits. (Adapted from Stanwood and Levitt, 2004.) (See plate 17.)

Balance of receptor signaling The amount of receptor stimulation by each of the monoamines, as well as the molecular identity of those receptors, determines the intracellular response to the extracellular cue. These molecules, however, have relatively slow modulatory actions, in contrast to fast-acting neurotransmitters such as glutamate and GABA. It has been hypothesized, therefore, that an even more important determinant of ultimate functional response than the absolute level of stimulation is the balance of receptor subtype activation (figure 6.3 and plate 17). This concept was initially proposed nearly 20 years ago to explain some of the actions of DA D1 and D2 receptor activation, but few studies have been conducted to test this hypothesis. Recently, two phenomena have returned attention to the idea. First, detailed analysis of the

stanwood and levitt: effects of monoamines

91

electrophysiological responses of prefrontal cortical neurons to application of DA and specific agonists have revealed complex “state” changes in firing patterns mediated distinctly by D1-like and D2-like receptors (Seamans and Yang, 2004). Second, a recent biochemical study revealed that interactions between DA receptors and specific G proteins are more complicated than previously believed. In this regard, D1 receptors couple to cAMP through Gs or Golf if expressed alone, and D2 receptors couple best with Gi or Go in the absence of D1 receptors. However, when the two receptors are coactivated in the same cell, an entirely new intracellular cascade, a rise in intracellular calcium mediated by Gq, is activated (Lee et al., 2004). Other monoamine receptors exhibit similar patterns of coupling to G proteins, activate diverse intracellular signaling pathways, and are often coexpressed by the same neurons within the brain. It is therefore likely that combinatorial rules of receptor balance are utilized by monoamines during development, as well as in the mature brain (Stanwood and Levitt, 2004).

Conclusions Alterations in biogenic amine availability can modify neurotransmitter systems and intracellular messengers both in the developing and adult brain, with likely very different outcomes. Monoamines are pleiotropic signaling molecules that serve to regulate specific aspects of central nervous system development. Pharmacological or genetic modulation of their receptors leads to specific, targeted changes in brain structure and circuitry, and results in permanent alterations in neural function. Such effects on the developing nervous system, before homeostatic regulatory mechanisms are properly calibrated, differ from their effects on mature systems. We suggest that mature, dysfunctional states acquired following disruption of monoamine signaling during development are defined in part by altered function of biogenic amines as nontraditional modulators of pre- and postnatal brain maturation, and more traditional influences on neurotransmission in the adult. It is likely that other environmental factors that engage the DA systems, such as other psychotropic drugs or environmental stressors, may contribute to dysregulation of circuit formation and function by altering distinct cell signaling systems in analogous ways. The sensitive periods for these insults will be defined by the relation between the timing of the environmental exposure and the timing of developmental modifications in the neural system upon which it acts. acknowledgments

The authors are supported in part by DA11165, DA017957, and P30HD15052. We thank Drs. Barbara Thompson, Alexandre Bonnin, and Kathie Eagleson for helpful conversations on the topics reviewed in this chapter.

92

REFERENCES Andersen, S. L., 2005. Stimulants and the developing brain. Trends Pharmacol. Sci. 26:237–243. Anderson, S. A., D. D. Eisenstat, L. Shi, and J. L. Rubenstein, 1997. Interneuron migration from basal forebrain to neocortex: Dependence on Dlx genes. Science 278:474–476. Arendt, R. E., E. J. Short, L. T. Singer, S. Minnes, J. Hewitt, S. Flynn, L. Carlson, M. O. Min, N. Klein, and D. Flannery, 2004. Children prenatally exposed to cocaine: Developmental outcomes and environmental risks at seven years of age. J. Dev. Behav. Pediatr. 25:83–90. Behar, T. N., C. A. Scott, C. L. Greene, X. Wen, S. V. Smith, D. Maric, Q. Y. Liu, C. A. Colton, and J. L. Barker, 1999. Glutamate acting at NMDA receptors stimulates embryonic cortical neuronal migration. J. Neurosci. 19:4449–4461. Bjartmar, L., A. D. Huberman, E. M. Ullian, R. C. Renteria, X. Liu, W. Xu, J. Prezioso, M. W. Susman, D. Stellwagen, C. C. Stokes, R. Cho, P. Worley, R. C. Malenka, S. Ball, N. S. Peachey, D. Copenhagen, B. Chapman, M. Nakamoto, B. A. Barres, and M. S. Perin, 2006. Neuronal pentraxins mediate synaptic refinement in the developing visual system. J. Neurosci. 26:6269–6281. Bonnin, A., W. Peng, W. Hewlett, and P. Levitt, 2006. Expression mapping of 5-HT1 serotonin receptor subtypes during fetal and early postnatal mouse forebrain development. Neuroscience 141:781–794. Boulanger, L. M., and C. J. Shatz, 2004. Immune signalling in neural development, synaptic plasticity and disease. Nature Rev. Neurosci. 5:521–531. Bourgeois, J.-P., P. S. Goldman-Rakic, and P. Rakic, 1999. Formation, elimination, and stabilization of synapses in the primate cerebral cortex. In The New Cognitive Neurosciences, 2nd ed., ed. M. S. Gazzaniga, 45–53. Cambridge, MA: MIT Press. Burrone, J., M. O’Byrne, and V. N. Murthy, 2002. Multiple forms of synaptic plasticity triggered by selective suppression of activity in individual neurons. Nature 420:414–418. Campbell, K., and M. Gotz, 2002. Radial glia: Multipurpose cells for vertebrate brain development. Trends Neurosci. 25:235–238. Carlezon, W. A., Jr., and C. Konradi, 2004. Understanding the neurobiological consequences of early exposure to psychotropic drugs: Linking behavior with molecules. Neuropharmacology 47 (Suppl 1):47–60. Cases, O., C. Lebrand, B. Giros, T. Vitalis, E. De Maeyer, M. G. Caron, D. J. Price, P. Gaspar, and I. Seif, 1998. Plasma membrane transporters of serotonin, dopamine, and norepinephrine mediate serotonin accumulation in atypical locations in the developing brain of monoamine oxidase A knock-outs. J. Neurosci. 18:6914–6927. Cases, O., T. Vitalis, I. Seif, E. De Maeyer, C. Sotelo, and P. Gaspar, 1996. Lack of barrels in the somatosensory cortex of monoamine oxidase A-deficient mice: Role of a serotonin excess during the critical period. Neuron 16:297–307. Caviness, V. S., D. N. Kennedy, J. F. Bates, and N. Makris, 1997. The developing human brain: A morphometric profile. In Developmental Neurobiology: Mapping the Development of Brain and Behavior, ed. R. W. Thatcher, G. R. Lyon, J. Rumsey, and N. Krasneger, 3–14. San Diego: Academic Press. Charman, T., 1999. Autism and the pervasive developmental disorders. Curr. Opin. Neurol. 12:155–159. Chelonis, J. J., M. P. Gillam, and M. G. Paule, 2003. The effects of prenatal cocaine exposure on reversal learning using a simple

fundamentals of developmental neurobiology

visual discrimination task in rhesus monkeys. Neurotoxicol. Teratol. 25:437–446. Dalley, J. W., R. N. Cardinal, and T. W. Robbins, 2004. Prefrontal executive and cognitive functions in rodents: Neural and neurochemical substrates. Neurosci. Biobehav. Rev. 28:771–784. Elston, G. N., 2003. Cortex, cognition and the cell: New insights into the pyramidal neuron and prefrontal function. Cerebral Cortex 13:1124–1138. Flavell, S. W., C. W. Cowan, T. K. Kim, P. L. Greer, Y. Lin, S. Paradis, E. C. Griffith, L. S. Hu, C. Chen, and M. E. Greenberg, 2006. Activity-dependent regulation of MEF2 transcription factors suppresses excitatory synapse number. Science 311:1008–1012. Gaspar, P., O. Cases, and L. Maroteaux, 2003. The developmental role of serotonin: News from mouse molecular genetics. Nature Rev. Neurosci. 4:1002–1012. Goldman-Rakic, P. S., 1998. The cortical dopamine system: Role in memory and cognition. Adv. Pharmacol. (NY) 42:707–711. Gotz, M., and W. B. Huttner, 2005. The cell biology of neurogenesis. Nature Rev. Mol. Cell Biol. 6:777–788. Gross, C., X. Zhuang, K. Stark, S. Ramboz, R. Oosting, L. Kirby, L. Santarelli, S. Beck, and R. Hen, 2002. Serotonin1A receptor acts during development to establish normal anxiety-like behaviour in the adult. Nature 416:396–400. Harvey, J. A., 2004. Cocaine effects on the developing brain: Current status. Neurosci. Biobehav. Rev. 27:751–764. Herlenius, E., and H. Lagercrantz, 2004. Development of neurotransmitter systems during critical periods. Exp. Neurol. 190 (Suppl 1):S8–21. Hyman, S. E., and R. C. Malenka, 2001. Addiction and the brain: The neurobiology of compulsion and its persistence. Nature Rev. Neurosci. 2:695–703. Jones, L. B., G. D. Stanwood, B. S. Reinoso, R. A. Washington, H. Y. Wang, E. Friedman, and P. Levitt, 2000. In utero cocaine-induced dysfunction of dopamine D1 receptor signaling and abnormal differentiation of cerebral cortical neurons. J. Neurosci. 20:4606–4614. Kanold, P. O., P. Kara, R. C. Reid, and C. J. Shatz, 2003. Role of subplate neurons in functional maturation of visual cortical columns. Science 301:521–525. Kellendonk, C., E. H. Simpson, H. J. Polan, G. Malleret, S. Vronskaya, V. Winiger, H. Moore, and E. R. Kandel, 2006. Transient and selective overexpression of dopamine D2 receptors in the striatum causes persistent abnormalities in prefrontal cortex functioning. Neuron 49:603–615. Kosofsky, B. E., 1998. Cocaine-induced alterations in neurodevelopment. Semi. Speech Lang. 19:109–121. Kriegstein, A. R., 2005. Constructing circuits: Neurogenesis and migration in the developing neocortex. Epilepsia 46 (Suppl 7):15–21. Lauder, J. M., 1993. Neurotransmitters as growth regulatory signals: Role of receptors and second messengers. Trends Neurosci. 16:233–240. Lauder, J. M., and F. E. Bloom, 1974. Ontogeny of monoamine neurons in the locus coeruleus, Raphe nuclei and substantia nigra of the rat. I. Cell differentiation. J. Comp. Neurol. 155:469–481. Lee, S. P., C. H. So, A. J. Rashid, G. Varghese, R. Cheng, A. J. Lanca, B. F. O’Dowd, and S. R. George, 2004. Dopamine D1 and D2 receptor coactivation generates a novel phospholipase C-mediated calcium signal. J. Biol. Chem. 279:35671– 35678.

Leech, S. L., G. A. Richardson, L. Goldschmidt, and N. L. Day, 1999. Prenatal substance exposure: Effects on attention and impulsivity of 6-year-olds. Neurotoxicol. Teratol. 21:109–118. Levitt, P., 2003. Structural and functional maturation of the developing primate brain. J. Pediatr. 143:S35–45. Levitt, P., K. L. Eagleson, and E. M. Powell, 2004. Regulation of neocortical interneuron development and the implications for neurodevelopmental disorders. Trends Neurosci. 27:400–406. Levitt, P., J. A. Harvey, E. Friedman, K. Simansky, and E. H. Murphy, 1997. New evidence for neurotransmitter influences on brain development. Trends Neurosci. 20:269–274. Levitt, P. and P. Rakic, 1982. The time of genesis, embryonic origin and differentiation of the brain stem monoamine neurons in the rhesus monkey. Brain Res. 256:35–57. Lewis, B. A., L. T. Singer, E. J. Short, S. Minnes, R. Arendt, P. Weishampel, N. Klein, and M. O. Min, 2004. Four-year language outcomes of children exposed to cocaine in utero. Neurotoxicol. Teratol. 26:617–627. Lewis, D. A., and P. Levitt, 2002. Schizophrenia as a disorder of neurodevelopment. Annu. Rev. Neurosci. 25:409–432. Lewis, M. W., S. Misra, H. L. Johnson, and T. S. Rosen, 2004. Neurological and developmental outcomes of prenatally cocaineexposed offspring from 12 to 36 months. Am. J. Drug. Alcohol Abuse 30:299–320. Lidow, M. S., 1998. Nonhuman primate model of the effect of prenatal cocaine exposure on cerebral cortical development. Ann. NY Acad. Sci. 846:182–193. Lidow, M. S., and F. Wang, 1995. Neurotransmitter receptors in the developing cerebral cortex. Crit. Rev. Neurobiol. 9:395– 418. Luo, X., A. M. Persico, and J. M. Lauder, 2003. Serotonergic regulation of somatosensory cortical development: Lessons from genetic mouse models. Dev. Neurosci. 25:173–183. Mactutus, C. F., 1999. Prenatal intravenous cocaine adversely affects attentional processing in preweanling rats. Neurotoxicol. Teratol. 21:539–550. Mactutus, C. F., A. S. Herman, and R. M. Booze, 1994. Chronic intravenous model for studies of drug (ab)use in the pregnant and/or group-housed rat: An initial study with cocaine. Neurotoxicol. Teratol. 16:183–191. Manunta, Y., and J. M. Edeline, 2004. Noradrenergic induction of selective plasticity in the frequency tuning of auditory cortex neurons. J. Neurophysiol. 92:1445–1463. Mayes, L. C., 2002. A behavioral teratogenic model of the impact of prenatal cocaine exposure on arousal regulatory systems. Neurotoxicol. Teratol. 24:385–395. Mayes, L. C., 2003. Genetics of childhood disorders: LV prenatal drug exposure. J. Am. Acad. Child Adolesc. Psychiatry 42:1258– 1261. Mayes, L. C., M. H. Bornstein, K. Chawarska, and R. H. Granger, 1995. Information processing and developmental assessments in 3-month-old infants exposed prenatally to cocaine. Pediatrics 95:539–545. McAllister, A. K., 2000. Cellular and molecular mechanisms of dendrite growth. Cerebral Cortex 10:963–973. Melnick, S. M., and D. L. Dow-Edwards, 2003. Blunted metabolic response to SKF 82958 in the mesolimbic system following preweaning cocaine treatment. Brain Res. Dev. Brain Res. 143:253–259. Morrow, B. A., J. D. Elsworth, and R. H. Roth, 2002. Male rats exposed to cocaine in utero demonstrate elevated expression of Fos in the prefrontal cortex in response to environment. Neuropsychopharmacology 26:275–285.

stanwood and levitt: effects of monoamines

93

Morrow, B. A., J. D. Elsworth, and R. H. Roth, 2003. Axoaxonic structures in the medial prefrontal cortex of the rat: Reduction by prenatal exposure to cocaine. J. Neurosci. 23: 5227–5234. Nadarajah, B., and J. G. Parnavelas, 2002. Modes of neuronal migration in the developing cerebral cortex. Nature Rev. Neurosci. 3:423–432. Naqui, S. Z., B. S. Harris, D. Thomaidou, and J. G. Parnavelas, 1999. The noradrenergic system influences the fate of CajalRetzius cells in the developing cerebral cortex. Brain Res. Dev. Brain Res. 113:75–82. Noctor, S. C., A. C. Flint, T. A. Weissman, R. S. Dammerman, and A. R. Kriegstein, 2001. Neurons derived from radial glial cells establish radial units in neocortex. Nature 409: 714–720. Ohtani, N., T. Goto, C. Waeber, and P. G. Bhide, 2003. Dopamine modulates cell cycle in the lateral ganglionic eminence. J. Neurosci. 23:2840–2850. Olney, J. W., C. Young, D. F. Wozniak, V. Jevtovic-Todorovic, and C. Ikonomidou, 2004. Do pediatric drugs cause developing neurons to commit suicide? Trends Pharmacol. Sci. 25:135–139. Olson, L., and A. Seiger, 1972. Early prenatal ontogeny of central monoamine neurons in the rat: Fluorescence histochemical observations. Z. Anat. Entwicklungsgeschichte 137:301–316. Persico, A. M., G. Di Pino, and P. Levitt, 2006. Multiple receptors mediate the trophic effects of serotonin on ventroposterior thalamic neurons in vitro. Brain Res. 1095:17–25. Rakic, P., 1995. A small step for the cell, a giant leap for mankind: A hypothesis of neocortical expansion during evolution. Trends Neurosci. 18:383–388. Richardson, G. A., M. L. Conroy, and N. L. Day, 1996. Prenatal cocaine exposure: Effects on the development of school-age children. Neurotoxicol. Teratol. 18:627–634. Rocha, B. A., A. N. Mead, and B. E. Kosofsky, 2002. Increased vulnerability to self-administer cocaine in mice prenatally exposed to cocaine. Psychopharmacology (Berl.) 163:221–229. Salichon, N., P. Gaspar, A. L. Upton, S. Picaud, N. Hanoun, M. Hamon, E. De Maeyer, D. L. Murphy, R. Mossner, K. P. Lesch, R. Hen, and I. Seif, 2001. Excessive activation of serotonin (5-HT) 1B receptors disrupts the formation of sensory maps in monoamine oxidase a and 5-ht transporter knock-out mice. J. Neurosci. 21:884–896. Seamans, J. K., and C. R. Yang, 2004. The principal features and mechanisms of dopamine modulation in the prefrontal cortex. Prog. Neurobiol. 74:1–58. Shalizi, A., B. Gaudilliere, Z. Yuan, J. Stegmuller, T. Shirogane, Q. Ge, Y. Tan, B. Schulman, J. W. Harper, and A. Bonni, 2006. A calcium-regulated MEF2 sumoylation switch controls postsynaptic differentiation. Science 311:1012–1017. Sibley, D. R., and F. J. Monsma, 1992. Molecular biology of dopamine receptors. Trends Pharmacol. Sci. 13:61–68. Singer, L. T., R. Arendt, S. Minnes, K. Farkas, A. Salvator, H. L. Kirchner, and R. Kliegman, 2002. Cognitive and motor outcomes of cocaine-exposed infants. JAMA 287:1952–1960.

94

Singer, L. T., S. Minnes, E. Short, R. Arendt, K. Farkas, B. Lewis, N. Klein, S. Russ, M. O. Min, and H. L. Kirchner, 2004. Cognitive outcomes of preschool children with prenatal cocaine exposure. JAMA 291:2448–2456. Slotkin, T. A., 1998. Fetal nicotine or cocaine exposure: Which one is worse? J.Pharmacol. Exp. Ther. 285:931–945. Stanwood, G. D., and P. Levitt, 2001. The effects of cocaine on the developing nervous system. In Handbook of Developmental Cognitive Neuroscience, ed. C. A. Nelson and M. Luciana, 519–536. Cambridge, MA: MIT Press. Stanwood, G. D., and P. Levitt, 2003. Repeated i.v. cocaine exposure produces long-lasting behavioral sensitization in pregnant adults, but behavioral tolerance in their offspring. Neuroscience 122:579–583. Stanwood, G. D., and P. Levitt, 2004. Drug exposure early in life: Functional repercussions of changing neuropharmacology during sensitive periods of brain development. Curr. Opin. Pharmacol. 4:65–71. Stanwood, G. D., J. P. Parlaman, and P. Levitt, 2005. Anatomical abnormalities in dopaminoceptive regions of the cerebral cortex of dopamine D(1) receptor mutant mice. J. Comp. Neurol. 487:270–282. Stanwood, G. D., R. A. Washington, and P. Levitt, 2001. Identification of a sensitive period of prenatal cocaine exposure that alters the development of the anterior cingulate cortex. Cerebral Cortex 11:430–440. Thadani, P. V., 2002. The intersection of stress, drug abuse and development. Psychoneuroendocrinology 27:221–230. Thompson, B. L., P. Levitt, and G. D. Stanwood, 2005. Prenatal cocaine exposure specifically alters spontaneous alternation behavior. Behav. Brain Res. 164:107–116. Trask, C. L., and B. E. Kosofsky, 2000. Developmental considerations of neurotoxic exposures. Neurol. Clin. 18:541–562. Tunbridge, E. M., P. J. Harrison, and D. R. Weinberger, 2006. Catechol-o-methyltransferase, cognition, and psychosis: Val158Met and beyond. Biol. Psychiatry 60:141–151. Weinberger, D. R., 1995. From neuropathology to neurodevelopment. Lancet 346:552–557. Whitaker-Azmitia, P. M., 2001. Serotonin and brain development: Role in human developmental diseases. Brain Res. Bull. 56:479–485. Xu, Y., Y. Sari, and F. C. Zhou, 2004. Selective serotonin reuptake inhibitor disrupts organization of thalamocortical somatosensory barrels during development. Brain Res. Dev. Brain Res. 150:151–161. Zhang, L., J. Bai, A. S. Undie, C. Bergson, and M. S. Lidow, 2005. D1 dopamine receptor regulation of the levels of the cell-cycle-controlling proteins, cyclin D, P27 and Raf-1, in cerebral cortical precursor cells is mediated through cAMP-independent pathways. Cerebral Cortex 15:74– 84. Zhen, X., C. Torres, H. Y. Wang, and E. Friedman, 2001. Prenatal exposure to cocaine disrupts D1A dopamine receptor function via selective inhibition of protein phosphatase 1 pathway in rabbit frontal cortex. J. Neurosci. 21:9160–9167.

fundamentals of developmental neurobiology

B. STRUCTURAL FOUNDATIONS OF SENSATION, PERCEPTION, AND COGNITION

7

Mechanisms of Auditory Reorganization during Development: From Sounds to Words RICHARD N. ASLIN, MEGHAN A. CLAYARDS, AND NEIL P. BARDHAN

Several weeks before the human newborn is first exposed to visual stimulation, auditory inputs are capable of stimulating hair cells on the basilar membrane, generating neural signals that travel from the auditory nerve to brain-stem, thalamic, and cortical areas of the central nervous system, triggering behavioral responses, and establishing rudimentary representations that can affect behavioral preferences several days or weeks later. Newborns exhibit the ability to orient toward sounds, to discriminate between some acoustic differences, and to show preferences for certain classes of auditory stimuli, only some of which could have been induced by prenatal experience. Given this state of the auditory system at birth, the challenge facing the developing infant is to determine which of the myriad of acoustic cues present in the proximal auditory environment carries information relevant to solving particular ecologically relevant tasks. For example, sound localization requires the extraction of particular spectral and temporal cues, the most important of which involve interaural differences. In contrast, speech perception involves the extraction of different spectral and temporal cues, most of which do not depend on binaural information. In addition to this problem of learning to extract the “right” acoustic cues, the infant is also faced with tremendous variability along the acoustic dimensions that are relevant for solving a particular task. For example, spectral cues are relatively unimportant for sound localization along the horizontal axis but play a crucial role along the vertical axis. Similarly, variations in pitch and duration are less important than spectral cues for the perception of consonants and vowels, but these pitch and duration cues play a crucial role in tone languages and in identifying a particular talker or determining the talker’s emotional state. We begin this chapter with a brief summary of the development of basic auditory capacities in infancy, which, as we will see, play a relatively minor role in constraining most auditory tasks that involve suprathreshold auditory stimuli. We then turn to underlying changes in auditory anatomy and physiology gleaned from invasive studies of nonhuman infants, as well as behavioral evidence for plasticity in these

underlying neural mechanisms in human infants and adults. Next we review how one class of auditory signals produced by the human vocal tract—speech sounds—are perceived by infants and how their underlying discriminative capacities influence the formation and maintenance of speech-sound categories in both infancy and adulthood. Finally, we review how speech sounds are mapped onto words by describing how infants begin to associate sounds with meanings and develop a lexical categorization system that rapidly expands in size and is robust to considerable dialectal variation and other nonphonemic speaker-dependent acoustic variability. We conclude with some speculations about how a cognitive neuroscience approach could reveal the underlying brain mechanisms that make this progression from sounds to words possible.

Fundamental auditory capacities in infancy The goal of studying basic auditory capacities in infancy is not only to document the normative course of auditory development and its underlying neural mechanisms, but also to determine whether higher levels of auditory processing are constrained by lower level immaturities. For example, the speech perception literature (see chapter 8 by Friederici in this volume) has highlighted the sophisticated discriminative skills of very young infants. However, it is unclear whether and how these discriminative capacities are used in a natural language context to form speech categories and organize these categories in auditory memory for use in a lexical (referential) context. Thus basic data on auditory detection and discrimination place a lower bound on when higher level auditory processing can begin and how this higher level processing compares to the exquisite sensitivity of adults to linguistically relevant stimuli. The Onset of Hearing Evidence from the intact fetus and from infants born prematurely suggests that auditory stimuli can elicit changes in heart rate (Lecanuet, GranierDeferre, and Busnel, 1988), eyeblinks (Birnholz and Benacerraf, 1983), gross motor responses (Kisilevsky, Muir,

97

and Low, 1992), and auditory brain-stem responses (ABRs; Hecox and Burkard, 1982) as early as the 28th week of gestation (when ABRs can first be recorded). Behavioral evidence using two techniques—conjugate reinforcement of sucking responses and preferential listening while fixating a visual stimulus—have shown that newborns exhibit auditory preferences (Cooper, 1997). Auditory preferences in newborns reveal two important facts about development. First, any evidence of listening preferences confirms the newborn’s ability to discriminate among two or more classes of sounds (of course, the absence of a preference is uninformative about discriminative capacities). Second, listening preferences in newborns could be due to the effects of auditory experience (and learning) in the womb, analogous to the effects of auditory experience in prehatchling birds (Gottlieb, 1976). The plausibility of auditory experience as a significant influence on the human fetus was bolstered by demonstrations that premature infants have a functional auditory system and by hydrophone recordings showing that both internal (maternal) and external sounds of high intensity and low frequency are detectable in the amniotic fluid surrounding the fetus’s ears (Querleu et al., 1988). DeCasper and Fifer (1980) reported that newborns suck differentially when given the opportunity to listen to their own mother’s voice over the voice of an unfamiliar mother. Thus not only are maternal sounds available to a functioning auditory system in utero, but also these sounds are apparently encoded during some portion of the prenatal period and retained across a 2–3 day perinatal period to affect the newborn’s listening preferences. When given a choice between a lowpass-filtered version of their mother’s voice (similar to the spectral content available in utero) and an unfiltered version of their mother’s voice, newborns prefer the low-pass-filtered version (Fifer and Moon, 1985), suggesting that they have learned the proximal characteristics of the most intense and frequent intrauterine sounds. In addition to a specific preference for the mother’s voice, newborns exhibit a general preference for highly familiar auditory materials. DeCasper and Spence (1986) showed that newborns prefer to listen longer to a familiar story that had been read aloud repeatedly by the mother during the final weeks of pregnancy than to a novel story. The stories were chosen to have different rhythmic structures, and the newborns’ preferences were exhibited despite the presentation of the test stories by an unfamiliar female voice. Similarly, Mehler and associates (1988) and Moon and associates (1993) have shown that newborns prefer their native language to a foreign language with different rhythmic properties. Taken together, these results suggest that newborns have extracted a number of prosodic (rhythmic and intonational) characteristics of auditory input during the last few days or weeks of prenatal development.

98

Newborns raised in nearly all languages also prefer to listen to a pattern of maternal speech called “motherese,” or infant-directed (ID) speech, over the same sentences spoken by the mother in an adult-directed (AD) register (Cooper and Aslin, 1990). Infant-directed speech is characterized by a slow speaking rate, a small number of words per utterance, and large pitch excursions (Fernald, 1985). In contrast to the early evidence for maternal voice and prosodic preferences, the preference for ID speech is not induced by prenatal experience. In the absence of prenatal exposure to ID speech (Cooper and Aslin, 1990) or ID singing in hearing newborns of deaf parents (Masataka, 1999), newborns show clear preferences for these sounds over AD sounds, and newborns prefer hyperarticulated ID speech (despite never being heard in utero) over normally articulated ID speech (Cooper and Cooper, 1999). Thus there are some acoustic characteristics of sounds that are intrinsically preferred by newborns in the absence of any inducing experience. Newborns also appear to spontaneously form categories for speech sounds along the linguistically relevant dimension of lexical form class: function words versus content words. Function words are generally shorter in duration and unstressed compared to the longer and stressed content words. Shi and associates (1999) showed that newborns who were habituated to a list of function words or a list of content words failed to discriminate a shift to a novel list of words from the same category, but succeeded in discriminating a shift to a list from the opposite category. Thus, when multiple acoustic cues are available to differentiate two lexical categories, even newborns can readily do so. However, as we will see in a later section, there are limits to category formation in infants and a paucity of methods for assessing how sophisticated these speech categories are as compared to adults. An important point in the domain of speech is that maternal input to young infants, although hyperarticulated to enhance some auditory cues (e.g., vowel differences; see Kuhl et al., 1997; Liu, Kuhl, and Tsao, 2003), cannot be exaggerated too much, or the exemplars will extend into a different perceptual category. Thus there are limits to how “simplified” the acoustic properties of maternal speech can be when it is presented to infants. Absolute Thresholds The foregoing evidence confirms that sounds of sufficient intensity are perceived not only by newborns but also prenatally. Thus, as long as sounds are above the infant’s hearing threshold, acoustic information along highly salient auditory dimensions is both detected and discriminated. However, if auditory sensitivity is poorer in infants than in adults, then many sounds that are easily detected by adults will not be available to infants for further processing. The available evidence from studies using complex sounds like speech suggests that auditory

fundamentals of developmental neurobiology

thresholds do not impose severe constraints on higher level processing. Three techniques have been used to assess sound thresholds in human infants: the conditioned head-turning procedure (CHP; Moore et al., 1975), the observer-based procedure (OBP; Olsho, Koch, Halpin, and Carter, 1987), and the auditory brain-stem response (ABR; Hecox, 1975). It is beyond the scope of this review to discuss these methodologies in detail, but the interested reader is referred to the works we have cited. The latter two techniques have documented significant improvements in absolute thresholds between birth and 6 months of age, and the conditioned head-turning technique has documented further improvements between 6 months and 2 years of age. The ABR in newborns is 10–15 dB less sensitive than in adults (Hecox, 1975), and the OBP in 2-week-olds (Werner and Gillenwater, 1990) is up to 50 dB less sensitive. By 3 months of age, OBP thresholds are 15–30 dB poorer than in adults (Olsho et al., 1988). The CHP has also provided estimates of absolute thresholds in 6- to 18-month-olds for pure-tone and octave-band noise stimuli (Berg and Smith, 1983; Nozza and Wilson, 1984; Schneider, Trehub, and Bull, 1979, 1980; Sinnott, Pisoni, and Aslin, 1983; Trehub, Schneider, and Endman, 1980). Although estimates vary across studies, by 6 months of age, absolute thresholds are approximately 10– 20 dB poorer than in adults, and these estimates do not differ from those obtained using OBP (Olsho et al., 1988). In summary, while there are significant postnatal improvements in absolute thresholds, many signals in the natural environment are accessible to young infants (and to the fetus) because they are well above threshold. Intensity, Frequency, and Duration Discrimination Given a sound above absolute threshold, one can ask whether the infant auditory system can make fine discriminations of intensity, frequency, and duration—the key acoustic parameters that are used to analyze auditory signals. Using the CHP, thresholds to detect an intensity increment improve from 6 dB to 4 dB between 6 and 12 months of age (Schneider, Bull, and Trehub, 1988; Sinnott and Aslin, 1985). Although these infant thresholds are 2–3 times worse than adults’, most of the critical information carried in speech and nonspeech environmental sounds contains intensity differences that are easily detected by infants. The CHP has also provided evidence that 6-month-old infants can discriminate an increment or decrement in frequency of approximately 2 percent for midfrequency tones (Olsho, 1984; Olsho et al., 1982; Sinnott and Aslin, 1985). The OBP has confirmed this 2 percent threshold in 6-montholds and shown that 3-month-olds have a slightly poorer threshold of 3 percent (Olsho, Koch, and Halpin, 1987). Again, it should be noted that while adult thresholds are slightly less than 1 percent, a 2–3 percent difference in fre-

quency is exceeded by nearly all relevant speech and nonspeech stimuli in the infant’s environment. Other studies using a masking technique to assess infants’ thresholds for detecting a signal (either an octave-band noise or a pure tone) in the presence of either a narrow or broadband masking noise (Bull, Schneider, and Trehub, 1981; Nozza and Wilson, 1984; Olsho, 1985; Schneider, Morrongiello, and Trehub, 1990, 1989; Spetner and Olsho, 1990) have shown that thresholds are elevated by the same relative amount in infants and adults, suggesting that frequency resolution is adultlike by 6 months of age. Perhaps more relevant to everyday listening, Bargones and Werner (1994) have shown that infants do not always appear to selectively attend to the frequency range within which signals are presented. How infants choose the specific acoustic cues to which their processing is directed will be discussed in a later section on cue weighting. Duration discrimination is also important for suprasegmental or prosodic aspects of speech (and phonemic discrimination in languages that make duration distinctions), and temporal acuity is important for the processing of binaural cues to sound localization. Morrongiello and Trehub (1987) used the CHP to assess duration discrimination and reported that 6-month-olds’ thresholds were 25 milliseconds (ms) whereas adults’ were 10 ms. Werner and associates (1992) used OBP to assess gap detection in 3- and 6-montholds and, by altering the low-frequency cutoff of the noise, to determine which frequencies were used by infants in this gap-detection task. Trehub and associates (1995) used the CHP to assess two-tone gap-detection thresholds in 6- and 12-month-olds. These two studies show that gap-detection thresholds decline from 4–5 times longer in 3-month-olds than adults to only a twofold difference by 12 months of age. However, several studies (Irwin et al., 1985; Wightman et al., 1989) have shown that gap detection thresholds continue to improve until at least 5 years and perhaps up to 10 years of age. Duration discrimination also continues to improve well into early childhood (Elfenbein, Small, and Davis, 1993; Jensen and Neff, 1993) and may play a role in the discrimination of stop consonants in medial syllable position where closure duration is a primary cue. A final aspect of duration discrimination is the processing of spectrotemporal cues. Aslin (1989) showed using the CHP that 6- to 9-month-olds’ thresholds for discriminating a rising or falling tone from a steady tone were quite good (similar to the 2% thresholds for frequency increments or decrements). However, when these tone sweeps were rapid (50 ms), thresholds increased by a factor of two, and when infants had to discriminate one rising (or falling) tone from another rising (or falling) tone, their thresholds increased by an additional factor of 3–4. Thus, in the context of broadband spectral information whose components are undergoing rapid changes over time, as in the time-varying

aslin, clayards, and bardhan: mechanisms of auditory reorganization

99

formants of speech, infants are quite poor at discriminating these spectrotemporal cues. This poor performance may not prevent infants from discriminating the major acoustic/phonetic cues that are used to make phonemic distinctions in natural languages, but it may reduce the robustness of speech discrimination under the less than ideal listening conditions outside the laboratory. In addition, many of the subtle timevarying cues in fluent speech may be inaccessible to the infant until spectrotemporal sensitivity and working memory have matured. Summary Fundamental aspects of auditory development undergo substantial improvements in the first few postnatal months. The processing of intensity and frequency information appears to reach values similar to those of adults by 6 months of age. However, temporal processing appears to have a more protracted development continuing at least into the preschool years, and spectrotemporal processing shows similar immaturities. How these basic capacities affect the processing of more complex sounds like speech remains unclear.

Neural specializations and reorganization The foregoing summary of basic acoustic sensitivities and preferences in newborns suggests that a quite sophisticated auditory system, albeit less mature than in adults, is functioning at birth. However, it is unclear whether newborns are analyzing auditory stimuli with the same neural subsystems used by the adult brain. The classic view of how the neocortex is organized in the adult brain is one of functional specialization; that is, separate and discrete regions are devoted to different sensory modalities and to the analysis of specialized features of sensory inputs. To some extent this classic view is correct, with separate pathways from the sensory periphery to modality-specific regions of primary cortex. However, the interaction among cortical areas is both extensive and far-reaching, suggesting that simple models of neural specialization are at best incomplete and perhaps misleading, particularly models that fail to consider the multitude of direct and indirect inputs to any given cortical area. A key developmental question in this debate about neural specialization is one of origins. Where do modality-specific and neural processing specializations come from? The two extremes of this question, of course, comprise the naturenurture debate, in which neural specializations are either an intrinsic property that arose from evolutionary processes or an experience-dependent property that emerged during ontogeny. Neither of these extremes is likely to be correct, and the debate has shifted to understanding whether neural specializations can be modified by interventions during development and perhaps even in adulthood.

100

Auditory Localization in Nonhumans A classic example of specialization in the auditory domain is localization. Most mammals that have a well-developed visual system combine information from both their visual and auditory systems to localize a sound source. Visual stimuli are essentially twodimensional projections of the external world onto a sensory surface (the retina), with very high acuity along the line of central sight for some species (e.g., predatory mammals and birds). Auditory stimuli do not have this spatiotopic representation at the sensory periphery (the cochlea), but rather compute sound location from intensity (sound level) and time of arrival differences at the two ears. In binaural animals, a sound originating from anywhere other than the midline will strike one ear before the other because the ears are separated on the head, producing an interaural time difference (ITD), which is the main cue for azimuthal localization (right-left along the horizontal plane). In addition, because the head sits between the ears of many binaural creatures, one ear lies in the sound shadow of the head, reducing the intensity of the sound at that ear and resulting in an interaural level difference (ILD). These differences, in combination with monaural influences on sound quality (spectral cues) produced by the shape of the external ears (pinnae) and head, provide the information used in sound localization. Perturbing either the ITD, ILD, or spectral cues (see Parsons et al., 1999) can disrupt the ability to localize a sound source. The visual and auditory systems form a useful combination of contrasting characteristics: the auditory system detects sound sources regardless of the orientation of the ears, and the visual system provides detailed information about a stimulus and its surrounding environment in the visual field. A major concern is how these two perceptual systems are integrated so that information from one is able to inform the other for stimulus localization. Visual and auditory information come together for the purpose of stimulus localization in the optic tectum in vertebrates and the superior colliculus in mammals. The tectum contains topographic maps of both visual and auditory space, as well as bimodal neurons that respond to both visual and auditory stimuli (Knudsen and Brainard, 1995). Thus the tectum provides a locus for crossmodal integration of auditory and visual information concerning the location of stimuli in the external world. An example of auditory localization that has clear parallels with humans comes from research on barn owls. In a series of perturbation studies, Knudsen and colleagues altered the correspondence between visual and auditory localization cues in barn owls, either by placing prisms in front of the eyes (Knudsen, 1988) or by monaural occlusion with an earplug (Knudsen, Esterly, and Knudsen, 1984; Knudsen, Knudsen, and Esterly, 1984). Wearing displacing prisms or plugging one of the animal’s ears produces significant alterations in the tectal alignment of receptive fields for

fundamentals of developmental neurobiology

locations in space using auditory (ITD and ILD) and visual cues. If the prism or plug is in place during early development, then a remapping occurs and spatial localization is recalibrated. However, altered experience during adulthood does not lead to a successful remapping. Thus there is a sensitive period for the alignment of auditory and visual spatial cues in the tectum that directly mediates spatial localization responses. If the prism or plug is removed prior to the end of the sensitive period, then a return to “normal” is achieved. However, if the prism or plug is removed after the end of the sensitive period, then realignment is not observed and a permanent mismatch results. Perhaps most importantly, Knudsen (1998) found that owls that had adapted to prisms during the sensitive period and subsequently readapted to normal vision before the close of the sensitive period were able as adults to show subsequent shifts in the tuning of tectal neurons after the reintroduction of the original prismatic displacement. Similar evidence was not found in control animals that had never experienced prismatic displacement. These results suggest that there is a trace of the former adaptive experience that remains in the nervous system long after a different set of adaptive parameters has been adopted, and that the neural pathways that were shaped originally by experience during the sensitive period can be resurrected during adulthood. Auditory Localization in Human Infants Auditory localization in human infants has been studied almost exclusively by noting the presence of overt head-turning responses to discrete sound sources (see review by Clifton, 1992). Despite early anecdotal evidence of auditory localization in a newborn (Wertheimer, 1961), Muir and Field (1979) were the first to report definitive evidence that newborns, under special circumstances, could reliably orient their head toward a sound. Critical factors included supporting the newborn’s head to compensate for poor neck muscle strength and using broadband (preferably high-pass) stimuli of at least 1 second in duration (Clarkson, Clifton, and Morrongiello, 1985, 1989; Morrongiello and Clifton, 1984). These localization responses in the newborn are only crudely spatial; that is, the sound source was located 90 degrees to the right or left of the head along the interaural axis, and the criterion response was any head turn in the direction of the stimulus. However, the presence of bidirectional head turns in newborns, and even in premature infants at 32 weeks of gestation (Muir, 1985), suggests that this crude localization response is not learned from visual feedback. The subsequent postnatal improvement in the accuracy of head turns (Muir, 1985; Muir, Clifton, and Clarkson, 1989) and their greater accuracy in the light (4 degrees) than in the dark (16 degrees) (Morrongiello and Rocca, 1987a) suggests

that visual information serves to improve localization accuracy. However, poor control of the motor system controlling head orientation could mask superior sensitivity to the sensory cues for auditory localization. Studies of infants’ ability to discriminate a change in the spatial location of a sound source (the minimum audible angle, or MAA) reveal a substantial postnatal improvement in the spatial resolution of the auditory system. Studies using both the OBP and conditioned head turning have shown that the MAA in the horizontal plane improves from nearly 30 degrees in 2-month-olds to 9 degrees in 11-month-olds (Ashmead et al., 1991; Morrongiello, 1988; Morrongiello, Fenwick, and Chance, 1990). Further improvements to adult levels (1–2 degrees) occur by 5 years of age (Litovsky, 1991). The MAA in the vertical plane, which is much greater in adults than in the horizontal plane because binaural cues are absent along the midline, is approximately the same as the horizontal MAA in 6-month-olds. However, like adults, the MAA in infants relies on the high-frequency spectral information that comes from the shape of the pinnae (Morrongiello and Rocca, 1987b; Morrongiello, 1987). Importantly, improvements in the horizontal MAA in 4- to 7-month-olds are not the result of improvements in detecting interaural timing differences (Ashmead et al., 1991), because the ability to resolve temporal cues is a factor of two better than the temporal cues that are present in infant MAAs. Rather it appears that the MAA improves because of a tighter sensorimotor mapping, perhaps because in early infancy, motor control of the head is so poor that the associated temporal cues cannot be learned with great reliability. Auditory localization in humans is susceptible to the effects of early deprivation, though detailed occlusion experiments like those conducted with barn owls cannot be performed with human infants. However, Morrongiello (1989) was able to test infants who had unilateral ear infections, both during the infection and after it had been successfully treated. As expected, the accuracy of auditory localization was systematically biased while the infant had a hearing loss in one ear. There was no long-term recalibration of auditory localization, presumably because these episodes of unilateral hearing loss were quite brief. Recalibration of Auditory Localization in Human Adults Although the studies of barn owls support a sensitive period for the calibration (and recalibration) of spatial localization using visual and auditory cues, there are no experimental studies to determine whether a similar sensitive period is present in human infants. However, two recent studies have addressed the question of auditory-visual recalibration in normal adults (Alais and Burr, 2004; Battaglia, Jacobs, and Aslin, 2003). Both these studies

aslin, clayards, and bardhan: mechanisms of auditory reorganization

101

compared the ability of adults to localize a bimodal stimulus consisting of a visual object and a briefly presented auditory noise. Under normal conditions, the spatial location of a visual object and the sound it generates are coincident in space. However, when these two sensory components are spatially dissociated, subjects judge the combined stimulus to be located in the direction of its visual component—a phenomenon called visual capture. The accepted explanation of visual capture is based on the higher spatial resolution of the visual system as compared to the auditory system, which renders judgments of spatial location more reliable, since they are, on average, based primarily on visual information. Alais and Burr (2004) and Battaglia and associates (2003) asked whether this visual dominance was fixed or whether it was susceptible, even in adults, to recalibration. They could not enhance the spatial resolution of the auditory system, so they degraded the visual stimulus so that its spatial location was rendered less reliable. Then they presented the combined visual-auditory stimulus to adults in a cue-conflict situation—the visual stimulus and the auditory stimulus were not located in the same position in space. The question of interest was whether the degree of visual capture was reduced. Both studies reported a significant reduction in the magnitude of visual capture after less than an hour of experience with a visual stimulus whose location in space was made less reliable. Although some of the quantitative aspects of the data from these two studies differed, the important point in the present context is that the weighting of auditory and visual cues for object location in space is still plastic and can be recalibrated in adulthood, at least over short time periods. Cortical Plasticity in Nonhumans The foregoing review of auditory localization provided substantial evidence for plasticity in both infants and adults, even though there may well be a sensitive period for recalibration in infancy. However, auditory localization is largely mediated by midbrain and thalamic neural areas. Because the discrimination and identification of more complex sounds, such as speech, are mediated by cortical areas, it is important to examine their developmental properties and their ability to undergo reorganization. A key question in this regard is whether the mapping of frequency from the auditory periphery to the cortex is invariant during development. The basilar membrane (BM) and the outer hair cells undergo a normal developmental process of stiffening and elongation that results in a shift of the best frequency that is represented at a given locus along the BM. As a result, a pure tone of a given frequency will be represented at a different location along the BM as the animal matures. Hyson and Rudy (1987) showed that rats conditioned to a pure tone respond selectively to that tone immediately after training, but they

102

respond selectively to a frequency-shifted tone several days after training. Thus the cortical representation of the tone is dependent not on its absolute frequency but on the location along the BM from which it originated. The shift in the frequency map between BM and cortex occurs very early in postnatal development and is unlikely to play an important role in humans because their auditory periphery is quite mature at birth. However, other studies have shown that the frequency map in auditory cortex is susceptible to the effects of altered inputs. Robertson and Irvine (1989) showed that the tonotopic map in the primary auditory cortex (A1) of adult guinea pigs undergoes reorganization in response to lesions to restricted areas of the cochlea. In particular, the area of cortex that represents the lesioned region of the cochlea was negatively affected. Immediately after the cochlea was lesioned, there was an increase in the minimum intensity needed to trigger firing in the cortical cells in that area. After several weeks, the area was found to have cells with normal thresholds that responded to frequencies adjacent to those represented in the lesioned area of the cochlea. This reorganization is similar to that found in the somatosensory cortex following amputation (Merzenich et al., 1984; Calford and Tweedale, 1988). Recanzone and associates (1993) investigated changes in the organization of the tonotopic map in primary auditory cortex of adult owl monkeys after extensive frequencydiscrimination training. Monkeys were trained on an auditory frequency discrimination task for several weeks and demonstrated an improvement in their performance with training. Compared to untrained controls and passively stimulated monkeys engaged in a tactile discrimination task, electrophysiological recordings in primary auditory cortex of trained monkeys revealed that the number of recording sites as well as the cortical area representing the frequencies used in training were larger than for frequencies not used for training. This increase in cortical area of representation was significantly correlated with behavioral improvement. However, it is unclear whether these changes in cortical organization reflect changes at the cortical level, the subcortical level, or both. Nevertheless, it seems that some aspects of the auditory system remain plastic into adulthood and that changes in behavior based on auditory experience may reflect changes at the neural level rather than simple response biases. Early acoustic environments play a major role in the organization of auditory cortex. Zhang and associates (2001) found in rat pups that repeated exposure to a pure tone led to a greater representation of frequencies near that tone, as well as a compensatory lack of sharpening in A1-neuron sensitivity at other frequencies. Exposure to a train of pulsed noise (Zhang, Bao, and Merzenich, 2002) yielded similarly broad tuning curves at all frequencies. Noise may also effectively delay the sensitive period for A1 sharpening. Chang

fundamentals of developmental neurobiology

and Merzenich (2003) exposed rat pups to continuous white noise in the time period during which the neurons would normally sharpen their tuning curves in response to relevant frequencies in the environment. When frequency-specific input followed noise exposure, sharpening occurred in adult rats as in naive younger rats. Thus the sensitive period can be extended by broadband early exposure. Another contributor to auditory plasticity is the presence of differential reward. Rutkowski and Weinberger (2005) showed that a conditioned tone stimulus (CS) paired with a water reward led to an expansion in representation of the CS tone within primary auditory cortex of adult rats. Control animals who received the tone but not paired with reward showed no increase in the A1 representation. Not only did the CS tone itself elicit changes in cortical mapping, but there was also an increase in representation for frequencies emitted by the reward delivery equipment. The measurable area of auditory cortex did not differ between trained rats and controls; only the distribution of representations differed. Similarly, Polley and associates (2006) found that the task-dependent properties of the stimulus determine the changes that occur in the cortical representation. In their study two groups of adult rats were exposed to identical stimuli and were trained to attend to either the intensity or the frequency of the sounds. The rats in the frequency condition experienced an expansion of neuronal representation for the frequency range of interest. Similarly, those in the intensity condition had enhanced representation for the particular intensities presented in training. Moreover, the degree of neural retuning was well predicted by each animal’s magnitude of perceptual learning performance. In summary, there is ample evidence from research with nonhumans that the auditory cortex has considerable plasticity, and that some plasticity is present even in adult cortex despite greater plasticity in infancy. We turn now to related findings from humans. Cortical Plasticity in Humans A number of recent studies using noninvasive neuroimaging have confirmed the presence of cortical plasticity in humans. These studies have approached the question of cortical plasticity by examining subjects who have experienced a variety of sensory deprivations, some beginning at birth and others acquired later in life, or by exposing normal adults to a specific auditory training regimen. Lessard and associates (1998) reported evidence of auditory compensation in humans similar to that found in binocularly deprived cats (Korte and Rauschecker, 1993; Rauschecker, 1995). They compared the performance of subjects with normal vision, residual peripheral vision, or total blindness in an auditory localization task. In the binaural condition, totally blind subjects were able to localize at least as well as normals,

and better than subjects with residual vision. In the monaural condition, the totally blind subjects fell into two different categories: those biased to judge a stimulus presented to the occluded ear as originating from the side of the unoccluded ear (a bias shared by normal subjects and subjects with residual vision) and those not. Biased blind subjects performed similar to normals, but showed increased variability for stimuli presented to the occluded ear. However, unlike sighted subjects and those with residual peripheral vision, biased blind subjects reported qualitative differences in sounds presented to the occluded ear as compared to those presented to the unoccluded ear. Unbiased blind subjects performed better with stimuli presented to the occluded ear than both normal subjects and subjects with residual peripheral vision. The fact that unbiased blind subjects outperformed sighted subjects in this condition, and that biased blind subjects reported differences in sound quality even though they did not localize to the correct side, suggests that people blind from birth are able to use monaural cues better than these other two groups in analyzing their auditory environment, perhaps because of neural changes similar to those Korte and Rauschecker (1993) found in binocularly deprived cats. Röder and associates (1999) combined behavioral and electrophysiological methods to investigate auditory spatial tuning in blind and sighted humans and found better auditory localization in blind subjects, but only in peripheral (lateral) auditory space. Normally sighted adults, after a period of recalibration to artificially altered external ears (pinnae), can also adjust their reliance on those spectral cues that are now most reliable for sound localization (Hofman, Van Riswick, and Van Opstal, 1998). Perhaps the most direct measure of auditory cortical plasticity in humans comes from the use of cochlear implants to restore hearing in the deaf. Although this topic is beyond the scope of the present chapter, and the literature on the outcomes of hearing loss are covered in chapter 26 of this volume by Mitchell, a number of key points about the impact of cochlear implants are clear. First, the success of cochlear implants is much higher in children and adults who have already acquired their native language and are then faced with a hearing loss. This result occurs because the implant introduces two types of changes in the mapping of external sound frequencies onto the hair cells that send their signals to the brain. These changes are caused by the inability of the multielectrode wire that is inserted into the cochlea to be positioned all the way to the tip (low-frequency end) of the basilar membrane. As a result, external sound frequencies below approximately 1000 Hz are unmapped, and all higher frequencies stimulate regions of the basilar membrane that would ordinarily be triggered by lower frequency sounds. As a result, the low-frequency sounds that carry crucial information for speech are missing, and the frequen-

aslin, clayards, and bardhan: mechanisms of auditory reorganization

103

cies that the brain does receive are frequency shifted. Thus a postlingual listener fitted with a cochlear implant can attempt to match the frequency-shifted input to stored representations of lexical items. In contrast, the prelingual listener fitted with an implant has no remapping task, but must make sense of stimuli that are missing crucial pieces of lowfrequency speech information. Most adults and postlingual children are able, after weeks or months of listening experience, to adapt to the frequencyshifted sounds and to understand speech. But some implant users never adapt. The situation is made much more complicated for prelingual children because they have never acquired their native language and so have no lexical representations to which they can compare the distorted speech sounds. Although some prelingual deaf children learn to perceive speech, these children are relatively rare (see Pisoni, 2005). More relevant to the present discussion about cortical plasticity is how normal hearing adults adapt to speech that is frequency shifted in a manner that mimics what an implant user would experience. Dorman and associates (1997) reported considerable difficulty in adapting to frequency-shifted speech, and Fitzgerald and associates (2006) showed that the sudden introduction of the frequency shift results in a longer period of adaptation than a gradual introduction. This latter result suggests that cortical mechanisms for adaptation to frequency-shifted speech can better handle small and progressive frequency shifts (much like prism adaptation in the visual-motor domain). What remains unclear is which regions of cortex are involved in this adaptation. There could be a remapping in primary auditory cortex, where detailed maps have been revealed by fMRI in both adult humans (Talavage et al., 2004) and monkeys (Petkov et al., 2006), or to changes in higher-level auditory areas as revealed in a pitch-memory task after five days of training (Gaab, Gaser, and Schlaug, 2006). We will return to this question of levels in the section on adaptation to speech dialects. Taken together, these studies of blind and deaf adults, as well as simulated frequency shifts in normal adults, suggest that both auditory and visual cortex can adjust to altered sensory input either by utilizing regions of cortex that would normally process inputs from a sensory periphery that is now silent or by remapping the sensory or lexical representations. Similar examples have been reported for the activation of visual cortex during tactile discrimination tasks by blind subjects (Sadato et al., 1996) and for the filling-in of primary visual cortex by patients with a congenital absence of three of the four classes of retinal photoreceptors (Baseler et al., 2002). In summary, we have reviewed evidence of behaviorally relevant adult cortical plasticity both on a large scale in humans (recruitment for a new modality) and on a small

104

scale in rats and primates (changes in frequency sensitivity in A1). It is not currently possible to study small-scale cortical plasticity at a neuroanatomical level in humans because of the invasive procedures required. However, we can study changes in human behavioral sensitivity to relevant stimuli in the environment. Changes in phonetic organization throughout the lifetime provide an excellent venue to study small-scale changes in sensitivity because (1) the stimuli are highly behaviorally relevant, (2) the variety of languages provides us with a natural way to contrast the role of the environment with the role of maturation, and (3) the processing of speech stimuli requires the listener to make finegrained distinctions, thereby allowing us to observe very small-scale changes.

Phonetic reorganization Universal Inventories (Human and Nonhuman) The voluminous literature on phonetic discrimination and identification in infants and adults is beyond the scope of this chapter, but it is covered by several excellent recent reviews (Jusczyk, 1997; Kuhl, 2004; Saffran, Werker, and Werner, 2006; Werker and Yeung, 2005), as well as chapter 8 by Friederici in this volume. We wish to make only two major points by way of introduction to the topic of phonetic reorganization. First, infants in all language environments, as well as several species of nonhuman animals exposed to English, show discontinuities in discriminability along many phonetic continua, including continua that are not used contrastively in their language input. This fact has been interpreted as evidence for an innate system of categorization—categorical perception (CP). However, none of these findings from infants include the requisite labeling data that would show a close correspondence between peaks in discriminability and labeling boundaries that defines CP. Moreover, when both discrimination and labeling data are available from nonhumans, they suggest that CP is not unique to humans, and therefore not unique to a linguistic analysis of the speech stimuli (because animals do not have a phonetic system). Second, an innate system of categorization could not match every native language and, therefore, must be adjusted to conform to the details of the native-language phonetic categories. This adjustment process requires exposure to the distributional properties of linguistic input carried by the words to which the infant is exposed. The earliest evidence of such adjustments is present by 6 months of age for vowel perception (Kuhl et al., 1992; Polka and Werker, 1994), followed by language-specific consonant perception at 10–12 months of age (Werker and Tees, 1984). Thus early phonetic reorganization is traditionally characterized as a shift from the language-universal sensitivities shared by all infants, and at least some animals, to language-specific sensitivities.

fundamentals of developmental neurobiology

Distributional Modifications We now turn to the question of what can account for the changes in auditory sensitivity that are observed in the first year of life and the language-specific sensitivities that are observed in adulthood. These changes have been described as a warping of the perceptual space by experience with a particular language environment (Kuhl, 1994; Iverson and Kuhl, 2000). The result is that listeners become more sensitive to particular dimensions of the speech stream (acoustic cues) and less sensitive to others. For example, a comparison of the distributions of acoustic cues from many tokens (instances of words) in English and Japanese reveals that in English, the F3 dimension is relevant for identifying the approximates /l/ and /r/, while in Japanese the F2 dimension is the more relevant (Lotto, Sato, and Diehl, 2004). Iverson and Kuhl (1996) found that speakers of English, German, and Japanese showed different sensitivities in an F2–F3 space for distinguishing English /r/ and /l/. The Japanese speakers exhibited expanded sensitivity in the F2 dimension and reduced sensitivity in the F3 dimension, while English and German speakers exhibited the opposite pattern of perceptual warping. There is still debate about whether the perceptual warping happens at the level of auditory processing or at higher levels that influence behavior but not perception. Evidence that adults can show sensitivity to nonnative contrasts under certain testing conditions that reduce memory load (e.g., Tees and Werker, 1984) suggests it is not a matter of auditory processing. There may be several mechanisms that could cause perceptual warping through exposure. In general they should direct the attention of the listener toward dimensions that are informative and away from dimensions that are not. This allows the listener to optimize his or her use of the acoustic information in identifying the target utterance of the speaker. One way of formalizing this optimization is to apply Bayes’ theorem: listeners should base their judgments on dimensions (cues) to the extent that they are reliable, and this reliability will depend on the variability of those cues in the environment (for a general overview of this Bayesian approach, see Ernst and Bulthoff, 2004). It has been shown in the visual-motor domain that the relative reliabilities of different cues influence how heavily they are weighted by adults (Atkins, Fiser, and Jacobs, 2001; Ernst and Banks, 2002). A similar differential weighting of cues has been shown in the perception of novel nonspeech sounds (Lotto and Holt, 2006; see Smits, Sereno, and Jongman, 2006, for a different categorization theory). Such an account predicts that in learning a native language, the task of the listener is to track the variability of multiple dimensions and develop a cue-weighting strategy that relies most heavily on the dimensions that are the least noisy. Since cue-weighting strategies are language specific, they must be learned. For

example, vowel length plays a greater role in the /i/ versus /I/ distinction in Scottish English than it does in British English (Escudera and Boersma, 2003), and the F3 dimension is more relevant for English approximates such as /r/ and /l/ than for Japanese (Lotto, Sato, and Diehl, 2004). When a language-specific cue-weighting strategy is then applied to speech from a new language that does not utilize the same dimensions, as in the case of native Japanese speakers listening to English /l/ and /r/, listeners may have great difficulty in making accurate judgments. Their cue-weighting strategy has led them to put the most weight on a dimension that is not informative (F2 in this case) and to pay relatively less attention to the dimension that is informative (F3 in this case). Distributional Sensitivity The foregoing cue-weighting account depends on sensitivity to the distributional frequency of tokens, a source of information that has long been known to be informative. Lisker and Abramson (1964) noted that the major cue distinguishing stop consonants along the voicing dimension was voice onset time (VOT). They found that the frequency distributions of VOT across any given language tended to group into 1–3 roughly normal distributions. It is also known that there is a close relationship between the distributional frequency of tokens and phonetic sensitivities. For any given language, the locations along the VOT continuum where very few tokens occur also correspond to category boundaries in that language. It is precisely in these regions of low token probability that we find increased discrimination sensitivity and labeling boundaries (Lisker and Abramson, 1964). The sensitivity of infants to distributional frequency was presumed to account for how innate phonetic categories were adjusted to a specific native language. The empirical demonstration by Saffran and associates (1996) that 8month-olds can perform distributional learning on the sequential properties of speech streams set the stage for a direct test of distributional learning of phonetic categories. Maye and associates (2002) familiarized 6- and 8-month-olds to a range of values along a /d/–/t/ continuum that were either distributed unimodally: with no region of low token probability, or bimodally: with a region of low token probability in the middle of the range of values. After exposing infants to passive listening to these distributions, the researchers found that infants who had been exposed to the bimodal distribution discriminated the phonetic contrast, whereas infants exposed to the unimodal distribution did not. Since all infants heard the same tokens during training, it can only be the frequency of occurrence of particular tokens that formed a distribution along the phonetic continuum that led one group of infants to show evidence of discrimination. Maye and associates (2008) extended these findings to a more difficult phonetic contrast. These results provide

aslin, clayards, and bardhan: mechanisms of auditory reorganization

105

empirical evidence for the theory of distributional learning suggested by Lisker and Abramson (1964). Furthermore, a recent study by Werker and associates (2007) found that Japanese- and English-speaking mothers addressing their 12-month-old infants produced just the kind of languagespecific distributional patterns used in the studies by Maye and associates. It is important to emphasize a key assumption of a distributional theory of phonetic category learning or adjustment: Infants must be sensitive to small (within-category) acoustic differences. If they were only sensitive to the category labels themselves, as presumed by a strong theory of CP, then there would be no distributional information available to the learning mechanism except for the frequency of each category. While this assumption of a distributional theory seems to fly in the face of empirical evidence that infants’ discriminative capacities for withincategory differences are poor, much as they are in adults, recent evidence from both adults and infants provides strong support for within-category sensitivity. McMurray and associates (2002) used an eye-tracking measure to confirm earlier research using rating scales and reaction times and showed that adults are sensitive to within-category differences along a /b/–/p/ continuum. Moreover, the likelihood of falsely labeling a given token along the continuum was monotonically related to the distance from the category boundary; that is, sensitivity was gradient within the phonetic category. Thus adults are sensitive to where along the phonetic dimension a given token resides. McMurray and Aslin (2005) found similar evidence for gradiency in the listening preferences of 8-month-olds exposed to words whose initial /b/ or /p/ consonant was in the center of the phonetic category (i.e., prototypical) versus near the category boundary. Further evidence of sensitivity to within-category differences comes from a set of ERP studies by Rivera-Gaxiola and colleagues showing early or preattentive sensitivity (mismatch negativity, late positive deflections, or N1/P2 components) to within-category differences as well as nonnative contrasts in adults (Rivera-Gaxiola, Johnson et al., 2000; Rivera-Gaxiola, Csibra et al., 2000) and infants at 7 and 11 months (Rivera-Gaxiola et al., 2005) using an oddball paradigm. The result with 11-month-old infants is particularly surprising because this is an age when behavioral sensitivity to nonnative contrasts is reduced and because previous work (Cheour et al., 1998) had not found evidence of nonnative contrast detection. These empirical results on gradient sensitivity provide the requisite evidence that a distributional theory of phonetic category learning and adjustment is plausible. These properties of gradient sensitivity and the role of distributional input in forming and adjusting phonetic categories have recently been captured in computational models by McMurray and associates (in press) and Vallabha and colleagues (2007).

106

Despite the evidence that infants are sensitive to the distributional frequency of tokens, it has not been shown experimentally that infants use distributional frequency to learn to weight a particular acoustic dimension more heavily than another in order to make a phonetic judgment. Little is known about how infants combine multiple acoustic cues or whether they show the same patterns of differential cue-weightings as adults. More is known about young children. Like adults, young children often weight some acoustic cues more heavily than others, but their cueweighting strategies are not always the same as adults and will change throughout childhood (Krause, 1982; Nittrouer and Studdert-Kennedy, 1987; Mayo and Turk, 2004; Morrongiello et al., 1984; Wadrip-Fruin and Peach, 1984). The cause of this developmental shift in cue weighting is unknown. One theory proposes that children process speech more globally and thus place more weight on cues such as formant transitions that occur over larger segments of the speech stream and less weight on cues such as release bursts that could be considered more local phenomena (Nittrouer et al., 2000). However, a study by Mayo and Turk (2004) found that children’s strategies (like adults) differ according to the particular phonetic context of the segment (e.g., /ta/–/da/ versus /ti/–/di/) and do not always favor formant transitions over other cues. Instead they may start out placing more weight than adults on strong (i.e., intrinsically salient) cues and less weight on weak cues, achieving adultlike strategies over time. Further investigation is necessary to determine whether distributional frequencies may be part of the developmental process. Distributional Sensitivity in a Second Language The dramatic changes seen in early infancy are by far the strongest cases of auditory reorganization by exposure to a specific language environment. Moreover, infants raised in a bilingual environment are able to develop and maintain two different underlying phonological systems (Bosch and Sebastian-Galles, 2003). In contrast, research on secondlanguage (L2) acquisition in adults shows that although some modification of sensitivities is possible, it fails to result in native-language proficiency and is very difficult to achieve. It is a well-established observation in second-language learning that some contrasts are hard to learn by adults whereas others are relatively easy (see Jusczyk, 1997, for a review) and that the phonological similarity of L1 and L2 is a major predictor of success on L2. A well-known example is the difficulty that speakers of Japanese have in learning to distinguish between English /r/ and /l/. Speakers of German (which does not have a sound like English /r/) find this contrast relatively easy. Similarly, speakers of English have little difficulty distinguishing between different Zulu clicks (Best et al., 1988) but find it relatively difficult to do the same task with Hindi dental and alveolar stops (Werker et al., 1981).

fundamentals of developmental neurobiology

A variety of models have been proposed to account for this complex pattern of results (Best et al., 1988; Flege, Schirru, and MacKay, 2003; Kuhl, 2004; Werker and Curtin, 2005), but all these models propose that the phonetic categories of the native language are established early in development and constrain the ability of older children and adults to acquire native phonetic proficiency in L2. Crucially, it is thought to be the particular acoustic dimensions involved in the native and nonnative contrasts that determine the amount of L1 interference on L2 (Iverson and Kuhl, 1995). Although it is difficult to distinguish some foreign contrasts, it is possible to improve on this ability through training. Tees and Werker (1984) found that English listeners could learn to discriminate the difficult Hindi dental alveolar stop contrast with training. Similarly, several studies with native Japanese speakers learning English /l/–/r/, show that intensive training can improve discrimination (Strange and Dittmann, 1984), or labeling performance (Lively, Logan, and Pisoni, 1993), but learning only transfers to acoustically different stimuli when training involves multiple tokens of many speakers (Logan, Lively, and Pisoni, 1991). This training regime (called high variability phonetic training or HVPT) is thought to work because subjects get a range of natural variability that allows them to determine the relevant dimensions. McCandliss and associates (2002) showed that by enhancing an acoustic cue to the /r/–/l/ distinction in a training study of native speakers of Japanese, and then gradually fading out this enhanced cue (coined perceptual fading), there was substantial enhancement of discrimination performance, even when subjects were given no feedback during training. Thus enhanced distributional cues may serve a useful role in drawing attention to otherwise difficult-to-discriminate speech contrasts. A recent study directly comparing these two training regimes found equal improvements in labeling performance for both HVPT and perceptual fading groups (Iverson, Hazan, and Bannister, 2005). Thus it is unclear whether high variability in the irrelevant dimension or increased attention to the meaningful variability in the relevant dimension is the more important factor in learning new contrasts. These studies do not explicitly test whether subjects are learning to attend more to one dimension over another. There is evidence, however, that subjects can learn to change their attentional weights. A series of studies by Escudera and colleagues has found that when learning a second language, listeners will adopt a range of cue-weighting strategies at first and may eventually arrive at a strategy that is more appropriate for the linguistic context (Brasileiro and Escudero, 2005; Escudero and Boersma, 2004). Further evidence comes from cue-weight-training studies. Francis and associates (2000) trained listeners to use either the burst cue or the formant transition cue when distinguishing between /ba/, /da/, and /ga/ or /bi/, /di/, and /gi/ syllables. Training

stimuli were cross-spliced so that one of the cues signaled one place of articulation and the other cue signaled a different place of articulation. During training, listeners were given feedback on the “correct” place of articulation. After training, they found an increase in responses consistent with the trained cue across all stimuli. A similar study replicated this finding using an /i/–/I/ vowel contrast and real words and found that individual variability in training success was correlated with sensitivity to the weaker dimension before training (Clayards, Tanenhaus, and Aslin, 2006). These studies demonstrate that listeners can change their cue weightings with training, but they manipulated the relationship between cues and labels, rather than the distribution of tokens. Maye and Gerken (2000) did study distributional learning of phonetic categories in adults and found that exposure to a bimodal distribution allowed for discrimination while exposure to a unimodal distribution did not, just as in infants. Similarly, Clarke and Luce (2005) found that exposure to a shifted distribution of VOT values in sentence context changed listeners’ categorization boundaries. Although no study has directly tested the role of distributional learning in adjusting cue-weighting strategies (rather than category boundaries), Escudera and Boersma (2003) have shown that the distribution of cues available to listeners in the dialect-learning study (Escudera and Boersma, 2004) could predict the optimal cue-weighting strategies that native listeners use. In summary, early universal auditory sensitivities are altered by language-specific phonetic reorganization in infancy. This reorganization is characterized by a warping of the perceptual space to produce increased sensitivity to some dimensions and reduced sensitivity to others. The likely cause of this cue-reweighting process is exposure to the distributional frequencies of tokens along those dimensions in the ambient language. In this way, children gradually develop an adultlike cue-weighting strategy tailored to their linguistic environment (though developing auditory capabilities may also play a role in the developmental progression). In adulthood, when faced with a new language for which the strategy is unsuitable, there is often initial difficulty in distinguishing contrasts. With enough exposure to the new distributions, however, new cue-weighting strategies can develop that at least partially accommodate the new language.

Reference and learning the lexicon Given that infants by the end of the first year have established their native-language phonetic categories, how do they deploy those categories in the service of learning words? It may seem straightforward to think of the phonetic categories as an acoustic alphabet whose elements are simply strung together into a unique sequence (e.g.,

aslin, clayards, and bardhan: mechanisms of auditory reorganization

107

the auditory word form ball) and then mapped onto a real-world referent (a spherical object). Two problems confront the infant listener in solving this seemingly simple task. First, speech discrimination has been assessed in isolated word (or syllable) contexts rather than the fluent multiword speech characteristic of natural linguistic input. Thus there are working memory constraints that influence whether the infant’s discriminative capacities are utilized effectively in recognizing maternal utterances. Second, there is substantial acoustic variability in how talkers (mothers and other adults) produce speech, including gender, speaking rate, prosody, and subtle stress patterns used for emphasis or to convey emotional states. Do infants readily solve these two problems? Evidence from Jusczyk and Aslin (1995) suggests that the first problem is solved by 8 months of age. Infants can extract auditory word forms from fluent speech and recognize these words when presented in isolation. Moreover, infants do not falsely recognize subtle acoustic variants of the words, thereby showing robust evidence for specificity in their auditory memory of those words. However, the second problem has proven more challenging for infants. Stager and Werker (1997) showed that 14-month-olds who easily discriminate a phonetic distinction in a nonreferential context (i.e., when only listening to the phonetic contrast) fail to map these same word forms onto two different visual objects. Initially it was thought that the referential context itself caused infants to relax their criterion for what counts as a meaningful phonetic difference. But more recent evidence suggests that 14month-olds are, indeed, sensitive to the relevant phonetic information, but they fail to associate it with the visual referents. Swingley and Aslin (2000, 2002) showed in an eyetracking paradigm that 14–20-month-olds detect mispronunciations of known words when they are presented with visual referents of those words. And Fennell (2006) showed that this word-referent mapping problem can be alleviated by rendering the visual objects familiar. That is, evidence now suggests that the mapping problem is most acute when both the word forms and the visual referents are unfamiliar. By making the word forms and/or the visual referents familiar, even in isolation, the associative mechanism that links sounds to objects can rapidly perform the mapping task. Thus it is not reference per se that is the bottleneck, but rather the facility with which associations can be formed between sounds and objects—unfamiliar sounds and objects are not easily encoded, and resources devoted to this encoding are not available to form associative linkages. Swingley and Aslin (2007) added support to this view of word learning by showing that new words for novel objects are only acquired rapidly if the words come from sparse regions of the infant’s lexicon. That is, a new word that could

108

be confused with a familiar word inhibits the lexical learning process. And Graf Estes and associates (2007) have shown that novel auditory word forms that have recently been segmented from fluent speech using distributional information can be mapped onto referents in a word-learning task. Thus there is growing evidence that infants have all the acoustic/ phonetic skills needed to map word forms onto referents, provided that the word forms are familiar to auditory memory and not easily confused with similar words already in the lexicon. Dialects and Multiple Representations One interpretation of the foregoing review of word learning is that candidate auditory word forms map transparently onto referents, provided that the word forms and the referents are both familiar perceptual objects. However, such a model presumes that word forms have little acoustic variability, an assumption that we know to be incorrect. Adult listeners are able to cope with widely varying acoustic information and normalize the signal to map it onto their phonological representations in the lexicon in order to access meaning. These nonphonemic acoustic parameters, although not used for decoding the speech signal itself, can provide so-called indexical cues to the speaker’s gender, identity, and accent. How does the infant determine which acoustic cues are phonemically relevant and which cues are indexical? Moreover, do indexical cues interfere with the extraction of phonemic cues? The indexical cue that has received the most attention in both adult and infant speech perception is talker variability. The voice quality of a specific talker is defined by subtle variations in pitch, harmonic structure, rhythm, and vowel color. These are the kinds of indexical cues that play little or no role in decoding the phonemes from the speech signal. However, Allen and Miller (2004) showed that listeners can learn a particular talker’s distribution of voice onset times and use that to identify the talker, provided that there is sufficient experience with a talker and a task that demands tracking VOT as a cue. This ability transferred to novel words they had not heard the speaker produce. McLennan and Luce (2005) discovered that indexical information (talker identity and speaking rate) was used late in processing, after early stages of phonetic processing. Creel and associates (2008) found that talker identity was processed earlier, but only for highly overlearned words and not for newly learned words. And Houston and Jusczyk (2000, 2003) found that infants have long-term memory for the indexical properties of word lists and are negatively affected by talker variability in a word-segmentation task. Thus indexical factors play a role in speech processing both early in development and in adulthood. Presumably, as in the case of cue weighting in phoneme identification, indexical cues are up- or down-weighted depending on their reli-

fundamentals of developmental neurobiology

ability in providing information about a specific property of speech. Talker-specific effects in speech perception are, by definition, linked to particular vocal tracts. A more pervasive indexical variable is the regional dialect to which the infant (or adult) is exposed (Labov, 1998). The ability of one American English listener to understand another from a different dialect implies that a listener can impose a shift, or rotation, of their perceptual vowel space to accommodate the speaker’s productions to arrive at the intended meaning. Maye and associates (in press) exposed listeners to a 20minute passage of an artificial, but plausible, American English dialect wherein front vowels were lowered (e.g., “witch” becomes “wetch”). After exposure, subjects were more likely to classify items within the dialect as words than they had before experiencing the dialect. This was true of items that they had heard in the exposure phase as well as other dialect items that they had not heard. Importantly, subjects learned only the specific direction and magnitude of the vowel shift rather than simply accepting any unusual vowel. Bardhan and associates (2006) found the same results when the spoken items were heard in isolation but paired with visual images, and this paradigm may be applicable to infants and children in the future. In related work, Clarke and Garrett (2004) studied adult listeners’ adaptation to foreign-accented speech. Native speakers of Chinese and Spanish spoke English sentences to which English listeners had to match the sentence-final word. Within a minute of exposure, listeners were as accurate at identifying the target word as those who listened to unaccented speech. These studies suggest that listeners are remarkably efficient in adapting to unfamiliar speech. Unfortunately, to date there are no studies of dialect adaptation in infants or young children, and the mechanism by which adaptation occurs is unknown. Sine-Wave and Noise-Band Speech An extreme form of dialect adaptation involves natural speech that has been degraded in various ways. One technique involves eliminating all the fine spectral detail in speech and replacing it with a parallel series of amplitude-modulated noise bands (Shannon et al., 1995). Adults are able to perform at nearly asymptotic levels in an identification task with these stimuli, provided that there are at least four noise bands. Another technique involves retaining the changing spectral information from the formants of natural speech but replacing these formants with a pure tone that follows the frequency contour of the center frequency of each formant (Remez et al., 1981). Again, adults are able to reliably identify these so-called sine-wave speech stimuli at high levels of accuracy, despite the fact that nearly all the spectral cues of speech have been removed.

An important fact about adults’ perception of sine-wave or noise-band speech is that they are being asked to make judgments of lexical identity from stimuli that are not good exemplars of natural speech. But as in the case of postlingually deafened adults who are fitted with a cochlear implant, these adults have a robust representation of the phonological system of their native language and a lexicon to which nonprototypical speech can be compared. Although adults are also able to recognize sine-wave speech versions of nonwords (Remez, Fellowes, and Nagel, 2007), this ability surely exists, in part, because of their overlearned phonological system. Importantly, many adults, upon first hearing an exemplar of sine-wave speech, perceive it as a set of varying tones and clicks that lack any language-like or phonetic quality. However, when the same exemplar is repeated or when subjects are given instructions that the stimuli are degraded speech, most listeners will suddenly hear them as speech. Again, this repetition effect is undoubtedly due, in part, to the top-down knowledge that adults bring to any speech perception task. In fact, Davis and associates (2005) have shown for noise-band speech that improvement in identification is driven by activation of lexical candidates. These studies of adults with a robust phonological and lexical system raise the question of what qualifies as phonetic to an infant, as well as the question of whether phonetic qualities differ fundamentally from indexical qualities. Vouloumanos and Werker (2004) found that infants as young as 2 months of age preferred to listen to natural speech over sine-wave speech. This finding suggests an early bias toward natural speech that could be based on an innate species-specific preference, a learned preference from early exposure to maternal speech, or a simple preference for broadband stimuli over narrowband stimuli.

Brain correlates of auditory plasticity and learning One of the most seductive, yet vexing, questions in cognitive neuroscience is what brain regions are responsible for language processing. This question will not be answered here, in part because no definitive studies have yet been published and in part because the topic of brain and language is so multifaceted. For example, in mature users of a natural language it is impossible to isolate only one of the many levels of language processing (from phonetics to pragmatics) using either spoken or written materials because even the simplest stimulus (phoneme or letter) is associated with each of these levels in natural language materials (Price, Thierry, and Griffiths, 2005; McCandliss, Cohen, and Dehaene, 2003). Thus activation of a particular brain area, at least in adults, is undoubtedly the outcome of a cascade of processing steps that are triggered by a simple stimulus. Four paradigms have been used in an attempt to gain a foothold on this seemingly

aslin, clayards, and bardhan: mechanisms of auditory reorganization

109

intractable problem. The first is to direct the subject’s attention to a particular level of processing. The second is to elicit repetition effects to passive exposure. The third is to examine brain responses in bilinguals to L1 and L2 speech or in monolinguals to native and nonnative speech. And the fourth is to study infants who have not yet attained higher levels of language processing. Each of these paradigms has advantages and disadvantages. Directing attention to a particular level of processing is fraught with difficulties because it depends on the ability of the subject to control attention and exclude other levels of processing. Moreover, the typical paradigm of using two different classes of stimuli (e.g., speech and music) can never eliminate the possibility that differential brain responses are due to the physical differences between the classes. A more fruitful type of stimulus for controlling the level of processing is one that is perceptually bistable. This paradigm has been used to great effect in the visual modality by eliciting binocular rivalry—a spontaneous fluctuation of two perceptual states—and asking the subject to report changes from one state to the other. Brain responses correlated with that perceptual shift are taken as evidence of separate (in fact, incompatible) attentional states (see Tong, Meng, and Blake, 2006). The sine-wave speech stimuli discussed earlier serve a similar role in the auditory domain. These stimuli do not oscillate between two perceptual states, as in the case of binocular rivalry, but for most listeners they begin as nonspeech and transition (typically rapidly) to intelligible but nonprototypical speech. Vouloumanos and associates (2001) showed that adults have greater activation in several portions of the left temporal lobe and right frontal lobe in response to natural speech than to sine-wave speech, with no areas responding more to sine-wave speech. This finding suggests that these stimuli are both being processed by the same brain regions. Unfortunately, none of the subjects interpreted the sine-wave tokens phonetically. Dehaene-Lambertz and associates (2005) found that the posterior superior temporal sulcus (pSTS) was indeed sensitive to mode of perception (speech versus nonspeech) while listening to sine-wave speech and that the supramarginal gyrus responded to phonemic changes. They suggest that while phonemic and nonphonemic auditory stimuli may be processed simultaneously, the phonemic processing system may inhibit purely acoustic processing. This inhibition presumably can happen only after subjects are aware of the speechlike quality of the stimuli. Möttönen and associates (2004) not only found a greater BOLD response to sine-wave speech in left pSTS in informed subjects (i.e., in speech mode), but also found that these subjects were able to integrate the phonetic information with a visual stimulus, demonstrating the McGurk effect. Liebenthal and associates (2003) found categorical percep-

110

tion for phonetic continua created with sine-wave speech, but only when subjects perceived them phonetically, with left medial and anterior STS activation to this phonetic difference in the speech mode. Finally, Benson and associates (2006) contrasted sine-wave speech with sine-wave nonspeech stimuli, and Uppenkamp and associates (2006) contrasted normal vowels with spectrally matched nonvowels. Again, left STS was differentially activated by the sine-wave speech and normal vowels. Thus the evidence from adults points fairly clearly to the left STS as the locus of phonetic processing, although many other brain regions are activated above baseline to speech and nonspeech stimuli. Repetition effects (decrements and recovery to novelty) are the hallmark of the infant habituation paradigm. Under the assumption of neural adaptation in brain regions involved in stimulus processing, repetition effects have been adapted for use in fMRI (Grill-Spector, Henson, and Martin, 2006). Two recent studies have used this paradigm to study speech perception (Joanisse, Zevin, and McCandliss, 2007; Zevin and McCandliss, 2005). Both these studies used a short-term habituation design in which subjects heard either four identical speech tokens (AAAA) or three identical tokens followed by a novel fourth token (BBBA). This design was based on the mismatch negativity response gathered from scalp electrodes (Dehaene-Lambertz, 1997; Naatanen et al., 1997). Zevin and McCandliss (2005) reported fMRI activations in a broad array of temporal cortex to these four-stimulus epochs, and a left hemisphere bias for the novel BBBA epochs. Novelty responses were also observed in the right frontal cortex and hippocampus, with the former most likely a general novelty response and not a speechspecific response. Joanisse and associates (2007) extended this paradigm to study between-category and withincategory phonetic contrasts. Greater activation was elicited by between- than by within-category contrasts in the BBBA epochs, and these activations were located in the left temporal cortex. This paradigm has great potential for use with infants and young children because it does not require the listener to perform a task and appears to tap basic discriminative capacities. Brain responses in bilinguals have been used to determine whether language fluency is predictive of differential patterns of activation. Kim and associates (1997) gathered fMRI data from both early and late bilinguals, with fluency in L2 greater in the early group. Activations to L1 and L2 speech did not differ in temporal cortex between the two groups, with both languages activating common areas. But activations in frontal cortex differed, with early bilinguals showing an overlapping region of activation to L1 and L2, but late bilinguals showing two separate but partially overlapping regions of activation. The Kim and associates report suggested a sensitive period for the formation

fundamentals of developmental neurobiology

of subcomponents of frontal brain regions for processing speech in different languages. However, Pallier and associates (2003) found no such segregation of L1 and L2 in adults who had acquired Korean as L1 and then were adopted into French homes and acquired L2 without the maintenance of L1. It is not clear if traces of L1 were present but below detection thresholds (as in the case of barn owl sound localization), or whether L2 in the absence of maintaining L1 simply used the same area of frontal cortex. Two other studies have reported that native speech contrasts elicit greater fMRI activations than nonnative contrasts (Callan et al., 2004; Jacquemot et al., 2003). Thus, when one phonological system dominates the other, brain activations may swamp the smaller nonnative (or L2) activations, rendering them invisible as a spatially separate region of activation. Finally, there are structural aspects of the brain, both in terms of white matter density and volumetric properties, that predict the facility with which adults improve their discrimination of nonnative speech contrasts during training (Golestani et al., 2007) and infants progress through the vocabulary spurt (Pujol et al., 2006). These and other structural measures, such as diffusion tensor imaging that estimates the organizational complexity of white matter tracts (see Dubois et al., 2006, and chapter 17 in this volume by Wozniak, Mueller, and Lim), offer the prospect of accounting for individual differences in language proficiency. Finally, ERP, fMRI, and near-infrared spectroscopy (NIRS) responses have been measured from normal infants (see recent review by Dehaene-Lambertz, Hertz-Pannier, and Dubois, 2006, and chapter 8 by Friederici in this volume). Two fMRI studies (Dehaene-Lambertz et al., 2002; Dehaene-Lambertz, Hertz-Pannier et al., 2006) are particularly impressive because they obtained data from normal 3-month-old infants in a 1.5T scanner, which entails solving problems of scanner noise (and hearing safety), movement artifacts, and data analysis in uncooperative subjects. Dehaene-Lambertz and associates (2002) reported greater activation to forward then to backward speech in left temporal areas and right frontal areas. Dehaene-Lambertz, Hertz-Pannier, and associates (2006) presented short sentences and recorded activations that unfolded over time in both an anterior and a posterior direction, emanating from primary auditory cortex. These results suggest that different levels of speech and language processing can be extracted from the same recording session in young infants and that these patterns of activation are similar to those observed in adults (Dehaene-Lambertz, Dehaene et al., 2006). However, many more studies will be required to map out the functional activity of brain areas that respond to language materials and to answer fundamental questions about how these areas are interconnected and change with experience.

Summary and conclusions Our coverage of auditory development and early language learning has admittedly been selective. We have attempted to outline the major events that enable the auditory system to access sounds, beginning prenatally, and select those acoustic components that are useful for solving a particular task. This process of selecting a subset of the available acoustic cues implies either the operation of innate constraints on learning or the ability to rapidly adjust the weights attached to these cues based on their functional utility. We outlined a Bayesian model of cue weighting in which the relative importance of a given cue is proportional to its reliability. Empirical evidence supports this model of cue weighting in the tasks of sound localization and speech perception. Furthermore, we speculate that as infants acquire the referential properties of their native language, they map sounds onto meanings using a similar Bayesian cue-weighting scheme. We also reviewed a number of neural mechanisms that are available to support auditory development. These mechanisms have been studied invasively in animals, but of course studies of language development must be limited to humans. Recent neuroimaging techniques have provided insights about where in the brain activity is present as speech stimuli are being processed. The challenge for the future is to understand how these neural systems develop and become elaborated as the phonological system matures in infancy and the lexicon expands rapidly in childhood. Finally, the mechanisms of adaptive plasticity that enable the lexicon to continue to expand in adulthood but limit nativelike acquisition of a second phonological system will prove crucial in addressing the role of experience in auditory learning. acknowledgments

Preparation of this chapter was made possible, in part, by grants from NIH (HD-37082 and DC05071) and the McDonnell and Packard Foundations to RNA, and by NIH training grant (DC-000035) to the University of Rochester that supported NB. REFERENCES

Alais, D., and D. Burr, 2004. The ventriloquist effect results from near-optimal bimodal integration. Curr. Biol. 14:257–262. Allen, J. S., and J. L. Miller, 2004. Listener sensitivity to individual talker differences in voice-onset-time. J. Acoust. Soc. Am. 115:3171–3183. Ashmead, D., D. Davis, T. Whalen, and R. Odom, 1991. Sound localization and sensitivity to interaural time differences in human infants. Child Dev. 62:1211–1226. Aslin, R. N., 1989. Discrimination of frequency transitions by human infants. J. Acoust. Soc. Am. 86:582–590. Atkins, J. E., J. Fiser, and R. A. Jacobs, 2001. Experiencedependent visual cue integration based on consistencies between visual and haptic percepts. Vision Res. 41:449–461.

aslin, clayards, and bardhan: mechanisms of auditory reorganization

111

Bardhan, N. P., R. N. Aslin, and M. K. Tanenhaus, 2006. Return of the weckud wetch: Rapid adaptation to a new accent. Acoustical Society of America meeting, Providence, RI, June. Bargones, J. Y., and L. A. Werner, 1994. Adults listen selectively; infants do not. Psychol. Sci. 5:170–174. Baseler, H. A., A. A. Brewer, L. T. Sharpe, A. B. Morland, H. Jägle, and B. A. Wandell, 2002. Reorganization of human cortical maps caused by inherited photoreceptor abnormalities. Nature Neurosci. 5:364–370. Battaglia, P. W., R. A. Jacobs, and R. N. Aslin, 2003. Bayesian integration of visual and auditory signals for spatial localization. J. Opt. Soc. Am. [A] 20:1391–1397. Benson, R. R., M. Richardson, D. H. Whalen, and S. Lai, 2006. Phonetic processing areas revealed by sinewave speech and acoustically similar nonspeech. NeuroImage 31:342–353. Berg, K. M., and M. C. Smith, 1983. Behavioral thresholds for tones during infancy. J. Exp. Child Psychol. 35:409–425. Best, C., G. McRoberts, and N. M. Sithole, 1988. Examination of perceptual reorganization for nonnative speech contrasts: Zulu click discrimination by English-speaking adults and infants. J. Exp. Psychol. [Hum. Percept.] 14:345–360. Birnholz, J. C., and B. R. Benacerraf, 1983. The development of human fetal hearing. Science 222:516–518. Bosch, L., and N. Sebastian-Galles, 2003. Simultaneous bilingualism and the perception of a language specific vowel contrast in the first year of life. Lang. Speech 46:217–244. Brasileiro, I., and P. Escudero, 2005. Perceptual development of a Dutch vowel contrast by Brazilian learners. Paper presented at Plasticity in Speech Perception workshop of the ISCA, London, June. Bull, D., B. A. Schneider, and S. E. Trehub, 1981. The masking of octave-band noise by broad band spectrum noise: A comparison of infant and adult thresholds. Percept. Psychophys. 30: 101–106. Calford, M. B., and R. Tweedale, 1988. Immediate and chronic changes in responses of somatosensory cortex in adult flying-fox after digit amputation. Nature 332:446–448. Callan, D. E., J. A. Jones, A. M. Callan, and R. AkahaneYamada, 2004. Phonetic perceptual identification by native- and second-language speakers differentially activates brain regions involved with acoustic phonetic processing and those involved with articulatory-auditory/orosensory internal models. NeuroImage 22:1182–1194. Chang, E. F., and M. M. Merzenich, 2003. Environmental noise retards auditory cortical development. Science 300:498–502. Cheour, M., R. Ceponien, A. Lehtokoski, A. Luuk, J. Allik, K. Alho, and R. Naatanen, 1998. Development of languagespecific phoneme representation in the infant brain. Nature Neurosci. 1:351–353. Clarke, C. M., and M. Garrett, 2004. Rapid adaptation to foreign accented speech. J. Acoust. Soc. Am. 116:3647–3658. Clarke, C. M., and P. A. Luce, 2005. Perceptual adaptation to speaker characteristics: VOT boundaries in stop voicing categorization. In C. T. McLennan, P. A. Luce, G. Mauner, and J. Charles-Luce, eds., University at Buffalo Working Papers on Language and Perception 2:362–366. Clarkson, M. G., R. K. Clifton, and B. A. Morrongiello, 1985. The effects of sound duration on newborns’ head orientation. J. Exp. Child Psychol. 39:20–36. Clarkson, M. G., R. K. Clifton, I. U. Swain, and E. E. Perris, 1989. Stimulus duration and repetition rate influences newborns’ head orientation toward sound. Dev. Psychol. 22:683– 705.

112

Clayards, M., M. K. Tanenhaus, and R. N. Aslin, 2006. Integration of multiple cues based on experience with variation. Tenth International Conference on Laboratory Phonology, Paris, June. Clifton, R. K., 1992. The development of spatial hearing in human infants. In L. A. Werner and E. W. Rubel, eds., Developmental Psychoacoustics, 135–157. Washington, DC: American Psychological Association. Cooper, R. P., 1997. An ecological approach to infants’ perception of intonation contours as meaningful aspects of speech. In C. Dent-Read and P. Zukow-Goldring, eds., Evolving Explanations of Development, 55–85. Washington, DC: American Psychological Association. Cooper, R. P., and R. N. Aslin, 1990. Preference for infantdirected speech in the first month after birth. Child Dev. 61: 1584–1595. Cooper, J. S., and R. P. Cooper, April 1999. Rate affects infants’ attention to speech: Two-month-old infants prefer unusually slow infant-directed speech. Poster presented at the Society for Research in Child Development, Albuquerque, NM. Creel, S. C., R. N. Aslin, and M. K. Tanenhaus, 2008. Heeding the voice of experience: The role of talker variation in lexical access. Cognition 106:633–664. Davis, M. H., I. S. Johnsrude, A. Hervais-Adelman, K. Taylor, and C. McGettigan, 2005. Lexical information drives perceptual learning of distorted speech: Evidence from the comprehension of noise-vocoded sentences. J. Exp. Psychol. [Gen.] 134: 222–241. DeCasper, A. J., and W. P. Fifer, 1980. Of human bonding: Newborns prefer their mothers’ voices. Science 208:1174–1176. DeCasper, A. J., and M. J. Spence, 1986. Prenatal maternal speech influences newborns’ perception of speech sounds. Infant Behav. Dev. 9:133–150. Dehaene-Lambertz, G., 1997. Electrophysiological correlates of categorical phoneme perception in adults. NeuroReport 8:914– 924. Dehaene-Lambertz, G., S. Dehaene, J.-L. Anton, A. Campagne, P. Ciuciu, G. P. Dehaene, I. Denghien, A. Jobert, D. LeBihan, M. Sigman, C. Pallier, and J.-B. Poline, 2006. Functional segregation of cortical language areas by sentence repetition. Hum. Brain Mapping 27:360–371. Dehaene-Lambertz, G., S. Dehaene, and L. Hertz-Pannier, 2002. Functional neuroimaging of speech perception in infants. Science 298:2013–2015. Dehaene-Lambertz, G., L. Hertz-Pannier, and J. Dubois, 2006. Nature and nurture in language acquisition: Anatomical and functional brain-imaging studies in infants. Trends Neurosci. 29:367–373. Dehaene-Lambertz, G., L. Hertz-Pannier, J. Dubois, S. Meriaux, A. Roche, M. Sigman, and S. Dehaene, 2006. Functional organization of perisylvian activation during presentation of sentences in preverbal infants. Proc. Natl. Acad. Sci. USA 103: 14240–14245. Dehaene-Lambertz, G., C. Pallier, W. Serniclaes, L. SprengerCharolles, A. Jobert, and S. Dehaene, 2005. Neural correlates of switching from auditory to speech perception. NeuroImage 24:21–33. Dorman, M. F., P. C. Loizou, and D. Rainey, 1997. Simulating the effect of cochlear-implant electrode insertion depth on speech understanding. J. Acoust. Soc. Am. 102:2993–2996. Dubois, J., L. Hertz-Pannier, G. Dehaene-Lambertz, Y. Cointepas, and D. LeBihan, 2006. Assessment of the early organization and maturation of infants’ cerebral white matter fiber bundles:

fundamentals of developmental neurobiology

A feasibility study using quantitative diffusion tensor imaging and tractography. NeuroImage 30:1121–1132. Elfenbein, J. L., A. M. Small, and M. Davis, 1993. Developmental patterns of duration discrimination. J. Speech Hear. Res. 36: 842–849. Ernst, M. O., and M. S. Banks, 2002. Humans integrate visual and haptic information in a statistically optimal fashion. Nature 415:429–433. Ernst, M. O., and H. H. Bulthoff, 2004. Merging the senses into a robust percept. Trends Cogn. Sci. 8:162–169. Escudero, P., and P. Boersma, 2003. Modeling the perceptual development of phonological contrasts with optimality theory and the gradual learning algorithm. Penn Working Papers in Linguistics 8.1:71–85. Escudero, P., and P. Boersma, 2004. Bridging the gap between L2 speech perception research an phonological theory. Stud. Second Lang. Acquisition 26:551–585. Fennell, C. T., 2006. Infants of 14 months use phonetic detail in novel words embedded in naming phrases. In D. Bamman, T. Magnitskaia, and C. Zaller, eds., Proceedings of the Boston University Child Language Development Conference, 178–189. Somerville, MA: Cascadilla Press. Fernald, A., 1985. Four-month-old infants prefer to listen to motherese. Infant Behav. Dev. 8:181–195. Fifer, W. P., and C. Moon, 1985. The effects of fetal experience with sound. In J. P. Lacanuet, N. A. Krasnegor, W. P. Fifer, and W. P. Smotherman, eds., Fetal Behavior: A Psychobiological Perspective, 351–366. Hillsdale, NJ: Erlbaum. Fitzgerald, M. B., T. A. Morbiwala, and M. A. Svirsky, 2006. Customized selection of frequency maps in an acoustic simulation of a cochlear implant. IEEE EMBS Conference, New York, August. Flege, J., C. Schirru, and I. MacKay, 2003. Interaction between the native and second language phonetic subsystems. Speech Commun. 40:467–491. Francis, A. L., K. Baldwin, and H. C. Nusbaum, 2000. Effects of training on attention to acoustic cues. Percept. Psychophys. 62: 1668–1680. Gaab, N., C. Gaser, and G. Schlaug, 2006. Improvement-related functional plasticity following pitch memory training. NeuroImage 15:255–263. Golestani, N., N. Molko, S. Dehaene, D. LeBihan, and C. Pallier, 2007. Brain structure predicts the learning of foreign speech sounds. Cerebral Cortex 17:575–582. Gottlieb, G., 1976. Conceptions of prenatal development: Behavioral embryology. Psychol. Rev. 83:215–234. Graf Estes, K., J. L. Evans, M. W. Alibali, and J. R. Saffran, 2007. Can infants map meaning to newly segmented words? Statistical segmentation and word learning. Psychol. Sci. 18: 254–260. Grill-Spector, K., R. Henson, and A. Martin, 2006. Repetition and the brain: Neural models of stimulus-specific effects. Trends Cogn. Sci. 10:14–23. Hecox, K., 1975. Electrophysiological correlates of human auditory development. In L. B. Cohen and P. Salapatek, eds., Infant Perception: From Sensation to Cognition, 151–191. New York: Academic Press. Hecox, K., and R. Burkard, 1982. Developmental dependencies of the human auditory evoked response. Ann. NY Acad. Sci. 388:538–556. Hofman, P. M., J. G. A. Van Riswick, and A. J. Van Opstal, 1998. Relearning sound localization with new ears. Nature Neurosci. 1:417–421.

Houston, D. M., and P. W. Jusczyk, 2000. The role of talkerspecific information in word segmentation by infants. J. Exp. Psychol. [Hum. Percept.] 26:1570–1582. Houston, D. M., and P. W. Jusczyk, 2003. Infants’ long-term memory for the sound pattern of words and voices. J. Exp. Psychol. [Hum. Percept.] 29:1143–1154. Hyson, R. L., and J. W. Rudy, 1987. Ontogenetic change in the analysis of sound frequency in the infant rat. Dev Psychobiol. 20:189–207. Irwin, R. J., A. K. R. Ball, N. Kay, J. A. Stillman, and J. Rosser, 1985. The development of temporal acuity in children. Child Dev. 56:614–620. Iverson, P., V. Hazan, and K. Bannister, 2005. Phonetic training with acoustic cue manipulations: A comparison of methods for teaching English /r/–/l/ to Japanese adults. J. Acoust. Soc. Am. 118:3267–3278. Iverson, P., and P. K. Kuhl, 1995. Mapping the perceptual magnet effect for speech using signal detection theory and multidimensional scaling. J. Acoust. Soc. Am. 97:553–562. Iverson, P., and P. K. Kuhl, 1996. Influences of phonetic identification and category goodness on American listeners’ perception of /r/ and /l/. J. Acoust. Soc. Am. 99:1130–1140. Iverson, P., and P. K. Kuhl, 2000. Perceptual magnet and phoneme boundary effects in speech perception: Do they arise from a common mechanism? Percept. Psychophys. 62:874– 886. Iverson, P., P. K. Kuhl, R. Akahane-Yamada, E. Diesch, Y. Tohkura, A. Kettermann, and C. Siebert, 2003. A perceptual interference account of acquisition difficulties for nonnative phonemes. Cognition 87:B47–B57. Jacquenot, C., C. Pallier, D. LeBihan, S. Dehaene, and E. Dupoux, 2003. Phonological grammar shapes the auditory cortex: A functional magnetic resonance imaging study. J. Neurosci. 23:9541–9546. Jensen, J. K., and D. L. Neff, 1993. Development of basic auditory discrimination in preschool children. Psychol. Sci. 4:104–107. Joanisse, M., J. D. Zevin, and B. D. McCandliss, 2007. Brain mechanisms implicated in the preattentive categorization of speech sounds revealed using fMRI and short interval habituation trials. Cerebral Cortex 17:2084–2093. Jusczyk, P. W., 1997. The Discovery of Spoken Language. Cambridge, MA: MIT Press. Jusczyk, P. W., and R. N. Aslin, 1995. Infants’ detection of the sound patterns of words in fluent speech. Cogn. Psych. 29:1–23. Kim, K. H. S., N. R. Relkin, K.-M. Lee, and J. Hirsch, 1997. Distinct cortical areas associated with native and second languages. Nature 388:171–174. Kisilevsky, B. S., D. W. Muir, and J. A. Low, 1992. Maturation of human fetal responses to vibroacoustic stimulation. Child Dev. 63:1497–1508. Knudsen, E. I., 1988. Sensitive and critical periods in the development of sound localization. In S. S. Easter, K. F. Barald, and B. M. Carlson, eds., Message to Mind: Directions in Developmental Neurobiology, 303–319. Sunderland, MA: Sinauer Associates. Knudsen, E. I., 1998. Capacity for plasticity in the adult owl auditory system expanded by juvenile experience. Science 279: 1531–1533. Knudsen, E. I., and M. S. Brainard, 1995. Creating a unified representation of visual and auditory space in the brain. Annu. Rev. Neurosci. 18:19–43. Knudsen, E. I., S. D. Esterly, and P. F. Knudsen, 1984. Monaural occlusion alters sound localization during a sensitive period in the barn owl. J. Neurosci. 4:1001–1011.

aslin, clayards, and bardhan: mechanisms of auditory reorganization

113

Knudsen, E. I., P. F. Knudsen, and S. D. Esterly, 1984. A critical period for the recovery of sound localization accuracy following monaural occlusion in the barn owl. J. Neurosci. 4:1012–1020. Korte, M., and J. P. Rauschecker, 1993. Auditory spatial tuning of cortical neurons is sharpened in cats with early blindness. J. Neurophysiol. 70:1717–1721. Krause, S. E., 1982. Vowel duration as a perceptual cue to postvocalic consonant voicing in young children and adults. J. Acoust. Soc. Am. 71:990–995. Kuhl, P. K., 1994. Learning and representation in speech and language. Curr. Opin. Neurobiol. 4:812–822. Kuhl, P. K., 2004. Early language acquisition: Cracking the speech code. Nature Rev. Neurosci. 5:831–843. Kuhl, P. K., J. E. Andruski, I. A. Chistovich, L. A. Chistovich, E. V. Kozhevnikova, V. L. Ryskina, E. I. Stolyarova, U. Sundberg, and F. Lacerda, 1997. Cross-language analysis of phonetic units in language addressed to infants. Science 277:684–686. Kuhl, P. K., K. A. Williams, F. Lacerda, K. N. Stevens, and B. Lindblom, 1992. Linguistic experience alters phonetic perception in infants by 6 months of age. Science 255:606–608. Labov, W., 1998. The three dialects of English. In M. D. Linn, ed., Handbook of Dialects and Language Variation, 39–81. San Diego: Academic Press. Lecanuet, J. P., C. Granier-Deferre, and M. C. Busnel, 1988. Fetal cardiac and motor responses to octave-band noises as a function of central frequency, intensity and heart rate variability. Early Hum. Dev. 18:81–93. Lessard, N., M. Pare, F. Lepore, and M. Lassonde, 1998. Earlyblind human subjects localize sound sources better than sighted subjects. Nature 395:278–280. Liebenthal, E., J. R. Binder, R. L. Piorkowski, and R. E. Remez, 2003. Short-term reorganization of auditory analysis induced by phonetic experience. J. Cogn. Neurosci. 15:549–558. Lisker, L., and A. S. Abramson, 1964. A cross-language study of voicing in initial stops: Acoustical measurements. Word 20: 384–422. Litovsky, R., 1991. Developmental changes in sound localization precision under conditions of the precedence effect. Unpublished doctoral dissertation, University of Massachusetts, Amherst. Liu, H. M., P. K. Kuhl, and F. M. Tsao, 2003. An association between mothers’ speech clarity and infants’ speech discrimination skills. Dev. Sci. 6:F1–F10. Lively, S. E., J. S. Logan, and D. B. Pisoni, 1993. Training Japanese listeners to identify English /r/ and /l/. II: The role of phonetic environment and talker variability in learning new perceptual categories. J. Acoust. Soc. Am. 94:1242–1255. Logan, J. S., S. E. Lively, and D. B. Pisoni, 1991. Training Japanese listeners to identify English /r/ and /l/: A first report. J. Acoust. Soc. Am. 89:874–886. Lotto, A. J., and L. L. Holt, 2006. Cue weighting in auditory categorization: Implications for first and second language acquisition. J. Acoust. Soc. Am. 119:3059–3071. Lotto, A. J., J. Sato, and R. L. Diehl, 2004. Mapping the task for the second language learner: The case of Japanese acquisition of /r/ and /l/. Paper presented at From Sound to Sense, MIT, June. C-181–186. Masataka, N., 1999. Preference for infant-directed singing in 2day-old hearing infants of deaf parents. Dev. Psychol. 35:1001– 1005. Maye, J., R. N. Aslin, and M. T. Tanenhaus, in press. The Weckud Wetch of the Wast: Lexical adaptation to a novel accent. Cogn. Sci.

114

Maye, J., and L. Gerken, 2000. Learning phoneme categories without minimal pairs. Proceedings of the 24th Annual Boston University Conference on Language Development, 522–533. Somerrille, MA: Cascadilla Press. Maye, J., D. J. Weiss, and R. N. Aslin, 2008. Statistical phonetic learning in infants: Facilitation and feature generalization. Dev. Sci. 11:122–134. Maye, J., J. F. Werker, and L. A. Gerken, 2002. Infant sensitivity to distributional information can affect phonetic discrimination. Cognition 82:B101–B111. Mayo, C., and A. Turk, 2004. Adult-child differences in acoustic cue weighting are influenced by segmental context: Children are not always perceptually biased toward transitions. J. Acoust. Soc. Am. 115:3184–3194. McCandliss, B. D., L. Cohen, and S. Dehaene, 2003. The visual word form area: Expertise for reading in the fusiform gyrus. Trends Cogn. Sci. 7(7):293–299. McCandliss, B. D., J. A. Fiez, A. Protopapas, M. Conway, and J. L. McClelland, 2002. Success and failure in teaching the [r]–[l] contrast to Japanese adults: Tests of a Hebbian model of plasticity and stabilization in spoken language perception. Cogn. Affective Behav. Neurosci. 2:89–108. McLennan, C. T., and P. A. Luce, 2005. Examining the time course of indexical specificity effects in spoken word recognition. J. Exp. Psychol. [Learn. Mem. Cogn.] 31:306–321. McMurray, B., and R. N. Aslin, 2005. Infants are sensitive to within-category variation in speech perception. Cognition 95: B15–B26. McMurray, B., R. N. Aslin, and M. K. Tanenhaus, 2002. Gradient effects of within-category phonetic variation on lexical access. Cognition 86:B33–B42. McMurray, B., R. N. Aslin, and J. Toscano, in press. Statistical learning of phonetic categories: Insights from a computational approach. Dev. Sci. Mehler, J., P. W. Jusczyk, G. Lambertz, N. Halsted, J. Bertoncini, and C. Amiel-Tison, 1988. A precursor of language acquisition in young infants. Cognition 29:144–178. Merzenich, M. M., R. J. Nelson, M. P. Stryker, M. S. Cynader, A. Schoppmann, and J. M. Zook, 1984. Somatosensory cortical map changes following digit amputation in adult monkeys. J. Comp. Neurol. 224:591–605. Moon, C., R. P. Cooper, and W. P. Fifer, 1993. Two-day-old infants prefer their native language. Infant Behav. Devel. 16: 495–500. Moore, J. M., G. Thompson, and M. Thompson, 1975. Auditory localization of infants as a function of reinforcement conditions. J. Speech Hear. Disorders 40:29–34. Morrongiello, B. A., 1987. Infants’ localization of sounds in the medial sagittal plane: Effects of signal frequency. J. Acoust. Soc. Am. 82:900–905. Morrongiello, B. A., 1988. Infants’ localization of sounds in the horizontal plane: Estimates of minimum audible angle. Dev. Psychol. 24:8–13. Morrongiello, B. A., 1989. Infants’ monaural localization of sounds: Effects of unilateral ear infections. J. Acoust. Soc. Am. 86:597–602. Morrongiello, B. A., and R. K. Clifton, 1984. Effects of sound frequency on behavioral and cardiac orienting in newborn and five-month-old infants. J. Exp. Child Psychol. 38:429– 446. Morrongiello, B. A., K. Fenwick, and G. Chance, 1990. Sound localization acuity in very young infants: An observer-based testing procedure. Dev. Psychol. 26:75–84.

fundamentals of developmental neurobiology

Morrongiello, G. A., R. C. Robson, C. T. Best, and R. K. Clifton, 1984. Trading relation in the perception of speech by five-year-old children. J. Exp. Child Psychol. 37:231–250. Morrongiello, B. A., and P. T. Rocca, 1987a. Infants’ localization of sounds in the horizontal plane: Effects of auditory and visual cues. Child Dev. 58:918–927. Morrongiello, B. A., and P. T. Rocca, 1987b. Infants’ localization of sounds in the median vertical plane: Estimates of minimum audible angle. J. Exp. Child Psychol. 43:181–193. Morrongiello, B. A., and S. E. Trehub, 1987. Age-related changes in auditory temporal perception. J. Exp. Child Psychol. 44:413–426. Möttönen, R., G. A. Calvert, I. P. Jääskeläinen, P. M. Matthews, T. Thesen, J. Tuomainen, and M. Sams, 2006. Perceiving identical sounds as speech or nonspeech modulates activity in left posterior superior temporal sulcus. NeuroImage 30: 563–569. Muir, D., 1985. The development of infants’ auditory spatial sensitivity. In S. E. Trehub and B. A. Schneider, eds., Auditory Development in Infancy, 51–84. New York: Plenum Press. Muir, D., R. K. Clifton, and M. G. Clarkson, 1989. The development of human auditory localization response: A U-shaped function. Can. J. Psychol. 43:199–216. Muir, D. and T. Field, 1979. Newborn infants orient to sounds. Child Dev. 50:431–436. Naatanen, R., A. Lehtokoski, M. Lennes, M. Cheour, M. Huotilainen, A. Iivolnen, M. Vaino, P. Alku, R. Ilmoniemi, A. Luuk, J. Allik, J. Sinkkonen, and K. Alho, 1997. Languagespecific phoneme representations revealed by electric and magnetic brain responses. Nature 385:432–434. Nittrouer, S., M. E. Miller, C. S. Crowther, and M. J. Manhart, 2000. The effect of segmental order on fricative labeling by children and adults. Percept. Psychophys. 62:266– 284. Nittrouer, S., and M. Studdert-Kennedy, 1987. The role of coarticulatory effects in the perception of fricatives by children and adults. J. Speech Hear. Res. 30:319–329. Nozza, R. J., and W. R. Wilson, 1984. Masked and unmasked pure tone thresholds of infants and adults: Development of auditory frequency selectivity and sensitivity. J. Speech Hear. Res. 27:613–622. Olsho, L. W., 1984. Infant frequency discrimination as a function of frequency. Infant Behav. Dev. 7:27–35. Olsho, L. W., 1985. Infant auditory perception: Tonal masking. Infant Behav. Dev. 8:371–384. Olsho, L. W., E. G. Koch, E. A. Carter, C. F. Halpin, and N. B. Spetner, 1988. Pure-tone sensitivity of human infants. J. Acoust. Soc. Am. 84:1316–1324. Olsho, L. W., E. G. Koch, and C. F. Halpin, 1987. Level and age effects in infant frequency discrimination. J. Acoust. Soc. Am. 82:454–464. Olsho, L. W., E. G. Koch, C. F. Halpin, and E. A. Carter, 1987. An observer-based psychoacoustic procedure for use with young infants. Dev. Psychol. 23:627–640. Olsho, L. W., C. Schoon, R. Sakai, V. Sperduto, and R. Turpin, 1982. Auditory frequency discrimination in infancy. Dev. Psychol. 18:721–726. Pallier, C., S. Dehaene, J.-B. Poline, D. LeBihan, A.-M. Argenti, E. Dupoux, and J. Mehler, 2003. Brain imaging of language plasticity in adopted adults: Can a second language replace the first? Cerebral Cortex 13:155–161. Parsons, C. H., R. G. Lanyon, J. W. H. Schnupp, and A. J. King, 1999. Effects of altering spectral cues in infancy on horizontal

and vertical sound localization by adult ferrets. J. Neurophysiol. 82:2294–2309. Petkov, C. I., C. Kayser, M. Augath, and N. K. Logothetis, 2006. Functional imaging reveals numerous fields in the monkey auditory cortex. PLoS Biol. 4:1213–1226. Pisoni, D. B., 2005. Speech perception in deaf children with cochlear implants. In D. B. Pisoni and R. E. Remez, eds., The Handbook of Speech Perception, 494–523. Malden, MA: Blackwell. Polka, L., and J. F. Werker, 1994. Developmental changes in perception of nonnative vowel contrasts. J. Exp. Psychol. [Hum. Percept.] 20:421–435. Polley, D. B., E. E. Steinberg, and M. M. Merzenich, 2006. Perceptual learning directs auditory cortical map reorganization through top-down influences. J. Neurosci. 26:4970–4982. Price, C., G. Thierry, and T. Griffiths, 2005. Speech-specific auditory processing: Where is it? Trends Cogn. Sci. 9:271–276. Pujol, J., C. Soriano-Mas, H. Ortiz, N. Sebastian-Galles, J. M. Losilla, and J. Deus, 2006. Myelination of language-related areas of the developing brain. Neurology 66:339–343. Querleu, D., X. Renard, F. Versyp, L. Paris-Delure, and G. Crepin, 1988. Fetal hearing. Eur. J. Obstet. Gynaecol. Reprod. Biol. 29:191–212. Rauschecker, J. P., 1995. Compensatory plasticity and sensory substitution in the cerebral cortex. Trends Neurosci. 18:36–43. Recanzone, G. H., C. E. Schreiner, and M. M. Merzenich, 1993. Plasticity in the frequency representation of primary auditory cortex following discrimination training in adult owl monkeys. J. Neurosci. 13:87–103. Remez, R. E., J. M. Fellowes, and D. S. Nagel, 2007. On the perception of similarity among talkers. J. Acoust. Soc. Am. 122: 3688–3696. Remez, R. E., P. E. Rubin, D. B. Pisoni, and T. D. Carrell, 1981. Speech perception without traditional speech cues. Science 212: 947–950. Remez, R. E., D. Shoretz Nagel, R. L. Piorkowski, J. M. Fellowes, C. Y. Yang, S. Wissig, and A. Batchelder, manuscript in preparation. On the role of auditory quality in consonant perception. Rivera-Gaxiola, M., G. Csibra, M. H. Johnson, and A. Karmiloff-Smith, 2000. Electrophysiological correlates of crosslinguistic speech perception in native English speakers. Behav. Brain Res. 111:13–23. Rivera-Gaxiola, M., M. H. Johnson, G. Csibra, and A. Karmiloff-Smith, 2000. Electrophysiological correlates of category goodness. Behav. Brain Res. 112:1–11. Rivera-Gaxiola, M., J. Silva-Pereyra, and P. K. Kuhl, 2005. Brain potentials to native and nonnative speech contrasts in 7- and 11-month-old American infants. Dev. Sci. 8: 162–172. Roberston, D., and D. R. F. Irvine, 1989. Plasticity of frequency organization in auditory cortex of guinea pigs with partial unilateral deafness. J. Comp. Neurol. 282:456–471. Röder, B., W. Teder-Salejarvl, A. Sterr, F. Rosler, S. A. Hillyard, and H. J. Neville, 1999. Improved auditory spatial tuning in blind humans. Nature 400:162–166. Rutkowski, R. G., and N. M. Weinberger, 2005. Encoding of learned importance of sound by magnitude of representational area in primary auditory cortex. Proc. Natl. Acad. Sci. USA 102:13664–13669. Sadato, N., A. Pascual-Leone, J. Grafman, V. Ibanez, M.-P. Deiber, G. Dold, and M. Hallett, 1996. Activation of the primary visual cortex by Braille reading in blind subjects. Nature 380:526–528.

aslin, clayards, and bardhan: mechanisms of auditory reorganization

115

Saffran, J. R., R. N. Aslin, and E. L. Newport, 1996. Statistical learning by 8-month-old infants. Science 274:1926–1928. Saffran, J. R., J. F. Werker, and L. A. Werner, 2006. The infant’s auditory world: Hearing, speech, and the beginnings of language. In R. Siegler and D. Kuhn, eds., Handbook of Child Development, 6th ed., 58–108. New York: Wiley. Schneider, B. A., D. Bull, and S. E. Trehub, 1988. Binaural unmasking in infants. J. Acoust. Soc. Am. 83:1124–1132. Schneider, B. A., B. A. Morrongiello, and S. E. Trehub, 1990. The size of the critical band in infants, children, and adults. J. Exp. Psychol. [Hum. Percept.] 16:642–652. Schneider, B. A., S. E. Trehub, and D. Bull, 1979. The development of basic auditory processes in infants. Can. J. Psychol. 33:306–319. Schneider, B. A., S. E. Trehub, and D. Bull, 1980. Highfrequency sensitivity in infants. Science 207:1003–1004. Schneider, B. A., S. E. Trehub, B. A. Morrongiello, and L. A. Thorpe, 1989. Developmental changes in masked thresholds. J. Acoust. Soc. Am. 86:1733–1742. Shannon, R. V., F.-G. Zeng, V. Kamath, J. Wygonsky, and M. Ekelid, 1995. Speech recognition with primarily temporal cues. Science 270:303–304. Shi, R., J. F. Werker, and J. L. Morgan, 1999. Newborn infants’ sensitivity to perceptual cues to lexical and grammatical words. Cognition 72:B11–B21. Sinnott, J. M., and R. N. Aslin, 1985. Frequency and intensity discrimination in human infants and adults. J. Acoust. Soc. Am. 78:1986–1992. Sinnott, J. M., D. B. Pisoni, and R. N. Aslin, 1983. A comparison of pure-tone auditory thresholds in human infants and adults. Infant Behav. Dev. 6:3–17. Smits, R., J. Sereno, and A. Jongman, 2006. Categorization of sounds. J. Exp. Psychol. [Hum. Percept.] 32:733–754. Spetner, N. B., and L. W. Olsho, 1990. Auditory frequency resolution in human infancy. Child Dev. 61:632–652. Stager, C. L., and J. F. Werker, 1997. Infants listen for more phonetic detail in speech perception than in word learning tasks. Nature 388:381–382. Strange. W., and S. Dittman, 1984. Effects of discrimination training on the perception of /r–l/ by Japanese adults learning English. Percept. Psychophys. 36:131–145. Swingley, D., and R. N. Aslin, 2000. Spoken word recognition and lexical representation in very young children. Cognition 75:1–20. Swingley, D., and R. N. Aslin, 2002. Lexical neighborhoods and word-form representations of 14-month-olds. Psychol. Sci. 13:480–484. Swingley, D., and R. N. Aslin, 2007. Lexical competition in young children’s word learning. Cogn. Psych. 54:99–132. Talavage, T. M., M. I. Sereno, J. R. Melcher, P. J. Ledden, B. R. Rosen, and A. M. Dale, 2004. Tonotopic organization in human auditory cortex revealed by progressions of frequency sensitivity. J. Neurophysiol. 91:1282–1296. Tees, R. C., and J. F. Werker, 1984. Perceptual flexibility: Maintenance or recovery of the ability to discriminate nonnative speech sounds. Can. J. Psychol. 38:579–590.

116

Tong, F., M. Meng, and R. Blake, 2006. Neural bases of binocular rivalry. Trends Cogn. Sci. 10:502–511. Trehub, S. E., B. A. Schneider, and M. Endman, 1980. Developmental changes in infants’ sensitivity to octave-band noises. J. Exp. Child Psychol. 29:282–293. Trehub, S. E., B. A. Schneider, and J. L. Henderson, 1995. Gap detection in infants, children, and adults. J. Acoust. Soc. Am. 98:2532–2541. Uppenkamp, S., I. S. Johnsrude, D. Norris, W. Marslen-Wilson, and R. D. Patterson, 2006. Locating the initial stages of speechsound processing in human temporal cortex. NeuroImage 31: 1284–1296. Vallabha, G. K., J. L. McClelland, F. Pons, J. F. Werker, and S. Amano, 2007. Unsupervised learning of vowel categories from infant-directed speech. Proc. Natl. Acad. Sci. USA 104: 13273–13278. Vouloumanos, A., K. A. Kiehl, J. F. Werker, and P. F. Liddle, 2001. Detection of sounds in the auditory stream: Event-related fMRI evidence for differential activation to speech and nonspeech. J. Cogn. Neurosci. 13:994–1005. Vouloumanos, A., and J. F. Werker, 2004. Tuned to the signal: The privileged status of speech for young infants. Dev. Sci. 7:270–276. Wadrip-Fruin, C., and S. Peach, 1984. Developmental aspects of the perception of acoustic cues in determining the voicing feature of final stop consonants. Lang. Speech 27:367–379. Werker, J. F., and S. Curtin, 2005. PRIMIR: A developmental model of speech processing. Lang. Learn. Dev. 1:197–234. Werker, J. F., J. H. V. Gilbert, G. K. Humphrey, and R. C. Tees, 1981. Developmental aspects of cross-language speech perception. Child Dev. 52:349–355. Werker, J. F., F. Pons, C. Dietrich, S. Kajikawa, L. Fais, and S. Amano, 2007. Infant-directed speech supports phonetic category learning in English and Japanese. Cognition 103:147–162. Werker, J. F., and R. C. Tees, 1984. Cross-language speech perception: Evidence for perceptual reorganization during the first year of life. Infant Behav. Dev. 7:49–63. Werker, J. F., and H. H. Yeung, 2005. Infant speech perception bootstraps word learning. Trends Cogn. Sci. 9:519–527. Werner, L. A., and J. M. Gillenwater, 1990. Pure-tone sensitivity of 2- to 5-week-old infants. Infant Behav. Dev. 13:355–375. Werner, L. A., G. C. Marean, C. F. Halpin, N. B. Spetner, and J. M. Gillenwater, 1992. Infant auditory temporal acuity: Gap detection. Child Dev. 63:260–272. Wertheimer, M., 1961. Psychomotor coordination of auditory and visual space at birth. Science 134:1692. Wightman, F., P. Allen, T. Dolan, D. Kistler, and D. Jamieson, 1989. Temporal resolution in children. Child Dev. 60:611–624. Zevin, J. D., and B. D. McCandliss, 2005. Dishabituation of the BOLD response to speech sounds. Behav. Brain Funct. 1;4:1–12. Zhang, L. I., S. Bao, and M. M. Merzenich, 2001. Persistent and specific influences of early acoustic environments on primary auditory cortex. Nature Neurosci. 4:1123–1130. Zhang, L. I., S. Bao, and M. M. Merzenich, 2002. Disruption of primary auditory cortex by synchronous auditory inputs during a critical period. Proc. Natl. Acad. Sci. USA 99:2309–2314.

fundamentals of developmental neurobiology

8

Brain Correlates of Language Processing during the First Years of Life ANGELA D. FRIEDERICI

Introduction Children first experience language by calculating phonological and positional regularities of items in speech input. Keen to extract rules from auditory input, the infant around the age of 7 to 8 months demonstrates the ability to abstract beyond the mere probability of occurrence of particular items in a sequence (for a review see Gomez and Gerken, 2000). With this ability, the foundation for acquisition of syntactic rules in the mother tongue is already partly laid. It has been argued, however, that the acquisition of syntax can only be successful if statistical learning is constrained by innate and domain-specific principles of linguistic structures. Parallel to the task of syntactic rule extraction, infants must learn that certain segments of the speech stream carry meaning that refers to objects and actions in the world around them. Since the relations between words and the objects and actions they refer to are arbitrary, this is not an easy task to accomplish. Infants are helped in this task by a number of cues ranging from extralinguistic ones such as pointing and gaze to innerlinguistic cues provided by functional categories such as function words and inflectional morphology. For example, studies have demonstrated that the learning of object names is eased by social cues (e.g., Baldwin and Moses, 2001; Tomasello, 2003). Also, functional categories help in determining whether a particular name belongs to an object (ball) or an action (rolling). (For example, when preceded by a determiner, the likelihood is high that the word refers to an object, “a ball,” but when preceded by an auxiliary the likelihood is high that the word refers to an action, “is rolling.”) In addition, languages differ in the extent to which the different lexical categories are marked by morphology or position in the sentence. (For example, some languages, such as English, have a strict word order, while others, such as German, have relatively free word order but more morphological

marking.) It has been shown that infants are sensitive to functional categories starting at 6 months of age (e.g., Höhle and Weissenborn, 2003; Höhle et al., 2004). I will not review the behavioral studies supporting the different theories (e.g., Wanner and Gleitman, 1982; Pinker, 1984; Jusczyk, 2000; Gleitman and Gleitman, 2000; Werker and Yeung, 2005), but instead will focus on the neurocognitive literature relevant to the topic of how the infant extracts his or her target language from auditory input. Therefore, the goal of the present chapter is to provide information concerning the brain basis underlying early semantic and syntactic processes. One of the important questions in this context is whether the brain mechanisms supporting prosodic, lexical, semantic, and syntactic processes between 4 months and 4 years of life are similar precursors to those of adults, or whether they are qualitatively different. Qualitative differences would indicate support for the discontinuity hypothesis claiming that processes underlying language performance differ between children and adults (Felix, 1994). In contrast, quantitative differences such as changes in latency and duration of processes would provide support for the continuity hypothesis that assumes that processes are principally the same during development and just change quantitatively with development (Gleitman and Wanner, 1982; Pinker, 1984; Weissenborn et al., 1992). Before reviewing the available neurophysiological literature that is relevant to these theories, the different methods used to investigate the brain bases of language processing in early development will be briefly reviewed.

Measuring brain activity in early development The goal of measuring the brain’s reaction to particular input can be achieved by different tools. The most frequently used measures in infants and young children are eventrelated brain potentials (ERPs) as registered with electro-

117

encephalography (EEG). Event-related potentials reflect the brain’s activity in response to a particular stimulus event with a high temporal resolution. Each time-locked average waveform typically shows several positive or negative peaks at particular latencies after stimulus onset, and each peak, or component, has a characteristic scalp distribution. The polarity (negative/positive inflection of the waveform relative to baseline), as well as the latency and scalp distribution of different components, allows us to dissociate cognitive processes associated with them. Changes within the dimensions of the ERP can indicate changes in the cognitive mechanisms they reflect. For example, changes can be interpreted to reflect a slowing down of a particular cognitive process (reflected in the latency), a reduction in the processing demands or efficiency (amplitude) of a positivity or negativity, or a change in the cortical tissue supporting a particular process (topography). A second method is near-infrared spectroscopy (NIRS), also called optical imaging (OI) (Villringer and Chance, 1997). It allows us to examine the cortical hemodynamic response in infants. This method relies on the spectroscopic determination of changes in hemoglobin concentrations in the cerebral cortex resulting from increased regional cerebral blood flow. It assesses the spectroscopic characteristics of the cerebral cortex through the scalp and skull. Changes in light attenuation at different wavelengths greatly depend on the concentration changes in oxygenated and deoxygenated hemoglobin ([oxy-Hb] and [deoxy-Hb]) in the cerebral cortex. This method’s temporal resolution is low as it reflects hemodynamic responses, which are relatively slow. Its spatial resolution depends upon the number of channels measured (Obrig and Villringer, 2003; Okamoto et al., 2004; Schroeter et al., 2004). Another method used to measure the metabolic demand due to neural signaling is functional magnetic resonance imaging (fMRI). The resulting changes in oxygenated hemoglobin, the blood-oxygen-level-dependent (BOLD) contrast is measured in fMRI. This method’s temporal resolution is similar to OT, but its spatial resolution is higher than that of OT. Until now, this measurement has been applied to infants only when they are asleep in the scanner. Thus no measurement tool has really been available for the study of infants and children that combines optimal temporal and spatial resolution as well as feasibility. Results from the respective measures, however, add up to provide insight into the brain bases of language development in its early stages.

Processing sentential prosody Sentential prosody can be defined as the overall intonational pattern of a sentence including the fall and rise of the fundamental frequency in the acoustic input, but also including linguistically relevant pauses inserted in the speech input

118

that mark intonational phrases. An infant’s first exposure to language is based on the processing of phonological information. One of the crucial abilities in this first step in language acquisition is the differentiation of speech from nonspeech auditory input. In an experiment using fMRI, DehaeneLambertz and associates (2002) measured the brain activity evoked by normal (forward) speech and reversed (backward) speech in 3-month-old sleeping and awake infants who were exposed to French language. Forward and backward speech elicited stronger activation in the left-temporal lobe compared to silence. The left-hemisphere activation ranged from the superior temporal gyrus, including Heschl’s gyrus extending to the superior temporal sulcus and the temporal pole, but no significant activation was found in the right temporal lobe. Left-hemispheric asymmetry between forward and backward speech was found for the angular gyrus and precuneus. These data suggest that infants as young as 3 months of age show a left-hemispheric dominance for the processing of fast transitions in auditory input such as speech. Separate analyses of awake infants revealed an additional activation for forward speech compared to backward speech in the right frontal cortex, which was interpreted to reflect higher attention for the normal speech condition in awake infants. While the left hemisphere is dominant for speech processing in right-handed adults, the right hemisphere is responsible for suprasegmental prosodic processing (for a review see Friederici and Alter, 2004). A study comparing normal speech and speech in which the intonation (e.g., the pitch contour) of sentences was removed, leaving formant and spectral information intact, found clear right-hemispheric activation in the temporal cortex for the processing of prosodic information in German-speaking adults (Meyer et al., 2004). Using the same paradigm in Japanese, Homae and associates (2006) investigated 3-month-old sleeping infants. In their study, they applied near-infrared optical topography and observed bilateral activation in the temporoparietal and frontal cortex for both normal and flattened speech. A direct comparison between normal and flattened speech revealed the right temporoparietal cortex as the region supporting pitch information processing. This finding suggests that the right-hemispheric dominance in the processing of sentential prosody seen in adults is established in babies as young as 3 months. One crucial aspect of sentential prosody is that it marks intonational phrase boundaries (e.g., “Peter knows # Mary loves her work;” the # indicates the prosodic break). This feature is important, because intonational phrase boundaries signal syntactic phrase boundaries. Every syntactic phrase boundary is not necessarily marked by prosody, but each intonational phrase boundary is a syntactic boundary. Therefore, once infants are able to detect intonational phrase boundaries, they are equipped with the information needed to structure incoming speech into phrases. In adults, a par-

fundamentals of developmental neurobiology

ticular ERP component has been found to correlate with intonational phrase boundary processing in spoken language—that is, the positive shift. As the intonational phrase boundary signals the end of a phrase, this shift is referred to as closure positive shift (CPS) (Steinhauer, Alter, and Friederici, 1999). The CPS has also been observed for hummed sentences and can, therefore, be taken to reflect purely prosodic processes (Pannekamp et al., 2005). In an ERP study with 8-month-old infants, a CPS similar to what has been observed in adults could be observed for intonational phrase boundaries in spoken language. This finding indicates that the brain mechanism supporting intonational phrase boundary processing is clearly established by the age of 8 months (Pannekamp, Weber, and Friederici, 2006) (figure 8.1). Future studies will have to show whether this mechanism is in place at an even earlier developmental stage.

Stress Patterns of Words Different languages have different rules according to how stress is assigned within multisyllabic words. English, like German, is a stress-based language and has a bias toward a stress-initial pattern for two-syllable words (Cutler and Carter, 1987). French, in

contrast, is a syllable-based language that tends to lengthen the word’s last syllable (Nazzi et al., 2006). Behavioral studies have demonstrated that infants learning English are able to segment disyllabic words with stress on the first syllable from speech input but not those with stress on the second syllable at the age of 7.5 months (Jusczyk, Houston, and Newsome, 1999). This ability was also reported for 9-month-olds learning Dutch (Houston et al., 2000). The ability to segment words with stress on the second syllable in various contexts, however, was only observed by the age of 10.5 months in Englishlearning infants (Jusczyk, Houston, and Newsome, 1999). Therefore, it appears that knowledge about the stress pattern of possible words in a language must be available at about 7 months in order to be used for segmentation. It has been shown behaviorally that infants learning English in particular acquire this knowledge between the ages of 6 and 9 months (Jusczyk, Cutler, and Redanz, 1993). However, as behavioral paradigms require the attention of infants during testing, it may well be that infants have already acquired this knowledge but attentional lapses make this fact undetectable. In contrast, neurophysiological studies using the method of ERPs suggest that infants are indeed sensitive to the preferred stress pattern of their target language as early as 4 to 5 months of age. In these studies, the so-called oddball paradigm or mismatch paradigm was used. In such a paradigm a string of identical stimuli or standard stimuli are presented, and a deviant stimulus is introduced at rare occasions. The brain reacts to the deviant auditory stimulus with a mismatch response that in adults is characterized by a more

Figure 8.1 The closure positive shift (CPS) as an index of processing intonational phrase boundaries. (a) Grand-average ERP for adults at electrode PZ. Vertical line indicates sentence onset. IPh1, IPh2, and IPh3 bars indicate the length of the two intonational phrases and the intervening IPh boundary in sentence type A ([Kevin verspricht Mama zu schlafen] IPh1 [und ganz lange lieb zu sein] IPh2 / [Kevin promises Mom to sleep] [and to be a good boy for a while]) represented as solid line (with one CPS), and the three intonational phrases and two intervening IPh boundaries in

sentence type B ([Kevin verspricht] IPh1 [Mama zu küssen] IPh2 [und ganz lange lieb zu sein] IPh3 / [Kevin promises] IPh1 [to kiss Mom] IPh2 [and to be a good boy for a while] IPh3) represented as dotted line (with CPS1 and CPS2). (b) Grand-average ERP for 8-month-old infants at electrode P4 for sentence type A (solid line) and sentence type B (line) (Modified with permission from Friederici, A. D., 2005. Neurophysiological markers of early language acquisition: From syllables to sentences. Trends Cogn. Sci. 9:481–488.)

Toward identifying the lexical form To achieve lexical knowledge, infants first have to segment words from the auditory stream. Both knowledge about a word’s stress pattern and information about a word’s possible phonotactic structure enable infants to successfully recognize words when they hear the target language.

friederici: brain correlates of language processing during the first years of life

119

negative going wave for the deviant as compared to the standard stimuli (Näätänen et al., 2001) and that in infants is sometimes expressed as a more positive going wave for the deviant stimulus (Weber et al., 2004). Studies that used the mismatch paradigm to investigate the infants’ ability to discriminate different phonemes early during development indicate this general ability very early on and start to show a language-specific discrimination response by the age of 6 months (for a review see Kuhl, 2004). In a recent ERP study (Friederici, Friedrich, and Christophe, 2007), groups of 5-month-old German- or French-learning infants were tested for their ability to discriminate between different stress patterns. A mismatch paradigm was used in which the standard stimuli were disyllabic words with stress on the first syllable (baaba) and the deviant stimuli had the stress on the second syllable (babaa). The data showed that both groups are able to discriminate

between the two types of stress patterns (figure 8.2). However, they differed in the amplitude of the brain response: infants learning German showed a larger effect for the languagenontypical iambic pattern (stress on the second syllable), whereas infants learning French demonstrated a larger effect for the language-nontypical trochaic pattern (stress on the first syllable). This finding suggests that a pattern that is nontypical for a particular language is considered deviant both within the experiment (i.e., a rare stimulus in the set) and with respect to an individual infant’s target language. This finding, in turn, presupposes that infants have already established knowledge about the dominant stress pattern of their target language by the age of 5 months. In an ERP study with infants learning Dutch, Kooijman and associates (2005) found that 10-month-olds recognized two-syllable words with stress on the first syllable when these were presented in continuous speech after they had heard the

Figure 8.2 The mismatch response (MMR) during processing of a word’s stress pattern expressed as a more positive going wave for the deviant stimulus condition (dotted line) as compared to the standard condition (solid line). (a) Grand-average ERPs for trochaic stress pattern (stress on the first syllable) as a deviant

(rarely occurring) stimulus in a train of stimuli with stress on the second syllable. (b) Grand-average ERPs for iambic stress pattern (stress on the second syllable) as a deviant stimulus in a train of stimuli with stress on the first syllable. The MMR is indicated by the arrow.

120

fundamentals of developmental neurobiology

words in isolation. Recognition was reflected in a greater negativity between 350 and 500 milliseconds (ms) over the left hemisphere for familiar words than for unfamiliar words. Phonotactic Knowledge and Lexical Form There are only a few ERP studies that have investigated infants’ phonotactic knowledge, that is, knowledge about the legal combinations of phonemes in syllables or words and about legal positions of syllables in words. Behavioral studies have shown that initial phonotactic knowledge is established by 9 months (Friederici and Wessels, 1993; Jusczyk and Luce, 1994). The so-called preferential looking paradigm used in these studies revealed a preference for auditory input containing stimuli that are phonotactically legal (“str” at word onset as in “street”) in the respective target language compared to those that are illegal (“str” at word offset which is not a possible word ending in English or in Dutch, the language in which in infants were tested by Friederici and Wessels, 1993). However, these behavioral studies cannot resolve whether this phonotactic knowledge is considered to be lexically relevant by infants. One way to test phonotactic knowledge and thereby the lexical status of a given stimulus is by means of the electrophysiological N400 component, that is, a negative-going waveform peaking at around 400 ms. In adults, an N400 effect is reflected in a larger amplitude for words that are semantically incongruous to a given context than for congruous words (see figure 8.3). Moreover, the N400 amplitude is larger for pseudowords than for words, whether they are phonotactically legal or not (for reviews, see Kutas and Federmeier, 2000; Kutas and Van Petten, 1994). A paradigm appropriate for both adults and children is one in which the participant is shown the picture of an object and at the same time is presented with an auditory stimulus that is either a word matching the object’s name or not, or one that is a pseudoword that is phonotactically legal or illegal. Using this paradigm, we observed a developmental change between the ages of 12 and 19 months (Friedrich and Friederici, 2005a, 2005b). The ERP effects in 19-month-olds are quite similar to those of adults, that is, an N400 effect for the incongruous words and phonotactically legal pseudowords but not for phonotactically illegal pseudowords. However, no N400 effects were observed in 12-month-olds. As the N400 is taken to reflect mechanisms of lexicalsemantic integration, these data suggest that at the age of 19 months, both real words and phonotactically legal pseudowords are considered as possible word candidates, but that phonotactically illegal pseudowords have already been excluded from the native language lexicon (Friedrich and Friederici, 2005a). The preceding behavioral and electrophysiological data indicate the presence of initial phonotactic knowledge at 9 months, but this knowledge is not actually used in lexical processing until several months later.

Phonological Familiarity of Words There is another ERP effect suggesting that children are already trying to map sounds onto objects (or pictures of objects) at about 11 months of age. This development has been indicated by a negativity around 200 ms reported for 11-month-olds in response to listening to familiar versus unfamiliar words (Thierry et al., 2003). There are some concerns about the statistical techniques used in this study. An ANOVA was performed to cover every millisecond of recording, and no correction for multiple comparisons was applied, so this fact challenges the authors’ interpretation to some extent. Using a picture-word priming paradigm, our group has found an early frontocentral negativity between 100 and 400 ms in 12- and 14-month-olds for auditory word targets that were congruous with a picture compared to incongruous words (Friedrich and Friederici, 2005a). This early effect was interpreted as a familiarity effect reflecting the fulfillment of a phonological (word) expectation after seeing the picture of an object. At this age, infants seem to have some lexical knowledge, but the specific word form referring to a given object might not be sharply defined, so phonetically similar words are still considered as possible word candidates. This interpretation is supported by the finding that 12- and 14month-olds showed an ERP difference between phonetically dissimilar words they knew, but not between words they knew and phonetically similar words (Friedrich and Friederici, 2005a). The available data thus indicate that phonological and semantic knowledge interact at around 12 months of age.

Semantic processes The adult N400 component is taken to reflect the integration of a lexical element into a semantic context (Kutas and Van Petten, 1994). In the study of semantic processes in infants and young children, the adult N400 has been used as an ERP template pattern against which the ERPs for semantic knowledge and processes during early development are compared. Word Level In a study on the processing of words whose meaning infants either knew or did not know, infants between 13 and 17 months old showed a bilateral negativity for unknown words, but 20-month-olds showed a lefthemispheric negativity (Mills, Coffey-Corina, and Neville, 1997). This result was interpreted as a developmental change toward a hemispheric specialization for word processing. In a more recent study, the effects of word experience (training) and vocabulary size (word production) were tested (Mills, Coffey-Corina, and Neville, 1997). In this word-learning paradigm, 20-month-olds acquired novel words either paired with a novel object or without an object. After training, the infants’ ERPs showed a repetition effect indicated by a

friederici: brain correlates of language processing during the first years of life

121

reduced N200–500 amplitude to familiar and novel unpaired words, whereas an increased bilaterally distributed N200– 500 was found for novel paired words. This finding is taken to indicate that the N200–500 is linked to word meaning. However, it is not entirely clear whether the N200–500 reflects semantic processes only or whether phonological familiarity also plays a role. The interpretation of this early effect as semantic is challenged, given that semantic effects in adults are observed later in reference to the N400. It is possible, however, that the early onset of this effect in infants as compared to adults is due to infants’ relatively small vocabularies. A small vocabulary results in a low number of phonologically possible alternative word forms, allowing the brain to react earlier, after hearing a word’s first phonemes (see earlier section on phonological familiarity). A clear semantic-context N400 effect at the word level has been demonstrated for 14- and 19-month-olds (Friedrich and Friederici, 2005b, 2004). The ERP to words in picture context showed a centroparietal, bilaterally distributed negative-going wave between 400 and 1400 ms, which was more negative for words that did not match the picture context than those that did (see figure 8.3). Compared to adults, this N400-like effect reached significance later and lasted longer. There were also small topographic differences

of the effect as children showed a stronger involvement of frontal electrode sites than adults did. The latency differences suggest slower lexical-semantic processes in children than in adults. The more frontal distribution could mean either that children’s semantic processes are still more image based (adults show a frontal distribution when pictures instead of words are processed; West and Holcomb, 2002) or that children may recruit frontal brain regions associated with attention in adults (Courchesne, 1990) in addition to those subserving semantic processing.

Figure 8.3 The N400 as an index of lexical-semantic processes, here showing a picture-word incongruity effect. Top: Grandaverage ERP at electrode PZ for the different age groups. Note the different mircovolt scales for the different age groups. Bottom: N400 effect (difference between congruent and incongruent words)

as a distributional map. Negativity is coded in dark gray. (Modified with permission from Friederici, A. D., 2005. Neurophysiological markers of early language acquisition: From syllables to sentences. Trends Cogn. Sci. 9:481–488.)

122

Sentence Level The processing of congruous and incongruous words in sentential context has only recently been investigated in children younger than 4 years of age (Silva-Pereyra, Rivera-Gaxiola, and Kuhl, 2005). Previous studies with 5- to 15-year-olds (Holcomb, Coffey, and Neville, 1992) and 6- to 13-year-olds (Hahne, Eckstein, and Friederici, 2004), as well as the study with 3- and 4-year-olds, reported N400-like negativities for semantically anomalous sentences in children of all age groups. In the study with 3and 4-year-olds, the children listened passively to sentence stimuli that were either semantically correct or anomalous (e.g., My uncle will blow the movie) while watching a puppet show (Silva-Pereyra, Rivera-Gaxiola, and Kuhl, 2005).

fundamentals of developmental neurobiology

Whereas 3-year-olds showed a first negativity between 300 and 500 ms, a second negativity between 500 and 800 ms, and a third negativity between 800 and 1000 ms, 4-year-olds demonstrated a first negativity peaking at around 400 ms and a second negativity between 500 and 800 ms. These negativities, which were anteriorly distributed, were interpreted to reflect different semantic mechanisms. The later negativities were taken to reflect sentence closure. The negativity in the time window relevant for the adult N400 was not given any specific interpretation but could be taken to reflect semantic processes. Using the same paradigm with slightly younger children at 30 months, a frontally distributed negativity between 300 and 500 ms with a statistically significant difference between semantically anomalous and nonanomalous sentences was reported (Silva-Pereyra, Klarman, et al., 2005). These data suggest that the semantic processes at the sentential level similar to those reflected by the adult N400 are present between the ages of 30 and 36 months. More recently, an N400-like semantic effect at the sentence level has been reported for children at 19 and 24 months (Friedrich and Friederici, 2005b). This study investigated children learning German and used sentences that were either semantically incorrect (e.g., Die Katze trinkt den Ball/The cat drinks the ball) or semantically correct (The cat drinks the milk). For 19-month-olds, a first negativity was observed between 400 and 500 ms followed by a sustained negativity between 600 and 1200 ms (figure 8.4). In 24month-olds, the negativity was found to start as early as 300 ms and lasted until 1200 ms. For adult listeners, the N400 effect was present between 300 and 800 ms. From these data, it is apparent that the N400 effect in children starts at around the same time as the adult N400, but extends longer, until 1200 ms. The longer duration of the children’s N400 effect suggests that the integration of the object noun into the sentence context requires enhanced efforts. The presence of the N400 effect indicates that before the age of two, children possess lexical representations of verbs specified with respect to their lexical restrictions and that brain mechanisms underlying semantic processes at the sentential level are established in an adultlike manner.

Syntactic processes One of the key questions in early syntax acquisition is whether infants rely on statistical cues of the input or whether

Figure 8.4 The N400 as an index of lexical-semantic integration processes at the sentence level, here showing a verb-object noun incongruity effect. Grand-average ERPs of N400 effect across different age groups. Right: Distributional map of the N400 effect (difference between correct and incorrect condition).

friederici: brain correlates of language processing during the first years of life

123

the acquisition of syntax is guided by some innate rule system during early syntax acquisition (for recent reviews see Gomez and Gerken, 2000; Saffran, Aslin, and Newport, 1996). Because ERP studies that focus on whether infants can detect rule deviations are not available in the literature, we cannot answer the question about the innateness of syntax. However, there are ERP studies that have investigated syntactic processing in 2- to 4-year-old children. These groups will be evaluated for whether and to what extent infants’ ERP patterns resemble those of adults. Studies of sentence processing in adults have shown that violations of syntactic rules are associated with two ERP components: a late, centroparietal positivity (P600) and an earlier left anterior negativity (ELAN). While the ELAN is seen as a fast response to violations of local syntactic rules and the LAN is seen in response to morphosyntactic errors, the P600 has been interpreted to reflect later, more controlled processes of syntactic revision that could either be a syntactic reanalysis of a correct but falsely analyzed syntactic structure or the repair of a syntactically incorrect structure (Friederici, 2002). Currently, only a few ERP studies on syntactic processing during early language acquisition are available. One paradigm has been used to study morphosyntactic violations in English, and the other has been used to examine phrase structure violations in German. The morphosyntactic paradigm, using sentences containing a morphosyntactic violation (My uncle will watching the movie), reported a P600-like positivity for 3- and 4-year-olds (Silva-Pereyra, RiveraGaxiola, and Kuhl, 2005). For slightly younger 30-monthold children, the positivity observed between 600 and 1000 ms did not reach significance (Silva-Pereyra, Klarman, et al., 2005). No LAN effects were observed at any of the ages tested.

However, a recent ERP study investigating the processing of phrase structure violations (e.g., Der Löwe im brüllt/The lion in-the roars) in children at 32 months demonstrated a biphasic ERP pattern consisting of a left-hemispheric negativity around 500 ms and a bilaterally distributed centroparietal late P600 for such syntactically incorrect sentences (Oberecker et al., 2005) (figure 8.5). The left lateralization of the children’s negativity suggests that this component can be interpreted as a child precursor to the ELAN observed in adults for phrase structure violations. Both components, the ELAN and the P600, started later and persisted longer than those observed in adults. The appearance of these syntax-related components indicates that the neural mechanisms of syntactic parsing are present in principle at 32 months, although the processes are clearly slower in children than in adults. When applying the same phrase structure violation paradigm to younger 2-year-olds, the ERPs revealed a late P600 but no ELAN component or any other left-lateralized negativity preceding the P600 (Oberecker and Friederici, 2006) (figure 8.5). Thus it appears that the automatic initial phrase structure building reflected in the ELAN is established later than the late integration processes reflected in the P600. The P600 also seems to precede the development of morphosyntactic processing reflected in the LAN (Silva-Pereyra, RiveraGaxiola, and Kuhl, 2005). Moreover, the data from the German and English studies suggest that the neural mechanisms for phrase structure building (ELAN) might be established earlier during development than those for morphosyntactic processes (LAN).

Figure 8.5 The ELAN-P600 pattern as an index of syntactic processes. ELAN stands for early left anterior negativity and P600 for a late centroparietal positivity. Grand-average ERPs at selected electrodes (F7, PZ) across the different age groups. Note the differ-

ent mircovolt scales between children and adults. (Modified with permission from Friederici, A. D., 2005. Neurophysiological markers of early language acquisition: From syllables to sentences. Trends Cogn. Sci. 9:481–488.)

124

Conclusions The literature reviewed in this chapter demonstrates that we are beginning to formulate a developmental neuroscience of

fundamentals of developmental neurobiology

language processing based on electrophysiological data. The methodological approach of recording ERPs while participants process language input of different types is based on the general assumption that different ERP patterns observed for different types of experimental conditions or different age groups can inform us about the underlying processes (Rugg and Coles, 1995). If qualitatively different patterns of brain activity are observeed, then the conclusion can be drawn that different neural structures and therefore different functional processes are involved, whereas pure quantitative differences are more likely to constitute evidence for the same neural structures and functional processes at different levels of engagement. In the case of ERPs, qualitative differences are correlated with either different spatial distributions of electrical activity or with distinct polarities. Quantitative differences, however, are correlated with different magnitudes of electrical activity (i.e., different amplitudes or time courses but an equal topographic distribution and polarity). The ERP studies investigating language development available thus far indicate that language-related ERP components that reflect lexical-semantic processes (N400), syntactic processes (ELAN–P600), and prosodic processes (CPS) appear to change in their latency and duration from childhood to adulthood, but not in their basic morphology, namely, their polarity or their main topographic distribution. According to the assumptions that we have specified, these ERP data suggest that the functional processes underlying lexical-semantic, syntactic, and prosodic aspects of language comprehension change in their level of engagement and timing, but not in their qualitative parameters. These neurophysiological findings may provide additional evidence for the theoretical debate on language development during which two hypotheses have been put forward. The “discontinuity hypothesis” holds that processes underlying language comprehension and production are qualitatively different in childhood and adulthood (Felix, 1994), and the “continuity hypothesis” assumes that language processes are similar during development and adulthood with observable differences only being quantitative in nature (Gleitman and Wanner, 1982; Pinker, 1984). The ERP findings reviewed could be taken to support the continuity hypothesis according to which the underlying principles of language processing develop continuously from childhood to adulthood and change only quantitatively over time. Moreover, they indicate that major parts of the brain system underlying adult language processing are already installed by the age of 2.5 years.

REFERENCES Baldwin, D. A., and L. J. Moses, 2001. Links between social understanding and early word learning: Challenges to current accounts. Soc. Dev. 10:309–329.

Courchesne, E., 1990. Chronology of postnatal human brain development: Event-related potential, positron emission tomography, myelogenesis, and synaptogenesis studies. In J. W. Rohrbaugh, R. Parasuraman, and R. Johnson, eds., Event-Related Brain Potentials: Basic Issues and Applications, 210–241. New York: Oxford University Press. Cutler, A., and D. Carter, (1987). The predominance of strong initial syllables in the English vocabulary. Comput. Speech Lang. 2:133–142. Dehaene-Lambertz, G., S. Dehaene, and L. Hertz-Pannier, 2002. Functional neuroimaging of speech perception in infants. Science 298:2013–2015. Felix, S., 1994. Maturational aspects of universal grammar. In A. Davis, C. Criper, and A. P. R. Howatt, eds., Interlanguage, 133– 161 Edinburgh: University Press. Friederici, A. D., 2002. Towards a neural basis of auditory sentence processing. Trends Cogn. Sci. 6:78–84. Friederici, A. D., and K. Alter, 2004. Lateralization of auditory language functions: A dynamic dual pathway model. Brain Lang. 89:267–276. Friederici, A. D., M. Friedrich, and A. Christophe, 2007. Brain responses in 4-month-old in fants are already language specific. Curr. Biol. 17:1208–1211. Friederici, A. D., and J. M. I. Wessels, 1993. Phonetic knowledge of word boundaries and its use in infant speech perception. Percept. Psychophys. 54:287–295. Friedrich, M., and A. D. Friederici, 2004. N400-like semantic incongruity effect in 19-months-old infants: Processing known words in picture contexts. J. Cogn. Neurosci. 16:1465–1477. Friedrich, M., and A. D. Friederici, 2005a. Phonotactic knowledge and lexical-semantic processing in one-year-olds: Brain responses to words and nonsense words in picture contexts. J. Cogn. Neurosci. 17:1785–1802. Friedrich, M., and A. D. Friederici, 2005b. Lexical priming and semantic integration reflected in the ERP of 14-month-olds. NeuroReport 16:653–656. Gleitman, L., and H. Gleitman, 2000. Bootstrapping a first vocabulary. In J. Weissenborn and B. Höhle, eds., Approaches to Bootstrapping: Phonological, Lexical, Syntactic and Neurophysiological Aspects of Early Language Acquisition, 79–96. Amsterdam/Philadelphia: John Benjamins. Gleitman, L., and E. Wanner, 1982. Language acquisition: The state of the state of art. In E. Wanner and L. Gleitman, eds., Language Acquisition: The State of the Art, 3–48. Cambridge, UK: Cambridge University Press. Gomez, R. L., and L. A. Gerken, 2000. Infant artificial language learning and language acquisition. Trends Cogn. Sci. 4:178–186. Hahne, A., K. Eckstein, and A. D. Friederici, 2004. Brain signatures of syntactic and semantic processes during children’s language development. J. Cogn. Neurosci. 16:1302–1318. Höhle, B., and J. Weissenborn, 2003. German-learning infants’ ability to detect unstressed closed-class elements in continuous speech. Dev. Sci. 6:122–127. Höhle, B., E. Weissenborn, D. Kiefer, A. Schulz, and M. Schmitz, 2004. Functional elements in infants’ speech processing: The role of determiners in the syntactic categorization of lexical elements. Infancy 5:341–353. Holcomb, P. J., S. A. Coffey, and H. J. Neville, 1992. Visual and auditory sentence processing: A developmental analysis using event-related brain potentials. Dev. Neuropsycho. 8:203–241. Homae, F., H. Watanabe, T. Nakano, K., Asakawa, and G. Taga, 2006. The right hemisphere of sleeping infant perceives sentential prosody. Neurosci. Res. 54:276–280.

friederici: brain correlates of language processing during the first years of life

125

Houston, D. M., P. W. Jusczyk, C. Kuijpers, R. Coolen, and A. Cutler, 2000. Cross-language word segmentation by 9-montholds. Psychon. Bull. Rev. 7:504–509. Jusczyk, P. W., 2000. Bootstrapping from the signal: Some further directions. In J. Weissenborn and B. Höhle, eds., Approaches to Bootstrapping: Phonological, Lexical, Syntactic and Neurophysiological Aspects of Early Language Acquisition, 3–23. Amsterdam/Philadelphia: John Benjamins. Jusczyk, P. W., A. Cutler, and N. J. Redanz, 1993. Infants’ preference for the predominant stress patterns of English words. Child Dev. 64:675–687. Jusczyk, P. W., D. M. Houston, and M. Newsome, 1999. The beginnings of word segmentation in English-learning infants. Cogn. Psych. 39:159–207. Jusczyk, P. W., and P. A. Luce, 1994. Infants’ sensitivity to phonotactic pattern in the native language. J. Mem. Lang. 33: 630–645. Kooijman, V., P. Hagoort, and A. Cutler, 2005. Electrophysiological evidence for prelinguistic infants’ word recognition in continuous speech. Cogn. Brain Res. 24:109–116. Kuhl, P., 2004. Early language acquisition: Cracking the speech code. Nature Rev. Neurosci. 5:831–843. Kutas, M., and K. D. Federmeier, 2000. Electrophysiology reveals semantic memory use in language comprehension. Trends Cogn. Sci. 4:463–470. Kutas, M., and C. K. van Petten, 1994. Psycholinguistics electrified: Event-related brain potential investigations. In M. A. Gernsbacher, ed., Handbook of Psycholinguistics, 83–143. San Diego, CA: Academic Press. Meyer, M., K. Steinhauer, K. Alter, A. D. Friederici, and D. Y. von Cramon, 2004. Brain activity varies with modulation of dynamic pitch variance in sentence melody. Brain Lang. 89:277–289. Mills, D. L., S. Coffey-Corina, and H. J. Neville, 1997. Language comprehension and cerebral specialization from 13 to 20 months. Dev. Neuropsychol. 13:397–445. Näätänen, R., M. Tervaniemi, E. Sussman, P. Päävilainen, and I. Winkler, 2001. Primitive intelligence in the auditory cortex. Trends Neurosci. 24:283–288. Nazzi, T., G. Iakimova, J. Bertoncini, S. Fredonie, and C. Alcantara, 2006. Early segmentation of fluent speech by infants acquiring French: Emerging evidence for crosslinguistic differences. J. Mem. Lang. 54:283–299. Oberecker, R., and A. D. Friederici, 2006. Syntactic ERP components in 24-month-olds’ sentence comprehension. NeuroReport 17:1017–1021. Oberecker, R., M. Friedrich, and A. D. Friederici, 2005. Neural correlates of syntactic processing in two-year-olds. J. Cogn. Neurosci. 17:1667–1678. Obrig, H., and A. Villringer, 2003. Beyond the visible—Imaging the human brain with light. J. Cereb. Blood Flow Metab. 23:1– 18. Okamoto, M., H. Dan, K. Shimizu, K. Takeo, T. Amita, I. Oda, I. Konishi, K. Sakamoto, S. Isobe, T. Suzuki, K. Kohyma, and

126

I. Dan, 2004. Multimodal assessment of cortical activation during apple peeling by NIRS and fMRI. NeuroImage 21: 1275–1288. Pannekamp, A., U. Toepel, K. Alter, A. Hahne, and A. D. Friederici, 2005. Prosody-driven sentence processing: An event-related brain potential study. J. Cogn. Neurosci. 17:407– 421. Pannekamp, A., C. Weber, and A. D. Friederici, 2006. Prosodic processing at sentence level in infants. NeuroReport 17:675–678. Pinker, S., 1984. Language Learnability and Language Development. Cambridge, MA: Harvard University Press. Rugg, M. D., and M. G. H. Coles, 1995. Electrophysiology of Mind: Event-Related Potentials and Cognition. New York: Oxford University Press. Saffran, J. R., R. N. Aslin, and E. L. Newport, 1996. Statistical learning by 8-month-old infants. Science 274:1926–1928. Schroeter, M. L., S. Zysset, M. Wahl, and D. Y. von Cramon, 2004. Prefrontal activation due to Stroop interference increases during development—An event-related fNIRS study. NeuroImage 23:1317–1325. Silva-Pereyra, J., M. Rivera-Gaxiola, and P. K. Kuhl, 2005. An event-related brain potential study of sentence comprehension in preschoolers: Semantic and morphosyntactic processing. Cogn. Brain Res. 23:247–258. Silva-Pereyra, J., L. Klarman, J. L. Lin, and P. Kuhl, 2005b. Sentence processing in 30-month-old children: An event-related potential study. NeuroReport 16:645–648. Steinhauer, K., K. Alter, and A. D. Friederici, 1999. Brain potentials indicate immediate use of prosodic cues in natural speech processing. Nature Neurosci. 2:191–196. Thierry, G., 2003. The affective content of words is primarily processed in the left hemisphere. J. Psychophysiol. 17:237. Thierry, G., M. Vihman, and M. Roberts, 2003. Familiar words capture the attention of 11-month-olds in less than 250 ms. NeuroReport 14:2307–2310. Tomasello, M., 2003. Origins of language. In Constructing a Language: A Usage-Based Theory of Language Acquisition, 8–42. Cambridge, MA: Harvard University Press. Villringer, A., and B. Chance, 1997. Noninvasive optical spectroscopy and imaging of human brain function. Trends Neurosci. 20:435–442. Wanner, E., and L. R. Gleitman, 1982. Language Acquisition: The State of the Art. Cambridge, UK: Cambridge University Press. Weber, C., A. Hahne, M. Friedrich, and A. D. Friederici, 2004. Discrimination of word stress in early infant perception: Electrophysiological evidence. Cogn. Brain Res. 18:149–161. Weissenborn, J., H. Goodluck, and T. Roeper, 1992. Theoretical Issues in Language Acquisition: Continuity and Change in Development. Hillsdale, NJ: Lawrence Erlbaum. Werker, J. F., and H. H. Yeung, 2005. Infant speech perception bootstraps word learning. Trends Cogn. Sci. 9:519–527. West, W. C., and P. J. Holcomb, 2002. Event-related potentials during discourse-level semantic integration of complex pictures. Cogn. Brain Res. 13:363–375.

fundamentals of developmental neurobiology

9

Brain-Behavior Relationships in Early Visual Development BOGDAN F. ILIESCU AND JAMES L. DANNEMILLER

Empirical and theoretical work on the relations between neural maturation and perceptual capabilities during infancy has focused on the major visual pathways. This work has addressed development at the retinal, subcortical, and cortical levels and has attempted to link observed data on visual function, whether behavioral or evoked electrical activity, to changes taking place within the underlying neural substrate. Theories of brain-behavior relations in early visual development typically make inferences either from the anatomical or physiological data to observed behavior or from observed behavior to the anatomical and physiological substrate. In the former case, data on the states of various neural elements within the visual system are used to explain observed behavioral data (e.g., acuity, color vision, or contrast sensitivity). In the latter case, observed behavioral data are used to make inferences about the presence and functioning of various subpopulations of feature-selective neurons within the visual pathway (e.g., orientation or direction selectivity). These theories are developmental examples of what Teller (1984) referred to as linking propositions in visual science, that is, formal propositions that link aspects of visual anatomy and physiology with observed visual capacities and behavior. At the neural level, this period is characterized by variable developmental timetables for processes such as neuroarchitectural transformations, physiological activity, biochemical events, and gene expression. Different structures and functional pathways within the visual system (and many other brain systems) develop at different rates. For example, Atkinson (1998) proposed that the magnocellular pathway may lag slightly behind the parvocellular pathway early in postnatal development. Johnson (1990) and Banton and Bertenthal (1997) based their models on data showing that different layers within primary visual cortex develop at different rates. The regions that process primary motor and sensory information mature earliest, whereas parietal and temporal association cortices that will serve as substrates for spatial attention and basic language skills mature later. Higher order association areas, such as prefrontal cortex responsible for integrative behavior

and decision-making processes, mature last (Gogtay et al., 2004). It is important to realize that in these theories, especially those referring to development at the cortical level, vision in the early postnatal human infant is not simply assumed to be a scaled version of adult vision. It is not simply that there is more noise in the young visual system or that the signals are generally weaker, but rather that the sensory and perceptual information processed by the infant brain is of a different quality than that available in normal human adults. The reasoning behind this claim is that some types of information that rely on specific brain structures for their processing or extraction will never be available to the infant because those structures are not yet capable of extracting that information. While it is well known that postnatal experience plays a large role in visual development, we also review important studies on the roles of intrinsic activity in setting up the neural circuits that experience will later mold. In some ways this information addresses two age-old developmental questions: What structure is present innately in the visual system, and how did that structure arise in the absence of visual experience? The important role played by inhibition in visual processing also provides an opportunity to understand how mature visual processing differs both quantitatively and qualitatively from visual processing early in postnatal life. The purpose of this chapter is to review the available data on the early postnatal development of the visual system as an example of how one might think about these complex brain-behavior relations. We will focus on the previously mentioned observation that different structures and processes in the infant brain develop at different times and with different rates. We will look at the progress made at different levels in understanding neural development and its relations with perception. We will review the anatomical, physiological, biochemical, and genetic data from human and animal visual development research without going into the extensive literature on plasticity and deprivation in early visual development. The interested reader might wish to consult chapter 25 in this volume by Maurer, Lewis, and Mondloch for a discussion on visual system plasticity.

127

Retinal development The absorption of photons of light by the photoreceptors is the first step in the visual cascade that eventually leads to seeing. Why is it important to focus on this first stage of information transmission in the visual pathway? There is a simple answer to this question. Information in the optical image that is lost at this first stage of processing can never be recovered. The only thing that subsequent stages in the visual pathway can do is to operate on the information that leaves the retina; they cannot recreate information (spatiotemporal contrast) that has been lost in the sampling and capture of light by the photoreceptor mosaic. If the anatomical data on which the models are based are accurate and representative, and if the assumptions are valid, then these models tell us importantly that vision can be quantitatively no better than what has been calculated; it can only be equal to or worse than this. It pays, therefore, to understand these models, because they tell us what spatial information is available to the remainder of the infant’s visual system for important tasks like object recognition and the perception of spatial layout. Two factors make it possible to construct reasonably rigorous, quantitative developmental models at the retinal level. First, the contributions of various structures in the eye to the amount of information in the pattern of photoreceptor absorptions are fairly well understood. For example, the diameter and length of photoreceptor inner and outer segments determine what proportion of the incident photons are likely to be absorbed by the photopigment. This is a purely physical/optical calculation like calculating the optical aperture of a telescope. Second, anatomical data exist on the characteristics of these optical and retinal structures. One can use these data to calculate estimates of information transmission to and including the point at which photons are actually absorbed by the photopigment. Examples of the models that have been proposed at this level are ones covering scotopic (night) vision (Brown, Dobson, and Maier, 1987; Hansen and Fulton, 1999) and photopic (day) vision (Banks and Bennett, 1988; Banks and Crowell, 1993; Brown, 1993; Brown, Dobson, and Maier, 1987; Candy, Crowell, and Banks, 1998; Wilson, 1988, 1993). The Hansen and Fulton model is somewhat simpler, because it is meant to explain only the large differences that exist between the absolute thresholds of infants and adults in the detection of light. Absolute threshold is simply a measure of the minimum amount of light that can be detected reliably under a set of fixed viewing conditions. We will turn next to the Hansen and Fulton (1999) model because it illustrates how anatomical measurements can be used to make reasonably accurate predictions of the development of visual sensitivity.

128

Absolute thresholds are significantly higher in young infants than in adults (see Brown, 1990, for a review). Why is this statement true? It would be possible to answer this question in many ways. The answer to this simple question illustrates one of the primary virtues of these models. One could point to the weaker attentional capacities of young infants or to their lack of motivation for behaving like practiced psychophysical observers as explanations for why their thresholds are higher. Instead, these models choose to answer this question by using known anatomical data to constrain best or “ideal” visual performance. In other words, if one is going to explain improvements in absolute threshold, perhaps the place to start is at the very beginning, in the photoreceptors themselves. Only after estimates have been derived using these anatomical data, should one then proceed to attribute any remaining infant/adult differences to higher level attentional or motivational processes. These models have the virtue of simplicity; more complex, psychological processes should only be entertained as explanations for some observed behavior after simpler, more peripheral explanations have been ruled out as being incomplete. With postmortem anatomical measurements Fulton and associates (1999) showed that the amount of rhodopsin—the photosensitive pigment in rods—is much lower in preterm and term infant eyes than it is in adult eyes, and that it increases rapidly during infancy. At birth, rhodopsin content is approximately 35 percent of its adult value, and it reaches 50 percent of its adult value by 5 weeks postnatally. It appears to reach its adult value sometime near 40 weeks of age. These increases in rhodopsin are likely to reflect the increases in the lengths of the rod outer segments (OS) in which the rhodopsin is held (Hendrickson, 1994). Hansen and Fulton (1999) noted that rod outer segment lengths at a peripheral retinal site in adults are approximately 2.3 times longer than they are in 5-day-old infants. Given that the amount of rhodopsin should be directly proportional to the length of the rod OS (assuming equal densities), this finding would predict a threshold increase for infants relative to adults of 0.4 log units. In their study of the development of absolute thresholds, Hansen and Fulton (1999) reported that the thresholds for 10-week-old infants were approximately 0.58 log units higher than they were in adults. At a more parafoveal retinal site, rod OS lengths in adults are approximately 9 times as long as they are in 5-day-old human infants. This figure would predict a difference of approximately one log unit. The 10-week-old absolute threshold at this parafoveal site was approximately 1.06 log units higher than the adult threshold. Finally, at 11 months of age, infant rod OS lengths are approximately 68 percent of the adult value, leading to the prediction that infant thresholds at this age should be only 0.16 log units higher than adult thresholds. Hansen and Fulton (1999) noted that no infant in their study who was 6 months of age

fundamentals of developmental neurobiology

or older had an absolute threshold that was more than 0.1 log units above the median adult threshold. These data show quite simply how increases in rod OS length and by inference in the amounts of photon-catching rhodopsin (Fulton et al., 1999) can be used to explain some of the differences between young infants and adults in their absolute thresholds and how those differences narrow over the course of postnatal development (see also Nusinowitz et al., 1998). The same strategy has been used by Banks and Bennett (1988), Banks and Crowell (1993), Brown, Dobson, and Maier (1987), Brown (1993), Candy, Crowell, and Banks (1998), and Wilson (1988, 1993) to model the development of visual acuity and contrast sensitivity. These models also include the spatial distribution of the cone photoreceptors because this distribution differs significantly across the retina in adults, and it has strong implications for photon capture and spatial contrast transfer from the retina. We will concentrate here on the models proposed by Banks and colleagues, because Banks and Crowell (1993) showed that many of the conclusions offered by these alternative models regarding front-end constraints on vision in early infancy are very similar. These models start with the optics of the eye. Prior to being absorbed by the photoreceptors, photons must of course pass through the optics of the eye. Several factors contribute to the optical quality of retinal images. For the purposes of this chapter it is sufficient to note two things. First, Candy, Crowell, and Banks (1998) and Banks and Bennett (1988) both assume that the optical transfer function of the neonatal eye is similar to that of the adult eye. More recent evidence suggests that there is a modest degradation of image quality caused by the optics in 5- to 7-week-old human infants (Wang and Candy, 2005). The optical transfer function represents the extent to which contrast is attenuated as it passes through the optics of the eye (cornea, lens, and ocular media) as a function of spatial frequency. Optical systems generally attenuate contrast more at high spatial frequencies than at low spatial frequencies. The second factor that is important for the purposes of this chapter is that the length of the infant’s eye is shorter than the length of the adult’s eye. This fact has two consequences: (1) the retinal image is spread over a smaller area in the infant’s eye, and (2) the number of photons falling on a small patch of retina (photons/degree2) would be higher in the infant’s eye if pupil sizes did not differ. The latter intensity effect is offset by the smaller pupil size of the infant’s eye, so that the actual number of incident photons per patch of retina is probably not very different between infants and adults, except for small differences resulting from the slightly higher media transmittance in the neonate (Candy, Crowell, and Banks, 1998). Candy, Crowell, and Banks (1998) estimated that the image of a small, distant object would be approximately 2/3 as large on the neonatal retina as on the adult

retina. If photoreceptor packing densities were the same at these two ages, and they most certainly are not (see discussion later in this section), then this difference in eye size alone would spread the image over fewer photoreceptors in the neonate’s eye, leading to a less detailed initial encoding. Once the optical properties of the cornea, lens, and media have been factored into the model, the most important remaining factors are (1) the morphologies of individual cone photoreceptors and (2) their spatial arrangements (packing densities). Data on the development of these properties and their mature, adult values are available from several sources (Abramov et al., 1982; Curcio, 1987; Curcio et al., 1990; Hendrickson, 1994; Hendrickson and Drucker, 1992; Yuodelis and Hendrickson, 1986). It should be noted, however, that these anatomical data used to model spatial contrast vision during early infancy come from one 5-dayold human infant (Yuodelis and Hendrickson, 1986), so the issue of the representativeness of this example must always be kept in mind (Candy, Crowell, and Banks, 1998). It is known that in adults there is considerable individual variability in some of the anatomical parameters used in these models (Curcio et al., 1990). The morphologies of individual cones in the neonatal retina are markedly different from those of adults (Youdelis and Hendrickson, 1986). As was true of rods, the outer segments of the cones are much shorter in the neonate than in the adult. Banks and Bennett (1988) modeled this difference as a factor of approximately 16 : 1 for foveal cones. Of course, this length difference implies that the amount of photosensitive material for capturing incident photons is that much less in the neonate, assuming equal densities. Additionally, the effective apertures of these cones through which photons must be funneled to be absorbed was estimated by Banks and Bennett (1988) to differ by a factor of approximately 1.88 (0.48/0.35)2 in favor of infants. The slightly larger aperture of the neonate’s cones means that slightly more photons will be absorbed per infant cone given the same flux, but it also means that higher spatial frequency information will be degraded more because of greater spatial averaging over the cone’s aperture. The cones are also packed much less densely in the newborn’s fovea. This packing density is important because it determines to a large extent the highest spatial frequency that can be reliably signaled from the photoreceptor catches across a small region of the retina. Sampling the image very finely permits information about high spatial frequencies (fine detail) to be transmitted well, while sampling the image very coarsely attenuates this high-spatial-frequency information significantly.1 Based on the spacings of the photoreceptor inner segment spacing, Candy, Crowell, and Banks (1998) calculated an acuity limit of 15 cycles per degree (cpd) for newborns (cf. 60 cpd for adults). Additionally, lower packing densities also imply fewer photon absorptions given

iliescu and dannemiller: brain-behavior relationships in early visual development

129

a fixed nodal distance because more of the photons pass through the retinal space not occupied by photoreceptors. When all of these factors were taken into account, Banks and Bennett (1988) estimated that if identical patches of light were to be presented to newborn and adult central retina, approximately 350 photons would be absorbed by the adult cone lattice for every one photon absorbed by the newborn’s cones. It is hard to overstate the significance of this factor for newborn vision. Fewer photon absorptions mean much noisier signals, because the emission of light is an inherently random process. Because the variance in the number of photons emitted from a source or reflected from a surface is proportional to the mean number of photons emitted or reflected (Poisson process), the signal-to-noise ratio expressed as the ratio of the mean to the standard deviation will increase with the square root of the mean. In other words, the availability of more photons ideally leads to an improvement in sensitivity that is proportional to the square root of the mean. The 350 : 1 ratio of absorbed photons for adults compared to infants means that the signal-to-noise ratio for infants would be at best 1/18 (square root behavior) as strong in newborns as in adults. Imagine trying to estimate a mean with a set of observations that was 18 times as variable as another set. Any visual task (e.g., intensity discrimination at an edge) that depends on pooling and averaging the responses from these isomerizations will surely suffer substantially by having such impoverished information on which to base performance. Recent evidence from the study of the development of retinal circuitry in other species suggests that it is not just the photoreceptors that undergo marked postnatal change. The ganglion cells of the retina collect signals from prior levels in the retina and pass this information to higher levels in the visual system in the form of action potentials. In mice, retinal ganglion cell dendritic arbors are subject to significant remodeling during development, eventually making contact with amacrine cells in very specific strata within the inner plexiform layer. Cholinergic amacrine cells probably guide and shape this remodeling of ganglion cell dendritic processes during postanatal development (Stacy and Wong, 2003), reflecting the postnatal maturation of this aspect of the retinal circuitry. In the mouse this class of ganglion cells corresponds to directionally selective ganglion cells in other species. While human retinal ganglion cells do not exhibit directional selectivity, these studies nonetheless show that intrinsic chemical signals within the retina play a role in the postnatal development of retinal circuitry in at least one mammalian species. Experimental techniques ranging from gene manipulation to two-photon microscopy helped in working out the details of this circuit (He and Masland, 1997; Yoshida et al., 2001). It is clear that the normal development of this functional relationship between these two

130

retinal cell types is essential for the appropriate computation of the direction of motion in mouse vision. Stacy and Wong suggest that amacrine–ganglion cell interactions are important in determining the stratification of ganglion cells into on and off laminae in the inner plexiform layer as well. on and off pathways in the visual system are important for encoding the visual signal into positive (increments) and negative (decrements) contrast levels. Models of retinal function based on extant data do not fully explain the poor contrast sensitivities and reduced acuities observed during early postnatal life. In other words, observed values for these measures are worse than would be predicted from a visual system with the optical and retinal characteristics included in these models. Although an obvious possibility is that we lack a comprehensive understanding of the retinal circuitry and its development beyond the photoreceptors, another possibility is that there are probably postretinal sites in the visual pathways where significant information loss occurs. The immaturity of these sites, both subcortical and cortical, could further limit visual function. In the remainder of this chapter we will consider the evidence for how structures in the visual pathway beyond the retina might inform our understanding of early postnatal brain–behavior relations.

A selective look at cortical development Because of the greater complexity of the visual circuitry beyond the retina, theories trying to make a direct connection between the neural substrate and visual behavior tend to be less quantitative than their retinal counterparts. Cortical theories generally make predictions about the presence or absence of a certain perceptual capacity, known or thought to depend on cortical processing (e.g., orientation sensitivity). We will focus mainly on two themes in considering development beyond the retina: (1) While there is incontrovertible evidence for postnatal experience playing an important role in shaping cortical development and instructing specific functional capabilities at least in the visual cortex, there is also clear evidence that intrinsic neural activity in the prenatal period has a crucial influence on the formation of a normal neural architecture and connectivity. (2) There is a very dynamic evolution of the biochemical landscape during this period; the excitatory and inhibitory neural mechanisms in particular that, in light of the latest data, account for important developmental phenomena during this period are particularly malleable. Before discussing these themes, we will review briefly the major visual pathways. For an overall view of the visual pathway up to primary visual cortex please refer to figure 9.1. Figure 9.2 is a schematic of several subcortical and cortical visual pathways, although this is certainly not an exhaustive diagram of such pathways. The major subcortical

fundamentals of developmental neurobiology

Figure 9.1 Schematic representation of the visual pathway with emphasis on the functional organization of the dorsal lateral geniculate nucleus (upper insert) and of the primary visual cortex. (Lower

inset reproduced from Grinvald et al., 1999, with kind permission of Springer Science and Business Media.)

pathway involves direct projections from the retina to the superior colliculus (SC). The axons of retinal ganglion cells terminate primarily in the superficial layers of the superior colliculus (Kaas and Huerta, 1988). The major cortical pathway runs from the retina to the lateral geniculate nucleus (LGN) to area V1 of visual cortex (VC). There are also numerous descending pathways from VC to subcortical structures (e.g., LGN, SC, pulvinar) that are not shown here. We will discuss selective aspects of these pathways, as they are relevant to the issues that we have raised. A good general discussion of these pathways as they exist in mature primates can be found in Rodiek (1998). The retinocortical pathway is thought to consist of two parallel and quasi-independent streams of processing: parvocellular and magnocellular (Livingstone and Hubel, 1988; Maunsell and Newsome, 1987; Van Essen, Anderson, and

Felleman, 1992). These streams are evident in distinct classes of ganglion cells in the retina, are segregated in layers within the LGN (four parvocellular and two magnocellular layers, figure 9.1, upper insert), and project to distinct layers with the recipient zone of layer 4C of primary visual cortex. The parvocellular stream, postulated to subserve mainly color and form vision and comprising 80 percent of all ganglion cells in the primate retina, projects mainly to layer 4Cβ, while the magnocellular stream projects to layer 4Cα. This division is important because, as noted earlier, several models advance the hypothesis that the differential maturation of cortical layers explains corresponding differences in certain aspects of visual performance. The division of this pathway into magnocellular and parvocellular streams persists to some extent into other cortical areas. The magnocellular or parietal-directed stream is

iliescu and dannemiller: brain-behavior relationships in early visual development

131

frontal

ITp

frontal

ITr IT

VPP

ITc MST

DLr

FSTd

FSTv DLc

DM

MT

MT

CO

Inter

V2 (area 18)

CO

V1 (area 17)

SC

K

M

P

LGN

retina Figure 9.2 Various subcortical and cortical visual pathways: M, magnocellular; P, parvocellular. Notice the retinal projections both to the superior colliculus and to the LGN. There are also extensive reciprocal connections between various extrastriate cortical areas

132

and the pulvinar not shown here. The K (koniocellular) pathway is not discussed in this chapter. (From Casagrande, 1994, by permission.)

fundamentals of developmental neurobiology

associated with areas V3, MT, and MST, and posterior parietal areas. The parvocellular or temporal-directed stream is associated with areas V4 and inferotemporal cortex. It is important to keep in mind that this division into two primary streams does not imply complete independence between the processing of information within these streams (Nassi, Lyon, and Callaway, 2006). Even in area V1, there is evidence for cross talk between these two streams. Many of the visual processes that are important in the mature brain only begin to appear postnatally. More importantly, these processes start at different postnatal ages, progress at different rates, and terminate at different ages (figure 9.3A). The latest relevant research shows that although the foundations of a healthy, normal cerebral circuitry are laid down before birth, evoked activity of neural circuits and certain biochemical transformations that take place after birth ensure the formation of highly specialized and essential cortical areas and connections. The idea that neuronal activity has a profound impact on the normal development of the visual system connectivity has its roots in the pioneering deprivation studies of Hubel, Wiesel, and later Shatz. They showed that closing one eye during a critical period in postnatal life rendered that eye permanently incapable of driving cortical cells (Wiesel and Hubel, 1965). This effect was accompanied by a marked reduction in the cortical area innervated by that eye and by a significant increase in the cortical territory allocated to the geniculocortical axons serving the nondeprived eye—that is, a massive reorganization of the brain due to the lack of input from one eye (Hubel, Wiesel, and LeVay, 1977; Shatz and Stryker, 1978). What factors are responsible for this postnatal reorganization of visual cortex? The mark of the adult lateral geniculate nucleus (LGN) and of the primary visual cortex is the presence of a highly organized structure. The LGN is segregated into eye-specific layers, while the visual cortex presents ocular dominance columns—collections of neurons driven more strongly by the input from one eye than from the other. The development of these structures is a good place to examine the roles that intrinsic (e.g., genetic) factors and extrinsic factors (e.g., specific sensorial inputs) play in defining the final form of cortical and subcortical visual areas. Consider the fact that the retina of infant ferrets is swept every minute by spatiotemporal waves of intrinsically generated neural activity (Katz and Crowley, 2002). Importantly, these waves occur independently in the two eyes (in contrast to evoked activity). In other words, neighboring cells in one retina will fire at nearly the same time, but the firing of corresponding ganglion cells in the two retinas will be uncorrelated. It has been suggested that this withineye correlated retinogeniculate activity and the uncorrelated activity between the two eyes will lead to the laminar segregation of the LGN (Huberman, Stellwagen, and Chapman,

2002; Huberman et al., 2003) following the Hebbian postulate that coactive inputs are preferentially stabilized relative to temporally uncorrelated inputs (“cells that fire together wire together”). This is a good example of neural structure being formed by spontaneous activity rather than evoked activity. In the case of the ocular dominance (OD) columns, the initial view was that visual stimulation was needed to form these columns in the primary visual cortex. However, subsequent results have shown that retinal activity is very unlikely to be the cause of OD band formation. Experimental data suggest that the initial formation of the OD columns is not influenced at all by the balance between the inputs from the two eyes. The work of Crowley and Katz (1999) showed the presence of OD columns in ferrets that were enucleated at birth. The OD structure in these animals closely resembled, in spatial cortical periodicity, that of normal animals. The same investigators (2000) also showed that changing the balance between retinal inputs through unilateral enucleation had no significant effect on the sizes of the columns mapping the deprived and the nondeprived eyes. Based on these data, it looks as though the segregation of the OD column in layer 4 is driven mostly by molecular cues, although the role of spontaneous activity from the LGN cannot be completely excluded. Although the instructing influence of spontaneous electrical activity on the formation of cortical columns is questionable, convincing data exist showing that electrical activity is essential in maintaining these columns (Chapman, 2000). These findings call to mind the important distinction between induction and maintenance in development as noted by Gottlieb (1976). Although thinking about the role of neural activity in forming OD columns has undergone recent serious revision, the role of this activity in constructing some receptive field properties has yet to be challenged. The work of Weliky and Katz (1999) in particular makes a strong case for a causal relation between spontaneous correlated activity from the LGN and the early appearance of orientation selectivity in the visual cortex. Electrophysiological data in neonatal kitten cortex suggest that some neurons have a weak orientation preference, with these preferences mostly concentrated around the cardinal axes (Fregnac and Imbert, 1978). Weliky and Katz (1999) showed that synchronous bursts of spontaneous activity occur in the LGN of ferrets before eye opening with a frequency similar to that of spontaneous retinal waves. Significant binocular correlations were present only when corticothalamic feedback was intact. Weliky and Katz (1999) also showed that disruption of these natural input patterns results in the degradation of early cortical orientation selectivity. Chronic stimulation with synchronous electrical activity (through a nerve cuff implanted postnatally on one of the optical nerves, the other eye being enucleated) disrupted the spontaneous retinogeniculate drive. The result was that,

iliescu and dannemiller: brain-behavior relationships in early visual development

133

Figure 9.3 (A) Time line illustrating many of the main events during the development of the visual cortex and its connections with the thalamus in ferrets. (Reprinted from Sur and Leamey, 2001, by permission from Macmillan Publishers Ltd.) (B, C) Developing neurons have a higher concentration of intracellular chloride than adult neurons. (B) The electrochemical equilibrium potential for Cl− decreases with age. (Reprinted from Ben-Ari, 2002, by permission from Macmillan Publishers Ltd.) (C) The Nerst equation relates the transmembrane chloride concentration gradient to the reversal potential. The curve is very steep at physiological concentrations; that is, a small change in the ion concentration is

134

sufficient to cause GABAA reversal potential to be either below or above the resting membrane potential or the threshold for action potential generation, that is, acting as an inhibitory or excitatory messenger. (Reprinted from Staley and Smith, 2001, by permission from Macmillan Publishers Ltd.) (D) Cortical synaptic density (left y-axis) and the percentage of infants displaying stereopsis (right yaxis) plotted against postnatal age (x-axis). Synaptic density is plotted with open symbols, and stereopsis is plotted with closed symbols. There is a close, time-lagged correlation between this anatomical measure and this functional measure. (From Wilson, 1993, by permission.)

fundamentals of developmental neurobiology

although the macroscopic aspects of the orientation preference maps in the visual cortex apparently were not very different than in the normal case, orientation selectivity was lower at both the population and single-cell levels. Additional evidence shows that visual experience plays an instructive role in the formation of orientation selectivity within the neuronal population of V1 (Sengspiel, Stawinsky, and Bonhoeffer, 1999). Here, then is an example of a visual property that appears to depend both on intrinsic and extrinsic neural activity to achieve its mature form. These uniform waves of spontaneous activity, present throughout the brain as a mark of developing networks, are but one example of the complexity confronting our understanding of the relations between brain and behavior during development. Another example can be found in the role played by the balance between excitation and inhibition that one finds at very early stages in development. The signature electrical activity of developing circuits is the presence of giant depolarizing potentials (GDPs) (for a comprehensive review please refer to Ben-Ari, 2001). These potentials generate large oscillations of intracellular calcium and lead to an activity-dependent modulation of neural growth and the formation of synapses. Gamma-aminobutyric acid (GABA), an inhibitory neurotransmitter in the adult brain, plays an important role in the generation of the excitatory GDPs (Ben-Ari, 2002) early in development. Although, as noted, it is the main inhibitory neurotransmitter in the adult brain, responsible for optimal information processing (any imbalances can lead to pathologic conditions, e.g., epilepsy or Tourette’s syndrome), GABA exerts an excitatory effect and generates significant trophic effects early in development (Owens et al., 1996; Hensh, 2005). Additional evidence shows that the marked difference in GABA activity in infancy and adulthood is due to yet another difference between immature and adult neurons: The concentration of the Cl− ion is higher in immature neurons with about 25 mM, a concentration sufficient to change the action of GABA from being inhibitory to excitatory (in adult neurons the intracellular chloride concentration is around 7 mM; see figure 9.3B). Furthermore, there are data that suggest that the same thing is true for the astrocytes of the developing but not adult optic nerve (Sakatani, Black, and Kocsis, 1992), possibly making this higher Cl− concentration a universal developmental signal throughout nervous system structures. Comparative research shows that GABA is a very well conserved developmental signal phylogenetically. GABA or its phylogenetic homologues have been described throughout the animal and vegetable kingdoms (from tomatoes, Gallego et al., 1995; to beetles, Wegerhoff, 1999; to Drosophila, Lee and O’Dowd, 1999). In the mammalian brain it has been shown in a variety of systems (spinal cord, Reichling et al., 1994; hypothalamus, Chen, Trombley, and Van Den Pol, 1996; hippocampus, Leinekugel et al., 1995;

cortex, Owens et al., 1996; and other areas) that GABA produces depolarization and an increase in intracellular calcium levels in the immature but not in the adult brain. Given this important developmental change in the action of GABA, it is also interesting to note that it is probably GABA itself that promotes the developmental switch of neuronal GABAergic responses from excitatory to inhibitory (Ganguly et al., 2001). Intimately related to the intracellular Cl− concentration (the main ion conducted by the GABA receptor channel) this shift changes the GABA-related stimulation from being a depolarizing signal to being a hyperpolarizing signal (i.e., inhibitory), with inhibitory GABA probably playing a very important role in the mature brain’s ability to adapt dynamically to evoked activity (see figure 9.3C). Considering the focus of this chapter, it is reasonable to ask whether these animal models of visual cortical development are also valid for the development of the human brain. Recent evidence from postmortem samples of postnatal V1 (Murphy et al., 2005) showed that the mature expression of the gene responsible for the formation of the GABAproducing enzyme and the conversion of GABA receptors to a state compatible with driving plastic changes within a network occurs slowly, over several years—a fact that is consistent with the extended length of the critical period for amblyopia in humans (Berardi, Pizzorusso, and Maffei, 2000). Additionally, particular injuries like brain lesions or retinal scotomas in adulthood produce, among other effects, a rearrangement of the excitatory-inhibitory balance to a more immature state (Arckens et al., 2000). The development of the brain function is a very dynamic process that involves both intrinsic and extrinsic factors. We have reviewed several of the intrinsic mechanisms that are important in the development of the visual system. An additional line of evidence offers striking support for the complementary idea that sensory experience itself (evoked neuronal activity) plays a significant role in defining the shape and the organization of the future mature brain. These studies look at the development of cerebral circuitry and resultant function in animals that have the afferents that carry information about one sensory modality redirected to cerebral targets that normally process different modalities. Some of the most direct evidence comes from the work of Sharma, von Melchner, and Sur (Sharma, Angelucci, and Sur, 2000; von Melchner, Pallas, and Sur, 2000), who showed in ferrets that rerouting retinal projections into the auditory pathway makes the neurons in what should be the primary auditory cortex respond to visual stimuli. The cortex becomes organized into orientation modules (a mark of visual cortex), and neurons show orientation tuning comparable to that of V1 neurons even though the orientation map is less orderly. Furthermore, when light stimuli were presented in the portion of the visual field seen by the “auditory” cortex,

iliescu and dannemiller: brain-behavior relationships in early visual development

135

“rewired” animals responded as though they perceived the stimuli as being visual rather than auditory. While it is undoubtedly true that intrinsic factors play a large role in setting up the initial circuitry in cortex, this work provides a clear if somewhat artificial example of how the final organization and function of a section of cortex depends critically on the incoming, stimulus-evoked afferent activity.

Cortically motivated developmental brain–behavior models We turn next to the enterprise of trying to construct models of behavioral development during infancy from what is known about the anatomical and physiological development of the visual pathways beyond the retina. One good example of such a model was proposed by Wilson (1993). In this model, Wilson attempted to relate cortical synaptic density as reported by Huttenlocher and associates (1983; see also Huttenlocher and de Courten, 1987) to the development of various functions most likely to be mediated cortically (e.g., orientation selectivity, binocular rivalry, and stereopsis). Figure 9.3D shows that there is a time-lagged but close correlation between stereopsis development and the increase in synaptic density over the first year of life. Of course, this correlation cannot be interpreted causally, mainly because it is difficult if not impossible at this point in our understanding of cortical circuitry to relate quantitatively a measure like synaptic density to a measure like stereoacuity. Figure 9.4 shows Conel’s (1939, 1951) renderings of sections of visual cortex from a newborn human infant (left) and a six-monthold human infant (right). Although the postnatal increase in dendritic complexity is evident even to the naked eye, it is another matter to try to relate this increase quantitatively to some aspect of visual performance. All that one can do at this point is to agree with Wilson (1993) that until adequate models of functions like stereopsis are available, it is best to be content with the apparent correlation between the brain and behavioral measures. Several other investigators have proposed models of early visual development that involve cortical brain-behavior relationships (Atkinson, 1984, 1992, 1998; Banton and Bertenthal, 1997; Bronson, 1974; Johnson, 1990). One theme common to these models is that early visual development may be characterized by differential rates of maturation both within subpathways and between subpathways in the visual system (see figure 9.2). The most common example of this type of model involves earlier maturation of subcortical function than of cortical function (e.g., Bronson, 1974; Atkinson, 1984). This theme is also evident in Atkinson’s (1992, 1998) proposals that the parvocellular pathway may lead the magnocellular pathway in maturation during early postnatal development, thereby leading to the earlier emergence of certain visual functions typically attributed to the parvocellular pathway. Finally, Johnson

136

(1990) and Banton and Bertenthal (1997) have used anatomical data showing differential maturation of the laminae in primary visual cortex to argue for the differential emergence of certain aspects of visual attention and motion processing. Bronson (1974) first proposed that early in postnatal life much of the visual capacity of the neonate could be explained by supposing that the subcortical pathway from retina to superior colliculus matured or was functional earlier than the cortical pathway from retina to visual cortex. Atkinson (1984) echoed this proposal that much of the observed visual behavior in the first month is controlled subcortically and added the additional postulate that the descending pathways from visual cortex to superior colliculus mature later, leading to changes around two months in certain visual capacities (e.g., attentional switching, binocular convergence). Using resting positron emission tomography (PET), Chugani and Phelps (1986) concluded that prior to three months of age, various subcortical areas (e.g., thalamus, midbrain–brain stem) were probably more functionally mature than various cortical areas (e.g., occipital, parietal, and temporal). Different cortical areas might also mature earlier than others (e.g., temporal prior to frontal, Erickson et al., 1998). Aspects of visual behavior such as orienting probably depend heavily on the superior colliculus, so it is natural to suppose that the visual orienting in newborns, however sluggish, reflects the function of this structure. Additionally, early postnatal asymmetries in optokinetic nystagmus (OKN), a reflexive visual tracking of large moving fields, and the disappearance of these asymmetries later in the first half year of life imply that the superior colliculus alone may mediate much of the visual behavior observed in the neonatal period. However, it is important when examining these models to determine just what vision is like with and without the operation of visual cortex. What visual capacities might we expect to be present when only subcortical structures are mediating vision or when vision is mediated by pathways other than the ones through primary visual cortex? Several reports exist that are relevant to this question (see Stoerig and Cowey, 1997, for a review). For example, color vision is typically thought to involve cortical processing within the parvocellular stream, although the substrate for this is surely set up initially within the color opponent ganglion cells of the retina. Despite this cortical involvement, Stoerig and Cowey (1992) and Barbur and associates (1998) reported color discrimination in individuals with significant primary cortical lesions or without primary visual cortices. Additionally, Braddick and associates (1992) have reported on several visual capacities present in unilaterally decorticate human infants. This report is important, because the absence of striate visual cortex does not necessarily imply that all

fundamentals of developmental neurobiology

Figure 9.4 Striate cortex from a newborn (left) and a 6-month-old (right) human infant. (From Conel, 1939 [left] and Conel, 1951 [right] by permission.)

iliescu and dannemiller: brain-behavior relationships in early visual development

137

visual function is then mediated subcortically; it is possible that projections to extrastriate visual cortex could also be involved. In the two cases reported in Braddick and associates (1992), however, both striate and extrastriate cortex were missing, so subcortical pathways must have mediated the remaining visual capacities. These infants could visually orient to a conspicuous target presented to the contralesional visual field as would be expected if such a function were subserved by the superior colliculus. Somewhat unexpectedly, however, binocularly symmetric OKN, which had been thought previously to be mediated by crossed subcortical pathways, was found to be asymmetric in both these infants. This observation implies that contrary to earlier ideas that cortical involvement need not be invoked to explain symmetric binocular OKN, such cortical involvement may indeed be necessary. This possibility is significant, because even in the newborn infant OKN is binocularly symmetric. One other recent report is relevant to the question of what vision might be like in the absence of striate visual cortex. Shewmon, Holmes, and Byrne (1999) reported clinical observations on multiple functions in four congenitally decorticate children. Of particular relevance are their reports that some of these children could actively track, albeit not very smoothly, moving objects such as faces, visually orient to objects moved into the peripheral visual field, and in one case fixate steadily. Additionally, several of these children were said to be able to recognize familiar adults, although such recognition probably occurred through other sensory modalities. All these children were assumed to be cortically blind. Shewmon, Holmes, and Byrne (1999) proposed the interesting idea that some of these visual capacities might represent “vertical” plasticity in subcortical structures. It is generally accepted that “horizontal” plasticity between different cortical areas may take place when tissue in one area is disrupted and its function is taken over or usurped by remaining cortical tissue (Baseler, Morland, and Wandell, 1999; Chugani, Muller, and Chugani, 1996; Cohen et al., 1997; Sadato et al., 1996). What Shewmon, Holmes, and Byrne (1999) propose is that subcortical structures may be “vertically” plastic and organize prenatally or reorganize to take over supposedly cortical functions in the absence of target occipital cortex (see also Kalil and Behan, 1987). Until more systematic visual tests are done on such children, it is difficult to say precisely how good these remaining visual capacities are. Nonetheless, observations such as these are valuable in giving some insight into what vision is like without cortical pathways. The remaining models of visual development have all proposed differential maturation of specific subpathways in the visual system as explanations for the emergence of various visual capacities. Atkinson (1992, 1998) has proposed that the parvocellular pathway is more mature early

138

in postnatal development than the magnocellular pathway. The behavioral evidence that is meant to be explained by this model is primarily pattern vision (e.g., orientation sensitivity). Johnson’s (1990) model is meant to explain primarily orienting and attentional behavior based on anatomical data from Conel (1939, 1951) that the primary visual cortex matures from deeper (levels 5 and 6) to more superficial layers (layers 1–3). Additionally, Johnson proposed that descending pathways from striate and extrastriate cortex to superior colliculus mature at different rates. Finally, Banton and Bertenthal (1997) also proposed that differential postnatal maturation of the laminar structure of striate cortex from deeper to more superficial layers may explain various aspects of the development of motion processing. We will use Atkinson’s (1992, 1998) model as an example of the theme of explaining early visual development by pointing to differential maturation of subpathways. This model is a good example of using behavioral evidence to infer the state of various cortical areas. As such it relies on modern visual neurophysiology with its major division of visual processing into parvocellular and magnocellular streams (see figure 9.2). This model affords a good opportunity to review the behavioral evidence on early visual development as well as the anatomical and neurophysiological evidence on the maturation of various areas in visual cortex. Atkinson’s model (1992, 1998) relies heavily on the idea that certain aspects of visual processing are the province of the parvocellular (ventral) stream while others are the province of the magnocellular (dorsal) stream. Additionally, development in this model consists of integration across these streams (the binding problem) so that the infant eventually comes to have a complete representation of objects in the world and of their spatial positions. This representation serves the purposes of object recognition and action. Atkinson argues that the ability of very young infants in the first few weeks of life to discriminate the orientations of grating patterns (Atkinson et al., 1988) and the temporal-frequency dependency shown by evoked responses to orientation changes are evidence of the early maturation of the parvocellular pathway (Braddick et al., 1986) because this function is subserved primarily by parvocellular neurons in adults (but see also Levitt, Lund, and Yoshioka, 1996). In contrast, motion processing that is selective for direction is supposed to be handled primarily by the magnocellular pathway. Excluding OKN, true evidence of direction-selective responding does not emerge until after 6–8 weeks of age, and it is evident at slow velocities before it is evident at faster ones (Wattam-Bell, 1991). Thus, unlike orientation discrimination, which is evident well before 6–8 postnatal weeks, this magnocellular characteristic—directional selectivity—is not evident until near

fundamentals of developmental neurobiology

the end of the second month. Atkinson (1998) concludes that magnocellular function lags parvocellular function early in development. Disparity sensitivity and binocular correlation detection do not emerge until approximately 3–4 months postnatally (Birch, Gwiazda, and Held, 1982; Fox et al., 1980). Depth processing is thought to be handled by the magnocellular pathway. In contrast, 8-week-olds (Hamer, Alexander, and Teller, 1982; Packer, Hartmann, and Teller, 1984) and possibly even neonates and 4-week-olds (Adams, Maurer, and Davis, 1986; Maurer and Adams, 1987) can detect pattern differences based on color alone, and color is thought to be processed primarily by the parvocellular pathway (although see Barbur et al., 1998). Once again, functions attributed mainly to the magnocellular stream (disparity, depth) appear to emerge slightly later than one attributed to the parvocellular stream (color). What do we know about the differential emergence of these two processing streams from anatomical and neurophysiological studies on humans or closely related species? The retinal ganglion cells that make up these two streams diverge early in embryonic development in the primate visual system, and they project to the appropriate laminae in the developing LGN well before any visual experience (Meissirel et al., 1997). The projections from relay neurons in LGN arrive at their targets in layer 4C in striate cortex prenatally, and the axons terminate in appropriate sublaminae within 4C, again before any visual experience is attained (Mates and Lund, 1983). Initially, it appears that LGN neurons may also project to layer 6 of striate cortex before birth (Rakic, 1976, 1977). As noted previously, intrinsic activity during prenatal development probably plays a major role in organizing the laminar characteristics, feature maps (e.g., orientation preferences, ocular dominance columns) and appropriate projections in the retinogeniculocortical pathway prior to extrinsic, visual experience (Hubener, 1998; Godecke and Bonhoeffer, 1996; Katz and Shatz, 1996; Shatz, 1996). Indeed, the laminar structure of visual cortex and its reciprocal connections with LGN even develop when slices of future visual cortex and LGN are cocultured in vitro (Toyama et al., 1991).2 Anatomical data on the development of visual cortical areas from humans and closely related primates present a mixed picture on whether or not one of the two processing streams is ahead of the other in terms of its state at birth and its postnatal maturation. One of the problems in using the anatomical data to make inferences about the presence of a particular function is that it is not clear exactly which characteristics should be used to infer effective function. One could examine myelination as a marker for effective function, or one could examine synaptic connectivity and morphology for clues to when a particular region appears to be functional. As others have noted (e.g., Banton and Ber-

tenthal, 1997), Conel (1939) pointed out that axons in the deeper laminae (5 and 6) of striate cortex appear to myelinate earlier than those in more superficial laminae. Thus extrastriate areas receiving projections from neurons in striate laminae 5 and 6 (e.g., area MT) might be expected to support effective function before those receiving projections for more superficial laminae. This reasoning is the basis for Banton and Bertenthal’s (1997) proposal that newborns may be expected to respond to translatory motion. If one examines other markers of anatomical development in striate cortex, differential gradients of maturation are either not as clear or are conflicting. For example, Lund and Holbach (1991) showed that the development of Type I dendritic spines that are associated with synaptic contacts reaches a peak in macaque monkeys approximately five weeks postnatally for neurons in lamina 4Cα (magnocellular), whereas a similar peak is not reached for neurons in lamina 4Cβ until eight weeks postnatally (see also LeVay, Wiesel, and Hubel, 1980; Lund, Holbach, and Chung, 1991). The same was also true of Type II inhibitory synapses in layer 4C with those in the magnocellular recipient layer developing slightly ahead of those in the parvocellular layer (Lund and Harper, 1991). It is also interesting to note that these presumed inhibitory synapses in layer 4C lag the development of excitatory synapses in the same laminae (Lund and Harper, 1991). Becker and associates (1984) showed that in humans, dendritic branching was more advanced in layer 5 than in layer 3 in visual cortex prenatally and during early postnatal development—a trend consistent with the myelination from deeper to more superficial laminae. Neuronal densities in human striate visual cortex also appear to follow the gradient noted earlier with deeper layers preceding more superficial layers until densities stabilize several months after birth (Leuba and Garey, 1987). Finally, Lund, Boothe, and Lund (1977) concluded that in macaque monkeys, maturation of the morphology of visual cortical neurons occurs at the same rate in all laminae with perhaps some advantage based on the size of the neuron; large pyramidal neurons in layer 5 tended to mature earlier than smaller neurons in other laminae. It is evident from this brief review that whether or not there is a noticeable advance in the maturation of parvocellular versus magnocellular neurons or by laminar layer depends on what aspect of cortical or neuronal morphology is being considered. However, we agree with Banton and Bertenthal (1997) that on many of these characteristics, magnocellular neurons appear to be slightly ahead of parvocellular neurons.

Subtleties of the differential maturation hypothesis Most of the models based on cortical development that we have discussed argue that the order of emergence of certain visual functions in early postnatal development reflects the

iliescu and dannemiller: brain-behavior relationships in early visual development

139

differential maturation of subpathways in the visual system. As noted earlier, it is clear that intrinsic factors and extrinsic, evoked activity combine to determine the precise development of the visual pathways. A great deal of attention has been paid over the years to the effects of postnatal experience on visual development. Perhaps as important, but somewhat neglected, is the emergence of cortical structure determined by intrinsic factors with very precise timing relations (e.g., triggering a sensitive period by the precise balance between excitatory and inhibitory neurons in a given cortical area). We will conclude with several subtleties and implicit assumptions that can arise when models are formulated regarding explicitly developmental brain–behavior relations in vision. 1. The degree of independence of the visual subpathways involved in these models may not be as great as postulated. For example, Dobkins and Albright (1994) showed that neurons in area MT of the macaque, thought to be the paradigm case for magnocellular processing, are capable of using color differences to signal the directions of motion. Maunsell, Nealey, and DePriest (1990) had reported similar results earlier showing that both magnocellular and parvocellular influences can be detected in the responses of area-MT neurons. Color processing by a strict division of labor would be expected to fall to the parvocellular system, so this result indicates less independence of these subpathways than may be indicated in some of the models discussed earlier. Gegenfurtner and Kiper (1996) showed that as early as cortical area V2, neurons were multiply selective for form, color, and motion, and argued that there is no functional segregation of these stimulus features at this level of the visual pathway. These neurophysiological studies are supported by anatomical studies showing interneurons as early as cortical areas V1 and V2 that merge the parvocellular and magnocellular streams (Levitt, Lund, and Yoshioka, 1996; Levitt, Yoshioka, and Lund, 1994; Lund, 1987; Lund and Yoshioka, 1991; Yoshioka, Levitt, and Lund, 1994). 2. A particular visual function that is ascribed to one subsystem in the mature adult may not necessarily be mediated by that subsystem when it first emerges in development. For example, Dobkins, Lia, and Teller (1997) and Dobkins, Anderson, and Lia (1999) have argued from psychophysical tests of chromatic and luminance temporal-contrast sensitivity that the magnocellular pathway may mediate detection of both luminance and chromatic spatiotemporal contrast. If this assumption were true, then it would differ substantially from the case later in development in which chromatic contrast sensitivity is signaled primarily by a temporally low-pass parvocellular channel and luminance contrast is signaled by a temporally band-pass magnocellular channel. Dobkins and Teller (1996) concluded that for 3-month-old infants, patterns defined by color contrast alone or patterns defined by luminance

140

contrast are both detected by mechanisms sensitive to the direction of motion. This situation contrasts strongly with the case in adults in which mechanisms sensitive to luminance contrast are selective for direction of motion, but those mediating detection of color contrast are not. Morton and Johnson (1991) have proposed that face processing, a function typically attributed to the ventral/ temporal cortical pathway, may first emerge during the newborn period mediated by subcortical structures. In all these cases, a visual function typically attributed to one level or subpathway within the mature visual system may emerge first in development being mediated by a different subsystem. If the hypothesis of differential rates of maturation of subpathways is taken seriously, then it should not be surprising that the most sensitive pathway for mediating detection of some stimulus feature early in development may not necessarily be the same pathway that mediates detection of that feature later in childhood or in adulthood (see also Banton and Bertenthal, 1997, for a similar suggestion). A corollary to this point is that infants and adults almost certainly differ in the effort required to execute visual tasks. Differences in effort or automaticity have been shown in adults to be related to differential activation of cortical versus subcortical areas (Alexander et al., 1999; Schneider, PimmSmith, and Worden, 1994). Thus the extent of cortical versus subcortical mediation for a given task early in development may differ substantially from that seen in mature adults. 3. Even if subpathways are quasi-independent in mature adults, when they first emerge there may be considerable overlap in how they process visual information. For example, Hawken, Blakemore, and Morley (1997) showed that in the monkey LGN, spatial contrast sensitivity functions of magnocellular and parvocellular neurons overlap quite substantially at birth. Only later in the first year do the contrast sensitivities of the magnocellular neurons improve substantially enough to separate them from the parvocellular neurons that showed little improvement in peak luminance-contrast sensitivity over 8 months. It may prove very difficult to attribute processing early in development to one subsystem or another simply because the subsystems may not show the degree of stimulus separation early in life that is characteristic of the mature visual system. This argument is similar to Johnson’s and Vecera’s (1996) “cortical parcellation” hypothesis in which the segregation of visual information processing into subsystems only occurs postnatally with many of these subsystems operating initially after birth unsegregated. A corollary to this point is that early in postnatal development there may be transient connections between cortical regions that disappear later in development (Huntley et al., 1988; Rodman and Consuelos, 1994). There is even evidence from the other end of the life

fundamentals of developmental neurobiology

span that the two major processing streams may become less distinct with age (Grady et al., 1992). 4. There may be significant temporal lags between the emergence of a given function at the single-unit level and behavioral evidence for that function. This possibility makes it very difficult to try to link the anatomical and neurophysiological data and the behavioral data. For example, Chino and associates (1997) demonstrated that binocular, disparity-sensitive neurons exist at or near birth in macaque area V1, although it is not for several weeks that stereopsis can be demonstrated behaviorally. 5. “Bottleneck” developmental theories (e.g., Banton and Bertenthal, 1997; Johnson, 1990) which propose that gradients of maturation within striate cortical laminae are responsible for the order of emergence of various behavioral functions may overlook the extensive alternative pathways to extrastriate cortex that do not go through striate cortex. For example, Benson, Guo, and Blakemore (1998) showed that many of the motion perception functions typically associated with cortical area MT (direction discrimination, perception of moving plaid direction, and coherent motion in random dot displays) remain after loss of primary (V1) visual cortex and its subsequent input to area MT. Benson, Guo, and Blakemore (1998) suggested that a subcortical pathway from superior colliculus through the inferior pulvinar to area MT may mediate such capacities in the absence of input/output from primary visual cortex. A similar suggestion was made by Baseler, Morland, and Wandell (1999). Differential maturation of laminae within primary visual cortex early in postnatal development would only be expected to impose an order on the emergence of visual capacities if these alternative routes from subcortex to extrastriate cortex were immature or if these extrastriate areas themselves were also immature.

Conclusions An immature and developing visual system is not necessarily an adult visual system that has been scaled down in terms of spatial and temporal processing, nor is it necessarily like an adult visual system in which some subpathways have simply been deleted. It is entirely possible that prior to the point at which these subpathways are segregated and subsequently integrated in a mature fashion they may interact and process information in ways that are not characteristic of the adult visual system. Thus the magnocellular pathway early in postnatal development may signal chromatic contrast to a much greater extent than in the adult, or scanning eye movements may produce enough spatiotemporal contrast to allow the magnocellular stream to process chromatic boundaries to the point at which discrimination is possible (Teller, 1998). Epelbaum and Teller (1995) showed that asymmetries in OKN in 2-month-olds are reversed, not eliminated, when isoluminant chromatic stimuli are used to drive the

eye movements—a result that is not found in adults. Unlike models based on photoreceptor morphology and geometry in which the brain–behavior relationships can be modeled based on well-understood physical and optical principles, our lack of understanding of complex cortical circuitry and of how various visual functions arise from such circuits makes it inherently more difficult to link brain and behavior when the brain side of the link is visual cortex. Nonetheless, much of the interesting development that occurs postnatally in the visual system undoubtedly involves striate and extrastriate cortical areas and inhibitory interactions within and between these areas. Our modeling should respect the additional, specifically developmental subtleties likely to arise when multiple subsystems develop “on-line” during the postnatal period. It should also be informed by the considerable cortical structure and connections present prior to the point at which visual experience begins to exert its effects on shaping cortical function. acknowledgment This research was supported by grant NICHD R01 HD32927 to JLD.

NOTES 1. We will not discuss the phenomenon of aliasing that occurs in undersampled systems. The interested reader can consult Thibos, Walsh, and Cheney (1987) for an example and a discussion of spatial aliasing in vision. 2. One note of caution is in order. Horton and Hocking (1996) reported the presence of clearly defined ocular dominance columns in newborn macaque monkeys, although Horton and Hedley-Whyte (1984) did not observe these columns in newborns, but only in a 6-month-old human. Possible species differences should always be kept in mind when generalizing from monkeys to humans.

REFERENCES Abramov, I., J. Gordon, A. Hendrickson, L. Hainline, V. Dobson, and E. LaBossiere, 1982. The retina of the newborn human infant. Science 217:265–267. Adams, R., D. Maurer, and M. Davis, 1986. Newborns’ discrimination of chromatic from achromatic stimuli. J. Exp. Child Psychol. 41:267–281. Alexander, G. E., M. J. Mentis, J. D. Van Horn, C. L. Grady, K. F. Berman, M. L. Furey, P. Pietrini, S. I. Rapoport, M. B. Schapiro, and J. R. Moeller, 1999. Individual differences in PET activation of object perception and attention systems predict face matching accuracy. NeuroReport 10:1965–1971. Arckens, L., G. Schweigart, Y. Qu, G. Wouters, D. V. Pow, F. Vandesande, U. T. Eysel, and G. A. Orban, 2000. Cooperative changes in GABA, glutamate and activity levels: The missing link in cortical plasticity. Eur. J. Neurosci. 12:4222–4232. Atkinson, J., 1984. Human visual development over the first 6 months of life: A review and a hypothesis. Hum. Neurobiol. 3:61– 74. Atkinson, J., 1992. Early visual development: Differential functioning of parvocellular and magnocellular pathways. Eye 6:129–135.

iliescu and dannemiller: brain-behavior relationships in early visual development

141

Atkinson, J., 1998. The “where and what” or “who and how” of visual development. In F. Simion and G. Butterworth, eds., The Development of Sensory, Motor and Cognitive Capacities in Early Infancy, 3–24. East Sussex, UK: Psychology Press. Atkinson, J., B. Hood, J. Wattam-Bell, S. Anker, and J. Tricklebank, 1988. Development of orientation discrimination in infancy. Perception 17:587–595. Banks, M. S., and P. J. Bennett, 1988. Optical and photoreceptor immaturities limit the spatial and chromatic vision of human neonates. J. Op. Soc. Am. [A] 5:2059–2079. Banks, M. S., and P. J. Bennett, 1991. Anatomical and physiological constraints on neonatal visual sensitivity and determinants of fixation behavior. In M. J. S. Weiss and P. R. Zelazo, eds., Newborn Attention: Biological Constraints and the Influence of Experience, 177–217. Norwood, NJ: Ablex. Banks, M. S., and J. A. Crowell, 1993. Front-end limitations to infant spatial vision: Examination of two analyses. In K. Simons, ed., Early Visual Development, Normal and Abnormal, 91–116. New York: Oxford University Press. Banton, T., and B. I. Bertenthal, 1997. Multiple developmental pathways for motion processing. Optom. Vis. Sci. 74:751–760. Barbur, J. L., A. Sahraie, A. Simmons, L. Weiskrantz, and S. C. R. Williams, 1998. Residual processing of chromatic signals in the absence of a geniculostriate projection. Vis. Res. 38: 3447–3453. Baseler, H. A., A. B. Morland, and B. A. Wandell, 1999. Topographic organization of human visual areas in the absence of input from primary cortex. J. Neurosci. 19:2619–2627. Becker L. E., D. L. Armstrong, F. Chan, and M. M. Wood, 1984. Dendritic development in human occipital cortical neurons. Dev. Brain Res. 13:117–124. Ben-Ari, Y., 2001. Developing networks play a similar melody. Nature Rev. Neurosci. 24:353–360. Ben-Ari, Y., 2002. Excitatory actions of GABA during development: The nature of the nurture. Nature Rev. Neurosci. 3:728– 739. Benson, P. J., K. Guo, and C. Blakemore, 1998. Direction discrimination of moving gratings and plaids and coherence in dot displays without primary visual cortex (V1). Eur. J. Neurosci. 10: 3767–3772. Berardi, N., T. Pizzorusso, and L. Maffei, 2000. Critical periods during sensory development. Curr. Opin. Neurobiol. 10:138– 145. Birch, E. E., J. Gwiazda, and R. Held, 1982. Steroacuity development for crossed and uncrossed disparities in human infants. Vis. Res. 22:507–513. Braddick, O., J. Atkinson, B. Hood, W. Harkness, G. Jackson, and F. Vargha-Khadem, 1992. Possible blindsight in infants lacking one cerebral hemisphere. Nature 360:461–463. Braddick, O. J., J. Wattam-Bell, and J. Atkinson, 1986. Orientation-specific cortical responses develop in early infancy. Nature 320:617–619. Bronson, G., 1974. The postnatal growth of visual capacity. Child Dev. 45:873–890. Brown, A. M., 1990. Development of visual sensitivity to light and color vision in human infants: A critical review. Vis. Res. 30: 1159–1188. Brown, A. M., 1993. Intrinsic noise and infant visual performance. In K. Simons, ed., Early Visual Development, Normal and Abnormal, 178–196. New York: Oxford University Press. Brown, A. M., V. Dobson, and J. Maier, 1987. Visual acuity of human infants at scotopic, mesopic and photopic luminances. Vis. Res. 27:1845–1858.

142

Candy, T. R., J. A. Crowell, and M. S. Banks, 1998. Optical, receptoral, and retinal constraints on foveal and peripheral vision in the human neonate. Vis. Res. 38:3857–3870. Casagrande, V. A., 1994. A third parallel visual pathway to primate area V1. Trends Neurosci. 17:305–310. Chapman, B., 2000. Necessity for afferent activity to maintain eyespecific segregation in ferret lateral geniculate nucleus. Science 287:2479–2482. Chen, G., P. Q. Trombley, and A. N. Van Den Pol, 1996. Excitatory actions of GABA in developing rat hypothalamic neurons. J. Physiol. 494:451–464. Chino, Y. M., E. L. Smith, S. Hatta, and H. Cheng, 1997. Postnatal development of binocular disparity sensitivity in neurons of the primate visual cortex. J. Neurosci. 17:296–307. Chugani, H. T., R. A. Muller, and D. C. Chugani, 1996. Functional brain reorganization in children. Brain Dev. 18: 347–356. Chugani, H. T., and M. E. Phelps, 1986. Maturational changes in cerebral function in infants determined by 18FDG positron emission tomography. Science 231:840–843. Cohen, L. G., P. Celnik, A. Pascual-Leone, B. Corwell, L. Faiz, J. Dambrosia, M. Honda, N. Sadato, C. Gerloff, M. D. Catala, and M. Hallett, 1997. Functional relevance of crossmodal plasticity in blind humans. Nature 389:180–183. Conel, J. L., 1939. The Postnatal Development of the Human Cerebral Cortex. Cambridge, MA: Harvard University Press. Conel, J. L., 1951. The Postnatal Development of the Human Cerebral Cortex, 4th ed. Cambridge, MA: Harvard University Press. Crowley, J. C., and L. C. Katz, 1999. Development of ocular dominance columns in the absence of retinal input. Nature Neurosci. 2:1125–1130. Crowley, J. C., and L. C. Katz, 2000. Early development of ocular dominance columns. Science 290:1321–1324. Curcio, C. A., 1987. Diameters of presumed cone apertures in human retina. J. Opt. Soc. Am. [A.] 4:P70. Curcio, C. A., K. R. Sloan, R. E. Kalina, and A. E. Hendrickson, 1990. Human photoreceptor topography. J. Comp. Neurol. 292:497–523. Dobkins, K. R., and T. D. Albright, 1994. What happens if it changes color when it moves? The nature of chromatic input to macaque visual area MT. J. Neurosci. 14:4854–4870. Dobkins, K. R., C. M. Anderson, and B. Lia, 1999. Infant temporal contrast sensitivity functions (tCSFs) mature earlier for luminance than for chromatic stimuli: Evidence for precocious magnocellular development? Vis. Res. 39:3223–3239. Dobkins, K. R., B. Lia, and D. Y. Teller, 1997. Infant color vision: Temporal contrast sensitivity functions for chromatic (red/green) stimuli in 3-month-olds [review]. Vis. Res. 37:2699– 2716. Dobkins, K. R., and D. Y. Teller, 1996. Infant motion : detection (M : D) ratios for chromatically defined and luminance-defined moving stimuli. Vis. Res. 36:3293–3310. Epelbaum, M., and D. Y. Teller, 1995. Infant eye movement asymmetries: Temporal-nasal asymmetry is reversed at isoluminance in 2-month-olds. Vis. Res. 35:1889–1895. Erickson, S. L., M. Akil, A. I. Levey, and D. A. Lewis, 1998. Postnatal development of tyrosine hydroxylase- and dopamine transporter-immunoreactive axons in monkey rostral entorhinal cortex. Cerebral Cortex 8:415–427. Fox, R., R. N. Aslin, S. L. Shea, and S. T. Dumais, 1980. Stereopsis in human infants. Science 207:323–324. Fregnac, Y., and M. Imbert, 1978. Early development of visual cortical cells in normal and dark-reared kittens: Relationship

fundamentals of developmental neurobiology

between orientation selectivity and ocular dominance. J. Physiol. 278:27–44. Fulton, A. B., J. Dodge, R. M. Hansen, and T. P. Williams, 1999. The rhodopsin content of human eyes. Invest. Ophthal. Vis. Sci. 40:1878–1883. Gallego, P. P., L. Whotton, S. Picton, D. Grierson, and J. E. Gray, 1995. A role for glutamate decarboxylase during tomato ripening: The characterization of a cDNA encoding a putative glutamate decarboxylase with a calmodulin-binding site. Plant Mol. Biol. 27:1143–1151. Ganguly, K., A. F. Schinder, S. T. Wong, and M. M. Poo, 2001. GABA itself promotes the developmental switch of neuronal GABAergic responses from excitation to inhibition. Cell 105: 521–532. Gegenfurtner, K. R., and D. C. Kiper, 1996. Processing of color, form, and motion in macaque area V2. Visual Neurosci. 13:161– 172. Godecke, I., and T. Bonhoeffer, 1996. Development of identical orientation maps for two eyes without common visual experience. Nature 379:251–254. Gogtay, N., et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101:8174–8179. Gottlieb, G., 1976. The roles of experience in the development of behavior and the nervous system. In G . Gottlieb, ed., Natural and Behavioral Specificity, 25–54. New York: Academic Press. Grady, C. L., J. V. Haxby, B. Horwitz, M. B. Shapiro, S. I. Rapoport, L. G. Ungerleider, M. Mishkin, R. E. Carson, and P. Herscovitch, 1992. Dissociation of object and spatial vision in human extrastriate cortex: Age-related changes in activation of regional cerebral blood flow measured with [15O]Water and positron emission tomography. J. Cogn. Neurosci. 4:23–34. Grinvald, A., D. Shoham, A. Shmuel, D. Glaser, I. Vanzetta, E. Shtoyermann, H. Slovin, A. Sterkin, C. Wijnbergen, R. Hildesheim, and A. Arieli, 1999. In-vivo optical imaging of cortical architecture and dynamics. In U. Windhorst and H. Johansson eds., Modern Techniques in Neuroscience Research. New York: Springer Verlag. Hamer, R. D., K. Alexander, and D. Y. Teller, 1982. Rayleigh discriminations in human infants. Vis. Res. 22:575–587. Hansen, R. M., and A. B. Fulton, 1999. The course of maturation of rod-mediated visual thresholds in infants. Invest. Ophthalmol. Vis. Sci. 40:1883–1886. Hawken, M. J., C. Blakemore, and J. W. Morley, 1997. Development of contrast sensitivity and temporal-frequency selectivity in primate lateral geniculate nucleus. Exp. Brain Res. 114:86–98. He, S., and R. H. Masland, 1997. Retinal direction selectivity after targeted laser ablation of starburst amacrine cells. Nature 389:378–382. Hendrickson, A. E., 1994. The morphologic development of human and monkey retina. In D. M. Albert and F. A. Jakobiec, eds., Principles and Practice of Ophthalmology: Basic Sciences, 561–577. Philadelphia: W. B. Saunders. Hendrickson, A. E., and D. Drucker, 1992. The development of parafoveal and mid-peripheral human retina. Behav. Brain Res. 49:21–31. Hensh, T., 2005. Cortical period plasticity in local cortical circuits. Nature Rev. Neurosci. 6:877–888. Horton, J. C., and E. T. Hedley-Whyte, 1984. Mapping of cytochrome oxidase patches and ocular dominance columns in human visual cortex. Philos. Trans. R. Soc. London [Biol.] 304:255– 272.

Horton, J. C., and D. R. Hocking, 1996. An adult-like pattern of ocular dominance columns in striate cortex of newborn monkeys prior to visual experience. J. Neurosci. 16:1791–1807. Hubel, D. H., T. N. Wiesel, and S. Levay, 1977. Plasticity of ocular dominance columns in monkey striate cortex. Philos. Trans. R. Soc. London [Biol.] 278:377–409. Hubener, M., 1998. Making maps in the dark. Curr. Biol. 8: R342–R345. Huberman, A. D., D. Stellwagen, and B. Chapman, 2002. Decoupling eye-specific segregation from lamination in the lateral geniculate nucleus. J. Neurosci. 22:9419–9429. Huberman, A. D., G. Y. Yang, L. C. Liets, O. A. Collins, B. Chapman, and L. M. Chalupa, 2003. Eye-specific retinogeniculate segregation independent of normal neural activity. Science 300:994–998. Huntley, G. W., S. H. C. Hendry, H. P. Killackey, L. M. Chalupa, and E. G. Jones, 1988. Temporal sequence of neurotransmitter expression by developing neurons of fetal monkey visual cortex. Dev. Brain Res. 43:69–96. Huttenlocher, P. R., and C. de Courten, 1987. The development of synapses in striate cortex of man. Hum. Neurobiol. 6: 1–9. Huttenlocher, P. R., C. de Courten, L. J. Garey, and H. Van Der Loos, 1983. Synaptogenesis in human visual cortex— Evidence for synapse elimination during normal development. Neurosci. Lett. 33:247–252. Johnson, M. H., 1990. Cortical maturation and the development of visual attention in early infancy. J. Cogn. Neurosci. 2:81–95. Johnson, M. H., and S. P. Vecera, 1996. Cortical differentiation and neurocognitive development—The parcellation conjecture. Behav. Process. 36:195–212. Kaas, J. H., and M. F. Huerta, 1988. The subcortical visual system of primates. In H. D. Steklis and J. Erwin, eds., Comparative Primate Biology, vol. 4, 327–391. New York: Alan R. Liss. Kalil, R. E., and M. Behan, 1987. Synaptic reorganization in the dorsal lateral geniculate nucleus following damage to visual cortex in newborn or adult cats. J. Comp. Neurol. 257:216– 236. Katz, L. C., and J. C. Crowley, 2002. Development of cortical circuits: Lessons from ocular dominance columns. Nature Rev. Neurosci. 3:34–42. Katz, L. C., and C. J. Shatz, 1996. Synaptic activity and the construction of cortical circuits. Science 274:1133–1138. Lee, D., and D. K. O’Dowd, 1999. Fast excitatory synaptic transmission mediated by nicotinic acetylcholine receptors in Drosophila neurons. J. Neurosci. 19:5311–5321. Leinekugel, X., V. Tseeb, Y. Ben Ari, and P. Bregestovski, 1995. Synaptic GABAA activation induces Ca2+ rise in pyramidal cells and interneurons from rat neonatal hippocampal slices. J. Physiol. 487:319–329. Leuba, G., and L. J. Garey, 1987. Evolution of neuronal numerical density in the developing and aging human visual cortex. Hum. Neurobiol. 6:11–18. LeVay, S., T. N. Wiesel, and D. H. Hubel, 1980. The development of ocular dominance columns in normal and visually deprived monkeys. J. Comp. Neurol. 191:1–51. Levitt, J. B., J. S. Lund, and T. Yoshioka, 1996. Anatomical substrates for early stages in cortical processing of visual information in the macaque monkey. Behav. Brain Res. 76:5–19. Levitt, J. B., T. Yoshioka, and J. S. Lund, 1994. Intrinsic cortical connections in macaque visual area V2: Evidence for interaction between different functional streams. J. Comp. Neurol. 342:551– 570.

iliescu and dannemiller: brain-behavior relationships in early visual development

143

Livingstone, M., and D. Hubel, 1988. Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science 240:740–749. Lund, J. S., 1987. Local circuit neurons of macaque monkey striate cortex: I. Neurons of laminae 4C and 5A. J. Comp. Neurol. 257: 60–92. Lund, J. S., R. G. Boothe, and R. D. Lund, 1977. Development of neurons in the visual cortex (area 17) of the monkey (Macaca nemestrina): A Golgi study from fetal day 127 to postnatal maturity. J. Comp. Neurol. 176:149–188. Lund, J. S., and T. R. Harper, 1991. Postnatal development of thalamic recipient neurons in the monkey striate cortex. III. Somatic inhibitory synapse acquisition by spiny stellate neurons of layer 4C. J. Comp. Neurol. 309:141–149. Lund, J. S., and S. M. Holbach, 1991. Postnatal development of thalamic recipient neurons in the monkey striate cortex. I. Comparison of spine acquisition and dendritic growth of layer 4C alpha and beta spiny stellate neurons. J. Comp. Neurol. 309:115– 128. Lund, J. S., S. M. Holbach, and W.-W. Chung, 1991. Postnatal development of thalamic recipient neurons in the monkey striate cortex. II. Influence of afferent driving on spine acquisition and dendritic growth of layer 4C spiny stellate neurons. J. Comp. Neurol. 309:129–140. Lund, J. S., and T. Yoshioka, 1991. Local circuit neurons of macaque monkey striate cortex. III. Neurons of laminae 4B, 4A, and 3B. J. Comp. Neurol. 311:234–258. Mates, S. L., and J. S. Lund, 1983. Neuronal composition and development in lamina 4C of monkey striate cortex. J. Comp. Neurol. 221:60–90. Maunsell, J., T. Nealey, and D. DePriest, 1990. Magnocellular and parvocellular contributions to responses in the middle temporal visual area (MT) of the macaque monkey. J. Neurosci. 10:3233–3334. Maunsell, J. H. R., and W. T. Newsome, 1987. Visual processing in monkey extrastriate cortex. Annu. Rev. Neurosci. 10: 363–401. Maurer, D., and R. Adams, 1987. Emergence of the ability to discriminate a blue from gray at one month of age. J. Exp. Child Psychol. 44:147–156. Meissirel, C., K. C. Wikler, L. M. Chalupa, and P. Rakic, 1997. Early divergence of magnocellular and parvocellular functional subsytems in the embryonic visual system. Proc. Natl. Acad. Sci. USA 94:5900–5905. Morton, J., and M. H. Johnson, 1991. CONSPEC and CONLERN: A two-process theory of infant face recognition. Psychol. Rev. 98:164–181. Murphy, K. M., B. R. Beston, P. M. Boley, and D. G. Jones, 2005. Development of human visual cortex: A balance between excitatory and inhibitory plasticity mechanisms. Dev. Psychobiol. 46:209–221. Nassi, J. J., D. C. Lyon, and E. M. Callaway, 2006. The parvocellular LGN provides a robust disynaptic input to the visual motion area MT. Neuron 50:319–327. Nusinowitz, S., D. G. Birch, and E. E. Birch, 1998. Rod photoresponses in 6-week and 4-month-old human infants. Vis. Res. 38:627–635. Owens, D. F., L. H. Boyce, M. B. E. Davis, and A. R. Kriegstein, 1996. Excitatory GABA responses in embryonic and neonatal cortical slices demonstrated by gramicidin perforatedpatch recordings and calcium imaging. J. Neurosci. 16:6414– 6423.

144

Packer, O., E. E. Hartmann, and D. Y. Teller, 1984. Infant color vision: The effect of test field size on Rayleigh discriminations. Vis. Res. 24:1247–1260. Rakic, P., 1976. Prenatal genesis of connections subserving ocular dominance in the rhesus monkey. Nature 261:467–471. Rakic, P., 1977. Prenatal development of the visual system in rhesus monkey. Philos. Trans. R. Soc. London [Biol.] 278:245–260. Reichling, D. B., A. Kyrozis, J. Wang, and A. B. Macdermott, 1994. Mechanisms of GABA and glycine depolarization-induced calcium transients in rat dorsal horn neurons in culture. Eur. J. Neurosci. 6:1275–1280. Rodiek, R. W., 1998. The First Steps in Seeing. Sunderland, MA: Sinauer. Rodman, H. R., and M. J. Consuelos, 1994. Cortical projections to anterior inferior temporal cortex in infant macaque monkeys. Visual Neurosci. 11:119–133. Sadato, N., A. Pascual-Leone, J. Grafman, V. Ibanez, M.-P. Deiber, G. Dold, and M. Hallett, 1996. Activation of the primary visual cortex of Braille reading in blind subjects. Nature 380:526–528. Sakatani, K., J. A. Black, and J. D. Kocsis, 1992. Transient presence and functional interaction of endogenous GABA and GABAA receptors in developmental optic nerve. Proc. R. Soc. Lond [Biol.] 247:155–161. Schneider, W., M. Pimm-Smith, and M. Worden, 1994. Neurobiology of attention and automaticity. Curr. Opin. Neurobiol. 4:177– 182. Sengspiel, F., P. Stawinsky, and T. Bonhoeffer, 1999. Influence of experience on orientation maps in cat visual cortex. Nature Neurosci. 2:727–732. Sharma, J., A. Angelucci, and M. Sur, 2000. Induction of visual orientation modules in auditory cortex. Nature 404:841– 847. Shatz, C. J., 1996. Emergence of order in visual system development. Proc. Natl. Acad. Sci. USA 93:602–608. Shatz, C. J., and M. P. Stryker, 1978. Ocular dominance in layer IV of the cat’s visual cortex and the effects of monocular deprivation. J. Physiol. 281:267–283. Shewmon, D. A., G. L. Holmes, and P. A. Byrne, 1999. Consciousness in congenitally decorticate children: Developmental vegetative state as self-fulfilling prophecy. Dev. Med. Child Neurol. 41:364–374. Stacy, R. C., and R. O. L. Wong, 2003. Developmental relationship between cholinergic amacrine processes and ganglion cell dendrites of the mouse retina. J. Comput. Neurosci. 456:154–166. Staley, K., and R. Smith, 2001. A new form of feedback at the GABAA receptor. Nature Neurosci. 4:674–676. Stoerig, P., and A. Cowey, 1992. Wavelength discrimination in blindsight. Brain 115:425–444. Stoerig, P., and A. Cowey, 1997. Blindsight in man and monkey. Brain 120:535–559. Sur, M., and C. A. Leamey, 2001. Development and plasticity of cortical areas and networks. Nat. Rev. Neurosci. 2:251–262. Teller, D. Y., 1984. Linking propositions. Vis. Res. 24:1233–1246. Teller, D. Y., 1998. Spatial and temporal aspects of infant color vision. Vis. Res. 38:3275–3282. Thibos, L. N., D. J. Walsh, and F. E. Cheney, 1987. Vision beyond the resolution limit: Aliasing in the periphery. Vis. Res. 27:2193–2197. Toyama, K., Y. Komatsu, N. Yamamoto, T. Kurotani, and K. Yamada, 1991. In vitro approach to visual cortical development and plasticity. Neurosci. Res. 12:57–71.

fundamentals of developmental neurobiology

Van Essen, D. C., C. H. Anderson, and D. J. Felleman, 1992. Information processing in the primate visual system: An integrated systems perspective. Science 225:419–423. Von Melchener, L., S. L. Pallas, and M. Sur, 2000. Visual behavior mediated by retinal projections directed to the auditory pathway. Nature 404:871–876. Wang, J., and T. R. Candy, 2005. Higher order monochromatic aberrations of the human infant eye. J. Vis. 23:543–555. Wattam-Bell, J., 1991. Development of motion-specific cortical responses in infancy. Vis. Res. 31:287–297. Wegerhoff, R., 1999. GABA and serotonine immunoreactivity during postembryonic brain development in the beetle Tenebrio molitor. Microsc. Res. Tech. 45:154–164. Weliky, M., and L. C. Katz, 1997. Disruption of orientation tuning in visual cortex by artificially correlated neuronal activity. Nature 386:680–685. Weliky, M., and L. C. Katz, 1999. Correlational structure of spontaneous neuronal activity in the developing lateral geniculate nucleus in vivo. Science 285:599–604.

Wiesel, T. N., and D. H. Hubel, 1965. Comparison of the effects of unilateral and bilateral eye closure on cortical unit responses in kittens. J. Neurophysiol. 28:1029–1040. Wilson, H. R., 1988. Development of spatiotemporal mechanisms in infant vision. Vis. Res. 28:611–628. Wilson, H. R., 1993. Theories of infant visual development. In K. Simons, ed., Early Visual Development, Normal and Abnormal, 560– 572. New York: Oxford University Press. Yoshida, K., D. Watanabe, H. Ishikane, M. Tachibana, I. Pastan, and S. Nakanishi, 2001. A key role of starburst amacrine cells in originating retinal directional selectivity and optokinetic eye movement. Neuron 30:771–780. Yoshioka, T., J. B. Levitt, and J. S. Lund, 1994. Independence and merger of thalamocortical channels within macaque monkey primary visual cortex: Anatomy of interlaminar projections. Visual Neurosci. 11:467–489. Yuodelis, C., and A. Hendrickson, 1986. A qualitative and quantitative analysis of the human fovea during development. Vision Res. 26:847–855.

iliescu and dannemiller: brain-behavior relationships in early visual development

145

10

Motor Systems Development ROSA M. ANGULO-BARROSO AND CHAD W. TIERNAN

Typically, motor behavior has received little attention in developmental psychology. Most textbooks include a small section or a chapter outlining the stereotyped sequence of expected gross and fine motor accomplishments of a child. For example, gross motor milestones include rolling over, crawling, standing up, and walking independently. The development of these motor milestones tends to be discussed in isolation with little apparent relevance to the psychological development of the child. Within this chapter, we wish to accomplish five major objectives. First, we stress the importance of the motor systems in the wholeness of child development. Second, we will present a theoretical approach that facilitates the understanding of motor development as a complex and dynamic process where exploration (flexibility) and selection (stability) of motor actions coexist to allow the acquisition of skilled motor behavior. Third, we will expose and delineate the differences between postural and movement control. Fourth, we will summarize the neural substrates related to the development and control of posture and the different aspects of movement such as initiation, speed, delays, coordination, and sequencing. Fifth, we will outline the existing relationships between the motorcognitive and motor-emotional domains. To end, we will draw conclusions in an effort to guide therapeutic interventions in the motor domain.

Relevance of movement in the developing human being When considering the most common activities performed by a human being, one finds the involvement of some form of motor action almost invariably in every activity. The production of speech involves the movement of the tongue, lips, and vocal cords; typing on a computer requires the action of the fingers and the maintenance of good postural control; walking, eating, writing, dancing, and even sleeping entail movement. When the motor system of an individual fails, life becomes difficult and limited in many ways. Would it not make sense then to think that movement is an essential part of development? In fact, motor behavior is at the core of development, having implications for attention, motivation, perception, memory, and planning. It is imperative to realize that every motor action not only generates the physical movement, but also generates perceptual information by way of the proprioceptive systems, and it also constitutes a means to learn about anticipation, memory, planning, and

consequences of action. As stated by von Hofsten (2004), “Motor development is not just a question of gaining control over the muscles; equally important are questions such as why a particular movement is made, how the movements are planned, and how children (or even infants) anticipate what is going to happen next” (266). In fact, von Hofsten proposes an action approach to motor development, according to which planning of movement and prediction of its consequences are critical to understanding the emergence of skilled motor behavior. He proposes that actions are fundamentally different from reflexes and that even newborns’ movements are never just reflexes. In fact, abundant evidence supports the view that most newborn motor behaviors are flexible goal-directed actions. From rooting to sucking, gaze, and hand orienting, these actions have been shown to be flexibly controlled and adaptive depending on the context (Craig and Lee, 1999; Farroni et al., 2002; Haith, 1980; Rochat and Hespos, 1997; von Hofsten, 1982). An example of such flexibility can be found in studies by van de Meer where newborns and infants modified their spontaneous arm movements, so (1) the arm functionally controlled a weight in order to keep the arm in view (van de Meer et al., 1995) and (2) the hand was kept within the range of a light beam so the infant could see it (van de Meer, 1997). If moving the weight had no functional consequences, the newborn stopped compensating for it. Similarly, when the location of the light beam was moved, infants also moved their preferred arm locations so they could see their hands. Such newborn and infant actions are, therefore, the early practice that an individual needs to learn about control of movement, motivation, planning, and prospective control (consequences of action). Furthermore, learning about motor planning, motor memory, and motor consequences has implicit effects in the development of cognitive skills. Hence, these early actions may constitute the core of higher order cognitive development. Only a few of the well-accepted psychological theories of human development stress the importance of movement as a core foundation to early development: Piagetian theory (Piaget, 1952) and dynamic systems approaches (DSA) (Thelen and Smith, 1994; Thelen and Ulrich, 1991). Moreover, recent dynamic systems proposals claim that cognition is embodied, that is, emergent from bodily actions in the real world, not only during early development but also throughout the entire life span (Thelen, 2000; Thelen and Bates, 2003).

147

Piaget (1952) proposed that infants initially discover interesting consequences of their actions by means of repetition (circular reactions). For instance, infants bang their hands or feet on a surface to make sounds. Using repetition, infants not only control their actions and become more skilled, but they also form action-perception relationships (sensorimotor schemata) that will be later used as a foundation for higher cognitive functions. Piaget called this period the sensorimotor phase, which lasts from birth to about one year of age. From this perspective, human cognition has its origins in the sensorimotor activities of the infant. With time, however, Piaget’s assumption is that the child can only form the abstract concept of, for example, an object once he/she has been able to disconnect (overcome) the bodily component. Similarly, ignoring the information of the senses and his/her initial egocentric references is critically necessary for the emergence of formal operations (Piaget, 1954). This cognitive developmental view of becoming progressively more disconnected from action and perception as the individual gains further cognitive skills has been recently challenged by DSA. Abundant research evidence points toward a relevant interaction between motion and cognition not only in early development but also throughout the life span, as we will see in more detail in a later section (“Interconnectedness . . .”).

Theoretical approaches to motor development Dynamic systems approaches (DSA) have been commonly used to explain the development and acquisition of motor

responses (Goldfield, Kay, and Warren, 1993; Heriza, 1991; Kelso, 1995; Thelen and Smith, 1994; Turvey and Fitzpatrick, 1993; Ulrich, 1997). One of the most important tenets of this approach is that new forms of behavior emerge in a nonprescribed fashion from the cooperative interaction of multiple subsystems. Therefore, human movement is defined as the emergent behavior of a complex and adaptive system (see figure 10.1). This definition means that human beings are open systems constantly exchanging energy and changing their levels of stability. Intrinsic factors (organismic, physiological, and psychological) as well as extrinsic factors (informational cues, surface of support, and context) cooperate to accomplish a unifying goal: the task. A complex and adaptive system creates variability, or diverse and coexisting alternatives, to initiate a process of exploration and selection. Through practice and action-perception coupling, the individual ultimately finds the most adaptive forms of motor patterns (i.e., the system retains relative flexibility to adapt to task demands). Change in movement activity can be seen in dynamic terms as a series of states and phase shifts, reflecting the probability that a pattern of movement will emerge under particular task constraints. From dynamic principles, one can predict that change is facilitated by the loss of stability. In other words, highly variable movement patterns might represent an exploration stage. Some subcomponents of the system must disrupt the currently stable movement pattern so that the system is free to change its state. The components of the system disrupting the current stability and thus

Complex and Adaptive System

Intrinsic Dynamics

Extrinsic Dynamics

- Organismic - Neurophysiological - Psychological

Instability

TASK

Creation of variability Exploration

Experience/Practice Stability

- Informational cues - Surface of support - Context

Action-perception coupling Selection Adaptive Motor Patterns

Figure 10.1 Schematic diagram of the process of motor development from a dynamic systems approach. Infants and children are complex adaptive systems with particular intrinsic and extrinsic

148

dynamics. Task-oriented activities drive development by means of the processes of exploration and selection so motor patterns emerge that are adaptive (flexible) to task demands.

fundamentals of developmental neurobiology

engendering change can only be known through careful empirical study. As Thelen (1995) stated, such study is difficult because these agents may be nonobvious and changing through time. For instance, growth or physiological factors, such as leg-mass-to-strength ratios, are important in early infancy for the development of functional leg movements. However, in situations later in development, experience or socioemotional conditions may have a greater influence. The task in question dictates the relevance of the contributing factors. There are critical implications of a dynamic system’s conceptualization of motor development: (1) A child is viewed as continually shifting his or her movement pattern stability: there will be times when some motor behaviors are very stable and other times when new patterns will emerge. (2) The stability of these motor patterns depends on the degree of coherence among the important parts. (3) Change in motor behavior cannot be explained on the basis of change in one subsystem or part, but rather by the interaction between the embedded components (i.e., nervous system, body) and environment. No one part of the child is privileged in this regard. (4) Finally, the current dynamic state of the child is a function of previous states and also serves as the basis for future states. From this perspective, one can say that children become skillful motor performers because they have the capacity to maintain flexibility in the degree of coupling among the nervous system, body, and environment at the same time that they are dynamically responsive to the task at hand.

Posture, movement, and overall level of motor activity Movement scientists may have different views about the definitions of posture, motor action/movement, and motor activity today. However, it is useful to agree about functional definitions of posture and movement so we can clarify their distinct contribution within motor development. We define posture as the maintenance of a specific body configuration that minimizes movement in some parts of the body (or all of it) while facilitating motor action efficiency in other parts of the body. In this case, a motor action could be maintaining posture itself, as we do when standing on an icy sidewalk while waiting for a bus on a winter day. However, the motor action could be the performance of any goal-directed movement such as reaching for an object. In general, movement could be defined as changes in the body configuration so a goal-directed action is accomplished. Finally, as scientists it is useful to estimate overall motor activity because it gives us a gross measure of all types of movements independent of goals. A large body of literature exists describing the orderly progression of postural control and movement acquisition observed in infants and children. Many of these studies

have also examined the underlying processes of developmental changes in the aforementioned areas (for reviews of motor development, see Angulo-Kinzler, 2001a; Bertenthal and Clifton, 1998; Fentress and McLeod, 1986; Jouen and Lepecq, 1990; Schmidt and Fitzpatrick, 1996; Thelen, 1995; Zelazo, 1998). Still other motor developmental studies have focused on how these motor actions become more efficient and finely tuned to environmental demands (Adolph, 1997; Angulo-Kinzler, 2001b; von Hofsten, 1979; Thelen et al., 1993). In this chapter we chose to present the development of reaching and walking as examples of how motor skills emerge and change through time. The Development of Reaching From 0 to about 4 months of age, infants demonstrate a level of poor arm and hand motor control. They are able to move their arms, but they are not successful at grasping an object. These initial movements, however, allow infants to practice and learn about their movements during this period when perceptionaction relationships are formed and selection of more effective action takes place. Infants learn to first move their arms to the vicinity of the object. Initial improvements in reaching skill come from advancements in postural trunk control (Spencer et al., 2000), eye-hand coordination (von Hofsten, 1982), and adjustments in arm velocity and muscle forces (Spencer and Thelen, 2000; Thelen et al., 1993). During this phase of spontaneous pre-reaching movements, infants explore a range of patterns and select those that take their arms close to the desired toy. Nevertheless, the range of patterns that infants explore is not infinite but rather is constrained by their initial preferences and other neurobiomechanical factors such as formation of linear synergies (i.e., coupling of two or more joints to reduce movement options) (Zaal et al., 1999). As infants transition to their first successful reaches, the muscle patterns change from a more predominant use of biceps and triceps to a preferred use of the deltoid muscle (Spencer and Thelen, 2000). Still, these reaches lack the smoothness in the hand path and the bell-shaped velocity profile observed in adult reaching (Morasso, 1981). Subsequent changes in the reaching abilities of infants continue during the second half of the first year. In fact, infants show increased hand skill in several dimensions. The 6-month-old generally grasps an object with the entire hand, showing only primitive aspects of coordination between the palm and fingers. By 9–10 months, the child begins to use the forefinger in grasping, and reaching and grasping are coordinated into one continuous movement. Although infants as young as 6–8 months are capable of precision grasping, the developmental trend is an increase in use of the pincer grasp over time along with progressive elimination of ulnar and palmar grasps. Younger infants do not

angulo-barroso and tiernan: motor systems development

149

exhibit a consistent grasp pattern with the smallest objects, while older infants use the inferior forefinger grasp or the pincer grasp. In addition, 6-month-old infants do not change grasp configurations as a function of object size, while 9month-olds adjust their hand opening (amplitude) accordingly. Also at 9 months, infants can correct hand trajectory to a moving target even after they started their reaching action. Initiating this adaptation, however, takes twice as much time as it does in adults (Berthier and Robin, 1998). In other words, the last half of the first year is characterized by the acquisition of a more stable form of reaching, which in turn allows the development of other forms that are more adaptive. Concomitant changes at the neuromotor systems level also occur during the first year of life. It has been proposed that direct connections between neurons in the cortex and motor neurons in the spinal cord by way of the corticospinal tracts are critical for the control of fine hand movements (Kuypers, 1982). These neurons and connections have a special role in precision grasping, as they are active in a pincer grasp but not in a palmar grasp (Muir and Lemon, 1983). Increasing hand skill parallels the decline in latencies of motor evoked potentials (motor responses resulting from stimuli over the motor cortex), increase in conduction velocity, and increase in myelination and axon diameter of the relevant corticospinal tracts (Eyre et al., 1991). However, other neuromotor areas are also important in the development of skillful reaching, since reaching involves motivation, attention, control of movement, planning, anticipation, and prospective control. The Development of Walking Infants undergo substantial changes during their transition to adultlike independent, upright locomotion over the first two years of life. Such transitions became the focal point of developmental research starting in the 1930s and 1940s with the work of Gessell and McGraw. McGraw (1945) suggested that infants progress sequentially through seven phases to achieve adultlike erect locomotion and ascribed maturation within the nervous system as a cause for these development acquisitions. The first phase was known as “reflexive stepping,” which occurs during the first two months of life. She described these movements as primitive and subcortically driven. The next period was classified as the “static phase.” During this time, there is an apparent decline in stepping behavior, which she attributed to cortical inhibitory processes. Third, the “transition phase” is marked by heightened variability in leg movements and a difficulty discerning whether or not the stepping movements are reflexive or deliberate. In the fourth phase, stepping is described as “deliberate”: infants take steps in an intentional manner resulting from more direct cortical participation. Next, rudimentary elements of “independent stepping” can be observed. These movements, although not yet refined, are

150

characterized by the integration of posture and forward propulsion of the legs. The sixth phase was referred to as “heel-toe progression,” where coordination improves and infants more consistently initiate stance with a definite heel strike and end with pushing their toes off the ground. The final stage, “integrated walking,” shows refined coordination and smooth and automatic independent upright locomotion during the second year of life. While the general development sequence described by McGraw is largely agreed upon, both the extent of the role of the nervous system and the deterministic nature of these characterizations of walking has been largely debated over the years. According to Thelen and Smith (1994), McGraw herself later acknowledged that histological changes in the brain and the assumption of localization of function were a much too simplistic view of development. Similarly, Gessell (1945) stated that it was more appropriate to describe such developmental processes as dynamic and nonlinear, rather than stagelike and maturationally driven. Regardless, neural accounts of walking have still been popular throughout the years. In particular, many researchers believe that “central pattern generators” (CPGs—a set of neurons in the spinal cord capable of a patterned neural activation which matches that needed for gait) drive the development of adultlike walking because these neural networks within the spinal cord can generate muscle-specific activations (i.e., alternating activity of flexors and extensors) (Grillner, 1975, 1981; Forssberg, 1985). Forssberg (1985) argued for a hierarchical explanation of walking, whereby CPGs were the basis for locomotor development. He contended that adultlike locomotion resulted from the refinement of these neural networks as higher brain centers develop. Much of the support for CPGs stems from animal research, including spinalized cats that have been shown to produce hind-limb movements similar to those of their nonspinalized counterparts (Forssberg, 1980a, 1980b). Neural network accounts have been challenged, however, by proponents of DSA who question the functional capabilities of neural networks and argue that the nervous system should not be seen as a privileged component within the system. For example, Thelen and Fisher (1983) examined muscle activation patterns of infants during early stepping movements and concluded that the observed activation patterns were far too complex to be explained by CPGs. In addition, the decline in stepping after two months of age has been shown to be influenced by other than purely neural factors. Research has shown that step frequency in early infancy is related to infant arousal and rate of weight gain (Thelen et al., 1982). Similarly, Thelen, Fisher, and Ridley-Johnson (1984) manipulated leg mass by placing young infants upright in warm water to decrease the load of their legs. Results showed that infants increased their step rate when compared to the out-of-water condition. Finally,

fundamentals of developmental neurobiology

What about Overall Level of Motor Activity: Does It Have Any Role in Development? In the past, quantifying the frequency of motor activity was beset with a number of problems, including questionable reliability of self-report measures, time-consuming and intrusive direct observations, and complicated physiological measures such as heart rate telemetry and energy expenditures (Sirard and Pate, 2001). Recent advances in the memory capacity and sensitivity of activity monitors have improved the reliability and ecological validity of activity-based research (Puyau et al., 2002). The small lightweight devices simply attach like a wristwatch to the limb or to the trunk of the participant and allow researchers the opportunity to collect less intrusive, ecologically valid, and objective motor activity data for extended periods of time in children (Trost et al., 2005). Traditionally, activity monitors were used in sleep research (for review see Sadeh and Acebo, 2002) but more recently these devices have been shown to be sensitive enough to detect differences between spontaneous motor activity levels in awake and alert infants with and without iron-deficiency anemia (Angulo-Kinzler et al., 2002a, 2002b). Additional work in our laboratory has examined the spontaneous leg motor activity of infants with and without Down syndrome (DS) from 3 to 6 months of age. We used activity monitors to objectively quantify the frequency of spontaneous motor activity for 48 continuous hours. Data analyses, controlling for movement artifacts, indicated that infants with DS spent more time in low-intensity activity during the day and the night (see figure 10.2). Furthermore, the level of low motor activity showed a significant relationship with the onset of locomotor activities such as crawling and walking in both groups. That is, infants who spent more time in low activity also showed later onset of locomotion (McKay and AnguloBarroso, 2006). These results are important for illustrating the relationship between the amount of motor activity and motor development as measured by the onset of important milestones.

Time in Low Intensity Activity (min)

treadmill training has been found to increase functional stepping patterns in infants who initially showed little or no stepping (Thelen, 1986; Vereijken and Thelen, 1997). Advocates of DSA acknowledge the critical role of nervoussystem maturation in locomotor development but argue that such findings illustrate the importance of other, nonneural factors (i.e., body composition, strength, motivation) that cooperate with the nervous system during locomotorrelated tasks. In addition, the aforementioned findings suggest that locomotor development is not fixed and can be influenced by early intervention, a possibility that has implications for individuals with locomotor delays. Recently, Ulrich and associates (2001) found that treadmill training facilitated the onset of independent walking in infants with Down syndrome.

500 450 400 350 300 250 200 150 100 50 0

*

*

*

DS TD

3

4

5 Months

6 Day data

Figure 10.2 Group means and SD values of time spent in lowintensity activity in infants with Down syndrome (DS) and typical development (TD) at 3 through 6 months of age. p < .05. (Data published by McKay and Angulo-Barroso in Infant Behavior and Development 29:153–168; reprinted with permission.)

Neural substrates of motor behavior As reviewed in the section on our theoretical approach to motor development, it is important to note that although the nervous system is an important contributor to development, it is not the only one. Previous research has shown that many other physical and contextual factors, such as body dimensions, proportions, postures, and inertial properties, contribute equally to motor behavior (Thelen and Fisher, 1982, 1983; Thelen, 1986; Ulrich et al., 2001). Although dynamic approaches to development make no explicit proposal of a neurobehavioral theory, they are perfectly compatible with neurodevelopmental theories that view the brain as a dynamic and complex system whose development complies with the similar dynamic principles proposed for motor development (Sporns and Edelman, 1993; Stiles et al., 2005). Currently, little is known about the specific changes occurring at the brain level that accompany the acquisition of new motor behaviors. The neural mechanisms that could explain the changes that have been described in the development of reaching or walking, for instance, are rather complex and involve many neuronal structures. A seemingly simple task such as reaching for an object requires motivation, goal identification, visual processing, postural control, planning the response, setting the appropriate amount of force and correct direction of movement, timing the different muscles involved, and so on . . . Since many of these processes occur in parallel, it is implied that many interconnected areas of the brain are activated. For instance, activity related to the visual cue of the object to be reached activates not only visual association areas but also areas related to motor planning and preexecution such as the supplementary motor/ premotor areas, cerebellum, and basal ganglia, all of which

angulo-barroso and tiernan: motor systems development

151

are involved in motor planning, motor memory, and learning ( Jeannerod, 1997; Schwartz, 1994). Similarly, the neurons that code for reaching direction involve not only those in the primary motor cortex (Georgopoulos, Kettner, and Schwartz, 1988) but also those in the parietal cortex. It seems that the distinction between what defines a visual cue, planning, motor memory, and execution of the action is becoming blurred as we learn more about the activation patterns of the entire brain (Anderson et al., 1997). Although accepting a wide distribution and multiple representations of motor skills in the young brain may seem problematic, it also brings the advantage of permitting exploration (flexibility) and selection (stability) in a child’s motor repertoire. In addition, it implies that brain areas are multiply and densely connected and that the child’s experiences with motor skills have a structural and functional impact on the development of the brain itself (experience-dependent plasticity, or more precisely in our case, activity-dependent plasticity; Jones, Kleim, and Greenough, 1996). From this perspective, brain development is a dynamic process. As stated by Stiles (2001): “the developing brain is a dynamic, responsive, and to some extent self-organizing system” (p. 266). One important aspect of a dynamic and complex system, such as the brain, is its capacity to self-organize and therefore generate stable patterns of action (Kelso, 1995). Throughout development, changes in the volume and activity of the different areas of the brain are not uniform, especially when comparing subcortical and cortical regions, or among the different cortical areas. For instance, peak density of synaptic connection occurs earlier in the visual cortex (4–12 months) compared to prefrontal cortex (after 12 months) (Huttenlocher, 1990; Huttenlocher et al., 1982; Huttenlocher and Dabholkar, 1997). Additionally, areas of higher activity (represented by glucose uptake) in the basal state in newborns are the sensorimotor cortex, thalamus, brain stem, and cerebellar vermis. In contrast, parietal, temporal, and occipital cortices, along with the basal ganglia and cerebellar cortex are most active at 3 months of age (Chugani, Phelps, and Mazziotta, 1987). Interestingly, the most active areas in the newborn phase are those that underlie motor execution, whereas those most active in the infancy phase are already capable of motor planning. Whether activitydependent experience is the guide for an already highly interconnected brain or the interconnections are formed as the child learns is still under debate ( Johnson, 2001). Further changes in brain connectivity and metabolism continue through development, including a temporary phase where these levels surpass those of adults. However, once the brain has reached a certain level of maturity, neuroscientists are more ready to ascribe specific function to particular areas of the brain. Table 10.1 summarizes the most relevant areas of the neuromotor system and their corresponding functions. The interconnectedness among these areas and

152

pathways is summarized in figure 10.3. Please note that these graphics only begin to depict the complexity of the neuromotor system. As children get older the efficiency of each of these pathways improves, but they do not necessarily do so at the same time. An example of these developmental improvements is discussed in the work of Garvey and associates (2003) where they examined the cortical correlates of neuromotor development in healthy children using focal transcranial magnetic stimulation (TMS). They demonstrated that the corticospinal pathway improved its efficiency with time as evidenced by decreases in the threshold of motor evoked potentials (MEP) and decreases in the ipsilateral silent period (iSP), with the latter defined as a transient interruption of ipsilateral voluntary muscle activation. These measures were recorded while children of different ages performed fingertapping movements. It is thought that MEP threshold denotes the developmental myelination stage. However, it is important to note that myelination does more than improve conduction velocity—it also helps provide nutritional and structural support for neurons, in addition to aiding neuronal and neurotransmitter activity modulation (Fields, 2004). White matter within the nervous system has gained more and more functional relevance in recent years. For instance, reductions in frontal white matter have been found in otherwise healthy, male children who exhibit complex stereotypies (Kates, Lanham, and Singer, 2005). However, iSP is thought to depend on transcallosal connectivity and is mediated by inhibitory circuits in the motor cortex. Interestingly, the iSP is absent in children under 6, but it is present in children older than 10 years of age (Heinen et al., 1998). Therefore, transcallosal pathways seem to be involved in the production of finger movements, but other motor behaviors such as bilateral hand movements and mirror movements are affected by the corpus callosum function as well (Rademaker et al., 2004). To this point we have focused on brain areas involved predominately in motor functioning. However, other brain areas often ascribed to motor behavior, namely, the basal ganglia and the cerebellum, are also relevant to cognitive and emotional development (Allin et al., 2001; Diamond, 2000; Middleton and Strick, 1994, 2000). For example, the basal ganglia have two major interconnected loops: one for motor activity (cortico-putamen-thalamo-motor cortical circuitry) and another for cognitive–emotional processing (cortico-caudate-thalamo-prefrontal cortical circuitry). Similarly, the neocerebellum is more dedicated to cognitive tasks, while the rest of the cerebellum is devoted to motor control. Interestingly, both the cerebellum and basal ganglia are somatotopically organized, similar to the organization seen in the motor and sensory cortices (i.e., adjacent or connected fields of these areas correspond to adjacent parts of the body). Both structures develop relatively late in neuronal

fundamentals of developmental neurobiology

Table 10.1 Motor systems: components, afferent and efferent pathways, and functions (only most relevant are highlighted) Motor Systems Function Muscles Executors of action Developmentally: from Functionally: better motor control Motor unit = motor neuron and (as internal forces, more to less muscle implies fewer muscle fibers per (MN) and muscle fibers Motor but external forces fibers per MN (i.e., MN that innervate Units also impact initially less capacity motion) for fine motor control) Spinal Efferent: Corticospinal and many other Reflexes, inhibitory Execution of motor action Cord Motor from cortex spinal efferents (see BS) interneurons, CPGs Assist in coordination, locomotion (SC) Afferent: Dorsal column–medial Proprioception, fine Sensory info to lemniscal pathway touch Convey sensory info from action to BS-cortex and Anterolateral pathway Coarse touch, pain cortex and CB cerebellum Spinocerebellar (see CB) Somatosensory Brain Stem Efferent To SC and cortex Vestibulospinal Flexor and extensor tone, balance, (BS) Reticulospinal skilled arm/hand movement (cortico) Rubrospinal Afferent To cortex Spinocerebellar Convey somatosensory info from To cerebellum (see CB) action to cortex and CB Efferent Cerebellum Afferent (CB) Vestibulocerebellar Flocculonodular To axial MN Axial motor control and balance (lateral vestibular nucleus) Spinocerebellar Vermis To reticular form., and Axial and proximal motor control, (medial) (fastigial nucleus) MI ongoing execution Spinocerebellar Intermediate part. hemisphere To red nucleus and MI Distal motor control, ongoing (lateral) (interposed nucleus) execution Cerebrocerebellar Lateral part. hemisphere To red nucleus and Initiation, planning, timing of (dentate nucleus) premotor cortex movement Cognitive components of motor action Basal Motor loops Putamen Direct: Facilitation of Selection of movement, bimanual Ganglia movement (excited by coordination, sequential (BG) dopamine) movements Indirect: Inhibition of movement (inhibited by dopamine) Cognitive-emotional components of Emotion-cognitive Caudate movement loops Somatotopic organization, Motor Primary motor (MI) Output to SC Execution, force, and direction of input from PM, SMA, SI, Cortices (corticospinal) movement CB (reticulospinal) Premotor (PM) Input from PPC, CB, and Output to MI, BS Planning, preparation for next SMA movement, postural orientation

Other Cortex Areas

Supplementary motor area (SMA)

Input from PPC and BG

Output to PM and MI

Primary sensory (SI)

Somatotopic organization, input via thalamus Input SI, PM, VC, and limbic areas of cortex

Homotopic connections Output to MI, PPC Output PM, SMA, lateral CB

Posteriorparietal (PPC)

Planning, bimanual coord., sequential movements, motor imagery, coordination, balance, and movement Perception of somatosensory information Motivation to move, visual-motor transformation, spatial-visual attention

Note: Additional functions CB: motor coordination, sensorimotor integration, fine adjustment of muscle tone, sensory discrimination, posture, motor learning, motor timing (lateral hemispheres; Keele and Ivry, 1990), comparison of intended versus actual movement (Ghez, 1991), anticipation (feed-forward control), cognitive aspects of motor behavior. Optimizing movement by monitoring outcome of movement (sensory information processing; i.e., afferent component of action). BG: bimanual coordination and sequential movements (planning series of actions), higher order aspects of visual processing, cognitive and emotional aspects of motor behavior. Activated when selection of movement (efferent component of action). Both, CB and BG: improving motor performance, motor learning. Primary and premotor cortical areas: movement force and direction, execution of movement, involved when decision about next movement. Prefrontal cortical areas and BG (striatum): when attention demands increase in the movement.

Temporal

Parietal

Prefrontal

PPC Sensory Cortices via

TL

Sensory recept.

VC

vi

PM vi

Eye Muscles

L aT

SI

(Eyes)

Motor Cortices

via

TL

SMA via TL

L aT

BG

MI

CB via TL

TL

L via T

Corticospinal

via

BS

SC Sensory receptors

Muscles

(muscle, skin, joint)

Figure 10.3 Motor systems: A simplified version is presented with only the most relevant connections of both afferent and efferent pathways. Pathways of the visual system are clearly incomplete. The vestibular influences in movement have been ignored in this diagram. Cerebral cortex is represented in the top layer of the diagram with the exception of five important sensorimotor areas that have been placed lower and separated to facilitate representation of connectivity. These areas are PM (premotor), SMA (supplementary motor area), MI (primary motor area), SI (primary sensory areas), and VC (visual cortices). In addition, the posteriorparietal cortex has been highlighted because of its important role

in motor motivation and visual-motor transformations. BG (basal ganglia), CB (cerebellum), TL (thalamus), BS (brain stem), SC (spinal cord). Note that BG receives input from most cortical areas, while CB receives input mainly from sensorimotor areas. Also note the rapid efferent path between MI and neurons in SC activating muscles (corticospinal) and the fast afferent path from SC to CB allowing consequences of action to be processed quickly. These two pathways have been represented with thicker arrow lines. Finally, note the important role of the thalamus as a hub to relay information.

ontogeny. As the infant and child mature, the development of the cerebellar cortical circuitry closely parallels motor coordination and motor learning (Swinny, van der Want, and Gramsbergen, 2005). Extrapolation from animal models indicated that cerebellar cortex circuitry starts developing in the last trimester of pregnancy and lasts until beyond the first year of life (Gramsbergen, 2003). Indirect estimates of synaptic formation suggest a significant increase during infancy and childhood in the gray matter of the cortex, thalamus, and cerebellum. In contrast, a constant high level of synaptic formation was found in the entire basal ganglia (Pouwels et al., 1999). The basal ganglia (striatum [caudate/ putamen], globus pallidus, subthalamic nucleus, and substantia nigra) are differentiated at birth, start myelinating prenatally (Chugani and Phelps, 1986; Chugani, Phelps, and Mazziotta, 1987), and mature rapidly over early postnatal years. Because the basal ganglia send output to the supplementary motor area, tasks that involve bimanual coordination and sequential movements are thought to be affected by BG dysfunction (Ronald et al., 1980).

Relative increases in the activation of all motor cortices and association areas seem to occur during rapid learning periods or when the task demands increase drastically. As the motor action becomes skilled and well learned, the activations of the primary/premotor cortex seem to take over. In such cases, the cognitive and even emotional load is reduced, so deactivation of nonmotor cortical areas is expected (Muller et al., 1998). When the motor task has a large learning component, evidence shows a shift from more anterior to more posterior cortex activation as the learning consolidates (Shadmehr and Holcomb, 1997). As we have seen, the brain dedicates multiple, largely interconnected, and in some ways redundant areas to the control of movement (Schwartz, 1994; Georgopoulos, 1995). Redundancy and interconnectedness in the neuromotor system at the cortical and subcortical level mean that there are multiple ways to achieve a given motor behavior (Passingham, 1993). As motor actions are explored and selected, concomitant changes in neural substrates also occur. For instance, Martin and associates (2004) showed

154

fundamentals of developmental neurobiology

that corticospinal development depends on motor experience. The relationship between neural substrates and motor behavior is bidirectional (Kleim, Jones, and Schallert, 2003). In the same way that ongoing activity may change a neural state, the reorganization of the neural structures may also have an impact on behavior. According to Stiles (2001), the capacity to reorganize and change (i.e., plasticity) is crucial to neural development. However, neural plasticity is not unique to the postnatal period. It is preserved during adulthood, although to a lesser extent. In fact, plasticity is a basic process that underlies not only neural but also motor and cognitive functioning (Stiles, 2001).

Interconnectedness between motor, cognitive, and emotional domains Motor activity is interrelated with several domains, including perception, cognition, and emotion. We use perceptual information to execute both gross motor activities, such as walking down the street, and fine motor skills like taking a cap off of a pen. The perceptual information generated allows us to adapt our step length to clear a puddle or adjust our finger force to uncap a pen efficiently. However, the motor activity we generate also influences our perceptual skills. This cyclical process is known as perception–action coupling, where correlations between the information generated from perceiving and acting exist (Edelman, 1987; Gibson, 1988). While this phenomenon is important for understanding development, it is also quite intuitive and well documented. For the purpose of this chapter, we will concentrate on the relation between the motor domain and two domains more relevant to psychology—cognition and emotion. Motor–Cognitive Relationships Based on observations of his children, Piaget (1954) proposed that self-produced locomotion was related to the development of spatial cognition. Since then, more direct evidence has strengthened this postulation by means of spatial-orientation and spatialsearch-performance tasks. Findings from Acredolo (1978) renewed interest in the possibility that locomotor experience may influence the development of spatial orientation. In her study, infants were first trained to locate an object in a window to either their right or left side. Following the training, infants were repositioned so they were facing the opposite direction and then asked to search for the object in the original window. Results showed that young infants almost always looked to the incorrect window (egocentric coding), whereas infants in their second year of life were much more successful in looking to the correct window (allocentric coding). Furthermore, a shift from egocentric to allocentric coding seemed to coincide with established locomotor trajectories. The idea was that crawling

experience might facilitate the development of allocentric coding because once infants begin locomoting, an egocentric coding strategy would have to be continuously updated, a process that is very inefficient. Eventually, infants would begin to adopt allocentric strategies that are independent of the infants’ orientations in space (see Bremner, 1978; Bremner and Bryant, 1977). Researchers have subsequently manipulated locomotor experience while controlling for age on the aforementioned spatial orientation task and found that, in fact, infants with more locomotor experience were more likely to use allocentric coding (Enderby, 1984; Bertenthal, Campos, and Barrett, 1984). Therefore, the literature in this area suggests that the development of spatial orientation is impacted by infants’ experience with locomotion. Similar conclusions regarding locomotion and spatial search have also been drawn. Briefly, the traditional A-not-B task requires infants to sit at a table and search for an object hidden in one of two locations in front of them (A or B). The object is first repeatedly hidden at A and then eventually hidden at B. When the object is hidden at B, young infants often fail to search at B, making the “A-not-B error.” An abundance of research suggests that experience with selfproduced locomotion correlates with success in the A-not-B task (Acredolo, 1985; Bell and Fox, 1992; Bertenthal and Campos, 1990; Horobin and Acredolo, 1986; Kermonian and Campos, 1988). Further, it appears that providing infants with experience in locomotion in otherwise prelocomotor infants using artificial walkers increases success on B trials (Bertenthal and Campos, 1990). It has been argued that locomotor experience may enhance spatial search capabilities in infants because it both demands and sets up contingencies that teach spatial discriminations (Smith et al., 1999). It should be noted that the role of locomotor experience in development of spatial cognition is deemed to be one of facilitation, rather than one of necessitation. In addition to the link between locomotion and spatial cognition, motor–cognitive relationships have been identified in other areas. One example involves studies regarding motor imagery. In a comprehensive review of this topic (see Decety, 1996), functional correlates of motor imagery have been determined. It was concluded that the timing of mentally simulated actions closely mimics actual movement times. Similarly, autonomic responses during actual exercise appear to rival those during motor imagery. Finally, cerebral blood flow increases were observed in the motor cortices involved in planning movements. According to Decety (1996), these results suggest that imagined and actual movements share the same neural substrates, at least to some extent. Further support for the interrelatedness between cognitive and motor behaviors and their neural bases has been suggested by others as well. For example, Diamond (2000) has suggested that both the prefrontal cortex and cerebellum

angulo-barroso and tiernan: motor systems development

155

are important for both cognitive and motor tasks based on the close activation of these two areas in functional neuroimaging, in addition to the fact that children with “cognitive” disorders, such as ADHD, demonstrate motor deficits in conjunction with the well-documented cognitive deficiencies. Further, children with ADHD have been shown to have abnormal development and activity in both the cerebellum and prefrontal cortex in numerous studies (Berquin et al., 1998; Castellanos et al., 1996; Amen, Paldi, and Thisted, 1993). Collectively, research suggests that there is a connectedness between motor activity and cognition to a higher degree than previously thought. However, the extent to which these two areas influence each other and in what manner warrants further investigation. Motor–Emotional Relationships The word emotion comes from an old French word, esmovoir, which means “to set in motion.” In fact, emotions can be understood as a drive for the generation of action and thought (Thelen and Smith, 1994, 314). Motor activity is an important component of behavioral expression that shares with emotion many of its neural pathways in addition to some of its primary physiological correlates—heart rate and cortisol responses. Since the time of Darwin, a number of links between motion and emotion have been proposed. In the adult literature on emotion, several links have been made between emotion and gait (Montepare, Goldstein, and Clausen, 1987) as well as between motion quality and rated personality characteristics (Grammer, Honda, and Juette, 1999). In infants or children, co-occurring links between motion and emotion have rarely been examined. However, we know from iron deficiency and other nutritional insults (Lozoff and Black, 2004) that children who have such nutritional deficiencies have both motor and emotional characteristics that differ from normally developing children. In our research together with Lozoff and her colleagues, iron-deficient children were less engaged emotionally with flatter affect to a wide range of situations. The same children were also delayed in their general motor development and showed decreased motor activity (Angulo-Kinzler et al., 2002a; Lozoff, 1991). Developmentally speaking, it is possible that newly acquired patterns of activity may have implications for arousal and a developing sense of self-control (Robertson, Bacher, and Huntington, 2001; Watson, 1966, 1972). In addition, behavioral (arm, facial) and heart rate responses of 4-month-old infants appear to be correlated such that infant learning led to increases in operant arm responses and expressions of positive emotions (Lewis, Hitchcock, and Sullivan, 2004). In this study, there was an inverse relationship between facial expression and heart rate—when frustrated, infants’ negative expressions and heart rate increased. During the first few months of life, infants also show an increasing degree of organized periodicity of movement and

156

better establishment of the activity/sleep circadian rhythm, and the establishment of this rhythmicity coincides with day-night periodicity of cortisol release (Gossel-Symank, Grimmer, and Siegmund, 2004; Kiess et al., 1995; Mantagos, Moustogiannis, and Vagenakis, 1998). Thus there are important links between motor activity, motor development, emotional tone, and physiological activation. Additional infancy research provides support for a motoremotional relationship. Fear in infancy has often been studied by examining infants’ wariness of heights during their performance on the visual cliff. After a series of studies, Campos and colleagues concluded that experiences generated by locomotion (both self-generated and artificial) make possible the development of wariness of heights (Campos, Bertenthal, and Kermonian, 1992). Similarly, locomotion is believed to be important for the development of attachment because it provides infants with physical proximity to the caregiver and allows them to move independently toward novel and possibly away from frightening situations. This locomotor experience also causes infants to be more in tune with their caregivers’ locations, show more distress during separation from them, and look toward them more often in ambiguous situations (Ainsworth et al., 1978; Bowlby, 1973; Campos, Bertenthal, and Kermonian, 1992). Finally, motor activity has also been linked to behavioral and mental health conditions. For example, early problems of overactivity, along with impulsivity and aggression, appear to be risk markers for a broad range of externalizing problems later in life, particularly ADHD (Barkley, 1998; Berger and Posner, 2000; Campbell, Pierce, and March, 1994). Among lower risk groups, infants who had high levels of motor activity and negative affect in response to novel auditory and visual stimuli at 4 months of age were likely to be behaviorally inhibited toddlers and described as “shy” with high cortisol levels at 4 years of age (Kagan and Snidman, 1991; Schmidt et al., 1997). These types of relations between motion and emotion have been argued to stem from overlapping pathways in the brain—specifically, in the basal ganglia and the hypothalamus, which rely on intact involvement of the dopaminergic pathways. The dopaminergic system and prefrontal cortex involvement in ADHD-related motor problems discussed earlier have long been suspected (Hoover and Strick, 1993; Swanson et al., 2000) and are known to be involved in motor and emotional development problems.

Guides for therapeutic interventions in the motor domain Early intervention programs designed to ameliorate motor developmental delays in infants and children have been applied for many years. However, the evidence collected thus far from such studies is largely inconclusive. The latest

fundamentals of developmental neurobiology

systematic review of the effects of early intervention on motor development (Blauw-Hospers and Hadders-Algra, 2005) concluded that a potential benefit for motor development exists from those early interventions using the Newborn Individualized Developmental Care and Assessment Program (NIDCAP) during the newborn intensive care unit (NICU) period. The NIDCAP is a holistic and naturalistic observation approach to best fit assessment and health care to the individual needs of each newborn. More robust effects were also found when utilizing interventions based on specific motor training or from general developmental programs, which yielded positive effects on motor development during the postnatal period. Nevertheless, the authors claim that further research is necessary to address issues such as what type of intervention is most beneficial or when to start the intervention. Interestingly, animal models of early intervention have demonstrated an interaction between the maturational process of the nervous system and timing of the training program. In these studies, rat pups were trained to recover good postural control before and/or after a cerebellectomy. If this surgery was conducted around 10 days of age, training either pre- or postoperative was ineffective. However, preoperative training was shown to be effective when surgery was done at 15 days of age, while postoperative training showed the greatest positive effects at 24 days of age (Zion et al., 1990). Additionally, these animal models have also suggested that the best combination for the timing of preand postoperative training is when there is training preinjury, then a rest period of a few days after the injury, followed by postinjury training (Caston, Jones, and Stelz, 1995). Additional research is needed to further delineate the link between early intervention and neuromotor development, particularly in humans. Such studies could have a tremendous impact for children with motor difficulties, rehab professionals, and scientists in this area. Based on the current knowledge of motor development and neuroscience, we propose the following practical points to be taken into consideration when designing an early intervention: 1. Active participation in therapy should be promoted to facilitate self-generated movements. 2. Diverse and challenging opportunities for exploration should be provided to increase repetitions of self-initiated movement. 3. Ability to change posture and movement patterns is more important than ability to sustain them. 4. Focus on periods/phases when motor patterns are less stable (more flexibly organized). 5. Concentrate on quality and quantity in relation to function. 6. Measure movement patterns in a natural environment whenever possible.

7. Focus on perceptions generated by movement and/or environment. 8. Identify perceptual deficits related to impaired movement patterns. 9. Determine variables (action, perceptual, cognitive, emotional) that constrain or facilitate change in movement patterns (both intrinsic and extrinsic).

Conclusions Based on the theoretical approach to motor development outlined in this chapter, as well as the evidence collected on early intervention programs designed to enhance motor development, the following theoretical guidelines to better understand motor development are proposed: First, changes in motor behavior are multicausal, context dependent, and selforganizing. Therefore, factors not directly or “obviously” related to the behavior may surprisingly affect change in the pattern. For instance, weight gain was critical in the appearance/disappearance of stepping. Both intrinsic and extrinsic factors may play important roles to drive change. Second, exploration and selection are two necessary processes that underlie the emergence of new behaviors. Without the creation of variability (exploration) and the establishment of a certain degree of stability (selection), advances in motor development are impossible. These two processes are not mutually exclusive, but rather they coexist throughout development. Third, stability/instability is a fundamental concept when assessing the functional quality of performance, the potential for change, and the effectiveness of a therapeutic intervention. Thus stable movement patterns must be identified in normal populations for comparative purposes. In addition, one ought to recognize that behaviors with high stability are very resistant to change. In such cases, strategies to induce some degree of instability may be necessary at the early stage of the intervention.

REFERENCES Acredolo, L. P., 1978. The development of spatial orientation in infancy. Dev. Psychol. 14:224–234. Acredolo, L. P., 1985. Coordinating perspectives on infant spatial orientation. In R. Cohen, ed., The Development of Spatial Cognition, 115–140. Hillsdale, NJ: LEA. Adolph, K. E., 1997. Learning in the development of infant locomotion. Monogr. Soc. Res. Child Dev. 62(3):i–vi, 1–158. Ainsworth, M., M. Blehar, E. Waters, and S. Wall, 1978. Patterns of Attachment. Hillsdale, NJ: Lawrence Erlbaum. Allin, M., H. Matsumoto, A. M. Santhouse, C. Nsarti, M. H. Alasady, A. L. Stewart, L. Rifkin, and R. M. Murray, 2001. Cognitive and motor function and the size of the cerebellum in adolescents born pre-term. Brain 124:60–66. Amen, D. G., F. Paldi, and R. A. Thisted, 1993. Brain SPECT imaging. J. Am. Acad. Child Adolesc. Psychiatry 32:1080–1081. Anderson, R. A., L. H. Snyder, D. C. Bradley, and J. Xing, 1997. Multimodal representation of space in the posterior parietal

angulo-barroso and tiernan: motor systems development

157

cortex and its use in planning movements. Annu. Rev. Neurosci. 20:303–330. Angulo-Kinzler, R. M., 2001a. Motor development: From spontaneous to functional action. Motriz 7(1):39–48. Angulo-Kinzler, R. M., 2001b. Exploration and selection of intralimb coordination patterns in three-month-old infants. J. Motor Behav. 33(4):363–376. Angulo-Kinzler, R. M., P. Peirano, E. Lin, M. Garrido, and B. Lozoff, 2002a. Motor activity in human infants with irondeficiency anemia. Early Hum. Dev. 66(2):67–79. Angulo-Kinzler, R. M., P. Peirano, E. Lin, C. Algarin, M. Garrido, and B. Lozoff, 2002b. Twenty-four-hour motor activity in human infants with and without iron-deficiency anemia. Early Hum. Dev. 70:85–101. Barkley, R. A., 1998. Attention-deficit hyperactivity disorder. Sci. Am. 279:66–71. Bell, M. A., and N. A. Fox, 1992. The relations between frontal brain electrical activity and cognitive development. Child Dev. 63(5):1142–1163. Berger, A., and M. I. Posner, 2000. Pathologies of brain attentional networks. Neurosci. Biobehav. Rev. 24:3–5. Berquin, P. C., J. N. Gidd, L. K. Jacobsen, S. D. Hamburger, A. L. Krain, J. L. Rapoport, and F. X. Castellanos, 1998. Cerebellum in attention-deficit hyperactivity disorder: A morphometric MRI study. Neurology 50:1087–1093. Bertenthal, B. I., and J. J. Campos, 1990. A systems approach to the organizing effects of self-produced locomotion during infancy. In C. Rovee-Collier and L. P. Lipsitt, eds., Advances in Infancy Research, vol. 6, 1–60. Norwood, NJ: Ablex. Bertenthal, B. I., J. J. Campos, and K. C. Barrett, 1984. Selfproduced locomotion: An organizer of emotional, cognitive, and social development in infancy. In R. N. Emde and R. J. Harmon, eds., Continuities and Discontinuities in Development, 175–209. New York: Plenum. Bertenthal, B. I., and R. K. Clifton, 1998. Perception and action. In D. Kuhn and R. S. Siegler, eds., Handbook of Child Psychology. Vol. 2: Cognition, Perception, and Language, 5th ed., 51– 102. New York: Wiley. Berthier, N. E., and D. J. Robin, 1998. Midreach correction in 7-month-olds. J. Motor Behav. 30(4):290–300. Blauw-Hospers, C. H., and M. Hadders-Algra, 2005. A systematic review of the effects of early intervention on motor development. Dev. Child Neurol. Child Dev. 47:421–432. Bowlby, J., 1973. Attachment and Loss: Separation, vol. 2. New York: Basic Books. Bremner, J. G., 1978. Egocentric versus allocentric spatial coding in nine-month-old infants: Factors influencing the choice of code. Dev. Psychol. 14:346–355. Bremner, J. G., and P. E. Bryant, 1977. Place versus response as the basis for spatial errors made by young infants. J. Exp. Child Psychol. 23:162–171. Campbell, S. B., E. W. Pierce, and C. L. March, 1994. Hard-tomanage preschool boys: Symptomatic behavior across contexts and time. Child Dev. 65(3):836–851. Campos, J. J., B. I. Bertenthal, and R. Kermonian, 1992. Early experience and emotional development: The emergence of wariness of heights. Psychol. Sci. 3(1):61–64. Castellanos, F. X., J. N. Giedd, W. L. Marsh, S. D. Hamburger, A. C. Vaituzis, D. P. Dickstein, S. E. Sarfatti, Y. C. Vauss, J. W. Snell, N. Lange, D. Kaysen, A. L. Krain, G. F. Ritchie, J. C. Rajapakse, and J. L. Rapoport, 1996. Quantitative brain magnetic resonance imaging in attention deficit/hyperactivity disorder. Arch. Gen. Psychiatry 53:607–616.

158

Caston, J., N. Jones, and T. Stelz, 1995. Role of preoperative and postoperative sensorimotor training on restoration of the equilibrium behavior in adult mice following cerebellectomy. Neurobiol. Learn. Mem. 64(3):195–202. Chugani, H. T., and M. E. Phelps, 1986. Maturational changes in cerebral function in infants determined by 18FDG positron emission tomography. Science 231:840–843. Chugani, H. T., M. E. Phelps, and J. C. Mazziotta, 1987. Positron emission tomography study of human brain functional development. Ann. Neurol. 22:487–497. Craig, C. M., and D. N. Lee, 1999. Neonatal control of sucking pressure: Evidence for an intrinsic tau-guide. Exp. Brain Res. 124:371–382. Decety, J., 1996. Do imagined and executed actions share the same neural substrate? Cogn. Brain Res. 3:87–93. Diamond A., 2000. Close interrelation of motor development and cognitive development of the cerebellum and prefrontal cortex. Child Dev. 71:44–56. Edelman, G. M., 1987. Neural Darwinism: The Theory of Neuronal Group Selection. New York: Basic Books. Enderby, S. F., 1984. The effects of self-produced locomotion on the development of spatial orientation in infancy. Unpublished honors thesis, University of Denver, Denver, CO. Eyre, J. A., S. Miller, and V. Ramesh, 1991. Constancy of central conduction delays during development in man: Investigation of motor and somatosensory pathways. J. Physiol. 434: 441–452. Farroni, T., et al., 2002. Eye contact detection in humans from birth. Proc. Natl. Acad. Sci. USA 99:9602–9605. Fentress, J. C., and P. McLeod, 1986. Motor patterns in development. In E. M. Blass, ed., Handbook of Behavioral Neurobiology: Developmental Processes in Psychobiology and Neurobiology, 35–97. New York: Plenum Press. Fields, R. D., 2004. The other half of the brain. Sci. Am. 290: 54–61. Forssberg, H., 1985. Ontogeny of human locomotor control. I. Infant stepping, supported locomotion, and transition to independent locomotion. Exp. Brain Res. 57:480–493. Forssberg, H., S. Grillner, and J. Halbertsma, 1980a. The locomotion of the low spinal cat. I. Coordination within a hindlimb. Acta Physiol. Scand. 108(3):269–281. Forssberg, H., S. Grillner, J. Halbertsma, and S. Rossignol, 1980b. The locomotion of the low spinal cat. II. Interlimb coordination. Acta Physiol. Scand. 108(3):283–295. Garvey, M. A., U. Zieman, J. J. Bartko, M. B. Denckla, C. A. Barker, and E. M. Wassermann, 2003. Cortical correlates of neuromotor development in healthy children. Clin. Neurophysiol. 114(9):1662–1670. Georgopoulos, A. P., 1995. Motor cortex and cognitive processing. In M. Gazzaniga, ed., The Cognitive Neurosciences, 507–517. Cambridge, MA: MIT Press. Georgopoulos, A. P., R. E. Kettner, and A. B. Schwartz, 1988. Primate motor cortex and free arm movements to visual targets in three-dimensional space. II. Coding of the direction of movement by a neuronal population. J. Neurosci. 8:2928–2937. Gessell, A., 1945. The Embryology of Behavior. New York: Harper. Ghez, C., 1991. The cerebellum. In E. R. Kandel, J. H. Schwartz, and T. M. Jessell, eds., Principles of Neural Science, 626–646. Norwalk, CT: Appleton & Lange. Gibson, E. J., 1988. Exploratory behavior in the development of perceiving, acting and the acquiring of knowledge. Annu. Rev. Psychol. 39:1–41.

fundamentals of developmental neurobiology

Goldfield, E. C., B. A. Kay, and W. H. Warren, Jr., 1993. Infant bouncing: The assembly and tuning of action systems. Child Dev. 64:1128–1142. Gossel-Symank, R., I. K. Grimmer, and R. Siegmund, 2004. Actigraphic monitoring of the activity-rest behavior of preterm and full-term infants at 20 months of age. Chronobiol. Int. 21(4–5): 661–671. Grammer, K., M. Honda, and A. Juette, 1999. Fuzziness of nonverbal courtship communication unblurred by motion energy detection. J. Pers. Soc. Psychol. 77(3):487–508. Gramsbergen, A., 2003. Clumsiness and disturbed cerebellar development: Insights from animal experiments. Neural Plast. 10:1–2, 129–140. Grillner, S., 1975. Locomotion in vertebrates: Central mechanisms and reflex interaction. Physiol. Rev. 55:247–304. Grillner, S., 1981. Control of locomotion in bipeds, tetrapods, and fish. In V. B. Brooks, ed., Handbook of Physiology, vol. 3: Motor Control, 1179–1236. Bethesda, MD: American Physiological Society. Haith, M., 1980. Rules That Babies Look By: The Organization of Newborn Visual Activity. Hillsdale, NJ: Erlbaum. Heinen, F., F. X. Glocker, U. Fietzek, B.-U. Meyer, C. H. Lucking, and R. Korinthenberg, 1998. Absence of transcallosal inhibition following focal magnetic stimulation in preschool children. Ann. Neurol. 43:608–612. Heriza, C. B., 1991. Implications of a dynamical systems approach to understanding infant kicking behavior. Phys. Ther. 71:222–235. Hoover, J. E., and P. L. Strick, 1993. Multiple output channels in the basal ganglia. Science 259:819–821. Horobin, K., and L. Acredolo, 1986. The role of attentiveness, mobility history, and separation of hiding sites on stage IV behavior. J. Exp. Child Psychol. 41:114–127. Huttenlocher, P. R., 1990. Morphometric study of human cerebral cortex development. Neuropsychologia 28:517–527. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Huttenlocher, P. R., C. de Courten, L. J. Garey, and H. Van der Loos, 1982. Synaptogenesis in human visual cortex— Evidence for synapse elimination during normal development. Neurosci. Lett. 33:247–252. Jeannerod, M., 1997. The Cognitive Neuroscience of Action. Oxford: Blackwell Publishers. Johnson, M. H., 2001. Functional brain development in humans. Nature Rev. Neurosci. 2:475–483. Jones, T. A., J. A. Kleim, and W. T. Greenough, 1996. Synaptogenesis and dendritic growth in the cortex opposite unilateral sensorimotor cortex damage in adult rats: A quantitative electron microscopic examination. Brain Res. 733:142–148. Jouen, F., and J.-C. Lepecq, 1990. Early perceptuo-motor development: Posture and locomotion. In C.-A. Hauert, ed., Developmental Psychology: Cognitive, Perceptuo-motor and Neuropsychological Perspective, 61–85. Amsterdam: Elsevier. Kagan, J., and N. Snidman, 1991. Temperamental factors in human development. Am. Psychol. 46(8):856–862. Kates, W. R., D. C. Lanham, and H. S. Singer, 2005. Frontal white matter reductions in healthy males with complex stereotypies. Pediatr. Neurol. 32(2):109–112. Keele, S. W., and R. Ivry, 1990. Does the cerebellum provide a common computation for diverse tasks? A timing hypothesis. Ann. NY Acad. Sci. 608:179–211. Kelso, J. A. S., 1995. Dynamic Patterns: The Self-Organization of Brain and Behavior. Cambridge, MA: MIT Press.

Kermonian, R., and J. J. Campos, 1988. Locomotor experience: A facilitator of spatial cognitive development. Child Dev. 59: 908–917. Kiess, W., A. Meidert, R. A. Dressendorfer, K. Schriever, U. Kessler, A. Konig, H. P. Schwarz, and C. J. Strasburger, 1995. Salivary cortisol levels throughout childhood and adolescence: Relation with age, pubertal stage, and weight. Pediatr. Res. 37(4):502–506. Kleim, J. A., T. A. Jones, and T. Schallert, 2003. Motor enrichment and the induction of plasticity before or after brain injury. Neurochem. Res. 28:1757–1769. Kuypers, H. G., 1982. A new look at the organization of the motor system. Prog. Brain Res. 57:381–403. Lewis, M., D. F. A. Hitchcock, and M. W. Sullivan, 2004. Physiological and emotional reactivity to learning and frustration. Infancy 6(2):121–143. Lozoff, B., 1991. Long-term developmental outcome of infants with iron deficiency. N. Engl. J. Med. 325(10):687–694. Lozoff, B., and M. Black, 2004. Impact of micronutrient deficiencies on behavioral development. In J. M. Pettifor and S. Zlotkin, eds., Micronutrient Deficiencies during the Weaning Period and the First Years of Life. Nestle Nutrition Workshop Series Pediatric Program, vol. 54, 119–135. Basel: Nestec Ltd. Vevey/S. Karger AG. Mantagos, S., A. Moustogiannis, and A. G. Vagenakis, 1998. Diurnal variation of plasma cortisol levels in infancy. J. Pediatr. Endocrinol. Metab. 11(4):549–553. Martin, J. H., M. Choy, S. Pullman, and Z. Meng, 2004. Corticospinal system development depends on motor experience. J. Neurosci. 24(9):2122–2132. McGraw, M. B., 1945. The Neuromuscular Maturation of the Human Infant. New York: Columbia University Press. McKay, S., and R. M. Angulo-Barroso, 2006. Spontaneous motor activity in infants with and without Down syndrome. Infant Behav. Dev. 29(2):153–168. Middleton, F. A., and P. L. Strick, 1994. Anatomical evidence for cerebellar and basal ganglia involvement in higher cognitive function. Science 266:458–461. Middleton, F. A., and P. L. Strick, 2000. Basal ganglia and cerebellar loops: Motor and cognitive circuits. Brain Res. Rev. 31: 236–250. Montepare, J. M., S. B. Goldstein, and A. Clausen, 1987. The identification of emotions from gait information. J. Nonverb. Behav. 11(1):33–42. Morasso, P., 1981. Spatial control of arm movements. Exp. Brain Res. 42:223–227. Muir, R. B., and R. N. Lemon, 1983. Corticospinal neurons with special role in precision grip. Brain Res. 261:312–316. Muller, R. A., R. D. Rohermel, M. E. Behen, O. Muzik, T. J. Mangner, and H. T. Chugani, 1998. Developmental changes of cortical and cerebellar motor control: A clinical positron emission tomography study with children and adults. J. Child Neurol. 13(11):550–556. Passingham, R. E., 1993. The Frontal Lobes and Voluntary Action. Oxford, UK: Oxford University Press. Piaget, J., 1952. The Origins of Intelligence in Children. New York: International Universities Press. Piaget, J., 1954. The Construction of Reality in the Child. New York: Basic Books. Pouwels, P. J., K. Brockman, B. Kruse, B. Wilken, M. Wick, F. Hanefeld, and J. Frahm, 1999. Regional age dependence of human brain metabolites from infancy to adulthood as detected by quantitative localized proton MRS. Pediatr. Res. 46(4): 474–485.

angulo-barroso and tiernan: motor systems development

159

Puyau, M. R., A. L. Adolph, F. A. Vohra, and N. F. Butte, 2002. Validation and calibration of physical activity monitors in children. Obes. Res. 10:150–157. Rademaker, K. J., J. N. Lam, I. C. Van Haastert, C. S. Uiterwaal, A. F. Lieftink, F. Groenendaal, D. E. Grobbee, and L. S. de Vries, 2004. Larger corpus callosum size with better motor performance in prematurely born children. Sem. Perinatol. 28(4):279–287. Robertson, S. S., L. F. Bacher, and N. L. Huntington, 2001. The integration of body movements and attention in young infants. Psychol. Sci. 12(6):523–526. Rochat, P., and S. J. Hespos, 1997. Differential rooting responses by neonates: Evidence for an early sense of self. Early Dev. Parenting 6:105–112. Ronald, P. E., B. Larsen, N. A. Lassen, et al., 1980. Supplementary motor area and other cortical areas in organization of voluntary movements in man. J. Neurophysiol. 43:118–136. Sadeh, A., and C. Acebo, 2002. The role of actigraphy in sleep medicine. Sleep Med. Rev. 6(2):113–124. Schmidt, L. A., N. A. Fox, K. H. Rubin, E. M. Sternberg, P. W. Gold, C. C. Smith, and J. Schulkin, 1997. Behavioral and neuroendocrine responses in shy children. Dev. Psychobiol. 30(2):127–140. Schmidt, R. C., and P. Fitzpatrick, 1996. Dynamical perspective on motor learning. In H. N. Zelaznik, ed., Advances in Motor Learning and Control, 195–223. Champaign, IL: Human Kinetics. Schwartz, A. B., 1994. Neuronal substrate for volitional movement. In K. M. B. Bennett and V. Castiello, eds., Insights into the Reach to Grasp Movement, 59–63. Amsterdam: Elsevier. Shadmehr, R., and H. Holcomb, 1997. Neural correlates of motor memory consolidation. Science 277:821–824. Sirard, J. R., and R. R. Pate, 2001. Physical activity assessment in children and adolescents. Sports Med. 31(6):439–454. Smith, L. B., E. Thelen, R. Titzer, and D. McLin, 1999. Knowing in the context of acting: The task dynamics of the A-not-B error. Psychol. Rev. 106(2):235–260. Spencer, J. P., and E. Thelen, 2000. Spatially-specific changes in infants’ muscle co-activity as they learn to reach. Infancy 1:275–302. Spencer, J. P., B. Vereijken, F. J. Diedrich, and E. Thelen, 2000. Posture and the emergence of manual skills. Dev. Sci. 3:216–233. Sporns, O., and G. M. Edelman, 1993. Solving Bernstein’s problem: A proposal for the development of coordinated movement by selection. Child Dev. 64:960–981. Stiles, J., 2001. Neural plasticity and cognitive development. Dev. Neuropsychol. 18(2):237–272. Stiles, J., J. Reilly, B. Paul, and P. Moses, 2005. Cognitive development following early brain injury: Evidence for neural adaptation. Trends Cogn. Sci. 9(3):136–143. Swanson, M., P. Flodman, J. Kennedy, M. A. Spence, R. Moyzis, S. Schuck, M. Murias, J. Moriarity, C. Barr, M. Smith, and M. Posner, 2000. Dopamine genes and ADHD. Neurosci. Biobehav. Rev. 24:21–25. Swinny, J. D., J. J. van der Want, and A. Gramsbergen, 2005. Cerebellar development and plasticity: Perspectives for motor coordination strategies, for motor skills, and for therapy. Neural Plast. 12(2):153–160. Thelen, E., 1986. Treadmill-elicited stepping in seven-month-old infants. Child Dev. 57:1498–1506. Thelen E., 1995. Motor development. A new synthesis. Am. Psychol. 50(2):79–95. Thelen, E., 2000. Grounded in the world: Developmental origins of the embodied mind. Infancy 1(1):3–28.

160

Thelen, E., and E. Bates, 2003. Connectionism and dynamic systems: Are they really different? Dev. Sci. 6(4): 378–391. Thelen, E., D. Corbetta, K. Kamm, J. P. Spencer, K. Schneider, and R. F. Zernicke, 1993. The transition to reaching: Mapping intention and intrinsic dynamics. Child Dev. 64:1058–1098. Thelen, E., and D. M. Fisher, 1982. Newborn stepping: An explanation for a “disappearing” reflex. Dev. Psychol. 18:760–775. Thelen, E., and D. M. Fisher, 1983. The organization of spontaneous leg movements in newborn infants. J. Motor Behav. 15:353–377. Thelen, E., D. Fisher, and R. Ridley-Johnson, 1984. The relationship between physical growth and a newborn reflex. Infant Behav. Dev. 7:479–493. Thelen, E., D. M. Fisher, R. Ridley-Johnson, and N. Griffin, 1982. The effects of body build and arousal on newborn infant stepping. Dev. Psychobiol. 15:447–453. Thelen, E., and L. B. Smith, 1994. A Dynamic Systems Approach to the Development of Cognition and Action. Cambridge, MA: MIT Press. Thelen, E., and B. D. Ulrich, 1991. Hidden skills: A dynamical systems analysis of treadmill stepping during the first year. Monogr. Soc. Res. Child Dev. vol. 56, no. 223. Trost, S. G., K. L. McIver, and R. R. Pate, 2005. Conducting accelerometer-based activity assessments in field-based research. Med. Sci. Sports Exerc. 37(11):S531–543. Turvey, M. T., and P. Fitzpatrick, 1993. Commentary: Development of perception-action systems and general principles of pattern formation. Child Dev. 64:1175–1190. Ulrich, B. D., 1997. Dynamic systems theory and skill development in infants and children. In K. J. Connolly and H. Forssberg, eds., Neurophysiology and Neuropsychology of Motor Development, 319–345. London: Mac Keith Press. Ulrich, D. A., B. D. Ulrich, R. M. Angulo-Kinzler, and J. Yun, 2001. Treadmill training of infants with Down syndrome: Evidence-based developmental outcomes. Pediatrics 108(5): 84–90. van der Meer, A. L. H., 1997. Keeping the arm in the limelight: Advanced visual control of arm movements in neonates. Eur. J. Paediatr. Neurol. 4:103–108. van der Meer, A. L. H., et al., 1995. The functional significance of arm movements in neonates. Science 267:693–695. Vereijken, B., and E. Thelen, 1997. Training infant treadmill stepping: The role of individual pattern stability. Dev. Psychobiol. 30:89–102. von Hofsten, C., 1979. Development of visually guided reaching: The approach phase. J. Hum. Move. Stud. 5:160–178. von Hofsten, C., 1982. Eye-hand coordination in newborns. Dev. Psychol. 18:450–461. von Hofsten, C., 2004. An action perspective on motor development. Trends Cogn. Sci. 8(6):266–272. Watson, J. S., 1966. The development of and the generalization of contingent awareness in early infancy. Merrill-Palmer Q. 10:223–228. Watson, J. S., 1972. Smiling, cooing and “the game.” Merrill-Palmer Q. 18:323–339. Zaal, F. T. M. J., K. Daigle, G. L. Gottlieb, and E. Thelen, 1999. An unlearned principle for controlling natural movements. J. Neurophysiol. 82:255–259. Zelazo, P. R., 1998. McGraw and the development of unaided walking. Dev. Rev. 18:449–471. Zion, C., N. Auvray, J. Caston, A. Reber, and T. Stelz, 1990. Effects of cerebellectomy at day 15 on the ontogenesis of equilibrium behavior in the rat. Brain Res. 515:1–2, 104–110.

fundamentals of developmental neurobiology

11

Neurodevelopment of Social Cognition MELISSA D. BAUMAN AND DAVID G. AMARAL

PART 1: A BRIEF OVERVIEW OF THE NEUROBIOLOGY OF SOCIAL BEHAVIOR AND SOCIAL COGNITION Humans are so embedded in, and dependent on, their social milieu that it seems only natural that large portions of their brains would be dedicated to mediating social behavior and social cognition. It is true that large portions of the frontal, temporal, and parietal cortices, as well as subcortical structures such as the amygdala, have been associated with social function. However, in attempting to carefully deconstruct the component processes of social behavior and social cognition, as well as incorporating potentially relevant research such as on the “mirror neuron” system, the definition and independence of the “social brain” become less apparent. Even the terms “social behavior” and “social cognition” have many layers of complexity. At one level, social behavior is akin to togetherness. Bargmann and colleagues (de Bono and Bargmann, 1998; de Bono et al., 2002; de Bono and Maricq, 2005) have demonstrated that certain wild-type Caenorhabditis elegans nematode colonies demonstrate either social or solitary eating of bacteria. The choice of whether to eat in groups or to eat alone is dependent on the activity of a few nociceptive neurons that are responsive to stressful or adverse environmental conditions. The activity of these neurons and thus of feeding behavior is determined by a single polymorphism in the neuropeptide receptor npr-1. The less active allele is found in animals that feed in clusters, whereas the more active allele promotes solitary feeding. Thus, in an animal model with a nervous system consisting of only 302 neurons, substantial genetically determined differences in social behavior can be observed. In species such as the vole, with many more neurons, equally dramatic and neurobiologically based species differences in social behavior have been observed (Insel, 1997; Insel et al., 1997; Young et al., 2001). In these animals, the neuropeptides oxytocin and vasopressin play important roles in social behaviors such as pair bonding, affiliation, and paternal care of offspring. The monogamous prairie voles, for example, have a higher density of oxytocin receptors in the nucleus accumbens than the nonmonogamous montane

vole. Evidence that blockade of receptors in this important reward-related portion of the brain interrupts partner preference formation in the prairie vole indicates that this neuropeptide system plays an important role in orchestrating this social behavior. As impressive as these examples of neural control of social behaviors are, the complexity of social behavior and social cognition increases dramatically when one considers the topic in humans and higher order nonhuman primate species. Here the interest is not so much in the area of basic biological functions such as eating, sexual behavior, parenting, and affiliation but on the perception and interpretation of social signals for the accurate performance of subtle social interactions. A seminal paper in the evolution of research on the neurobiology of the social brain was published by Brothers in 1990. Brothers stated, “Primate social cognition is the processing of information which culminates in the accurate perception of the dispositions and intentions of other individuals” (Brothers, 1990, 28). In her view, dispositions are related to evaluations of the identity of other individuals, their posture, and direction of their movements, as well as their facial expression. Even more fundamentally, social cognition relies on the ability of an individual to build an internal representation of the psychological states of others in order to predict their intentions and actions. Premack and Woodruff (1978) in the course of studying cognition in chimpanzees referred to this ability of modeling the mind of another as Theory of Mind (ToM). Thus, developing a neurobiology of human social cognition would require a number of levels of analysis. In addition to the direct approach of using imaging and lesion techniques to study neurobiological correlates of human social cognition, an early step would be to decide what, if any, animal model is appropriate for the analysis of several components of human social cognition. While one would probably receive little argument that C. elegans will not contribute much to the understanding of human social cognition, there is still ample debate about the contribution of nonhuman primate studies (Gottlieb and Lickliter, 2004). Beyond identifying the appropriate species for study, one would need to determine whether certain brain regions are selectively responsible for the perception of social signals

161

such as facial expressions, body postures, and such, whether other brain regions are involved in the evaluation of these social signals to determine intentionality, and finally how appropriate social responses are generated. Since the publication of Brothers, a number of authors have addressed these issues, and candidate brain regions have been identified (Adolphs, 1999, 2003). However, it is important to note that there remains enormous ambiguity about the neural circuits that underlie social cognition. Thus several sections of this chapter deal with fundamental issues such as how one would identify brain regions particularly associated with social cognition. It is also important to point out that while there is certainly a literature related to candidate brain regions for social processing in the mature brain, there is almost no published information on the development of these brain regions and of the correlation between anatomical or physiological development and the emergence of social cognition.

demonstrate some forms of social cognition. Through a series of experiments in which a dominant and subordinate chimpanzee compete for food, Tomasello and colleagues have demonstrated that chimpanzees have knowledge of what others see and that they know something about intention in action (Tomasello et al., 2003). In spite of these abilities, it is clear that chimpanzees do not have the same ToM capabilities that are seen in young children that require an understanding of attention, perspective, and communicative intentions. Our ability to study social cognition is clearly limited by our inability to directly ask nonhuman primates (and even young children) what they understand about the psychological states of others. However, it is clear that they understand much about the behavior of conspecifics and can respond in an appropriate social fashion. Consequently, the expression of overt social behavior is the level of analysis that we can most readily relate among adult, young child, and nonhuman primate models of social processing.

What is social behavior and social cognition? Model of social processing Before we attempt to outline the neural components of the social brain, it is important to attempt to define what is meant by the terms “social behavior” and “social cognition.” Broadly defined, social behavior is the use of species-typical social signals, including body postures, vocalizations, and facial expressions, for the purpose of interaction with conspecifics (members of the same species). Social behavior relies on the ability of the individual to recognize speciestypical social signals and to have at least a rudimentary understanding of their relationships within the social network. However, there is considerable variability in the complexity and awareness of social interactions across species, and here lies the basis for the differentiation of social behavior from social cognition. Brothers states, “While many non-primates (for example, ants) can interact in highly specific ways with others of their kind, it appears that primates, especially those closely related to ourselves, have developed a unique capacity to perceive psychological facts (dispositions and intentions) about other individuals” (Brothers, 1990, 28). It is this capacity to link first- and third-person social experiences that differentiates social behavior from social cognition (Gallese et al., 2004). Elements of social cognition have been described in nonhuman primates ranging from macaque monkeys to chimpanzees. Macaque monkeys, for example, can use information about where an experimenter is looking (at, versus away from, a food item) in order to successfully “steal” the food (Flombaum and Santos, 2005). This task capitalizes on a macaque monkey’s typical evaluation of the location and attention of conspecifics to determine its chances of gaining food in a normal social situation. Chimpanzees, perhaps more than any other nonhuman primate species,

162

Given the complexity of primate social behavior, we have found it useful to generate a schematic of social information processing in order to clearly define the component processes of social interactions. Our schematic breaks down the interpretation and production of social behavior into component processes in order to more clearly define the essential and modulatory brain functions underlying social interactions (figure 11.1A and plate 18). In order to process the information conveyed in a social stimulus, such as a facial expression, the expression must first be perceived as an important source of information. The meaning of that particular expression must be evaluated, and finally an appropriate social response must be generated. There is, of course, the overarching precondition that the individual must be motivated to engage in social interactions or to interpret the gestures of others as social communication. Beyond this obligatory social impetus, the fundamental components of social processing include (1) perception of the social stimulus, (2) evaluation of its social significance, and (3) production of a species-appropriate response. The species-appropriate response could be either the production of a social response gesture or the interpretation of the disposition or intention of the other individual. The nature of that response (or whether a response will even be generated) will depend on several modulatory factors including (1) whether emotions such as fear are generated that might modulate the response, (2) the motivation to respond, and (3) whether the context is conducive to a social response. Given adequate perception, evaluation, motivation, emotion, and context, the brain must then execute an appropriate response that may take the form of thought or deed.

fundamentals of developmental neurobiology

Figure 11.1 Model of social processing. (A) Simplified diagram of the component processes essential for social behavior. An organism must first have an overarching motivation to interact with conspecifics. Given this proviso, the brain must be able to perceive a social stimulus through sensory processing. Once the social stimulus has been perceived, there must be an evaluative process in order to determine its intent. The intent of the social stimulus will have different consequences depending on the context in which it takes place. Context includes situational variables as well as the characteristics of the particular conspecific. Whether a social behavior is executed will depend not only on the development of a motor plan but also on whether modulatory influences are consistent with the

implementation of the behavior. (B) Amygdalocentric network mediating danger detection. A variety of sensory stimuli indicative of dangers are perceived by the ventral visual processing system. This information is directed to the lateral nucleus of the amygdala, where an evaluation of the potential danger of the stimulus is carried out. Internal connections within the amygdala convey information from the lateral to the basal nucleus, which, in turn, receives a prominent input from the orbitofrontal cortex. This input may provide context information that is utilized to determine whether an escape behavior is executed. It is not surprising that the amygdala, as a multipurpose danger detector, is attuned to facial stimuli such as fear expressions. (See plate 18.)

Is there a social brain?

cortices as key structures mediating declarative memory (Suzuki and Amaral, 2004). A central question for the field of social neuroscience is to determine whether the social brain is organized in a similar fashion. If it is, one would expect that the elimination of candidate social brain regions would dramatically alter normal social behavior. More than 15 years ago, Brothers proposed that the social brain was composed of several brain regions, including the amygdala, anterior cingulate cortex, orbitofrontal cortex, and temporal cortex (Brothers, 1990). However, the exact role that these structures play in social processing, or even if these structures are essential for social behavior, remains unclear to this day. Moreover, recent evidence indicates that regions of the brain not previously implicated in social

Our discussion of the neurobiology of social development is based on the assumption that regions of the brain are indeed specialized for processing social information and can therefore be identified and studied. This approach has been useful in identifying regions of the brain that are involved in other complex cognitive processes such as learning and memory. For example, it is well accepted that structures in the medial temporal lobe are essential for forming declarative memories (conscious memories for facts and events) (Squire et al., 2004). A combination of both neuroanatomical studies and lesion research has specifically identified the hippocampal formation along with the perirhinal and parahippocampal

bauman and amaral: neurodevelopment of social cognition

163

processing may indeed play an important role (e.g., the mirror neuron system of the ventral premotor and inferior parietal cortices). Before delving into the roles that these specific structures may play in social processing, it is first important to reconsider the logic that led to the identification of these regions as portions of the “social brain.” Brother’s seminal review included a summary of arguments supporting the existence of neural networks specialized for social processing. These included evidence (1) for a common evolution of social behavior, (2) that social knowledge is distinct from other knowledge, (3) that there is a well-defined developmental progression for the emergence of social behavior, and (4) that social behavior can be selectively disrupted in disorders such as autism or following damage to specific brain regions. In the following section, we will briefly summarize these main lines of evidence and incorporate new findings that lend support to the existence of a social brain. From an evolutionary perspective, the ability to interpret accurately and produce appropriate social behavior is of paramount importance for humans and other group-living primates. Sophisticated social interactions form the basis for primate societies and are necessary for forming and maintaining long-lasting relationships with other group members, acquiring resources, maintaining protection from predators (and competitors), and ultimately ensuring propagation of one’s genetic material (Cheney et al., 1986). In her review, Brothers highlights the phylogeny of facial expressions in primates as evidence of a common evolution of social processing, beginning with the following quote from Charles Darwin: “The community of certain expressions in distinct though allied species, as in the movements of the same facial muscles during laughter by man and by various monkeys, is rendered somewhat more intelligible, if we believe in their descent from a common progenitor” (Darwin, 1872; Brothers, 1990, 29). Despite the accuracy of his observations, Darwin’s work on emotional expressions, and in particular his emphasis on biological determinants of socioemotional behavior, was overlooked for many years. Nearly a century later, Ekman and colleagues provided evidence that a set of emotional expressions are universal to the human species (Ekman and Friesen, 1971; Ekman, 1993). The commonality of emotional expression across cultures suggests that these behaviors are innate components of our behavioral repertoire. Moreover, it appears that closely related species of nonhuman primates display similarities in facial muscular structure and produce facial expressions believed to be homologous to several human expressions (Parr et al., 2005). Brothers suggests that as this capacity to produce social signals evolved, it is likely that neural networks evolved to facilitate the accurate processing of this information. What remains unknown is the precise relationship between pressures from the social domain and

164

corresponding changes in neural circuitry (Dunbar, 1998; Reader and Laland, 2002; Barrett et al., 2003). Another line of evidence proposed by Brothers is that social knowledge is operationally distinct from other domains of knowledge. She suggests that certain cognitive functions, such as making transitive inferences, operate most strongly in the context of social processing. In transitive inference, overlapping pairs of items are trained (e.g., A+B−, B+C−, C+D−, and D+E−, where + and − indicate correct and incorrect choices). During training, terms B and D are correct and incorrect equally often. Later, participants who choose B over D when presented with novel pair BD are said to demonstrate transitive inference. Although transitive inference has been demonstrated in several primate species in the laboratory using inanimate objects as the “terms,” it requires an enormous amount of training to reach criteria (McGonigle and Chalmers, 1977). In contrast, many primate species live in well-defined dominance hierarchies and are able to quickly and accurately determine the status of other group members in relation to one another (Cheney and Seyfarth, 1990). In a very real sense, they are using conspecifics as the “terms” of a transitive inference paradigm and are clearly much more adept at performing these complex logical operations within a naturalistic social context. This difference suggests that at least some aspects of primate intelligence evolved specifically to solve the challenges of interacting with conspecifics (Cheney et al., 1986). The increasing complexity of social groups may have forced the evolution of ever more sophisticated social cognitive processing. It is reasonable to presume that producing and interpreting social gestures was a first step to subtly applying this expertise for manipulation and deception of conspecifics. This reasoning raises the question of whether the same or different brain region(s) subserve the affiliative lip smack in macaque monkeys and the poker face in humans. Brothers also argued that the distinctive and characteristic trajectory that characterizes primate social development is also evidence for the existence of a social brain. Newborn human infants, for example, display visual preferences for facelike stimuli (Goren et al., 1975; Johnson et al., 1991; Morton and Johnson, 1991) and are capable of imitating adult facial gestures immediately after birth (Meltzoff and Moore, 1977, 1983). Interestingly, social smiling is present even in blind infants, adding further support to the notion that there is an innate mechanism for mediating early social predispositions (Freedman, 1964). It has recently been demonstrated that infant monkeys display a transient ability to imitate facial gestures in the first week after birth that is similar to the transient abilities reported in human infants (Ferrari et al., 2006). These innate social predispositions suggest that components of the human and nonhuman primate genetic endowment generate brain regions that are specialized to mediate complex social interactions.

fundamentals of developmental neurobiology

Maestripieri and colleagues have provided compelling evidence to support this view. For example, infant monkeys removed from their biological mothers and reared by foster mothers develop a pattern of social behavior (e.g., rates of social contact and aggression) that resembles their biological mothers rather than their foster mothers (Maestripieri, 2003). This study highlights a mode of direct genetic inheritance of complex behavioral traits, lending further support to the notion of biological predispositions for social behavior. Brothers further argues that selective changes in social behavior as a result of a developmental disorder, such as autism, support the existence of a social brain. Autism is a neurodevelopmental disorder characterized by deficits in social interactions, impaired communication, and restricted patterns of behaviors, interests, and activities. Of these three domains, the alteration in social behavior is often considered the “hallmark” feature. Indeed, DSM-IV criteria for autism diagnosis must include a qualitative impairment in social interaction, including two of the following characteristics: (1) marked impairment in the use of nonverbal behaviors that regulate social interactions, (2) failure to develop peer relationships, (3) a lack of spontaneous seeking to share enjoyments, interests or achievements, and (4) lack of social or emotional reciprocity (APA, 1994). Brothers (1990) suggests that the “inborn selective absence of social cognition” that characterizes autism provides evidence that social processing can selectively be disrupted and therefore supports the existence of a specialized neural substrate of social behavior. Unfortunately, the neuropathology of autism is still quite uncertain. But, information concerning which brain regions are impaired in autism may ultimately provide suggestive evidence concerning what regions subserve normal social cognition. Recent research on other neurodevelopmental disorders that alter social behavior may provide further insight into the social brain. For example, Williams syndrome is a rare genetic disorder caused by a hemizygous deletion in chromosome band 7q11.23. In contrast to the diminished social interest that is characteristic of patients with autism, patients with Williams syndrome display hypersociability (Bellugi et al., 1999). These changes in social behavior resulting from a genetic alteration provide additional evidence of a genetic/ biological basis of social development (Doyle et al., 2004) and support of the notion that there is indeed a social brain. After establishing the plausibility of a social brain, Brothers concludes her paper with a more direct line of evidence—the existence of socially responsive cells in the macaque monkey cortex. Brothers proposes that neurons responding to social stimuli have evolved to enable the interpretation of information about other individuals. Indeed, cells that respond to faces, body movements, gaze, and the like are found throughout the adult macaque temporal lobe,

concentrated in both the inferior temporal gyrus and on the banks of the superior temporal sulcus (Gross et al., 1972; Desimone et al., 1984; Baylis et al., 1987; Hasselmo et al., 1989; Perrett et al., 1992) and in the amygdala (Rolls, 1984; Leonard et al., 1985; Brothers et al., 1990; Brothers and Ring, 1993). In the past two decades, additional regions of the brain that may contribute to social processing have been identified, including the mirror neuron system and a region of the human fusiform gyrus known as the fusiform face area (FFA). In the following section we will briefly summarize neurobiological and functional evidence implicating certain brain regions in social function.

What are the putative structures of the social brain? Amygdala The amygdala has long been implicated in the mediation of emotional and social behavior. Even early lesion studies in primates (Rosvold et al., 1954) suggested that the amygdala might be essential for normal social behavior in macaque monkeys. However, our own work has questioned the notion that the amygdala is an essential component of the social brain. We view it as a modulatory influence on social and other behaviors. Here is the background evidence: Neuroanatomy. The amygdala is a cytoarchitectonically complex structure located in the anterior temporal lobe (see figure 11.2). The neuroanatomical connections of the amygdala give it a unique position to combine highly processed sensory information from all modalities with contextual information needed to evaluate stimuli and to elicit appropriate behavioral responses (Emery and Amaral, 2000). The amygdala receives high-level information from all sensory modalities, which is further processed through the complex intrinsic connections of the 13 amygdaloid nuclei (Amaral et al., 1992). The amygdala receives projections from a variety of cortical areas, including medial and orbitofrontal regions of the frontal lobe, anterior portions of inferotemporal cortex, superior temporal cortex, perirhinal cortex, and anterior cingulate cortex. In turn, the amygdala projects to numerous brain areas capable of influencing behavioral output. The amygdala can influence relatively early stages in cortical sensory processing (Freese and Amaral, 2005), and extensive projections from the central nucleus of the amygdala innervate many autonomic and visceral control regions of the brain stem. There are additional projections to the striatum and the hippocampal formation. Taken together, the amygdala is capable of interacting with widespread brain regions to orchestrate an appropriate behavioral response to a provocative stimulus. But, what types of stimuli does the amygdala respond to? Functional studies. The amygdala has been implicated in a wide range of behavioral and psychological processes,

bauman and amaral: neurodevelopment of social cognition

165

Figure 11.2 Schematic overview of neural regions implicated in social processing. (A) Lateral view: Anterior and posterior components of the human mirror neuron system (MNS) are shown. An anterior area with mirror neuron properties is located in the inferior frontal gyrus and adjacent ventral premotor cortex. A posterior area with mirror neuron properties is located in the rostral portion of the inferior parietal lobule. Cortex surrounding the superior temporal sulcus (STS) is also highlighted. (B) Ventral view: The fusiform gyrus forms the posterior portion of the occipitotemporal

gyrus, bounded by the collateral sulcus medially and the lateral occipitotemporal sulcus laterally. A region of the fusiform gyrus, selective for face stimuli (the fusiform face area, FFA) is generally located in the middle lateral fusiform gyrus. The orbitofrontal cortex (OFC) lies on the ventral surface of the frontal lobe just above the eye orbits. (C) Medial view: Anterior cingulate cortex (ACC) is the frontal part of the cingulate cortex and includes Brodmann’s areas 24 (ventral ACC) and 32 (dorsal ACC). The amygdala is located in the anterior portion of the temporal lobe.

including fear processing (Davis, 1992; LeDoux, 1998, 2000; Davis and Whalen, 2001), reward association (Malkova et al., 1997; Baxter and Murray, 2002; Gottfried et al., 2003), memory modulation (Cahill and McGaugh, 1998; Canli et al., 2000), and social behavior (Brothers, 1990; Kling, 1992; Adolphs, 1999; Bachevalier, 2000; Emery and Amaral, 2000). The specific contributions of the amygdala to some of these processes (e.g., fear processing) have been well characterized. However, the precise role of the amygdala in other behaviors, such as reward evaluation and social behavior, remains controversial.

Electrophysiological recordings from the nonhuman primate amygdala have indicated that it is responsive to many socially relevant stimuli, including faces, specific facial expressions and direct eye contact (Rolls, 1984; Leonard et al., 1985; Brothers et al., 1990; Brothers and Ring, 1993). In nonhuman primates, neural responses to appeasing faces are often marked by significant decreases of firing rates, whereas responses to threatening faces are associated with increased firing rate, suggesting that global activation in the amygdala might be larger to threatening faces (Gothard et al., 2007).

166

fundamentals of developmental neurobiology

Noninvasive functional neuroimaging experiments have indicated that the human amygdala is preferentially activated by a variety of meaningful social signals. Facial expressions are one of the most replicable elicitors of human amygdala activation, particularly facial expressions depicting fear (Morris et al., 1996; Morris, Friston, et al., 1998). Other facial expressions depicting negative emotions, such as disgust, anger, and sadness (Morris, Ohman, et al., 1998; Blair et al., 1999), as well as several positive emotions (Breiter et al., 1996; Canli et al., 2002), have also resulted in amygdala activation, but not as consistently as fearful expressions. The amygdala is also activated when subjects make complex social judgments, such as determining whether an individual is trustworthy (Winston et al., 2002), discerning what someone may be thinking (Baron-Cohen et al., 1999), or evaluating social content during a social attribution task that involves perception of humanlike interactions among simple geometric shapes (R. Schultz et al., 2003). Amygdala activation has also been reported during nonsocial testing paradigms, including fear conditioning (Buchel et al., 1998; LaBar et al., 1998; Cheng et al., 2003), viewing pictures of phobia-related stimuli (Dilger et al., 2003), anticipation of aversive stimuli (e.g., shock) (Phelps et al., 2001), and viewing threatening and fearful nonsocial stimuli (Hariri et al., 2003). One unifying hypothesis is that the amygdala, at least in part, functions as a danger detector (Amaral, 2003). It uses incoming sensory information to evaluate the environment for potential threats (figure 11.1B). In this role as a danger detector, the amygdala would evaluate a vast array of potentially fear-inducing stimuli, ranging from threats that are common to most vertebrates (e.g., snakes or fire) to speciesspecific displays of social aggression (e.g., open-mouthed threat facial expressions of macaque monkeys). Since indications of danger are conveyed socially through facial expressions and other biological gestures, it would not be surprising that the amygdala is particularly attuned to these “social” danger signals. One would not expect, however, that the evaluations carried out by the amygdala deal with the general category of others’ intentions or dispositions but only those that might present a danger. Thus ready activation of the amygdala by fearful faces (which convey the fact that there is a proximal danger) or when evaluating whether another individual is trustworthy (i.e., not likely to cause harm) would be consistent with its acting as a danger detector. It is clear from lesion studies in humans that bilateral lesions of the amygdala do not remove social function, nor do they gravely impair an individual’s ability to judge the intentions or dispositions of others. Lesion studies. Human patients with bilateral amygdala damage are very rare, yet they provide invaluable information regarding the function of the amygdala. One of the

most extensively studied patients with bilateral amygdala damage, patient S.M., developed her lesion during adolescence from Urbach-Wiethe disease—a rare syndrome associated with selective bilateral amygdala calcification and atrophy. Although patient S.M.’s social behavior remains remarkably intact (i.e., she has a high school education, lives independently, is married, holds a job, and is raising a family), she does have areas of impaired function. In general, patient S.M. shows the most consistent deficits related to fear processing. For example, patient S.M. was not able to recognize fearful expressions (Adolphs et al., 1994) and was impaired in judging how much to trust another person after viewing the person’s face (Adolphs et al., 1998). Patient S.M.’s deficits in identifying fearful expressions may be explained by an inability to make normal use of information from the eye region of the face, which is critical for identifying fear (Adolphs et al., 2005). Though her deficits are primarily related to fear processing, it is important to note that patient S.M. has shown abnormalities in her ability to assess social stimuli not related to fear, such as the social attribution task described earlier (Heberlein and Adolphs, 2004). In this case, patient S.M. fails to normally attribute social intentions to videos of inanimate objects. While the reason for this failure is unknown, it does seem to be clear that patient S.M. does not demonstrate autistic symptomatology, can interpret and produce the majority of social signals, and is at least fairly astute at engaging in social cognition. Patient H.M. is another famous example of an individual who has a complete bilateral loss of his amygdala and hippocampal formation (Corkin et al., 1997). Yet, despite his dramatic and near complete loss of the ability to form new episodic memories, his daily social interactions have been generally quite normal. The collective results from previous nonhuman primate studies indicated that bilateral amygdala lesions disrupt species-typical social behavior, generally resulting in decreased affiliative behavior and subsequent social isolation when tested in socially complex environments (Kling, 1992). However, recent studies utilizing more selective lesioning techniques combined with quantitative behavioral assessments have found that amygdala damage in adult monkeys does not preclude social interactions (Emery et al., 2001). In these animals, behavioral changes appeared more closely related to deficits in fear processing rather than specific impairments in social behavior, suggesting that the amygdala plays a modulatory rather than an essential role in social behavior. One caveat to this interpretation is that these experiments were conducted in mature animals, thus leaving open the possibility that the amygdala may play a role in acquiring social behavior at earlier ages. We discuss the results of neonatal lesions of the amygdala in part 2 of this chapter.

bauman and amaral: neurodevelopment of social cognition

167

Summary. The amygdala has been proposed to play an essential role in the production of normal social behavior. While there are intriguing neuroimaging and behavioral data indicating that the human amygdala is responsive to a variety of social signals, many of the evaluations carried out by the amygdala are driven by signals that may convey potential danger. Moreover, amygdala lesion studies both in rhesus monkeys and human subjects indicate that the amygdala is not essential for the expression of species-typical social interaction. These observations compel us to conclude that if the amygdala is a component of the social brain, it must be a nonessential component. Orbitofrontal Cortex Whereas the role of the amygdala in social function remains equivocal, there is a substantial and converging body of evidence that the orbitofrontal cortex is a central component of the social brain. Neuroanatomy. The frontal cortex of humans and apes occupies a larger percentage of total cortical volume as compared to smaller primates (gibbons and monkeys) (Semendeferi et al., 2002). The orbitofrontal cortex (OFC) lies on the ventral surface of the frontal lobe just above the eye orbits. This brain area has strong connections to the amygdala, cingulate cortex, and somatosensory areas (Ongur and Price, 2000). It is generally agreed that the cytoarchitecture of the orbital cortex is similar in humans and nonhuman primates (Petrides and Pandya, 1994). Functional studies. Traditionally the OFC has been implicated in processing a variety of sensory information relating to reward and punishment (Rolls, 1996). It is thought that the OFC plays an important role in processing socially relevant information by regulating responses to positive and negative reinforcers, thereby influencing a wide range of socioemotional behaviors (Kringelbach and Rolls, 2004; Rolls, 2004). Indeed, data from functional imaging studies indicate that the OFC is activated in response to complex social decisions, such as rating the attractiveness of a face (O’Doherty et al., 2003), using facial expressions to guide behavior in a reversal learning paradigm (Kringelbach and Rolls, 2003), choosing to cooperate with another person (Rilling et al., 2002), making social judgments of emotionally evocative or morally conflicted statements (Moll et al., 2002), and determining whether social norms have been violated (Berthoz et al., 2002). Lesion studies. Data from human lesion studies further support the role of the OFC in social processing. Human patients with bilateral lesions of the OFC show deficits in the identification of facial and vocal emotional expression (Hornak et al., 1996, 2003) and have abnormalities in regulating social behavior (Rolls et al., 1994; Angrilli et al., 1999; Blair and Cipolotti, 2000) and impulsivity (Berlin et al.,

168

2004, 2005), despite performing well on conventional intelligence tests. A case study of a patient sustaining damage to the OFC reported profound changes in personality that resulted in the patient engaging in risky behavior, alienating friends and family, and acting out in many socially inappropriate ways (Eslinger and Damasio, 1985). More recently, Mah and Grafman (Mah et al., 2004) demonstrated that subjects with lesions of the orbitofrontal cortex were unable to make accurate judgments of videotaped interpersonal interactions on the basis of nonverbal information. These data indicate that damage to the OFC produces deficits in self-regulation and an inability to use affective and social information to guide social behavior and social decisions. Converging evidence from nonhuman primates with experimentally induced damage to the OFC support the role of the OFC in complex social interactions. Lesions of the OFC in macaque monkeys are associated with decreased aggressive behaviors and increased avoidance behaviors (Butter et al., 1970), as well as decreased social abilities in vervet monkeys (Raleigh et al., 1979). Social dominance, which serves as an important indicator of social competency in macaque monkeys, is also affected by OFC damage. Previously high ranking macaque monkeys that received damage to the OFC were unable to maintain their high social rank following the surgery (Butter and Snyder, 1972). A more recent study utilizing selective OFC lesions found that damage to the OFC was associated with decreased affiliative personality traits, increased avoidant personality traits, and atypical responses to specific social signals (Machado and Bachevalier, 2006). Summary. The orbitofrontal cortex appears to use socially relevant information combined with indicators of positive and negative consequences to define a social course of action that supports both self-interest and social compatibility. Social judgment and planning appear to be impaired following damage to the OFC. Yet it is not clear whether this impairment leads to deficits in judging the dispositions and intentions of others. Deficits in these types of “theory of mind” functions have often been attributed to damage in other brain regions discussed later in this chapter. Anterior Cingulate Cortex Neuroanatomy. The cingulate cortex is recognized as a complex collection of cortical subregions that subserve a variety of cognitive, emotional, and motor functions (Vogt et al., 1992; Bush et al., 2002). The anterior cingulate cortex (ACC) forms a large region surrounding the rostrum of the corpus callosum and can be distinguished from the posterior cingulate cortex based on cytoarchitecture (the ACC lacks

fundamentals of developmental neurobiology

layer IV) and connectivity patterns (Vogt et al., 2005). The human ACC has been divided into “affective” and “cognitive” domains, with the more ventral affective division sharing extensive connections with other regions capable of influencing social behavior, including the amygdala and autonomic brain stem nuclei (Devinsky et al., 1995; Bush et al., 2000). Interestingly, a morphologically unique class of large, bipolar cells has been found in layer V of anterior cingulate cortex in humans and apes (Vogt et al., 1995; Nimchinsky et al., 1999; Hof et al., 2001; Allman et al., 2005). The concentration of these spindle-shaped neurons (also known as Von Economo neurons) is greatest in humans, followed by chimpanzees, gorillas, and then orangutans (Nimchinsky et al., 1999). Though spindle cells have not been found in any other primate species, recent evidence indicates that several species of cetaceans do have spindle cells (Hof and Van Der Gucht, 2007). The possible parallel evolution of spindle cells in highly social species of hominids and cetaceans characterized by a very large brain and a large body size raises intriguing questions regarding the function of these neurons (Allman et al., 2005; Watson et al., 2006). Functional studies. The ACC has been linked to a number of cognitive processes, including affective and social behavior. For example, in nonhuman primates the ACC plays a critical role in the voluntary initiation and suppression of species-typical vocalizations ( Jurgens, 2002). In humans, activity within the ventral ACC is closely linked with aspects of socioemotional processing (Bush et al., 2000) such as empathy ( Jackson et al., 2005), deception (Spence et al., 2004), and receiving social feedback (Somerville et al., 2006). Lesion studies. Though there are a number of reports in the literature of patients with damage to the ACC that demonstrate deficits in social behavior, these cases have been complicated by the fact that the patients also sustained damage to the orbitofrontal cortex (Bechara et al., 1998). Lesion studies of the anterior cingulate cortex in nonhuman primates have produced inconsistent findings. Historical studies have reported that lesions of the ACC produce profound changes in monkey social behavior (Ward, 1948; Glees et al., 1950), while other studies claim that social behavior was essentially unchanged following damage to the ACC (Pribram and Fulton, 1954; Mirsky et al., 1957). A more recent study reported that pairs of cynomolgus macaques with cingulate lesions interacted less with one another and produced fewer vocalizations compared to control pairs (Hadland et al., 2003). Rudebeck and colleagues have compared lesions of the orbitofrontal cortex with lesions of the anterior cingulate cortex (Rudebeck et al., 2006). They found that animals with the anterior cingulate damage demonstrated alterations in social inter-

est, whereas those with orbitofrontal lesions had fear-related impairments not unlike those seen in animals with amygdala damage. Summary. While much additional work is necessary to evaluate the precise role of the anterior cingulate cortex in social behavior, existing literature certainly supports the consideration of this region as a component of the social brain. Temporal Cortex Single-unit electrophysiological studies in the primate as well as fMRI studies in the human have indicated that regions of temporal lobe are involved in social perception (Allison et al., 2000; Puce and Perrett, 2003). We will discuss these areas more generally in this section but then deal with the fusiform face area, in particular, in the following section. These regions are challenging as one attempts to determine whether they should be included as part of high-level sensory processing apparatuses or as part of the “social brain.” Neuroanatomy. Inferior temporal (IT) cortex is a region of visual association cortex that, in the monkey, occupies the inferior temporal gyrus and adjacent portions of the superior temporal sulcus (STS). Converging evidence from humans and nonhuman primates suggests that these regions play a role in social perception, though direct comparisons are difficult because homologies between human and nonhuman primate temporal cortex are not completely characterized. Functional studies. Cells responsive to social stimuli are found throughout the macaque temporal lobe, concentrated in both the inferior temporal gyrus and on the banks of the STS. Cells in the macaque STS respond to specific body movements (Perrett et al., 1985a), and some responses appear related to the object or goal of the movements (e.g., reaching for or walking toward a specific place) (Perrett et al., 1989). Much attention has been devoted to cells in the temporal cortex that respond preferentially to faces (Perrett et al., 1982; Desimone et al., 1984; Rolls, 1984) and are sensitive to biologically important characteristics such as face identity or expression (Perrett et al., 1984; Hasselmo et al., 1989), as well as direction of gaze (Perrett, Smith et al., 1985b). A highly face-selective region along the upper bank of the STS was recently identified using fMRI (Tsao et al., 2003) and subsequently confirmed to be almost entirely face selective by using fMRI-guided single-unit recordings of neurons (Tsao et al., 2006). This selective face region in the macaque may be homologous to the human FFA (to be discussed later), though additional research is needed to clarify this relationship. The causal relationship between the activity of face-selective neurons in temporal cortex and face perception has recently been demonstrated by artificially activating small clusters of IT neurons while

bauman and amaral: neurodevelopment of social cognition

169

monkeys judged whether visual images belonged to “face” or “nonface” categories (Afraz et al., 2006). Microstimulation of face-selective sites within IT cortex, but not other sites, strongly biased the monkeys’ decisions toward the face category. Neuroimaging studies in humans further support the role of the temporal cortex in processing socially relevant information. An array of biologically/socially relevant stimuli activate the STS, including body movement/biological motion (Bonda et al., 1996; Grossman et al., 2000), mouth movements in faces (Puce et al., 1998), and gaze direction (Hoffman and Haxby, 2000). Interestingly, it appears that the STS is involved not only in the low-level processing of gaze, but also in the processing of more sophisticated social information. Stronger STS activity is elicited by gaze shifts that violate the viewer’s expectations (Pelphrey et al., 2003) or by gaze shifts that simulate eye contact, as opposed to gaze aversion, with a stranger (Pelphrey et al., 2004). Indeed, other studies have suggested that the STS is concerned with more than basic perception of biologically important stimuli. For example, the posterior STS shows a greater response to animations of moving geometric shapes that demonstrate social interactions or complex goal-directed movements as opposed to animations depicting random motion (Castelli et al., 2000; J. Schultz et al., 2004). Moreover, a region in the temporoparietal junction (in the region of the posterior STS) is activated when subjects are required to perceive intentions and/or attribute mental states to others (Gallagher et al., 2000; Saxe and Kanwisher, 2003; Saxe et al., 2004). These studies are consistent with the proposal that the STS plays a role in the perception of biological motion as well as a possible role in making inferences about the mental states of others (Allison et al., 2000). Lesion studies. A recent case study indicates that a patient with a circumscribed lesion to the right superior temporal gyrus shows a diminished capacity to maintain eye contact during conversations (Akiyama et al., 2006a). Interestingly, this same patient was impaired in her ability to utilize biological directional information such as gaze, but showed no impairments for using nonbiological counterparts (arrows) (Akiyama et al., 2006b). Though nonhuman primates with bilateral lesions of the STS are also impaired in their ability to perceive eye gaze and to differentiate the angle of faces (Campbell et al., 1990; Heywood and Cowey, 1992), it is not clear if these deficits are the result of a more general impairment in visual discrimination learning (Eacott et al., 1993). Summary. Collectively, the studies from humans and nonhuman primates indicate that cells in the temporal cortex play a clear role in the perception of socially relevant stimuli. These initial perceptual stages are essential components of

170

the cascade of information processing that culminates in the accurate analysis of the dispositions and intentions of other individuals. However, it is not entirely clear at what point basic perceptual processing transitions to more sophisticated representations of socially relevant concepts (i.e., at what point does the sensory brain end and the social brain begin?). Converging evidence from humans and nonhuman primates indicates that cells in the temporal cortex may be involved with more sophisticated aspects of social processing, including evaluations of goal directedness, predictability, and intention of action. These findings lend support to the proposal that some cells in the STS are essential building blocks in the capacity to understand goals and intentions of an individual (Jellema and Perrett, 2005). Fusiform Gyrus Research related to the fusiform gyrus has forced investigators to question whether a brain region dedicated to responding to a possible social stimulus, such as the face of a conspecific, provides evidence for a differentiated component of the “social brain” or an example of a highly refined cortex specialized for perception of classes of complex objects. Neuroanatomy. The fusiform gyrus forms the posterior portion of the occipitotemporal gyrus, bounded by the collateral sulcus medially and the lateral occipitotemporal sulcus laterally. A region of the fusiform gyrus selective for face stimuli (the fusiform face area; FFA), is generally located in the middle lateral fusiform gyrus (Kanwisher et al., 1997). Functional studies. In recent years, the FFA has been at the heart of the debate over specialized modules of processing versus a more distributed processing approach. The “domain-specific” side of the debate argues that the FFA is specialized for processing faces (Kanwisher, 2000). This interpretation is based on the findings that FFA activation for faces is twice as strong as for nonface stimuli, such as letter strings, objects, and animals (Puce et al., 1996; Kanwisher et al., 1997, 1999) and that FFA activation is highly correlated with detection and identification of faces, but not of objects (Grill-Spector et al., 2004). In contrast, the “domain-general” view of FFA function suggests that the FFA is part of a distributed object recognition system (Bukach et al., 2006). This view is supported by studies indicating that the FFA shows a statistically significant response to nonface objects driven by expertise (Gauthier et al., 1999, 2000; Haxby et al., 2001). According to this model, faces activate the fusiform area because humans are “experts” at identifying faces, not because faces constitute a special class of information. Though the debate over FFA function is ongoing (Gauthier and Bukach, 2007; McKone and Robbins, 2007), new research methods are providing insight into the finescale functional organization of the FFA and may prove

fundamentals of developmental neurobiology

useful in addressing the larger issue of neural specialization. For example, results from a recent high-resolution fMRI study indicate that the FFA is actually composed of faceselective cortical patches that are intermixed with cortical patches selective for processing other objects (Grill-Spector et al., 2006). These findings challenge the hypothesis that the FFA is a uniform brain region composed of face-selective neurons. Lesion studies. Lesions of the fusiform gyrus result in deficits in face processing (Farah, 1996; Farah et al., 1998). Moreover, bilateral lesions of the ventral occipitotemporal cortex, involving the fusiform gyri, are associated with prosopagnosia, a deficit characterized by the inability to recognize faces of known individuals (Sergent and Signoret, 1992; Wada and Yamamoto, 2001). Summary. Given the importance of faces in conveying social information, it is likely that the FFA contributes to the early stages of social perception. Indeed, Haxby and colleagues have proposed a model in which the FFA is part of a core system for the visual perception of faces, particularly invariant aspects of face perception such as identity (Haxby et al., 2000, 2002). Since the neuroanatomical connectivity of the FFA is not known, it is currently unclear which other brain regions might build on this face processing in order to create models of the thoughts and intentions of others. Mirror Neuron System Neuroanatomy. Mirror neurons are a recently discovered class of visuomotor neurons identified using single-unit recordings in macaque monkeys (for reviews see Rizzolatti et al., 2001; Rizzolatti and Craighero, 2004; Iacoboni and Dapretto, 2006). Mirror neurons in the ventral premotor cortex and inferior parietal cortex of nonhuman primates fire with both the execution and observation of an action (Gallese et al., 1996; Rizzolatti et al., 1996) and have received substantial interest because of their proposed roles in understanding the actions of others (Rizzolatti and Craighero, 2004) and imitation (Rizzolatti et al., 2001; Iacoboni, 2005). Indeed, functional imaging studies in human subjects provide indirect evidence that humans also possess a mirror neuron system located in the pars opercularis of the inferior frontal gyrus and another region in the rostral posterior parietal cortex (Iacoboni et al., 1999). Functional studies. A defining feature of mirror neurons is the ability to fire both when a monkey does a particular action and when it passively observes an individual performing a similar action (Gallese et al., 1996; Rizzolatti et al., 1996). The majority of nonhuman primate mirror neurons require an interaction between a biological effector (e.g., a hand or

mouth) and an object, as the presentation of either stimulus alone is not sufficient to evoke activity, though a small subset of mirror neurons appear to respond to communicative gestures (Ferrari et al., 2003). It has been speculated that the motor-neuron-mediated type of action understanding in the macaque mirror system may have evolved into a neural system that encodes cognitive understanding of others (Iacoboni and Dapretto, 2006). In support of this hypothesis, the putative mirror neuron system in humans is activated in response to a wide range of cognitive functions, including imitation, empathy, and theory of mind (Carr et al., 2003; Gallese et al., 2004; Iacoboni et al., 2004, 2005). Moreover, new evidence suggests that the mirror neuron system may be dysfunctional in individuals with autism (Oberman et al., 2005; Dapretto et al., 2006). Lesion studies. The mirror neuron system in nonhuman primates includes large portions of the parietal and premotor cortices. The extensive lesions of these cortices that would be needed to disrupt mirror neuron function would most likely lead to other cognitive deficits. Although the lesion technique has not been used to evaluate the function of the mirror neuron system in nonhuman primates, reports of deficits in emotion recognition tasks in a patient with damage to the left frontal operculum lends support to the possibility that the human mirror neuron system contributes to some aspects of social processing (Adolphs et al., 2002). Summary. The discovery of the mirror neuron system provides a plausible neurophysiological mechanism for the development of imitation and other important social behaviors. Though much additional work is necessary to evaluate how the mirror neuron system may interact with other regions of the brain involved with social cognition (Keysers and Perrett, 2004), the current literature certainly supports the consideration of this region as a component of the social brain. Other Brain Regions It has been suggested that several other regions of the brain play a role in particular aspects of social processing, though the evidence implicating these regions as part of the social brain is less substantial than for the regions that we have described in detail. For example, Adolphs and colleagues have demonstrated that recognizing emotions from visually presented facial expressions requires right somatosensory-related cortices (Adolphs et al., 2000). These findings support the suggestions that we may recognize another individual’s emotional state by internally generating somatosensory representations that simulate how the other individual would feel when displaying a certain facial expression. Likewise, recent evidence suggests that the insula is activated both during the experience of disgust and the observation of the facial expression of disgust in others

bauman and amaral: neurodevelopment of social cognition

171

(Wicker et al., 2003). These data lend support to the idea that there is a common mechanism for understanding an emotion in others and feeling the same emotion in ourselves. Finally, given our interest in the neural basis of social cognition, it is important to include preliminary neuroimaging studies that have begun to identify neural networks involved with high-order cognitive processes, such as ToM. Though lesion studies have not yet provided compelling evidence of any one region that is essential for ToM abilities (Apperly et al., 2005; Griffin et al., 2006), several regions have been activated in ToM tasks, including the amygdala, OFC and posterior cingulate cortex, medial prefrontal regions (anterior paracingulate cortex), the temporoparietal junction found within the superior temporal sulcus, and the temporal poles. Of these regions, three areas appear most consistently activated across a wide range of ToM paradigms: (1) medial prefrontal regions, (2) the temporoparietal junction, and (3) the temporal poles (Gallagher and Frith, 2003). Several different ToM paradigms have consistently reported activity in these regions, including viewing animated shapes moving with “intentions” (Castelli et al., 2000), reading stories that require mental-state attribution (Fletcher et al., 1995; Vogeley et al., 2001; Saxe and Kanwisher, 2003), and performing tasks requiring competition or cooperation with a human partner as opposed to a computer (McCabe et al., 2001; Gallagher et al., 2002). Recent studies indicate that specific regions may contribute to different components of ToM, though additional research is needed to further explore these relationships (Saxe, 2006; Singer, 2006).

PART 2: DEVELOPMENT OF THE SOCIAL BRAIN Given the continuing ambiguities in the definition of the social brain, it is all the more challenging for social neuroscience to evaluate the neurodevelopment of social cognition. Critical questions for this field of research include the following: How and when do neural networks become specialized for processing social information? Is the specialization innate or determined by experience? How can early social experiences alter the development of the social brain? These questions remain open-ended, as we are at the very early stages of understanding how the brain develops the capacity to process social information. In part 1 of this chapter, we outlined regions of the brain most commonly thought to play a role in adult social behavior, including the amygdala, anterior cingulate cortex, orbital frontal cortex, regions of the temporal cortex, and the mirror neuron system. Much of the evidence linking these particular brain regions with social processing has been obtained from functional-imaging research or lesion studies on adult

172

human subjects or from animal models using mature subjects (Adolphs, 1999, 2003). In many cases, there is little if any evidence evaluating the contributions of these structures to the development of social behavior early in life. It is problematic to rely on information obtained solely from mature subjects because it is becoming increasingly evident that the developing brain is not simply an immature version of the adult brain. Indeed, converging evidence from humans and animal models has clearly shown that the brain undergoes profound changes in structural and functional maturation that extend into early adulthood. Using the amygdala as one example, research has demonstrated that young children show unique patterns of amygdala growth (Giedd et al., 1996; Schumann et al., 2004) and that their amygdala activation in response to social stimuli is different from the activation seen in adults (Thomas et al., 2001). Moreover, animal models have indicated that the amygdala may play a different functional role in the early postnatal period compared to adulthood (Amaral et al., 2003; Moriceau and Sullivan, 2005). Although the behavioral significance of these findings is not completely understood, these data emphasize the fact that the developing brain is structurally and functionally different from the adult brain. Thus it is quite possible that regions of the social brain may play very different roles during the acquisition of social behavior early in life as opposed to the production of social behavior later in development. Though the field of developmental psychology has provided a rich literature on the social development of infants and young children, these studies tell us little about the underlying neural substrates of these behaviors. Approaches that combine behavior and indices of neural function, such as recording event-related potentials (ERPs) as infants view social stimuli, provide important insight into the changes in neural response properties during development (de Haan and Nelson, 1997, 1999). However, the relatively low spatial resolution associated with these techniques does not facilitate the identification of specific brain regions in which the neural changes occur. Complementary techniques, such as functional magnetic resonance imaging (fMRI) and positron emission tomography (PET), which are commonly used in adult subjects, are logistically challenging and pose ethical dilemmas for use in infants and children. Thus there have been very few functional imaging studies evaluating the development of the social brain in infants and children (Tzourio-Mazoyer et al., 2002). Given the challenges of conducting neuroscience research on human infants, alternative approaches, including the use of animal models, may be important in evaluating the specific neural underpinnings of social development. Unfortunately, there has been a relatively limited use of developmental animal models specifically designed to evaluate the emergence of species-typical social behavior. Moreover,

fundamentals of developmental neurobiology

there are methodological issues specific to developmental studies that must be taken into consideration. In order to study social behavior using animal models, it is essential to provide conditions that facilitate the development of speciestypical social behaviors. Providing such conditions is especially important for studies with immature subjects because the developing brain is particularly susceptible to influences from the environment (Meaney, 2001). Aspects of social development appear to rely on “experience-expectant” processes (Nelson, 2001), where the development of skills and abilities that are characteristic of a species are dependent on exposure to specific experiences at certain time points in development (Greenough et al., 1987). In human populations, the damaging effects of early institutional rearing on the social development of young children have been well documented (Gunnar, 2001; Parker and Nelson, 2005). Likewise, it is known that restricting social access of infant nonhuman primates has a profound effect on behavioral development (Capitanio, 1986). Though historical research focused on the profound deficits in social development that result from rearing nonhuman primates in total social isolation (Harlow et al., 1965), we now know that much less severe social restrictions (e.g., nursery or peer rearing of nonhuman primates) can alter both behavioral and neurobiological development (Winslow et al., 2003; Capitanio et al., 2005; Ichise et al., 2006). As we learn more about the interactions between the brain and the environment, it will become increasingly important to evaluate the social rearing conditions of the animals in order to determine whether the behavioral and neural development of these models accurately reflects typical developmental processes.

Macaque monkeys: An animal model of social development In order to understand the underlying neurobiology of human social development, it will be important to examine similar social processes in animal models that are more amenable to traditional neuroscience methodologies. The rich repertoire of social behavior shared by many nonhuman primates makes them a particularly good model in which to study the neural basis of social behavior. Macaque monkeys display remarkable similarities to humans in both social behavior complexity and neuroanatomical organization (Machado and Bachevalier, 2003). These animals are therefore often considered the model of choice for studying the neural bases of complex cognitive processes, such as social behavior. Despite these similarities, it is important to note that the last common ancestor of humans and macaques dates back more than 25 million years (Kay et al., 1997); thus homologous relationships among brain regions are not always clear (Sereno and Tootell, 2005). Moreover, not all aspects of human social cognition can be modeled and

studied in nonhuman primates. However, we are able to study fundamental aspects of social development common to many group-living primates, including the use of speciestypical social signals, the motivation to interact with group members, and the ability to form and maintain lifelong relationships with group members. In this section we will briefly outline the social development of macaque monkeys as a potential model for understanding the component processes of social behavior. Though there are clear behavioral differences among different species of macaques (Thierry, 1985a, 1985b; Maestripieri, 2005), most macaques share common social organization and utilize similar speciestypical social signals. Although it is an oversimplification, we will refer to all species of macaques simply as “macaques” in the following section. Macaque monkeys live in large and cohesive social groups where they form long-lasting relationships with other group members. In general, females will remain in their natal group for the duration of their lives, while males emigrate into a new group when they are 3–5 years of age (Altmann, 1967). As a result, macaques have a strong matrilineal structure in which females from several generations live together and form long-lasting social networks, or matrilines (Wrangham, 1980). Most species of macaques demonstrate well-defined dominance hierarchies in both free-ranging (Drickamer, 1975) and captive social groups (Bernstein and Mason, 1963). Prediction of social rank is closely linked to the dominance status of kin, with high-ranking mothers producing high-ranking offspring (Sade, 1967; Missakian, 1972). In general adult daughters will acquire rank just below their mothers and above their older sisters (de Waal, 1977; Datta, 1984). Macaques utilize a variety of social signals, including vocalizations, facial expressions, and body postures, to communicate with other members of their group. Like humans, infant macaques must rapidly learn to interpret and produce these social signals in order to interact successfully with members of their social group. This wellcharacterized sequence of social development (figure 11.3) in many ways parallels that of human infants, though at a maturational rate approximately four times faster (Suomi, 1999) (i.e., a one-month-old monkey is roughly comparable developmentally to a four-month-old human). As is the case for most primate species, infant macaque monkeys are born with largely functional sensory systems and display an array of reflexive motor responses (Mowbray and Cadell, 1962; Mendelson, 1982a). The mother is their primary source of social stimuli at these early time points, as infant macaques spend almost all of their time in ventral contact or nursing during the first two postnatal weeks (Hansen, 1966; Berman, 1980). This early period of development is characterized by frequent social interactions with their mother including face-to-face communication (e.g., mutual exchanges of lip smacks—an affiliative social signal)

bauman and amaral: neurodevelopment of social cognition

173

Figure 11.3 Social developmental milestones for infant macaque monkeys. The social development of macaque monkeys parallels that of human infants, though at a rate approximately four times faster.

(Hinde and Spencer-Booth, 1967; Ferrari et al., 2006). Indeed, infant macaques show a clear preference for facelike stimuli very early in development (Lutz et al., 1998; Kuwahata et al., 2004) and will produce lip smacks in response to pictures of unfamiliar monkey faces (Mendelson, 1982b; Mendelson et al., 1982). Beginning around 2 weeks of age, infant macaques start to explore the surrounding environment with brief trips away from the mother. Like human infants, the macaque infant appears to use the mother as a secure base and will return to her immediately if alarmed or distressed (Hinde et al., 1964; Berman, 1980). As infants spend more time exploring away from their mothers, they will begin to

174

interact with other members of the social group, primarily their mother’s close female kin and their offspring (Berman, 1982). This period marks a critical stage in development when the infant must acquire the ability to correctly evaluate social signals, particularly signals that may convey potential aggression. At one week of age, infant monkeys do not respond differently to faces of conspecifics staring directly ahead as opposed to faces looking away. However, by 3 weeks of age, infants make fewer fixations on faces looking straight at them (Mendelson et al., 1982). Given that prolonged direct eye contact is often a threatening gesture for macaques (Altmann, 1967), the ability to correctly evaluate the meaning of direct eye contact and respond appropriately

fundamentals of developmental neurobiology

represents a critical developmental milestone. Between 2 and 4 months of age infant monkeys begin to respond to fearful stimuli with a species-typical expression of fear/subordination, the fear grimace (Sackett, 1966; Suomi, 1999) and develop the ability to regulate fear reactions to threatening stimuli (e.g., freezing to remain undetected in the presence of danger) (Kalin and Shelton, 1989). At approximately the same time as the emergence of the fear grimace, infant macaques develop a fear of unfamiliar conspecifics, a behavior that is possibly akin to the “stranger anxiety” observed in human infants between 8 and 12 months of age (Suomi, 1999). By three months of age infants regularly explore away from their mothers, spending approximately 50 percent of their time out of physical contact with her (Hinde and Spencer-Booth, 1967). As infants become more adept at social interactions, they become more fully integrated into the larger social group. At 4–8 months of age social play becomes the dominant activity of macaque infants, taking several forms including physical bouts of “rough and tumble” play, “approach withdrawal” play, and nonsocial play with objects (Ruppenthal et al., 1974). Social play is essential for practicing and refining social behaviors that will be required later in adulthood. Likewise, social grooming is another critical skill that must be acquired. Many primate species participate in social grooming, which plays an important role in establishing and maintaining social relationships (Matheson and Bernstein, 2000). During the second half of the first year, infants begin to consistently initiate grooming of their mothers and other members of their immediate social group (Hinde and Spencer-Booth, 1967). The mother-infant relationship begins to transition into weaning at 5–7 months (Hansen, 1966), though there remains considerable overlap between the social networks of mothers and their infants (Berman, 1982). In order to obtain their appropriate social rank, infant macaques must learn which individuals they outrank and which individuals outrank them. Some time in the second half of the first year of life, infants will begin to direct aggression to adult females (and their offspring) who are lower in rank than their mothers, while deferring to adults (and the offspring) of individuals who are higher ranking than their mothers (Datta, 1984). Thus the social context in which a behavior occurs begins to play an increasingly important role in producing an appropriate response (i.e., was the threat received from a dominant or subordinate group member?). Following the birth of a sibling, macaques continue to develop more independence from their mothers and progress from infant to juvenile social behavior (Devinney et al., 2001). This well-defined sequence of macaque monkey social development provides a rich context in which to study the neurobiology of social behavior. It will be necessary to apply tools of basic neuroscience (e.g., neuroanatomy,

neurophysiology, lesion research, etc.) in order to identify regions of the brain that underlie the acquisition of speciestypical social behavior. In the final section of the chapter we will examine how tools of basic neuroscience can be combined with behavioral studies to evaluate the development of the social brain. It should be apparent that establishing a causal role between neural structures and specific aspects of social behavior will require multiple methodological approaches.

Techniques for studying the neural bases of developmental processes One approach to understanding the developmental neurobiology of social behavior is to relate the functional maturation of the brain with well-defined social milestones, such as those outlined in figure 11.3. Unfortunately, we observe only a rudimentary correlation between overall neuroanatomical changes and the corresponding development of primate social behavior (Levitt, 2003; Machado and Bachevalier, 2003). Moreover, the idea that new social skills are expressed as components of the social brain come “online” may be too simplistic. (That is, how do we define that a region has become fully functional?) It will therefore be necessary to utilize a variety of neuroscience methods in order to identify and study regions of the brain that are essential for acquiring social behavior early in life. In the following subsections we will briefly examine several basic neuroscience techniques and highlight studies that have focused on the development of the social brain using macaque monkeys as a model. Neuroanatomical Studies Neuroanatomy provides a powerful technique to study the developmental progression and maturation of specific regions of interest. Several regions of the macaque brain implicated in social processing in adult animals appear to mature at relatively early developmental time points. For example, we know that neurogenesis of the macaque monkey amygdala begins around embryonic day 33 and is complete by embryonic day 56 of the 165-day gestation period (Kordower et al., 1992). Injections of neuroanatomical tracers during the early postnatal period indicate that macaque amygdalocortical connections already closely resemble connections in the mature subjects by two weeks of postnatal age (Amaral and Bennett, 2000). Likewise, adultlike projections from inferior temporal areas TE and TEO to both amygdala and orbitofrontal areas have been observed in one-week-old macaques (Webster et al., 1991, 1994). Though little is known regarding development of specific neurochemical systems within these regions, it appears that the distribution of opiate receptors within the amygdala and cingulate cortex is comparable to adult patterns as early as one week of age (Bachevalier et al., 1986) and that the pattern of serotonergic innervation of the

bauman and amaral: neurodevelopment of social cognition

175

amygdala resembles the adult pattern within the first postnatal month (Prather and Amaral, 2000). Clearly much more work is needed to provide a comprehensive assessment of neuroanatomical development. Nonetheless, these studies indicate that several regions implicated in social processing appear to mature very early in postnatal development and may therefore play a critical role in the emergence of speciestypical social behavior. Despite these adultlike properties, some aspects of neural development within these circuits may undergo postnatal maturation. As described in part 1 of this chapter, inferior temporal (IT) cortex is critical for visual pattern recognition in adult primates and contains cells highly selective for specific stimuli such as faces (Bruce et al., 1981; Fujita et al., 1992). Detailed studies on both neuroanatomical and physiological development of IT cortex have revealed that it undergoes an extended period of postnatal development and may not be functionally mature until the end of the first year of life (Rodman, 1994). Indeed, cortical inputs and outputs of the IT cortices undergo considerable refinement during the first postnatal months. Interestingly, infant IT cortex receives transient inputs and gives rise to transient outputs, forming connections that are not found in adult monkeys (Webster et al., 1991; Rodman, 1994). Although the functional significance of this unique maturational process is not known, the extended period of postnatal neuroanatomical development provides complementary information for developmental neurophysiological data (see next subsection). In addition to refinement of connections, another aspect of brain development that is protracted is the process of myelination. Although axons can propagate impulses prior to myelination, the process of myelination enhances functional efficiency and specificity. By 3 to 6 months, subcortical regions of the macaque brain are well myelinated, and most cortical regions contain some myelin. However, cortical layers continue to acquire myelin until at least 3.5 years of age in macaque monkeys (Gibson, 1991). Myelination of the cortex follows a similar pattern in humans and monkeys, beginning first and proceeding most rapidly in primary sensory and motor areas, followed by the more protracted myelination of association cortices. Although the general pattern of myelination has been examined, less is known about the progression of myelination of axons within and across brain regions implicated in social processing and how this patterning may relate to functional changes. We do know that myelination of axons within IT cortex has not reached adultlike levels in 7-month-old macaque monkeys (Rodman, 1994), and that myelination within the macaque orbitofrontal cortex may take 1–2 years to reach adultlike levels (Gibson, 1991). Additional studies on the pattern of myelination may provide insight into the functional significance of this prolonged aspect of neural development. Taken together, these neuroanatomical studies

176

indicate that several key structures implicated in social processing are well developed at the time of birth, yet continue to undergo refinement during the first postnatal year of macaque development. Neurophysiology and Functional Imaging Neuroanatomical studies provide critical information regarding the putative processing capabilities and developmental trajectory of specific brain regions. However, this information alone does not imply a functional contribution to social processing. While human research has relied heavily on EEG/ERP and fMRI as indices of brain function, these techniques have not been commonly used in macaque-monkey models of social processing. In recent years PET neuroimaging has been used to identify regions of the adult macaque brain that are activated in response to social challenge (Rilling et al., 2004) or species-specific vocalizations (Gil-da-Costa et al., 2006). Though PET neuroimaging has not been extensively used in developing animals, the potential use of this technique has been demonstrated in a PET study that identified regions of the brain underlying maternal separation behaviors in juvenile macaque monkeys (Rilling et al., 2001). Nonhuman primate research has benefited from the ability to directly record from neurons in awake/behaving animals, providing more direct evidence that a region of interest is involved in social processing. As discussed in the first section of this chapter, cells that respond to faces, body movements, gaze, and so on are found throughout the adult macaque temporal lobe, concentrated in both the inferior temporal gyrus and along the banks of the STS (Gross et al., 1972; Desimone et al., 1984; Baylis et al., 1987; Hasselmo et al., 1989; Perrett et al., 1992) and in the amygdala (Rolls, 1984; Leonard et al., 1985; Brothers et al., 1990; Brothers and Ring, 1993). It would be of interest to developmental social neuroscience to evaluate whether these response properties are established at birth or whether social experience plays a role in developing these highly selective response properties. Unfortunately, little is known regarding how and when these neurons become specialized for processing social information because of the many challenges of conducting this type of research in infant monkeys. We are aware of only one series of studies that has extensively examined response properties of neurons in infant monkeys (Rodman, 1994). These labor-intensive studies revealed that within the second month of life, individual IT neurons show response selectivity for faces, though cells in the infant monkeys show lower response magnitudes and longer response latencies compared to adults (Rodman et al., 1991, 1993). These data suggest that adultlike face selectivity is present at early developmental time points, but undergoes considerable postnatal refinement. These studies highlight the potential use of neurophysiology in understanding how and when the brain becomes specialized for processing social information.

fundamentals of developmental neurobiology

Lesion Research Lesion research provides a means of evaluating whether a neural structure is essential for specific aspects of social behavior. A general approach that we have taken in our nonhuman primate studies is to explore the dependency of component processes of social behavior on putative brain regions. This purpose is achieved experimentally by making selective lesions of the regions of interest and then carrying out detailed behavioral observations in order to define how the behavioral repertoire of the subjects has been altered. As reviewed in part 1 of this chapter, the lesion technique has been extensively used in mature subjects in order to evaluate the contributions of specific brain regions to social processing (Kling, 1992). However, relatively few studies have evaluated the effects of producing similar brain lesions early in development. In this section we will briefly summarize lesion research on regions of the brain implicated in social processing, highlighting studies that have focused on the developmental contributions of these structures. Our discussion will focus primarily on regions of the temporal lobe implicated in social processing, as the contributions of other regions of the social brain (e.g., OFC, ACC, etc.) have not been extensively examined using neonatal lesions (Bowden et al., 1971). In order to evaluate previous lesion studies in neonatal nonhuman primates, it is first important to consider methodological issues that often complicate interpretation of these data. For example, previous lesion research has often utilized aspiration lesions, which may cause unintended collateral damage to surrounding structures and pathways (Meunier et al., 1999). Thus it is possible that behavioral changes are due to unintended collateral damage, rather than to damage to the target structure. Another methodological issue is the quality of behavioral observations. As described earlier, macaque monkeys display a sophisticated repertoire of social behavior. In order to associate changes in behavior to the experimental procedure, it is necessary to utilize a comprehensive behavioral ethogram and to evaluate social development over time and under a variety of testing conditions. Finally, an important consideration for developmental studies is the rearing conditions of the experimental subjects. The majority of previous developmental studies utilized peer- or isolate-reared monkeys, a practice that is standard animal husbandry procedure at many primate facilities. There is, however, reason to believe that this manipulation is sufficient to produce animals with atypical social behavior and altered neurobiological development (Suomi, 1997; Winslow et al., 2003). In order to assess the effects of early lesion damage on social development, it is necessary to provide the experimental subjects with a rearing environment that will ensure species-typical development. Thus any changes in behavior can be more confidently attributed to the experimental lesion and not to a combination of brain damage and restricted rearing conditions.

These methodological issues are important to bear in mind when interpreting the results of lesion research. Lesions of the temporal lobe have historically been implicated in behavioral changes related to socioemotional processing (Brown and Shafer, 1888; Kluver and Bucy, 1939). Indeed, this is one of the few regions extensively studied in developing animals. For example, peer-reared infant monkeys that receive large aspiration lesions of medial temporal lobe structures (including the amygdala, hippocampus, and surrounding cortices) demonstrate fewer social contacts and more withdrawals from attempted social advances compared to unoperated controls (Bachevalier et al., 2001). Although much of the lesion research has focused on larger lesions of the medial temporal lobe, it is beneficial to have information on more discrete lesions of temporal lobe structures in order to associate behavioral deficits with specific brain regions. For example, infant monkeys that sustained damage to the inferior temporal visual area TE within the first postnatal month display less social contact compared to controls at 6 months of age, but do not show deficits in other aspects of social behavior such as eye contact and approach/withdrawal (Bachevalier et al., 2001). Early damage to area TE is also associated with abnormal vocal responses to separation from mothers (Newman and Bachevalier, 1997). The amygdala has long been implicated as a key structure in social processing (Kling, 1992), though recent lesion studies in adult monkeys suggest that the amygdala is not needed to produce species-typical social behavior (Emery et al., 2001). The results from developmental amygdala lesion studies have produced conflicting results, most likely as a result of methodological differences. While early studies on neonatal amygdala damage had reported few changes in behavior (Kling and Green, 1967), subsequent studies reported pronounced changes in fear behaviors (Thompson, 1968; Thompson et al., 1969, 1977; Thompson, 1981) or deficits in social development (Bachevalier, 1994). Given that impaired social communication and a lack of social interest is the hallmark of autism, it was proposed that lesions of the medial temporal lobe, specifically the amygdala, might provide an animal model of autism (Bachevalier, 1994, 1996). However, macaque monkeys that are reared in a social environment and receive selective amygdala lesions at two weeks of age do not demonstrate profound impairments in social development within the first year of life. These subjects were able to produce and respond to a variety of species-typical social signals and did not differ from controls in the amount of their social interactions (Bauman, 2004a, 2004b). These monkeys did, however, show abnormal behavior in fear regulation (e.g., heightened fear of nonthreatening conspecifics and absence of fear to normally fear-inducing objects) (Prather et al., 2001; Bauman et al., 2004b). Our interpretation of these data is that the amygdala

bauman and amaral: neurodevelopment of social cognition

177

does not play an essential role in the development of social behavior but does play a modulatory role by means of regulating emotions (e.g., fear responses) within a social context. Indeed, over time the amygdala-lesioned infants have displayed changes in social behavior, such as decreased social dominance, that may be due to an inability to regulate fear responses (Bauman et al., 2006). Although lesion studies provide a means of assessing whether a given structure is essential for particular aspects of behavior, there are potential limitations to behavioral lesion research that warrant further discussion. We presume that behaviors that are significantly altered following the surgical procedure are normally dependent on structures that have been removed. In contrast, we presume that behaviors that are unchanged following the lesion are not dependent upon those structures. The research presented here is based on permanent, destructive lesions of brain regions implicated in social processing. However, these regions do not function in isolation from other neural structures. Consequently, neonatal damage will most likely affect other brain regions that share connections with these structures. For example, neonatal temporal lobe lesions have previously been associated with delayed maturation of the prefrontal cortex (Bertolino et al., 1997). These changes in brain development are characterized by dysregulation of prefrontal-striatal dopamine transmission that is not observed following similar lesions in adult animals (Saunders et al., 1998; Heinz et al., 1999). These findings of brain reorganization following early brain damage have implications for results obtained through lesion research. Moreover, it is possible that early damage triggers compensatory changes in the brain, recruiting structures that are not normally involved in social behavior to carry out these functions. Thus any sparing of social behavior may be due to compensatory mechanisms of other brain structures not normally involved in social behavior.

Final comments We started the preparation of this chapter with the goal of discussing the development of neural systems involved in social behavior and social cognition. The neurobiology of social cognition is a relatively new area of inquiry spurred on by the seminal paper of Brothers in 1990. However, we have found that the definition of social cognition is still in need of refinement. Moreover, the stages in social behavior from perception to evaluation to behavioral output have not been clearly associated with particular brain regions. In addition, the highest level of social cognition, the building of models of another’s intentions and dispositions—that is, theory of mind—has only recently begun to be evaluated with cognitive neuroscience tools. Thus this chapter does not provide a definitive statement concerning the development

178

of social cognition but represents much more a first step in developing a framework for thinking about and studying developmental social cognition. This will remain an exciting and productive area of research for decades to come. This research will undoubtedly have important implications not only for understanding and nurturing normal social development but also for understanding the impairments of social function that are the hallmark of ever more common neurodevelopmental disorders such as autism spectrum disorders. acknowledgments

Original research described in this chapter was supported by a grant from the National Institute of Mental Health (R37MH57502) and through the Early Experience and Brain Development Network of the MacArthur Foundation. Our research is conducted, in part, at the California National Primate Research Center (RR0069).

REFERENCES Adolphs, R., 1999. Social cognition and the human brain. Trends Cogn. Sci. 3(12):469–479. Adolphs, R., 2003. Cognitive neuroscience of human social behaviour. Nature Rev. Neurosci. 4(3):165–178. Adolphs, R., H. Damasio, and D. Tranel, 2002. Neural systems for recognition of emotional prosody: A 3-D lesion study. Emotion 2(1):23–51. Adolphs, R., H. Damasio, D. Tranel, G. Cooper, and A. R. Damasio, 2000. A role for somatosensory cortices in the visual recognition of emotion as revealed by three-dimensional lesion mapping. J. Neurosci. 20(7):2683–2690. Adolphs, R., F. Gosselin, T. W. Buchanan, D. Tranel, P. Schyns, and A. R. Damasio, 2005. A mechanism for impaired fear recognition after amygdala damage. Nature 433(7021):68– 72. Adolphs, R., D. Tranel, and A. R. Damasio, 1998. The human amygdala in social judgment. Nature 393(6684):470–474. Adolphs, R., D. Tranel, H. Damasio, and A. Damasio, 1994. Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdala. Nature 372(6507): 669–672. Afraz, S. R., R. Kiani, and H. Esteky, 2006. Microstimulation of inferotemporal cortex influences face categorization. Nature 442(7103):692–695. Akiyama, T., M. Kato, T. Muramatsu, F. Saito, R. Nakachi, and H. Kashima, 2006a. A deficit in discriminating gaze direction in a case with right superior temporal gyrus lesion. Neuropsychologia 44(2):161–170. Akiyama, T., M. Kato, T. Muramatsu, F. Saito, S. Umeda, and H. Kashima, 2006b. Gaze but not arrows: A dissociative impairment after right superior temporal gyrus damage. Neuropsychologia 44(10):1804–1810. Allison, T., A. Puce, and G. McCarthy, 2000. Social perception from visual cues: Role of the STS region. Trends Cogn. Sci. 4(7):267–278. Allman, J. M., K. K. Watson, N. A. Tetreault, and A. Y. Hakeem, 2005. Intuition and autism: A possible role for Von Economo neurons. Trends Cogn. Sci. 9(8):367–373.

fundamentals of developmental neurobiology

Altmann, S., 1967. The structure of primate social communication. Social Communication among Primates, ed. S. Altmann. Chicago, University of Chicago Press. Amaral, D. G., 2003. The amygdala, social behavior, and danger detection. Ann. NY Acad. Sci. 1000:337–347. Amaral, D. G., M. D. Bauman, J. P. Capitanio, P. Lavenex, W. A. Mason, and M. L. Mauldin-Jourdain, 2003. The amygdala: Is it an essential component of the neural network for social cognition? Neuropsychologia 41:517–522. Amaral, D. G., and J. Bennett, 2000. Development of amygdalo-cortical connections in the macaque monkey: Program no. 644.10. Abstract Viewer and Itinerary Planner. Washington DC: Society for Neuroscience, 2003. Online. Amaral, D. G., J. L. Price, A. Pitkanen, and S. T. Carmichael, 1992. Anatomical organization of the primate amygdaloid complex. The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction, ed. J. P. Aggleton. New York, WileyLiss. Angrilli, A., D. Palomba, A. Cantagallo, A. Maietti, and L. Stegagno, 1999. Emotional impairment after right orbitofrontal lesion in a patient without cognitive deficits. NeuroReport 10(8):1741–1746. APA, 1994. Diagnostic and Statistical Manual of Mental Disorders (DSMIV). Washington, DC: American Psychiatric Association. Apperly, I. A., D. Samson, and G. W. Humphreys, 2005. Domainspecificity and theory of mind: Evaluating neuropsychological evidence. Trends Cogn. Sci. 9(12):572–577. Bachevalier, J., 1994. Medial temporal lobe structures and autism—A review of clinical and experimental findings. Neuropsychologia 32(6):627–648. Bachevalier, J., 1996. Brief report—Medial temporal lobe and autism—A putative animal model in primates. J. Autism Dev. Disord. 26(2):217–220. Bachevalier, J., 2000. The amygdala, social cognition, and autism. The Amygdala: A Functional Analysis, ed. J. P. Aggleton, 509–543. New York, Oxford University Press. Bachevalier, J., L. Malkova, and M. Mishkin, 2001. Effects of selective neonatal temporal lobe lesions on socioemotional behavior in infant rhesus monkeys (Macaca mulatta). Behav. Neurosci. 115(3):545–559. Bachevalier, J., L. G. Ungerleider, J. B. O’Neill, and D. P. Friedman, 1986. Regional distribution of [3H] naloxone binding in the brain of a newborn rhesus monkey. Brain Res. 390:302–308. Baron-Cohen, S., H. A. Ring, S. Wheelwright, E. T. Bullmore, M. J. Brammer, A. Simmons, et al., 1999. Social intelligence in the normal and autistic brain: An fMRI study. Eur. J. Neurosci. 11(6):1891–1898. Barrett, L., P. Henzi, and R. Dunbar, 2003. Primate cognition: From “what now?” to “what if?” Trends Cogn. Sci. 7(11): 494–497. Bauman, M. D., P. Lavenex, W. A. Mason, J. P. Capitanio, and D. G. Amaral, 2004a. The development of mother-infant interactions after neonatal amygdala lesions in rhesus monkeys. J. Neurosci. 24(3):711–721. Bauman, M. D., P. Lavenex, W. A. Mason, J. P. Capitanio, and D. G. Amaral, 2004b. The development of social behavior following neonatal amygdala lesions in rhesus monkeys. J. Cogn. Neurosci. 16(8):1388–1411. Bauman, M. D., J. E. Toscano, W. A. Mason, P. Lavenex, and D. G. Amaral, 2006. The expression of social dominance following neonatal lesions of the amygdala or hippocampus in rhesus monkeys (Macaca mulatta). Behav. Neurosci. 120(4):749–760.

Baxter, M. G., and E. A. Murray, 2002. The amygdala and reward. Nature Rev. Neurosci. 3(7):563–573. Baylis, G. C., E. T. Rolls, and C. M. Leonard, 1987. Functional subdivisions of the temporal lobe neocortex. J. Neurosci. 7(2):330–342. Bechara, A., H. Damasio, D. Tranel, and S. W. Anderson, 1998. Dissociation of working memory from decision making within the human prefrontal cortex. J. Neurosci. 18(1):428– 437. Bellugi, U., L. Lichtenberger, D. Mills, A. Galaburda, and J. R. Korenberg, 1999. Bridging cognition, the brain and molecular genetics: Evidence from Williams syndrome. Trends Neurosci. 22(5):197–207. Berlin, H. A., E. T. Rolls, and U. Kischka, 2005. Borderline personality disorder, impulsivity, and the orbitofrontal cortex. Am. J. Psychiatry 162(12):2360–2673. Berlin, H. A., E. T. Rolls, and S. D. Iversen, 2004. Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions. Brain 127(part 5): 1108–1126. Berman, C. I., 1980. Mother-infant relationships among freeranging rhesus monkeys on Cayo Santiago: A comparison with captive pairs. Anim. Behav. 28:860–873. Berman, C. M., 1982. The ontogeny of social relationships with group companions among free-ranging rhesus monkeys. I. Social networks and differentiation. Anim. Behav. 30:149–162. Bernstein, I. S., and W. A. Mason, 1963. Group formation by rhesus monkeys. Anim. Behav. 11(1):28–31. Berthoz, S., J. L. Armony, R. J. Blair, and R. J. Dolan, 2002. An fMRI study of intentional and unintentional (embarrassing) violations of social norms. Brain 125(part 8):1696–1708. Bertolino, A., R. C. Saunders, V. S. Mattay, J. Bachevalier, J. A. Frank, and D. R. Weinberger, 1997. Altered development of prefrontal neurons in rhesus monkeys with neonatal mesial temporo-limbic lesions: A proton magnetic resonance spectroscopic imaging study. Cerebral Cortex 7(8):740–748. Blair, R. J., and L. Cipolotti, 2000. Impaired social response reversal: A case of “acquired sociopathy.” Brain 123(part 6): 1122–1141. Blair, R. J., J. S. Morris, C. D. Frith, D. I. Perrett, and R. J. Dolan, 1999. Dissociable neural responses to facial expressions of sadness and anger. Brain 122(part 5):883–893. Bonda, E., M. Petrides, D. Ostry, and A. Evans, 1996. Specific involvement of human parietal systems and the amygdala in the perception of biological motion. J. Neurosci. 16(11):3737– 3744. Bowden, D. M., P. S. Goldman, H. E. Rosvold, and R. L. Greenstreet, 1971. Free behavior of rhesus monkeys following lesions of the dorsolateral and orbital prefrontal cortex in infancy. Exp. Brain Res. 12(3):265–274. Breiter, H. C., N. L. Etcoff, P. J. Whalen, W. A. Kennedy, S. L. Rauch, R. L. Buckner, et al., 1996. Response and habituation of the human amygdala during visual processing of facial expression. Neuron 17(5):875–887. Brothers, L., 1990. The social brain: A project for integrating primate behavior and neurophysiology in a new domain. Concepts Neurosci. 1:27–51. Brothers, L., and B. Ring, 1993. Mesial temporal neurons in the macaque monkey with responses selective for aspects of social stimuli. Behav. Brain Res. 57(1):53–61. Brothers, L., B. Ring,and A. Kling, 1990. Response of neurons in the macaque amygdala to complex social stimuli. Behav. Brain Res. 41(3):199–213.

bauman and amaral: neurodevelopment of social cognition

179

Brown, S., and E. A. Shafer, 1888. An investigation into the functions of the occipital and temporal lobes of the monkey’s brain. Philos. Trans. R. Soc. London [Biol.] 179:303–327. Bruce, C., R. Desimone, and C. G. Gross, 1981. Visual properties of neurons in a polysensory area in superior temporal sulcus of the macaque. J. Neurophysiol. 46(2):369–384. Buchel, C., J. Morris, R. J. Dolan, and K. J. Friston, 1998. Brain systems mediating aversive conditioning: An event-related fMRI study. Neuron 20(5):947–957. Bukach, C. M., I. Gauthier, and M. J. Tarr, 2006. Beyond faces and modularity: The power of an expertise framework. Trends Cogn. Sci. 10(4):159–166. Bush, G., P. Luu, and M. I. Posner, 2000. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn. Sci. 4(6):215– 222. Bush, G., B. A. Vogt, J. Holmes, A. M. Dale, D. Greve, M. A. Jenike, et al., 2002. Dorsal anterior cingulate cortex: A role in reward-based decision making. Proc. Natl. Acad. Sci. USA 99(1): 523–528. Butter, C. M., and D. R. Snyder, 1972. Alterations in aversive and aggressive behaviors following orbital frontal lesions in rhesus monkeys. Acta Neurobiol. Exp. (Warsz.) 32(2):525–565. Butter, C. M., D. R. Snyder, and J. A. McDonald, 1970. Effects of orbital frontal lesions on aversive and aggressive behaviors in rhesus monkeys. J. Comp. Physiol. Psychol. 72(1):132–144. Cahill, L., and J. L. McGaugh, 1998. Mechanisms of emotional arousal and lasting declarative memory. Trends Neurosci. 21(7): 294–299. Campbell, R., C. A. Heywood, A. Cowey, M. Regard, and T. Landis, 1990. Sensitivity to eye gaze in prosopagnosic patients and monkeys with superior temporal sulcus ablation. Neuropsychologia 28(11):1123–1142. Canli, T., H. Sivers, S. L. Whitfield, I. H. Gotlib, and J. D. Gabrieli, 2002. Amygdala response to happy faces as a function of extraversion. Science 296(5576):2191. Canli, T., Z. Zhao, J. Brewer, J. D. Gabrieli, and L. Cahill, 2000. Event-related activation in the human amygdala associates with later memory for individual emotional experience. J. Neurosci. 20(19):RC99. Capitanio, J., 1986. Behavioral pathology. Comparative Primate Biology: Behavior, Conservation, and Ecology, ed. G. Mitchell and J. Erwin, 2A:411–454. New York, Alan R. Liss. Capitanio, J. P., S. P. Mendoza, W. A. Mason, and N. Maninger, 2005. Rearing environment and hypothalamicpituitary-adrenal regulation in young rhesus monkeys (Macaca mulatta). Dev. Psychobiol. 46(4):318–330. Carr, L., M. Iacoboni, M. C. Dubeau, J. C. Mazziotta, and G. L. Lenzi, 2003. Neural mechanisms of empathy in humans: A relay from neural systems for imitation to limbic areas. Proc. Natl. Acad. Sci. USA 100(9):5497–5502. Castelli, F., F. Happe, U. Frith, and C. Frith, 2000. Movement and mind: A functional imaging study of perception and interpretation of complex intentional movement patterns. NeuroImage 12(3):314–325. Cheney, D. L., and R. M. Seyfarth, 1990. The representation of social relations by monkeys. Cognition 37(1–2):167–196. Cheney, D., R. Seyfarth, and B. Smuts, 1986. Social relationships and social cognition in nonhuman primates. Science 234 (4782):1361–1366. Cheng, D. T., D. C. Knight, C. N. Smith, E. A. Stein, and F. J. Helmstetter, 2003. Functional MRI of human amygdala activity during Pavlovian fear conditioning: Stimulus processing versus response expression. Behav. Neurosci. 117(1):3–10.

180

Corkin, S., D. G. Amaral, R. G. Gonzalez, K. A. Johnson, and B. T. Hyman, 1997. H.M.’s medial temporal lobe lesion: Findings from magnetic resonance imaging. J. Neurosci. 17(10): 3964–3979. Dapretto, M., M. S. Davies, J. H. Pfeifer, A. A. Scott, M. Sigman, S. Y. Bookheimer, et al., 2006. Understanding emotions in others: Mirror neuron dysfunction in children with autism spectrum disorders. Nature Neurosci. 9(1):28–30. Darwin, C., 1872. The Expression of the Emotions in Man and Animals. London: John Murray. Datta, S. B., 1984. Relative power and the acquisition of rank. Primate Social Relationships: An Integrated Approach, ed. R. A. Hinde, 93–103. Oxford, UK: Blackwell’s Scientific. Davis, M., 1992. The role of the amygdala in fear and anxiety. Annu. Rev. Neurosci. 15:353–375. Davis, M., and P. J. Whalen, 2001. The amygdala: Vigilance and emotion. Mol. Psychiatry 6(1):13–34. de Bono, M., and C. I. Bargmann, 1998. Natural variation in a neuropeptide Y receptor homolog modifies social behavior and food response in C. elegans. Cell 94(5):679–689. de Bono, M., and A. V. Maricq, 2005. Neuronal substrates of complex behaviors in C. elegans. Annu. Rev. Neurosci. 28:451– 501. de Bono, M., D. M. Tobin, M. W. Davis, L. Avery, and C. I. Bargmann, 2002. Social feeding in Caenorhabditis elegans is induced by neurons that detect aversive stimuli. Nature 419(6910): 899–903. de Haan, M., and C. A. Nelson, 1997. Recognition of the mother’s face by six-month-old infants: A neurobehavioral study. Child Dev. 68(2):187–210. de Haan, M., and C. A. Nelson, 1999. Brain activity differentiates face and object processing in 6-month-old infants. Dev. Psychol. 35(4):1113–1121. Desimone, R., T. D. Albright, C. G. Gross, and C. Bruce, 1984. Stimulus-selective properties of inferior temporal neurons in the macaque. J. Neurosci. 4(8):2051–2062. Devinney, B. J., C. M. Berman, and K. L. Rasmussen, 2001. Changes in yearling rhesus monkeys’ relationships with their mothers after sibling birth. Am. J. Primatol. 54(4):193–210. Devinsky, O., M. J. Morrell,and B. A. Vogt, 1995. Contributions of anterior cingulate cortex to behaviour. Brain 118 (part 1):279–306. de Waal, F. B., 1977. The organization of agonistic relations within two captive groups of Java-monkeys (Macaca fascicularis). Z. Tierpsychol. 44(3):225–282. Dilger, S., T. Straube, H. J. Mentzel, C. Fitzek, J. R. Reichenbach, H. Hecht, et al., 2003. Brain activation to phobia-related pictures in spider phobic humans: An event-related functional magnetic resonance imaging study. Neurosci. Lett. 348(1):29– 32. Doyle, T. F., U. Bellugi, J. R. Korenberg, and J. Graham, 2004. Everybody in the world is my friend hypersociability in young children with Williams syndrome. Am. J. Med. Genet. [A] 124(3): 263–273. Drickamer, L. C., 1975. Quantitative observation of behavior in free-ranging Macaca mulatta: Methodology and aggression. Behaviour 55(3–4):209–236. Dunbar, R. I. M., 1998. The social brain hypothesis. Evol. Anthropol. 178:178–190. Eacott, M. J., C. A. Heywood, C. G. Gross, and A. Cowey, 1993. Visual discrimination impairments following lesions of the superior temporal sulcus are not specific for facial stimuli. Neuropsychologia 31(6):609–619.

fundamentals of developmental neurobiology

Ekman, P., 1993. Facial expression and emotion. Am. Psychol. 48(4):384–392. Ekman, P., and W. V. Friesen, 1971. Constants across cultures in the face and emotion. J. Pers. Soc. Psychol. 17(2):124–129. Emery, N. J., and D. G. Amaral, 2000. The role of the primate amygdala in social cognition. Cognitive Neuroscience of Emotion, ed. R. D. Lane and L. Nadel, 156–191. Series in affective science. New York: Oxford University Press. Emery, N. J., J. P. Capitanio, W. A. Mason, C. J. Machado, S. P. Mendoza, and D. G. Amaral, 2001. The effects of bilateral lesions of the amygdala on dyadic social interactions in rhesus monkeys (Macaca mulatta). Behav. Neurosci. 115(3):515–544. Eslinger, P. J., and A. R. Damasio, 1985. Severe disturbance of higher cognition after bilateral frontal lobe ablation: Patient EVR. Neurology 35(12):1731–1741. Farah, M. J., 1996. Is face recognition “special”? Evidence from neuropsychology. Behav. Brain Res. 76(1–2):181–189. Farah, M. J., K. D. Wilson, M. Drain, and J. N. Tanaka, 1998. What is “special” about face perception? Psychol. Rev. 105(3): 482–498. Ferrari, P. F., V. Gallese, G. Rizzolatti, and L. Fogassi, 2003. Mirror neurons responding to the observation of ingestive and communicative mouth actions in the monkey ventral premotor cortex. Eur. J. Neurosci. 17(8):1703–1714. Ferrari, P. F., E. Visalberghi, A. Paukner, L. Fogassi, A. Ruggiero, and S. J. Suomi, 2006. Neonatal imitation in rhesus macaques. PLoS Biol. 4(9). Fletcher, P. C., F. Happe, U. Frith, S. C. Baker, R. J. Dolan, R. S. Frackowiak, et al., 1995. Other minds in the brain: A functional imaging study of “theory of mind” in story comprehension. Cognition 57(2):109–128. Flombaum, J. I., and L. R. Santos, 2005. Rhesus monkeys attribute perceptions to others. Curr. Biol. 15(5):447–452. Freedman, D. G., 1964. Smiling in blind infants and the issue of innate vs. acquired. J. Child Psychol. Psychiatry 47:171–184. Freese, J. L., and D. G. Amaral, 2005. The organization of projections from the amygdala to visual cortical areas TE and V1 in the macaque monkey. J. Comp. Neurol. 486(4):295–317. Fujita, I., K. Tanaka, M. Ito, and K. Cheng, 1992. Columns for visual features of objects in monkey inferotemporal cortex. Nature 360(6402):343–346. Gallagher, H. L., and C. D. Frith, 2003. Functional imaging of “theory of mind.” Trends Cogn. Sci. 7(2):77–83. Gallagher, H. L., F. Happe, N. Brunswick, P. C. Fletcher, U. Frith, and C. D. Frith, 2000. Reading the mind in cartoons and stories: An fMRI study of “theory of mind” in verbal and nonverbal tasks. Neuropsychologia 38(1):11–21. Gallagher, H. L., A. I. Jack, A. Roepstorff, and C. D. Frith, 2002. Imaging the intentional stance in a competitive game. NeuroImage 16(3, part 1):814–821. Gallese, V., L. Fadiga, L. Fogassi, and G. Rizzolatti, 1996. Action recognition in the premotor cortex. Brain 119(part 2): 593–609. Gallese, V., C. Keysers, and G. Rizzolatti, 2004. A unifying view of the basis of social cognition. Trends Cogn. Sci. 8(9): 396–403. Gauthier, I., and C. Bukach, 2007. Should we reject the expertise hypothesis? Cognition 103(2):322–330. Gauthier, I., P. Skudlarski, J. C. Gore, and A. W. Anderson, 2000. Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neurosci. 3(2):191–197. Gauthier, I., M. J. Tarr, A. W. Anderson, P. Skudlarski, and J. C. Gore, 1999. Activation of the middle fusiform “face area”

increases with expertise in recognizing novel objects. Nature Neurosci. 2(6):568–573. Gibson, K. R., 1991. Myelination and behavioral development: A comparative perspective on questions of neoteny, altriciality and intelligence. Brain Maturation and Cognitive Development: Comparative and Cross-Cultural Perspectives, ed. K. R. Gibson and A. C. Peterson, 29–63. New York: Aldine De Gruyter. Giedd, J. N., A. C. Vaituzis, S. D. Hamburger, N. Lange, J. C. Rajapakse, D. Kaysen, et al., 1996. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: Ages 4–18 years. J. Comp. Neurol. 366(2): 223–230. Gil-da-Costa, R., A. Martin, M. A. Lopes, M. Munoz, J. B. Fritz, and A. R. Braun, 2006. Species-specific calls activate homologs of Broca’s and Wernicke’s areas in the macaque. Nature Neurosci. 9(8):1064–1070. Glees, P., J. Cole, C. W. Whitty, and H. Cairns, 1950. The effects of lesions in the cingular gyrus and adjacent areas in monkeys. J. Neurol. Neurosurg. Psychiatry 13(3):178–190. Goren, C. C., M. Sarty, and P. Y. Wu, 1975. Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics 56(4):544–549. Gothard, K. M., F. P. Battaglia, C. A. Erickson, K. M. Spitler, and D. G. Amaral, 2007. Neural responses to facial expression and face identity in the monkey amygdala. J. Neurophysiol. 97(2):1671–1683. Gottfried, J. A., J. O’Doherty, and R. J. Dolan, 2003. Encoding predictive reward value in human amygdala and orbitofrontal cortex. Science 301(5636):1104–1107. Gottlieb, G., and R. Lickliter, 2004. The various roles of animal models in understanding human development. Soc. Dev. 13(2): 311–325. Greenough, W. T., J. E. Black, and C. S. Wallace, 1987. Experience and brain development. Child Dev. 58(3):539–559. Griffin, R., O. Friedman, J. Ween, E. Winner, F. Happe, and H. Brownell, 2006. Theory of mind and the right cerebral hemisphere: Refining the scope of impairment. Laterality 11(3): 195–225. Grill-Spector, K., N. Knouf, and N. Kanwisher, 2004. The fusiform face area subserves face perception, not generic withincategory identification. Nature Neurosci. 7(5):555–562. Grill-Spector, K., R. Sayres, and D. Ress, 2006. Highresolution imaging reveals highly selective nonface clusters in the fusiform face area. Nature Neurosci. Gross, C. G., C. E. Rocha-Miranda, and D. B. Bender, 1972. Visual properties of neurons in inferotemporal cortex of the macaque. J. Neurophysiol. 35(1):96–111. Grossman, E., M. Donnelly, R. Price, D. Pickens, V. Morgan, G. Neighbor, et al., 2000. Brain areas involved in perception of biological motion. J. Cogn. Neurosci. 12(5):711–720. Gunnar, M. R., 2001. Effects of early deprivation: Findings from orphanage-reared infants and children. Handbook of Developmental Cognitive Neuroscience, ed. C. A. Nelson and M. Luciana, 617–629. Cambridge, MA: MIT Press. Hadland, K. A., M. F. Rushworth, D. Gaffan, and R. E. Passingham, 2003. The effect of cingulate lesions on social behaviour and emotion. Neuropsychologia 41(8):919–931. Hansen, E. W., 1966. The development of maternal and infant behavior in the rhesus monkey. Behaviour 27(1–2):107– 149. Hariri, A. R., V. S. Mattay, A. Tessitore, F. Fera, and D. R. Weinberger, 2003. Neocortical modulation of the amygdala response to fearful stimuli. Biol. Psychiatry 53(6):494–501.

bauman and amaral: neurodevelopment of social cognition

181

Harlow, H. F., R. O. Dodsworth, and M. K. Harlow, 1965. Total social isolation in monkeys. Proc. Natl. Acad. Sci. USA 54(1):90–97. Hasselmo, M. E., E. T. Rolls, and G. C. Baylis, 1989. The role of expression and identity in the face-selective responses of neurons in the temporal visual cortex of the monkey. Behav. Brain Res. 32(3):203–218. Haxby, J. V., M. I. Gobbini, M. L. Furey, A. Ishai, J. L. Schouten, and P. Pietrini, 2001. Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science 293(5539):2425–2430. Haxby, J. V., E. A. Hoffman, and M. I. Gobbini, 2000. The distributed human neural system for face perception. Trends Cogn. Sci. 4(6):223–233. Haxby, J. V., E. A. Hoffman, and M. I. Gobbini, 2002. Human neural systems for face recognition and social communication. Biol. Psychiatry 51(1):59–67. Heberlein, A. S., and R. Adolphs, 2004. Impaired spontaneous anthropomorphizing despite intact perception and social knowledge. Proc. Natl. Acad. Sci. USA 101(19):7487–7491. Heinz, A., R. C. Saunders, B. S. Kolachana, D. W. Jones, J. G. Gorey, J. Bachevalier, et al., 1999. Striatal dopamine receptors and transporters in monkeys with neonatal temporal limbic damage. Synapse 32(2):71–79. Heywood, C. A., and A. Cowey, 1992. The role of the “face-cell” area in the discrimination and recognition of faces by monkeys. Philos. Trans. R. Soc. London [Biol.] 335(1273):31–37; discussion 37–38. Hinde, R. A., T. E. Rowell, R. A. Hinde, T. E. Rowell, and Y. Spencer-Booth, 1964. Behavior of socially living rhesus monkeys in their first six months. Proc. Zool. Soc. Lond. 143:609– 649. Hinde, R. A., and Y. Spencer-Booth, 1967. The behaviour of socially living rhesus monkeys in their first two and a half years. Anim. Behav. 15(1):169–196. Hof, P. R., E. A. Nimchinsky, D. P. Perl, and J. M. Erwin, 2001. An unusual population of pyramidal neurons in the anterior cingulate cortex of hominids contains the calcium-binding protein calretinin. Neurosci. Lett. 307(3):139–142. Hof, P. R., and E. Van Der Gucht, 2007. Structure of the cerebral cortex of the humpback whale, Megaptera novaeangliae (Cetacea, Mysticeti, Balaenopteridae). Anat. Rec. (Hoboken) 290(1): 1–31. Hoffman, E. A., and J. V. Haxby, 2000. Distinct representations of eye gaze and identity in the distributed human neural system for face perception. Nature Neurosci. 3(1):80–84. Hornak, J., J. Bramham, E. T. Rolls, R. G. Morris, J. O’Doherty, P. R. Bullock, et al., 2003. Changes in emotion after circumscribed surgical lesions of the orbitofrontal and cingulate cortices. Brain 126(part 7):1691–1712. Hornak, J., E. T. Rolls, and D. Wade, 1996. Face and voice expression identification in patients with emotional and behavioural changes following ventral frontal lobe damage. Neuropsychologia 34(4):247–261. Iacoboni, M., 2005. Neural mechanisms of imitation. Curr. Opin. Neurobiol. 15(6):632–637. Iacoboni, M., and M. Dapretto, 2006. The mirror neuron system and the consequences of its dysfunction. Nature Rev. Neurosci. 7(12):942–951. Iacoboni, M., M. D. Lieberman, B. J. Knowlton, I. MolnarSzakacs, M. Moritz, M. C. J. Throop, et al., 2004. Watching social interactions produces dorsomedial prefrontal and medial

182

parietal BOLD fMRI signal increases compared to a resting baseline. NeuroImage 21(3):1167–1173. Iacoboni, M., I. Molnar-Szakacs, V. Gallese, G. Buccino, J. C. Mazziotta, and G. Rizzolatti, 2005. Grasping the intentions of others with one’s own mirror neuron system. PLoS Biol. 3(3): e79. Iacoboni, M., R. P. Woods, M. Brass, H. Bekkering, J. C. Mazziotta, and G. Rizzolatti, 1999. Cortical mechanisms of human imitation. Science 286(5449):2526–2528. Ichise, M., D. C. Vines, T. Gura, G. M. Anderson, S. J. Suomi, J. D. Higley, et al., 2006. Effects of early life stress on [11C]DASB positron emission tomography imaging of serotonin transporters in adolescent peer- and mother-reared rhesus monkeys. J. Neurosci. 26(17):4638–4643. Insel, T. R., 1997. A neurobiological basis of social attachment. Am. J. Psychiatry 154(6):726–735. Insel, T. R., L. Young, and Z. Wang, 1997. Molecular aspects of monogamy. Ann. NY Acad. Sci. 807:302–316. Jackson, P. L., A. N. Meltzoff, and J. Decety, 2005. How do we perceive the pain of others? A window into the neural processes involved in empathy. NeuroImage 24(3):771–779. Jellema, T., and D. I. Perrett, 2005. Neural basis for the perception of goal-directed actions. The Cognitive Neuroscience of Social Behavior, ed. N. J. Emery and A. Easton, 81–111. New York: Psychology Press. Johnson, M. H., S. Dziurawiec, H. Ellis, and J. Morton, 1991. Newborns’ preferential tracking of face-like stimuli and its subsequent decline. Cognition 40(1–2):1–19. Jurgens, U., 2002. Neural pathways underlying vocal control. Neurosci. Biobehav. Rev. 26(2):235–258. Kalin, N., and S. Shelton, 1989. Defensive behaviors in infant rhesus monkeys: Environmental cues and neurochemical regulation. Science 243:1718–1721. Kanwisher, N., 2000. Domain specificity in face perception. Nature Neurosci. 3(8):759–763. Kanwisher, N., J. McDermott, and M. M. Chun, 1997. The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17(11):4302–4311. Kanwisher, N., D. Stanley, and A. Harris, 1999. The fusiform face area is selective for faces not animals. NeuroReport 10(1): 183–187. Kay, R. F., C. Ross, and B. A. Williams, 1997. Anthropoid origins. Science 275(5301):797–804. Keysers, C., and D. I. Perrett, 2004. Demystifying social cognition: A Hebbian perspective. Trends Cogn. Sci. 8(11):501–507. Kling, A. S., 1992. The amygdala and social behavior. The Amygdala: Neurobiological Aspects of Emotion, Memory, and Dysfunction, 353–378, ed. J. P. Aggleton. New York: Wiley-Liss. Kling, A. S., and P. C. Green, 1967. Effects of neonatal amygdalectomy in the maternally reared and maternally deprived macaque. Nature 213(5077):742–743. Kluver, H., and P. Bucy, 1939. Preliminary analysis of functions of the temporal lobes in monkeys. Arch. Neurol. Psychiatry 42:979– 1000. Kordower, J. H., P. Piecinski, and P. Rakic, 1992. Neurogenesis of the amygdaloid nuclear complex in the rhesus monkey. Brain Res. Dev. Brain Res. 68(1):9–15. Kringelbach, M. L., and E. T. Rolls, 2003. Neural correlates of rapid reversal learning in a simple model of human social interaction. NeuroImage 20(2):1371–1383. Kringelbach, M. L., and E. T. Rolls, 2004. The functional neuroanatomy of the human orbitofrontal cortex: Evidence from

fundamentals of developmental neurobiology

neuroimaging and neuropsychology. Prog. Neurobiol. 72(5):341– 372. Kuwahata, H., I. Adachi, K. Fujita, M. Tomonaga, and T. Matsuzawa, 2004. Development of schematic face preference in macaque monkeys. Behav. Process. 66(1):17–21. LaBar, K. S., J. C. Gatenby, J. C. Gore, J. E. LeDoux, and E. A. Phelps, 1998. Human amygdala activation during conditioned fear acquisition and extinction: A mixed-trial fMRI study. Neuron 20(5):937–945. LeDoux, J., 1998. Fear and the brain: Where have we been, and where are we going? Biol. Psychiatry 44(12):1229–1238. LeDoux, J. E., 2000. Emotion circuits in the brain. Annu. Rev. Neurosci. 23:155–184. Leonard, C. M., E. T. Rolls, F.A.W. Wilson, and G. C. Baylis, 1985. Neurons in the amydgala of the monkey with responses selective for faces. Behav. Brain Res. 1985:159–176. Levitt, P., 2003. Structural and functional maturation of the developing primate brain. J. Pediatr. 143(4 suppl):S35– 45. Lutz, C. K., J. S. Lockard, V. M. Gunderson, and K. S. Grant, 1998. Infant monkeys’ visual responses to drawings of normal and distorted faces. Am. J. Primatol. 44(2):169–174. Machado, C. J., and J. Bachevalier, 2003. Nonhuman primate models of childhood psychopathology: The promise and the limitations. J. Child Psychol. Psychiatry 44(1):64–87. Machado, C. J., and J. Bachevalier, 2006. The impact of selective amygdala, orbital frontal cortex, or hippocampal formation lesions on established social relationships in rhesus monkeys (Macaca mulatta). Behav. Neurosci. 120(4):761–786. Maestripieri, D., 2003. Similarities in affiliation and aggression between cross-fostered rhesus macaque females and their biological mothers. Dev. Psychobiol. 43(4):321–327. Maestripieri, D., 2005. Gestural communication in three species of macaques (Macaca mulatta, M. nemestrina, M. arctoides): Use of signals in relation to dominance and social context. Gesture (5):57–73. Mah, L., M. C. Arnold, and J. Grafman, 2004. Impairment of social perception associated with lesions of the prefrontal cortex. Am. J. Psychiatry 161(7):1247–1255. Malkova, L., D. Gaffan, and E. A. Murray, 1997. Excitotoxic lesions of the amygdala fail to produce impairment in visual learning for auditory secondary reinforcement but interfere with reinforcer devaluation effects in rhesus monkeys. J. Neurosci. 17(15):6011–6020. Matheson, M. D., and I. S. Bernstein, 2000. Grooming, social bonding, and agonistic aiding in rhesus monkeys. Am. J. Primatol. 51(3):177–186. McCabe, K., D. Houser, L. Ryan, V. Smith, and T. Trouard, 2001. A functional imaging study of cooperation in two-person reciprocal exchange. Proc. Natl. Acad. Sci. USA 98(20):11832– 11835. McGonigle, B. O., and M. Chalmers, 1977. Are monkeys logical? Nature 267(5613):694–696. McKone, E., and R. Robbins, 2007. The evidence rejects the expertise hypothesis: Reply to Gauthier and Bukach. Cognition. 103(2):331–336. Meaney, M. J., 2001. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu. Rev. Neurosci. 24:1161–1192. Meltzoff, A. N., and M. K. Moore, 1977. Imitation of facial and manual gestures by human neonates. Science 198(4312):74– 78.

Meltzoff, A. N., and M. K. Moore, 1983. Newborn infants imitate adult facial gestures. Child Dev. 54(3):702–709. Mendelson, M. J., 1982a. Clinical examination of visual and social responses in infant rhesus monkeys. Dev. Psychol. 18(5):658– 664. Mendelson, M. J., 1982b. Visual and social responses in infant rhesus monkeys. Am. J. Primatol. 3:333–340. Mendelson, M. J., M. M. Haith, and P. S. Goldman-Rakic, 1982. Face scanning and responsiveness to social cues in infant rhesus monkeys. Dev. Psychol. 18(2):222–228. Meunier, M., J. Bachevalier, E. A. Murray, L. Malkova, and M. Mishkin, 1999. Effects of aspiration versus neurotoxic lesions of the amygdala on emotional responses in monkeys. Eur. J. Neurosci. 11(12):4403–4418. Mirsky, A. F., H. E. Rosvold, and K. H. Pribram, 1957. Effects of cingulectomy on social behavior in monkeys. J. Neurophysiol. 20(6):588–601. Missakian, E. A., 1972. Genealogical and cross-genealogical dominance relations in a group of free ranging rhesus monkeys (Macaca mulatta) on Cayo Santiago. Primates 13(2):169–180. Moll, J., R. de Oliveira-Souza, I. E. Bramati, and J. Grafman, 2002. Functional networks in emotional moral and nonmoral social judgments. NeuroImage 16(3, part 1):696–703. Moriceau, S., and R. M. Sullivan, 2005. Neurobiology of infant attachment. Dev. Psychobiol. 47(3):230–242. Morris, J. S., K. J. Friston, C. Buchel, C. D. Frith, A. W. Young, A. J. Calder, et al., 1998. A neuromodulatory role for the human amygdala in processing emotional facial expressions. Brain 121(part 1):47–57. Morris, J. S., C. D. Frith, D. I. Perrett, D. Rowland, A. W. Young, A. J. Calder, et al., 1996. A differential neural response in the human amygdala to fearful and happy facial expressions. Nature 383(6603):812–815. Morris, J. S., A. Ohman, and R. J. Dolan, 1998. Conscious and unconscious emotional learning in the human amygdala. Nature 393(6684):467–470. Morton, J., and M. H. Johnson, 1991. CONSPEC and CONLERN: A two-process theory of infant face recognition. Psychol. Rev. 98(2):164–181. Mowbray, J. B., and T. E. Cadell, 1962. Early behavior patterns in rhesus monkeys. J. Comp. Physiol. Psychol. 55:350–357. Nelson, C. A., 2001. The development and neural basis of face recognition. Infant Child Dev. 10:3–18. Newman, J. D., and J. Bachevalier, 1997. Neonatal ablations of the amygdala and inferior temporal cortex alter the vocal response to social separation in rhesus macaques. Brain Res. 758(1–2):180–186. Nimchinsky, E. A., E. Gilissen, J. M. Allman, D. P. Perl, J. M. Erwin, and P. R. Hof, 1999. A neuronal morphologic type unique to humans and great apes. Proc. Natl. Acad. Sci. USA 96(9):5268–5273. Oberman, L. M., E. M. Hubbard, J. P. McCleery, E. L. Altschuler, V. S. Ramachandran, and J. A. Pineda, 2005. EEG evidence for mirror neuron dysfunction in autism spectrum disorders. Brain Res. Cogn. Brain Res. 24(2):190–198. O’Doherty, J., J. Winston, H. Critchley, D. Perrett, D. M. Burt, and R. J. Dolan, 2003. Beauty in a smile: The role of medial orbitofrontal cortex in facial attractiveness. Neuropsychologia 41(2):147–155. Ongur, D., and J. L. Price, 2000. The organization of networks within the orbital and medial prefrontal cortex of rats, monkeys and humans. Cerebral Cortex 10(3):206–219.

bauman and amaral: neurodevelopment of social cognition

183

Parker, S. W., and C. A. Nelson, 2005. An event-related potential study of the impact of institutional rearing on face recognition. Dev. Psychopathol. 17(3):621–639. Parr, L. A., B. M. Waller, and J. Fugate, 2005. Emotional communication in primates: Implications for neurobiology. Curr. Opin. Neurobiol. 15(6):716–720. Pelphrey, K. A., J. D. Singerman, T. Allison, and G. McCarthy, 2003. Brain activation evoked by perception of gaze shifts: The influence of context. Neuropsychologia 41(2): 156–170. Pelphrey, K. A., R. J. Viola, and G. McCarthy, 2004. When strangers pass: Processing of mutual and averted social gaze in the superior temporal sulcus. Psychol. Sci. 15(9):598–603. Perrett, D. I., M. H. Harries, R. Bevan, S. Thomas, P. J. Benson, A. J. Mistlin, et al., 1989. Frameworks of analysis for the neural representation of animate objects and actions. J. Exp. Biol. 146:87–113. Perrett, D. I., J. K. Hietanen, M. W. Oram, and P. J. Benson, 1992. Organization and functions of cells responsive to faces in the temporal cortex. Philos. Trans. R. Soc. London [Biol.] 335(1273): 23–30. Perrett, D. I., E. T. Rolls, and W. Caan, 1982. Visual neurones responsive to faces in the monkey temporal cortex. Exp. Brain Res. 47(3):329–342. Perrett, D. I., P. A. Smith, D. D. Potter, A. J. Mistlin, A. S. Head, A. D. Milner, et al., 1984. Neurones responsive to faces in the temporal cortex: Studies of functional organization, sensitivity to identity and relation to perception. Hum. Neurobiol. 3(4):197–208. Perrett, D. I., P. A. Smith, A. J. Mistlin, A. J. Chitty, A. S. Head, D. D. Potter, et al., 1985a. Visual analysis of body movements by neurones in the temporal cortex of the macaque monkey: A preliminary report. Behav. Brain Res. 16(2–3):153–170. Perrett, D. I., P. A. Smith, D. D. Potter, A. J. Mistlin, A. S. Head, A. D. Milner, et al., 1985b. Visual cells in the temporal cortex sensitive to face view and gaze direction. Proc. R. Soc. London [Biol.] 223(1232):293–317. Petrides, M., and D. Pandya, 1994. Comparative cytoarchitectonic analysis of the human and the macaque frontal cortex. Handbook of Neuropsychology, ed. F. Boller and J. Grafman, 9:17– 58. Amsterdam: Elsevier. Phelps, E. A., K. J. O’Connor, J. C. Gatenby, J. C. Gore, C. Grillon, and M. Davis, 2001. Activation of the left amygdala to a cognitive representation of fear. Nature Neurosci. 4(4): 437–441. Prather, M. D., and D. G. Amaral, 2000. The development and distribution of serotonergic fibers in the macaque monkey amygdala. Abstract Viewer and Itinerary Planner. Washington, DC: Society for Neuroscience, 2003. Online. Prather, M. D., P. Lavenex, M. L. Mauldin-Jourdain, W. A. Mason, J. P. Capitanio, S. P. Mendoza, S. P., et al., 2001. Increased social fear and decreased fear of objects in monkeys with neonatal amygdala lesions. Neuroscience 106(4): 653–658. Premack, D., and G. Woodruff, 1978. Chimpanzee problemsolving: A test for comprehension. Science 202(4367):532–535. Pribram, K. H., and J. F. Fulton, 1954. An experimental critique of the effects of anterior cingulate ablations in monkey. Brain 77(1):34–44. Puce, A., T. Allison, M. Asgari, J. C. Gore, and G. McCarthy, 1996. Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance imaging study. J. Neurosci. 16(16):5205–5215.

184

Puce, A., T. Allison, S. Bentin, J. C. Gore, and G. McCarthy, 1998. Temporal cortex activation in humans viewing eye and mouth movements. J. Neurosci. 18(6):2188–2199. Puce, A., and D. Perrett, 2003. Electrophysiology and brain imaging of biological motion. Philos. Trans. R. Soc. London [Biol.] 358(1431):435–445. Raleigh, M. J., H. D. Steklis, F. R. Ervin, A. S. Kling, and M. T. McGuire, 1979. The effects of orbitofrontal lesions on the aggressive behavior of vervet monkeys (Cercopithecus aethiops sabaeus). Exp. Neurol. 66(1):158–168. Reader, S. M., and K. N. Laland, 2002. Social intelligence, innovation, and enhanced brain size in primates. Proc. Natl. Acad. Sci. USA 99(7):4436–4441. Rilling, J., D. Gutman, T. Zeh, G. Pagnoni, G. Berns, and C. Kilts, 2002. A neural basis for social cooperation. Neuron 35(2):395–405. Rilling, J. K., J. T. Winslow, and C. D. Kilts, 2004. The neural correlates of mate competition in dominant male rhesus macaques. Biol. Psychiatry 56(5):364–375. Rilling, J. K., J. T. Winslow, D. O’Brien, D. A. Gutman, J. M. Hoffman, and C. D. Kilts, 2001. Neural correlates of maternal separation in rhesus monkeys. Biol. Psychiatry 49(2):146–157. Rizzolatti, G., and L. Craighero, 2004. The mirror-neuron system. Annu. Rev. Neurosci. 27:169–192. Rizzolatti, G., L. Fadiga, V. Gallese, and L. Fogassi, 1996. Premotor cortex and the recognition of motor actions. Brain Res. Cogn. Brain Res. 3(2):131–141. Rizzolatti, G., L. Fogassi, and V. Gallese, 2001. Neurophysiological mechanisms underlying the understanding and imitation of action. Nature Rev. Neurosci. 2(9):661–670. Rodman, H. R., 1994. Development of inferior temporal cortex in the monkey. Cerebral Cortex 4(5):484–498. Rodman, H. R., S. P. Scalaidhe, and C. G. Gross, 1993. Response properties of neurons in temporal cortical visual areas of infant monkeys. J. Neurophysiol. 70(3):1115–1136. Rodman, H. R., J. P. Skelly, and C. G. Gross, 1991. Stimulus selectivity and state dependence of activity in inferior temporal cortex of infant monkeys. Proc. Natl. Acad. Sci. USA 88(17): 7572–7575. Rolls, E. T., 1984. Neurons in the cortex of the temporal lobe and in the amygdala of the monkey with responses selective for faces. Hum. Neurobiol. 3:209–222. Rolls, E. T., 1996. The orbitofrontal cortex. Philos. Trans. R. Soc. London [Biol.] 351(1346):1433–1443; discussion 1443–1444. Rolls, E. T., 2004. The functions of the orbitofrontal cortex. Brain Cogn. 55(1):11–29. Rolls, E. T., J. Hornak, D. Wade, and J. McGrath, 1994. Emotion-related learning in patients with social and emotional changes associated with frontal lobe damage. J. Neurol. Neurosurg. Psychiatry 57(12):1518–1524. Rosvold, H. E., A. F. Mirsky, and K. H. Pribram, 1954. Influence of amygdalectomy on social behavior in monkeys. J. Comp. Physiol. Psychol. 47:173–178. Rudebeck, P. H., M. J. Buckley, M. E. Walton, and M. F. Rushworth, 2006. A role for the macaque anterior cingulate gyrus in social valuation. Science 313(5791):1310–1312. Ruppenthal, G. C., M. K. Harlow, C. D. Eisele, H. F. Harlow, and S. J. Suomi, 1974. Development of peer interactions of monkeys reared in a nuclear-family environment. Child Dev. 45(3):670–682. Sackett, G. P., 1966. Monkeys reared in isolation with pictures as visual input: Evidence for an innate releasing mechanism. Science 154(3755):1468–1473.

fundamentals of developmental neurobiology

Sade, D. S., 1967. Determinants of dominance in a group of free-ranging rhesus monkeys. Social Communication among Primates, ed. S. A. Altmann, 99–114. Chicago: University of Chicago Press. Saunders, R. C., B. S. Kolachana, J. Bachevalier, and D. R. Weinberger, 1998. Neonatal lesions of the medial temporal lobe disrupt prefrontal cortical regulation of striatal dopamine. Nature 393(6681):169–171. Saxe, R., 2006. Uniquely human social cognition. Curr. Opin. Neurobiol. 16(2):235–239. Saxe, R., and N. Kanwisher, 2003. People thinking about thinking people: The role of the temporo-parietal junction in “theory of mind.” NeuroImage 19(4):1835–1842. Saxe, R., D. K. Xiao, G. Kovacs, D. I. Perrett, and N. Kanwisher, 2004. A region of right posterior superior temporal sulcus responds to observed intentional actions. Neuropsychologia 42(11):1435–1446. Schultz, J., H. Imamizu, M. Kawato, and C. D. Frith, 2004. Activation of the human superior temporal gyrus during observation of goal attribution by intentional objects. J. Cogn. Neurosci. 16(10):1695–1705. Schultz, R. T., D. J. Grelotti, A. Klin, J. Kleinman, C. Van der Gaag, R. Marois, et al., 2003. The role of the fusiform face area in social cognition: Implications for the pathobiology of autism. Philos. Trans. R. Soc. London [Biol.] 358(1430):415–427. Schumann, C. M., J. Hamstra, B. L. Goodlin-Jones, L. J. Lotspeich, H. Kwon, M. H. Buonocore, et al., 2004. The amygdala is enlarged in children but not adolescents with autism; the hippocampus is enlarged at all ages. J. Neurosci. 24(28):6392– 6401. Semendeferi, K., A. Lu, N. Schenker, and H. Damasio, 2002. Humans and great apes share a large frontal cortex. Nature Neurosci. 5(3):272–276. Sereno, M. I., and R. B. Tootell, 2005. From monkeys to humans: What do we now know about brain homologies? Curr. Opin. Neurobiol. 15(2):135–144. Sergent, J., and J. L. Signoret, 1992. Functional and anatomical decomposition of face processing: Evidence from prosopagnosia and PET study of normal subjects. Philos. Trans. R. Soc. London [Biol.] 335(1273):55–61; discussion 61–62. Singer, T., 2006. The neuronal basis and ontogeny of empathy and mind reading: Review of literature and implications for future research. Neurosci. Biobehav. Rev. 30(6):855–863. Somerville, L. H., T. F. Heatherton, and W. M. Kelley, 2006. Anterior cingulate cortex responds differentially to expectancy violation and social rejection. Nature Neurosci. 9(8):1007– 1008. Spence, S. A., M. D. Hunter, T. F. Farrow, R. D. Green, D. H. Leung, C. J. Hughes, et al., 2004. A cognitive neurobiological account of deception: Evidence from functional neuroimaging. Philos. Trans. R. Soc. London [Biol.] 359(1451): 1755–1762. Squire, L. R., C. E. Stark, and R. E. Clark, 2004. The medial temporal lobe. Annu. Rev. Neurosci. 27:279–306. Suomi, S. J., 1997. Early determinants of behaviour: Evidence from primate studies. Br. Med. Bull. 53(1):170–184. Suomi, S. J., 1999. Attachment in rhesus monkeys. Handbook of Attachment: Theory, Research, and Clinical Applications, ed. J. Cassidy and P. R. Shaver, 181–197. New York: Guilford Press. Suzuki, W. A., and D. G. Amaral, 2004. Functional neuroanatomy of the medial temporal lobe memory system. Cortex 40(1): 220–222.

Thierry, B., 1985a. Patterns of agonistic interactions in three species of macaque (Macaca mulatta, M. fascicularis, M. tonkeana). Aggressive Behav. 11:223–233. Thierry, B., 1985b. Social development in three species of macaque (Macaca mulatta, M. fascicularis, M. tonkeana): A preliminary report on the first ten weeks of life. Behav. Process. 11:89– 95. Thomas, K. M., W. C. Drevets, P. J. Whalen, C. H. Eccard, R. E. Dahl, N. D. Ryan, et al., 2001. Amygdala response to facial expressions in children and adults. Biol. Psychiatry 49(4): 309–316. Thompson, C. I., 1968. Social development of infant rhesus monkeys following amygdaloid lesions, University of Wisconsin. Thompson, C. I., 1981. Long-term behavioral development of rhesus monkeys after amygdalectomy in infancy. The Amygdaloid Complex, ed. Y. Ben-Ari, 259–270. Amsterdam: Elsevier. Thompson, C. I., R. M. Bergland, and J. T. Towfighi, 1977. Social and nonsocial behaviors of adult rhesus monkeys after amygdalectomy in infancy or adulthood. J. Comp. Physiol. Psychol. 91(3):533–548. Thompson, C. I., J. S. Schwartzbaum, and H. F. Harlow, 1969. Development of social fear after amygdalectomy in infant rhesus monkeys. Physiol. Behav. 4(2):249–254. Tomasello, M., J. Call, and B. Hare, 2003. Chimpanzees understand psychological states—The question is which ones and to what extent. Trends Cogn. Sci. 7(4):153–156. Tsao, D. Y., W. A. Freiwald, T. A. Knutsen, J. B. Mandeville, and R. B. Tootell, 2003. Faces and objects in macaque cerebral cortex. Nature Neurosci. 6(9):989–995. Tsao, D. Y., W. A. Freiwald, R. B. Tootell, and M. S. Livingstone, 2006. A cortical region consisting entirely of face-selective cells. Science 311(5761):670–674. Tzourio-Mazoyer, N., S. De Schonen, F. Crivello, B. Reutter, Y. Aujard, and B. Mazoyer, 2002. Neural correlates of woman face processing by 2-month-old infants. NeuroImage 15(2):454– 461. Vogeley, K., P. Bussfeld, A. Newen, S. Herrmann, F. Happe, P. Falkai, et al., 2001. Mind reading: Neural mechanisms of theory of mind and self-perspective. NeuroImage 14(1, part 1): 170–181. Vogt, B. A., D. M. Finch, and C. R. Olson, 1992. Functional heterogeneity in cingulate cortex: The anterior executive and posterior evaluative regions. Cerebral Cortex 2(6):435– 443. Vogt, B. A., E. A. Nimchinsky, L. J. Vogt, and P. R. Hof, 1995. Human cingulate cortex: Surface features, flat maps, and cytoarchitecture. J. Comp. Neurol. 359(3):490–506. Vogt, B. A., L. Vogt, N. B. Farber, and G. Bush, 2005. Architecture and neuro-cytology of monkey cingulate gyrus. J. Comp. Neurol. 485(3):218–239. Wada, Y., and T. Yamamoto, 2001. Selective impairment of facial recognition due to a haematoma restricted to the right fusiform and lateral occipital region. J. Neurol. Neurosurg. Psychiatry 71(2):254–257. Ward, A. A., 1948. The cingular gyrus: Area 24. J. Neurophysiol. 11:13–23. Watson, K. K., T. K. Jones, and J. M. Allman, 2006. Dendritic architec-ture of the Von Economo neurons. Neuroscience 141(3): 1107–1112. Webster, M. J., J. Bachevalier, and L. G. Ungerleider, 1994. Connections of inferior temporal areas TEO and TE with parietal and frontal cortex in macaque monkeys. Cerebral Cortex 4(5):470–483.

bauman and amaral: neurodevelopment of social cognition

185

Webster, M. J., L. G. Ungerleider, and J. Bachevalier, 1991. Connections of inferior temporal areas TE and TEO with medial temporal-lobe structures in infant and adult monkeys. J. Neurosci. 11(4):1095–1116. Wicker, B., C. Keysers, J. Plailly, J. P. Royet, V. Gallese, and G. Rizzolatti, 2003. Both of us disgusted in My insula: The common neural basis of seeing and feeling disgust. Neuron 40(3):655–664. Winslow, J. T., P. L. Noble, C. K. Lyons, S. M. Sterk, and T. R. Insel, 2003. Rearing effects on cerebrospinal fluid oxytocin

186

concentration and social buffering in rhesus monkeys. Neuropsychopharmacology 28(5):910–918. Winston, J. S., B. A. Strange, J. O’Doherty, and R. J. Dolan, 2002. Automatic and intentional brain responses during evaluation of trustworthiness of faces. Nature Neurosci. 5(3):277–283. Wrangham, R. W., 1980. An ecological model of female-bonded primate groups. Behaviour 75:262–300. Young, L. J., M. M. Lim, B. Gingrich, and T. R. Insel, 2001. Cellular mechanisms of social attachment. Horm. Behav. 40(2):133–138.

fundamentals of developmental neurobiology

12

Pre- and Postnatal Morphological Development of the Human Hippocampal Formation LÁSZLÓ SERESS AND HAJNALKA ÁBRAHÁM

The hippocampal formation plays an important role in the process of memory formation both in nonprimate mammals and in primates, including the human. This chapter concentrates on human studies; therefore, experimentation is excluded, and the developmental processes are presented in a descriptive manner by monitoring the formation of mature structures from simpler, undifferentiated stages. Among the fundamental issues regarding development, we discuss neuronal cell formation, cell death, cell migration, and the morphological and neurochemical development of neuronal types. Since axonal pathfinding and synapse formation are very difficult to verify in preand postnatal human tissue, information about these developmentally important issues will not be presented. The possible influence of premature birth on cell formation, cell migration, and neurochemical maturation will also be discussed.

Germinal matrices and neuronal cell formation in the hippocampal formation The developmental pattern of cell formation is similar in most regions of the central nervous system of various species. However, in a few areas, such as the hippocampal formation, substantial differences are seen between rodents and primates, in addition to a certain level of resemblance. In all mammalian species, the germinal matrices (the sites of cell proliferation) are comparable. Pyramidal cells of Ammon’s horn are formed in the ventricular zone immediately under the ventricular wall. Significant difference can be seen between the archi- and neocortical ventricular germinal matrices. In the archicortex only the ventricular zone exists, whereas in all areas of the neocortex a wide subventricular zone participates in cell proliferation. Principal cells of the dentate gyrus, the granule cells, also start to proliferate in the ventricular zone. However, laterborn granule cells are generated in a secondary germinal matrix that is located in the hilus of the dentate gyrus. The formation of this secondary matrix will be discussed later.

In rodents, principal cells of Ammon’s horn, the pyramidal cells, are generated in the second half of the embryonic period, and granule cells of the dentate gyrus start to be formed only a few days before birth. As a consequence, all pyramidal cells are formed prenatally, whereas 85 percent of the dentate granule cells are formed postnatally (Angevine, 1975; Bayer, 1980). In contrast, neurogenesis in the primate hippocampal formation takes place relatively early in prenatal development. In rhesus monkeys, the first neurons appear almost simultaneously in the different subregions of the hippocampal formation, from the entorhinal cortex to the dentate gyrus, between embryonic days 36 and 38 (Rakic and Nowakowski, 1981). Except for the dentate gyrus, cell formation ceases during the first half of pregnancy, between embryonic days 62 and 65. Granule cell formation lasts until the end of the first postnatal month, with only 15 percent of the granule cells being formed postnatally (Rakic and Nowakowski, 1981). Granule cell formation has also been found in small numbers in the dentate gyrus of adult monkeys (Kornack and Rakic, 1999). Classic descriptions of the developing hippocampal fissure are centered on the morphological changes of the cytoarchitectonics of the human hippocampal formation from the time it is recognizable using the conventional histological stains such as toluidin blue, erythrosine, and hematoxylineosin (Hines, 1922; Humphrey, 1967). Previously, appropriate cell proliferation markers were not available; therefore, only recent studies addressed the question of local cell formation in the germinal layers of the human cerebral cortex as well as in different areas of the hippocampal formation (Zaidel, 1999; Seress et al., 2001; Tiu, Chan, and Yew, 2004; Curtis et al., 2005). Subfields of the hippocampal formation, such as the entorhinal cortex, subiculum, and Ammon’s horn, are discernable at the 10th gestational week (GW), whereas the dentate granule cell layer appears around the 11.5th GW (Humphrey, 1967). At the 25th GW the cytoarchitectonic characteristics of all divisions of the hippocampal formation are similar to what is observed in adults (Arnold and Trojanowski, 1996a; Humphrey, 1967).

187

The human brains used in our study (table 12.1) were obtained from autopsy (24 GW old or older) and from legal abortions (14–22 GW). The gestational age is based on postovulation time and on somatic measurements (crown-rump length). For the description of the hippocampal formation the suggested terminology of Amaral (1990) is followed. The term “hippocampal formation” includes the dentate gyrus, Ammon’s horn, subicular subregions (subiculum,

presubiculum, parasubiculum), and the entorhinal cortex. In our study of cell formation, we used the MIB-1 clone of the Ki-67 cell proliferation marker that is a commercially available monoclonal antibody (DAKO A/S, Glostrup, Denmark) and is widely used in conventional histopathology (Verheijen, Kuijpers, Schlingemann et al., 1989; Verheijen, Kuijpers, van Driel et al., 1989; Gerdes et al., 1992; Rose et al., 1994).

Table 12.1 Gender and clinical diagnosis verified by autopsy of the cases used in this study Case Number Gender Age Diagnosis 1 Female 14 gestational weeks Legal abortion (maternal disorder) 2 Male 15 weeks Spontaneous abortion 3 Male 16 weeks Legal abortion (maternal disorder) 4 Female 16 weeks Legal abortion (maternal disorder) 5 Female 17 weeks Spontaneous abortion 6 Female 18 weeks Spontaneous abortion 7 Female 20 weeks Legal abortion (maternal Hodgkin disease) 8 Female 22 weeks Spontaneous abortion 9 Female 24 weeks IRDS 10 Female 28 weeks IRDS, CHD 11 Female 30 weeks IRDS 12 Male 32 weeks Sepsis, pneumonia, CHD 13 Female 34 weeks Pneumoniua, IRDS 14 Female 36 weeks IRDS, CHD 15 Female 38 weeks CHD, IRDS 16 Male 38 weeks Esophageal atresia, pneumonia 17 Male 39 weeks Respiratory distress, asphyxia 18 Male 39 weeks Pneumonia, asphyxia 19 Female 40 weeks Respiratory distress 20 Male 40 weeks CHD 21 Female 1 postnatal week SIDS 22 Male 1 week BPD 23 Female 1 week CHD 24 Male 1 month Pneumothorax, pneumonia 25 Male 2 months Sepsis 26 Male 3 months Pneumonias, muscular distrophy 27 Male 3 months Leukemia 28 Female 3 months CHD 29 Male 5 months Respiratory distress, asphyxia 30 Female 5 months Agenesis of pulmonary arteries 1 31 Male 8/2 months Pneumonia 32 Female 11 months Ileus 33 Female 2 years Pneumonia 34 Female 8 years Leukemia 35 Female 10 years Leukemia 36 Male 47 years Heart attack Abbreviations: BPD, bronchopulmonary dysplasia; CHD, congenital heart disease; IRDS, infant respiratory distress syndrome; SIDS, sudden infant death syndrome

188

fundamentals of developmental neurobiology

In the early descriptions of the hippocampal fissure, it was demonstrated that hippocampal formation and the future dentate gyrus form a straight line. The dentate gyrus is located at the peak that starts to bend away from the ventricular wall at the 12th GW (see figure 51 in Hines, 1922, and figure 8 in Humphrey, 1967). At this age, the ventricular germinal zone along the hippocampal formation is a continuous layer where immature cells are formed and move to the pyramidal layer of Ammon’s horn and to the granule cell layer of the dentate gyrus (see the cell-dense future stratum oriens and hilar region in contrast to the practically cell-free future stratum radiatum in figure 8, Humphrey, 1967). With the aid of the mitotic marker MIB-1 it has been shown that both germinative matrices, the ventricular zone and hilus, contain a large number of proliferating cells as

early as the 14th GW (the earliest age examined in our study). At the 14th GW the hippocampal ventricular zone is relatively thin and does not include a subventricular germinal layer. In contrast, the temporal neocortical germinal layer includes a thick ventricular zone and wide subventricular zone displaying high mitotic activity. At the age of 14 GW, the dentate gyrus is already bent toward the CA1 area of Ammon’s horn and is localized far from the ventricular germinal layer (figure 12.1A). Therefore, at this age, similarly to in the 15- and 16-week-old fetuses (figure 12.2A), the continuous germinative zone from the ventricular wall to the dentate gyrus is not outlined as a straight line, but a stream of MIB-1-positive cells along the pyramidal cell layer of the CA3 area and below the hilar region indicates the presence of a germinal layer. In the 16-week-old fetus, the ventricular

Figure 12.1 Changes of the cytoarchitectonics of the fetal hippocampal formation. Camera lucida drawings of the hippocampal formation in (A) 14-week-old, (B) 18-week-old, (C) 20-week-old,

and (D) 22-week-old fetuses. CA1–3, subfields of Ammon’s horn; h, hilus of the dentate gyrus; p, pyramidal cell layer. Arrows point to the hippocampal fissure. Calibration bar, 500 μm.

seress and ábrahám: pre- and postnatal human hippocampal formation

189

Figure 12.2 Cell formation in the hippocampal formation of a 16-week-old fetus. Photomicrographs of MIB-1-labeled cells in cresyl-violet–counterstained coronal sections of the hippocampal formation of a 16-week-old fetus. (A) The ventricular zone (vz) along the CA1 area of Ammon’s horn and the region below the dentate gyrus (DG) and CA3 area is fully packed with labeled cells. There are far fewer labeled cells in the hilus (h) below the granule

cell layer (g). The area marked with a single asterisk is shown with higher magnification in B, whereas the region marked with double asterisks is shown in C. Arrows in A and C point to the section of the ventricular wall, where the proliferative germinal layer terminates. Calibration bars, 250 μm for A and 75 μm for B and C. (See plate 19.)

germinal layer terminates at that zone of the ventricular wall where the CA3 pyramidal layer (CA3/CA2 border) of Ammon’s horn bends away from the ventricular surface (figure 12.2B,C). Groups of dispersed MIB-1-positive cells occur close to the ventricular wall, but the major stream of the proliferating cells locates alongside the CA3 pyramidal cells (figure 12.2A). In addition, large groups of proliferating cells occur in the hilus, below the suprapyramidal blade of the dentate gyrus (figure 12.3A). Six weeks later, in a 22week-old fetus, the proliferating cells disappear from the future stratum oriens of the CA3 area of Ammon’s horn, and groups of MIB-1 positive cells are only visible in the hilar region (figure 12.3B). (See also plates 19 and 20.) These developmental changes of the cytoarchitectonics (figure 12.1B–D) and the location of the groups of MIB-1positive proliferating cells indicate that the formation of the hilar secondary germinal matrix can be explained with the following developmental events. The hilar region and the dentate gyrus originally form the peak of the growing hippocampal formation that starts to bend at the 12th GW, probably because of an increasing number of fibers in the fimbria-fornix that pass between Ammon’s horn and the ventricular surface. The growing axonal bundle of afferent and efferent fibers pushes the dentate gyrus away from the ventricular wall and increases the distance between the ventricular germinal layer and the dentate gyrus. The passing fibers separate the stream of migrating cells from the ventricular wall, and the originally continuous ventricular germinal layer is disrupted at the zone where the fimbria-fornix starts from the CA3 area (figure 12.2A). The ventricular germinal zone is becoming continuously thinner after the 16th week and progressively disappears along the CA1–3 areas. In the case of the earlier-maturing CA3 area, only remnants of the ventricular germinal zone remain by the 20th gestational week (figure 12.1B). In the CA1 area, the ventricular germinal zone persists longer (figure 12.1C,D). In the 22-week-old fetus the ventricular germinal layer along the CA1 area still contains proliferating cells, although fewer in number than in the adjacent neocortical germinal layer (figure 12.4A–D and plate 21). In the 28-week-old fetus the ventricular germinal layer of the hippocampal formation is thin and only occasionally contains dividing cells. A similar case can be found in the full-term newborn, where the ventricular germinal layer contains 5–6 layers of loosely packed cells and rarely displays MIB-1 immunoreactivity (figure 12.4D). The adjacent subventricular zone of the temporal neocortex is thicker and still contains a few proliferating cells in the ventricular zone—and even more in the subventricular zone (figure 12.4D). At 3 months of age, the ventricular germinal layer along the hippocampal formation completely disappears, whereas in the temporal neocortex the subventricular germinal layer still contains scattered groups of MIB-1-labeled cells. The temporal cortical subventricular

zone is still recognizable in the 5-month-old child and disappears by 1 year of age (Seress, 2001). In the hilus of the dentate gyrus a high rate of cell proliferation is observed between the 16th and 22nd postnatal weeks (figure 12.3A,B). The number of dividing cells decreases rapidly after the 24th GW (Seress, 2001), but in a low percent (>0.1%) MIB-1 positive cells can be observed during the first six postnatal months (figure 12.5C). Only a portion of the labeled cells might be neuronal precursors, whereas the others are glial and endothelial cells (Seress et al., 2001; Ábrahám et al., 2004). The MIB-1 antibody against the Ki-67 nuclear protein labels cells in G1-S-G2-M phases of cell cycle independently of their nature (neurons, glial cells, endothelial cells). (See also plate 22.) Since the proteins that help to identify cells of the central nervous system appear postmitotically, the exact determination of the fate of MIB-1-labeled cells is not possible. Therefore, it has to be considered that not all proliferating cells are neurons, and a proportion of them are probably glial or endothelial cells. We suggest that before and around midgestation, most of the MIB-1-positive cells in the germinative zones become neurons, and parallel with the decreasing proliferative rate the proportion of the dividing neuronal precursors is decreased. It also has to be noted that neuronal precursors are exclusively formed in the germinal zones, and, therefore, proliferating cells in layers other than the germinal zones, according to our present knowledge, are all glial and endothelial cells. Recent observations emphasize long-lasting granule cell proliferation in the dentate gyrus of the human hippocampus (Eriksson et al., 1998; Roy et al., 2000). However, in surgically removed adult hippocampi we failed to detect more than a few dividing cells (>0.01%), independently of the cause of operation—for example, epilepsy or a benign tumor that did not invade the hippocampus itself. We only occasionally found MIB-1-labeled granule cells in the dentate gyrus of children who were older than 1 year (figure 12.5D), suggesting that a few neurons may preserve their capability of proliferation, but the frequency of such cells is extremely low in humans. This statement may appear negativistic in light of previously published data. However, if one carefully compares the data across studies it will be evident that experimental results in mice and rats support the idea of adult neuro-genesis (Kuhn, Dickinson-Anson, and Gage, 1996; Gould and Gross, 2002), but statements about results in monkeys are very carefully formulated (Rakic, 1998; Kornack and Rakic, 1999). In addition, only one publication indicates granule cell formation in the adult human dentate gyrus (Eriksson et al., 1998), and the other published reports are based on in vitro studies (Murell et al., 1996; Roy et al., 2000). Similarly, only one study suggested neurogenesis for the adult primate neocortex (Gould, Reeves, Graziano, et al., 1999), but it is clear now that neurogenesis in the primate neocortex is an

seress and ábrahám: pre- and postnatal human hippocampal formation

191

Figure 12.3 Cell formation in the dentate gyrus of 16- and 22week-old fetuses. Photomicrographs of MIB-1-labeled and cresylviolet–counterstained coronal sections of the dentate gyrus of (A) 16-week-old and (B) 22-week-old fetuses. Equally large numbers of

192

MIB-1-labeled cells appear in the hilus (h), below the granule cell layer (g), while only a few labeled cells are in the molecular layer (m) and in Ammon’s horn (CA3). Calibration bars, 75 μm. (See plate 20.)

fundamentals of developmental neurobiology

Figure 12.4 Cell formation in the ventricular germinal layers. Photomicrographs of cresyl-violet–counterstained sections (A,B,C) from 22-week-old fetus and (D) from newborn child. In the hippocampal ventricular zone (v) there are only a few MIB-1-labeled cells, whereas both the ventrical zone (v) and the subventricular zone (SVZ) of the temporal cortex is fully packed with MIB-1positive cells. The area marked with an asterisk is shown with

higher magnification from an adjacent section in B, whereas a similar area marked with double asterisks is shown in C. In the newborn child MIB-1-positive proliferating cells are not visible in the hippocampal ventricular germinal zone (v) and are sparse in the zone of the temporal neocortex (D). Calibration bars, 100 μm for A and 50 μm for B, C, and D. (See plate 21.)

seress and ábrahám: pre- and postnatal human hippocampal formation

193

Figure 12.5 Postnatal cell proliferation and migration. Photomicrographs of coronal sections showing the border between the granule cell layer (g) and the hilus (h) of the dentate gyrus. The subgranular zone contains large numbers of migrating cells (a few pointed out by arrows) in a newborn (A), whereas such cells are absent in a one-year-old child (B). In a one-year-old child the

neurons in the granule cell layer (g) are easily distinguishable from astroglial (arrows) and oligodendroglial (open arrow) cells (B). In the newborn child a few MIB-1-positive cells are in the hilus (arrows) (C). Labeled cells (arrow) are rare in the hilus of the oneyear-old child (D). Calibration bars, 20 μm. (See plate 22.)

early embryonic phenomenon (Kornack and Rakic, 2001; Rakic, 2002; Bhardwaj et al., 2006). Moreover, data concerning neurogenesis in the epileptic dentate gyrus are also controversial in rodents and primates, because both enhanced and reduced neurogenesis was found in epileptic rodent models (Parent et al., 1997; McCabe et al., 2001), but neurogenesis was not changed in the resected hippocampi of epileptic patients (Fahrner et al., 2007; Heinrich et al., 2006). Similar results were found in our laboratory, where no indication of enhanced neurogenesis was found in the hippocampi of epileptic patients (Seress et al., 2001). Very limited possibilities for neuronal repair were found in humans after stroke and irradiation (Price, 2001; Snyder and Park, 2002; Arvidsson et al., 2002), although lesion-induced neurogenesis was reported in rodents (Gould and Tanapat, 1997), and a recent report indicates that neurogenesis may occur in the human brain after stroke (Jin et al., 2006). In mental disorders, reports about neurogenesis are controversial, because stem-cell proliferation was found to be decreased in schizophrenia but not in depression, and in Alzheimer’s disease glial and endothelial cell proliferation but not neurogenesis was found (Boekhoorn, Joels, and Lucassen, 2006; Reif et al., 2006). In conclusion, very few data indicate that neurogenesis would occur in the adult human hippocampal formation or in the adult human brain in general.

Evidence of cell death Since cell formation is accompanied by cell death, we examined the frequency of occurrence of pyknotic cell nuclei, which indicate that cell death has occurred in the hippocampal formation. Pyknotic cell nuclei are inevitable morphological signs, and pyknosis was a reliable marker of acute cell death in experimental models (Seress, 1977), although recent apoptotic markers provide a more modern approach. However, in the postmortem human brain one may expect a variable number of apoptotic cells, especially because the sensitivity of that method is great. In those cases where other morphological signs (large perineuronal and pericapillary spaces, etc.) also indicated strong hypoxia, pyknotic nuclei were frequent. However, in the brains of fetuses and young infants, where postmortem delay was short and fixation was good, only a few pyknotic cells were visible inside the granule cell layer, and almost no pyramidal cells were pyknotic in Ammon’s horn. These results are in harmony with our previous results from newborn monkeys where pyknotic nuclei were rarely observed (approx. 0.3%) in the granule cell layer, although granule cells in monkeys are still formed postnatally (Rakic and Nowakowski, 1981). In conclusion, neuronal cell death is not a necessary requirement for the formation of cytoarchitectonic layers of Ammon’s horn and that of the dentate gyrus in humans, and probably the overwhelming majority of the newly formed neurons survive. This finding

is in obvious contradiction with findings in experimental animals that indicate that programmed death of neurons and glia is always found in the postnatal rat brain (Siman et al., 1999). In addition, continuous apoptotic cell death may explain why an increased number of granule cells of the dentate gyrus cannot be found in old rodents when compared with young adults, although it would be expected following a constant granule cell generation in adulthood (Biebl et al., 2000; Cameron and McKay, 2001; Dayer et al., 2003).

Cell migration in the hippocampal formation Postmitotic pyramidal cells and GABAergic interneurons find their final position in the developing cortex through two different migrational processes. Excitatory cells that originate from the ventricular/subventricular zones of the telencephalon migrate through two forms of radial migration: somal translocation and radial glia-guided migration (Rakic, 1971; Nadarajah and Parnavelas, 2002). Pyramidal cells build the cerebral cortex according to an inside-out migrational grandient that results in the positioning of earlygenerated neurons in the deeper layers (VI, V) and, younger, later-born neurons migrating through the deeper, alreadyformed layers and forming superficial layers (II, III). In rodents, virtually all cortical inhibitory local circuit neurons, including hippocampal interneurons, are generated in the ganglionic eminences and migrate to their destination by means of tangential migration and ventricle-guided migration (Anderson et al., 1997; Nadarajah and Parnavelas, 2002; Nadarajah et al., 2002). However, in humans, only 35 percent of the neocortical interneurons have been shown to originate from the ganglionic eminences, and 65 percent of GABAergic cells are formed in the neocortical ventricular and subventricular zones and migrate radially guided by glial processes (Letinic, Zoncu, and Rakic, 2002). Experimental studies in rodents and primates have shown that the ventricular zone underlying the hippocampal formation is the source of neurons (Bayer, 1980; Nowakowski and Rakic, 1981). Migrating principal cells in Ammon’s horn, subiculum, and entorhinal cortex bypass previously generated neurons on their way to the superficial limits of the developing cortical plate (Nowakowski and Rakic, 1981). The inside-out migration pattern is similar to that in neocortex. The exception is cell migration of the dentate gyrus, where the granule cell layer is formed according to an outside-in pattern (Bayer, 1980; Nowakowski and Rakic, 1981). The dentate gyrus receives neurons both from the ventricular zone and from the hilus, which is a separate proliferative zone after the 16–18th GW as indicated by a dynamics of cell formation that is very similar in the hilus and in the subventricular cortical layer (Bayer, 1980; Seress, 1977). In nonhuman primates, groups of small cells with thin cytoplasm and dark cell nuclei persist in the subgranular

seress and ábrahám: pre- and postnatal human hippocampal formation

195

zone of the dentate gyrus throughout the first postnatal year (Eckenhoff and Rakic, 1988). Since no neurons were labeled with 3H-thymidine-autoradiography after the first postnatal month, these cells had to be formed in the prenatal or early postnatal period. Not all of the undifferentiated, immaturelooking cells are positive for glial fibrillary acidic protein in the 6-month- and 1.5-year-old monkeys, indicating that these immature cells may be neurons and may differentiate into granule cells later in development (Eckenhoff and Rakic, 1988). Indeed, recent observations suggest that neurogenesis and cell migration occur in the dentate gyrus of adult monkeys, although the initiator factors of cell proliferation from these “dormant” progenitor cells are not clear (Gould, Reeves, Fallah, et al., 1999; Kornack and Rakic, 1999). Prenatally, in 14–22-week-old fetuses and in 24–30-weekold infants, a large number of migrating cells are visible below the granule cell layer of the dentate gyrus (figure 12.3A) as well as in the intermediate zone (future stratum oriens) between the ventricular wall and pyramidal cell layer. In 32-week-old and older infants, CA1–3 areas of Ammon’s horn lack migrating cells (Arnold and Trojanowski, 1996a). In contrast, large numbers of immature cells still persist in the subgranular zone of the dentate gyrus in older (32–36week-old) infants and even in neonates (Seress, 1992; Seress et al., 2001). In cresyl-violet–stained preparations, relatively large numbers of immature cells display a dark, elongatedovoid nucleus and a thin cytoplasmic rim in neonates (figure 12.5A). In a 5-month-old child, the proportion of immature cells is much lower, although clusters of dark, immature cells occur in the subgranular zone of the 8- or 11-month-old children. After the first postnatal year, the hilar border of the granule cell layer and the deep hilus are free of immature cells (figure 12.5B), and the cytoarchitectonic features of the dentate gyrus appear to be adultlike. This finding correlates well with the observation of an extremely low number of proliferating granule cells (>0.01%) of the dentate gyrus in children and in adults, suggesting that granule cell proliferation in adults may occur from localized pools of progenitor cells that may locate in islands along the longitudinal axis of the dentate gyrus. So far we were unable to detect such a pool in the human dentate gyrus, either in surgically removed hippocampi of adults or in autopsy material of children.

Cajal-Retzius cells in the developing hippocampal formation Precise regulatory mechanisms of cell migration are required for the development of a normally functioning cerebral cortex. One of the most studied regulatory pathways of neuronal migration is controlled by reelin. Reelin is an extracellular matrix glycoprotein expressed by several classes of different cells, particularly by large neurons of the developing marginal zone (MZ, future layer I) of the cerebral

196

cortex (D’Arcangelo et al., 1995). These neurons are called, after the first descriptions provided by Cajal and Retzius, Cajal-Retzius cells (Cajal, 1891; Retzius, 1893). Reelin secreted by these cells is responsible for the correct lamination of the cerebral cortex, which follows the inside-out migrational pattern. The importance of reelin is highlighted by the phenotype of a naturally occurring mouse-mutant reeler caused by a mutation of the reelin gene (D’Arcangelo et al., 1995). This mutation results in cytoarchitectonic abnormalities that are pronounced in the cerebral cortex and hippocampal formation, in addition to the cerebellum. In the absence of reelin, young neurons are unable to migrate through the layers of their predecessors, resulting in an altered neocortical structure. In the archicortical hippocampal formation the abnormalities are mainly restricted to the dentate gyrus, although the arrangement of pyramidal cells of the CA1 region is also affected (Stanfield and Cowan, 1979; Drakew et al., 2002). Instead of one single layer, pyramidal cells of CA1 form two distinct layers (Deller et al., 1999), while granule cells of the dentate gyrus are not arranged in one layer but are dispersed in the hilus with their dendrites oriented in all directions (Stanfield and Cowan, 1979; Drakew et al., 2002). Investigations of the reelin signaling pathway show that the protein plays a role in the radial glia-guided migration in the cerebral cortex, a finding that explains the cytoarchitectionical abnormalities found in the reeler mutant (for review see Lambert de Rouvroit and Goffinet, 2001). In the dentate gyrus, reelin has been shown to promote the differentiation and orientation of radial glia and, therefore, to direct the migration of granule cells (for review see Forster et al., 2006). In addition, in the cerebral cortex it functions as a stop signal for the migrating neurons as they reach the border of the MZ and most superficial layer of the cortical plate. Mutation of the reelin gene in humans results in a low or undetectable level of this extracellular protein and causes autosomal recessive lyssencephaly (Hong et al., 2000). In addition to the robust neocortical and cerebellar malformations found in this disease, the hippocampal formation appears flattened, lacking its normal folded shape and definable upper and lower blades, indicating the importance of the reelin in the development of the human archicortex. Moreover, abnormal reelin signaling in the hippocampus was reported to be associated with neurological and mental illnesses such as epilepsy, schizophrenia, bipolar disorders, depression, and autism (Bartlett et al., 2005; Fatemi, Earle, and McMenomy, 2000; Haas et al., 2002). Reelin-secreting Cajal-Retzius cells can be visualized using immunhistochemical methods (Ogawa et al., 1995; Meyer and Goffinet, 1998). Independently of the reelin secretion, most Cajal-Retzius cells express a calcium-binding protein, calretinin, and a few of them calbindin (Soriano et al., 1994; Weisenhorn, Prieto, and Celio, 1994; Ábrahám

fundamentals of developmental neurobiology

and Meyer, 2003). However, both reelin and calretinin are also produced by interneuronal populations that, in the hippocampal formation, make the clear identification and quantification difficult (Schiffmann, Bernier, and Goffinet, 1997; Hof et al., 1999). A nuclear transcription factor p73 that belongs to the family of the tumor-suppressor protein p53 was found to be expressed exclusively by Cajal-Retzius cells in the archi- and neocortex of both rodents and humans (Yang et al., 2000; Meyer et al., 2002; Ábrahám and Meyer, 2003). For the exact morphology of Cajal-Retzius cells we have to consult the drawings of Cajal (1891) and Retzius (1893), who used Golgi impregnation for the visualization of these cells in the developing cerebral neocortex. Since the method developed by Golgi allows the clear visualization of the whole cell with its processes, Ramón y Cajal, in his original work (1891), described and drew bipolar and fusiform large neurons with dendrites and filopodia-like appendages in the MZ of the fetal human cortex. In addition, Retzius (1893) drew large cells in the MZ, which display profuse dendritic and axonal arborization with dendritic side branches ramifying perpendicular to the main dendrites and running toward the pial surface. These perpendicular dendritic side branches endow these cells with a bizarre appearance that can never be observed in other neuronal cell types. Regarding the differences between the morphology of reelin-labeled and calretinin-labeled Cajal-Retzius cells, reelin immunohistochemistry reveals only somata and only the proximal parts of the main dendrites of these cells, whereas calretinin labels longer segments of dendrites as well as the axons of CajalRetzius cells. In the cerebral cortex, because of these unique morphological characteristics, Cajal-Retzius cells can be clearly identified even after reelin or calretinin immunohistochemistry that allows a description of morphological changes of Cajal-Retzius cells in different stages of cortical development (Meyer, Goffinet, and Fairen, 1999). Unfortunately, Cajal (1891) and Retzius (1893) described these cells only in the neocortex and not in the hippocampal formation. Being a selective marker of the Cajal-Retzius cells in the neocortex, p73 immunoreactivity indicates that reelincontaining early-generated cells of the hippocampal marginal zone can be considered as the archicortical equivalents of the neocortical Cajal-Retzius cells (Ábrahám, PerezGarcia, and Meyer, 2004). Compared to the neocortical Cajal-Retzius cells, they form a rather homogeneous population displaying bipolar or fusiform shapes. In most cases, reelin and calretinin immunohistochemistry visualizes neither the perpendicularly running dendritic side branches nor filopodia. However, Golgi impregnation reveals a large number of small dendritic appendages running mostly perpendicular to the main dendrites of hippocampal Cajal-Retzius cells (figure 12.6D). Although these processes are smaller and less pronounced than in the neocortex

(figure 12.7E,F), their arrangement and the localization of the cells exclude any morphological similarity with any other neuronal cell type. (See also plates 23 and 24.) Cajal-Retzius cells form an early neuronal population in the marginal zone of the developing fetal cerebral cortex (Meyer and Goffinet, 1998). They can already be found at the 7th–10th GW expressing p73 and reelin in the cerebral cortex including the hippocampal primordium, and their number and morphological diversities increase afterward (Ábrahám, Perez-Garcia, and Meyer, 2004). The proposed sites of origin of Cajal-Retzius cells, where they invade the marginal zone through tangential migration, are the retrobulbar basal forebrain and the cortex-choroid plexus boundary, the so-called cortical hem (Meyer and Wahle, 1999; Zecevic and Rakic, 2001; Meyer et al., 2002). Based on morphological characteristics and on spatial vicinity, it was suggested that Cajal-Retzius cells of the hippocampal formation are derived from the ventricular epithelium of the cortical hem (Ábrahám, Perez-Garcia, and Meyer, 2004), which is a putative signaling center of cortical patterning (Grove et al., 1998; Grove and Tole, 1999). Parallel with the folding and maturation of the hippocampus, the fibers of the fimbria fornicis grow between the choroid plexus and the ventricular zone of the hippocampus, shifting the birthplace of Cajal-Retzius cells near the dentate-fimbrial boundary. From the 15th–16th GW and onward, Cajal-Retzius cells proliferate at the dentate-fimbrial boundary often marked by an indentation of the lateral ventricle. Immunocytochemical investigation of dividing cells both in the cortical hem and dentate-fimbrial boundary using the Ki-67 cell proliferation marker showed that Cajal-Retzius cells start to express p73 after exiting from the cell cycle. Reelin immunostaining appears later, when the p73-immunoreactive cells approach the marginal zone. Their migrational route is located medially from the proposed migrational path of the granule cell precursors. Therefore, Cajal-Retzius cells appear in the outer region of marginal zone of the dentate gyrus (future stratum moleculare). Parallel with the decrease in number of proliferating precursors in the hippocampal ventricular zone between the 21st and 25th GW, the number of p73-positive cells is also decreasing at the dentate-fimbrial boundary. While in the early fetal stages, Cajal-Retzius cells are numerous in the marginal zone of Ammon’s horn and of the dentate anlage that corresponds to the suprapyramidal or dorsal blade of dentate gyrus described in rodents. Later-born Cajal-Retzius cells populate mostly the infrapyramidal (ventral) blade of the dentate gyrus, following in time the developmental gradient of Ammon’s horn and supra- and infrapyramidal blades of the dentate gyrus. Parallel with this regional shift, reelin-immunoreactive p73-negative cells that are morphologically identical to the reelin-positive hippocampal interneurons (Alcantara et al., 1998; Ábrahám and Meyer, 2003)

seress and ábrahám: pre- and postnatal human hippocampal formation

197

Figure 12.6 Cajal-Retzius cells in the developing hippocampal formation. (A) Calretinin-immunoreactive Cajal-Retzius cell, displaying characteristic filopodia (open arrows) at the hippocampal fissure of a newborn child. (B) Reelin-immunostained Cajal-Retzius cells (arrows) at the hippocampal fissure of a newborn child. (C) A large, bipolar calretinin-immunoreactive Cajal-Retzius cell with

long dendrites running along the hippocampal fissure in a 2-yearold child. (D) Photomontage of a Golgi-impregnated large, bipolar Cajal-Retzius type cell at the hippocampal fissure of a newborn child. Filopodium-like processes (open arrows) are on both the soma and the dendrites. Calibration bars, 25 μm. (See plate 23.)

Figure 12.7 Cajal-Retzius cells in layer I of the temporal neocortex. (A) Photomicrographs of calretinin-positive Cajal-Retzius cells (arrows) in the temporal neocortex of a newborn child. Arrowheads point to calretinin-immunoreactive axonal plexus deeper in layer I, which may correspond to the axonal plexus of the Cajal-Retzius cells as shown on (F). (B, C) Calretinin-positive Cajal-Retzius cell (arrow) of different morphology in 5-month-old infant. (D) Reelinimmunoreactive Cajal-Retzius cells (arrows) and reelin-positive interneuron (curved arrow) in a 3-month-old infant. (E, F) Golgiimpregnated Cajal-Retzius-type cells (arrows) in the temporal

cortex of a newborn child corresponding to the cells first described by Retzius. (E) One of the cells has a long dendrite that runs parallel with the pial surface displaying the characteristic side branches (open arrows) that can be observed on many drawings of Retzius. (F) The other cell shows a fuzzy cell body (arrow) with several dendritic branches that run in different directions. The thin side branches leave the main dendrites perpendicularly to them (open arrows). Arrowheads show axonal plexus of Cajal-Retzius cells in the deep layer I. Calibration bars, 25 μm for A–C and E, 20 μm for D, 50 μm for F. (See plate 24.)

start to appear in the inner marginal zone overlying the ammonic plate, while p73 reelin-immunoreactive cells predominate in the outer marginal zone. With further development, the number of reelin-immunoreactive interneurons increases in the strata lacunosum-moleculare and radiatum of Ammon’s horn. Cajal-Retzius cells form a predominant cell population in the developing archi- and neocortex approximately until midgestation. After that, their relative number continuously decreases, and they are reported to disappear from the cerebral cortex after cortical neurons have found their final position (Meyer et al., 1999). However, in newborns large numbers of Cajal-Retzius cells can be identified in the hippocampal formation (figure 12.6A,B,D). Their number gradually decreases with age, although many of them are visible in older infants (figure 12.6C) or adolescents. Moreover, a few Cajal-Retzius cells persist in the adult hippocampal formation and temporal neocortex (Meyer et al., 2002; Ábrahám and Meyer, 2003). In the postnatal archicortex a higher number of CajalRetzius cells can be observed than in the neocortex. Based on reelin immunostaining, it was demonstrated that only small, bipolar cells can be found in layer I of the neocortex after the migration period that terminates around the 30th gestational week (Meyer, Goffinet, and Fairen, 1999). However, in addition to Golgi impregnation (figure 12.7E,F), calretinin-immunohistochemistry also reveals large cells with bizarre morphology in the temporal neocortex of newborns (figure 12.7A), as well as later in the first few postnatal months (figure 12.7B,C). Similar large cells are reelin immunoreactive, indicating that they may also belong to the Cajal-Retzius cells (figure 12.7D). In most cases they are found in the vicinity of small reelin-containing interneurons (figure 12.7D). The prolonged formation of the granule cell layer of the dentate gyrus may explain why Cajal-Retzius cells persist postnatally (figure 12.6C,D) in the hippocampal formation. It might be hypothesized that their presence in the adult hippocampus would verify the proposed postnatal neurogenesis of the dentate granule cells. However, Cajal-Retzius type cells can also be found in the perinatal neocortex, and rarely in the mature neocortex, where neurogenesis and migration are not assumed. Therefore, the functional importance of Cajal-Retzius cells in the postnatal human cerebral cortex may be complex, and these cells may support possible plastic changes both in the neocortex and in the hippocampal formation. This theory is supported by the study of Del Rio and associates (2002) using in vitro slice culture of the mouse hippocampal formation to show that Cajal-Retzius cells promote the regeneration of entorhino-hippocampal fibers after transection of the perforant pathway. Although the identity of the axonal growth-promoting signals is still unknown, they are effective in young adult nervous tissue

200

with an inhibitory nonpermissive milieu for the regrowth of injured axons. In addition, Cajal-Retzius cells are found in large number in the human hippocampal formation of preschool children and adolescents, and their number significantly decreases after that age (Ábrahám and Meyer, 2003). Developmentally important events (as will be discussed later in this chapter), such as axonal growth, synapse formation, and reorganization, occur in the hippocampal formation in children that may require the contribution of Cajal-Retzius cells.

Development of excitatory and inhibitory neurons of the human hippocampus Granule cells, hilar mossy cells, and CA3 pyramidal cells of monkeys are in an advanced stage of development at birth (Seress and Ribak, 1995a, 1995b). In contrast, in humans, a large proportion of the principal cells are in an early stadium of dendritic and spine development at birth (Purpura, 1975; Seress, 1992; Seress and Mrzljak, 1992). However, fully matured granule cells displaying densely spiny dendrites and an axon that gives rise to several collaterals in the hilus can be found in neonates (Seress, 1992). In contrast, other granule cells display varicose, stubby, short, and spineless dendrites that terminate in growth cones (Seress, 1992). These latter granule cells are still growing and are supposed to be those that are formed in the perinatal period. Therefore, the diversity in maturation of granule cell development seen by Purpura (1975) in the 33-week-old fetus is still observable at birth. A few immature-looking granule cells are still seen in the 15-month-old child, suggesting that granule cells exhibit a prolonged period of cell proliferation and maturation (Seress, 1992). The long-lasting development of principal cells of the dentate gyrus can also be illustrated with the neurochemical maturation of granule cells. Calbindin, a calcium-binding protein, is a marker of granule cells and can be visualized with immunohistochemistry (Seress et al., 1993). During fetal development, granule cells start to express calbindin relatively early. At the 22nd GW a few granule cells of the dentate gyrus are already calbindin-positive (figure 12.8B). This first calbindin-positive cell group is found in the area of granule cell layer that locates between the CA3 and CA1 regions, and corresponds to the dorsal blade of the rodent dentate gyrus. Granule cells form the granule cell layer in an outside-in migrational pattern because older granule cells locate more superficially than the younger ones. Correspondingly, the granule cells locating closer to the molecular layer are calbindin positive, whereas the deeper locating cells at the hilar border are calbindin negative (figure 12.8B,C,D). At birth a large number of granule cells are calbindin positive, but the majority of granule cells of the ventral blade are calbindin negative (figure 12.8C). In addition, the granule

fundamentals of developmental neurobiology

Figure 12.8 Calbindin immunoreactivity in granule cells of the dentate gyrus and in the pyramidal cells of Ammon’s horn. (A) Pyramidal cells that are located closer to the ventricular germinal layer express calbindin at the 16th GW in the pyramidal cell layer of CA1 area. (B) At the 21st GW, virtually all pyramidal cells of the CA1 area are calbindin immunoreactive. In contrast, only the oldest granule cells are calbindin positive at the border of the granule cell (g) and molecular layers (m). (C) At term, in the dorsal blade (arrow) of the granule cell layer (g) most cells display strong immunoreactivity for calbindin, while in the ventral blade (open curved arrows) only cells in the outer part of the granule cell layer

(g) close to the molecular layer (m) are immunoreactive. In the hilus (h) of the dentate gyrus a few scattered calbindin-positive cells are probably interneurons, whereas in the pyramidal layer of the CA3 area all cells are immunonegative. (D) Calbindin-immunoreactive cells in granule cell layer (g) of the dentate gyrus in a 5-month-old child. Many of the cells display immunoreactive dendrites that run toward the molecular layer (m). In contrast, large numbers of cells with elongated cell nuclei (one pointed out by an arrow) are immunonegative at the hilar border (h). These are probably the newly generated, still-migrating granule cells. Calibration bars, 50 μm for A, 100 μm in B and C, 20 μm for D. (See plate 25.)

cells at the hilar border of the granule cell layer are calbindin negative both in the dorsal and the ventral blades (figure 12.8C). As late as the 5th postnatal month, the granule cells at the hilar border are still calbindin negative, indicating arrival of new granule cells from the hilus (figure 12.8D). (See also plate 25.) Formation of complex spines of hilar mossy cells occurs exclusively postnatally, because no mossy cells with thorny excrescences could be observed in humans at birth (Seress and Mrzljak, 1992). The first small thornlike excrescences appeared on human mossy cells by the third postnatal month. At the age of seven months thorny excrescences are frequent on mossy cells, but their number and size continue to increase up to the third year (Seress and Mrzljak, 1992). The thorny excrescences of the adult human mossy cells are much larger than the complex spines of mossy cells in monkeys (Frotscher et al., 1991). There is a similarly extended postnatal period of development of thorny excrescenses of CA3 pyramidal cells. A period of maximum dendritic growth for hippocampal pyramidal cells was observed between the 20th and 28th weeks of gestation (Purpura, 1975). Pyramidal cells of the CA3 area display only a few small spines (spicules) and filopodia on their dendrites in a 22-week-old fetus. In the 33-week-old fetus, the first thornlike excrescences appear on the dendrites of the CA3 pyramidal cells (Purpura, 1975). Therefore, the first complex spines appear 3–4 months earlier on the dendrites of CA3 pyramidal cells, which project to the CA1 region, than on the dendrites of mossy cells, which innervate the granule cells. CA1 pyramidal cells also show considerable postnatal morphological changes, although their neurochemical maturation starts early in fetal development. Pyramidal cells of the CA1 area express calbindin, similarly to the granule cells (Seress, Gulyás, and Freund, 1992). The first calbindin-immunoreactive cells in the CA1 pyramidal layer can be observed at the 16th GW (figure 12.8A). First, cells that locate closer to the ventricular zone are immunoreactive, indicating the inside-out migrational pattern of pyramidal cells. Although virtually all pyramidal cells are calbindin positive at birth, they have few basal dendritic branches and poorly developed side branches of the apical dendrites. A similar pattern has been found in neocortex, where the pyramidal cells of the newborn child display a few varicose, short basal dendrites and a few poorly developed side branches of the apical dendrite (Purpura, 1975; Seress, 2001). There are only a few spines on the apical and basal dendrites of pyramidal cells of the newborn infant, whereas the equivalent portions of the dendrites in the adult neocortex and hippocampus are fully covered with spines. In rodents, similar to the neocortical local circuit neurons, virtually all hippocampal interneurons are generated in the ganglionic eminences (Anderson et al., 1997). In humans, however, the possibility cannot be excluded that inhibitory

202

neurons originate from the hippocampal ventricular zone (Letinic, Zoncu, and Rakic, 2002). Interneurons are formed early, parallel with the proliferation of pyramidal cells of Ammon’s horn. However, at the site of proliferation they cannot be differentiated from pyramidal cells of Ammon’s horn because of the late expression of their neurochemical markers. Large numbers of GABAergic cells of the human hippocampal formation express calcium-binding proteins— calretinin, calbindin, or parvalbumin (Freund and Buzsáki, 1996). In addition, reelin-immunoreactive interneurons form a distinct interneuronal population, although reelin is partly colocalized with calretinin and calbindin (Schiffmann, Bernier, and Goffinet, 1997; Ábrahám and Meyer, 2003). In adults, calretinin-positive interneurons occur in high numbers at the border between the strata radiatum and lacunosummoleculare of the CA1 region. The dendritic orientation of these small bipolar cells is perpendicular to the hippocampal fissure. In addition, large multipolar calretinin-containing interneurons are also found in all layers of the hippocampal formation. As early as the 14th GW, reelin-positive non-Cajal-Retzius cells appear in the inner marginal zone of the hippocampal formation. These cells might form the first interneuronal population. Their number increases afterward, and they gradually display a mature phenotype. Calretininimmunoreactive interneurons are recognizable in all layers of Ammon’s horn at the 16th GW. At around midgestation (22GW) their characteristic location is prominent at the border of the strata lacusonum-moleculare and radiatum (figure 12.9B). In addition, their morphology corresponds to that found in adults, although at this age cells are smaller and display fewer and shorter dendritic branches. Parallel with the appearance of calretinin in the layers of Ammon’s horn, small immature cells express calretinin in the ventricular zone, below the ventricular epithelium. Their size, shape, and location indicate that they are young, postmitotic cells and that they might migrate from this position to CA1–3 areas (figure 12.9C). In fact, a population of calretininimmunoreactive cells with immature cell body and with one or two small dendritic processes can be seen in the future strata oriens, pyramidale, and radiatum of Ammon’s horn at the age of 16th–18th GW, indicating a probable migration of these cells to their final destination. It has to be mentioned, however, that a transient expression of calretinin in ventricular germinal cells cannot be excluded. Another calcium-binding protein, calbindin, expressed by granule cells of the dentate gyrus and pyramidal cells of Ammon’s horn, is also a neurochemical marker for a characteristic group of interneurons in the hilus and molecular layer of the dentate gyrus and in the strata radiatum and oriens of Ammon’s horn (Seress et al., 1993). Cabindinpositive interneurons initially appear in the fetal period, although slightly later than the calretinin-positive interneu-

fundamentals of developmental neurobiology

rons. At 16th GW only a few very immature immunoreactive cells can be observed. The first calbindin-immunoreactive cells with morphology and location identical to matured hippocampal interneurons are seen only at the 20th–22nd GW, both in Ammon’s horn and in the dentate gyrus (figure 12.9A). However, at birth, calbindin-positive interneurons still have an immature appearance with short and rarely branched dendritic trees. (See also plate 26.) The third calcium-binding protein, parvalbumin, is contained by interneurons providing axo-somatic or axo-axonic inhibition to glutamatergic principal cells (Freund and Buzsáki, 1996). Parvalbumin-positive cells are usually large and multipolar and are located in close proximity to the population of their target cells—for example, inside or at the borders of the pyramidal or granule cell layers. In contrast to the previous two populations of interneurons, parvalbumin expression starts late in development. In the late fetal period, parvalbumin-positive cells could not be found in the human hippocampal formation. At birth, in Ammon’s horn only a few, immature small parvalbumin-positive somata are seen displaying short, rarely branching dendrites. At birth, parvalbumin-positive cells are not visible in the dentate gyrus. In a 1-month-old infant parvalbumin-immunoreactive cells still display poorly developed dendritic branches (figure 12.10A), but already a few parvalbuminpositive axons and a few axon-terminal-like boutons can be observed (figure 12.10A, 12.11A). In the next few months (before the second year of age) both dendritic and axonal arborization expand, although the developmental delay between Ammon’s horn and the dentate gyrus is still visible (figures 12.10B, 12.11B). Even in a 2-year-old child, the morphology of the dendrites and axonal branching as well as the number of terminal boutons of the parvalbuminimmunoreactive cells are less developed than in 8- or 10year-old children. In our material, the adultlike morphology of parvalbumin-containing cells and the adultlike pattern of parvalbumin-containing axonal network in the principal cell layers of Ammon’s horn and the dentate gyrus appeared in an 8-year-old child (figures 12.10C, 12.10D, 12.11C, 12.11D). However, we still have no specimens between the 2nd and 8th years; therefore, we suggest that maturation of these cells is completed during this period, similarly to the development of hilar mossy cells (Seress and Mrzljak, 1992). It has to be emphasized that lack of parvalbumin in inhibitory neurons does not mean that perisomatic inhibitory terminals would be completely missing in the hippocampal formation at birth. Our unpublished electron microscopic observations revealed that perisomatic inhibitory synapses exist on somata of both granule and pyramidal cells in neonates, although those axon terminals are not parvalbumin immunoreactive. (See also plates 27 and 28.) Interneurons play a crucial role in the processes of memory formation through their role in the γ-oscillation

of the hippocampal neuronal network induced by the synchronized activity of GABAergic local-circuit neurons (Wang and Buzsaki, 1996; Wallenstein and Hasselmo, 1997). Recent investigations showed that parvalbumincontaining local-circuit neurons are critical in this process. Hippocampal parvalbumin-positive neurons form a syntitium through their dendro-dendritic gap junctions (Fukuda and Kosaka, 2003) and supposedly mediate inhibition-based coherent γ–rhythms (Tamás et al., 2000). In the hippocampal formation of adult parvalbumin-deficient mice, in which perisomatic inhibitory neurons exist but do not express parvalbumin, inhibition-based γ-oscillation increases, resulting in a lower threshold for the development of epileptiform activity (Schwaller et al., 2004). In addition, the lack of parvalbumin in interneurons may affect the higher cognitive functions associated with γ–oscillation (Vreugdenhil et al., 2003). We propose that the long-lasting postnatal maturation of parvalbumin-containing axo-somatic inhibitory cells in humans might increase the susceptibility of the hippocampal formation in newborns or in infants for stimuli that later cannot induce a similar effect. This hypothesis might clarify why a high fever may cause febrile seizures in young infants and may also explain why such early generated seizures disrupt the normal development of the inhibitory circuitry, resulting in manifestations of epileptic activity and epileptic morphological changes in the hippocampal formation later in life. However, the delayed expression of parvalbumin in the perisomatic inhibitory cells coincides with the prolonged maturation of the principal cells, such as the granule and mossy cells of the dentate gyrus (Seress, 1992; Seress and Mrzljak, 1992). The long postnatal development of both principal and inhibitory cells may offer an explanation for the long-lasting cognitive development of children.

Cell proliferation, migration, and neurochemical maturation in preterm infants The better chance of survival of preterm infants is coupled with the growing concern about the neurodevelopmental outcomes of the infants (Hack and Fanaroff, 1999; Vohr et al., 2000). The lower IQ scores and educational difficulties of preterm children can be correlated with the reduced size of several brain areas, including the hippocampus. The neurobehavioral outcome of preterm infants has been reported to worsen with younger gestational age at birth and with lower birth weight (Hack, Friedman, and Fanaroff, 1996; McCormick, Workman-Daniels, and Brooks-Gunn, 1996). A plausible reason for reduction, in case of extreme preterms, would be the reduced cell formation following premature birth or some defect in the migratory pathway that would guide the neurons to a wrong place. Although in

seress and ábrahám: pre- and postnatal human hippocampal formation

203

Figure 12.9 Development of the calbindin- or calretinincontaining interneurons. (A) Calbindin-immunoreactive local circuit neurons (arrows) with immature morpholgy in the hilus (h), the stratum moleculare (m) of the dentate gyrus, and the stratum radiatum (r) of the CA3 area of the hippocampal formation of a 22-week-old fetus. (B) Large numbers of calretinin-positive interneurons at the border of the strata radiatum (r) and lacunosum-

moleculare (l-m) of the CA1 region of a 22-week-old fetus. The location is similar to that found in the adult, although the cellular morphology of these cells is still immature. (C) Small calretininimmunoreactive cells in the ventricular germinative zone (vz) along the CA1 area of the hippocampus. These calretinin-positive cells may migrate later to the strata oriens (o) and pyramidale (p) of Ammon’s horn. Calibration bars, 50 μm. (See plate 26.)

Figure 12.10 Parvalbumin immunoreactive interneurons in the CA1 region. (A) Large parvalbumin-positive cell with immature morphology and sparse dendritic branches in a 1-month-old infant. Arrowheads point to parvalbumin-immunoreactive axons with terminal-like axonal swellings. (B) A multipolar, large parvalbuminimmunoreactive interneuron with lightly stained branching dendrites in a 3-month-old infant. Arrowheads point to rare terminal-like boutons. (C) Large, multipolar parvalbumin-positive cell in the pyramidal layer (p) of Ammon’s horn with dendrites running

both toward stratum oriens and through stratum radiatum (r) in an 10-year-old child. The morphology of this cell is comparable to parvalbumin-positive cells found in adults. The parvalbuminpositive axonal network is confined to the pyramidal cell layer (p). (D) High-magnification photomicrograph of the parvalbumin-positive axonal network in the pyramidal layer (p). Arrowheads point to axonal swellings (terminal-like boutons) that appear to surround individual neurons (n). Calibration bars, 20 μm for A and D, 25 μm for B, and 50 μm for C. (See plate 27.)

Figure 12.11 Parvalbumin-immunoreactive interneurons in the dentate gyrus. (A) Parvalbumin-immunoreactive axonal branches displaying large swellings (arrowheads) in the granule cell layer (g) and in the hilus (h) of the dentate gyrus in a 1-month-old infant. (B) Soma and main dendrites of a parvalbumin-immunoreactive cell in the granule cell layer of the dentate gyrus in a 3-month-old infant. Axonal branching (arrowheads) is similarly sparse as in 1month-old child. (C) Large-magnification photomicrograph of parvalbumin-immunoreactive axon terminals in the granule cell layer of an 10-year-old child. The axons (arrowheads) display large

numbers of boutons that appear to surround somata of granule cells (n). (D) A large parvalbumin-immunoreactive hilar (h), neuron with long dendrites that cross the granule cell layer (g). Inside the granule cell layer (g) the dense parvalbumin-immunoreactive axonal network appears to be denser at the hilar border (h), suggesting uneven perisomatic innervation of granule cells in the width of the granule cell layer. Density of axonal branches is low in the hilus (h), corresponding with a lower cellular density in the hilus than in the granule cell layer. Calibration bars, 20 μm for A, 25 μm for B, 10 μm for C, 50 μm for D. (See plate 28.)

preterm infants the hippocampus is reduced in volume, reduced cell formation was never demonstrated. To study a possible change in cell proliferation, migration, or neurochemical development we compared the hippocampi of preterm infants with age-matched controls (table 12.2). Different groups were created to compare the effects of preterm delivery in extremely preterm infants (gestational age at birth <30 weeks) with that in those who were born after the 30th gestational week. There was no difference in the cytoarchitectonics of dentate gyrus and Ammon’s horn of the hippocampal formation of preterm infants and full-term age-matched controls (Ábrahám et al., 2004). In preterm infants, including those with extremely low birth weight, the rate of cell proliferation was slightly higher in all layers of the dentate gyrus than in their controls. The proliferating cells were mainly found in the hilar region of the dentate gyrus (Ábrahám et al., 2004). A large proportion of the MIB-1-positive dividing cells were not neuronal precursors but developing astrocytes and endothelial cells of capillaries. In fact, the highest number of proliferating endothelial cells was found in the molecular layer of the dentate gyrus and in the strata lacunosum-moleculare and radiatum of Ammon’s horn, the places where neuronal dendritic growth is the greatest. The values were slightly higher in preterm infants than in controls. In contrast, in the hilus, where cell proliferation is the largest, the ratio of labeled endothelial cells was low and slightly decreased in preterms, whereas the proportion of dividing astrocytes was increased in premature infants

(Ábrahám et al., 2004). At the same time, there was no sign of an abnormal glial or neuronal accumulation in any area of the hippocampal formation. Neurochemical development appeared to be similar both in the principal cells and in the local circuit neurons in preterm infants and in age-matched controls. Interestingly enough, the number of calretinin- and reelin-positive Cajal-Retzius cells appeared to be higher along the hippocampal fissure in preterm infants than in controls, especially along the subiculum and entorhinal cortex as well as in the temporal neocortex. Since Cajal-Retzius cells and their product reelin are responsible for the radial organization of the mammalian cortex (Bar, Lambert de Rouvroit, and Goffinet, 2000), our finding is exciting regarding the fact that extremely preterm infants, when measured at 38–42 postconceptional weeks of age, had a smaller cortical surface that was less complex than in the normal infants born at term (Ajayi-Obe et al., 2000). Although the mechanism is not clear, it can be suggested that reelin secreted by Cajal-Retzius cells may play a role in the formation of gyrification of the cerebral cortex. In conclusion, rate of cell proliferation correlates with the postconceptual age and is not reduced in preterms after birth. It appears that Cajal-Retzius cells persist in larger numbers both in the hippocampal formation and in the temporal neocortex of preterm infants, although cytoarchitectonic differences could not be detected either in the dentate gyrus or in Ammon’s horn. Neurochemical differentiation of individual neurons correlates with the postconceptual age and not with the time of birth. The lower

Cases Control Preterm Preterm

Table 12.2 Personal data and clinical diagnosis verified by autopsy of preterm and full-term children included in this study Gestational Age Postconceptual Age Gender Birth Weight at Birth (weeks) Postnatal Life Time at Death Female 2,650 g 38 2 days 38 weeks Male 1,620 g 34 4 weeks 38 weeks Male 2,610 g 36 2 weeks 38 weeks

Control Preterm Preterm

Male Male Male

2,800 g 780 g 720 g

39 27 24

1 day 14 weeks 21 weeks

39 weeks 41 weeks 45 weeks

CRI CRI, IRDS CRI, IRDS

Control Preterm Preterm

Female Female Male

3,050 g 1,890 g 3,050 g

39 35 37

13 weeks 16 weeks 12 weeks

52 weeks 51 weeks 49 weeks

CI CRI CRI

Cause of Death CRI CRI CRI

Control Male 2,300 g 38 22 weeks 60 weeks CRI Preterm Male 990 g 27 39 weeks 66 weeks CI, IRDS Preterm Male 980 g 27 32 weeks 59 weeks CRI, IRDS Abbreviations: CRI, cardiorespiratory insufficiency; CI, cardial insufficiency; IRDS, infant respiratory distress syndrome

seress and ábrahám: pre- and postnatal human hippocampal formation

207

volume of the hippocampal formation in preterm infants is unlikely to be the result of reduced neuronal or glial cell formation.

Conclusion and functional implications The major bulk of the neuronal cell formation of the human hippocampus occurs within a relatively short period of time, with the exception of granule cells. The major cell formation for entorhinal cortex, subicular subregions, and Ammon’s horn ceases before the 14th or 15th GW. Cell migration is rapid in these regions, because an adultlike cytoarchitectonics can be observed at the 24th GW. Formation of granule cells should start around the 11th week, because at the 12th week the rudiment of the dentate granular layer is formed. The majority of granule cells are formed until the 28th GW although they continue proliferating afterward. Consequently, granule cell migration from the hilus to the granule cell layer lasts for eight to ten months. Such a long period of cell migration may explain the presence of immaturelooking granule cells in the dentate granular layer of an 18month-old child (Seress, 1992). At the end of the first year, migrating cells disappear from the hilus. This period coincides with neurochemical maturation, since it is the first time when all granule cells express calbindin. It is also likely that the young granule cells form their synapses later than the earlier-generated granule cells, a fact that may influence the maturation of their postsynaptic cells, such as the mossy cells, CA3 pyramidal neurons, and parvalbumin-positive interneurons. The adultlike function of the hippocampal formation cannot be expected without an established mature trisynaptic circuitry among the principal cells of the dentate gyrus and Ammon’s horn. In addition, the inhibitory cells and the inhibitory circuitry that provide GABAergic inhibition for the principal cells should also be adultlike. In this respect, it is especially important that inhibitory cells providing dendritic inhibition mature earlier but that the perisomatic inhibitory cells that have very strong effect on principal cell activity develop late in childhood. Morphological and neurochemical development of principal and GABAergic neurons suggests that neuronal connectivity in the human hippocampus reaches an adultlike complexity between the 2nd and 8th years. Therefore, hippocampus may be involved in memory formation in young infants (de Haan et al., 2006), but hippocampus-related adultlike memory formation in humans may not be expected earlier than in early childhood (3–5 years of age). acknowledgments

The authors wish to thank Dr Béla Veszprémi, M.D., Ph.D. (Department of Obstetrics and Gynecology), Dr Éva Gömöri, M.D., Ph.D. (Department of Pathology), and Dr Tamás Tornóczky (Department of

208

Pathology) for their collaboration in collecting the material. The authors acknowledge Mrs. Emese Papp for her excellent technical assistance in the histological preparations of the tissue. This work was supported by the Hungarian National Science Fund (OTKA) with grant T047109.

REFERENCES Ábrahám, H., and G. Meyer, 2003. Reelin-expressing neurons in the postnatal and adult human hippocampal formation. Hippocampus 13:715–727. Ábrahám, H., C. G. Perez-Garcia, and G. Meyer, 2004. p73 and Reelin in Cajal-Retzius cells of the developing human hippocampal formation. Cerebral Cortex 14:484–495. Ábrahám, H., T. Tornóczky, Gy. Kosztolányi, and L. Seress, 2004. Cell proliferation correlates with the postconceptual and not with the postnatal age in the hippocampal dentate gyrus, temporal neocortex and cerebellar cortex of preterm infants. Early Hum. Dev. 78:29–43. Ajayi-Obe, M., N. Saeed, F. M. Cowan, M. A. Rutherford, and A. D. Edwards, 2000. Reduced development of cerebral cortex in experimentally preterm infants. Lancet 356:1162–1163. Alcantara, S., M. Ruiz, G. D’Arcangelo, F. Ezan, L. de Lecea, T. Curran, C. Sotelo, and E. Soriano, 1998. Regional and cellular patterns of reelin mRNA expression in the forebrain of the developing and adult mouse. J. Neurosci. 18:7779–7799. Amaral, D. G., 1990. Hippocampal formation. In The Human Nervous System, ed. G. Paxinos, 711–755. New York: Academic Press. Anderson, S. A., D. D. Eisenstat, L. Shi, and J. L. Rubenstein, 1997. Interneuron migration from basal forebrain to neocortex: Dependence on Dlx genes. Science 278:474–476. Angevine, J. B., 1975. Development of the hippocampal formation. In The Hippocampus, ed. R.L. Isaacson and K. H. Pribram, 61–90. New York and London: Plenum Press. Arnold, S. E., and J. Q. Trojanowski, 1996a. Human fetal hippocampal development. I. Cytoarchitecture, myeloarchitecture and neuronal morphologic features. J. Comp. Neurol. 367:274– 292. Arnold, S. E., and J. Q. Trojanowski, 1996b. Human fetal development. II. The neuronal cytoskeleton. J. Comp. Neurol. 367:293– 307. Arvidsson, A., T. Collin, D. Kirik, Z. Kokaia, and O. Lindvall, 2002. Neuronal replacement from endogenous precursors in the adult brain after stroke. Nature Med. 8:963–970. Bar, I., C. Lambert de Rouvroit, and A. M. Goffinet, 2000. The evolution of cortical development: A hypothesis based on the role of Reelin signaling pathway. Trends Neurosci. 23:633– 638. Bartlett, C. W., N. Gharani, J. H. Millonig, and L. M. Brzustowicz, 2005. Three autism candidate genes: A synthesis of human genetic analysis with other disciplines. Int. J. Dev. Neurosci. 23:221–234. Bayer, S. A., 1980. The development of the hippocampal region in the rat. I. Neurogenesis examined with H3-thymidine autoradiography. J. Comp. Neurol. 190:87–114. Bhardwaj, R. D., M. A. Curtis, K. L. Spalding, B. A. Buchholz, D. Fink, T. Bjork-Eriksson, C. Nordborg, F. H. Gage, H. Druid, P. S. Ericsson, and J. Frisen, 2006. Neocortical neurogenesis in humans is restricted to development. Proc. Natl. Acad. Sci. USA 103:12564–12568.

fundamentals of developmental neurobiology

Biebl, M., C. M. Cooper, J. Winkler, and H. G. Kuhn, 2000. Analysis of neurogenesis and programmed cell death reveals a self-renewing capacity in the adult rat brain. Neurosci. Lett. 291:17–20. Boekhoorn, K., M. Joels, and P. J. Lucassen, 2006. Increased proliferation reflects glial and vascular associated changes, but not neurogenesis in the presenile Alzheimer hippocampus. Neurobiol. Dis. 24:1–14. Cajal, S. R., 1891. Sur la structure de l’ecorce cérébrale de quelques mammiféres. La Cellule 7:123–176. English translation in Cajal on the Cerebral Cortex, ed. J. DeFelipe and E. G. Jones, 23–54. Oxford, UK: Oxford University Press, 1988. Cameron H. A., and R. D. G. McKay, 2001. Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. J. Comp. Neurol. 435:406–417. Curtis, M. A., E. B. Penney, J. Pearson, M. Dragunow, B. Connor, and R. L. M. Faull, 2005. The distribution of progenitor cells in the subependymal layer of the lateral ventricle in the normal and Huntington’s disease human brain. Neuroscience 132: 777–788. D’Arcangelo, G., G. G. Miao, S. C. Chen, H. D. Soares, J. I. Morgan, and T. Curran, 1995. A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature 374:719–723. Dayer, A. G., A. A. Ford, K. M. Cleaver, M. Yassaee, and H. A. Cameron, 2003. Short-term and long-term survival of new neurons in the rat dentate gyrus. J. Comp. Neurol. 460:563–572. de Haan, M., M. Mishkin, T. Baldeweg, and F. Vargha-Khadem, 2006. Human memory development and its dysfunction after early hippocampal injury. Trends Neurosci. 29:374–381. Deller, T., A. Drakew, B. Heimrich, E. Forster, A. Tielsch, and M. Frotscher, 1999. The hippocampus of the reeler mutant mouse: Fiber segregation in area CA1 depends on the position of the postsynaptic target cells. Exp. Neurol. 156:254–267. Del Rio, J. A., M. Solé, V. Borrell, A. Martinez, and E. Soriano, 2002. Involvement of Cajal-Retzius cells in robust and layer-specific regeneration of the entorhino-hippocampal pathway. Eur. J. Neurosci. 15:1881–1890. Drakew, A., T. Deller, B. Heimrich, C. Gebhardt, D. Del Turco, A. Tielsch, E. Forster, J. Herz, and M. Frotscher, 2002. Dentate granule cells in reeler mutants and VLDLR and ApoER2 knockout mice. Exp. Neurol. 176:12–24. Eckenhoff, M. F., and P. Rakic, 1988. Nature and fate of proliferative cells in the hippocampal dentate gyrus during life span of the rhesus monkey. J. Neurosci. 8:2729–2747. Eriksson, P. S., E. Perfilieva, T. Bjork-Eriksson, A. M. Alborn, C. Nordborg, D. E. Peterson, and F. H. Gage, 1998. Neurogenesis in the adult human hippocampus. Nature Med. 4:1313– 1317. Fahrner, A., G. Kann, A. Flubacher, C. Heinrich, T. M. Freiman, J. Zentner, M. Frotscher, and C. A. Haas, 2007. Granule cell dispersion is not accompanied by enhanced neurogenesis in temporal lobe epilepsy patients. Exp. Neurol. 203:320–332. Fatemi, S. H., J. A. Earle, and T. McMenomy, 2000. Reduction in Reelin immunoreactivity in hippocampus of subjects with schizophrenia, bipolar disorder and major depression. Mol. Psychiatry 5:654–663. Forster, E., Y. Jossin, S. Zhao, X. Chai, M. Frotscher, and A. M. Goffinet, 2006. Recent progress in understanding the role of Reelin in radial neuronal migration, with specific emphasis on the dentate gyrus. Eur. J. Neurosci. 23:901–909. Freund, T. F., and G. Buzsáki, 1996. Interneurons of the hippocampus. Hippocampus 6:345–470.

Frotscher, M., L. Seress, W. K. Schwerdtfeger, and E. Buhl, 1991. The mossy cells of the fascia dentata: A comparative study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol. 312:145–163. Fukuda, T., and T. Kosaka, 2003. Ultrastructural study of gap junctions between dendrites of parvalbumin-containing GABAergic neurons in various neocortical areas of the adult rat. Neuroscience 120:5–20. Gerdes, J., M. H. G. Becker, and G. Key, 1992. Immunohistological detection of tumor growth fraction (Ki-67) in formalinfixed and routinely processed tissues. J. Pathol. 168:85–87. Gould, E., and C. G. Gross, 2002. Neurogenesis in adult mammals: Some progress and problems. J. Neurosci. 22:619–623. Gould, E., A. J. Reeves, M. Fallah, P. Tapanat, C. G. Gross, and E. Fuchs, 1999. Hippocampal neurogenesis in adult Old World primates. Proc. Natl. Acad. Sci. USA 96:5263–5267. Gould, E., A. J. Reeves, M. S. Graziano, and C. G. Gross, 1999. Neurogenesis in the neocortex of adult primates. Science 286: 548–552. Gould, E., and P. Tanapat, 1997. Lesion-induced proliferation of neuronal progenitors in the dentate gyrus of the adult rat. Neuroscience 80:427–436. Grove, E. A., and S. Tole, 1999. Patterning events and specification signals in the developing hippocampus. Cerebral Cortex 9: 551–561. Grove, E. A., S. Tole, J. Limon, L. Yip, and C. W. Ragsdale, 1998. The hem of the embryonic cerebral cortex is defined by the expression of multiple Wnt genes and is compromised in Gli3-deficient mice. Development 125:2315–2325. Haas, C. A., O. Dudeck, M. Kirsch, C. Huszka, G. Kann, S. Pollak, J. Zentner, and M. Frotscher, 2002. Role for reelin in the development of granule cell dispersion in temporal lobe epilepsy. J. Neurosci. 22:5797–5802. Hack, M., and A. A. Fanaroff, 1999. Outcomes of children of extremely low birthweight and gestational age in 1990’s. Early Hum. Dev. 53:193–218. Hack, M., H. Friedman, and A. A. Fanaroff, 1996. Outcomes of extremely low birth weight infants. Pediatrics 98:931–937. Heinrich, C., N. Nitta, A. Flubacher, M. Muller, A. Fahrner, M. Kirsch, T. Freiman, F. Susuki, A. Depaulis, M. Fotscher, and C. A. Haas, 2006. Reelin deficiency and displacement of mature neurons, but not neurogenesis, underlie the formation of granule cell dispersion in the epileptic hippocampus. J. Neurosci. 26:4701–4703. Hines, H., 1922. Studies in the growth and differentiation of the telencephalon in man: The fissura hippocampi. J. Comp. Neurol. 34:73–170. Hof, P. R., I. I. Glezer, F. Conde, R. A. Flagg, M. B. Rubin, E. A. Nimchinsky, D. Vogt, and M. Weisenhorn, 1999. Cellular distribution of the calcium-binding proteins parvalbumin, calbindin, and calretinin in the neocortex of mammals: Phylogenetic and developmental patterns. J. Chem. Neuroanat. 16:77–116. Hong, S. E., Y. Y. Shugart, D. T. Huang, S. A. Shahwan, P. E. Grant, J. O. Hourihane, N. D. Martin, and C. A. Walsh, 2000. Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nature Genet. 26:93–96. Humphrey, T., 1967. The development of the human hippocampal fissure. J. Anat. 101:655–676. Jin, K., X. Wang, L. Xie, X. O. Mao, W. Zhu, Y. Wang, J. Shen, Y. Mao, S. Banwait, and D. A. Greenberg, 2006. Evidence for stroke-induced neurogenesis in the human brain. Proc. Natl. Acad. Sci. USA 103:13198–13202.

seress and ábrahám: pre- and postnatal human hippocampal formation

209

Kornack, D. R., and P. Rakic, 1999. Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Proc. Natl. Acad. Sci. USA 96:5768–5773. Kornack, D. R., and P. Rakic, 2001. Cell proliferation without neurogenesis in adult primate neocortex. Science 294:2127– 2130. Kuhn, H. G., H. Dickinson-Anson, and F. H. Gage, 1996. Neurogenesis in the dentate gyrus of the adult rat: Age-related decrease of neuronal progenitor proliferation. J. Neurosci. 16:2027–2033. Lambert de Rouvroit, C., and A. M. Goffinet, 2001. Neuronal migration. Mech. Dev. 105:47–56. Letinic, K., R. Zoncu, and P. Rakic, 2002. Origin of GABAergic neurons in the human neocortex. Nature 417:645–649. McCabe, B. K., D. C. Silveira, M. R. Cilio, B. H. Cha, X. Liu, Y. Sogawa, and G. L. Holmes, 2001. Reduced neurogenesis after neonatal seizures. J. Neurosci. 21:2094–2103. McCormic, M., K. Workman-Daniels, and J. Brooks-Gunn, 1996. The behavioral and emotional well-being of school-age children with different birth weight. Pediatrics 97:18–25. Meyer, G., and A. M. Goffinet, 1998. Prenatal development of reelin-immunoreactive neurons in the human neocortex. J. Comp. Neurol. 397:29–40. Meyer, G., A. M. Goffinet, and A. Fairen, 1999. What is a CajalRetzius cell? A reassessment of a classical cell type based on recent observations in the developing neocortex. Cerebral Cortex 9:765–775. Meyer, G., C. G. Perez-Garcia, H. Ábrahám, and D. Caput, 2002. Expression of p73 and Reelin in the developing human cortex. J. Neurosci. 22:4973–4986. Meyer, G., and P. Wahle, 1999. The paleocortical ventricle is the origin of reelin-expressing neurons in the marginal zone of the foetal human neocortex. Eur. J. Neurosci. 11:3937–3944. Murell, W., G. R. Bushell, J. Livesey, J. McGrath, K. P. A. McDonald, P. R. Bates, and A. Mackay-Sim, 1996. Neurogenesis in adult human. NeuroReport 7:1189–1194. Nadarajah, B., P. Alifragis, R. O. Wong, and J. G. Parnavelas, 2002. Ventricle-directed migration in the developing cerebral cortex. Nature Neurosci. 5:218–224. Nadarajah, B., and J. G. Parnavelas, 2002. Modes of neuronal migration in the developing cerebral cortex. Nature Rev. Neurosci. 3:423–432. Nowakowski, R. S., and P. Rakic, 1981. The site of origin and route and rate of migration of neurons to the hippocampal region of the rhesus monkey. J. Comp. Neurol. 196:129–154. Ogawa, M., T. Miyata, K. Nakajima, K. Yagyu, M. Seike, K. Ikenaka, H. Yamamoto, and K. Mikoshiba, 1995. The reeler gene-associated antigen on Cajal-Retzius neurons is a crucial molecule for laminar organization of cortical neurons. Neuron 14:899–912. Parent, J. M., T. W. Yu, R. T. Leibowitz, D. A. Geschwind, R. A. Sloviter, and D. H. Lowenstein, 1997. Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult hippocampus. J. Neurosci. 17:3727–3738. Price, J., 2001. Neuronal stem cells—Where are you? Nature Med. 7:998–999. Purpura, D. S., 1975. Normal and aberrant neuronal development in the cerebral cortex of human fetus and young infant. In Brain Mechanisms in Mental Retardation, ed. M. Brazier, 141–169. UCLA Forum in Medical Sciences. New York: Academic Press. Rakic, P., 1971. Guidance of neurons migrating to the fetal monkey neocortex. Brain Res. 33:471–476.

210

Rakic, P., 1998. Young neurons for old brains? Nature Neurosci. 1:645–647. Rakic, P., 2002. Neurogenesis in adult primate neocortex: An evaluation of the evidence. Nature Rev. Neurosci. 3:65–71. Rakic, P., and R. S. Nowakowski, 1981. The time of origin of neurons in the hippocampal region of the rhesus monkey. J. Comp. Neurol. 196:99–128. Reif, A., S. Fritzen, M. Finger, A. Strobel, M. Lauer, A. Schmitt, and K. P. Lesch, 2006. Neural stem cell proliferation is decreased in schizophrenia, but not in depression. Mol. Psychiatry 11:514–522. Retzius, G., 1893. Die Cajal’schen Zellen der Grosshirnrinde beim Menschen und bei Säugethieren: Biologische Untersuchungen von prof. Dr. Gustav Retzius, Neue Folge, vols. 1–8, Stockholm. Rose, D. S. C., P. H. Maddox, and D. C. Brown, 1994. Which proliferation markers for routine immunohistology? A comparison of five antibodies. J. Clin. Pathol. 47:1010–1014. Roy, N. S., S. Wang, L. Jiang, J. Kang, A. Benraiss, C. HarrisonRestelli, R. A. Fraser, W. T. Couldwell, A. Kawaguchi, H. Okano, M. Nedergaard, and S. A. Goldman, 2000. In vitro neurogenesis by progenitor cells isolated from the adult human hippocampus. Nature Med. 6:271–277. Schiffmann, S. N., B. Bernier, and A. M. Goffinet, 1997. Reelin mRNA expression during mouse brain development. Eur. J. Neurosci. 9:1055–1071. Schwaller, B., I. V. Tetko, P. Tandon, D. C. Silveira, M. Vreugdenhil, T. Henzi, M. C. Potier, M. R. Celio, and A. E. Villa, 2004. Parvalbumin deficiency affects network properties resulting in increased susceptibility to epileptic seizures. Mol. Cell Neurosci. 25:650–663. Seress, L., 1977. The postnatal development of the rat dentate gyrus and the effect of early thyroid hormone treatment. Anat. Embryol. (Berl.) 151:335–339. Seress, L., 1992. Morphological variability and developmental aspects of monkey and human granule cells: Differences between the rodent and primate dentate gyrus. In The Dentate Gyrus and Its Role in Seizures, ed. C. E. Ribak, C. M. Gall, and I. Mody. Epilepsy Res. (Suppl.) 7:3–28. Seress, L., 2001. Morphological changes of the human hippocampal formation from midgestation to early childhood. In Handbook of Developmental Cognitive Neuroscience, ed. C. E. Nelson and M. Luciana, 45–58. Cambridge, MA: MIT Press. Seress, L., H. Ábrahám, T. Tornóczky, and G. Kosztolányi, 2001. Cell formation in the human hippocampal formation from mid-gestation to the late postnatal period. Neuroscience 105:831– 843. Seress, L., A. Gulyás, I. Ferrer, T. Tunon, E. Soriano, and T. F. Freund, 1993. Distribution, morphological features, and synaptic connections of parvalbumin- and calbindin D28kimmunoreactive neurons in the human hippocampal formation. J. Comp. Neurol. 337:208–230. Seress, L., A. Gulyás, and T. F. Freund, 1992. Pyramidal neurons are immunoreactive for calbindin D28k in the CA1 subfield of the human hippocampus. Neurosci. Lett. 138:257–260. Seress, L., and L. Mrzljak, 1992. Postnatal development of mossy cells in the human dentate gyrus: A light microscopic Golgi study. Hippocampus 2:127–142. Seress, L., and C. E. Ribak, 1995a. Postnatal development CA3 pyramidal neurons and their afferents in the Ammon’s horn of Rhesus monkeys. Hippocampus 5:217–231. Seress, L., and C. E. Ribak, 1995b. Postnatal development and synaptic connections of hilar mossy cells in the hippocampal dentate gyrus of Rhesus monkeys. J. Comp. Neurol. 355:93–110.

fundamentals of developmental neurobiology

Siman, R., D. Bozyczko-Coyne, S. L. Meyer, and R. V. Ad Bhat, 1999. Immunolocalization of caspase proteolysis in situ: Evidence for widespread caspase-mediated apoptosis of neurons and glia in postnatal rat brain. Neuroscience 92:1425–1442. Snyder, E. Y., and K. I. Park, 2002. Limitations in brain repair. Nature Med. 7:928–930. Soriano, E., J. A. Del Rio, A. Martinez, and H. Super, 1994. Organization of the embryonic and early postnatal murine hippocampus. I. Immunocytochemical characterization of neuronal populations in the subplate and marginal zone. J. Comp. Neurol. 342:571–595. Stanfield, B. B., and W. M. Cowan, 1979. The morphology of the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol. 185:393–422. Tamas, G., E. Buhl, A. Lorincz, and P. Somogyi, 2000. Proximally targeted GABAergic synapses and gap junctions synchronize cortical interneurons. Nature Neurosci. 3:366–371. Tiu, S. C., W.Y. Chan, and D. T. Yew, 2004. Cell proliferation in the developing human cerebral cortex. Neuroembryology 5:27– 35. Verheijen, R., H. J. H. Kuijpers, R. O. Schlingemann, A. L. M. Boehmer, R. vanDriel, G. J. Brakenhoff, and F. C. S. Ramaekers, 1989. Ki-67 detects a nuclear matrix-associated proliferation-related antigen. I. Intracellular localization during interphase. J. Cell Sci. 92:123–130. Verheijen, R., H. J. H. Kuijpers, R. van Driel, J. L. M. Beck, J. H. vanDierendock, G. J. Brakenhoff, and F. C. S. Ramaekers, 1989. Ki-67 detects a nuclear matrix-associated proliferation related antigen. II. Localization in mitotic cells and association with chromosomes. J. Cell Sci. 92:531–540.

Vohr, B. R., L. L. Wright, A. M. Dusick, L. Mele, J. Verter, and J. J. Steichen, 2000. Neurodevelopmental and functional outcomes of extremely low birth weight infants in the National Institute of Child Health and Human Development Neonatal Research Network, 1993–1994. Pediatrics 105:1216–1226. Vreugdenhil, M., J. G. Jerrerys, M. R. Celio, and B. Schwaller, 2003. Parvalbumin-deficiency facilitates repetitive IPSCs and gamma oscillations in hippocampus. J. Neurophysiol. 89:1414– 1422. Wallenstein, G. V., and M. E. Hasselmo, 1997. GABAergic modulation of hippocampal population activity: Sequence learning, place field development, and the phase precession effect. J. Neurophysiol. 78:393–408. Wang, X. J., and G. Buzsaki, 1996. Gamma oscillation by synaptic inhibition in a hippocampal interneuronal network model. J. Neurosci. 16:6402–6413. Weisenhorn, D. M., E. W. Prieto, and M. R. Celio, 1994. Localization of calretinin in cells of layer I (Cajal-Retzius cells) of the developing cortex of the rat. Brain Res. Dev. Brain Res. 82:293– 297. Yang, A., N. Walker, R. Bronson, M. Kaghad, M. Oosterwegel, J. Bonnin, C. Vagner, H Bonnet, P. Dikkes, A. Sharpe, F. McKeon, and D. Caput, 2000. p73-deficient mice have neurological, pheromonal and inflammatory defects but lack spontaneous tumours. Nature 404:99–103. Zaidel D. W., 1999. Quantitative morphology of human hippocampus early neuron development. Anat. Rec. 254:87– 91. Zecevic, N., and P. Rakic, 2001. Development of layer I neurons in the primate cerebral cortex. J. Neurosci. 21:5607–5619.

seress and ábrahám: pre- and postnatal human hippocampal formation

211

13

Structural Development of the Human Prefrontal Cortex IVICA KOSTOVIC´, MILOŠ JUDAŠ, AND ZDRAVKO PETANJEK

Architecture and connectivity of the adult human prefrontal cortex Delineation and Cytoarchitectonics The prefrontal cortex is the largest and most rostral (anterior) portion of the human cerebral cortex and together with underlying white matter is the largest portion of the human frontal lobe. The human frontal lobe can be loosely delineated at the lateral surface of the cerebrum as the portion in front of (rostral to) the central sulcus. From the central sulcus toward the frontal pole, one can distinguish the coronally situated precentral gyrus corresponding to the motor cortex; in front of the gyrus precentralis, there are three convolutions (gyrus frontalis superior, medius, and inferior), which run in a rostral-caudal (anterior-posterior) direction. The caudal portions of these three gyri correspond to the premotor cortex, while larger, rostral portions correspond to the prefrontal cortex. On the medial aspect of the cerebrum, the prefrontal cortex occupies a position rostral to the premotor and cingulate cortices (the anterior cingulate cortex is considered by some authors as functionally part of the prefrontal cortex). From several cytoarchitectonic maps used in neuroanatomy, the most widely used is Brodmann’s map (Brodmann, 1909; Rajkowska and Goldman-Rakic, 1995) in which the prefrontal cortex corresponds to areas 9, 10, 11, 46, and 47, mostly situated in the superior, medial, and inferior gyri and frontal operculum of the insula. The main cytoarchitectonic feature of the prefrontal (frontal granular) cortex is the presence of large pyramidal neurons in layer IIIc and prominent granularity of layer IV. The granularity of the prefrontal cortex is especially well developed in the human frontal lobe, well pronounced in nonhuman primates, present in carnivora, and poorly developed in rodents. In contrast to the prefrontal cortex, granular layer IV is poorly developed in the premotor cortex and virtually absent in the motor cortex. Neuronal Organization and Intrinsic Circuitry The layer III pyramidal neurons are major projection neurons for corticocortical associative and callosal/commissural pathways. Knowledge about synaptology of layer III pyramidal neurons is based on electron microscope and

immunocytochemical studies in monkeys (Goldman-Rakic, 1999; Lund and Lewis, 1993; Melchitzky and Lewis, 2003). Central elements of both intrinsic and extrinsic circuitry are pyramidal neurons of layer III (figure 13.1 and plate 29). It was shown that these glutamatergic neurons play a central role in conveying information both between and within the cortical regions (Lewis et al., 2002). In contrast, GABA-containing neurons serve as strictly local circuitry neurons (interneurons) that control input and output of pyramidal neurons (figure 13.1). These two classes of neurons (pyramidal and nonpyramidal) interact physiologically but differ in basic properties: pyramidal neurons generate waves known as “regular spiking,” while nonpyramidal neurons generate “fast-spiking” waves. The principal axons of layer III pyramidal neurons (1) project through the white matter to other cortical regions (associational and commissural projections), (2) send longrange (3–4 mm) axon collaterals through gray matter to the nearby cortex (intrinsic projections), and (3) give local collaterals that arborize within 300 micrometers of the cell body (Lewis, Melchitzky, and Burgos, 2002; Pucak et al., 1996; Kritzer and Goldman-Rakic, 1995). The reconstruction of these projections has shown different aspects of modular organization: associational projections to other prefrontal cortical regions form clusters of axonal terminals spanning layers I–VI, while intrinsic long-range projections form discrete clusters that span layers I–III. Both associational and long-range intrinsic axonal terminals form parallel elongated stripes (Pucak et al., 1996). The neurons that contribute axonal collaterals in this circuitry are also arranged in a stripelike fashion (Kritzer and Goldman-Rakic, 1995; Pucak et al., 1996). The modular arrangement of axon terminals is coregistered with those of pyramidal cell bodies (Pucak et al., 1996) and forms by means of reciprocal connections in neuronal networks (Goldman-Rakic, 1999). Neurophysiological studies have shown that the majority of layer-III pyramidal neurons with long-range intrinsic connections link clusters of cells with similar functional properties (Pucak et al., 1996; Lewis, Melchitzky, and Burgos, 2002). Goldman-Rakic (1995, 1999) has proposed that long-range intrinsic connections in the prefrontal cortex link clusters of cells sharing memory fields.

213

interneurons serves to constrain the propagation of pyramidal cell excitation in local, long-range intrinsic and associational projections. There are key questions related to the intrinsic circuitry of the prefrontal cortex: (1) How does the intrinsic circuitry functionally respond to the input from other parts of the cortex and subcortical structures? (2) How does the intrinsic circuitry regulate modular organization of the prefrontal cortex? (3) How and to what extent does the functional architecture of the intrinsic circuitry underlie prefrontal cortex functions, especially working memory? At the moment, experimental evidence suggests that the memory field of a prefrontal neuron is a product of both excitatory and inhibitory influences—long tract projections providing sensory information to the cell by way of excitatory afferents and local circuits providing inhibitory influences from neurons with nonpreferred orientations (Goldman-Rakic, 1999). It should be noted, however, that a major class of GABA neurons is spatially tuned within the memory field. Figure 13.1 Schematic illustration of the local circuitry of monkey prefrontal cortex formed by the axons of layer III pyramidal neurons (blue) and parvalbumin (red)- and calretinin (green)containing cells. Collaterals of pyramidal axons make excitatory synapses (filled blue circles) on dendrites of interneurons, and axons of specific classes of interneurons (A, B, C) make synapses on specific parts of pyramidal somata and dendrites (open red and green circles). (See plate 29.)

The key components of the intrinsic circuitry of the prefrontal cortex are GABA interneurons that innervate dendritic shafts of layer-III pyramidal neurons and serve as inhibitory-function neurons (Melchitzky et al., 2001). By regulating pyramidal cell excitability, interneurons influence or control both their input and output. GABA neurons can be divided into different subclasses based on certain morphological, biochemical, and electrophysiological criteria (see figure 13.1). The first subclass consists of double-bouquet neurons that synthesize calretinin (A), a calcium-binding protein. In the second subclass, chandelier neurons, GABA coexists with parvalbumin (B); in the third subclass, large (wide-arbor) basket cells contain parvalbumin (C). It is difficult to identify the exact functional role of interneurons in local, long-range intrinsic axonal and associational connectivity. However, several generalizations may be drawn. First, excitatory connections of layer-III pyramidal cells are preferentially directed to the parvalbumin-containing class of GABA neurons, forming Gray’s Type 1 synapses (Melchitzky and Lewis, 2003). Second, a great proportion of parvalbumin-positive neurons contact the proximal portions of dendrites of pyramidal neurons. Chandelier cells contact initial segments of axons exhibiting inhibitory input to pyramidal neurons (Anderson et al., 1995). In summary, activation of

214

Extrathalamic (Monoaminergic and Cholinergic) Afferents The prefrontal cortex is innervated by three major chemically identified afferent systems originating in the midbrain tegmentum: (1) dopaminergic (DA) fibers originating from the ventral tegmental area (cell group A10) and forming the mesocortical DA system; (2) noradrenergic (NA) fibers originating from the locus coeruleus; and (3) serotonergic (5-HT) fibers originating in raphe nuclei. All these monoaminergic afferents run through the lateral hypothalamus within the so-called medial forebrain bundle, reach the basal forebrain-septal-preoptic continuum, turn rostrally around the genu of the corpus callosum, and innervate the prefrontal cortex with the network of poorly myelinated varicose fibers. Dopaminergic system. Using immunostaining of different components of the dopamine synthetic chain (DA membrane transporter, DAT; tyrosine hydroxylase, TOH), receptor autoradiography, and electron microscopy on both human and experimental primate material, Goldman-Rakic and colleagues (1992) have described the anatomy of dopamine in the primate prefrontal cortex. Dopamine afferents innervate the prefrontal cortex very densely, showing bilaminar distribution with high concentration in layers I, II, and upper III and in deep layers V and VI (Goldman-Rakic et al., 1992). Dopamine axon terminals have one principal target: pyramidal neurons. The postsynaptic elements for DA terminals on pyramidal neurons are predominantly dendrites and spines, where they form symmetric synapses. Goldman-Rakic and colleagues (1992) have noticed that postsynaptic spines with DA terminals also contain another prospectively excitatory (glutamatergic) asymmetric synapse, calling this synaptic complex a “triad.” Goldman-Rakic and

fundamentals of developmental neurobiology

colleagues (1992) also proposed that a triad is an anatomical substrate for D1/glutamate interactions in the cortex. Distribution of dopaminergic receptors is one indication of DA synaptic functions in the prefrontal cortex and the basis for action of different psychotropic drugs. Receptor autoradiography (Goldman-Rakic et al., 1992) has shown that the D1 dopamine receptor is present in the highest relative concentration in superficial layers I, II, and III and layers V and VI, while D2 receptors show high concentration in deep layer V. This bilaminar distribution of DA axons in the prefrontal cortex reflects the bilaminar distribution seen in the prenatal cortex. Basal forebrain–prefrontal cholinergic system. The major source of cholinergic innervation of the prefrontal cortex is the nucleus basalis of Meynert. Topographically, the nucleus is situated below the putamen and reaches the lateral part of the anterior commissure. The external landmark for the position of the nucleus basalis of Meynert in the brain corresponds to the area in front of the optic tract, the so-called anterior perforated substance. This prominent nucleus is a part of the magnocellular cholinergic chain of nuclei situated in the basal forebrain, which are consecutively termed Ch1–Ch4 (Mesulam et al., 1983). Cholinergic neurons of the basal nucleus are the largest in the brain, but the nucleus basalis also contains parvocellular noncholinergic components. The main fiber output from the nucleus basalis is directed laterally and runs through the external capsule in the external sagittal stratum, situated in the most superficial part of the prefrontal white matter. The cholinergic fibers then innervate densely all cortical layers, forming an elaborate network (Mesulam et al., 1983). They also innervate layer III pyramidal neurons (Mrzljak and Goldman-Rakic, 1992), which are important for cognitive activity. These neurons also show dense acetylcholinesterase activity; this is the enzyme that degrades ACh (Kostovic´, Škavic´, and Strinovic´, 1988). Another source of cholinergic fibers to the frontal cortex, especially the rostral part of Broca’s area, is the nucleus subputaminalis (Šimic´ et al., 1999), which extends very rostrally along and below the junction of lateral and ventral aspects of the putamen, being in the “strategic” position to innervate cognitively important parts of the granular speech area (Brodmann’s area 45). Thalamocortical Afferents Major thalamic input to the prefrontal cortex originates in the mediodorsal thalamic nucleus (MD). Afferents to the prefrontal cortex originate from the thalamus, pass through the anterior limb of the internal capsule, and form the anterior thalamic peduncle, which fans out in the thalamocortical radiation–corona radiata system and densely innervates layer IV of the prefrontal cortex. The mediodorsal projection is glutamatergic and excitatory in nature.

Corticothalamic projections. Pyramidal cells of layers V and VI in the prefrontal cortex project with glutamatergic fibers to the mediodorsal nucleus of the thalamus (Giguere and Goldman-Rakic, 1988) and form two types of synapses: (1) large terminals of corticothalamic axons that contact primary dendrites of thalamocortical relay neurons, forming the glomerulus together with GABA-containing (inhibitory) presynaptic elements, and (2) smaller terminal boutons that form asymmetric synapses on GABAergic inhibitory interneurons. Connections with the amygdala, striatum, and hypothalamus. In the human brain, the amygdaloid body is a large nuclear complex situated in the rostral part of the temporal lobe, just in front of the inferior horn of the lateral ventricle. From this “strategic” position, the large basolateral nuclear group of the amygdaloid body projects into the association frontal cortex through massive amygdalocortical pathways. The amygdaloid body is connected with the auditory cortex, and this connection is important for conveying necessary information for fear reaction. Therefore, the temporal-lobe– amygdala–frontal-cortex circuit may serve as a framework for emotional behavior. The prefrontal cortex is essentially an association cortex, and the majority of prefrontal pathways serve to connect it with other parts of the cerebral cortex. However, in order to control emotional behavior and related motor and visceral expressions, the prefrontal cortex projects significantly to the corpus striatum by way of corticostriatal projections and to the hypothalamus by way of corticohypothalamic connections. In addition, prefrontal executive functions may be expressed by way of prospective frontopontine pathways, although they originate predominantly from the premotor frontal cortex. Prefrontal corticostriatal fibers project to both the caudate nucleus and the putamen. However, the projection to the caudate is more extensive, a difference that is related to the involvement of the caudate nucleus in cognitive behavior (Goldman-Rakic, 1981, 1995, 1999). The origin of prefrontocaudate projection is in pyramidal neurons of layers V and VI (Goldman-Rakic, 1981). The termination of corticostriatal fibers is not random, and terminal fields reflect modular organization of the striatum and surround cytoarchitectonic modular compartments (Goldman-Rakic, 1981). Corticohypothalamic Connections Corticohypothalamic connections were shown to exist in the primate brain. They connect the frontal cortex with the lateral hypothalamus, which is known as a gateway to the brain-stem tegumentum and the amygdala (Nauta and Haymaker, 1969). Corticocortical Connections The prefrontal cortex is typically thought of as an association cortex and possesses

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

215

an extremely rich number of corticocortical associative and commissural connections that are necessary for its cognitive, behavioral, and executive functions (Goldman-Rakic, 1995; Mesulam, 1998). For this type of associative cortex Mesulam has proposed the term “heteromodal” because it receives various modality-specific inputs. Together with paralimbic and limbic cortices, heteromodal areas constitute distributed but integrated critical gateways for assessment of relevant distributed information (Mesulam, 1998; figure 13.2 and plate 32). The major cortical input to the prefrontal cortex originates in the posterior parietal cortex (Goldman-Rakic, 1995, 1999), which is a cortical center for processing spatial information. These connections are bilateral and transmit visuospatial information involved in delayed-response performance. The fibers originating in the parietal cortex terminate in the prefrontal cortex as distinct vertically oriented columns 0.5 mm in width (Goldman-Rakic, 1995, 1999). These columns alternate with columns of fibers from the contralateral prefrontal cortex that cross the midline as a part of the corpus callosum. Both callosal and associative fibers terminate in layers I and IV, while associational fibers terminate in layer VI. The other important cortical connection is to the temporal cortex (Germuska et al., 2006) and to the parahippocampal gyrus, entorhinal cortex, and presubiculum (structures of the hippocampal formation), subserving storage-recall mechanisms for mnemonic aspects of prefrontal functions (figure 13.2).

Early development of the human prefrontal cortex The description of the structural development of the prefrontal cortex given in this chapter will include data on (1) major neurogenetic events, (2) development of neuronal phenotypes and postsynaptic elements, (3) sequential development of afferent pathways, (4) development of intrinsic circuitry, (5) description of circuitry organization at different developmental periods, and (6) evidence on frontal lobe asymmetry. Phases/periods. Although there is a continuity in structural development, one can say that prenatal development is dominated by major neurogenetic events, whereas postnatal development is dominated by the formation of synapses and fine intrinsic circuitry. The most dramatic reorganization occurs perinatally. Therefore, we divided developmental events by those occurring prenatally and postnatally. It will be evident that cellular developmental events in the prefrontal cortex begin early in fetal life and show prolonged (protracted) maturation throughout the postadolescent period of life. Neurogenetic Cellular Events and Early Laminar Development Neurogenetic cellular events (proliferation,

216

migration, aggregation, cell differentiation, growth of axons, and the establishment of neuronal connections) proceed in the cerebral cortex according to the specific timetable within transient fetal zones that do not have an equivalent in the adult brain (Rakic, Ang, and Breunig, 2004). This is the rule in every cortical region. In the anlage of prefrontal cortex, production and migration of neurons, together with their differentiation and establishment of connections, begin surprisingly early in fetal life and last remarkably long. We have observed the last changes in the chemically identified innervation of pyramidal neurons of layer III in the adolescent and young-adult prefrontal cortex (Kostovic´, Škavic´, and Strinovic´, 1988). Neurons destined for the prefrontal cortex are being produced from neuroepithelial stem cells in the ventricular zone at the end of embryonic period, around the 7th postconceptional week when the frontal pole of cerebral vesicles becomes clearly distinguished. At this early period, the cerebral wall consists of three layers: ventricular zone, intermediate zone (mantle), and marginal zone (figure 13.3 and plate 33). Very soon, another embryonic zone, the subventricular zone (SVZ), starts to produce cortical neurons. All neurons born in the ventricular zone and subventricular zone move to their final destination by means of cell migration. There are two basic modes of migration: radial and tangential. The first mode is radial migration along radial glia fibers (Rakic, Ang, and Breunig, 2004) across the intermediate zone; it is the main mechanism of moving for the principal, pyramidal neurons of the cortex. The second mechanism of migration is tangential migration, which seems to be the main mechanism for cortical interneurons (Wonders and Anderson, 2006). The tangential migration of neurons was observed in the subventricular zone, intermediate zone, and marginal zone. However, tangentially oriented migratory neurons were seen in other transient cerebral zones, too (Mrzljak et al., 1988). The origin of tangentially migrating GABAergic interneurons in the human brain is predominantly in the ventricular-subventricular zone of the cerebral wall. Only 20 percent of interneurons (Letinic, Zoncu, and Rakic, 2002) originate in the ganglionic eminence (the enlarged bulging structure, composed of proliferative cells, that appears as a basal continuation of the ventricular zone). In the rodent brain the vast majority of cortical interneurons originate from the ganglionic eminence (Wonders and Anderson, 2006). The process of migration, together with new production, is very intense throughout midgestation and stops after 22 postconceptional weeks. Immature, postmigratory neurons form a new layer below the marginal zone at the so-called cortical plate (CP; figure 13.3) when the embryo is 8 postconceptional weeks old and its crown-rump length is 20–22 mm. However, even before formation of the cortical plate by postmigratory neurons, some special, large fetal neurons are born and situated below

fundamentals of developmental neurobiology

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

217

Figure 13.2 Wiring diagram of prefrontal connections. The vast majority of connections shown are monosynaptic; however, some may be polysynaptic. Terminology for cortical areas according to Mesulam (1998): green, modulatory pathways; red, glutamatergic excitatory pathways; black, inhibitory neurons and pathways; blue, cholinergic projection; interconnection between different heteromodal areas is not completely shown. (See plate 32.)

218

fundamentals of developmental neurobiology

Figure 13.3 Laminar development of cytoarchitectonic layers in the frontal cortex, from the stage before the appearance of the cortical plate to the newborn stage. All layers are transient, and their developmental changes reflect neurogenetic events (proliferation, migration, differentiation, ingrowth of afferent pathways). (A) Seven postconceptional weeks; (B) 10 postconceptional weeks; (C) 13 postconceptional weeks; (D) 23 postconceptional weeks; (E) 34 postconceptional weeks; (F) newborn. Abbreviations for this and subsequent figures: C, caudate nucleus; CP, cortical plate; G, ganglionic eminence; IZ, intermediate zone; MZ, marginal zone; P, putamen; SP, subplate zone; SPf, the subplate zone in formation; SVZ, subventricular zone; SVZf, subventricular fibrillar zone; VZ, ventricular zone; WM, white matter; I–VI, cortical layers I–VI. Arabic numerals in Figure 13.5E correspond to central (1), intermediate (2), and corona radiata (3), segments of the white matter. (See plate 33.)

the pia or at the outer border of the (mantle) intermediate zone of His (Rakic, Ang, and Breunig, 2004). Below the pia is a special population of fetal Cajal-Retzius neurons in the intermediate (mantle) zone of His. Cajal-Retzius cells are some of the early-born neurons of the so-called preplate (the widely accepted term in current neurohistology, although it corresponds to classical descriptions under the term “mantle layer of His”). At the end of this early fetal period, a special laminar event occurs around 13 postconceptional weeks: the formation of the subplate zone (SP; Kostovic´ and Rakic, 1990), also called the second cortical plate (Kostovic´ and Rakic, 1990). This event is caused by the spread of cells from the deep cortical plate that become more loosely arranged and lose their radial orientation (Kostovic´ and Rakic 1990). This event leads to the development of new, prominent lamina, the subplate zone, which becomes the thickest layer of the fetal cortex, containing postmigratory neurons with the most advanced neuronal differentiation, an abundance of extracellular matrix (ECM), and synapses and axons in plexiform arrangement. The development of the subplate zone from the deep cortical plate was overlooked by most scientists working with experimental mammals, probably because that event lasts briefly in other mammals and the subplate zone is less developed. A similar pattern of subplate-zone development can be reconstructed from the paper of Luskin and Shatz (1985) in the cat, where they distinguished the upper and lower parts of the subplate zone. Typical fetal lamination with a well-pronounced transient fetal subplate zone and a condensed cortical plate is present between 14 and 24 postconceptional weeks. During this period, fetal zones can be easily delineated on Nisslstained sections, histochemical preparations, and in vitro (Kostovic´ et al., 2002) and in vivo (Judasˇ et al., 2005; Maas, Mukherjee, and Carballido-Gamio, 2004) MR images. Starting from the ventricles toward the pia, one can distin-

guish the following zones: (1) the ventricular zone (germinal matrix) of high cell-packing density, which corresponds to high MRI signal intensity, (2) the subventricular zone with lower cell-packing density, abundance of periventricular fiber-rich zone, and low MRI signal intensity, (3) the intermediate zone with axonal strata and migrating waves of neurons and moderate MRI signal intensity, (4) the subplate zone of randomly scattered cells, abundance of extracellular matrix, and low MRI signal intensity, (5) the cortical plate composed of tightly packed columns of radially arranged cells and high MRI signal intensity, and (6) the cell-sparse and narrow marginal zone (prospective layer I) which is too thin to be visualized (figures 13.4 and 13.5 and plates 30 and 34). The first lamination in the cortical plate occurs around 21 postconceptional weeks and is the sign of elaboration of dendritic arborizations (see page 221) and ingrowth of thalamocortical afferents (see pages 222–224). The six-layered lamination characteristic of isocortex appears relatively late in the preterm cerebrum around 34 postconceptional weeks. However, despite the six-layered appearance, laminar organization of the preterm cortex shows substantially different cytoarchitecture than that of, for example, a 3-month-old infant. First of all, the subplate zone is very prominent (figure 13.3), and layer VI does not directly border the white matter in cortical gyri and sulci. Second, the marginal zone shows sublamination that is not present in the postnatal brain. Third, layer III pyramidal neurons, together with granularity of layer IV, cytoarchitectonic hallmarks of the prefrontal cortex, are not visible until 36 postconceptional weeks, indicating that growth of somata of layer III pyramidal neurons achieves newborn values after 32 postconceptional weeks (Mrzljak et al., 1992). The prefrontal cortex shows the same basic pattern of laminar development and arrangement in axonal strata as other lateral neocortical areas. However, several specificities

Figure 13.4 Histological laminar development correlated with postmortem MR images: (A) cresyl violet staining; (B,C,E) acetylcholinesterase histochemistry; (D) T1-weighted MRI scan; (A,B) 18–19 postconceptional weeks; (C,D) 23 postconceptional weeks;

(E) 29 postconceptional weeks. Note the prominent subplate zone (5) in the prefrontal cortex. (Reproduced with permission from Kostovic´ et al., 2002.) (See plate 30.)

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

219

Figure 13.5 Laminar development of the frontal lobe revealed by in vitro (A) and in vivo (C, E) imaging, correlated with Nisslstained histological sections (B, D, F). Early consolidation of cortical plate at 10 postconceptional weeks (A, B). Typical fetal

220

lamination with prominent subplate zone at 21 postconceptional weeks (C, D). Gradual disappearance of the subplate zone with appearance of gyri, corona radiata, and “mature” pattern at 34–35 postconceptional weeks (E, F). (See plate 34.)

fundamentals of developmental neurobiology

were observed. First, there are prominent cell rows in the intermediate zone during the period of neuronal migration (15–21 postconceptional weeks). These indicate the presence of more numerous waves of migratory neurons than in other cortices. Second, there is higher cell-packing density in the subplate zone of the prefrontal cortex than in other cortices. The upper subplate zone of the frontal lobe appears as sublaminated. Finally, the subplate zone is very prominent in the prefrontal cortex and is 6–8 mm thick in the preterm cortex (Kostovic´, 1990). The remnants of the subplate zone exist in the prefrontal cortex longer than in other parts and are seen as late as the sixth postnatal month (Kostovic´, 1990). The development and transformation of transient cellular zones has received significant attention in current neuroimaging studies (Glenn and Barkovich, 2006). In vivo analysis of developing transient zones (Kostovic´ et al., 2006) offers a new insight into cellular development of the cerebral cortex and provides indirect evidence of maturation of neuronal connectivity (figure 13.5). Prenatal Development of Neuronal Phenotypes (Special Fetal Cells, Pyramidal Neurons,and Interneurons) The earliest postmigratory neurons appear in the telencephalic wall before the appearance of the cortical plate. One class of the earliest-generated neurons is the early constituent of the marginal zone; these are known as Cajal-Retzius cells. These characteristically large fetal neurons show early cytological and ultrastructural differentiation (Kostovic´ and Rakic, 1990) and can be identified in immunocytochemical preparations owing to the presence of the protein reelin (Meyer and Goffinet, 1998; Rakic and Zecevic, 2003), which is important for the mechanism of termination of neuron migration. The other class of early-maturing postmigratory neurons consists of cells at the border of the intermediate zone (mantle layer of His), recently called preplate neurons (Meyer and Goffinet, 1998). During their transitory accumulation at the border of the intermediate zone/mantle layer (Meyer and Goffinet, 1998) they are called pioneering neurons. After the formation of the cortical plate, these cells are situated in the narrow presubplate zone (Kostovic´ and Rakic, 1990). These early-maturing cells have multipolar, polymorphic shape and show early GABA reactivity, are peptidergic, and probably are sites (Kostovic´ and Rakic, 1990) of earliest cortical synapses (Molliver, Kostovic´, and Van der Loos, 1973). These early preplate/presubplate cells are the earliest interneurons of the fetal cortex. The earliest differentiated neurons in the developing cortical plate of the fetal prefrontal cortex were described by Mrzljak et al. (1988). During the 11th postconceptional week, the neurons in the cortical plate are immature, showing bipolar morphology with vertically (radially) oriented ascending and descending processes. However, these miniature

processes show early branching: the ascending process bifurcates in the marginal zone, resembling early apical dendrite branching. The descending processes are nonbranched or form rootlike arborization in the presubplate zone. Around 13 postconceptional weeks, when the subplate zone is formed from the deep, loose part of the cortical plate (“second” cortical plate), numerous polymorphic neurons are seen in the zone of the subplate zone formation at the interface between the cortical plate and the intermediate zone. The neurons that are in the superficial, radially oriented, dense part of the cortical plate are immature pyramidal neurons and show initial differentiation of basal dendrites and axons passing at different angles in the subplate zone and turning into the intermediate zone on the way to subcortical centers. Between 17 and 22 postconceptional weeks, there is substantial differentiation of the multipolar polymorphic population of subplate neurons. Subplate neurons show remarkable chemical maturity displaying GABA (Zecevic and Milosevic, 1997), somatostatin (Kostovic´ et al., 1991), neuropeptide Y (Delalle et al., 1997), and NADPH-diaphorase activity (Judasˇ, Šestan, and Kostovic´, 1999). After 23 postconceptional weeks, pyramidal neurons of the deep cortical plate develop characteristic morphology with visible basal and apical dendrites. The first nonpyramidal morphologies, characteristic for interneurons, appear around 26 PW. These neurons are “double bouquet” neurons with typical “columnar” and axonal dendritic shape (Mrzljak et al., 1988). The first immature basket interneurons are not observed until 30 postconceptional weeks. They have stellate distribution of dendrites and sparsely developed axonal plexus, indicating that typical “pericellular baskets” will form later in the development. Sequential Development of Afferent Pathways There are two main features of development of afferent pathways in the cortex. The first one is spatial, with the afferents arranged in axonal strata and showing a laminar preference in their distribution, above and below the cortical plate. The second feature is temporal: pathways show sequential but overlapping ingrowth (Kostovic´ and Judaš, 2002). Within the cerebral wall, axonal strata are arranged as follows: The most superficial fibers are monoaminergic afferents, followed by cholinergic afferents located in the external capsule—that is, the outermost part of the intermediate zone. Thalamocortical afferents are located deep beneath the external capsule, while motor fibers represent the deepest stratum of the intermediate zone. Associative pathways develop later and occupy a deep portion of the fetal white matter. The deepest fiber system is situated adjacent to the cerebral ventricle and is composed of commissural (callosal) fibers.

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

221

Development of extrathalamic (monoaminergic, cholinergic, and amygdala) afferents. Monoaminergic afferents from the brain stem nuclei arrive to the frontal part of the cerebral cortex very early at around 9 postconceptional weeks (Nobin and Björklund, 1973; Verney, 1999; Verney et al., 1993). Dopaminergic fibers originating in the ventral tegmental area were shown to pass the cerebral stalk at the border between diencephalon and telencephalon. The dopaminergic afferents run above and below the cortical plate, and their distribution thus corresponds to the bilaminar distribution of synapses. Serotonergic (5HT) fibers originating from cells of raphe nuclei in the brain stem grow very early and approach basal telencephalon and septal levels, turn toward the developing telencephalic wall, and distribute above and below the cortical plate, showing a bilaminar pattern of distribution. The nucleus basalis of Meynert is one of the earliest differentiating subcortical structures in the developing telencephalon. Its development may be followed using acetylcholinesterase (AChE) histochemistry. As early as 9 postconceptional weeks, one can see a histochemically “hot” area below the anlage of the striatum (figure 13.6 and plate 31). During the next week (10 postconceptional weeks), fibers of the external capsule grow outward from this area, approaching the telencephalic wall through the outermost stratum of the developing intermediate zone (fetal white matter). After the formation of the subplate zone around 13 postconceptional weeks, AChE-reactive fibers from the nucleus basalis spread throughout the subplate zone of the developing frontal lobe (Kostovic´, 1986). In the late fetal stage, fibers from the basal telencephalon overlap with thalamocortical afferents in the superficial subplate zone between 21 and 23 postconceptional weeks (Kostovic´ and GoldmanRakic, 1983) and after 24 postconceptional weeks in the cortical plate. Between the preterm and neonatal periods, there are substantial shifts in laminar and cellular distribution of “cholinergic” fibers originating in the basal telencephalon (Kostovic´, 1990). Individual axon networks innervating all cortical layers develop during the first postnatal months (Kostovic´, Skavic´, and Strinovic´, 1988). However, a fine network of AChE-reactive fibers around the pyramidal layer III neurons starts to develop late (after the first postnatal year) and shows prolonged maturation during adolescence and young adulthood (Kostovic´, Škavic´, and Strinovic´, 1988). Development of afferents from the amygdala in the human fetus is not known. However, the amygdala shows very early cytoarchitectonic maturation, and by 15 postconceptional weeks characteristic cytoarchitectonic modules were found in the amygdala of the human fetal brain (Nikolic´ and Kostovic´, 1986). The early origin of amygdala cells and nuclei was documented in experimental primates (Kordower, Piecínski, and Rakic, 1992). This early maturation may be

222

Figure 13.6 Early origin of cholinergic afferents (arrows) from the nucleus basalis area (dark area that surrounds the asterisk). (See plate 31.)

a sign of unexpected early projection to the prefrontal cortex and early maturation of the emotional system as stated by Machado and Bachevalier (2003). Development of thalamocortical pathways. During the development of thalamocortical pathways, one can distinguish the following phases: axonal outgrowth, pathfinding, target selection, and address selection (Rakic, Ang, and Breunig, 2004). The initial outgrowth of thalamocortical axons from the mediodorsal nucleus of the thalamus occurs early in fetal life, and these fibers, after pathfinding in the internal capsule, can be demonstrated in the intermediate zone as early as 9 postconceptional weeks (Kostovic´ and Goldman-Rakic, 1983). The thalamocortical fibers (Kostovic´ and Judasˇ, 2002) run below the cortical plate and parallel to the formation of the subplate zone at 11 postconceptional weeks, and spread throughout the subplate zone between 10.5 and 12 PW, reaching also dorsomedial aspects of the prefrontal cortex (figure 13.7 and plate 35). Thus the thalamocortical fibers form a plexiform area, waiting in the subplate zone for a prolonged period of at least 2 months. As already described, an enlarged subplate zone is visible on both histochemical and in vitro and in vivo MRI preparations (Kostovic´ et al., 2002; Judasˇ et al., 2005; Maas, Mukherjee, and CarballidoGamio, 2004) and may serve as a parameter of the development of thalamocortical pathways (figure 13.5).

fundamentals of developmental neurobiology

Figure 13.7 Sequential development of afferents from the dorsomedial nucleus of the thalamus (acetylcholinesterase-stained histological sections, with superimposed line drawings). Early outgrowth (A, 10.5 postconceptional weeks), spread through the subplate zone

(B, 20 postconceptional weeks), waiting in the upper part of the subplate zone (C, 23 postconceptional weeks), and penetration into the cortical plate (D, 28 postconceptional weeks) during ingrowth and address selection are visible. (See plate 35.)

The accumulation of thalamocortical afferents in the superficial part of the subplate zone occurs between 22 and 24 postconceptional weeks, during the process of target selection. This special phase of development of the thalamocortical pathways is visible on immunocytochemical preparations for demonstration of extracellular matrix molecules and on in vivo (in utero) MR images (Kostovic´ et al., 2006). Thalamocortical fibers from the mediodorsal nucleus penetrate the cortical plate between 24 and 28 postconceptional

weeks (Kostovic´, 1990), simultaneously with thalamocortical fibers destined to somatosensory (Kostovic´ et al., 1980), auditory (Krmpotic´-Nemanic´ et al., 1983), and visual cortices (Kostovic´ and Rakic, 1984). During the intracortical elaboration of thalamocortical fibers they express transient columnar distribution (Kostovic´, 1990; Kostovic´ and Goldman-Rakic, 1983). As stated previously, thalamocortical fibers interact synaptically with subplate neurons (Kostovic´ and Rakic, 1990;

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

223

Penn and Shatz, 1999; Hanganu, Kilb, and Luhmann, 2002) and represent a major input to the fetal cortex. The interaction of thalamocortical fibers with neurons of the subplate zone is both morphogenetic and functional (Kostovic´ and Judasˇ, 2006). The subplate zone is rich in extracellular matrix and axonal guidance molecules and thus has a morphogenetic role as both the substrate and a gradient zone for navigation of thalamocortical axons (Kostovic´ et al., 2002). With respect to function, glutamatergic thalamocortical afferents establish synapses with subplate neurons (Hanganu, Kilb, and Luhmann, 2002) and may have an inductive role. Similar inputs are probably present in human preterm infants, because after 24 postconceptional weeks numerous synapses are still present in the subplate zone, whereas the initial synaptogenesis occurs in the deep part of the cortical plate (Molliver, Kostovic´, and Van der Loos, 1973). The establishment of thalamocortical connections within the cortical plate in the somatosensory (Kostovic´ and Rakic, 1990), auditory (Krmpotic´-Nemanic´ et al., 1983), and visual cortices (Kostovic´ and Rakic, 1984) coincides with the appearance of evoked and event-related potentials (Graziani et al., 1974; Kostovic´ and Judasˇ, 2002, 2006). While the establishment of thalamocortical synapses in the sensory cortices after 24 postconceptional weeks is of paramount significance for consideration of interactions of the human fetus and preterm infants with the environment, the functional significance of thalamocortical connections in the prefrontal cortex remains obscure. However, it is important to realize that this connection is well developed in utero and may convey some premature functional interactions in the case of preterm birth. Development of corticocortical (associative and callosal) connections. Here we consider the development of long corticocortical commissural and associative connections—specifically, the development of the corpus callosum. The earliest commissural callosal fibers in the human telencephalon appear around 11 postconceptional weeks in the commissural plate, which develops in the telencephalon impar. During the first, early phase of callosal formation, pioneering fibers cross the telencephalic midline. It was shown that axonal guidance molecules (robo, slit, ephrins, semaphorins) discovered in different species are also active during the guidance of the human callosal axons (Ren et al., 2006). It was also confirmed that both neuronal and glial structures situated along the telencephalic midline are crucial cellular elements for production of axonal guidance molecules. The development of the corpus callosum during the later midfetal and late fetal periods can be successfully followed by magnetic resonance imaging. The increase in cross-sectional area of the corpus callosum before myelinization is a good indicator of the increase in the number and (possibly) size of callosal axons. However,

224

the size of the cross-sectional area does not reveal anything about the actual situation with axonal termination in the opposite hemisphere. Fetal increase in the cross-sectional area of the corpus callosum (Innocenti and Price, 2005) is followed by moderate decrease after 34 postconceptional weeks (Innocenti and Price, 2005). This was interpreted as a reduction in the number of exuberant axons (Innocenti and Price, 2005). It is unresolved as to whether exuberant callosal axons terminate around ventricles in white matter, the subplate zone, or the deep part of the cortical plate. According to Schwartz and Goldman-Rakic (1991), callosal projection to the cortical plate in late fetal monkeys has already begun an adultlike pattern of columnar distribution. This finding means that exuberant, supernumerary axons probably never reach the cortical plate. By all means, the developmental interaction between two hemispheres begins in utero, and callosal axons are important constituents of the transient subplate zone (Kostovic´ and Rakic, 1990), residing in the subplate zone longer than thalamocortical fibers, that is, after 32 postconceptional weeks (Kostovic´ and JovanovMilosˇevic´, 2006). This finding also means that callosal axons show growth potential in the late preterm infant and that the major reduction of exuberant callosal axons in humans occurs during the perinatal period. The exact developmental window in which reduction of callosal axons occurs is not known, but it may last until the sixth postnatal month (Innocenti and Price, 2005). In monkeys, the process of reducing the number of callosal axons takes place postnatally, leading ultimately to the pronounced reduction of the number of callosal axons (from 180 million to 50 million). It is important to emphasize that the reduction of callosal axons occurs before the reduction in number of synapses and probably reflects developmental, nonsynaptic interaction between hemispheres. Long associative connections. Long associative corticocortical connections are also established during the prenatal period. Experimental evidence in developing monkeys has shown that the parietal cortex projects to the frontal cortex as early as embryonic days E95/E105 (Schwartz and GoldmanRakic, 1991) and that its termination in columnar fashion resembles adult patterns (Schwartz and Goldman-Rakic, 1991). An indirect parameter of growth of long associative fibers may serve as a proxy for the presence of the subplate zone in the late preterm period, which may be gradually reduced because of ingrowth of callosal and association fibers (Kostovic´ and Rakic, 1990). In conclusion, sequential development of cortical fiber pathways, during prenatal life, leads to the basic “wiring.” This process is obviously under intrinsic control where genetic molecular mechanisms regulate the production of

fundamentals of developmental neurobiology

axonal guidance molecules, growth substrate in the extracellular matrix, and receptor-ligand interactions between growth cones and their cellular targets. During the period of growth, cortical pathways are located in vulnerable periventricular regions. One of the most vulnerable locations is the main crossroad of pathways located external to the lateral angle of the frontal horn of lateral vesicles (Judasˇ et al., 2005). Hypoxic/ischemic lesions in these periventricular areas cause damage to the fetal white matter known as periventricular leucomalacia.

Immature cortex of the newborn and early postnatal development of circuitry elements Laminar Organization and Cytoarchitectonics The prefrontal areas show a clear six-layered neocortical pattern of lamination, with prominent granularity of layer IV and visible pyramidal neurons of layers III and V. However, the reorganization into the mature postnatal pattern is not finished, and there are essential differences when compared to the cortex of a 1-year-old child, such as the presence of well-developed subplate neurons below layer VI at the border with white matter. Neuronal Differentiation A comprehensive review on neuronal development in the human prefrontal cortex during postnatal development has been presented by Mrzljak and colleagues (1990) and Petanjek and associates (2007). This study demonstrated that the progressive differentiation of pyramidal neurons continues after birth in both supragranular and infragranular layers. Pyramidal neurons of layer III exhibit well-developed basal and apical dendritic trees with growth of dendritic spines and axonal collaterals emanating from the main axonal arbor (figure 13.8). Pyramidal neurons in layer V of the newborn have larger cell bodies and dendritic trees than pyramidal neurons in the lower half of layer III. It is not until 9 postnatal months that basal dendrites of layer III reach the size of basal dendrites of layer-V neurons. The enlargement of basal dendrites of pyramidal layer-III and layer-V neurons may be a considerable factor in changing the cortical “dipole” toward the depth of the cortex, causing changes of polarity of cortical surface response (Kostovic´ and Judasˇ, 2002, 2006). Indeed, the predominantly negative surface potential of the late preterm period changes to a predominantly positive (adultlike) surface potential in the neonatal brain. During the second postnatal year, significant changes occur, and pyramidal neurons in the deep part of layer III become larger than those in layer V, a phenomenon that is the hallmark of “magnopyramidal” appearance. The largest pyramidal neurons are located in the deepest part of layer III, which is designated as layer IIIc.

Development of Dendritic Spines It is generally accepted that dendritic spines are the major postsynaptic elements in the cerebral cortex. In the fetal prefrontal cortex, dendritic spines develop relatively late, after 34 postconceptional weeks. After birth, there is rapid and massive production of dendritic spines. Rapid spinogenesis was observed on apical and basal dendrites of both layer III and layer V pyramidal neurons (figure 13.9). The number of dendritic spines increases steadily and reaches its maximum around 2.5 years. The number of dendritic spines remains high throughout late childhood, puberty, adolescence, and young adulthood (developmental plateau). After the third decade, however, there is the decrease in the total number of dendritic spines. Essentially, the curve corresponds to data of Huttenlocher and Dabholkar (1997) and Bourgeois, Goldman-Rakic, and Rakic (1994) (see also chapter 10 of this handbook), with some differences regarding the peak and the duration of the plateau. However, these data are based on more specimens from adolescence and childhood, and may better show developmental trends of synapse formation. Development of Interneurons Data about differentiation of interneurons in the human prefrontal cortex, as with primates, are very limited. The most complete study of localcircuit neurons in the postnatally developing and mature prefrontal cortex was done in the macaque monkey, which seems to be a good model for intrinsic circuitry (Lund and Lewis, 1993). Regarding morphological and chemical criteria, Lund and Lewis (1993) define 13 different classes of interneurons. As no further details were mentioned regarding changes in morphology and chemical expression during postnatal development, it might be suggested that interneurons do not undergo dramatic morphological and chemical differentiation during the postnatal period. This suggestion seems to be in accordance with data obtained on Golgi material of the human prefrontal cortex (Mrzljak et al., 1988, 1990), where it was found that axonal and dendritic differentiation of Golgi-impregnated interneurons (figure 13.10) was occurring mostly during the prenatal period, and it was suggested that differentiation of interneurons precedes the differentiation of projection neurons (Mrzljak et al., 1988, 1990). However, during the early postnatal period, interneurons are transiently covered by long, hairlike dendritic spines; later, during childhood and adolescence, no other changes were described (Mrzljak et al., 1988, 1990). Anderson and colleagues (1995) examined the postnatal development of layer-III pyramidal neurons in areas 9 and 46 of the rhesus monkey prefrontal cortex. In this study, quantitative reconstructions of Golgi-impregnated layer-III pyramidal neurons showed that the spine density increased by 50 percent during the first two postnatal months, remained at a plateau through 1.5 years of age, and

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

225

Figure 13.8 Development of pyramidal neurons from the early fetal period to 3 postnatal years. Golgi impregnation; all images are at the same magnification (A, 10 postconceptional weeks; B, 13 postconceptional weeks; C, 21 postconceptional weeks; D, 26 post-

226

conceptional weeks; E, 32 postconceptional weeks; F, newborn; G, 3 months postnatally; H, 3 years). Note the acceleration of dendritic development after 21 postconceptional weeks and the late appearance of dendritic spines.

fundamentals of developmental neurobiology

Figure 13.9 Histogram of dendritic spine development in the human prefrontal cortex. The development of dendritic spines is represented as estimated total spine number on basal dendritic tree of large layer-IIIc and layer-V pyramidal neurons in the human

prefrontal cortex in correlation with different postnatal ages; the total number of dendritic spines is estimated according to the data on total dendritic length and dendritic spine density.

then decreased over the peripubertal age range until stable adult levels were achieved. Anderson and colleagues (1995) determined that the density of parvalbumin-immunoreactive axon terminals (cartridges) belonging to the chandelier class of local-circuit neurons exhibited a temporal pattern of change that exactly paralleled the changes in dendritic spine density.

after the establishment and reshaping of long (extrinsic) cortical connections during the late fetal and perinatal periods, synaptogenesis in the child brain is the predominant event related to the development of intracortical circuitry.

Development of Intrinsic (Intracortical, Local) Circuitry Very little is known about the postnatal development of intrinsic cortical circuitry in the human prefrontal cortex. The studies of local circuitry development in the human visual cortex reveal that intracortical connections begin to develop during the fetal period (Hevner, 2005) but show prolonged maturation and reach mature form sometime before 15 months of age (Burkhalter, Bernardo, and Charles, 1993). Since the prefrontal cortex shows prolonged maturation by all available parameters (Chugani, Phelps, and Mazziotta, 1987; Huttenlocher and Dabholkar, 1997; Petanjek et al., 1994; Fuster, 2002; Sowell et al., 1999, 2003; Giedd et al., 1999), it is logical to assume prolonged development of local circuitry during early childhood, particularly during the second and third years of life. Our finding of the highest number of dendritic spines around 2.5 years is in accordance with this concept. Namely,

Development of circuitry organization: An overview Fetal (Endogenous) Circuitry Synapses, strategic points of neuronal interaction, are present from the 9th postconceptional week in two laminae: above and below the cortical plate. After the formation of the subplate zone, synapses are present throughout these prominent laminae. The majority of synapses at that very early age are asymmetric and are located on dendrites and somata of subplate neurons and on dendrites of unknown origin. The early differentiation of subplate neurons, together with the presence of numerous synapses, represents an early substrate of fetal functions. The neurotransmitters of subplate neurons are GABA and glutamate with the presence of numerous peptidergic modulators as described for the carnivora (Penn and Shatz, 1999), monkey (Rakic, Ang, and Breunig, 2004), and human cortices (Kostovic´ and Rakic, 1990; Dellale et al., 1997; Kostovic´ et al., 1991; Zecevic and Milosevic, 1997). The function of early fetal circuitry is endogenous (spontaneous)

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

227

Figure 13.10 Development of nonpyramidal neurons from the fetal period to adulthood . (A) Interneuron in the adult prefrontal cortex; dendrites are thin and not covered with spines; arrow indicates the origin of axon. (B) Double-bouquet interneuron (32 postconceptional weeks); arrow indicates the axon. (C) Basket type of interneuron with long horizontal and vertical axon branches

228

(arrows) in the newborn cortex. (D) The chandelier-type interneuron with extensive axonal arborization (arrows) in three-month-old infant cortex. (E) The double-bouquet type interneuron with vertically organized axonal and dendritic arborizations (arrows) in the one-month-old infant cortex.

fundamentals of developmental neurobiology

activity as described in experimental mammals (Kanold et al., 2003; Penn and Shatz, 1999; Hanganu, Kilb, and Luhmann, 2002). Preterm Infants and Newborns: Coexistence of Endogenous (Spontaneous) and Sensory-Driven Circuitry In the preterm infant, there is continuation of the fetal pattern of endogenous (spontaneous) circuitry, concentrated on early differentiating subplate neurons and postsynaptic elements in the marginal zone. The major developmental change is the accumulation of thalamocortical fibers below the cortical plate after 21 postconceptional weeks and their penetration of the cortical plate after 24 postconceptional weeks (Kostovic´ and Goldman-Rakic, 1983; Kostovic´ and Judasˇ, 2002). Parallel to this event, intensive synaptogenesis begins in the deep part of the cortical plate (Molliver, Kostovic´, and Van der Loos, 1973). The onset of synaptogenesis in the cortical plate, between ingrowing thalamocortical fibers and postsynaptic elements in the cortical plate, is crucial for the new phase of circuitry development, which may be sensory driven because thalamic fibers carry impulses from the sensory periphery to the cortex and may activate intracortical synapses. This establishment of permanent thalamocortical circuitry is the anatomical basis for the first appearance of the first evoked potentials (Kostovic´ and Judasˇ, 2002, 2006). It was not known whether pain stimuli could activate the cortex by way of thalamocortical pathways at this early stage (S. Lee et al., 2005). Very recently, Slater and colleagues (2006) have shown cortical response to pain using infrared monitoring (optical imaging) of cortical blood flow after the application of pain stimuli. This indicates that the first permanent thalamocortical synaptic link is not only active in initiating general somatosensory stimuli but also may underlie cortical activation through pain stimuli. Coexistence of both transient endogenous (spontaneously active) circuitry and permanent sensory-driven circuitry from 24 postconceptional weeks throughout the preterm period (figure 13.11 and plate 36) is a salient feature of the human preterm cortex (Kostovic´ and Judasˇ, 2006, 2007). It was previously described in the fetal cat (Penn and Shatz, 1999) and rat (Hanganu, Kilb, and Luhmann, 2002), but this phenomenon seems to be even more significant in humans because of the extreme richness of corticocortical connections. This process may be especially prominent in the prefrontal cortex, where the transient subplate zone is particularly thick (Kostovic´, 1990) and corticocortical pathways are very abundant. If exuberant axons (Innocenti and Price, 2005) contribute synaptically to this transient circuitry, then we may have a very complex transient circuitry arrangement in the preterm cortex. Some of the neurotransmitters in this transient circuitry are known from experimental studies (Penn and Shatz, 1999; Hanganu, Kilb, and Luhmann,

2002). The two main transmitters in subplate neurons, GABA and glutamate, are modulated by glutamatergic fibers from the thalamus by way of two types of receptors (NMDA and AMPA) and GABA fibers from other subplate neurons. The same thalamic afferent fiber may contact subplate neurons and cells in the cortical plate (Penn and Shatz, 1999). The early transient and permanent circuitry activity in the subplate zone is probably an important factor in the development of neuronal damage by means of glutamatergic neurotoxic sequelae after hypoxia-ischemia. The participation of other neurotransmitters in the preterm circuitry is largely unknown. It was found that cholinergic afferent fibers from the basal forebrain, being present from the early fetal phases, overlap in their laminar distribution with thalamocortical fibers (Kostovic´ and GoldmanRakic, 1983). According to this observation, cholinergic afferents from the basal forebrain overlap with thalamocortical afferents from the mediodorsal nucleus in the subplate zone, accumulate below the cortical plate around 21–23 postconceptional weeks, and penetrate the cortical plate after 24 postconceptional weeks. The exact postsynaptic site for cholinergic afferents in the subplate zone is yet to be determined, but the strong modulatory effect of activated cholinergic receptors on the early transient network was documented in a recent experimental study (Hanganu, Kilb, and Luhmann, 2002). Monoaminergic systems are also present in the preterm cortex, although there is no direct anatomical evidence for their organization in the preterm cortex. This conclusion can be reached on the basis of their early arrival in the fetal cortex (Verney et al., 1993; Verney, 1999; Nobin and Björklund, 1973) and prolonged maturation (Benes, 2001). In conclusion, the organization of cortical circuitry in the preterm infant is characterized by the coexistence of transient endogenous (spontaneously active) and permanent sensory-driven circuitry with gradual elaboration of input to the cortical plate. Connectivity of the preterm infant cortex also undergoes dynamic changes because major cortical pathways actively grow and relocate during the phase of cortical target selection. Neonatal circuitry. The main characteristic of the frontal circuitry is the disappearance of a transient pattern of organization, cessation of growth of major pathways, and intensive synaptogenesis in the cortical plate layers. The transient subplate zone is reduced to a transitional lamina between layer VI and white matter, but subplate neurons are still visible and contain synapses (Kostovic´ and Rakic, 1980, 1990). Thalamic projection begins intracortical elaboration within layer IV with the development of different classes of interneurons (Mrzljak et al., 1988, 1990). There are laminar shifts in somatostatin (Kostovic´ et al., 1991) and neuropeptide Y neurons (Delalle et al., 1997). Acetylcholinesterase-

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

229

Figure 13.11 Transient circuitry of the human fetal cortex coexists with permanent circuitry elements. CP, cortical plate; PYR, pyramidal neurons; SP, subplate zone. (See plate 36.)

reactive fibers from the basal forebrain develop dense networks throughout all cortical layers (Kostovic´, 1990; Kostovic´, Škavic´, and Strinovic´, 1988). The postsynaptic elements on pyramidal neurons are predominantly dendrites and dendritic spines (Petanjek et al., 1994), which begin to rapidly develop during the neonatal phase. The reorganization of circuitry and the disappearance of transient patterns of organization in the neonatal period are widened through two additional events: first, the dynamic growth of short corticocortical connections (Burkhalter, Bernardo, and Charles, 1993), and second, exuberance and reduction of callosal axons (Innocenti and Price, 2005). As stated previously, the indirect sign of growth of short corti-

230

cocortical connections is the prolonged existence of the subplate zone in the prefrontal cortex. Infancy and Childhood: Overproduction of Circuitry Elements and Interaction with Environment After laminar reorganization in the neonatal period, the development of circuitry in childhood proceeds at synaptic levels in both qualitative and quantitative terms. In quantitative terms, there is rapid increase in number of postsynaptic dendritic spines (figure 13.9). The maximum number of dendritic spines was seen around 2.5 years during language-cognitive maturation. These quantitative changes in the number of dendritic spines occur in parallel

fundamentals of developmental neurobiology

to the change in the total number of synapses. In addition to the general quantitative parameters (Travis, Ford, and Jacobs, 2005), there are changes in the number of individual types of postsynaptic elements (dendritic spines) and asymmetric synapses, indicating selective changes in synaptic processing throughout childhood. The “cognitive” pyramidal neurons of layer III, the hallmark of the prefrontal cortex, develop their dendritic trees up to 3 years of age (Petanjek et al., 1994, 2007). It is not known when in the human brain the transient developmental excitatory function of GABA interneurons (Vanhatalo et al., 2005) changes into the “classical” inhibitory function of the adult brain. The postnatal development of the neurotransmitter systems in the human prefrontal cortex is largely unexplored. Indirect evidence can be derived from pharmacological and clinical trials. Correlation of reward-seeking behavior and brain dopamine activity suggests an accelerated maturation of the dopaminergic system during the first postnatal year (Goldman-Rakic et al., 1992; Goldman-Rakic, 1999; Luciana, 2001). In conclusion, the circuitry in childhood shows the highest level of synaptic contacts on dendritic spines, and this developmental increase reaches a maximum around 2.5 years of age, during cognitive function of the prefrontal cortex (Nelson, 1999; Nelson et al., 2000). Synaptogenesis and spinogenesis occur in the conditions of intensive interactions with the environment. This fact opens a possibility of environmental influence in the process of maintenance and elimination of synapses. Adolescent and Postadolescent Period: Prolonged Plasticity and Reorganization Little is known about the developmental changes in circuitry organization during adolescence and early postadolescence and the structural properties of neural pathways (Paus et al., 1999; Casey et al., 2005). Until recently, the best known change was myelinization, which extends well into the third decade of life. However, research during the last 15 years has shown other substantial changes: changes in the number of synapses and dendritic spines, protracted chemical maturation (Kostovic´, Škavic´, and Strinovic´, 1988), general maturation of gray matter “composition” (Sowell et al., 1999; Giedd et al., 1999), and overall changes of the intrinsic circuitry (Woo et al., 1997). The high production and turnover of synapses during childhood and early adolescence correspond to increased brain activity in the prefrontal cortex, which underlies the development of visuospatial working memory (Klingberg, Forssberg, and Westerberg, 2002; Luciana and Nelson, 1998; Nelson et al., 2000). Despite the attempts to correlate cognitive development with the maturation of cortical circuitry in the frontal lobe, the exact neurological substrate of cognitive development remains obscure. However, studies on different neurobiological parameters

should be continued. In this respect, longitudinal studies are very promising, as they permit the study of the structurefunction relationship within the same individuals during cognitive development (Sowell et al., 2004). We have already described in this chapter (figure 13.9) changes in the number of dendritic spines, indicating that the production and number of dendritic spines is still high in the adolescent and young postadolescent brain. These data are complementary to the general curve of the number of synapses (Huttenlocher and Dabholkar, 1997; Burgeois, Goldman-Rakic, and Rakic, 1994) based on electronmicroscopic quantification of synapses, where an adequate number of specimens was not available for the late stages of development. At the level of neurotransmitter-related innervation of cortical neurons, we have no direct evidence for changes in the human prefrontal cortex. However, it was shown that dopamine D1 receptors increase in the rhesus monkey prefrontal cortex until puberty (Lewis et al., 2004). Since the dopaminergic ventral tegmental area system interacts with several types of receptors, we can expect changes in dopaminergic postsynaptic receptors. Modern neuroimaging methods offer a new opportunity for in vivo studies of gray matter. Sowell and colleagues (1999, 2004) reported a reduction in gray matter density between adolescence and young adulthood, while Giedd and colleagues (1999) observed a decline in gray matter size during postadolescence. These changes are in accordance with electrophysiologic and cerebral glucose metabolism studies of gray matter (Chugani, Phelps, and Mazziotta, 1987).

Development of microstructural asymmetry and hemispheric specialization in the frontal lobe Surprisingly little is known about the microstructural development of cytoarchitectonic asymmetry and hemispheric specialization of the human frontal lobe, and the few existing studies have been almost exclusively focused on the development of Broca’s region (i.e., opercular and triangular parts of the inferior frontal gyrus, corresponding to Brodmann’s areas 44 and 45) as a putative structural correlate of the maturation of language dominance (Simonds and Scheibel, 1989; Amunts et al., 2003). Simonds and Scheibel (1989) followed the maturation of motor speech areas and their adjacent orofacial motor zones in the right and left hemispheres of the human infant from 3 months to 72 months of age. They quantitatively studied basilar dendrite patterns of layer V pyramidal neurons and found that at 3 postnatal months dendrites are more developed in orofacial motor zones than in Broca’s region and in the right hemisphere than in the left hemisphere (Simonds and Scheibel, 1989). However, during subsequent development, basal dendritic

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

231

arborizations become longer and more complex in Broca’s region than in the orofacial cortex, and the total length of dendrite systems in the left hemisphere finally exceeds those in the right hemisphere. However, it should be noted that even at 72 postnatal months, the length of distal dendritic segments in the right Broca’s region still exceeds that of the left, suggesting that maturation of Broca’s region is not complete at this stage (Simonds and Scheibel, 1989). In a recent cytoarchitectonic developmental study of areas 44 and 45, Amunts and colleagues (2003) found that (1) asymmetry was already present in 1-year-old infants, but tended to increase with age (which was significant in area 45, but not in area 44); (2) an adultlike, left-larger-than-right asymmetry in the volume fraction of cell bodies (the so-called gray level index, GLI) was reached at approximately 5 years in area 45 and 11 years in area 44; and (3) interhemispheric asymmetry in the cytoarchitecture of areas 44 and 45 continues to change throughout life. As the decrease in GLI value signifies an increase in the volume of neuropil, these findings suggest that (1) compared with infancy and childhood, areas 44 and 45 showed relatively more neuropil during adulthood, that is, greater volume proportion of dendrites, axons, and synapses as the basis of cortical connectivity; (2) the development of neuropil in areas 44 and 45 is most intense during infancy and early childhood (4–5 years) and only marginally so in late childhood and adulthood; and (3) an adultlike cytoarchitectonic asymmetry (with relative neuropil volume significantly higher in the dominant, left hemisphere) developed relatively late in childhood (Amunts et al., 2003). Taken together, these data suggest that the microstructural development of Broca’s region roughly parallels a dramatic increase in language abilities during the first 2 years of life (Kuhl et al., 1997); these data also coincide with functional data demonstrating left-hemispheric language lateralization in children at 7 (B. Lee et al., 1999) or 8 years of age (Balsamo et al., 2002). However, Amunts and colleagues (2003) justly point out that developmental changes in interhemispheric asymmetry did not stop by the end of childhood but reached into adulthood, as suggested previously by Jacobs and Scheibel (1993). Thus the evidence available at present suggests an influence of language practice on brain anatomy and leads to the hypothesis that the delayed maturation is the microstructural basis of the development of language abilities and that microstructural plasticity of Broca’s region endures throughout almost the whole life span (Amunts et al., 2003).

REFERENCES Amunts, K., A. Schleicher, A. Ditterich, and K. Zilles, 2003. Broca’s region: Cytoarchitectonic asymmetry and developmental changes. J. Comp. Neurol. 465:72–89.

232

Anderson, S. A., J. D. Classey, F. Condé, J. S. Lund, and D. A. Lewis, 1995. Synchronous development of pyramidal neuron dendritic spines and parvalbumin-immunoreactive chandelier neuron axon terminals in layer III of monkey prefrontal cortex. Neuroscience 67(1):7–22. Balsamo, L. M., B. Xu, C. B. Grandin, J. R. Petrella, S. H. Braniecki, T. K. Elliott, and W. D. Gaillard, 2002. A functional magnetic resonance imaging study of left hemisphere language dominance in children. Arch. Neurol. 59:1168– 1174. Benes, F. M., 2001. The development of prefrontal cortex: The maturation of neurotransmitter systems and their interactions. In Handbook of Developmental Cognitive Neuroscience, ed. C. A. Nelson and M. Luciana, 79–92. Cambridge, MA: MIT Press. Bourgeois, J.-P., P. S. Goldman-Rakic, and P. Rakic, 1994. Synaptogenesis in the prefrontal cortex of rhesus monkeys. Cerebral Cortex 4:78–96. Brodmann, K., 1909. Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Leipzig: Johann Ambrosius Barth. Burkhalter, A., K. L. Bernardo, and V. Charles, 1993. Development of local circuits in human visual cortex. J. Neurosci. 13(5):1916–1931. Casey, B. J., N. Tottenham, C. Liston, and S. Durston, 2005. Imaging the developing brain: What have we learned about cognitive development? Trends Cogn. Sci. 9(3):104–110. Chugani, H. T., M. E. Phelps, and J. C. Mazziotta, 1987. Positron emission tomography study of human brain functional development. Ann. Neurol. 22(4):487–497. Delalle, I., P. Evers, I. Kostovic´, and H. B. M. Uylings, 1997. Laminar distribution of neuropeptide Y-immunoreactive neurons in human prefrontal cortex during development. J. Comp. Neurol. 379:515–522. Fuster, J. M., 2002. Frontal lobe and cognitive development. J. Neurocytol. 31:373–385. Germuska, M., S. Saha, J. Fiala, and H. Barbas, 2006. Synaptic distinction of laminar-specific prefrontal-temporal pathways in primates. Cerebral Cortex 16:865–875. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, T. Paus, A. C. Evans, and J. L. Rapoport, 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2(10):861–863. Giguere, M., and P. S. Goldman-Rakic, 1988. Mediodorsal nucleus: Areal, laminar and tangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys. J. Comp. Neurol. 277:195–213. Glenn, O. A., and A. J. Barkovich, 2006. Magnetic resonance imaging of the fetal brain and spine: An increasing important tool in prenatal diagnosis, Part 1. AJNR Am. J. Neuroradiol. 27:1604–1611. Goldman-Rakic, P. S., 1981. Prenatal formation of cortical input and development of cytoarchitectonic compartments in the neostriatum of the rhesus monkey. J. Neurosci. 1(7):721–735. Goldman-Rakic, P. S., 1995. Cellular basis of working memory. Neuron 14:477–485. Goldman-Rakic, P. S., 1999. The physiological approach: Functional architecture of working memory and disordered cognition in schizophrenia. Biol. Psychiatry 46:650–661. Goldman-Rakic, P. S., M. S. Lidow, J. F. Smiley, and M. S. Williams, 1992. The anatomy of dopamine in monkey and human prefrontal cortex. J. Neural Transm. Suppl. 36:163–177. Graziani, L. J., L. Katz, R. Q. Cracco, J. B. Cracco, and E. D. Weitzman, 1974. The maturation and interrelationship of EEG

fundamentals of developmental neurobiology

patterns and auditory evoked responses in premature infants. Electroencephalogr. Clin. Neurophysiol. 36:367–375. Hanganu, I. L., W. Kilb, and H. J. Luhmann, 2002. Functional synaptic projections onto subplate neurons in neonatal rat somatosensory cortex. J. Neurosci. 22:7165–7176. Hevner, R. F., 2005. Development of connections in the human visual system during fetal midgestation: A DiI-tracing study. J. Neuropathol. Exp. Neurol. 59(5): 385–392. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387(2):167–178. Innocenti, G. M., and D. J. Price, 2005. Exuberance in the development of cortical networks. Nature Rev. Neurosci. 6:955–965. Jacobs, B., and A. B. Scheibel, 1993. A quantitative dendritic analysis of Wernicke’s area in humans. I. Lifespan changes. J. Comp. Neurol. 327:384–396. Judasˇ, M., M. Radosˇ, N. Jovanov-Milosˇevic´, P. Hrabacˇ, R. Štern-Padovan, and I. Kostovic´, 2005. Structural, immunocytochemical and MRI properties of periventricular crossroads of growing cortical pathways in preterm infants. AJNR Am. J. Neuroradiol. 26:2671–2784. Judasˇ, M., N. Šestan, and I. Kostovic´, 1999. Nitrinergic neurons in the developing and adult human telencephalon: Transient and permanent patterns of expression in comparison to other mammals. Microsc. Res. Tech. 45:401–419. Kanold, P. O., P. Kara, R. C. Reid, and C. J. Shatz, 2003. Role of subplate neurons in functional maturation of visual cortical columns. Science 301:521–525. Klingberg, T., H. Forssberg, and H. Westerberg, 2002. Incrreased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14(1):1–10. Kordower, J. H., P. Piecinski, and P. Rakic, 1992. Neurogenesis of the amygdaloid nuclear complex in the rhesus monkey. Dev. Brain Res. 68:9–15. Kostovic´, I., 1986. Prenatal development of nucleus basalis complex and related fiber systems in man: A histochemical study. Neuroscience 17:1047–1077. Kostovic´, I., 1990. Structural and histochemical reorganization of the human prefrontal cortex during perinatal and postnatal life. Prog. Brain Res. 85:223–239. Kostovic´, I., and P. S. Goldman-Rakic, 1983. Transient cholinesterase staining in the mediodorsal nucleus of the thalamus and its connections in the developing human and monkey brain. J. Comp. Neurol. 219:431–447. Kostovic´, I., and N. Jovanov-Milosˇevic´, 2006. The development of cerebral connections during the first 20–45 weeks’ gestation. Sem. Fetal Neonatal. Med. 11:415–422. Kostovic´, I., and M. Judasˇ, 2002. Correlation between the sequential ingrowth of afferents and transient patterns of cortical lamination in preterm infants. Anat. Rec. 267:1–6. Kostovic´, I., and M. Judasˇ, 2006. Prolonged coexistence of transient and permanent circuitry elements in the developing cerebral cortex of fetuses and preterm infants. Dev. Med. Child Neurol. 48:388–393. Kostovic´, I., and M. Judasˇ, 2007. Transient patterns of cortical lamination during prenatal life: Do they have implications for treatment? Neurosci. Biobehav. Rev. 31:1157–1168. Kostovic´, I., M. Judasˇ, and N. Bogdanovic´, 1987. Postnatal maturation of human frontal cortex: New cytoarchitectonic criteria. Soc. Neurosci. Abstracts 13:1099. Kostovic´, I., M. Judasˇ, M. Radosˇ, and P. Hrabacˇ, 2002. Laminar organization of the human fetal cerebrum revealed by histo-

chemical markers and magnetic resonance imaging. Cerebral Cortex 12:536–544. Kostovic´, I., M. Judasˇ, S. Škrablin-Kucˇic´, R. Štern-Padovan, M. Radosˇ, 2006. In vivo MR imaging of transient subplate zone in the human fetal telencephalon. Society for Neuroscience Meeting (Atlanta) Abstracts, No. 96.10. Kostovic´, I., Z. Kelovic´, J. Krmpotic´-Nemanic´, and I. Kracˇun, 1980. Development of the human somatosensory cortex: Laminar distribution and vertical organization of the acetylcholinesterase positive fibres during fetal life. Neurosci. Lett. (Suppl.) 5:S484. Kostovic´, I., and P. Rakic, 1984. Development of prestriate visual projections in the monkey and human fetal cerebrum revealed by transient cholinesterase staining. J. Neurosci. 4:25– 42. Kostovic´, I., and P. Rakic, 1990. Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain. J. Comp. Neurol. 297: 441–470. Kostovic´, I., J. Škavic´, and D. Strinovic´, 1988. Acetylcholinesterase in the human frontal associative cortex during the period of cognitive development: Early laminar shifts and late innervation of pyramidal neurons. Neurosci. Lett. 90:107–112. Kostovic´, I., A. Štefulj-Fucˇic´, L. Mrzljak, S. Jukic´, and I. Delalle, 1991. Prenatal and perinatal development of the somatostatin-immunoreactive neurons in the human prefrontal cortex. Neurosci. Lett. 124:153–156. Kritzer, M. F., and P. S. Goldman-Rakic, 1995. Intrinsic circuit organization of the major layers and sublayers of the dorsolateral prefrontal cortex in the rhesus monkey. J. Comp. Neurol. 359:131– 143. Krmpotic´-Nemanic´, J., I. Kostovic´, Z. Kelovic´, D. Nemanic´, and L. Mrzljak, 1983. Development of the human fetal auditory cortex: Growth of afferent fibers. Acta Anat. 116:69–73. Kuhl, P. K., J. E. Andruski, I. A. Chistovich, L. A. Chistovich, E. V. Kozhevnikova, V. L. Ryskina, E. I. Stolyarova, U. Sundberg, and F. Lacerda, 1997. Cross-language analysis of phonetic units in language addressed to infants. Science 277: 684–686. Lee, B. C., K. Kuppusamy, R. Grueneich, O. El-Ghazzawy, R. E. Gordon, W. Lin, and E. M. Haacke, 1999. Hemispheric language dominance in children demonstrated by functional magnetic resonance imaging. J. Clin. Neurophysiol. 14:78–82. Lee, S. J., H. J. Ralston, E. A. Drey, J. C. Partridge, and M. A. Rosen, 2005. Fetal pain: A systematic multidisciplinary review of the evidence. JAMA 294(8):947–954. Letinic, K., R. Zoncu, and P. Rakic, 2002. Origin of GABAergic neurons in the human neocortex. Nature 417:645–649. Lewis, D. A., D. Cruz, S. Eggan, and S. Erickson, 2004. Postnatal development of prefrontal inhibitory circuits and the pathophysiology of cognitive dysfunction in schizophrenia. Ann. NY Acad. Sci. 1021:64–76. Lewis, D. A., D. S. Melchitzky, and G. G. Burgos, 2002. Specificity in the functional architecture of primate prefrontal cortex. J. Neurocytol. 31:265–276. Luciana, M., 2001. Dopamine-opiate modulations of rewardseeking behavior: Implications for the functional assessment of prefrontal development. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 647–662. Cambridge, MA: MIT Press. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36:273–293.

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

233

Lund, J. S., and D. A. Lewis, 1993. Local circuit neurons of developing and mature macaque prefrontal cortex: Golgi and immunocytochemical characteristics. J. Comp. Neurol. 328: 282–312. Luskin, M. B., and C. J. Shatz, 1985. Studies of the earliest generated cells of the cat’s visual cortex: Cogeneration of subplate and marginal zones. J. Neurosci. 5(4):1062–1075. Maas, L. C., P. Mukherjee, J. Carballido-Gamio, et al., 2004. Early laminar organization of the human cerebrum demonstrated with diffusion tensor imaging in extremely premature infants. NeuroImage 22:1134–1140. Machado, C. J., and J. Bachevalier, 2003. Non-human primate models of childhood psychopathology: The promise and the limitations. J. Child Psychol. Psychiatry 44(1):64–87. Melchitzky, D. S., G. Gonzalez-Burgos, G. Barrionuevo, and D. A. Lewis, 2001. Synaptic targets of the intrinsic axon collaterals of supragranular pyramidal neurons in monkey prefrontal cortex. J. Comp. Neurol. 430:209–221. Melchitzky, D. S., and D. A. Lewis, 2003. Pyramidal neuron local axon terminals in monkey prefrontal cortex: Differential targeting of subclasses of GABA neurons. Cerebral Cortex 13:452– 460. Mesulam, M. M., 1998. From sensation to cognition. Brain 121(6): 1013–1052. Mesulam, M. M., E. J. Mufson, A. I. Levey, and B. H. Wainer, 1983. Cholinergic innervation of cortex by the basal forebrain: Cytochemistry and cortical connections of the septal area, diagonal band nuclei, nucleus basalis (substantia innominata) and hypothalamus in the rhesus monkey. J. Comp. Neurol. 214:170– 197. Meyer, G., and A. M. Goffinet, 1998. Prenatal development of reelin-immunoreactive neurons in the human neocortex. J. Comp. Neurol. 397:29–40. Molliver, M. E., I. Kostovic´, and H. Van der Loos, 1973. The development of synapses in cerebral cortex of the human fetus. Brain Res. 50:403–407. Mrzljak, L., and P. S. Goldman-Rakic, 1992. Acetylcholinesterase reactivity in the frontal cortex of human and monkey: Contribution of AChE-rich pyramidal neurons. J. Comp. Neurol. 324:261–281. Mrzljak, L., H. B. M. Uylings, I. Kostovic´, and C. G. Van Eden, 1988. Prenatal development of neurons in the human prefrontal cortex. I. A qualitative Golgi study. J. Comp. Neurol. 271:355– 386. Mrzljak, L., H. B. M. Uylings, I. Kostovic´, and C. G. Van Eden, 1992. Prenatal development of neurons in the human prefrontal cortex. II. A quantitative Golgi study. J. Comp. Neurol. 316: 485–496. Mrzljak, L., H. B. M. Uylings, C. G. Van Eden, and M. Judasˇ, 1990. Neuronal development in human prefrontal cortex in prenatal and postnatal stages. Prog. Brain Res. 85:185–222. Nauta, W. J. H., and W. Haymaker, 1969. Hypothalamic nuclei and fiber connections. In W. Haymaker, E. Anderson, and W. J. H. Nauta, eds., The Hypothalamus, 136–209. Springfield, IL: Charles C. Thomas. Nelson, C. A., 1999. Change and continuity in neurobehavioral development: Lessons from the study of neurobiology and neural plasticity. Infant Behav. Dev. 22(4):415–429. Nelson, C. A., C. S. Monk, J. Lin, L. J. Carver, K. M. Thomas, and C. L. Truwit, 2000. Functional neuroanatomy of spatial working memory in children. Dev. Psychol. 36:109–116. Nikolic´, I., and I. Kostovic´, 1986. Development of the lateral amygdaloid nucleus in the human fetus: Transient presence of

234

discrete cytoarchitectonic units. Anat. Embryol. (Berl.) 174:355– 360. Nobin, A., and A. Björklund, 1973. Topography of the monoamine neuron systems in the human brain as revealed in fetuses. Acta Physiol. Scand. (Suppl.) 388:1–40. Paus, T., A. Zijdenbos, K. Worsley, D. L. Collins, J. Blumenthal, J. N. Giedd, J. L. Rapoport, and A. C. Evans, 1999. Structural maturation of neural pathways in children and adolescents: In vivo study. Science 283:1908–1911. Penn, A. A., and C. J. Shatz, 1999. Brain waves and brain wiring: The role of endogenous and sensory-driven neural activity in development. Pediatr. Res. 45:447–458. Petanjek, Z., M. Judasˇ, I. Kostovic´, and H. B. M. Uylings, 2007. Lifespan alterations of basal dendritic trees of pyramidal neurons in the human prefrontal cortex: A layer-specific pattern. Cerebral Cortex Advance Access published July 25, 2007 (doi:10.1093/cercor/bhm124). Petanjek, Z., G. Šimic´, M. Judasˇ, E. Radonic´, and I. Kostovic´, 1994. Postnatal development of dendritic spines on layer IIIc pyramidal neurons in the human prefrontal cortex (Abstract Book of the 17th ENA Annual Meeting in Vienna, 4–8 September). Eur. J. Neurosci. (Suppl No. 7):49. Pucak, M. L., J. B. Levitt, J. S. Lund, and D. A. Lewis, 1996. Patterns of intrinsic and associational circuitry in monkey prefrontal cortex. J. Comp. Neurol. 376:614–630. Rajkowska, G., and P. S. Goldman-Rakic, 1995. Cytoarchitectonic definition of prefrontal areas in the normal human cortex. 1. Remapping of areas 9 and 46 using quantitative criteria. Cerebral Cortex 5:307–322. Rakic, P., S. B. C. Ang, and J. Breunig, 2004. Setting the stage for cognition: Genesis of the primate cerebral cortex. In M. S. Gazzaniga, ed. The Cognitive Neurosciences III, 33–49. New York: MIT Press. Rakic, P., and N. Zecevic, 2003. Emerging complexity of layer I in human cerebral cortex. Cerebral Cortex 13:1072–1083. Ren, T., A. Anderson, W. B. Shen, H. Huang, C. Plachez, J. Zhang, S. Mori, S. L. Kinsman, and L. J. Richards, 2006. Imaging, anatomical, and molecular analysis of callosal formation in the developing human fetal brain. Anat. Rec. (Part A) 288A:191–204. Schwartz, M. L., and P. S. Goldman-Rakic, 1991. Prenatal specification of callosal connections in rhesus monkey. J. Comp. Neurol. 307:144–162. Shaw, P., D. Greenstein, J. Lerch, I. Clasen, R. Lenroot, N. Gogtay, A. Evans, J. Rapaport, and J. Giedd, 2006. Intellectual ability and cortical development in children and adolescents. Nature 440(7084):676–679. Šimic´, G., L. Mrzljak, A. Fucˇic´, B. Winblad, H. Lovric´, and I. Kostovic´, 1999. Nucleus subputaminalis (Ayala): The still disregarded magnocellular component of the basal forebrain may be human specific and connected with the cortical speech area. Neuroscience 89(1):73–89. Simonds, R. J., and A. B. Scheibel, 1989. The postnatal development of the motor speech area: A preliminary study. Brain Lang. 37:42–58. Slater, R., A. Cantarella, S. Gallella, A. Worley, S. Boyd, J. Meek, and M. Fitzgerald, 2006. Cortical pain responses in human infants. J. Neurosci. 26(14):3662–3666. Sowell, E. R., B. S. Peterson, P. M. Thompson, S. E. Welcome, A. L. Henkenius, and A. W. Toga, 2003. Mapping cortical change across the human life span. Nature Neurosci. 6(3):309–315. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 1999. In vivo evidence for postadolescent brain

fundamentals of developmental neurobiology

maturation in frontal and striatal regions. Nature Neurosci. 2(10): 859–861. Sowell, E. R., P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, and A. W. Toga, 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24(38):8223–8231. Travis, K., K. Ford, and B. Jacobs, 2005. Regional dendritic variation in neonatal human cortex: A quantitative Golgi study. Dev. Neurosci. 27:277–287. Vanhatalo, S., J. M. Palva, S. Andersson, C. Rivera, J. Voipio, and K. Kaila, 2005. Slow endogenous activity transients and developmental expression of K+Cl− cotransporter 2 in the immature human cortex. Eur. J. Neurosci. 22:2799–2804. Verney, C., 1999. Distribution of the catecholaminergic neurons in the central nervous system of human embryos and fetuses. Microsc. Res. Tech. 46:24–47.

Verney, C., A. Milosevic, C. Alvarez, and B. Berger, 1993. Immunocytochemical evidence of well-developed dopaminergic and noradrenergic innervations in the frontal cerebral cortex of human fetuses at midgestation. J. Comp. Neurol. 336:331– 344. Wonders, C. P., and S. A. Anderson, 2006. The origin and specification of cortical interneurons. Nature Rev. Neurosci. 7:687– 696. Woo, T. U., M. L. Pucak, C. H. Kye, C. V. Matus, and D. A. Lewis, 1997. Peripubertal refinement of the intrinsic and associational circuitry in monkey prefrontal cortex. Neuroscience 80: 1149–1158. Zecevic, N., and A. Milosevic, 1997. Initial development of gamma-aminobutyric acid immunoreactivity in the human cerebral cortex. J. Comp. Neurol. 380:495–506.

kostovic´, judasˇ, and petanjek: development of the human prefrontal cortex

235

14

White Matter Maturation and Cognitive Development during Childhood TORKEL KLINGBERG

Myelination is an important process that starts during gestation and continues throughout childhood and early adulthood. This chapter will summarize some of the neuroimaging studies that have mapped white matter maturation during this period, as well as the functional consequences of this structural maturation for cognitive function, especially for development of reading ability and working-memory capacity.

Development of white matter One of the most important processes in the maturation of the brain is the gradual increase in myelin around axons in the central nervous system. This myelination does not occur at the same rate all over the central nervous system (CNS), but with different onset times and at different rates in the various pathways, which could have consequences for how different functions mature during childhood and adolescent development. The first evidence of this behavioral-structural correlation can be seen during gestation, with the myelination of the spinal tracts for sensory and motor pathways underlying simple reflexes. Other motor and sensory systems of the CNS then follow according to their functional complexity (Brody et al., 1987; Kinney et al., 1988). Some of the last fibers to myelinate are the association fibers at the border of gray and white matter, as well as the pathway from the hippocampus to the neocortex (Benes, 1989). Paul Flechsig first described differences in myelination of the part of the axons closest to the cortex and provided a map of the progression of myelination among brain areas (figure 14.1) (Flechsig, 1920). Primary sensory and motor regions are myelinated first, and parts of the association cortex in frontal and parietal lobes are the last to myelinate. Yakovlev and Lecours (1967) largely confirm the map from Flechsig and note, “The intracortical neuropil of the anterolateral convexity of the frontal lobe, of the inferior parietal lobule and of the basolateral convexity of the temporal lobe seems to maintain potencies of myelination the longest.” They

suggested that these regions continue to myelinate even after 20 years of age. Structural MRI has provided a tool for evaluating the maturation of white matter in vivo. Consistent with the histological studies, MRI studies have confirmed that the volume of white matter increases at least until 20 years of age (Pfefferbaum et al., 1994; Caviness et al., 1996; Reiss et al., 1996; Paus et al., 1999; Giedd et al., 1999; Sowell et al., 1999; De Bellis et al., 2001; Castellanos et al., 2002). Some regional specificity has also been noted. Paus and associates, for example, showed the more pronounced development (as defined by higher white matter intensity on T1-weighted images) of the internal capsule and a region that could correspond to the arcuate fasciculus, which contains fibers between language regions in temporal and frontal lobes (Paus et al., 1999). An interesting finding is the nonlinear relationship between changes of cortical thickness, as measured by T1-weighted MRI, and changes in IQ (Shaw et al., 2006). The cellular basis of this cortical thinning is not known. It could reflect pruning and neuronal loss, but another interpretation is that the gradual myelination of the part of the axons close to and within the cortex affects the apparent white/gray-matter border, thus giving the impression of cortical thinning (Paus, 2005). This pattern would explain the observed relationship between development of IQ and regional myelination of cortical areas. Another MR method that can be used to investigate white matter is diffusion tensor MR imaging (DTI). In contrast to conventional T1- or T2-weighted MR imaging, DTI is based on the fact that the diffusion of water in the white matter of the brain is anisotropic (Moseley et al., 1990), so that it is faster along the axons than perpendicular to them. Some of this directional preference is provided by the axonal membrane itself, and possibly an intracellular component, even without myelin around the axons (Wimberger et al., 1995; Gulani et al., 2001). However, myelination of the axons further increases the anisotropy, as shown in studies comparing anisotropy with histological findings (Wimberger et al., 1995) comparing anisotropy

237

Figure 14.1 Flechsig’s map of myelination. Black regions are myelinated first, then gray, and finally white (Flechsig, 1920).

in normal mice with that of knockout mice lacking myelin (Gulani et al., 2001), as well as in human studies of demyelination (Werring et al., 1999). The degree of anisotropy can be quantified as fractional anisotropy (FA) which ranges from 0, corresponding to free diffusion, to 1, which is the hypothetical case of diffusion along a single line (Basser and Pierpaoli, 1996; Pierpaoli and Basser, 1996). In addition, FA can be affected by microstructural properties of the pathways, such as how densely the axons are packed and how regular their arrangement seems to be. For instance, a higher number of axon crossings in a given region decreases the FA value in that region. In order to image such crossing within a single voxel, diffusion has to be estimated from a high number of directions (Tuch et al., 2003). However,

238

there is no evidence that the basic architecture of axonal connections changes in a systematic way during normal development. The increase in FA during childhood is thus mainly affected by myelination and thickening of axons. Diffusion tensor imaging has been used to map the maturation of white matter during childhood (Klingberg et al., 1999; Mukherjee et al., 2001; Schmithorst et al., 2002; Snook et al., 2005). Klingberg and associates (1999) first demonstrated differences in the myelination of the frontal lobe in children compared to adults. Mukherjee and associates (2001) measured changes in anisotropy in children from 0 to 11 years of age and found changes in anisotropy not only in white matter regions, but also in the basal ganglia and the thalamus.

fundamentals of developmental neurobiology

Figure 14.2 Imaging of white matter structure with diffusion tensor imaging (Klingberg et al., 2000). The direction of axons is color coded: yellow, up/down; red, anterior/posterior; blue, left/ right. The encircled region (VOI) is where subjects with dyslexia

have a disturbance of white matter. STG, superior temporal gyrus. IC, internal capsule. Inset, a lateral view of the white matter region affected in dyslexia. (See plate 37.)

Functional consequences of myelination

Another aspect of myelination is that it has been shown to increase the probability that a particular action potential from one neuron reaches the next (Zhou and Chiu, 2001). This increased probability means that the influence of one neuron on the next increases, and this increase is functionally equivalent to increasing the connection strength between these neurons. Both increased speed and increased connection strength are thus possible functional consequences of myelination. How, then, do gross white matter changes correspond to differences in cognitive function? Although several studies mapped the development of white matter during childhood, only a few have correlated this with cognitive development. The first study to relate behavior to the fractional anisotropy measured by DTI demonstrated a relationship between reading ability and myelination (Klingberg et al., 2000); see figure 14.2 and plate 37. Dyslexic subjects were observed to have lower

One of the functional consequences of myelination of a previously unmyelinated axon is that transduction of the signal will be saltatory, spreading from node to node—a way of transmitting a signal that is faster and less energy consuming (Morell, Quarles, and Norton, 1989). Increased myelination of an already myelinated axon further increases the speed of transmission, with the speed of transmission being proportional to the total thickness of the axons (Waxman, 1980). Exactly how speed translates to improved information processing within the brain is not known, but it could provide faster processing and more precise timing. This increment in processing speed could be of importance for synchronization between brain regions of the fast oscillatory activity that has been suggested to be an important basis of communication between brain regions (Singer, 1993).

klingberg: white matter maturation and cognitive development

239

anisotropy, indicating less myelin, in a temporoparietal region where the axons presumably connect posterior and anterior language regions. Furthermore, a correlation between anisotropy and reading ability was also seen within the control group, suggesting that this structure-function relationship could be viewed as a continuum, with dyslexic individuals at the far end of the distribution. The relationship between myelination and reading ability could be related to the theories that conceptualize dyslexia as a deficit in the processing of rapidly changing information, in many sensory modalities (Farmer and Klein, 1995), which presumably relies on fast axonal conduction. Later studies have confirmed the relationship between temporoparietal myelination and reading ability also in 8–12-year-old children with varying reading ability, some of them with dyslexia (Beaulieu et al., 2005). Niogi and McCandliss (2006) also found a correlation between reading ability and anisotropy in 6–10year-old children in the same temporoparietal region. Again, this structure-function relationship was evident for both children within the normal reading range and those with below-normal reading ability. The most straightforward effect of myelination would be for it to decrease reaction time. Consistent with this effect, Tuch and associates found reaction times in young adults to negatively correlate with myelination in the optic radiations (Tuch et al., 2005). Diffusion in striatofrontal tracts was also found to predict reaction times in a go-no-go task (Liston et al., 2006). Nagy, Westerberg, and Klingberg (2004) studied the relationship between myelination during childhood and development of cognitive functions. Diffusion tensor imaging was used to estimate diffusion in white matter in 23 children between 8 and 18 years of age (mean age 11.9, SD 3.1, 14 boys), and fractional anisotropy was used as an indicator of white matter maturation, including myelination and thickening of axons. Behavioral measures included assessments of visuospatial working-memory capacity and reading of word lists. In an exploratory analysis, we searched the brain for voxels in which fractional anisotropy values correlated with working-memory capacity and reading scores across individuals. We found that the development of workingmemory capacity was positively correlated with fractional anisotropy in the central white matter of the left frontal lobe, in the anterior part of the corpus callosum, and in a region between the superior frontal and parietal cortices. Reading ability, however, was only significantly correlated with fractional anisotropy in the left temporal lobe, in the same white matter region in which adults with reading disability were previously shown to have lower fractional anisotropy (Klingberg et al., 2000). The study by Nagy and associates thus shows that myelination is regionally specific and that maturation of specific tracts is associated with the development of specific cognitive functions.

240

In two previous studies, the same visuospatial workingmemory tasks were used in order to evaluate how brain activity correlates with development of working-memory capacity during childhood (Klingberg, Forssberg, and Westerberg, 2002a; Olesen et al., 2003). The first study (Klingberg, Forssberg, and Westerberg, 2002a) included 13 children (age 9–18, mean age 13.4, 9 boys). This sample was later extended to 23 children (age 8–18, mean age 11.9, 14 boys). Children were scanned while performing a visuospatial working-memory task and a baseline task. Subtraction of brain activity recorded during the control task from that recorded during the working-memory task resulted in a measurement of working-memory-related activity. Workingmemory capacity was found to correlate with workingmemory-related brain activity in the posterior part of the superior frontal sulcus (−26, 8, 56), in the intra- and inferior parietal cortex (−36, −50, 56), and in the head of the caudate nucleus in the left hemisphere (Olesen et al., 2003). The superior frontal region and parietal regions were located close to the frontoparietal white matter tracts where myelination correlates with working-memory capacity development. Myelination of specific white matter tracts might thus be directly related to developmental changes in the cortical regions to which they connect. If this assumption is correct, one would expect fractional anisotropy and BOLD signals in these regions to be correlated. This association was investigated by Olesen and associates (2003); see figure 14.3 and plate 38. The subjects and the DTI data in the study by Olesen and associates (2003) were identical to those reported

Figure 14.3 A frontoparietal network involved in development of working memory (Olesen et al., 2003). Red: frontal and parietal regions where brain activity (BOLD signal) correlates with development of working-memory capacity; white: region where myelination (FA) correlates with development of working-memory capacity. Lines indicate correlations between brain activity and myelination. (See plate 38.)

fundamentals of developmental neurobiology

by Nagy, Westerberg, and Klingberg (2004), and the group from Klingberg, Forssberg, and Westerberg (2002a) was extended with fMRI measurements from 10 additional subjects. White matter regions that showed a developmental trend were identified by using working-memory scores from the children as a covariate and selecting white matter regions in which there was a positive correlation between workingmemory scores and FA. We then extracted the FA values from these regions for each individual and used them as covariates in an exploratory analysis of BOLD activity. Corresponding analyses were also done starting with BOLD response values in gray matter regions that were correlated with performance and performing a correlation with FA values. This second analysis (BOLD to FA) was primarily performed in order to confirm findings from the first analysis. A general positive correlation between all measures (age, working memory, FA, BOLD) is to be expected during development. The question was whether we could detect any sign of regional specificity in these data, with some regions showing stronger correlations depending on the functional networks in which they participate. Fractional anisotropy values in frontoparietal white matter were positively correlated with the BOLD response in closely located gray matter in the superior frontal sulcus (x = −26, y = 6, z = 56) and intraparietal cortex (x = −34, y = −68, z = 52). The correlation of FA values in frontoparietal white matter with the BOLD response in the superior frontal sulcus was confirmed in the converse analysis, where BOLD response values were used as covariates on FA maps. This correlation is primarily explained by the age-related maturation of white and gray matter, since working-memory scores did not correlate with FA values or the BOLD response in these regions when age-related variance was removed. Taken together, these studies (Klingberg, Forssberg, and Westerberg, 2002a; Olesen et al., 2003; Westerberg et al., 2004; Nagy, Westerberg, and Klingberg, 2004) thus suggest that myelination and developmental changes in brain activity are associated and, more specifically, that there is a superior frontal–intraparietal network in which brain activity, myelination, and development of visuospatial working-memory capacity are related during childhood and early adulthood. The mechanisms by which myelination affects brain activity as well as cognitive functions remain to be elucidated. One possible way to approach this question is to use neural network models. A recent study (Edin et al., 2007) used a neural network model with neurons with realistic inputoutput functions, based on the Hodgin-Huxley model of the membrane potential (Tegner, Compte, and Wang, 2002), in order to simulate the neural activity in two connected brain areas during the delay in a working-memory task. These simulations suggested that increased speed does not necessarily lead to higher brain activity and higher BOLD

response. However, a stronger functional connectivity does lead to higher BOLD response, as well as higher stability of the delay activity that represents the information. If myelination increases functional connectivity, as suggested by the results of Zhou and Chiu (2001), this could be one mechanism by which myelination leads to improved function, as well as higher BOLD activity. Another mechanism by which myelination might affect brain function is by enhancing between-region synchronization of fast oscillatory activity, an aspect of cortical processing that was not specifically investigated by Edin and associates. It has been shown that the BOLD response is affected by synchronized gamma oscillations (Niessing et al., 2005). Such synchronicity is dependent on precision within the millisecond range and is presumably enhanced by faster axonal connectivity.

The role of experience One way of interpreting the maturation of the frontoparietal network is to assume a genetically programmed maturation of white matter that affects the neural activity in the frontal and parietal regions, and in turn determines workingmemory capacity and the BOLD response. However, it is important to keep in mind the correlational nature of these studies and the possible direct effect of experience on both capacity and myelination. There are several animal studies suggesting that activity in the axons affects the myelination of the same axons (Demerens et al., 1996; Stevens and Fields, 2000). Studies using DTI have also demonstrated increased myelination in pianists after extensive training that spans several years (Schmithorst and Wilke, 2002; Bengtsson et al., 2005). If we look specifically at working memory, it has previously generally been assumed that working-memory capacity is relatively unaffected by experience. Recent studies, however, have suggested that practice of working-memory tasks can result in improved performance, not only for the trained tasks, but also for nontrained working-memory tasks, given that the practice is done daily, intensively, with adaptive algorithms, and during several weeks (Klingberg, Forssberg, and Westerberg, 2002b; Klingberg et al., 2005). In a subsequent study, fMRI was used to measure changes in brain activity induced by practice of working memory in young healthy adults. Several weeks of daily practice increased the working-memory-related activity in the intraparietal cortex and middle frontal gyrus (Olesen, Westerberg, and Klingberg, 2004). This pattern of findings suggests that the neural systems underlying working memory are plastic and formed by experience, and that the effect of experience must be considered even when evaluating such basic cognitive functions as working-memory capacity. Moreover, the part of intraparietal cortex that is affected by training overlaps with the regions in which a developmental

klingberg: white matter maturation and cognitive development

241

change is present. This overlap suggests that the neural basis of development and practice have some similarities and that the developmental changes we have observed may be the results of experience and practice in everyday life during development. However, the fact that similar regions are involved, and that increases in BOLD are observed in both cases, does not necessarily mean that the same mechanisms underlie the two processes. An analysis of white matter changes in the subjects from the study by Olesen, Westerberg, and Klingberg (2004) did not show any significant changes in fractional anisotropy. It is possible that there are different neuronal mechanisms that result in an identical behavioral change, as well as a similar change in the BOLD signal. In order to resolve these questions we need a deeper understanding of the relationships between cellular mechanisms, the BOLD signal, and information processing in the brain. In conclusion, maturation of white matter is an important process that has been shown to relate to the development of both reading ability and improvement of working-memory capacity during childhood and early adulthood. These white matter tracts form functional networks together with the cortical areas to which they are connected. Knowledge about this normal development could provide insight into the neural basis of neurodevelopmental disorders. Indeed, the regions implicated in development of reading ability are among those affected in dyslexia. Future research might provide a more detailed map of how maturation of different networks underlies development of various cognitive functions and the role of experience for such development. REFERENCES Basser, P. J., and C. Pierpaoli, 1996. Microstructural and physiological features of tissues elucidated by quantitativediffusion-tensor MRI. J. Magn. Reson. Imaging Series B 111: 209–219. Beaulieu, C., C. Plewes, L. A. Paulson, D. Roy, L. Snook, L. Concha, et al., 2005. Imaging brain connectivity in children with diverse reading ability. NeuroImage 25:1266–1271. Benes, F. M., 1989. Myelination of cortical-hippocampal relays during late adolescence. Schizophr. Bull. 15:585–593. Bengtsson, S. L., Z. Nagy, S. Skare, L. Forsman, H. Forssberg, and F. Ullen, 2005. Extensive piano practicing has regionally specific effects on white matter development. Nature Neurosci. 8:1148–1150. Brody, B. A., H. C. Kinney, A. S. Kloman, and F. H. Gilles, 1987. Sequence of central nervous system myelination in human infancy. I. An autopsy study of myelination. J. Neuropathol. Exp. Neurol. 46:283–301. Castellanos, F. X., P. P. Lee, W. Sharp, N. O. Jeffries, D. K. Greenstein, L. S. Clasen, et al., 2002. Developmental trajectories of brain volume abnormalities in children and adolescents with attention-deficit/hyperactivity disorder. JAMA 288:1740–1748. Caviness, V. S., D. N. Kennedy, C. Richelme, J. Rademacher, and P. A. Filipek, 1996. The human brain age 7–11 years:

242

A volumetric analysis based on magnetic resonance images. Cerebral Cortex 6:726–736. De Bellis, M. D., M. S. Keshavan, S. R. Beers, J. Hall, K. Frustaci, A. Masalehdan, et al., 2001. Sex differences in brain maturation during childhood and adolescence. Cerebral Cortex 11:552–557. Demerens, C., B. Stankoff, M. Logak, P. Anglade, B. Allinquant, F. Couraud, et al., 1996. Induction of myelination in the central nervous system by electrical activity. Proc. Natl. Acad. Sci. USA 93:9887–9892. Edin, F., J. Macoveanu, P. Olesen, J. Tegner, and T. Klingberg, 2007. Stronger synaptic connectivity as a mechanism behind development of working memory–related brain activity during childhood. J. Cogn. Neurosci. 19(5):750–760. Farmer, M. E., and R. M. Klein, 1995. The evidence for a temporal processing deficit linked to dyslexia: A review. Psychon. Bull. & Rev. 2:460–493. Flechsig, P., 1920. Anatomie des Menschlichen Gehirns und Rückenmarks auf Myelogenetischer Grundlage. Leipzig: Thieme. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, et al., 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2:861–863. Gulani, V., A. G. Webb, I. D. Duncan, and P. C. Lauterbur, 2001. Apparent diffusion tensor measurements in myelindeficient rat spinal cords. Magn. Reson. Med. 45:191–195. Kinney, H. C., B. A. Brody, A. S. Kloman, and F. H. Gilles, 1988. Sequence of central nervous system myelination in human infancy. J. Neuropathol. Exp. Neurol. 47:217–234. Klingberg, T., E. Fernell, P. Olesen, M. Johnson, P. Gustafsson, K. Dahlström, et al., 2005. Computerized training of working memory in children with ADHD—A randomized, controlled trial. J. Am. Acad. Child Adolesc. Psychiatry 44:177–186. Klingberg, T., H. Forssberg, and H. Westerberg, 2002a. Increased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14:1–10. Klingberg, T., H. Forssberg, and H. Westerberg, 2002b. Training of working memory in children with ADHD. J. Clin. Exp. Neuropsychol. 24:781–791. Klingberg, T., M. Hedehus, E. Temple, T. Salz, J. D. E. Gabrieli, M. E. Moseley, R. A. Poldrack, 2000. Microstructure of temporo-parietal white matter as a basis for reading ability: Evidence from diffusion tensor magnetic resonance imaging. Neuron 25:493–500. Klingberg, T., C. J. Vaidya, J. D. E. Gabrieli, M. E. Moseley, and M. Hedehus, 1999. Myelination and organization of the frontal white matter in children: A diffusion tensor MRI study. NeuroReport 10:1–5. Liston, C., R. Watts, N. Tottenham, M. C. Davidson, S. Niogi, A. M. Ulug, et al., 2006. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cerebral Cortex 16:553–560. Morell, P., R. H. Quarles, and W. T. Norton, 1989. Formation, structure, and biochemistry of myelin. In G. Siegel, B. Agranoff, R. W. Albers, and P. Molinoff, eds., Basic Neurochemistry, 4th ed., 109–136. New York: Raven Press. Moseley, M. E., J. Cohen, J. Kucharczyk, H. S. Mintorovitch, M. F. Asgari, J. Wendland, et al., 1990. Diffusion-weighted MR imaging of anisotropic water diffusion in cat central nervous system. Radiology 176:439– 445.

fundamentals of developmental neurobiology

Mukherjee, P., J. H. Miller, J. S. Shimony, T. E. Conturo, B. C. Lee, C. R. Almli, et al., 2001. Normal brain maturation during childhood: Developmental trends characterized with diffusion-tensor MR imaging. Radiology 221:349–358. Nagy, Z., H. Westerberg, and T. Klingberg, 2004. Regional maturation of white matter during childhood and development of function. J. Cogn. Neurosci. 16:1227–1233. Niessing, J., B. Ebisch, K. E. Schmidt, M. Niessing, W. Singer, and R. A. Galuske, 2005. Hemodynamic signals correlate tightly with synchronized gamma oscillations. Science 309:948– 951. Niogi, S. N., and B. D. McCandliss, 2006. Left lateralized white matter microstructure accounts for individual differences in reading ability and disability. Neuropsychologia 44:2178– 2188. Olesen, P., Z. Nagy, H. Westerberg, and T. Klingberg, 2003. Combined analysis of DTI and fMRI data reveals a joint maturation of white and grey matter in a fronto-parietal network. Cogn. Brain Res. 18:48–57. Olesen, P., H. Westerberg, and T. Klingberg, 2004. Increased prefrontal and parietal brain activity after training of working memory. Nature Neurosci. 7:75–79. Paus, T., 2005. Mapping brain maturation and cognitive development during adolescence. Trends Cogn. Sci. 9:60–68. Paus, T., A. Zijdenbos, K. Worsley, D. L. Collins, J. Blumenthal, J. N. Giedd, et al., 1999. Structural maturation of neural pathways in children and adolescents: In vivo study. Science 283:1908–1911. Pfefferbaum, A., D. H. Mathalon, E. V. Sullivan, J. M. Rawles, R. B. Zipursky, and K. O. Lim, 1994. A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch. Neurol. 51:874–887. Pierpaoli, C., and P. J. Basser, 1996. Toward a quantitative assessment of diffusion. Magn. Reson. Med. 36:893–906. Reiss, A. L., M. T. Abrams, H. S. Singer, J. L. Ross, and M. B. Denckla, 1996. Brain development, gender and IQ in children—A volumetric imaging study. Brain 119:1763– 1774. Schmithorst, V. J., and M. Wilke, 2002. Differences in white matter architecture between musicians and nonmusicians: A diffusion tensor imaging study. Neurosci. Lett. 321:57–60. Schmithorst, V. J., M. Wilke, B. J. Dardzinski, and S. K. Holland, 2002. Correlation of white matter diffusivity and anisotropy with age during childhood and adolescence: A cross-sectional diffusion-tensor MR imaging study. Radiology 222:212–218.

Shaw, P., D. Greenstein, J. Lerch, L. Clasen, R. Lenroot, N. Gogtay, et al., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440:676–679. Singer, W., 1993. Synchronization of cortical activity and its putative role in information processing and learning. Annu. Rev. Physiol. 55:349–374. Snook, L., L. A. Paulson, D. Roy, L. Phillips, and C. Beaulieu, 2005. Diffusion tensor imaging of neurodevelopment in children and young adults. NeuroImage 26:1164–1173. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 1999. In vivo evidence for postadolescent brain maturation in frontal and striatal regions. Nature Neurosci. 2:859–861. Stevens, B., and R. D. Fields, 2000. Response of Schwann cells to action potentials in development. Science 287:2267–2271. Tegner, J., A. Compte, and X. J. Wang, 2002. The dynamical stability of reverberatory neural circuits. Biol. Cybern. 87:471– 481. Tuch, D. S., T. G. Reese, M. R. Wiegell, and V. J. Wedeen, 2003. Diffusion MRI of complex neural architecture. Neuron 40:885–895. Tuch, D. S., D. H. Salat, J. J. Wisco, A. K. Zaleta, N. D. Hevelone, and H. D. Rosas, 2005. Choice reaction time performance correlates with diffusion anisotropy in white matter pathways supporting visuospatial attention. Proc. Natl. Acad. Sci. USA 102:12212–12217. Waxman, S. G., 1980. Determinants of conduction velocity in myelinated nerve fibers. Muscle Nerve 3:141–150. Werring, D. J., C. A. Clark, G. J. Barker, A. J. Thomson, and D. H. Miller, 1999. Diffusion tensor imaging of lesions and normal-appearing white matter in multiple sclerosis. Neurology 52:1626–1632. Westerberg, H., T. Hirvikoski, H. Forssberg, and T. Klingberg, 2004. Visuo-spatial working memory: A sensitive measurement of cognitive deficits in ADHD. Child Neuropsychol. 10:155–161. Wimberger, D. M., T. P. Roberts, A. J. Barkovich, L. M. Prayer, M. E. Moseley, and J. Kucharczyk, 1995. Identification of “premyelination” by diffusion-weighted MRI. J. Comput. Assist. Tomogr. 19:28–33. Yakovlev, P. I., and A.-R. Lecours, 1967. The myelogenetic cycles of regional maturation of the brain. In A. Minkowski, ed., Regional Development of the Brain in Early Life, 3–65. Oxford, UK: Blackwell Scientific. Zhou, L., and S. Y. Chiu, 2001. Computer model for action potential propagation through branch point in myelinated nerves. J. Neurophysiol. 85:197–210.

klingberg: white matter maturation and cognitive development

243

II METHODOLOGICAL PARADIGMS

15

Electrophysiological Methods in Studying Infant Cognitive Development GERGELY CSIBRA, ELENA KUSHNERENKO, AND TOBIAS GROSSMANN

Cognitive neuroscience, through its various neuroimaging techniques, enables us to look at the living brain at work and thus provides us with tools to investigate the neural underpinnings of developmental behavioral change. One class of these neuroimaging methods relies on the noninvasive and painless recording of brain electrical activity measured by electrodes placed on the scalp. The recorded signal, the electroencephalogram (EEG), carries information about the ongoing brain activation at the millisecond time scale, and derived measurements, like the event-related potentials (ERPs) and event-related oscillations (EROs), allow us to relate this information to the cognitive processes that the brain is engaged in. This technique is especially popular for measuring functional brain activation in developmental populations because it is relatively easy to record and the signal is relatively robust. For example, it is less sensitive to artifacts created by movement than is fMRI, and thus is better suited for studying awake, attentive children and infants (de Haan and Thomas, 2002). Furthermore, there is a long-standing tradition in its use with infants and young children, and its excellent temporal resolution can reveal information about the timing of neurocognitive processes that occur while an infant or child is engaged in some cognitive activity rather than providing information only about the final behavioral outcome of these processes.

The electroencephalogram The ongoing electrical brain activity can be recorded simultaneously from a number of electrodes attached to the scalp. These electrodes pick up the voltage changes that occur when a large number of cerebral neurons are activated in close proximity and in high synchrony. The electrical potential changes measured on the scalp primarily reflect the postsynaptic depolarization of cell dendrites and not the action potentials generated by the neurons. It is also important to note that in order to generate a large enough electrical field to be measurable on the scalp, the depolarization has to occur on many synapses that are more or less aligned in the same direction. The closer the activated neuron popu-

lation is to the surface of the cortex, the more likely it is that its activation will be reflected in the ongoing EEG. The placement of electrodes on the scalp conventionally follows the international 10/20 system (Jasper, 1958), in which electrodes are placed at certain distances from each other along the anterior-posterior and the lateral axes. The electrodes in this system are labeled according to cerebral lobe (frontal, temporal, parietal) under the location, odd numbers assigned to the left hemisphere and even numbers to the right, with Z indicating the midline. The relationships of the standard 10/20 system to the infant cerebral cortex differ slightly from those to the adult (Blume, 1974) because of differences in relative proportions of parietal and frontal lobes in infants and adults and incomplete opercularization of the temporal lobe in infants. Although it is most frequently used, the 10/20 system offers relatively poor spatial resolution of the signal. Over the years, this system has been extended to a 10/10 (Nuwer et al., 1998) and a 10/5 system (Oostenweld and Praamstra, 2001) to provide better spatial sampling on the surface of the head. Recently, high-density electrode systems, such as caps or nets comprising 64 or 128 electrodes, have become popular. A relatively new technique, the geodesic sensor net (GSN) allows a large number of electrodes to be quickly applied to the scalp surface (Tucker, 1993); this technique is especially useful in studies with infants, children, and special populations (Johnson et al., 2001). Even a 64-electrode GSN recording yields a sampling density of less than 3 cm with infants. The ongoing EEG is traditionally analyzed in the frequency domain. One of the main parameters that modulate the rate and amplitude of the EEG waves is the general alertness of the person. The alpha-band activity (8–13 Hz) can be measured from the occipital region in an awake and resting person when the eyes are closed. Beta waves (13–30 Hz) are characteristic of strong mental activation and are detectable over the parietal and frontal lobes. Delta-band waves (0.5–4 Hz) are detectable in infants and sleeping adults, while theta activity (4–8 Hz) is obtained from children and sleeping adults. The EEG also contains

247

activation in the gamma band (24–100 Hz) in both children and adults. However, the magnitude of this activity is so small compared to other frequency bands that it cannot normally be detected without special signal-analysis techniques. The EEG of children varies with age. During the first weeks of life, low-amplitude, poorly defined activity has been observed, composed of random-frequency waves in the delta, theta, and alpha ranges. The dominant activity in resting state is 4–6 Hz in 6-month-olds and 5–7 Hz by 1 year of age. At 4 years of age, the 7–8-Hz waves predominate, and by 8 years, the majority of children show alpha-band activity within adult frequency range. Distinct runs of beta activity are seldom observed in children, whereas episodic theta activity may be observed in frontal lobes until the age of 20 (Kooi, 1971). Some researchers have attempted to relate cognitive and brain development by correlating activity in certain frequency bands with behavioral variables in infants (Bell and Fox, 1992; Bell, 2001; Mundy, Card, and Fox, 2000). The majority of studies using electrophysiological measures, however, attempt to link cognitive processes to brain activations not by means of interindividual variability but by directly manipulating those cognitive processes. Studies with adult participants normally require them to solve various cognitive tasks; the studies then compare the brain electrical activation during the short time periods when the participants are engaged in the different conditions. For obvious reasons, preverbal infants cannot be assigned tasks to solve, but they can be exposed to different stimuli that require different cognitive processes to deal with. In this chapter, we discuss some techniques that have been used to analyze the EEG data collected during such cognitive studies with human infants. All the methods employ the same general idea for extracting the signal from the continuous EEG that is related to specific cognitive processes: repetition. Inducing the replication of the same cognitive event several times can help us to separate the electrical signal reflecting brain activation related to those events from other ongoing neural activity. We do not discuss in detail a class of techniques called steady-state and sweep evoked potentials, which employ rapid repetitions of stimuli at regular intervals and assess their effect by frequency analysis. Although they are rarely used in developmental cognitive neuroscience research, they provide a good example of how to use electrophysiological techniques to answer developmental questions that do not aim at infant cognition directly but are essential to resolve in order to know the constraints that determine what young infants can see and can hear. Eventrelated potentials and event-related oscillations also rely on repeated events and reduce the effect of background or nonrelated brain activations by time-locked averaging of the EEG. We do not discuss the methodological details of these

248

methodological paradigms

techniques here (for methodological guidance, see recent excellent handbooks: e.g., Handy, 2005; de Haan, 2007). Instead, we summarize how these techniques have been used to answer questions of cognitive development in human infants.

Event-related potentials Event-related potentials are calculated as time-locked averages of the EEG signal. The event that defines the EEG segment must be defined with millisecond-range accuracy and can be a stimulus of any modality, an omission of an expected stimulus, or a response generated by the participant. Event-related potentials also differ from traditional stimulus-evoked potentials in that the brain activation related to the event may also precede, and not just follow, the event. This is frequently the case when the event is a response or an external stimulus with well-predictable timing. The averaged ERPs are composed of a series of negativegoing and positive-going waves. An ERP component, according to the definition of Näätänen and Picton (1987), is the contribution of a generator process to the ERP waveform. If temporally overlapping components are of opposite polarities, they cancel one another either partially or totally. Traditionally, ERP components are divided into two categories: exogenous and endogenous. Exogenous (or sensory or obligatory) components can be elicited by any detectable stimulus and represent brain response to the occurrence of the stimulus. Exogenous components typically occur within the first 100–200 ms after stimulus onset and are, to a limited extent, sensitive to the physical features of the stimulus, such as intensity, frequency, and rate of stimulus presentation (Näätänen, 1992). Endogenous components mainly reflect internally generated mental events related to the cognitive assessment of the stimulus. The endogenous components, occurring after 100–200 ms from stimulus onset, not only reflect the processing of physical stimulus features, but also, depending on the paradigm and task, can index several stimulus-related cognitive processes. Because more and more recent studies suggest that even the early ERP responses can be modulated by top-down processes, like selective attention, the distinction between exogenous and endogenous components is not as meaningful as it once seemed to be. The majority of ERP waves are thought to reflect the synchronous activity of neural systems generated by excitatory and inhibitory postsynaptic potentials. Thus the maturational changes in ERP morphology might to a large extent involve changes in intracortical synaptic organization and synaptic density (Eggermont, 1988; Vaughan and Kurtzberg, 1992). Vaughan and Kurtzberg (1992) suggested that the ERP amplitude is proportional to the magnitude of synaptic activation. Indeed, the sequence of changes in synaptic

density parallels changes in the ERP amplitude, which follow an inverted U-shaped function, with rapid increase of the ERP amplitudes during infancy followed by a gradual decline during childhood. The sequence of synaptogenesis has been described as following a similar U-shaped function: rapid increase in synaptic density during infancy is followed by gradual decline to the mature adult levels at puberty (Huttenlocher, 1979). The striking parallel between the course of synaptogenesis previously reported by Huttenlocher and Dabholkar (1997) and the ERP-peak amplitudes was observed for auditory (Kushnerenko et al., 2002a) and visual modality in infants (Vaughan and Kurtzberg, 1992), for auditory modality in children (Ponton et al., 2000), and for both the visual and auditory modalities together (Courchesne, 1990). The increase in consistency of brain response with age (Thomas and Crow, 1994), resulting in decrease of the trial-to-trial latency variability, also contributes to the shortening of the ERP peak latencies. The latency changes in one ERP peak might also be affected by the maturational changes in another, overlapping peak (Ponton et al., 2000; Kushnerenko et al., 2002a). What has emerged as a general finding from developmental ERP studies is that young infants do not show as many well-defined peaked ERP responses as adults, but they do show greater slow wave activity (Nelson and Luiciana, 1998). The greater slow-wave activity during the first two years of life has been attributed to reduced synaptic efficiency. Eventrelated-potential waveforms with well-defined peaks over the frontal cortex, which are typical for adults, begin to emerge around 4 years but continue to develop well beyond that age (Nelson and Luciana, 1998; Taylor, Batty, and Itier, 2004). We discuss the typical findings of ERP research in infants grouped by the events that are used to elicit those potentials. First we review infant ERPs to auditory and visual stimuli, then we briefly summarize some studies with multimodal stimuli and eye-movement-related analyses.

ERPs to auditory stimuli The P1-N1-P2 Complex In adults, auditory ERPs start with a small P1 (or P50) deflection that peaks at about 50 ms. The P1 is followed by a usually larger N1 response, peaking at about 100 ms and further by a P2 component peaking at approximately 180–200 ms from stimulus onset (Näätänen, 1992; Ponton et al., 2000). The P2 peak is often followed by a negativity, labeled N2 (Picton et al., 1974). This peak has an adult latency of 220–270 ms (Ponton et al., 2000). The N2 elicited by frequent repetitive stimuli was reported mostly in children (Ceponiene, Cheour, and Näätänen, 1998; Enoki et al., 1993; Karhu et al., 1997; Korpilahti and Lang, 1994), but it was also shown in adults (Ceponiene et al., 2001; Karhu et al., 1997; Kushnerenko et al., 2001; Picton et al., 1974; Ponton et al., 2000).

The adult P50-N100-P200 (P1-N1-P2) complex is not readily identifiable in infants and children before about 10 years of age (Ponton et al., 2000; Courchesne, 1990). Most of the ERP studies in infants have reported a large positive deflection at midline electrodes, with a maximum amplitude at about 300 ms, followed by a negativity at about 600 ms (Barnet et al., 1975; Graziani et al., 1974; Ohlrich et al., 1978; Pasman et al., 1992; Rotteveel et al., 1987; Shucard et al., 1987). Novak and colleagues (1989) followed the maturation of the auditory ERPs to speech stimuli (/da/ and /ta/ syllables) from birth to 6 months. The P2-N2 complex recorded at birth changed in morphology by the age of 3 months. The authors discerned two positive peaks in the latency range of the infantile P2 (P1m and P2m) with different scalp predominance: the P1m was larger frontally than centrally, whereas the P2m was largest centrally. A discontinuity (negative trough) between these two positive peaks, at about 160– 200 ms, was termed N1m by the authors. The N1m became prominent by the age of 6 months. During the first 6 months of life, the P1m and P2m increased in amplitude and gradually decreased in latency. Further, peak amplitudes did not increase linearly, but followed an inverted-U function with a maximum at 3 months in a study by Barnet and colleagues (1975) and at 6 months in studies by Vaughan and Kurtzberg (1992) and Kushnerenko and colleagues (2002a). The amplitude of the second major positive peak (P2m or P350) markedly decreased between 6 and 9 months, while the amplitude of the preceding negativity (N1m or N250) increased (Kurtzberg et al., 1986; Kushnerenko et al., 2002a). A longitudinal study performed in the same infants from birth to 12 months of age employing a spectrally rich harmonic tone showed that an analogue to the N1m (labeled N250 according to its peak latency) could be identified already at birth and consolidates by the age of 6 months. By the age of 1 year, the waveform morphology observable through early childhood is attained (figure 15.1). However, as can be seen in figure 15.1, the N250 in newborns appears to be sensitive to stimulus duration and probability. A negative peak within the N1 latency range was obtained by Wunderlich, Cone-Wesson, and Shepherd (2006) from birth in response to words but not to tones. Thus tones do not seem to be perfect stimuli for auditory ERP research with very young infants (see also Kushnerenko et al., 2006). The P350 (P2) is also dependent upon stimulus probability and interstimulus interval (for compatible evidence in older children see Ceponiene, Rinne, and Näätänen, 2002; Kurtzberg et al., 1995). Throughout the early school years, children’s ERP in response to auditory stimuli presented at a fast rate consists of the P100, N250, and N450 peaks (Ceponiene, Cheour, and Näätänen, 1998; Ceponiene et al., 2001; Ceponiene,

csibra, kushnerenko, and grossmann: infant cognitive development

249

Figure 15.1 Auditory event-related potentials as a function of age, stimulus probability, and duration. The left panel (A) represents ERP responses to harmonic tones of three different frequencies presented randomly mixed with probability of 0.33 each (see

Kushnerenko et al., 2002a). The right panel (B) shows the ERPs to the harmonic tones presented with the same ISI but with probability of 0.85. Negativity is plotted upward.

Rinne, and Näätänen, 2002). With a slower presentation rate, however, an adultlike P1-N1-P2-N2 pattern can be observed (Ceponiene, Cheour, and Näätänen, 1998; Ceponiene, Rinne, and Näätänen, 2002; Karhu et al., 1997). The N1 can be elicited from the age of 3 years with a slow stimulation rate (Paetau et al., 1995; Sharma et al., 1997), a finding that suggests longer refractory periods of N1 generators in children. Mismatch Responses In order to study the capacity to discriminate or categorize auditory stimuli, many researchers employ the so-called auditory oddball paradigm. In this paradigm one stimulus (the “standard”) is repeated frequently (about 60% to 90% of trials) and another (the “deviant”) occurs infrequently. The oddball paradigm can either be passive (unattended), when no response is required from the participant, or active (attended), when the subject is supposed to react to the deviant, “target” stimuli. The most extensively studied ERP component, the P3b (or P300), is elicited in response to such target stimuli under attended oddball conditions. Among the endogenous components elicited in passive auditory oddball paradigms in adults are the mismatch negativity peaking about 150–200 ms, the P3a (250–350 ms), and the late negativity, Nc (commencing at about 500 ms). The mismatch negativity (MMN) was isolated from the N2 wave by Näätänen, Gaillard, and Mantysalo (1978). The MMN is generated by a neural matching process between a deviant sensory input and the neural representation, or “sensory memory trace,” formed by the repetitive standard sound. This auditory-cortex activation presumably reflects an automatic preattentional change-detection process, comparing the new auditory input with information stored in auditory sensory memory (Näätänen, 1992). Several auditory-change detection components have been described in infants. In the majority of the studies, a positivity peaking at about 300 ms was observed (DehaeneLambertz and Pena, 2001; Dehaene and Gliga, 2004; Friedrich, Weber, and Friederici, 2004; Dehaene-Lambertz, 2000; Dehaene-Lambertz and Baillet, 1998; DehaeneLambertz and Dehaene, 1994; Winkler et al., 2003). An early negativity peaking at about 150 ms was obtained only in response to frequency change with grossly deviating stimuli (Ceponiene, Kushnerenko, et al., 2002; Morr et al., 2002; Kushnerenko et al., 2002b) and was suggested to be related to spectral change of acoustic parameters (Kushnerenko et al., 2006). A broad long-lasting later negativity (270–400 ms) was found in response to relatively small auditory contrasts in newborns and even prematurely born infants (e.g., to the difference between Finnish vowels /y/ and /i/, CheourLuhtanen et al., 1995, 1996; Cheour et al., 1999). Leppanen and colleagues (2004) attempted to explain the discrepancy

between the polarity of the change-detection response in infants and adults by the maturational level of the newborn, because it had been established that immature neonates display inverse-polarity ERPs (Kurtzberg and Vaughan, 1985). Thus the broad negativity obtained in prematurely born infants may be due to an immature neural response. In full-term newborns, however, no immature ERPs were found in response to broadband stimuli and large spectral changes (Kushnerenko et al., 2006). In contrast, both a large-amplitude early negativity and a central positivity were elicited reliably across neonates in response to these stimuli. The requirement of large spectral deviation suggests an incomplete refinement of frequency-specific pathways and is consistent with evidence showing that frequency resolution and fine frequency tuning improves during the first 6 months of life (Abdala and Folsom, 1995; Werner, 1996). Further in development, negative and positive mismatch responses appear to overlap and mask each other, resulting in a predominantly positive deflection before 1 year of age, and even in the absence of deviant-standard difference between 1 and 4 years of age with relatively small acoustic contrasts (Morr et al., 2002). Accordingly, it was shown that the relative strength of positive and negative mismatch responses varied from age to age and from infant to infant (Kushnerenko et al., 2002b). Another component that can be obtained in the passive oddball paradigm is the P3a, a frontocentrally maximal positivity elicited by stimuli that catch attention. Squires, Squires, and Hillyard (1975) proposed that the P3a was the central electrophysiological marker of the orienting response (see also Sokolov et al., 2002). “Novel” sounds (random mixture of mechanical or environmental noises) among pure tones are often used to elicit the P3a. Such grossly deviating stimuli typically elicit a large P3a response in children (Gumenyuk et al., 2004; Ceponiene et al., 2004) and adults (Escera et al., 2000). Surprisingly, newborns also show a similar pattern of response to “novel” sounds (Kushnerenko et al., 2002b, 2006). It has been argued, however, that the major part of the P3a in newborns is elicited by the spectral richness of the novel sounds, which recruit new afferent neurons into the response pool. The P3a is sometimes followed by a frontal negativity at 500–600 ms latency in children’s and infants’ auditory ERPs (Gumenyuk et al., 2004; Ceponiene et al., 2004; Kushnerenko et al., 2002b, 2006). This late negativity is larger in amplitude in younger than in older children—the same maturational profile that has previously been reported for the negative component Nc (Courchesne, 1983). This negative component has been suggested to be a sign of enhanced auditory and visual attention (see also later section “ERPs to visual stimuli”), since it was elicited in response to surprising, interesting, or important stimuli (Courchesne, 1978, 1990).

csibra, kushnerenko, and grossmann: infant cognitive development

251

A similar negativity was also found when participants had to reorient their attention back to a task after distraction by “novel” sounds (Escera, Yago, and Alho, 2001) or in response to unexpected frequency changes in auditory stimuli (Schröger and Wolff, 1998; Schröger, Giard, and Wolff, 2000). This negativity was called the reorienting negativity (RON) by Schröger and Wolff (1998). Being of comparable latency and scalp topography, the Nc and RON might, in fact, reflect the same neural process. Interestingly, as noted by Courchesne (1990), the maturational time course of the Nc (amplitude increase across infancy and early childhood followed by a gradual decline through preadolescence) is parallel to the synaptic density changes in the frontal cortex reported by Huttenlocher (1979), and to the metabolic activity changes as reported by Chugani, Phelps, and Mazziotta (1987). Thus the Nc might reflect the development of the higher-order cognitive functions associated with the frontal cortex. An Nc-like, frontally maximal negativity was found not only in response to surprising or “novel” stimuli, but also to nonnovel speech-syllable contrasts in newborns and very young infants (Dehaene-Lambertz and Dehaene, 1994; Kurtzberg et al., 1984; Friederici, Friedrich, and Weber, 2002). This finding might indicate that for a newborn infant any stimulus change might be “novel” or surprising, whereas with increasing age a capacity to respond only to the most attention-getting stimuli matures (Courchesne, 1990). In infants, negative and positive slow waves (NSW and PSW) were observed to follow the Nc under certain circumstances. Deregnier and colleagues (2000) obtained NSW in response to a stranger’s voice compared with the maternal voice in sleeping newborns. In Courchesne’s early work (1978), a long-latency PSW was also observed to infrequently presented stimuli. Nelson and colleagues have speculated that these waves are typically invoked by stimuli that the infant has only partially encoded and indicate detection of novel stimuli against a background of familiar stimuli (de Haan and Nelson, 1997; Nelson, 1994; see Nelson and Monk, 2001, for discussion). Components Related to Lexical and Syntactic Processing Further auditory ERP paradigms are related to linguistic processing of acoustic stimuli. In these paradigms, words, nonsense words, or sentences are presented to participants that are either appropriate in the semantic or syntactic context or violate some linguistic aspects (for recent reviews see Friederici, 2002, 2005). Mills and colleagues (2004) have shown that 14- and 20-month-old infants responded with a larger amplitude of N200–N400 to known words than to nonsense words. In addition, Friedrich and Friederici (2005) have shown an N400-like semantic incongruity effect in 19-month-old infants. The N400 is usually elicited by sentences that end with semantically

252

methodological paradigms

inappropriate words (Kutas and Hillyard, 1983). In the study of Friedrich and Friederici (2005), object words were presented either matching the concurrent visual stimulus or not. In response to semantic incongruity a slow negative wave was observed starting from about 400 ms and reaching significance between 800 and 1,400 ms. In addition, the congruous words elicited more negative responses than the incongruous words in the shorter latency range, matching the finding of Mills and colleagues (2004) that known words elicited more negative response than unknown or nonsense words. Korpilahti and colleagues (2001) also reported that the second negativity (denominated as late mismatch negativity) was significantly larger in 4- to 7-year-old children for words than for pseudowords, a finding that led the authors to propose that this late MMN might reflect the detection of semantic anomaly. The ERP studies of sentence-structure processing in adults have shown that syntactic (grammatical) violations are associated with two ERP components: an early left anterior negativity (ELAN) and a late, centroparietal positivity (P600) (Friederici, 2002). Recently, Oberecker, Friedrich, and Friederici (2005) have shown that children below three years of age also responded with an early left negativity and late positivity to syntactic violations. These deflections, however, peak later and persist longer in children than in adults.

ERPs to visual stimuli Event-related potentials have proven to be a very useful tool in studying the development of visual processing. The goal of this section is to give an overview of the primary ERP components that have been used to study the development of visual processes in human infants. It is beyond the scope of this chapter to provide an exhaustive review of the development of all components previously reported in visual ERP studies (DeBoer, Scott, and Nelson, 2004; de Haan, Johnson, and Halit, 2003; Nelson, 1994; Nelson and Monk, 2001; Taylor, Batty, and Itier, 2004). We will therefore mainly focus on the well-studied components observed during face processing, which will be discussed in the order of their temporal occurrence in the waveform. The P1 Component Visual stimuli reliably elicit a positivegoing component between 90 and 150 ms, called P1, in individuals of all ages (de Haan, Johnson, and Halit, 2003). It has been shown that from 4 years of age, P1 latency is shorter to upright than inverted faces and, similarly, the P1 is shorter to faces than to objects (Taylor, Batty, and Itier, 2004). There is also evidence to suggest global effects of facial expressions of emotion on the P1 (Batty and Taylor, 2006). However, some of the effects observed could not be replicated (Rossion et al., 1999), and it has been argued that low-level physical differences, which were not controlled for

in these studies, or more general attentional top-down processes might influence the properties of this early visual ERP component (for a discussion see de Haan, Johnson, and Halit, 2003). The N170/N290 and P400 Components In adults, human faces elicit an N170 response, which is most prominent over posterior temporal sites and is larger in amplitude and longer in latency to inverted than to upright faces (Bentin et al., 1996; de Haan, Pascalis, and Johnson, 2002). Other kinds of objects evoke a similar response around this latency in adults (generally called N1). This component is not modulated by the inversion of monkey faces (de Haan, Pascalis, and Johnson, 2002), nor when upright objects are compared to inverted objects (Bentin et al., 1996). This selective effect has been taken as evidence for a special face-processing mechanism generating the N170. From studies examining the influence of stimulus inversion on infants’ ERP responses to faces, it has been suggested that the infant N290 is a precursor to the adult N170. Like the N170, the infant N290 is a negative-going deflection observed over posterior electrodes. Its peak latency decreases from 350 ms at 3 months to 290 ms at 12 months of age (Halit, de Haan, and Johnson, 2003). In the studies that measured ERPs to upright and inverted human and monkey faces (de Haan, Pascalis, and Johnson, 2002; Halit, de Haan, and Johnson, 2003) the amplitude of the infant N290 at 12 months of age, like the adult N170, enhanced to inverted human but not to inverted monkey faces when compared to upright faces. However, the amplitude of the N290 was not affected by stimulus inversion at an earlier age (3 and 6 months). These younger infants showed an inversion effect on the amplitude of the P400 that follows the N290. The P400 that follows the N290 is a positive deflection most prominent over lateral posterior electrodes, and its peak latency decreases from 450 to 390 ms between 3 and 12 months of age (Halit, de Haan, and Johnson, 2003). This component is similar to the adult N170 in two ways: first, like the adult N170, the P400 is more prominent at lateral electrodes (de Haan, Pascalis, and Johnson, 2002; Halit, de Haan, and Johnson, 2003), and second, like the adult N170, the peak latency of the P400 is shorter to faces than to objects (de Haan and Nelson, 1999). However, unlike the adult N170, the modulation of the amplitude of the P400 at 3 and 6 months is not specific to inverted human faces, since it was also observed in response to inverted monkey faces (de Haan, Pascalis, and Johnson, 2002; Halit et al., 2003). By 12 months of age, infants’ P400, like the adult N170, appears to be longer in latency to inverted human faces, but does not differ between upright and inverted monkey faces. This finding suggests that, like the N290, the P400 becomes more finely tuned to human faces toward the end of the first year. However, it is important to note that similar amplitude

Figure 15.2 Event-related potentials to faces and matched visual noise stimuli in 3-month-old infants. Note that positivity is plotted upward. (Adapted from Halit et al., 2004.)

enhancement for faces when compared to matched visual noise was observed in the infant N290 and adult N170, whereas the elicited P400 only showed a latency effect (figure 15.2), which makes it unlikely that the P400 is the main precursor of the adult N170 (Halit et al., 2004). The development of the brain processes reflected in the N170/N290 continues well beyond infancy (for a review see Taylor, Batty, and Itier, 2004). While the latency of the adult N170 is delayed by inversion, no such effect has been observed for the latency of the infant N290 at any age (de Haan, Pascalis, and Johnson, 2002; Halit, de Haan, and Johnson, 2003). There is evidence that suggests that this latency effect is not apparent until 8 to 11 years (Taylor, Batty, and Itier, 2004). Another important developmental finding is that while the amplitude of the adult N170 is larger to the monkey faces, infants’ N290 shows the opposite pattern. An adultlike modulation of the amplitude of the N170 has not been reported until 13 to 14 years (Taylor, Batty, and Itier, 2004). Furthermore, while the amplitude of the adult N170 is not affected by direction of gaze (Grice et al., 2005; Taylor et al., 2001), it has been shown that the amplitude of the N290 in 4-month-old infants is modulated by eye gaze (Farroni et al., 2002; Farroni, Csibra, and Johnson, 2004). In these studies, faces with direct gaze compared to faces with averted gaze elicited an enhanced N290 in infants’ ERPs, which might indicate that a face with direct gaze is the perceptually more “prototypical” for 4-month-olds than a face with averted gaze. This finding suggests that face and eye gaze share common patterns of cortical activation early in ontogeny, which later partially dissociate and become more specialized. It is important to note that the development of face processing is associated not only with progressive specialization

csibra, kushnerenko, and grossmann: infant cognitive development

253

of cortical populations, but also with a broadening of certain representations. Using an adaptation paradigm (GrillSpector and Malach, 2001), Gliga and Dehaene-Lambertz (2007) showed that view-specific (front-view and profile) and view-invariant face representation are accessed in adults at the level of the N170. On the contrary, only viewdependent representations are employed by 4-month-old infants, as reflected by an N290 response suppression when front-view faces are repeated but not when faces with different orientations are repeated. Negative Component (Nc) The Nc is one of the most studied components in infant ERP research. This component is a negative deflection that occurs between 400 and 800 ms after stimulus onset and is most prominent over frontal and central electrode sites. The Nc has been thought of as an obligatory attentional response sensitive to stimulus familiarity (Courchesne, Ganz, and Norcia, 1981; Quinn, Westerlund, and Nelson, 2006; Snyder, Webb, and Nelson, 2002) that is observed in response not only to visual stimuli but also to stimuli in other modalities (see the earlier section “ERPs to auditory stimuli” and also Grossmann, Striano, and Friederici, 2005, 2006; Purhonen et al., 2004). Dipole modeling has revealed that the cortical sources of the Nc can be localized in the anterior cingulate and other prefrontal regions (Reynolds and Richards, 2005). The Nc has been observed in a series of studies using a visual oddball paradigm (Ackles and Cook, 1998; Courchesne, Ganz, and Norcia, 1981; Karrer and Ackles, 1987; Karrer and Monti, 1995). The infant Nc has consistently been found to be greater in its amplitude to the infrequent stimulus event when compared to the frequent stimulus. In this context, the Nc has been interpreted as reflecting either infants’ allocation of attention, with the greater negativity to the infrequently presented stimulus indexing orientation toward the novel or more unexpected event (Courchesne, Ganz, and Norcia, 1981; Nelson, 1994), or as a more generalized arousal elicited by novel or infrequent stimuli (Richards, 2002). However, the Nc is also thought to reflect recognition processes, as the Nc is greater in its amplitude to the mother’s than to a stranger’s face (de Haan and Nelson, 1997), and it is also greater to familiar than to novel toys (de Haan and Nelson, 1999) when faces and objects are presented with equal probability. A recent longitudinal investigation of infants’ visual ERPs to novel and familiar faces and objects revealed that the exact response properties of the Nc and other ERP components undergo complex development throughout the first year of life (Webb, Long, and Nelson, 2005). In this study, the Nc was found to decrease in its latency and increase in its amplitude toward the end of the first year, a pattern that reflects the general developmental pattern observed during infancy.

254

methodological paradigms

Alternatively, the Nc may reflect processing of semantic and/or emotional information, since its amplitude has also been found to be modulated by the emotional content of a face (Nelson and de Haan, 1996). Furthermore, the direction of the difference in the amplitude between a mother’s and a stranger’s face changes with age. Namely, children younger than 24 months show a larger Nc to a mother’s face, but children older than 45 months show a larger Nc to a stranger’s than to mother’s face (Carver et al., 2003). Carver and colleagues (2003) interpreted this finding as indicating that the caregiver’s face is particularly salient during the first two years of life, as children are forming their relationship and bond with the caregiver, but that these are well enough established by 4 years that these children can begin to allocate more resources to processing strangers’ faces. This interpretation supports the view that the Nc is related to the relative “emotional/semantic” salience of a stimulus (Nelson and de Haan, 1996; Carver et al., 2003). All in all, the available evidence seems to suggest that the Nc reflects attentional processes that are affected by the familiarity, recognition, and emotional content of the stimulus. Negative and Positive Slow Waves (NSW and PSW) Just like auditory stimuli, visual stimuli have also been found to elicit slow-wave activity in infants’ ERPs following the Nc. The amplitude of these slow waves varies as a function of stimulus familiarity and presentation probability. For example, in a visual oddball paradigm (Nelson and Collins, 1991), 6-month-old infants were familiarized to two faces and were then presented with one of the familiar faces frequently (60%), the second familiar face infrequently (20%), and a group of novel faces infrequently (20%). In this study, only the brain activation following the Nc differed between conditions. The infrequently presented novel face elicited a long-latency negative slow wave (NSW), which was interpreted as reflecting processes related to novelty detection. The infrequently presented familiar face elicited a longlatency positive slow wave (PSW), which, according to the authors, reflected processes related to updating a decaying memory. Event-related potentials to the frequently presented familiar stimulus returned to baseline, indicating the recognition of a well-encoded face for which memory updating was no longer necessary. The view that the PSW reflects how much a visual stimulus is encoded is further supported by the finding that its amplitude decreases with stimulus repetition throughout an experimental session (Snyder, Webb, and Nelson, 2002). Based on these and other findings (see Nelson, 1994), it has been argued that infants’ PSW might be a precursor to adults’ P300, which is thought to be involved in context updating (Donchin and Coles, 1988; Friedman, 1991; Nelson and Collins, 1991), whereas the NSW might be specific to infants, since it has not been observed beyond infancy.

ERPs to multimodal stimuli Most developmental ERP studies have concentrated on examining the neural correlates of processing stimuli presented only in a single modality. However, it is of great interest to understand how the human brain that develops in a multimodal world uses and integrates information from different senses. Only a few attempts have been made to assess brain processing in multimodal designs that have revealed insights into the neural underpinnings of infants’ cross-modal integration abilities from the haptic to the visual modality (Nelson, Henschel, and Collins, 1993) and from the visual to the auditory modality (Grossmann, Striano, and Friederici, 2006; Friedrich and Friederici, 2005). For example, 7-month-old infants’ processing of emotionally congruent and incongruent face-voice pairs was investigated using ERP measures (Grossmann, Striano, and Friederici, 2006). Infants watched facial expressions (happy or angry) and heard a word spoken with either an emotionally congruent or incongruent tone of voice. The ERP data revealed that the amplitude of a negative component (Nc) and a subsequent positive component (Pc) in infants’ ERPs varied as a function of cross-modal emotional congruity (figure 15.3).

Emotionally incongruent face-voice pairs elicited a larger Nc in infants’ ERPs than emotionally congruent pairs. Conversely, the amplitude of infants’ Pc was found to be larger to emotionally congruent words than to incongruent words. Based on previous work that has shown that an attenuation of the negative component and an enhancement of the later positive component in infants’ ERPs reflects the recognition of an item (Nelson et al., 1998), it was suggested that 7-month-olds integrate emotional information across modalities and recognize common affect in the face and voice. Interestingly, presenting multimodal stimuli allowed the use of an unusually high number of trials for analysis, suggesting an advantage of multimodal over unimodal stimuli in capturing infants’ attention (Grossmann, Striano, and Friederici, 2006). This is of special interest considering the generally low signal-to-noise ratio (SNR) of infant ERP studies. Moreover, a recent study employed a novel interactive paradigm to assess the neural correlates of joint attention in 9-month-old infants, and found that neural processing in infants is enhanced when learning takes place in the context of a joint attention interaction (Striano, Reid, and Hoehle, 2006).

Figure 15.3 Seven-month-old infants’ ERP responses to emotionally congruent (solid) and incongruent (dotted) face-voice pairs. Negativity is plotted upward. (Adapted from Grossmann, Striano, and Friederici, 2006.)

csibra, kushnerenko, and grossmann: infant cognitive development

255

The amplitude of the elicited ERP component was substantially larger in this study than that seen in previous literature, a difference that may be due to the new paradigm’s employing live interaction. The usage of multimodal stimuli and interactive paradigms has the advantage of an increased social significance for the child and of a higher ecological validity when compared with those ERP paradigms utilized in the past. It is experimentally challenging but worthwhile to continue to improve the paradigm on ERP studies by modifying the tasks so that they are in accordance with the child’s world.

Saccade-related ERPs The electrical brain activity can also be analyzed in relation to events that are not externally but internally generated by the participants. In particular, response-related ERPs are calculated by time locking the EEG to the manual responses (e.g., key presses) performed during the task. Of course, infants do not usually participate in tasks that require manual responses, but they are quite proficient in performing another type of action: eye movements. Saccade-related potentials are usually time locked to the initiation of eye movements, though one can also calculate ERPs to the termination of the saccades (i.e., the fixation) as well. These time points can be identified from the horizontal and vertical electro-oculograms (EOGs) that are usually coregistered with the EEG. Saccades are usually preceded by characteristic presaccadic components, like the sharp spike potential (SP), the presaccadic positivity (PSP), both maximal over parietal areas, and the presaccadic negativity (PSN) measured over the anterior cortex (Balaban and Weinstein, 1985; Csibra, Johnson, and Tucker, 1997). The spike potential is absent in young infants’ saccade-related potentials, and its development is discussed in another chapter of this handbook (Johnson, Mareschal, and Csibra, chapter 28) in the context of the early development of the visual pathways (see also Csibra, Tucker, and Johnson, 1998; Csibra, Tucker, et al., 2000). The other presaccadic potentials also differ between infants and adults. Richards (2000) identified a presaccadic positivity over anterior areas in infants that preceded the start of the eye movement by about 50 ms. This component occurred only when infants made a saccade toward a precued target, and it did not emerge before 20 weeks of age. A similar presaccadic positivity in a later study was localized to the superior frontal gyrus (Richards, 2005). While the presaccadic potentials differ markedly between infants and adults, the postsaccadic lambda wave behaves functionally the same way in infants as in later ages. The lambda wave is a sharp potential appearing over visual cortical areas that is generated when a peripheral target stimulus is foveated (Kurtzberg and Vaughan, 1977). This wave is

256

methodological paradigms

essentially a visual ERP to the newly fixated stimulus, and it can be detected in both 6- and 12-month-olds (Csibra, Tucker, and Johnson, 1998; Csibra, Tucker, et al., 2000).

Event-related oscillations Neurons have the inherent capacity to spontaneously produce oscillatory activity at frequencies above 20 Hz (Llinás, 1988). Sensory stimuli in several modalities (visual, auditory, olfactory) can elicit such oscillations in the gammaband frequency range (20–80 Hz, most commonly around 40 Hz). When a large number of neurons fire synchronously at the same frequency, these oscillations can be recorded from the scalp by conventional EEG techniques. Recently several laboratories have started to analyze human EEG signals in terms of bursts of oscillatory activities and interpret them in relation to the cognitive functions that the participants performed while their brain waves were recorded (e.g., Tallon-Baudry and Bertrand, 1999). Oscillatory neural activities are usually restricted both in time and frequency content; therefore, analyses only in the time or frequency domain tend to be blind to them. To reveal task-related bursts of oscillatory activities, especially if they occur at higher frequency ranges, we need to perform a time-frequency analysis that tracks how amplitude (or power) varies at different frequencies over time. There are two types of EROs: evoked EROs are oscillations that are phase locked to the corresponding event and can be recovered from averaged, nonfiltered ERP waveforms; induced EROs are not phase locked and are obtained from raw EEGs before averaging. Evoked oscillations are usually short-latency responses, while induced oscillations can occur both close to and farther away from the corresponding events (Csibra, Davis, et al., 2000; Herrmann and Mecklinger, 2000). For detailed guidance as to the calculation of these oscillations, see Herrmann, Grigutsch, and Busch (2005) and Csibra and Johnson (2007). We illustrate the use of gamma-band oscillations by recent studies that explored neural correlates of one of the most debated phenomena of infant cognition: the representation of hidden objects. Sustained responses in neural circuits have been identified as a mechanism for maintaining representations of objects during a period of occlusion (Rainer and Miller, 2000). In particular, in human adults gamma-band (∼40 Hz) activity has been associated with maintaining an object/location in mind (Tallon-Baudry et al., 1998). We measured infants’ electrophysiological responses to occlusion events at the age where reaching behavior does not yet show evidence of understanding “object permanence” (Kaufman, Csibra, and Johnson, 2003). Six-month-old infants were shown sequences of videorecorded and digitally edited events depicting an object (a train engine) appearing or failing to appear from under a tunnel when it should or

Figure 15.4 Gamma-band EROs time locked to a tunnel-lifting event in 6-month-old infants. In the “unexpected-disppearance” condition the infants had just seen a train entering the tunnel. The

difference map represents the scalp distribution of the oscillatory activity. (Adapted from Kaufman, Csibra, and Johnson, 2003.) (See plate 39.)

should not have been there. We hypothesized that gammaband oscillatory activity may be present in the infant brain during object occlusion. The results are illustrated in figure 15.4 and plate 39. Statistical analyses on the average gammaband EROs (20–60 Hz) revealed higher activity in the unexpected than in the expected-disappearance condition both before and after the hand lifted the tunnel. Comparing gamma power in each of the two conditions to the preceding baseline revealed that, prior to the tunnel’s being lifted, gamma power was reduced in the expected-disappearance condition. These ERO changes were largely restricted to the right temporal area. These results demonstrate a sustained period during which gamma power over the right temporal region was consistently higher during an event where infants represented an object despite its being occluded. If this sustained gamma activity is related to representation of nonvisible objects, it should also be evident in an ordinary event of temporary hiding, like the expected-appearance event. Indeed, we found no significant increase in gamma activity over right-temporal channels time locked to the unexpected-appearance event. In a recent experiment we have also demonstrated that the right temporal gamma-band activation does not simply reflect a memory trace of the disappearing object but rather its active maintenance (Kaufman, Csibra, and Johnson, 2005). Six-month-old infants displayed a higher activation when an object disappeared by deletion (consistent with being occluded) than when it disappeared by disintegration. This result supports the view that derives young infants’ competence with moving objects from perceptual routines that track objects through space and time (Scholl and Leslie, 1999). Whatever the exact neural basis of these effects, the

finding that increased gamma-band activity is associated with the representation of hidden objects will inform fundamental issues about how infants process their visual world.

Future directions Ten years ago only a handful of pioneering laboratories recorded and analyzed infant EEGs and ERPs to study the neural bases of cognitive development. Today many researchers use these methods, and we expect that the number of infant electrophysiological laboratories will increase further in the future. However, we also expect that, beyond this horizontal extension, the techniques that applied to these recordings will also improve significantly. Here we discuss just one example of such developments, which has already proven to be a successful tool for understanding the neural bases of cognitive development in infancy. Several attempts have been made to localize ERP components and ERO signals to specific anatomical structures in adults. However, so far there is no generally agreed solution for such an inverse problem. Applying the techniques developed for source localization of adult ERPs to recordings from infants is hampered by two factors. First, these techniques require clean data, which are free from “noise” (brain activity unrelated to the eliciting event), and preferably reflect the activation of a single neural source or very few sources. It is very difficult to obtain such clean recordings from human infants, primarily because of the low number of averageable trials in infant studies. Second, source localization techniques depend on assumptions about anatomical structures like the skull, the cerebrospinal fluid, and the cortical convolutions. The physical parameters, as well as

csibra, kushnerenko, and grossmann: infant cognitive development

257

the maturation, of these anatomical details are not well known for young infants, and applying the parameters used in adult research could lead to mislocalization in infants. A potential way to overcome these difficulties is to apply some kind of statistical method to separate underlying sources behind ERPs in the statistical, rather than in the physical, space. Such techniques may allow us to explore and compare the activity of functionally independent neural sources without committing ourselves to their exact anatomical location. However, if such a source-separation procedure successfully isolates the activation of a neural source, it will be more likely to reflect the functioning of a single structure than are ERP components, and it will be more easily localizable in the brain itself as well. One such technique, which has been applied to infant ERPs, is called independent component analysis (ICA, Bell and Sejnowski, 1995). ICA attempts to decompose the raw EEG signal into the sum of independently generated signals by making assumptions about the statistical distribution of the neural activation of the sources. This kind of decomposition has been shown to successfully isolate neural activity related to face and gaze processing (Johnson et al., 2001, 2005) and allocation of spatial attention (Richards, 2005) in infant ERPs. For example, Johnson and colleagues (2005) found that ICA components reflecting face processing in the occipital and temporal cortices were larger in amplitude when the eyes of the face displayed direct, rather than averted, gaze. This is consistent with earlier ERP reports by Farroni and colleagues (2002) and Farroni, Csibra, and Johnson (2004). However, the ICA analysis also identified further sources that were sensitive to gaze direction, and a subsequent localization attempt estimated that these sources originated from the prefrontal (possibly orbitofrontal) cortex. This result illustrates the possible power of statistical source-separation methods, since such an effect had not been uncovered in traditional ERP analyses. Other neuroimaging methods could help further to clarify the interpretation of ERP and ERO results. Although source localization of infant electrophysiological findings is difficult, mapping these activations onto brain images collected by structural MRI in infants (preferably the same ones who provided the electrophysiological data) reduces the degrees of freedom considerably. Alternatively, magnetoencephalography (MEG), a technique that is much more expensive and technically more challenging than EEG, can also be used with infants (e.g., Imada et al., 2006). Magnetoencephalography recording is less affected by the blurring effect of the several layers of tissues between the cerebral cortex and the sensors, and hence provides a better signal for localization. Another possibility is combining electrophysiological measures with neuroimaging techniques that target functional hemodynamic changes in the brain. Espe-

258

methodological paradigms

cially promising is the combination of EEG and fNIRS methods, because this latter optical imaging technique is noninvasive and requires less cooperation from young infants than does fMRI. Such coregistration is possible (e.g., Koch et al., 2006) and offers a unique opportunity to uncover the functioning of cortical mechanisms of human infants. Although electrophysiological techniques do not offer accurate high-resolution images of the brain, they are valuable tools in assessing the neural underpinnings of the tremendous cognitive development that humans go through during the first years of life. This is why ERPs and related measurements have become ineliminable methods in the cognitive neuroscience of human infants. acknowledgments

The writing of this chapter was partly supported by the UK MRC program grant G9715587, by a Pathfinder grant from the European Commission (CALACEI), and by the Academy of Finland (project 213672). We thank Sarah Fox and Dr. Teodora Gliga for their comments and editorial help.

REFERENCES Abdala, C., and R. C. Folsom, 1995. The development of frequency resolution in humans as revealed by the auditory brainstem response recorded with notched-noise masking. J. Acoust. Soc. Am. 98:921–930. Ackles, P. K., and K. Cook, 1998. Stimulus probability and eventrelated potentials of the brain in 6-month-old human infants: A parametric study. Int. J. Psychophysiol. 29:115–143. Balaban, C. D., and J. M. Weinstein, 1985. The human presaccadic spike potential: Influences of a visual target, saccade direction, electrode laterality and instructions to perform saccades. Brain Res. 347:49–57. Barnet, A. B., E. S. Ohlrich, I. P. Weiss, and B. Shanks, 1975. Auditory evoked potentials during sleep in normal children from ten days to three years of age. Electroencephalogr. Clin. Neurophysiol. 39:29–41. Batty, M., and M. J. Taylor, 2006. The development of emotional face processing during childhood. Dev. Sci. 9:207–220. Bell, M. A., 2001. Brain electrical activity associated with cognitive processing during a looking version of the A-not-B task. Infancy 2:311–330. Bell, M. A., and N. A. Fox, 1992. The relations between frontal brain electrical activity and cognitive development during infancy. Child Dev. 63:1142–1163. Bell, A. J., and T. J. Sejnowksi, 1995. An informationmaximization approach to blind separation and blind deconvolution. Neural Comput. 7:1129–1159. Bentin, S., T. Allison, A. Puce, A. Perez, and A. McCarthy, 1996. Electrophysiological studies of face perception in humans. J. Cogn. Neurosci. 8:551–565. Blume, W. T. 1974. Anatomic correlates of the ten-twenty electrode placement system in infants. Electroencephalogr. Clin. Neurophysiol. 36:303–307. Carver, L. J., G. Dawson, H. Panagiotides, A. N. Meltzoff, J. McPartland, J. Gray, and J. Munson, 2003. Age-related

differences in neural correlates of face recognition during the toddler and preschool years. Dev. Psychobiol. 42:148–159. Ceponiene, R., M. Cheour, and R. Näätänen, 1998. Interstimulus interval and auditory event-related potentials in children: Evidence for multiple generators. Electroencephalogr. Clin. Neurophysiol. 108:345–354. Ceponiene, R., E. Kushnerenko, V. Fellman, M. Renlund, K. Suominen, and R. Näätänen, 2002. Event-related potential features indexing central auditory discrimination by newborns. Cogn. Brain Res. 13:101–113. Ceponiene, R., T. Lepisto, M. Soininen, E. Aronen, P. Alku, and R. Näätänen, 2004. Event-related potentials associated with sound discrimination versus novelty detection in children. Psychophysiology 41:130–141. Ceponiene, R., T. Rinne, and R. Näätänen, 2002. Maturation of cortical sound processing as indexed by event-related potentials. Clin. Neurophysiol. 113:870–882. Ceponiene, R., A. Shestakova, P. Balan, P. Alku, K. Yiaguchi, and R. Näätänen, 2001. Children’s auditory event-related potentials index sound complexity and “speechness.” Int. J. Neurosci. 109:245–260. Cheour, M., R. Ceponiene, J. Hukki, M.-L. Haapanen, R. Näätänen, and K. Alho, 1999. Brain dysfunction in neonates with cleft palate revealed by the mismatch negativity (mmn). Clin. Neurophysiol. 110:324–328. Cheour-Luhtanen, M., K. Alho, T. Kujala, K. Sainio, K. Reinikainen, M. Renlund, O. Aaltonen, O. Eerola, and R. Näätänen, 1995. Mismatch negativity indicates vowel discrimination in newborns. Hear. Res. 82:53–58. Cheour-Luhtanen, M., K. Alho, K. Sainio, T. Rinne, K. Reinikainen, M. Pohjavuori, M. Renlund, O. Aaltonen, O. Eerola, and R. Näätänen, 1996. The ontogenetically earliest discriminative response of the human brain. Psychophysiology, Special Report, 33:478–481. Chugani, H. T., M. E. Phelps, and J. C. Mazziotta, 1987. Positron emision tomography study of human brain functional development. Ann. Neurol. 22:487–497. Courchesne, E., 1978. Neurophysiological correlates of cognitive development: Changes in long-latency event-related potentials from childhood to adulthood. Electroencephalogr. Clin. Neurophysiol. 45:468–482. Courchesne, E., 1983. Cognitive components of the event-related brain potential: Changes associated with development. In Tutorials in ERP Research: Endogenous Components, ed. A. W. K. Gaillard and W. Ritter, 329–344. Amsterdam: North-Holland Elsevier. Courchesne, E., 1990. Chronology of postnatal human brain development: Event-related potential, positron emission tomography, myelinogenesis, and synaptogenesis studies. In EventRelated Potentials: Basic Issues and Applications, ed. R. Rohrbaugh, R. Parasuroman, and R. Johnson, 210–241. New York: Oxford University Press. Courchesne, E., L. Ganz, and M. E. Norcia, 1981. Event-related brain potentials to human faces in infants. Child Dev. 52:804– 811. Csibra, G., G. Davis, M. W. Spratling, and M. H. Johnson, 2000. Gamma oscillations and object processing in the infant brain. Science 290:1582–1585. Csibra, G., and M. H. Johnson, 2007. Investigating event-related oscillations in infancy. In Infant EEG and Event-Related Potentials, ed. M. de Haan, 289–304. Hove, England: Psychology Press. Csibra, G., M. H. Johnson, and L. Tucker, 1997. Attention and oculomotor control: A high-density ERP study of the gap effect. Neuropsychologia 35:855–865.

Csibra, G., L. A. Tucker, and M. H. Johnson, 1998. Neural correlates of saccade planning in infants: A high-density ERP study. Int. J. Psychophysiol. 29:201–215. Csibra, G., L. A. Tucker, A. Volein, and M. H. Johnson, 2000. Cortical development and saccade planning: The ontogeny of the spike potential. NeuroReport 11:1069–1073. Deboer, T., L. Scott, and C. A. Nelson, 2004. Event-related potentials in developmental populations. In Event-Related Potentials: A Methods Handbook, ed. T. Handy, 263–297. Cambridge, MA: MIT Press. de Haan, M., ed., 2007. Infant EEG and Event-Related Potentials. Hove, England: Psychology Press. de Haan, M., M. H. Johnson, and H. Halit, 2003. Development of face-sensitive event-related potentials during infancy: A review. Int. J. Psychophysiol. 51:45–58. de Haan, M., and C. A. Nelson, 1997. Recognition of the mother’s face by six-month-old infants: A neurobehavioral study. Child Dev. 68:187–210. de Haan, M., and C. A. Nelson, 1999. Brain activity differentiates face and object processing by 6-month-old infants. Dev. Psychol. 34:1114–1121. de Haan, M., O. Pascalis, and M. H. Johnson, 2002. Specialization of neural mechanisms underlying face recognition in human infants. J. Cogn. Neurosci. 14:199–209. de Haan, M., and K. M. Thomas, 2002. Applications of ERP and fMRI techniques to developmental science. Dev. Sci. 5:335– 343. Dehaene-Lambertz, G., 2000. Cerebral specialization for speech and non-speech stimuli in infants. J. Cogn. Neurosci. 12:449– 460. Dehaene-Lambertz, G., and S. Baillet, 1998. A phonological representation in the infant brain. NeuroReport 9:1885–1888. Dehaene-Lambertz, G., and S. Dehaene, 1994. Speed and cerebral correlates of syllable discrimination in infants. Nature 28:293–294. Dehaene-Lambertz, G., and T. Gliga, 2004. Common neural basis for phoneme processing in infants and adults. J. Cogn. Neurosci. 8:1375–1387. Dehaene-Lambertz, G., and M. Pena, 2001. Electrophysiological evidence for automatic phonetic processing in neonates. NeuroReport 12:3155–3158. Deregnier, R. A., C. A. Nelson, K. M. Thomas, S. Wewerka, and M. K. Georgieff, 2000. Neurophysiologic evaluation of auditory recognition memory in healthy newborn infants and infants of diabetic mothers. J. Pediatr. 137:777–784. Donchin, E., and M. G. H. Coles, 1988. Is the P300 component a manifestation of context updating? Behav. Brain Sci. 11:357– 374. Eggermont, J. J., 1988. On the rate of maturation of sensory evoked potentials. Electroencephalogr. Clin. Neurophysiol. 70:293–305. Enoki, H., S. Sanada, H. Yoshinada, E. Oka, and S. Ohtahara, 1993. The effects of age on the N200 component of the auditory event-related potentials. Cognit. Brain Res. 1: 161–167. Escera, C., K. Alho, E. Schröger, and I. Winkler, 2000. Involuntary attention and distractibility as evaluated with eventrelated brain potentials. Audiol. Neurootol. 5:151–166. Escera, C., E. Yago, and K. Alho, 2001. Electrical responses reveal the temporal dynamics of brain events during involuntary attention switching. Eur. J. Neurosci. 14:877–883. Farroni, T., G. Csibra, and M. H. Johnson, 2004. Mechanisms of eye gaze perception during infancy. J. Cogn. Neurosci. 16: 1320–1326.

csibra, kushnerenko, and grossmann: infant cognitive development

259

Farroni, T., G. Csibra, F. Simion, and M. H. Johnson, 2002. Eye contact detection in humans from birth. Proc. Natl. Acad. Sci. USA 99:9602–9605. Friederici, A. D., 2002. Towards a neural basis of auditory sentence processing. Trends Cogn. Sci. 6:78–84. Friederici, A. D., 2005. Neurophysiological markers of early language acquisition: From syllables to sentences. Trends Cogn. Sci. 9:481–488. Friederici, A. D., M. Friedrich, and C. Weber, 2002. Neural manifestation of cognitive and precognitive mismatch detection in early infancy. NeuroReport 13:1251–1254. Friedman, D., 1991. The endogenous scalp-recorded brain potentials and their relation to cognitive development. In Handbook of Cognitive Psychophysiology: Central and Autonomic Nervous System Approaches, ed. J. R. Jennings and M. G. H. Coles, 621–656. New York: John Wiley. Friedrich, M., and A. D. Friederici, 2005. Phonotactic knowledge and lexical-semantic processing in one-year-olds: Brain responses to words and nonsense words in picture contexts. J. Cogn. Neurosci. 17:1785–1802. Friedrich, M., C. Weber, and A. Friederici, 2004. Electrophysiological evidence for delayed mismatch response in infants at-risk for specific language impairment. Psychophysiology 41: 772–782. Gliga, T., and G. Dehaene-Lambertz, 2007. Development of a view-invariant representation of the human head. Cognition. 102:261–288. Graziani, L. J., L. Katz, Q. Cracco, J. B. Cracco, and E. D. Weitzman, 1974. The maturation and interrelationship of EEF patterns and auditory evoked response in premature infants. Electroencephalogr. Clin. Neurophysiol. 36:367–375. Grice, S. J., H. Halit, T. Farroni, S. Baron-Cohen, P. Bolton, and M. H. Johnson, 2005. Neural correlates of eye-gaze detection in young children with autism. Cortex 41:342– 353. Grill-Spector, K., and R. Malach, 2001. fMRI-adaptation: A tool for studying the functional properties of human cortical neurons. Acta Psychol. (Amst.) 107:293–321. Grossmann, T., T. Striano, and A. D. Friederici, 2005. Infants’ electric brain responses to emotional prosody. NeuroReport 16: 1825–1828. Grossmann, T., T. Striano, and A. D. Friederici, 2006. Crossmodal integration of emotional information from face and voice in the infant brain. Dev. Sci. 9:309–315. Gumenyuk, V., O. Korzyukov, K. Alho, C. Escera, and R. Näätänen, 2004. Effects of auditory distraction on electrophysiological brain activity and performance in children aged 8–13 years. Psychophysiology 41:30–36. Halit, H., M. DeHaan, and M. H. Johnson, 2003. Cortical specialisation for face processing: Face-sensitive event-related potential components in 3- and 12-month-old infants. NeuroImage 19:1180–1193. Halit, H., G. Csibra, A. Volein, and M. H. Johnson, 2004. Facesensitive cortical processing in early infancy. J. Child Psychol. Psychiatry 45:1228–1234. Handy, T. C., ed., 2005. Event-Related Potentials. Cambridge, MA: MIT Press. Herrmann, C. S., M. Grigutsch, and N. A. Busch, 2005. EEG oscillations and wavelet analysis. In Event-Related Potentials: A Methods Handbook, ed. T. Handy, 229–259. Cambridge, MA: MIT Press. Herrmann, C. S., and A. Mecklinger, 2000. Magnetoencephalographic responses to illusory figures: Early evoked gamma is

260

methodological paradigms

affected by processing of stimulus features. Int. J. Psychophysiol. 38:265–281. Huttenlocher, P. R., 1979. Synaptic density in human frontal cortex—Developmental changes and effects of aging. Brain Res. 163:195–205. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Imada, T., et al., 2006. Infant speech perception activates Broca’s area: A developmental magnetoencephalography study. NeuroReport 17:957–962. Jasper, H. H., 1958. The ten-twenty electrode system of the international federation of societies for electroencephalography: Appendix to report of the committee on methods of clinical examinations in electroencephalography. Electroencephalogr. Clin. Neurophysiol. 10:371–375. Johnson, M. H., M. DeHaan, A. Oliver, W. Smith, H. Hatzakis, L. A. Tucker, and G. Csibra, 2001. Recording and analyzing high-density event-related potentials with infants: Using the geodesic sensor net. Dev. Neuropsychol. 19:295–323. Johnson, M. H., R. Griffin, G. Csibra, H. Halit, T. Farroni, M. DeHaan, L. Tucker, S. Baron-Cohen, and J. Richards, 2005. The emergence of the social brain network: Evidence from typical and atypical development. Dev. Psychopathol. 17: 599–619. Karhu, J., E. Herrgard, A. Paakkonen, L. Luoma, E. Airaksinen, and J. Partanen, 1997. Dual cerebral processing of elementary auditory input in children. NeuroReport 8:1327–1330. Karrer, R., and P. K. Ackles, 1987. Visual event-related potentials of infants during a modified oddball procedure. Electroencephalogr. Clin. Neurophysiol. 49 (suppl.): 603–608. Karrer, R., and L. A. Monti, 1995. Event-related potentials of 4–7 week old infants in a visual recognition memory task. Electroencephalogr. Clin. Neurophysiol. 94:414–424. Kaufman, J., G. Csibra, and M. H. Johnson, 2003. Representing occluded objects in the human infant brain. Proc. R. Soc. Lond [Biol.] 270:140–143. Kaufman, J., G. Csibra, and M. H. Johnson, 2005. Oscillatory activity in the infant brain reflects object maintenance. Proc. Natl. Acad. Sci. USA 102:15271–15274. Koch, S. P., J. Steinbrink, A. Villringer, and H. Obrig, 2006. Synchronization between background activity and visual evoked potential is not mirrored by focal hyperoxigenization: Implications for the interpretation of vascular brain imaging. J. Neurosci. 26(18):4940–4948. Kooi, K. A., 1971. Fundamentals of Electroencephalography (1st ed.). New York: Harper and Row. Korpilahti, P., and H. A. Lang, 1994. Auditory ERP components and mismatch negativity in dysphasic children. Electroencephalogr. Clin. Neurophysiol. 91:256–264. Korpilahti, P., K. M. Krause, I. Holopainen, and H. Lang, 2001. Early and late mismatch negativity elicited by words and speech-like stimuli in children. Brain Lang. 76:332–339. Kurtzberg, D., P. L. Hilpert, J. A. Kreuzer, and H. G. J. Vaughan, 1984. Differential maturation of cortical auditory evoked potentials to speech sounds in normal fullterm and very low-birthweight infants. Dev. Med. Child Neurol. 26: 466–475. Kurtzberg, D., and H. G. Vaughan, Jr., 1977. Electrophysiological observation on the visuomotor system and visual neurosensorium. In Visual Evoked Potentials in Man: New Developments, ed. J. E. Desmedt, 314–331. Oxford, UK: Clarendon Press.

Kurtzberg, D., and H. G. Vaughan, Jr., 1985. Electrophysiologic assessment of auditory and visual function in the newborn. Clin. Perinatol. 12:277–297. Kurtzberg, D., H. G. Vaughan, Jr., J. A. Kreuzer, and K. Z. Fliegler, 1995. Developmental studies and clinical application of mismatch negativity: Problems and prospects. Ear Hear. 16:105–117. Kurtzberg, D., H. G. Vaughan, Jr., and G. P. Novak, 1986. Discriminative brain responses to speech sounds in the newborn high risk infant. In Maturation of the CNS and Evoked Potentials, ed. V. Gallai, 253–259. Amsterdam: Elsevier. Kushnerenko, E., R. Ceponiene, P. Balan, V. Fellman, M. Huotilainen, and R. Näätänen, 2002a. Maturation of the auditory event-related potentials during the 1st year of life. NeuroReport 13:47–51. Kushnerenko, E., R. Ceponiene, P. Balan, V. Fellman, and R. Näätänen, 2002b. Maturation of the auditory change-detection response in infants: A longitudinal ERP study. NeuroReport 13: 1843–1848. Kushnerenko, E., R. Ceponiene, V. Fellman, M. Huotilainen, and I. Winkler, 2001. Event-related potential correlates of sound duration: Similar pattern from birth to adulthood. NeuroReport 12:3777–3781. Kushnerenko, E., I. Winkler, J. Horváth, R. Näätänen, I. Pavlov, V. Fellman, et al., 2006. Newborn infants process contextual novelty. Manuscript submitted for publication. Kutas, M., and S. A. Hillyard, 1983. Event-related brain potentials to grammatical errors and semantic anomalies. Mem. Cogn. 11:539–550. Leppanen, P., T. Guttorm, E. Pihko, S. Takkinen, K. Eklund, and H. Lyytinen, 2004. Maturational effects on newborn ERPs measured in the mismatch negativity paradigm. Exp. Neurol. 190:91–101. Llinas, R. R. 1988. The intrinsic electrophysiological properties of mammalian neurons: Insights into central nervous system function. Science (Amst.) 242:1654–1664. Mills, D. L., C. Prat, R. Zangl, C. L. Stager, H. J. Neville, and J. F. Werker, 2004. Language experience and the organization of brain activity to phonetically similar words: ERP evidence from 14- and 20-month-olds. J. Cogn. Neurosci. 16: 1452–1464. Morr, M. L., V. L. Shafer, J. A. Kreuzer, and D. Kurtzberg, 2002. Maturation of mismatch negativity in typically developing infants and preschool children. Ear Hear. 23:118–136. Mundy, P., J. Card, and N. Fox, 2000. Fourteen month cortical activity and different infant joint attention skills. Dev. Psychobiol. 36:325–338. Näätänen, R., 1992. Attention and Brain Function. Hillsdale, NJ: Lawrence Erlbaum. Näätänen, R., A. W. K. Gaillard, and S. Mantysalo, 1978. Early selective attention effect on evoked potential reinterpreted. Acta Psychol. (Amst.) 42:313–329. Näätänen, R., and T. Picton, 1987. The N1 wave of the human electric and magnetic response to sound: A review and an analysis of the component structure. Psychophysiology 24:375–425. Nelson, C. A., 1994. Neural correlates of recognition memory in the first postnatal year of life. In Human Behavior and the Developing Brain, ed. G. Dawson and K. Fischer, 269–313. New York: Guilford Press. Nelson, C. A., and P. F. Collins, 1991. Event-related potential and looking time analysis of infants’ responses to familiar and novel events: Implications for visual recognition memory. Dev. Psychol. 27:50–58.

Nelson, C. A., and M. de Haan, 1996. Neural correlates of infants’ visual responsiveness to facial expression of emotion. Dev. Psychobiol. 29:577–595. Nelson, C. A., M. Henschel, and P. F. Collins, 1993. Neural correlates of crossmodal recognition memory by 8-month-old human infants. Dev. Psychol. 29:411–420. Nelson, C. A., and M. Luciana, 1998. Electrophysiological studies. II. Evoked potentials and event-related potentials. In Textbook of Neuropsychiatry, ed. C. E. Coffey and R. A. Brumbeck, 331–356. Washington, DC: American Psychiatric Press. Nelson, C. A., and C. Monk, 2001. The use of event-related potentials in the study of cognitive development. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 125–136. Cambridge, MA: MIT Press. Nelson, C. A., K. M. Thomas, M. DeHaan, and S. Wewerka, 1998. Delayed recognition memory in infants and adults as revealed by event-related potentials. Int. J. Psychophysiol. 29:145–165. Novak, G. P., D. Kurtzberg, J. A. Kreuzer, and H. G. J. Vaughan, 1989. Cortical responses to speech sounds and their formants in normal infants: Maturational sequence and spatiotemporal analysis. Electroencephalogr. Clin. Neurophysiol. 73:295–305. Nuwer, M. R., G. Comi, R. Emerson, A. Fuglsang-Frederiksen, J. M. Guerit, H. Hinrichs, A. Ikeda, F. J. Luccas, and P. Rappelsburger, 1998. ICFN standards for digital recording of clinical EEG. Electroencephalogr. Clin. Neurophysiol. 106:259–261. Oberecker, R., M. Friedrich, and A. D. Friederici, 2005. Neural correlates of syntactic processing in two-year-olds. J. Cogn. Neurosci. 17:1667–1678. Ohlrich, E. S., A. B. Barnet, I. P. Weiss, and B. L. Shanks, 1978. Auditory evoked potential development in early childhood: A longitudinal study. Electroencephalogr. Clin. Neurophysiol. 44: 411–423. Oostenveld, R., and P. Praamstra, 2001. The five percent electrode system for high-resolution EEG and ERP measurements. Clin. Neurophysiol. 112:713–719. Paetau, R., A. Ahonen, O. Salonen, and M. Sams, 1995. Auditory evoked magnetic fields to tones and pseudowords in healthy children and adults. J. Clin. Neurophysiol. 12:177–185. Pasman, J. W., J. J. Rotteveel, R. DeGraaf, D. F. Stegeman, and Y. M. Visco, 1992. The effect of preterm birth on brainstem, middle latency and cortical auditory evoked responses (BMC AERs). Early Hum. Dev. 31:113–129. Picton, T. W., S. A. Hillyard, H. I. Krausz, and R. Galambos, 1974. Human auditory evoked potentials. I. Evaluation of components. Electroencephalogr. Clin. Neurophysiol. 36:179–190. Ponton, C. W., J. J. Eggermont, B. Kwong, and M. Don, 2000. Maturation of human central auditory system activity: Evidence from multi-channel evoked potentials. Clin. Neurophysiol. 111: 220–236. Purhonen, M., R. Kilpeläinen-Lees, M. Valkonen-Korhonen, J. Karhu, and J. Lehtonen, 2004. Cerebral processing of mother’s voice compared to unfamiliar voice in 4-month-old infants. Int. J. Psychophysiol. 52:257–266. Quinn, P. C., A. Westerlund, and C. A. Nelson, 2006. Neural markers of categorization in 6-month-old infants. Psychol. Sci. 17:59–66. Rainer, G., and E. K. Miller, 2000. Effects of experience on the representation of objects in the prefrontal cortex. Neuron 27: 179–189. Reynolds, G. D., and J. E. Richards, 2005. Familiarization, attention, and recognition memory in infancy: An ERP and cortical source analysis study. Dev. Psychol. 41:598–615.

csibra, kushnerenko, and grossmann: infant cognitive development

261

Richards, J. E., 2000. Localizing the development of covert attention in infants using scalp event-related-potentials. Dev. Psychol. 36:91–108. Richards, J. E., 2002. The development of visual attention and the brain. In The Cognitive Neuroscience of Development, ed. M. de Haan and M. H. Johnson, 73–98. Hove, UK: Psychology Press. Richards, J. E., 2005. Localizing cortical sources of event-related potentials in infants’ covert orienting. Dev. Sci. 8:255–278. Rossion, B., S. Campanella, C. M. Gomez, A. Delinte, D. Debatisse, L. Liard, S. Dubois, R. Bruyer, M. Cromelinck, and J. M. Guerit, 1999. Task modulation of brain activity related to familiar and unfamiliar face processing: An eventrelated potential study. Biol. Psychol. 50:173–189. Rotteveel, J. J., E. J. Colon, D. F. Stegeman, and Y. M. Visco, 1987. The maturation of the central auditory conduction in preterm infants until three months post term. IV. Composite group averages of the cortical auditory evoked responses (ACRS). Hear. Res. 27:85–93. Scholl, B. J., and A. M. Leslie, 1999. Explaining the infant’s object concept: Beyond the perception/cognition dichotomy. In What Is Cognitive Science? ed. E. Lepore and Z. Pylyshyn, 26–73. Oxford, UK: Blackwell. Schröeger, E., and C. Wolff, 1998. Attentional orienting and re-orienting is indicated by human event-related brain potentials. NeuroReport 9:3355–3358. Schröeger, E., M. H. Giard, and C. Wolff, 2000. Auditory distraction: Event-related potential and behavioral indices. Clin. Neurophysiol. 111:1450–1460. Sharma, A., N. Kraus, T. J. McGee, and T. G. Nicol, 1997. Developmental changes in P1 and N1 central auditory responses elicited by consonant-vowel syllables. Electroencephalogr. Clin. Neurophysiol. 104:540–545. Shucard, D. W., J. L. Shucard, and D. G. Thomas, 1987. Auditory event-related potentials in waking infants and adults: A developmental perspective. Electroencephalogr. Clin. Neurophysiol. 68:303–310. Snyder, K., S. J. Webb, and C. A. Nelson, 2002. Theoretical and methodological implications of variability in infant brain response during recognition memory paradigm. Infant Behav. Dev. 25: 466–494. Sokolov, E. N., J. A. Spinks, R. Näätänen, and H. Lyytinen, 2002. The Orienting Response in Information Processing. Mahwah, NJ: Lawrence Erlbaum.

262

methodological paradigms

Squires, K. C., N. K. Squires, and S. A. Hillyard, 1975. Decision-related cortical potentials during an auditory signal detection task with cued observation intervals. J. Exp. Psychol. [Hum. Percept.] 1:268–279. Striano, T., V. M. Reid, and S. Hoehle, 2006. Neural mechanisms of joint attention in infancy. Eur. J. Neurosci. 23: 2819–2823. Tallon-Baudry, C., and O. Bertrand, 1999. Oscillatory gamma activity in humans and its role in object representation. Trends. Cogn. Sci. 3:151–162. Tallon-Baudry, C., O. Bertrand, F. Perronnet, and J. J. Pernier, 1998. Induced gammaband activity during the delay of a visual short-term memory task in humans. Neuroscience 18:4244–4254. Taylor, M. J., M. Batty, and R. J. Itier, 2004. The faces of development: A review of early face processing over childhood. J. Cogn. Neurosci. 16:1426–1442. Taylor, M. J., R. J. Itier, T. Allison, and G. E. Edmonds, 2001. Direction of gaze effects on early face processing: Eyes-only vs. full faces. Cogn. Brain Res. 10:333–340. Thomas, D. G., and C. D. Crow, 1994. Development of evoked electrical brain activity in infancy. In Human Behavior and the Developing Brain, ed. G. Dawson and K. W. Fisher, 207–231. New York: Guilford Press. Tucker, D. M., 1993. Spatial sampling of head electrical fields: The geodesic sensor net. Electroencephalogr. Clin. Neurophysiol. 87:154–163. Vaughan, H. G. J., and D. Kurtzberg, 1992. Electrophysiologic indices of human brain maturation and cognitive development. In Minnesota Symposia on Child Psychology, ed. M. R. Gunnar and C. A. Nelson, vol. 24, pp. 1–36. Hillsdale, NJ: Erlbaum. Webb, S. J., J. D. Long, and C. A. Nelson, 2005. A longitudinal investigation of visual event-related potentials in the first year of life. Dev. Sci. 8:605–616. Werner, L. A., 1996. The development of auditory behavior (or what the anatomists and psysiologists have to explain). Ear Hear. 17:438–445. Winkler, I., E. Kushnerenko, J. Horvath, R. Ceponiene, V. Fellman, M. Huotilainen, M. R. Naatanen, and E. Sussman, 2003. Newborn infants can organize the auditory world. Proc. Natl. Acad. Sci. USA 100:11812–11815. Wunderlich, J. L., B. K. Cone-Wesson, and R. Shepherd, 2006. Maturation of the cortical auditory evoked potential in infants and young children. Hear. Res. 212:185–202.

16

Eye Tracking Studies of Normative and Atypical Development CANAN KARATEKIN

The eyes have been viewed as a window to the mind in typical and clinical populations for more than a century. We take in the world through our eyes, and almost everything we do during the time we are awake involves eye movements. Furthermore, although we are not aware of it, our pupils dilate rapidly all the time in response to a large array of cognitive and emotional stimuli, whether internally or externally generated. Thus our eyes reveal a great deal about what we are thinking and feeling, and eye tracking measures can harness this potential to improve our understanding of the mind and its development. Eye tracking measures have been used to elucidate a wide variety of cognitive processes, from visual-spatial attention to object perception, memory, and language. They can also be helpful in examining socioemotional processes, such as motivation, responses to different types of rewards, and aspects of social information processing. Because the neural substrates of eye movements are fairly well established, eye tracking has also been used to make inferences about the brain. Although there is a very large and sound body of research on eye tracking in adults and nonhuman primates, this research has so far been vastly underutilized in research with children and adolescents. This article will provide an overview of eye tracking studies in healthy and clinical populations of children and adolescents with the hope that these measures can be added to the tool kit of developmentalists as they seek to understand cognitive and social processes and their neural substrates in typical and atypical development. In the first section, I will briefly review key eye tracking measures to provide some basic background information for the rest of the paper. The next two sections will summarize studies in which these measures have been used in typical and atypical populations of children and adolescents. In the final section, I will evaluate studies that have used eye tracking as a research tool, emphasizing points that may be of general interest to developmentalists, list limitations of eye tracking as a tool, and end with its potentials for addressing developmental questions. The goal of the paper is not to review how eye tracking measures have

contributed to understanding of specific developmental or clinical phenomena or models and theories of normative and atypical development (which would require placing these studies in the context of a broader body of research that does not involve eye tracking or children), but to provide a sense of why and how these measures have been used in typical and atypical populations of children and adolescents and to appraise these measures as tools for probing development. Eye movements in children have been reviewed by Hainline (1988), and this paper will focus on research conducted since 1988. I will also limit the review to studies of participants from 4 or 5 years of age through adolescence. For reviews of eye tracking in infancy, the reader is referred to Haith (2004), Simion and Butterworth (1998), and von Hofsten (2004). To limit the scope of the manuscript, I will not cover research on eye movements during reading (for reviews see Liversedge and Findlay, 2000; Rayner, 1998; Starr and Rayner, 2001).

Eye tracking measures A summary of common eye tracking tasks, typical measures used on these tasks, and what the measures assess is in table 16.1. Pictures of eye monitors can be seen in figure 16.1. Saccades Saccades are ballistic eye movements aimed at bringing objects into foveal vision. Thus a saccade to an object normally coincides with an overt shift of visual-spatial attention to that object. Saccades can be divided roughly into externally versus internally guided saccades. Externally guided saccades are often assessed on visually guided saccade tasks (labeled prosaccades in the rest of this paper), where participants are instructed to look at a visual stimulus as soon as it appears. In a number of the studies reviewed in this chapter, saccade tasks included conditions in which there was a 200-ms gap or overlap between fixation and the target. The reduction of saccadic response times (RTs) in the gap condition compared to those in a typical condition with no gap (sometimes

263

Task All saccade tasks

Gap and overlap tasks

Antisaccade task Memory-guided saccade task Predictive saccade task Active fixation tasks Pursuit

Scene/face perception tasks

Pupillary dilation tasks

Table 16.1 Summary of common eye tracking tasks and measures Common Measures What Is Measured? Duration, peak velocity and amplitude of Basic dynamics of saccades the saccade Gain (saccade amplitude/target amplitude) Spatial accuracy of the saccade (hypometric saccades undershoot the target, whereas hypermetric saccades overshoot the target) Latency to initiate the saccade (RT) Speed of processing and movements of visualspatial attention Variability of saccadic RTs Variability of speed of processing and movements of visual-spatial attention Frequency of express saccades (saccades Disengagement of visual-spatial attention with very short RTs, i.e., between 80 and 130 ms) Corrective saccades (saccades to the correct Ability to monitor performance, perceive and location after an initial error) correct errors Premature saccades (saccades prior to Ability to inhibit disallowed saccades target onset despite instructions to fixate) Gap effect (reduction in average saccadic Disengagement of visual-spatial attention and RT when there is a temporal gap nonspecific response preparation between fixation and target) Increase in average saccadic RT when Engagement of visual-spatial attention at fixation there is a temporal overlap between fixation and target) Accuracy (whether the saccade was in the Ability to inhibit a saccade to the disallowed correct direction) location Spatial accuracy of the saccades (distance Accuracy of visual-spatial working memory error) Frequency of predictive saccades for targets Ability to form an internal representation of the whose location and/or timing is target and to predict its occurrence predictable Intrusive saccades during fixation Ability to maintain fixation, sustained attention Root-mean-square error (the difference Overall efficiency of the pursuit system and its between target and gaze position during interaction with the saccadic system pursuit) Position gain (gaze position/target position) Overall efficiency of the pursuit system and its interaction with the saccadic system Velocity gain (gaze velocity/target velocity) Efficiency of pursuit, independent of catch-up and intrusive saccades Compensatory saccades (to catch up with Inefficiency in pursuit, compensated for by the the target) saccadic system Intrusive saccades (saccades that anticipate Disruption of pursuit by the saccadic system the target’s location and “square wave resulting from failure to inhibit saccades jerks”) Initiation phase (pursuit during the first Visually guided, dependent on bottom-up 100–120 ms) information by onset of target movement Maintenance phase (pursuit after the first Internally guided, dependent on top-down 100–120 ms) information about target velocity Location and sequencing of fixations Allocation of visual-spatial attention across the scene/face Duration of fixations Duration of processing foveal and parafoveal information and deciding where to look next Distance between fixations Breadth of visual-spatial attention Allocation of resources Peak pupilary dilation

Latency to peak pupillary dilation Abbreviation: RT, response time.

Speed of processing

A

B

C

Figure 16.1 Remote tabletop and head-mounted eye monitors. (From Trueswell et al., 1999; reprinted with permission from Elsevier.)

called the null condition) is termed the gap effect. It has been hypothesized that the reduced saccadic RTs typically observed in the gap task reflect early disengagement of attention from fixation and nonspecific response preparation processes resulting from the warning provided by the offset of fixation, whereas elevated RTs reflect the longer period of time needed to disengage attention from fixation (e.g., Fischer and Weber, 1993; Kingstone and Klein, 1993; Pratt, Bekkering, and Leung, 2000; Spantekow et al., 1999). Internally guided saccades are executed in the absence of a visual stimulus. Saccades to a location opposite from a visual stimulus (antisaccades), or to the predicted (predictive

saccades) or remembered location of a visual stimulus (memory-guided saccades) fall under this category. The neurobiological bases of prosaccades are depicted in figure 16.2. Internally guided saccades are mediated by additional regions (for reviews see Carpenter, 1988; Everling and Fischer, 1998; Hikosaka, Takikawa, and Kawagoe, 2000; Leigh and Zee, 1999; Leigh and Kennard, 2004; Pierrot-Deseilligny et al., 1997). The main measures extracted from saccades include duration, peak velocity, amplitude, gain (saccade amplitude/stimulus amplitude), and latency to initiate the saccade.

karatekin: eye tracking studies of normative and atypical development

265

Figure 16.2 Lateral view of human cerebral cortex and projections to superior colliculus (SC) involved in saccade triggering. (A) Cortical and subcortical areas involved in oculomotor control, with excitatory and inhibitory pathways depicted in solid and broken lines, respectively. Direct excitatory pathways to the SC shown are from the dorsolateral prefrontal cortex (DLPFC), frontal eye fields (FEF), parietal eye fields (PEF), and supplementary eye fields (SEF). Indirect cortical input from the DLPFC and FEF are through the caudate nucleus (CN), which inhibits the substantia nigra pars

reticulata (SNpr) and which, in turn, inhibits the SC. Cerebellar connections and pontine connections to the SC are also shown. (B) Retinotopic map of left SC depicting fixation neurons in rostral portions and saccade neurons in caudal portions. Directional coding is shown with upward direction in superior and downward direction in inferior regions. (Figure and figure caption reprinted from Reilly et al., 2005, with permission from the Society of Biological Psychiatry.)

Pursuit To track small objects that move relatively slowly and smoothly, we use smooth-pursuit eye movements (Fukushima, 2003). These are smooth, nonballistic movements that match gaze velocity to target velocity and keep the object within foveal vision. The neural substrates of smooth-pursuit eye movements are displayed in figure 16.3 and plate 40 (for reviews see Fukushima, 2003; Krauzlis, 2005; Krauzlis and Stone, 1999; Thier and Ilg, 2005). Although the pursuit and saccade systems have separable neural substrates, they nevertheless work in an integrated fashion in tracking moving objects (e.g., Fukushima, 2003; Krauzlis and Stone, 1999; Liston and Krauzlis, 2003; Missal and Keller, 2002). Smooth-pursuit eye movements are usually assessed by instructing participants to visually track a small stimulus that moves at a relatively slow and predictable velocity along a horizontal path. There is controversy regarding the most appropriate measures to assess the integrity of the pursuit system (Hutton and Kennard, 1998). The main quantitative measure used in most recent studies is gain (defined as peak [or mean] gaze velocity divided by peak [or mean] target velocity). Low gain scores suggest difficulty in matching gaze to target velocity, suggesting inefficiency in the functioning of the pursuit system. Performance can also be assessed by

root-mean-square error (RMSE; Clementz, Iacono, and Grove, 1996), which is based on the difference between gaze and target position throughout tracking. RMSE is calculated by taking the square of the difference between target and gaze positions at each artifact-free point during pursuit, averaging the squares, and then taking the square root of this average. Additional measures of performance include compensatory and intrusive saccades during tracking (Hutton and Kennard, 1998; Ross et al., 1996). Pursuit performance can also be divided into initiation and maintenance phases (e.g., Fukushima, 2003; Avila et al., 2006; Thier and Ilg, 2005).

266

methodological paradigms

Eye Movements During Scene and Face Perception Whereas simple paradigms requiring participants to look at a few stimuli are informative regarding the cognitive and neural correlates of saccades, eye movements during scene or face perception are more useful for examining information processing in more naturalistic contexts. We normally make 3 to 4 saccades a second and pause in between (fixate) for 300–400 ms at a time, to take in the information at the fovea and to decide where to fixate next. As demonstrated decades ago by Buswell (1935) and Yarbus (1965/1967), people look

Figure 16.3 The major substrates of smooth-pursuit eye movements and their connections. Broken lines indicate connections that are still hypothetical or have not been elucidated in sufficient detail. The scheme considers observations, not discussed in the main text, that suggest that signals for horizontal and vertical smooth pursuit are dealt with by different parts of the vestibular complex: namely, horizontal smooth pursuit by medial vestibular nuclei; and vertical smooth pursuit by the y-group—a small cell group that caps the

inferior cerebellar peduncle and that, similar to vestibular complex neurons, receives primary vestibular afferents. Abbreviations: FEF, frontal eye field; LGN, lateral geniculate nucleus; MST, middle superior temporal; MT, middle temporal; NRTP, nucleus reticularis tegmenti pontis; PN, pontine nuclei; SEF, supplementary eye field; V1, primary visual cortex; VN, vestibular nuclei. (Figure and figure caption reprinted from Thier and Ilg, 2005, with permission from Elsevier.) (See plate 40.)

at informative regions when shown a picture of a scene or of a face. Furthermore, when given more time to look at the picture, they return again and again to these informative regions rather than covering the whole area of the picture (e.g., see figure 16.4 and plate 41). It is especially under these conditions that it becomes obvious that eye movements do not reflect a passive type of perception but represent active, goal-directed movements. Reviews of recent research on eye tracking during scene perception can be found in Hayhoe and Ballard (2005), Henderson (2003), Henderson and Hollingworth (1999), and Land and Furneaux (1997). The main measures used in scene/face perception paradigms include the location, duration, and sequencing of fixations, and distance between fixations. The location and sequencing of fixations are used to infer what individuals are attending to and in which order. The duration of fixations is used as a measure of speed of processing foveal and extrafoveal information, and distance between fixations is used to

estimate the width of the attentional spotlight. In general, fixation duration increases and saccadic amplitude decreases as task difficulty and the need to gather more fine-grained information increases. Pupillary Dilation The main factor that regulates pupillary diameter is amount of light. However, pupillary diameter also varies as a function of task-specific recruitment of cognitive resources. The relation between pupillary diameter and task-specific pupillary dilation has been likened to that between spontaneous electroencephalogram records and event-related potentials (ERPs; Beatty, 1982). Tonic changes in pupillary diameter are influenced by general factors, such as level of arousal, anxiety, and stress. Task-specific pupillary dilations, however, are phasic changes in pupillary diameter time locked to the onset of stimuli or to responses. Correlations between pupillary dilations and indices of autonomic function (e.g., heart rate, galvanic skin

karatekin: eye tracking studies of normative and atypical development

267

Figure 16.4 Fixations made by an observer while making a peanut butter and jelly sandwich. Images were taken from a camera mounted on the head, and a composite image mosaic was formed by integrating over different head positions using a method described in Rothkopf and Pelz (2004). (The reconstructed panorama shows artifacts due to the incomplete imaging model that

does not take the translational motion of the subject into account.) Fixations are shown as yellow circles, with diameter proportional to fixation duration. Red lines indicate the saccades. Note that almost all fixations fall on task-relevant objects. (Figure and figure caption reprinted from Hayhoe and Ballard, 2005, with permission from Elsevier.) (See plate 41.)

response) are not high, consistent with the idea that neural control of pupillary dilation lies at the intersection of the autonomic and central nervous systems (Loewenfeld, 1993). Similarly, although pupillary dilations and ERPs covary, they are not perfectly correlated, indicating that they reflect different aspects of processing (Steinhauer and Hakerem, 1992). Pupillary dilations show a remarkable sensitivity to working-memory load on tasks such as the digit span task, where dilations increase linearly with each increase in memory load, reaching a peak just before participants repeat back the digits, and level off or decrease when the number of digits to be remembered exceeds memory span (Granholm et al., 1997; Kahneman and Beatty, 1966; Kahneman, Onuska, and Wolman, 1968; Peavler, 1974). In mental rotation tasks, pupillary dilation also increases linearly with angular disparity (Just, Carpenter, and Miyake, 2003). Phasic pupillary dilations increase with task difficulty

in many other tasks (for a review, see Beatty and LuceroWagoner, 2000). Importantly, the pupil dilates even to the absence of expected stimuli, indicating clearly that dilations are not passive reactions to perceptual stimuli (Qiyuan et al., 1985). Task-specific pupillary dilations are mediated by rapid interactions among the frontal cortex, the thalamus, and the reticular activating system, resulting in activation of sympathetic pathways and inhibition of parasympathetic pathways that terminate in the muscles controlling pupillary diameter (Beatty and Lucero-Wagoner, 2000). These dilations “likely reflect the cortical modulation of the reticular core” and level of arousal in accordance with task demands (Beatty, 1982, p. 290). The effect of task difficulty on pupillary dilation is hypothesized to be mediated primarily by cortical inhibition of the parasympathetic pathway, rather than activation of the sympathetic pathway (Steinhauer et al., 2004), and increasing dilation with parametric increases in task

268

methodological paradigms

difficulty on a digit-sorting task is paralleled by increasing activation in the middle frontal gyrus, measured through functional brain imaging (Siegle, Steinhauer, Stenger et al., 2003). Task-specific pupillary dilations are quantified through peak dilation and pupillary waveforms during a brief period (e.g., 1–2 s) after stimulus onset or prior to response. Depending on the task, different components of pupillary waveforms over longer periods of time (e.g., 10–20 s) can also be analyzed (e.g., Granholm and Verney, 2004; Siegle, Steinhauer, Carter et al., 2003). Another measure is latency to peak dilation, an index of processing speed (Beatty and LuceroWagoner, 2000).

Eye tracking in children and adolescents: normative development saccades There has been a great deal of research on the development of saccades, especially pro- and antisaccades, in childhood and adolescence (summarized in table 16.2). The main concern of a number of these studies has been to delineate the shape of the developmental trajectory on saccade tasks, using large samples and statistical techniques that go beyond simple correlations. In general, results indicate that the basic dynamics of prosaccades (peak velocity and duration) are mature by age 4–6 years. In contrast, prosaccade RTs decrease gradually through at least adolescence, and antisaccade RTs and errors show much steeper rates of improvement during this period (see figure 16.5). The shape of the trajectory and the age at which performance reaches maturity differ across studies for both pro- and antisaccades, probably because of differences in sample sizes, task parameters, and the method of analyzing age-related changes. In one study, developmental data were used to address a question formulated in research with adults (Klein and Fischer, 2005a). It had been argued previously that express saccades (saccades with RTs of 80–130 ms) can be distinguished from regular saccades in terms of their neural substrates and that they reflect the state of attentional engagement. Principal components analyses indicated that both express prosaccades and “express errors” on the antisaccade task loaded on the same factor, which showed little developmental change. Antisaccade errors and prosaccade RTs loaded on another factor, which was not related to express errors and which showed developmental change. The authors used these developmental differences to bolster the argument that express saccades are distinct from regular saccades. The primary goal in several other studies was to draw inferences about cognitive development that go beyond the oculomotor system. These include studies of visual-spatial attention in the presence of internal or external distractors

(Ross, Radant, et al., 1994), top-down and bottom-up control of inhibition (Kramer, Gonzalez de Sather, and Cassavaugh, 2005), and the relations among inhibition, working memory, and speed of processing (Luna et al., 2004). Only three studies have examined the neural correlates of saccades in children. Luna and colleagues (2001) compared three age groups on pro- and antisaccade tasks while they underwent functional brain imaging. There were no agerelated changes in brain activation in the supplementary eye fields, insula, precuneus, or anterior cingulate during the antisaccade task. However, compared to adults, there was less activation in children and adolescents in the superior frontal eye fields, intraparietal sulcus, thalamus, cerebellum, and superior colliculus. The younger children showed less activation in the basal ganglia than adolescents or adults. However, children had greater activation in the supramarginal gyrus than adults. Surprisingly, the adolescents showed greater activation than either children or adults in the dorsolateral prefrontal cortex. Results were interpreted as indicating that age-related changes in antisaccade performance are “influenced by the maturation of integrated function among the neocortex, striatum, thalamus, and cerebellum,” possibly through synaptic pruning and myelination (p. 791). In a subsequent study, these researchers (Scherf, Sweeney, and Luna, 2006) compared participants drawn from those who took part in the previous study on a memory-guided saccade task. A complex set of neuroimaging results was obtained, as 22 regions were analyzed in each hemisphere, using three different methods of analysis. The authors interpreted the findings as indicating that there are both quantitative and qualitative changes with age in the neural substrates of visual-spatial working memory, possibly because of synaptic pruning and myelination. Finally, Klein and Feige (2005) examined age-related changes in the contingent negative variation (CNV) to the warning stimulus on pro- and antisaccade tasks in 7- to 18-year-olds. When the warning predicts the imperative stimulus with certainty, the CNV reflects contingency formation and response preparation. The topography of the CNV changed with age on both tasks, with a lateralposterior source in 7- to 11-year-olds, an anterior-central source in 17- to 18-year-olds, and transitional patterns in 12- to 16-year-olds. The authors concluded that “the cognitive functions supported by the anterior-central CNV generating structures are supported by different cortical regions (possibly located in inferior parietal lobe) in children” (p. 8). Pursuit In pursuit studies with healthy children, eye tracking has been used as a tool to make inferences about the development of the smooth-pursuit system and its

karatekin: eye tracking studies of normative and atypical development

269

Table 16.2 Saccadic eye movement studies in normative development Analysis of Age-Related Taska Changes Prosac (100 tr) Linear stepwise regression

Resultsb

Study Salman, Sharpe, Eizenmann, et al., 2006

Age 8–19

N 39

Fischer, Biscaldi, and Gezeck, 1997

8–70

281

Pro-o (200 tr), anti-g (200 tr) Task order not rep

Participants classifed into 10 age groups, increasing bin width with age Omnibus ANOVAs, followed by visual inspection

Prosac RT: ↓ until 15–20, ↑ after 30 Antisac RT: steep ↓ between 9 and 15, continued to ↓ until 25, ↑ after Antisac errors: ↓ until 20, ↑ after Frequency of express sac on pro: ns

Klein, 2001

6–28

199

Pro-g/o, anti-g/o (100 tr/condition) Tasks in counterbalanced order

Participants classified into 1year age groups (N = 1– 16/bin); multiple regression

Prosac RT: ↓ Antisac RT and errors: ↓ to a greater extent than prosac RTs Difference between anti and prosac RTs: ↓ Gap effect: ↓ Express sac: ns

Klein and Fischer, 2005a

Same as Klein, 2001

Munoz et al., 1998

5–79

168

Pro-g/o (120 tr), antig/o (240 tr) Prosac always followed by antisac

Participants classifed into 11 age groups increasing bin width with age (N = 8–28/bin) Omnibus Kruskal-Wallis, followed by visual inspection

Sac RTs: U-shaped curve, shortest in 18–22 5- to 8-year-olds: more sac RT variability, express sac in overlap, antisac errors, prosac hypometria; larger gap effect and difference between pro and antisac RTs Peak velocity: ns Sac duration: ns until age 60

Fukushima, Hatta, and Fukushima, 2000c

4–13

99

Prosac (60 tr)

Sac RT: plateaued by 12

20–38

22

Sample sizes ranged from 3 to 26 across ages T tests (uncorrected), correlations

7–10, 12 Adults 6–8 Adults

59 15 10 11

Antisac (60 tr)

11

674

Model fitting (along with heritability estimates)

17

616

Anti (20 tr; 17 tr for 167 participants) Pro always followed by anti

7–15

53

Pro (≤30 s), predictive (≤30 s), fixation (≤30 s) Tasks always in this order

Participants classified into four 2-year age groups ANOVAs, followed by posthoc tests Regression for predictive and fixation tasks to find best-fitting curves and age of maturation

Malone and Iacono, 2002

Ross Radant, Young, and Hommer, 1994

270

Sac RT: ↓ Gain and peak velocity: ns

Express prosac and express antisac errors: ns Antisac errors and prosac RTs: ↓

Pro and antisac with warning (0, 300, 600 or 1000 ms; 40 tr/ task) Task order not rep

methodological paradigms

Amplitude and peak velocity of sac: ns Antisac errors and RTs: ↓ Adults and children benefited similarly from warning, but children benefited less on both pro- and antisac Antisac errors ↓ Contributions of genetic and environmental influences similar between ages Sac RT on all tasks: ↓ linearly Premature sac on predictive task ↓ until 12 Sac during fixation: ↓ until 10

Study

Age

N

Taska

Table 16.2—continued Analysis of Age-Related Changes

Resultsb

Luna et al., 2004

8–30

245

Prosac (54 tr), antisac (36 tr), MGSs (1, 2, 4 or 8 s; 6 tr/delay) Tasks always in this order

Participants divided into 7 age groups. First 6 groups spanned 2 years each (N = 20–30/bin), 7th group spanned 6 years (N = 62) Multiple regression, model fitting

Inverse regression best fit for sac RTs, antisaccade errors, distance errors of MGS Antisac errors and spatial error of initial MGS mature at 14, sac RTs at 15, distance errors of final MGSs at 19 Peak velocity and gain of prosac: ns

Kramer, Gonzalez de Sather, and Cassavaugh, 2005

8–9 10–12 13–15 16–18 19–25

25 25 25 25 25

Prosac (60 tr), antisac (120 tr), oculomotor captured (180 tr) Pro- and antisac presented on a different day than capture Task order counterbalanced

ANOVA, followed up with Tukey’s HSD

Antisac errors: equally high in the 3 younger groups, lower and similar in older groups Oculomotor capture errors: ns Prosac errors: infrequent and ns Prosac RTs: ↓ until 13–15, ns after Antisac RTs: ↓ until 16–18, ns after

Luna et al., 2001

8–13 14–17 18–30

11 15 10

Pro (54 tr), anti (36 tr) Task order not rep Participants administered tasks in blocks on the day before scan, in trials alternating with fixation during scan; behavioral results recorded only on tasks administered prior to scan

Inverse curve fit

Antisac errors: ↓ Antisac RTs: ns Peak velocity, spatial accuracy, amplitude, and duration of antisac: ns

Scherf, Sweeney, and Luna, 2006

10–13 14–17 29.5 (10.6)

9 13 8

Prosac (40 tr in scanner), MGS (5 s, 40 tr in scanner) Participants administered slightly different versions of the task prior to and during scan; behavioral results recorded only on tasks administered prior to scan

ANOVA, inverse curve fit

Prosac RTs and spatial accuracy: ns MGS RTs: main effect of age, but pairwise comparisons ns Accuracy of initial MGS: children < adolescents children = adults Accuracy of final MGS: ns Inverse curve fit: accuracy of initial MGS: ↑ Accuracy of final MGS: ↑ Sac RT: ns

7–8 12 Pro-o (100 tr), anti-o ANOVA Pro- and antisac RTs: ↓ 10–11 11 (100 tr) (no age × task interaction) 13–14 11 Task order Antisac errors: ↓ after 10–11 15–16 12 counterbalanced 17–18 12 Notes: a. The numbers in parentheses refer to task parameters (number of trials, task duration, or warning or delay period). b. All results refer to analyses of age-related differences. c. On the antisaccade task, a subgroup of the 7- to 10-year-olds and 12-year-olds who took part in the antisaccade study were compared to a subgroup of the adults. In the study on the effects of warning signals, the participants were a subgroup of the participants in study 1. d. In the oculomotor capture task, small gray circles with the number 8 on them were presented on an imaginary circle for 1000 ms, after which all but one of the circles changed to red and the numbers changed to letters. The participants were instructed to report the orientation of the letter on the circle whose color had not changed. On most of the trials, a new red circle appeared simultaneously with the color change. This new circle did not have any informational value but was expected to “capture” attention because of its sudden onset. On the remaining trials, the additional circle remained on screen throughout the trial. An error was counted if the initial saccade was made to the distractor. Abbreviations: ↑, increased with age. ↓, decreased with age. Anti-g/o: Antisaccade task with gap and overlap conditions. Anti-o: Antisaccade task with overlap condition only. HSD, honestly significant difference. MGS, memory-guided saccade. Not rep, not reported. Ns, not significant. Pro-g/o: prosaccade task with gap and overlap conditions. Pro-o: prosaccade task with overlap condition only. RT, response time. Sac, saccade(s) or saccadic. Tr, trials. Klein and Feige, 2005

karatekin: eye tracking studies of normative and atypical development

271

B

methodological paradigms 0

0

0

29

34 ms

15

18% 10 20 30 40 50 60 70

% Errors

10 20 30 40 50 60 70

35 ms

30 40 50 60 70

27

AG-SRT (ms)

10 20

20 19 35 43 33

37 22

0

20

40

60

200

300

400

500

200

300

400

0

0

0

40

20

20

40

40

60

60

60

Antisaccade Task

20

80

80

80

0

10

20

30

40

50

60

70

80

90

275 250 225 200 175

525 500 475 450 425 400 375 350 325 300

125

225 200 175 150

425 400 375 350 325 300 275 250

Age

0

10

20

30

40

50

60

70

80

90

525 500 475 450 425 400 375 350 325 300 275 250 225 200 175

125

150

375 350 325 300 275 250 225 200 175

425 400

475 450

Age

336 312 288 264 240 216 192 168 144 120 96 72 336 312 288 264 240 216 192 168 144 120 96 72

336 312 288 264 240 216 192 168 144 120 96 72

SRT (ms)

Figure 16.5 Age-related changes in (A) prosaccade RTs, (B) antisaccade RTs, and (C) antisaccade errors in three studies. The first column presents data from Fisher, Biscaldi, and Gezeck (1997), the second from Munoz and colleagues (1998), and the third from Klein (2001). (The figure from Fischer, Biscaldi, and Gezeck, 1997, is reprinted with permission from Elsevier, and the figures from Munoz et al., 1998, and Klein, 2001, are used with kind permission of Springer Science and Business Media.)

0

20

40

60

150

200

250

300

150

200

250

500

Prosaccade Task SRT PG SRT AG

475 450

SRT PO SRT AO

300

PO-SRT (ms)

336 312 288 264 240 216 192 168 144 120 96 72

C

A

272 PDE AG

336 312 288 264 240 216 192 168 144 120 96 72

PDE AO

336 312 288 264 240 216 192 168 144 120 96 72

interactions with the saccadic system. The results of pursuit studies in children are summarized in table 16.3. As can be seen in this table, there are quite a few discrepancies across studies in terms of target velocity and whether it was constant or not, operational definitions of pursuit performance (particularly for intrusive and compensatory saccades), and the method of analyzing age-related differences. With these caveats in mind, the studies reviewed in this section suggest that there are either small or no differences in gain between young children (around age 7) and adults for slow targets, but that performance continues to improve through adolescence for faster targets. Results are inconsistent for intrusive and catch-up saccades. The results of a pursuit study on 8- to 19-year-olds are depicted in figure 16.6. Eye Movements during Scene and Face Perception Because there are no standard paradigms for assessing eye movements during scene or face perception, there is a great deal of diversity in the studies reviewed here and less of a sense of accumulating knowledge about a specific research question. All the developmental studies in this area have included 1–4 age groups, with small to medium sample sizes. Eye movements were used in two studies examining how children resolve linguistic ambiguities (Sekerina, Stromswold, and Hestvik, 2004; Trueswell et al., 1999). Most tasks used with adults in research on this topic involve reading and are not suitable for children. Therefore, eye movements were particularly useful in extending this research to younger ages. In both studies, the fixations of 4- to 7-year-olds and young adults on visual stimuli were compared as they were listening to spoken instructions regarding the stimuli. In both studies, the sentences contained ambiguities early on that could be resolved with the help of the contextual information provided by the stimuli. In both studies, the adults took into account the contextual information as they were listening to the sentence and revised their original interpretation when necessary, whereas the children appeared to commit themselves to only one interpretation at the beginning. The findings were interpreted as indicating that children rely on local linguistic information to resolve ambiguities, and have difficulty revising their initial interpretations and coordinating multiple sources of linguistic and contextual information as they process language. Two studies of eye movements in children were aimed at examining the role of shared storybook reading in emergent literacy (Evans and Saint-Aubin, 2005; Justice et al., 2005). In both studies, the eye movements of 4- to 6-year-olds were examined as they looked at storybooks with pictures while they were listening to adults read the stories. Contrary to common assumptions about shared reading having a specific effect on promoting children’s print knowledge, the results of both studies indicated that the children’s fixations fell

overwhelmingly on the pictures, regardless of the relative salience or layout of the pictures. Two developmental studies examined eye movements during face perception. In one study (Schwarzer, Huber, and Dümmler, 2005), two experiments were conducted to examine holistic and analytic modes of processing faces in 6- to 8-year-olds, 9- to 10-year-olds, and adults. There were few differences between ages for schematic pictures of faces, which were processed analytically. For realistic photographs of faces, however, there was a shift from a holistic to an analytic mode of processing with increasing age. Although adults made fewer fixations than children (consistent with holistic processing), there were no age-related differences in gaze time to different facial regions in either experiment. The development of facial expression perception was examined in another series of studies comparing 8-year-olds, 12-year-olds, and adults (Marcus, 2005). Although adults performed differently than children on behavioral measures (showing better accuracy for identifying emotions in inverted faces, sharper distinctions between emotional categories, and a stronger bias away from angry faces), there were few differences between the age groups in terms of eye movements to faces depicting different emotions. Finally, we used eye tracking to examine incidental and intentional spatial sequence learning on a serial reaction time task in 8- to 10-year-olds (35), 11- to 13-year-olds (N = 28), 14- to 17-year-olds (N = 13), and young adults (N = 24)] (Karatekin, Marcus, and White, 2007). Along with behavioral measures, we used oculomotor anticipations (looks to the target location prior to target onset) as an index of sequence learning. As in our previous study with adults (Marcus, Karatekin, and Markiewicz, 2006), participants spontaneously tried to anticipate the target location from the beginning of the task, regardless of age, and oculomotor anticipations and RTs showed learning effects similar to those in the manual modality. There were few age-related differences in incidental sequence learning indices in either the manual or oculomotor modalities, whereas all measures showed age-related differences for intentional learning. Results suggested that the search for regularities and the ability to rapidly learn a sequence incidentally are mature by ages 8–10. In contrast, the ability to learn a sequence intentionally, which requires cognitive resources and strategies, continues to develop through adolescence. Pupillary Dilation To my knowledge, there are only three studies in which pupillary dilation was used as an index of cognitive processes in children and adolescents. In a 1970 study, pupillary dilation was examined as a function of difficulty of mental arithmetic problems in 10 “educable retardates” (mean age = 10.6, SD = 0.11) and 10 controls (mean age = 11.0, SD = 0.6) (Boersma et al., 1970). As expected, dilation was greater for difficult than for easy

karatekin: eye tracking studies of normative and atypical development

273

274

methodological paradigms 24 8

10 10

38

5–7 Undergraduates

7–12 30–38

8–19

Langaas et al., 1998d

Accardo et al., 1995

Salman, Sharpe, Lillakas et al., 2006e

3–4 5 6 22–37

N 9 12 7 5

Agea

Study Haishi and Kokubun, 1995

0.25 (15.5°/s) (sinusoidal) 0.5 (31°/s) (sinusoidal)

0.2, 0.4, 0.8, 1.0, 1.2 Hz (cosinusoidal)

0.3 Hz/11.3°/s (sinusoidal)

20 cycles/freq

5 min total

Two 30-s cycles

Gain and phase

Velocity and position gain

Gain, saccades

Table 16.3 Smooth-pursuit eye movement studies in normative development Target Velocityb Duration Measures 0.3, 0.5, 0.7 Hz 15 s/frequency Power ratio (an (sinusoidal) estimate of how smooth the eye movements are) Phase difference

Correlation

Visual inspection

Gain: Mann-Wh U Saccades: ANOVA and Scheffe

Analysisc ANOVA

Gain: ∼1 in both groups, ns Sac (≥0.5°): children > adults Sac (<0.5°): ns Velocity gain: adults > children for all frequencies See text for more information Gain ↑ for both 0.25 and 0.5 Hz Gain reached adult level by midadolescence Phase: ns

Resultsc Power ratio: no effect and no interaction between age and frequency among the three child groups Phase: no effect of age within child groups, but interaction; eye movements lagged behind target in 3- to 4-year-olds, ahead of target in 5- and 6-year-olds

karatekin: eye tracking studies of normative and atypical development

275

11.4 (7–15)

Ross, Radant, and Hommer, 1993 53

62 39 36

6°/s, 12°/s (constant)

0.4 Hz (sinusoidal)

<30 s/frequency

30 s

Gain, saccades

Gain, RMSE, saccades

Correlation

MANOVA, chi-sq, post-hoc tests Levene test for homogeneity of variance

Compared to the older groups, the youngest group had lower gain, higher RMSE, more anticipatory (≥5°) sac, greater variability in gain and RMSE Catch-up sac: ns 6°/s target. Sac: ns 12°/s target. Intrusive sac: ns 12°/s target. gain ↑ and catch-up sac ↓ Initiation phase: ns M. gain for oppositedirection targets: ↑ M. gain for samedirection targets: ns

Ross, Radant, Young, and Hommer, 1994

11.5 (8–15)

51

9°/s (constant); target Initiation and Correlation, <30 s/task stepped to right or maintenance ANOVA left, then moved gain smoothly in the same or opposite direction 94 4 to 32°/s (constant) Foveofugal: 32 tr Maintenance gain Inverse regression Inverse regression Takarae et al., M = 19.3 (SD = Foveofugal and pure Pure: 40 tr models yielded results 2004g 11.3) step-ramp tasks; Oscillating: 22 s at for maintenance gain oscillating target each of 4 on all tasks task frequencies Notes: a. Numbers within parentheses refer to standard deviation or range, when available. b. Target velocity refers to peak target velocity in studies in which velocity was not constant. c. All analyses and results refer to those involving age-related differences. d. Participants were controls compared to children born prematurely or with developmental coordination disorder. e. The authors also examined vertical pursuit. Results showed a great deal of intersubject variability and weaker developmental changes than for horizontal pursuit. f. The younger two groups consisted of the firstborns of male twin pairs. Eleven of the adults were parents of the younger participants. g. Participants were controls compared to children with autism, reviewed in the section on pervasive developmental disorders. Abbreviations: ↑ = increased with age. ↓ = decreased with age. Chi-sq, chi-square. M, mean. M. gain, maintenance gain. Mann-Wh U, Mann-Whitney U. Ns = not significant. RMSE, root-mean-square error. Sac: saccade(s).

11–12 17–18 34–63

Katsanis, Iacono, and Harris, 1998f

Horizontal target motion at 0.5 Hz. r = 0.41, p = 0.012

1.1

1.1

1.0

1.0

0.9

0.9

Smooth Pursuit Gain

Smooth Pursuit Gain

Horizontal target motion at 0.25 Hz. r = 0.46, p = 0.004

0.8 0.7 0.6 0.5 0.4 0.3

0.7 0.6 0.5 0.4 0.3

0.2

0.2

0.1

0.1 8

A

0.8

10

12

14

16

18

8

20

Age

B

10

12

14

16

18

20

Age

Figure 16.6 Smooth pursuit gain as a function of age and target motion (Salman, Sharpe, Lillakas, et al., 2006). (Reprinted with kind permission of Springer Science and Business Media.)

problems, with dilation increasing during the response period of 20 s in controls. Although there were no group differences in dilation before the presentation of the problems or during the first part of the response period, the “retardate” group had smaller dilations than controls during the rest of the response period, particularly for the difficult questions. In addition, increases in dilation were related to accuracy in both groups. The authors attributed the results in the retarded group to a greater degree of attentional fluctuation. We used pupillary dilation in two studies of attention and working memory. In one study (Karatekin, 2004), top-down control over attention was investigated on a dual task in 10-year-olds (N = 15) and adults (N = 21). The tasks were an auditory digit span (with three sequence lengths) and a simple visual RT task. In four conditions, participants performed neither (no-task), one (digit span or RT only), or both tasks (dual). Dependent variables were digit-span accuracy, manual RT, and pupillary dilation to digits. The pupillary results are depicted in figure 16.7. At both ages, the slopes of the functions relating pupillary dilation to the presentation of the digits were flat in the no-task and RT-only conditions. In contrast, the slopes were linear and increasing in the digit-span-only and dual conditions, indicating that resource recruitment increased linearly with each increase in workingmemory load. In addition, slopes were shallower in the dual than in the digit-span-only condition, suggesting that some of the resources were diverted into the RT task. Finally, the slopes were shallower in children than in adults in both digit span only and dual. Thus, although the 10-year-olds allocated their attention between tasks and across increasing memory loads in a manner similar to that of adults, their

276

methodological paradigms

ability to recruit sufficient resources at higher loads was not yet fully mature. In a second study, we examined regulation of cognitive resources in 10-year-olds (N = 32) and young adults (N = 72) on spatial n-back tasks assessing sustained attention and spatial working memory (Karatekin, Marcus, and Couperus, 2007). Performance was assessed with behavioral measures (accuracy, RT) and pupillary dilation. Repeated administration of 0-back led to a decrease in pupillary dilation and increase in RT variability, revealing a subtle vigilance decrement. Effects of repeated administration of 0-back were similar between ages. Compared to adults, children’s sensitivity (d' ) and RTs were not disproportionately affected by 1-back. However, they showed a disproportionately higher response bias (c) and larger pupillary dilations to hits on 1-back, suggesting that they were not as effective as adults in extracting information about target frequency under the high-working-memory-load condition. Thus, on relatively simple tasks of sustained attention and working memory, 10-year-olds appeared to recruit resources in a manner similar to young adults.

Eye tracking in children and adolescents: atypical development In this section, eye tracking studies will be reviewed in three disorders in which a fair number of such studies have accumulated. There is a growing body of research on eye movements in learning disabilities (e.g., Desroches, Joanisse, and Robertson, 2006; Fischer, Hartnegg, and Mokler, 2000; Fukushima et al., 2005; Hutzler et al., 2006). To limit the scope of the chapter, this research will not be reviewed here.

Adults: 4 Digits

120% 115%

4-DS only 4-RT only 4-Dual

110% 105% 100%

115% 110% 105%

105%

120%

“digit “digit “digit “digit 1” 2” 3” 4” Children: 6 Digits 6-DS only 6-RT only 6-Dual

“go”

“digit “digit “digit “digit “digit “digit 1” 2” 3” 4” 5” 6”

“go”

115% 110% 105% 100%

“digit “digit “digit “digit “digit “digit 1” 2” 3” 4” 5” 6” 125%

“go”

Children: 8 Digits

125%

Adults: 8 Digits 8-DS only 8-RT only 8-Dual

Pupillary dilation

Pupillary dilation

110%

125%

Adults: 6 Digits 6-DS only 6-RT only 6-Dual

100%

115% 110% 105% 100%

A

115%

“go”

Pupillary dilation

Pupillary dilation

125%

120%

120%

4-DS only 4-RT only 4-Dual

100% “digit “digit “digit “digit 1” 2” 3” 4”

120%

Children: 4 Digits

125% Pupillary dilation

Pupillary dilation

125%

120%

8-DS only 8-RT only 8-Dual

115% 110% 105% 100%

“digit “digit “digit “digit “digit “digit “digit “digit 1” 2” 3” 4” 5” 6” 7” 8”

“go”

B

“digit “digit “digit “digit “digit “digit “digit “digit 1” 2” 3” 4” 5” 6” 7” 8”

“go”

Figure 16.7 Pupillary dilation (percent increase over no-task) to the auditory stimuli as a function of condition and sequence length in (A) adults versus (B) 10-year-olds. (Figure and figure caption reprinted from Karatekin, 2004, with permission from Elsevier.)

Eye tracking has been used in several other disorders in children and adolescents, including developmental coordination disorder and prematurity (Langaas et al., 1998), neurofibromatosis (Lasker, Denckla, and Zee, 2003), obsessive-compulsive disorder (Rosenberg et al., 1997), and depression and anxiety (Jazbec et al., 2005). Schizophrenia One of the most robust findings in the schizophrenia literature is an impairment of smooth-pursuit eye movements, not only in actively psychotic individuals but also in remitted patients and in unaffected relatives of individuals with schizophrenia. These findings indicate that pursuit impairments reflect a genetic vulnerability to the

disorder (for reviews see Broerse, Crawford, and den Boer, 2001; Holzman, 2000; Hutton and Kennard, 1998; Reuter and Kathmann, 2004; Trillenberg, Lencer, and Heide, 2004). Thus studies of youth with schizophrenia-spectrum disorders (who have a more severe and genetically loaded, but not qualitatively different, form of the disorder; for a review see Asarnow and Karatekin, 2000) and high-risk offspring of parents with schizophrenia have focused mostly on pursuit abnormalities. As can be seen in table 16.4, the results of pursuit studies in youth with schizophrenia-spectrum disorders or at risk for schizophrenia have shown consistently that schizophrenia is associated with reduced gain (Jacobsen et al., 1996; Kumra

karatekin: eye tracking studies of normative and atypical development

277

278

methodological paradigms

29 20 26 18

13 19

13 19 10 14 20 28 10 49 60 80

45 58 84

Kumra et al., 2001e

Ross et al., 1996

Ross et al., 1999

Ross et al., 2005

Ross, 2003

N 17 18 22

Study Jacobsen et al., 1996

Table 16.4 Eye tracking studies in children and adolescents with schizophrenia-spectrum disorders or at risk for schizophrenia M:F IQb Taskc Resultsd Groups Agea COS 14.5 (10–18) 59 : 41 85 (17) Pursuit (11°/s, constant, Gain: COS < ADHD = C ADHD-III-R 12.6 (9–15) 94 : 6 111 (17) 5 cycles) RMSE: COS > ADHD > C Control 13.5 (9–18) 73 : 27 116 (18) Anticipatory (≥4°) sacc: COS > ADHD = C Back-up sac: COS > ADHD = C Catch-up sac: ns COS 14.6 (2.5) 55 : 45 7.3 (3.7) Pursuit (17°/s, constant, Gain: COS and Psy NOS < C Control COS 14.6 (2.3) 55 : 45 11.2 (2.1) 10 cycles) RMSE: COS and Psy NOS > C Psy NOS 13.7 (3.5) 85 : 15 7.8 (3.1) Catch-up sac: COS > C Control NOS 13.1 (3.4) 72 : 28 15.8 (2.5) All anticipatory sac: ns Large (>4°) anticipatory sac: ns At risk 10.6 (6–15) 39 : 61 9.3 (2.6) Pursuit (12°/s, constant, Gain: At-risk < C Control 11.1 (6–15) 47 : 53 11.9 (2.8) not rep) RMSE: At-risk > C Small (∼2°) anticipatory sac: at-risk > C Catch-up sac: ns Anticipatory sac: COS, adult-onset sz, and At risk 10.6 (6–15) 39 : 61 9.3 (2.6) Pursuit (12°/s or 17°/s,f constant, 3 ming) both sets of parents > age-matched C. Control 11.1 (6–15) 47 : 53 11.9 (2.8) COS group had more anticipatory sac than COS 10.3 (7–15) 70 : 30 Not rep adult-onset sz and at-risk children. Adult-onset sz 21 (16–29) Not rep Not rep Bilineality (pursuit impairments in both Parents of COS (28–53) Not rep Not rep parents) greater in parents of COS than Parents of adult sz (40–81) Not rep Not rep in parents of adult-onset sz. Control adults (22–45) Not rep Not rep COS 10.4 (2.5) 71 : 29 Not rep Pursuit (17°/s, constant, Gain: COS < at-risk = C At risk 10.5 (2.4) 48 : 52 Not rep not rep) Large (>4°) anticipatory sac: COS > at-risk Controls 11.5 (2.5) 47 : 53 Not rep =C Small (1–4°) anticipatory sac COS > at-risk >C Catch-up sac: ns COS 10.4 (2.4) 73 : 27 Not rep MGSs (1 and 3 s, tr Premature MGSs: COS > at risk = C At risk 11.2 (3.1) 53 : 47 Not rep not rep) Spatial accuracy at 1 s: COS < at-risk Control 11.3 (2.7) 46 : 54 Not rep =C Spatial accuracy at 3 s: ns Sac RT: ns

karatekin: eye tracking studies of normative and atypical development

279

13 28 38

13 30 26

Karatekin and Asarnow, 1998

Karatekin and Asarnow, 1999 COS ADHD-III-R Control

COS ADHD-III-R Control

At risk Control

14.5 (3.3) 13.8 (3.1) 13.2 (2.4)

14.4 (3.3) 13.9 (3.1) 14.1 (2.7)

13.0 (2.6) 13.0 (2.7)

54 : 46 63 : 37 46 : 54

54 : 46 68 : 32 47 : 53

48 : 52 48 : 52

Visually guided saccades (targets stepped 10°– 60°; 109 tr)

Sac of at-risk group were hypometric, particularly for 40° to 60° targets. Frequency of hypometric sac ↓ in C but not in at-risk. 89 (17) Visually guided saccades RT of visually guided sac: ns 105 (16) (72 tr/task) Parallel search rate: ns 106 (17) Parallel and serial search Serial search rate: COS = ADHD > C (216 tr/task) RT of first sac in parallel search: ns RT of first sac in serial search: ADHD > COS = C 89 (17) Eye movements during COS looked at fewer relevant, but not more 106 (15) scene perception, irrelevant regions than C 105 (18) (5 pictures, 8 s/picture) COS stared more for global, but not for structured questions ADHD had shorter fixations than C for questions requiring detailed analysis deviation or range, when available. Age was used as a covariate in Jacobsen et al. (1996), Ross

116 (16) 117 (12)

Notes: a. Ages refer to mean ages; numbers within parentheses refer to standard et al. (1996), Ross (2003), and Ross et al. (2005). b. IQ was estimated from vocabulary subtest scores in Kumra et al. (2001), Ross et al. (1996), and Ross et al. (1999). c. Information in parentheses refers to task characteristics, including number and duration of trials. In pursuit studies, the information in parentheses refers to, in order, target speed, whether it was constant or not, and total duration or cycles for the task. d. All results refer to those involving group differences. e. Different control groups were used for the schizophrenia and psychosis NOS groups to provide matching on age and gender. f. The task was the same as in Ross et al. (1996); however, the parents of the two schizophrenia groups were presented with targets moving at 16.7° rather than 12°/s. g. Pursuit performance was recorded in 60-s intervals for the adults and 30-s intervals for the children in repeated trials until 3 min of usable data were obtained. Abbreviations: III-R, the children were diagnosed using the DMS-III-R, which did not specify ADHD subtypes. ADHD, attention-deficit/hyperactivity disorder. C, controls. COS, childhood-onset schizophrenia. MGS, memory-guided saccade. NOS, not otherwise specified. Not rep, not reported. Ns, not significant. Psy NOS, psychotic disorder not otherwise specified. RMSE, root-mean-square error. RT, response time. Sac, saccade(s) or saccadic. Sz, schizophrenia. Tr, trials.

21 21

Schreiber et al., 1997

30 degrees

et al., 2001; Ross, 2003; Ross et al., 1996, 1999), with one exception (Ross et al., 2003). Figures 16.8 and 16.9 depict data from two of these studies. These results have been taken to indicate support for continuity between the adult- and childhood-onset forms of schizophrenia and for the greater severity of childhood-onset schizophrenia compared to the adult-onset form. However, the operational definitions and findings related to intrusive and catch-up saccades are not as consistent. Two studies demonstrated impairments in memoryguided saccades in children with schizophrenia but not in those at risk (Ross et al., 2005) and hypometria for highamplitude targets in adolescent offspring of schizophrenic parents (Schreiber et al., 1997). Asarnow and I used eye tracking to examine visual-spatial attention during visual search (Karatekin and Asarnow, 1998) and scene perception (Karatekin and Asarnow, 1999) in childhood-onset schizophrenia. Taken together, these studies indicated that lower levels of visual attention (e.g., basic control of eye movements) were intact in schizophrenic children. In contrast, they appeared to have difficulty with top-down control of selective attention in the service of selfguided behavior.

For comparisons to normative development, it should be noted that all the pursuit studies reviewed in this section included targets moving at a constant velocity of 11 to 17°/s, whereas the velocity and the nature of the target’s motion differed considerably across normative studies. In addition, given the rarity of schizophrenia in children and the difficulties involved in recruiting at-risk children, most of these studies included participants across a wide age range, and their data were averaged in group comparisons. Pervasive Developmental Disorders (PDD) Table 16.5 summarizes eye tracking studies in PDD. Because profound social difficulties are among the defining characteristics of PDD, several studies of PDD focused on fixation patterns to socially relevant stimuli. There are discrepancies in the results of these studies. In two studies that likely included the same participants, there was no difference between controls and children with PDD in fixation patterns to cartoonlike drawings of scenes (Van der Geest, Kemner, Camfferman, et al., 2002b) or to upright photographs of faces (Van der Geest, Kemner, Verbaten, et al., 2002). However, the controls spent less time looking at inverted than upright faces, whereas the fixation durations of the PDD group did

BL AS BL BL

BL

AS AS AS

8.0 seconds

30 degrees

8.0 seconds

BS AS

AS

A

2.0 seconds

B

2.0 seconds

Figure 16.8 (A) Eye-tracking record from a 15-year-old medication-free subject with childhood onset schizophrenia. Top, 16-sec segment. Bottom, enlargement of the first descending ramp. (B) Eye tracking record from a 10-year-old subject with ADHD. Top, 16-sec segment. Bottom, enlargement of the first descending

280

8.0 seconds

methodological paradigms

C

2.0 seconds

ramp. (C) Eye-tracking record from a 15-year-old normal subject. Top, 16-sec segment. Bottom, enlargement of the first descending ramp. AS, anticipatory saccade; BS, back-up saccade; BL, blink artifact. (Figure and figure caption reprinted from Jacobsen et al., 1996, with permission from the Society of Biological Psychiatry.)

40

Percentage of Total Eye Movements Due to Saccades

35

30

25

20

15

10

5

0

Normal children

Childhood-onset schizophrenia

Nonpsychotic children of schizophrenic parents

Normal adults

Adult-onset schizophrenia

Figure 16.9 Percentage of total eye movements due to anticipatory saccades in normal children, nonpsychotic children of schizophrenic parents, childhood-onset schizophrenic probands, adult-onset schizophrenic probands, and normal adults. Values above the dotted line at 2.5 percent are abnormal. Childhood-onset

probands who are currently adults are in open circles. Eye movements recorded during a constant velocity 12°/sec task. (Figure and figure caption reprinted from Ross et al., 1999, with permission from John Wiley & Sons, Inc.)

not vary as a function of orientation. These results were interpreted as indicating that the abnormal gaze behavior in autism in everyday life may be due to the demands of social interactions, and that autistic children may not be processing faces in a holistic fashion. However, in two other investigations, children with PDD looked less at the eyes than the controls when viewing photographs of faces (Dalton et al., 2005) and video clips of scenes depicting social interactions (Klin et al., 2002). In addition, while viewing the video clips, the autistic group fixated more than controls on the mouths, bodies, and objects. Measures of social competence were positively correlated with fixations on the mouth (suggesting a focusing of attention on speech rather than the social cues from the eyes) and negatively correlated with fixations on objects in the PDD group. In one of these studies (Dalton et al., 2005), participants also underwent functional magnetic resonance imaging while viewing faces. Compared to controls, the autistic group showed greater activation in the amygdala in response

to facial stimuli and greater activation in both the amygdala and orbitofrontal gyrus for emotional faces. In addition, duration of time fixating the eyes was correlated with amygdala activation in the autistic group. The authors suggested that hyperactivation in neural circuits mediating emotions causes “negatively valenced hyperarousal” and heightened sensitivity to social stimuli in autism, which leads to reduced fixations on eyes, which helps to reduce overarousal. Studies of saccades on gap and overlap tasks have also yielded conflicting results, with two studies pointing to a weak attentional engagement system (Kemner et al., 1998; van der Geest et al., 2002b), one study suggesting difficulties with disengaging attention from fixation (Landry and Bryson, 2004), and one study showing no impairment in disengaging attention (Goldberg et al., 2002). There were also conflicting findings regarding prosaccade RTs, with elevated RTs in the PDD group in one study (Goldberg et al., 2002), but not in others (Minshew, Luna, and Sweeney, 1999; Rosenhall,

karatekin: eye tracking studies of normative and atypical development

281

Table 16.5 Eye tracking studies in autism M:F IQ Agea 10.6 (2.1) Not rep 93 (17) 9.9 (1.5) 97 (10)

Study Van der Geest et al., 2002b

N 16 14

Groups PDD (10 HFA) Control

Van der Geest et al., 2002b

17 17

10 HFA, 7 NOS Controls

10.6 (2.1) 10.1 (1.3)

94 : 6 94 : 6

95 (15) 98 (11)

Dalton et al., 2005

11 12 16 16

Autism or Asp Control Autism or Asp Control

15.9 17.1 14.5 14.5

(4.7) (2.8) (4.6) (4.6)

100 : 0 100 : 0 100 : 0 100 : 0

94 (19) Not assessed 92 (28) 123 (13)

Klin et al., 2002

15 15

HFA Controls

15.4 (7.2) 17.9 (5.6)

100 : 0 100 : 0

101 (25) 103 (20)

Kemner et al., 1998

10 10 10 10

Autism ADHD-III Dyslexia Control

10.3 (1.3) 8.9 (1.6) 10.0 (1.4) 10.7 (1.3)

80 : 20 100 : 0 100 : 0 80 : 20

72 94 96 98

Van der Geest et al., 2001

16 15

10.9 (2.2) 10.3 (1.4)

100 : 0 100 : 0

98 (16) 97 (10)

Landry and Bryson, 2004

15 13 13

10 HFA, 6 NOS Control 13 aut, 2 asp Down syndrome Control

5.6 (3.8–7.6) 5.5 (3.5–8.0) 3.6 (2.1–6.2)e

Not rep

70 (29) 65 (15) 110 (22)

282

methodological paradigms

(15) (10) (7) (9)

Taskb Scan drawings of 25 scenes (10 s/scene)

Scan 16 faces depicting emotions Scan 12 upright, 12 inverted faces (10 s/face) Emotion discrimination from 40 faces (3 s/face) Facial recognition (10 familiar, 10 not) Face/scene perception during 5 video clips (30–60 s/ clip) Visual oddball task (80% frequent, 10% infrequent, 10% novel stimuli; 1 s/ stimulus; 140 tr)

Pro-g/o (60 tr/ cond), order counterbalanced Shift (10 tr) and disengage (10 tr), presented in mixed order

Resultsc,d Duration and number of fixations: ns Average and total scanpath length: ns Time to look at human figure: ns Total duration and number of fixations on human figure: ns Fixations to different regions for upright: ns Fixation duration for upright > inverted in C but not in PDD Accuracy: PDD < C Duration of fixations on eyes: PDD < C Fixations on other regions: ns

Proportion of time on eyes: PDD < C Proportion of time on mouths, bodies, objects: PDD > C Sac to frequent stimuli: PDD > ADHD, C Sac during 2 intertrial intervals: PDD > ADHD, C C: more sac to novel than to other stimuli; no difference in PDD or ADHD Dyslexia: fewer sac to infrequent than to frequent stimuli Sac RT in gap and overlap: ns Gap effect: PDD < C Sac RT on Shift: ns Sac RT on Disengage: PDD > C

Groups

Agea

Table 16.5—continued M:F IQ

Taskb

Resultsc,d

Study

N

Goldberg et al., 2002

11 11

HFA Control

13.8 (1.5) 14.4 (1.5)

73 : 27 73 : 27

99 (11) 113 (14)

Predictive (90 tr), anti (42 tr) MGSs (1.5–3 s, 40 tr) Pro-g/n/o, 25 tr/cond) Tasks always presented in that order

Express sac on gap: PDD < C Antisac errors: PDD > C Premature MGSs: PDD >C Predictive sac: PDD < C Sac RT on gap/null/ overlap and MGSs: PDD > C Gap effect: ns RT, amplitude, and peak velocity of predictive sac.: ns RT, velocity, and spatial accuracy of antisac: ns Velocity and spatial accuracy of MGSs: ns

Minshew, Luna, and Sweeney, 1999

26 26

HFA Control

20.2 (8.5) 20.0 (8.7)

96 : 4 96 : 4

105 (13) 101 (18)

Nowinski et al., 2005

52 52

HFA Control

17 (8–46) 18 (8–45)

92 : 8 92 : 8

106 (13) 109 (12)

Prosac (54 tr), antisac (36 tr), MGSs (1, 2, 4, 8 s; 6 tr/delay) Tasks always presented in that order Active fixation (15–30 s)

Rosenhall, Johansson, and Gillberg, 1988

11

55 : 45

60–100

26

8 autism 13 (9–16) 3 “autistic-like” Control 10 (7–13)

Antisac errors: PDD > C Premature MGSs: PDD >C Spatial accuracy of MGSs: PDD < C Peak velocity, duration, RT of all sac: ns Rate of intrusive sac: ns Intrusive sac in PDD had larger amplitudes and shorter latency to return to fixation than in C Sac RT: ns Pursuit: failure to complete task

Scharre and Creedon, 1992 Takarae et al., 2004

34

Autism

94 : 6

Ave to sev. retarded

60 94

HFA Control

88 : 12 84 : 16

102 (16) 108 (13)

Median 7.5 (2–11) 20.1 (11.2) 19.3 (11.3)

Prosac (90 tr) Pursuit (10, 20, 30 or 40°/s; 12 or more tr/velocity) Order not rep Pursuit (follow a cube moved by experimenter) Pursuit (4–32°/s) 3 tasks (32 tr, 40 tr, 22 s; constant) Task order not rep

Pursuit: only 15% could perform the task well

Initiation gain for targets moving into right visual field: PDD < C; left visual field: ns Maintenance gain on all tasks: PDD < C Notes: a. Ages refer to mean ages, numbers within parentheses refer to standard deviation or range, when available. b. Information in parentheses refers to task characteristics, including number and duration of trials. In pursuit studies, the information in parentheses refers to, in order, target speed, whether it was constant or not, and total duration or cycles for the task. c. All results refer to those involving group differences. d. All participants are referred to as PDD in the results column to ease comparison across studies. e. In Landry and Bryson (2004), the control group was significantly younger than the other two groups. However, age was not entered into the analyses. In all other studies, age was reported to not differ significantly between groups, and none of these studies used age as a covariate in the analyses. Abbreviations: ADHD-III: attention-deficit/hyperactivity disorder, diagnosed based on DSM-III. Asp, Asperger’s syndrome. Ave. to sev. retarded, average to severely retarded. C, control. HFA, high-functioning autism. MGS, memory-guided saccade. Ns, not significant. NOS, pervasive developmental disorder not otherwise specified. Not rep, not reported. PDD, pervasive developmental disorder. Pro-g/n/o, prosaccades, with gap, null, and overlap conditions. RT, response time. Sac, saccade(s) or saccadic. Tr, trials.

karatekin: eye tracking studies of normative and atypical development

283

Johansson, and Gillberg, 1988; van der Geest, Kemner, Verbaten, et al., 2001). A more consistent finding is that children and adolescents with PDD have impairments in internally guided saccades and that they make more antisaccade errors and premature saccades than controls (Goldberg et al., 2002; Minshew, Luna, and Sweeney, 1999). Duration and peak velocity of saccades appear to be intact in PDD (Goldberg et al., 2002; Minshew, Luna, and Sweeney, 1999). Studies have also consistently shown impaired smooth pursuit (Rosenhall, Johansson, and Gillberg, 1988; Scharre and Creedon, 1992; Takarae et al., 2004). It is important to note that in most of these studies (1) the samples were quite small, (2) average IQ in the PDD group was in the normal range, whereas the majority of individuals with PDD have low IQs, (3) the overwhelming majority of the participants were male, and (4) as in the schizophrenia studies, the age ranges of the samples were quite large. Attention-Deficit/Hyperactivity Disorder (ADHD) Difficulties with inhibition and “executive functions” are among the central features of ADHD. Thus many of the eye tracking studies in children and adolescents with ADHD have compared the integrity of internally versus externally guided saccades, especially antisaccades. As can be seen in table 16.6, the most consistent finding is that individuals with ADHD make more premature saccades on a number of different tasks and more errors on the antisaccade tasks, reflecting difficulties with inhibition. Peak velocity of saccades was also found to be reduced in two out of three studies. There are inconsistencies across studies in terms of saccadic RTs. Although most of these studies used traditional saccadic paradigms, one investigation (Cairney et al., 2001) focused on a task assessing ability to use contextual information to modulate oculomotor responses. Results were interpreted as indicating that the ADHD group “had difficulty inhibiting saccades only when they were required to use context to increase the level of tonic inhibition within their saccadic system. Thus the presence of inhibitory deficits in ADHD depends on the context in which the individual’s current behavioral goals are set” (p. 516). In a recent study (Karatekin, 2006), I examined the effects of task manipulations on improving antisaccade accuracy and RTs of adolescents with ADHD, age-matched controls, 10-year-olds, and young adults. Order effects were tested by administering the task at the beginning and end of the session. Other manipulations involved a visual landmark to reduce demands on working memory and internal generation of saccades, spatially specific and nonspecific cues at three intervals, and central engagement of attention through perceptual and cognitive means at three intervals. As expected, the ADHD group was impaired in terms of accu-

284

methodological paradigms

racy and saccadic RT on the first administration of the task. Although their accuracy improved with most of the manipulations, it did not improve disproportionately compared to controls. Nevertheless, with most of the manipulations, members of this group achieved the same level of accuracy as unaided controls on the first administration. In contrast, their saccadic RTs came close to normal under several conditions, indicating that their elevated antisaccade RTs may have been related to attention. Thus cognitive scaffolds can ameliorate at least some of the inhibition deficits in adolescents with ADHD. Pursuit was tested in three studies of ADHD (one study by Jacobsen et al., 1996, is summarized in the schizophrenia section). Two studies showed reduced RMSE (Castellanos et al., 2000; Jacobsen et al., 1996), and one study that included an attentional manipulation designed to enhance performance showed no difference between control and ADHD groups (Bylsma and Pivik, 1989). No impairments were observed in gain in the two studies in which it was measured. Two studies with relatively large samples investigated group differences in developmental trends between ADHD and control participants. In Munoz and colleagues (2003), performance on both antisaccade and prolonged-fixation tasks improved through adolescence in controls, reaching adult levels only at age 16 (see figure 16.10). The developmental trajectory of the ADHD group lagged behind that of controls for both tasks and reached an asymptote later. In another study (Klein, Raschke, and Brandenbusch, 2003), both pro- and antisaccade RTs declined with age in controls, although the decline in antisaccade RTs was steeper. In the ADHD group, however, the functions relating age to RT had similar slopes on the two tasks. Furthermore, although the frequency of premature responses decreased with age in both groups, the rate of this reduction was slower in the ADHD group. As can be seen in table 16.6, results of studies on medication effects are inconsistent, probably as a result of selection bias, order effects, and small samples. Other factors that should be noted in the studies in this section are that (1) as in studies of PDDs, most participants were male and (2) studies differed in the extent to which participants had different subtypes of ADHD and comorbid conditions. These subtypes and comorbid conditions were not always reported, and their effects were usually not analyzed.

Review of studies of eye tracking in normative and atypical development The main goals of the saccade and pursuit studies in typical development have been to chart the developmental trajectories of different indices of performance and to examine the effects of several factors (especially gap and overlap condi-

karatekin: eye tracking studies of normative and atypical development

285

32 20

Castellanos et al., 2000

8.8 (6–13) 9.6 (1.7)

(10–12) (10–12) 11.2 (10–12)

At risk + ADHD At risk Control

12 55 12

Habeych et al., 2006g

ADHD-C Control

11.7 (7.8–14.3) 10.8 (8.4–14.6) 10.6 (8.1–12.6)

TS + ADHD-III-R TS Control

Mostofsky, Lasker, Singer, et al., 2001

11 14 10

20 21

Rothlind, Posner, and Schaughency, 1991

11.3 (7–15) 11.3 (7–15)

10.5 (6.9–13.9) 9.9 (2.8)

ADHD Control

46 46

Klein, Raschke, and Brandenbusch, 2003f

10.3 10.7 31.0 34.1

Agea (0.3) (0.3) (1.4) (1.1)

ADHD-III-R Control

Groups ADHD Control ADHD Control

N 76 75 38 105

Study Munoz et al., 2003e

Not rep Not rep

111 (12)

67 : 33

0 : 100 0 : 100

Not rep

Not rep Not rep Not rep

99 (10) 111 (13)

102 (11) 106 (16)

Not rep

100 : 0 100 : 0 100 : 0

100 : 0 100 : 0

89 : 11 83 : 17

Pursuit (17°/s, constant 5 cycles) Go/no-go (12 tr/condition) MGSs (1.2 s, 13 tr) Tasks always presented in this order

Anti-g/n/o (99 tr/cond) Task order not rep

Pro- and antisac (60 tr/condition), MGSs (4.5–5 s; 60 tr) Task order not rep

Pro-null/o, anti-null/o (10 tr/condition)

Pro-g/o, anti-g/o (100 tr/condition) Order counterbalanced

Table 16.6 Eye tracking studies in ADHD IQ (or M:F est’d IQ) Taskb 80 : 20 Not rep Pro-g/o (80–120 tr), anti-g/o 53 : 47 Not rep (160–240 tr), active fixation 47 : 53 Not rep (300–450 tr) 44 : 56 Not rep Pro- and antisac administered in first session, fixation in second Prosac always followed by anti

Table continued

Prosac RT: TS, TS + ADHD > C Prosac RT variability: TS + ADHD > TS, C Antisac errors and premature MGSs: TS + ADHD > TS Antisac errors on gap and null: at risk + ADHD > at-risk Antisac RTs in null: at risk + ADHD > at risk Antisac RTs in gap and overlap: ns Peak velocity of antisac in null: At risk + ADHD > at-risk Peak velocity of antisac in gap and overlap: ns Frequency of MGSs: ADHD < C Premature MGSs: ADHD > C Commission errors to no-go stimuli: ADHD > C Intrusive saccades in go/no-go: ADHD > C

Prosac and antisac RTs: ADHD > C Premature sac on all tasks: ADHD > C Antisac errors: ADHD > C Corrective sac on antisac task: ADHD
Resultsc,d Prosac: RT, CV of RT, duration: ADHD > C Prosac RT: peak velocity: ADHD < C Gap effect: ns Prosac: proportion of express saccades: ns Antisac: errors, RT, CV of RT: ADHD >C Active fixation: intrusive sac: ADHD > C

286

methodological paradigms 24 29 26 18 13 15 15 15

Gould et al., 2001

12.1 (1.2) 12.1 (1.2)

ADHDi Control ADHD ADHD Control

22 22 8 11 25

Mostofsky, Lasker, Cutting, et al., 2001j

10.8 (7.2–17.9)

11.2 (1.3) 11.5 (1.0)

ADHD-III-R Control

(12–18) (11–19) (9–11) (18–20)

13 10

14.3 15.0 10.3 19.2

Ross, Hommer, Breiger, et al., 1994 Aman, Roberts, and Pennington, 1998

ADHD-C Control Control Control

10.0 (2.0) 8.8 (1.6) 10.3 (1.5) 9.4 (1.7) 8.2 (1.7) 8.9 (1.9) 8.5 (6–11) 25.5 (21–38)

Agea

10 15 15 18

ADHD-C ADHD-C Control Control ADHD-C ADHD-C Control Control

Groups

Karatekin, 2006

Cairney et al., 2001

N

Study

52 : 48

100 : 0 100 : 0

100 : 0 50 : 50

80 : 20 60 : 20 47 : 53 11 : 89

100 : 0 0 : 100 100 : 0 0 : 100 77 : 23 93 : 7 87 : 13 40 : 60

Not rep Not rep Not rep

111 (8) 110 (7)

107 (12) 116 (16)

12 (2)h 13 (2) 14 (3) 12 (2)

Not rep Not rep Not rep Not rep 99 (15) 96 (12) 101 (12) Not rep

Table 16.6—continued IQ (or M:F est’d IQ)

Modified prosac (42 tr) and antisac (42 tr) Pro always followed by anti Prosac (60 tr), antisac (60 tr), MGSs (4.5–5 s, 60 tr) Task order not rep

MGSs (800 ms, 31.5 s)

Pro- and antisac, modified antisac tasks (32 tr/task) Anti always followed by pro; order of antisac tasks counterbalanced

Contextual modulation of sac (200 sac tr, 200 catch tr)

Active fixation (21 s)

Taskb

Antisac errors: ADHD (med = unmed) >C Premature MGS: ADHD (med = unmed) > C Prosac RTs: ns CV of prosac RTs and RT of MGSs: ADHD (unmed) > ADHD (med) = C

FOE for high-probability targets: ns FOE for low-probability targets: ADHD < C Premature and inappropriate sac: ADHD > C Prosac RT: ns Antisac errors: ADHD > age-matched and younger C Antisac RT on 1st administration: ADHD > age-matched C Antisac RT on 2nd administration: ns Premature sac: ADHD > age-matched and younger C Corrective sac: ADHD < age-matched and younger C Premature MGS: ADHD > C Sac RTs: ns Spatial accuracy of MGS: ns Prosac errors: ns Antisac errors: ns

Large (> 4°) sac: ADHD > C Floor effects, poor test-retest reliability

Resultsc,d Spatial accuracy of MGSs: ADHD < C (trend) Anticipatory and catch-up sac in pursuit: ns Maintenance gain in pursuit: ns RMSE in pursuit: ADHD > C (p = .09)

karatekin: eye tracking studies of normative and atypical development

287

10 12 10

O’Driscoll et al., 2005 ADHD-C ADHD-I Control

ADHD ADHD Control

8 11 25

Mostofsky, Lasker, Cutting, et al., 2001j

12.4 (0.6) 12.7 (0.6) 12.7 (0.6)

10.8 (7.2–17.9)

12.1 (1.2) 12.1 (1.2)

ADHDi Control

22 22

Aman, Roberts, and Pennington, 1998

9.6 (1.7) 9.5 (1.7)

12.4 (0.6) 12.7 (0.6) 12.7 (0.6)

11.2 (1.3) 11.5 (1.0)

13 10

Effects of Medications Ross, Hommer, Breiger, et al., 1994

ADHD-III-R Control

ADHD-C ADHD-I Control

ADHD-III-R Control

20 20

10 12 10

Bylsma and Pivik, 1989

O’Driscoll et al., 2005

100 : 0 100 : 0 100 : 0

52 : 48

100 : 0 100 : 0

100 : 0 50 : 50

85 : 15 55 : 45

100 : 0 100 : 0 100 : 0

108 (9) 109 (11) 110 (11)

Not rep Not rep Not rep

111 (8) 110 (7)

107 (12) 116 (16)

Not rep Not rep

108 (9) 109 (11) 110 (11)

Prosac (48 tr), antisac (48 tr) predictive (tr not rep) task switch (48 tr) Order counterbalanced Double-blind cross-over trial, ADHD tested once at baseline, then on placebo and methylphenidate, in counterbalanced order, with at least 3 weeks in between Controls tested once

MGSs (800 ms, 31.5 s) ADHD tested on placebo and methylphenidate, in random order, with 1 week in between Controls also tested twice Modified prosac (42 tr) and antisac (42 tr) Pro always followed by anti ADHD tested on psychostimulant med first, off med second Controls also tested twice Intersession interval: 1 week Prosac (60 tr), antisac (60 tr), MGSs (4.5–5 s, 60 tr) Task order not rep Children were not assigned randomly to medication

Pursuit (.45 Hz or 29°/s, sinusoidal, 15–20 oscillations; attentional manipulation)

Prosac (48 tr), antisac (48 tr) predictive (tr not rep) task switch (48 tr) Order counterbalanced

Table continued

No difference between med and unmed ADHD in prosac RTs, antisac errors and premature MGSs CV of prosac RTs and RT of MGSs: med ADHD > unmed ADHD Methylphenidate led to faster proand antisac RTs, more predictive sac, fewer antisac errors, fewer task-switching errors

Correct prosac: close to perfect in both groups in both sessions Antisac errors: decreased in controls from 1st to 2nd session; no difference in ADHD from 1st (med) to 2nd (unmed) session

No effect of methylphenidate on MGSs

RT, amplitude, peak velocity of prosac: ns Predictive saccades for predictable direction: ns Predictive saccades for predictable direction and timing: ADHD-C < ADHD-I = C Antisac errors: ADHD-C > C; ADHDI between ADHD-C and C and did not differ significantly from either Antisac RTs: ns Task switching: ns Velocity arrest scores: ADHD > C RMSE: ns

288

methodological paradigms ADHD-III-R Control

ADHD-C

Groups

9.6 (1.7) 9.5 (1.7)

12.6 (10–15)

Agea

85 : 15 55 : 45

100 : 0

Not rep Not rep

Not rep

Pro-g/o, anti-g/o (tr and order reported elsewhere) Children tested on and off methylphenidate in counterbalanced order, with 1 week in between

Taskb

Methylphenidate led to shorter proand antisac RTs, fewer antisac errors, more corrective antisac, shorter RTs for corrective sac, more express sac Interaction between medication and order effects: When tested on placebo first and med 2nd, performance improved on most measures When tested on med first, little difference between sessions Only two measures showed a main effect of medication but no medication × order interaction: children made more express prosac and corrected their antisac errors more frequently when on medication No effect of methylphenidate on velocity arrest scores or RMSE Velocity arrests: Med ADHD = C

Resultsc,d

Pursuit (.45 Hz or 29°/s, sinusoidal, 15–20 oscillations; attentional manipulation) Controls tested once, ADHD tested twice, on and off methylphenidate, with 1 week in between Notes: a. Ages refer to mean ages, numbers within parentheses refer to standard deviation or range, when available. Age was used as a covariate in Mostofsky, Lasker, Cutting, et al. (2001) and in Klein, Raschke, and Brandenbusch (2003). b. Information in parentheses refers to task characteristics, including number and duration of trials. In pursuit studies, the information in parentheses refers to, in order, target speed, whether it was constant or not, and total duration or cycles for the task. c. All results refer to those involving group differences. d. Studies examining the effects of medications are listed twice. In the first section, results refer to the unmedicated state to facilitate comparisons across studies. In the second section, details of the design relevant to medications are added, and only the medication effects are listed under results. e. The control group overlapped with that used in Munoz et al. (1998). f. All the controls in the study were included in Klein (2001). g. Children were at risk for substance use disorders by virtue of having a father with a lifetime diagnosis of alcohol abuse or dependence. ADHD was diagnosed based on questionnaires filled out by the mother and child. The ages, gender ratios, and IQs of the at-risk children with and without ADHD were not reported separately. Average age for the at-risk group was 11.0 years (range, 10–12), M : F gender ratio was 55 : 45, and average IQ was 108 (SD = 17). h. IQ was estimated from vocabulary subtest scores. i. Seventy-three percent of the sample had the combined subtype, 18% had the inattentive subtype, and 9% had the hyperactive/impulsive subtype. j. The ages, gender ratios, and subtypes of the medicated and unmedicated children were not reported separately. Mean age of the whole ADHD group was 11.3 years (range, 7.1– 16.1), M : F ratio was 58 : 42, 32% were diagnosed with the combined subtype, and 42% with the inattentive subtype. Twenty-six percent of the sample could not be assigned a subtype because of discrepancies between parent and teacher reports. Abbreviations: ADHD, attention-deficit/hyperactivity disorder, subtype not specified. ADHD-III-R, attention-deficit/hyperactivity disorder, diagnosis based on DSM-III-R, which did not specify subtypes. Anti-g/n/o, antisaccade task with gap, null (no gap), and overlap conditions. C, control. CV, coefficient of variation. FOE, fixation offset effect. Med, medicated or medications. MGS, memory-guided saccade. Not rep, not reported. Ns, not significant. RT, response time. Sac, saccade(s) or saccadic. Tr, trials. TS, Tourette’s syndrome. Unmed, unmedicated.

20 20

27

Klein et al., 2002

Bylsma and Pivik, 1989

N

Study

Table 16.6—continued IQ (or M:F est’d IQ)

Control ADHD

Direction Errors (%)

80

60

40

20

10

A

20

30

40

50

1.5 Intrusive Saccade Rate (saccades/s)

100

1.0

0.5

10

60

Age (years)

Control ADHD

B

20

30

40

50

60

Age (years)

Figure 16.10 (A) Antisaccade errors and (B) intrusive saccades during fixation as a function of age in individuals with ADHD and healthy controls. (From Munoz et al., 2003, reprinted with permission of the American Physiological Society.)

tions in saccades and target velocity in pursuit) on children versus adults. Studies on atypical development have focused on delineating impairments in disorders and making inferences about the neural bases of these impairments. At a global level, there are some impressive consistencies across these studies. These replications across different labs and samples clearly demonstrate the reliability of the findings. In normative development, the consistent findings include the observations that the dynamics of prosaccades (peak velocity, duration) do not change substantially after age 4, that prosaccades and antisaccades have different developmental trajectories, and that pursuit gain continues to improve through adolescence for fast targets. In atypical development, the most consistent findings are pursuit impairments in schizophrenia, elevated antisaccade error rates in ADHD and PDDs, and premature saccades on a variety of tasks in schizophrenia, ADHD, and PDDs. At a more detailed level, however, there are some discrepancies across studies. For instance, different studies have come to different conclusions about the precise developmental trajectories of pro- and antisaccades and pursuit gain. There are also discrepancies in terms of the intrusive and compensatory saccades in pursuit, express saccades and the gap effect, and spatial accuracy of memory-guided saccades. Some of these discrepancies are due to idiosyncratic factors that are not of general concern. Some of the reasons, however, may be relevant to the field of developmental cognitive neuroscience in general. First, charting “the” developmental trajectory of performance on a task may not be a feasible goal. As with many other tasks (e.g., Kagan, 2003), performance on eye tracking tasks is a function of not only task difficulty, but also of con-

textual factors, including the task parameters and state variables such as anxiety and fatigue. For example, introducing a 200-ms gap or overlap between fixation and stimulus on a prosaccade task substantially changes the nature of the resulting saccades. As a result, different developmental trajectories are obtained for the gap versus the overlap versions of a prosaccade task (Klein, 2001). Different trajectories are also obtained in pursuit studies depending on the target velocity (Accardo et al., 1995) and in face perception depending on whether the stimuli are schematic drawings or photographs (Schwarzer, Huber, and Dümmler, 2005). The precise wording of the instructions (“follow the lights” versus “move your eyes in time with the lights”) affects the characteristics of predictive saccades in adults (Isotalo, Lasker, and Zee, 2005), and the saliency of task instructions affects manifestation of antisaccade impairments in clinical populations of adults (Nieuwenhuis et al., 2004). The characteristics and neural bases of prosaccades differ depending on whether they are presented in blocks or intermixed with antisaccades (Cornelissen et al., 2002). In a large sample of “healthy” adults, levels of anxiety and depression as measured on a checklist affect antisaccade performance (Smyrnis et al., 2003). Like other cognitive tasks, performance on eye tracking tasks is also tied to temporal factors unfolding over short periods of time. For instance, studies in adults show timeon-task (Smyrnis et al., 2002) and practice effects (Dyckman and McDowell, 2005) on antisaccades. In ADHD, elevated antisaccade RTs come down to normal levels when the task is readministered (Karatekin, 2006). It is important to note that a factor that varies quite a bit across the studies that we have reviewed is task length.

karatekin: eye tracking studies of normative and atypical development

289

Estimating the effect of task length on performance is not easy. Of course, an inadequate number of trials reduces the reliability of the measure. In the studies reviewed, only Malone and Iacono (2002) have addressed this issue statistically. In addition, when a task includes relatively few trials, it is likely that it is tapping not only the specific construct it is purported to tap, but also the ability of the participants to adapt to a novel situation. Shorter tasks are also less likely to induce fatigue. Longer tasks, however, assess participants’ abilities to improve their performance with practice (learning to learn), which may differ depending on age and clinical status. For instance, in their study of a 100-trial antisaccade task in ADHD, Klein, Raschke, and Brandenbusch (2003) conducted post hoc analyses to examine the effect of task length. They report that differences in antisaccade errors between the ADHD and control groups were larger for the whole task and for the second 50 trials than for the first 50 trials. Whereas the error rate declined slightly in the control group from the first to the second half, it increased slightly in the ADHD group. Furthermore, this increase in error rate was more pronounced in the younger than in the older ADHD participants. Thus the effect of task length may differ depending on the nature of the task as well as the age and clinical status of the participants. It is, therefore, disconcerting to see that task length does not differ randomly across these studies—in general (and for good reason), tasks tend to be shorter in studies that include younger children or clinical populations of children. It is also disconcerting that order effects were not controlled for and that task order was not even reported in a number of the studies reviewed in this chapter. So it is clear that what is being measured on these eye tracking tasks cannot be viewed as unitary, static constructs that exist independently of context and time. Not only do contextual and temporal factors affect performance, but they also interact with age and clinical status. Thus it seems like an illusory goal to try to chart “the” developmental trajectory of a “pure” process and to pinpoint the age at which performance reaches maturity. Instead, it might be more informative to incorporate contextual and dynamic factors into the definition of the constructs and the design of the studies and to collect more empirical evidence on the factors that facilitate or hinder performance as a function of age and clinical status (cf. Karatekin, 2006; Klein, 2001; Klein, Raschke, and Brandenbusch, 2003; Nieuwenhis et al., 2004). This approach could also make it easier to compare results across studies and allow for stronger inferences. Another reason for the discrepancies across studies has to do with the method of analyzing age-related changes in normative studies and the method of addressing developmental issues in clinical studies. For the same data set, it is possible to conclude that performance matures earlier if an

290

methodological paradigms

ANOVA is used or that it continues to develop through that age range if a correlational or regression approach is used and whether researchers correct for multiple comparisons or not. This discrepancy can be clearly seen in the study by Scherf, Sweeney, and Luna (2006) in which both an ANOVA and inverse-curve fit were used to analyze saccadic data, and different conclusions emerged concerning age-related changes from the different statistical methods. There are also inconsistencies in clinical studies regarding methods of addressing age differences across groups. First, the width of the age range varies substantially across studies, as can be seen in tables 16.3 to 16.6. Second, when the sample sizes are small, age differences do not reach significance. Thus the groups are reported to not differ in age. The developmental studies, however, indicate that depending on the age range and task, seemingly small age differences can, in fact, affect the results. For instance, Klein, Raschke, and Brandenbusch (2003) note that in a previous study (Rothlind et al., 1991) in which an ADHD sample was found to make more antisaccade errors than controls, “patients were 8 months younger than controls (125 months old). In our sample of 199 participants aged 6–28 years, we found that this age difference alone can explain [antisaccade] errors” (p. 26). Third, some studies use age as a covariate when comparing clinical to control groups. This approach assumes that age and the dependent variable are linearly related in both groups and that the strength of the relationship does not differ across groups. However, in only two of the studies reviewed (Klein, Raschke, and Brandenbusch, 2003; Ross et al., 2005) were these assumptions explicitly tested. As with most other measures, the shape of the developmental trajectory and the extent of differences between clinical and control groups depend on the psychometric characteristics of the measures, including their variability, reliability, discriminating power, and whether they are subject to ceiling or floor effects (e.g., Knight and Silverstein, 2001; Meier and Perrig, 2000; Miller et al., 1995). Given two measures purported to assess the same construct, the measure with greater reliability and discriminating power is more likely to show age-related or clinical differences. The problem becomes thornier when different age or clinical groups are compared on tasks with different psychometric properties assessing different constructs (e.g., antisaccades and memory-guided saccades). In these cases, psychometric properties are confounded with what is being measured, and interpretations of differential developmental trajectories and differential deficit become suspect. The reliability of the measures also affects the extent to which they are intercorrelated, complicating interpretations of factor analyses and correlational and regression analyses to examine the interdependence of cognitive processes.

Thus, to interpret normative and clinical data, it is important to be aware of the psychometric properties of these measures and to test if they differ as a function of age or clinical status. Reliability of eye tracking measures was examined in several of the studies reviewed. Internal consistency (measured with Cronbach’s alpha) of antisaccade errors on a 20-trial task in a large sample of healthy children was found to be .81 for 11-year-olds and .82 for 17-year-olds (Malone and Iacono, 2002). Intraclass correlations (ICCs) for split-half reliability of RMSE in a pursuit task in adolescents with schizophrenia-spectrum disorders and controls ranged from .75 to .96 (Kumra et al., 2001). Test-retest reliability (measured with ICCs) over 3–6 weeks in a sample of 22 children with ADHD was .79 for antisaccade errors and .62 for predictive saccades for which both the direction and timing of the target were predictable (O’Driscoll et al., 2005). However, test-retest reliability of intrusive saccades during active fixation (measured through Pearson r) was only .16 over 3–9 weeks in 23 girls with ADHD, which probably reflects the floor effects on this measure (Gould et al., 2001). In the largest developmental study of reliability of saccadic measures in children, internal consistency and split-half reliability of pro- and antisaccades were computed with the Pearson r in 327 healthy 9- to 88-year-olds (Klein and Fischer, 2005b). With age partialed out, instrumental reliabilities were high for pro- and antisaccade RTs (.91–.96 for odd-even reliability, .81–.90 for split-half reliability) and antisaccade errors (.95 for odd-even, .83 for split-half). Testretest reliability over 19 months was computed only for 6- to 18-year-olds. Reliability estimates were moderately high for pro- and antisaccade RTs (.65–.66), but low for antisaccade errors (.43), perhaps because of individual differences in the rate of maturation of the neural substrates of antisaccades. The authors further noted that the instrumental (withinsession) and test-retest reliability estimates were not affected by age. These issues regarding the method of data analysis, lack of statistical power, and psychometric properties of the tasks matter for making inferences about the integrity and maturation of brain-behavior relations. These inferences are based not only on the positive findings, but also on the total pattern of positive and negative findings considered jointly. If enough confidence cannot be placed in the negative findings, one cannot interpret the positive findings with much confidence, either. In addition to the reasons for discrepancies across studies, it is important to note that all of the developmental studies reviewed in this paper used a cross-sectional approach. The risks of drawing inferences about development from crosssectional data are well documented (e.g., Kramer et al., 2000; Schneider et al., 2004; Siegler, 1998). For instance, the data we have reviewed point to the conclusion that

development of pro- and antisaccades occurs in a gradual, nonlinear manner through middle childhood and adolescence, whereas a different picture could well emerge in a longitudinal study. The problems associated with making inferences about development from cross-sectional data are compounded in clinical samples, where age can easily be confounded with other variables, such as severity, gender, comorbidity, confidence in the diagnosis, and duration of pharmacological treatment. Some of these problems could be overcome if researchers provided more information on these variables when comparing different age groups. Otherwise, caution needs to be exercised when making inferences about differences in developmental trajectories between clinical and control groups. Finally, a weakness in some of the normative and clinical studies reviewed is that insufficient consideration has been given to the role of attention in performance. There is a close relationship between saccades and visual-spatial attention, although the precise nature of the relationship between orienting of visual-spatial attention and programming of saccades is still being debated (e.g., Doré-Mazars, Pouget, and Beauvillain, 2004; Godjin and Theeuwes, 2003; Juan, Shorter-Jacobi, and Schall, 2004). On the one hand, this relationship provides an advantage for researchers interested in attention because eye movements provide an excellent means of investigating attention. On the other hand, it also means that eye movement data cannot be interpreted without taking attentional factors into account. Although this observation may be true for most cognitive tasks, it is especially true for eye movement tasks. An eye movement to a stimulus implies overt orienting of visualspatial attention to that location (e.g., Hoffman and Subramaniam, 1995; Peterson, Kramer, and Irwin, 2004; but also see Mokler and Fischer, 1999), and regions involved in the control of eye movements, such as the frontal eye fields, lateral intraparietal area, and superior colliculus, are also involved in covert visual-spatial attention (e.g., Moore and Fallah, 2001; Müller, Philiastides, and Newsome, 2004; Murthy, Thompson, and Schall, 2001; Powell and Goldberg, 2000; Wardak et al., 2006). Indeed, some researchers have questioned the meaningfulness of the distinction between visual-spatial attention and visualspatial working memory (e.g., Awh and Jonides, 2001; Medendorp, Goltz, and Vilis, 2006). Thus, for instance, it is likely that the improvements seen in prosaccade RTs through at least adolescence are in part related to attentional factors, and not just improvements in general processing speed or oculomotor programming (Hainline, 1988; Ross and Ross, 1983). Similarly, immaturity or impairments on tasks assessing antisaccades, predictive saccades, or memory-guided saccades cannot be interpreted solely in terms of inhibition, anticipation, or working memory,

karatekin: eye tracking studies of normative and atypical development

291

respectively, without considering the role of visual-spatial attention in performance.

Limitations of eye tracking as a tool One limitation of eye tracking is the extent to which inferences can be drawn from the oculomotor system about other motor systems. The neural bases of eye movements on laboratory tasks are established at a higher level of detail than the neural bases of other kinds of movements (Schall, Hanes, and Taylor, 2000). In addition, oculomotor and skeletomotor processes are hypothesized to be organized in parallel frontostriatal loops (Alexander, DeLong, and Strick, 1986). Furthermore, certain experimental manipulations have similar effects on oculomotor and skeletomotor systems (Schall et al., 2000), and eye and hand movements share a common reference frame when reaching for a target (Scherberger, Goodale, and Andersen, 2003). Thus it is tempting to use eye movements as a simple model of motor control. However, other evidence suggests that the two systems may not have a parallel organization. For example, pointing away and looking away from visual targets recruit overlapping but separable regions and may be accomplished through somewhat different mechanisms (Connolly et al., 2005). In addition, phenomena observed in the manual modality are not always replicated in the oculomotor modality (e.g., Jas´kowski and Sobieralska, 2004; Pratt, Shen, and Adam, 2004). Therefore, it is important to be careful about making generalizations across systems. Second, eye movements in the lab may not necessarily behave like eye movements in the real world. Many saccade tasks require a single saccade to a single stimulus. In a naturalistic context, however, we make sequences of saccades in more complex environments. The characteristics and neural bases of sequences of saccades (e.g., Caspi, Beutter, and Eckstein, 2004; Shima and Tanji, 1998; Van Loon, Hooge, and Van der Berg, 2002) or in perceptually complex environments (e.g., Deubel and Frank, 1991; Ilg et al., 2006; Schiller and Kendall, 2004) differ from those elicited during typical laboratory tasks. The characteristics of monkeys’ eye movements during free-viewing visual search also differ from fixations during traditional search tasks, in which practice trials are provided and accurate performance is rewarded (Shen and Paré, 2006). Naturalistic contexts also tax to a greater extent the decision-making processes involved in where and when to look and elicit more proactive and anticipatory eye movements (Land and Furneaux, 1997). Brainbehavior relationships may also differ between typical laboratory tasks and more naturalistic settings (Ipata et al., 2006). More direct comparisons between laboratory and naturalistic contexts, perhaps using head-mounted eye monitors, can shed more light on the degree to which results obtained in the lab are applicable to the real world.

292

methodological paradigms

Third, because even the simplest prosaccades and smoothpursuit eye movements are mediated by distributed neural networks, eye tracking by itself is of limited use when the researchers’ goal is to make inferences about specific brain regions in typically developing children or in psychiatric disorders that do not involve focal lesions. Finally, there are limitations in making inferences about the sources of developmental and clinical differences. Especially when only one or two outcome variables are used to measure performance, it cannot be assumed that the task is being performed in the same way—with the same cognitive and neural substrates—across ages or clinical groups. Many of the tasks used in eye tracking, though usually simpler than more traditional tasks, still involve a multiplicity of cognitive processes. Therefore, it is possible that the lower level of performance observed in children compared to adults is due to one reason at one stage of development and another reason at another stage. In fact, in the only developmental study of eye tracking that used electrophysiological measures, Klein and Feige (2005) demonstrated that 7- to 11– year-olds were recruiting different regions than adults prior to target onset on the antisaccade task, and that the neural substrates shifted with age. Similarly, in clinical populations, it cannot be assumed without evidence that the lower performance evidenced by the clinical groups is attributable to the main construct that is purported to be assessed by the task.

Conclusions regarding potentials of eye tracking as a tool Despite these limitations, eye tracking measures have much to offer to developmentalists. One important benefit of eye tracking measures is that they open up alternative ways of examining development. They are easily amenable to a process-oriented approach, through examination of the dynamics of saccades, the course of pursuit performance, the shapes of pupillary waveforms, and the evolving pattern of fixations during scene and face perception. By emphasizing the active, goal-directed nature of eye movements, eye tracking studies also highlight the fact that perception involves control over action. By emphasizing the dynamic nature of resource recruitment as a function of task demands, organismic priorities, and motivational factors, pupillary dilation studies highlight the top-down, active, and flexible nature of attentional control (e.g., Meyer and Kieras, 1997), as opposed to a boxology approach in which attention is viewed as more of a static construct (e.g., Baddeley, 1996). In addition, many of the tasks used in eye tracking studies are relatively simple and require no reading skills or complex motor skills. Thus eye tracking measures provide a powerful means to compare performance across a wide range of ages and clinical groups without these confounding factors.

The fundamental characteristics of basic eye tracking measures (saccades, pursuit, fixations during scene perception, pupillary dilation) are well delineated, there is consensus on the operational definitions of the key measures, and the values of most of these measures can reasonably be expected to fall within a relatively narrow range. In addition, there is a large body of excellent research on saccades and smooth-pursuit eye movements in adults and nonhuman primates. This research has detailed which kinds of factors affect which aspects of eye movements and how they are related to specific cognitive and motor processes. This work provides a strong foundation for further research. So far, however, it has been vastly underutilized in studies with children and adolescents. Much more work can be done examining age-related and clinical differences in the effects of contextual and temporal factors on eye movements as well as processes such as visual-spatial attention and its guidance by bottom-up and top-down factors (e.g., Gagnon et al., 2002; Iba and Sawaguchi, 2003; Shen and Paré, 2006; Thompson, Bichot, and Sato, 2005), and motor control (e.g., Gold and Shadlen, 2000; Sinha, Brown, and Carpenter, 2006; Vaziri, Diedrichsen, and Shadmehr, 2006). There is also, of course, a very large literature on saccade tasks assessing inhibition (e.g., Klein, 2001) and working memory (e.g., Ross et al., 2005). The neural bases of different types of eye movements have been examined in many studies with human adults and nonhuman primates, but only four studies so far have used electrophysiological (Klein and Feige, 2005) or brainimaging methods (Dalton et al., 2005; Luna et al., 2001; Scherf, Sweeney, and Luna, 2006) to examine the maturation or integrity of the neural bases of eye movements in children and adolescents. Thus the potentials of eye tracking measures to probe the maturation and integrity of the oculomotor circuitry in these populations have barely been tapped. In studies with specific a priori hypotheses and careful attention to contextual and temporal factors, eye tracking can be very useful in investigating this circuitry. In addition, despite the vast amount of research using singlecell recording and lesion methods to probe neural substrates of eye movements in nonhuman primates, very little research has been conducted on young animals. Extensions of primate research on saccades to developing animals can be invaluable in addressing questions of interest to developmental psychologists. There is a smaller but equally interesting body of research in human adults and nonhuman primates on eye movements during face and scene perception. Eye movements can be especially useful for addressing questions about implicit cognition (e.g., Dragoi and Sur, 2006; Tseng and Li, 2004) and problem solving (e.g., Grant and Spivey, 2003; Hodgson et al., 2002).

Pupillary dilation provides a direct psychophysiological measure of resource recruitment and effort that can be time locked to stimuli and responses in an exquisitely sensitive manner. Thus, in research on normative development, pupillary dilation can play a key role in providing an empirical foundation for resource theories of development (e.g., Case, 1991; Pascual-Leone, 2000; Swanson, 1999). For the most part, the tasks in the studies reviewed above have been taken from the cognitive science or neuroscience literatures and used to investigate cognitive processes in children and adolescents. However, eye tracking is a tool that can also be used effectively in the service of elucidating emotional processes and social information processing. The studies on eye movements to faces and social stimuli in typically developing and autistic samples are excellent examples of this approach. One could imagine many more studies examining eye movements during face and scene perception that could yield valuable information regarding socioemotional development and the influence of different factors on how typical and atypical populations of children process social information. There are a growing number of eye movement studies in adults in cognition, emotion, and motivation (e.g., Platt, 2002; Polli et al., 2005; Roesch and Olson, 2005), including the effects of rewards on saccades and the neural bases of these effects (e.g., Campos et al., 2005; Hikosaka, Nakamura, and Nakahara, 2006; Hodgson et al., 2000). In a recent study extending this line of research to adolescents, investigators showed that monetary rewards and punishments had greater effects on antisaccade parameters in adolescents than in adults (Jazbec et al., 2006). Pupillary dilation measures also provide an excellent means of testing the effects of rewards and difficulty on effort (Steinhauer and Hakerem, 1992) and examining processing of emotionally laden stimuli (Bitsios, Szabadi, and Bradshaw, 2004; Partala and Surakka, 2003). Finally, eye tracking measures can be useful in different kinds of translational research with typical and atypical populations. As reviewed earlier, for instance, eye tracking measures have been used to examine the effects of medications in ADHD. They can also be used to examine the effects of cognitive and psychosocial interventions in educational and clinical settings. For instance, eye movements were used to examine the effects of monetary rewards and punishments on antisaccades in adolescents with depression or anxiety (Jazbec et al., 2005), and how adults with spider phobia process pictures that include spiders (Rinck and Becker, 2006). Pupillary dilations were used to demonstrate that depression in adults is associated with ruminatory tendencies for negative or personally relevant stimuli (Siegle, Steinhauer, Carter, et al., 2003). This research can easily be extended to examine the effects of interventions. In a study on dyslexia, where abnormalities in eye movements

karatekin: eye tracking studies of normative and atypical development

293

themselves are believed to contribute to reading difficulties, children with dyslexia were trained to exert greater control over internally guided eye movements (Fischer and Hartnegg, 2000). Finally, advances in technology have made it feasible to use head-mounted cameras to determine where participants are fixating in a scene. It is especially easy to imagine studies where this technology could be used to translate research from the lab to the real world (e.g., Hayhoe and Ballard, 2005). In conclusion, studies of eye tracking in normative and atypical populations of children and adolescents have already yielded reliable and intriguing findings regarding a wide variety of cognitive processes and their neural substrates. Eye tracking offers researchers many more possibilities for studying cognitive and socioemotional development. REFERENCES Accardo, A. P., S. Pensiero, S. Da Pozzo, and P. Perissutti, 1995. Characteristics of horizontal smooth pursuit eye movements to sinusoidal stimulation in children of primary school age. Vis. Res. 35:539–548. Alexander, G. E., M. R. Delong, and P. L. Strick, 1986. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. 9:357–381. Aman, C. J., R. J. Roberts, and B. F. Pennington, 1998. A neuropsychological examination of the underlying deficit in attention deficit hyperactivity disorder: Frontal lobe versus right parietal lobe theories. Dev. Psychol. 34:956–969. Asarnow, R. F., and C. Karatekin, 2000. Children with schizophrenia: A neurobehavioral approach. In H. E. Remschmidt, ed., Schizophrenia in Children and Adolescents, 135–167. Cambridge, UK: Cambridge University Press. Avila, M. T., E. Hong, A. Moates, K. A. Turano, and G. K. Thaker, 2006. Role of anticipation in schizophrenia-related pursuit initiation deficits. J. Neurophysiol. 95:593–601. Awh, E., and J. Jonides, 2001. Overlapping mechanisms of attention and spatial working memory. Trends Cogn. Sci. 5:119–126. Baddeley, A., 1996. The fractionation of working memory. Proc. Natl. Acad. Sci. USA 93:13468–13472. Beatty, J., 1982. Task-evoked pupillary responses, processing load, and the structure of processing resources. Psychol. Bull. 91:276– 292. Beatty, J., and B. Lucero-Wagoner, 2000. The pupillary system. In J. T. Cacioppo, L. G. Tassinary, and G. G. Berntson, eds., Handbook of Psychophysiology, 2nd ed., 142–162. New York: Cambridge University Press. Bitsios, P., E. Szabadi, and C. M. Bradshaw, 2004. The fearinhibited light reflex: Importance of the anticipation of an aversive event. Int. J. Psychophysiol. 52:87–95. Boersma, F., K. Wilton, R. Barham, and W. Muir, 1970. Effects of arithmetic problem difficulty on pupillary dilation in normals and educable retardates. J. Exp. Child Psychol. 9:142–155. Broerse, A., T. J. Crawford, and J. A. den Boer, 2001. Parsing cognition in schizophrenia using saccadic eye movements: A selective overview. Neuropsychologia 39:742–756. Buswell, G. T., 1935. How People Look at Pictures. Chicago: University of Chicago Press. Bylsma, F. W., and R. T. Pivik, 1989. The effects of background illumination and stimulant medication on smooth pursuit eye

294

methodological paradigms

movements of hyperactive children. J. Abnorm. Child Psychol. 17:73–90. Cairney, S., P. Maruff, A. Vance, R. Barnett, E. Luk, and J. Currie, 2001. Contextual abnormalities in saccadic inhibition in children with attention deficit hyperactivity disorder. Exp. Brain Res. 141:507–518. Campos, M., B. Breznen, K. Bernheim, and R. A. Andersen, 2005. Supplementary motor area encodes reward expectancy in eyemovement tasks. J. Neurophysiol. 94:1325–1335. Carpenter, R. H. S., 1988. Movements of the Eyes. London: Pion. Case, R., 1991. The Mind’s Staircase: Exploring the Conceptual Underpinnings of Children’s Thought and Knowledge. Hillsdale, NJ: Erlbaum. Caspi, A., B. R. Beutter, and M. P. Eckstein, 2004. The time course of visual information accrual guiding eye movement decisions. Proc. Natl. Acad. Sci. USA 101:13086–13090. Castellanos, F. X., F. F. Marvasti, J. L. Ducharme, J. M. Walter, M. E. Israel, A. Krain, C. Pavlovsky, and D. W. Hommer, 2000. Executive function oculomotor tasks in girls with ADHD. J. Am. Acad. Child Adolesc. Psychiatry 39:644–650. Clementz, B. A., W. G. Iacono, and W. M. Grove, 1996. The construct validity of root-mean-square error for quantifying smooth-pursuit eye tracking abnormalities in schizophrenia. Biol. Psychiatry 39:448–450. Connolly, J. D., M. A. Goodale, H. C. Goltz, and D. P. Munoz, 2005. fMRI activation in the human frontal eye field is correlated with saccadic reaction time. J. Neurophysiol. 94:605–611. Cornelissen, F. W., H. Kimmig, M. Schira, R. M. Rutschmann, R. P. Maguire, A. Broerse, J. A. Den Boer, and M. W. Greenlee, 2002. Event-related fMRI responses in the human frontal eye fields in a randomized pro- and antisaccade task. Exp. Brain Res. 145:270–274. Dalton, K. M., B. M. Nacewicz, T. Johnstone, H. S. Schaefer, M. A. Gernsbacher, H. H. Goldsmith, A. L. Alexander, and R. J. Davidson, 2005. Gaze fixation and the neural circuitry of face processing in autism. Nature Neurosci. 8:519–526. Descroches, A. S. M. F. Joanisse, and E. K. Robertson, 2006. Specific phonological impairments in dyslexia revealed by eyetracking. Cognition 100:B32–B42. Deubel, H., and H. Frank, 1991. The latency of saccadic eye movements to texture-defined stimuli. In R. Schmid and D. Zambarbieri, eds., Oculomotor Control and Cognitive Processes, 369– 384. New York: Elsevier Science. Doré-Mazars, K., P. Pouget, and C. Beauvillain, 2004. Attentional selection during preparation of eye movements. Psychol. Res. 69:67–76. Dragoi, V., and M. Sur, 2006. Image structure at the center of gaze during free viewing. J. Cogn. Neurosci. 18:737–748. Dyckman, K. A., and J. E. McDowell, 2005. Behavioral plasticity of antisaccade performance following daily practice. Exp. Brain Res. 162:63–69. Evans, M. A., and J. Saint-Aubin, 2005. What children are looking at during shared storybook reading: Evidence from eye movement monitoring. Psychol. Sci. 16:913–920. Everling, S., and B. Fischer, 1998. The antisaccade: A review of basic research and clinical studies. Neuropsychologia 36:885– 899. Fischer, B., M. Biscaldi, and S. Gezeck, 1997. On the development of voluntary and reflexive components in human saccade generation. Brain Res. 754:285–297. Fischer, B., and K. Hartnegg, 2000. Effects of visual training on saccade control in dyslexia. Perception 29:531–542. Fischer, B., K. Hartnegg, and A. Mokler, 2000. Dynamic visual perception of dyslexic children. Perception 29:523–530.

Fischer, B., and H. Weber, 1993. Express saccades and visual attention. Behav. Brain Sci. 16:553–610. Fukushima, K., 2003. Frontal cortical control of smooth-pursuit. Curr. Opin. Neurobiol. 13:647–654. Fukushima, J., T. Hatta, and K. Fukushima, 2000. Development of voluntary control of saccadic eye movements. I. Age-related changes in normal children. Brain Dev. 22:173–180. Fukushima, J., S. Tanaka, J. D. Williams, and K. Fukushima, 2005. Voluntary control of saccadic and smooth-pursuit eye movements in children with learning disorders. Brain Dev. 27: 579–588. Gagnon, D., G. A. O’Driscoll, M. Petrides, and G. B. Pike, 2002. The effect of spatial and temporal information on saccades and neural activity in oculomotor structures. Brain 125: 123–139. Godjin, R., and J. Theeuwes, 2003. Parallel allocation of attention prior to the execution of saccade sequences. J. Exp. Psychol. [Hum. Percept.] 29:882–896. Gold, J. I., and M. N. Shadlen, 2000. Representation of a perceptual decision in developing oculomotor commands. Nature 404:390–394. Goldberg, M. C., A. G. Lasker, D. S. Zee, E. Garth, A. Tien, and R. J. Landa, 2002. Deficits in the initiation of eye movements in the absence of a visual target in adolescents with high functioning autism. Neuropsychologia 40:2039–2049. Gould, T. D., T. M. Bastain, M. E. Israel, D. W. Hommer, and F. X. Castellanos, 2001. Altered performance on an ocular fixation task in attention-deficit/hyperactivity disorder. Biol. Psychiatry 50:633–635. Granholm, E., S. K. Morris, A. J. Sarkin, R. F. Asarnow, and D. V. Jeste, 1997. Pupillary responses index overload of working memory resources in schizophrenia. J. Abnorm. Psychol. 106:458– 467. Granholm, E., and S. P. Verney, 2004. Pupillary responses and attentional allocation problems on the backward masking task in schizophrenia. Int. J. Psychophysiol. 52:37–51. Grant, E. R., and M. J. Spivey, 2003. Eye movements and problem solving: Guiding attention guides thought. Psychol. Sci. 14:462– 466. Habeych, M. E., M. M. Folan, B. Luna, and R. E. Tarter, 2006. Impaired oculomotor response inhibition in children of alcoholics: The role of attention deficit hyperactivity disorder. Drug Alcohol Depend. 82:11–17. Hainline, L., 1988. Normal lifespan developmental changes in saccadic and pursuit eye movements. In C. W. Johnston and F. J. Pirozzolo, eds., Neuropsychology of Eye Movements, 31–64. Hillsdale, NJ; LEA. Haishi, K., and M. Kokubun, 1995. Developmental trends in pursuit eye movements among preschool children. Percept. Mot. Skills 81:1131–1137. Haith, M. M., 2004. Progress and standardization in eye movement work with human infants. Infancy 6:257–265. Hayhoe, M., and D. Ballard, 2005. Eye movements in natural behavior. Trends Cogn. Sci. 9:188–194. Henderson, J. M., 2003. Human gaze control during real-world scene perception. Trends Cogn. Sci. 7:498–504. Henderson, J. M., and A. Hollingworth, 1999. High-level scene perception. Annu. Rev. Psychol. 50:243–271. Hikosaka, O., K. Nakamura, and H. Nakahara, 2006. Basal ganglia orient eyes to reward. J. Neurophysiol. 95:567–584. Hikosaka, O., Y. Takikawa, and R. Kawagoe, 2000. Role of the basal ganglia in the control of purposive saccadic eye movements. Physiol. Rev. 80:953–978.

Hodgson, T. L., C. Golding, D. Molyva, C. R. Rosenthal, and C. Kennard, 2000. Eye movements during task switching: Reflexive, symbolic, and affective contributions to response selection. J. Cogn. Neurosc. 16:318–330. Hodgson, T. L., B. Tiesman, A. M. Owen, and C. Kennard, 2002. Abnormal gaze strategies during problem solving in Parkinson’s disease. Neuropsychologia 40:411–422. Hoffman, J. E., and B. Subramaniam, 1995. The role of visual attention in saccadic eye movements. Percept. Psychophys. 57:787– 795. Holzman, P. S., 2000. Eye movements and the search for the essence of schizophrenia. Brain Res. Brain Res. Rev. 31:350– 356. Hutton, S., and C. Kennard, 1998. Oculomotor abnormalities in schizophrenia: A critical review. Neurology 50:604–609. Hutzler, F., M. Kronbichler, A. M. Jacobs, and H. Wimmer, 2006. Perhaps correlational but not causal: No effect of dyslexic readers’ magnocellular system on their eye movements during reading. Neuropsychologia 44:637–648. Iba, M., and T. Sawaguchi, 2003. Involvement of the dorsolateral prefrontal cortex of monkeys in visuospatial target selection. J. Neurophysiol. 89:587–599. Ilg, U. J., Y. Jin, S. Schumann, and U. Schwarz, 2006. Preparation and execution of saccades: The problem of limited capacity of computational resources. Exp. Brain Res. 171:7–15. Ipata, A. E., A. L. Gee, M. E. Goldberg, and J. W. Bisley, 2006. Activity in the lateral intraparietal area predicts the goal and latency of saccades in a free-viewing visual search task. J. Neurosci. 26:3656–3661. Isotalo, E., A. G. Lasker, and D. S. Zee, 2005. Cognitive influences on predictive saccadic tracking. Exp. Brain Res. 165:461– 469. Jacobsen, L. K., W. L. Hong, D. W. Hommer, S. D. Hamburger, F. X. Castellanos, J. A. Frazier, J. N. Giedd, C. T. Gordon, B. I. Karp, K. McKenna, and J. L. Rapoport, 1996. Smooth pursuit eye movements in childhood-onset schizophrenia: Comparison with attention-deficit hyperactivity disorder and normal controls. Biol. Psychiatry 40:1144–1154. Jas´kowski, P., and K. Sobieralska, 2004. Effect of stimulus intensity on manual and saccadic reaction time. Percept. Psychophys. 66:535–544. Jazbec, S., M. G. Hardin, E. Schroth, E. McClure, D. S. Pine, and M. Ernst, 2006. Age-related influence of contingencies on a saccade task. Exp. Brain Res. 174:754–762. Jazbec, S., E. McClure, M. Hardin, D. S. Pine, and M. Ernst, 2005. Cognitive control under contingencies in anxious and depressed adolescents: An antisaccade task. Biol. Psychiatry 58: 632–639. Juan, C., S. M. Shorter-Jacobi, and J. D. Schall, 2004. Dissociation of spatial attention and saccade preparation. Proc. Natl. Acad. Sci. USA 101:15541–15544. Just, M. A., P. A. Carpenter, and A. Miyake, 2003. Neuroindices of cognitive workload: Neuroimaging, pupillometric, and eventrelated potential studies of brain work. Theor. Issues Ergonomic Sci. 4:56–88. Justice, L. M., L. Skibbe, A. Canning, and C. Lankford, 2005. Pre-schoolers, print and storybooks: An observational study using eye movement analysis. J. Res. Reading 28:229– 243. Kagan, J., 2003. Biology, context, and developmental inquiry. Annu. Rev. Psychol. 54:1–23. Kahneman, D., and J. Beatty, 1966. Pupil diameter and load on memory. Science 154:1583–1585.

karatekin: eye tracking studies of normative and atypical development

295

Kahneman, D., L. Onuska, and R. E. Wolman, 1968. Effects of grouping on the pupillary response in a short-term memory task. Q. J. Exp. Psychol. 20:309–311. Karatekin, C., 2004. Development of attentional allocation on the dual task paradigm. Int. J. Psychophysiol. 52:7–22. Karatekin, C., 2006. Improving antisaccade performance in adolescents with attention-deficit/hyperactivity disorder (ADHD). Exp. Brain Res. 174:324–341. Karatekin, C., and R. F. Asarnow, 1998. Components of visual search in childhood-onset schizophrenia and attention-deficit/ hyperactivity disorder (ADHD). J. Abnorm. Child Psychol. 26: 367–380. Karatekin, C., and R. F. Asarnow, 1999. Exploratory eye movements to pictures in childhood-onset schizophrenia and attention-deficit/hyperactivity disorder (ADHD). J. Abnorm. Child Psychol. 27:35–49. Karatekin, C., D. J. Marcus, and J. W. Couperus, 2007. Regulation of cognitive resources during sustained attention and working memory in 10-year-olds and adults. Psychophysiology 44:128–144. Karatekin, C., D. J. Marcus, and T. J. White, 2007. Oculomotor and manual indices of incidental and intentional spatial sequence learning in middle childhood and adolescence. J. Exp. Child Psychol. 96:107–130. Katsanis, J., W. G. Iacono, and M. Harris, 1998. Development of oculomotor functioning in preadolescence, adolescence, and adulthood. Psychophysiology 35:64–72. Kemner, C., M. N. Verbaten, J. M. Cuperus, G. Camfferman, and H. van Engeland, 1998. Abnormal saccadic eye movements in autistic children. J. Autism Dev. Disord. 28:61–67. Kingstone, A., and R. M. Klein, 1993. Visual offsets facilitate saccadic lateny: Does predisengagement of visuospatial attention mediate this gap effect? J. Exp. Psychol. [Hum. Percept.] 19:1251– 1265. Klein, C., 2001. Developmental functions for saccadic eye movement parameters derived from pro- and antisaccade tasks. Exp. Brain Res. 139:1–17. Klein, C., and B. Feige, 2005. An independent components analysis (ICA) approach to the study of developmental differences in the saccadic contingent negative variation. Biol. Psychol. 70:105– 114. Klein, C., and B. Fischer, 2005a. Developmental fractionation and differential discrimination of the antisaccadic direction error. Exp. Brain Res. 165:132–138. Klein, C., and B. Fischer, 2005b. Instrumental and testretest reliability of saccadic measures. Biol. Psychol. 68:201– 213. Klein, C., B. Fischer, B. Fischer, and K. Hartnegg, 2002. Effects of methylphenidate on saccadic responses in patients with ADHD. Exp. Brain Res. 145:121–125. Klein, C., A. Raschke, and A. Brandenbusch, 2003. Development of pro- and antisaccades in children with attention-deficit hyperactivity disorder (ADHD) and healthy controls. Psychophysiology 40:17–28. Klin, A., W. Jones, R. Schultz, F. Volkmar, and D. Cohen, 2002. Visual fixation patterns during viewing of naturalistic social situations as predictors of social competence in individuals with autism. Arch. Gen. Psychiatry 59:809–816. Knight, R. A., and S. M. Silverstein, 2001. A process-oriented approach for averting confounds resulting from general performance deficiencies in schizophrenia. J. Abnorm. Psychol. 110: 15–30.

296

methodological paradigms

Kramer, A. F., J. C. M. Gonzalez de Sather, and N. D. Cassavaugh, 2005. Development of attentional and oculomotor control. Dev. Psychol. 41:760–772. Kramer, H. C., J. A. Yesavage, J. L. Taylor, and D. Kupfer, 2000. How can we learn about developmental processes from cross-sectional studies, or can we? Am. J. Psychiatry 157:163– 171. Krauzlis, R. J., 2005. The control of voluntary eye movements: New perspectives. Neuroscientist 11:124–137. Krauzlis, R. J., and L. S. Stone, 1999. Tracking with the mind’s eye. Trends Neurosci. 22:544–550. Kumra, S., A. Sporn, D. W. Hommer, R. Nicolson, G. Thaker, E. Israel, M. Lenane, J. Bedwell, L. K. Jacobsen, P. Gochman, and J. L. Rapoport, 2001. Smooth-pursuit eye-tracking impairment in childhood-onset psychotic disorders. Am. J. Psychiatry 158:1291–1298. Land, M. F., and S. Furneaux, 1997. The knowledge base of the oculomotor system. Philos. Trans. Royal Soc. London [Biol.] 352:1231–1239. Landry, R., and S. E. Bryson, 2004. Impaired disengagement of attention in young children with autism. J. Child Psychol. Psychiatry 45:1115–1122. Langaas, T., M. Mon-Williams, J. P. Wann, E. Pascal, and C. Thompson, 1998. Eye movements, prematurity, and developmental coordination disorder. Vis. Res. 38:1817–1826. Lasker, A. G., M. B. Denckla, and D. S. Zee, 2003. Ocular motor behavior of children with neurofibromatosis 1. J. Child Neurol. 18:348–355. Leigh, R. J., and C. Kennard, 2004. Using saccades as a research tool in the clinical neurosciences. Brain 127:460–477. Leigh, R. J., and D. S. Zee, 1999. The Neurology of Eye Movements, 3rd ed. Philadelphia: F. A. Davis. Liston, D., and R. J. Krauzlis, 2003. Shared response preparation for pursuit and saccadic eye movements. J. Neurosci. 23:11305– 11314. Liversedge, S. P., and J. M. Findlay, 2000. Saccadic eye movements and cognition. Trends Cogn. Sci. 4:6–14. Loewenfeld, I. E., 1993. The Pupil: Anatomy, Physiology, and Clinical Applications, vol. 1. Ames: Iowa State University Press/Detroit: Wayne State University Press. Luna, B., K. E. Garver, T. A. Urban, N. A. Lazar, and J. A. Sweeney, 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75:1357–1372. Luna, B., K. R. Thulborn, D. P. Munoz, E. P. Merriam, K. E. Garver, N. J. Minshew, M. S. Keshavan, C. R. Genovese, W. F. Eddy, and J. A. Sweeney, 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 13:786–793. Malone, S. M., and W. G. Iacono, 2002. Error rate on the antisaccade task: Heritability and developmental change in performance among preadolescent and late-adolescent female twin youth. Psychophysiology 39:664–673. Marcus, D. J., 2005. Processing of facial expression information in children and adults: A behavioral and psychophysiological study. Doctoral dissertation, University of Minnesota, 2005. Dissertation Abstracts International-B, 66/04, 2325. Marcus, D. J., C. Karatekin, and S. W. Markiewicz, 2006. Oculomotor evidence of sequence learning on the serial reaction time (SRT) task. Mem. Cogn. 34:420–432. Medendorp, W. P., H. C. Goltz, and T. Vilis, 2006. Remapping the remembered target location for antisaccades in human posterior parietal cortex. J. Neurophysiol. 94:734–740.

Meier, B., and W. J. Perrig, 2000. Low reliability of perceptual priming: Consequences for the interpretation of functional dissociations between explicit and implicit memory. Q. J. Exp. Psychol. [A] 53:211–233. Meyer, D. E., and D. E. Kieras, 1997. A computational theory of executive cognitive processes and multiple-task performance: Part 1. Basic mechanisms. Psychol. Rev. 104:3–65. Miller, M. B., J. P. Chapman, L. J. Chapman, and J. Collins, 1995. Task difficulty and cognitive deficits in schizophrenia. J. Abnorm. Psychol. 104:251–258. Minshew, N. J., B. Luna, and J. A. Sweeney, 1999. Oculomotor evidence for neocortical systems but not cerebellar dysfunction in autism. Neurology 52:917–922. Missal, M., and E. L. Keller, 2002. Common inhibitory mechanism for saccades and smooth-pursuit eye movements. J. Neurophysiol. 88:1880–1892. Mokler, A., and B. Fischer, 1999. The recognition and correction of involuntary prosaccades in an antisaccade task. Exp. Brain Res. 125:511–516. Moore, T., and M. Fallah, 2001. Control of eye movements and spatial attention. Proc. Natl. Acad. Sci. USA 98:1273–1276. Mostofsky, S. H., A. G. Lasker, L. E. Cutting, M. B. Denckla, and D. S. Zee, 2001. Oculomotor abnormalities in attention deficit hyperactivity disorder: A preliminary study. Neurology 57:423–430. Mostofsky, S. H., A. G. Lasker, H. S. Singer, M. B. Denckla, and D. S. Zee, 2001. Oculomotor abnormalities in boys with Tourette syndrome with and without ADHD. J. Am. Acad. Child Adolesc. Psychiatry 40:1464–1472. Müller, J. R., M. G. Philiastides, and W. T. Newsome, 2004. Microstimulation of the superior colliculus focuses attention without moving the eyes. Proc. Natl. Acad. Sci. USA 102:524– 529. Munoz, D. P., I. T. Armstrong, K. A. Hampton, and K. D. Moore, 2003. Altered control of visual fixation and saccadic eye movements in attention-deficit hyperactivity disorder. J. Neurophysiol. 90:503–514. Munoz, D. P., J. R. Broughton, J. E. Goldring, and I. T. Armstrong, 1998. Age-related performance of human subjects on saccadic eye movement tasks. Exp. Brain Res. 121:391– 400. Murthy, A., K. G. Thompson, and J. D. Schall, 2001. Dynamic dissociation of visual selection for saccade programming in frontal eye field. J. Neurophysiol. 86:2634–2637. Nieuwenhuis, S., A. Broerse, M. M. A. Nielen, and R. de Jong, 2004. A goal activation approach to the study of executive function: An application to antisaccade tasks. Brain Cogn. 56: 198–214. Nowinski, C. V., N. J. Minshew, B. Luna, Y. Takarae, and J. A. Sweeney, 2005. Oculomotor studies of cerebellar function in autism. Psychiatry Res. 137:11–19. O’Driscoll, G. A., L. Dépatie, A. V. Holahan, T. SavionLemieux, R. G. Barr, C. Jolicoeur, and V. I. Douglas, 2005. Executive functions and methylphenidate response in subtypes of attention-deficit/hyperactivity disorder. Biol. Psychiatry 57: 1452–1460. Partala, T., and V. Surakka, 2003. Pupil size variation as an indication of affective processing. Int. J. Hum.-Comput. Stud. 59:185–198. Pascual-Leone, J., 2000. Reflections on working memory: Are the two models complementary? J. Exp. Child Psychol. 77:138– 154.

Peavler, W. S., 1974. Individual differences in pupil size and performance. In M.P. Janisse ed., Pupillary Dynamics and Behavior, 159–175. New York: Plenum Press. Peterson, M. S., A. F. Kramer, and D. E. Irwin, 2004. Covert shifts of attention precede involuntary eye movements. Percept. Psychophys. 66:398–405. Pierrot-Deseilligny, C., B. Gaymard, R. Müri, and S. Rivaud, 1997. Cerebral ocular motor signs. J. Neurol. 244:65–70. Platt, M. L., 2002. Neural correlates of decisions. Curr. Opin. Neurobiol. 12:141–148. Polli, F. E., J. J. S. Barton, M. S. Cain, K. N. Thakkar, S. L. Rauch, and D. S. Manoach, 2005. Rostral and dorsal anterior cingulate cortex make dissociable contributions during antisaccade error commission. Proc. Natl. Acad. Sci. USA 102:15700– 15705. Powell, K. D., and M. E. Goldberg, 2000. Response of neurons in the lateral intraparietal area to a distractor flashed during the delay period of a memory-guided saccade. J. Neurophysiol. 84: 301–310. Pratt, J., H. Bekkering, and M. Leung, 2000. Estimating the components of the gap effect. Exp. Brain Res. 130:258–263. Pratt, J., J. Shen, and J. Adam, 2004. The planning and execution of sequential eye movements: Saccades do not show the one target advantage. Hum. Movement Sci. 22:679–688. Qiyuan, J., F. Richer, B. L. Wagoner, and J. Beatty, 1985. The pupil and stimulus probability. Psychophysiology 22:530–534. Rayner, K., 1998. Eye movements in reading and information processing: 20 years of research. Psychol. Bull. 124:372–422. Reilly, J. L., M. S. H. Harris, M. S. Keshavan, and J. A. Sweeney, 2005. Abnormalities in visually guided saccades suggest corticofugal dysregulation in never-treated schizophrenia. Biol. Psychiatry 57:145–154. Reuter, B., and N. Kathmann, 2004. Using saccade tasks as a tool to analyze executive dysfunctions in schizophrenia. Acta Psychol. 115:255–269. Rinck, M., and E. S. Becker, 2006. Spider fearful individuals attend to threat, then quickly avoid it: Evidence from eye movements. J. Abnorm. Psychol. 115:231–238. Roesch, M. R., and C. R. Olson, 2005. Neuronal activity dependent on anticipated and elapsed delay in macaque prefrontal cortex, frontal and supplementary eye fields, and premotor cortex. J. Neurophysiol. 94:1469–1497. Rosenberg, D. R., D. H. Averbach, K. M. O’Hearn, A. Seymour, B. Birmaher, and J. A. Sweeney, 1997. Oculomotor response inhibition abnormalities in pediatric obsessive-compulsive disorder. Arch. Gen. Psychiatry 54:831–838. Rosenhall, U., E. Johansson, and C. Gillberg, 1988. Oculomotor findings in autistic children. J. Laryngol. Otol. 102:435–439. Ross, R. G., 2003. Early expression of a pathophysiological feature of schizophrenia: Saccadic intrusions into smooth-pursuit eye movements in school-age children vulnerable to schizophrenia. J. Am. Acad. Child Adolesc. Psychiatry 42:468–476. Ross, R. G., S. Heinlein, G. O. Zerbe, and A. Radant, 2005. Saccadic eye movement task identifies cognitive deficits in children with schizophrenia, but not in unaffected child relatives. J. Child Psychol. Psychiatry 46:1354–1362. Ross, R. G., D. Hommer, D. Breiger, C. Varley, and A. Radant, 1994. Eye movement task related to frontal lobe functioning in children with attention deficit disorder. J. Am. Acad. Child Adolesc. Psychiatry 33:869–874. Ross, R. G., D. Hommer, A. Radant, M. Roath, and R. Freedman, 1996. Early expression of smooth-pursuit eye movement

karatekin: eye tracking studies of normative and atypical development

297

abnormalities in children of schizophrenic parents. J. Am. Acad. Child Adolesc. Psychiatry 35:941–949. Ross, R. G., A. Olincy, J. G. Harris, A. Radant, M. Hawkins, L. E. Adler, and R. Freedman, 1999. Evidence for bilineal inheritance of physiological indicators of risk in childhood-onset schizophrenia. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 88:188– 199. Ross, R. G., A. D. Radant, and D. W. Hommer, 1993. A developmental study of smooth pursuit eye movements in normal children from 7 to 15 years of age. J. Am. Acad. Child Adolesc. Psychiatry 32:783–791. Ross, R. G., A. D. Radant, D. A. Young, and D. W. Hommer, 1994. Saccadic eye movements in normal children from 8 to 15 years of age: A developmental study of visuospatial attention. J. Autism Dev. Disord. 24:413–431. Ross, S. M., and L. E. Ross, 1983. The effects of onset and offset warning and post-target stimuli on the saccade latency of children and adults. J. Exp. Child Psychol. 36:340–355. Rothkopf, C. A., and J. B. Pelz, 2004. Head movement estimation for wearable eye tracker. In Proceedings ACM SIGCHI: Eye Tracking Research and Applications, 123–130, ACM Press. Rothlind, J. S., M. I. Posner, and E. A. Schaughency, 1991. Lateralized control of eye movements in attention deficit hyperactivity disorder. J. Cogn. Neurosci. 3:377–381. Salman, M. S., J. A. Sharpe, M. Eizenmann, L. Lillakas, C. Westall, T. To, M. Dennis, and M. J. Steinbach, 2006. Saccades in children. Vis. Res. 46:1432–1439. Salman, M. S., J. A. Sharpe, L. Lillakas, M. Dennis, and M. J. Steinbach, 2006. Smooth pursuit eye movements in children. Exp. Brain Res. 169:139–143. Schall, J. D., D. P. Hanes, and T. L. Taylor, 2000. Psychol. Res. 63:299–307. Scharre, J. E., and M. P. Creedon, 1992. Assessment of visual function in autistic children. Optom. Vis. Sci. 69:433–439. Scherberger, H., M. A. Goodale, and R. A. Andersen, 2003. Target selection for reaching and saccades share a similar behavioral reference frame in the macaque. J. Neurophysiol. 89:1456– 1466. Scherf, K. S., J. A. Sweeney, and B. Luna, 2006. Brain basis of developmental change in visuospatial working memory. J. Cogn. Neurosci. 18:1045–1058. Schiller, P. H., and J. Kendall, 2004. Temporal factors in target selection with saccadic eye movements. Exp. Brain Res. 154:154– 159. Schneider, W., V. Kron, M. Hünnerkopf, and K. Krajewski, 2004. The development of young children’s memory strategies: First findings from the Würzburg longitudinal memory study. J. Exp. Child Psychol. 88:193–209. Schreiber, H., G. Stolz-Born, J. Born, J. Rothmeier, A. Rothenberger, R. Jürgens, W. Becker, and H. H. Kornhuber, 1997. Visually-guided saccadic eye movements in adolescents at genetic risk for schizophrenia. Schizophr. Res. 25:97–109. Schwarzer, G., S. Huber, and T. Dümmler, 2005. Gaze behavior in analytical and holistic face processing. Mem. Cogn. 33:344–354. Sekerina, I. A., K. Stromswold, and A. Hestvik, 2004. How do adults and children process referentially ambiguous pronouns? J. Child Lang. 31:123–152. Shen, K., and M. Paré, 2006. Guidance of eye movements during visual conjunction search: Local and global contextual effects on target discriminability. J. Neurophysiol. 95:2845–2855. Shima, K., and J. Tanji. 1998. Both supplementary and presupplementary motor areas are crucial for the temporal organization of multiple movements. J. Neurophysiol. 80:3247–3260.

298

methodological paradigms

Siegle, G. J., S. R. Steinhauer, C. S. Carter, W. Ramel, and M. E. Thase, 2003. Do the seconds turn into hours? Relationships between sustained pupil dilation in response to emotional information and self-reported rumination. Cogn. Ther. Res. 27:365– 382. Siegle, G. J., S. R. Steinhauer, V. A. Stenger, R. Konecky, and C. S. Carter, 2003. Use of concurrent pupil dilation assessment to inform interpretation and analysis of fMRI data. NeuroImage 20:114–124. Siegler, R. S., 1998. Emerging Minds: The Process of Change in Children’s Thinking. New York: Oxford University Press. Simion, F., and G. Butterworth, eds., 1998. The Development of Sensory, Motor and Cognitive Capacities in Early Infancy: From Perception to Cognition. East Sussex, UK: Psychology Press. Sinha, N., J. T. G. Brown, and R. H. S. Carpenter, 2006. Task switching as a two-stage decision process. J. Neurophysiol. 95: 3146–3153. Smyrnis, N., I. Evdokimidis, N. C. Stefanis, D. Avramopoulos, T. S. Constantinidis, A. Stavropoulos, and C. N. Stefanis, 2003. Antisaccade performance of 1,273 men: Effects of schizotypy, anxiety, and depression. J. Abnorm. Psychol. 112:403– 414. Smyrnis, N., I. Evdokimidis, N. C. Stefanis, T. S. Constantinidis, D. Avramopoulos, C. Theleritis, C. Paximadis, C. Efstratiadis, G. Kastrinakis, and C. N. Stefanis, 2002. The antisaccade task in a sample of 2,006 young males. II. Effects of task parameters. Exp. Brain Res. 147:53–63. Spantekow, A., P. Krappmann, S. Everling, and H. Flohr, 1999. Event-related potentials and saccadic reaction times: Effects of fixation point offset or change. Exp. Brain Res. 127:291–297. Starr, M. S., and K. Rayner, 2001. Eye movements during reading: Some current controversies. Trends Cogn. Sci. 5:156– 163. Steinhauer, S. R., and G. Hakerem, 1992. The pupillary response in cognitive psychophysiology and schizophrenia. Ann. NY Acad. Sci. 658:182–204. Steinhauer, S. R., G. J. Siegle, R. Condray, and M. Pless, 2004. Sympathetic and parasympathetic innervation of pupillary dilation during sustained processing. Int. J. Psychophysiol. 52:77– 86. Swanson, H. L., 1999. What develops in working memory? A lifespan perspective. Dev. Psychol. 35:986–1000. Takarae, Y., N. J. Minshew, B. Luna, C. M. Krisky, and J. A. Sweeney, 2004. Pursuit eye movement deficits in autism. Brain 127:2584–2594. Thier, P., and U. J. Ilg, 2005. The neural basis of smooth-pursuit eye movements. Curr. Opin. Neurobiol. 15:645–652. Thompson, K. G., N. P. Bichot, and T. R. Sato, 2005. Frontal eye field activity before visual search errors reveals the integration of bottom-up and top-down salience. J. Neurophysiol. 93: 337–351. Trillenberg, P., R. Lencer, and W. Heide, 2004. Eye movements and psychiatric disease. Curr. Opin. Neurol. 17:43–47. Trueswell, J. C., I. Sekerina, N. M. Hill, and M. L. Logrip, 1999. The kindergarten-path effect: Studying on-line sentence processing in young children. Cognition 73:89–134. Tseng, Y., and C. R. Li, 2004. Oculomotor correlates of contextguided learning in visual search. Percept. Psychophys. 2004: 1363–1378. Van der Geest, J. N., C. Kemner, G. Camfferman, M. N. Verbaten, and H. van Engeland, 2001. Eye movements, visual attention, and autism: A saccadic reaction time study using the gap and overlap paradigm. Biol. Psychiatry 50:614–619.

Van der Geest, J. N., C. Kemner, G. Camfferman, M. N. Verbaten, and H. van Engeland, 2002a. Looking at images with human figures: Comparison between autistic and normal children. J. Autism Dev. Disord. 32:69–75. Van der Geest, J. N., C. Kemner, M. N. Verbaten, and H. van Engeland, 2002b. Gaze behavior of children with pervasive developmental disorder toward human faces: A fixation time study. J. Child Psychol. and Psychiatry 43:669–678. Van Loon, E. M., I. T. C. Hooge, and A. V. Van den Berg, 2002. The timing of sequences of saccades in visual search. Proc. R. Soc. Lond. [Biol.] 269:1571–1579.

Vaziri, S., J. Diedrichsen, and R. Shadmehr, 2006. Why does the brain predict sensory consequences of oculomotor commands? Optimal integration of the predicted and the actual sensory feedback. J. Neurosci. 26:4188–4197. Von Hofsten, C., 2004. An action perspective on motor development. Trends Cogn. Sci. 8:266–272. Wardak, C., G. Ibos, J. Duhamel, and E. Olivier, 2006. Contribution of the monkey frontal eye field to covert visual attention. J. Neurosci. 26:4228–4235. Yarbus, A. L., 1967. Eye Movements and Vision, trans. B. Haigh. New York: Plenum. (Original work published 1965.)

karatekin: eye tracking studies of normative and atypical development

299

17

Diffusion Tensor Imaging JEFFREY R. WOZNIAK, BRYON A. MUELLER, AND KELVIN O. LIM

Principles of diffusion tensor imaging

DTI data acquisition

Unlike traditional anatomical magnetic resonance imaging (MRI) techniques, which provide information about brain macrostructure, such as brain size and specific tissue volumes, diffusion tensor imaging (DTI) provides microstructural information about brain tissue. Diffusion tensor imaging is sensitive to developmental change in tissue microstructure and damage at the microstructural level. This level of sensitivity exceeds that provided by traditional anatomical imaging techniques. Because it provides measures of tissue organization at the microstructural level, DTI is particularly applicable to the study of normal white matter development as well as abnormal neurodevelopment, because it can detect differences in the normally highly ordered structure of white matter fibers. Diffusion tensor imaging is a technique that depends on measuring the diffusion of water molecules in tissue. In free water, the random Brownian movement of water molecules is equal in all directions and is referred to as isotropic. In contrast, when barriers to diffusion, such as cell membranes, fibers, and myelin are present, diffusion occurs more easily in the direction of the fibers rather than perpendicular to the fibers, and the diffusion is referred to as anisotropic (Beaulieu, 2002). In the 1960s it was discovered that magnetic resonance is particularly well suited to the quantitative and noninvasive measurement of water diffusion (Stejskal and Tanner, 1965). However, the measurement of diffusion with magnetic resonance was not developed into an imaging technique until later when Le Bihan and colleagues (1986) first described the measurement of the apparent diffusion coefficient (ADC). Early animal experiments using diffusion methods demonstrated that ADC is higher when measured in the direction parallel to white matter fibers compared with perpendicular to the fibers, thus providing information about tissue organization (Moseley et al., 1990). Understanding the relationship between diffusion and fiber direction, together with the ability to detect localized diffusion differences with MRI, formed the basis for the development of DTI. Diffusion-tensor-imaging protocols rely on measurements of the magnitude and direction of water-molecule diffusion in three dimensions and subsequently reconstruct the data into a three-dimensional volume (Basser, Mattiello, and Le Bihan, 1994).

A complete description of the physics of diffusion imaging is beyond the scope of this chapter, but can be found in several review articles in the literature (Beaulieu, 2002; Le Bihan, 2003; Mori and Zhang, 2006). Briefly, diffusion is the random, translational motion of molecules caused by collisions with other surrounding molecules. Diffusion imaging sequences use specially designed pulse sequences to measure the magnitude and direction of water diffusion in tissue. The measured diffusion property of tissue is referred to as the apparent diffusion coefficient (ADC) because it differs from the true, microscopic diffusion value. This difference exists because the complex structure of tissue hinders diffusion, reducing the measured value. To put this description of diffusion in perspective, it is worth noting that the average distance traveled by a free water molecule during the diffusion-encoding part of the imaging sequence is about 10 microns (Le Bihan, 2003; Mori and Zhang, 2006). Axons in the human corpus callosum have a median diameter of 0.6 to 1 micron with a range of 0.2 to 10 microns (Aboitiz et al., 1992). These measurements illustrate how water diffusion within an axon can be several times greater in the direction of fibers than in the direction perpendicular to fibers. It is also worth pointing out that the actual spatial resolution of the DTI image is orders of magnitude larger than the scale of the individual axon. For example, a typical DTI voxel (a single threedimensional “volume pixel” in the resultant image) is 2.5 × 2.5 × 2.5 millimeters, or 2,500 microns per side and 6.25 million square microns in cross section. In this corpus callosum example, therefore, the measured diffusion is a result of millions of axons that pass through that single voxel. A diffusion-imaging sequence creates contrast (typically viewed as shades of gray in an image) by causing the MR signal within the voxel to decrease when water molecules move along a specific direction. To understand how this is accomplished, it is necessary to first understand how magnetic field gradients affect the MR signal. The frequency of the MR signal, ω, is proportional to the magnetic field strength, B0, through the Larmor relationship ω = γB0, where γ is a Larmor constant, a number specific to each molecule. Magnetic resonance scanners have special gradient hardware that can create a linear change in the strength of the magnetic field, resulting in a linear relationship between MR

301

signal frequency and position, ω(X) = γB0 + γG · X. For example, a gradiant can be applied such that the magnetic field is temporarily strongest at the right side of the head and decreased slightly over the distance to the left side of the head. Mapping spatial location to MR signal frequency is one of the fundamental elements of creating a magnetic resonance image. A second critical element is mapping position with the phase of the MR signal. The MR signal from each voxel is the vector sum of the signals from all the molecules in that voxel. Initially, when the MR signal is first produced, the proton spins from water molecules in the voxel are pointing in the same direction, rotating together. The spins have the same phase, and the MR signal is at maximum. With time, tiny differences in the local magnetic field cause the spins to rotate at different speeds. These magnetic field inhomogeneities cause the spins to point in different directions, or dephase, thus reducing the vector sum of the spins and producing a reduction in the MR signal from the voxel. A diffusion-imaging sequence creates image contrast by dephasing spins in the water molecules that move/diffuse relative to stationary water molecules. While many MR imaging sequences can be modified to perform diffusion imaging (Bammer, 2003), the spin echo sequence is most often used in clinical DTI. In this sequence, diffusionsensitive gradients with the exact same amplitude and duration are applied both before and after the 180-degree radio frequency (RF) refocusing pulse. The refocusing pulse inverts the phase accumulation caused by the earlier gradient, causing the second gradient to exactly rewind and cancel the accumulated phase for all of the spins that remain stationary. For water molecules that move during the application of the gradients, the phase is not perfectly rewound at the conclusion of the second gradient; the phase between molecules differs, resulting in a reduction in the signal from the voxel. The signal loss is an exponential function S = S0 exp(−bDADC) that depends on the apparent diffusion coefficient (DADC) of water along the direction of the diffusion encoding gradient and on b, the value of the magnitude of the diffusion-sensitizing gradient (b value), which depends on the specific properties of the pulse sequence (Bammer, 2003). A DTI data set includes one or more images collected without the application of any diffusion gradients and multiple diffusion-weighted images, obtained with the diffusionsensitized gradients applied in noncollinear directions. A minimum of six noncollinear images is needed, but, often, more images are collected in order to increase the accuracy of the measures. Using simulations, Jones (2004) demonstrated that at least 20 unique sampling orientations are required for a robust estimation of anisotropy, while 30 may be required for robust determination of diffusion direction and mean diffusivity (MD).

302

methodological paradigms

DTI data processing Most of the computation of DTI measures is performed using specialized software programs run on postprocessing computers, utilizing either in-house software or relying on freely available software packages. In most cases, processing begins by converting the data from the native scanner image format (vender specific, but often DICOM) into the image format (NIFTI, Analyze, others) that is required by the chosen processing program. An eddy current correction algorithm should typically be applied to the data to remove the image distortion (stretching, shearing) induced by transient current loops produced in the scanner hardware by the rapidly switching diffusion gradients (Haselgrove and Moore, 1996). Maps of apparent diffusion coefficient are then generated by dividing each diffusion volume (the noncollinear images) by the volume acquired without the application of a diffusion gradient. The diffusion tensor elements (a 3 × 3 matrix) are computed from the ADC maps using linear regression or more sophisticated techniques, and the eigenvalues and eigenvectors are computed from the diffusion matrix through the use of matrix diagonalization (Basser and Pierpaoli, 1998). Three eigenvectors and three eigenvalues are computed. The first and largest of the eigenvalues, λ1, is representative of the diffusion occurring in parallel to the axonal fibers and is called axial diffusivity (Basser and Pierpaoli, 1995). The second and third eigenvalues, λ2 and λ3, represent diffusion in the planes orthogonal to the long axis of the axon. Radial diffusivity is the average of the two smaller eigenvalues (λ2 + λ3)/2 while trace is the sum of all three eigenvalues (λ1 + λ2 + λ3). A less commonly used measure, relative anisotropy (RA), is the ratio of the anisotropic component of the tensor to the isotropic component. In contrast, fractional anisotropy (FA) represents the fraction of the magnitude of tensor that is due to anisotropic water diffusion (Basser, Mattiello, and Le Bihan, 1994; Masutani et al., 2003). Fractional anisotropy is perhaps the most commonly reported measure in studies utilizing DTI. The formula that follows is for the computation of FA. In this equation, λM represents the mean diffusivity, or (λ1 + λ2 + λ3)/3. FA=

3 2

( λ1 − λ M )2 + ( λ 2 − λ M )2 + ( λ 3 − λ M )2 λ12 + λ 22 + λ 32

Once the maps of the scalar quantities such as FA are computed for each voxel in the three-dimensional volume, there are several methods available to the researcher to investigate diffusion changes in development. If the investigator is interested in gross-level analyses, mean regional or whole-brain values of the scalar measure can be obtained. Anatomical MR images collected during the scan session can be segmented and used to identify the tissue type (white

matter, gray matter, etc.) for each voxel. The DTI data are then aligned to the anatomical image, a tissue-type “mask” is used to identify specific areas of the brain, and the mean value in that region and tissue type is then computed. For example, all the white matter in the brain or all voxels in the corpus callosum might be selected and the mean DTI measures then computed. Investigators can also utilize a “regions of interest” (ROI) approach, in which a specific area in the brain is identified by hand for each subject. Regions can be identified by placing circular or square ROIs on the image or by hand drawing regions on the image itself. When the ROI method is employed, the analysis focuses on a mean value of the scalar measure derived from only those voxels within the ROI. A third method of DTI analysis utilizes the voxelbased approach in which each voxel is considered to be a data point itself or is “clustered” with neighboring voxels based on statistical techniques. In a voxel-based analysis, the data from each subject are first transformed or aligned onto a template image utilizing a nonlinear alignment method. The alignment process is necessary to compensate for the anatomical variability between subjects, to insure that a particular voxel represents the brain structure across subjects. Statistical methods are used to identify voxels or clusters of voxels that are different between populations or to represent correlations with some other variable (e.g., age). Voxel-based methods are complicated by the need to correct for the large number of statistical tests and are, thus, often considered exploratory or hypothesis generating in nature. Last, tractography can be employed to identify specific anatomic regions that roughly correspond to underlying white matter tracts. Scalar measures, such as FA, can then be derived from the voxels that are identified as within a specific “tract.” Each of these methods is associated with challenges, but each offers unique advantages. In many cases, analysis at more than one of these levels is appropriate.

Anatomical and physiological correlates of DTI measures In a DTI image, the contrast is primarily a function of the differential diffusion of water molecules in gray matter, white matter, cerebrospinal fluid (CSF), and other tissue. In CSF, diffusion is essentially unrestricted and random in all directions. The anisotropy is zero and MD is high. In gray matter, diffusion is restricted, but the restriction is not directional in nature. Anisotropy is low, and MD is lower than that for CSF. In white matter regions with highly ordered structure, such as the corpus callosum, diffusion is directional, anisotropy is high, and MD is lower than CSF but nearly the same as in gray matter. In general, DTI measures are not especially sensitive to variations in gray matter microstructure (either normal developmental or pathological variations),

but DTI measures are highly sensitive to changes in white matter microstructure. Currently, the physiological underpinnings of DTI measures such as anisotropy are not yet fully understood. Initially, myelin was thought to be the primary contributor to anisotropy in axons (Rutherford et al., 1991). However, several studies investigating the role of myelin have shown that myelinated and unmyelinated axons actually have similar anisotropy (Beaulieu and Allen, 1994a, 1994b). These studies of excised, nonmammalian axons may be limited in terms of generalizability to intact mammalian systems, but the basic finding has been corroborated by others including Gulani and colleagues (2001), who reported only a modest anisotropy difference between myelinated and unmyelinated rat spinal cords. In that study, tissue from myelin-deficient rats showed significant anisotropy. Also, studies of “premyelinated” neurons in mammalian species, including rat pups, have corroborated the general finding that myelin is not necessary in order for white matter fibers to display significant diffusion anisotropy (Prayer et al., 2001; Wimberger et al., 1995). Studies of human newborns that examine white matter regions prior to significant myelination also draw the same conclusion (Huppi et al., 1998; Inder and Huppi, 2000). Thus, it is now well established that myelin is not the sole determining factor in white matter anisotropy, but alterations in myelin are reflected in altered anisotropy. Other potential sources of diffusion anisotropy in white matter include the axonal membrane itself and the longitudinally oriented microtubules and neurofilaments that are part of the structure of the axon (Beaulieu, 2002; Neil et al., 2002). It has also been proposed that fast axonal transport may accentuate diffusion, as measured by DTI, in the direction of the axon (Beaulieu, 2002). Last, extra-axonal water diffusion between densely packed axons is believed to be another important factor contributing to anisotropy in white matter. As axons mature and myelinate, this extra-axonal space decreases, resulting in increasingly restricted diffusion and increased anisotropy with maturation. For a very thorough discussion of the basis of anisotropy in neural fibers, see Beaulieu’s review (2002). While FA has been the most commonly reported DTI measure in studies of development and pathology, alternative measures such as axial, radial, and mean diffusivity may be useful (alone or in combination with each other) in characterizing different aspects of white matter status. For example, Song and colleagues (2002) examined the effects of dysmyelination using shiverer mice, which have a mutation in myelin basic protein that results in incomplete myelination, but otherwise normal axons. In that study, dysmyelination was reflected in radial diffusivity abnormalities, but no differences in axial diffusivity. Animals without myelin showed nearly the same level of axial diffusivity as those with normal myelin, but their radial diffusivity was higher than

wozniak, mueller, and lim: diffusion tensor imaging

303

control mice. Interestingly, the authors noted that the nearly complete lack of myelin in the mice only resulted in a 20 percent increase in radial diffusivity, again clearly indicating the importance of nonmyelin factors in white matter diffusion properties as measured by DTI. Further evidence for the specificity of DTI measures comes from a study of retinal ischemia in mice (Song et al., 2003). In that study, optic nerve axonal degeneration, which occurred quickly after ischemia, was associated with a decrease in axial diffusivity. Demyelination, which occurred later, was reflected in a subsequent increase in radial diffusivity. Thus the axial diffusivity measure may be particularly sensitive to damage to the axon itself, whereas radial diffusivity is sensitive to abnormalities in the myelin sheath surrounding the axon. Song and colleagues (2005) have also shown that radial diffusivity functions as a real-time measure of myelin status in experimental manipulations. Radial diffusivity was sensitive to both demyelination and remyelination in mice that were fed cuprizone, a demyelinating agent. Demyelination was reflected in increased radial diffusivity, and remyelination, after removal of dietary cuprizone, was reflected in the return of radial diffusivity values to normative levels. A follow-up paper by the same group further clarified that axial diffusivity reflected damage to axons early on in the course of cuprizone-induced white matter changes and that radial diffusivity reflected the loss of myelin that occurred later (Sun et al., 2006). In summary, the various component measures from a DTI series may provide unique information about the type of white matter damage or abnormality present. It is likely that examinations of radial and axial diffusivity will also prove useful in the study of normal brain development.

Methodological challenges in DTI data acquisition Because DTI is designed to be sensitive to molecular motion, it is also sensitive to bulk motion, which can result in image movement artifacts if a subject moves. Strategies that have been developed to reduce subject movement include physical means, such as immobilization padding or providing real-time feedback to the subject to signal and prevent further movement. Single-shot imaging is generally used, and postprocessing registration can correct some motion artifacts. The rapid rise of the magnetic gradient during the acquisition of diffusion-weighted images contributes to the formation of eddy currents in the scanner hardware. The eddy current produces distortions in the highly uniform magnetic field of the scanner. Although eddy currents occur in other MRI acquisitions, DTI is particularly prone to eddy-current artifacts because of the large and frequent gradients that are used. One method of eddy-current correction involves collecting a map of the magnetic field and, sub-

304

methodological paradigms

sequently, correcting image distortions based on the observed distortions in the field ( Jezzard, Barnett, and Pierpaoli, 1998). This is not commonly used because of the time required to collect the additional field map scan. More commonly, eddy-current correction is done in postprocessing, using image-registration techniques. The least distorted image, typically the image collected without the application of the diffusion gradient, is used as the “target image” to which the others are spatially aligned. A variety of alignment methods can be used, some of which also help correct the images for subject motion (Rohde et al., 2004). An additional source of magnetic field inhomogeneity is the human head itself. Specifically, in anterior, inferior frontal and temporal regions, the transition between tissue and air-filled sinuses results in significant disturbances in the magnetic field. “Susceptibility artifacts” result from this disturbance and make imaging the inferior frontal lobes quite difficult. Common artifacts include image distortion, signal pileup (bright regions), and image dropout (areas without signal). These artifacts are a result of the rapid imaging sequence, and typically echo planar imaging (EPI), which is highly sensitive to magnetic field inhomogeneities. Distortion effects can be corrected by applying algorithms based on the magnetic field distortion maps discussed earlier. Regions with signal dropout and pileup, however, cannot be recovered. One promising approach to this problem is the use of parallel imaging that uses multiple coils to reduce the amount of time required to acquire the signal, resulting in a reduction in artifact (Bammer et al., 2001). Ultimately, a DTI data set is dependent on a number of important decisions concerning acquisition parameters. These decisions reflect compromises that must be made based on the available equipment, the available time for acquisition, the desired quality of the resulting data, and the experimental needs.

Developmental correlates of DTI measures Because of their sensitivity to microstructural tissue changes, diffusion-imaging techniques are proving very useful in the study of brain development during infancy, childhood, and adolescence (Neil et al., 1998). Water diffusion in the brain changes dramatically during early developmental stages, reflecting changes in the underlying tissue structure and overall water content in the brain (Rutherford et al., 1991; Sakuma et al., 1991). For an excellent visual review of the changes that occur in MR images, including diffusion images, during very early brain development, see Miller and colleagues (2003). One diffusion measure, ADC, has been shown to be significantly higher in the brains of newborns than in those of older infants. During the first 3 or 4 months of life, ADC drops by more than 25 percent (Morriss et al., 1999). By age

three, ADC values are similar to adult values. Some aspects of very early development are reflected in fractional anisotropy decreases, such as those seen in newborns (Gilmore et al., 2004). Within a longer developmental timeframe, mean diffusivity generally decreases and FA increases with brain maturation as a result of development of the axon structure, changes to the axonal membrane, and myelination (Neil et al., 1998; Nomura et al., 1994; Sakuma et al., 1991). In the myelin sheath itself, water diffusion becomes increasingly anisotropic with development, also contributing to changes in MD and FA. In one early DTI investigation of development, Klingberg and colleagues (1999) reported lower anisotropy in frontal white matter in children (mean age 10) who were compared to adults (mean age 27). In a study of children between the ages of 5 and 18, Schmithorst and colleagues (2002) found significant inverse correlations between age and white matter trace values. They also observed positive correlations between FA measures and age in a number of different white matter regions. Similarly, Snook and colleagues (2005) observed positive correlations between age and FA in the corpus callosum, corona radiata, and several other regions in a sample of children 8 to 12 years of age. Those authors noted that previous studies have not shown volumetric changes in those same regions. Thus DTI potentially promises to provide a set of metrics for characterizing the developmental status and integrity of white matter tissue in the brain at a level not previously available to researchers interested in understanding brainbehavior relationships in the context of development. Early efforts to develop “normative” data for brain white matter using DTI measures now appear in the literature (Hermoye et al., 2006). In the near future, much larger databases of this type will almost certainly be available to researchers and will serve a critical role in future studies of neurocognitive development. As DTI tractography methods improve (see the next section for a thorough discussion), it becomes increasingly possible to study the development of individual white matter tracts and, potentially, specific insults to those tracts. Huang and colleagues (2006) employed these methods with fetal tissue samples, neonates, and children, demonstrating the ability to track the emergence and development of limbic fibers, then associational fibers, and finally commissural and projection fibers. They suggest that the methodology may be particularly useful in studying neurodevelopmental insults that occur during a specific time frame such as prenatal insults resulting in cerebral palsy. These methods may prove useful in studying the impact that various potential neurodevelopmental insults (toxic exposure, maternal stress and illness, malnutrition, etc.) have on the developing fetus’s brain and eventual cognitive status. Diffusion-tensor-imaging studies of cognition are beginning to appear in the literature. In one study of cognitive

development in children ages 7 to 18, Nagy, Westerberg, and Klingberg (2004) reported significant correlations between reading and working-memory measures and FA in specific left-frontal and left-temporal white matter regions. These authors also reported significant relationships between age and FA in these subjects. Their data suggest that DTI is useful in understanding the relationship between structural brain changes and the development of specific cognitive skills. Similarly, Mabbott and colleagues (2006) investigated relationships between age, cognition, and DTI measures (FA and ADC) in several white matter tracts. Significant increases in FA across age (6 to 17) were seen in white matter throughout both hemispheres and in the posterior portion of the corpus callosum. Fewer age-related changes in FA were found in commissural pathways and major projection bundles compared to hemispheric white matter. Certain cognitive domains, including speed of visual search, were correlated with FA in hemispheric white matter. In that study ADC did not correlate significantly with cognitive measures. Liston and colleagues (2006) used DTI to examine the neural circuits underlying elements of cognitive control development in subjects 7 to 31 years old. Fibertracking techniques were used to identify frontrostriatal tracts. Radial diffusivity was found to be inversely correlated with age in this region. Radial diffusivity was also found to be inversely correlated with faster reaction times during a decision-making task, even after age was removed during the examination of the partial correlation. These relationships were not observed in a “control” white matter tract, a corticospinal tract that was assumed to be unrelated to the cognitive function in question. A critical next phase of DTI studies of normal and abnormal development will incorporate longitudinal studies. At this point, the majority of data are from cross-sectional investigations. While these studies provide much-needed insight into processes, time course, and potential insults to brain development, they have limitations, including the methodological challenges associated with alignment and transformation of the brain images in order to perform group analyses. These types of manipulations are particularly challenging against the backdrop of a developing brain. Longitudinal studies employing DTI will provide muchneeded data about individual brain development and will likely allow for more detailed structural analyses.

Advanced DTI methods Diffusion-tensor-imaging protocols result in the computation of vectors for each voxel in the brain. These vectors indicate the predominant orientation for the diffusion in three-dimensional space. Tractography uses the vector information derived from DTI and related imaging techniques to visualize, identify, and quantify putative white

wozniak, mueller, and lim: diffusion tensor imaging

305

matter tracts in the brain. It produces realistic pictures that closely match our notion of the brain’s fiber tracts based on knowledge of neuroanatomy and white matter dissections. However, it is important to remember that the technique does not directly image the actual fibers. Instead, it allows for a detailed modeling of the interaction of the water molecules with the tissue. While the qualitative characteristics of tractography images can be compelling, there are issues about the use of the data for the quantitative measurements required for clinical and research investigations ( JohansenBerg and Behrens, 2006). Despite the quality of the images produced, tractography has important limitations, the first of which is the low spatial resolution of the DTI data on which it is based. Typical high-resolution magnetic resonance anatomical images have voxels on the order of 1 mm × 1 mm × 1 mm, or 1 mm3. Data from DTI have a typical voxel size of 2.5 mm × 2.5 mm × 2.5 mm, or 16 mm3, more than an order of magnitude larger in volume than for anatomical imaging, with the potential for millions of axons passing through the voxel. The low resolution is due to technical limitations of the methods used to acquire the DTI data. In order to produce the tractography images, the DTI data are interpolated by factors of 10 to 20× to produce the samplings necessary for computing and displaying tracts. In figure 17.1 and plate 42 the left panel shows the cingulum bundle as rendered by a streams-based tractography program (Kim et al., 2006). On the right panel is the uninterpolated coronal fractional anisotropy image from the same data set passing through the pons and corpus callosum. The cingulum is barely discernable above the corpus callosum because of the limited spatial resolution.

Another limitation of DTI tractography is the lack of a measurement gold standard because the great majority of the data are collected in vivo. Although some experimental phantoms for modeling DTI phenomena have been developed (Perrin et al., 2005), there are no generally accepted and available phantoms that an investigator can utilize to quantitatively verify techniques and measurements. The poor spatial resolution and the lack of validation cloud the interpretation of DTI-based tractography measurements. In one case report of tractography employed in neurosurgical planning for tumor resection, the method was found to underestimate the volume of a motor tract, resulting in hemiplegia for the patient (Kinoshita et al., 2005). Fortunately, use of tractography for research does not pose such high risk for subjects, but it does point out the potential inaccuracy and limitations of the methods. Recently the test-retest reliability and intersubject variability of FA, MD, and volume of specific white matter tracts based on tractography were reported along with the power calculations for typical comparisons (Heiervang et al., 2006). Reliability was found to be highest for FA and MD in the fiber tracts and lowest for the volume of the tracts. The authors calculated that, because of the error in the tractography measurement, at least 105 subjects per group would be required in order to detect a 10 percent volume difference between groups. Clearly there are serious implications for studies employing tractography-based volume measurements, most of which have far fewer than this number of subjects. A key hope for tractography is that it can provide information about anatomical connectivity between brain regions, complementing functional information obtained from fMRI.

Figure 17.1 Sagittal view of the cingulum bundle as rendered by stream-based tractography (left panel) and un-interpolated DTI fractional anisotropy image of the same data in the coronal plane (right panel) illustrating the low spatial resolution of DTI. (See plate 42.)

306

methodological paradigms

Tractographic methods employ various algorithms to determine whether adjacent, highly interpolated voxels are likely part of the same fiber tract. The direction of the primary eigenvector is assumed to represent the direction of the fibers in the voxel and is used to build a voxel-to-voxel path, connecting regions of the brain. Unfortunately, the fact that there are many regions in the brain where fibers cross can result in ambiguous tract identification from the available data (primary eigenvectors). This is one of many such issues that pose challenges to investigators using tractographybased methods. New methods are emerging to address these challenges. Diffusion spectral imaging (DSI) aims to determine probability diffusion function (PDF), the diffusion probability distribution as a function of distance, as well as direction, of water in each voxel (Basser, 2002; Tuch et al., 2003). While measuring the PDF does allow for the resolution of multiple fiber tracts within a single voxel and DSI does provide unique information about water diffusion in the brain, there are limitations to this methodology. Because a DSI acquisition consists of hundreds of diffusion volumes and takes an order of magnitude longer to acquire than a traditional DTI examination, DSI may not be a practical method for imaging young subjects (Tuch et al., 2003). High-angular-resolution diffusion imaging (HARDI) methods aim to measure the orientation distribution function (ODF), the probability that a water molecule will diffuse in a particular angular direction, without regard to the magnitude of the diffusion (Alexander, 2005). A HARDI acquisition consists of diffusion images acquired with many directions (anywhere from 60 to 252 equally spaced volumes) using only a single, large b value. Large numbers of equally spaced directions increase the ability to resolve crossing fibers with shallower angular separation. Studies using 60– 252 uniformly sampled angles have been published, requiring diffusion acquisition times from 10 to 30 minutes (Tuch et al., 2003). The high-angular sampling allows for the use of reconstruction methods beyond DTI. Many reconstruction methods have been proposed for HARDI data (Ramnani et al., 2004), including spherical harmonic deconvolution (Hess et al., 2006; Tournier et al., 2004), generalized diffusion tensor imaging (Ozarslan and Mareci, 2003), and the q-ball approach (Tuch, 2004). Although these methods show promise, there are limitations to the number of crossing and/or divergent fibers that can be accounted for by the models. In a recent in vivo study, Cheng and colleagues (2006) used tractography to map the thalamic connections to the cerebellum and performed a validation and reliability study on these results. This study and others make it clear that the ability to resolve a voxel containing crossed fibers depends on the details of the algorithm, the number of independent diffusion volumes collected in the acquisition, the signal to

noise in the acquisition, the volume fraction of the crossed fibers, and artifactual effects such as motion and flow. Much work remains in understanding and solving the challenges inherent in tractography, but this promising technique remains an area of active research that likely will continue to evolve and improve, and should contribute greatly to our understanding of the connectivity within the brain.

Conclusions The dramatic increase in published studies using DTI over the last few years and the increasing availability of the methods to investigators is promising. Diffusion-tensorimaging methods are particularly well suited for application to the study of normal and abnormal brain development. The new tools will make it possible to study brain-behavior relationships at a level not previously possible. In addition to benefiting from advances in DTI methodology, developmental neuroscientists are in a unique position to help refine the techniques, establish their reliability and validity, and demonstrate their application in uncovering the complexities of human brain development.

REFERENCES Aboitiz, F., A. B. Scheibel, R. S. Fisher, and E. Zaidel, 1992. Fiber composition of the human corpus callosum. Brain Res. 598(1–2):143–153. Alexander, D. C., 2005. Multiple-fiber reconstruction algorithms for diffusion MRI. Ann. NY Acad. Sci. 1064:113–133. Bammer, R., 2003. Basic principles of diffusion-weighted imaging. Eur. J. Radiol. 45(3):169–184. Bammer, R., S. L. Keeling, M. Auer, K. P. Pruessmann, P. Roeschmann, R. Stollberger, et al., 2001. Diffusion tensor imaging using sense-single-shot epi. Paper presented at the Proceedings of the International Society of Magnetic Resonance in Medicine, Glasgow, Scotland. Basser, P. J., 2002. Relationships between diffusion tensor and qspace MRI. Magn. Reson. Med. 47(2):392–397. Basser, P. J., J. Mattiello, and D. Le Bihan, 1994. MR diffusion tensor spectroscopy and imaging. Biophys. J. 66(1):259–267. Basser, P. J., and C. Pierpaoli, 1995. Elucidating tissue structure by diffusion tensor MRI. Paper presented at the Proceedings of the International Society for Magnetic Resonance in Medicine, Third Scientific Meeting and Exhibition, Nice, France, August 19–25. Basser, P. J., and C. Pierpaoli, 1998. A simplified method to measure the diffusion tensor from seven MR images. Magn. Reson. Med. 39(6):928–934. Beaulieu, C., 2002. The basis of anisotropic water diffusion in the nervous system—A technical review. NMR Biomed. 15(7–8): 435–455. Beaulieu, C., and P. S. Allen, 1994a. Determinants of anisotropic water diffusion in nerves. Magn. Reson. Med. 31(4):394–400. Beaulieu, C., and P. S. Allen, 1994b. Water diffusion in the giant axon of the squid: Implications for diffusion-weighted MRI of the nervous system. Magn. Reson. Med. 32(5):579–583.

wozniak, mueller, and lim: diffusion tensor imaging

307

Cheng, P., V. A. Magnotta, D. Wu, P. Nopoulos, D. J. Moser, J. Paulsen, et al., 2006. Evaluation of the GTRACT diffusion tensor tractography algorithm: A validation and reliability study. NeuroImage 31(3):1075–1085. Gilmore, J. H., G. Zhai, K. Wilber, J. K. Smith, W. Lin, and G. Gerig, 2004. 3-tesla magnetic resonance imaging of the brain in newborns. Psychiatry Res. 132(1):81–85. Gulani, V., A. G. Webb, I. D. Duncan, and P. C. Lauterbur, 2001. Apparent diffusion tensor measurements in myelin-deficient rat spinal cords. Magn. Reson. Med. 45(2):191–195. Haselgrove, J. C., and J. R. Moore, 1996. Correction for distortion of echo-planar images used to calculate the apparent diffusion coefficient. Magn. Reson. Med. 36(6):960–964. Heiervang, E., T. E. Behrens, C. E. Mackay, M. D. Robson, and H. Johansen-Berg, 2006. Between-session reproducibility and between-subject variability of diffusion MR and tractography measures. NeuroImage 33(3):867–877. Hermoye, L., C. Saint-Martin, G. Cosnard, S. K. Lee, J. Kim, M. C. Nassogne, et al., 2006. Pediatric diffusion tensor imaging: Normal database and observation of the white matter maturation in early childhood. NeuroImage 29(2):493–504. Hess, C. P., P. Mukherjee, E. T. Han, D. Xu, and D. B. Vigneron, 2006. Q-ball reconstruction of multimodal fiber orientations using the spherical harmonic basis. Magn. Reson. Med. 56(1):104– 117. Huang, H., J. Zhang, S. Wakana, W. Zhang, T. Ren, L. J. Richards, et al., 2006. White and gray matter development in human fetal, newborn and pediatric brains. NeuroImage 33(1): 27–38. Huppi, P. S., S. Warfield, R. Kikinis, P. D. Barnes, G. P. Zientara, F. A. Jolesz, et al., 1998. Quantitative magnetic resonance imaging of brain development in premature and mature newborns. Ann. Neurol. 43(2):224–235. Inder, T. E., and P. S. Huppi, 2000. In vivo studies of brain development by magnetic resonance techniques. Ment. Retard. Dev. Disabil. Res. Rev. 6(1):59–67. Jezzard, P., A. S. Barnett, and C. Pierpaoli, 1998. Characterization of and correction for eddy current artifacts in echo planar diffusion imaging. Magn. Reson. Med. 39(5):801–812. Johansen-Berg, H., and T. E. Behrens, 2006. Just pretty pictures? What diffusion tractography can add in clinical neuroscience. Curr. Opin. Neurol. 19(4):379–385. Jones, D. K., 2004. The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: A monte carlo study. Magn. Reson. Med. 51(4):807–815. Kim, M., M. Ducros, T. Carlson, I. Ronen, S. He, K. Ugurbil, et al., 2006. Anatomical correlates of the functional organization in the human occipitotemporal cortex. Magn. Reson. Imaging. 24(5):583–590. Kinoshita, M., K. Yamada, N. Hashimoto, A. Kato, S. Izumoto, T. Baba, et al., 2005. Fiber-tracking does not accurately estimate size of fiber bundle in pathological condition: Initial neurosurgical experience using neuronavigation and subcortical white matter stimulation. NeuroImage 25(2):424–429. Klingberg, T., C. J. Vaidya, J. D. Gabrieli, M. E. Moseley, and M. Hedehus, 1999. Myelination and organization of the frontal white matter in children: A diffusion tensor MRI study. NeuroReport 10(13):2817–2821. Le Bihan, D., 2003. Looking into the functional architecture of the brain with diffusion MRI. Nature Rev. Neurosci. 4(6):469– 480. Le Bihan, D., E. Breton, D. Lallemand, P. Grenier, E. Cabanis, and M. Laval-Jeantet, 1986. MR imaging of intravoxel inco-

308

methodological paradigms

herent motions: Application to diffusion and perfusion in neurologic disorders. Radiology 161(2):401–407. Liston, C., R. Watts, N. Tottenham, M. C. Davidson, S. Niogi, A. M. Ulug, et al., 2006. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cerebral Cortex 16(4):553–560. Mabbott, D. J., M. Noseworthy, E. Bouffet, S. Laughlin, and C. Rockel, 2006. White matter growth as a mechanism of cognitive development in children. NeuroImage 33(3):936–946. Masutani, Y., S. Aoki, O. Abe, N. Hayashi, and K. Otomo, 2003. MR diffusion tensor imaging: Recent advance and new techniques for diffusion tensor visualization. Eur. J. Radiol. 46(1): 53–66. Miller, J. H., R. C. McKinstry, J. V. Philip, P. Mukherjee, and J. J. Neil, 2003. Diffusion-tensor MR imaging of normal brain maturation: A guide to structural development and myelination. AJR Am. J. Roentgenol. 180(3):851–859. Mori, S., and J. Zhang, 2006. Principles of diffusion tensor imaging and its applications to basic neuroscience research. Neuron 51(5): 527–539. Morriss, M. C., R. A. Zimmerman, L. T. Bilaniuk, J. V. Hunter, and J. C. Haselgrove, 1999. Changes in brain water diffusion during childhood. Neuroradiology 41(12):929–934. Moseley, M. E., Y. Cohen, J. Kucharczyk, J. Mintorovitch, H. S. Asgari, M. F. Wendland, et al., 1990. Diffusion-weighted MR imaging of anisotropic water diffusion in cat central nervous system. Radiology 176(2):439–445. Nagy, Z., H. Westerberg, and T. Klingberg, 2004. Maturation of white matter is associated with the development of cognitive functions during childhood. J. Cogn. Neurosci. 16(7):1227–1233. Neil, J. J., J. Miller, P. Mukherjee, and P. S. Huppi, 2002. Diffusion tensor imaging of normal and injured developing human brain—A technical review. NMR Biomed. 15(7–8):543– 552. Neil, J. J., S. Shiran, R. McKinstry, G. Schefft, A. Snyder, C. Almli, et al., 1998. Normal brain in human newborns: Apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology 209(1):57–66. Nomura, Y., H. Sakuma, K. Takeda, T. Tagami, Y. Okuda, and T. Nakagawa, 1994. Diffusional anisotropy of the human brain assessed with diffusion-weighted MR: Relation with normal brain development and aging. AJNR Am. J. Neuroradiol. 15(2):231– 238. Ozarslan, E., and T. H. Mareci, 2003. Generalized diffusion tensor imaging and analytical relationships between diffusion tensor imaging and high angular resolution diffusion imaging. Magn. Reson. Med. 50(5):955–965. Perrin, M., C. Poupon, B. Rieul, P. Leroux, A. Constantinesco, J. F. Mangin, et al., 2005. Validation of q-ball imaging with a diffusion fibre-crossing phantom on a clinical scanner. Philos. Trans. R. Soc. London [Biol.] 360(1457):881–891. Prayer, D., A. J. Barkovich, D. A. Kirschner, L. M. Prayer, T. P. Roberts, J. Kucharczyk, et al., 2001. Visualization of nonstructural changes in early white matter development on diffusion-weighted MR images: Evidence supporting premyelination anisotropy. AJNR Am. J. Neuroradiol. 22(8):1572–1576. Ramnani, N., T. E. Behrens, W. Penny, and P. M. Matthews, 2004. New approaches for exploring anatomical and functional connectivity in the human brain. Biol. Psychiatry 56(9):613–619. Rohde, G. K., A. S. Barnett, P. J. Basser, S. Marenco, and C. Pierpaoli, 2004. Comprehensive approach for correction of motion and distortion in diffusion-weighted MRI. Magn. Reson. Med. 51(1):103–114.

Rutherford, M. A., F. M. Cowan, A. Y. Manzur, L. M. Dubowitz, J. M. Pennock, J. V. Hajnal, et al., 1991. MR imaging of anisotropically restricted diffusion in the brain of neonates and infants. J. Comput. Assist. Tomogr. 15(2):188–198. Sakuma, H., Y. Nomura, K. Takeda, T. Tagami, T. Nakagawa, Y. Tamagawa, et al., 1991. Adult and neonatal human brain: Diffusional anisotropy and myelination with diffusion-weighted MR imaging. Radiology 180(1):229–233. Schmithorst, V. J., M. Wilke, B. J. Dardzinski, and S. K. Holland, 2002. Correlation of white matter diffusivity and anisotropy with age during childhood and adolescence: A crosssectional diffusion-tensor MR imaging study. Radiology 222(1): 212–218. Snook, L., L. A. Paulson, D. Roy, L. Phillips, and C. Beaulieu, 2005. Diffusion tensor imaging of neurodevelopment in children and young adults. NeuroImage 26(4):1164–1173. Song, S. K., S. W. Sun, W. K. Ju, S. J. Lin, A. H. Cross, and A. H. Neufeld, 2003. Diffusion tensor imaging detects and differentiates axon and myelin degeneration in mouse optic nerve after retinal ischemia. NeuroImage 20(3):1714–1722. Song, S. K., S. W. Sun, M. J. Ramsbottom, C. Chang, J. Russell, and A. H. Cross, 2002. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. NeuroImage 17(3):1429–1436.

Song, S. K., J. Yoshino, T. Q. Le, S. J. Lin, S. W. Sun, A. H. Cross, et al., 2005. Demyelination increases radial diffusivity in corpus callosum of mouse brain. NeuroImage 26(1):132–140. Stejskal, E. O., and J. E. Tanner, 1965. Spin diffusion measurements: Spin echos in the presence of time-dependent field gradient. J. Chem. Phys. 42:288–292. Sun, S. W., H. F. Liang, K. Trinkaus, A. H. Cross, R. C. Armstrong, and S. K. Song, 2006. Noninvasive detection of cuprizone-induced axonal damage and demyelination in the mouse corpus callosum. Magn. Reson. Med. 55(2):302– 308. Tournier, J. D., F. Calamante, D. G. Gadian, and A. Connelly, 2004. Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution. NeuroImage 23(3):1176–1185. Tuch, D. S., 2004. Q-ball imaging. Magn. Reson. Med. 52(6):1358– 1372. Tuch, D. S., T. G. Reese, M. R. Wiegell, and V. J. Wedeen, 2003. Diffusion MRI of complex neural architecture. Neuron 40(5):885–895. Wimberger, D. M., T. P. Roberts, A. J. Barkovich, L. M. Prayer, M. E. Moseley, and J. Kucharczyk, 1995. Identification of “premyelination” by diffusion-weighted MRI. J. Comput. Assist. Tomogr. 19(1):28–33.

wozniak, mueller, and lim: diffusion tensor imaging

309

18

Functional MRI Methods in Developmental Cognitive Neuroscience KATHLEEN M. THOMAS AND ANGELA TSENG

Philosophers have long pondered intricacies of the mindbody problem. An essential component of human nature is our ability to internally, and often covertly, represent a complex system of thoughts and emotional processes. A lingering mystery is how these phenomena can be instantiated in the biological system of the human body. The fields of cognitive and affective neuroscience have emerged in response to such questions regarding the relationship between thought or behavior and the brain. Over the past two decades, substantial progress has been made in addressing these philosophical and conceptual issues with the development of new methods for assessing brain function. Although physiologists and psychologists have long used biological signals such as heart rate or galvanic skin response as downstream indices of behavioral state changes, the advent of methods to assess changes in brain metabolism and energy consumption have forever altered our understanding of the relationship between cognition and brain. Techniques like positron emission tomography (PET) have demonstrated that cognitive and behavioral events are associated with regionally localized changes in glucose consumption and blood flow (Fox and Raichle, 1984; Pardo, Fox, and Raichle, 1991). The extension of magneticresonance-imaging techniques to questions of temporally limited changes in energy utilization have opened the door for noninvasive measures of brain-behavior relations with both spatial and relative temporal precision. Functional magnetic resonance imaging (MRI) has emerged as a primary method for examining cognitive neuroscience questions. Over the past two decades, hundreds of papers have appeared in the literature that have used this technique to address the brain regions engaged during specific cognitive processes and to dissect high-level cognitive computations into smaller elements with specific neural correlates. However, these issues gain a significant degree of complexity when development is added to the picture. Elegant work from developmental neurobiology and related basic neuroscience disciplines demonstrates the complex set of processes that occur during embryologic

brain development (Lumsden and Kintner, 2003; Monk, Webb, and Nelson, 2001). These mechanisms are not hardwired. That is, the outcome at each stage is necessarily affected not only by the genetic code of the developing organism, but also by the influence of local and distal environmental factors during the developmental process. Factors such as nutrition or temperature as well as cell-to-cell interactions can significantly alter the course of development, with both positive and negative consequences (Levitt, 2003). Similar mechanisms are assumed to occur at a macro level in human behavioral development, such that both genetic and environmental influences, as well as the interaction of the two, are essential in determining the long-term cognitive and emotional development of a child and, ultimately, the adult. The field of developmental cognitive neuroscience addresses the developmental course of specific brain-behavior relationships across the human life span. Importantly, functional MRI methods are proving to be a useful tool in approaching these developmental questions. The added complexity of change over time renders developmental research particularly challenging. However, with a consideration of the specific hurdles posed by developmental questions, as well as the limitations of our methods, functional MRI may prove to be a critical technique for exploring this exciting field. In this chapter we introduce a basic description of functional MRI methods as well as necessary considerations in the application of these methods to developmental questions and pediatric populations. This introduction includes an overview of the principles underlying the magnetic resonance signal, the design of appropriate behavioral paradigms in fMRI research, and a consideration of issues specific to research with child and/or clinical populations. The second portion of the chapter addresses selected examples from the existing developmental fMRI literature to illustrate the current state of knowledge in our field, including progress thus far as well as recommendations for the future. This overview includes example domains in which significant fMRI work has begun, but it should not be viewed as a

311

review, comprehensive or otherwise, of the developmental neuroimaging literature. Our goal is to highlight major themes in the area and identify some of the remaining gaps in our use of this methodology.

Basic principles of functional magnetic resonance imaging (fMRI) The Physics of Magnetic Resonance Imaging Magnetic resonance imaging (MRI) is a tomographic imaging technique based on the principles of nuclear magnetic resonance (NMR), a spectroscopic method used to study structural and dynamical properties of atoms and molecules (Jezzard and Clare, 2001). At its inception, MRI was actually denoted as nuclear magnetic resonance imaging. However, in order to circumvent the negative, albeit erroneous, association with ionizing radiation exposure, the term “nuclear” has been commonly removed. Because MRI is effectively noninvasive, this methodology allows for the in vivo study of the anatomy, function, and metabolism of the living human brain. In essence, MRI utilizes the interaction between radio frequency pulses, a strong magnetic field, and tissue properties to acquire images of planes from inside the body (McRobbie et al., 2003). Magnetic resonance imaging capitalizes on the observation that the molecules and atoms that compose the body are sensitive to the magnetic properties of their environment. In the presence of an ambient magnetic field, the constantly spinning and moving atomic particles tend to become highly organized and assume specific behavioral states. For example, protons like the one comprising the hydrogen nucleus possess a property called spin, which can be thought of as a small local magnetic field that causes the nucleus to produce an NMR signal. Protons naturally align their spins parallel to the direction of the magnetic field. Magnetic resonance imaging entails ascertaining this known behavioral state and then continually perturbing the nuclei. Both the time required for the protons to return to the steady state and the energy alterations associated with this shift are then measured (Purcell, Torrey, and Pound, 1946). Since the human body is comprised primarily of fat and water, which both contain hydrogen, the human body is approximately 63 percent hydrogen atoms. Given the prevalence and properties of hydrogen atoms, magnetic resonance imaging in humans primarily images the NMR signal from hydrogen nuclei (Huettel, Song, and McCarthy, 2004; Mansfield and Pykett, 1978). The magnetic resonance scanner supplies a strong magnetic field of uniform strength and direction, termed a static magnetic field. The intensity of this field varies with the MRI scanner, but commonly ranges in human studies from 1 to 4 tesla (T) in magnetic field strength. Clinical MRI scanners frequently have static magnetic fields of 1.5 T,

312

methodological paradigms

whereas many research institutions now use higher field magnets, such as 3.0 or 4.0 T or even stronger fields, for research imaging. Higher magnetic fields yield greater sensitivity to signal changes with behavior but may incur additional complications as well (Voss, Zevin, and McCandliss, 2006). Considerable energy is necessary to initially generate this static field; however, the use of superconducting materials allows the field to sustain itself with no significant decay for long periods of time. Thus this magnetic field is omnipresent, even when the MRI scanner is not in use (Huettel, Song, and McCarthy, 2004). Images of the brain or other tissue compartments are created by applying a series of brief energy pulses, termed radio frequency (RF) pulses, through the tissue. If the frequency of the RF pulse corresponds to the resonant, or preferred, frequency of the tissue molecules, energy is absorbed into the molecule, momentarily disrupting or tilting the protons out of the aligned or steady state induced by the static magnetic field. Measurements of the resultant change from one energy state to another (termed relaxation) are used to create images of the tissue. Different body tissue types demonstrate different electromagnetic properties or relaxation times, with lipids showing longer relaxation times than water (Bloch, 1946; Hahn, 1950). These relaxation differences can be mapped to produce structural images of the tissue, for example, by emphasizing the difference between brain regions high in water compared to regions high in fat. Visually dissimilar images of the same tissue can be produced by alternate indices of relaxation and different sequences of RF pulses. For example, so-called T1-weighted or T2-weighted images emphasize different measures of relaxation. In T1-weighted images, cerebrospinal fluid appears dark because it shows very low signal using this measure of relaxation (figure 18.1A). However, in T2-weighted images, cerebrospinal fluid appears bright or white, indicating a high signal using this measure (figure 18.1B). The level of contrast that is available in the images is determined by these differences in measurement technique or the electromagnetic pulse sequence (Buxton et al., 1987). Physiological Basis of fMRI Importantly, in addition to structural information, MRI techniques can also be used to assess functional brain activity. Functional MRI is contingent on the assumption that a change in brain activity is associated with a local hemodynamic change in blood flow and oxygenation. That is, activity of neurons in a particular brain region is assumed to be correlated with local increases in blood flow and blood oxygenation. Although the mechanism is not entirely clear, the increase in blood flow is greater than the oxygen demand of the active neurons (Fox et al., 1988; Raichle, 1987), resulting in a local increase in oxygenated compared to deoxygenated hemoglobin.

A

B

Figure 18.1 Structural magnetic resonance images of the brain. (A) T1-weighted images of the brain at three axial slice locations from inferior to superior. (B) T2-weighted images of the brain at the same three axial slice locations presented in figure 18.1A.

The most frequent type of functional MRI, bloodoxygenation-level-dependent (BOLD) imaging, is based on the observation that oxygenated and deoxygenated hemoglobin behave differently when pulsed by an RF signal (Ogawa et al., 1990; Turner et al., 1991; Kwong et al., 1992). Deoxygenated hemoglobin is paramagnetic and causes distortions in the local magnetic field, leading the spinning protons in that region to drop out of phase with one another. That is, the various protons spin at different rates. This change in energy state from the in-phase to the out-of-phase spin states (called T2* relaxation) can then be mapped to indicate the relative neuronal activity across the brain, as indexed by changes in blood oxygenation. Highly oxygenated areas of the brain will show stronger MR signal (less disruption or inhomogeneity) in T2*-weighted images than less oxygenated regions. Magnetic resonance signals may be mapped across the brain to provide an indirect measure of regional differences in neuronal activity during the activated state. This hemodynamic or blood oxygenation change is not instantaneous. The observed signal change is delayed in time from the presumed neuronal activity. In fact,

the peak measurable signal change would be expected to occur as late as 5 to 6 seconds after the presumed eventrelated neuronal response (Miezin et al., 2000). The signal change observed from the activated to the deactivated state is quite small when using this BOLD contrast (e.g., approximately 1% change in a 1.5-tesla MRI scanner), and therefore the subtle signal differences expected between behavioral states of the individual are even more difficult to detect. Figure 18.2A illustrates a T2*-weighted image acquired during a single acquisition of BOLD fMRI scanning. The hemodynamic signal is apparent throughout the brain; however, regional differences in signal strength are expected depending on behavioral state and activity. This activity difference is measured across multiple exemplars of each activity state (e.g., motor activity versus rest) to increase the power to detect small differences in an MR response. Significant regions of differential activity are assessed statistically after completion of the entire scanning session, resulting in the colorful statistical maps of activity that are more commonly presented in published fMRI studies (figure 18.2B). The hemodynamic signal change can

thomas and tseng: functional mri methods

313

A

B

Figure 18.2 Functional magnetic resonance images of the brain. (A) T2*-weighted blood-oxygenation-level-dependent (BOLD) magnetic resonance signal overlaid on T1-weighted structural image at the same axial slice location. (B) Regions demonstrating a statisti-

cally significant difference in activation for an experimental condition relative to a baseline condition. Color changes indicate the magnitude of the F-statistic. Only pixels showing an F-value significant at p = 0.005 or better are displayed. (See plate 43.)

be amplified by increasing the strength of the ambient magnetic field (e.g., 3.0-T or 4.0-T static field). However, increased field strength is also associated with increased artifacts and other potential confounds in functional MRI (Huettel, Song, and McCarthy, 2004). (See also plate 43.)

inhomogeneous. This type of artifact presents a significant difficulty near the borders between tissue and air—for example, around the edges of the brain and near the sinus cavities. Such borders present an abrupt change in magnetic resonance signal from the high signal values of the tissue to the immediately adjacent low signal value of the air, making the boundary difficult to measure. Susceptibility artifact appears as signal loss on the functional scan and makes some brain regions more difficult to image than others. The orbital frontal cortex and medial temporal poles can be particularly difficult to image given their location near the sinus cavities. Additional sources of signal artifact include the mechanical motion of the chest during respiration and blood flow through draining veins, which in surface features may be indistinguishable from the hemodynamic response elicited by neuronal activity (Jezzard, 1999). These sources of noise are unavoidable but may be measured to assess their impact on the concurrent functional data of interest.

Sources of Signal Artifact As with any type of measurement, the quality of the obtained magnetic resonance images is determined predominantly by the signal-to-noise ratio of the data. Although noise is determined in part by the physical scanner that is used, both conventional MRI and functional MRI are also subject to data artifacts produced by subject movement and physiological sources (Brammer, 2001). The degree of allowable subject motion is quite small (on the order of 1–3 mm). Initially this standard may appear impossible to meet, and indeed, head motion represents the primary source of artifact for functional brainimaging research, particularly in pediatric populations, as child participants tend to move more than adult participants. However, a decade of pediatric functional imaging research has now demonstrated that with a sufficiently comfortable environment and some creative head-restraint techniques, healthy developing children as young as 5–6 years of age can be successfully scanned in the awake state using fMRI (Thomas and Casey, 2003). Functional MRI has the additional problem of physiological artifacts. While these include motion induced by cardiac pulse and respiration, one of the most insidious is magnetic susceptibility artifact. This artifact appears as a loss of functional signal in regions in which the magnetic field is

314

methodological paradigms

Safety Considerations for fMRI Studies Magnetic resonance imaging is an extremely safe medical and research procedure. However, as with any medical procedure, one must be aware of the specific risks involved before determining whether MRI is the most appropriate method to use. The most basic safety concerns relate to the strong static magnetic fields involved in MR imaging. Individuals with metal in or on their bodies are at risk of injury in the MRI environment (Kanal et al., 2004). For example, the strong magnetic field may cause an implanted device to be dislodged or pulled through tissue, or it may cause a metal device to become

heated during the imaging process, resulting in a burn. Additionally, the MR environment must be completely protected to prevent individuals from inadvertently introducing metal objects that could become projectiles in the face of the strong static field. Although these are very serious safety concerns, all reputable MR imaging facilities have safety measures in place, including careful screening of participants and standardized procedures for staff training. Other less dramatic safety concerns arise from the imaging sequences used in scanning. The rapid image acquisition during functional scanning involves the application of gradient magnetic fields to allow for spatial coding in the image. The rapid oscillation of these magnetic gradients produces a loud auditory artifact. For example, one common functional imaging pulse sequence, echo planar imaging (EPI), produces a loud, high-pitched, repetitive beeping in its most common variants (Huettel, Song, and McCarthy, 2004). Subjects are provided with headphones or earplugs to attenuate this noise, but depending on the type of scan and the gradients used, the noise intensity could be harmful to hearing. In certain configurations, rapid gradient oscillations also can induce electrical currents in the body, producing peripheral nerve stimulation, and absorption of radio frequency energy can lead to elevated body temperature and tissue heating. However, guidelines exist for determining the acceptable specific absorption rate (SAR) for different subject populations and the maximal rate and amplitude of gradient field changes to avoid these problems. Such rapid field changes can also induce heating in otherwise safe materials in the scanner environment. For example, nonferrous metals such as aluminum, which do not pose a projectile risk in the static magnetic field, may nonetheless be dangerous as they could heat to the point of burning. Other materials may show similar behavior, including some inks used in tattoos (Kanal et al., 2004). Behavioral Paradigms in fMRI Research A critical component of any functional imaging method is the behavioral probe task used to elicit activity. Even the most sophisticated scanning environment will produce low-quality data if the brain is not stimulated in a meaningful way. Experimental task design is critical here, allowing the variable of interest to be captured. However, elegant experimental paradigms are not sufficient in and of themselves. Any behavioral paradigm relies on the assumption that manipulations of the variables of interest will be associated with measurable differences in brain activity. Given the subtle manipulations common in psychological research, this assumption deserves careful consideration when designing behavioral paradigms. A multitude of behavioral paradigms have been successfully utilized in adult imaging research. Nevertheless, it is

often insufficient to simply transfer these tasks for use with pediatric populations. While school-age children are capable of a wide range of behavioral repertoires, they have yet to reach adult levels of performance in many areas. For example, behavioral observation has shown that children often use visual checking methods (i.e., looking down at the button box and their hand) to insure that they are correctly responding in a behavioral task (Thomas and Casey, 1999). Yet, when lying prone in the scanner, it is difficult to look down at one’s hands without producing considerable movement artifact. Moreover, fine motor skills are still developing in the early school years, and tasks requiring multiple or complicated button presses may not be feasible. Thomas and colleagues (1999) found that young subjects do not always perform as well behaviorally in the scanner as they do outside the imaging environment. While adults demonstrated behavioral improvements in spatial working memory with time in the scanner (or time on task), children showed an increase in behavioral errors as well as an increase in movement artifact with time in the scanner (Thomas and Casey, 2003). Clearly, it is essential to employ age-appropriate tasks that allow young participants to attain adequate levels of performance. Children may become discouraged or frustrated with tasks that are too difficult or not sufficiently engaging, affecting both the cognitive processes of interest and the attrition rate of the study. The overall length and structure of the scanning session are equally important factors, as attention tends to wane more quickly in children than in most adults. Children must be able to maintain focus long enough to allow collection of ample data. Our experience with child participants suggests that behavioral errors as well as head and body movement increase significantly after 45– 60 minutes in the scanner. These developmental considerations, along with factors such as the number of conditions to be tested and the number of data points required for sufficient statistical power, are critical in effective study design (Thomas and Casey, 2003). As discussed, elegant experimental task design allows subtle manipulation of the variables of interest while minimizing potential confounds. One such source of confound is participant anxiety in the scanner environment. Many people experience some level of discomfort or anxiety in unfamiliar settings, particularly when being tested or when performance is measured. A critical part of the fMRI scanning process is to help alleviate or circumvent potential anxiety-related brain activity by creating a comfortable and supportive environment. In work with children, this process may be as simple as providing toys and activities to place the child at ease and provide distraction while parents complete consent and other necessary paperwork. The MRI setting should enable the children to relax and spend time with a researcher prior to the testing session. Researchers must be

thomas and tseng: functional mri methods

315

attentive to signs of anxiety throughout the testing procedure and continue to subsequent testing levels only when the child appears comfortable with the current activity. Given the reliance on hemodynamic responses, increased participant anxiety would be expected to activate anxiety-related regions of the brain during performance of the intended cognitive probe task. Such anxiety-related activity will be impossible to separate from the probe task activity if anxiety varies as a function of task demands (Huettel, Song, and McCarthy, 2004). As neuroimaging research exposes participants to an extremely novel setting, including loud and potentially startling equipment, taking adequate steps to familiarize participants with the environment is critical. Simulated MRI scanners have become an invaluable tool for the successful acclimation of child participants (Rosenberg et al., 1997). Simulators can provide an early indicator of anxiety or claustrophobia that may be experienced in the tunnellike bore of the magnet. Simulation sessions offer a less intimidating way to introduce a participant to the sensory experiences of the scanner setting, as well as providing the child (and parents) with an opportunity to ask questions about the scanning procedure. The simulator experience can also be utilized for training participants on a behavioral task and can be used as an opportunity to provide feedback about head and body movement in the scanner. Both MRI and fMRI are sensitive to artifacts produced by participant motion. Even minor movements of the head and neck can drastically decrease image quality. Understandably, these artifacts are more likely when working with young children. Several methods used to minimize motion in adult subjects have also been shown to be effective with children (Thomas and Casey, 1999). These methods may involve head restraints (e.g., pillows, foam padding, head and chin straps or bite bars) and intermittent reminders to minimize motion (e.g., “Remember to lie as still as a statue!”). Training techniques have also been employed to decrease motion by helping children recognize when they are moving. In such paradigms, children are presented with a movie or other visual stimulus while in the simulator to hold their attention and help distract them from scanner noise and the like. A motion-detection system in the head coil can be connected to the video presentation system such that a disruption of the video display occurs in response to excessive head motion. By gradually reducing the acceptable range of motion, behavior can be shaped such that children learn to remain still to avoid disruptions in the video presentation. In our own laboratory, simulation and behavioral training experiences have been very successful in reducing attrition due to behavioral noncompliance, motion artifact, or early scan termination. Given the enclosed nature of the magnet, behavioral monitoring of the subjects by simple observation is not

316

methodological paradigms

practical. Yet, not infrequently, participants may fall asleep during the scan if left to their own devices. However, interaction with the participant can be difficult during data collection. This potential problem is enhanced if researchers cannot monitor behavioral performance during the scanning session. The inclusion of some form of behavioral response from the subject during scanning, even during otherwise passive viewing tasks, may be beneficial to ensure that the subject is actually attending to and performing the tasks as directed. Online performance monitoring can allow for correction of simple problems, such as a child having her fingers on the wrong response buttons or inadvertently reversing the task instructions. In addition, changes in behavioral performance across the scanning session may be indicative of altered comfort in the scanner environment or may signal an unspoken withdrawal of consent in child participants.

Functional MRI studies of developmental cognitive neuroscience History of Developmental fMRI Research Functional MRI techniques were first applied to developmental populations only a decade ago. In an initial pediatric fMRI report, Hertz-Pannier and colleagues (1995, abstract) examined language-related brain activity in children and adolescents to assess whether similar laterality effects were observed between adults and children. This study reported more diffuse and less lateralized patterns of activity in the child sample. However, this sample was largely made up of children with chronic epilepsy or other medical issues and cannot be considered normative. In the same year, Casey and colleagues (1995) published the first pediatric fMRI paper, examining the brain regions engaged during verbal working memory functions in 9- to 11-year-old typically developing children. This study was pioneering in several key ways. First, it demonstrated that hemodynamic responses could be measured in school-age children in essentially the same manner as measured in adults. Although this may seem an obvious point, the technique had never been used with children, and it was not clear whether the hemodynamic functions of the child brain would be sufficiently similar to the mature brain to be measured by this technique. Research since that time has demonstrated that the onset and peak activity of the hemodynamic response is essentially equivalent to adult functions at least as young as 7 years of age (Richter and Richter, 2003). Several studies suggest that this response function may differ for infants in the first postnatal year of life (e.g., Born et al., 1996, 1998; Martin et al., 1999), perhaps even showing an inverted form. However, adultlike hemodynamic responses have been observed in at least some brain systems in children as young as a few months of age. For example, DeHaene-Lambertz, DeHaene, and

Hertz-Pannier (2002) demonstrated evidence of lefthemisphere activity in response to natural language stimuli in 2- to 3-month-old nonsedated infants. In this case, models of the hemodynamic response function appeared to mimic adult responses. Studies examining the utility of functional MRI methods in understanding infant perceptual and cognitive development are becoming more common under the leadership of researchers such as DeHaene-Lambertz and others in France (e.g., Dehaene-Lambertz, Hertz-Pannier, and Dubois, 2006; Dehaene-Lambertz et al., 2006). The results of the Casey and colleagues (1995) experiment demonstrated that, using a cognitive task developed for adult participants, school-age children showed local changes in hemodynamic activity that were similar in location to previous adult findings. That is, verbal working memory as measured by an N-back working-memory task (e.g., “Does the current stimulus match the one 2 trials back?”) was associated with increased activity in the dorsal lateral prefrontal cortex relative to activity during a visuomotor control task. This pattern was similar to results observed in adult participants on the same task (Cohen et al., 1994). Another important advance of this study was the demonstration that such methods are feasible in healthy children. These children clearly were able to perform the cognitive task with sufficient accuracy, even while lying in an MRI scanner, and were able to lie still enough to produce reliable images of functional brain activity. Previous work had already demonstrated that structural MRI measures were reasonable with typical pediatric populations, even when sedation was not used (e.g., Giedd et al., 1994; Castellanos et al., 1994; longitudinal data in Giedd et al., 1999). However, this study was the first to test the feasibility of functional MRI, a technique that is extremely sensitive to motion-induced artifact. Finally, this initial cognitive investigation by Casey and colleagues suggested that, despite obvious differences in overall cognitive level between adults and children, the brain regions activated during specific aspects of cognitive function (in this case, verbal working memory) may be quite similar across developmental periods. This result may be even more significant for the question that it does not address than the one it does. That is, if the regions of brain activity are similar across development, then what accounts for observed developmental differences in behavior? Developmental fMRI research is constantly challenged by such issues. Viewed more broadly, developmental fMRI studies must tackle the problem of dissociating performance differences from maturational differences. If two groups of adults are performing differentially on the same cognitive task, it is possible that we would attribute these differences to maturational effects, but more likely that we would attribute them to individual differences among the two groups (e.g., IQ or attentional abilities). A recurring challenge in the developmental

literature is to define the meaning of “development” in each context. Whereas some researchers use the term to reflect maturational biological changes, others use the term “development” to refer to changes in cognitive structure or behavior that occur over time. Functional MRI studies will require different designs and interpretations depending on the nature of the developmental question. It may be important to demonstrate cases of equivalent brain activity in the presence of behavioral differences, as well as the opposite case of equivalent behavior accompanied by differential patterns of brain activity. Brain Development and f MRI Since 1995 the number of published fMRI studies of typical cognitive or emotional development has increased dramatically. These papers have addressed the brain systems engaged during such diverse behavioral functions as memory (Menon, Boyett-Anderson, and Reiss, 2005; Scherf, Sweeney, and Luna, 2006), cognitive and inhibitory control (Casey et al., 2004; Davidson et al., 2006), emotion recognition (Hare et al., 2005; Thomas, Drevets, Whalen, et al., 2001), and language processing and reading (Chou et al., 2006; Szaflarski et al., 2006). Two general trends have emerged among the many findings that are specific to each topic area or behavioral paradigm. First, when comparing younger children to older children or adults, activity in cortical regions, particularly the prefrontal cortex, appears to become more focal with age. That is, younger children demonstrate more diffuse activity, including activation of multiple distinct cortical regions relative to adults performing the same task. An example of this age effect was observed by Casey and colleagues (1997) in a study of inhibitory control. Children ages 7–12 years and adults performed a go-no-go task in the scanner environment. The task required participants to press a response button as quickly as possible for letter stimuli presented on the screen. However, on 25 percent of trials, a no-go stimulus was presented (the letter X). Participants were instructed to withhold responses to the no-go stimulus. The high percentage of “go” trials insured that participants accrued a general tendency to respond. This study and others (e.g., Casey et al., 2001; Crone, Wendelken, et al., 2006; Crone, Donohue, et al., 2006) have emphasized a role for the prefrontal cortex in inhibitory control. In the Casey and colleagues (1997) study, children showed a larger volume of activity in prefrontal cortex than adults did. Specifically, this age difference was limited to dorsolateral regions of prefrontal cortex. While the two age groups showed similar activity in ventral prefrontal cortex, children showed extended activity in dorsolateral regions (figure 18.3). This increased activity may reflect less specificity of engagement in the younger age group, or perhaps increased activity as a function of task difficulty. In fact, children made twice as many false alarms as adults (incorrectly pressing for the X). However, the

thomas and tseng: functional mri methods

317

Figure 18.3 In a study of inhibitory control, child participants showed more extensive regions of activity in the middle frontal gyrus than adult participants. (Adapted from Casey et al., 1997.)

children with the largest volumes of activation were not the ones with the highest error rates, but instead the ones who performed most similarly to adult participants. This relatively diffuse activity in younger compared to older participants has now been replicated across a number of studies and behavioral domains (see Durston et al., 2006). A second general finding across studies has been a shift from more subcortically supported behavior to more cortically driven behavior with age. While both adults and children tend to activate subcortical regions including the basal ganglia and thalamus as well as cortical regions such as dorsal and ventral prefrontal cortex during cognitive and motor performance, the relative activity in the two systems differs across age. In a study of implicit sequence learning (a form of unconscious or incidental memory), 7- to 11-yearold children showed significantly greater activity in the putamen than adults did, whereas adults showed significantly greater activity in motor cortex (Thomas et al., 2004). In fact, across subjects, the level of activation in one region was inversely related to activation in the other (figure 18.4). Similar differences in the relative activation of subcortical versus cortical regions have been observed in studies of cognitive control as well (e.g., Casey et al., 2002, 2004). Behavioral Development and f MRI An important goal in neuroimaging research is to identify the relationship between physiological changes and behavior. Biological measures of physiological state (heart rate, vagal tone, EEG, blood flow, etc.) are not very helpful if variability in the measure is unrelated to behavioral stimulation or behavioral performance. In fMRI studies, findings of differential brain activity are most informative when the level or extent of such activity varies with behavioral measures of cognitive function. For example, in the Casey and colleagues (1997) study of the go-no-go task discussed previously, many brain regions showed differential activation during inhibitory periods

318

methodological paradigms

Figure 18.4 During performance of an implicit motor sequencelearning task, the extent of activity in primary motor cortex was inversely related to the extent of activity in the putamen, suggesting a trade-off in recruitment of these regions. (Adapted from Thomas et al., 2004.)

compared to “go” periods. However, only two brain regions showed activity that varied as a function of behavioral performance on the task. Signal intensity in the anterior cingulate gyrus (ACC) increased as a function of false alarm rate, such that children having the most difficulty on the task were engaging the ACC to the greatest extent (figure 18.5a). In contrast, the extent of activity in orbitofrontal cortex (OFC) was negatively related to behavioral performance. That is, the OFC was most active in children who were performing well and making very few inhibitory errors (figure 18.5b). The significant relationship between activity and behavioral performance lends support in interpreting the role of this activity in inhibitory processing. While other brain regions may be necessary to the cognitive process, we are unable assess the role of such regions in the overall behavior. The relationship between ACC activity and inhibitory errors reflects individual differences in brain activity and behavioral performance. However, one can also examine the role of various brain systems by probing within-subject signal variability. By parametrically varying the inhibitory load within individual trials, Durston and colleagues (2002) demonstrated that activity in the prefrontal cortex increased as a function of inhibitory demand in adult participants, even when few or no errors were made. Under conditions of high demand, activity in PFC was high, whereas under low demand, PFC signal was low. This parametric effect was not observed for children, who showed a high level of prefrontal activity for all inhibitory loads compared to “go” trials (Durston et al., 2003). Similar parametric manipulations have been applied to domains such as working-memory load in adult samples (e.g., Braver et al., 1997), as well as other aspects of cognitive control in both adults and children (Casey et al., 2000, 2001; Durston et al., 2003).

A

B

Figure 18.5 Changes in magnetic resonance signal as a function of behavioral performance on a test of inhibitory control. (A) The magnitude of the MR signal in the anterior cingulate gyrus was significantly positively correlated with the number of false alarms

To date, the majority of pediatric fMRI research has focused on describing patterns of developmental change and comparing brain activity across age groups. More recent research has begun to address the impact of individual differences on patterns of brain activity. In a study of brain responses to facial emotion, Thomas, Drevets, Dahl, and colleagues (2001) described group differences in the amygdala response to fearful faces in typically developing 8- to 15-year-old children and children with diagnoses of generalized anxiety disorder. Anxious children showed enhanced amygdala activity to fearful faces relative to neutral faces, but typically developing children showed the opposite pattern, with neutral faces eliciting the strongest amygdala response. These data provide evidence that the diagnostic category of generalized anxiety disorder is associated with altered amygdala function, but do not clarify whether particular behavior patterns or symptoms of the disorder are associated with the group difference in amygdala recruitment. Further analysis suggested that individual differences in reported everyday anxiety were significantly related to the amygdala response. That is, high levels of self-reported everyday anxiety were associated with a stronger amygdala response to fearful faces compared to neutral faces (figure 18.6). Children with low self-reported anxiety scores tended to show greater amygdala response to the neutral faces. Since the anxiety-disordered group scored higher on measures of everyday anxiety than the typically developing group, it is unclear whether additional features of the disorder contribute to the group differences in amygdala response, or whether individual reports of everyday anxiety are sufficient to account for variations in the amygdala response to fearful and neutral faces. Clearly, personality characteristics, temperament, IQ, or other individual differences may be strong predictors of particular patterns of brain activity.

committed in the go-no-go task. (B) The extent of activation in orbital frontal cortex was significantly negatively correlated with the number of go-no-go false alarms. (Adapted from Casey et al., 1997.)

Figure 18.6 Individual differences in temperament and behavior as predictors of MR signal differences. Activity differences in the right amygdala while viewing fearful relative to neutral facial expressions were positively correlated with self-reports of everyday anxiety in 8- to 15-year-old children and adolescents with or without anxiety disorders. (Adapted from Thomas, Drevets, Dahl, et al., 2001.)

Intervention or Treatment Effects Using f MRI A particularly promising arena for fMRI research is in the evaluation of intervention or treatment efforts. This technique has been applied not only in assessing the efficacy of both pharmacotherapy and behavioral interventions in psychiatric illness, but also in intervention settings to examine whether behavioral improvements following remediation are associated with specific changes in brain activity. In an important study of treatment for attention-deficit/ hyperactivity disorder (ADHD), Vaidya and colleagues (1998) examined the effects of stimulant medication on brain responses to a cognitive attention task. Employing the standard go-no-go paradigm used by Casey and colleagues

thomas and tseng: functional mri methods

319

(1997), Vaidya and colleagues examined group differences in recruitment of prefrontal cortex and basal ganglia systems in children with and without ADHD diagnoses. In the absence of pharmacological treatment (placebo condition), the ADHD group showed significantly less activation of basal ganglia regions than the typically developing group, accompanied by high numbers of inhibitory errors (incorrect presses to the X). When the two groups were tested under medication (methylphenidate treatment), the group with ADHD showed significant improvement in behavioral performance and an accompanying increase in basal ganglia activity during the inhibitory portions of the task. Medication treatment did not affect performance in the typically developing control group, although basal ganglia activity was decreased in this group under medication treatment. This pre-post intervention design may help clarify the mechanisms of action of various treatment approaches and may provide an additional quantitative measure for assessing treatment efficacy (e.g., percent signal change in a particular brain region). Functional MRI methods can also be used to assess change under nonpharmacological interventions. One active area of inquiry using this approach is in the remediation of reading disorders such as phonological dyslexia. For example, Shaywitz and colleagues (2004) provided two forms of reading intervention to children with reading disabilities. The experimental intervention was specially designed to improve phonological knowledge (understanding of the sound structures underlying spoken language) and increase awareness of the relationship between phonology and orthography (how spoken speech sounds relate to written letters and words). A second group of children received the standard interventions typically provided by schools. Both groups were compared to children without reading disabilities. Functional MRI scans were acquired during phonological processing both before and after the 1-year behavioral reading intervention. In the fMRI task, children were presented with an auditory letter (e.g., B) and then asked to choose the matching written letter (e.g., B or T). Difficulty was varied by providing similar or dissimilar choice stimuli (e.g., B T versus B K). Comparisons of fMRI data before and after intervention indicated that children in the experimental intervention group showed increased activity in left inferior frontal gyrus and left middle temporal gyrus (areas previously linked to phonological processing and reading fluency) relative to the group receiving the standard school-based reading intervention. This change in brain activity from pre- to postintervention was accompanied by a significant improvement in reading fluency in the experimental intervention group. Such approaches may advance studies of treatment efficacy by exploring whether interventions impact a global or specific mechanism of action.

320

methodological paradigms

Conclusions and future directions This chapter has provided a basic introduction to functional magnetic resonance imaging (fMRI) methods, and several illustrations of the value added by this technique when combined with other measures of cognitive development. Although the application of fMRI methods to pediatric populations and developmental questions is relatively new, the past decade has seen an explosion of published work in this area. Importantly, investigators have recognized that simple descriptions of the differences in brain activity between groups will be insufficient to advance our understanding of the neurological processes underlying cognitive and emotional function. A current challenge for developmental neuroimaging methods is to move beyond descriptions of behavioral change and closer to explanations of developmental change. As described in this overview, several approaches may facilitate this goal. First, fMRI studies must relate changes in brain activity to changes in behavior or individual differences. Group differences in brain activity (between developmental time points or between clinical and typically developing samples) may arise for a multitude of reasons, not all of which are related to the behavior of interest. It will be important to link brain differences to hypothesized mechanisms of developmental change. Second, a large portion of the developmental fMRI literature addresses activity differences between adults and children. Although this work provides an essential groundwork for relating pediatric studies to the existing adult literature, significant work is needed to examine cognitive and emotional changes within a developmental context, whether longitudinally in the same group of children or cross-sectionally. The inclusion of two age groups does not make a study “developmental,” particularly if the age groups are “adult” and “child.” Our ability to describe and explain developmental change would be enhanced by data sets allowing us to map a developmental trajectory across childhood and adolescence. Of course, it must be acknowledged that such large data sets incur equally large costs associated with functional MR scanning. Finally, as discussed in other chapters in this volume, the utility of fMRI methods may be significantly enhanced by combination with other measures of development. Clearly, well-defined cognitive or behavioral tasks are essential for fMRI research. In addition, converging evidence from other methodologies provides the strongest test of complex developmental hypotheses. An exciting direction for developmental cognitive neuroscience research is the combination of multiple imaging methods with elegant behavioral probes. For instance, combining the spatial resolution of fMRI measures with data from electrophysiological techniques such as event-related potentials (ERPs) could

enhance understanding of the temporal order of activity within various neural systems, as well as the overall time course of cognitive or behavioral events. Similarly, other magnetic resonance imaging measures, including structural measures like anatomical MRI or diffusion tensor imaging (see Chapter 17), can provide complementary information regarding brain maturation and organization. Advances in statistical analysis are also expanding the range of questions that can be addressed with functional MRI data, including questions of functional connectivity among brain regions or the role of genotypic differences in explaining variability in brain activity. These combined methodologies have the potential to advance understanding of developmental processes as never before. acknowledgments

This work was supported by a career development award to the first author (NIMH K01 MH02024) and a predoctoral traineeship to the second author (NICHD T32 HD007151). REFERENCES

Bloch, F., 1946. Nuclear induction. Phys. Rev. 70:460–474. Born, P., E. Rostrup, H. Leth, B. Peitersen, and H. C. Lou, 1996. Change of visually induced cortical activation patterns during development. Lancet 347:543. Born, P., H. Leth, M. J. Miranda, E. Rostrup, A. Stensgaard, B. Peitersen, H. B. Larsson, and H. C. Lou, 1998. Visual activation in infants and young children studies by functional magnetic resonance imaging. Pediatr. Res. 44(4):578–583. Brammer, M. J., 2001. Head motion and its correction. In P. Jezzard, P. M. Matthews, and S. M. Smith, eds, Functional MRI: An Introduction to Methods, 243–250. New York: Oxford. Braver, T. S., J. D. Cohen, L. E. Nystrom, J. Jonides, E. E. Smith, and D. C. Noll, 1997. A parametric study of prefrontal cortex involvement in human working memory. NeuroImage 5:49–62. Buxton, R. B., R. R. Edelman, B. R. Rosen, G. L. Wismer, and T. J. Brady, 1987. Contrast in rapid MR imaging: T1- and T2-weighted imaging. J. Comput. Assist. Tomogr. 11:7–16. Casey, B. J., J. D. Cohen, P. Jezzard, R. Turner, D. C. Noll, R. J. Trainor, J. Giedd, D. Kaysen, L. Hertz-Pannier, and J. L. Rapoport, 1995. Activation of prefrontal cortex in children during a nonspatial working memory task with functional MRI. NeuroImage 2(3):221–229. Casey, B. J., M. C. Davidson, Y. Hara, K. M. Thomas, A. Martinez, A. Galvan, J. M. Halperin, C. E. RodriguezAranda, and N. Tottenham, 2004. Early development of subcortical regions involved in non-cued attention switching. Dev. Sci. 7(5):534–542. Casey, B. J., S. D. Forman, P. Franzen, A. Berkowitz, T. S. Braver, L. E. Nystrom, K. M. Thomas, and D. C. Noll, 2001. Sensitivity of prefrontal cortex to changes in target probability: A functional MRI study. Hum. Brain Mapping 13:26– 33. Casey, B. J., K. M. Thomas, M. C. Davidson, K. Kunz, and P. L. Franzen, 2002. Dissociating striatal and hippocampal function developmentally with a stimulus-response compatibility task. J. Neurosci. 22(19):8647–8652.

Casey, B. J., K. M. Thomas, T. F. Welsh, R. D. Badgaiyan, C. H. Eccard, J. R. Jennings, and E. A. Crone, 2000. Dissociation of response conflict, attentional selection, and expectancy with functional magnetic resonance imaging. Proc. Natl. Acad. Sci. USA 97:8728–8733. Casey, B. J., R. J. Trainor, J. L. Orendi, A. B. Schubert, L. E. Nystrom, J. N. Giedd, F. X. Castellanos, J. V. Haxby, D. C. Noll, J. D. Cohen, S. D. Forman, R. E. Dahl, and J. L. Rapoport, 1997. A developmental functional MRI study of prefrontal activation during performance of a go-no-go task. J. Cogn. Neurosci. 9:835–847. Castellanos, F. X., J. N. Giedd, P. Eckburg, W. L. Marsh, A. C. Vaituzis, D. Kaysen, S. D. Hamburger, and J. L. Rapoport, 1994. Quantitative morphology of the caudate nucleus in attention deficit hyperactivity disorder. Am. J. Psychiatry 151(12):1791–1796. Chou, T. L., J. R. Booth, T. Bitan, D. D. Burman, J. D. Bigio, N. E. Cone, D. Lu, and F. Cao, 2006. Developmental and skill effects on the neural correlates of semantic processing to visually presented words. Hum. Brain Mapping 27(11):915–924. Cohen, J. D., S. D. Forman, T. S. Braver, B. J. Casey, D. Servan-Schreiber, and D. C. Noll, 1994. Activation of the prefrontal cortex in a nonspatial working memory task with functional MRI. Hum. Brain Mapping 1:293–304. Crone, E. A., S. E. Donohue, R. Honomichi, C. Wendelken, and S. A. Bunge, 2006. Brain regions mediating flexible rule use during development. J. Neurosci. 26(43):11239–11247. Crone, E. A., C. Wendelken, S. Donohue, L. van Leijenhorst, and S. A. Bunge, 2006. Neurocognitive development of the ability to manipulate information in working memory. Proc. Natl. Acad. Sci. USA 103(24):9315–9320. Davidson, M. C., D. Amso, L. C. Anderson, and A. Diamond, 2006. Development of cognitive control and executive functions from 4 to 13 years: Evidence from manipulations of memory, inhibition, and task switching. Neuropsychologia 44(11):2037– 2078. Dehaene-Lambertz, G., S. Dehaene, and L. Hertz-Pannier, 2002. Functional neuroimaging of speech perception in infants. Science 298(5600):2013–2015. Dehaene-Lambertz, G., L. Hertz-Pannier, and J. Dubois, 2006. Nature and nurture in language acquisition: Anatomical and functional brain-imaging studies in infants. Trends Neurosci. 29(7):367–373. Dehaene-Lambertz, G., L. Hertz-Pannier, J. Dubois, S. Meriaux, A. Roche, M. Sigman, and S. Dehaene, 2006. Functional organization of perisylvian activation during presentation of sentences in preverbal infants. Proc. Natl. Acad. Sci. USA 103(38): 14240–14245. Durston, S., M. C. Davidson, N. Tottenham, A. Galvan, J. Spicer, J. A. Fossella, and B. J. Casey, 2006. A shift from diffuse to focal cortical activity with development. Dev. Sci. 9(1):1–8. Durston, S., K. M. Thomas, M. S. Worden, Y. Yang, and B. J. Casey, 2002. The effect of preceding context on inhibition: An event-related fMRI study. NeuroImage 16(2):449–453. Durston, S., N. T. Tottenham, K. M. Thomas, M. C. Davidson, I. M. Eigsti, Y. Yang, A. M. Ulug, and B. J. Casey, 2003. Differential patterns of striatal activation in young children with and without ADHD. Biol. Psychiatry 53:871–878. Fox, P. T., and M. E. Raichle, 1984. Stimulus rate dependence of regional cerebral blood flow in human striate cortex, demonstrated by positron emission tomography. J. Neurophysiol. 51(5):1109–1120.

thomas and tseng: functional mri methods

321

Fox, P. T., M. E. Raichle, M. A. Mintun, and C. Dence, 1988. Nonoxidative glucose consumption during focal physiologic neural activity. Science 241:1445–1448. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, T. Paus, A. C. Evans, and J. L. Rapoport, 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2(10):861–863. Giedd, J. N., F. X. Castellanos, B. J. Casey, P. Kozuch, A. C. King, S. D. Hamburger, and J. L. Rapoport, 1994. Quantitative morphology of the corpus callosum in attention deficit hyperactivity disorder. Am. J. Psychiatry 151(5):665–669. Hahn, E. L., 1950. Spin echoes. Phys. Rev. 80:580–594. Hare, T. A., N. Tottenham, M. C. Davidson, G. H. Glover, and B. J. Casey, 2005. Contributions of amygdala and striatal activity in emotion regulation. Biol. Psychiatry 57(6):624– 632. Hertz-Pannier, L., W. D. Gaillard, S. Mott, S. Weinstein, J. Conry, W. H. Theodore, and D. Le Bihan, 1995 (abstract). Functional MRI of language tasks: Frontal diffuse activation patterns in children. Hum. Brain Mapping 1:S231. Huettel, S. A., A. W. Song, and G. McCarthy, 2004. Functional Magnetic Resonance Imaging. Sunderland, MA: Sinauer Associates. Jezzard, P., 1999. Physiological noise: Strategies for correction. In C. Moonen and P. A. Bandettini, eds., Medical Radiology: Functional MRI, 173–182. New York: Springer-Verlag. Jezzard, P., and S. Clare, 2001. Principles of nuclear magnetic resonance and MRI. In P. Jezzard, P. M. Matthews, and S. M. Smith, eds., Functional MRI: An Introduction to Methods, 67–92. New York: Oxford University Press. Kanal, E., J. P. Borgstede, A. J. Barkovich, C. Bell, W. G. Bradley, S. Etheridge, J. P. Felmlee, J. W. Froelich, J. Hayden, E. M. Kaminski, J. W. Lester, Jr., E. A. Scoumis, L. A. Zaremba, and M. D. Zinninger, 2004. American College of Radiology white paper on MR safety: 2004 update and revisions. AJR Am. J. Roentgenol. 182(5):1111–1114. Kwong, K. K., J. W. Belliveau, D. A. Chesler, I. E. Goldberg, R. M. Weisskoff, B. P. Poncelet, D. N. Kennedy, B. E. Hoppel, M. S. Cohen, R. Turner, H.-M. Cheng, T. J. Brady, and B. R. Rosen, 1992. Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc. Natl. Acad. Sci. USA 89:5675–5679. Levitt, P., 2003. Structural and functional maturation of the developing primate brain. J. Pediatr. 143(4 suppl):S35–S45. Lumsden, A., and C. Kintner, 2003. Neural induction and pattern formation. In L. R. Squire, F. E. Bloom, S. K. McConnell, J. L. Roberts, N. C. Spitzer, and M. J. Zigmond, eds., Fundamental Neuroscience, 2nd ed., 363–390. New York: Academic Press. Mansfield, P., and I. L. Pykett, 1978. Biological and medical imaging by NMR. J. Magn. Reson. 29:355–373. Martin, E., P. Joeri, T. Loenneker, D. Ekatodramis, D. Vitacco, J. Hennig, and V. L. Marcar, 1999. Visual processing in infants and children studied using functional MRI. Pediatr. Res. 46(2):135–140. McRobbie, D. W., E. A. Moore, M. J. Graves, and M. R. Prince, 2003. MRI: From Picture to Proton. Cambridge, MA: Cambridge University Press. Menon, V., J. M. Boyett-Anderson, and A. L. Reiss, 2005. Maturation of medial temporal lobe response and connectivity during memory encoding. Cogn. Brain Res. 25:379–385. Miezin, F. M., L. Maccotta, J. M. Ollinger, S. E. Petersen, and R. L. Buckner, 2000. Characterizing the hemodynamic response: Effects of presentation rate, sampling procedure, and

322

methodological paradigms

the possibility of ordering brain activity based on relative timing. NeuroImage 11(6, part 1):735–759. Monk, C. S., S. J. Webb, and C. A. Nelson, 2001. Prenatal neurobiological development: Molecular mechanisms and anatomical change. Dev. Neuropsychol. 19:211–236. Ogawa, S., T. S. Lee, A. S. Nayak, and P. Glynn, 1990. Oxygenation-sensitive contrast in magnetic responance image of rodent brain at high magnetic fields. Magn. Reson. Med. 26:68–78. Pardo, J. V., P. T. Fox, and M. E. Raichle, 1991. Localization of a human system for sustained attention by positron emission tomography. Nature 349(6304):61–64. Purcell, E. M., H. C. Torrey, and R. V. Pound, 1946. Resonance absorption by nuclear magnetic moments in a solid. Phys. Rev. 69:37–38. Raichle, M. E., 1987. Circulatory and metabolic correlates of brain functions in normal humans. In V. B. Mountcastle, F. Plum, and S. R. Geiger, eds., Handbook of Physiology: The Nervous System, Vol. 5, 643–674. Bethesda, MD: American Psychological Society. Richter, W., and M. Richter, 2003. The shape of the fMRI BOLD response in children and adults changes systematically with age. NeuroImage 20(2):1122–1131. Rosenberg, D. R., J. A. Sweeney, J. S. Gillen, J. Kim, M. J. Varanelli, K. M. O’Hearn, P. A. Erb, D. Davis, and K. R. Thulborn, 1997. Magnetic resonance imaging of children without sedation: Preparation with simulation. J. Am. Acad. Child Adolesc. Psychiatry 36(6):853–859. Scherf, K. S., J. A. Sweeney, and B. Luna, 2006. Brain basis of developmental change in visuospatial working memory. J. Cogn. Neurosci. 18(7):1045–1058. Shaywitz, B. A., S. E. Shaywitz, B. A. Blachman, K. R. Pugh, R. K. Fulbright, P. Skudlarski, W. E. Mencl, R. T. Constable, J. M. Holahan, K. E. Marchione, J. M. Fletcher, G. R. Lyon, and J. C. Gore, 2004. Development of left occipitotemporal systems for skilled reading in children after a phonologically-based intervention. Biol. Psychiatry 55(9):926–933. Szaflarski, J. P., V. J. Schmithorst, M. Altaye, A. W. Byars, J. Ret, E. Plante, and S. K. Holland, 2006. A longitudinal functional magnetic resonance imaging study of language development in children 5 to 11 years old. Ann. Neurol. 59(5): 796–807. Thomas, K. M., and B. J. Casey, 1999. Functional MRI in pediatrics. In C. Moonen and P. A. Bandettini, eds., Medical Radiology: Functional MRI, 513–523. New York: Springer-Verlag. Thomas, K. M., and B. J. Casey, 2003. Methods for imaging the developing brain. In M. de Haan and M. H. Johnson, eds., The Cognitive Neuroscience of Development, 19–41. East Sussex, UK: Psychology Press. Thomas, K. M., W. C. Drevets, R. E. Dahl, N. D. Ryan, B. Birmaher, C. H. Eccard, D. Axelson, P. J. Whalen, and B. J. Casey, 2001. Amygdala response to fearful faces in anxious and depressed children. Arch. Gen. Psychiatry 58(11):1057–1063. Thomas, K. M., W. C. Drevets, P. J. Whalen, C. H. Eccard, R. E. Dahl, N. D. Ryan, and B. J. Casey, 2001. Amygdala response to facial expressions in children and adults. Biol. Psychiatry 49:309–316. Thomas, K. M., R. H. Hunt, N. Vizueta, T. Sommer, S. Durston, Y. Yang, and M. S. Worden, 2004. Evidence of developmental differences in implicit sequence learning: An fMRI study of children and adults. J. Cogn. Neurosci. 16(8):1339–1351. Thomas, K. M., S. W. King, P. L. Franzen, T. F. Welsh, A. L. Berkowitz, D. C. Noll, V. Birmaher, and B. J. Casey, 1999.

A developmental functional MRI study of spatial working memory. NeuroImage 10:327–338. Turner, R., D. Le Bihan, C. T. W. Moonen, D. Despres, and J. Frank, 1991. Echo-planar time course MRI of cat brain oxygenation changes. Magn. Reson. Med. 22:159–166. Vaidya, C. J., G. Austin, G. Kirkorian, H. W. Ridlehuber, J. E. Desmond, G. H. Glover, and J. D. Gabrieli, 1998. Selective

effects of methylphenidate in attention deficit hyperactivity disorder: A functional magnetic resonance study. Proc. Natl. Acad. Sci. USA 95(24):14494–14499. Voss, H. U., J. D. Zevin, and B. D. McCandliss, 2006. Functional MR imaging at 3.0 T versus 1.5 T: A practical review. Neuroimaging Clin. N. Am. 16:285–297.

thomas and tseng: functional mri methods

323

19

Mechanisms of Language Acquisition: Imaging and Behavioral Evidence JACQUES MEHLER, MARINA NESPOR, JUDIT GERVAIN, ANSGAR ENDRESS, AND MOHINISH SHUKLA

The attempt to explain the uniqueness of language is as old as our own cultural memory. Among the great linguists, Panini investigated the structure of Sanskrit nearly 2500 years ago. Grammarians such as Spinoza pursued the exploration of language structure further and speculated on ways in which various phonological categories are used. Descartes and the Port Royal grammarians made specific proposals about the endowment that allows humans to learn natural language. More than a century later, von Humboldt followed in their footsteps. More recently, Troubetzkoy and the various structuralist schools took a more empiricist stance, deriving much of language structure from the distributional information found in natural languages. It was during the 20th century that these major theoretical traditions developed into rival theories. On the one hand, psychologists were responsible for the popularizations of some of the most radical versions of empiricism, namely, behaviorism and its more sophisticated contemporary versions, such as connectionism (Elman et al., 1996). On the other hand, Chomsky (1965, 1980) proposed the most developed characterization of universal grammar and the principles and parameters theory, which generative grammarians developed to explain how infants acquire the natural language spoken in our surrounds. Chomsky’s main contribution was to provide the first formulation of the type of linguistic theory that is adequate linguistically, psychologically, and biologically. Rather than trying to describe normatively the well-formed utterances of a language, he explicitly stated that the aim of a grammatical theory is to offer the underlying formulas that explain why only the utterances that are grammatical will be generated. Indeed, it is possible to show that native speakers of a language know implicitly the underlying structures that are implemented by the grammar. Last but not least, Chomsky explicitly tied the value of a particular linguistic theory to its ability to account for language acquisition, that is, why it is that any uninjured infant, born into the community, will acquire language with great speed and facility—an ability

that generally escapes most adults who are trying to acquire a new language. Interestingly, theories of language acquisition were explored only from a functionalist perspective. The notion that brain mechanisms as studied in cognitive neuroscience could, at one point, become another source of information for attaining a better understanding of language acquisition seemed preposterous to many. Yet our viewpoint is consistent with the notion that if human languages arose as the result of a unique endowment characteristic of our species, then a cognitive neuroscience approach to this question is likely to enlighten our research. For instance, Neville and Bavelier (1999) state: A general hypothesis that may account for the different patterns of plasticity within both vision and language is that systems employing fundamentally different learning mechanisms display different patterns of developmental plasticity. It may be that systems displaying experience-dependent change throughout life—including the topography of sensory maps, . . . lexical acquisition . . . and the establishment of form, face and object representations . . . —rely upon general, associative mechanisms that permit learning and adaptation throughout life. This type of developmental evidence can contribute to fundamental descriptions of the architecture of different cognitive systems.

This was a position that is reminiscent of the one adopted by Eric Lenneberg (1967) in his Biological Foundations of Language. Lenneberg reviews whether the claims that the higher generic learning capacity, as suggested by behaviorists such as Skinner (1957), can account for the facts, and concludes that, contrary to “commonsense” accounts, general intelligence is not correlated with language. More recent studies have strengthened Lenneberg’s early writings. In particular, Gleitman and her students (Landau and Gleitman, 1985) have observed no language acquisition delays in the blind, contrary to what learning theories would suggest. Likewise, Goldin-Meadow and Mylander (1998) have shown that deaf infants raised in a surround that does not afford linguistic

325

input will spontaneously generate a sign language similar to the already-existing sign languages that deaf communities use. Today, the naturalistic approach, as well as the genetic and anatomical information that Lenneberg and others have claimed to be essential to gain understanding about language development, is one that is being actively researched. In this chapter, we also argue that an adequate theory of language acquisition needs to take into consideration some of the basic properties of language—namely, productivity, partial input, and the ability to acquire multiple natural languages simultaneously. Briefly, productivity refers to the capacity to understand and generate any well-formed sentence in the language if the lexicon is available. We can transform any thought into a sentence of the language, if we are so inclined, even if we have to invent new lexical terms, as is done continually in science. Partial input refers to the capacity of humans to learn the language spoken in their milieu on the basis of a limited amount of fragmentary input. Last, the ability to acquire multiple natural languages simultaneously refers to the ability of a young child growing up in a multilingual environment to create different files for the various languages spoken in the surrounds without suffering interferences, delay, or other problems that affect adults in similar situations. In the past, psycholinguists working on language acquisition did not pay sufficient attention to the resilience of the ability to learn language despite great deficits. More recently, linguists and psycholinguists formulated theories of language acquisition in which learning had little or no role. But now the pendulum has again shifted, and it would be fair to state that during the last two decades attention has been focused on how statistical machines extract regularities embodied in the linguistic input. Such machines are often taken as providing realistic models of how humans converge on the language spoken in their surrounds (see Hayes and Clark, 1970, and Rumelhart, McClelland, and Group, 1986, but see also Yang, 2004). Unfortunately, we often forget that while arbitrary statistical machines might explain, a posteriori, how the properties of the linguistic signals can shape the native speakers’ behavior, they do not address the problem of why it is that nonhuman primates, which often succeed in statistical learning tasks (Hauser et al., 2001), nevertheless fail to learn human languages, even after prolonged exposure to linguistic stimuli. The evolutionary accounts of how language arose in humans have been a taboo subject for many decades. However, in the last few years there have been several proposals comparing humans to apes (Hauser, Chomsky, and Fitch, 2002; Fitch and Hauser, 2004; Fitch, Hauser, and Chomsky, 2005; Jackendoff and Pinker, 2005; Pinker and Jackendoff, 2005). For instance, Hauser, Chomsky, and Fitch proposed that to understand the evolution of language it is best to split the study of language into the broad

326

methodological paradigms

language properties that humans share with other animals and the narrow language properties that may only be present in humans. Concretely, the conjecture that Hauser and collegues propose to evaluate is that only humans are capable of performing recursive operations. This view has been challenged by Pinker and Jackendoff, who argue that Hauser and colleagues neglect adaptation as the most likely mechanism capable of explaining the evolution of grammar. While admitting the importance of evolutionary explanations and related cross-species comparisons, our own stance is that the study of the biological foundations of language in contemporary humans—for instance, through the investigation of genetic language deficits or genetically endowed language abilities in infants—can provide equally relevant evidence about evolutionary issues. Moreover, the study of prelinguistic infants can greatly clarify what the unlearned precursors are, explaining some of the phonological and morphosyntactic properties of natural languages. In other words, modern techniques make it possible to explore whether the specific abilities to learn in humans are what shapes the form of natural languages. It may seem paradoxical that most of the work presented in this chapter is based on the learning of artificial grammars. However, since many of the experiments attempt to explore both infants and adults, simplification of the materials is desirable. The first section tries to highlight the brain structures that underlie the dispositions to acquire language that are being detected in the neonate. Imaging methods are many, and we focus on near-infrared spectroscopy (NIRS), also known as optical topography (OT). Next, we present data suggesting that rule extraction, statistical learning, and perceptual primitives intervene in the acquisition and processing of language, and we argue for their integration into comprehensive models.

Language dispositions in very young infants: NIRS studies Behavioral studies of neonates’ perception, attention, and learning abilities have relied on demanding methods to obtain the highly informative database that we now possess. Indeed, we have a fairly good understanding of how the neonate begins to process faces (Pascalis, de Haan, and Nelson, 2002), colors (Bornstein, Kessen, and Weiskopf, 1976), and aspects of speech ( Jusczyk, 1985; Mehler et al., 1988). These discoveries are all the more astonishing considering that large numbers of infants had to be discarded from the experiments because of crying, fussing, and several other reasons. Nonnutritive sucking, the most widely used method to test neonates, was notorious. Usually, more than half of the tested participants failed to complete the experiments. Three-month-olds and older infants are usually

tested using a variety of head- or eye-turning methods. It is, however, difficult or impossible to test neonates with these methods (see Aslin, Jusczyk, and Pisoni, 1998). Behavioral investigations continue to be important for the study of infant development, since they have already provided a large body of replicable data, and methods continue to improve. However, the search for supplementary methods suitable to study behavior and also inform us about the underlying brain mechanisms responsible for the infants’ behaviors is under constant development. Moreover, empirical results should be cross-validated using several methodologies. Thus it is not surprising that investigators are trying to expand the panoply of methods that developmental cognitive neuroscientists can use; some of them are exemplified in other chapters (Friederici, chapter 8 in this volume). For well over half a century, developmental science has used physiological measures like EEG and ERP for research purposes. More recently, researchers have begun using modern functional imaging techniques with very young infants. Notice, however, that fMRI is rather noisy and that immobility is required to obtain data, two considerations that render this methodology quite difficult to use with young infants. Nevertheless, a number of studies have been reported to explore the onset of language learning. For instance, some highly informative fMRI studies of language processing have been conducted with 3-month-olds (Dehaene-Lambertz, Dehaene, and Hertz-Pannier, 2002; Dehaene-Lambertz et al., 2006). The first study compared the processing of normal and reversed speech in 3-month-olds uncovering a left-hemisphere (LH) advantage in temporal areas and in the angular gyrus, much as we observe in adults. Likewise, in the second study with the same age group the authors explored the temporal sequence of activations taking place in different brain areas. The participating infants listened to utterances in their native language while they were being imaged using event-related fMRI. The authors found that Heschel’s gyrus was the first locus displaying increased activity. Some time later, both more posterior and anterior areas, including Broca’s area, also displayed increased activation (Dehaene-Lambertz et al., 2006). In this section we focus on recent discoveries made possible by NIRS in the domain of language acquisition in neonates and very young infants. Near-infrared spectroscopy relies on the differential absorption of near-infrared light by brain tissue. Near-infrared light incident on the skull is scattered, reflected, and absorbed to varying extent by various brain tissues. Changes in intensity between the emitted and the recorded light can be related to neural activity, which produces hemodynamic changes, that is, an increase in oxyhemoglobin (oxyHb) and a decrease of the deoxyhemoglobin (deoxyHB; see Jobsis, 1977; Villringer

and Chance, 1997; Yamashita, Maki, and Koizumi, 1999; Obrig and Villringer, 2003). In fact, the extent to which light is absorbed by a medium depends on the wavelength of the near-infrared light. The absorption coefficient is a measure of the relative absorbance of light given a particular medium and the wavelength. Choosing the two optimal wavelengths licenses the simultaneous estimation of changes in both oxyHb and deoxyHb. A number of laboratories have already adopted this technology to study the cognitive neuroscience of language development (e.g., Peña et al., 2003; Taga et al., 2003; Bortfeld, Wruck, and Boas, 2006). The silence with which NIRS operates is one of the greatest advantages for students of language acquisition in populations of very young infants. Moreover, movements are less critical, since the fiber optics move with the head of the participant. Unfortunately, NIRS only measures emerging photons in a given part of the head, the quantity of which relates to the functionally triggered hemodynamic response, without providing a good enough characterization of the underlying brain anatomy, because in most cases the optical probes are placed on the head using surface landmarks, such as the vertex or the ears. Experiments based on NIRS, like several of the previously mentioned fMRI studies, have observed responses to speech stimulation suggesting that the brains of young infants are already organized into areas with functions similar to those observed in older children or adults. For instance, Peña and colleagues (2003) have shown that infants’ brains respond to normal speech differently than to reversed speech, a result that is in many ways comparable to the previously mentioned fMRI study and to a behavioral study (Ramus et al., 2000). There are, however, a number of differences as well. Although the NIRS study tested newborns, 3-montholds were tested in the fMRI study. Furthermore, the newborns were mostly sleeping, whereas the sleeping babies in the fMRI study failed to show activations in some areas that displayed activity when they were awake. Moreover, the NIRS study found that the channels overlaid on the temporal, perisylvian regions of the LH are significantly more activated than the corresponding channels in the RH for normal compared to reversed speech. A more recent unpublished study (summarized in Shukla, 2006) attempted to replicate Peña and colleagues (2003) using a more sophisticated OT machine. This study found basically the same pattern of results, although the evidence in favor of an LH superiority in response to speech was restricted to a few channels. These results mesh well with results reported with deaf infants (Holowka and Pettito, 2002). Other studies have expanded the ages of the infants that OT can track. Indeed, Bortfeld, Wruck, and Boas (2006) used a sequence of speech plus visual animation interspersed with only visual animation. These blocks were separated using a blank screen presented in total silence. The authors

mehler et al.: language acquisition

327

report activations in L-temporal areas during the speech sequence and in occipital regions during exposure to visual animations. In an investigation with 3-month-olds, Homae and colleagues (2006) found that regions of the right hemisphere became activated when infants processed sentential prosody. The authors used short Japanese sentences from a previous behavioral study (Nazzi, Bertoncini, and Mehler, 1998) under two conditions. In one condition, the original sentences, which were pronounced normally, were used, while in the other condition, infants listened to the same sentences, this time with flattened prosody. The authors reported that the infants showed bilateral activation to the normal sentences. However, when they compared the activation of the normal sentences to the flattened sentences, they reported that the channels with the greatest activation were located in the RH temporal-parietal cortex. However, at the individual level, 15 infants showed a greater activation in channel 16 in the RH, while 10 infants showed greater activation in the homologous LH. In a yet-unpublished experiment, Gervain and colleagues (submitted) showed, using NIRS, that neonates process a string of structured items differently from an otherwise very similar list of items that contain no detectable structure. The structured list consisted of trisyllabic sequences with a syllable followed by a pair of identical syllables, in short, an ABB grammar. The other grammar contained no repetitions, that is, had an ABC configuration. The anterior areas of the LH showed greater overall activation (as measured by changes in the oxyHB concentration) when the neonates were listening to the ABB grammar as compared to listening to the ABC grammar. Moreover, the difference between ABB and ABC grew during the time course of the experiment. Indeed, the concentration of oxyHB became higher for the ABB grammar toward the second part of the experiment, suggesting that infants build abstract representations only for the structured grammar. As we shall see, these results can be interpreted from the perspective of purely symbolic computations, as in Marcus and colleagues (1999) or from that of configurational perceptual primitives that favor the salience of repetitions in edge positions (Endress, Scholl, and Mehler, 2005).

The interaction of statistics and prosodic structures Since the early 1970s, psycholinguists have proposed that distributional properties embodied in natural languages are used to extract words and possibly other structural regularities (Hayes and Clark, 1970). Indeed, statistical strategies were proposed for the segmentation of words, based on distributional properties over sublexical units like phonemes or syllables (e.g., Brent and Cartwright, 1996; Batchelder, 2002).

328

methodological paradigms

Ten years ago, Saffran and her colleagues reported a stunning result, namely, babies segment an artificial grammar composed of trisyllabic “words” defined by high transition probabilities1 (TPs) from one syllable to the next. A TP dip between “words” was the only cue available to the 8-monthold infants to parse the continuous string. Details and other work are reported in Aslin, Clayards, and Bardhan (chapter 7 in this volume). Statistical parsing or grouping is observed in the auditory, visual, and motor domains and in different species. The original studies by Saffran, Aslin, and others simplified their stream by disregarding prosodic cues. Johnson and Jusczyk (2001), however, provided evidence for an interaction between various cues. They reported that English 8month-olds weigh stress and coarticulatory cues more heavily than statistical cues. More recently, Thiessen and Saffran (2003) pitted TPs against stress patterns in English-learning infants and found that 7-month-olds group bisyllables according to TPs, so a coherent bisyllable is weak-strong, although in English strong syllables are typically word-initial. In contrast, for 9-month-old infants, the stress cues take precedence, and they consider strong-weak, low-TP bisyllables as coherent. Collectively, the various findings suggest that by 9 months of age, infants utilize and integrate multiple cues to word boundaries. However, stress is not the only cue to prosodic structure in spoken language. Thus sensitivity to larger prosodic constituents can signal the edges of words. Indeed, Gout, Christophe, and Morgan (2004) showed that 10- and 12.5-month-olds do not attempt lexical access on syllable sequences that span phonological phrases (see also Soderstrom et al., 2003). In addition, young infants have been shown to use intonational phrases in organizing fluent speech (e.g., Mandel, Jusczyk, and Nelson, 1994). For example, Nazzi and colleagues (2000) showed that 6-month-olds could detect previously heard word sequences in fluent speech only if the sequence did not contain an intonational phrase boundary inside it. Different cues, such as statistics and prosody, are present simultaneously in fluent speech. Indeed, several researchers have examined how various cues might interact in segmenting speech into words. More recently, we have examined possible models for how cues interact in speech segmentation. In particular, we asked how the detection of intonational phrases in fluent speech impacts the extraction of statistical regularities (Shukla, 2006; Shukla, Nespor, and Mehler, 2007). In these experiments, adults were exposed to carefully controlled artificial speech streams. In this novel paradigm, distributionally coherent (high-TP) trisyllabic nonce words were placed at different locations with respect to artificially generated (intonational) “phrases.” Thus, while some words occurred “phrase”-internally, others straddled such “phrases” (figure 19.1 and plate 44).

Figure 19.1 The artificial speech streams used in Shukla, Nespor, and Mehler (2007). The upper panel shows the design of a monotonous speech stream with nonce words inserted at the indicated locations. The lower panel shows an intonated stream, obtained by

overlaying prosody (i.e., intonational phrase [IP] contours) on the previously monotonous syllable string. Now the nonce words fall within a contour or straddling the boundary of two adjacent contours. (See plate 44.)

We found that in the absence of prosody all the nonce words are recognized, while in the presence of prosody only the “phrase”-internal words are subsequently recognized. These experiments allowed us to ask: Do prosodic boundaries inhibit the computation of TPs across them? We found this not to be the case. Under certain conditions, participants successfully recalled even the contour-straddling words. Thus we proposed that distributional information is computed independently of the presence of prosodic break points. Only at a later stage do the two cues interact— prosody acts as a filter, disallowing sequences that are aligned with prosodic edges.2

consonants or vowels, constrain the extraction of statistical regularities. Let us first review the different functions of consonants and vowels, as established by linguistic theory, in order to gain insight into how they might interact with statistical learning. The main generalization, supported by numerous empirical observations (Nespor, Mehler, and Peña, 2003), claims that consonants tend to carry the lexical meanings of words, while vowels express grammatical and morphological functions. Almost universally, languages have more consonants than vowels. Consequently, consonants allow for greater diversity and can encode more information. Thus they are more adequate than vowels to subserve the storage of a large number of distinctions, characteristic of the lexicon. Vowels, however, are less numerous, thus less distinct, and even tend to harmonize in certain languages, like Turkish or Hungarian. Importantly, the domain over which vowels harmonize is larger than just the lexical word and usually encompasses the morphological and some of the syntactic dependents of a word as well. More direct evidence for the division-of-labor hypothesis comes from Semitic languages, in which lexical roots are made up of consonants only, which thus define a basic meaning (-k-t-b- is the root of words related to “writing”), whereas the vowels indicate the morphological features of words. These linguistic observations had been backed up by results from several other domains of research. In psycholinguistics, it had been established that consonants cue the lexicon more than vowels do. In an experiment, Cutler and colleagues (2000) found that participants prefer to keep the consonants rather than the vowels constant in nonsense

The interaction of distributional information and linguistic categories The interplay between domain general mechanisms, such as statistical learning, and representations specific to language is an emerging research area. In particular, research is focusing on the nature of the unit(s) over which statistics are computed by human learners. The original statistical learning experiments used artificial streams in which the transitional probabilities were equally informative between units of different kinds—for example, syllables, consonants, and vowels. In natural languages, however, these units play different roles (Nespor, Mehler, and Peña, 2003). Moreover, cross-linguistic variation in their relative importance and function is also considerable. Therefore, in order to understand how statistical learning might scale up from artificial grammars to the acquisition of a natural language, it is crucial to investigate how linguistic representations, such as

mehler et al.: language acquisition

329

words that allow both the change of a consonant and the change of a vowel to yield an existing lexical item (e.g., kebra is more often changed into cobra than into zebra). Studies in language acquisition showed that infants lose the discrimination of nonnative vowels earlier than that of nonnative consonants (Werker and Tees, 1984; Kuhl et al., 1992). Language pathologies also provided evidence for the asymmetry between consonants and vowels. Caramazza and colleagues (2000) reported a double dissociation between them, evidenced by two aphasic patients, one of whom exhibited selective impairment for consonants, while the other showed impairment for vowels. If true, the division-of-labor hypothesis makes rather direct predictions about the selective role of consonants and vowels in statistically based segmentation. Since consonants are claimed to carry lexical meaning, it is not unreasonable to expect that they are preferred over vowels for the purposes of statistical segmentation, one of the main uses of which is to assist word learning. Indeed, in the past years, a considerable body of evidence has accrued, suggesting that statistics might be preferentially computed over consonants, but not over vowels. The initial investigations yielded mixed results. While Newport and Aslin (2004) found that participants segment with equal ease using statistical information over consonants and vowels, Bonatti and colleagues (2005) obtained segmentation over consonants only. There are, however, a number of differences between the methodologies and materials used by the two groups, possibly explaining their diverging results. Newport and Aslin (2004), for instance, used only two consonantal and vocalic frames as opposed to the three frames of Bonatti and colleagues (2005). Moreover, the former authors allowed immediate repetitions of the same frame in the familiarization, while the latter ones did not. The smaller number of frames and the repetitions in Newport and Aslin’s (2004) experiments might be partly or even fully responsible for successful segmentation with vowels. This conclusion has been confirmed by further investigations. Toro and colleagues (2008) have found that different mechanisms operate over consonants and vowels in artificial-grammar-learning situations. While consonants allowed segmentation but not generalization, the vowels of the same speech stream readily subserved the extraction of regularities. This observation was true even when the generalization concerning consonants was made very simple (identity) and the information about it was highly redundant. Unpublished work by Shukla and colleagues has further shown that such simple generalizations (identity) over vowels were easy for participants to learn and actually prevailed over consonantal TPs. Taken together, these results argue for the view that the different cues available in language interact with each other. Specifically, the general learning mechanism of TP compu-

330

methodological paradigms

tations is constrained in language by the nature of the different types of cues present in the input. Some of these representations, such as consonants, readily undergo TP computations, because their linguistic function—that is, encoding lexical distinctions—is compatible with the output of TP computations—that is, potential word candidates. The last two sections have addressed the difficult problem of how such a powerful mechanism as statistical computations interacts with other salient properties of natural languages. We saw that while intonational phrases and statistics interact to disallow the statistical nonce words that straddle boundaries, prosody cannot suppress the automatic statistical computations. We also saw that consonants are a more suitable category of speech upon which to compute statistical dependencies than are vowels. It is premature to say whether this conclusion indicates that speakers utilize the knowledge of their native language, which, in most cases has many more consonants than vowels, to select the former over the latter to carry out the parsing routines. It could be the case that an unlearned disposition in humans results in languages that have more consonants than vowels because they are more learnable and the lexicon of such language leads to improved lexical access routines. We are currently conducting NIRS experiments with neonates and 4-month-olds to clarify which of the preceding options might be correct. We have illustrated the function of a powerful learning mechanism and how it interacts with other properties of languages. We also saw how categories of speech can constrain which of these mechanisms operates best. Now we are going to illustrate other mechanisms and constraints that play an important role in language acquisition. Indeed, it is conceivable that the properties of natural languages honor the functional characteristics of our perceptual organs— most particularly, audition.

Perceptual primitives Recent research has uncovered two mechanisms to highlight auditory units regardless of whether they are frequent or statistically salient. One of these mechanisms is the highlighting of edges and the other is the detection of repetitions. As we shall argue, neither of these mechanisms seems to result from learning. Edges of domains in speech may modulate how words are segmented, but they may also determine what kinds of generalizations can be extracted from speech streams. One case in point comes from a study by Peña and colleagues (2002) showing that the inclusion of subliminal silences between words, in otherwise continuous speech streams such as the ones described earlier, induces participants to extract generalizations. Peña and colleagues (2002) familiarized participants with a sequence of nonsense words in which the first syllable always predicted the last one while the middle syl-

lable was variable. The predictive relation between the first and the last syllable could be used in at least two ways. On the one hand, participants could use this relation as a cue to word boundaries and use this statistical relation to segment the speech stream into its constituent words. Peña and colleagues (2002) showed that participants do indeed have this capacity. On the other hand, participants may also generalize this relation to new items; in this case, they should accept items as legal if they conform to the dependencies between the first and the last syllables, although they have a different middle syllable. After familiarization with a continuous speech stream, participants did not accept these generalizations, even when familiarized with a stream of 30 minutes. However, when words were separated by subliminal silences, a 2-minute familiarization was sufficient for inducing the generalizations. Indeed, participants preferred items that had never occurred during the speech stream but that respected the configuration of the edge syllables.3 What are the mechanisms underlying this generalization? To address this issue, Endress and Mehler (under review) used pentasyllabic words and asked whether participants would learn generalizations only when the crucial syllables were in the edge positions (that is, the first and the fifth ones) or also in middle positions (the second and the fourth ones). When the critical syllables were in edges, participants readily learned to generalize. In contrast, when the critical syllables were in nonedge positions, participants showed no evidence for the generalizations. Unlike the generalizations, however, statistical processes worked well also in middles. The latter results also suggest that the edge advantage for the generalizations cannot be explained only in terms of the salience of the edges. If it were so, one would expect also statistical processes to break down in middles, which, in fact, they do not do. Hence, edges seem to play a different role for generalization than merely to highlight particular syllables. Another case in point for the importance of edges in artificial grammar learning has come from phonotactic generalizations. Languages differ in their permissible sound sequences; for example, most consonant clusters would be illegal in Japanese but frequent in Polish. Chambers, Onishi, and Fisher (2003) showed that young infants can learn constraints on permissible sound patterns from very short exposure. They familiarized participants with CVC (consonant-vowel-consonant) words in which they restricted the consonants that could occur in the first or the last position, respectively. In other words, the first and the last consonants had to come from two distinct sets. After such a familiarization, the infants applied the constraints to new words, thus generalizing them to new instances. Again, the crucial consonants were placed in the edges of words. To ask whether this feature of the experiments was crucial to the generalizations, Endress and Mehler (under review) asked whether adults could learn similar constraints

in longer CVCCVC words. Again, participants had to learn that two consonants had to come from two distinct sets. However, the crucial consonants were in the edges (that is, the first and the last ones) for half of the participants, and in middles for the other participants (that is, the second and the third consonants). Participants readily generalized the constraints when the crucial consonants were in the edges but not when they were in middles. This outcome may have occurred because participants simply do not perceive middle consonants well. However, Endress and Mehler (under review) also showed that participants can discriminate words that differ only in their middle consonants perfectly well; hence, a global impairment for processing middle consonants is unlikely to be the only explanation of the edge advantage for generalizations. The importance of edges for generalizations in artificial grammar learning can also be demonstrated by considering the experiments by Marcus and colleagues (1999). In their experiments, young infants were familiarized with syllable sequences conforming to one of the grammars ABA, AAB, or ABB (e.g., a sequence like “wo-fe-fe′” would conform to ABB). The infants generalized these grammars to new syllables they had not heard before. Marcus and colleagues (1999) argued that these generalizations were evidence for algebraic-like rules in very young infants. While algebraiclike rules definitely play a significant role in language acquisition, the particular stimuli that Marcus et al. (1999) used might be learned through other mechanisms. First, their structures used repetitions, and we have argued elsewhere that repetition-based structures may be generalized by a simple, specialized operation rather than by a more general rule-extraction mechanism. Second, the repetitions also occurred in sequence edges. To test the role of the edges in this context, Endress, Scholl, and Mehler (2005) used sevensyllable sequences (rather than the triplets in Marcus et al.’s 1999 experiments) to ask whether repetition-based structures would be generalized as easily in edges as in middles. They showed that participants generalized repetition-based grammars much more readily when the critical syllables were in edges than when they were in middles; for example, they readily generalized the structure ABCDEFF, but they failed to generalize the structure ABCDDEF. One may be tempted to attribute this result to perceptual difficulties for processing middle syllables. Endress, Scholl, and Mehler’s (2005) control experiments show that this explanation of the edge advantage is unlikely. They asked participants to discriminate sequences that differed either only in middles or in edges; participants still had to process middle (or edge) syllables but were no longer required to abstract the underlying structure. Participants could discriminate both types of stimuli well above chance. These results suggest that the generalization of such grammars is constrained independently of psychophysical problems for processing middle syllables.

mehler et al.: language acquisition

331

A plausible explanation is thus that only edges have proper positional codes, while other positions can be encoded only relative to such anchor points. One may ask whether the biases that we have reviewed may be useful also for linguistic phenomena or only for artificial grammars. While learning syntax obviously entails much more than an edge detector, even such an operation may be important for some aspects of grammar. The location of word stress in phonology is a first example. Word stress is located relative to either the left or to the right edge; it may be initial, final, or, otherwise, on a syllable counted from the right edge. In contrast, no language has been observed that appeals to word middles, for example, by locating stress on the middle syllable (e.g., Halle and Vergnaud, 1987; Hayes, 1995). Morphology also often appeals to edges. Suffixes and prefixes have been observed in many languages, while infixes are rare across languages (e.g., Greenberg, 1957). Another important function of edges may be to interface different levels of representation. For example, morphosyntactic and phonological representations are both hierarchical but have distinct hierarchies—for example, some morphemes. In such cases, the constituents of the two hierarchies do not coincide; however, at least one of the edges of the constituents must be aligned (Nespor and Vogel, 1986; McCarthy and Prince, 1993). Edges thus seem to help integrate different hierarchies and levels of representation, and to coordinate them. Surprisingly, mechanisms as simple as an “edge detector” may thus be important for hierarchical processing, a property that has been considered crucial for human cognition (e.g., Fodor, 1983; Gallistel, 1990, 2000; Marr, 1982; Marr and Nishihara, 1992). It also highlights the fact that some perceptual biases may have been recruited by the language faculty both for word learning and for more abstract, structural computations.

Discussion and conclusion We have attempted to show that theorists who focus on one mechanism to the detriment of other mechanisms with which the first interacts may limit our understanding of development. Since the publication of Saffran, Aslin, and Newport (1996), it has been recognized that infants rely on distributional cues to segment speech streams. In the section on the interaction of statistics and prosodic structures, we presented data that corroborated the importance and automaticity of statistical computations during speech processing. However, we also showed that when other sources of information are made available in the input, complementary mechanisms provide a complete processing account. This finding suggests that studies of language acquisition, while relying on past discoveries, must also understand how different processing components mesh with one another,

332

methodological paradigms

helping to elaborate more naturalistic explorations of language acquisition. We believe that working with artificial grammars will still prove very useful. However, the more we succeed in scaling up to naturalistic stimuli, the more we are going to learn. For instance, consider the ability of neonates to respond differently to the ABB as compared to the ABC grammar, as described earlier. Clearly, making the grammars more complex generates richer models, yielding testable predictions. For instance, comparing grammars containing adjacent repetitions to others with nonadjacent repetitions might instruct us about how working memory develops during the first months of life. That is, making repetitions more and more distant, memory span can be tested. Likewise, introducing prosody in a grammar-learning context may also allow us to track in greater detail which cues intervene to constrain the underlying computations. We also want to stress the importance of the research reviewed in the section on perceptual primitives. Kimball (1973) and Bever (1970) claim that perceptual processes are essential to our understanding of how the language user parses novel sentences. Indeed, psycholinguists have experimentally documented the reality of several such claims. Nevertheless, the influence of perception on language acquisition has only recently turned into an active research area. In that section, we presented research suggesting that repetitions are detected through a primitive identity detector. Gervain and associates (submitted) showed that even newborn infants detect adjacently repeated syllables. Furthermore, Shukla (2006) has demonstrated that the closer a word recurs the more it is highlighted. Lindblom and Lacerda (personal communication) have shown that motherese across many languages contains an unsuspected number of word repetitions. Endress has argued that edges of such items as words, phrases, and sentences tend to be far more salient than middles. These and other such perceptual primitives should not be ignored. Indeed, those primitives are well documented in the domain of auditory sequential processing. Endress, Dehaene-Lambertz, and Mehler (in press) showed that repetition detectors function with tones as well as syllables. Whether such perceptual primitives can be attested for visual simultaneous or sequential processing is still an open question. In brief, we do not think that generalizations, statistics, or perceptual primitives should be considered as singletons. Rather, we believe that the language acquisition device (LAD) uses all these mechanisms to make language learnable by humans. Whereas Chomsky (1975) formulated the LAD as a framework within which language learning ought to be conceived, the time is ripe to fill in the details giving an outline of how each of the mechanisms fulfills its prespecified roles. Even the most detailed linguistic theory of how language might be acquired, the principles and parameters

theory, will ultimately be judged by how well it can integrate all these mechanisms to explain how an infant goes from signals to abstract grammatical representations. Bootstrapping theories of language acquisition (e.g., Morgan and Demuth, 1996) have isolated some perceptual properties in the speech signal that correlate with abstract grammatical properties. For instance, Nespor (1995) and Nespor, Guasti, and Christophe (1996) have argued that OV (object-verb) and VO languages place the prosodic prominence at opposite edges of phrases. If so, abstract properties of grammar might be signaled by the prosodic structure of the linguistic data. Since there exist numerous languages that have both OV and VO constructions, it is possible that the frequency of these constructions, together with prosody, might select some grammatical properties for a particular language. Notice, however, that the prosodic bootstrapping hypothesis requires that the infant be already endowed with alternative possible grammars (“parameters”). Some properties might arise from the signal plus constraints proper to the perceptual mechanisms of the modality through which language is transmitted. Thus, in general, prefixing and suffixing are far more frequent as morphological positions than infixing. This fact might arise from the salience of edges in auditory signals. Likewise, grammatical markers tend to appear in edges rather than in middles of constituents. In conclusion, we have argued in favor of a linguistically informed cognitive neuroscience model of language acquisition. Although we have mostly presented data concerning very basic processes, we have done so considering that the human mind is endowed with the specific disposition to acquire a grammatical system with its appropriate categories. The details of how the human endowment interfaces with the psychological mechanisms that go from universal grammar to particular grammars is still a matter of active investigation. Last but not least, the progress achieved over the past decade or two in brain imaging has made it possible to explore the endowment for grammar from birth through the first year of life with a facility that was previously unimaginable. Our understanding of the mature brain is constantly increasing, making it possible to view the infant’s brain from a perspective of greater ontogenetic continuity than our predecessors had fathomed. NOTES 1. TP(A→B) = P(AB)/P(A), where A and B are units of language, e.g., segments, syllables, etc.; AB is the co-occurrence of A and B; and P(X) is the probability of the occurrence of unit X. 2. What drives the perception of such prosodic edges in fluent speech? It is known that the boundaries of prosodic units are associated with acoustic cues like final lengthening and pitch decline (e.g., Beckman and Pierrehumbert, 1986). Indeed, such cues have also been shown to be important in detecting “phrases” in music. For example, Krumhansl and Jusczyk (1990) used a

pause-detection paradigm with 4.5- and 6-month-olds and showed that even the younger infants perceived musical phrases as being defined by a pitch decline-reset at phrase boundaries and by a relatively longer final tone. These results suggest that prosodic contours are marked by acoustic patterns that might not be language specific. We can thus consider a decline in pitch as a perceptual primitive: a preexisting capacity that is engaged even in a languagespecific task—segmenting fluent speech. 3. More recent research has shown that participants actually extract a regularity entailing syllable classes from such subliminally segmented speech streams; see Endress and Bonatti (2007). The authors showed that participants learn that the first and the last syllables of words have to belong to different syllable classes (the classes being the sets of syllables that can occur in these positions). REFERENCES Aslin, R. N., P. W. Jusczyk, and D. B. Pisoni, 1998. Speech and auditory processing during infancy: constraints on and precursors to language. In Handbook of Child Psychology, ed. D. K. R. Siegler, 147–198. New York: Wiley. Beckman, M. E., and J. B. Pierrehumbert, 1986. Intonational structure in Japanese and English. Phonol. Yearb. 3:255–309. Batchelder, E. O., 2002. Bootstrapping the lexicon: A computational model of infant speech segmentation. Cognition 83(2):167– 206. Bever, T. G., 1970. The cognitive basis for linguistic structures. In J. R. Hayes, ed., Cognition and the Development of Language, 279–362. New York: Wiley. Bonatti, L., M. Peña, M. Nespor, and J. Mehler, 2005. Linguistic constraints on statistical computations: The role of consonants and vowels in continuous speech processing. Psychol. Sci. 16(6): 451–459. Bornstein, M. H., W. Kessen, and S. Weiskopf, 1976. Color vision and hue categorization in young human infants. J. Exp. Psychol. 2:115–119. Bortfeld, H., E. Wruck, and D. A. Boas, 2006. Assessing infants’ cortical response to speech using near-infrared spectroscopy. NeuroImage 34(1):407–415. Brent, M., and T. Cartwright, 1996. Distributional regularity and phonotactic constraints are useful for segmentation. Cognition 61:93–125. Caramazza, A., D. Chialant, R. Capasso, and G. Miceli, 2000. Separable processing of consonants and vowels. Nature, 403:159– 160. Chambers, K. E., K. H. Onishi, and C. Fisher, 2003. Infants learn phonotactic regularities from brief auditory experience. Cognition 87(2):B69–77. Chomsky, N., 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Chomsky, N., 1975. Reflections on Language. New York: Pantheon Books. Chomsky, N., 1980. Rules and Representations. New York: Columbia University Press. Cutler, A., N. Sebastián-Gallés, O. Soler-Vilageliu, and B. Van Ooijen, 2000. Constraints of vowels and consonants on lexical selection: Cross-linguistic comparisons. Mem. Cogn. 28(5): 746–755. Dehaene-Lambertz, G., S. Dehaene, and L. Hertz-Pannier, 2002. Functional neuroimaging of speech perception in infants. Science 298:2013–2015.

mehler et al.: language acquisition

333

Dehaene-Lambertz, G., et al. 2006. Functional segregation of cortical language areas by sentence repetition. Hum. Brain Mapping 27(5):360–371. Elman, J. L., E. A. Bates, M. H. Johnson, A. Karmiloff-Smith, D. Parisi, and K. Plunkett, 1996. Rethinking Innateness: A Connectionist Perspective on Development. Cambridge, MA: MIT Press. Endress, A. D., and L. L. Bonatti, 2007. Rapid learning of syllable classes from a perceptually continuous speech stream. Cognition. 105(2):247–299. Endress, A. D., G. Dehaene-Lambertz, and J. Mehler, in press. Perceptual constraints and the learnability of simple grammars. Cognition. Endress, A. D., and J. Mehler (under review). Perceptual constraints in phonotactic learning. J. Mem. Lang. Endress, A. D., B. J. Scholl, and J. Mehler, 2005. The role of salience in the extraction of algebraic rules. J. Exp. Psychol. [Gen.] 134(3):406–419. Fitch, W. T., and M. D. Hauser, 2004. Computational constraints on syntactic processing in a nonhuman primate. Science 303:377– 380. Fitch, W. T., M. D. Hauser, and N. Chomsky, 2005. The evolution of the language faculty: Clarifications and implications. Cognition 97:179–210. Fodor, J. A., 1983. The Modularity of Mind. Cambridge, MA: MIT Press. Gallistel, C., 1990. The Organization of Learning. Cambridge, MA: MIT Press. Gallistel, C., 2000. The replacement of general-purpose learning models with adaptively specialized learning modules. In M. Gazzaniga, ed., The Cognitive Neurosciences, 2nd ed., 1179–1191. Cambridge, MA: MIT Press. Gervain, J., F. Macagno, S. Cogoi, M. Pena, and J. Mehler, submitted. The neonate brain extracts speech structure. Goldin-Meadow, S., and C. Mylander, 1998. Spontaneous sign systems created by deaf children in two cultures. Nature 391: 279–281. Gout, A., A. Christophe, and J. Morgan, 2004. Phonological phrase boundaries constrain lexical access. II. Infant data. J. Mem. Lang. 51:547–567. Greenberg, J. H., 1957. Essays in Linguistics. Chicago: University of Chicago Press. Halle, M., and J.-R. Vergnaud, 1987. An Essay on Stress. Cambridge, MA: MIT Press. Hauser, M. D., N. Chomsky, and W. T. Fitch, 2002. The faculty of language: What is it, who has it, and how did it evolve? Science 298:1569–1579. Hauser, M. D., E. L. Newport, and R. N. Aslin, 2001. Segmentation of the speech stream in a nonhuman primate: Statistical learning in cotton top tamarins. Cognition 78:B53–B64. Hayes, B., 1995. Metrical Stress Theory: Principles and Case Studies. Chicago: University of Chicago Press. Hayes, J. R., and H. H. Clark, 1970. Experiments on the segmentation of an artificial speech analogue. In Cognition and the Development of Language, ed. J. R. Hayes, 221–234. New York: Wiley. Holowka, S., and L. A. Petitto, 2002. Left hemisphere cerebral specialization for babies while babbling. Science 297: 1515. Homae, F., H. Watanabe, T. Nakano, K. Asakawa, and G. Taga, 2006. The right hemisphere of sleeping infants perceives sentential prosody. Neurosci. Res. 54(4):276–280. Jackendoff, R., and S. Pinker, 2005. The nature of the language faculty and its implications for evolution of language (reply to Fitch, Hauser, and Chomsky). Cognition 97:211–225.

334

methodological paradigms

Jobsis, F., 1977. Noninvasive infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science 198:1264–1267. Johnson, E. K., and P. W. Jusczyk, 2001. Word segmentation by 8-month-olds: When speech cues count more than statistics. J. Mem. Lang. 44:548–567. Jusczyk, P. W., 1985. On characterizing the development of speech perception. In Neonate Cognition: Beyond the Blooming, Buzzing Confusion, ed. J. Mehler and R. Fox, 199–229. Hillsdale, NJ: Erlbaum. Kimball, J. P., 1973. Seven principles of surface structure parsing in natural language. Cognition 2(1):15–47. Krumhansl, C. L., and P. W. Jusczyk, 1990. Infants’ perception of phrase structure in music. Psychol. Sci. 1:70–73. Kuhl, P. K., K. A. Williams, F. Lacerda, K. N. Stevens, and B. Lindblom, 1992. Linguistic experience alters phonetic perception in infants by 6 months of age. Science 255:606–608. Landau, B., and L. R. Gleitman, 1985. Language and Experience: Evidence from the Blind Child. Cambridge, MA: Harvard University Press. Lenneberg, E. H., 1967. The Biological Foundations of Language. New York: Wiley. Mandel D. R., P. W. Jusczyk, and D. G. Kemler Nelson, 1994. Does sentential prosody help infants to organize and remember speech information? Cognition 53:155–180. Marcus, G. F., S. Vijayan, S. B. Rao, and P. Vishton, 1999. Rule learning by seven-month-old infants. Science 283(5398): 77–80. Marr, D., 1982. Vision. San Francisco: W. H. Freeman. Marr, D., and H. K. Nishihara, 1992. Visual information processing: Artificial intelligence and the sensorium of sight. In S. M. Kosslyn and R. A. Andersen, eds., Frontiers in Cognitive Neurosceince, 165–186. Cambridge, MA: MIT Press. McCarthy, J. J., and A. Prince, 1993. Generalized alignment. In G. Booij and J. van Marle, eds., Yearbook of Morphology, 1993, 79–153. Boston: Kluwer. Mehler, J., et al., 1988. A precursor of language acquisition in young infants. Cognition 29:143–178. Morgan, J., and K. Demuth, 1996. Signal to Syntax: Bootstrapping from Speech to Grammar in Early Acquisition. Mahwah, NJ: Lawrence Erlbaum. Nazzi, T., J. Bertoncini, and J. Mehler, 1998. Language discrimination by newborns: Toward an understanding of the role of rhythm. J. Exp. Psychol. [Hum. Percept.] 24:756– 766. Nazzi, T., D. Kemler Nelson, P. Jusczyk, and A. M. Jusczyk, 2000. Six-month-olds’ detection of clauses embedded in continuous speech: Effects of prosodic well-formedness. Infancy 1:123– 147. Nespor, M., 1995. The role of the phonological phrase for syntactic access. Proc. ABRALIN 1994:97–100. Nespor, M., and I. Vogel, 1986. Prosodic Phonology. Foris: Dordrecht. Nespor, M., M. T. Guasti, and A. Christophe, 1996. Selecting word order. In U. Kleinhenz, ed., Interfaces in Phonology, 1–26. Berlin: Akademie Verlag. Nespor, M., J. Mehler, and M. Peña, 2003. On the different role of vowels and consonants in language processing and language acquisition. Lingue e Linguaggio 221–247. Neville, H. J., and D. Bavelier, 1999. Specificity and plasticity in neurocognitive development in humans. In M. Gazzaniga, ed., The New Cognitive Neurosciences, 2nd ed., 83–98. Cambridge, MA: MIT Press.

Newport, E. L., and R. N. Aslin, 2004. Learning at a distance. I. Statistical learning of nonadjacent dependencies. Cogn. Psych. 48:127–162. Obrig, H., and A. Villringer, 2003. Beyond the visible—Imaging the human brain with light. J. Cereb. Blood Flow Metab. 23(1):1–18. Pascalis, O., M. de Haan, and C. A. Nelson, 2002. Is face processing species-specific during the first year of life? Science 296:1321–1323. Peña, M., L. L. Bonatti, M. Nespor, and J. Mehler, 2002. Signaldriven computations in speech processing. Science 298(5593): 604–607. Peña, M., et al., 2003. Sounds and silence: An optical topography study of language recognition at birth. Proc. Natl. Acad. Sci. USA 100(20):11702–11705. Pinker, S., and R. Jackendoff, 2005. The faculty of language: What’s special about it? Cognition 95:201–236. Ramus, F., M. D. Hauser, C. Miller, D. Morris, and J. Mehler, 2000. Language discrimination by human newborns and by cotton-top tamarin monkeys. Science 288:349–351. Rumelhart, D. E., J. L. McClelland, and P. R. Group, 1986. Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Cambridge, MA: MIT Press. Saffran, J. R., R. N. Aslin, and E. L. Newport, 1996. Statistical learning by 8-month-old infants. Science 274:1926–1928. Schulz, R. W., 1955. Generalization of serial position in rote serial learning. J. Exp. Psychol. 49(4):267–272. Shukla, M., 2006. Prosodic constraints on statistical strategies in segmenting fluent speech. PhD dissertation, SISSA, Trieste, Italy.

Shukla, M., M. Nespor, and J. Mehler, 2007. An interaction between prosody and statistics in the segmentation of fluent speech. Cogn. Psych. 54(1):1–32. Skinner, B. F., 1957. Verbal Behavior. New York: Appleton-CenturyCrofts. Soderstrom, M., A. Seidl, D. Kemler Nelson, and P. Jusczyk, 2003. The prosodic bootstrapping of phrases: Evidence from prelinguistic infants. J. Mem. Lang. 49:249–267. Taga, G., K. Asakawa, A. Maki, Y. Konishi, and H. Koizumi, 2003. Brain imaging in awake infants by near-infrared optical topography. Proc. Natl. Acad. Sci. USA 100:10722– 10727. Thiessen, E. D., and J. R. Saffran, 2003. When cues collide: Use of stress and statistical cues to word boundaries by 7- to 9-month-old infants. Dev. Psychol. 39:706–716. Toro, J. M., M. Nespor, J. Mehler, and L. Bonatti, 2008. Finding words and rules in a speech stream: functional differences between vowels and consonants. Psychol. Sci. 19(2). Villringer, A., and B. Chance, 1997. Noninvasive optical spectroscopy and imaging of human brain function. Trends Neurosci. 20:435–442. Werker, J., and R. Tees, 1984. Cross-language speech perception evidence for perceptual reorganization during the first year of life. Infant Behav. Dev. 7:49–63. Yamashita, Y., A. Maki, and H. Koizumi, 1999. Measurement system for noninvasive dynamic optical topography. J. Biomed. Opt. 4(4):414–417. Yang, C. D., 2004. Universal grammar, statistics, or both? Trends Cogn. Sci. 8:451–456.

mehler et al.: language acquisition

335

20

Magnetic Resonance Spectroscopy of Developing Brain BALASRINIVASA RAO SAJJA AND PONNADA A. NARAYANA

Our understanding of developmental changes in brain has so far been mainly based on cell cultures, animal models, and postmortem samples. However, these techniques are limited to providing information at one time point. In addition, such measurements are affected by tissue excision and sample preparation. Therefore, it is highly desirable to follow developmental changes in vivo. Rapid neuronal organization, cell proliferation and differentiation, and myelination associated with maturation alter the metabolic and biochemical profiles during early development. The inherently quantitative nature of magnetic resonance spectroscopy (MRS) allows us to follow age-dependent metabolic profiles in vivo noninvasively. The normal age-dependent metabolic profile is altered in the presence of even subtle pathology. Thus MRS should help investigate the effect of neurological disorders, including developmental disorders, on the normal trajectory of age-dependent metabolic profile (Scarabino et al., 1999). In this chapter, whenever possible, we rely on recent reviews rather than providing individual references to older publications. For example, Vigneron (2006) has provided a broad review of the role of proton MRS imaging (MRSI) in neurodevelopment. Some of the older literature has also been reviewed in a number of excellent publications on proton MRS of childhood neurodevelopment (Grodd et al., 1991; van der Knaap et al., 1990; Tzika et al., 1993; Kreis, Ernst, and Ross, 1993; Huppi et al., 1995). In this review we concentrated only on normal human brain development and the effects of various neurodevelopmental disorders on metabolic profiles, as assessed by proton MRS. We did not include the effects of acquired injuries on the metabolic profiles. Instead the reader is referred to some of the excellent publications in this area (Groenendaal et al., 1994; Robertson et al., 1999; Peden et al., 1993; Barkovich et al., 1999; Roelants–Van Rijn et al., 2001; Amess et al., 1999; Frahm and Hanefeld, 1997; Cady, 2001; Peden et al., 1990; Robertson et al., 2001; Barkovich et al., 2006).

Magnetic resonance spectroscopy Magnetic resonance imaging (MRI) exhibits superb softtissue contrast and provides exquisite anatomical and

morphological information. Conventional MRI maps the density distribution of hydrogen nuclei (protons or 1H) in water molecules in the volume of interest, modulated by relaxation times. In contrast, MRS detects tissue biochemicals or metabolites noninvasively. It is important to point out at the outset that not all neurochemicals can be detected with MRS. Magnetic resonance spectrum in the frequency domain exhibits a series of peaks or resonances arising from different molecules within the tissue. The locations or frequencies of the metabolite resonances depend on the local biochemical environment. The area under the spectral peak is proportional to the corresponding metabolite concentration. The concentrations of cerebral metabolites and their positions are altered by various factors that include normal brain maturation and tissue pathology. The potentially rich biochemical information that can be gained through MRS, which may not be apparent on conventional MRI, helps us to understand the metabolic changes during typical development and the effect of developmental disorders on the normal metabolic profiles. With the development of high-field MR scanners for human studies and improvements in hardware, software, and spectral analysis techniques, MRS has been transformed from a highly specialized to a relatively routine technique. Magnetic resonance spectroscopy can be performed using a variety of MR sensitive nuclei such as 1H, 13C, 19F, 23Na, and 31P. The biologically important nuclei with adequate concentrations in tissues are 1H, 23Na, and 31P. Among these three nuclei, 1H and 31P are most commonly used in biological MRS. Phosphorus spectroscopy has great biological importance because it allows one to probe metabolites such as phosphocreatine (PCr), adenosine triphosphate (ATP), and inorganic phosphate (Pi) that play important roles in tissue energetics (Cady, 1990; Ross and Bluml, 2001). The metabolite 31P MRS demonstrated the ability to detect significant changes in neonatal hypoxic-ischemic insults (Cady, 2001). In spite of its importance, phosphorus MRS is limited in its clinical applicability because of its inherently low sensitivity, which limits the minimum voxel size to approximately 20 cm3 for adequate signal-to-noise ratio (SNR). This relatively large voxel introduces significant partial volume

337

averaging from different tissue types. In addition, 31P MRS studies require special hardware that may not be commonly available on all clinical MR scanners. Proton MRS has attracted considerable attention for a number of reasons. These include high MR sensitivity of 1H, access to a large number of tissue biochemicals such as Nacetyl aspartate (NAA), creatine + phosphocreatine (Cr), choline + choline containing compounds (Cho), lactate (Lac), glutamine and glutamate (Glx), gamma-aminobutyric acid (GABA), and alanine, among others. In addition, proton MRS can be performed on most clinical MR imaging scanners without additional hardware requirements. This review focuses only on proton MRS. MRS Acquisition Magnetic resonance spectroscopy is most commonly acquired in either single-voxel or multivoxel mode. In the single-voxel mode, MRS is acquired from one region at a time. The volume of this region is typically 1 to 8 cm3. Single-voxel spectroscopy involves shorter acquisition and processing times. A high-quality spectrum is relatively easy to acquire in this mode. It can be routinely performed on a clinical scanner by technologists. Multivoxel mode, also referred to as MRSI, chemical shift imaging (CSI), or simply spectroscopic imaging (SI), allows simultaneous acquisition of MRS data from multiple voxels and provides information about regional distribution of metabolite concentrations. Multivoxel MRS can be acquired in two or three dimensions and has superior spatial resolution (<1 cm3). However, MRSI involves relatively long acquisition and processing times. The concentration of metabolites is generally in the millimolar range. Thus MRS has relatively low sensitivity and requires larger voxels (poor spatial resolution) and longer acquisition times for adequate SNR. Proton MRS is dominated by the tissue water signal that has a concentration of about 50 M. In order to visualize the resonances from metabolites with relatively low concentration, it is common to selectively suppress the water signals using chemically selective radio frequency (RF) pulses. In addition, even a small contamination from lipids from the surrounding nonneural tissues would adversely affect spectral quantification. The spectral contamination from extrameningeal tissues is generally minimized by prelocalizing the volume of interest using localization techniques, such as point-resolved spectroscopy (PRESS) (Bottomley, 1987), simulated echo acquisition mode (STEAM) (Frahm, Merboldt, and Hanicke, 1987), and outer volume suppression (Posse, DeCarli, and LeBihan, 1994). The PRESS technique provides superior SNR compared to the STEAM localization. However, STEAM allows shorter echo times (TE) relative to PRESS. Proton MRS The major metabolites in neural tissue that are observed in proton MRS include NAA, Cr, Cho, glutamate, myoinositol (mI), and Lac. Lactate is normally

338

methodological paradigms

seen only in the presence of certain pathologies. Of all these metabolites, NAA has attracted the greatest attention, since it is exclusively located in neurons (Birken and Oldendorf, 1989) and is considered to be a putative marker of neuronal/ axonal integrity. The observed Cr resonance arises from creatine and phosphocreatine. The concentration of Cr is not significantly affected by pathology and is often used as an internal standard relative to which various 1H-MRSobserved metabolite concentrations are expressed. Cho signals have contributions from multiple molecules that include phosphorylcholine, glycerophoshorylcholine, and choline plasmalogen and a very minor contribution from acetylcholine and choline (Frahm and Hanefeld, 1997). Regional Cho levels approximately parallel the degree of myelination, and therefore Cho is generally considered to be a marker of demyelination. However, an elevated Cho level may arise from either elevated phosphorylcholine— from increased membrane turnover and during cell proliferation—or accumulation of breakdown products such as glycerophoshorylcholine from myelin and other cell membranes. Brain osmolyte mI appears to be glia specific but is not found in neurons (Brand, Richter-Landsberg, and Leibfritz, 1993). Also, mI is a precursor of phospholipid membrane constituents, and its concentration is affected by the formation and breakdown of myelin. It exhibits two resonances, at 3.5 and 4.06 ppm. Both mI and glycene contribute to the resonance at 3.5 ppm. However, in normal brain, the glycene concentration is fairly low, and therefore the 3.5-ppm peak is attributed to mI. The resonance at 4.06 ppm does not have any contribution from glycene. However, because of its proximity to the tissue water resonance, the amplitude of the 4.06-ppm peak is affected by the degree of water suppression and is not generally considered to be a reliable indicator of mI concentration. Lactate is the end product of anaerobic glycolysis and is not commonly observed in normal brain. However, its presence generally indicates tissue pathology. Short and Long Echo Proton MRS Proton MRS can be acquired at either short or long echo time (TE). The advantage of long-TE proton MRS is the well-defined spectral baseline that allows robust spectral quantification. This is achieved at the expense of loss of signals from metabolites with short transverse relaxation time (T2) such as lipids, glutamate, and the like that contain a wealth of metabolic information (Wolinsky and Narayana, 2002). However, short-TE data contain broad resonances from macromolecules that introduce baseline distortions. Figure 20.1 shows brain spectra acquired on a normal volunteer at 3 Tesla (T) from deep white matter at TE = 35 ms and TE = 144 ms. The flatter baseline in the long-TE spectrum and rich spectral information in the short-TE spectrum can be appreciated on this figure.

Figure 20.1 Representative single-voxel MR spectrum acquired at 3 Tesla from deep white matter of a normal adult brain with TE = 35 ms (left) and TE = 144 ms (right). The flat baseline at TE = 144 ms makes the metabolite quantification relatively simple

and more robust. The short-TE spectrum has rich information content, since it preserves the information from short T2 metabolites.

Spectral Quantification The inherently quantitative nature is a major advantage of MRS and allows one to monitor tissue state and response to therapeutic intervention objectively. Quantitative spectroscopy information is commonly presented as the ratio of metabolite peak areas relative to an internal standard such as Cr or as absolute metabolite concentrations. Expressing the metabolite concentration relative to an internal standard (Cr is the most commonly used internal standard) is relatively straightforward and is robust. However, the interpretation of the metabolite ratios becomes equivocal if the concentration of the internal standard varies with pathology (Narayana et al., 1998; also see the section on attention-deficit/hyperactivity disorder). This problem can be overcome to some extent by expressing the metabolite concentrations relative to tissue water. However, this may not be very accurate if the tissuewater content is altered by pathology. Whenever possible, it is best to estimate the absolute concentrations of the metabolites for unequivocal interpretation of the MRS results. A number of commercial software packages such as LC model (Provencher, 1993) or free software packages such as MRUI (Naressi et al., 2001) are commonly used for the estimation of absolute concentrations. In addition, a number of research groups have implemented their own versions for the analysis of MRS (Doyle et al., 1995; Maudsley et al., 1994; Soher et al., 1998; Narayana et al., 1998; Bhat, Sajja, and Narayana, 2006).

separately. As expected, there is some overlap in the data across different age groups.

Proton MRS of normal brain development The metabolic profiles exhibit a highly nonlinear age dependence (Vigneron, 2006). In order to highlight these changes, we review the metabolic changes in normal brain development in fetuses, neonates, and children and adolescents

Fetuses In vivo fetal MRS studies are technically very challenging for a number of reasons that include general nonavailability of dedicated fetal-brain-imaging coils (Brunel et al., 2004), large distance between the fetal brain and the RF receive coils, and fetal motion during data acquisition (Heerschap, Kok, and van den Berg, 2003). In addition, scanning fetuses for nonclinical reasons poses severe ethical problems. It is therefore not surprising that normal patterns and normal variants in fetal brain are not yet fully determined (Garel, 2006). The first study demonstrating the feasibility of in vivo acquisition of proton MRS in fetal brain was reported by Heerschap and van den Berg (1994). These authors acquired spectra on six fetuses with gestational age (GA) between 32 and 38 weeks. The first quantitative MRS analysis in fetal brain was published by Kok and colleagues (2001). Their study included 21 normal fetuses with GA of 36–41 weeks. Resonances from NAA, Cho, Cr, and mI were clearly observed in fetal brains. Girard, Fogliarini, and colleagues (2006) and Girard, Gouny, and associates (2006) have reported MRS of fetal brains from 22 to 39 weeks GA. Figure 20.2 shows the metabolic changes with maturation in the human fetal brain from 22 to 39 weeks of GA at short echo time (TE = 30 ms) and long echo time (TE = 135 ms). At 22 weeks, the spectrum is characterized by two prominent resonances that are assigned to mI and Cho. A weak NAA peak is also detected at this age. Both mI and Cho tend to decrease with increasing age, while NAA and Cr resonances have become better defined and more intense. At 34 weeks, the metabolic pattern is very similar to the neonatal spectrum, with three dominant

sajja and narayana: magnetic resonance spectroscopy

339

TE = 30 ms NAA

myo-ins + gly tCr tCr Cho

TE = 135 ms

NAA Glx

Mobile lipids

Cho tCr

*

39 weeks Acetate

Aspartate

34 weeks

31 weeks

24 weeks

22 weeks 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00

4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00

ppm

ppm

* Taurine + scyllo-Inositol

Figure 20.2 Typical fetal brain spectra obtained with the PRESS sequence at short (left panel) and long (right panel) echo times from 22 to 39 weeks GA. (Reproduced from Girard, Gouny, et al., 2006.)

resonances, Cho, Cr, and NAA, at long echo time and five dominant resonances, mI, Cho, Cr, NAA, and Glx, at short echo time (Girard, Fogliarini, et al., 2006; Girard, Gouny, et al., 2006). A progressive increase in the NAA peak with GA was attributed to maturation of the brain with the development of dendrites and synapses. The Cho peak decreases as myelination progresses (Garel, 2006). Kok and colleagues (2002) used proton MRS to monitor cerebral metabolite tissue levels in 35 normal fetuses during development at 30–41 weeks GA. A volume of interest (15– 43 cc) of brain tissue was selected for 1H MRS. Spectral localization was achieved with two pulse sequences— STEAM at TE = 20 ms and PRESS at TE = 135 ms. Consistent with the previously published results, the MR spectra of the brain showed resonances from mI, Cho, Cr, and Nacetyl (NA). The ratios of NA/Cr and NA/Cho increased with GA, whereas the ratio of Cho/Cr decreased, reflecting maturation of the brain. Neonates Perhaps the first in vivo quantitative MRS studies in neonates were reported by Huppi and colleagues

340

methodological paradigms

(1991). Their study included preterm and term newborns. The range of postconceptional age was 31–45 weeks. Singlevoxel 1H MRS of newborns was obtained from the cerebellum regions. By using the biochemically measured creatine concentrations as internal standard, the in vivo concentrations of NAA, Cho, taurine (Tau), and mI were presented. The results showed a significant increase of NAA from 1.8 mM in preterm infants to 3.1 mM in term infants. Similar variation in Tau concentration, 1.1 mM in preterm infants and 2.3 mM in term infants, was observed. Cho and mI did not show any significant changes between preterm and term infants. Kries, Ernst, and Ross (1993) have investigated MRS changes over a relatively large age range of 34.5 to 926 weeks of GA. Spectra were acquired using the STEAM sequence from predominantly white matter in the parietooccipital region and gray matter in the occipital cortex. At birth, the mI peak was the most dominant peak with a concentration of 12 mM/kg, while Cho was strongest in older infants (2.5 mM/kg). Both Cr and NAA were significantly lower relative to adults. These studies suggest that NAA and Cr concentrations correlated strongly with

GA whereas mI showed the strongest correlation with postnatal age. Regional metabolic changes in neonates (postnatal age of 34–40 weeks) were investigated by Cady and colleagues (1996) and Penrice and colleagues (1996). Spectra were acquired from the thalamic region (predominantly gray matter) and occipital region (predominantly white matter). These studies suggest that the concentrations of Cho, Cr, and NAA are higher in thalamus than in the occipitoparietal white matter. The higher NAA concentration in thalamus was attributed to the greater neuronal cellularity in this predominantly gray matter region. However, the reasons for higher Cho and Cr in thalamus are not clear. The concentration of lactate is similar in both regions, but higher than in older subjects. In a seminal study, Huppi and colleagues (1995) have investigated the regional metabolic changes in early brain development using in vivo proton MRS and chromatography in autopsied samples. The in vivo studies were performed on 14 preterm infants (GA of 27–34 weeks) and 14 term infants (GA of 37–40 weeks). The autopsy data included five preterm infants (GA 29.2 ± 3.7 weeks) and five term infants (39.6 ± 1.1 weeks). Consistent with other reports, an increase in NAA and decrease in mI with age were observed. However, Cho levels appeared to be relatively stable during this period. The autopsy studies that were performed at much higher spatial resolution showed much larger Cr concentration in thalamus and basal ganglia than in the precentral area or the frontal lobe. However, this concentration distribution was not evident on in vivo MRS studies because of partial volume-averaging effect resulting from relatively poor spatial resolution. In an initial study to detect the spatial distribution of proton MRS-detectable compounds in premature and term infants, Vigneron and colleagues (2001) performed 3D MRSI on nine premature and eight term neonates. Figure 20.3 is representative of MRSI data from a premature and a term neonate. The multivoxel spectra along with the image from a 30-week (postconceptional age) premature infant with normal neurologic outcome is shown in figure 20.3A. The corresponding data from a 41-week term neonate are shown in figure 20.3B. This study demonstrated significant spectral differences among anatomic locations and between the premature and term groups. Figure 20.4 shows representative spectra from 1-cm3 voxels from three regions in the brain of a premature neonate (postconceptional age, 30 weeks) and a term neonate (postconceptional age, 40 weeks). In premature babies, regions that mature earliest, such as the thalamus, demonstrated the highest levels of NAA, whereas the latermaturing frontal white matter showed the lowest levels. The basal ganglia spectra demonstrated the largest increase in NAA in term neonates compared to premature infants, consistent with rapid maturation over this period. A similar

metabolite pattern in frontal white matter that is consistent with later maturation of this region was also observed. These studies suggest that the metabolite profiles are different between the term and preterm infants. Further, the temporal changes in metabolic profiles are different between these two groups, and these differences depend on the brain regions. Based on studies of 21 newborns with GA ranging from 32 to 43 weeks, Kreis and colleagues (2002) have shown that the brain maturation was associated with significant increases in NAA, Glx, and Cr and reduction of mI and lactate, among other metabolites. Although NAA was observed to be significantly reduced with an elevation in Cho levels in premature babies (32 weeks GA) compared to term newborns, the total brain metabolite content was not significantly different between these two groups, indicating that premature birth did not substantially affect the biochemical brain maturation. This finding is somewhat inconsistent with the results reported by Huppi and colleagues (1991) and 3D MRSI results reported by Vigneron and colleagues (2001). However, the study by Kreis and colleagues (2002) did not include infants of extremely low birth weight. All these studies demonstrate that MRS in neonates presents a significantly different appearance from those of an adult brain. In newborns, mI is the dominant resonance, and the NAA peak amplitude is much smaller than the Cho resonance. In contrast, in the adult brain the NAA peak is much higher than the Cho peak. The metabolite concentrations and their ratios in neonatal brain are region dependent and change nonlinearly with age. Children and Adolescents Proton MRS of brain in 47 healthy children and in six healthy adults was reported by Hashimoto and colleagues (1995). Resonances from NAA, Cho, and Cr were observed in all cases. In the right parietal region NAA/Cho and NAA/Cr increased and Cho/Cr decreased with age. The most rapid changes were noted between 1 and 3 years of age. Also, 1H-MRS of the right frontal region was performed in 21 cases (20 children and one adult). In the right frontal region, NAA/Cho showed an increase, while Cho/Cr decreased with age. However, developmental changes were not observed in NAA/Cr. The ratios of both NAA/Cho and NAA/Cr were lower in the right frontal region than in the right parietal region. As indicated in the previous section, the metabolite concentrations vary from region to region during development, and these variations need to be considered in interpreting the MRS results of neurological disorders. Pouwels and colleagues (1999) were perhaps the first to report the age-dependent changes in the absolute concentrations of metabolites in different brain regions using quantitative single-voxel MRS. The group in this study ranged in age from 0 to 18 years. Unlike many previous publications, these authors were also able to separately quantify NAAG

sajja and narayana: magnetic resonance spectroscopy

341

Figure 20.3 Representative MR imaging data from a premature and a term neonate. (A) Images and spectral array from the brain of a premature, 30-week postconceptional age neonate. (B) Images

342

methodological paradigms

and spectral array from the brain of a term, 41-week postconceptional age neonate. (Reproduced from Vigneron et al., 2001.)

Figure 20.4 Representative spectra from three regions in a premature neonate (postconceptional age, 30 weeks) and a term neonate (postconceptional age, 40 weeks). Note higher relative N-acetylaspartate levels in the thalamus and especially the basal

ganglia of the term infant. Note also a similar metabolite pattern in the frontal white matter, which is consistent with later maturation of this region. (Reproduced from Vigneron et al., 2001.)

and PCr (phosphocreatine). These studies revealed an agedependent increase in NAA and a decrease of Tau in gray matter. The concentration of NAA was observed to be constant. However, an increase of NAAG in white matter (WM), an increase of tNAA and a decrease of Tau in cerebellum, and an increase of tNAA and NAAG in thalamus were observed. None of the regions studied showed age-related changes in glutamate (Glu). Glutamine (Gln) decreased by 50 percent only in WM. Since Gln is synthesized from Glu in astrocytes, this finding may reflect either a functional change or density of glia in WM. Taurine (Tau) was observed to have the highest concentration in cerebellum and gray matter in infancy. Tau has been implicated in neuronal differentiation, dendritic arborization, and synaptic connections (Sturman, 1993). Perhaps this finding explains increased NAA and concomitant decrease in Tau during brain development. Horska and colleagues (2002) examined the age dependence of metabolite concentrations during postnatal brain development in 15 normal, healthy subjects aged between 3 and 19 years. This study documented small but significant changes in regional cerebral metabolism from childhood to adolescence. The observed nonlinear age-related changes of NAA/Cho in frontal and parietal areas may be associated with dendritic and synaptic development and regression. The linear changes of NAA/Cho in white matter with age are also in agreement with the age-related increases in white matter volumes, and may reflect progressive increase in axonal diameter and myelination.

In summary, the differences in the acquisition and analysis techniques and choice of different regions of interest (ROIs) in the brain have produced metabolic profiles that are not always consistent. However, a few consistent observations in the age-related metabolic profiles have been reported. These include (1) low NAA levels at birth that reach almost the adult levels by 3 years of age, (2) decrease in Cho during this age period, (3) high levels of mI at birth that rapidly fall to the adult levels by the age of 3–4 months, and (4) all these changes occurring both in gray and white matter. In typical development, NAA and Cho levels correlate with gestational age, while mI correlates with the postnatal age. However, caution should be exercised in the interpretation of age-related changes in the metabolic profiles, since they can be affected by various factors such as changes in nutrition, medication, and functional state.

Developmental disorders The effect of developmental disorders on metabolic profiles is reviewed in this section. The effect of acquired injuries such as hypoxic ischemic injury on metabolic profiles is not included in this review. Rather the focus was on various neurodevelopmental disorders such as developmental delays, autism, attention-deficit/hyperactivity disorder (ADHD), and schizophrenia. Developmental Delays Mild developmental delay or isolated developmental delay affects approximately 15

sajja and narayana: magnetic resonance spectroscopy

343

percent of school-age children (Shevell, 1998). These children suffer from restricted deficiencies in language, learning, or motor skills. Their academic performance is below that of their peers and educational level. Between 35 and 43 percent of subjects diagnosed with developmental delay are normal on conventional MRI (Shevell et al., 2003). Delayed myelination is thought to be associated with developmental delay. Filippi and colleagues (2002) have evaluated metabolic changes in children with mild developmental delay. The subjects were divided into two age groups: children under 2 years (6 months to 2 years) and older than 2 years (3 to 10 years). The rationale for this division was that the MR signal changes from myelin maturation are almost complete by 2 years. Children were diagnosed with developmental delay if they failed to meet one or more typical developmental milestones that include speech and language skills, motor skills, behavioral development, and learning. Spectra were acquired from frontal and parieto-occipital subcortical white matter. In children under 2 years, there were no differences in the NAA/Cr and Cho/Cr ratios between normals and those diagnosed with developmental delays. However, in older children (>2 years), the delayed de-velopmental group had decreased NAA/Cr and increased Cho/Cr in the frontal and parieto-occipital white matter compared to the agematched controls. Lack of significant changes in the metabolite ratios in the younger group could be due to small sample size and methodological deficiencies in these studies. The altered metabolic ratios in older group suggest decreased synaptic density (reduced NAA/Cr) and/or hypomyelination (elevated Cho/Cr). These studies suggest a potential role for proton MRS as a modality that could provide an objective marker of developmental delay. More recent studies by Fayed and colleagues (2006) suggest that the reduced NAA/ Cr ratio could serve as a marker of developmental delay. Yeo and colleagues (2000) investigated the effect of known genetic and environmental influences on development and working-memory ability in children from 7 to 12 years of age. Working memory was assessed by the Visual Two-Back test, and developmental instability (DI) was assessed by a composite score that was created by measuring minor physical anomalies, fluctuating asymmetry of body characteristics, and fluctuating dermatoglyphic features. These studies demonstrated that lower Cho and Cr was a strong predictor of DI. The lower NAA and Cr values showed correlation with working memory. Thus proton MRS appears to have a role in assessing working-memory and developmental instabilities in children. Disturbances in cortical development during the fetal period result in developmental delays or mental retardation and sometimes cause epileptic seizures. In a small number of children with lissencephaly, gray matter heterotopia, and cortical dysplasia, Kaminaga, Kobayashi, and Abe (2001) quantitatively evaluated the proton MRS-derived

344

methodological paradigms

metabolite levels in the affected cortex. The results were compared with the MRS of occipital cortex in age-matched children. In all the patient groups, NAA was observed to be lower in the affected cortex relative to the controls. Significantly lower levels of Cho were observed in lissencephaly relative to controls. The levels of Cr and mI were not different in patients compared to controls. These authors interpreted the reduced NAA levels in these disorders as an indication of immature neurons in the affected cortex that result in reduced numbers of axons and dendrites forming synapses. These authors suggest that reduced Cho in lissencephaly could be the result in hypomyelination. Thus proton MRS appears to provide important information about cortical development. Autism Autism or autistic spectrum disorder (ASD) is a heterogeneous developmental disorder characterized by social deficits, learning disabilities, and language and communication difficulties. Neuroimaging has provided significant information about the neuroanatomical changes in autism (Lainhart, 2006; Courchesne et al., 2005; Herbert, 2004). A number of proton MRS studies were performed to understand the metabolic changes in ASD (Friedman et al., 2003; Hashimoto et al., 1997; Otsuka et al., 1999; Mori et al., 2001; Fayed and Modrego, 2005; Levitt et al., 2003; Friedman et al., 2006). Hashimoto and colleagues (1997), using a single-voxel, long-echo MRS, did not find any changes in the NAA/Cr, Cho/Cr, and NAA/Cho ratios in the white matter of the right parieto-occipital region in autistic children and age-matched healthy controls. Similarly, Zeegers and colleagues (2007) failed to detect any differences in metabolites in the frontal subcortical white matter and amygdala-hippocampus complex between autistic and normal children. Other studies (Otsuka et al., 1999; Mori et al., 2001) reported reduced NAA in subjects in the age range of 2–21 years. Short-echo MRS studies (Fayed and Modrego, 2005) did not reveal any changes in the neurochemical concentrations in the white matter of centrum semiovale. Friedman and colleagues (2003), using proton echoplanar spectroscopic imaging (PEPSI), a fast spectroscopic imaging sequence, reported region-dependent changes in NAA, Cr, Cho, and mI in 3- to 4-year-old ASD subjects. A subsequent, multislice MRSI at long TE revealed decreased Cho in the left anterior cingulate, left caudate, and occipital regions (Levitt et al., 2003). The same study also reported increased Cho and Cr in the right caudate nucleus. More recently, Friedman and colleagues (2006) performed crosssectional MRSI studies in 3- to 4-year-old ASD children and compared the results with age-matched children with delayed development (DD) and typical development (TD). These studies revealed decreased neurochemical concentration in gray matter in ASD compared to TD and DD. However, the changes in white matter appear to be less specific to

ASD but may be more related to general developmental pace. These observations were interpreted as suggestive of decreased cellularity or density in the early phase of ASD. Murphy and colleagues (2002) also reported an increased NAA in frontal lobe in Asperger syndrome, an autistic disorder. Interestingly, Chugani and colleagues (1999) also implicated altered metabolism in autism. However, these studies were performed on a relatively few patients, and caution should be exercised in interpreting these results. Social interaction and language difficulties are common in ASD. It is therefore possible that metabolic disturbances occur in the Wernicke’s area, the speech center, and the prefrontal area that is involved in social behavior. Hisaoka and colleagues (2001) acquired single-voxel, long-TE proton MRS from Broadmann’s areas 41 and 42, frontal lobe, temporal lobe, brain stem, and the cingulate. These studies were performed on a relatively large number of autistic subjects, 55, in the age range of 2–21 years and 51 healthy controls in the age range of 3 months to 15 years. These authors observed that although the NAA concentration increased in Broadmann’s area with age in both groups, the agedependent changes were much slower in patients than in controls. The autistic subjects had significantly lower NAA only in the left temporal and right temporal lobes relative to controls. The levels of Cho and Cr did not differ significantly between these two groups. These findings demonstrate neuronal impairment or dysfunction in Wernicke’s area that correlates with the language problems in this disorder. The absence of any differences in the Cho and Cr concentrations suggests that nonneuronal cells, including astrocytes, are not remarkably affected in autism. Published results of the proton MRS studies on ASD do not appear to be completely consistent. This lack of consistency is at least in part due to the differences in the selection of the ROIs in the brain and the heterogeneity of this disorder. Attention-Deficit/Hyperactivity Disorder Attentiondeficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder whose pathophysiologic basis is not well understood. The prevalence rate of ADHD in children is estimated to be between 3 and 5 percent (NIMH, 1996). However, more recent studies report significantly different prevalence rates that vary anywhere between 2.2 and 17.8 percent (Brown et al., 2001; Skounti, Philalithis, and Galanakis, 2007). This large variation is mainly due to the modifications in the definition of the disorder in the Diagnostic and Statistical Manual of Mental Disorders (DSM) and differences in clinical impairment evaluation. It is generally believed that the frontal-striatal circuitry, particularly the right side, is involved in ADHD (Barkley, Grodzinsky, and DuPaul, 1992; Himelstein, Newcorn, and Halperin, 2000). The prefrontal cortex plays an important role in

impulsivity and emotional control and may be involved in hyperactive behavior (van Elst et al., 2001). The typical frontal asymmetry that is normally found in healthy children is generally absent in ADHD, perhaps suggesting underdevelopment of white matter (Filipek et al., 1997). Magnetic resonance imaging studies indicate decreased volumes in different regions of brain (Castellanos et al., 2002). The most consistent finding in ADHD, based on neuroimaging studies, is the significant volume reduction in dorsolateral prefrontal cortex, caudate, palladium, corpus callosum, and cerebellum (Seidman, Valera, and Makris, 2005). Proton MRS in a small number of adult ADHD subjects has shown reduced NAA in the left dorsolateral prefrontal cortex compared to normals as well as patients without hyperactivity (ADD patients) (Hesslinger et al., 2001). There were no differences in other metabolites. This finding indicates left prefrontal neuropathology in ADHD in adults. Interestingly, these studies did not reveal any differences in the NAA levels in striatum between ADHD and controls. The first proton MRS of ADHD in children was reported by Jin and colleagues (2001) to determine metabolic alterations in the striatum and whether methylphenidate treatment has any effect on the metabolic profile. Twelve children (10–16 years of age) diagnosed with ADHD and 10 controls (11–15 years of age) were included in these studies. The initial scans were performed on drug-naïve children, and a second scan was performed after 2 weeks or later within 1.5 to 2 hours after the administration of methylphenidate. Single-voxel, short-TE MRS was acquired from the globus pallidus. The NAA/Cr ratio was significantly reduced bilaterally in the striatum in the ADHD subjects compared to controls. In contrast, the Cho/Cr ratio showed a slight increase unilaterally in the ADHD subjects relative to the controls. The treatment did not alter the NAA/Cr or Cho/ Cr ratio significantly. Based on these studies, the authors concluded that the striatum is involved bilaterally in the ADHD and that about 20–25 percent of neurons in this structure may have been dysfunctional. Functional studies of ADHD demonstrated reduced activity in both striatum and cingulate gyrus, suggesting neuronal dysfunction in these regions (Bush et al., 1999; Teicher et al., 2000; Vaidya et al., 1998; Rubia et al., 1999). MacMaster and colleagues (2003), using a long-TE, singlevoxel MRS, studied nine ADHD subjects (7–16 years of age) and age- and gender-matched controls. The ADHD subjects were medication free with an early onset of disease (3.67 ± 1.41 years). Spectra were acquired from the right prefrontal cortex and left striatum. Unlike the previous reports (Jin et al., 2001), this study failed to document any change in the NAA/Cr in the striatum. In another study, Yeo and colleagues (2003) also failed to demonstrate any changes in the metabolite ratios in the frontal lobe in ADHD.

sajja and narayana: magnetic resonance spectroscopy

345

In a recent study, Courvoisie and colleagues (2004) investigated the correlation between MRS and neuropsychological assessment (NEPSY), an age-normed standardized battery of tasks (Korkman and Peltomaa, 1991). This study included ADHD children (6–12 years) without affective or learning comorbidity (ADHD-H), and single-voxel MRS was acquired at short TE from left and right frontal lobe areas. Children with ADHD showed higher NAA/Cr, Cho/ Cr, and Glu/Cr in the right frontal lobe compared to normals. However, in the left frontal lobe, only Glu/Cr appeared to be significantly higher compared to the normals, while the differences in NAA/Cr approached the significance level. Based on the comparison of the metabolite resonance areas, the elevated NAA/Cr and Cho/Cr values appeared to be mainly due to reduced Cr levels. However, the depressed Cr levels did not completely account for the observed increases in Glu/Cr. These authors concluded that in ADHD subjects Cr levels were depressed in the right frontal lobe and Glu increased bilaterally relative to controls. These studies also demonstrated a strong correlation between the NAA/Cr in the right frontal lobe and regulation of the sensorimotor, attention/executive, and learning. Since the Cr resonance at 3 ppm has contributions from GABA (gamma-aminobutyric acid), these authors suggested that, based on the reduced level of the 3-ppm peak, the ADHD group has (1) either decreased GABA levels or hypometabolism (reduced Cr) and (2) an increased glutamate level. The increased Glu level also appears to be consistent with the previously reported higher Glx resonances in the frontalstriatal in ADHD (MacMaster et al., 2003; Carrey et al., 2003). However, the spectral quantification in the 2.3-ppm region is notoriously difficult, particularly at long echo times employed by MacMaster and colleagues (2003). At lower fields such as 1.5 T at which these data were acquired, it is very difficult to separate glutamate from glutamine. These technical problems, combined with the relatively small number of subjects in these studies, require a high degree of caution in interpreting these results. Recently, Fayed and Modrego (2005), based on singlevoxel short-TE MRS, observed increased NAA/Cr in the centrum semiovale region in the ADHD children relative to controls. They interpreted this as an increase in the NAA levels and suggested hypermetabolism in mitochondria, since there is some suggestion that NAA could be an indicator of energy metabolism (Clark, 1998). However, it is unclear that the increased NAA/Cr is the result of increased NAA and not decreased Cr, as reported by Courvoisie and colleagues (2004). This observation underscores the need for determining the absolute concentrations rather than metabolite ratios. Sun and colleagues (2005) have recently reported differences in the metabolic profiles between subjects with predominantly inattentive type (ADHD-I) and the combined

346

methodological paradigms

subtype (ADHD-C). They observed a significant reduction in the NAA/Cr values in lenticular nucleus in the ADHD-C group compared to the ADHD-I subtype. However, no differences in any of the other metabolite ratios were observed between these two subtypes and controls. Based on these studies, it appears possible to distinguish these two ADHD subtypes on the proton MRS. Overall, proton MRS results in ADHD do not appear to be consistent. A number of reasons, including differences in the acquisition of MRS, analysis strategies, and ROI placement, could be responsible for these inconsistencies. Another shortcoming of these studies is that majority of these studies determined the metabolite ratios rather than absolute concentrations. This makes unambiguous interpretation of these results difficult. Schizophrenia Schizophrenia is considered to be a neurodevelopmental disorder. The pathophysiological basis for this disease is still not well known. Previous MRI studies in adult schizophrenia have documented structural changes in a number of brain regions, including anterior cingulate, frontal cortex, thalamus, and striatum (Wright et al., 2000). Many of these regions also exhibit metabolic abnormalities on MRS (Keshavan, Stanley, and Pettegrew, 2000; Stanley, 2002). The metabolic abnormalities include depressed NAA and elevated Cho and Cr levels. These observations suggest compromised neuronal integrity and disturbed membrane turnover. O’Neill and colleagues (2004) argued that these metabolic disturbances should be more prominent in childhood-onset schizophrenia, considered to be a more severe form of the disorder. These authors have performed proton MRSI studies of childhood-onset schizophrenia at long echo times in a small group of children and adolescents. Based on the determination of absolute concentrations, they observed higher Cr in superior anterior cingulate in patients relative to controls. Increased Cho levels were observed in the superior anterior cingulate, frontal cortex, and caudate head. However, NAA did not show a generalized decrease, but it appears to be lower in males, but not in females, in the thalamus. These observations suggest abnormal local energy demands and/or phospholipids membrane disturbance. However, this study is based on a small number of patients and needs to be confirmed.

Limitations of MRS of developing brain In vivo MRS studies of developing brain are technically very challenging. There are a number of practical issues that affect acquisition of MRS data in children. The first major problem is lack of cooperation on the part of subjects. It is quite difficult to expect children to lie still in the magnet during the relatively long MRS scans. This problem can be overcome by sedating these subjects. This poses significant

ethical problems with normal children. Children who are scanned for clinical management may be too sick to be anesthetized. There are also concerns that anesthesia may affect the MRS data (van der Knapp, Ross, and Valp, 1994). Thus the best way to scan these subjects is during their natural sleep time. This approach invariably results in a significant number of failed scans. Ethical considerations also prevent us from developing an adequate database that documents age-dependent metabolic profiles in a large number of normal children. The results are not always consistent due to differences in (1) the choice of ROI, (2) acquisition methods, and (3) analysis techniques. Another limitation of the published studies is that majority of the studies are cross-sectional rather than longitudinal. As demonstrated in the elegant study by Huppi and colleagues (1995), the relatively large spectroscopy voxel masks some of the true regional differences in the concentrations of metabolites. Similarly, a majority of the studies are based on a single-voxel MRS, providing limited brain coverage. Even though proton MRS can provide information on a large number of neurochemicals, much of the information has been derived from a relatively few resonances such as NAA, Cr, Cho, and to some extent mI. Another limitation of proton MRS is that many of the observed resonances have contributions from multiple molecules that limit the specificity. In spite of these limitations, proton MRS has provided unique information about the metabolic changes in typical development and neurodevelopment disorders. With the introduction of high-field scanners and improvements in hardware, software, and analysis techniques, MRS is expected to play an increasingly important role in providing critical information on normal neurodevelopment and developmental disorders. REFERENCES Amess, P. N., J. Penrice, M. Wylezinska, A. Lorek, J. Townsend, J. S. Wyatt, C. Amiel-Tison, E. B. Cady, and A. Stewart, 1999. Early brain proton magnetic resonance spectroscopy and neonatal neurology related to neurodevelopmental outcome at 1 year in term infants after presumed hypoxic-ischaemic brain injury. Dev. Med. Child Neurol. 41:436–445. Barkley, R. A., G. Grodzinsky, and G. J. DuPaul, 1992. Frontal lobe functions in attention deficit disorder with and without hyperactivity: A review and research report. J. Abnorm. Child Psychol. 20:163–188. Barkovich, A. J., K. Baranski, D. Vigneron, J. C. Partridge, D. K. Hallam, B. L. Hajnal, and D. M. Ferriero, 1999. Proton MR spectroscopy for the evaluation of brain injury in asphyxiated, term neonates. AJNR Am. J. Neuroradiol. 20:1399– 1405. Barkovich, A. J., S. P. Miller, A. Bartha, N. Newton, S. E. Hamrick, P. Mukherjee, O. A. Glenn, D. Xu, J. C. Partridge, D. M. Ferriero, and D. B. Vigneron, 2006. MR imaging, MR spectroscopy, and diffusion tensor imaging of sequential studies

in neonates with encephalopathy. AJNR Am. J. Neuroradiol. 27: 533–547. Bhat, H., B. R. Sajja, and P. A. Narayana, 2006. Fast quantification of proton magnetic resonance spectroscopic imaging with artificial neural networks. J. Magn. Reson. 183:110–122. Birken, D. L., and W. H. Oldendorf, 1989. N-acetyl-L-aspartic acid: A literature review of a compound prominent in 1H-NMR spectroscopic studies of brain. Neurosci. Biobehav. Rev. 13:23–31. Bottomley, P. A., 1987. Spatial localization in NMR spectroscopy in vivo. Ann. NY Acad. Sci. 508:333–348. Brand, A., C. Richter-Landsberg, and D. Leibfritz, 1993. Multinuclear NMR studies on the energy metabolism of glial and neuronal cells. Dev. Neurosci. 15:289–298. Brown, R. T., W. S. Freeman, J. M. Perrin, M. T. Stein, R. W. Amler, H. M. Feldman, K. Pierce, and M. L. Wolraich, 2001. Prevalence and assessment of attention-deficit/hyperactivity disorder in primary care settings. Pediatrics 107:E43. Brunel, H., N. Girard, S. Confort-Gouny, A. Viola, K. Chaumoitre, C. D’Ercole, D. Figarella-Branger, C. Raybaud, P. Cozzone, and M. Panuel, 2004. Fetal brain injury. J. Neuroradiol. 31:123–137. Bush, G., J. A. Frazier, S. L. Rauch, L. J. Seidman, P. J. Whalen, M. A. Jenike, B. R. Rosen, and J. Biederman, 1999. Anterior cingulate cortex dysfunction in attention-deficit/hyperactivity disorder revealed by fMRI and the Counting Stroop. Biol. Psychiatry 45:1542–1552. Cady, E. B., 1990. Clinical Magnetic Resonance Spectroscopy. New York: Plenum Press. Cady, E. B., 2001. Magnetic resonance spectroscopy in neonatal hypoxic-ischaemic insults. Childs Nerv. Syst. 17:145–149. Cady, E. B., J. Penrice, P. N. Amess, A. Lorek, M. Wylezinska, R. F. Aldridge, F. Franconi, J. S. Wyatt, and E. O. Reynolds, 1996. Lactate, N-acetylaspartate, choline and creatine concentrations, and spin-spin relaxation in thalamic and occipitoparietal regions of developing human brain. Magn. Reson. Med. 36:878–886. Carrey, N., F. P. MacMaster, J. Fogel, S. Sparkes, D. Waschbusch, S. Sullivan, and M. Schmidt, 2003. Metabolite changes resulting from treatment in children with ADHD: A 1H-MRS study. Clin. Neuropharmacol. 26:218–221. Castellanos, F. X., P. P. Lee, W. Sharp, N. O. Jeffries, D. K. Greenstein, L. S. Clasen, J. D. Blumenthal, R. S. James, C. L. Ebens, J. M. Walter, A. Zijdenbos, A. C. Evans, J. N. Giedd, and J. L. Rapoport, 2002. Developmental trajectories of brain volume abnormalities in children and adolescents with attentiondeficit/hyperactivity disorder. JAMA 288:1740–1748. Chugani, D. C., B. S. Sundram, M. Behen, M. L. Lee, and G. J. Moore, 1999. Evidence of altered energy metabolism in autistic children. Prog. Neuropsychopharmacol. Biol. Psychiatry 23:635– 641. Clark, J. B., 1998. N-acetyl aspartate: A marker for neuronal loss or mitochondrial dysfunction. Dev. Neurosci. 20:271–276. Courchesne, E., E. Redcay, J. T. Morgan, and D. P. Kennedy, 2005. Autism at the beginning: Microstructural and growth abnormalities underlying the cognitive and behavioral phenotype of autism. Dev. Psychopathol. 17:577–597. Courvoisie, H., S. R. Hooper, C. Fine, L. Kwock, and M. Castillo, 2004. Neurometabolic functioning and neuropsychological correlates in children with ADHD-H: Preliminary findings. J. Neuropsychiatry Clin. Neurosci. 16:63–69. Doyle, T. J., R. Pathak, J. S. Wolinsky, and P. A. Narayana, 1995. Automated proton spectroscopic image processing. J. Magn. Reson. [B] 106:58–63.

sajja and narayana: magnetic resonance spectroscopy

347

Fayed, N., and P. J. Modrego, 2005. Comparative study of cerebral white matter in autism and attention-deficit/hyperactivity disorder by means of magnetic resonance spectroscopy. Academic Radiol. 12:566–569. Fayed, N., H. Morales, P. J. Modrego, and J. Munoz-Mingarro, 2006. White matter proton MR spectroscopy in children with isolated developmental delay: Does it mean delayed myelination? Academic Radiol. 13:229–235. Filipek, P. A., M. Semrud-Clikeman, R. J. Steingard, P. F. Renshaw, D. N. Kennedy, and J. Biederman, 1997. Volumetric MRI analysis comparing subjects having attention-deficit hyperactivity disorder with normal controls. Neurology 48:589– 601. Filippi, C. G., A. M. Ulug, M. D. Deck, R. D. Zimmerman, and L. A. Heier, 2002. Developmental delay in children: Assessment with proton MR spectroscopy. AJNR Am. J. Neuroradiol. 23:882– 888. Frahm, J., and F. Hanefeld, 1997. Localized proton magnetic resonance spectroscopy of brain disorders in childhood. In H. S. Bachelard, ed., Advances in Neurochemistry, vol. 18, 329–402. New York: Plenum. Frahm, J., K.-D. Merboldt, and W. Hanicke, 1987. Localized proton spectroscopy using stimulated echoes. J. Magn. Reson. 72: 502–508. Friedman, S. D., D. W. Shaw, A. A. Artru, G. Dawson, H. Petropoulos, and S. R. Dager, 2006. Gray and white matter brain chemistry in young children with autism. Arch. Gen. Psychiatry 63:786–794. Friedman, S. D., D. W. Shaw, A. A. Artru, T. L. Richards, J. Gardner, G. Dawson, S. Posse, and S. R. Dager, 2003. Regional brain chemical alterations in young children with autism spectrum disorder. Neurology 60:100–107. Garel, C., 2006. New advances in fetal MR neuroimaging. Pediatr. Radiol. 36:621–625. Girard, N., C. Fogliarini, A. Viola, S. Confort-Gouny, Y. L. Fur, P. Viout, F. Chapon, O. Levrier, and P. Cozzone, 2006. MRS of normal and impaired fetal brain development. Eur. J. Radiol. 57:217–225. Girard, N., S. C. Gouny, A. Viola, Y. Le Fur, P. Viout, K. Chaumoitre, C. D’Ercole, C. Gire, D. Figarella-Branger, and P. J. Cozzone, 2006. Assessment of normal fetal brain maturation in utero by proton magnetic resonance spectroscopy. Magn. Reson. Med. 56:768–775. Grodd, W., I. Krageloh-Mann, U. Klose, and R. Sauter, 1991. Metabolic and destructive brain disorders in children: Findings with localized proton MR spectroscopy. Radiology 181:173– 181. Groenendaal, F., R. H. Veenhoven, J. van der Grond, G. H. Jansen, T. D. Witkamp, and L. S. de Vries, 1994. Cerebral lactate and N-acetyl-aspartate/choline ratios in asphyxiated fullterm neonates demonstrated in vivo using proton magnetic resonance spectroscopy. Pediatr. Res. 35:148–151. Hashimoto, T., M. Tayama, M. Miyazaki, E. Fujii, M. Harada, H. Miyoshi, M. Tanouchi, and Y. Kuroda, 1995. Developmental brain changes investigated with proton magnetic resonance spectroscopy. Dev. Med. Child Neurol. 37: 398–405. Hashimoto, T., M. Tayama, M. Miyazaki, Y. Yoneda, T. Yoshimoto, M. Harada, H. Miyoshi, M. Tanouchi, and Y. Kuroda, 1997. Differences in brain metabolites between patients with autism and mental retardation as detected by in vivo localized proton magnetic resonance spectroscopy. J. Child Neurol. 12:91–96.

348

methodological paradigms

Heerschap, A., R. D. Kok, and P. P. van den Berg, 2003. Antenatal proton MR spectroscopy of the human brain in vivo. Childs Nerv. Syst. 19:418–421. Heerschap, A., and P. P. van den Berg, 1994. Proton magnetic resonance spectroscopy of human fetal brain. Am. J. Obstet. Gynecol. 170:1150–1151. Herbert, M. R., 2004. Neuroimaging in disorders of social and emotional functioning: What is the question? J. Child Neurol. 19:772–784. Hesslinger, B., T. Thiel, L. Tebartz van Elst, J. Hennig, and D. Ebert, 2001. Attention-deficit disorder in adults with or without hyperactivity: Where is the difference? A study in humans using short echo (1)H-magnetic resonance spectroscopy. Neurosci. Lett. 304:117–119. Himelstein, J., J. H. Newcorn, and J. M. Halperin, 2000. The neurobiology of attention-deficit hyperactivity disorder. Front. Biosci. 5:D461–478. Hisaoka, S., M. Harada, H. Nishitani, and K. Mori, 2001. Regional magnetic resonance spectroscopy of the brain in autistic individuals. Neuroradiology 43:496–498. Horska, A., W. E. Kaufmann, L. J. Brant, S. Naidu, J. C. Harris, and P. B. Barker, 2002. In vivo quantitative proton MRSI study of brain development from childhood to adolescence. J. Magn. Reson. Imaging 15:137–143. Huppi, P. S., C. Fusch, C. Boesch, R. Burri, E. Bossi, M. Amato, and N. Herschkowitz, 1995. Regional metabolic assessment of human brain during development by proton magnetic resonance spectroscopy in vivo and by high-performance liquid chromatography/gas chromatography in autopsy tissue. Pediatr. Res. 37:145–150. Huppi, P. S., S. Posse, F. Lazeyras, R. Burri, E. Bossi, and N. Herschkowitz, 1991. Magnetic resonance in preterm and term newborns: 1H-spectroscopy in developing human brain. Pediatr. Res. 30:574–578. Jin, Z., Y. F. Zang, Y. W. Zeng, L. Zhang, and Y. F. Wang, 2001. Striatal neuronal loss or dysfunction and choline rise in children with attention-deficit hyperactivity disorder: A 1H-magnetic resonance spectroscopy study. Neurosci. Lett. 315:45–48. Kaminaga, T., M. Kobayashi, and T. Abe, 2001. Proton magnetic resonance spectroscopy in disturbances of cortical development. Neuroradiology 43:575–580. Keshavan, M. S., J. A. Stanley, and J. W. Pettegrew, 2000. Magnetic resonance spectroscopy in schizophrenia: Methodological issues and findings—Part II. Biol. Psychiatry 48:369– 380. Kok, R. D., A. J. van den Bergh, A. Heerschap, R. Nijland, and P. P. van den Berg, 2001. Metabolic information from the human fetal brain obtained with proton magnetic resonance spectroscopy. Am. J. Obstet. Gynecol. 185:1011–1015. Kok, R. D., P. P. van den Berg, A. J. van den Bergh, R. Nijland, and A. Heerschap, 2002. Maturation of the human fetal brain as observed by 1H MR spectroscopy. Magn. Reson. Med. 48:611–616. Korkman, M., and K. Peltomaa, 1991. A pattern of test findings predicting attention problems at school. J. Abnorm. Child Psychol. 19:451–467. Kreis, R., T. Ernst, and B. D. Ross, 1993. Development of the human brain: In vivo quantification of metabolite and water content with proton magnetic resonance spectroscopy. Magn. Reson. Med. 30:424–437. Kreis, R., L. Hofmann, B. Kuhlmann, C. Boesch, E. Bossi, and P. S. Huppi, 2002. Brain metabolite composition during early human brain development as measured by quantitative in vivo

1H magnetic resonance spectroscopy. Magn. Reson. Med. 48: 949–958. Lainhart, J. E., 2006. Advances in autism neuroimaging research for the clinician and geneticist. Am. J. Med. Genet. (Sem. Med. Genet.) 142:33–39. Levitt, J. G., J. O’Neill, R. E. Blanton, S. Smalley, D. Fadale, J. T. McCracken, D. Guthrie, A. W. Toga, and J. R. Alger, 2003. Proton magnetic resonance spectroscopic imaging of the brain in childhood autism. Biol. Psychiatry 54:1355–1366. MacMaster, F. P., N. Carrey, S. Sparkes, and V. Kusumakar, 2003. Proton spectroscopy in medication-free pediatric attention-deficit/hyperactivity disorder. Biol. Psychiatry 53:184– 187. Maudsley, A. A., Z. Wu, D. J. Meyerhoff, and M. W. Weiner, 1994. Automated processing for proton spectroscopic imaging using water reference deconvolution. Magn. Reson. Med. 31: 589–595. Mori, K., T. Hashimoto, M. Harada, Y. Yoneda, S. Shimakawa, E. Fujii, T. Yamaue, M. Miyazaki, T. Saijo, and Y. Kuroda, 2001. [Proton magnetic resonance spectroscopy of the autistic brain.] No To Hattatsu 33:329–335. Murphy, D. G., H. D. Critchley, N. Schmitz, G. McAlonan, T. Van Amelsvoort, D. Robertson, E. Daly, A. Rowe, A. Russell, A. Simmons, K. C. Murphy, and P. Howlin, 2002. Asperger syndrome: A proton magnetic resonance spectroscopy study of brain. Arch. Gen. Psychiatry 59:885–891. Narayana, P. A., T. J. Doyle, D. Lai, and J. S. Wolinsky, 1998. Serial proton magnetic resonance spectroscopic imaging, contrast-enhanced magnetic resonance imaging, and quantitative lesion volumetry in multiple sclerosis. Ann. Neurol. 43:56– 71. Naressi, A., C. Couturier, I. Castang, R. de Beer, and D. Graveron-Demilly, 2001. Java-based graphical user interface for MRUI, a software package for quantitation of in vivo/ medical magnetic resonance spectroscopy signals. Comput. Biol. Med. 31:269–286. National Institute of Mental Health (NIMH), 1996 [revised 2003]. Attention Deficit Hyperactivity Disorder. Bethesda, MD: National Institute of Mental Health, National Institutes of Health, US Department of Health and Human Services. (NIH Publication Number NIH 3572.) 48 pages. Available from http://www. nimh.nih.gov/publicat/adhd.cfm. O’Neill, J., J. Levitt, R. Caplan, R. Asarnow, J. T. McCracken, A. W. Toga, and J. R. Alger, 2004. 1H MRSI evidence of metabolic abnormalities in childhood-onset schizophrenia. NeuroImage 21:1781–1789. Otsuka, H., M. Harada, K. Mori, S. Hisaoka, and H. Nishitani, 1999. Brain metabolites in the hippocampus-amygdala region and cerebellum in autism: An 1H-MR spectroscopy study. Neuroradiology 41:517–519. Peden, C. J., F. M. Cowan, D. J. Bryant, J. Sargentoni, I. J. Cox, D. K. Menon, D. G. Gadian, J. D. Bell, and L. M. Dubowitz, 1990. Proton MR spectroscopy of the brain in infants. J. Comput. Assist. Tomogr. 14:886–894. Peden, C. J., M. A. Rutherford, J. Sargentoni, I. J. Cox, D. J. Bryant, and L. M. Dubowitz, 1993. Proton spectroscopy of the neonatal brain following hypoxic-ischaemic injury. Dev. Med. Child Neurol. 35:502–510. Penrice, J., E. B. Cady, A. Lorek, M. Wylezinska, P. N. Amess, R. F. Aldridge, A. Stewart, J. S. Wyatt, and E. O. Reynolds, 1996. Proton magnetic resonance spectroscopy of the brain in normal preterm and term infants, and early changes after perinatal hypoxia-ischemia. Pediatr. Res. 40:6–14.

Posse, S., C. DeCarli, and D. Le Bihan, 1994. Three-dimensional echo-planar MR spectroscopic imaging at short echo times in the human brain. Radiology 192:733–738. Pouwels, P. J., K. Brockmann, B. Kruse, B. Wilken, M. Wick, F. Hanefeld, and J. Frahm, 1999. Regional age dependence of human brain metabolites from infancy to adulthood as detected by quantitative localized proton MRS. Pediatr. Res. 46:474–485. Provencher, S. W., 1993. Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn. Reson. Med. 30:672–679. Robertson, N. J., I. J. Cox, F. M. Cowan, S. J. Counsell, D. Azzopardi, and A. D. Edwards, 1999. Cerebral intracellular lactic alkalosis persisting months after neonatal encephalopathy measured by magnetic resonance spectroscopy. Pediatr. Res. 46: 287–296. Robertson, N. J., R. H. Lewis, F. M. Cowan, J. M. Allsop, S. J. Counsell, A. D. Edwards, and I. J. Cox, 2001. Early increases in brain myo-inositol measured by proton magnetic resonance spectroscopy in term infants with neonatal encephalopathy. Pediatr. Res. 50:692–700. Roelants–Van Rijn, A. M., J. van der Grond, L. S. de Vries, and F. Groenendaal, 2001. Value of (1)H-MRS using different echo times in neonates with cerebral hypoxia-ischemia. Pediatr. Res. 49:356–362. Ross, B., and S. Bluml, 2001. Magnetic resonance spectroscopy of the human brain. Anat. Rec. 265:54–84. Rubia, K., S. Overmeyer, E. Taylor, M. Brammer, S. C. Williams, A. Simmons, and E. T. Bullmore, 1999. Hypofrontality in attention deficit hyperactivity disorder during higher-order motor control: A study with functional MRI. Am. J. Psychiatry 156:891–896. Scarabino, T., T. Popolizio, A. Bertolino, and U. Salvolini, 1999. Proton magnetic resonance spectroscopy of the brain in pediatric patients. Eur. J. Radiol. 30:142–153. Seidman, L. J., E. M. Valera, and N. Makris, 2005. Structural brain imaging of attention-deficit/hyperactivity disorder. Biol. Psychiatry 57:1263–1272. Shevell, M. I., 1998. The evaluation of the child with a global developmental delay. Sem. Pediatr. Neurol. 5:21–26. Shevell, M., S. Ashwal, D. Donley, J. Flint, M. Gingold, D. Hirtz, A. Majnemer, M. Noetzel, and R. D. Sheth, 2003. Practice parameter: Evaluation of the child with global developmental delay: Report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 60:367– 380. Skounti, M., A. Philalithis, and E. Galanakis, 2007. Variations in prevalence of attention deficit hyperactivity disorder worldwide. Eur. J. Pediatr. 166:117–123. Soher, B. J., K. Young, V. Govindaraju, and A. A. Maudsley, 1998. Automated spectral analysis. III. Application to in vivo proton MR spectroscopy and spectroscopic imaging. Magn. Reson. Med. 40:822–831. Stanley, J. A., 2002. In vivo magnetic resonance spectroscopy and its application to neuropsychiatric disorders. Can. J. Psychiatry 47:315–326. Sturman, J. A., 1993. Taurine in development. Physiol. Rev. 73: 119–147. Sun, L., Z. Jin, Y. F. Zang, Y. W. Zeng, G. Liu, Y. Li, L. J. Seidman, S. V. Faraone, and Y. F. Wang, 2005. Differences between attention-deficit disorder with and without hyperactivity: A 1H-magnetic resonance spectroscopy study. Brain. Dev. 27:340–344.

sajja and narayana: magnetic resonance spectroscopy

349

Teicher, M. H., C. M. Anderson, A. Polcari, C. A. Glod, L. C. Maas, and P. F. Renshaw, 2000. Functional deficits in basal ganglia of children with attention-deficit/hyperactivity disorder shown with functional magnetic resonance imaging relaxometry. Nature Med. 6:470–473. Tzika, A. A., D. B. Vigneron, W. S. Ball, Jr., R. S. Dunn, and D. R. Kirks, 1993. Localized proton MR spectroscopy of the brain in children. J. Magn. Reson. Imaging 3:719–729. Vaidya, C. J., G. Austin, G. Kirkorian, H. W. Ridlehuber, J. E. Desmond, G. H. Glover, and J. D. Gabrieli, 1998. Selective effects of methylphenidate in attention deficit hyperactivity disorder: A functional magnetic resonance study. Proc. Natl. Acad. Sci. USA 95:14494–14499. van der Knapp, M., B. Ross, and J. Valk, 1994. Uses of MR in inborn errors of metabolism. In J. Kucharczyk et al., eds., Magnetic Resonance Neuroimaging, 246–309. Boca Raton, FL: CRC Press. van der Knaap, M. S., J. van der Grond, P. C. van Rijen, J. A. Faber, J. Valk, and K. Willemse, 1990. Age-dependent changes in localized proton and phosphorus MR spectroscopy of the brain. Radiology 176:509–515. van Elst, L. T., M. R. Trimble, D. Ebert, and L. T. van Elst, 2001. Dual brain pathology in patients with affective aggressive episodes. Arch. Gen. Psychiatry 58:1187–1188. Vigneron, D. B., 2006. Magnetic resonance spectroscopic imaging of human brain development. Neuroimaging Clin. N. Am. 16:75–85: viii.

350

methodological paradigms

Vigneron, D. B., A. J. Barkovich, S. M. Noworolski, M. von dem Bussche, R. G. Henry, Y. Lu, J. C. Partridge, G. Gregory, and D. M. Ferriero, 2001. Three-dimensional proton MR spectroscopic imaging of premature and term neonates. AJNR Am. J. Neuroradiol. 22:1424–1433. Wolinsky, J. S., and P. A. Narayana, 2002. Magnetic resonance spectroscopy in multiple sclerosis: Window into the diseased brain. Curr. Opin. Neurol. 15:247–251. Wright, I. C., S. Rabe-Hesketh, P. W. Woodruff, A. S. David, R. M. Murray, and E. T. Bullmore, 2000. Meta-analysis of regional brain volumes in schizophrenia. Am. J. Psychiatry 157: 16–25. Yeo, R. A., D. Hill, R. Campbell, J. Vigil, and W. M. Brooks, 2000. Developmental instability and working memory ability in children: A magnetic resonance spectroscopy investigation. Dev. Neuropsychol. 17:143–159. Yeo, R. A., D. E. Hill, R. A. Campbell, J. Vigil, H. Petropoulos, B. Hart, L. Zamora, and W. M. Brooks, 2003. Proton magnetic resonance spectroscopy investigation of the right frontal lobe in children with attention-deficit/hyperactivity disorder. J. Am. Acad. Child Adolesc. Psychiatry 42:303– 310. Zeegers, M., J. van der Grond, E. van Daalen, J. Buitelaar, and H. van Engeland, 2007. Proton magnetic resonance spectroscopy in developmentally delayed young boys with or without autism. J. Neural Transm. 114:289–295.

21

The Integration of Neuroimaging and Molecular Genetics in the Study of Developmental Cognitive Neuroscience ESSI VIDING, DOUGLAS E. WILLIAMSON, ERIKA E. FORBES, AND AHMAD R. HARIRI

Recent advances in both molecular genetics and human neuroimaging have begun to provide the tools necessary to explore how individual differences in complex cognitive and emotional behaviors emerge and how such differences may confer vulnerability to psychopathology. With publication of the reference human genome sequence (Lander et al., 2001; Venter et al., 2001), a major effort is under way to exhaustively identify common variations in this sequence that impact on gene function (i.e., functional polymorphisms) and subsequently to understand how such functional variations alter human biology. Since approximately 70 percent of all genes are expressed in the brain, many of these functional polymorphisms will influence how the brain processes information and, as a consequence, regulates both cognitive and affective behaviors. Human neuroimaging (e.g., MRI, fMRI, EEG, MEG, PET), because of its capacity to assay detailed brain structure and function within individuals, has unique potential as a tool for characterizing functional genetics in neural circuitry. The goals of this chapter are to (1) describe the conceptual basis for imaging genetics; (2) outline some major findings in imaging genetics to highlight the effectiveness of this strategy in delineating biological pathways and mechanisms by which individual differences in brain function emerge and potentially bias risk for psychiatric illness; (3) discuss the importance of applying the imaging genetics framework to study development and the emergence of psychopathology; (4) illustrate designs of developmental imaging genetics studies; and (5) summarize the anticipated short- and long-term rewards of this program of research, with an emphasis on the translational impact of this approach. As a first step, we will consider behavior association studies relevant to the genetics of emotion regulation in healthy and clinical populations. These studies, with their challenges and limitations, provide a foundation for the imaging genetics approach,

which aims to examine mechanisms leading to individual variability in emotion regulation that is more proximal to genetic factors.

Behavior genetics Why Study Genes? Genes represent the “go” square on the monopoly board of life. They are the biological toolbox with which one negotiates the environment. While most human behaviors including emotion regulation cannot be explained by genes alone, and certainly much variance in aspects of brain information processing will not be genetically determined, variations in genetic sequence that impact gene function will contribute some variance to these more complex brain and behavioral phenotypes. This conclusion is implicit in the results of studies of twins, which have revealed heritabilities of 40–70 percent for various aspects of cognition, temperament, and personality (McGuffin, Riley, and Plomin, 2001). Our proposed approach to understanding the nature of individual differences in emotion regulation revolves around genes because these constructs have unparalleled potential impact on all levels of biology. In the context of disease states, particularly behavioral disorders, genes not only transcend phenomenological diagnosis, but also represent mechanisms of disease. Moreover, genes offer the potential to identify at-risk individuals and biological pathways for the development of new treatments. In the case of psychiatric illness, genes appear to be the only consistent risk factors that have been identified across populations, and the lion’s share of susceptibility to major psychiatric disorders is accounted for by inheritance (Moldin and Gottesman, 1997). Though the strategy for finding susceptibility genes for complex disorders, by traditional linkage and association methods, may seem relatively straightforward (albeit not easily achieved), developing a useful and comprehensive understanding of the

351

mechanisms by which such genes increase biological risk is a much more daunting challenge. How many genes contribute to a particular complex behavior or complex disease state? What genetic overlap exists across behaviors and diseases? How large are the effects of candidate genes on particular brain functions? And, perhaps most importantly, how does a gene affect brain information processing to increase risk for a disorder of behavior? Traditional Association Studies The “candidate gene association approach” has been a particularly popular strategy for attempting to answer these questions. Genetic association is a test of a relationship between a particular phenotype and a specific allele of a gene. This approach usually begins with selecting a biological aspect of a particular condition or disease, then identifying variants in genes thought to impact on the candidate biological process, and next searching for evidence that the frequency of a particular variant (“allele”) is increased in populations having the disease or condition. A significant increase in allele frequency in the selected population is evidence of association. When a particular allele is significantly associated with a particular phenotype, it is potentially a causative factor in determining that phenotype. There are caveats to the design and interpretation of genetic association studies, such as linkage disequilibrium with other loci and ancestral stratification, that are beyond the scope of this review and have been discussed at length elsewhere (Emahazion et al., 2001). A starting point for personality or psychiatric genetics is the findings of genetic influence through traditional behavioral genetics studies such as twin studies. By comparing the similarity of monozygotic and dizygotic twins on individual difference variables such as extraversion, such studies examine the extent to which variability is due to heritable (i.e., genetic) factors or to environmental factors. These studies indicate that genetic influences shape personality traits and do so across the life span (Nigg and Goldsmith, 1998; Plomin and Nesselroade, 1990; Rutter and Plomin, 1997; Viken et al., 1994). From these broad estimates of heritability stem more circumscribed association studies of personality that attempt to identify relationships between specific allelic variants or genotypes of distinct genes and personality or temperament factors. These studies, which take a molecular genetics approach, identify candidate genes of interest and examine their relation to a phenotype. The candidate genes included in these studies are typically those related to the patterns of brain function that putatively underlie personality style. Association Studies of Genes Involved in Emotion Regulation Behavioral and molecular genetics approaches have not been applied to questions of particular emotionregulation responses as defined in studies of behavior or

352

methodological paradigms

physiology. For instance, it would be a stretch to examine the behavioral, cognitive, or physiological components of the emotion-regulation strategy of situation modification in relation to a specific gene variant. Instead, typical research in genetics has addressed the association between genes and proxy variables for emotion regulation. These proxy variables represent broad individual differences in emotional style or tendency and have generally been in the areas of personality or affective disorders. Although these variables are related to emotion-regulation constructs, they are more broad and heterogeneous. Behavioral and molecular genetics approaches have been applied to two topics that are relevant to stable emotion regulatory style: personality and affective disorders. Personality refers to stable normal individual differences, many of which pertain to emotional experience and expression. Affective disorders, while more in the realm of abnormal emotional experience, can be considered examples of pathological emotion dysregulation. These disorders—which include intense and long-duration depressed, manic, or anxious emotional states—involve reduced emotional flexibility. Presumably, difficulty with modulating the frequency, intensity, or duration of affective states underlies these disorders. For example, depression is characterized by sustained sadness and unusually low-frequency, low-intensity positive affect. The genes that predispose people to experience the disorders therefore may constitute genetic influences on effective, healthy emotion regulation. Molecular genetics approaches to emotion regulation often focus on polymorphisms leading to variability in neurotransmitter availability or neurotransmitter receptor function. For example, extraversion’s characteristics of dominance, novelty seeking, and reward sensitivity are thought to be driven by variability in function of the dopamine (DA) system. There are many neurotransmitter systems, each of which has a complex function and influence on brain and behavior. In addition, the influence of the various neurotransmitter systems on emotion regulation is presumably complex and interrelated. Research on the association between neurotransmitter genes and emotion-regulationrelated characteristics has focused on narrow aspects of specific neurotransmitter systems. Two particular systems appear to be especially relevant to questions of emotion regulation, however: the serotonin (5-HT) system and the dopamine system. Serotonin has been implicated in the generation and regulation of emotional behavior (Lucki, 1998), and manipulation of serotonin activity has effects on behaviors such as impulsivity and aggression (Manuck et al., 1999). The dopamine system plays a critical role in reward processing and has been linked to normal individual differences in reward traits (Depue and Collins, 1999) as well as to disorders involving enhanced reward seeking such as addiction (Kalivas and Volkow, 2005).

We will address both neurotransmitter systems in our review of association studies, which follows, and we will focus specifically on the 5-HT system in our discussion of imaging genetics in the remainder of the chapter. In addition, while we address genetic factors in both normal and abnormal individual differences in the review of association studies, we emphasize normal individual differences in our treatment of imaging genetics. As we will explain, the conceptual foundation for imaging genetics lends itself best to first examining normal variability in neural function. Personality Extraversion is likely to involve approach toward goals despite setbacks and assertion that serves to modify a current situation. Consequently, studies of the genetic underpinnings of extraversion have focused on polymorphisms related to DA function (Ebstein et al., 2002; Noblett and Coccaro, 2005). Specifically, genetic variants of DA receptor subtypes, such as the D2 and D4 receptors, which mediate the myriad neuromodulatory effects of DA, as well as the dopamine transporter, which facilitates the active reuptake of DA from the extracellular space, have been examined in relation to the broad trait of extraversion and to one of its facets, novelty seeking. More recently, studies have begun to examine other genes that influence broader DA and other catecholamine availability, including catechol-O-methyltransferase (COMT) and monoamine oxidase A (MAOA). As recent reviews and meta-analyses have noted, the associations between specific DA polymorphisms and complex measures of personality have been inconsistent across studies, with null findings relatively common (Ebstein, 2006; Schinka, Letsch, and Crawford, 2002; Strobel et al., 2003). Another significant line of related research from the field of personality genetics is the examination of serotonin (5-HT) subsystem polymorphisms on negative emotional behaviors such as neuroticism, impulsivity, and aggression. A gene of particular interest has been a relatively frequent length variant in the promoter or regulatory region of the 5-HT transporter (5-HTT) gene. Numerous studies have indicated that the short (S) variant of this gene, resulting in relatively reduced 5-HTT availability, is associated with higher levels of temperamental anxiety. Other investigators have established links between variation in 5-HT genes controlling biosynthesis, receptor function, and metabolic degradation with additional dimensional measures of negative emotionality such as impulsive aggression (Manuck et al., 1999, 2000) and suicidality (Bellivier, Chaste, and Malafosse, 2004). Despite some replication, these lines of investigation have also been marked by null findings (Glatt and Freimer, 2002), with several reports, including meta-analyses, emphasizing that the ability to detect associations depends on the personality instruments used, with “broad bandwidth” personality measures (e.g., extraversion) typi-

cally representing constructs that are too heterogeneous to map meaningfully onto biological systems (Munafo, Clark, and Flint, 2005). Affective Disorders The leap from studies of genetic influences on dimensional indices of normal variability in personality and temperament to studies of genetic influences on affective disorders such as depression and anxiety is understandable given the correlation of these indices with symptoms of these disorders and the genetic influences on such correlations (Carey and DiLalla, 1994). For example, depression and the personality trait of neuroticism appear to share genetic influence, and in addition, the correlation between depression and neuroticism appears to be influenced by genetic factors (Kendler et al., 1993). Such attempts to link polymorphisms directly with clinical syndromes has been fueled by the suggestion that genes might have more detectable influence at extreme, pathological ends of the emotional trait distribution. While any specific gene in isolation is unlikely to serve as a predisposition to a complex disorder such as major depressive disorder, the influence of a particular gene is more likely to be detected in a clinical population than in individuals with lower levels of the emotional dysfunction involved in the disorder. If neuroticism and depression share genetic influence (Kendler et al., 1993) and if depression can be seen as an extreme version of high neuroticism, then influences of 5-HT polymorphisms, for instance, may be more clear when depression is the target construct. Studies of genetic influences on depression and anxiety in humans have emphasized the role of genes related to 5-HT and hypothalamic-pituitary-adrenal (HPA) axis function (see Leonardo and Hen, 2006, for a more thorough review). Both the 5-HT and HPA systems play a critical role in emotional reactivity and regulation and are thus prime candidates for studies of these mood disorders. Many candidate polymorphisms in these systems have been linked to increased risk for mood disorders. Moreover, the existence of an association has been demonstrated to be moderated by the environment. In particular, social stress, such as maltreatment during childhood or divorce in adulthood, appears to unmask genetic vulnerability for depression and anxiety. Limitations of Behavioral Association Studies All the findings from traditional behavioral association studies have been inconsistent, with an impressive number of null findings for each gene studied. In many ways, this inconsistency underscores the argument that in the context of behavior and psychiatric illness there are only susceptibility genes and not disease genes, which clearly and specifically determine affective disorders. Association studies have important limitations, not least of which is the long chain of events

viding et al.: neuroimaging and molecular genetics

353

from gene function to personality or psychiatric disorder. Additional limitations include the specificity of findings to particular personality instruments, the reliance on self-report rather than observed behavior, the failure to account for developmental effects, and the difficulties of defining and examining gene-by-environment effects.

Conceptual basis of imaging genetics Genes have unparalleled potential impact on all levels of biology. In the context of disease states, particularly behavioral disorders, genes represent the cornerstone of mechanisms that either directly or in concert with environmental events ultimately result in disease. Moreover, genes offer the potential to identify at-risk individuals and biological pathways for the development of new treatments. While most human behaviors cannot be explained by genes alone, and certainly much variance in aspects of brain information processing will not be genetically determined directly, it is anticipated that variations in genetic sequence that impact gene function will contribute an appreciable amount of variance to these resultant complex behavioral phenomena. This conclusion is implicit in the results of studies of twins that have revealed heritabilities ranging from 40 to 70 percent for various aspects of cognition, temperament, and personality (Plomin, Owen, and McGuffin, 1994). In the case of psychiatric illness, genes appear to be the only consistent risk factors that have been identified across populations, and the majority of susceptibility for major psychiatric disorders is accounted for by inheritance (Moldin and Gottesman, 1997). Traditionally, the impact of genetic polymorphisms on human behavior has been directly examined using clinical evaluations, personality questionnaires, and neuropsychological batteries. Genetic epidemiological investigations have directly examined the relationship between specific genetic polymorphisms and behaviors and have reported equivocal results (Malhotra and Goldman, 1999). These are not surprising for at least two reasons. First, there is considerable individual variability in dimensions of observable behavior as well as subjectivity in the assessment of behavior necessitating very large samples, often exceeding several hundred subjects, to identify even small gene effects (Glatt and Freimer, 2002). Moreover, it is apparent that there are etiological subgroups within any given disease that obscure effects at the broader group level. Second and perhaps most importantly, the effects of genes are not expressed directly at the level of behavior. As we will discuss in detail, gene effects on behavior are mediated by their molecular and cellular effects on information processing in brain. Thus examining gene effects on brain represents a critical step in understanding their ultimate contribution to variability in behavior.

354

methodological paradigms

Since genes are directly involved in the development and function of brain regions subserving specific cognitive and emotional processes, functional polymorphisms in genes may be strongly related to the function of these specific neural systems and, in turn, mediate/moderate their involvement in behavioral outcomes. This is the underlying assumption of our investigations examining the relation between genes and neural systems, which we initially called imaging genomics (Hariri and Weinberger, 2003b) and more recently describe as imaging genetics (Hariri, Drabent, and Weinberger, 2006), because this approach is utilized to explore variation in specific genes and not the genome broadly. The potential for marked differences at the neurobiological level underscores the need for a direct assay of brain function. Accordingly, imaging genetics within the context of a “candidate gene association approach” provides an ideal opportunity to further our understanding of biological mechanisms potentially contributing to individual differences in behavior and personality. Moreover, imaging genetics provides a unique tool with which to explore and evaluate the functional impact of brain-relevant genetic polymorphisms with the potential to understand their impact on behavior. Of course, the relevance of imaging genetics findings for disease vulnerability will only be made once the variants under study are further associated with disease risk directly or if their impact on brain function is manifest (or even exaggerated) in the diseases of interest. Neuroimaging techniques, especially those that are noninvasive like MRI and EEG/MEG, typically require no more than a few minutes of subject participation to acquire substantial data sets, reflecting the acquisition of many hundreds of repeated measures of brain structure or function within a single subject. The efficiency of these techniques allows for the ability to investigate the specificity of gene effects by examining their influence on multiple functional systems (e.g., prefrontal, striatal, limbic) in a single subject in one experimental session. This capacity to rapidly assay differences in brain structure and information processing with enhanced power and sensitivity places neuroimaging at the forefront of available tools for the in vivo study of functional genetic variation. The protocol for imaging genetics typically involves first identifying a meaningful variation in the DNA sequence within a candidate gene. For the variant to be meaningful, it should have an impact at the molecular and cellular levels in gene or protein function (i.e., be a functional variant), and the distribution of such effects at the level of brain systems involved in specific forms of information processing should be predictable. For example, a genetic variation in the gene for the serotonin transporter that impacts the availability of synaptic serotonin would be expected to affect amygdala function because serotonin is important in amygdala physiology (discussed later). Alternatively, recent imaging genetics

studies have taken the lead in exploring the functionality of candidate variants by first describing in vivo effects at the level of brain systems (Brown et al., 2005). As such, imaging genetics can provide the initial impetus for further characterization of molecular and functional effects of specific candidate genes in brain systems thought to be involved in behavior. In this manner, the contributions of abnormalities in these systems to complex behaviors and emergent phenomena, possibly including psychiatric syndromes, can then be understood from the perspective of their biological origins.

Selected overview of imaging genetics findings Converging evidence from animal and human studies has revealed that serotonin (5-hydroxytryptamine; 5-HT) is a critical neuromodulator in the generation and regulation of emotional behavior (Lucki, 1998). Serotonergic neurotransmission has also been an efficacious target for the pharmacological treatment of mood disorders including depression, obsessive-compulsive disorder, anxiety, and panic (Blier and de Montigny, 1999). Moreover, genetic variation in several key 5-HT subsystems, presumably resulting in altered central serotonergic tone and neurotransmission, has been associated with various aspects of personality and temperament (Munafo, Clark, and Flint, 2005; Schinka, Busch, and Robichaux-Keene, 2004; Sen, Burmeister, and Ghosh, 2004) as well as susceptibility to affective illness (Murphy et al., 1998; Reif and Lesch, 2003). However, enthusiasm for the potential of such genetic variation to affect behaviors and especially disease liability has been tempered by weak, inconsistent, and failed attempts at replication of specific associations with psychiatric syndromes (Glatt and Freimer, 2002). The inability to substantiate such relationships through consistent replication in independent cohorts may simply reflect methodological issues such as inadequate control for population stratification, insufficient power, and inconsistency in the methods applied. Alternatively, and perhaps more importantly, such inconsistency may reflect the underlying biological nature of the relationship between allelic variants in serotonin genes, each of presumably small effect, and observable behaviors in the domain of mood and emotion that typically reflect complex functional interactions and emergent phenomena. Given that the biological impact of variation in a gene traverses an increasingly divergent path from cells to neural systems to behavior, the response of brain regions subserving emotional processes in humans (e.g., amygdala, hippocampus, prefrontal cortex, anterior cingulate gyrus) represents a critical first step in their impact on behavior. Thus functional polymorphisms in 5-HT genes may be strongly related to the integrity of these underlying neural systems and mediate/moderate

their ultimate effect on behavior (Hariri and Weinberger, 2003a). The 5-HT transporter plays an important role in serotonergic neurotransmission by facilitating reuptake of 5-HT from the synaptic cleft. In 1996 a relatively common polymorphism was identified in the human 5-HTT gene located on chromosome 17q11.1-q12 (Heils et al., 1996). The polymorphism is a variable repeat sequence in the promoter region (5-HTTLPR) resulting in two common alleles: the short (S) variant comprising 14 copies of a 20–23-base-pair repeat unit and the long (L) variant comprising 16 copies. In populations of European ancestry, the frequency of the S allele is approximately 0.40, and the genotype frequencies are in Hardy-Weinberg equilibrium (L/L = 0.36, L/S = 0.48, S/S = 0.16). These relative allele frequencies, however, can vary substantially across populations (Gelernter, Kranzler, and Cubells, 1997). Following the identification of this polymorphism, Lesch and colleagues demonstrated in vitro that the 5-HTTLPR alters both gene transcription and level of 5-HTT function (Lesch et al., 1996). Cultured human lymphoblast cell lines homozygous for the long allele have higher concentrations of 5-HTT mRNA and express nearly twofold greater 5-HT reuptake in comparison to cells possessing either one or two copies of the short allele. Subsequently, both in vivo imaging measures of radioligand binding to 5-HTT (Heinz et al., 2000) and postmortem calculation of 5-HTT density (Little et al., 1998) in humans reported nearly identical reductions in 5-HTT binding levels associated with the short allele as observed in vitro (but see Patkar et al., 2004; Shioe et al., 2003; van Dyck et al., 2004). These data are consistent with ß-CIT SPECT studies in humans and nonhuman primates reporting an inverse relationship between 5-HTT availability and CSF concentrations of 5-hydroxyindoleacetic acid (5-HIAA), a 5-HT metabolite (Heinz et al., 1998, 2002), and indicate that the 5-HTTLPR is functional and impacts on serotonergic neurotransmission. In their initial study, Lesch and colleagues also demonstrated that individuals carrying the short allele are slightly more likely to display abnormal levels of anxiety in comparison to L/L homozygotes (Lesch et al., 1996). Since their original report, others have confirmed the association between the 5-HTTLPR short allele and heightened anxiety (Du, Bakish, and Hrdina, 2000; Katsuragi et al., 1999; Mazzanti et al., 1998; Melke et al., 2001), and have also demonstrated that individuals possessing the short allele more readily acquire conditioned fear responses (Garpenstrand et al., 2001) and develop affective illness (Lesch and Mossner, 1998) in comparison to those homozygous for the long allele. Recent studies utilizing pharmacological challenge paradigms of the 5-HT system suggest that these differences in affect, mood, and temperament may reflect 5-HTTLPRdriven variation in 5-HTT expression and subsequent

viding et al.: neuroimaging and molecular genetics

355

changes in synaptic concentrations of 5-HT (Moreno et al., 2002; Neumeister et al., 2002; Whale, Clifford, and Cowen, 2000). Furthermore, reduced 5-HTT availability, as putatively indexed by the 5-HTTLPR short allele, has been associated with mood disturbances including major depression (Caspi et al., 2003; Malison et al., 1998) and the severity of depression and anxiety in various psychiatric disorders (Eggers et al., 2003; Heinz et al., 2002; Willeit et al., 2000). Intriguingly, it appears that exposure to stressful life events moderates the impact of the 5-HTTLPR for the development of depression (Caspi et al., 2003). The amygdala is a central brain structure in the generation of both normal and pathological emotional behavior, especially fear (LeDoux, 2000). Furthermore, the amygdala is densely innervated by serotonergic neurons, and 5-HT receptors are abundant throughout amygdala subnuclei (Azmitia and Gannon, 1986; Sadikot and Parent, 1990; Smith et al., 1999). Thus the activity of this subcortical region may be uniquely sensitive to alterations in serotonergic neurotransmission, and any resulting variability in amygdala excitability is likely to contribute to individual differences in emergent phenomena such as mood and temperament. However, it is essential to appreciate the importance of a distributed and interconnected network of cortical and subcortical brain regions for the generation, integration, and modulation of emotional behavior. Results from a series of landmark imaging studies (Beauregard, Levesque, and Bourgouin, 2001; Hariri, Bookheimer, and Mazziotta, 2000; Keightley et al., 2003; Lange et al., 2003; Nakamura et al., 1998; Narumoto et al., 2000) suggest that the dynamic interactions of the amygdala and prefrontal cortex may be critical in regulating emotional behavior (Hariri et al., 2003). Although the potential influence of genetic variation in 5-HTT function on human mood and temperament was bolstered by subsequent studies demonstrating increased anxiety-like behavior and abnormal fear conditioning in 5HTT knockout mice (Holmes et al., 2003), the underlying neurobiological correlates of this functional relationship remain unknown. Because the physiologic response of the amygdala during the processing of fearful or threatening stimuli temporally precedes the subjective experience of emotionality, the 5-HTTLPR may have a more obvious impact at the level of amygdala biology. In 2002 our research group at the NIMH utilized an imaging genetics strategy with fMRI to directly explore the neural basis of the apparent relationship between the 5HTTLPR and emotional behavior (Hariri et al., 2002b). Specifically, we hypothesized that 5-HTTLPR short-allele carriers, who presumably have relatively lower 5-HTT function and higher synaptic concentrations of 5-HT (analogous to the 5-HTT knockout mice) and have been reported to be more anxious and fearful, would exhibit greater amygdala activity in response to fearful or threatening stimuli than

356

methodological paradigms

those homozygous for the long allele, who presumably have lower levels of synaptic 5-HT and have been reported to be less anxious and fearful (analogous to the contrasting wild type mice). In our initial study, subjects from two independent cohorts (n = 14 in each) were divided into equal groups based on their 5-HTTLPR genotype, with the groups matched for age, gender, IQ, and task performance. During scanning, the subjects performed a simple perceptual processing task involving the matching of fearful and angry human facial expressions. Importantly, this task has been effective at consistently engaging the amygdala across multiple subject populations and experimental paradigms (Hariri, Bookheimer, and Mazziotta, 2000; Hariri et al., 2002a; Hariri, Tessitore, et al., 2002; Tessitore et al., 2002). Consistent with our hypothesis, we found that subjects carrying the less efficient 5-HTTLPR short allele exhibited significantly increased amygdala activity in comparison with subjects homozygous for the L allele (Hariri et al., 2002b). In fact, the difference in amygdala activity between 5-HTTLPR genotype groups in this study was nearly fivefold, accounting for 20 percent of the total variance in the amygdala response. This initial finding suggested that the increased anxiety and fearfulness associated with individuals possessing the 5HTTLPR S allele may reflect the hyperresponsiveness of their amygdala to relevant environmental stimuli. Recently, five independent functional imaging studies have reported identical 5-HTTLPR S allele–driven amygdala hyperreactivity in cohorts of healthy German (Heinz et al., 2005), Italian (Bertolino et al., 2005), and American (Canli et al., 2005) adult volunteers as well as Dutch patients with social phobia (Furmark et al., 2004) and German patients with panic disorder (Domschke et al., 2005). Moreover, we have also replicated our initial finding of 5HTTLPR S effects on amygdala reactivity in a large, independent cohort of adult volunteers (n = 92) who were carefully screened to exclude individuals with a past history of psychiatric illness or treatment. We again observed that 5-HTTLPR S allele carriers exhibited significantly increased right amygdala activation in response to our fMRI challenge paradigm (Hariri et al., 2005). In addition, our latest data revealed that 5-HTTLPR S allele–driven amygdala hyperresponsivity is equally pronounced in both sexes and independent of S allele load. The equivalent effect of one or two S alleles on amygdala function is consistent with the original observations of Lesch and colleagues (1996) on the influence of the 5-HTTLPR on in vitro gene transcription efficiency and subsequent 5-HTT availability. The absence of sex differences suggests that the increased prevalence of mood disorders in females may be related to factors other than the direct risk effect of the 5-HTTLPR S allele. A third study from the large NIMH cohort (n = 114) further captured the dynamic effects of the 5-HTTLPR on

genes, brain, and behavior by examining effects on brain structure and corticolimbic functional connectivity (Pezawas et al., 2005). We found that, in comparison to the LL genotype subjects, S allele carriers showed significantly reduced gray matter volume of the perigenual anterior cingulate cortex (pACC) and amygdala. Moreover, there was a positive correlation (“structural covariation”) between amygdala and pACC volume, and S carriers showed significantly lower structural covariation between amygdala and subgenual anterior cingulate than LL individuals. This finding suggests that pACC and amygdala represent a functional circuit the morphological development of which is modulated by genetic variation in the serotonergic system. We next explored the impact of these observed structural effects on the functional interactions of the amygdala and pACC in our fMRI data set. Independent of 5-HTTLPR genotype status, we found that the amygdala and pACC were significantly functionally connected. Two distinct regions of functional connectivity were identified within the pACC—a positive coupling between the amygdala and the subgenual cingulate, and a negative coupling between the amygdala and the supragenual cingulate. This pattern of functional connectivity is consistent with anatomical tracing studies in nonhuman primates that have defined a feedback circuit from amygdala to rostral cingulate and then from dorsal cingulate back to amygdala. These intrinsic cingulate regions also showed strong positive connectivity with each other, suggesting that the corticolimbic feedback loop is closed by means of local processing within the cingulate cortex. This intrinsic cingulate connection also is consistent with anatomical studies in nonhuman primates. Remarkably, 5-HTTLPR S allele carriers showed a significant reduction of amygdala-pACC functional connectivity in comparison to LL homozygotes. This difference was most pronounced in the coupling of the amygdala and subgenual ACC. These findings suggest that a disruption of this amygdala-pACC feedback circuitry could underlie the earlier observation of increased amygdala activity in S carriers during the processing of biologically salient stimuli (Hariri et al., 2005, 2002b). More specifically, the data suggest that the overactivation of the amygdala associated with the 5HTTLPR short allele may reflect more a relative failure of regulation of the amygdala response than an abnormal primary response per se. Taken together these data show that 5-HTTLPR genotype affects the structure and putative wiring of a core region within the limbic system thought to be crucial for anxiety-related temperamental traits and depression (Mayberg, 2003a, 2003b; Phillips et al., 2003a, 2003b). The collective results of these imaging genetics studies reveal that the 5-HTTLPR S allele has a robust effect on human amygdala structure and function, as well as the functional interactions of corticolimbic circuitry implicated in

both normal and pathological mood states. Importantly, the absence of group differences in age, gender, IQ, and ethnicity in each of these studies indicates that the observed effects are not likely due to a bias resulting from population stratification. Rather, the data suggest that heritable variation in 5-HT signaling associated with the 5-HTTLPR results in structural alterations of the amygdala and pACC, accompanied by biased amygdala reactivity and functional coupling with pACC in response to salient environmental cues. Furthermore, the emergence of these effects in samples of ethnically matched volunteers carefully screened to exclude any lifetime history of psychiatric illness or treatment argues that they represent genetically determined biological traits that are not altered by the presence of a psychiatric illness. In contrast to the striking imaging genetics findings of 5HTTLPR short-allele effects on amygdala reactivity and limbic circuitry dynamics, initial attempts to link these effects on brain function with measures of emergent behavioral phenomena, namely, the personality trait of harm avoidance, have failed to detect any significant direct relationships. Specifically, in both our initial (Hariri et al., 2002b) and replication studies (Hariri et al., 2005) we did not find any significant 5-HTTLPR genotype association with subjective behavioral measures of anxiety-like or fear-related traits as indexed by the Harm Avoidance (HA) component of the Tridimensional Personality Questionnaire, a putative personality measure related to trait anxiety and 5-HT function (Cloninger, 1986; Cloninger, Svrakic, and Przybeck, 1993). This failure to find a behavioral association is not surprising given the relatively small sample sizes of each study, given that participants were screened not to have extreme scores on psychiatric inventories, and thus given the limited power of the studies to detect likely small (e.g., 1–5%) genetically mediated differences in behavior—as well as the theses of this chapter: namely, that genes do not directly predict behavior and their effect on behavior is mediated/ moderated by their effects on distinct brain circuitry. A convergence of evidence from animal and human studies clearly demonstrates that emotional behaviors, especially those as complex as HA, are likely influenced by a densely interconnected and distributed cortical and subcortical circuitry of which the amygdala is only one component. Thus we were compelled to examine the relationship between HA and the observed 5-HTTLPR effects on the functional connectivity of the amygdala and pACC. We reasoned that if functional uncoupling of the amygdala-pACC affective circuit underlies reported associations of 5-HTTLPR with emotional phenotypes, functional connectivity indices between these regions should predict normal variation in temperamental trait measures related to anxiety and depression such as HA. These analyses revealed a striking pattern wherein nearly 30 percent of the variance in HA scores was predicted by our measure of amygdala-pACC functional

viding et al.: neuroimaging and molecular genetics

357

connectivity (Pezawas et al., 2005). This finding is particularly remarkable as about 50 percent of individual differences in HA are estimated to be due to genetic effects (Keller et al., 2005), thus suggesting that the 5-HTTLPR effects on the amygdala-pACC coupling account for 60 percent of the genetically driven variance on HA. Consistent with our previous studies (Hariri et al., 2005, 2002b), functional (or structural) measures of single brain regions (i.e., amygdala or pACC) were of no predictive value. Thus 5-HTTLPRmediated corticolimbic functional connectivity alterations are manifested in anxiety-related temperamental traits, possibly reflecting inadequate regulation and integration of amygdala-mediated arousal, leading to an increased vulnerability for persistent negative affect and eventually depression in the context of accumulating environmental adversity. While investigations of localized structural and functional abnormalities have provided insights about depression, our data underscore the importance of studying genetic mechanisms of complex brain disorders at the level of dynamically interacting neural systems. We suggest that such relationships capture more proximally the functional consequences of neurodevelopmental processes altering circuitry function implicated in human temperament and psychiatric disorders.

Importance of imaging genetics for understanding development and developmental psychopathology It is important to emphasize that the 5-HTTLPR S allele effect on amygdala structure, reactivity, and connectivity in our studies as well as those by Heinz, Bertolino, and Canli and their colleagues exists in samples of healthy adult volunteers with no history of affective or other psychiatric disorders. On one hand, this point is consistent with a recent fMRI study reporting that while amygdala hyperexcitability reflects a stable, heritable trait associated with inhibited behavior, it does not by itself predict the development of affective disorders (Schwartz et al., 2003). On the other hand, more and more evidence is accumulating that indicates that the majority of psychopathology is rooted early in life, first emerging during childhood and adolescence (e.g., Kim-Cohen et al., 2003). Thus it is possible that the relevance of 5-HTTLPR S allele effects on corticolimbic brain circuitry will be more manifest during the development of individuals predisposed to psychopathology. Moreover, it is likely that exposure to environmental stressors impacts this gene-brain pathway, in turn increasing one’s risk of developing psychopathology. The hallmark study of Caspi and colleagues (2003) and subsequent replication studies (Eley et al., 2004; Kaufman et al., 2004; Kendler et al., 2005) suggest that the existence of significant stressors in the environment of individuals carrying the 5-HTTLPR S allele is necessary to further tip the balance toward the development of

358

methodological paradigms

psychopathology. Similarly, abnormal social behavior (Champoux et al., 2002) and 5-HT metabolism (Bennett et al., 2002) have been reported in rhesus macaques with the 5-HTTLPR S allele homologue, but only in peer-reared, and thus environmentally stressed, individuals. Emerging data from studies of the 5-HTT knockout mouse implicate similar early developmental phenomena interacting with genetically driven variation in 5-HT in shaping the neurobiological landscape, contributing to emotional behaviors (Ansorge et al., 2004; Esaki et al., 2005; Holmes and Hariri, 2003). It is pertinent to note that in many of these examples the genetic vulnerability has manifested as a consequence of environmental stressors that have occurred early in development. This shift from normal to pathological behaviors and when during the life span this shift occurs may reflect the effects of cumulative environmental stress on brain regions, most notably the prefrontal cortex, critical in the regulation of amygdala activity (Hariri et al., 2003; Keightley et al., 2003; Rosenkranz, Moove, and Grace, 2003). For example, repeated exposure to environmental insults before the maturation of relatively late developing prefrontal regulatory circuits (Lewis, 1997) may result in further biased amygdala drive in S allele carriers. Such relative hyperamygdala and hypoprefrontal activity has been documented in affective disorders (Phillips et al., 2003b; Siegle et al., 2002) and thus may represent a critical pathway or predictive biological marker for the future development of psychopathology. Recent imaging studies (Heinz et al., 2005; Pezawas et al., 2005) demonstrating altered functional coupling of the amygdala and medial prefrontal cortex during affective processing in adult healthy S allele carriers underscore that complex dynamic interactions of the amygdala and prefrontal cortex may be critical for normal behavioral responses in individuals possessing this risk allele. These results suggest that individual differences in indices of complex, emergent behaviors, such as harm avoidance, reflect the effects of genetic variation on a distributed brain system involved in not only mediating physiologic and behavioral arousal (e.g., amygdala), but also regulating and integrating this arousal in the service of adaptive responses to environmental challenges (e.g., prefrontal cortex). As Rutter and colleagues suggest, some environmental adversity during early development (and likely throughout one’s lifetime) is necessary to enable an individual to cope with the stresses and challenges of everyday life (Rutter, Kim-Cohen, and Maughan, 2006; Rutter, Moffitt, and Caspi, 2006). However, some experiences are beyond the range of normative environmental stress, and continually encountering such experiences may, in the long run, compromise an individual’s ability to respond to environmental events in an adaptive way. The potent neurobiological effects of such adversity have been detailed in studies by McEwen

and colleagues illustrating the critical role of gonadal and adrenal hormones in mediating the impact of environmental stress on brain structure and function (McEwen, 2001) as well as those by Meaney and colleagues revealing the profound impact of maternal care on the developmental of neural circuitries involved in meeting environmental challenges (Meaney and Szyf, 2005). It is likely that genetic makeup influences the “fingerprint” of how an individual reacts to stress as well as determines the person’s resiliency (i.e., adaptive functioning in the face of adversity) to chronic stressors. The biological “fingerprint” of resiliency is likely to vary by disorder and as a function of environmental circumstances (Luthar, Cicchetti, and Becker, 2000). Developmental imaging genetics represents one unique approach by which resiliency can be investigated within a translational research paradigm. As mentioned previously, adult psychiatric disorders rarely arise de novo, without any warning of childhood problems, including full-blown episodes of psychopathology (Kim-Cohen et al., 2003). Some psychiatric disorders may lie dormant until a crucial point in development or until a critical environmental stressor is present to precipitate their onset. An individual may have a vulnerability to a disorder from birth, but this vulnerability may only manifest later in life. Depression and schizophrenia are two disorders that traditionally have not been diagnosed before adolescence or early adulthood, although sufficient evidence has accumulated to suggest that aspects of each originate in early development (Cicchetti and Cannon, 1999; Kim-Cohen et al., 2003). Other adult disorders represent the eventual manifestation of a behavioral difficulty visible since early childhood. Antisocial personality disorder is an example where one of the diagnostic criteria includes evidence of childhood conduct problems. In all cases, it is vital to understand the developmental course of psychopathologies that inform our efforts to understand the etiology, progression, and treatment of psychiatric disease. As part of such a broader developmental research program, imaging genetics has great potential yield in identifying component etiologic processes of psychopathology as well as the emergence of normal individual differences in behavior. Accordingly, we speculate that adult imaging genetics findings represent windows into systems whose current structure and function resulted from developmental alterations during unique periods of plasticity, long before the physiological associations were captured by means of neuroimaging in adulthood (i.e., they represent “ghosts in the machine”). One avenue of potential fruitful research into these developing systems is the use of longitudinal studies beginning in childhood. This approach represents an ideal way to examine the impact of genetic and environmental effects on the developing neural circuitry supporting behavior and conferring risk for psychopathology. Such an approach will allow for

the determination of genetically driven variation on structural and functional brain development during windows of time that reflect critical maturational processes (e.g., myelination, synaptic pruning). Longitudinal studies in at-risk populations prior to the development of psychopathology will further allow for a more accurate and rigorous assessment and recording of environmental events and experiences as well as their interplay with genetically mediated risks. Moreover, a longitudinal approach will facilitate charting the behavioral consequences, including risk for psychopathology, of such genetically and environmentally driven variation in brain structure and function. Such studies will not only serve to collect important baseline structural and functional data during critical stages of development, but also enable us to disentangle how genetic and environmental factors converge during development to bias brain systems and, in turn, mediate risk for psychopathology. We know of one existing longitudinal study of brain structure that also includes twin data (Wallace et al., 2006). This study is already yielding interesting information about substantial genetic influences on individual differences in brain maturation. As the developmental trajectories vary considerably for different brain structures and across individuals (Lenroot and Giedd, 2006), potential clues about developmental vulnerability periods for particular illnesses are starting to accumulate. With regard to development in general, although age is broadly considered as an index of development, it may also be one of the most imprecise because it simultaneously serves as an index of physical maturation, hormonal state, cognitive level, social circumstances, and life experiences (Rutter, Kim-Cohen, and Maughan, 2006). It is therefore crucial to attempt to disentangle which of these features associated with age is likely to have an influence with respect to the development of psychopathology. For some behaviors, this endeavor will be made easier by a rich research base that can identify specific age-related variables crucial for understanding the target developmental change and its relevance in moderating risk or resiliency for psychopathology. For other phenomena, and often for brain function in particular, one will not have the luxury of a rich empirical database and will have to start with carefully formulated biologically constrained hypotheses about genetic influences on normal and abnormal brain function. There are, of course, some limitations to this experimental strategy. For example, imaging genetics has only been used to observe the effects of a few polymorphic genes (or large chromosomal deletions; e.g., Meyer-Lindenberg et al., 2005) on brain structure and function in humans. However, a number of nonvarying genes are likely to be very important in initiating developmental cascades that have profound impact on brain structure/function and ultimately behavior. Moreover, our understanding of the timing and biological

viding et al.: neuroimaging and molecular genetics

359

impact of gene expression is conceptually simplistic, and the interacting influences of other genes as well as environmental factors on the temporal and spatial patterns of gene expression are poorly described. At present, such effects are only tractable in animal models (Meaney and Szyf, 2005). These types of genetic effects, which we cannot test for in developmental imaging genetics studies at the present time, are likely to provide “background noise” in our assessments of specific candidate genes and gene-environment interactions. Despite the fact that longitudinal developmental imaging data will, by nature, require concerted and protracted efforts and are not useful in studying the effects of nonpolymorphic genes on human brain development, there remains tremendous value in developmental imaging genetics “here and now.” For example, it is of interest to investigate whether the 5-HTTLPR effects on the functional connectivity of the amygdala and pACC can be seen prior to the maturation of prefrontal cortex in adolescence, a period of increased risk for mood disorders, and can influence temperament in similar ways as in adults. Cross-sectional developmental data would be adequate in initially addressing this question.

Developmental imaging genetics: Ongoing investigations We will briefly describe two of our ongoing developmental studies in an effort to illustrate potential applications of translational research that incorporates a developmental imaging genetics approach to examine the biological pathways to normal and pathological behavior. These two transcontinental studies are aimed at (1) identifying pathways to antisocial behavior (AB) and (2) elucidating an affective pathway to the development of alcohol use disorders (AUD) emerging in adolescence. Development of Antisocial Behavior Children with early-onset antisocial behavior are at risk to develop chronic persistent antisocial behavior during the course of their lifetime (Eley, Lichtenstein, and Moffitt, 2003). Elucidating different developmental pathways to persistent antisocial behavior is of major importance because individuals with early-onset antisocial behavior are 10 times more costly to society than controls (Scott et al., 2001). In this context, the aim of a developmental imaging genetics approach would be to increase our understanding of both the genetic and neurobiological mechanisms of vulnerability in antisocial children and ultimately to identify pathways for the developmental of extreme behavioral problems. One risk factor that appears to have predictive importance for life-course persistent antisocial behavior is the presence of callous-unemotional (CU) traits (e.g., lack of guilt and empathy). Adult psychopaths (individuals with severe overt antisocial behavior coupled with CU traits) are chronic,

360

methodological paradigms

versatile, and serious offenders who have early-emerging behavioral problems (Hare, Hart, and Harpur, 1991). Children with early-onset antisocial behavior coupled with CU traits are more likely to persist with antisocial conduct and commit more serious acts of antisocial behavior than their antisocial peers with nonelevated levels of CU traits (Frick et al., 2003). For both adults and children with antisocial behavior and CU traits there is evidence of amygdala hyporeactivity (Blair, 2005). We recently demonstrated that CU traits delineate an etiologically distinct subgroup of children with early-onset antisocial behavior (Viding et al., 2005) as indexed by high heritability estimates of 81 percent among the AB children with elevated levels of CU traits compared to only 30 percent among the AB children without elevated levels of CU traits (Viding et al., 2005). Given the level of the genetic contribution to AB with CU traits and the demonstrated impact of genetic polymorphisms on neural systems supporting affective behaviors, we reasoned that there was a high likelihood that genetically driven biological systems contribute to the emergence of these traits. In short, from our work we know that the AB + CU subtype is likely to be under genetic mediation, as evidenced by high heritability, and likely involves alterations in the corticolimbic circuitry mediating arousal and vigilance. As such, we reasoned that the selection of candidate genes on the basis of previous investigations (e.g., 5-HTTLPR and MAOA) would potentially help to elucidate genetic moderation of alterations in this corticolimbic circuitry. For example, is the amygdala hyporeactivity seen in the AB + CU subtype in part mediated by the 5-HTTLPR LL genotype? Accordingly, we have undertaken a developmental imaging genetics project to document (1) whether preadolescent children with AB + CU have distinct corticolimbic dysfunction, (2) whether this dysfunction is heritable, and (3) what specific genes (e.g., 5-HTTLPR and MAOA) can be found to be associated with this dysfunction. Our investigation is being conducted in two steps. First, children with AB + CU (n = 12) are being compared to those who have AB without CU (n = 12) and ability-matched controls (n = 12). This comparison will enable us to determine whether there is unique brain dysfunction associated with AB + CU, rather than with AB in general. (We are predicting that amygdala hyporeactivity to others’ distress will be a unique marker for AB + CU). We will then recruit a larger sample of identical and nonidentical twins (132 twins in total), in concordant and discordant AB + CU pairs, as well as control (unaffected) pairs. This procedure will enable us to estimate the net contribution of genes generally, as well as candidate polymorphisms specifically, on brain activation differences between AB + CU and control children (heritability). In this initial investigation, children aged 10–11 years are being assessed cross-sectionally. However, we will follow these

children longitudinally to see whether the pattern of amygdala reactivity will predict later antisocial behavior and whether this is moderated by CU traits as well as individual experiences with environmental stressors/precipitators. An Affective Pathway for the Development of Alcohol Use Disorders Alcohol use disorders (AUD) constitute a multifactorial disorder in which the environment interacts with genetic predisposition to produce the final level of risk (Schuckit, 1998). It has been shown that stress and affective responses to stress, including depression, influence alcohol drinking and relapse (Kreek and Koob, 1998). Although the relation between stress and alcohol drinking in humans (Pohorecky, 1991) and laboratory animals (Pohorecky, 1990) is complex, it is known that, in some individuals, increases in alcohol consumption are associated with stress. While such stress-induced alcohol drinking and relapse behavior apparently have a significant genetic component (LaForge, Yuferov, and Kreek, 2000; Wand et al., 2002), the neurobiological pathways underlying stress-induced alcohol drinking and relapse behavior are still obscure. Adolescence is a crucial period for the development of affect regulation, affective disorders, and the onset of alcohol use. Although the relation between major depressive disorder (MDD) and AUD has been widely observed, there is currently little understanding of the shared etiology and diathesis of these two disorders—especially their initial emergence during adolescence. As part of our ongoing developmental work examining depression, we have recently investigated the emergence of AUD during adolescence among subjects initially studied in childhood (10–14 years) and followed through early adulthood (16–20 years). Analyses revealed that children with MDD as well as high familial loading for MDD had increased risk for developing AUD during adolescence in comparison with children with low familial loading for MDD. The risk for developing AUD did not differ between MDD and high familial load children. In addition to group status effects, age at intake, gender, and follow-up months were all significant predictors of time to AUD. Our preliminary data strongly suggest that depression early in life and having high familial loading for depression significantly increase one’s risk of developing AUD during adolescence. Coupled with the existing literature, these preliminary studies converge to highlight the need to understand the interplay between genes and environment on the development of neurobiolo-gical systems in adolescence as they impact one’s risk to develop AUD. As a result of our preliminary work, we have begun a project to disentangle several critical risk factors involved in the development of AUD, since these risk factors appear to be moderated by preexisting risk for depression. Neuroimaging, neurobehavioral, and genetic measures are being used

to achieve this goal, and the study design has three phases. First, adolescents between the ages of 12 years and 14 years 11 months who have high (n = 150) and low (n = 150) familial loading for depression are being recruited. Initial assessments include measures of genes (primarily 5-HT subsystems like 5-HTTLPR), alcohol use, emotional behavior (depression/negative affect, temperament), and environmental stress (acute/chronic negative parent-child relationships), and imaging genetics assessments of brain function (amygdala reactivity and prefrontal regulatory control). A second phase of the proposed research project will consist of yearly follow-ups of all subjects to reassess behavior, environmental stress, and alcohol use outcomes. The final phase of our project will investigate the relationship of the development of AUD on neural circuitry central to the generation and regulation of arousal and affect. Translational Aspects of Developmental Imaging Genetics Although imaging genetics in and of itself provides a powerful new approach to the study of genes, brain, and behavior, its true potential will only be realized by aggressively expanding the scope and scale of the experimental protocols within a developmental framework, especially one that is focused on examining the developmental origins of behavior and disease. Although gene effects on brain function can be readily documented in samples of adults, the contributions of these genes acting in response to variable environmental pressures across development (when these systems are arguably most malleable) must be assessed in order to understand the biological pathways that bias behavior and risk for psychiatric illness. Combining these different strands of scientific inquiry (genetic, brain imaging, developmental, social) in a way that translates into public health benefits for affected individuals is in line with the current NIMH research direction (http:// www.nimh.nih.gov/strategic/strategicplanmenu.cfm). Translational research will be able to capitalize on developmental-imaging-genetics findings, and in the future we will be able to document the markers that are truly predictive for developmental outcome and disease progression as well as allow for the early identification of individuals at greater risk for emotional regulatory problems that can have long-term health-related implications. The short-term rewards of a developmental-imaging-genetics program are identifying the contributions of selected genetic, environmental, behavioral, and neurobiological factors for the emergence of individual differences in behavior and risk for disease. Whereas such short-term rewards can be realized with cross-sectional data, the long-term rewards of this approach, and thus the truly translational aspects of this work, will depend on longitudinal data. With such data we hope to gain greater understanding of how the genetic and environmental effects on circumscribed neuroanatomical

viding et al.: neuroimaging and molecular genetics

361

circuits operate across development. Consequently, we can better document crucial time points for providing interventions targeted to specific neural processes. These interventions can also be tailored to take into account an individual’s genetic and environmental risk profiles. Thus for some cases pharmacological intervention may be judged most appropriate, while for others cognitive and behavioral approaches would be better suited to counter aberrant neural development.

Summary This chapter has outlined an experimental strategy by which genetic effects on brain function can be explored using neuroimaging, namely, imaging genetics. We discussed the effectiveness of this strategy for delineating biological pathways and mechanisms by which individual differences in brain function emerge and potentially bias behavior and risk for psychiatric illness. The main focus of this chapter was to highlight the importance of applying an imaging genetics framework to the study of psychopathology within a developmental framework, which we called developmental imaging genetics. We argued that by beginning to move toward a systems-level approach to understanding pathways to behavioral outcomes, as well as integrating the developmental angle, we will move closer to understanding the complexities of the specific mechanisms involved in the etiology of psychiatric disease. Numerous challenges lie ahead for developmental imaging genetics, but we believe that this approach has the potential to yield highly informative results that will translate to public health benefits for people with psychiatric disorders. acknowledgments

Portions of this article were published in Development and Psychopathology (Viding et al., “Developmental Imaging Genetics: Challenges and Promises for Translational Research” [2006], 18:877–892) and the Handbook of Emotion Regulation (Hariri and Forbes, “Genetics of Emotion Regulation” [2006], ed. J. J. Gross). This research was supported in part by funding from the United Kingdom National Program on Forensic Mental Health Research and Development (Dr. Viding); the Medical Research Council G0401170 (Dr. Viding); the National Institute of Mental Health grants K01-MH001957 (Dr. Williamson), K01-MH072837 (Dr. Hariri), and K01MH074769 (Dr. Forbes); and a NARSAD Young Investigator Award (Drs. Forbes and Hariri).

REFERENCES Ansorge, M. S., M. Zhou, A. Lira, R. Hen, and J. A. Gingrich, 2004. Early-life blockade of the 5-HT transporter alters emotional behavior in adult mice. Science 306(5697):879–881.

362

methodological paradigms

Azmitia, E. C., and P. J. Gannon, 1986. The primate serotonergic system: A review of human and animal studies and a report on Macaca fascicularis. Adv. Neurol. 43:407–468. Beauregard, M., J. Levesque, and P. Bourgouin, 2001. Neural correlates of conscious self-regulation of emotion. J. Neurosci. 21(18):RC165. Bellivier, F., P. Chaste, and A. Malafosse, 2004. Association between the tph gene A218C polymorphism and suicidal behavior: A meta-analysis. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 124(1):87–91. Bennett, A. J., K. P. Lesch, A. Heils, J. C. Long, J. G. Lorenz, S. E. Shoaf, et al., 2002. Early experience and serotonin transporter gene variation interact to influence primate cns function. Mol. Psychiatry 7(1):118–122. Bertolino, A., G. Arciero, V. Rubino, V. Latorre, M. De Candia, V. Mazzola, et al., 2005. Variation of human amygdala response during threatening stimuli as a function of 5-HTTLPR genotype and personality style. Biol. Psychiatry 57(12):1517–1525. Blair, R. J., 2005. Applying a cognitive neuroscience perspective to the disorder of psychopathy. Bev. Psychopathol. 17:865–891. Blier, P., and C. de Montigny, 1999. Serotonin and drug-induced therapeutic responses in major depression, obsessive-compulsive and panic disorders. Neuropsychopharmacology 21(2 Suppl.): 91S–98S. Brown, S. M., E. Peet, S. B. Manuck, D. E. Williamson, R. E. Dahl, R. E. Ferrell, et al., 2005. A regulatory variant of the human tryptophan hydroxylase-2 gene biases amygdala reactivity. Mol. Psychiatry 10(9):884–888, 805. Canli, T., K. Omura, B. W. Haas, A. Fallgatter, R. T. Constable, and K. P. Lesch, 2005. Beyond affect: A role for genetic variation of the serotonin transporter in neural activation during a cognitive attention task. Proc. Natl. Acad. Sci. USA 102(34):12224–12229. Carey, G., and D. L. DiLalla, 1994. Personality and psychopathology: Genetic perspectives. J. Abnorm. Psychol. 103(1): 32–43. Caspi, A., K. Sugden, T. E. Moffitt, A. Taylor, I. W. Craig, H. Harrington, et al., 2003. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301(5631):386–389. Champoux, M., A. Bennett, C. Shannon, J. D. Higley, K. P. Lesch, and S. J. Suomi, 2002. Serotonin transporter gene polymorphism, differential early rearing, and behavior in rhesus monkey neonates. Mol. Psychiatry 7(10):1058– 1063. Cicchetti, D., and T. D. Cannon, 1999. Neurodevelopmental processes in the ontogenesis and epigenesis of psychopathology. Dev. Psychopathol. 11(3):375–393. Cloninger, C. R., 1986. A unified biosocial theory of personality and its role in the development of anxiety states. Psychiatr. Dev. 4(3):167–226. Cloninger, C. R., D. M. Svrakic, and T. R. Przybeck, 1993. A psychobiological model of temperament and character. Arch. Gen. Psychiatry 50(12):975–990. Depue, R. A., and P. F. Collins, 1999. Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behav. Brain Sci. 22(3):491–517. Domschke, K., M. Braun, P. Ohrmann, T. Suslow, H. Kugel, J. Bauer, et al., 2005. Association of the functional −1019C/G 5-HT1A polymorphism with prefrontal cortex and amygdala activation measured with 3 T FMRI in panic disorder. Int. J. Neuropsychopharmacol. 1–7.

Du, L., D. Bakish, and P. D. Hrdina, 2000. Gender differences in association between serotonin transporter gene polymorphism and personality traits. Psychiatr. Genet. 10(4):159–164. Ebstein, R. P., 2006. The molecular genetic architecture of human personality: Beyond self-report questionnaires. Mol. Psychiatry 11(5):427–445. Ebstein, R. P., A. H. Zohar, J. Benjamin, and R. H. Belmaker, 2002. An update on molecular genetic studies of Human personality traits. Appl. Bioinformatics 1(2):57–68. Eggers, B., W. Hermann, H. Barthel, O. Sabri, A. Wagner, and S. Hesse, 2003. The degree of depression in Hamilton rating scale is correlated with the density of presynaptic serotonin transporters in 23 patients with Wilson’s disease. J. Neurol. 250(5):576–580. Eley, T. C., P. Lichtenstein, and T. E. Moffitt, 2003. A longitudinal behavioral genetic analysis of the etiology of aggressive and nonaggressive antisocial behavior. Dev. Psychopathol. 15(2):383–402. Eley, T. C., K. Sugden, A. Corsico, A. M. Gregory, P. Sham, P. McGuffin, et al., 2004. Gene-environment interaction analysis of serotonin system markers with adolescent depression. Mol. Psychiatry 9(10):908–915. Emahazion, T., L. Feuk, M. Jobs, S. L. Sawyer, D. Fredman, D. St Clair, et al., 2001. SNP association studies in Alzheimer’s disease highlight problems for complex disease analysis. Trends Genet. 17(7):407–413. Esaki, T., M. Cook, K. Shimoji, D. L. Murphy, L. Sokoloff, and A. Holmes, 2005. Developmental disruption of serotonin transporter function impairs cerebral responses to whisker stimulation in mice. Proc. Natl. Acad. Sci. USA 102(15):5582–5587. Frick, P. J., A. H. Cornell, C. T. Barry, S. D. Bodin, and H. E. Dane, 2003. Callous-unemotional traits and conduct problems in the prediction of conduct problem severity, aggression, and selfreport of delinquency. J. Abnorm. Child Psychol. 31(4):457–470. Furmark, T., M. Tillfors, H. Garpenstrand, I. Marteinsdottir, B. Langstrom, L. Oreland, et al., 2004. Serotonin transporter polymorphism linked to amygdala excitability and symptom severity in patients with social phobia. Neurosci. Lett. 362(1): 1–4. Garpenstrand, H., P. Annas, J. Ekblom, L. Oreland, and M. Fredrikson, 2001. Human fear conditioning is related to dopaminergic and serotonergic biological markers. Behav. Neurosci. 115(2):358–364. Gelernter, J., H. Kranzler, and J. F. Cubells, 1997. Serotonin transporter protein (SLC6A4) allele and haplotype frequencies and linkage disequilibria in African- and European-American and Japanese populations and in alcohol-dependent subjects. Hum. Genet. 101(2):243–246. Glatt, C. E., and N. B. Freimer, 2002. Association analysis of candidate genes for neuropsychiatric disease: The perpetual campaign. Trends Genet. 18(6):307–312. Hare, R. D., S. D. Hart, and T. J. Harpur, 1991. Psychopathy and the DSM-IV criteria for antisocial personality disorder. J. Abnorm. Psychol. 100(3):391–398. Hariri, A. R., S. Y. Bookheimer, and J. C. Mazziotta, 2000. Modulating emotional responses: Effects of a neocortical network on the limbic system. NeuroReport 11(1):43–48. Hariri, A. R., E. M. Drabant, K. E. Munoz, B. S. Kolachana, V. S. Mattay, M. F. Egan, et al., 2005. A susceptibility gene for affective disorders and the response of the human amygdala. Arch. Gen. Psychiatry 62(2):146–152. Hariri, A. R., E. M. Drabant, and D. R. Weinberger, 2006. Imaging genetics: Perspectives from studies of genetically driven

variation in serotonin function and corticolimbic affective processing. Biol. Psychiatry 59(10):888–897. Hariri, A. R., V. S. Mattay, A. Tessitore, F. Fera, W. G. Smith, and D. R. Weinberger, 2002a. Dextroamphetamine modulates the response of the human amygdala. Neuropsychopharmacology 27(6):1036–1040. Hariri, A. R., V. S. Mattay, A. Tessitore, F. Fera, and D. R. Weinberger, 2003. Neocortical modulation of the amygdala response to fearful stimuli. Biol. Psychiatry 53(6):494–501. Hariri, A. R., V. S. Mattay, A. Tessitore, B. Kolachana, F. Fera, D. Goldman, et al., 2002b. Serotonin transporter genetic variation and the response of the human amygdala. Science 297(5580):400–403. Hariri, A. R., A. Tessitore, V. S. Mattay, F. Fera, and D. R. Weinberger, 2002. The amygdala response to emotional stimuli: A comparison of faces and scenes. NeuroImage 17(1):317–323. Hariri, A. R., and D. R. Weinberger, 2003a. Functional neuroimaging of genetic variation in serotonergic neurotransmission. Genes Brain Behav. 2(6):314–349. Hariri, A. R., and D. R. Weinberger, 2003b. Imaging genomics. Br. Med. Bull. 65:259–270. Heils, A., A. Teufel, S. Petri, G. Stober, P. Riederer, D. Bengel, et al., 1996. Allelic variation of human serotonin transporter gene expression. J. Neurochem. 66(6):2621–2624. Heinz, A., D. F. Braus, M. N. Smolka, J. Wrase, I. Puls, D. Hermann, et al., 2005. Amygdala-prefrontal coupling depends on a genetic variation of the serotonin transporter. Nature Neurosci. 8(1):20–21. Heinz, A., J. D. Higley, J. G. Gorey, R. C. Saunders, D. W. Jones, D. Hommer, et al., 1998. In vivo association between alcohol intoxication, aggression, and serotonin transporter availability in nonhuman primates. Am. J. Psychiatry 155(8):1023– 1028. Heinz, A., D. W. Jones, G. Bissette, D. Hommer, P. Ragan, M. Knable, et al., 2002. Relationship between cortisol and serotonin metabolites and transporters in alcoholism. Pharmacopsychiatry 35(4):127–134. Heinz, A., D. W. Jones, C. Mazzanti, D. Goldman, P. Ragan, D. Hommer, et al., 2000. A relationship between serotonin transporter genotype and in vivo protein expression and alcohol neurotoxicity. Biol. Psychiatry 47(7):643–649. Holmes, A., and A. R. Hariri, 2003. The serotonin transporter gene–linked polymorphism and negative emotionality: Placing single gene effects in the context of genetic background and environment. Genes Brain Behav. 2(6):332–335. Holmes, A., Q. Lit, D. L. Murphy, E. Gold, and J. N. Crawley, 2003. Abnormal anxiety-related behavior in serotonin transporter null mutant mice: The influence of genetic background. Genes Brain Behav. 2(6):365–380. Kalivas, P. W., and N. D. Volkow, 2005. The neural basis of addiction: A pathology of motivation and choice. Am. J. Psychiatry 162(8):1403–1413. Katsuragi, S., H. Kunugi, A. Sano, T. Tsutsumi, K. Isogawa, S. Nanko, et al., 1999. Association between serotonin transporter gene polymorphism and anxiety-related traits. Biol. Psychiatry 45(3):368–370. Kaufman, J., B. Z. Yang, H. Douglas-Palumberi, S. Houshyar, D. Lipschitz, J. H. Krystal, et al., 2004. Social supports and serotonin transporter gene moderate depression in maltreated children. Proc. Natl. Acad. Sci. USA 101(49):17316–17321. Keightley, M. L., G. Winocur, S. J. Graham, H. S. Mayberg, S. J. Hevenor, and C. L. Grady, 2003. An fMRI study investigating cognitive modulation of brain regions associated with

viding et al.: neuroimaging and molecular genetics

363

emotional processing of visual stimuli. Neuropsychologia 41(5): 585–596. Keller, M. C., W. L. Coventry, A. C. Heath, and N. G. Martin, 2005. Widespread evidence for non-addictive gene variation in Cloninger’s and Eysenck’s personality dimensions using a twin plus sibling design. Behav. Genet. 35(6):707–721. Kendler, K. S., J. W. Kuhn, J. Vittum, C. A. Prescott, and B. Riley, 2005. The interaction of stressful life events and a serotonin transporter polymorphism in the prediction of episodes of major depression: A replication. Arch. Gen. Psychiatry 62(5): 529–535. Kendler, K. S., M. C. Neale, R. C. Kessler, A. C. Heath, and L. J. Eaves, 1993. A longitudinal twin study of personality and major depression in women. Arch. Gen. Psychiatry 50(11):853– 862. Kim-Cohen, J., A. Caspi, T. E. Moffitt, H. Harrington, B. J. Milne, and R. Poulton, 2003. Prior juvenile diagnoses in adults with mental disorder: Developmental follow-back of a prospective-longitudinal cohort. Arch. Gen. Psychiatry 60(7):709–717. Kreek, M. J., and G. F. Koob, 1998. Drug dependence: Stress and dysregulation of brain reward pathways. Drug Alcohol Depend. 51(1–2):23–47. LaForge, K. S., V. Yuferov, and M. J. Kreek, 2000. Opioid receptor and peptide gene polymorphisms: Potential implications for addictions. Eur. J. Pharmacol. 410(2–3):249–268. Lander, E. S., L. M. Linton, B. Birren, C. Nusbaum, M. C. Zody, J. Baldwin, et al., 2001. Initial sequencing and analysis of the human genome. Nature 409(6822):860–921. Lange, K., L. M. Williams, A. W. Young, E. T. Bullmore, M. J. Brammer, S. C. Williams, et al., 2003. Task instructions modulate neural responses to fearful facial expressions. Biol. Psychiatry 53(3):226–232. LeDoux, J. E., 2000. Emotion circuits in the brain. Annu. Rev. Neurosci. 23:155–184. Lenroot, R. K., and J. N. Giedd, 2006. Brain development in children and adolescents: Insights from anatomical magnetic resonance imaging. Neurosci. Biobehav. Rev. 30:718–729. Leonardo, E. D., and R. Hen, 2006. Genetics of affective and anxiety disorders. Annu. Rev. Psychol. 57:117–137. Lesch, K. P., D. Bengel, A. Heils, S. Z. Sabol, B. D. Greenberg, S. Petri, et al., 1996. Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 274(5292):1527–1531. Lesch, K. P., and R. Mossner, 1998. Genetically driven variation in serotonin uptake: Is there a link to affective spectrum, neurodevelopmental, and neurodegenerative disorders? Biol. Psychiatry 44(3):179–192. Lewis, D. A., 1997. Development of the prefrontal cortex during adolescence: Insights into vulnerable neural circuits in schizophrenia. Neuropsychopharmacology 16(6):385–398. Little, K. Y., D. P. McLaughlin, L. Zhang, C. S. Livermore, G. W. Dalack, P. R. McFinton, et al., 1998. Cocaine, ethanol, and genotype effects on human midbrain serotonin transporter binding sites and mRNA levels. Am. J. Psychiatry 155(2): 207–213. Lucki, I., 1998. The spectrum of behaviors influenced by serotonin. Biol. Psychiatry 44(3):151–162. Luthar, S. S., D. Cicchetti, and B. Becker, 2000. The construct of resilience: A critical evaluation and guidelines for future work. Child Dev. 71(3):543–562. Malhotra, A. K., and D. Goldman, 1999. Benefits and pitfalls encountered in psychiatric genetic association studies. Biol. Psychiatry 45(5):544–550.

364

methodological paradigms

Malison, R. T., L. H. Price, R. Berman, C. H. van Dyck, G. H. Pelton, L. Carpenter, et al., 1998. Reduced brain serotonin transporter availability in major depression as measured by [123i]–2 beta–carbomethoxy–3 beta–(4-iodophenyl)tropane and single photon emission computed tomography. Biol. Psychiatry 44(11):1090–1098. Manuck, S. B., J. D. Flory, R. E. Ferrell, K. M. Dent, J. J. Mann, and M. F. Muldoon, 1999. Aggression and anger-related traits associated with a polymorphism of the tryptophan hydroxylase gene. Biol. Psychiatry 45(5):603– 614. Manuck, S. B., J. D. Flory, R. E. Ferrell, J. J. Mann, and M. F. Muldoon, 2000. A regulatory polymorphism of the monoamine oxidase-a gene may be associated with variability in aggression, impulsivity, and central nervous system serotonergic responsivity. Psychiatry Res. 95(1):9–23. Mayberg, H. S., 2003a. Modulating dysfunctional limbic-cortical circuits in depression: Towards development of brain-based algorithms for diagnosis and optimised treatment. Br. Med. Bull. 65:193–207. Mayberg, H. S., 2003b. Positron emission tomography imaging in depression: A neural systems perspective. Neuroimaging Clin. N. Am. 13(4):805–815. Mazzanti, C. M., J. Lappalainen, J. C. Long, D. Bengel, H. Naukkarinen, M. Eggert, et al., 1998. Role of the serotonin transporter promoter polymorphism in anxiety-related traits. Arch. Gen. Psychiatry 55(10):936–940. McEwan, B. S., 2001. From molecules to mind: Stress, individual differences, and the social environment. Ann. NY Acad. Sci. 935:42–49. McGuffin, P., B. Riley, and R. Plomin, 2001. Genomics and behavior: Toward behavioral genomics. Science 291(5507): 1232–1249. Meaney, M. J., and M. Szyf, 2005. Environmental programming of stress responses through DNA methylation: Life at the interface between a dynamic environment and a fixed genome. Dialogues Clin. Neurosci. 7(2):103–123. Melke, J., M. Landen, F. Baghei, R. Rosmond, G. Holm, P. Bjorntorp, et al., 2001. Serotonin transporter gene polymorphisms are associated with anxiety-related personality traits in women. Am. J. Med. Genet. 105(5):458–463. Meyer-Lindenberg, A., A. R. Hariri, K. E. Munoz, C. B. Mervis, V. S. Mattay, C. A. Morris, et al., 2005. Neural correlates of genetically abnormal social cognition in Williams syndrome. Nature Neurosci. 8(8):991–993. Moldin, S. O., and I. I. Gottesman, 1997. At issue: Genes, experience, and chance in schizophrenia—Positioning for the 21st century. Schizophr. Bull. 23(4):547–561. Moreno, F. A., D. C. Rowe, B. Kaiser, D. Chase, T. Michaels, J. Gelernter, et al., 2002. Association between a serotonin transporter promoter region polymorphism and mood response during tryptophan depletion. Mol. Psychiatry 7(2):213–216. Munafo, M. R., T. Clark, and J. Flint, 2005. Does measurement instrument moderate the association between the serotonin transporter gene and anxiety-related personality traits? A metaanalysis. Mol. Psychiatry 10(4):415–419. Murphy, D. L., A. M. Andrews, C. H. Wichems, Q. Li, M. Tohda, and B. Greenberg, 1998. Brain serotonin neurotransmission: An overview and update with an emphasis on serotonin subsystem heterogeneity, multiple receptors, interactions with other neurotransmitter systems, and consequent implications for understanding the actions of serotonergic drugs. J. Clin. Psychiatry 59(Suppl. 15):4–12.

Nakamura, K., R. Kawashima, S. Nagumo, K. Ito, M. Sugiura, T. Kato, et al., 1998. Neuroanatomical correlates of the assessment of facial attractiveness. NeuroReport 9(4):753–757. Narumoto, J., H. Yamada, T. Iidaka, N. Sadato, K. Fukui, H. Itoh, et al., 2000. Brain regions involved in verbal or nonverbal aspects of facial emotion recognition. NeuroReport 11(11): 2571–2576. Neumeister, A., A. Konstantinidis, J. Stastny, M. J. Schwarz, O. Vitouch, M. Willeit, et al., 2002. Association between serotonin transporter gene promoter polymorphism (5-HTTLPR) and behavioral responses to tryptophan depletion in healthy women with and without family history of depression. Arch. Gen. Psychiatry 59(7):613–620. Nigg, J. T., and H. H. Goldsmith, 1998. Developmental psychopathology, personality, and temperament: Reflections on recent behavioral genetics research. Hum. Biol. 70(2):387–412. Noblett, K. L., and E. F. Coccaro, 2005. Molecular genetics of personality. Curr. Psychiatry Rep. 7(1):73–80. Patkar, A. A., W. H. Berrettini, P. Mannelli, R. Gopalakrishnan, M. R. Hoehe, L. Bilal, et al., 2004. Relationship between serotonin transporter gene polymorphisms and platelet serotonin transporter sites among African-American cocainedependent individuals and healthy volunteers. Psychiatr. Genet. 14(1):25–32. Pezawas, L., A. Meyer-Lindenberg, E. M. Drabant, B. A. Verchinski, K. E. Munoz, B. S. Kolachana, et al., 2005. 5HTTLPR polymorphism impacts human cingulate-amygdala interactions: A genetic susceptibility mechanism for depression. Nature Neurosci. 8(6):828–834. Phillips, M. L., W. C. Drevets, S. L. Rauch, and R. Lane, 2003a. Neurobiology of emotion perception. I. The neural basis of normal emotion perception. Biol. Psychiatry 54(5):504–514. Phillips, M. L., W. C. Drevets, S. L. Rauch, and R. Lane, 2003b. Neurobiology of emotion perception. II. Implications for major psychiatric disorders. Biol. Psychiatry 54(5):515–528. Plomin, R., and J. R. Nesselroade, 1990. Behavioral genetics and personality change. J. Pers. 58(1):191–220. Plomin, R., M. J. Owen, and P. McGuffin, 1994. The genetic basis of complex human behaviors. Science 264(5166):1733– 1739. Pohorecky, L. A., 1990. Interaction of ethanol and stress: Research with experimental animals—An update. Alcohol Alcohol. 25(2–3):263–276. Pohorecky, L. A., 1991. Stress and alcohol interaction: An update of human research. Alcohol. Clin. Exp. Res. 15(3):438–459. Reif, A., and K. P. Lesch, 2003. Toward a molecular architecture of personality. Behav. Brain Res. 139(1–2):1–20. Rosenkranz, J. A., H. Moore, and A. A. Grace, 2003. The prefrontal cortex regulates lateral amygdala neuronal plasticity and responses to previously conditioned stimuli. J. Neurosci. 23(35):11054–11064. Rutter, M., J. Kim-Cohen, and B. Maughan, 2006. Continuities and discontinuities in psychopathology between childhood and adult life. J. Child Psychol. Psychiatry 47(3–4): 276–295. Rutter, M., T. E. Moffitt, and A. Caspi, 2006. Gene-environment interplay and psychopathology: Multiple varieties but real effects. J. Child Psychol. Psychiatry 47(3–4):226–261. Rutter, M., and R. Plomin, 1997. Opportunities for psychiatry from genetic findings. Br. J. Psychiatry 171:209–219. Sadikot, A. F., and A. Parent, 1990. The monoaminergic innervation of the amygdala in the squirrel monkey: An immunohistochemical study. Neuroscience 36(2):431–447.

Schinka, J. A., R. M. Busch, and N. Robichaux-Keene, 2004. A meta-analysis of the association between the serotonin transporter gene polymorphism (5-HTTLPR) and trait anxiety. Mol. Psychiatry 9(2):197–202. Schinka, J. A., E. A. Letsch, and F. C. Crawford, 2002. DRD4 and novelty seeking: Results of meta-analyses. Am. J. Med. Genet. 114(6):643–648. Schuckit, M. A., 1998. Biological, psychological and environmental predictors of the alcoholism risk: A longitudinal study. J. Stud. Alcohol 59(5):485–494. Schwartz, C. E., C. I. Wright, L. M. Shin, J. Kagan, and S. L. Rauch, 2003. Inhibited and uninhibited infants “grown up”: Adult amygdalar response to novelty. Science 300(5627):1952– 1953. Scott, S., M. Knapp, J. Henderson, and B. Maugham, 2001. Financial cost of social exclusion: Follow up study of antisocial children into adulthood. Br. Med. J. 323:191. Sen, S., M. Burmeister, and D. Ghosh, 2004. Meta-analysis of the association between a serotonin transporter promoter polymorphism (5-HTTLPR) and anxiety-related personality traits. Am. J. Med. Genet. 127B(1):85–89. Shioe, K., T. Ichimiya, T. Suhara, A. Takano, Y. Sudo, F. Yasuno, et al., 2003. No association between genotype of the promoter region of serotonin transporter gene and serotonin transporter binding in human brain measured by PET. Synapse 48(4):184–188. Siegle, G. J., S. R. Steinhauer, M. E. Thase, V. A. Stenger, and C. S. Carter, 2002. Can’t shake that feeling: Event-related fMRI assessment of sustained amygdala activity in response to emotional information in depressed individuals. Biol. Psychiatry 51(9):693–707. Smith, H. R., J. B. Daunais, M. A. Nader, and L. J. Porrino, 1999. Distribution of [3H]citalopram binding sites in the nonhuman primate brain. Ann. NY Acad. Sci. 877:700–702. Strobel, A., F. M. Spinath, A. Angleitner, R. Riemann, and K. P. Lesch, 2003. Lack of association between polymorphisms of the dopamine D4 receptor gene and personality. Neuropsychobiology 47(1):52–56. Tessitore, A., A. R. Hariri, F. Fera, W. G. Smith, T. N. Chase, T. M. Hyde, et al., 2002. Dopamine modulates the response of the human amygdala: A study in Parkinson’s disease. J. Neurosci. 22(20):9099–9103. van Dyck, C. H., R. T. Malison, J. K. Staley, L. K. Jacobsen, J. P. Seibyl, M. Laruelle, et al., 2004. Central serotonin transporter availability measured with [123i]beta-CIT SPECT in relation to serotonin transporter genotype. Am. J. Psychiatry 161(3):525–531. Venter, J. C., M. D. Adams, E. W. Myers, P. W. Li, R. J. Mural, G. G. Sutton, et al., 2001. The sequence of the human genome. Science 291(5507):1304–1351. Viding, E., R. J. Blair, T. E. Moffitt, and R. Plomin, 2005. Evidence for substantial genetic risk for psychopathy in 7-yearolds. J. Child Psychol. Psychiatry 46(6):592–597. Viken, R. J., R. J. Rose, J. Kaprio, and M. Koskenvuo, 1994. A developmental genetic analysis of adult personality: Extraversion and neuroticism from 18 to 59 years of age. J. Pers. Soc. Psychol. 66(4):722–730. Wallace, G. L., E. J. Schmitt, R. Lenroot, E. Viding, S. Ordaz, M. A. Rosenthal, E. A. Molloy, L. S. Clasen, K. S. Kendler, M. C. Neale, and J. N. Giedd, 2006. A pediatric twin study of brain morphometry. Behav. Genet. 35(6):707–721. Wand, G. S., M. McCaul, X. Yang, J. Reynolds, D. Gotjen, S. Lee, et al., 2002. The mu-opioid receptor gene polymorphism

viding et al.: neuroimaging and molecular genetics

365

(A118G) alters HPA axis activation induced by opioid receptor blockade. Neuropsychopharmacology 26(1):106–114. Whale, R., E. M. Clifford, and P. J. Cowen, 2000. Does mirtazapine enhance serotonergic neurotransmission in depressed patients? Psychopharmacology (Berl.) 148(3):325–326.

366

methodological paradigms

Willeit, M., N. Praschak-Rieder, A. Neumeister, W. Pirker, S. Asenbaum, O. Vitouch, et al., 2000. [123I]-beta-CIT SPECT imaging shows reduced brain serotonin transporter availability in drug-free depressed patients with seasonal affective disorder. Biol. Psychiatry 47(6), 482–489.

22

Neural Network Models of Cognitive Development YUKO MUNAKATA, JENNIFER MERVA STEDRON, CHRISTOPHER H. CHATHAM, AND MARIA KHARITONOVA

This chapter covers neural network modeling (also known as connectionist or parallel-distributed processing modeling) as a tool for studying developmental cognitive neuroscience. Neural network models provide a powerful method for exploring the complex relation between brain development and cognitive development. This chapter reviews what neural network models consist of, why modeling is useful, and how models have helped to address fundamental questions about development. Important challenges for this methodology are also discussed, along with productive directions for future work within the neural network modeling framework. Neural network models provide a powerful tool in the study of developmental cognitive neuroscience. Such models implement neural processes in computer simulations, in the form of mathematical equations that characterize neural activity and learning. Neural network simulations thus allow an exploration of the role of neural processes in behavior. The modeling methodology provides an important complement to other methods, by building upon findings from other studies and pointing the way toward new studies to advance our understanding of the relation between brain and behavior. In this chapter, we cover neural network models of cognitive development from the perspective of answering three critical methodological questions: why, what, and how? More specifically, we explain why it is important to use neural network models in the study of developmental cognitive neuroscience and to explain more about what the nuts and bolts of neural network models entail. We then describe how neural network models have been used to address fundamental developmental questions about the origins of knowledge and how change occurs. We also discuss challenges relevant to each of these issues of the why, what, and how of neural network modeling. This chapter aims to confer an appreciation of the potential contributions of neural network models to the advancement of developmental cognitive neuroscience, as well as the ability to critically evaluate both the over- and underselling of this methodology.

Why First, we describe some of the benefits of neural network modeling (adapted from O’Reilly and Munakata, 2000; see also Seidenberg, 1993; Rumelhart and McClelland, 1986; Elman et al., 1996). All these benefits are demonstrated by specific models covered later in the “How” section, and they support a productive interchange between modeling work and other methodologies. Some of these benefits are arguably conferred to some degree by purely verbal theories; however, implementing a working model of a theory is both more demanding and more powerful than simply stating the theory, and so provides greater benefits. Models Allow Control Models can be manipulated, lesioned, tested, and observed much more precisely than the thing being modeled (whether the thing is a single neuron, a small collection of neurons, a human infant, a monkey, and so on). Such control enables a clearer picture of the causal role of different factors. For example, in this chapter, we will see how such control allows an assessment of longterm effects of word frequencies in language learning. Models Help Us to Understand Behavior With such control, we can watch a model in action to get a sense of why behavior unfolds as it does. Seemingly unrelated or even contradictory behaviors can be related to one another in nonobvious ways through common neural network mechanisms. Further, neural network models can provide an important bridge between neural and cognitive aspects of behavior. Lesioned models can also provide insight into behavior following specific types of brain damage and, in turn, into normal functioning. In this chapter, we will see how models can help us to understand various potentially puzzling aspects of children’s behavior, including nonlinear trajectories in their development. Models Deal with Complexity Complex, emergent phenomena (the brain is more than the sum of its parts) can be captured in models in principled, satisfying ways. Such

367

emergent phenomena arise from the complex interactions of multiple elements of a model, without being obviously present in the behavior of the individual elements. Without the models and the principles, such complexity might otherwise be lost in vague, verbal arguments. In this chapter, we will see how models have provided insight into the emergence of complex phenomena in domains as diverse as infant object processing and children’s semantic development. Models Are Explicit Creating an implemented model forces you to be explicit about your assumptions. For example, what do children encode about a particular task, and how? How do they subsequently process this information? What kinds of mechanisms support their learning in this task? Explicitness about such assumptions confers many potential benefits, including the generation of novel, empirically testable predictions and the deconstruction of black box constructs. In this chapter we will see the advantages of such explicitness in a model that deconstructs the notion of an object-permanence concept and leads to novel predictions subsequently tested in infants.

What We will see all these benefits in action when we consider how models have been used to explore issues in cognitive development, in the next section. First, we consider the nuts and bolts of what neural network models are, which will provide the foundation for understanding their contributions. Here, we focus on five critical elements of neural network models: units, weights, net input and activation functions, and learning algorithms (for more extensive treatments of these and other nuts and bolts of neural network models, see Elman et al., 1996; O’Reilly and Munakata, 2000; Rumelhart and McClelland, 1986). Each of these elements maps onto neural constructs while capturing important aspects of psychological processing, allowing neural network models to provide an important step in understanding the relation between neural and cognitive development. Units and Weights Neural network models consist of two basic elements: units and weights (figure 22.1a). In models most closely tied to the underlying biology, each unit corresponds to a neuron, the activity of each unit corresponds to the spiking of a neuron, and each weight corresponds to a synapse (the strength of the weight corresponds to the efficacy of the synapse). Models of psychological phenomena are much more scaled down; single units correspond to collections of neurons or even entire brain regions, the activity of each unit corresponds to the overall firing rates of these neurons, and the weights between units represent synapses between the groups of neurons.

368

methodological paradigms

Units communicate with one another by means of their weights. Each unit receives activity from other units by way of its weights, and if enough such input is received, the unit becomes active. The unit then sends this activity to other units by means of its weights to those units, an action that in turn influences the activity of those units. In most network simulations of behavior, units are organized into layers. An input layer (or layers) receives information that reflects the external world, in the form of patterns of activity on the units in the layer. Networks are described as “perceiving” their environments when they receive this input information, with the particular type of perception (seeing versus hearing, and so on) depending on the modality that the input layer represents. In the simplistic example shown in figure 22.1b, the network sees the word “dog” when its input units for the letters d, o, and g are activated. An output layer (or layers) produces patterns of activity that are interpreted in terms of some response behavior. For example, the network in figure 22.1b can say either “cat” or “dog,” by activating the corresponding output unit. (For much more realistic models of word reading, which incorporate semantic representations and more complex phonological and orthographic representations, see Harm and Seidenberg, 2004; Plaut et al., 1996; O’Reilly and Munakata, 2000.) Additionally, some number of hidden layers may sit between the input and the output layers, providing the network with the capability to transform input information in useful ways to support meaningful behavior. Units can be connected by means of their weights in a variety of ways (figure 22.1a). Feedforward weights connect units in the input layer(s) to units in the hidden layer(s), and units in the hidden layer(s) to units in the output layer(s). Feedback weights may connect the units in the reverse direction (output to hidden to input). Lateral weights connect units to other units in the same layer. Recurrent weights connect units to themselves. In addition to these different directions of connectivity, weights may vary in whether they are excitatory (increasing the input to the receiving unit) or inhibitory (decreasing the input to the receiving unit). Recurrent weights that are excitatory allow units to maintain their activity by continuing to excite themselves. Lateral weights that are inhibitory lead units within a layer to compete with one another for activity, also helping active units to maintain their activity by inhibiting the activity of competing units. As will be discussed further, “knowledge” in the neural network framework takes the form of patterns of activity across the processing units and patterns of connectivity in the weights. Knowledge is thus embodied in the processing machinery (in contrast with the traditional computer metaphor, in which knowledge structures [RAM] are separable from processing [CPU]). This embodied character of knowledge in the neural network framework makes it a

Units and Weights Lateral weight Output units Feedback weight Feedforward weight Hidden units

Recurrent weight Feedforward weights

A

Input units

Net Inputs, Activations, and Learning 3. Network says “dog.” ‘cat’

‘dog’ 4. Weights change to increase activation of correct “dog” response and to decrease activation of incorrect “cat” response via error-driven learning.

4. Weights between coactive units increase via self-organizing learning.

2. Activations of units computed as a function of their net inputs.

c

B

a

t

d

o

g

1. Network sees the word “dog.”

Figure 22.1 A diagram of representative neural network architecture (A) and processing (B). Circles indicate units, and their shading indicates activation levels; arrows indicate weights.

particularly useful methodology for developmental cognitive neuroscience, given the focus of this field on understanding how knowledge is embodied by the brain, and given the parallels between principles of neural communication and relations among units and weights in neural network models. Net Input and Activation Functions The process of computing a unit’s activity is broken down into two steps: computing the net input to the unit and then computing the unit’s activity as a function of the net input. The inputs to a unit are weighted by the strength of the connections from the sending units; the stronger the connection, the more the sending unit activity contributes to the net input to the receiving unit. Mathematically, the net input to a unit j (ηj) is expressed as η j = ∑ wi j ai

where wij is a weight from unit i to unit j, and ai is the activity of unit i. The activation function specifies how the units in a network update their activity as a function of this net input. Activation functions are typically S-shaped (figure 22.2), based on a sigmoidal activation function of the following form: 1 aj = 1− e −η j where aj is the activation of the unit and η j is its net input. This S shape reflects two important aspects of neural activity, regarding the nonlinear response of neurons in relation to their inputs. First, the unit is not guaranteed to become active just because it is receiving some amount of input. As indicated by the lower-left part of the S-shaped curve, this net input must get above a certain threshold for the unit to become very active. Second, once the unit is active to some

i

munakata et al.: neural network models of cognitive development

369

Sigmoidal Activation Function 1.0

Activation

0.8 0.6 0.4 0.2 0.0 –4

–2

0

2

4

Net input

Figure 22.2 The sigmoidal activation function, reflecting the nonlinear response of neurons in relation to their inputs.

degree, it is not guaranteed to become much more active with increasing amounts of input. As indicated by the upperright part of the S-shaped curve, a unit cannot substantially increase its activity level beyond a certain point, even with further net input. This nonlinearity in the activation function allows multiple layers of units to carry out complex computations that are not possible with units using linear activation functions. Learning Algorithms Learning in neural networks takes the form of changes to the weights, which are viewed as corresponding to changes in the efficacy of synapses. Such changes occur as a result of a network’s experience with its environment, and they affect how the network responds to subsequent inputs. Because weights may take a value of zero (which is equivalent to no connection), this learning process allows for the possibility of adding new connections (when a zero weight is increased) and pruning away existing connections (when a weight goes to zero) (cf. Shultz, 2006, for more specialized mechanisms for adding and pruning connections). Here, we consider two of the primary types of learning algorithms used in neural network models—selforganizing and error driven. Self-organizing algorithms are so named because they govern learning without specifying a particular target performance; that is, they lead units to organize their weights themselves based on their local inputs, rather than in terms of meeting particular goals. One of the most common selforganizing algorithms is a Hebbian algorithm (Hebb, 1949), whereby units that are simultaneously active increase the weight between them. Mathematically, the basic form of this learning rule is Δwij = e ai a j where Δwij reflects the change in the weight from unit i to unit j, and e reflects a learning rate parameter. This form of learning has typically been used by modelers focused on

370

methodological paradigms

biological plausibility (e.g., Miller, Keller, and Stryker, 1989), because the algorithm is grounded in the known biological learning mechanisms of long-term potentiation and long-term depression (Artola, Brocher, and Singer, 1989; Bear and Malenka, 1994). However, the algorithm is not very good at solving complex tasks (whereas humans are), a fact that has led other modelers to turn to more powerful, error-driven algorithms. Error-driven algorithms are so named because they govern learning based on the discrepancy between a network’s performance and its target performance. One of the most common error-driven algorithms is the backpropagation algorithm (Rumelhart, Hinton, and Williams, 1986), whereby the difference between a unit’s activity and its target activity is computed and propagated backward through the network, so that the resulting weight changes reduce the unit’s error. In this way, the backpropagation algorithm allows networks to learn to solve complex tasks, a necessary criterion for the modeling of human behavior. Mathematically, the backpropagation learning algorithm is Δwij = e δ j ai where δj reflects the contribution of a given unit to a network’s error. Although the backpropagation algorithm has been criticized for being biologically implausible in the details of its implementation (e.g., in the backward propagation of error terms for which there is no neural evidence), biologically plausible versions have been implemented (Hinton and McClelland, 1988; Hinton, 1989; Movellan, 1990; O’Reilly, 1996; O’Reilly and Munakata, 2000). These versions avoid the implausibility problems of backpropagation by indirectly communicating error information through the standard mechanisms of neural communication, the passing of activity signals by means of weights. Further, these activity signals reflect events in the world and networks’ expectations regarding the events, such that error information can be computed based on the discrepancies between expectations and outcomes, without requiring an explicit teacher that provides target signals. Such errordriven algorithms thus allow for the continued exploration of simulating performance on complex tasks. Further, the existence of such functionally similar algorithms suggests that models using backpropagation, while biologically implausible in their detailed implementation, should not simply be discounted; lessons from them are likely to prove relevant to the biologically plausible, functionally similar implementations. As we have shown, learning algorithms can be specified in precise mathematical terms; however, it is important to note that it can nonetheless be difficult to predict exactly how networks will come to solve tasks and how they will develop, given the complex, nonlinear interactions between network units and the environment. Similarly, even with a

precise specification of how synaptic changes occur in the brain, we would not necessarily be able to explain, for example, the complex neural bases of how children learn to read. Understanding changes at the level of the synapse/ weight does not translate directly into understanding behavior. Thus, even with a precisely specified learning algorithm, it can be very difficult to predict the behavior of networks of any complexity. We can therefore gain insights into the neural bases of behavior by exploring why networks develop as they do.

How Now that we have some sense of why we might want to use neural network models as a methodology, and what they consist of, we are in a position to consider how they have contributed to the study of developmental cognitive neuroscience. Neural networks have been used to address many different facets of cognitive development, including individual differences and disorders (Joanisse and Seidenberg, 2003; MacDonald and Christiansen, 2002; Morton and Munakata, 2005; Oliver et al., 2000; Thomas and KarmiloffSmith, 2002, 2003), constructivist mechanisms of development (Schlesinger and Parisi, 2001; Shultz, 2003, 2006), the coordination of separate specialized brain systems (Jacobs, 1999; Mareschal, Plunkett, and Harris, 1995; Munakata, 2004; Westerman and Miranda, 2004), the influence of early perceptual and motor development on cognition (Jacobs and Dominguez, 2003; Westermann and Mareschal, 2004), and the development of hierarchically organized brain regions (Shrager and Johnson, 1996). Many neural network models focus more on the role of learning in development than on maturational changes (cf. Shrager and Johnson, 1996). In fact, many aspects of development that may appear to be maturational, such as critical periods, arise in neural networks as a result of learning (Ellis and Lambon-Ralph, 2000; Elman, 1993; McClelland et al., 1999; Rohde and Plaut, 1999; Seidenberg and Zevin, 2006). Similarly, many biological changes that may appear to be hardwired, such as a reduction in the plasticity of synapses across development, have been shown to depend on experience and can be reversed if experience is withheld (E. Quinlan, Olstein, and Bear, 1999). Here, we focus on neural network explorations of two fundamental issues in cognitive development: the origins of our knowledge and mechanisms of change. We aim to convey an overall sense of how neural network models can speak to these developmental issues, but because of space constraints, we can only briefly cover two examples within each of these areas. Origins Where does our knowledge come from? Questions of origins (whether of knowledge, life, the universe, etc.) form

the basis for some of the most interesting, challenging, and hotly debated issues. In the context of the origins of our knowledge, the debate has taken the form of nature versus nurture, and more recently of specifying the nature of the interactions between them. Neural network models have been used to explore the origins of knowledge in a variety of domains, including language (e.g., Rumelhart and McClelland, 1986; Plunkett and Sinha, 1991; Elman, 1993; Harm and Seidenberg, 1999; Plaut and Kello, 1999; Onnis and Christiansen, 2005), numerical understanding (Dehaene and Changeux, 1993; Verguts and Fias, 2004), and problem solving (McClelland, 1989, 1995; Shultz, Mareschal, and Schmidt, 1994). Here, we focus on models exploring the origins of knowledge of objects, specifically, their continuity and their permanence. Object continuity. Young infants appear to be sensitive to the continuity of object motion, the fact that objects move only on connected paths, never jumping from one place to another without traveling a path in between. For example, infants as young as 2.5 months look longer at events in which objects appear to move discontinuously than at otherwise similar events in which the same objects move continuously (Spelke et al., 1992). Such longer looking times are taken as an indication that infants find the discontinuous events unnatural, and so possess some understanding of object continuity. What are the origins of such knowledge? Some researchers have concluded that an understanding of object continuity is part of our innate core knowledge, given infants’ very early sensitivity to it, and the apparent difficulty in learning such information given that objects are rarely continuously visible in our environment (Spelke et al., 1992). However, as many researchers have noted, it is not clear what the label “innate” really tells us about the nature of the origins of knowledge (Elman et al., 1996; Thelen and Smith, 1994; Smith, 1999). That is, does calling infants’ sensitivity to the continuity of object “innate” tell us anything about how infants come to be sensitive to this principle, or about the mechanisms underlying such sensitivity? In contrast, the neural network approach focuses attention on exactly these kinds of issues, because such mechanisms must actually be implemented in a working model for the account to be considered successful. One such model was devised in the study of imprinting behavior in chicks and of object recognition more generally (O’Reilly and Johnson, 1994, 2002). This model viewed a simplified environment in which objects moved continuously. Based on this experience, the model developed receptive field representations of objects that encoded continuous locations in space, thereby demonstrating a sensitivity to object continuity. What were the origins of the model’s sensitivity to object continuity? First, the network had recurrent excitatory connections and lateral inhibitory connections that allowed

munakata et al.: neural network models of cognitive development

371

active units to remain active; specifically, active units continued to send activation to themselves by way of the recurrent excitatory connections, and they prevented other competing units from becoming active by way of the lateral inhibitory connections. Thus, when an object was presented as input to the network, certain hidden units became active, and they tended to stay active even as the object moved around in the input. Second, the network learned according to a Hebbian learning rule, which led the model to associate this hidden unit pattern of activity with the object in different locations in the input. Thus, whenever the object appeared in any of these locations, the network came to activate the same units, or the same object representation. In this way, with exposure to events in the world that conformed to the principle of continuity, the model developed receptive field representations of objects that encoded continuous locations in space, and so learned to “recognize” objects that moved continuously in its environment. One might argue that this model was innately predisposed to understand the continuity of objects (Spelke and Newport, 1997), given that the network was structured “from birth” with recurrent excitatory and lateral inhibitory connections and a Hebbian learning rule—all it needed was the typical experience of viewing objects moving continuously in its environment. That is, the model required experience only in a generic sense, to support an experience-expectant process (Greenough, Black, and Wallace, 1987) that would naturally unfold for all members of a species given the normal environment available throughout evolutionary history. However, again, it is not clear what benefits would be conferred by calling the developmental time course of the model “innate.” In contrast, the benefits of the model should be clear in providing an explicit, mechanistic account of the potential origins of our sensitivity to object continuity. Object permanence. Several models have been proposed to account for infants’ apparent sensitivity to the permanence of objects, with very different assumptions about the origins of object-permanence knowledge. At one extreme, such knowledge has been built into a network, with target signals specifying from birth that hidden objects continue to exist when they are hidden (Mareschal, Plunkett, and Harris, 1995). At the other extreme, a model has demonstrated limited sensitivity to the permanence of objects without ever actually developing the ability to represent hidden objects, based on the simple origins of the goal of keeping objects in view (Schlesinger and Barto, 1999). Here, we discuss a model that lies between these two extremes, in which objectpermanence knowledge developed without being prespecified (Munakata et al., 1997). The model viewed a simplified environment in which objects disappeared from view behind occluders and reappeared after the occluders were removed. Based on this experience, the model became sensitive to the

372

methodological paradigms

permanence of objects, continuing to represent objects even after they were hidden. What were the origins of the model’s knowledge of object permanence? As in the object-recognition model described earlier, the object-permanence model had recurrent excitatory connections that allowed active units to remain active. Unlike the object-recognition model, the object-permanence model also had a goal of predicting what would happen next in its environment. Through error-driven learning, the network adjusted its weights if its predictions were incorrect, for example, if the network predicted that an occluded object would not reappear when the occluder was removed, and then the object did in fact reappear. So, when a visible object moved out of view, the network gradually learned to use its recurrent connections to maintain a representation of the object, allowing the network to accurately predict its environment (and the reappearance of such hidden objects). In this way, with exposure to events that conformed to the principle of object permanence, the model provided an explicit, mechanistic account of the potential origins of our sensitivity to the permanence of objects and demonstrated how object-permanence knowledge could develop without being innately specified. The model also led to the novel prediction that infants should show greater sensitivity to the permanence of familiar objects than of novel objects. The model showed this behavior because it formed stronger representations for familiar objects, based on changes to its connection weights from repeatedly processing those objects. Those changes to the connection weights allowed the model to generalize its knowledge of object permanence to novel objects, but its representations for those novel objects were not as strong as those for familiar objects. This prediction was confirmed in infants, who searched more for familiar objects than for novel objects after they were hidden, despite showing robust preferences for novel objects over familiar objects when they were visible (Shinskey and Munakata, 2005). Change How does change occur? As many researchers have noted (e.g., Flavell, 1984; Fischer and Bidell, 1991; Siegler, 1989), this question is one of the most fundamental yet unanswered questions in the study of cognitive development. For example, how do children develop complex, higher level cognitive abilities in relatively short periods of time? Why do children sometimes show nonlinear trajectories in their development, such as stagelike progressions, sensitive periods for learning, and U-shaped learning curves? The issue of change is not mutually exclusive from the previously discussed issue of origins. Providing an explicit model of origins entails specifying mechanisms of change (unless the model assumes full-fledged knowledge from the start, an assumption that is inconsistent with the neural network framework and with brain development, not

to mention with the theories of even the most extreme nativists). So all the models described and cited in the previous section also have something to say about change as well as about origins. In the neural network framework, change can take place at multiple levels, including in the activity of units as activations are propagated through the network, the connection weight changes that occur during learning, and the emergence of new forms that arise from the complex interactions of elements in the network (as described previously for the development of representations of object continuity and permanence). Neural network models have been used to explore the mechanisms underlying many aspects of developmental change, including stagelike progressions (McClelland, 1989, 1995; Raijmakers, van Koten, and Molenaar, 1996; Thomas, 2004; P. Quinlan et al., 2007), sensitive periods in learning (Ellis and Lambon-Ralph, 2000; Elman, 1993; McClelland et al., 1999; Rohde and Plaut, 1999; Seidenberg and Zevin, 2006), and U-shaped learning curves (Munakata, 1998; Plunkett and Sinha, 1991; Rogers, Rakison, and McClelland, 2004; Rumelhart and McClelland, 1986). Here, we focus on models exploring emergent effects in language and conceptual development, specifically how relatively low-level processes can lead to the development of higher-level cognitive abilities and nonlinear developmental trajectories. We consider models exploring how such changes can explain numerous aspects of children’s word learning. Learning about words and semantic categories. How do children learn new words and form appropriate semantic categories for objects in the world? Children show particular behaviors that have led many researchers to posit high-level, conceptual (and possibly innate) structures that guide children’s learning; however, neural network models have demonstrated how more basic learning mechanisms could explain these patterns in children’s behavior. For example, when learning the name for a new solid object, 2- to 3-year-olds will reliably extend that name to other solid objects that are similar in shape (a behavior known as the “shape bias”); in contrast, after learning the name for a nonsolid substance, children will extend that name to other nonsolid things that are similar in the material they are made from (the “material bias”) (Landau, Smith, and Jones, 1988). Moreover, very early in life, children are able to differentiate abstract semantic categories, such as animals versus artifacts (Mandler and McDonough, 1993). Some researchers explain this behavior in terms of children’s early (and possibly innate) high-level, conceptual understanding about different ontological kinds of things (Carey, 2000; Keil, 1989), such as animates, objects, and substances (Booth and Waxman, 2002, 2003; Gergely et al., 1995). In contrast, other researchers have used neural network models to explore the possibility that these patterns of behavior could result from

more basic learning mechanisms that can extract higherorder regularities, among stimuli with the same labels (e.g., Colunga and Smith, 2005) and among objects from the same category (e.g., Mareschal and French, 2000; McClelland and Rogers, 2003; Quinn and Johnson, 1997; Rogers and McClelland, 2004, 2005). To test the possibility that simple learning mechanisms could lead to the emergence of word-learning biases, a neural network model was trained to associate perceptual stimuli with their labels, and then tested on its shape and material biases for solids versus nonsolids (Colunga and Smith, 2005). The model was presented with stimuli in terms of perceptual inputs that represented their shape, material, and solidity. Through error-driven learning, the model was trained to produce the correct name for each stimulus on the output layer. The vocabulary the model was trained on captured several aspects of children’s vocabulary: the number of words for solids was greater than the number of words for nonsolids, nonsolids had a more restricted range of shapes than solids, and there were strong but imperfect correlations between solid objects and names based on shape, and between nonsolid things and names based on material. After the network learned how to name 24 stimuli, it was presented with novel solids and nonsolids so that its biases to attend to shape or material could be assessed. With each novel stimulus, the network was also presented with groups of other novel stimuli that were either similar in shape to it or similar in material. The network’s shape and material biases were measured in terms of the internal representations the network formed for objects in the hidden layer, which was bidirectionally connected to both the word and the perceptual layers and was also recurrently connected to itself. Like children, the network demonstrated a clear shape bias for solids and a clear material bias for nonsolids. Specifically, the network’s internal representations for two solid objects with the same shape but different material were more similar than the internal representation for two solids with different shapes but the same material; the opposite pattern was found for nonsolids. Thus, after simply learning to associate specific words and specific perceptual features, the network formed abstract, generalized expectations about the way different stimuli could be characterized, which correspond to the types of biases observed in young children’s word learning. These simulations thus demonstrate how basic low-level learning mechanisms could lead to the development of abstract higher-order generalizations, such that one need not invoke possibly innate, conceptual structures. Similar basic learning mechanisms may support children’s acquisition of semantic categories, such as animals versus plants. Neural networks have demonstrated how such semantic categories can be formed through the learning of statistical structure in the environment (e.g., Mareschal and French, 2000; Quinn and Johnson, 1997), in particular,

munakata et al.: neural network models of cognitive development

373

through patterns of coherent covariation across objects from the same category (McClelland and Rogers, 2003; Rogers and McClelland, 2004, 2005). When objects have similar representations and share many properties (e.g., all animals move and make sound on their own, while plants do neither), the properties shared by these items will be maximally coherent and will be a strong force driving learning, because they drive changes to connection weights in the same direction. In contrast, idiosyncratic properties (e.g., the fact that some animals can fly but cannot swim, and others do the reverse) drive weights in conflicting directions that tend to cancel each other out early in learning. Overall, this process leads the most coherent properties among categories to be learned earliest, and it can explain how children progress from more coarse to more fine-grained levels of differentiation in their category learning. These coherent properties do not need to be perceptually salient (e.g., the fact that animals can grow); as long as they covary coherently, statistical learning mechanisms can use them to guide category learning. In this way, coherent covariation of properties can lead perceptually distinct items (such as birds and fish) to be viewed as part of the same category, without requiring high-level concepts of animacy. Moreover, these neural networks also illustrate how nonlinear developmental progressions (such as U-shaped learning curves) can develop simply through sensitivity to statistical regularities. In many complex tasks, such as learning to correctly apply regular or irregular verb past-tense construction, children often seem to “unlearn” a correct behavior (e.g., saying “goed” after correctly saying “went”) before eventually achieving complete mastery (Ervin, 1964). These U-shaped patterns of development have elicited explanations in terms of qualitative shifts between different abstract, highlevel rule-based systems (Marcus et al., 1992). However, basic learning mechanisms that detect coherent covariation between properties of different objects can also lead to such U-shaped patterns of development (Rogers, Rakison, and McClelland, 2004). For example, the semantic categorization networks described earlier show U-shaped progressions in how they categorize unusual animals, such as bats (which are unusual because they do not have feathers, unlike other exemplars that fly). The networks first correctly categorize bats as animals with fur, then incorrectly characterize them as animals with feathers, and ultimately characterize them correctly again. This U-shaped behavior reflects the coarseto-fine property of category development. Early on, the networks learn that most animals have fur, thus attribute this coherent characteristic to all animal exemplars, and correctly identify bats as animals with fur. As the networks learn progressively finer distinctions, such as the fact that some animals fly, they again attribute coherent characteristics of flying (e.g., flying animals have feathers) to all flying exemplars. Thus, at this point, the networks incorrectly categorize

374

methodological paradigms

bats as flying animals with feathers. Finally, as the networks learn idiosyncratic properties of each animal, they become able to correctly identify bats as flying animals with fur and not feathers. Thus these neural networks show how simple learning mechanisms that pick up on statistical regularities in the environment can lead to complex nonlinear behaviors and to the development of high-level conceptual categories, like those observed in children. Age-of-acquisition effects. As in the category-learning example, language acquisition is frequently characterized by nonlinear developmental trajectories. In contrast to theories that invoke innate or high-level structures in learning, neural network models have demonstrated how these developmental trends can be explained by relatively simple, low-level mechanisms. Another example of nonlinear change in language acquisition concerns the age of acquisition (AoA) effect, a phenomenon in which words learned early in life are recognized and pronounced faster than words learned later in life. Such patterns might suggest the existence of specialized, time-sensitive learning systems. However, the age at which children learn particular words is typically confounded with word frequency and word length, making it difficult to interpret AoA effects and their implications for how children learn (Zevin and Seidenberg, 2002). Neural network modeling, however, allows for factors such as cumulative frequency and frequency trajectory of words to be manipulated independently of one another, thereby permitting a detailed analysis of each factor’s potential impact on lexical development. To assess whether words learned early in life might enjoy processing benefits not shared by words learned later, a series of recurrent backpropagation neural networks were trained to read nearly 3,000 monosyllabic words derived from natural language corpora (Zevin and Seidenberg, 2002). The networks consisted of three feedforward layers: an orthographic input layer, a hidden layer, and a phonological output layer. A fourth layer, bidirectionally connected with the phonological layer, served to improve the accuracy of the network’s phonological outputs. Training consisted of 1 million word presentations, in which the network might receive an input like “FIST” and would then be required to produce the phonemes corresponding to that word. The frequency of some words in the training set was manipulated, such that some words were most frequently presented early in training, while other words were most frequently presented later in training. Critically, the network’s cumulative exposure to words on both the “early” and the “late” lists was equivalent by the end of training. In these models, age of acquisition was predicted by frequency trajectory, such that words from the “early” list were learned more quickly than those in the “late” lists, which

were less prevalent at the beginning of training. However, by the end of training, both lists were learned equally well, and accuracy was at ceiling, thus yielding no lasting AoA effect. Why should this outcome occur? During training, the network learned to extract regularities in the orthographyto-phonology mappings that are characteristic of English. As a result, any benefits conveyed by learning a word early were also passed on to words learned later in training. This finding provides support to the idea that the AoA effects seen in behavioral research may result from the confound of cumulative word frequency with age of acquisition. However, an extreme reduction in the similarity between the orthography-to-phonology mappings of “early” list words to “late” list words did yield a reliable AoA effect. In this case, the regularities extracted by the network based on the “early” list—and the resulting changes in connection weights—actually disadvantaged the learning of words with a very different orthography-to-phonology mapping. The network was never able to produce the later words as accurately as it had learned the earliest words, because the regularities extracted by early learning could not be passed on to words prevalent later in training. In effect, the connection weights in the network became specialized for representing the early set of orthography-to-phonology mappings; despite later experience with a very different set of mappings, this early specialization could never be completely overcome. In this case, neural network models allowed for an examination of linguistic factors that are normally very difficult to dissociate: cumulative word frequency and frequency trajectory. The results suggested that for natural languages with reliable orthography-to-phonology mappings, cumulative frequency should influence ultimate levels of skilled reading, and frequency trajectory should affect age of acquisition without any lasting AoA effects on ultimate levels of skilled reading. In other words, children should first acquire those words they encountered with the highest frequency. However, this early learning will convey benefits to many words experienced later, such that this later learning is also facilitated by the child’s earlier experience. This process serves to wash out age-of-acquisition effects, such that words acquired later share the same processing benefits enjoyed by words acquired at earlier ages. Children should only show a disproportionate advantage for producing words learned early in life if they are experienced with greater total frequency. These predictions were subsequently confirmed in behavioral research where cumulative frequency and frequency trajectory were explicitly dissociated from other characteristics that might influence the ease with which words are pronounced (Zevin and Seidenberg, 2004). Ultimately, these models allowed for an initial investigation of the factors underlying purported AoA effects before they were cleanly dissociated in a behavioral experiment. More specifically,

they permitted direct insight into how, and in what particular situations, early experience with language might result in lasting effects on linguistic behavior. Summary of How Neural Network Models Have Contributed As preceding sections have illustrated, relatively basic principles of neural network modeling can have profound implications for a variety of developmental questions, including questions of the origins of knowledge and mechanisms of developmental change. As we have seen, neural network models demonstrate how knowledge of object permanence and object continuity—knowledge that is sometimes considered to be innate—can actually arise naturally from an interaction between early life experiences and the basic beginning state of a system (e.g., in terms of initial excitatory and inhibitory connectivity). We have also described models that reproduce specific features of language learning, including word-learning biases and nonlinear developmental trajectories, using simple learning mechanisms that pick up on statistical regularities in the environment. The relevance of neural network modeling to such vastly different phenomena is a testament to this framework’s flexibility and importance. Furthermore, these models have demonstrated how specific neural mechanisms can account for a variety of developmental phenomena, in addition to generating testable (and subsequently confirmed) predictions about children’s behavior.

Challenges to the why, what, and how As in all active areas of science, each of the aspects of neural network modeling that we have discussed has been challenged in some way. Here, we focus on one important criticism within each of the areas of why, what, and how (see also discussion in Elman et al., 1996; McClelland and Plaut, 1999; O’Reilly and Munakata, 2000; Seidenberg 1993; Seidenberg and Zevin, 2006). Challenges to Why Models Are Important A common criticism of neural network models is that they can do anything, solve any task, and so on; therefore, their ability to simulate human behavior is uninteresting. That is, there are so many parameters that can be manipulated in a network that it is guaranteed to work eventually. Because getting it to work is guaranteed, this process tells us nothing. Further ammunition for this criticism comes from the fact that several different neural network models may succeed in simulating the same human behavior. They all work, and yet they can’t all be right, indicating that neural networks are simply too powerful, so a successful simulation proves nothing. Before countering this criticism using specific examples of neural network models, we first emphasize a general

munakata et al.: neural network models of cognitive development

375

response: Criticisms about too much power and too many parameters are relevant to any attempts at scientific theorizing, and are not unique to the neural network modeling endeavor. One could easily level the same criticisms at verbal theories of behavior, for example. Across a range of domains (attention, memory, language, etc.), multiple competing theories can account for the same behavioral data. And, these verbal theories are typically powerful enough to encompass any new piece of behavioral data that comes along, thanks to the vagueness of constructs and the existence of multiple free parameters (in the form of new limitations or capabilities that can be incorporated into the theory). Thus verbal theories can be constructed to explain anything, and multiple competing theories can account for the same data, so the process of developing theories tells us nothing. Most people probably would not accept this conclusion in the domain of scientific theorizing, and yet many believe it to pose a fundamental problem for neural network models. We believe that the counterargument to this criticism as applied to neural network models is similar to the counterargument to this criticism as applied to scientific theorizing more generally. Competing theories and models can be evaluated by many criteria other than simply by accounting for a set of data. People generally know when a theory feels unsatisfying, even if it is able to account for some data. For example, if a theory needs to add a new component to account for each new piece of data, it will seem more arbitrary than a more unified theory that requires no such adjustments. Or if a theory accounts for data by relying on unspecified constructs, it will seem less compelling than a more fully specified theory. In this way, the plausibility and specificity of underlying assumptions, as well as the ease with which data can be accounted for and predicted, can be evaluated to compare competing theories. The same holds true for evaluating competing models. The neural network framework may support relatively rapid progress along these lines, because the models require the underlying assumptions to be made explicit and because the assumptions are constrained by both bottom-up (biological) and top-down (psychological) information. A second counterargument to this criticism is that the number of parameters in neural network simulations may accurately reflect the diversity of underlying mechanisms that contribute to behavior, so that neural network models provide a useful tool for exploring these mechanisms. For example, in the context of developmental disorders, the same behavioral deficit may arise from any of a number of distinct underlying causes (Thomas, 2003). Moreover, this problem is not specific to disorders; any group of individuals may behave similarly and yet differ in how those behaviors are produced. The capacity of neural network models to simulate this phenomenon can thus be viewed as an important strength. Such models allow us to formally analyze multiple causality in a way that is not possible with purely

376

methodological paradigms

behavioral measures, by allowing us to independently manipulate and assess the factors contributing to emergent behavior. Finally, it is important to note that many neural network models have made their contributions by not working, that is, by not simply simulating a particular behavior that they were designed to simulate. For example, neural network models have been lesioned to simulate (and provide insight into) the behavior of patients with brain damage (Plaut et al., 1996; Farah, O’Reilly, and Vecera, 1993; Cohen et al., 1994; Farah and McClelland, 1991; Plaut, 1995; Allen and Seidenberg, 1999; Heinke and Humphreys, 2003). Such lesions are performed by removing or damaging units or their connections. In these cases, the models are not trained to simulate such atypical performance. Instead, the models are trained to perform correctly, and then they are lesioned. Altered patterns of performance emerge from the basic properties of the models following damage. Lesions or other alterations can also be performed during the course of models’ development; such models have elucidated the possible causes of various developmental disorders (Thomas and Karmiloff-Smith, 2002, 2003; Thomas, 2003; Triesch et al., 2006; Williams and Dayan, 2005). In addition, failures of neural network models (e.g., in remembering both specific events and generalizing across multiple events) have provided insight into neural divisions of labor (e.g., between the hippocampus and neocortex) (McClelland, McNaughton, and O’Reilly, 1995). Thus neural network models, though powerful, can nonetheless fail and can provide insights when they do, and they (like purely verbal theories of behavior) can be evaluated on grounds other than simply accounting for a set of data. Challenges to What Models Comprise Many challenges have been issued regarding the nuts and bolts that we have described, specifically how well the various elements of neural models map onto elements in the brain. Critics argue that the elements of models are simplistic, missing essential aspects of neural communication that render their use misguided at best. We believe that there are no quick and definitive answers to this challenge, but rather preliminary responses to be further tested and elaborated over the coming years, as part of important progress in the neural network framework. One response is simply, “Simple is good.” That is, the simplified elements of neural network models capture the essential aspects of neural communication, thereby providing a critical methodological tool for exploring the complexities of the relation between brain and behavior. We would otherwise get bogged down in details not particularly relevant to understanding cognition. An analogy may be found in the technique of creating mosaic images from a large collection of smaller images. The details of each of the smaller

images (one is a flower, another is a landscape, etc.) are not particularly relevant, and in fact, one could easily lose sight of the point of the image by focusing on these details. Instead, it is more appropriate to stand back and see the overall image at a simplified level. Neural networks may similarly provide a useful simplification of details, to allow an understanding of neural function and its relevance to cognition. For example, we can understand the efficacy of a synapse in terms of a simplified, single value of a connection weight, much as we can understand a collection of small images in terms of the simplified, overall mosaic. Without such a simplification, we might otherwise get bogged down in all the details of how synaptic efficacies are determined at the biological level (the number of vesicles of neurotransmitter released by the presynaptic neuron, the alignment and proximity of release sites and receptors, the efficacy of channels on the postsynaptic neuron, etc.). This degree of biological detail might cloud the picture of the brain-behavior relation, which is instead clarified by the simplifications of the neural network framework. Of course, the simple-is-good argument assumes that the neural network simplifications capture the essential computational properties of the biological details. This assumption can be tested by including further details in models and exploring their computational significance. In addition, the appropriateness of the simplifications can be tested by developing models at more than one level of complexity. For example, one simplification in most neural network models is the units’ continuous-valued activation term (computed from the net input as described in the “What” section), meant to approximate the rate of firing of discrete spikes. Comparisons of this simplification with more detailed models that actually fire discrete spikes have indicated that the continuous-valued activations do in fact closely approximate the firing rates of the more detailed models (O’Reilly and Munakata, 2000). Thus the simplified nature of neural network models may allow for a clearer picture of the brain-behavior relation, and the validity of these simplifications can be tested by developing models at more than one level of complexity and by testing the functional relevance of biological details. Challenges to How Models Have Contributed to Developmental Cognitive Neuroscience Various aspects of the specific models we have elaborated, together with their associated claims, have been challenged (e.g., Baillargeon and Aguiar, 1998; Marcus, 1998; Smith et al., 1999; Stadthagen-Gonzalez, Bowers, and Damian, 2004; Ghyselinck, Lewis, and Brysbaert, 2004; Booth, Waxman, and Huang, 2005). In general, we believe that many of these challenges will lead to progress in developing better models. Further, the issuing of such challenges points to a strength of the modeling framework—instantiated models can be

subsequently tested on a range of measures, highlighting their potential limitations and suggesting necessary elaborations and revisions, as well as suggesting critical empirical tests to contrast competing models. Here we focus on one criticism that has been applied to a range of models, namely, their failure to generalize (Marcus, 1998; Pinker and Prince, 1988). According to this criticism, neural network models may mimic some aspects of human performance, but the bases for human and network behavior differ vastly. Specifically, humans use rules to govern their behavior (e.g., to form the past tense of most words, add “ed”), and so they can generalize to new instances (e.g., to know that the past tense of “blicket” must be “blicketed”). In contrast, neural network models use associations to govern their behavior (e.g., “walk” is associated with “walked”), and so they cannot generalize to new instances. Therefore, although neural networks may mimic certain aspects of human performance across a range of domains, these models fail to generalize to new instances in these domains in the ways that humans can, indicating a fundamental limitation to the models. We discuss three responses to this generalization criticism (see also McClelland and Plaut, 1999; Munakata and O’Reilly, 2003; Seidenberg and Elman, 1999). The first two responses suggest that the discrepancy between human and network generalization has been exaggerated, and the third response highlights important mechanisms for generalization in the neural network framework. First, it is not clear how much of cognition is driven by rules as we have just described. Although one might sometimes be able to characterize a person’s behavior in terms of rules, this fact does not mean that those rules are explicitly instantiated in and consulted by the person (McClelland and Plaut, 1999; McClelland, 1989; Rumelhart and McClelland, 1986; Munakata et al., 1997; Thelen and Smith, 1994). This point is particularly relevant for developmental cognitive neuroscience, where test populations are often preverbal, nonverbal, or limited in their linguistic skills, and therefore unable to explicitly indicate whether they are in fact using rules to govern their behavior. Thus the assumption that rules govern behavior, and that the neural network framework must therefore incorporate rules to be considered valid, is questionable. Nonetheless, with or without rules, humans are certainly able to generalize their knowledge to new instances, so failures of neural network models to do so would seem damning. However, the claim that neural networks cannot generalize to new instances has been based predominantly on misguided testing methods (Marcus, 1998). In such tests, neural networks are trained on a particular task, but one set of input units are never activated during this training. At test, those units are activated for the first time, and the network is tested on its ability to generalize what it has learned from training

munakata et al.: neural network models of cognitive development

377

(i.e., to respond appropriately to these new instances). This type of test is misguided for two reasons. First, this test assumes that when we are presented with new instances (e.g., the word “blicket”), this action activates neurons in our brains that have never fired before. No evidence supports this assumption. Instead, extensive evidence indicates that neural patterns of firing reflect the similarity of inputs (e.g., Desimone and Ungerleider, 1989; Tanaka, 1996), suggesting that generalization to new instances would occur through the overlap between patterns of firing to the new instances and patterns of firing in previous experiences. Second, as described in the “What” section, the basic nuts and bolts of neural network (and neural) processing dictate that units must become active to support learning and meaningful behavior. Therefore, it is not particularly informative to run simulations to test the performance of units that have never become active. In sum, no evidence supports the idea that generalizing an existing ability to a new stimulus involves the activation of a pool of never-fired neurons and that neurons must become active to support learning. However, tests of network generalization have assumed such never-fired pools and consisted of testing the performance of units that have never become active. Under more plausible testing conditions, networks can generalize to new instances. For example, networks can be presented with a set of stimuli and then tested on their ability to generalize to new instances that activate novel combinations of units that have been active before. Neural networks have been shown to generalize to new instances under such circumstances across a range of domains (Colunga and Smith, 2005; Hinton, 1986; Munakata et al., 1997; O’Reilly and Munakata, 2000; Plaut et al., 1996; Rougier et al., 2005; cf. Marcus, 1998). A key factor in networks’ successful generalization (and presumably in humans’ as well) is the overlap in representations, or the extent to which a new instance is represented in a way that overlaps with previously experienced instances, guiding how to respond to the new instance. Importantly, this overlap may be present in the input-level representation to the network (e.g., as one might expect in the auditory input patterns for the new instance of “blicket” and the familiar instance of “picket”) or in higher-level rerepresentations of the input (e.g., in patterns of activity indicating that a word is a verb). Such higher-level representations can function like categories, such that once a new instance is represented appropriately at these higher levels, the network can generalize all its knowledge about the category (verbs, males, objects, etc.) to the new instance. In this way, the learning mechanisms that build on associations in neural network models support more than simple stimulus-response kinds of learning; higher-level representations allow stimuli to be encoded in more abstract and meaningful ways. Further progress in this area will likely depend upon the exploration of the factors that influence networks’ abilities

378

methodological paradigms

to form systematic representations at appropriate levels of abstraction, which can then be used to support meaningful generalizations across different tasks.

Conclusions In this chapter we have considered why neural network modeling is an important methodology for developmental cognitive neuroscience, what neural network models are, and how neural network models have contributed to addressing two fundamental issues in the study of development—the origins of knowledge and how change occurs. In addition, we have covered criticisms of neural network modeling within each of these areas of why, what, and how. In this section we will briefly review how models have offered a unique opportunity to gain insight into cognitive development. We will close with thoughts about the most productive avenues for future work in neural network modeling. As described in the “Why” section, models provide many potential advantages, including (1) allowing control, (2) helping us to understand behavior, (3) dealing with complexity, and (4) being explicit. All the models in this chapter tap each of these advantages; here, we highlight one example for each of these advantages. First, the ability to control the frequencies of words that a model was exposed to provided insight into sources of apparent age-of-acquisition effects in children’s word learning (Zevin and Seidenberg, 2002). This ability to manipulate the training environment in such a controlled manner and to observe the long-term effects on language learning is unique to the modeling framework. Second, the ability to watch representations develop in a model provided an understanding of how children might progress from more coarse to more fine-grained semantic categories and how this process could lead to U-shaped patterns of development (McClelland and Rogers, 2003; Rogers & McClelland, 2004, 2005; Rogers, Rakison, and McClelland, 2004). This ability to watch learning unfold in networks can help us to understand behavior at a more mechanistic level than would otherwise be possible. Third, the ability to deal with complexity allowed a model to provide a principled account of the potential origins of infants’ sensitivity to object continuity (O’Reilly and Johnson, 1994, 2002). A purely verbal description of the complex process of developing receptive fields that encode continuous locations in space would probably appear vague; the model instead shows how this process can emerge naturally in a network. Finally, the need to be explicit about various assumptions in implementing a working model led to the deconstruction of the object permanence concept into specific learning mechanisms and resulting representations (Munakata et al., 1997) and motivated novel behavioral predictions that were subsequently confirmed (Shinskey and Munakata, 2005). Without the forcing function of

explicitness found in the modeling framework, such constructs often remain only black boxes in purely verbal theoretical accounts. Of course, all these advantages of the neural network modeling methodology rely on the existence of careful empirical studies, which lay out the important phenomena to be addressed and help test competing models. Models cannot stand alone and are meant to be put forth as complementary (rather than superior) to empirical studies, for the reasons elaborated previously. While this point may seem obvious, some criticisms of modeling have seemed to assume that the modeling methodology must be held to a higher standard than empirical work. Specifically, one criticism is that each parameter is not varied and systematically tested in neural network modeling, so that it can be hard to know which parameters are crucial to a network’s behavior (McCloskey, 1991; Mandler, 1998). However, the same criticism can be applied to empirical work. Typically the parameter of interest (e.g., delay in a memory task) is varied and its effects measured. Other parameters (e.g., the size of the testing room) are viewed as less relevant and are not varied. In both modeling and empirical methodologies, further progress can be made by subsequently testing such assumptions about which factors are relevant. Such progress has been made more rapidly with empirical methodologies, because the same testing paradigms are often employed by multiple different researchers, helping to isolate which factors are relevant to behavior. As the field of modeling continues to develop, with new models replicating and building on prior models, similar progress in isolating critical factors should result. This argument brings us to our final point, which focuses on the most productive way to proceed with neural network modeling as a methodology. We believe it will be most fruitful if researchers appreciate both the strengths and the limitations of neural network models (and recognize that some of the limitations are equally applicable to empirical work and to verbal theorizing), such that subsequent models can be developed that build on the strengths and begin to address the limitations. Although, again, this point may seem obvious, the field has tended to miss this kind of balance, instead oscillating between extreme hype (models should be fully accepted simply because they work) and extreme skepticism (models should be completely rejected simply because someone shows some limitation in them). As a caution against extreme hype, we have emphasized specific contributions from neural network models to our understanding of the processes of cognitive development (not simply touting the fact that a model works), and we have tried to underscore the need to evaluate models (like theories) on a range of criteria other than simply working. As a caution against extreme skepticism, we note that all models involve simplifications and, in turn, limitations, so it is not particularly constructive

to simply point out limitations and argue that models should thus be discounted. Rather, it will be most productive if an understanding of limitations can support the development of alternative models, which can then be evaluated on similar grounds. Again, it may be useful to consider the parallels with more traditional empirical work and verbal theorizing. Researchers rarely critique theories without providing alternatives or run studies simply to disprove others’ theories. Rather, researchers typically put forth alternate theories to account for the data, theories that are on the same playing field as the original theories, equally susceptible to criticism, testing, and so on. We believe that this same process would greatly benefit progress in the modeling endeavor. That is, we will make the most progress by specifying alternative models that build on existing strengths and begin to address limitations. In this way, better models will be developed that tap the unique advantages of this methodology, continuing to advance our understanding of developmental cognitive neuroscience. acknowledgments

Preparation of this chapter was supported by research grants from NIMH (MH59066-01), NICHD (HD37163), and NSF (IBN-9873492). We thank Eliana Colunga, Randy O’Reilly, Rob Roberts, Marshall Haith, and members of the Cognitive Development Center for useful comments and discussions.

REFERENCES Allen, J., and M. S. Seidenberg, 1999. The emergence of grammaticality in connectionist networks. In B. MacWhinney, ed., The Emergence of Language, 115–151. Mahwah, NJ: Lawrence Erlbaum. Artola, A., S. Brocher, and W. Singer, 1989. Different voltagedependent thresholds for inducing long-term depression and long-term potentiation in slices of rat visual cortex. Nature 347:69–72. Baillargeon, R., and A. Aguiar, 1998. Toward a general model of perseveration in infancy. Dev. Sci. 1:190–191. Bear, M. F., and R. C. Malenka, 1994. Synaptic plasticity: LTP and LTD. Cur. Opin. Neurobiol. 4:389–399. Booth, A., and S. Waxman, 2002. Word learning is “smart”: Evidence that conceptual information affects preschoolers’ extension of novel words. Cognition 84:B11–B22. Booth, A., and S. Waxman, 2003. Mapping words to the world in infancy: Infants’ expectations for count nouns and adjectives. J. Cogn. Dev. 4:357–381. Booth, A., S. R. Waxman, and Y. T. Huang, 2005. Conceptual information permeates word learning in infancy. Dev. Psychol. 41(3):491–505. Carey, S., 2000. The origins of concepts. J. Cogn. Dev., 1:37– 42. Cohen, J. D., R. D. Romero, M. J. Farah, and D. ServanSchreiber, 1994. Mechanisms of spatial attention: The relation of macrostructure to microstructure in parietal neglect. J. Cogn. Neurosci. 6(4):377–387.

munakata et al.: neural network models of cognitive development

379

Colunga, E., and L. B. Smith, 2005. From the lexicon to expectations about kinds: A role for associative learning. Psychol. Rev. 112:347–382. Dehaene, S., and J.-P. Changeux, 1993. Development of elementary numerical abilities: A neuronal model. J. Cogn. Neurosci. 5:390–407. Desimone, R., and L. G. Ungerleider, 1989. Neural mechanisms of visual processing in monkeys. In F. Boller and J. Grafman, eds., Handbook of Neuropsychology, vol. 2, chap. 14, pp. 267–299. New York: Elsevier Science. Ellis, A. W., and M. Lambon-Ralph, 2000. Age of acquisition effects in adult lexical processing reflect loss of plasticity in maturing systems: Insights from connectionist networks. J. Exp. Psychol. [Learn. Mem. Cogn.] 26:1103–1123. Elman, J. L., 1993. Learning and development in neural networks: The importance of starting small. Cognition 48(1):71–79. Elman, J., E. Bates, A. Karmiloff-Smith, M. Johnson, D. Parisi, and K. Plunkett, 1996. Rethinking Innateness: A Connectionist Perspective on Development. Cambridge, MA: MIT Press. Ervin, S., 1964. Imitation in children’s language. In E. H. Lenneberg, ed., New Directions in the Study of Language, 163–198. Cambridge, MA: MIT Press. Farah, M. J., and J. L. McClelland, 1991. A computational model of semantic memory impairment: Modality specificity and emergent category specificity. J. Exp. Psychol. [Gen.] 120:339– 357. Farah, M. J., R. C. O’Reilly, and S. P. Vecera, 1993. Dissociated overt and covert recognition as an emergent property of a lesioned neural network. Psychol. Rev. 100:571–588. Fischer, K. W., and T. Bidell, 1991. Constraining nativist inferences about cognitive capacities. In S. Carey and R. Gelman, eds., The Epigenesis of Mind, chap. 7, pp. 199–236. Hillsdale, NJ: Lawrence Erlbaum. Flavell, J. H., 1984. Discussion. In R. J. Sternberg, ed., Mechanisms of Cognitive Development. New York: Freeman. Gergely, G., Z. Nadasdy, G. Csibra, and S. Biro, 1995. Taking the intentional stance at 12 months of age. Cognition 56:165– 193. Greenough, W. T., J. E. Black, and C. S. Wallace, 1987. Experience and brain development. Child Dev. 58:539–559. Ghyselinck, M., M. B. Lewis, and M. Brysbaert, 2004. Age of acquisition and the cumulative-frequency hypothesis: A review of the literature and a new multi-task investigation. Acta Psychologica 115:43–67. Harm, M. W., and M. S. Seidenberg, 1999. Phonology, reading acquisition, and dyslexia: Insights from connectionist models. Psychol. Rev. 106:491–528. Harm, M. W., and M. S. Seidenberg, 2004. Computing the meanings of words in reading: Cooperative division of labor between visual and phonological processes. Psychol. Rev. 111:662–720. Hebb, D. O., 1949. The Organization of Behavior. New York: John Wiley and Sons. Heinke, D., and G. W. Humphreys, 2003. Attention, spatial representation and visual neglect: Simulating emergent attentional processes in the Selective Attention for Identification Model (SAIM). Psychol. Rev. 110:29–87. Hinton, G. E., 1986. Learning distributed representations of concepts. Proceedings of the Eighth Annual Conference of the Cognitive Science Society, 1–12. Hillsdale, NJ: Lawrence Erlbaum. Hinton, G. E., 1989. Deterministic Boltzmann learning performs steepest descent in weight-space. Neural Comput. 1:143–150. Hinton, G. E., and J. L. McClelland, 1988. Learning representations by recirculation. In D. Z. Anderson, ed., Neural Information

380

methodological paradigms

Processing Systems, 1987, 358–366. New York: American Institute of Physics. Jacobs, R. A., 1999. Computational studies of the development of functionally specialized neural modules. Trends Cogn. Sci. 3: 31–38. Jacobs R. A., and M. Dominguez, 2003. Visual development and the acquisition of motion velocity sensitivities. Neural Comput. 15(4):761–781. Joanisse, M., and M. S. Seidenberg, 2003. Phonology and syntax in specific language impairment: Evidence from a connectionist model. Brain Lang. 86:40–56. Keil, F., 1989. Concepts, Kinds, and Cognitive Development. Cambridge, MA: MIT Press. Landau, B., L. B. Smith, and S. S. Jones, 1988. The importance of shape in early lexical learning. Cogn. Dev. 3(3):299– 321. MacDonald, M., and M. Christiansen, 2002. Reassessing working memory: A reply to Just and Carpenter and Waters and Caplan. Psychol. Rev. 109:35–54 Mandler, J. M., 1998. On theory and modeling. Dev. Sci. 2(1):196– 197. Mandler, J. M., and L. McDonough, 1993. Concept formation in infancy. Cogn. Dev. 8:291–318. Marcus, G. F., 1998. Rethinking eliminative connectionism. Cogn. Psych. 37:243. Marcus, G. F., S. Pinker, M. Ullman, M. Hollander, T. J. Rosen, and F. Xu, 1992. Overregularization in language acquisition. Monogr. Soc. Res. Child Dev. 4 (Serial No. 228). Mareschal, D., and R. M. French, 2000. Mechanisms of categorization in infancy. Infancy 1:59–76. Mareschal, D., K. Plunkett, and P. Harris, 1995. Developing object permanence: A connectionist model. Proceedings of the 17th Annual Conference of the Cognitive Sciience Society, 170–175. Hillsdale, NJ: Lawrence Erlbaum. McClelland, J. L., 1989. Parallel distributed processing: Implications for cognition and development. In R. G. M. Morris, ed., Parallel Distributed Processing: Implications for Psychology and Neurobiology, chap. 2, pp. 8–45. New York: Oxford University Press. McClelland, J. L., 1995. A connectionist perspective on knowledge and development. In T. J. Simon, and G. S. Halford, eds., Developing Cognitive Competence: New Approaches to Process Modeling, 157–204. Hillsdale, NJ: Erlbaum. McClelland, J. L., B. L. McNaughton, and R. C. O’Reilly, 1995. Why there are complementary learning systems in the hippocampus and neocortex: Insights from the successes and failures of connectionist models of learning and memory. Psychol. Rev. 102:419–457. McClelland, J., and D. Plaut, 1999. Does generalization in infant learning implicate abstract algebra-like rules? Trends Cogn. Sci. 3:166–168. McClelland, J. L., and T. T. Rogers, 2003. The parallel distributed processing approach to semantic cognition. Nature Rev. Neurosci., 4:310–322. McClelland, J. L., A. Thomas, B. D. McCandliss, and J. A. Fiez, 1999. Understanding failures of learning: Hebbian learning, competition for representational space, and some preliminary experimental data. Prog. Brain Res. 121:75–80. McCloskey, M. 1991. Networks and theories: The place of connectionism in cognitive science. Psychol. Sci. 2(6):387–395. Miller, K. D., J. B. Keller, and M. P. Stryker, 1989. Ocular dominance column development: Analysis and simulation. Science 245:605–615.

Morton, J. B., and Y. Munakata, 2005. What’s the difference? Contrasting modular and neural network approaches to understanding developmental variability. J. Dev. Behav. Pediatr. 26: 128–139. Movellan, J. R., 1990. Contrastive Hebbian learning in the continuous Hopfield model. In D. S. Touretzky, G. E. Hinton, and T. J. Sejnowski, eds., Proceedings of the 1989 Connectionist Models Summer School, 10–17. San Mateo, CA: Morgan Kaufman. Munakata, Y., 1998. Infant perseveration and implications for object permanence theories: A PDP model of the AB task. Dev. Sci. 1:161–184. Munakata, Y., 2004. Computational cognitive neuroscience of early memory development. Dev. Rev. 24:133–153. Munakata, Y., J. L. McClelland, M. H. Johnson, and R. Siegler, 1997. Rethinking infant knowledge: Toward an adaptive process account of successes and failures in object permanence tasks. Psychol. Rev., 104(4):686–713. Munakata, Y., and R. C. O’Reilly, 2003. Developmental and computational neuroscience approaches to cognition: The case of generalization. Cogn. Stud. 10:76–92. Oliver, A., M. H. Johnson, A. Karmiloff-Smith, and B. Pennington, 2000. Deviations in the emergence of representations: A neuroconstructivist framework for analysing developmental disorders. Dev. Sci. 3:1–23. Onnis, L., and M. H. Christiansen, 2005. Happy endings for absolute beginners: Psychological plausibility in computational models of language acquisition. In Proceedings of the 27th Annual Meeting of the Cognitive Science Society, 1678–1683. Mahwah, NJ: Lawrence Erlbaum. O’Reilly, R. C., 1996. Biologically plausible error-driven learning using local activation differences: The generalized recirculation algorithm. Neural Comput. 8(5):895–938. O’Reilly, R. C., and M. H. Johnson, 1994. Object recognition and sensitive periods: A computational analysis of visual imprinting. Neural Comput. 6(3):357–389. O’Reilly, R. C., and M. H. Johnson, 2002. Object recognition and sensitive periods: A computational analysis of visual imprinting. In M. H. Johnson, R. O. Gilmore, and Y. Munakata, eds., Brain Development and Cognition: A Reader, 2nd ed. Oxford, UK: Blackwell. O’Reilly, R. C., and Y. Munakata, 2000. Computational Explorations in Cognitive Neuroscience. Cambridge, MA: MIT Press. Pinker, S., and A. Prince, 1988. On language and connectionism: Analysis of a parallel distributed processing model of language acquisition. Cognition 28:73–193. Plaut, D. C., 1995. Double dissociation without modularity: Evidence from connectionist neuropsychology. J. Clin. Exp. Neuropsychol. 17(2):291–321. Plaut, D., and C. Kello, 1999. The emergence of phonology from the interplay of speech comprehension and production: A distributed connectionist approach. In B. MacWhinney, ed., The Emergence of Language, 381–415. Mahwah, NJ: Lawrence Erlbaum. Plaut, D. C., J. L. McClelland, M. S. Seidenberg, and K. E. Patterson, 1996. Understanding normal and impaired word reading: Computational principles in quasi-regular domains. Psychol. Rev. 103:56–115. Plunkett, K., and C. Sinha, 1991. Connectionism and developmental theory. Psykologisk Skriftserie Aarhus 16(1):1–34. Quinlan, E. M., D. H. Olstein, and M. F. Bear, 1999. Bidirectional, experience-dependent regulation of N-methylD-aspartate receptor subunit composition in the rat visual

cortex during postnatal development. Proc. Natl. Acad. Sci. USA 96:12876–12880. Quinlan, P. T., H. L. J. van der Maas, B. R. J. Jansen, O. Booij, and M. Rendell, 2007. Rethinking stages of cognitive development: An appraisal of connectionist models of the balance scale task. Cognition 103(3):413–459. Quinn, P. C., and M. H. Johnson, 1997. The emergence of category representations in infants: A connectionist analysis. J. Exp. Child Psychol. 66:236–263. Raijmakers, M. E., S. van Koten, and P. C. Molenaar, 1996. On the validity of simulating stagewise development by means of PDP networks: Application of catastrophe analysis and an experimental test of rule-like network performance. Cogn. Sci. 20: 101–136. Rogers, T. T., and J. L. McClelland, 2004. Semantic Cognition: A Parallel Distributed Processing Approach. Cambridge, MA: MIT Press. Rogers, T. T., and J. L. McClelland, 2005. A parallel distributed processing approach to semantic cognition: Applications to conceptual development. In L. Gershkoff-Stowe and D. Rakison, eds., Building Object Categories in Developmental Time: 32nd Carnegie Symposium on Cognition. Mahwah, NJ: Erlbaum. Rogers, T. T., D. H. Rakison, and J. L. McClelland, 2004. Ushaped curves in development: A PDP approach. J. Cogn. Dev. 5:137–145. Rohde, D., and D. Plaut, 1999. Language acquisition in the absence of explicit negative evidence: How important is starting small? Cognition 72:67–109. Rougier, N. P., D. Noelle, T. S. Braver, J. D. Cohen, and R. C. O’Reilly, 2005. Prefrontal cortex and the flexibility of cognitive control: Rules without symbols. Proc. Natl. Acad. Sci. USA 102:7338–7343. Rumelhart, D. E., G. E. Hinton, and R. J. Williams, 1986. Learning representations by back-propagating errors. Nature 323:533–536. Rumelhart, D. E., and J. L. McClelland, 1986. PDP models and general issues in cognitive science. In D. E. Rumelhart, J. L. McClelland, and PDP Research Group, eds., Parallel Distributed Processing, vol. 1: Foundations, chap. 4, pp. 110–146. Cambridge, MA: MIT Press. Schlesinger, M., and A. Barto, 1999. Optimal control methods for simulating the perception of causality in young infants. In M. Hahn and S. C. Stoness, eds., Proceedings of the 21st Annual Conference of the Cognitive Science Society, 625–630. Hillsdale, NJ: Lawrence Erlbaum. Schlesinger, M., and D. Parisi, 2001. The agent-based approach: A new direction for computational models of development. Dev. Rev. 21:121–146. Seidenberg, M., 1993. Connectionist models and cognitive theory. Psychol. Sci. 4(4):228–235. Seidenberg, M. S., and J. L. Elman, 1999. Do infants learn grammar with algebra or statistics? Science 284:435–436. Seidenberg, M. S., and J. D. Zevin, 2006. Connectionist models in developmental cognitive neuroscience: Insights about critical periods. In Y. Munakata and M. H. Johnson, eds., Processes of Change in Brain and Cognitive Development: Attention and Performance XXI, 585–612. Oxford, UK: Oxford University Press. Shinskey, J. L., and Y. Munakata, 2005. Familiarity breeds searching: Infants reverse their novelty preferences when reaching for hidden objects. Psychol. Sci. 16:596–600. Shrager, J., and M. H. Johnson, 1996. Dynamic plasticity influences the emergence of function in a simple cortical array. Neural Net. 9:1119.

munakata et al.: neural network models of cognitive development

381

Shultz, T. R., 2003. Computational Developmental Psychology. Cambridge, MA: MIT Press. Shultz, T., 2006. Constructive learning in the modeling of psychological development. In Y. Munakata and M. H. Johnson, eds., Processes of Change in Brain and Cognitive Development: Attention and Performance XXI, 62–86. Oxford, UK: Oxford University Press. Shultz, T., D. Mareschal, and W. Schmidt, 1994. Modeling cognitive development on balance scale phenomena. Machine Learn. 16:57–86. Siegler, R., 1989. Mechanisms of cognitive development. Annu. Rev. Psych. 40:353–379. Smith, L. B., 1999. Do infants possess innate knowledge structures? The con side. Dev. Sci. 2(2):133–144. Smith, L. B., E. Thelen, B. Titzer, and D. McLin, 1999. Knowing in the context of acting: The task dynamics of the A-not-B error. Psychol. Rev. 106:235–260. Spelke, E., K. Breinlinger, J. Macomber, and K. Jacobson, 1992. Origins of knowledge. Psychol. Rev. 99:605–632. Spelke, E., and E. Newport, 1997. Nativism, empiricism, and the development of knowledge. In R. M. Lerner, ed., Theoretical Models of Human Development. In W. Damon, series ed., Handbook of Child Psychology, 5th ed. New York: Wiley. Stadthagen-Gonzalez, H., J. S. Bowers, and M. F. Damian, 2004. Age-of-acquisition effects in visual word recognition: Evidence from expert vocabularies. Cognition 93(1): B11–26. Tanaka, K., 1996. Inferotemporal cortex and object vision. Annu. Rev. Neurosci. 19:109–139. Thelen, E., and L. B. Smith, 1994. A Dynamic Systems Approach to the Development of Cognition and Action. Cambridge, MA: MIT Press.

382

methodological paradigms

Thomas, M. S. C., 2003. Multiple causality in developmental disorders: Methodological implications from computational modeling. Dev. Sci. 6(5):537–556. Thomas, M. S. C., 2004. How do simple connectionist networks achieve a shift from “featural” to “correlational” processing in categorization? Infancy 5(2):199–208. Thomas, M. S. C., and A. Karmiloff-Smith, 2002. Are developmental disorders like cases of adult brain damage? Implications from connectionist modeling. Behav. Brain Sci. 25:727– 788. Thomas, M. S. C., and A. Karmiloff-Smith, 2003. Modeling language acquisition in atypical phenotypes. Psychol. Rev. 110(4):647–682. Triesch, J., C. Teuscher, G. O. Deák, and E. Carlson, 2006. Gaze following: Why (not) learn it? Dev. Sci. 9(2):125–147. Verguts, T., and W. Fias, 2004. Representation of number in animals and humans: A neural model. J. Cogn. Neurosci. 16(9): 1493. Westerman, G., and D. Mareschal, 2004. From parts to wholes: Mechanisms of development in infant visual object processing. Infancy 5(2):131–151. Westerman, G., and E. R. Miranda, 2004. A new model of sensorimotor coupling in the development of speech. Brain Lang. 89(2):393–400. Williams, J., and P. Dayan, 2005. Dopamine, learning, and impulsivity: A biological account of attention-deficit/hyperactivity disorder. J. Child Adolesc. Psychopharmacol. 15(2):160–179. Zevin, J., and M. Seidenberg, 2002. Age of acquisition effects in reading and other tasks. J. Mem. Lang. 47:1–29. Zevin, J., M. Seidenberg, 2004. Age of acquisition effects in reading aloud: Tests of cumulative frequency and frequency trajectory. Mem. Cogn. 32:31–38.

III NEURAL PLASTICITY IN DEVELOPMENT

23

Early Brain Injury, Plasticity, and Behavior BRYAN KOLB, WENDY COMEAU, AND ROBBIN GIBB

The goal of this chapter is to illustrate some of the general principles underlying how injury to the developing brain, particularly to the cerebral cortex, can lead to alterations in brain and behavioral development and how such alterations can be modulated. In principle, there are three ways that the brain could show plastic changes that might support recovery during development. First, there could be changes in the organization of the remaining intact circuits in the brain. These would likely involve the generation of new synapses in extant pathways. Second, there could be a development of new circuitry that is novel to the injured brain. Third, there could be a generation of neurons and glia to replace at least some lost neurons and glia. In fact, the developing brain makes use of all of these changes, although the details vary with the precise developmental age at the time of injury. Furthermore, each of these outcomes can be influenced by various modulating factors, including especially experience, neuromodulators, and gonadal hormones. We will begin by considering how the normal brain changes in response to experience during development and then consider each of the three types of changes that occur in the developing, injured brain. Finally, we consider the way in which the plastic changes, and ultimately the behavior, are modulated. Because most of what we know about the changes in the normal and injured brain comes from studies of the rat brain, our discussion will focus on a consideration of the rat. We are confident, however, that our results will generalize to other mammalian species, especially humans.

Changes in the normal brain In order to understand how the brain can be changed to support functional restitution, we shall first consider how the normal brain can be changed. The general logic is that the nervous system is conservative and that plastic changes that normally occur during development are likely to be recruited in an attempt to repair the abnormal brain. Neuronal Changes during Development There are two aspects of neural development that are especially important in the current context. First, neurons of the neocortex of rats

are generated from about embryonic day 12 to 21, with birth being on day 22. Thus the genesis of cortical neurons begins about two-thirds of the way through gestation and is complete at about the time of birth. Neurons and glia arise from neural stem cells that reside in the ventricular zone of the developing brain. Neural stem cells can divide symmetrically, to produce two stem cells, or asymmetrically, to produce a stem cell and a progenitor cell. Stem cells can be thought of as multipotent cells that have the potential to reproduce themselves continuously throughout the lifetime of an organism. In contrast, progenitor cells have a limited capacity for reproduction and are destined to produce neurons or glia. Stem cells are located in the subependymal zone and remain active throughout life, although they may have a finite number of divisions before they die. It is believed that progenitor cells that can divide to produce neurons and/or glia can migrate away from the subependymal zone and may lie quiescently in the white or gray matter. While in these locations, they can be activated to produce neurons and/or glia. These cells likely form the basis of at least one form of postnatal neurogenesis, especially after injury. (For useful reviews, see Becq et al., 2005; Gregg, Shingo, and Weiss, 2001.) One challenge with stem and progenitor cells is to find the “switch” to turn on controlled cell production when they are needed after an injury. There is now considerable evidence that the mammalian brain, including the primate brain, can generate neurons destined for the olfactory bulb, hippocampal formation, and possibly even the neocortex of the frontal and temporal lobes (Eriksson et al., 1998; Gould et al., 1999; Kempermann and Gage, 1999). The reason for this generation is not at all clear at present, although it may function to enhance brain plasticity, particularly with respect to processes underlying learning and memory. It turns out, however, that although the generation of these neurons occurs continuously in the normal brain, the generation rate is influenced by injury in the developing brain, which, as we shall see, can lead to enhanced cortical regeneration and functional recovery (Kolb, Gibb, et al., 1998). The second important stage of neural development is the formation of synapses. As neurons migrate to their final destinations, they begin to develop axons and dendrites that will form synapses. Synapses do not form randomly on

385

cortical pyramidal neurons but show important characteristics. One feature is that most excitatory inputs (up to 95%) synapse on spines, which are found on the dendrites but rarely on or near the cell bodies. This means that it is possible to estimate excitatory synaptic numbers by estimating the number of spines, which is typically done by estimating both the length of dendritic material and the density of spines on the dendrites. Both dendrites and spines grow rapidly during development and show remarkable plasticity in adulthood as dendrites can form spines and axons can form new axon terminations in hours and possibly even minutes after some experience (Greenough and Chang, 1988). This process is most expedient during development and is modified by changes in the developing brain, including experience and injury. Experience-Dependent Changes One of the key principles of behavioral neuroscience is that experience can modify brain structure long after brain development is complete. Indeed, it is generally assumed that structural changes in the brain accompany memory storage (Bailey and Kandel, 1993; Kolb and Whishaw, 1998). We can now identify a large range of neural changes associated with experience. These include increases in brain size, cortical thickness, neuron size, dendritic branching, spine density, synapses per neuron, and glial numbers (Greenough and Chang, 1988). The magnitude of these changes should not be underestimated. For example, in our own studies of the effects of housing rats in enriched environments, we consistently see changes in overall brain weight in the order of 7–10 percent after 60 days in young animals (Kolb, 1995). This increase in brain weight represents increases in glia, blood vessels, neuron soma size, dendritic elements, and synapses. It would be difficult to estimate the total number of increased synapses, but it is probably in the order of 20 percent in the cortex, which is an extraordinary change. The magnitude of the effect is easily seen by examining the gross morphology of cortical neurons using Golgi-type stains, as illustrated in figure 23.1. Most studies of experience-dependent change manipulate experience either by providing special training or experiences or by housing animals in specific types of environments. For example, laboratory animals can be trained to make specific complex movements, such as reaching through a slot for food, or they can be placed in complex environments. In the former case there are changes in the morphology of cells in specific regions, such as primary motor cortex, whereas in the latter case the cellular changes are more global, presumably reflecting the more global activation of cerebral structures. Studies on the effects of experience in the normal developing brain have shown that experience can have very different effects upon the brain at different ages. For example, when rats are placed into complex

386

neural plasticity in development

Figure 23.1 Golgi-stained pyramidal neuron from parietal cortex. Analysis of such cells allows inferences regarding changes in neural networks’ underlying behavior.

environments for 3 months beginning either at weaning (21 days of age) or young adulthood (4 months of age), there is a qualitative difference. Young adult rats show a large increase in dendritic length in cortical neurons and an associated increase in spine density. In contrast, juvenile rats show a similar increase in dendritic length but a decrease in spine density. That is, the young animals show a qualitatively different change in the distribution of synapses on pyramidal neurons compared to older animals (Kolb, Gibb, and Gorny, 2003). Although infant animals cannot be given the same type of experience, we followed a procedure first used by Schanberg and Field (1987) and found that if infant rats are given tactile stimulation with a small brush for just 15 min, three times per day for 10 days, there is no change in dendritic length, but there is, paradoxically, a drop in spine density that is still present in adulthood. A drop in spine density implies that there is a drop in the number of synapses per neuron, which is a curious result. (It is not without precedent, however, as young chicks show a similar drop in spine density when they are imprinted on specific environmental stimuli; Wallhousser and Scheich, 1987.) We must also hasten to point out that a drop in synapses per neuron does not necessarily mean that there is an overall drop in synapses in the cortex. It is possible that the experience either stimulates the production of more neurons or,

more likely, the experience retards the normal programmed loss of neurons during development. In either case, the actual number of synapses in the cortex could be increased rather than decreased with the early experience. This possibility is consistent with our observation that the drop in spine density is not benign but rather is associated with chronic behavioral changes. For example, tactilely stimulated animals have enhanced fine motor and spatial navigation skills in adulthood. This result is remarkable because it shows that two weeks of tactile stroking in infancy significantly enhanced motor learning in adulthood. The mechanism for this enhanced learning is as yet unknown but we have found that the early stroking leads to a change in the expression of basic fibroblast growth factor (FGF-2) in both the skin and cortex (Gibb and Kolb, 2008b). Direct administration of the FGF-2 produces similar changes, leading us to the conclusion that one route to influencing brain development is through the skin. But what is perhaps even more intriguing is that the early tactile experience is especially effective in stimulating recovery from early brain injury, a result that we will discuss in more detail later. One of the surprises in our studies was to discover that prenatal experiences also could produce large postnatal changes in brain structure. For example, we have found recently that housing pregnant females in complex environments leads to increased spine density in cortical neurons of adult rats while at the same time decreasing overall dendritic length (Gibb, Gonzalez, and Kolb, 2008). In a parallel study we tactilely stimulated pregnant females and again found chronic changes in the structure of the cortex of the offspring. The changes mirrored the changes in animals in the complex environments: there was a decrease in dendritic length but an increase in spine density in adulthood (Gibb and Kolb, 2008a). In addition, we found a 25 percent increase in acetylcholine levels in the offspring of the treated mothers. Although we do not yet understand the mechanism(s) of these changes, it is our hunch that either the prenatal treatments are altering later gene expression by means of some sort of epigenetic effect or the treatments are increasing the production FGF-2 as we found in the postnatal treatments. In summary, it has long been assumed in the psychological literature that experiences in early childhood have greater effects on later behavior than do similar experiences in adulthood. Our analysis of behavioral and dendritic effects of experience-dependent changes following exposure to specific experiences during development suggests that there is a structural basis to this differential effect of early experience on behavior. Our results lead us to several conclusions. First, “enriched” experience can have very different effects upon the brain at different ages. Second, experience not only leads to “more” but can also lead to “less.” That is, although there is a temptation to presume that experiences lead to increased

numbers of synapses and probably to increases in glia, it appears that there may be either increases or decreases, the details varying with age at experience. Third, changes in dendritic length and dendritic spine density are clearly dissociable. It is not immediately clear what the differences mean in terms of neuronal function, but it is clear that experience can alter these two measures independently and in different ways at different ages. Fourth, prenatal experiences can alter brain and behavioral development, a conclusion that is somewhat surprising given that the experiences are occurring at a time when the brain is so immature. Finally, because the changes in neural structure that are associated with experience are correlated with more proficient production of a variety of behaviors, it is reasonable to expect that similar functional and morphological changes might be observed in animals with cerebral injuries. Note, however, that there is not a single change to look for in the injured brain but rather several different types of changes.

Behavioral sequelae of early brain injury Perhaps the best-known studies on the effects of early brain injury on behavior were those performed by Margaret Kennard in the late 1930s (Kennard, 1942). She made unilateral motor cortex lesions in infant and adult monkeys. The behavioral impairments in the infant monkeys were milder than those in the adults, leading Kennard to hypothesize that there had been a change in cortical organization in the infants and that these changes supported the behavioral recovery. In particular, she hypothesized that if some synapses were removed as a consequence of brain injury, “others would be formed in less usual combinations” and that “it is possible that factors which facilitate cortical organization in the normal young are the same by which reorganization is accomplished in the imperfect cortex after injury” (Kennard, 1942, 239). Although Kennard had much to say regarding the limitations of functional recovery after early brain injury (see a review by Finger and Almli, 1988), it was her demonstration that the consequences of motor cortex lesions in infancy were less severe than similar injury in adulthood that is usually associated with her name; in fact, it is commonly referred to as the Kennard principle. Kennard was aware that early brain damage might actually produce more severe deficits than expected, but it was Hebb (1947, 1949) who emphasized this possibility. On the basis of his studies of children with frontal lobe injuries, Hebb concluded that an early injury might prevent the development of some intellectual capacities that an equally extensive injury, at maturity, would not have destroyed. Hebb believed that this outcome resulted from a failure of initial organization of the brain, thus making it difficult for the child to develop many behaviors, especially socioaffective behaviors. The difference between the views of Kennard

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

387

and Hebb is important in the current context, for it provides an important starting point for studies looking for a relationship between synaptic change and behavior. Thus, whereas Kennard hypothesized that recovery from early brain damage was associated with a reorganization into novel neural networks, Hebb postulated that the failure to recover was correlated with a failure of initial organization. Extensive studies of both cats and rats with cortical injuries have shown that both views are partially correct (Kolb, 1995; Villablanca et al., 1993). Studies of both rats and cats with cortical lesions have shown that age at injury is the variable that predicts the Kennard or Hebb outcome. The relationship between age and behavior is illustrated in the following example. We have removed the frontal cortex of different groups of rats at various ages ranging from embryonic day 18 (recall that the gestation period of a rat is about 22 days) through infancy and adolescence (Kolb, 1987; Kolb and Whishaw, 1989). The behavioral results can be illustrated by the skilled reaching behavior of rats with removal of the frontal cortex on embryonic day 18 (E18) or postnatal day 1, 5, 10, or 90 (i.e., adult). Rats are trained to reach through bars to retrieve bits of chicken feed (Whishaw et al., 1991). Rats with lesions in adulthood or on day 1 show severe deficits relative to control animals, whereas rats with lesions on day 5 or 10 show intermediate deficits, and those with E18 lesions show no deficit at all. Functional outcome thus clearly varies with precise age at injury.

On the basis of our studies, we have concluded that damage during the period of neurogenesis, which in the rat cortex is from about E12 to E20, appears to be associated with a good functional outcome (see also Hicks and D’Amato, 1961); damage in the first week of life, which is a time of neural migration and the initiation of synaptic formation, is associated with a dismal outcome; damage in the second week of life, which is a time of maximal astrocyte generation and synapse development, results in a good functional outcome; and, damage after 2 weeks leads to progressively more severe chronic behavioral loss (figure 23.2). A similar pattern of results can be seen in parallel studies of the effects of cortical lesions in kittens by Villablanca and colleagues (1993). A key point here is that birth date is irrelevant. It is the developmental stage of the brain at injury that is critical. Thus, because rats and kittens are born at an embryologically younger age than primates, including humans, the time scale for functional outcome must be adjusted to match the neural events that are under way at the time of injury (figure 23.3). Because neural generation is most intense during the second trimester in humans and is largely complete by the third trimester, the second trimester is probably most similar to the last week of gestation in the rat. Similarly, because the third trimester in humans is a time of active cell migration and the beginning of differentiation, the third trimester of humans parallels the first week of life in the infant rat. From these observations we would predict that the worst time for

Figure 23.2 Top: Main cellular events related to cortical plasticity. Bars mark the approximate beginning and ending of different processes. The shaded area illustrates the time of

maximum activity. Bottom: Summary of the time-dependent differences in cortical plasticity.

388

neural plasticity in development

HUMAN E5m

P1m

P8m

P12m

P7d

P12d

P35d

Most intense dendrite growth

Maximum synaptic density

h RAT P1d b

Conception

Neuron division ends

Migration ends

Figure 23.3 Schematic illustration of the comparable developmental ages of the brain of the rat and human. E, embryonic day; P, postnatal day; b, day of birth. Note that the day of birth in the

rat is much earlier in embryonic development than the day of birth in the human. Rhesus monkeys are born even more developed than humans.

injury in the human brain would likely be the third trimester, whereas there should be relatively good compensation for injuries during the second trimester. It is interesting in this regard that one of the most common causes of epilepsy is now thought to be abnormalities of neural migration, which would occur in the third trimester. We should note here that infants born in the second trimester might not fare well, in part because the very premature brain is being exposed to such an abnormal ex utero environment (see chapter 24 by Ewing-Cobbs, Prasad, and Hasan).

earlier the injury, the smaller the brain and the thinner the cortical mantle. Thus rats with perinatal lesions have very small brains, whereas those with lesions at day 10 have larger brains. Curiously, however, the day-10 brains still are markedly smaller than the brains of rats with lesions later in life, such as day 25, even though the behavioral outcome is far better (Kolb and Whishaw, 1981). Therefore, it must be the organization of the brain rather than its size that predicts recovery in the day-10 animal. Changes in organization can be inferred from an analysis of synaptic numbers, cortical connectivity, and evidence of neuro- and gliogenesis.

Behavioral Measurement Although there is a tendency to see all behaviors as equivalent in assessing functional outcome from early injury, this is not the case. As a rule of thumb, recovery of cognitive/executive functions is much greater than recovery of species-typical and motor behaviors. Thus animals with early prefrontal injuries in the second week show almost complete recovery of cognitive/executive functions (e.g., working memory, strategy formation), partial recovery of motor behaviors, and no recovery of social/ affective behaviors (Kolb and Whishaw, 1981; Pellis et al., 2006). This behavioral difference may be similar to the difference between the recovery of at least certain cognitive functions (such as language) in human infants who at the same time have severe chronic deficits in social/affective behavior (Kolb and Whishaw, 2003).

Brain development after early brain injury In the course of studying the functional effects of early cortical injury in rats, we noticed that brain size in adulthood was directly related to the postnatal age at injury: the

Synaptic Space By staining brains with a Golgi-type stain it is possible to measure and to quantify the dendritic structure of neurons. Although dendritic length does not provide a direct measure of synapse number, we noted earlier that dendritic length and spine density can be used as a reasonable estimate of synaptic number. Golgi analyses of cortical neurons of rats with perinatal lesions consistently show a general atrophy of dendritic arborization and a drop in spine density across the cortical mantle (Kolb, Gibb, and van der Kooy, 1994). In contrast, rats with cortical lesions around 10 days of age show an increase in dendritic arbor and an increase in spine density relative to normal control littermates. Thus animals with the best functional outcome show the greatest synaptic increase, whereas animals with the worst functional outcome have a decrease in synapses relative to control animals. Furthermore, factors that act to increase dendritic space also enhance functional outcome, whereas those that act to decrease dendritic space act to retard functional outcome.

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

389

Changes in Cortical Connectivity Perhaps the most extensive studies of changes in cortical connectivity are those showing that after unilateral motor cortex lesions in infant rats or cats there is a major expansion of the ipsilateral corticospinal pathway from the undamaged hemisphere, which is correlated with partial recovery of skilled forelimb use (Castro, 1990; Whishaw and Kolb, 1988). The initial studies concluded that these aberrant connections were advantageous and likely provided an explanation for the Kennard effect. It appears, however, that the anomalous corticospinal projections may be formed at a significant cost. For example, when we compared the effects of motor cortex lesions on postnatal days 4, 10, and 90, we found that although it was only the youngest animals that showed the enhanced ipsilateral connections, it was the day-10 animals that showed the best functional outcome. Furthermore, animals with day-4 lesions showed unexpected deficits on cognitive tasks, such as the acquisition of a spatial navigation task. It thus seems likely that the aberrant corticospinal pathway interfered with the normal functioning of cortical areas that would not ordinarily be involved in motor function. In a parallel set of studies, we showed that damage to the medial prefrontal region produced a similar result: animals with frontal lesions on day 1 showed massive changes in cortical connectivity, but these animals had the worst behavioral outcome (Kolb, Gibb, and van der Kooy, 1994). Furthermore, we showed that the abnormal pathways did not reflect the creation of new connections so much as they reflected a failure of pruning of connections that are normally discarded during development. This point was demonstrated by our finding that newborn animals have extensive aberrant pathways that die off during the first week of life. If the cortex is damaged during this time, however, some of these pathways fail to die off, leaving the animal with apparently novel circuitry that is not seen in normally developing animals. Some of this unusual circuitry could prove helpful after an injury, but given that normal animals shed such circuitry, it seems likely that it could prove equally disadvantageous to maintain such circuitry. Indeed, both hypotheses are confirmed. Rats with infant motor cortex lesions show sparing of some motor skills (Whishaw and Kolb, 1988), but apparently at the price of impairments in other cognitive functions (Kolb, Cioe, and Whishaw, 2000a, 2000b). It therefore seems likely that the presence of abnormal corticofugal pathways after early cortical injury may be as disruptive as it is helpful. This possibility has been termed “crowding” to reflect the idea that the normal functions of a cortical region can be crowded out by the development of abnormal connections (Teuber, 1975). We would be remiss to leave the reader with the impression that all anomalous wiring after early cortical injury is detrimental. We already have noted the benefits to motor behavior with the expansion of the ipsilateral corticospinal

390

neural plasticity in development

tracts. Another compelling example of advantageous change comes from the study of cats with damage to the primary visual cortical areas (Payne and Cornwell, 1994; Payne and Lomber, 2001). These studies, which are among the most extensive and sophisticated in the field, have shown a major rewiring of thalamocortical and corticocortical connections of the visual system. Although there are certainly behavioral deficits in these animals, there is an impressive visual functional sparing. The contrast between the effects of postinjury rewiring in primary motor and sensory systems versus prefrontal systems may be instructive. It appears that it may be easier to initiate beneficial changes in circuitry after injury to lower-order cortical regions than after injury to multimodal regions like the prefrontal, posterior parietal, and anterior temporal regions. This idea deserves further study. Changes in Neurogenesis In the course of studies of the effect of restricted lesions of the medial frontal cortex or olfactory bulb, we discovered that, in contrast to lesions elsewhere in the cerebrum, midline telencephalic lesions on postnatal days 7–12 led to spontaneous regeneration of the lost regions, or at least partial regeneration of the lost regions. Similar injuries either before or after this temporal window did not produce such a result. Analysis of the medial frontal region showed that the area contained newly generated neurons that formed at least some of the normal connections of this region (Kolb, Gibb, et al., 1998). Furthermore, animals with this regrown cortex appeared virtually normal on many, although not all, behavioral measures (Kolb, Petrie, and Cioe, 1996). Additional studies showed that if we blocked regeneration of the tissue with prenatal injections of the mitotic marker bromodeoxyuridine (BrdU), the lost frontal tissue failed to regrow, and there was no recovery of function (Kolb, Pedersen, et al., 1999; Kolb Pedersen, and Gibb, 2006), a result that implies that the regrown tissue was supporting recovery. Parallel studies in which we removed the regrown tissue found complementary results: removal of the tissue eliminated the functional recovery (Dallison and Kolb, 2003). Thus, in the absence of the regrown tissue, either because we blocked the growth or because we removed the tissue, function was lost. Although the spontaneous regeneration of tissue after medial prefrontal injury appears to be a unique phenomenon because it is not observed after cortical injuries elsewhere in the cortex, it is possible to stimulate neurogenesis in other regions. We recently found that exogenous administration of FGF-2 after day-10 lesions of the motor cortex leads to regeneration of the lost tissue (Monfils, Driscoll, et al., 2006). This tissue does not appear normal in lamination, but it does have connections with the spinal cord that are functional, the cells have fairly normal patterns of spontaneous discharge, and, like the spontaneous regeneration of medial frontal

cortex after day-10 lesions, the FGF-2-stimulated regeneration of motor cortex is blocked by prenatal treatment with BrdU (Monfils, Kolb, and Faoud, in press). One question that arises is whether the regeneration of lost brain tissue during infancy influences later plastic events in the brain. For example, does the generation of so many new cells during infancy compromise the brain’s ability to generate cells for the olfactory bulb or hippocampus in adulthood? To test this possibility we made frontal lesions in mice at P7, and later in adulthood we removed the stem cells from the subventricular zone and placed them in vitro with neurotrophic factors (Kolb, Gibb, et al., 1999). When stem cells are placed in such a medium, they normally divide rapidly, producing large numbers of new stem cells and progenitor cells (Fisher, 1997; Reynolds and Weiss, 1992). In contrast to the cells from control mice, which produced thousands of new cells in vitro, the cells from the brains of the animals with previous P7 lesions produced few new cells. The early brain damage followed by regeneration of the lost tissue appears to have used up the proliferative potential of the subventricular stem cells, leading to an abnormal response in adulthood. We do not yet know what implications this result has for the normal endogenous production of new neurons in adulthood, but it seems likely that it will not be normal. Once again, there appears to be a price to be paid for plastic changes in the infant brain. In sum, it appears that neurogenesis can be reinitiated after cortical lesions on postnatal days 7–12. Why this time and place is special is unclear, but these results show that regeneration of lost tissue is possible. This regeneration may not be without cost to later plastic changes in the brain, however. In conclusion, studies of laboratory rats with cortical lesions at different developmental ages have shown that there are a variety of morphological changes that follow early cortical injury. These changes include either increases or decreases in synaptic space, alterations in corticofugal connectivity, and the regeneration of cortical tissue. Functional recovery correlates with increases in cortical synaptic space and the generation of new neurons, but changes in corticofugal connections may be as disruptive as helpful in stimulating functional recovery.

Manipulation of endogenous changes We have seen that if a cerebral injury is followed by an increase in dendritic space, there is a good functional outcome, whereas if an injury leads to an atrophy of dendritic space, then there is a poor functional outcome. It follows that if we can potentiate dendritic growth in animals showing poor recovery of function, we should enhance functional recovery. The treatments for the potentiated growth range from behavioral therapy to the application of some

sort of pharmacological treatment. The pharmacological treatments could be of various forms including growth factors (e.g., nerve growth factor), hormones (e.g., sex steroids), or chemicals that influence transmitters, especially the neuromodulators such as acetylcholine and noradrenaline. We will focus here on behavioral therapy, FGF-2, neuromodulators, gonadal hormones, and psychoactive drugs. Behavioral Therapy Although it is generally assumed that behavioral therapies will improve recovery from cerebral injury in humans, there have been few direct studies of how these might work, when the optimal time for therapy might be, or even whether it is actually effective (Kwakkel et al., 1997). Furthermore, as we try to develop animal models of cognitive or motor therapies, we are left with the problem of determining what an appropriate therapy might be. There have been many studies of the effects of various types of experience on functional outcome after cerebral injury in laboratory animals, but the results have been inconsistent and generally disappointing (for reviews see Shulkin, 1989; Will and Kelche, 1992). One difficulty with these studies is that few have actually measured neuronal morphology. Rather, most studies have focused primarily on functional outcome with different environmental manipulations. Thus it may be that treatments fail to potentiate recovery because they are ineffective in stimulating brain plasticity. Our approach has been somewhat different. We have chosen behavioral manipulations that we knew were capable of changing the brain of intact animals and then exposed our brain-injured animals, especially those with poor functional outcomes, to the same experiences. We will illustrate with a few examples from our work with rats with infant lesions. We noted earlier that the developing brain is influenced by both pre- and postnatal tactile stimulation and that the young animal is influenced by both pre- and postnatal housing in enriched environments. Because the animal with a cortical lesion in the first days of life is functionally devastated in adulthood, and because it shows atrophy of cortical neurons, we anticipated that such animals would benefit the most from early intervention. In one series of studies, animals were given frontal or posterior parietal lesions at 4 days of age, followed by tactile stimulation (stroking) until weaning. Tactile stimulation for 15 minutes, three times per day, for just 10 days permanently alters the morphological and neurochemical structure of the cortex of normal animals but has even bigger effects on the cortex of an injured brain (Gibb and Kolb, 2008b; Kolb and Gibb, 2008). Rats with tactile stimulation show an unexpectedly large attenuation of the behavioral deficits of cerebral injury as a result of this rather brief “therapy.” In fact, the rats with frontal lesions on day 4 showed a nearly complete recovery of performance in various motor tasks, such as skilled reaching. This is a stunning reversal of a devastating

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

391

functional loss normally seen in animals with such injuries at this age. Analysis of the brains showed a reversal of the atrophy of the remaining cortical neurons normally associated with such early lesions. Thus a treatment that reversed the dendritic atrophy after perinatal lesions also reversed the severe functional disturbance from the early lesion. Given that we had found that prenatal tactile stimulation also altered brain development, we reasoned that this treatment might also facilitate recovery from day-4 frontal injury. It does. Like tactile stimulation postinjury, prenatal tactile stimulation reversed the cognitive and motor deficits normally associated with such injuries (Gibb and Kolb, 2008a). This is an exciting finding because it shows that prenatal experiences can influence the outcome from brain injury that occurs some time later. We will return to this idea. In parallel studies we placed rats that had cortical lesions in the first week of life in complex environments for 3 months, beginning at the time of weaning or in adulthood. The animals placed in the environments as juveniles showed a dramatic reversal of functional impairments that was correlated with increased cortical thickness (Kolb and Elliott, 1987). The dramatic improvement in the animals with the earliest injuries carries an important message, for it suggests that even the young animal with substantial neural atrophy and behavioral dysfunction is capable of considerable neuroplasticity and functional recovery in response to

behavioral therapy. When animals with similar injuries were placed into enriched environments as adults, they showed a less impressive reversal of functional losses, although they did show marked reversal of dendritic atrophy (Comeau et al., 2008; Kolb, Gibb, and Gorny, 2008). The brain thus appears to be capable of considerable environment-mediated modification after early injury, although the timing of the postinjury therapy does appear to make some difference (table 23.1). Perhaps the synaptic organization of the remaining brain can be more easily remodeled if the reorganization occurs while the brain is still developing, whereas it is more difficult to remodel a brain with infant injuries once the synaptic organization has stabilized in adulthood. Following the logic of the prenatal tactile-stimulation study discussed earlier, we placed either pregnant dams in complex environments for the duration of their pregnancies or placed adult male rats in complex environments for 3 months before they were placed with receptive females. Both types of treatments facilitated their offsprings’ recovery from frontal injury on day 4 (Gibb and Kolb, 2008b). Perhaps the real surprise was that the experience of the male parent could influence recovery from cortical injury in the offspring. The simplest explanation for the male effect is that the complex housing produced some type of change in gene expression in the sperm. The possibility that experiences of males can influence brain and behavioral development in

Table 23.1 Modification of the effects of early frontal cortical injury Treatment Result Basic Reference Postinjury tactile stimulation Functional recovery after P4 injury; Gibb and Kolb, 2008b dendritic changes Prenatal tactile stimulation Functional recovery after P4 injury; Gibb and Kolb, 2008a dendritic changes Complex housing at weaning Functional recovery after P1–5 injury; Kolb and Elliott, 1987 dendritic growth Complex housing in adulthood No functional recovery after P1–5 injury Comeau et al., 2008 Complex housing prenatally Functional recovery after P4 injury Gibb, Gonzalez, and Kolb, 2008 Postinjury FGF-2 Functional recovery after P4 injury; Comeau et al., 2007 dendritic growth Prenatal FGF-2 Functional recovery after P4 injury Comeau et al., 2008 Postinjury FGF-2 Functional recovery after P10 injury; Monfils, Driscoll, et al., 2006 neurogenesis NA depletion before day-7 lesion Blocked recovery; dendritic atrophy; drop Kolb and Sutherland, 1992 in spine density High choline diet before and after Stimulated recovery; enhanced dendritic Halliwell, Tees, and Kolb, 2008 day-4 lesion growth GDX before day-7 lesion Reduced recovery; reduced dendritic Kolb and Stewart, 1995 growth Prenatal fluoxetine Blocks recovery after P10 injury; small Day, Gibb, and Kolb, 2003 brain Abbreviations: ACh, acetylcholine; GDX, castration on day of birth; NA, noradrenaline; P1–5, postnatal days 1–5; P4, P7, P10, postnatal day 4, 7, 10.

392

neural plasticity in development

their offspring is novel and clearly the grist for considerable further study. One additional question that we asked was how experience might modify the cortex of animals with medial frontal removals on postnatal day 7. On the one hand, these animals would normally show significant functional recovery in any event, and so one prediction would be that the brain would not show much response to the experience. On the other hand, if we gave the animals experience during the time that they would normally be reorganizing synaptic organization in order to support recovery, we might anticipate a more efficient reorganization and better functional recovery. The results showed that there was a small but significant added functional benefit of such experience (Kolb, Gibb, and Gorny, 2008). The cortical neurons in the brains of these animals also showed experience-dependent modifications, although these changes were reduced relative to complexhoused control animals. We expected the anatomical results to be particularly interesting because we knew from our earlier studies that the effects of the lesion and experience would be in opposite directions. That is, rats with frontal lesions on P7 normally show an increase in spine density in response to the lesion, an anatomical adaptation that is presumably related to the functional recovery (Kolb, Stewart, and Sutherland, 1997). In contrast, normal control rats placed in complex environments at weaning normally show a drop in spine density. Thus rats with P7 lesions would be placed in a situation where one force (the lesion) would stimulate the growth of synapses while the other (complex housing) would stimulate the pruning of synapses. What actually happened was that the lesion animals showed intermediate effects: The complex housing reduced the lesioninduced spine density increase. There was, however, little change in the functional recovery. Thus the experience did change the cells, but on the behavioral measures we used there were no obvious behavioral consequences. The spine results suggest that the synaptic organization of animals with P7 lesions may be modified in different ways depending upon the experience, whether it be injury or complex housing. It appears, however, that the different modifications are still able to support functional recovery. As with the experiments showing a decrease in spine density, and thus synapses per neuron, after early experiences, the P7 complex-rearing studies leave us with the conundrum of why a reduction in the number of synapses enhances functional outcome. Again, it may be that the early experience encourages the genesis of more neurons and/or glia or that the experience attenuates the normal process of cell death during development. In either case, the animals would have more neurons and thus actually have more total synapses. Although really just conjecture at this point, such a result would account for why a decrease in spine density appears functionally advantageous.

One reasonable question we might ask is just how complex or enriched environments for rats relate to environmental stimulation for children. In particular, it seems likely that standard laboratory housing for rats is rather sterile compared to the usual environment of feral rats; so if we try to generalize to humans, we are left wondering how stimulating an environment would have to be in order to induce functional changes in human infants. One way to answer this question is to consider what animals do in the complex environments. They are provided with novel objects to interact with every few days; they get a lot of activity as they move about the compounds; they have extensive sensory stimulation including visual, auditory, tactile, and olfactory stimulation; and they have considerable social experience as they live in groups of 6–8 other animals. The experience is thus perceptually and socially stimulating, motorically demanding, and continuous for several weeks or months. Although we do not know which aspects of this experience are the most important, our guess is that all contribute because they stimulate different types of brain activity. Thus our best guess as we generalize to human infants is that therapies would include perceptual stimulation, including novelty, as well as social interaction and motor activities. We are currently manipulating olfactory and social experiences in young laboratory animals to see whether these factors might alter normal cortical development and/or recovery from cortical injury. Our preliminary results show that daily exposure to novel olfactory stimulation does indeed facilitate functional recovery from early brain injury (C. Gonazalez and B. Kolb, unpublished observations). Neurotrophic Factors We have seen that FGF-2 is upregulated in both skin and brain of animals with tactile stimulation and that FGF-2 can stimulate neurogenesis after cortical lesions on day 10. We also have examined the effects of FGF-2 administration on recovery from frontal or parietal cortical injuries on postnatal day 4, which is a time at which spontaneous recovery is normally very poor. Given that we knew that it was possible to influence recovery from cortical injuries with prenatal treatments, we gave animals FGF-2 either prenatally, postinjury, or both (Comeau, Hastings, and Kolb, 2007, 2008; Gibb and Kolb, 2008b). All three treatment regimes proved to be beneficial, with the combined pre- and postnatal treatments being the most advantageous (figure 23.4). Curiously, although the anatomical studies have not been completed, it appears that the FGF-2 treatment produces a unique set of morphological changes at each age. For example, prenatal FGF reversed the loss in brain weight in the lesion rats, whereas postnatal FGF reversed the lesion-induced decrease in cortical thickness. Combined pre- and postinjury treatment reversed the reductions in both brain weight and cortical thickness.

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

393

No Treat FGF-2

*

Sum daily total latencies (sec)

250

* 200

150

100

50

0

Control

mPFC

PPC

Figure 23.4 Summary of Morris water task performance of rats given both pre- and postnatal FGF-2. There is complete recovery on this task. (After Comeau, Hastings, and Kolb, 2006b.)

Neuromodulators Both acetylcholine and noradrenaline have been implicated in various forms of cortical plasticity (Bear and Singer, 1986), and, as we shall see, they both modulate recovery from early cortical injury. We will consider each separately. In the course of doing our experiments on tactile stimulation and recovery, we noticed that tactile stimulation during infancy produced a chronic enhancement of acetylcholinesterase density in the cortex (Kolb et al., 1999). In addition, this increase was significantly greater in the animals with frontal removals than in sham controls. This result led us to speculate that the effect of tactile stimulation might be mediated through increased acetylcholine levels. We reasoned that if this speculation is accurarate, and if we increase acetylcholine levels more directly, we might also enhance recovery. Because it had been shown by others that choline-enriched diets could alter cortical plasticity (Pyapali et al., 1998), we placed pregnant dams on a choline-enhanced diet during the last 2 weeks of pregnancy and kept the lactating mothers on the diet until weaning (Halliwell, Tees, and Kolb, 2006). The pups received medial frontal lesions on postnatal day 4 and were assessed on various behavioral tasks in adulthood. Choline treatment significantly ameliorated the expected functional deficits, and, in addition, there was a significant increase in dendritic branching and spine density in the choline-treated lesion animals. There are now many studies showing that noradrenaline (NA) is necessary for various forms of experience-dependent cortical plasticity (e.g., Kasamatsu, Pettigrew, and Ary, 1979). We therefore depleted newborn rats of forebrain NA, which can be accomplished with subdural doses of the neurotoxin 6-hydroxydopamine in the young animal. This procedure is possible because the forebrain blood-brain barrier is permeable to the toxin in the first few days of life and the noradrenergic terminals are destroyed selectively, leaving

394

neural plasticity in development

dopaminergic terminals and cells undisturbed. Animals then received large frontal lesions at 7 days of age and were later tested in various behavioral tasks before undergoing a Golgi analysis of dendritic and spine morphology (Sutherland et al., 1982; Kolb and Sutherland, 1992; Kolb, Stewart, and Sutherland, 1997). The results showed that in the absence of NA there was no functional recovery. Analysis of the dendritic and spine changes showed that the NA depletion produced a drop in dendritic morphology in both sham control and lesion animals relative to saline-treated controls. The drop in dendritic length was greater in the lesion animals, as might be expected given that they essentially had two injuries—NA depletion and direct cortical injury. The sham NA-depleted animals, which had no obvious behavioral disturbance, apparently compensated for the dendritic loss with an increase in spine density. A similar increase was observed in the lesion animals as well, but these animals apparently had too great a synaptic deficit to reverse and thus failed to show recovery. In sum, we have seen that increased cholinergic activity facilitates recovery whereas decreased cholinergic or noradrenergic activity inhibits recovery. We can predict that increasing noradrenergic activity should stimulate recovery. It would be interesting to determine how manipulations of neuromodulators might interact with behavioral treatments such as tactile stimulation or complex rearing. Gonadal Hormones There is accumulating evidence that male and female brains differ in their structure, respond differently to environmental events, and respond differently to injury. For example, we have found that there are hormone-related structural differences in the prefrontal cortex of rats. Cells in the medial prefrontal regions have more dendritic arborizations in animals exposed to testosterone, whereas cells in the insular prefrontal regions have more dendritic arborizations in animals not exposed to testosterone (Kolb and Stewart, 1991). These differences led us to wonder if early brain injury might influence the pattern of sex-related differences in cortical morphology. It does. As we have seen, rats with medial frontal lesions in postnatal day 7 show marked functional recovery, but, in addition, this recovery is sexually dimorphic. In particular, males show enhanced recovery of spatial-navigation behaviors, whereas females show enhanced recovery of skilled forelimb reaching (Kolb and Cioe, 1996; Kolb, Gibb, and Gorny, 2008; Kolb and Stewart, 1995). These behavioral differences are associated with differences in cortical plasticity: Males with frontal lesions show a greater increase in spine density, whereas females show a larger increase in dendritic arborization (Kolb and Stewart, 1995). Thus, although both sexes show functional recovery, there are differences, and these differences are related to morphological differences. To complicate matters more, there appear to

be sex differences in the effect of tactile experience or complex housing on cortical neurons (Kolb, Gibb, and Gorny, 2008), and so the question arises as to whether sex, experience, and lesion might all interact. This is an important question if we are to design treatments for children with brain injuries. Psychoactive Drugs Research over the past decade has found that repeated exposure to a wide range of psychoactive drugs, including amphetamine, cocaine, nicotine, morphine, and marijuana, produces chronic changes in the structure of neurons in the striatum and prefrontal cortex of both adult and developing animals (Robinson and Kolb, 2004). Given that exposure to such drugs, as well as prescription drugs like selective serotonin reuptake inhibitors (SSRIs), anxiolytics, and antipsychotics, is common during pregnancy in humans, we have begun to investigate the effects of prenatal exposure to such drugs on brain development and recovery from cortical injury. The most dramatic finding to date is that prenatal exposure to fluoxetine (Prozac) has devastating consequences not only on the development of the normal brain but even larger effects on the perinatally injured brain (Day, Gibb, and Kolb, 2003). When pregnant dams are given a dose that is roughly equivalent to a standard dose in humans, the brains of the offspring are abnormally small, and if the animals have later brain injury at day 10, the preexposure to fluoxetine completely blocks recovery from day-10 prefrontal injury, including a complete block of the expected spontaneous neurogenesis. It is clear that the drug is acting as a teratogen to the developing brain. This is a disturbing, and unexpected, finding and is leading us to evaluate a wider range of prenatal drug exposures.

Summary One of the most intriguing questions in behavioral neuroscience concerns the manner in which the brain, and especially the neocortex, can modify its structure and ultimately its function throughout one’s lifetime. As the preceding review has suggested, the cortex can be changed dramatically by events during early development, especially early brain damage. Several basic conclusions can be extracted regarding the nature of the relationship between experience, brain plasticity, and behavior. 1. The effects of experience vary both qualitatively and quantitatively at different times in development. For example, “enriched” experiences during the immediate postnatal period have no measurable effect on dendritic length and lead to a chronic decrease in spine density in cortical neurons. Enriched experiences during the juvenile and adolescent period produce chronic increases in dendritic length but decreases in spine density. Similar experiences later in life produce chronic increases in both dendritic length and spine

density in cortical pyramidal cells. Finally, prenatal experience (by way of maternal experience) also alters later brain development and leads to a chronic decrease in dendritic length but an increase in spine density. All these experiences have behavioral sequelae in adulthood. The age-dependent plastic changes in the cortex presumably reflect the differential sensitivity of the child’s brain to experience during development. It is not yet known why there are such differences in response to experience. 2. There are three ways that the brain changes after injury during development: (a) endogenous changes in the organization of the remaining, intact, circuits in the brain; (b) the generation of new circuitry, including circuitry that is novel to the injured brain; and (c) the generation of neurons and glia to replace at least some lost neurons. Details of these changes vary with the precise nature of the developmental events under way at the time at injury. 3. When the cortex is injured during the time of neurogenesis, there is an apparent capacity to replace the lost cells, although cortical organization is not entirely normal. There is an associated sparing of cortical function. 4. Cortical injury in the period immediately following the completion of neurogenesis results in the development of a small brain and thin cortex, and significant reorganization of cortical connectivity. These latter changes are reflected especially in abnormalities in corticofugal connections. In addition, there is an atrophy of dendritic arborization and reduced spine density in cortical neurons. The behavioral outcome is especially poor after injury at this time, as animals show severe and chronic behavioral impairments. 5. Injury to the cortex while astrocytes are maximally developing and cortical neurons are maximally differentiating also produces a somewhat small brain and thin cortex, but there is a compensatory increase in dendritic arborization and an increase in spine density. These changes presumably reflect changes in the intrinsic organization of cortical circuits. 6. Injury-induced changes in cortical structure are modified by both pre- and postnatal experience. Furthermore, brains that show the least change in response to a cortical injury appear to be the most responsive to experience. For instance, animals with neonatal brain injuries show a poor functional outcome and an atrophy of cortical neurons, but these animals show dramatic functional recovery and marked synaptic growth in response to environmental manipulations. Behavioral manipulations that are initiated in the immediate postinjury period appear to be the most effective in stimulating functional recovery. The response of the developing brain to postinjury manipulations of experience is encouraging, for it suggests that behavioral therapies should be especially helpful in reversing some of the devastating consequences of brain damage in the latter periods of prenatal development in human infants.

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

395

7. Neurotrophic factors, especially FGF-2, influence recovery from early brain injury. The FGF levels can be manipulated indirectly by treatments such as tactile stimulation or directly by exogenous administration. The FGF-2 can stimulate neurogenesis after injury at day 10 but not after injury at day 3. In contrast, after day-3 lesions the FGF treatment leads to changes in neuronal morphology that correlate with functional recovery. 8. Neuromodulators, including both acetylcholine and noradrenaline, influence recovery from early brain injury. Treatments that act to increase the cortical levels of neuromodulators stimulate cortical plasticity and functional recovery, whereas treatments that act to reduce the cortical levels of neuromodulators act to retard recovery. 9. Gonadal hormones influence synaptic organization of the cortex throughout the lifetime of an animal. This influence is particularly strong during development, and the gonadal hormone effects interact with the plastic changes that follow cortical injury, at least at certain stages of development. Perhaps the biggest sex-related difference is that males show larger changes in spine density after cortical injury whereas females are more likely to show changes in dendritic arborization. 10. Prenatal experiences can influence recovery from brain injury that does not occur until infancy. Maternal and paternal experience in complex environments, tactile stimulation of pregnant dams, and FGF-2 all facilitate recovery from later cortical injury. Prenatal exposure to fluoxetine completely blocks recovery from day-10 cortical injuries. 11. There are limits to the amount that a brain can change. We still do not know what determines the limits, nor, in most cases, what the limits might be. Nonetheless, it does appear that when the brain changes “spontaneously” after injury, there is a reduction in the plasticity of the affected regions. This implies that behavioral therapies should be initiated early in the postinjury period to ensure that the “spontaneous” changes can be influenced in such a way as to maximize functional recovery. acknowledgments

This research was supported by Natural Science and Engineering Research Council of Canada and Canadian Institute for Health Research grants to BK and RG. The authors wish to thank the late Grazyna Gorny for her technical help in many of the experiments described here. REFERENCES

Bailey, C. H., and E. R. Kandel, 1993. Structural changes accompanying memory storage. Annu. Rev. Physiol. 55:397–426. Bear, M. F., and W. Singer, 1986. Modulation of visual cortical plasticity by acetylcholine and noradrenaline. Nature 320:172– 176.

396

neural plasticity in development

Becq, H., I. Jorquera, Y. Ben-Ari, S. Weiss, and A. Represa, 2005. Differential properties of dentate gyrus and CA1 neural precursors. J. Neurobiol. 62:243–261. Castro, A., 1990. Plasticity in the motor system. In B. Kolb and R. Tees, eds, Cerebral Cortex of the Rat, 563–588. Cambridge, MA: MIT Press. Comeau, W., R. Gibb, E. Hastings, J. Cioe, and B. Kolb, in press. Therapeutic effects of complex rearing or bFGF after perinatal frontal lesions. Dev. Psychobiol. Comeau, W., E. Hastings, and B. Kolb, 2007. Differential effect of pre- and postnatal FGF-2 following medial prefrontal cortical injury. Behav. Brain Res. In press. Comeau, W. L., E. Hastings, and B. Kolb, 2008. Pre- and postinjury FGF-2 act synergistically in enhancing functional outcome in rats with perinatal injury. Manuscript in submission. Dallison, A., and B. Kolb, 2003. Recovery from infant frontal cortical lesions in rats can be reversed by cortical lesions in adulthood. Behav. Brain Res. 146(1–2):57–63. Day, M., R. Gibb, and B. Kolb, 2003. Prenatal fluoxetine impairs functional recovery and neuroplasticity after perinatal frontal cortex lesions in rats. Soc. Neurosci. Abstracts 29.459.10. Eriksson, P. S., E. Perfilieva, T. Bjork-Eriksson, A. M. Alborn, C. Nordborg, D. A. Peterson, and F. H. Gage, 1998. Neurogenesis in the adult human hippocampus. Nature Med. 4: 1313–1317. Finger, S., and C. R. Almli, 1988. Margaret Kennard and her “principle” in historical perspective. In S. Finger, T. E. Le Vere, C. R. Almli, and D. G. Stein, eds., Brain Injury and Recovery: Theoretical and Controversial Issues, 117–132. New York: Plenum. Fisher, L. J., 1997. Neural precursor cells: Applications for the study and repair of the central nervous system. Neurobiol. Dis. 4:1–22. Gibb, R., C. L. R. Gonzalez, and B. Kolb, 2008. Complex environmental experience during pregnancy facilitates recovery from perinatal cortical injury in the offspring. Manuscript in submission. Gibb, R., and B. Kolb, 2008a. Prenatal tactile stimulation alters brain and behavioral development and facilitates recovery from perinatal brain injury. Manuscript in submission. Gibb, R., and B. Kolb, 2008b. Tactile-stimulation that facilitates functional recovery after perinatal cortical injury is mediated by FGF-2. Manuscript in submission. Gould, E., P. Tanapat, N. B. Hastings, and T. J. Shors, 1999. Neurogenesis in adulthood: A possible role in learning. Trends Cogn. Sci. 3:186–191. Greenough, W. T., and F. F. Chang, 1988. Plasticity of synapse structure and pattern in the cerebral cortex. In A. Peters and E. G. Jones, eds, Cerebral Cortex, vol. 7, pp. 391–440. New York: Plenum Press. Gregg, C. T., T. Shingo, and S. Weiss, 2001. Neural stem cells of the mammalian forebrain. Symp. Soc. Exp. Biol. 53:1–19. Halliwell, C., R. Tees, and B. Kolb, 2008. Prenatal choline treatment enhances recovery from perinatal frontal injury in rats. Manuscript in submission. Hebb, D. O., 1947. The effects of early experience on problem solving at maturity. Am. Psychol. 2:737–745. Hebb, D. O., 1949. The Organization of Behaviour. New York: McGraw-Hill. Hicks, S., and C. J. D’Amato, 1961. How to design and build abnormal brains using radiation during development. In W. S. Fields and M. M. Desmond, eds, Disorders of the Developing Nervous System, 60–79. Springfield, IL: Thomas.

Kasamatsu, T., J. D. Pettigrew, and M. Ary, 1979. Restoration of visual cortical plasticity by local microperfusion of norepinephrine. J. Comp. Neurol. 185:163–182. Kempermann, G., and F. H. Gage, 1999. New nerve cells for the adult brain. Sci. Am. 280:48–53. Kennard, M. 1942. Cortical reorganization of motor function. Arch. Neurol. 48:227–240. Kolb, B., 1987. Recovery from early cortical damage in rats. I. Differential behavioral and anatomical effects of frontal lesions at different ages of neural maturation. Behav. Brain Res. 25: 205–220. Kolb, B., 1995. Brain Plasticity and Behavior. Mahwah, NJ: Erlbaum. Kolb, B., and J. Cioe, 1996. Sex-related differences in cortical function after medial frontal lesions in rats. Behav. Neurosci. 110:1271–1281. Kolb, B., J. Cioe, and I. Q. Whishaw, 2000a. Is there an optimal age for recovery from unilateral motor cortex lesions? Behavioural and anatomical sequelae of unilateral motor cortex lesions in rats on postnatal days 1, 10, and in adulthood. Restor. Neurol. Neurosci. 17:61–70. Kolb, B., J. Cioe, and I. Q. Whishaw, 2000b. Is there an optimal age for recovery from motor cortex lesions? Behavioural and anatomical sequelae of bilateral motor cortex lesions in rats on postnatal days 1, 10, and in adulthood. Brain Res. 882: 62–74. Kolb, B., and W. Elliott, 1987. Recovery from early cortical damage in rats. II. Effects of experience on anatomy and behavior following frontal lesions at 1 or 5 days of age. Behav. Brain Res. 26:47–56. Kolb, B., and R. Gibb, 2008. Tactile stimulation after cortical injury in infant rats stimulates functional recovery and altered cortical morphology. Manuscript in submission. Kolb, B., R. Gibb, and G. Gorny, 2003. Experience-dependent changes in dendritic arbor and spine density in neocortex vary with age and sex. Neurobiol. Learn. Mem. 79:1–10. Kolb, B., R. Gibb, and G. Gorny, 2008. Therapeutic effects of enriched rearing after frontal lesions in infancy vary with age and sex. Manuscript in submission. Kolb, B., R. Gibb, G. Gorny, and I. Q. Whishaw, 1998. Possible brain regrowth after cortical lesions in rats. Behav. Brain Res. 91:127–141. Kolb, B., R. Gibb, D. J. Martens, B. Coles, and D. van der Kooy, 1999. Proliferation of neural stem cells in vitro and in vivo is reduced by infant frontal cortex lesions or prenatal BrdU. Soc. Neurosci. Abstracts, 25. Kolb, B., R. Gibb, and D. van der Kooy, 1994. Neonatal frontal cortical lesions in rats alter cortical structure and connectivity. Brain Res. 645:85–97. Kolb, B., B. Pedersen, M. Ballerman, R. Gibb, and I. Q. Whishaw, 1999. Embryonic exposure to BrdU produces chronic changes in brain and behavior in rats. J. Neurosci. 19: 2337–2346. Kolb, B., B. Pedersen, and R. Gibb, 2008. Recovery from frontal cortex lesions in infancy is blocked by embryonic pretreatment with bromodeoxyuridine. Manuscript in submission. Kolb, B., B. Petrie, and J. Cioe, 1996. Recovery from early cortical damage in rats. VII. Comparison of the behavioural and anatomical effects of medial prefrontal lesions at different ages of neural maturation. Behav. Brain Res. 79:1–13. Kolb, B., and J. Stewart, 1991. Sex-related differences in dendritic branching of cells in the prefrontal cortex of rats. J. Neuroendocrinol. 3:95–99.

Kolb, B., and J. Stewart, 1995. Changes in neonatal gonadal hormonal environment prevent behavioral sparing and alter cortical morphogenesis after early frontal cortex lesions in male and female rats. Behav. Neurosci. 109:285–294. Kolb, B., J. Stewart, and R. J. Sutherland, 1997. Recovery of function is associated with increased spine density in cortical pyramidal cells after frontal lesions and/or noradrenaline depletion in neonatal rats. Behav. Brain Res. 89:61–70. Kolb, B., and R. J. Sutherland, 1992. Noradrenaline depletion blocks behavioral sparing and alters cortical morphogenesis after neonatal frontal cortex damage in rats. J. Neurosci. 12:2221–2330. Kolb, B., and I. Q. Whishaw, 1981. Neonatal frontal lesions in the rat: Sparing of learned but not species-typical behavior in the presence of reduced brain weight and cortical thickness. J. Comp. Physiol. Psychol. 95:863–879. Kolb, B., and I. Q. Whishaw, 1989. Plasticity in the neocortex: Mechanisms underlying recovery from early brain damage. Prog. Neurobiol. 32:235–276. Kolb, B., and I. Q. Whishaw, 1998. Brain plasticity and behavior. Annu. Rev. Psychol. 49:43–64. Kolb, B., and I. Q. Whishaw, 2003. Fundamentals of Human Neuropyschology, 5th ed. New York: Worth. Kwakkel, G., R. C. Wagennar, T. W. Koelman, G. J. Lankhorst, and J. Koetsier, 1997. Effects of intensity of rehabilitation after stroke: A research synthesis. Stroke 28:1550–1556. Monfils, M.-H., I. Driscoll, H. Kamitakahara, B. Wilson, C. Flynn, G. C. Teskey, J. A. Kleim, and B. Kolb, 2006. FGF-2induced cell proliferation stimulates anatomical, neurophysiological, and functional recovery from neonatal motor cortex injury. Eur. J. Neurosci. 24:739–749. Monfils, M.-H., B. Kolb, and K. Faoud, in press. FGF-2 promotes restoration of functional corticospinal neurons after neonatal motor cortex injury. Exp. Brain Res. Payne, B. R., and P. Cornwell, 1994. System-wide repercussions of damage of immature visual cortex. Trends Neurosci. 17: 126–130. Payne, B. R., and S. Lomber, 2001. Reconstructing functional systems after lesions of the cerebral cortex. Nature Rev. Neurosci. 2:911–919. Pellis, S. M., E. Hastings, T. Takeshi, H. Kamitakahara, J. Komorowska, M. L. Forgie, and B. Kolb, 2006. The effects of orbital frontal cortex damage on the modulation of defensive responses by rats in playful and non-playful social contexts. Behav. Neurosci. 120:72–84. Pyapali, G. K., D. A. Turner, C. L. Williams, W. H. Meck, and H. S. Swartzwelder, 1998. Prenatal dietary choline supplementation decreases the threshold for induction of long-term potentiation in young adult rats. J. Neurophysiol. 79:1790–1796. Reynolds, B., and S. Weiss, 1992. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255:1727–1710. Robinson, T. E., and B. Kolb, 2004. Structural plasticity associated with drugs of abuse. Neuropharmacology 47(Suppl 1):33–46. Schanberg, S. M., and T. M. Field, 1987. Sensory deprivation stress and supplemental stimulation in the rat pup and preterm human neonate. Child Dev. 58:1431–1447. Shulkin, J., ed., 1989. Preoperative Events: Their Effects on Behavior Following Brain Damage. Hillsdale, NJ: Erlbaum. Sutherland, R. J., B. Kolb, J. B. Becker, and I. Q. Whishaw, 1982. Cortical noradrenaline depletion eliminates sparing of spatial learning after neonatal frontal cortex damage in the rat. Neurosci. Lett. 32:125–130.

kolb, comeau, and gibb: early brain injury, plasticity, and behavior

397

Teuber, H.-L., 1975. Recovery of function after brain injury in man. In Outcome of Severe Damage to the Nervous System, Ciba Foundation Symposium 34. Amsterdam: Elsevier North-Holland. Villablanca, J. R., D. A. Hovda, G. F. Jackson, and C. Infante, 1993. Neurological and behavioral effects of a unilateral frontal cortical lesion in fetal kittens. II. Visual system tests, and proposing a “critical period” for lesion effects. Behav. Brain Res. 57: 79–92. Wallhausser, E., and H. Scheich, 1987. Auditory imprinting leads to differential 2-deoxyglucose uptake and dendritic spine loss in the chick rostral forebrain. Dev. Brain Res. 31:29–44. Whishaw, I. Q., and B. Kolb, 1988. Sparing of skilled forelimb reaching and corticospinal projections after neonatal motor

398

neural plasticity in development

cortex removal or hemidecortication in the rat: Support for the Kennard doctrine. Brain Res. 451:97–114. Whishaw, I. Q., S. M. Pellis, B. P. Gorny, and V. C. Pellis, 1991. The impairments in reaching and the movements of compensation in rats with motor cortex lesions: An endpoint, videorecording, and movement notation analysis. Behav. Brain Res. 42:77–91. Will, B., and C. Kelche, 1992. Environmental approaches to recovery of function from brain damage: A review of animal studies (1981 to 1991). In F. D. Rose and D. A. Johnson, eds., Recovery from Brain Damage: Reflections and Directions, 79–104. New York: Plenum.

24

Developmental Plasticity and Reorganization of Function Following Early Diffuse Brain Injury LINDA EWING-COBBS, MARY R. PRASAD, AND KHADER M. HASAN

Developmental outcomes following early brain injury are affected by a host of interrelated biological and environmental factors. At the present time, it is unclear which neural mechanisms support intrinsic reorganization and contribute to favorable outcomes and which factors potentiate the deleterious effects of early brain injury. The central theme of this chapter is that the consequences of early brain injury may reflect the type and severity of injury, the stage of brain development at the time of injury, age at the time of assessment, and experiential and socioenvironmental factors. Brain structures that are in a stage of rapid development may be particularly vulnerable to varied biological and environmental insults and risks, resulting in preferential disruption of specific skills dependent upon the maturity of components of developing networks. Recent concepts of plasticity emphasize that the normal, ongoing state of the nervous system throughout the life span reflects continuous changes in response to alterations in either afferent inputs or output targets (Pascual-Leone et al., 2005). Following brain injury, a time-linked sequence of mechanisms is activated. Initial changes serve to minimize damage. Rapid behavioral changes may occur as neural components that are not damaged but are rendered dysfunctional by local or systemic processes resolve and as some partially damaged neural elements are repaired. Subsequent processes involve relearning rather than recovery and involve a two-stage process of initial unmasking and strengthening of existing pathways and the development of new structural changes (Pascual-Leone et al., 2005). Although there are differing opinions regarding the degree and timing of potential developmental effects on plasticity, a central issue is identification of mechanisms that account for patterns of vulnerability and sparing of specific abilities in relation to CNS development across the life span (Rosenzweig, 2003). Developmental outcomes following early brain injury are complexly determined and reflect interactions between genetic and premorbid characteristics impacting levels of

cognitive reserve, nature and timing of the brain insult, and environmental conditions. Greenough, Black, and Wallace (2002) described two types of neural changes in sensory systems reflecting interactions of the brain and environment that may be differentially related to changes in synaptic exuberance and pruning. Experience-expectant change results from exposure to critical types of inputs and experiences that are necessary for normal developmental adaptation, such as basic visual, auditory, and sensorimotor experiences, as well as from exposure to varied environmental conditions ranging from stimulating to deprived. Specific inputs are required during sensitive periods of development; environments or experiences that occur after the sensitive period are not likely to alter brain structure or function. Experience-expectant processes may sculpt behavior during periods of synaptic exuberance. Examples include the alterations in neuronal receptive fields in individuals with congenital sensory impairment (Neville and Bavelier, 2002) and reduced integrity of the uncinate fasciculus in children experiencing early socioemotional deprivation (Eluvathingal et al., 2006). Experience-dependent change is viewed as the neurobiological basis of learning and results in more localized changes resulting from the individual’s acquisition of specific information, and it may correspond to periods of synaptic pruning. Experience-dependent processes allow for dynamic changes in neural and behavioral development based on new experiences that may modify early learning. Experience-dependent processes are illustrated by the regional changes in white matter organization related to experience, such as changes in callosal, projection, and association pathways related to extensive piano practicing at different life stages (Bengtsson et al., 2005). Animal studies examining recovery from focal and diffuse lesions suggest complex, nonlinear relations of age, environmental conditions, and outcome. Based on studies of outcome from focal lesions in rats, Kolb and colleagues inferred that developmental age at the time of brain injury may predict

399

functional outcomes in children. Focal injury sustained during late stages of neurogenesis may be fully compensated. Injury during periods of cell migration may be associated with minimal plasticity. Perinatal injuries occurring during periods of dendritic growth and massive synaptogenesis may yield partial functional recovery, while damage in adolescence and adulthood during periods of synaptic pruning may yield chronic deficits (Kolb, Gibb, and Gorny, 2000; Luciana, 2003). The relation of timing of environmental changes on recovery from brain insult may also be nonlinear. For example, institutionalized children adopted prior to 6 months of age had substantially higher IQ scores than children adopted after 2 years of age when assessed at age 6 but not at age 11. Although there was some attenuation of IQ impairment between ages 6 and 11, the majority of increase in IQ scores occurred within 2 years of adoption, suggesting time-sensitive processes of socioenvironmental impact (Beckett et al., 2006). Hebb (1942) inferred that injury to the developing brain might have different and potentially more significant cognitive sequelae than similar injuries to the mature brain. Rapidly developing neural networks may be particularly susceptible to disruption by injury, resulting in different consequences of injuries sustained at different developmental stages (Ewing-Cobbs et al., 1987). Because brain injury often results in damage to neural networks supporting learning and memory, injury to the developing brain might interfere with the child’s ability to acquire, maintain, and develop new skills. Given the limited behavioral repertoires of children born preterm or children who sustain insults during infancy or preschool years, damage to learning mechanisms would theoretically produce slower rates of skill acquisition and reduced levels of development of new cognitive skills (Taylor and Alden, 1997). Acquisition and maintenance of skills should be examined longitudinally relative to constructs permitting dissociation of the sequence and rate of acquisition, eventual level of performance, and retention (Dennis et al., 2006). Understanding how the developing brain accommodates changes due to focal or diffuse injury requires examination of change in the expected developmental trajectory of skills or behaviors over time in relation to neuroimaging studies and socioenvironmental factors (Ewing-Cobbs, Barnes, and Fletcher, 2003; Stiles et al., 2005). To date, there are few prospective, longitudinal studies that integrate cognitive and neuroimaging findings in a developmental model.

Type and timing of brain injury: Relations with specific neuropsychological outcomes The most common causes of acquired brain injury and disability in children result in some cases of preterm birth occurring at less than 37 weeks postconceptional age and

400

neural plasticity in development

from traumatic brain injury (TBI). Each injury has a unique etiology and pathophysiology. Although each injury may have focal components, both are associated with an accumulation of diffuse secondary insults, including hypoxicischemic injury, perfusion-reperfusion injury, apoptosis, inflammatory processes, and cascades of excitatory neurotransmitters. First, we will examine recent neuroimaging studies that provide insight into mechanisms of injury and reorganization of function following preterm birth and TBI with special emphasis on studies of the corpus callosum and thalamocorticostriatal pathways. Second, we will review studies that highlight the impact of age at the time of CNS insult, severity of neurobiological risk, and age at assessment on specific neuropsychological outcomes. Given our interest in developmental trajectories of skill development following injuries sustained at varying developmental stages, we emphasize longitudinal studies of IQ as well as studies of executive processes and attention deficits that are functionally related to these neural networks. Finally, we examine the impact of socioenvironmental factors that may potentiate adverse outcomes or enhance positive outcomes in children experiencing neurobiological risk factors.

Neurodevelopmental changes associated with preterm birth Developing cells in the immature nervous system are selectively vulnerable; with preterm birth, there may be loss of oligodendrocytes, the progenitors of myelin, and disruption of projections of transient subplate neurons located immediately underneath the developing cortex that are essential in formation of thalamocortical connections (McQuillen and Ferriero, 2005). Subplate development peaks at the time of greatest risk for periventricular white matter injury due to preterm birth. At 24 weeks gestation, hemorrhage of the germinal matrix may affect the migrating glial precursor cells and adversely impact subsequent brain development through periventricular white matter injury (Volpe, 2005). Preterm birth during the late second and the third trimester coincides with extensive neural development occurring from 29 to 41 weeks postconceptional age. During this interval, brain volume increases 2.7 times, cortical gray matter increases 4 times, and myelinated white matter increases 5 times (Gupta et al., 2005; Huppi et al., 1998). Prenatal and perinatal insults occurring at different developmental stages may disproportionately influence developing structures and connectivity. Developing thalamocortical connections are vulnerable to the timing of disruption; early insult results in failure of thalamocortical innervation, while later cell death disrupts the refinement of connectivity into mature circuits (McQuillen and Ferriero, 2005). Hypoxicischemic insult may preferentially affect the dorsolateral caudate nucleus, parts of the hippocampus, and the reticular nucleus of the thalamus (Nosarti et al., 2005). In preterm

infants, global insult is associated with damage to the subcortical periventricular white matter; significant anatomical and functional pathology is present in thalamocorticostriatal networks (Gimenez et al., 2006; Nosarti et al., 2006; Woodward et al., 2005). In term infants, hypoxic-ischemic encephalopathy preferentially disrupts neurons in the deep gray nuclei and perirolandic cortex/perisagittal watershed regions (Volpe, 2005). The resulting damage to hippocampal regions is associated with significant deficits in episodic or everyday memory in conjunction with relative sparing of semantic memory (Isaacs et al., 2003; Vargha-Khadem et al., 2003). Neuroimaging Studies of Children Born Preterm Recent advances in central nervous system neuroimaging techniques highlight the significant alterations in both gray and white matter macrostructure and microstructure in children born preterm and scanned near term dates. Preterm infants born at less than 27 weeks had a greater reduction of deep nuclear gray matter than those born after 27 weeks gestation (Inder et al., 2005). The most significant regional volume loss occurs within thalamic and lentiform nuclei, particularly in infants with damage to overlying white matter and in those with decreasing postconceptional age (Counsell and Boardman, 2005). Preterm birth disrupts development of the corpus callosum, which grows in a linear trajectory from 20 to 40 weeks gestation. The posterior region develops rapidly during the third trimester and may be particularly vulnerable to disruption. In very-low-birth-weight infants, the growth trajectory assessed at term equivalence is nearly half the expected rate of development in utero (N. Anderson et al., 2005). Gyrification continues during the first postnatal year, with cortical folding occurring last in the temporal and frontal lobes. The temporal lobes, in which synaptogenesis begins during the third trimester and peaks at 3 to 4 months postterm (Huttenlocher and Dabholkar, 1997b), are particularly vulnerable to increased gyrification (Kesler et al., 2006). In children born preterm, reduced global and regional gray and white matter volumes are present at birth, persist into adolescence, and are related to a variety of neurobehavioral outcomes. Reduced volumes in parieto-occipital, sensorimotor, and inferior occipital cortices obtained from neonatal MRIs predict less favorable cognitive and motor outcomes at 18 to 20 months corrected age (B. Peterson et al., 2003). Bilateral reductions are noted in basal ganglia nuclei, amygdalae, and hippocampi (B. Peterson et al., 2000). The pulvinar, lateral, and medial geniculate nuclei, and ventral posterior lateral thalamic nuclei are smaller in preterm children and are associated with poorer semantic and phonetic fluency (Gimenez et al., 2006). Although caudate volumes in children and adolescents were not statistically distinguishable in some studies of term and preterm children (Nosarti et al., 2005), other studies identified reduced caudate and hippocampal volumes that were

related to lower scores on IQ (Abernethy, Cooke, and Foulder-Hughes, 2004) and on everyday memory tasks (Isaacs, Lucas, and Chong, 2000). Reduced left caudate volume was associated with increased hyperactivity and social adjustment difficulties for preterm boys (Nosarti et al., 2005). Even in samples of preterm youth with good performance on response inhibition and memory tasks, functional MRI (fMRI) studies documented different patterns of activation in thalamostriatocortico networks relative to term comparison groups (W. G. Curtis et al., 2006). Working memory performance was related to reduced volumes in dorsolateral prefrontal cortex as well as in sensorimotor, parietooccipital, and premotor regions (Woodward et al., 2005). Arrested growth of the corpus callosum at term equivalency was related to poorer general cognitive and motor outcomes assessed at 2 years (N. Anderson et al., 2005) and with reduced Verbal IQ and fluency during adolescence (Nosarti et al., 2004). At age 8, IQ scores were significantly correlated with concurrent volumes of the corpus callosum and the premotor, subgenual, sensorimotor, midtemporal, and parieto-occipital regions (B. Peterson et al., 2000). Diffusion tensor MRI (DTI) studies have revealed significant alterations in brain microstructure in children born preterm. These studies measure the self-diffusion of water molecules within tissues, which is a function of the degree to which directionally organized tissues are developing or losing their normal integrity. As tissue matures, anisotropy increases and diffusivity decreases (see Klingberg, chapter 14 in this volume). Serial DTI studies completed near birth and near term showed increasing fractional anisotropy in all regions of interest in infants with minimal white matter injury. In infants with moderate white matter injury, anisotropy either did not increase or decreased in frontal white matter and visual association areas (Miller et al., 2002). Relative to termequivalent infants with normal-appearing white matter and term infants, those with diffuse excessive high-signal intensities on conventional MRI showed reduced diffusivities in the posterior limb of the internal capsule, posterior corpus callosum, and centrum semiovale, as well as in white matter in frontal, periventricular, and occipital regions (Counsell et al., 2006). The diffusion abnormalities noted in varied white matter structures in these infants may represent widespread abnormalities in myelination and axonal pathology consistent with delayed or deficient wrapping of oligodendrocytes around axons, which may produce increased membrane permeability and decreased axonal diameter (Counsell et al., 2006). In summary, preterm birth is often associated with major alterations in the volume and integrity of cortical and subcortical tissue. It is possible that preterm birth occurring during the second trimester is associated with disproportionate disruption of subcortical gray matter nuclei and possibly the cortex itself and that birth during the third trimester

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

401

preferentially disrupts development and myelination of white matter pathways. Even in relatively healthy preterm infants without additional biomedical risks or neurosensory impairments, preterm birth has been associated with altered brain development (Ferriero, 2004). To examine brain-behavior relations in children born preterm, future imaging studies should examine the relations between preterm birth per se and specific complications of preterm birth in relation to the developmental trajectory of nuclei and pathways implicated in memory, attention, and related executive processes. Long-Term Neuropsychological Outcome of Children Born Preterm Variations in birth weight and gestational age are related to multiple outcomes through their influence on processes regulating brain development and connectivity. Although long-term longitudinal follow-up studies of preterm infants are sparse within the existing literature, several groups have reported on outcomes that extend into adolescence and young adulthood. Most prospective longterm outcome studies following infants into adolescence note a gradient of sequelae, with less favorable outcomes associated with lower birth weight, with lower postconceptional age, and with complications including intraventricular hemorrhage, periventricular leukomalacia, and respiratory difficulties (Hack et al., 1994; see review by Luciana, 2003). The cumulative consequences of brain injury associated with preterm birth are reflected in reduced head circumference and concomitant reduction in multiple cognitive and behavioral domains (J. Peterson et al., 2006). In the elegant series of studies by the Ohio group, the developmental trajectories of multiple outcomes were examined in a predominantly inner-city cohort of very-low-birthweight children (<750 g, 750–1,499 g) and term children from mixed socioeconomic backgrounds at intervals from early school age to young adulthood. Difficulties in cognition, behavior, and executive functions identified at ages 5 to 9 persisted at ages 7 to 14 (Hack et al., 1994; Taylor, Minich, Bangert, et al., 2004). After controlling for IQ, the lowest-weight group had selective cognitive sequelae; they scored lower on measures of language-processing, listlearning, perceptual-motor, and spatial-organizational skills, and also had a slower rate of acquisition of perceptual-motor and working-memory/set-shifting skills than the full-term children. Areas of sparing included naming, delayed recall of word lists, and an index of set shifting and mental flexibility. The rate of development across time in the <750 g group was similar to the higher-birth-weight and term groups in other areas (Taylor, Minich, Klein, et al., 2004). The selective sequelae are consistent with involvement of the periventricular region, including the basal ganglia, hippocampus, and frontal-striatal circuits. Follow-up at age 16 identified continued greater vulnerability of the lowest-weight group relative to term comparison children on measures of visual-

402

neural plasticity in development

motor, memory, and executive functions than on language tests. In low-birth-weight children weighing more than 750 g, sequelae were limited to lower motor proficiency and reduced visual memory relative to the controls, reflecting some sparing of function (Taylor, Minich, Bangert, et al., 2004). In the Minnesota cohort, which followed children from more advantaged backgrounds, specific areas of reduced performance in preterm relative to term comparison children at age 7 to 9 included spatial-memory span, poorer working-memory strategy, and longer planning times on a spatial problem-solving task (Luciana et al., 1999). Curtis and colleagues followed a subset of this cohort, additional children who had required neonatal intensive care, and term comparison children at age 10–14 years. The children either born prematurely or with neonatal complications showed residual areas of difficulty, including shorter spatial-memory spans and more forgetting errors on a self-ordered search task. Indices of psychomotor speed, planning, and set shifting were similar across groups (W. J. Curtis et al., 2002). The authors hypothesized that the right dorsal and ventrolateral prefrontal cortex and reciprocal subcortical connections supporting spatial-memory span and problem-solving strategies were particularly sensitive to effects associated with preterm birth. This hypothesis was supported by recent fMRI findings of reduced activation on a response-inhibition task in the cerebellum, right caudate, thalamus, left inferior prefrontal, and left anterior cingulate gyrus, in conjunction with increased activation in temporal regions in adolescents born preterm (Nosarti et al., 2006) and by altered activation patterns of the head of the caudate related to reduced ability to execute a planned motor sequence (W. G. Curtis et al., 2006). Attention appears to be an area of vulnerability in preterm children. Visual attentional difficulties have been noted during infancy (Rose, Feldman, and Janowski, 2001). Increased behavioral symptoms of hyperactivity were apparent at age 8; symptoms of inattention but not hyperactivity persisted at age 20 (Hack et al., 2004). In addition to increased rates of internalizing and externalizing behaviors, preterm birth is associated with more than twice the risk of developing attention-deficit/hyperactivity disorder (ADHD) symptoms relative to term-born comparison children (Bhutta, Cleves, and Casey, 2002). In some studies, reduced hippocampal volumes have been related to ADHD symptoms (Abernethy, Cooke, and Foulder-Hughes, 2004). Extended follow-up studies have identified different skill trajectories in different samples. Meta-analysis indicated that general cognitive ability scores neither increased nor decreased with age at assessment, suggesting a stable deficit of approximately 10 IQ points in children born preterm who were tested from ages 5 to 14 (Bhutta, Cleves, and Casey, 2002). Studies employing comparison groups and growth-

modeling techniques showed that children born preterm evidenced slower gains in general cognitive and social skills (Landry et al., 1998) as well as in selected cognitive domains, including perceptual-motor planning, attention shifting, and speed of processing (Taylor, Minich, Klein, et al., 2004) than full-term comparison children. In addition, these latter studies also identified moderating effects of environmental variables on the level and rate of skill development (see later section “Moderators influencing outcome . . . ”). Overall, minimal sparing has been reported for children weighing more than 750 g at birth; they continue to score lower than full-term comparison children on nearly every test across multiple follow-up intervals into later adolescence and early adulthood (Taylor, Minich, Klein, et al., 2004). Particular vulnerability of children with extremely low birth weight is evident in visual-motor, memory, and executive processes with relative sparing of naming ability (Taylor, Minich, Bangert, et al., 2004). In summary, attempts to establish brain-behavior relations and examine possible mechanisms of reorganization in preterm children have yielded inconsistent findings. Although some studies noted limited catch-up growth and others reported decline in specific areas, the majority of studies report generally stable patterns of deficit. The persisting difficulties in IQ, attention, memory span, working memory, motor speed, and strategy use that are commonly identified across ages and samples suggest significant limits to reorganization of function. These behavioral difficulties appear to be mediated by thalamocorticostriatal networks shown to be affected in structural and functional imaging studies examining brain development from birth to young adulthood in children born preterm. The relations of the specific developmental stage at birth and its impact on the rate of development in specific cognitive and behavioral areas require further investigation. Children born preterm are an important group to study, as these children sustain insults during distinct phases of brain development, such as late stages of corticogenesis in the second trimester as well as organization/differentiation and myelination during the third trimester, that may exert unique influences on the spatial and temporal progression of gray and white matter formation and connectivity (Aylward, 2005). Neuroimaging studies highlight increasing abnormalities in gray matter nuclei and cortical gray matter with decreasing gestational age (Counsell et al., 2006; Inder et al., 2005). Counsell and Boardman (2005) noted that tissue contraction within the thalamus and lentiform nuclei reflected the greatest regional volume loss in preterm children scanned at term dates. Prefrontal structures undergo extensive and protracted development throughout childhood and adolescence (Sowell et al., 2003; Toga, Thompson, and Sowell, 2006). If rapidly developing tissue is particularly vulnerable to disruption, then developmental stage at the time of

injury should be related to outcomes. Therefore, it would be interesting to see if executive attention deficits were more strongly predicted by integrity and function of striatal structures in the youngest cohorts of children and by integrity of prefrontal pathways in children with increasing gestational age.

Traumatic brain injury Neurodevelopmental Processes Disrupted by TBI Traumatic brain injury sustained during childhood may interfere with the maturational changes in brain morphology that are related to synaptic exuberance and pruning. Dynamic cortical mapping based on longitudinal MRI studies obtained from ages 4 to 21 years shows that gray matter density increases first in primary sensorimotor cortices and in frontal and occipital poles. The remaining regions mature from back to front in parietal to frontal regions, with the association areas of the superior temporal and dorsolateral prefrontal cortices developing last (Giedd et al., 1999b; Gogtay et al., 2004). In children ages 5–11 years, Sowell and colleagues noted gray matter thinning in the right dorsal frontal and biparietal regions, in conjunction with increased gray matter in the left frontal and temporal-parietal language areas (Sowell et al., 2003). Nonlinear increases in gray matter peak in different regions, with increases in the frontal and parietal regions at about age 12 and in the temporal lobe at about age 16, followed by gray matter loss (Giedd et al., 1999b, 1999a). White matter increases in a more linear pattern in frontal, temporal, and parietal regions. The protracted development of white matter appears to be related to myelination and is noted up to the third decade in adulthood (Hasan et al., 2007; Sowell et al., 2003; Toga, Thompson, and Sowell, 2006). The corpus callosum develops in a rostral to caudal pattern (Giedd et al., 1996; Rajapakse et al., 1996). Longitudinal studies identified rapid developmental changes in children ages 3–6 in the anterior regions, while fiber systems in the isthmus and splenium grew more rapidly from 5 to 18 years of age (Giedd et al., 1996; Thompson et al., 2000). Traumatic brain injury typically reflects a combination of focal and diffuse brain insults. Focal insults include mass lesions, including contusions, intracerebral hematomas, and extra-axial hematomas. These lesions produce behavioral changes as a result of direct tissue damage in addition to remote mass effects, such as midline shifts and herniations. The primary pathophysiological consequence of TBI is traumatic axonal injury (Gennarelli, Thibault, and Graham, 1998). In contrast to focal injuries in which the brain sustains a direct impact, axonal injury results from angular and rotational head movements of high magnitude that occur with or without impact (Gennarelli, Thibault, and Graham, 1998). Diffuse pathologic changes reflect the combined

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

403

effects of direct injury to both axons and cell bodies as well as effects of secondary reactive pathology related to anoxic, contusion, hemorrhagic, perfusion, and reperfusion mechanisms, as well as cascades of excitatory neurotransmitters (Singleton and Povlishock, 2004). Pathologic studies of diffuse TBI have revealed microscopic features corresponding to Wallerian-type axonal degeneration that most prominently affect the subcortical white matter, corpus callosum, and dorsolateral aspect of the upper brainstem (Adams et al., 1977). Neuroimaging Studies of Children with TBI Posttraumatic changes in cerebral macrostructure have been assessed using volumetric analysis of regional white and gray matter as well as lesion volumes. In children with TBI, indices of atrophy and lesion volume from structural MRIs are significant predictors of cognitive and functional outcomes. The distribution of focal injuries varies with age at injury; subdural, subarachnoid, and intraparenchymal hemorrhage as well as edema occur frequently in infants and young children (Ewing-Cobbs et al., 2000). Focal areas of enhanced signal intensity occur in approximately 75 percent of school-aged children and adolescents; the dorsolateral, orbitofrontal, and frontal lobe white matter are the most common sites of focal damage (Levin et al., 1997). Although volumetric studies uniformly report significantly reduced whole-brain volume in children with TBI relative to control children, findings are inconsistent regarding regional volume reduction. While some studies reported reductions in frontal white matter but not gray matter (Serra-Grabulosa et al., 2005), others report reductions in both gray and white matter in the superior medial and ventromedial prefrontal and temporal regions and selective reduction of white matter in the lateral frontal region (Wilde et al., 2005). Tasker and colleagues found that severe TBI associated with elevated intracranial pressure affected both brain growth and morphology assessed approximately 5 years postinjury. In one subgroup, periventricular white matter loss was associated with reduced brain volume relative to head circumference consistent with atrophy; the other subgroup had expected volume relative to head circumference, suggesting minimal head growth postinjury but no atrophy (Tasker et al., 2005). Several studies have examined posttraumatic changes to components of thalamostriatocortical networks. The impact of TBI on hippocampal volumes has been inconsistent; one study examining a younger cohort with a mean of 8 years noted a trend for reduced hippocampal volume (Di Stefano et al., 2000); other studies examining adolescents reported reduced right hippocampal volume corrected for hemisphere volume with normal perihippocampal volumes (Tasker et al., 2005), and another identified bilateral reduction (Serra-Grabulosa et al., 2005). Striatal volumes were within

404

neural plasticity in development

normal limits (Serra-Grabulosa et al., 2005). Focal tissue loss was noted in gray and white matter in the superior medial and ventromedial frontal regions, in white matter in lateral frontal regions, and in both gray and white matter in the temporal lobe (Wilde et al., 2005). In our sample of infants and preschoolers with TBI, we examined volumes of gray and white matter in precallosal, midcallosal, and retrocallosal regions in relation to behavioral outcomes in children 2–9 years of age at the time of scan. After controlling for age at testing, verbal and visual measures of general cognitive ability correlated predominantly with bilateral midcallosal and retrocallosal gray matter volumes; self-ordered pointing scores correlated with bilateral precallosal gray and white matter volumes (EwingCobbs, unpublished data). Studies of school-aged children and adolescents indicated that corpus callosum volumes were strongly related to indices of diffuse axonal injury as well as to cognitive measures assessing semantic memory, mental flexibility, processing speed, and concept formation (Benavidez et al., 1999; Verger et al., 2001). The rate of growth of the corpus callosum was significantly reduced during the first 3 years after severe pediatric TBI, reflecting distortion of subsequent neural developmental processes (Levin et al., 2000). Recent microstructural studies employing DTI to examine posttraumatic changes have identified reduced fractional anisotropy in the internal capsule, anterior commissure, and corpus callosum (Ewing-Cobbs et al., 2006; Wilde, Chu, et al., 2006; Wilde, Bigler, et al., 2006), even in patients with normal conventional MRI (Rugg-Gunn et al., 2001). We completed preliminary analyses of sensitivity of DTI metrics in multiple regions of interest representing gray matter nuclei and projection, commissural, and association white matter pathways in youth sustaining TBI at 0–15 years of age relative to a comparison group. The transverse diffusivities, which have been related to abnormalities in myelination (Sun et al., 2006), were sensitive to posttraumatic changes in gray matter (caudate), projection (corticospinal tract, posterior limb internal capsule), commissural (genu, isthmus, and splenium), and association fibers (frontal minor, arcuate, uncinate, cingulate) above and beyond the variance accounted for by the longitudinal diffusivity that has been related to axonal integrity (Sun et al., 2006). This pattern of findings suggests (1) that disruption of myelination is a primary pathologic finding following moderate to severe TBI and (2) that frontolimbic structures (uncinate, cingulate, genu of the corpus callosum, and frontal minor regions) are vulnerable to disruption by TBI (Hasan and Ewing-Cobbs, unpublished data). Figure 24.1 and plate 45 show application of diffu-sion tensor tractography to identification of posttraumatic changes in commissural, association, and projection fibers in two children scanned at age 9 who had sustained TBI during infancy.

Figure 24.1 Conventional MRI (top row) and diffusion tensor tractography (bottom row) in three 9-year-old children. Images show white matter pathways in a healthy child (left) and in two children who sustained early TBI (middle and right). The middle image is from a child sustaining bilateral subdural hematomas, subarachnoid hemorrhage, and frontal lobe contusions at 2 months of age; follow-up MRI disclosed focal atrophy in left parietal cortex, deep white periventricular atrophy with associated thinning of the corpus callosum, and compensatory enlargement of the lateral ventricles. Tractography shows reduced callosal fibers (left-right fibers

indicated in red), diminution of the corticospinal tract (inferiorsuperior fibers in blue), and diminution of association fibers (anterior-posterior fibers in green) in prefrontal regions and in arcuate and superior longitudinal fasciculi. Images on the right side are from a child sustaining a parietal depressed skull fracture with underlying subdural hematoma at 3 months of age; follow-up MRI revealed left parietal encephalomalacia and moderate thinning of the posterior body of the corpus callosum. Tractography shows focal reduction in posterior callosal and association fibers consistent with focal parietal lobe injury. (See plate 45.)

Long-Term Neuropsychological Outcome Studies of Children with TBI Follow-up studies have identified strong relations of both severity of TBI and age at TBI on a variety of outcomes. Traumatic brain injury may preferentially affect specific executive processes because of the frequent occurrence of diffuse traumatic axonal damage and focal frontal lobe damage (Levin et al., 1993). Difficulties in behavioral self-regulation, attention, and working memory occur frequently following moderate to severe TBI and likely reflect alteration of thalamocorticostriatal pathways. Preinjury diagnosis of attention-deficit/hyperactivity disorder (ADHD) is present in approximately 20 percent of children (Gerring et al., 1998). The new onset of ADHD

symptoms after TBI occurs in approximately 15 to 20 percent of children with severe TBI (Gerring et al., 1998; Max et al., 2005b). The onset of secondary ADHD symptoms is more common in children with thalamic and basal ganglia injury (Gerring et al., 2000) or lesions of the orbitofrontal gyrus (Max et al., 2005a). Low socioeconomic status and psychosocial adversity prior to the injury were associated with onset of symptoms between 6 and 24 months after TBI, reflecting the interplay of injury and environmental factors (Max et al., 2005b). Parent ratings indicated the persistence of clinically significant attention problems in 26 percent of a sample of youth with moderate to severe TBI; childhood TBI exacerbated premorbid attentional deficits

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

405

(Yeates et al., 2005). In long-term follow-up studies of children and adolescents with severe TBI, Levin and colleagues identified reductions in updating information in phonological working memory. Longitudinal analysis revealed that the severe TBI group’s working-memory scores assessed using an n-bask task declined from 1 to 2 years after injury, while growth in scores was noted in children with less severe injuries and comparison children (Levin et al., 2004). Primary effects of injury severity in other studies revealed that severe TBI reduced inhibitory control in children, particularly on tasks with increased demand for resisting interference or inhibiting a prepotent response (Levin et al., 2004) and on multidetermined tasks assessing working memory, response inhibition, and shifting (Yeates et al., 2005). Performance on go-no-go and stop-signal tasks was reduced in children with severe TBI relative to community controls in some studies (Konrad et al., 2000; Levin et al., 1993) and only in children who developed secondary attention-deficit/hyperactivity disorder in other samples (Schachar et al., 2004). Impact of Age at Injury Some studies suggest interactions of age at injury and severity of injury, with greater impact often noted in children injured at younger ages. Longitudinal outcome studies examining recovery of IQ scores after pediatric TBI are consistent with the hypothesis that skills in

a rapid stage of development may be particularly vulnerable to disruption by brain injury. Composite or Full-Scale IQ scores are lower in infants and young children than in school-aged children and adolescents (V. Anderson et al., 2005; Ewing-Cobbs et al., 1997). Recovery curves showing change in composite IQ scores over the first year after TBI indicate lower initial scores and/or less acceleration of scores in infants and preschool-aged children than in older children (Ewing-Cobbs, Barnes, and Fletcher, 2003; see figure 24.2). The pattern of component scores appears to vary with age at injury. Across a 1- to 2-year interval, younger children with mild to severe TBI show relative vulnerability in the skills tapped by the Verbal IQ score (V. Anderson et al., 2005); older children and adolescents may demonstrate greater vulnerability of abilities assessed by PIQ scores (Chadwick et al., 1981). Significant heterogeneity in IQ scores and in recovery patterns reflects differences in subject selection as well as variables such as cognitive reserve and socioenvironmental variables that moderate outcomes (Taylor et al., 1999; Yeates et al., 1997). Because networks involved in executive processes have a protracted developmental course (Huttenlocher and Dabholkar, 1997b; Rubia et al., 2006), specific processes may be selectively vulnerable during infancy, preschool, school-aged, and adolescent periods. Extended growth of both dorsolateral and orbitofrontal regions is implicated in

105

IQ Score

100 95 90 85 80 baseline

12

Months After Brain Injury Infant Mild-Mod Presch Mild-Mod Sch/Adol Mild-Mod

Infant Mod-Sev Presch Mod-Sev Sch/Adol Mod-Sev

Figure 24.2 Composite or Full-Scale IQ scores abstracted from the literature for infants, preschoolers, and school-aged children and adolescents with differing levels of TBI severity suggest reduced plasticity in younger children. IQ scores of infants and preschoolers

406

neural plasticity in development

Infant Severe Presch Severe Sch/Adol Severe

are lower and show less positive acceleration over 1 year than recovery curves depicting change over time in school-aged children and adolescents.

the lengthy trajectory of development of both cognitive and behavioral self-regulation as well as in longitudinal measures of cortical gray matter volumes (Galvan et al., 2006; Gogtay et al., 2004). Prefrontal regions develop rapidly in infants and preschool-aged children; peak synaptic density in medial prefrontal cortex occurs at ages 3 to 4 (Huttenlocher et al., 1997b) and is associated with corresponding increases noted in working memory, inhibitory control, and social cognition (Diamond, 2002; Hughes, Dunn, and White, 1998). Significant pruning occurs during mid- to late adolescence, reflecting the extended developmental trajectory of frontal lobe maturation (Huttenlocher and Dabholkar, 1997a). Therefore, early childhood may represent a developmental stage during which executive processes are particularly vulnerable to disruption by TBI. Injuries sustained during infancy and preschool years that involve orbitofrontal and dorsolateral prefrontal regions should preferentially disrupt emotional regulation, social cognition, working memory, and inhibitory control. Our preliminary studies provide some support for this hypothesis. In infants with TBI, joint attention, which is mediated by orbitofrontal pathways, is significantly reduced (Landry et al., 2004). Moderate to severe TBI sustained prior to age 6 disrupted performance on self-ordered pointing and delayed response tasks that require holding mental representations over a delay and inhibiting prepotent responses and that show rapid development during this age. In contrast, TBI did not reduce performance on set-shifting tasks that show less developmental change during this period (Ewing-Cobbs et al., 2004). When tested 5 years after TBI, children 2–5 years old at the time of injury had poorer age-adjusted performance than those injured during infancy on self-ordered pointing-and-response conflict tasks (EwingCobbs et al., 2006). Similarly, V. Anderson and colleagues (2006) identified deficits persisting 2.5 years after TBI sustained at ages 2–7 in visual memory span, story recall, spatial learning, and everyday memory. In school-aged children and adolescents, younger age at injury was associated with less adequate performance in several executive processes, including more perseverative errors on a card-sorting task and with reduced word fluency (Levin et al., 1993, 2001; Slomine et al., 2002). Vigilance, distractibility, and response modulation scores indicated increased difficulties in children injured at younger ages (Dennis et al., 2001).

Moderators influencing outcome from early brain injury: experience-expectant and experience-dependent processes Cognitive development following early brain injury raises issues regarding biological constraints on development and the role of psychosocial interventions. Studies of environmental influences on outcome in children with early brain injuries support a strong relation between outcome and environ-

mental factors. Taylor, Minich, Klein, and Hack (2004) found that environmental influences moderated the growth rate of several cognitive tasks in a longitudinal study of children born preterm. Children who weighed less than 750 g progressed more slowly than controls if they were from a family with a high socioeconomic level. Interestingly, preterm children who weighed more than 750 g and were from a low socioeconomic background had faster rates of skill development on several cognitive tasks than normal controls. Taylor and colleagues postulate that children with early brain injury may be constrained in their ability to benefit from environmental stimulation, consistent with the experience-expectant model. Thus preterm children from more advantaged backgrounds may not be initially constrained in their development by their environment but, as they age, may be biologically constrained from benefiting from a more enriched environment. In contrast, preterm children who are of heavier weight (>750 g) may have not been impacted by the socioeconomic disadvantage. Although socioeconomic status is an important predictor of outcome, several studies have demonstrated that other aspects of the home and family environment may influence outcome from early brain injuries. Although family income was related to cognitive ability in low-birth-weight premature infants, the effect was mediated through the provision of stimulating experiences in the home (Linver, BrooksGunn, and Kohen, 2002). Maternal emotional distress and parenting practices mediated the relation between income and children’s behavior problems. In a longitudinal study of preterm children, Smith, Landry, and Swank (2006) found that preterm children who experienced responsive parenting through infancy and preschool had higher levels of developmental outcome at 10 years of age than children who had less responsive parenting. The quality of the home environment has also been found to be a greater predictor of outcome than medical risk factors associated with very low birth weight (Molfese, DiLalla, and Bunce, 1997). The impact of environment on outcome has not been extensively studied in young children with TBI. Anderson and colleagues (V. Anderson et al., 2006) found that young children with severe TBI who had lower skills preinjury and who had greater family dysfunction had worse outcome than children with milder injuries. The impact of environmental factors on outcome from TBI in school-aged children is consistent with the findings we have reviewed. After accounting for injury severity, Yeates and colleagues (2002) found that preinjury family environment significantly predicted cognitive and behavioral outcome 1 year postinjury for children and adolescents with TBI. Taylor and colleagues (2001) identified that higher parent stress at 6 months postinjury predicted more child behavioral difficulties at 12 months postinjury and that more child behavior problems at 6 months predicted worse family outcomes at 12 months

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

407

postinjury. In essence, they found a bidirectional influence of child and family on outcomes following brain injury. In a long-term follow-up study conducted an average of 4 years postinjury, Taylor and colleagues (2002) found that social disadvantage for children with TBI predicted greater behavioral difficulties and poorer cognitive outcome. Although the moderating impact of environment in studies of children with early brain injury provides some evidentiary support for experience-dependent programming, it is not known whether positive/enriched home environments accelerate learning or simply allow learning to occur, whereas families with limited financial resources and greater stress impede the natural “recovery” process that may occur in children with early brain injuries. To understand whether a brain-injured child’s developmental trajectory can be truly altered, well-designed studies of cognitive and psychosocial interventions that utilize brain imaging are needed. Two recent studies provide neuroimaging evidence for experience-dependent and experience-expectant processes. One such study was conducted by Als and colleagues (2004), who randomized 30 preterm infants in the neonatal intensive care unit to an intervention program that provided developmentally supportive care or a control group that received standard care. The treatment was initiated within 72 hours of admission to the unit and continued until the infant was 2 weeks of age, corrected for prematurity. Neurobehavioral testing conducted at 2 weeks of age found the treatment group had better motor development and better self-regulation. These gains were maintained at 9 months of age, with the treatment group demonstrating significantly higher cognitive and motor scores on developmental measures. At 2 weeks of age EEG revealed increased coherence in the fast alpha and beta frequency bands between the left frontal region and occipital and parietal regions for the treatment group. Increased frontal-occipital coherence was significantly correlated with motor skills and attention. Scans conducted using DTI at 2 weeks of age indicated significant increase in overall fractional anisotropy and regional increases in the left internal capsule in the treatment group. Less behavioral hypersensitivity was associated with more mature development of the internal capsule and frontal white matter. The authors postulate that these findings highlight the vulnerability of the frontal lobes to environmental experiences. These findings indicate that time-limited experiences after birth can alter both brain function and brain structure and provide convergent imaging and behavioral data that support the occurrence of experience-dependent plasticity (Als et al., 2004). Another intriguing neuroimaging study examined children adopted from Romanian orphanages who experienced significant socioemotional deprivation: DTI studies revealed reduced fractional anisotropy values in the left uncinate fas-

408

neural plasticity in development

ciculus in adoptees relative to controls, suggesting adverse impact of environmental conditions on the organization of specific neural pathways (Eluvathingal et al., 2006). The adopted children had decreased metabolism in the orbital frontal gyrus, infralimbic prefrontal cortex, medial temporal structures, lateral temporal cortex, and brain stem (Chugani et al., 2001). Cumulatively, the findings illustrate how experience-expectant processes influenced by socioemotional deprivation adversely impact development and how experience-dependent processes associated with provision of positive environmental conditions during key developmental stages may alter neural development and skill trajectories over a protracted time frame.

Conclusions Developmental outcomes after early brain injury are complexly determined. Although recent studies have enhanced our understanding of neurodevelopmental outcomes in children born preterm and those sustaining TBI during different developmental stages, several issues remain outstanding. These issues include understanding the main and/or interactive effects of type and severity of injury, the developmental stage at which the injury is sustained and at which outcomes are evaluated, and the impact of socioenvironmental conditions on outcomes. Reorganization appears to be less favorable following conditions producing diffuse brain insult, such as TBI and preterm birth, than from focal brain lesions sustained during childhood. Despite different mechanisms of injury, brain injury associated with preterm birth and TBI may share features of secondary brain damage and may serve as models of primarily diffuse insult occurring at different developmental stages. In both preterm and TBI populations, the severity of the brain insult is a major predictor of outcome from initial stages of recovery through late follow-up. These conditions have both been conceptualized as primarily white matter disorders. Diffuse microscopic lesions in cerebral and cerebellar white matter are the primary pathophysiological substrate in TBI (Gennarelli, Thibault, and Graham, 1998), while noncystic diffuse injury to white matter is the most common finding in preterm children (Olesen et al., 2003). However, recent volumetric and DTI studies implicate loss and alteration of gray matter as a significant concern in both populations (Counsell and Boardman, 2005; Wilde et al., 2005). Children experiencing either preterm birth or severe TBI often show reduced velocity of brain growth as evidenced by reduced head circumference, reduced volumes of gray and white matter, and arrested development of specific structures, such as the corpus callosum. However, decreased volume may not always be accompanied by comparable behavioral deficit because compensatory changes, including dendritic arborization and density, may partially attenuate

the effects of injury (W. G. Curtis et al., 2006). Thalamocorticostriatal pathways appear to be vulnerable to injury in both groups of children, with associated deficits in cognitive and behavioral control. Recent fMRI studies highlight mechanisms of neural reorganization following preterm birth and early unilateral lesions. In adolescents born preterm who show normal behavioral performance on memory tasks tapping thalamostriatocortical pathways, neural activation patterns differed from those of typically developing adolescents. The altered activation appeared related to changes in related pathways rather than to the activation of the hippocampus (W. G. Curtis et al., 2006). Similarly, studies of children with unilateral lesions have documented mechanisms of both intrahemispheric and interhemispheric reorganization of spatial and somatosensory functions (Chu et al., 2000; Stiles et al., 2003). The apparent impact of a brain injury on a given skill may depend on the developmental stage at which the lesion is sustained and at which the skill is evaluated (Goldman, 1974; Kolb, Pellis, and Robinson, 2004). Some skills may show a stable deficit over time, others may show a transient lag and partial catch-up growth, and others may show increasing deficit as the damaged substrate required for mature expression of a given skill cannot support the skill and the child falls farther behind age expectations. Diffuse brain injury resulting from preterm birth or TBI significantly reduces the level of performance in most areas. The rate of development is typically parallel across groups, implying that the injured group continues to acquire new skills at the new lower rate— for example, at the level of a child with general cognitive scores of 85 as opposed to 100. Does the framework outlined by Kolb, Gibb, and Gorny (2000) regarding the impact of focal lesions acquired during different developmental stages on outcomes apply to cases of predominantly diffuse injury? Based on this framework, injuries that occur during periods of cell migration (second to third trimester) and synaptic pruning (adolescence) should be associated with less favorable outcomes than those occurring during periods of dendritic growth and massive synaptogenesis (which peaks at 4 months in visual cortex, at 3–4 years for prefrontal and superior temporal cortex). At present, the literature suggests that general cognitive functions such as IQ scores are lower in children who sustain injuries during infancy and preschool years than in older children. A variety of executive processes presumably mediated by thalamocorticostriatal pathways appear to be more affected by injuries sustained by preschoolers than infants and in injuries sustained by school-aged children than adolescents when evaluated several years after injury. Therefore, vulnerability may be greater during periods of peak synapse formation, consistent with the rapid-development hypothesis. Future studies fusing structural and functional

imaging methods and examining developmental change in well-defined measures of cognitive and motor processes involving the thalamic-striatal-cortical network may provide insight into mechanisms through which specific nuclei and pathways are affected by complications associated with preterm birth and TBI occurring at different developmental stages. Time since injury is a critically important variable. In children with TBI, time-linked mechanisms of plasticity described by Pascual-Leone and colleagues (2005) are apparent in the rapid recovery of school-aged children and adolescents in the first few months after TBI. However, subsequent learning that is dependent upon the damaged substrate is adversely affected. The impact of time since injury is best characterized using models that depict longitudinal posttraumatic changes in the level, rate of change, and eventual performance level of a given ability or skill and allow for assessment of possible moderating effects of experience-expectant and experience-dependent processes. In conclusion, additional studies characterizing structural and functional mechanisms supporting intrinsic reorganization of ability structures are needed. Outcomes appear to be less favorable following diffuse injury sustained early in life than in focal injuries and diffuse injuries sustained later in childhood and adolescence. The severity of brain insult, as indexed by gestational age, birth weight, and medical complications in children born preterm and by the depth and duration of impaired consciousness in children sustaining TBI, is a robust predictor of outcome. At present, how diffuse brain injury sustained at different stages of brain maturation is expressed in different ability structures, when assessed years following the insult, is unknown. For example, preterm birth occurring during the second trimester may be associated with greater injury to subcortical gray matter structures than either preterm birth occurring later during gestation or following brain injury sustained at later stages of development. Growth models that examine the level of skill development and rate of change, and that allow dissociation of constructs of lag, delay, deficit, and arrest, will assist in characterizing the ultimate impact of early brain injury on the trajectory and endpoint of developing skills. In addition, growth models can examine interactions of brain injury and socioenvironmental and experiential factors on diverse outcomes. Understanding of mechanisms of reorganization will be enhanced with the use of similar measures of brain structure and function and behavioral outcome measures in children who sustain focal and diffuse insults at different developmental stages and who are followed longitudinally. These types of studies will allow dissociation of maturational, injury-related, and experience-dependent changes in brain growth and connectivity as they relate to neurobehavioral development.

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

409

REFERENCES Abernethy, L. J., R. W. Cooke, and L. Foulder-Hughes, 2004. Caudate and hippocampal volumes, intelligence, and motor impairment in 7-year-old children who were born preterm. Pediatr. Res. 55:884–893. Adams, J. H., D. E. Mitchell, D. I. Graham, and D. Doyle, 1977. Diffuse brain damage of the immediate impact type. Brain 100:489–502. Als, H., F. H. Duffy, G. B. McAnulty, M. J. Rivkin, S. Vajapeyam, R. V. Mulkern, et al., 2004. Early experience alters brain function and structure. Pediatrics 113:846–857. Anderson, N. G., I. Laurent, N. Cook, L. Woodward, and T. E. Inder, 2005. Growth rate of corpus callosum in very premature infants. AJNR Am. J. Neuroradiol. 26:2685–2690. Anderson, V. A., C. Catroppa, P. Dudgeon, S. A. Morse, F. Haritou, and J. V. Rosenfeld, 2006. Understanding predictors of functional recovery and outcome 30 months following early childhood head injury. Neuropsychology 20:42–57. Anderson, V., C. Catroppa, S. Morse, F. Haritou, and J. Rosenfeld, 2005. Functional plasticity or vulnerability after early brain injury? Pediatrics 116:1374–1382. Aylward, G. P., 2005. Neurodevelopmental outcomes of infants born prematurely. J. Dev. Behav. Pediatr. 26:427–440. Beckett, C., B. Maughan, M. Rutter, J. Castle, E. Colvert, C. Groothues, et al., 2006. Do the effects of early severe deprivation on cognition persist into early adolescence? Findings from the English and Romanian adoptees study. Child Dev. 77:696– 711. Benavidez, D. A., J. M. Fletcher, J. H. Hannay, S. T. Bland, S. E. Caudle, D. Mendelsohn, et al., 1999. Corpus callosum damage and interhemispheric transfer of information following closed head injury in children. Cortex 35:315–336. Bengtsson, S. L., Z. Nagy, S. Skare, L. Forsman, H. Forssberg, and F. Ullen, 2005. Extensive piano practicing has regionally specific effects on white matter development. Nature Neurosci. 8:1148–1150. Bhutta, A. T., M. A. Cleves, and P. H. Casey, 2002. Cognitive and behavioral outcomes of school-aged children who were born preterm. JAMA 288:728–737. Chadwick, O., M. Rutter, G. Brown, D. Shaffer, and M. Traub, 1981. A prospective study of children with head injuries. II. Cognitive sequelae. Psychol. Med. 11:49–61. Chu, D., P. R. Huttenlocher, D. N. Levin, and V. L. Towle, 2000. Reorganization of the hand somatosensory cortex following perinatal unilateral brain injury. Neuropediatrics 31:63–69. Chugani, H. T., M. E. Behen, O. Muzik, C. Juhasz, F. Nagy, and D. C. Chugani, 2001. Local brain functional activity following early deprivation: A study of postinstitutionalized Romanian orphans. NeuroImage 14:1290–1301. Counsell, S. J., and J. P. Boardman, 2005. Differential brain growth in the infant born preterm: Current knowledge and future developments from brain imaging. Sem. Fetal Neonatal Med. 10:403–410. Counsell, S. J., Y. Shen, J. P. Boardman, D. J. Larkman, O. Kapellou, P. Ward, et al., 2006. Axial and radial diffusivity in preterm infants who have diffuse white matter changes on magnetic resonance imaging at term-equivalent age. Pediatrics 117:376–386. Curtis, W. J., L. L. Lindeke, M. K. Georgieff, and C. A. Nelson, 2002. Neurobehavioural functioning in neonatal intensive care unit graduates in late childhood and early adolescence. Brain 125:1646–1659.

410

neural plasticity in development

Curtis, W. G., J. Zhuang, E. L. Townsend, X. Hu, and C. A. Nelson, 2006. Memory in early adolescents born prematurely: A functional magnetic resonance imaging investigation. Dev. Neuropsychol. 29:341–377. Dennis, M., S. Guger, C. Roncadin, M. A. Barnes, and R. Schachar, 2001. Attentional-inhibitory control and socialbehavioral regulation after childhood closed head injury: Do biological, developmental, and recovery variables predict outcome? J. Int. Neuropsychol. Soc. 7:683–692. Dennis, M., K. O. Yeates, H. G. Taylor, and J. M. Fletcher, 2006. Brain reserve capacity, cognitive reserve capacity, and age-based functional plasticity after congenital and acquired brain injury in children. In Y. Stern, ed., Cognitive Reserve, 53–83. New York: Psychology Press. Di Stefano, G., J. Bachevalier, H. S. Levin, J. Song, R. S. Scheibel, and J. M. Fletcher, 2000. Volume of focal brain lesions and hippocampal formation in relation to memory function after closed head injury in children. J. Neurol. Neurosurg. Psychiatry 69:210–216. Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: Cognitive functions, anatomy, and biochemistry. In D. T. Stuss and R. T. Knight, eds., Frontal Lobe Function, 466–503. New York: Oxford University Press. Eluvathingal, T. J., H. T. Chugani, M. E. Behen, C. Juhasz, O. Muzik, M. Maqbool, et al., 2006. Abnormal brain connectivity in children after early severe socioemotional deprivation: A diffusion tensor imaging study. Pediatrics 117:2093–2100. Ewing-Cobbs, L., M. A. Barnes, and J. M. Fletcher, 2003. Early brain injury in children: Development and reorganization of cognitive function. Dev. Neuropsychol. 24(2&3):669–704. Ewing-Cobbs, L., J. M. Fletcher, H. S. Levin, D. J. Francis, K. Davidson, and M. E. Miner, 1997. Longitudinal neuropsychological outcome in infants and preschoolers with traumatic brain injury. J. Int. Neuropsychol. Soc. 3:581–591. Ewing-Cobbs, L., K. Hasan, M. Prasad, L. Kramer, and J. Bachevalier, 2006. Corpus callosum diffusion anisotropy correlates with neuropsychological outcomes in twins disconcordant for traumatic brain injury. AJNR Am. J. Neuroradiol. 27:879– 881. Ewing-Cobbs, L., H. S. Levin, H. M. Eisenberg, and J. M. Fletcher, 1987. Language functions following closed-head injury in children and adolescents. J. Clin. Exp. Neuropsychol. 9:575–592. Ewing-Cobbs, L., M. Prasad, L. Kramer, P. T. Louis, J. Baumgartner, J. M. Fletcher, et al., 2000. Acute neuroradiologic findings in young children with inflicted or noninflicted traumatic brain injury. Childs Nerv. Syst. 16:25–34. Ewing-Cobbs, L., M. Prasad, S. H. Landry, L. Kramer, and R. DeLeon, 2004. Executive functions following traumatic brain injury in young children: A preliminary analysis. Dev. Neuropsychol. 26:487–512. Ferriero, D. M., 2004. Neonatal brain injury. N. Engl. J. Med. 351:1985–1995. Galvan, A., T. A. Hare, C. E. Parra, J. Penn, H. Voss, G. Glover, et al., 2006. Earlier development of the accumbens relative to orbitofrontal cortex might underlie risk-taking behavior in adolescents. J. Neurosci. 26:6885–6892. Gennarelli, T. A., L. E. Thibault, and D. I. Graham, 1998. Diffuse axonal injury: An important form of traumatic brain damage. Neuroscientist 4:202–215. Gerring, J. P., K. D. Brady, A. Chens, C. Quinn, E. Herskovits, K. Bandeen-Roches, et al., 2000. Neuroimaging variables related to development of secondary attention deficit hyperactiv-

ity disorder after closed head injury in children and adolescents. Brain Inj. 14:205–218. Gerring, J. P., K. D. Brady, A. Chen, R. Vasa, M. Gardos, K. J. Bandeen-Roche, et al., 1998. Premorbid prevalence of ADHD and development of secondary ADHD after closed head injury. J. Am. Acad. Child Adolesc. Psychiatry 37:647–654. Giedd, J. N., J. D. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, et al., 1999a. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2:861–863. Giedd, J. N., J. D. Blumenthal, N. O. Jeffries, J. C. Rajapakse, A. C. Vaituzis, H. Liu, et al., 1999b. Development of the human corpus callosum during childhood and adolescence: A longitudinal MRI study. Prog. Neuropsychopharmacol. Biol. Psychiatry 23:571–588. Giedd, J. N., J. M. Rumsey, F. X. Castellanos, J. C. Rajapakse, D. Kaysen, A. C. Vaituzis, et al., 1996. A quantitative MRI study of the corpus callosum in children and adolescents. Dev. Brain Res. 91:274–280. Gimenez, M., C. Junque, A. Narberhaus, F. Botet, N. Bargallo, and J. M. Mercader, 2006. Correlations of thalamic reductions with verbal fluency impairment in those born prematurely. NeuroReport 17:463–466. Gogtay, N., J. N. Giedd, L. Lusk, K. M. Hayashi, D. K. Greenstein, A. C. Vaituzis, et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101:8174–8179. Goldman, P. S., 1974. An alternative to developmental plasticity: Heterology of CNS structures in infants and adults. In D. G. Stein, J. J. Rosen, and N. Butters, eds., Plasticity and Recovery of Function in the Central Nervous System, 149–174. New York: Academic. Greenough, W. T., J. E. Black, and C. S. Wallace, 2002. Experience and brain development. In M. H. Johnson, Y. Munakata, and R. O. Gilmore, eds., Brain Development and Cognition, 2nd ed., 186–216. Malden, MA: Blackwell. Gupta, R. K., K. Hasan, R. Trivedi, M. Pradhan, V. Das, N. A. Parikh, et al., 2005. Diffusion tensor imaging of the developing human cerebrum. J. Neurosci. Res. 81:172–178. Hack, M., H. G. Taylor, N. Klein, R. Eiben, C. Schatschneider, and N. Mercuri-Minich, 1994. School-age outcomes in children with birth weights under 750 g. N. Engl. J. Med. 331:753–759. Hack, M., E. A. Youngstrom, L. Cartar, M. Schluchter, H. G. Taylor, D. Flannery, et al., 2004. Behavioral outcomes and evidence of psychopathology among very low birth weight infants at age 20 years. Pediatrics 114:932–940. Hasan, K., C. Halphen, A. Sankar, T. J. Eluvathingal, L. Kramer, K. K. Stuebing, et al., 2007. Diffusion tensor imaging– based tissue segmentation: Validation and application to the developing child and adolescent brain. NeuroImage 34(4): 1497–1505. Hebb, D. O., 1942. The effect of early and late brain injury upon test scores, and the nature of normal adult intelligence. Proc. Am. Philos. Soc. 85:275–292. Hughes, C., J. Dunn, and A. White, 1998. Trick or treat? Uneven understanding of mind and emotion and executive dysfunction in “hard-to-manage” preschoolers. J. Child Psychol. Psychiatry 39:981–994. Huppi, P. S., S. E. Maier, S. Peled, G. P. Zientara, P. D. Barnes, F. A. Jolesz, et al., 1998. Microstructural development of human newborn cerebral white matter assessed in vivo by diffusion tensor magnetic resonance imaging. Pediatr. Res. 44:584–590.

Huttenlocher, P. R., and A. S. Dabholkar, 1997a. Developmental anatomy of prefrontal cortex. In G. R. Lyon and P. S. Goldman-Rakic, eds., Development of the Prefrontal Cortex: Evolution, Neurobiology, and Behavior, 69–84. Baltimore: Paul H. Brooks. Huttenlocher, P. R., and A. S. Dabholkar, 1997b. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Inder, T. E., S. K. Warfield, H. Wang, P. S. Huppi, and J. J. Volpe, 2005. Abnormal cerebral structure is present at term in premature infants. Pediatrics 115:286–294. Isaacs, E., A. Lucas, and W. K. Chong, 2000. Hippocampal volume and everyday memory in children of very low birth weight. Pediatr. Res. 47:713–720. Isaacs, E., F. Vargha-Khadem, K. E. Watkins, A. Lucas, M. Mishkin, and D. G. Gadian, 2003. Developmental amnesia and its relationship to degree of hippocampal atrophy. Proc. Natl. Acad. Sci. USA 100:13060–13063. Kesler, S. R., B. Vohr, K. C. Schneider, K. H. Katz, R. W. Makuch, A. L. Reiss, et al., 2006. Increased temporal lobe gyrification in preterm children. Neuropsychologia 44:445–453. Kolb, B., R. Gibb, and G. Gorny, 2000. Cortical plasticity and the development of behavior after early frontal cortical injury. Dev. Neuropsychol. 18:423–444. Kolb, B., S. Pellis, and T. E. Robinson, 2004. Plasticity and functions of the orbital frontal cortex. Brain Cogn. 55:104–115. Konrad, K., S. Gauggel, A. Manz, and M. Scholl, 2000. Inhibitory control in children with traumatic brain injury (TBI) and children with attention deficit/hyperactivity disorder (ADHD). Brain Inj. 14:859–875. Landry, S. H., K. E. Smith, C. L. Miller-Loncar, and P. R. Swank, 1998. The relation of change in maternal interactive styles with infants’ developing social competence across the first three years of life. Child Dev. 69:105–123. Landry, S. H., P. R. Swank, K. Steubing, M. Prasad, and L. Ewing-Cobbs, 2004. Social competence in young children with inflicted traumatic brain injury. Dev. Neuropsychol. 26:707–733. Levin, H. S., D. Benavidez, K. Verger-Maestre, N. Perachio, J. Song, D. Mendelsohn, et al., 2000. Reduction of corpus callosum growth after severe traumatic brain injury in children. Neurology 54:647–653. Levin, H. S., K. A. Culhane, D. Mendelsohn, M. A. Lilly, D. A. Bruce, J. M. Fletcher, et al., 1993. Cognition in relation to magnetic resonance imaging in head-injured children and adolescents. Arch. Neurol. 50:897–905. Levin, H. S., D. Mendelsohn, M. A. Lilly, J. Yeakley, J. Song, R. S. Scheibel, et al., 1997. Magnetic resonance imaging in relation to functional outcome of pediatric closed head injury: A test of the Ommaya-Gennarelli model. Neurosurgery 40:432– 441. Levin, H. S., J. Song, L. Ewing-Cobbs, S. B. Chapman, and D. Mendelsohn, 2001. Word fluency in relation to severity of closed head injury, associated frontal brain lesions, and age at injury in children. Neuropsychologia 39:122–131. Levin, H. S., L. Zhang, M. Dennis, L. Ewing-Cobbs, R. Schachar, J. E. Max, et al., 2004. Psychosocial outcome of TBI in children with unilateral frontal lesions. J. Int. Neuropsychol. Soc. 10:305– 316. Linver, M. R., J. Brooks-Gunn, and D. E. Kohen, 2002. Family processes as pathways from income to young children’s development. Dev. Psychol. 38:719–734. Luciana, M., 2003. Cognitive development in children born preterm: Implications for theories of brain plasticity following early injury. Dev. Psychopathol. 15:1017–1047.

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

411

Luciana, M., L. Lindeke, M. Georgieff, M. Mills, and C. A. Nelson, 1999. Neurobehavioral evidence for working memory deficits in school-aged children with histories of prematurity. Dev. Med. Child Neurol. 41:521–533. Max, J. E., D. L. Schachar, H. S. Levin, L. Ewing-Cobbs, S. B. Chapman, M. Dennis, et al., 2005a. Predictors of attention deficit/hyperactivity disorder within 6 months after pediatric traumatic brain injury. J. Am. Acad. Child Adolesc. Psychiatry 44: 1032–1040. Max, J. E., D. L. Schachar, H. S. Levin, L. Ewing-Cobbs, S. B. Chapman, M. Dennis, et al., 2005b. Predictors of secondary attention deficit/hyperactivity disorder in children and adolescents 6 to 24 months after traumatic brain injury. J. Am. Acad. Child Adolesc. Psychiatry 44:1041–1049. McQuillen, P. S., and D. M. Ferriero, 2005. Perinatal subplate neuron injury: Implications for cortical development and plasticity. Brain Pathol. 15:250–260. Miller, S. P., D. B. Vigneron, R. G. Henry, M. A. Bohland, C. Ceppi-Cozzio, C. Hoffman, et al., 2002. Serial quantitative diffusion tensor MRI of the premature brain: Development in newborns with and without injury. J. Magn. Reson. Imaging 16: 621–632. Molfese, V. J., L. F. DiLalla, and D. Bunce, 1997. Prediction of the intelligence test scores of 3- to 8-year-old children by home environment, socioeconomic status, and biomedical risks. MerrillPalmer Q. 43:219–234. Neville, H., and D. Bavelier, 2002. Human brain plasticity: Evidence from sensory deprivation and altered language experience. Prog. Brain Res. 138:177–188. Nosarti, C., M. P. Allin, S. Frangou, L. Rifkin, and R. M. Murray, 2005. Hyperactivity in adolescents born very preterm is associated with decreased caudate volume. Biol. Psychiatry 57: 661–666. Nosarti, C., K. Rubia, A. B. Smith, S. Frearson, S. C. Williams, L. Rifkin, et al., 2006. Altered functional neuroanatomy of response inhibition in adolescent males who were born very preterm. Dev. Med. Child Neurol. 48:265–271. Nosarti, C., T. M. Rushe, P. W. Woodruff, A. L. Stewart, R. Rifkin, and R. M. Murray, 2004. Corpus callosum size and very preterm birth: Relationship to neuropsychological outcome. Brain 127:2080–2089. Olesen, P. J., Z. Nagy, H. Westerberg, and T. Klingberg, 2003. Combined analysis of DTI and fMRI data reveals a joint maturation of white and grey matter in a fronto-parietal network. Cogn. Brain Res. 18:48–57. Pascual-Leone, J., A. Amedi, F. Fregni, and L. B. Merabet, 2005. The plastic human brain cortex. Annu. Rev. Neurosci. 28:377–401. Peterson, B. S., A. W. Anderson, R. Ehrenkranz, L. H. Staib, M. Tageldin, E. Colson, et al., 2003. Regional brain volumes and their later neurodevelopmental correlates in term and preterm infants. Pediatrics 111:939–948. Peterson, B. S., B. Vohr, L. H. Staib, C. J. Cannistraci, A. Dolberg, K. C. Schneider, et al., 2000. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA 284:1939–1947. Peterson, J., H. G. Taylor, N. Minich, N. Klein, and M. Hack, 2006. Subnormal head circumference in very low birth weight children: Neonatal correlates and school-age consequences. Early Hum. Dev. 82:325–334. Rajapakse, J. C., J. N. Giedd, J. M. Rumsey, A. C. Vaituzis, S. D. Hamburger, and J. L. Rapoport, 1996. Regional MRI measurements of the corpus callosum: A methodological and developmental study. Brain Dev. 18:379–388.

412

neural plasticity in development

Rose, S. A., J. F. Feldman, and J. J. Janowski, 2001. Attention and recognition memory in the 1st year of life: A longitudinal study of preterm and full-term infants. Dev. Psychol. 37:135– 151. Rosenzweig, M. R., 2003. Effects of differential experience on the brain and behavior. Dev. Neuropsychol. 24:523–540. Rubia, K., A. B. Smith, J. Woolley, C. Nosarti, I. Heyman, E. Taylor, et al., 2006. Progressive increase of frontostriatal brain activation from childhood to adulthood during event-related tasks of cognitive control. Hum. Brain Mapping, published online 8 May 2006. Rugg-Gunn, F. J., M. R. Symms, G. J. Barker, J. S. Greenwood, and J. S. Duncan, 2001. Diffusion imaging shows abnormalities after blunt head trauma when conventional magnetic resonance imaging is normal. J. Neurol. Neurosurg. Psychiatry 70: 530–533. Schachar, R., H. S. Levin, J. E. Max, and K. Purvis, 2004. Attention deficit hyperactivity disorder and inhibition deficit after closed head injury in children: Do preinjury behavior, injury severity, and recovery variables predict outcome? Dev. Neuropsychol. 25:179–198. Serra-Grabulosa, J. M., C. Junque, K. Verger, P. SalgadoPineda, C. Maneru, and J. M. Mercader, 2005. Cerebral correlates of declarative memory dysfunctions in early traumatic brain injury. J. Neurol. Neurosurg. Psychiatry 76:129–131. Singleton, R. H., and J. T. Povlishock, 2004. Identification and characterization of heterogeneous neuronal injury and death in regions of diffuse brain injury: Evidence for multiple independent injury phenotypes. J. Neurosci. 24:3543–3553. Slomine, B. S., J. P. Gerring, M. A. Grados, R. Vasa, K. D. Brady, J. R. Christensen, et al., 2002. Performance on measures of executive function following pediatric traumatic brain injury. Brain Inj. 16:759–772. Smith, K. E., S. H. Landry, and P. R. Swank, 2006. The role of early maternal responsiveness in supporting school-aged cognitive development for children who vary in birth status. Pediatrics 117:1608–1617. Sowell, E. R., B. S. Peterson, P. M. Thompson, S. E. Welcome, A. L. Henkenius, and A. W. Toga, 2003. Mapping cortical change across the human life span. Nature Neurosci. 6:309– 315. Stiles, J., P. Moses, K. Roe, N. A. Akshoomoff, D. Trauner, J. Hesselink, et al., 2003. Alternative brain organization after prenatal cerebral injury: Convergent fMRI and cognitive data. J. Int. Neuropsychol. Soc. 9:604–622. Stiles, J., J. S. Reilly, B. Paul, and P. Moses, 2005. Cognitive development following early brain injury: Evidence for neural adaptation. Trends Cogn. Sci. 9:136–143. Sun, S.-W., H.-F. Liang, K. Trinkaus, A. H. Cross, R. C. Armstrong, and S.-K. Song, 2006. Noninvasive detection of cuprizone induced axonal damage and demyelination in the mouse corpus callosum. Magn. Reson. Med. 55:302–308. Tasker, R. C., C. H. Salmond, A. G. Westland, A. Pena, J. H. Gillard, B. J. Sahakian, et al., 2005. Head circumference and brain and hippocampal volume after severe traumatic brain injury in childhood. Pediatr. Res. 58:302–308. Taylor, H. G., and J. Alden, 1997. Age-related differences in outcomes following childhood brain insults: An introduction and overview. J. Int. Neuropsychol. Soc. 3:555–567. Taylor, H. G., N. Minich, B. Bangert, P. A. Filipek, and M. Hack, 2004. Long-term neuropsychological outcomes of very low birth weight: Associations with early risks for periventricular brain insults. J. Int. Neuropsychol. Soc. 10:987–1004.

Taylor, H. G., N. Minich, N. Klein, and M. Hack, 2004b. Longitudinal outcomes of very low birth weight: Neuropsychological findings. J. Int. Neuropsychol. Soc. 10:149– 163. Taylor, H. G., K. O. Yeates, S. L. Wade, D. Drotar, and S. Klein, 1999. Influences on first-year recovery from traumatic brain injury in children. Neuropsychology 13:76–89. Taylor, H. G., K. O. Yeates, S. L. Wade, D. Drotar, T. Stancin, and C. Burant, 2001. Bidirectional child-family influences on outcomes of traumatic brain injury in children. J. Int. Neuropsychol. Soc. 7:755–767. Taylor, H. G., K. O. Yeates, S. L. Wade, D. Drotar, T. Stancin, N. Minich, 2002. A prospective study of short-and long-term outcomes after traumatic brain injury in children: Behavior and achievement. Neuropsychology 16:15–27. Thompson, P. M., J. N. Giedd, R. P. Woods, D. MacDonald, A. C. Evans, and A. W. Toga, 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 404:190–193. Toga, A. W., P. M. Thompson, and E. R. Sowell, 2006. Mapping brain maturation. Trends Neurosci. 29:148–159. Vargha-Khadem, F., C. H. Salmond, K. E. Watkins, K. J. Friston, D. G. Gadian, and M. Mishkin, 2003. Developmental amnesia: Effect of age at injury. Proc. Natl. Acad. Sci. USA 100:10055–10060. Verger, K., C. Junque, H. S. Levin, M. A. Jurado, M. PerezGomez, D. Bartres-Faz, et al., 2001. Correlation of atrophy measures on MRI with neuropsychological sequelae in children and adolescents with traumatic brain injury. Brain Inj. 15:211– 221.

Volpe, J. J., 2005. Encephalopathy of prematurity includes neuronal abnormalities. Pediatrics 116:221–225. Wilde, E.A., E. D. Bigler, J. M. Haider, Z. Chu, H. S. Levin, X. Li, et al., 2006. Vulnerability of the anterior commissure in moderate to severe pediatric traumatic brain injury. J. Child Neurol. 21:769–776. Wilde, E. A., Z. Chu, E. D. Bigler, J. V. Hunter, M. A. Fearing, G. Hanten, et al., 2006. Diffusion tensor imaging in the corpus callosum in children after moderate to severe traumatic brain injury. J. Neurotrauma 23:1412–1423. Wilde, E. A., J. V. Hunter, M. R. Newsome, R. S. Scheibel, E. D. Bigler, J. L. Johnson, et al., 2005. Frontal and temporal morphometric findings on MRI in children after moderate to severe traumatic brain injury. J. Neurotrauma 22:333–344. Woodward, L. J., J. O. Edgin, D. Thompson, and T. E. Inder, 2005. Object working memory deficits predicted by early brain injury and development in the preterm infant. Brain 128:2578– 2587. Yeates, K. O., K. Armstrong, J. Janusz, H. G. Taylor, S. Wade, T. Stancin, et al., 2005. Long-term attention problems in children with traumatic brain injury. J. Am. Acad. Child Adolesc. Psychiatry 44:574–584. Yeates, K. O., H. G. Taylor, D. Drotar, S. L. Wade, S. Klein, T. Stancin, et al., 1997. Preinjury family environment as a determinant of recovery from traumatic brain injuries in schoolage children. J. Int. Neuropsychol. Soc. 3:617–630. Yeates, K. O., H. G. Taylor, S. L. Wade, D. Drotar, T. Stancin, and N. Minich, 2002. A prospective study of short- and longterm neuropsychological outcomes after traumatic brain injury in children. Neuropsychology 16:514–523.

ewing-cobbs, prasad, and hasan: plasticity and function following brain injury

413

25

Plasticity of the Visual System DAPHNE MAURER, TERRI L. LEWIS, AND CATHERINE J. MONDLOCH

Newborns can see but there are serious limitations on their visual perception. Rapid improvements occur during infancy, but some aspects of visual perception continue to improve even into adolescence. In this chapter we describe briefly the normal developmental trajectory and then discuss its plasticity, specifically how it is altered by visual experience. The main evidence comes from our studies of children who missed visual experience during some period of development because of dense, central cataracts in one or both eyes. Supplementary evidence comes from studies of the reorganization of the visual pathway in the congenitally blind, from studies of the effect of variations in the details of visual experience in children with normal eyes, and from studies of residual visual plasticity in adulthood. We end the chapter with a discussion of the general principles about visual plasticity that can be deduced from the findings. Newborns’ vision is severely limited. They respond only to large objects of high contrast that are located near the central visual field. Their visual acuity is 40 times worse than that of an adult with normal eyes, and their sensitivity to contrast is 50 times worse (Atkinson, Braddick, and Moar, 1977; Banks and Salapatek, 1978; Brown and Yamamoto, 1986; Courage and Adams, 1990; van Hof–van Duin and Mohn, 1986; reviewed in Maurer and Lewis, 2001a, 2001b). Not surprisingly, they usually fixate the bold external contour of objects, including faces, rather than the smaller internal details (Bronson, 1990; Hainline, 1978; Haith, Bergman, and Moore, 1977; Maurer, 1983; Maurer and Salapatek, 1976; Salapatek, 1968; Salapatek and Kessen, 1966, 1973), and they respond to stimuli as a collection of unrelated elements, rather than as an integrated, global form (Cohen and Younger, 1984). Their visual fields are also restricted: even with large, bright stimuli, they do not respond to targets farther than 30° in the periphery (Lewis and Maurer, 1992). They are apparently insensitive to motion (Wattam-Bell, 1991; reviewed in Braddick, 1993) and stereoscopic cues to depth (reviewed in Birch, 1993). Despite these visual limitations, newborns can discriminate between the internal features of two faces if the external contour is occluded (Turati et al., 2006), look longer at attractive than at unattractive faces—based on either the external contour or internal features (Slater et al., 2000), are attracted toward ovals with more elements in the top half, such as human faces (Macchi Cassia, Turati, and Simion, 2004; Simion et al., 2001, 2006, 2002; reviewed in Turati, 2004), and look longer at faces

with direct than averted gaze (Farroni et al., 2002; Farroni, Menon, and Johnson, 2006). They also show some sensitivity to global form in nonface stimuli (e.g., Farroni et al., 2000). For example, after being habituated to an array of local elements consisting of horizontal rows of small black squares alternating with horizontal rows of small white squares, newborns treat vertical black-and-white stripes as novel and horizontal black-and-white stripes as familiar (Farroni et al., 2000). Presumably, they integrate the elements with the same luminance into a global percept of horizontal stripes, as predicted by the Gestalt principle of luminance similarity (Farroni et al., 2000). Over the next 6 months, there are dramatic improvements in infants’ vision. Acuity improves fivefold, there is some improvement in contrast sensitivity, and visual fields expand to adultlike proportions for large bright stimuli (reviewed in Maurer and Lewis, 1998, 2001a, 2001b). Sensitivity to motion and stereoscopic depth cues typically emerges around 10–20 weeks of age (Birch, 1993; Braddick, 1993; Wattam-Bell, 1991), and stereoacuity improves rapidly to near adult values by 6 months of age (Birch, 1993). At 2–3 months, babies begin to scan visual objects extensively, concentrating on the internal features and, in the case of faces, the eyes (Bronson, 1990; Hainline, 1978; Haith, Bergman, and Moore, 1977; Maurer, 1983; Maurer and Salapatek, 1976; Salapatek, 1975). By 3 months, they show evidence of forming face prototypes, representing the average value of faces recently seen and comparing novel faces to it (de Haan et al., 2001). By 4 months, they show increased sensitivity to the global configuration of stimuli. For example, they treat a face composed of the external contour of one familiar face and the internal features of another familiar face as completely novel (Cashon and Cohen, 2003, 2004). Despite the advances during infancy, visual development continues for many years. Grating acuity and contrast sensitivity do not reach adult levels until 4–7 years of age (Ellemberg, Lewis, Liu, and Maurer, 1999; Mayer and Dobson, 1982; but see Gwiazda et al., 1997). Sensitivity to pinpoints of light in the periphery continues to improve until 7 years of age (Bowering, 1992; Bowering et al., 1996; reviewed in Maurer and Lewis, 1998), stereoacuity does so until 5–9 years of age (Romano, Romano, and Puklin, 1975; Tomac and Altay, 2000), and for some parameters, sensitivity to local motion improves until sometime after 5 years of age (Ellemberg et al., 2003). Integrative processing at higher

415

levels of the visual system also shows a long developmental trajectory, with the details dependent on the aspect of vision and the parameters tested. For example, at least under some conditions, the integration of local motion cues into a global direction of motion is adultlike as early as 3 years of age (Parrish et al., 2005), but sensitivity to biological motion does not reach adult levels until about 9 years of age (Freire et al., 2006). The integration of small oriented elements into a global contour or form reaches adult levels at about 9 years of age on some measures (Lewis et al., 2002), but not until after 14 years of age on other measures (Kovács, 2000). Some aspects of face processing are also slow to develop: holistic processing (processing the facial features as a gestalt) and identifying faces based on local features or the shape of the external contour are adultlike, or nearly so, by 4–6 years of age, but identifying faces from different points of view or based on the spacing of internal features continues to improve into adolescence (Carey and Diamond, 1994; de Heering, Houthuys, and Rossion, 2007; Mondloch et al., 2003; Mondloch, Le Grand, and Maurer, 2002, 2003; Mondloch et al., 2007; Pellicano and Rhodes, 2003; Tanaka et al., 1998). In the rest of this chapter, we discuss the plasticity of visual development: how the developmental trajectory just described is altered by visual experience and how plasticity varies with age. We begin with the evidence from our studies of children who suffered a period of visual deprivation because of dense, central cataracts in one or both eyes. A cataract is an opacity in the lens of the eye. When, as in the children we selected for study, the cataract is large and dense, it allows only diffuse light to reach the retina. Pattern deprivation continues until the defective natural lens of the eye is removed surgically and the eye is fitted with a compensatory optical correction (e.g., a contact lens). By comparing the visual development of children treated for cataract to that of children with normal eyes, we have been able to deduce the role of patterned visual input in driving normal visual development. By comparing children with the onset of deprivation from cataracts at different ages, we have been able to deduce the sensitive periods during which visual experience tunes the development of different visual capabilities. By comparing children treated for bilateral and unilateral cataract with the same age of onset and similar duration of deprivation, we have been able to deduce the effect of deprivation per se (shared by the bilaterally and unilaterally deprived eyes) versus the additional adverse effects of uneven competition between the eyes for cortical connections (experienced only by unilaterally deprived eyes). Our results indicate that the extent of visual plasticity and the parameters of its timing vary for different aspects of vision (reviewed in Maurer, Lewis, and Mondloch, 2005). We will begin with the findings for low-level vision, that is, visual capabilities mediated mainly by the pathway from the

416

neural plasticity in development

retina through the lateral geniculate nucleus to the primary visual cortex: acuity, contrast sensitivity, peripheral vision, and sensitivity to local motion. We will then consider the findings for higher-level vision, that is, visual capabilities requiring integration of local information by extrastriate cortical areas: global motion, global form, face processing, and visual attention. Finally, we will make comparisons across visual domains to evaluate general principles about human visual plasticity.

Low-level vision Acuity Immediately after treatment for congenital cataract in one or both eyes, grating acuity is no better than that of a visually normal newborn, even if treatment is as late as 9 months of age (Maurer et al., 1999). Because grating acuity improves so rapidly during the first 6 months of life in visually normal children (Mayer et al., 1995; reviewed in Maurer and Lewis, 2001a, 2001b), the later the treatment, the larger the initial deficit. However, grating acuity improves rapidly after the onset of visual experience: after only one hour of patterned visual input, grating acuity improves as much as it does during the entire first month of life in visually normal infants. Grating acuity continues to develop at such a fast pace that by 1 year of age, patients’ acuity falls within the normal range (Lewis, Maurer, and Brent, 1995; Maurer et al., 1999). Thereafter, improvements fail to keep pace with normal development, and acuity falls outside the normal range by 2 years of age and remains abnormal thereafter, as the grating acuity of visually normal children improves to near adult functional levels by age 5 (Lewis, Maurer, and Brent, 1995). The left panel of figure 25.1 shows the grating acuity of eight patients treated for bilateral congenital cataract who had been deprived for the first 3–8 months of life and who, at the time of the test, ranged in age from 5 to 18 years. Relative to comparably aged controls, all 16 eyes of these patients had reduced grating acuity, averaging 3.5 times (1.8 octaves) worse than normal (Ellemberg et al., 2002). The deficits were not related to the age of treatment, at least when treatment occurred between 3 and 8 months of age. Overall, deficits in grating acuity for deprived eyes are larger after monocular deprivation than after binocular deprivation, but these differences disappear if, in unilateral cases, the good eye is patched regularly throughout early childhood (Ellemberg et al., 2002, 2000; Ellemberg, Lewis, Maurer et al., 1999; Lewis, Maurer, and Brent, 1995). Evidence for permanent deficits in asymptotic grating acuity also have been reported in other human cohorts (Birch et al., 1998; Mioche and Perenin, 1986) and in form-deprived monkeys (Harwerth et al., 1989, 1991). The results for later linear letter acuity mirror those for grating acuity: deficits in every patient are larger after mon-

a period of consolidation after acuity has reached adult functional levels (Lewis and Maurer, 2005).

Figure 25.1 Grating acuity of patients treated for bilateral congenital cataract (first panel, n = 8), unilateral congenital cataract (second panel, n = 14), bilateral developmental cataract (third panel, n = 6), or unilateral developmental cataract (fourth panel, n = 9). Circles represent the data from the better eyes of bilateral cases (determined from clinical history of alignment and Snellen acuity) and the nondeprived eyes of unilateral cases. Triangles represent the data from the worse eyes of bilateral cases and the deprived eyes of unilateral cases. Bilateral cases with equal alignment and acuity histories for the two eyes had one eye assigned randomly to each category. The dashed line represents the mean of 14 comparably aged participants with normal vision. (Reprinted from Vision Research, vol. 42, D. Ellemberg, T. L. Lewis, D. Maurer, S. Brar, and H. P. Brent, “Better perception of global motion after monocular than after binocular deprivation,” p. 175. Copyright 2002, with permission from Elsevier.)

ocular than after binocular deprivation unless, after monocular deprivation, uneven competition between a deprived eye and a good eye is reduced by aggressive patching of the good eye (Birch et al., 1998; Lundvall and Kugelberg, 2002a, 2002b; Magnusson, Abrahamsson, and Sjöstrand, 2002; reviewed in Maurer and Lewis, 1993, and in Maurer, Lewis, and Brent, 1989). The only exceptions are a few isolated reports of patients who achieved normal 20/20 acuity after early treatment for congenital cataract: 2 cases treated at 6–8 days of age for bilateral congenital cataract (Kugelberg, 1992), 1 of 13 cases treated before 7 weeks of age for bilateral congenital cataract (Magnusson, Abrahamsson, and Sjöstrand, 2002), and 2 of 16 cases treated by 6 weeks of age for unilateral congenital cataract with subsequent aggressive patching of the good eye (Birch and Stager, 1996). Studies of children who developed cataracts postnatally indicate that visual deprivation lasting more than a few days at any time during the first 10 years of life prevents the development of normal acuity (reviewed in Lewis and Maurer, 2005). Thus, even after acuity has achieved adult values (around 4–6 years of age), deprivation can still cause permanent damage, a result suggesting that visual experience is important during

Contrast Sensitivity When tested first after age 4, children treated for congenital cataract have nearly normal contrast sensitivity at low spatial frequencies (wide stripes), and their sensitivity improves at normal or faster-thannormal rates, so that any small deficit remains constant or decreases (Maurer, Ellemberg, and Lewis, 2006, in preparation; see also Birch et al., 1998; Birch and Stager, 1996; Birch et al., 1993; Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2000; Tytla et al., 1988; but see Mioche and Perenin, 1986). They also have normal contrast sensitivity at high temporal frequencies (fast rates of flicker) (Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2000; Tytla et al., 1988; but see Mioche and Perenin, 1986). However, the developmental trajectory is different for the parts of spatial and temporal contrast sensitivity that are especially immature during infancy (Maurer and Lewis, 2001a; Regal, 1981). Children treated for congenital cataract later fail to see high spatial frequencies even at maximum contrast. They can see mid spatial frequencies (medium stripe widths) and low temporal frequencies (slow flicker rates) but their sensitivity is poor: they require about 30 times more contrast than normal to detect stripes of 5 cycles per degree and about 5 times more contrast than normal to detect 5-Hz flicker (Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2000). The contrast sensitivity of visually normal children increases twofold for mid and high spatial frequencies between 5 and 7 years of age, whereas patients’ spatial contrast sensitivity in this range of spatial frequencies does not improve after 5 years of age (Ellemberg, Lewis, Liu, and Maurer, 1999; Maurer, Ellemberg, and Lewis, 2006, in preparation; reviewed in Maurer and Lewis, 2001a, 2001b). Thus patients’ deficits become larger and larger, leaving the typical patient with the spatial contrast sensitivity of a visually normal toddler (Gwiazda et al., 1997). (Similar longitudinal studies of recovery after treatment have not been performed for temporal contrast sensitivity.) As for acuity, deficits for spatial and temporal contrast sensitivity are worse after monocular than after binocular deprivation, unless the good eye was patched aggressively after monocular deprivation (Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2000; Tytla et al., 1988), and there is some evidence for a loss of sensitivity to mid spatial frequencies after age 5, perhaps as a result of the tapering of patching of the nondeprived eye (Maurer, Ellemberg, and Lewis, in preparation; see also Murphy and Mitchell, 1987). Peripheral Vision Children treated for congenital cataract later have severely restricted fields, even when the deprivation began as late as 6 years of age (Bowering 1992; Bowering et al., 1996; Bowering et al., 1997; reviewed in Maurer and

maurer, lewis, and mondloch: plasticity of the visual system

417

Lewis, 1998). The losses are largest in the temporal field, the part of the field that is slowest to reach a full adult extent for a pinpoint of light (although, as we will note, sensitivity in the near temporal field develops relatively quickly). Figure 25.2 shows the losses for various parts of the visual field and also illustrates the larger losses in eyes that were deprived for more than 6 months than in eyes that were deprived for a shorter time. The difference was significant in all parts of the field except the superior field, the part of the field that is first to reach an adult size. For the dimmer light, the losses in the deprived eye were greater after monocular deprivation than after binocular deprivation of comparable duration. Thus visual deprivation interferes with the normal development of the edges of the field, with the largest effect on the part of the field that is slowest to develop. The large restrictions after binocular deprivation indicate that visual deprivation interferes with the normal development of peripheral vision. The fact that the restrictions were even greater after monocular deprivation indicates that unfair competition between the eyes can have an additional adverse effect. That conclusion is bolstered by the finding

Mean field restriction ± 1 S.E. (°)

50

40

30

20

10

0 0

45

90

135 180 225 270 315

Meridian (°)

Figure 25.2 Mean field restrictions and standard errors along eight meridia in children treated for dense and central cataracts. The temporal field includes 0° and 315°; the superior field, 45°, 90°, and 135°; the nasal field, 180°; and the inferior field, 225° and 270°. Data are for children with less than 6 months deprivation tested with a pinpoint of light (0.11° in diameter) with a luminance of 31.8 cd/m2 (filled squares) and of 318 cd/m2 (open squares) and for children with more than 6 months deprivation tested with the less luminant (filled circles) and more luminant light (open circles). (Adapted from Journal of Pediatric Ophthalmology and Strabismus, vol. 34, E. R. Bowering, D. Maurer, T. L. Lewis, and H. P. Brent, “Constriction of the visual field of children after early visual deprivation,” p. 351. Copyright 1997, with permission from SLACK Incorporated, 6900 Grove Road, Thorofare, NJ 08086.)

418

neural plasticity in development

that the restrictions were smaller when there had been extensive patching of the fellow eye, but only in the far temporal field, the part of the field that is slowest to develop (Bowering, 1992; Bowering et al., 1996, 1997; reviewed in Maurer and Lewis, 1998). (Mean restrictions of the edge of the field temporally were 24°, 37°, and 35° for good, fair, and poor patchers, respectively.) Together, the results suggest that different aspects of peripheral vision develop at different rates and that the more slowly developing aspects are most dependent on normal visual input during infancy. Studies of peripheral sensitivity in other parts of the visual field indicate that visual deprivation also interferes with sensitivity to targets anywhere in the visual field, not just its edges. With pinpoints of light, there are large reductions throughout the field even when the deprivation began in the early teenage years, so late that acuity and contrast sensitivity are normal, showing no adverse effect of the deprivation (reviewed in Lewis and Maurer, 2005). Sensitivity to the contrast of sine wave gratings is depressed in children treated for congenital cataracts, not only in the center of the field, but all along the horizontal meridian (Maurer and Lewis, 1993; Mioche and Perenin, 1986; Tytla et al., 1991). In every case, the contrast had to be higher for the patient to detect the grating than for an age-matched normal control, with a larger difference at higher spatial frequencies. For example, in five of six patients treated for bilateral congenital cataract, we found that, at every eccentricity tested, the patient required 1–1.5 log units more contrast than normal to see the highest spatial frequency the eye could resolve (Tytla et al., 1991). Thus deprivation from birth interferes with sensitivity to targets throughout the visual field. In patients treated for unilateral congenital cataract, there are also losses throughout the entire visual field, not just at the edges, but the extent of the loss is larger in the near nasal field (e.g., left visual field when looking with the right eye) than in the near temporal field (e.g., right visual field when looking with the right eye). This is true both for tests of the contrast necessary to see a sine wave grating (Tytla et al., 1991) and in a comparison of threshold sensitivity when a light is presented in the nasal field (at 20°) versus the temporal field (at 30°) (Bowering et al., 1993) (see figure 25.3). These results parallel our findings during infancy: there is slower growth of the near nasal field than the near temporal field, and initially lower sensitivity for targets at 20° in the nasal field than for targets at 30° in the temporal field (Lewis and Maurer, 1992; Lewis, Maurer, and Blackburn, 1985). Like the results for the far edges of the temporal field, they suggest that slowly developing parts of peripheral vision are most affected by visual deprivation or unfair competition between the eyes. Moreover, the development of peripheral vision depends on visual experience until at least the teenage years.

15

Bilateral congenital Unilateral congenital Control

Threshold (Mean ± 1 S.E.)

Loss Scores (db) + 1 S.E.

0.1

10

5

0 1.5

6

12

Velocity (deg/sec)

0 20° Nasal Field

30° Temporal Field

Figure 25.3 Mean losses and standard errors in peripheral light sensitivity for patients relative to age-matched visually normal controls tested at 20° in the nasal visual field and at 30° in the temporal visual field. Data are for children treated for unilateral congenital cataract (filled bars), unilateral noncongenital cataract (vertically striped bars), bilateral congenital cataract (open bars), and bilateral noncongenital cataract (horizontally striped bars). Although all groups showed reduced light sensitivity, only the children treated for unilateral congenital cataract showed significantly greater losses at 20° in the nasal field than at 30° in the temporal field. (Reprinted with permission from figure 2 in E. R. Bowering, D. Maurer, T. L. Lewis, and H. P. Brent (1993), “Sensitivity in the nasal and temporal hemifields in children treated for cataract,” Investigative Ophthalmology and Visual Science, 34:3506, © Association for Research in Vision and Ophthalmology (ARVO).)

Sensitivity to Local Motion Children treated for congenital cataract later show deficits not only in low-level spatial vision but also in low-level motion perception. When tested after age 5 with sine wave gratings of 1 cycle per degree moving at slower (1.5° s1), medium (6° s1), and faster (12° s1) velocities, children who had been treated for congenital cataract needed more contrast than normal to perceive accurately whether the stripes were moving up or down (Ellemberg et al., 2005). The deficits were specific to motion perception and not to difficulties in seeing the spatial structure of the patterns: when asked to determine the orientation (horizontal or vertical) of similar stationary patterns, the patients had normal contrast thresholds, consistent with their normal contrast sensitivity at low spatial frequencies. As shown in figure 25.4, the deficits were comparable after unilateral and after bilateral deprivation, likely because the unilaterally deprived cases had patched the good eye sufficiently to reduce any additional deleterious effects of uneven competition between the eyes. For both unilaterally and bilaterally deprived patients, the deficits decreased as velocity increased, from patients needing 6 percent more

Figure 25.4 Threshold contrast for correctly identifying the direction of local motion in patients treated for bilateral congenital cataract (black bars), for unilateral congenital cataract (white bars), and age-matched controls (shaded bars). Patient’s deficit was largest at the slowest velocity.

amplitude modulation than controls at the slowest velocity (1.5° s−1) to needing only 1 percent more amplitude modulation at the fastest velocity (12° s−1) (see figure 25.4). In contrast, for controls, thresholds did not change with velocity. The pattern of deficits appears to be related to the normal pattern of development: losses were greater at the slower velocities than at the faster velocities, and sensitivity to slower velocities develops more slowly during infancy (Aslin and Shea, 1990; Bertenthal and Bradbury, 1992; Freedland and Dannemiller, 1987; Kaufmann, 1995; Roessler and Dannemiller, 1997; Volkmann and Dobson, 1976). Thus our findings of greater losses at slower than at faster velocities are consistent with the hypothesis that aspects of vision that are slower to mature are especially compromised by early visual deprivation. Summary for Low-Level Vision In summary, three patterns of results seem to emerge for low-level vision. First, most aspects of low-level vision are compromised after early visual deprivation, even when it was limited to the first 1–2 months of life and even when the capability takes many years to develop in children with normal eyes. Second, aspects of vision that are slower to mature are especially compromised by early visual deprivation. Sensitivity to high spatial frequencies and to low temporal frequencies is relatively slow to mature, as are sensitivity in the far temporal visual field and in the near nasal visual field and sensitivity to slower velocities of local motion—and it is these aspects of vision that are especially affected by early visual deprivation. Each of these involves the processing of local detail and depends on the tuning of neurons at lower levels of the visual cortex (V1/V2) based on input from the retina and lateral geniculate nucleus and feedback from higher cortical areas.

maurer, lewis, and mondloch: plasticity of the visual system

419

For example, neurons in V1 of the primate are highly selective and/or narrowly tuned for spatial frequency, direction of motion, spatial position, binocular disparity, and orientation (Hubel and Wiesel, 1962, 1968). Third, vision can be as good after monocular deprivation as after binocular deprivation if the good eye is patched aggressively throughout early childhood. In the absence of such patching, the deficit is usually greater. These patterns do not hold for higher-level vision, aspects of vision that require the integration of local information by networks of neural structures extending into extrastriate cortical areas. In the next section, we discuss four examples— global motion, global form, face processing, and visual attention—and show not only that the patterns for higherlevel vision are different from those for lower-level vision, but also that the patterns vary across different higher-level functions.

milliseconds (“limited-lifetime dots”) so that the predominant direction of motion cannot be determined by looking at any one dot. A typical measure of sensitivity, called the coherence threshold, is the minimum percentage of the dots needed to move in the same direction, among randomly moving dots, for the subject to accurately perceive that predominant direction of motion. To measure sensitivity to global motion after visual deprivation, we measured coherence thresholds in the same group of patients whose grating acuity was described previously and in figure 25.1, when patients were at least 6 years old. The first panel in figure 25.5 shows the coherence thresholds for each eye of the eight patients treated for bilateral congenital cataract. All 16 eyes of these patients had abnormal coherence thresholds, averaging 5 times worse than normal (Ellemberg et al., 2002). Surprisingly, coherence thresholds were worse after binocular deprivation than after monocular deprivation: monocularly deprived patients showed deficits, but their coherence thresholds were only 1.6 times worse than normal, three times better than those of binocularly deprived patients (see figure 25.5). The deficits were not related to the amount of time that the good eye had been patched after monocular deprivation or to age of treatment in monocularly or binocularly deprived cases. Moreover, in unilateral cases, the deficits were as pronounced in the good eye as in the deprived eye. The better performance after monocular than after binocular deprivation and the compa-

Higher-level vision Sensitivity to Global Motion The perception of global motion requires the integration of local cues into a global percept specifying the overall direction of motion. It is typically tested with random-dot kinematograms in which dots move in random direction, except for a proportion of signal dots moving in the same direction. To prevent the use of local cues, individual dots are replaced every few

Global motion

Clearance threshold (%)

80

Better eye Worse eye

60

40

20

0 Bilateral

Unilateral

Congenital

Figure 25.5 Global coherence thresholds of patients treated for bilateral congenital cataract (first panel, n = 8), unilateral congenital cataract (second panel, n = 14), bilateral developmental cataract (third panel, n = 6), or unilateral developmental cataract (fourth panel, n = 9). Symbols as in figure 25.1, where the acuity of the

420

neural plasticity in development

Bilateral

Unilateral

Developmental

same patients is shown. (Reprinted from Vision Research, vol. 42, D. Ellemberg, T. L. Lewis, D. Maurer, S. Brar, and H. P. Brent. “Better perception of global motion after monocular than after binocular deprivation,” p. 174. Copyright 2002, with permission from Elsevier.)

rable deficits in both eyes suggest that the deficits originated in higher areas of the visual pathway where there is convergence of inputs from the two eyes across large areas of the visual field that might allow a relative sparing of function after monocular deprivation. The pathway likely involves the middle temporal cortex in the dorsal visual stream (Maunsell and Newsome, 1987; Newsome and Pare, 1988; reviewed in Ellemberg et al., 2002). A second surprising finding was the short period during development when abnormal visual experience caused later deficits in sensitivity to global motion. When we tested coherence thresholds in patients who developed cataracts in one or both eyes between the ages of 4 months and 15 years, we were surprised to find that, for every single deprived eye, coherence thresholds were normal, even when the onset of deprivation was as early as 4 months of age (see figure 25.5). Yet every one of these eyes had abnormal grating acuity (see figure 25.1). Combined, our results indicate that the sensitive period for damage to the system mediating sensitivity to global motion is very short, ending during early infancy. These results contradict the traditional belief that sensitive periods are longer at higher than at lower levels of the visual system (Daw, 2003). Rather, the sensitive period for the damage of global motion appears to be very short, much shorter than that for grating acuity (which lasts until at least 5 years of age), for Snellen acuity (which lasts until about 10 years of age), and for peripheral vision (which lasts until at least the early teenage years). The conclusion of a very short sensitive period for global motion is supported by findings that an individual (MM) who had had deprivation lasting almost 40 years beginning at age 3 had poor acuity and contrast sensitivity but normal sensitivity to global motion (Fine et al., 2003). Interestingly, for global motion, the sensitive period for damage seems to end at an age when sensitivity to global motion is still very immature in visually normal children (about 7 times worse than in adults at 4 months of age—Wattam-Bell, 1994) and long before the age of functional maturity (between 3 and 6 years of age— Parrish et al., 2005; Ellemberg et al., 2002). Together, the findings of a very short sensitive period and of better sensitivity after monocular than after binocular deprivation indicate that, beyond the primary visual cortex, competitive interactions between the eyes can give way to collaborative interactions that enable a relative sparing of visual function after monocular deprivation or after postnatal deprivation. Sensitivity to Global Form To investigate whether the pattern of deficit observed for global motion (involving primarily the extrastriate dorsal stream) can be generalized to perception of global form (involving primarily the extrastriate ventral stream), we tested sensitivity to “Glass” patterns (Glass, 1969) like those shown in figure 25.6a. The pattern on the left of figure 25.6a has 100 percent signal: all

Figure 25.6 Test for ability to integrate elements into a global percept of form. (a) Examples of the patterns used. In the pattern on the left, 100 percent of the dots are paired to form a global swirl, whereas in the pattern on the right, there are 50 percent paired dots and 50 percent randomly paired noise dots. Threshold was calculated as the maximum number of noise dots that could be tolerated and still allow the subject to accurately discriminate a stimulus containing some paired signal dots from a stimulus comprised entirely of noise dots. (b) Mean threshold for patients treated for bilateral congenital cataract, patients treated for unilateral congenital cataract, and age-matched controls. (Reprinted from Trends in Cognitive Sciences, vol. 9, D. Maurer, T. L. Lewis, and C. Mondloch, “Missing sights: Consequences for visual cognitive development,” p. 148. Copyright 2005, with permission from Elsevier.)

dots are arranged in pairs such that the orientation of the pair is always tangent to a circle centered on the pattern. The pattern on the right has 50 percent signal: the global form was degraded by replacing 50 percent of the signal dot pairs with an equal number of randomly spaced noise dots that were the same size and shape as the signal dots. Thresholds for detecting global structure were defined as the minimum percent signal necessary to accurately discriminate signal from noise patterns. We found that children treated for congenital cataract have deficits in integrating local elements into a global percept of form, a capability that depends on an extrastriate network including area V4v in the ventral visual stream (reviewed in H. Wilson, 1999). For global form, as for global

maurer, lewis, and mondloch: plasticity of the visual system

421

motion, thresholds in the deprived eyes were significantly worse after binocular deprivation than after monocular deprivation of comparable duration, even when there was little patching of the nondeprived eye (see figure 25.6b). However, the deficits for global form are much smaller than those for global motion, despite the fact that sensitivity to global form matures much later (cf. Ellemberg et al., 2002; Lewis et al., 2004; Parrish et al., 2005) and we had predicted that aspects of vision that develop later would be more affected by early visual deprivation than those developing earlier, as is true for low-level vision (see the earlier section “Low-level vision”; Maurer and Lewis, 1993). Together, the results for global form and global motion suggest that the effect of early visual input on high-level vision involving networks in the extrastriate cortex is quite different from its effects on low-level vision involving the primary visual cortex. Jeffrey, Wang, and Birch (2004) also reported deficits in global form perception after treatment for congenital cataract but tested only two bilaterally deprived patients and therefore could not do a systematic comparison of the effects of binocular versus monocular deprivation. Face Perception The ability to recognize individual faces is a highly specialized skill that emerges during infancy, continues to develop throughout childhood, and becomes adultlike in late adolescence. Adults have the remarkable ability to detect faces, even in the absence of normal facial features. They readily detect faces in paintings in which faces are composed of objects such as an arrangement of fruit, vegetables, or rocks (Moscovitch, Winocur, and Behrman, 1997) or when presented with a two-tone Mooney face (Kanwisher, Tong, and Nakayama, 1998), at least when the stimuli are upright. Face detection is facilitated by the fact that all faces share the same first-order relations (two eyes above a nose, which is above a mouth). As a result, face images can be superimposed, or averaged, and the resulting stimulus remains recognizably facelike (Diamond and Carey, 1986). While adults are proficient at face detection, they recognize faces less often at this basic level (e.g., “That’s a face”) and more often at the subordinate level (e.g., “That’s Wayne Gretzky”), and can do so rapidly and accurately (Tanaka, 2001), even when the person is at a distance, is in poor lighting, has a new hairdo, or is a former schoolmate who has not been seen for more than 20 years (Bahrick, Bahrick, and Wittlinger, 1975). Several mechanisms underlie this expert face recognition: adults process faces holistically (i.e., as a gestalt), and they are sensitive to subtle differences among faces that indicate facial identity—the shape of the external contour, the shape of individual features, and the spacing among features. Face detection. Newborns orient preferentially toward a head-shaped stimulus containing three squares arranged as

422

neural plasticity in development

facial features, config, when it is paired with the same stimulus with the arrangement of squares inverted (Goren, Sarty, and Wu, 1975; Johnson et al., 1991; Mondloch et al., 1999; Valenza et al., 1996). This preference disappears by 6 weeks of age, but by 12 weeks of age a new preference has emerged (Dannemiller and Stephens, 1988; Mondloch et al., 1999): 12-week-old infants look longer at a positive-contrast face than at a phase-reversed version. These postnatal changes are driven by experience. We tested five infants treated for bilateral congenital cataract within two hours of their first receiving visual input; all these babies were at least 6 weeks old at the time of treatment (i.e., were at an age where no normal control has oriented preferentially toward config), and three of these babies were at least 12 weeks old (i.e., were at an age where virtually all controls look preferentially toward the positive contrast face). These infants behaved like normal newborns: three of them oriented preferentially toward config, and not a single baby looked preferentially toward the positive-contrast face (unpublished data). Despite patients’ face detection being abnormal at the time of treatment, they appear to catch up by adulthood. Individuals aged 9–20 years (mean = 14 years) who were deprived of early visual experience by bilateral congenital cataracts (mean length of deprivation = 3.9 months from birth) are as fast and accurate as age-matched controls at detecting facial structure in upright Mooney faces (Mondloch, Le Grand, and Maurer, 2003; see figure 25.7). Thus the mechanisms underlying face detection can be trained equally well by visual input from birth or after a delay. In adults, upright Mooney faces activate the face-sensitive fusiform face area (FFA) (Kanwisher, Tong, and Nakayama, 1998) and elicit a large N170 waveform in electrophysiological studies (Segalowitz et al., 2005; see chapter 31 in this volume by de Haan); we do not know whether the neural correlates of face detection are the same for patients as those found in visually normal individuals. Holistic face processing. A compelling demonstration of holistic processing is the composite face effect (Carey and Diamond, 1994; Hole, 1994; Hole, George, and Dunsmore, 1999; Le Grand et al., 2004; Michel et al., 2006; Young, Hellawell, and Hay, 1987). Adults find it difficult to recognize that the top half of a face is the same when it is recombined with the bottom half of another face, unless holistic processing is disrupted by misaligning the top and bottom halves (see figure 25.8). In addition, adults recognize the features from an individual’s face more easily in the context of the whole face (e.g., Larry’s nose in Larry’s face) than in isolation (the whole/ part advantage) (Tanaka and Farah, 1993). These findings demonstrate that facial features not only are represented individually, but also are integrated into a holistic representation that interferes with access to any representation of the individual features. Children show an adultlike whole-part

Figure 25.7 Test for face detection. Example of a Mooney face and a scrambled Mooney face formed by transforming all luminance values to black or white in order to obscure facial features (top). Mean accuracy and reaction time for discriminating intact

and scrambled Mooney faces for 11 patients treated for bilateral congenital cataract (black bars) and 11 age-matched controls (white bars) (bottom). Patients performed normally on both measures.

maurer, lewis, and mondloch: plasticity of the visual system

423

Figure 25.8 Test stimuli for the composite face effect. On “same” trials, the top halves of the two faces are the same, but they are combined with different bottom halves. The subjects’ task is to indicate that the tops are the same. When the halves are aligned in upright faces (bottom panel), adults find the task difficult because

holistic processing creates the impression of two different faces. When the halves are misaligned to break holistic processing, the task is much easier. (Reprinted from Developmental Science, vol. 10, D. Maurer, C. Mondloch, and T. L. Lewis, “Sleeper effects.” Copyright 2007.)

advantage by 4 years of age and adultlike composite face effect by 4–6 years of age (the youngest age tested) (Carey and Diamond, 1994; de Heering, Houthuys, and Rossion, 2007; Mondloch et al., 2007; Pellicano and Rhodes, 2003; Tanaka et al., 1998). No such whole/part advantage or composite face effect occurs for inverted faces (Hole, 1994; Tanaka and Farah, 1993; Young, Hellawell, and Hay, 1987) or, to the limited extent tested, nonface objects (Robbins and McKone, 2007; Tanaka and Farah, 1993). Unlike face detection, early visual deprivation prevents the later development of holistic processing as measured by the composite face effect. Patients (age 9–23 years) deprived of early visual experience by bilateral congenital cataracts perform just as well when composite faces are

aligned as they do when composite faces are misaligned (Le Grand et al., 2004). This result is particularly striking as the patients’ impairment in holistic processing is demonstrated by enhanced performance relative to normals when the top halves are the same and the faces are aligned. The details of later experience also impact holistic processing: adults process other-race faces less holistically than same-race faces, as measured by either the composite face effect or the whole-part advantage (Michel, Caldara, and Rossion, in press; Michel et al., 2006; Tanaka, Kiefer, and Bukach, 2004). The results demonstrate that this important aspect of processing, which distinguishes face from object processing (reviewed in Maurer, Le Grand, and Mondloch, 2002), is not prespecified, but rather depends on early visual input

424

neural plasticity in development

to set up or maintain its neural substrate and on later experience to refine it. Recognizing facial identity. Adults are able to recognize the identity of hundreds of faces, despite all faces sharing firstorder relations. Adults’ expert face recognition depends on their being sensitive to a variety of subtle differences among individual faces including the shape of the external contour, the shape of individual internal features (e.g., eyes, mouth), and differences in the spacing among features (e.g., distance between the eyes) or second-order relations. Adults’ expertise is limited to the category of faces that they differentiate regularly: they are much better at recognizing faces of individuals from their own race compared to faces of a different race—the other-race effect (reviewed in Meissner and Brigham, 2001), an effect that is modulated by the individual’s differential frequency of exposure to own versus other-race faces and that is apparent when faces differ only in features or only in second-order relations (Rhodes, Hayward, and Winkler, 2006). This perceptual-narrowing begins during infancy: 6-month-olds can discriminate between individual monkey faces as readily as between individual human faces, whereas 9-month-olds tested with the same method discriminate between the human but not the monkey faces (Pascalis, de Haan, and Nelson, 2002), unless they are regularly exposed to pictures of monkeys between 6 and 9 months of age (Pascalis et al., 2005; see Hannon and Trehub, 2005 for a similar finding in music perception and Kuhl, Tsao, and Liu, 2003 for a similar finding in phonemic perception). Likewise, 3-month-old infants, but not newborns, look longer at own-race faces than at otherrace faces with which they are paired (Kelly et al., 2005), unless they were exposed to other-race faces during the first three months of life (Bar-Haim et al., 2006), and they recognize the identity of own-race faces but not other-race faces to which they have been habituated, a difference that is eliminated when they are familiarized with several exemplars of other-race faces prior to the test (Sangrigoli and de Schonen, 2004b). Experience favoring familiar categories of faces continues to influence the face recognition system in childhood. Five-year-olds are better at recognizing the identity of human faces compared to faces from nonhuman categories such as sheep and monkeys (Pascalis et al., 2001), and 3- to 9-year-old children are better at recognizing ownrace than other-race faces (Corenblum and Meissner, 2006; Pezdek, Bladnon-Gitlin, and Moore, 2003; Sangrigoli and de Schonen, 2004a). Despite abundant exposure to faces during infancy and childhood, children do not reach adult levels of expertise in recognizing facial identity until adolescence. Six- and 8-year-olds have difficulty matching facial identity when two versions of the same face differ in facial expression, clothing, or lighting (reviewed in Carey, Diamond, and

Woods, 1980; Bruce et al., 2000; Mondloch et al., 2003). Even 10-year-olds have difficulty recognizing that identity has not changed when they see a face from two different points of view (e.g., en face and turned 45° to the right; Mondloch et al., 2003). Several studies have shown that the slow development of adultlike expertise in face recognition is attributable to the slow development of sensitivity to second-order relations. Whereas 6-year-olds are (nearly) adultlike when making same/different judgments about face pairs that differ in the shape of internal features or the external contour (Mondloch, Le Grand, and Maurer, 2002, 2003; see also Campbell and Tuck, 1995; Campbell, Walker, and Baron-Cohen, 1995 for evidence of early sensitivity to external contour), sensitivity to second-order relations for facial identity emerges after 4 years of age (Freire and Lee, 2001, see analysis of data for separate age groups in McKone and Boyer, 2006, 137; Mondloch, Leis, and Maurer, 2006), and even 14-year-olds make more errors than adults in discriminating faces that differ only in secondorder relations (Mondloch, Le Grand, and Maurer, 2002, 2003). Early visual deprivation has no apparent effect on the later development of recognition of facial identity based on featural processing. Deprived patients can easily distinguish faces that differ only in the shape of individual features (Le Grand et al., 2001), even with stimulus sets of typical or above-average difficulty for adults (unpublished observations) and they can match faces based on emotional expression, vowel being mouthed, and direction of eye gaze (Geldart et al., 2002), all of which are tasks that can be performed by processing local features. They can also distinguish faces that differ only in the external contour of the face (Le Grand et al., 2001). In contrast, they have deficits in distinguishing faces that differ only in the spacing among features such as the distance between the eyes, even though the stimulus sets covered most of the natural variability among faces (Le Grand et al., 2001, 2003). They also have deficits in matching faces’ identity when the matching face is presented from a novel point of view (Geldart et al., 2002). Both of these are tasks that require sensitivity to secondorder relations (Mondloch et al., 2003). Early visual input to specifically the right hemisphere is necessary for the development of normal sensitivity to second-order relations: patients deprived of input mainly to the right hemisphere or to both hemispheres show impairment at distinguishing faces that differ in the spacing of features, while patients whose early deprivation was mainly to the left hemisphere perform normally (Mondloch et al., 2003). In fact, visual deprivation to the right hemisphere lasting as little as the first 2 months after birth is sufficient to cause these deficits. This finding is especially interesting given that sensitivity to second-order relations is particularly slow to develop in visually normal children. In contrast, recognition of facial

maurer, lewis, and mondloch: plasticity of the visual system

425

identity based on the shape of internal features or the external contour develops more rapidly and does not require early visual experience. Visual Spatial Attention Adults are adept at directing attention toward the location of salient information in the visual field and at focusing attention on relevant information and ignoring distracters. Adults can do so even without making eye movements. Although the ability to orient attention covertly emerges during infancy (e.g., Johnson, Posner, and Rothbart, 1994; Johnson and Tucker, 1996), it takes many years for children to become as good as adults at redirecting attention quickly and at ignoring distracting information (e.g., Enns and Akhtar, 1989; Enns and Brodeur, 1989; Goldberg, Maurer, and Lewis, 2001a; Ridderinkhof and van der Molen, 1995). In monkeys, binocular deprivation beginning shortly after birth causes a large reduction in the sensitivity of cells in the posterior parietal cortex to visual stimuli, much larger than the effect on cells in the primary visual cortex (Carlson, Pertovaara, and Tanila, 1987; Hyvarinen and Hyvarinen, 1979, 1983; Hyvarinen, Hyvarinen, and Linnankoski, 1981). Studies of both monkeys and humans indicate that the posterior parietal cortex plays a major role in regulating visual spatial attention (e.g., Corbetta, 1998; Coull et al., 2000; Petersen, Robinson, and Currie, 1989; Posner et al., 1987; Steinmetz and Constantinidis, 1995). To test visual spatial attention in children treated for cataracts, we adapted two tasks that have been used to document developmental changes during childhood and impairments after lesions involving the posterior parietal cortex: covert orienting to the expected location of an upcoming target based on an informative central cue (termed an endogenous cue in the literature), either in a simple reaction time task or a harder discrimination task with added incompatible distracters (Goldberg, Maurer, and Lewis, 2001b). On both tasks, patients treated for bilateral congenital cataract showed subtle deficits. On the simple reaction task, they showed normal evidence of orienting covertly toward the target when the cue appeared shortly before it (100 or 400 msec) but not when the cue preceded it by 800 msec, especially if the deprivation had lasted more than the first 4 months of life. That result suggests that patients treated for bilateral congenital cataract have difficulty sustaining attention. On the harder discrimination task, they had more difficulty than the control group in ignoring irrelevant distracters and in shifting attention covertly to the upper visual field. Together, the results indicate that early binocular deprivation does not prevent the later development of visual spatial attention, but that it leads to subtle deficits that become apparent under specific conditions, such as the need to make difficult discriminations in a field crowded with irrelevant or misleading information.

426

neural plasticity in development

When we tested the deprived eye of children treated for unilateral congenital cataract with the same methods (Goldberg, 1998), they were normal on all aspects of both tasks: their reaction times indicated that they were normal at orienting covertly to the expected location of an upcoming target whether it came 100, 400, or 800 msec after the cue and whether it was the simple detection task or the harder discrimination task. As well, they were able to ignore distracters as well as the control group. Like the results for global form and global motion, these findings suggest that visual deprivation has less effect on the development of visual spatial attention if the deprivation is monocular rather than binocular.

Summary of effects of visual deprivation from cataract The studies of children treated for congenital cataract indicate that patterned visual input during infancy is necessary for the development of some, but not all, aspects of visual perception. Some visual capabilities are spared: sensitivity to high temporal frequencies and low spatial frequencies, face detection, and featural processing of facial identity. Only two of these capabilities have been studied longitudinally (sensitivity to low spatial frequencies and face detection), and in both cases there was an initial deficit that was overcome with age and, presumably, more years of delayed visual experience. These findings imply that there is a period during which the visual system is plastic enough to recover from the adverse effects of earlier deprivation. For all other visual capabilities that have been studied, there are permanent deficits from early visual deprivation. At least for low-level vision (comparable studies have not been done for high-level vision), the extent of the deficit varies as a function of stimulus parameters. Thus the deficit in contrast sensitivity is largest for high spatial frequencies and low temporal frequencies, the extent of the visual field is most restricted in the far temporal field, sensitivity to the direction of local motion is most degraded for slowly moving stimuli, and, after unilateral deprivation, sensitivity is degraded more in the near nasal field than in the near temporal field. In general, the low-level visual deficits are worse if the deprivation was monocular than if it was binocular, unless the unfair competition in monocular cases was offset by extensive patching of the nondeprived eye. There are also deficits in higher-level visual perception: sensitivity to global direction of motion, sensitivity to global form, holistic processing of faces, sensitivity to second-order relations in faces that are critical to recognizing facial identity, and visual spatial attention. However, the deficits for global motion, global form, and visual spatial attention are smaller in the deprived eye after monocular than after binocular deprivation (monocular studies have not been done on face processing). Together, the results point toward a

surprising sparing of higher-level visual function after monocular deprivation. Poorer sensitivity in a deprived eye after binocular than after monocular deprivation implies that extrastriate areas involved in at least some aspects of higher-level vision are not affected by competitive interactions between the eyes for cortical connections of the type well documented for the primary visual cortex. Rather, there appears to be another mechanism by which the eyes can interact, namely, complementary interactions. In extrastriate areas that integrate form signals, motion signals, and information relevant to visual spatial attention, these complementary interactions appear to replace the competitive interactions evident in primary visual cortex.

Evidence from blindness Additional information about visual plasticity comes from studies of the visual cortex of blind individuals. In adults who were blind from an early age, the visual cortex responds to tactile and auditory stimuli (Théoret, Meerabet, and PascualLeone, 2004). For example, the visual cortex is activated when these blind adults read Braille or discriminate complex tactile patterns, and the level of activation is almost as high as that in the somatosensory cortex (Burton, Snyder, Conturo, et al., 2002; Gizewski et al., 2003; Sadato et al., 2002). When the visual cortical activity is disrupted temporarily by transcranial magnetic stimulation (TMS), adults blind from an early age make more errors and report that Braille dots do not make sense, that some are missing, and that they feel extraneous phantom dots (Cohen et al., 1999). Auditory input also activates the visual cortex after early blindness: spoken sentences cause fMRI activation of primary and higher visual cortical areas (Röder et al., 2002), and deviant sounds such as an incongruous word or unexpected pitch elicit a response over the visual cortex (Bavelier and Neville, 2002; Kujala et al., 1995; Röder, Rösler, and Neville, 2000), although some of these effects may reflect higher-order language processing rather than auditory processing per se (Amedi et al., 2003; Burton, Diamond, and Mcdermott, 2003). In the cat, if the eyes are removed at birth, neurons in the primary visual cortex later respond to auditory stimuli (Yaka, Yinon, and Wollberg, 1999). These results complement the findings from congenital cataract in indicating that, in the absence of visual input during early childhood, the visual cortex does not develop its normal visual sensitivity.

Developmental changes in visual plasticity It has traditionally been assumed that the visual system becomes hardwired about the time that acuity reaches adult levels, around 6 years of age. After that age it is assumed

that visual perturbations like cataract, strabismus (misaligned eyes), or blindness will not cause functional or neural changes. After that age, it also appears to be too late to offset the effects of earlier visual perturbations by active training or by patching the good eye after monocular deprivation (reviewed in Birnbaum, Koslowe, and Sanet, 1977 and in MintzHittner and Fernandez, 2000; see also American Academy of Ophthalmology, 2002). However, our studies of children who developed cataracts postnatally indicate that the sensitive period during which visual deprivation can cause permanent deficits varies across visual functions and can last much less or much more than 6 years. For sensitivity to global motion, the sensitive period for damage is quite short, ending during infancy. In two studied cases, acuity and peripheral light sensitivity, it lasts longer than the period of normal development, namely, until about age 10 and into early adolescence, respectively, although the deficits are increasingly smaller with later onset of deprivation. There is also evidence that the visual system remains sufficiently plastic after age 6 for some deficits to be offset by training or patching. Most of the evidence comes from studies of children with reduced vision (amblyopia) caused by abnormal early visual input because the eyes were crossed or had unequal refractive errors. The standard therapy is to correct the peripheral problem (realign the eyes by muscle surgery; correct the refractive errors) and then to force usage of the affected eye by patching the good eye. This standard therapy is usually not attempted if the child is older than about 6 because it is assumed to be too late for remediation. However, a few systematic studies suggest it is not too late (e.g., Birnbaum, Koslowe, and Sanet, 1977; Oliver et al., 1986). The largest study to date involved 507 children with amblyopia at 49 clinical sites (Pediatric Eye Disease Investigation Group, 2005). There was a gradual and substantial improvement in the acuity of roughly 25 percent of the amblyopic eyes merely from prescribing the optimal glasses. The success rate was similar for children 7–8, 9–10, 11–12, and 13–17 years old, with even greater improvement if the child had not been treated previously for amblyopia. When the experimenters added up to 6 months of patching, detail work at a near distance, and pharmacological blurring of the sound eye, the amount of improvement was larger at all ages. Even with adults, training has been successful in inducing improvements in amblyopic eyes (e.g., Kupfer, 1957; Levi and Polat, 1996; Levi, Polat, and Hu, 1997; Li and Levi, 2004; Polat et al., 2004; Simmers and Gray, 1999; Zhou et al., 2006; reviewed in Levi, 2005). For example, amblyopic adults who are given feedback about the accuracy of their judgments about small misalignments between two lines or among three small oriented elements show improvements over a number of sessions on the practiced task and improvements in letter acuity. The improvements sometimes persist over many months, and if they do not, are easily

maurer, lewis, and mondloch: plasticity of the visual system

427

reestablished with a small amount of practice (reviewed in Levi, 2005). Training with feedback can even improve the vision of adults without eye problems, with at least some of the changes attributable to changes in the visual nervous system (reviewed in Fine and Jacobs, 2002). Similarly, adults with normal eyes who play action video games surpass those who do not on a number of visual measures: they have a larger useful field of view, process a stream of rapidly presented letters more accurately, have better central and peripheral acuity, and have less interference from peripheral distracters or between concurrent central and peripheral tasks (Green and Bavelier, 2003, 2006, 2007). Most of these effects are apparent even when adults are randomly assigned to play action video games for 2–6 weeks as part of a well-controlled experiment (Green and Bavelier, 2003, 2006, 2007). Evidence for plasticity after 5 years of age also comes from studies of face processing and of cortical reorganization after blindness. As noted earlier, face processing begins to specialize for familiar categories of faces (same race, same species) during infancy (Bar-Haim et al., 2006; Kelly et al., 2005; Pascalis, de Haan, and Nelson, 2002; Pascalis et al., 2005). However, Korean children adopted into European families before age 9 become better at discriminating Caucasian than Asian faces, suggesting that the other-race effect can be reversed completely by later experience (Sangrigoli et al., 2005). In blind individuals, the visual cortex responds to auditory and tactile inputs, and perhaps even language, not only in those blind from an early age, but also in adults who became blind as late as adolescence (Cohen et al., 1999; Sadato et al., 2002) and, to a lesser extent, in adults who became blind after 18 years of age or who were simply blindfolded for 5 days in the laboratory (Burton, Snyder, Conturo, et al., 2002; Burton, Sinclair, and McLaren, 2004; Burton, Snyder, Diamond, and Raichle, 2002; PascualLeone and Hamilton, 2001). Such rapid plasticity after the removal of visual stimulation suggests that tactile inputs have synapses to the visual cortices that are not normally active because of stronger visual inputs and/or active inhibition. Only when the visual input is removed do they become manifest. Additional evidence for plasticity in the adult visual system comes from case reports. A 48-year-old woman with an early history of crossed eyes was able to achieve, and slowly refine, binocular stereoscopic vision when her eyes were perfectly aligned, probably for the first time, by a prism and eyetraining exercises (Barry, 2006; Sacks, 2006). A 29-year-old blind man, born without natural lenses, slowly gained the ability to perceive unified objects after he was first given glasses to compensate for his inability to focus visual input (Mandavilli, 2006). There are also case reports of rapid improvements of vision in the amblyopic eye after the fellow “good” eye was lost to injury or disease—as if the amblyopic

428

neural plasticity in development

eye had formed neural connections that were being inhibited by the fellow eye (Polat et al., 2004; Rahl et al., 2002; Vereecken and Brabant, 1984; M. Wilson, 1992; see also Fronius et al., 2006; Leonards and Sireteanu, 1993; Mallach, Chakravarthy, and Hart, 2000). Evidence for hidden cortical connections with a previously deprived eye has also been reported in cats (Hoffmann and Lippert, 1982; Ohzawa and Freeman, 1988). There is converging evidence from a recent study of rats following long-term monocular deprivation: when the deprived eye was reopened in adulthood, long after the end of the critical period, there was a loss of acuity in the nondeprived eye from supranormal to normal values, followed by gradual improvement over 2 months in the acuity of the previously deprived eye from no pattern vision to values about one-third of normal (Iny et al., 2006). When the originally deprived eye was closed once again, the same process was repeated: enhanced acuity of the open eye while the deprived eye was closed; a drop in its acuity when the closed eye was reopened; and gradual improvement in the acuity of the previously deprived eye. These trade-offs indicate that there is visual plasticity in the adult rat that is modulated in part by trade-offs between the eyes. Collectively, these results suggest that there is considerable visual plasticity in the adult brain and that some of the deficits seen when early visual input was abnormal may be caused by inhibitory interactions that prevent intact connections from functioning.

General principles Early visual deprivation can impair the neural architecture for later learning: sleeper effects. Children treated for bilateral congenital cataracts before 2–3 months of age fail to develop normal contrast sensitivity for mid and high spatial frequencies, normal holistic face processing, or normal sensitivity to the spacing of facial features (Ellemberg, Lewis, Maurer, et al., 1999; Le Grand et al., 2001, 2004). Yet the first signs of these capabilities do not appear in the visually normal child until later during development (Cashon and Cohen, 2003; Maurer and Lewis, 2001a, 2001b; Mondloch, Leis, and Maurer, 2006), and thereafter they are refined over many years. Contrast sensitivity is not adultlike until 7 years of age (Ellemberg, Lewis, Liu, and Maurer, 1999), and sensitivity to the spacing of facial features becomes adultlike after 14 years of age (Mondloch, Le Grand, and Maurer, 2003). Longitudinal studies of acuity and contrast sensitivity provide striking evidence of this sleeper effect: by the first birthday, children treated during early infancy for bilateral congenital cataract have overcome their initial deficit in acuity and perform within normal limits. However, their acuity and contrast sensitivity for mid spatial frequencies hit an asymptote by age 4–5, while that of visually normal children continues to improve

until about age 6–7, leaving the patients with the permanent deficit in acuity described previously (Ellemberg et al., 2001; Lewis and Maurer, 2005; Maurer, Ellemberg, and Lewis, 2006; Maurer, Mondloch, and Lewis, 2007). Collectively, the results indicate that early visual input sets up or preserves the neural architecture that will be refined by later experience. Without visual input, that hardware may be recruited for other functions like touch or hearing, as it appears to be in the congenitally blind, for whom it supports later tactile learning of Braille characters. Thus, in the visually normal child, early visual input is necessary to preserve the neural infrastructure for later visual learning, even for visual capabilities that will not appear until later in development. For low-level vision, visual deprivation has a larger impact on abilities that develop more slowly, as predicted by the Detroit principle, but there is no obvious relationship between rate of development and visual plasticity for higher-level visual functions. One unifying principle about developmental plasticity that has been proposed is the Detroit model: “last hired, first fired” or “last to develop, most damaged” (Levi, 2005). For low-level visual capabilities, the data from congenital cataract fit the Detroit principle well. Thus, after early binocular deprivation, there are larger deficits in sensitivity to slow than to fast rates of flicker (Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2000), in sensitivity to high than to low spatial frequencies (Ellemberg, Lewis, Maurer, et al., 1999), in the far temporal field than in other parts of the visual field (Bowering et al., 1997), in the direction of local motion at slow than at fast velocities (Ellemberg et al., 2005), and to local motion defined by contrast rather than luminance (Ellemberg et al., 2005). In each case, the greater deficit is for the aspect of vision that is slower to mature in visually normal children (Bowering et al., 1996, 1997; Ellemberg, Lewis, Liu, and Maurer, 1999; Ellemberg, Lewis, Maurer, et al., 1999; Ellemberg et al., 2003). However, the effects of early binocular deprivation do not always follow the Detroit principle. Sensitivity to the direction of global motion becomes adultlike earlier than sensitivity to global form in both visually normal children and monkeys, and brain areas involved in processing global motion mature exceptionally early in the monkey (Bourne and Rosa, 2006; Bourne, Warner, and Rosa, 2005; Ellemberg et al., 2002; Lewis et al., 2004). However, early binocular deprivation in humans causes much larger deficits in sensitivity to global motion than in sensitivity to global form (cf. figures 25.5 and 25.6) (Ellemberg et al., 2002; Lewis et al., 2002). The Detroit principle also does not hold for the sensitive period for damage: grating acuity and sensitivity to global motion are similarly immature during infancy and, at least under some conditions, are still immature at 5 years of age (Ellemberg et al., 2004; Maurer and Lewis, 2001b; Wattam-Bell, 1994), but postnatal visual deprivation impairs

acuity if the onset is before about age 10 but has no effect on global motion if the onset is after the first few months of life. Similarly, the use of the visual cortex for tactile and auditory processing in the congenitally blind occurs for both capabilities that develop rapidly during infancy (e.g., auditory localization, processing of relative pitch) and those that develop slowly in the visually normal child (e.g., word learning). Both the dorsal (where) and ventral (what) streams are vulnerable. To a large extent, visual stimuli are processed in parallel by the dorsal visual stream, which favors the peripheral visual field and is specialized for information about motion and space, and by the ventral visual stream, which favors the central visual field and is specialized for information about form and color (Neville, 2006). Some have hypothesized that the dorsal pathway is more plastic than the ventral pathway both for possible enhancement (as in the congenitally deaf, where the absence of competitive auditory inputs promotes supranormal development of sensitivity to nonauditory inputs) and for deficits (as in children with dyslexia, Williams syndrome, or hemiplegic cerebral palsy), possibly because the dorsal pathway has a more protracted period of development (Atkinson, 2000; Gunn et al., 2002; Neville, 2006). Consistent with this hypothesis, children treated for bilateral congenital cataract have larger deficits in integrating local signals to perceive the direction of global motion (dorsal) than to perceive global form (ventral) (cf. figures 25.5 and 25.6) (Ellemberg et al., 2002; Lewis et al., 2002). However, contrary to the expectation of larger deficits in peripheral (dorsal) than central (ventral) vision, the deficits in contrast sensitivity in children treated for bilateral congenital cataract are of constant magnitude from the center of the visual field to 30° in the periphery. In children treated for unilateral cataract, the contrast sensitivity deficit is larger in the central visual field (ventral) than in the periphery (dorsal) (Maurer and Lewis, 1998). Deficits in configural processing of faces after early binocular deprivation (Le Grand et al., 2001, 2004; Mondloch, Le Grand, and Maurer, 2003) also indicate that the ventral pathway is not spared. Analogously, the visual cortex of the congenitally blind is recruited both for spatial localization of auditory stimuli (dorsal) and for recognition of word form (ventral) (Amedi et al., 2003; Burton, Diamond, Mcdermott, 2003; Leclerc et al., 2000). Overall, the results suggest that early visual input plays a large role in the development of both the dorsal and ventral streams. The predictions are complicated by recent evidence that the two pathways are not well segregated in infants (Dobkins, 2006) and that parts of the dorsal stream mature relatively early (Ellemberg et al., 2002; Bourne, Warner, and Rosa, 2005; Bourne and Rosa, 2006).

maurer, lewis, and mondloch: plasticity of the visual system

429

mus, but also for the crystallization of that substrate for a number of years after it is formed. Daw (1998, 2003) has argued that sensitive periods are longer for visual capabilities mediated at higher levels of the nervous system and hence predicts that the sensitive period for damage from bilateral cataracts should be longer for higher-level vision than for low-level vision. The data for global motion, the only high-level visual capability for which a sensitive period has been deduced, contradict the prediction (see figure 25.9). Sensitivity to the direction of global motion, which is mediated largely by the middle temporal complex in the dorsal stream, is normal even when the deprivation began at 4–12 months of age (Ellemberg et al., 2002), early enough to have led to permanent deficits in acuity and long before sensitivity approaches adult levels (reviewed in Lewis and Maurer, 2005). Additional evidence for the resilience of sensitivity to global motion comes from the case of MM described briefly in the section on global motion. This subject lost one eye and suffered blinding corneal damage to the other eye at age 3.5 years. After a corneal transplant at age 43, MM performed normally on many motion tasks and showed normal activation of the middle temporal complex (Fine et al., 2003).

The sensitive period for damage varies across visual functions and is not always longer for higher-level capabilities. Studies of children who developed cataracts at different ages indicate that the sensitive period during which visual deprivation can cause damage differs for different visual functions. For the low-level peripheral light sensitivity and acuity, it is surprisingly long, lasting longer than the period of normal functional development (see figure 25.9). For example, a short period of visual deprivation beginning any time before age 10 causes permanent deficits in acuity, even through acuity reaches adult functional levels by age 6 (Lewis and Maurer, 2005). Similarly, a short period of deprivation beginning in early adolescence causes permanent deficits in peripheral light sensitivity, even though it reaches adult functional levels by age 7 (Bowering et al., 1993). There is a similar pattern for some aspects of optokinetic (OKN) eye movements (see figure 25.9), the series of pursuit and saccadic eye movements that occur when a pattern moves across the visual field (reviewed in Lewis and Maurer, 2005). These results indicate that visual input is necessary not only for the normal development of the neural substrate mediating acuity, peripheral visual acuity, and optokinetic nystag-

Normal development Damage Letter Acuity

?

Light Sensitivity Midperiphery

Global Motion

Asymmetry of OKN for Large High-contrast Patterns

0

2

4

6

8

10

12

14

Age (years)

Figure 25.9 Comparison of sensitive periods for four aspects of vision. Gray bars represent the period of normal development for each visual function, and black bars represent the sensitive period during which visual deprivation causes damage. Ages for the end of normal development are approximate and for the conditions under which cataract patients were tested. The comparisons indi-

430

neural plasticity in development

cate that the sensitive period can be shorter (global motion) or longer (letter acuity, peripheral light sensitivity, OKN) than the period of normal development. (Reprinted from Developmental Psychobiology, vol. 46, T. L. Lewis and D. Maurer, “Multiple sensitive periods in human visual development: Evidence from visually deprived children,” p. 178. Copyright 2005.)

In contrast, he had severe deficits in acuity, form identification, and face processing, and abnormal activation of the primary visual cortex (area V1) and other high-level visual cortical areas (V2, V3, fusiform and lingual gyri). Collectively, these data indicate that there are different sensitive periods during which visual deprivation can lead to permanent deficits and that the sensitive periods for higher-level visual functions are not necessarily longer than those for low-level vision. The sensitive period for recovery cannot be predicted from the sensitive period for damage. The evidence for visual plasticity that extends into adolescence and even adulthood (see the section “Evidence from blindness”) indicates that the sensitive period for recovery extends long beyond the sensitive period for damage and the period of normal development. In fact, recovery may be possible even in adulthood, and there may be no sensitive period after which the potential for recovery is lost. Thus, for sensitivity to the direction of global motion, the sensitive period for damage is very short, ending during infancy, and normal development is relatively rapid, but MM, the individual who was blinded by corneal damage at age 3, was able to achieve normal sensitivity to global motion when treated at age 43 (Fine et al., 2003). In contrast, MM had severely reduced spatial contrast sensitivity and difficulty on all but the simplest measures of form perception, with no improvement over the 2 years following treatment. The contrasting results indicate that the sensitive period for recovery differs for different visual capabilities. There is converging evidence from our longitudinal results with children treated for bilateral congenital cataract. For contrast sensitivity, which reaches adult levels in children with normal eyes at age 7, our results indicate that the sensitive period for recovery for medium and high spatial frequencies asymptotes before age 4, leaving the child with a permanent deficit, but that the sensitive period for recovery for low spatial frequencies persists until a later age, such that children eventually overcome the deficit (Maurer, Ellemberg, and Lewis, 2006). Yet visual deprivation beginning as late as age 10 causes a permanent deficit in sensitivity to high spatial frequencies (reviewed in Lewis and Maurer, 2005). Collectively, the evidence indicates that recovery from visual deficits needs to be studied in its own right: conclusions about the potential for recovery and for timing constraints on recovery cannot be generalized across visual functions or from information about the way the system is damaged by abnormal visual experience at different ages. Implications for Understanding Normal Development The results described here indicate that patterned visual input immediately after birth plays a vital role in the construction and/or preservation of the neural architecture

that will later mediate visual function. Specifically, the low spatial frequencies that newborns can see set up a system, probably in the right hemisphere, for later learning about configural properties of faces and a system, probably in both hemispheres, for later development of other visual skills. These systems are refined by later visual experience—but only if their basic architecture was set up (or maintained) by stimulating the crude vision of the newborn. When the timing of that crude early visual experience is delayed until cataracts are removed, some visual capabilities will show sleeper effects: they will fail to emerge at a later point in childhood, perhaps because the requisite neural architecture is no longer available. Nevertheless, sufficient visual plasticity persists in some parts of the system to allow partial recovery or reorganization. acknowledgments

Our studies of children treated for cataract were supported by grants from the Natural Science and Engineering Council (Canada), the Medical Research Council (Canada), the Canadian Institutes of Health Research, the Social Science and Humanities Research Council (Canada), the Human Frontiers Foundation, the March of Dimes, and the National Institutes of Health (U.S.). REFERENCES

Amedi, A., N. Raz, P. Pianka, R. Malach, and E. Zohary, 2003. Early “visual” cortex activation correlates with superior verbal memory performance in the blind. Nature Neurosci. 6:758–766. American Academy of Ophthalmology, 2002. Preferred Practice Pattern: Amblyopia. San Francisco, CA: American Academy of Ophthalmology, 2002. Aslin, R. N., and S. L. Shea, 1990. Velocity thresholds in human infants: Implications for the perception of motion. Dev. Psychol. 26:589–598. Atkinson, J., 2000. The Developing Visual Brain. Oxford, UK: Oxford University Press. Atkinson, J., O. Braddick, and K. Moar, 1977. Contrast sensitivity of the human infant for moving and static patterns. Vis. Res. 17:1045–1047. Bahrick, H. P., P. O. Bahrick, and R. P. Wittlinger, 1975. Fifty years of memory for names and faces: A cross-sectional approach. J. Exp. Psychol. [Gen.] 104:54–75. Banks, M. S., and P. Salapatek, 1978. Acuity and contrast sensitivity in 1-, 2-, and 3-month-old human infants. Invest. Ophthalmol. Vis. Sci. 17:361–365. Bar-Haim, Y., T. Ziv, D. Lamy, and R. Hodes, 2006. Nature and nurture in own-race face processing. Psychol. Sci. 17:159–163. Barry, S., 2006. Steropsis. CVNet e-mail list, posting of August 21, 2006, by Sue Barry. Bavelier, D., and H. Neville, 2002. Cross-modal plasticity: Where and how? Nature Rev. Neurosci. 3:443–452. Bertenthal, B. I., and A. Bradbury, 1992. Infants’ detection of shearing motion in random-dot displays. Dev. Psychol. 28:1056– 1066. Birch, E. E., 1993. Stereopsis in infants and its developmental relation to visual acuity. In K. Simons, ed., Early Visual

maurer, lewis, and mondloch: plasticity of the visual system

431

Development, Normal and Abnormal, 224–236. Committee on Vision, Commission on Behavioral and Social Sciences and Education, National Research Council. New York: Oxford University Press. Birch, E. E., and D. R. Stager, 1996. The critical period for surgical treatment of dense congenital unilateral cataract. Invest. Ophthalmol. Vis. Sci. 37:1532–1538. Birch, E. E., D. R. Stager, J. Leffler, and D. Weakley, 1998. Early treatment of congenital unilateral cataract minimizes unequal competition. Invest. Ophthalmol. Vis. Sci. 39:1560–1566. Birch, E. E., W. H. Swanson, D. R. Stager, M. Woody, and M. Everett, 1993. Outcome after very early treatment of dense congenital unilateral cataract. Invest. Ophthalmol. Vis. Sci. 34: 3687–3699. Birnbaum, M., K. Koslowe, and R. Sanet, 1977. Success in amblyopia therapy as a function of age: A literature review. Am. J. Optom. Physiol. Opt. 54:269–275. Bourne, J. A., and M. G. Rosa, 2006. Hierarchical development of the primate visual cortex, as revealed by neurofilament immunoreactivity: Early maturation of the middle temporal area (MT). Cerebral Cortex 16:405–414. Bourne, J., C. E. Warner, and M. G. Rosa, 2005. Topographic and laminar maturation of striate cortex in early postnatal marmoset monkeys, as revealed by neurofilament immunohistochemistry. Cerebral Cortex 15:740–748. Bowering, E. R., 1992. The peripheral vision of normal children and of children treated for cataracts. Doctoral dissertation, McMaster University. Dissertation Abstracts International. Bowering, E. R., D. Maurer, T. L. Lewis, and H. P. Brent, 1993. Sensitivity in the nasal and temporal hemifields in children treated for cataract. Invest. Ophthalmol. Vis. Sci. 34:3501–3509. Bowering, E. R., D. Maurer, T. L. Lewis, and H. P. Brent, 1997. Constriction of the visual field of children after early visual deprivation. J. Pediatr. Ophthalmol. Strabismus 34:347–356. Bowering, E. R., D. Maurer, T. L. Lewis, H. P. Brent, and P. Riedel, 1996. The visual field in childhood: Normal development and the influence of deprivation. Developmental Cognitive Neuroscience Technical Report No. 96.1. Braddick, O. J., 1993. Orientation- and motion-selective mechanisms in infants. In K. Simons, ed., Early Visual Development, Normal and Abnormal, 163–177. Committee on Vision, Commission on Behavioral and Social Sciences and Education, National Research Council. New York: Oxford University Press. Bronson, G., 1990. Changes in infants’ visual scanning across the 2- to 14-week age period. J. Exp. Child Psychol. 49:101– 125. Brown, A., and M. Yamamoto, 1986. Visual acuity in newborn and preterm infants measured with grating acuity cards. Am. J. Ophthalmol. 102:245–253. Bruce, V., R. N. Campbell, G. Doherty-Sneddon, A. Import, S. Langton, S. McAuley, and R. Wright, 2000. Testing face processing skills in children. Br. J. Dev. Psychol. 18:319–333. Burton, H., J. Diamond, and K. McDermott, 2003. Dissociating cortical regions activated by semantic and phonological tasks: A fMRI study in blind and sighted people. J. Neurophysiol. 90:1965–1982. Burton, H., R. Sinclair, and D. McLaren, 2004. Cortical activity to vibrotactile stimulation: An fMRI study in blind and sighted individuals. Hum. Brain Mapping 23:210–238. Burton, H., A. Snyder, T. Conturo, E. Akbudak, J. Ollinger, and M. Raichle, 2002. Adaptive changes in early and late blind: fMRI study of Braille reading. J. Neurophysiol. 87:589– 607.

432

neural plasticity in development

Burton, H., A. Snyder, J. Diamond, and M. Raichle, 2002. Adaptive changes in early and late blind: fMRI study of verb generation to heard nouns. J. Neurophysiol. 88:3359–3371. Campbell, R., and M. Tuck, 1995. Recognition of parts of famousface photographs by children: An experimental note. Perception 24:451–456. Campbell, R., J. Walker, and S. Baron-Cohen, 1995. The development of differential use of inner and outer face features in familiar face identification. J. Exp. Child Psychol. 59:196–210. Carey, S., and R. Diamond, 1994. Are faces perceived as configurations more by adults than by children? Visual Cogn. 1:253– 274. Carey, S., R. Diamond, and B. Woods, 1980. Development of face recognition: A maturational component? Dev. Psychol. 16:257– 269. Carlson, S., A. Pertovaara, and H. Tanila, 1987. Late effects of early binocular visual deprivation on the function of Brodmann’s area 7 of monkeys (Macaca arctoides). Dev. Brain Res. 33:101–111. Cashon, C., and L. Cohen, 2003. The construction, deconstruction, and reconstruction of infant face perception. In O. Pascalis and A. Slater, eds., The Development of Face Processing in Infancy and Early Childhood, 55–68. New York: Nova. Cashon, C., and L. Cohen, 2004. Beyond U-shaped development in infants’ processing of faces: An information-processing account. J. Cogn. Dev. 5:59–80. Cohen, L., R. Weeks, N. Sadao, P. Celnik, K. Ishii, and M. Hallett, 1999. Period of susceptibility for cross-modal plasticity in the blind. Ann. Neurol. 45:451–460. Cohen, L. B., and B. A. Younger, 1984. Infant perception of angular relations. Infant Behav. Dev. 7:37–47. Corbetta, M., 1998. Frontoparietal cortical networks for directing attention and the eye to visual locations: Identical, independent, or overlapping neural systems? Proc. Natl. Acad. Sci. USA 95: 831–838. Corenblum, B., and C. A. Meissner, 2006. Recognition of faces of ingroup and outgroup children and adults. J. Exp. Child Psychol. 93:187–206. Coull, J. T., C. D. Frith, C. Buchel, and A. C. Nobre, 2000. Orienting attention in time: Behavioural and neuroanatomical distinction between exogenous and endogenous shifts. Neuropsychologia 38:808–819. Courage, M., and R. Adams, 1990. Visual acuity assessment from birth to three years using the acuity card procedure: Crosssectional and longitudinal samples. Optom. Vis. Sci. 67:713–718. Dannemiller, J. L., and B. R. Stephens, 1988. A critical test of infant pattern preference models. Child Dev. 59:210–216. Daw, N. W., 1998. Critical periods and amblyopia. Arch. Ophthalmol. 116:502–505. Daw, N. W., 2003. Critical periods in the visual system. In B. Hopkins and S.P. Johnson, eds., Neurobiology of Infant Vision, 43– 103. Westport, CT: Praeger. de Haan, M., M. Johnson, D. Maurer, and D. Perrett, 2001. Recognition of individual faces and average face prototypes by 1- and 3-month-old infants. Cogn. Dev. 16:659–678. de Heering, A., S. Houthuys, and B. Rossion, 2007. Holistic face processing is mature at 4 years of age: Evidence from the composite face effect. J. Exp. Child Psychol. 96:57–70. Diamond, R., and S. Carey, 1986. Why faces are and are not special: An effect of expertise. J. Exp. Psychol. [Gen.] 115:107– 117. Dobkins, K., 2006. Enhanced red/green color input to motion processing in infancy: Evidence for increasing dissociation of

color and motion information during development. In M. H. Johnson and Y. Munakata, eds., Attention and Performance XXI: Processes of Change in Brain and Cognitive Development, 403–425. Oxford, UK: Oxford University Press. Ellemberg, D., T. L. Lewis, N. Defina, D. Maurer, H. P. Brent, J.-P. Guillemot, and F. Lepore, 2005. Greater losses in the sensitivity to second-order local motion than to first-order local motion after early visual deprivation in humans. Vis. Res. 44: 2877–2884. Ellemberg, D., T. L. Lewis, M. Dirks, B. Lee, D. Maurer, T. Ledgway, J.-P. Guillemot, and F. Lepore, 2004. Putting order into the development of sensitivity to global motion. Vis. Res. 44:2403–2411. Ellemberg, D., T. L. Lewis, C. H. Liu, and D. Maurer, 1999. The development of spatial and temporal vision during childhood. Vis. Res. 39:2325–2333. Ellemberg, D., T. L. Lewis, D. Maurer, S. Brar, and H. P. Brent, 2002. Better perception of global motion after monocular than after binocular deprivation. Vis. Res. 42:169–179. Ellemberg, D., T. L. Lewis, D. Maurer, and H. P. Brent, 2000. Influence of monocular deprivation during infancy on the later development of spatial and temporal vision. Vis. Res. 40:3283– 3295. Ellemberg, D., T. L. Lewis, D. Maurer, and H. P. Brent, 2001. The role of visual input in setting up spatial filters in the human visual system. Proceedings of the Society for Imaging Science and Technology, PICS 2001: Image Processing, Image Quality, Image Capture Systems Conference, Imaging Science and Technology, Springfield, VA, pp. 421–425. Ellemberg, D., T. L. Lewis, D. Maurer, C. H. Liu, and H. P. Brent, 1999. Spatial and temporal vision in patients treated for bilateral congenital cataracts. Vis. Res. 39:3480– 3489. Ellemberg, D., T. L. Lewis, K. S. Meghji, D. Maurer, J.-P. Guillemot, and F. Lepore, 2003. Comparison of sensitivity to first- and second-order local motion in 5-year-olds and adults. Spatial Vis. 16:419–428. Enns, J. T., and N. Akhtar, 1989. A developmental study of filtering in visual attention. Child Dev. 60:1188–1199. Enns, J. T., and D. A. Brodeur, 1989. A developmental study of covert orienting to peripheral visual cues. J. Exp. Child Psychol. 48:171–189. Farroni, T., G. Csibra, F. Simion, and M. Johnson, 2002. Eye contact detection in humans from birth. Proc. Natl. Acad. Sci. USA 99:9602–9605. Farroni, T., E. Menon, and M. Johnson, 2006. Factors influencing newborns’ preference for faces with eye contact. J. Exp. Child Psychol. 95:298–308. Farroni, T., E. Valenza, F. Simion, and C. Umiltà, 2000. Configural processing at birth: Evidence for perceptual organisation. Perception 29:355–372. Fine, I., and R. A. Jacobs, 2002. Comparing perceptual learning across tasks: A review. J. Vis. 2:190–203. Fine, I., A. R. Wade, A. Brewer, M. G. May, D. F. Goodman, G. M. Boyton, B. A. Wandell, and D. I. MacLeod, 2003. Long-term deprivation affects visual perception and cortex. Nature Neurosc. 6:915–916. Freedland, R. L., and J. L. Dannemiller, 1987. Detection of stimulus motion in 5-month-old infants. J. Exp. Psychol. [Hum. Percept.] 13:566–576. Freire, A., and K. Lee, 2001. Face recognition in 4- to 7-year-olds: Processing of configural, featural, and paraphernalia information. J. Exp. Child Psychol. 80:347–371.

Freire, A., T. L. Lewis, D. Maurer, and R. Blake, 2006. The development of sensitivity to biological motion in noise. Perception 35:647–657. Fronius, M., L. Cirina, C. Kuhli, A. Cordey, and C. Orloff, 2006. Training the adult amblyopic eye with “perceptual learning” after vision loss in the nonamblyopic eye. Strabismus 14:75–79. Geldart, S., C. J. Mondloch, D. Maurer, S. de Schonen, and H. Brent, 2002. The effect of early visual deprivation on the development of face processing. Dev. Sci. 5:490–501. Gizewski, E., T. Gasser, A. de Greiff, A. Boehm, and M. Forsting, 2003. Cross-modal plasticity for sensory and motor activation patterns in blind subjects. NeuroImage 10: 968–975. Glass, L., 1969. Moiré effect from random dots. Nature 223: 578–580. Goldberg, M. C., 1998. Influence of binocular deprivation on the development of attention. Unpublished Ph.D. dissertation, McMaster University. Goldberg, M. C., D. Maurer, and T. L. Lewis, 2001a. Developmental changes in attention: The effects of endogenous cueing and of distractors. Dev. Sci. 4:209–219. Goldberg, M. C., D. Maurer, and T. L. Lewis, 2001b. The influence of binocular visual deprivation on the development of visual-spatial attention. Dev. Neuropsychol. 19:53–81. Goren, C., M. Sarty, and P. Wu, 1975. Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics 56:544–549. Green, C., and D. Bavelier, 2003. Action video game modifies visual selective attention. Nature 423:534–537. Green, C. S., and D. Bavelier, 2006. Effect of action video games on the spatial distribution of visuo-spatial attention. J. Exp. Psychol. [Hum. Percept.] 32:1465–1478. Green, C. S., and D. Bavelier, 2007. Action video game experience alters the spatial resolution of vision. Psychol. Sci. 18: 88–94. Gunn, A., E. Cory, J. Atkinson, O. Braddick, J. Wattam-Bell, A. Guzzetta, and G. Cioni, 2002. Dorsal and ventral stream sensitivity in normal development and hemiplegia. NeuroReport 13:843–847. Gwiazda, J., J. Bauer, F. Thorn, and R. Held, 1997. Development of spatial contrast sensitivity from infancy to adulthood: Psychophysical data. Optom. Vis. Sci. 74:785–789. Hainline, L., 1978. Developmental changes in visual scanning of face and nonface patterns by infants. J. Exp. Child Psychol. 25:90–115. Haith, M., T. Bergman, and M. Moore, 1977. Eye contact and face scanning in early infancy. Science 198:853–855. Hannon, E. E., and S. E. Trehub, 2005. Metrical categories in infancy and adulthood. Psychol. Sci. 16:48–55. Harwerth, R. S., E. L. Smith, M. L. J. Crawford, and G. K. von Noorden, 1989. The effects of reverse monocular deprivation in monkeys. I. Psychophysical experiments. Exp. Brain Res. 74:327–337. Harwerth, R. S., E. L. Smith, A. D. Paul, M. L. J. Crawford, and G. K. von Noorden, 1991. Functional effects of bilateral form deprivation in monkeys. Invest. Ophthalmol. Vis. Sci. 32:2311–2327. Hoffmann, K. P., and P. Lippert, 1982. Recovery of vision with the deprived eye after the loss of the non-deprived eye in cats. Hum. Neurobiol. 1:45–48. Hole, G., 1994. Configurational factors in the perception of unfamiliar faces. Perception 23:65–74.

maurer, lewis, and mondloch: plasticity of the visual system

433

Hole, G. J., P. A. George, and V. Dunsmore, 1999. Evidence for holistic processing of faces viewed as photographic negatives. Perception 28:341–359. Hubel, D. H., and T. N. Wiesel, 1962. Receptive fields, binocular interaction and function architecture in the cat’s visual cortex. J. Physiol. 160:106–154. Hubel, D. H., and T. N. Wiesel, 1968. Receptive fields and functional architecture of monkey striate cortex. J. Physiol. 195: 215–243. Hyvarinen, J., and L. Hyvarinen, 1979. Blindness and modification of association cortex by early binocular deprivation in monkeys. Child Care Health Dev. 5:385–387. Hyvarinen, J., and L. Hyvarinen, 1983. Higher functions and plasticity in visual pathways. Acta Ophthalmol. Suppl. 157:9–17. Hyvarinen, J., L. Hyvarinen, and I. Linnankoski, 1981. Modification of parietal association cortex and functional blindness after binocular deprivation in young monkeys. Exp. Brain Res. 42:1– 8. Iny, K., A. Heynen, E. Sklar, and M. Bear, 2006. Bidirectional modifications of visual acuity induced by monocular deprivation in juvenile and adult rats. J. Neurosci. 26:7368–7374. Jeffrey, B. G., Y. Z. Wang, and E. E. Birch, 2004. Altered global shape discrimination in deprivation amblyopia. Vis. Res. 44: 167–177. Johnson, M. H., S. Dziurawiec, H. D. Ellis, and J. Morton, 1991. Newborns’ preferential tracking of face-like stimuli and its subsequent decline. Cognition 40:1–19. Johnson, M. H., M. I. Posner, and M. K. Rothbart, 1994. Facilitation of saccades toward a covertly attended location in early infancy. Psychol. Sci. 5:90–93. Johnson, M. H., and L. A. Tucker, 1996. The development and temporal dynamics of spatial orienting in infants. J. Exp. Child Psychol. 63:171–188. Kanwisher, N., F. Tong, and K. Nakayama, 1998. The effect of face inversion on the human fusiform face area. Cognition 68: B1–B11. Kaufmann, F., 1995. Development of motion perception in early infancy. Eur. J. Pediatr. 154:S48–S53. Kelly, D. J., P. C. Quinn, A. M. Slater, K. Lee, A. Gibson, M. Smith, L. Ge, and O. Pascalis, 2005. Three-month-olds, but not newborns, prefer own-race faces. Dev. Sci. 8:F31– F36. Kovács, I., 2000. Human development of perceptual organisation. Vis. Res. 40:1301–1310. Kugelberg, U., 1992. Visual acuity following treatment of bilateral congenital cataracts. Doc. Ophthalmol. 82:211–215. Kuhl, P. K., F.-M. Tsao, and H.-M. Liu, 2003. Foreign-language experience in infancy: Effects of short-term exposure and social interaction on phonetic learning. Proc. Natl. Acad. Sci. USA 100: 9096–9101. Kujala, T., M. Huotilainen, J. Sinkkonen, A. Ahonen, K. Alho, M. Hamalainen, et al., 1995. Visual cortex activation in blind humans during sound discrimination. Neurosci. Lett. 183:143– 146. Kupfer, C., 1957. Treatment of amblyopia ex anopsia in adults. Am. J. Ophthalmol. 43:918–922. Leclerc, C., D. Saint-Amour, M. Lavoie, M. Lassonde, and F. Lepore, 2000. Brain functional reorganziation in early blind humans revealed by auditory event-related potentials. NeuroReport 11:545–550. Le Grand, R., C. J. Mondloch, D. Maurer, and H. P. Brent, 2001. Early visual experience and face processing. Nature 410: 890.

434

neural plasticity in development

Le Grand, R., C. Mondloch, D. Maurer, and H. Brent, 2003. Expert face processing requires visual input to the right hemisphere during infancy. Nature Neurosci. 6:1108–1112. Le Grand, R., C. J. Mondloch, D. Maurer, and H. P. Brent, 2004. Impairment in holistic face processing following early visual deprivation. Psychol. Sci. 15:762–768. Leonards, U., and R. Sireteanu, 1993. Interocular suppression in normal and amblyopic subjects: The effect of unilateral attenuation with neutral density filters. Percept. Psychophys. 54:65–74. Levi, D., 2005. Perceptual learning in adults with amblyopia: A reevaluation of critical periods in human vision. Dev. Psychobiol. 46:223–233. Levi, D. M., and U. Polat, 1996. Neural plasticity in adults with amblyopia. Proc. Natl. Acad. Sci. USA 93:6830–6834. Levi, D., U. Polat, and Y.-S. Hu, 1997. Improvement in vernier acuity in adults with amblyopia. Invest. Ophthalmol. Vis. Sci. 38:1493–1510. Lewis, T. L., D. Ellemberg, D. Maurer, M. Dirks, F. Wilkinson, and H. Wilson, 2004. A window on the normal development of sensitivity to global form in Glass patterns. Perception 33:409– 418. Lewis, T. L., D. Ellemberg, D. Maurer, F. Wilkinson, H. R. Wilson, M. Dirks, and H. P. Brent, 2002. Sensitivity to global form in Glass patterns after early visual deprivation in humans. Vis. Res. 42:939–948. Lewis, T. L., and D. Maurer, 1992. The development of the temporal and nasal visual fields during infancy. Vis. Res. 32: 903–911. Lewis, T. L., and D. Maurer, 2005. Multiple sensitive periods in human visual development: Evidence from visually deprived children. Dev. Psychobiol. 46:163–183. Lewis, T. L., D. Maurer, and K. Blackburn, 1985. The development of young infants’ ability to detect stimuli in the nasal visual field. Vis. Res. 25:943–950. Lewis, T. L., D. Maurer, and H. P. Brent, 1995. The development of grating acuity in children treated for unilateral or bilateral congenital cataract. Invest. Ophthalmol. Vis. Sci. 36:2080– 2095. Li, R., and D. Levi, 2004. Characterizing the mechanisms of improvement for position discrimination in adult amblyopia. J. Vis. 4:476–487. Lundvall, A., and U. Kugelberg, 2002a. Outcome after treatment for congenital bilateral cataract. Acta Ophthalmol. Scand. 80:593–597. Lundvall, A., and U. Kugelberg, 2002b. Outcome after treatment of congenital unilateral cataract. Acta Ophthalmol. Scand. 80:588–592. Macchi Cassia, V., C. Turati, and F. Simion, 2004. Can a nonspecific bias toward top-heavy patterns explain newborns’ face preference? Psychol. Sci. 15:379–383. Magnusson, G., M. Abrahamsson, and J. Sjöstrand, 2002. Changes in visual acuity from 4 to 12 years of age in children operated for bilateral congenital cataracts. Br. J. Ophthalmol. 86:1385–1398. Mallach, M., U. Chakravarthy, and P. Hart, 2000. Amblyopia: Is visual loss permanent? Br. J. Ophthalmol. 84:952–956. Mandavilli, A., 2006. Look and learn. Nature 441:271–272. (News feature.) Maunsell, J. H. R., and W. T. Newsome, 1987. Visual processing in monkey extrastriate cortex. Annu. Rev. Neurosci. 10:363–401. Maurer, D., 1983. The scanning of compound figures by young infants. J. Exp. Child Psychol. 35:437–448.

Maurer, D., D. Ellemberg, and T. L. Lewis, 2006. Repeated measurements of contrast sensitivity reveal limits to visual plasticity after early binocular deprivation in humans. Neuropsychologia 44:2104–2112. Maurer, D., D. Ellemberg, and T. L. Lewis, in preparation. Repeated measures of contrast sensitivity after early monocular deprivation in humans. Maurer, D., R. Le Grand, and C. Mondloch, 2002. The many faces of configural processing. Trends Cogn. Sci. 6:255–260. Maurer, D., and T. L. Lewis, 1993. Visual outcomes after infantile cataract. In K. Simons, ed., Early Visual Development, Normal and Abnormal, 454–484. Committee on Vision, Commission on Behavioral and Social Sciences and Education, National Research Council. New York: Oxford University Press. Maurer, D., and T. L. Lewis, 1998. Overt orienting toward peripheral stimuli: Normal development and underlying mechanisms. In J. Richards, ed., Eye Movements, Attention and Development, 51–102. Mahwah, NJ: Erlbaum. Maurer, D., and T. L. Lewis, 2001a. Visual acuity and spatial contrast sensitivity: Normal development and underlying mechanisms. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 237–251. Cambridge, MA: MIT Press. Maurer, D., and T. L. Lewis, 2001b. Visual acuity: The role of visual input in inducing postnatal change. Clin. Neurosci. Res. 1:239–247. Maurer, D., T. L. Lewis, and H. P. Brent, 1989. The effects of deprivation on human visual development: Studies of children treated for cataracts. In F. J. Morrison, C. E. Lord, and D. P. Keating, eds., Applied Developmental Psychology, 139–227. San Diego: Academic Press. Maurer, D., T. L. Lewis, H. P. Brent, and A. V. Levin, 1999. Rapid improvement in the acuity of infants after visual input. Science 286:108–110. Maurer, D., T. L. Lewis, and C. Mondloch, 2005. Missing sight: Consequences for visual cognitive development. Trends Cogn. Sci. 9:144–151. Maurer, D., C. J. Mondloch, and T. L. Lewis, 2007. Sleeper effects. Dev. Sci. 10:40–47. Maurer, D., and P. Salapatek, 1976. Developmental changes in the scanning of faces by young infants. Child Dev. 47:523–527. Mayer, D. L., A. S. Beiser, A. F. Warner, E. M. Pratt, K. N. Raye, and J. M. Lang, 1995. Monocular acuity norms for the Teller acuity cards between ages one month and four years. Invest. Ophthalmol. Vis. Sci. 36:671–685. Mayer, D. L., and V. Dobson, 1982. Visual-acuity development in infants and young children, as assessed by operant preferential looking. Vis. Res. 22:1141–1151. McKone, E., and B. Boyer, 2006. Sensitivity of 4-year-olds to relational changes in face distinctiveness. J. Exp. Child Psychol. 94:134–162. Meissner, C. A., and J. C. Brigham, 2001. Thirty years of investigating the own-race bias in memory for faces. Psychol. Public Policy Law 7:3–35. Michel, C., R. Caldara, and B. Rossion, in press. Same-race faces are perceived more holistically than other-race faces. Visual Cogn. Michel, C., B. Rossion, J. Han, C.-S. Chung, and R. Caldara, 2006. Holistic processing is finely tuned for faces of our own race. Psychol. Sci. 17:608–615. Mintz-Hittner, H., and K. Fernandez, 2000. Successful amblyopia therapy initiated after age 7. Arch. Ophthalmol. 118:1535– 1541.

Mioche, L., and M. Perenin, 1986. Central and peripheral residual vision in humans with bilateral deprivation amblyopia. Exp. Brain Res. 62:259–272. Mondloch, C. J., S. Geldart, D. Maurer, and R. Le Grand, 2003. Developmental changes in face processing skills. J. Exp. Child Psychol. 86:67–84. Mondloch, C. J., R. Le Grand, and D. Maurer, 2002. Configural face processing develops more slowly than featural face processing. Perception 31:553–566. Mondloch, C. J., R. Le Grand, and D. Maurer, 2003. Early visual experience is necessary for the development of some—but not all—aspects of face processing. In O. Pascalis and A. Slater, eds., The Development of Face Processing in Infancy and Early Childhood, 99–117. New York: Nova. Mondloch, C. J., A. Leis, and D. Maurer, 2006. Recognizing the face of Johnny, Suzy, and me: Insensitivity to the spacing among features at four years of age. Child Dev. 77:234–243. Mondloch, C. J., T. L. Lewis, D. Budreau, R., D. Maurer, J. L. Dannemiller, B. R. Stephens, and K. A. KleinerGathercoal, 1999. Face perception during early infancy. Psychol. Sci. 10:419–422. Mondloch, C. J., T. Pathman, D. Maurer, R. Le Grand, and S. de Schonen, 2007. The composite face effect in six-year-old children: Evidence of adultlike holistic face processing. Visual Cogn. 15:564–577. Moscovitch, M., G. Winocur, and M. Behrman, 1997. What is special about face recognition? Nineteen experiments on a person with visual object agnosia and dyslexia but normal face recognition. J. Cogn. Neurosci. 9:555–604. Murphy, K., and D. Mitchell, 1987. Reduced visual acuity in both eyes of monocularly deprived kittens following a short or long period of reverse occlusion. J. Neurosci. 7:1526– 1536. Neville, H., 2006. Variability in human brain plasticity. In M. H. Johnson and Y. Munakata, eds., Attention and Performance XXI: Processes of Change in Brain and Cognitive Development, 287–314. Oxford, UK: Oxford University Press. Newsome, W. T., and E. B. Pare, 1988. A selective impairment of motion perception following lesions of the middle temporal visual area. J. Neurosci. 8:2201–2211. Ohzawa, I., and R. D. Freeman, 1988. Binocularly deprived cats: Binocular tests of cortical cells show regular patterns of interaction. J. Neurosci. 8:2507–2516. Oliver, M., R. Neumann, Y. Chairmovitch, N. Gotesman, and M. Shimshoni, 1986. Compliance and results of treatment for amblyopia in children more than 8 years old. Am. J. Ophthalmol. 102:340–345. Parrish, E. E., D. E. Giaschi, C. Boden, and R. Dougherty, 2005. The maturation of form and motion perception in school age children. Vis. Res. 45:827–837. Pascalis, O., M. de Haan, and C. A. Nelson, 2002. Is face processing species specific during the first year of life? Science 296:1321–1323. Pascalis, O., E. Demont, M. de Haan, and R. Campbell, 2001. Recognition of faces of different species: A developmental study between 5 and 8 years of age. Infant Child Dev. 10:39–45. Pascalis, O., L. S. Scott, D. J. Kelly, R. W. Shannon, E. Nicholson, M. Coleman, and C. A. Nelson, 2005. Plasticity of face processing in infancy. Proc. Natl. Acad. Sci. USA 102:5297– 5300. Pascual-Leone, A., and R. Hamilton, 2001. The metamodal organization of the brain. Prog. Brain Res. 134:427–445.

maurer, lewis, and mondloch: plasticity of the visual system

435

Pediatric Eye Disease Investigative Group, 2005. Randomized trial of treatment of amblyopia in children aged 7 to 17 years. Arch. Opthalmol. 123:437–447. Pellicano, E., and G. Rhodes, 2003. Holistic processing of faces in preschool children and adults. Psychol. Sci. 14:618– 622. Petersen, S. E., D. L. Robinson, and J. N. Currie, 1989. Influences of lesions of parietal cortex on visual spatial attention in humans. Exp. Brain Res. 76:267–280. Pezdek, K., I. Blandon-Gitlin, and C. Moore, 2003. Children’s face recognition memory: More evidence for the cross-race effect. J. Appl. Psychol. 88:760–763. Polat, U., T. Ma-Naim, M. Belkin, and D. Sagi, 2004. Improving vision in adult amblyopia by perceptual learning. Proc. Natl. Acad. Sci. USA 101:6692–6697. Posner, M. I., J. A. Walker, F. A. Friedrich, and R. D. Rafal, 1987. How do the parietal lobes direct covert attention? Neuropsychologia 25:135–145. Rahl, J., S. Logan, M. Borja, C. Timms, I. Russell-Eggitt, and D. Taylor, 2002. Prediction of improved vision in the amblyopic eye after visual loss in the non-amblyopic eye. Lancet 360: 621–622. Regal, D. M., 1981. Development of critical flicker frequency in human infants, Vis. Res. 21:549–555. Rhodes, G., W. Hayward, and C. Winkler, 2006. Expert face coding: Configural and component coding of own-race and other-race faces. Psychon. Bull. & Rev. 13:499–505. Ridderinkhof, K. R., and M. W. Vandermolen, 1995. A psychophysiological analysis of developmental differences in the ability to resist interference. Child Dev. 66:1040–1056. Robbins, R., and E. McKone, 2007. No facelike processing for objects of expertise in three behavioral tasks. Cognition 103: 34–79. Röder, B., F. Rösler, and H. Neville, 2000. Event-related potentials during auditory language processing in congenitally blind and sighted people. Neuropsychologia 38:1482–1502. Röder, B., O. Sock, S. Bien, H. Neville, and F. Rösler, 2002. Speech processing activates visual cortex in congenitally blind humans. Eur. J. Neurosci. 16:930–936. Roessler, J. S., and J. L. Dannemiller, 1997. Changes in human infants’ sensitivity to slow displacements over the first 6 months. Vis. Res. 37:417–423. Romano, P. E., J. A. Romano, and J. E. Puklin, 1975. Stereoacuity development in children with normal binocular single vision. Am. J. Ophthalmol. 79:966–971. Sacks, O., 2006. Stereo Sue: Why two eyes are better than one. New Yorker, June 19, pp. 64–73. Sadato, N., T. Okada, M. Honda, and Y. Yonekura, 2002. Critical period for cross-modal plasticity in blind humans: A functional MRI study. NeuroImage 16:389–400. Salapatek, P., 1968. Visual scanning of geometric figures by the human newborn. J. Comp. Physiol. Psychol. 66:247–258. Salapatek, P., 1975. Pattern perception in early infancy. In L. B. Cohen and P. Salapatek, eds., Infant Vision: From Sensation to Cognition, New York: Academic Press. Salapatek, P., and W. Kessen, 1966. Visual scanning of triangles by the human newborn. J. Exp. Child Psychol. 3:155–167. Salapatek, P., and W. Kessen, 1973. Prolonged investigation of a plane geometric triangle by the human newborn. J. Exp. Child Psychol. 15:22–29. Sangrigoli, S., and S. de Schonen, 2004a. Effect of visual experience on face processing: A developmental study of inversion and nonnative effects. Dev. Sci. 7:74–87.

436

neural plasticity in development

Sangrigoli, S., and S. de Schonen, 2004b. Recognition of ownrace and other-race faces by three-month-old infants. J. Child Psychol. Psychiatry 45:1219–1227. Sangrigoli, S., C. Pallier, A. M. Argenti, V. A. G. Ventureyra, and S. de Schonen, 2005. Reversibility of the other-race effect in face recognition during childhood. Psychol. Sci. 16:440– 444. Segalowitz, S. J., C. J. Mondloch, A. Freire, D. Maurer, and J. Dywan, 2005 (April). The N170 response to standard versus scrambled Mooney faces: The eyes don’t have it! Poster presented to the meeting of Cognitive Neuroscience Society, New York. Simion, F., V. Macchi Cassia, C. Turati, and E. Valenza, 2001. The origins of face perception: Specific versus nonspecific mechanisms. Infant Child Dev. 10:59–65. Simion, F., C. Turati, E. Valenza, and I. Leo, 2006. The emergence of cognitive specialization in infancy: The case of face preference. In Y. Munakata and M. Johnson, eds., Attention and Performance. XXI: Processes of Change in Brain and Cognitive Development, 189–208. Oxford, UK: Oxford University Press. Simion, F., E. Valenza, V. Macchi Cassia, C. Turati, and C. Umiltà, 2002. Newborns’ preference for up-down asymmetrical configurations. Dev. Sci. 5:427–434. Simmers, A., and L. Gray, 1999. Improvement of visual function in an adult amblyope. Optom. Vis. Sci. 76:82–87. Slater, A., G. Bremner, S. Johnson, P. Sherwood, R. Hayes, and E. Brown, 2000. Newborn infants’ preference for attractive faces: The role of internal and external facial features. Infant Child Dev. 1:265–274. Steinmetz, M. A., and C. Constantinidis, 1995. Neurophysiological evidence for a role of posterior parietal cortex in redirecting visual attention. Cerebral Cortex 5:448–456. Tanaka, J. W., 2001. The entry point of face recognition: Evidence for face expertise. J Exp. Psychol. [Gen.] 130:534–543. Tanaka, J. W., and M. J. Farah, 1993. Parts and wholes in face recognition. Q. J. Exp. Psychol. [A] 46:225–245. Tanaka, J. W., J. B. Kay, E. Grinnell, B. Stansfield, and T. Szechter, 1998. Face recognition in young children: When the whole is greater than the sum of its parts. Visual Cogn. 5:479– 496. Tanaka, J. W., M. Kiefer, and C. M. Bukach, 2004. A holistic account of the own-race effect in face recognition: Evidence from a cross-cultural study. Cognition 93:B1–B9. Théoret, H., L. Meerabet, and A. Pascual-Leone, 2004. Behavioral and neuroplastic changes in the blind: Evidence for functionally relevant cross-modal interactions. J. Physiol. (Paris) 98:221–233. Tomac, S., and Y. Altay, 2000. Near stereoacuity: Development in preschool children; normative values and screening for binocular vision abnormalities; a study of 115 children. Binocular Vis. Strabismus Q. 15:221–228. Turati, C., 2004. Why faces are not special to newborns: An alternative account of the face preference. Curr. Dir. Psychol. Sci. 13:5–8. Turati, C., V. Macchi Cassia, F. Simion, and I. Leo, 2006. Newborns’ face recognition: Role of inner and outer facial features. Child Dev. 77:297–311. Tytla, M., T. L. Lewis, D. Maurer, and H. P. Brent, 1991. Peripheral contrast sensitivity in children treated for congenital cataract. Invest. Ophthalmol. Vis. Sci. 32:819. Tytla, M. E., D. Maurer, T. L. Lewis, and H. P. Brent, 1988. Contrast sensitivity in children treated for congenital cataract. Clin. Vis. Sci. 2:251–264.

Valenza, E., F. Simion, V. M. Cassia, and C. Umiltà, 1996. Face preference at birth. J. Exp. Psychol. [Hum. Percept.] 22:892– 903. Van Hof–van Duin, J., and G. Mohn, 1986. The development of visual acuity in normal full-term and preterm infants. Vis. Res. 26:909–916. Vereecken, E., and P. Brabant, 1984. Prognosis for vision in amblyopia after the loss of the good eye. Arch. Ophthalmol. 102: 220–224. Volkmann, F. C., and M. V. Dobson, 1976. Infant responses of ocular fixation to moving visual stimuli. J. Exp. Child Psychol. 22:86–99. Wattam-Bell, J., 1991. Development of motion-specific cortical responses in infancy. Vis. Res. 31:287–297. Wattam-Bell, J., 1994. Coherence thresholds for discrimination of motion direction in infants. Vis. Res. 34:877–883.

Wilson, H. R., 1999. Non-Fourier cortical processes in texture, form, and motion perception. In P. S. Ulinski, E. G. Jones, and A. Peters, eds., Cerebral Cortex, vol. 13, pp. 445–447. New York: Plenum Publishers. Wilson, M., 1992. Adult amblyopia reversed by contralateral cataract formation. J. Pediatr. Ophthalmol. Strabismus 29:100– 102. Yaka, R., U. Yinon, and Z. Wollberg, 1999. Auditory activation of cortical visual areas in cats after early visual deprivation. Eur. J. Neurosci. 11:1301–1312. Young, A. W., D. Hellawell, and D. C. Hay, 1987. Configurational information in face perception. Perception 16:747–759. Zhou, Y., C. Huang, P. Xu, L. Tao, Z. Oiu, X. Li, and Z. Lu, 2006. Perceptual learning improves contrast sensitivity and visual acuity in adults with anisometropic amblyopia. Vis. Res. 56:739–750.

maurer, lewis, and mondloch: plasticity of the visual system

437

26

Cross-Modal Plasticity in Development: The Case of Deafness TERESA V. MITCHELL

Multisensory interactions are pervasive in brain structure and function. Early hypotheses posited that the senses functioned essentially independently, yet substantial research from humans and other animals now shows that the senses are quite interdependent, regardless of the significant differences between them in the nature of the signal and signal transduction (Ghazanfar and Schroeder, 2006). For example, a surprisingly large number of neurons in cat and primate visual cortex are responsive to auditory input and roughly code for auditory space (Morrell, 1972; Rockland and Ojima, 2003). Such interconnections within the brain enable multisensory integration, whereby information from one sensory modality affects perception of information from another. Auditory spatial localization can be strongly influenced by visual information. Exposure to multiple presentations of a visual flash and a simultaneous auditory tone that are consistently spatially displaced by a few degrees influences subjects’ ability to localize auditory tones immediately following that exposure period (Recanzone, 1998). If the flash is always presented leftward of the tone during the exposure period, then subjects will tend to mislocalize simple auditory tones in the leftward direction during the following test period. Auditory information can similarly influence visual temporal rate perception. For example, if a subject is presented with three flashes of a light accompanied by five auditory tones, he or she behaves as if more than three flashes were presented (Recanzone, 2003). In sum, these extensive interactions between vision and audition within the nervous system affect behavior and development. If multisensory interactions are so prevalent, then the absence of one sensory modality is likely to affect the structure and function of the remaining modalities. In this chapter we will consider how the absence of auditory information affects the development of visual attention and perception. Interactions between audition and vision allow organisms with healthy sensory systems to divide attention to the environment between the sensory modalities. If a task requires focused visual attention, for example, audition can assume the responsibility of alerting the organism to events outside that attentional spotlight. When audition is absent, however,

this division of labor becomes untenable—vision becomes responsible both for focused attention and for monitoring events elsewhere in the visual field. In this instance, narrowly focusing visual attention on task-related information may be too costly because of the loss of information about unattended events. Thus deafness imposes specific and unique functional pressures on the visual system. The question is whether these functional pressures over time result in plastic changes to the structure and function of the remaining, intact sensory systems. The literature on blindness shows that analogous functional pressures placed on the auditory system affect tactile and auditory perception and attention as well as the structure and function of the visual cortex. In some instances, compensatory changes occur such that auditory functioning (Heil et al., 1991; Rauschecker and Kniepert, 1994; Röder et al., 1999), and tactile functioning (Grant, Thiagarajah, and Sathian, 2000; Van Boven et al., 2000; Hotting and Röder, 2004) exceed what is typically observed in normally sighted organisms. Furthermore, visual cortex in blind individuals is activated by auditory stimuli (Kujala et al., 1995) and tactile stimuli (Gizewski et al., 2003), as well as spoken language (Röder et al., 2002) and Braille (Sadato et al., 1996). This literature strongly suggests that functional needs imposed by the absence of visual information result in changes in the remaining modalities. This chapter will present research documenting specifically how deafness affects the development of visual processing and its neural organization. This research suggests that visual plasticity is an important adaptation that has implications for the development of individuals with hearing loss.

Effects of deafness on processing of visual motion Behavioral, electrophysiological, and brain imaging studies have all shown that deafness affects the processing of visual motion and the organization of its neural substrates. Early studies (Neville and Lawson, 1987a, 1987b, 1987c) employed the event-related potential (ERP) technique to examine whether deafness or native use of sign language affects the

439

processing of visual motion. Hearing and congenitally, genetically deaf subjects were presented with small white rectangles either in the center of a computer monitor or 18 degrees to the left or right periphery. Deaf and hearing subjects were similarly fast and accurate to report the direction of motion of the squares when presented in the central visual field (CVF), but deaf adults were faster and more accurate to report the direction of motion of the squares in the peripheral visual field (PVF). Results of the ERP data showed that stimuli presented in the CVF elicited similar amplitudes and latencies of the N1 component across the subject groups. By contrast, PVF stimuli evoked a significantly larger N1 in deaf than in hearing subjects. Furthermore, while N1 amplitudes were largest over occipital cortex in hearing adults, amplitudes were largest over more anterior temporal and parietal sites in deaf adults. This difference in distribution of the N1 suggested that visual stimuli elicited activity in a more distributed network in deaf adults, which may include typical auditory brain regions. These population differences were not observed in typically hearing adults who were born into deaf families and learned American Sign Language (ASL) as their native language (Neville and Lawson, 1987c). Thus these phenomena are related to deafness itself and not to the use of a visuospatial language. Results analogous to the Neville and Lawson (1987a, 1987b, 1987c) findings have been documented with functional MRI, a technique that allows for greater localization of brain activity. Moving random dot fields elicit greater activity in deaf than in hearing adults when attention is directed to the PVF, but not when attention is directed to the CVF or across the entire visual field (Bavelier et al., 2000, 2001). When attending to the periphery of the flow field, deaf adults displayed more activation than hearing adults within the middle temporal gyrus (MT), and displayed additional activation in the posterior parietal cortex (PPC) and the superior temporal sulcus (STS). These three areas are all involved in processing spatial dynamics. Area MT is highly responsive to any type of visual motion (Tootell et al., 1995; Beauchamp, Cox, and DeYoe, 1997), and the PPC is implicated in a variety of tasks requiring spatial analysis (Anderson, Essick, and Siegel, 1985; Bushara et al., 1999). Additional research has also documented that visual motion in the PVF activates temporal brain regions in deaf adults (Finney, Fine, and Dobkins, 2001; Fine et al., 2005). This temporal cortex activation suggests that auditory brain regions undergo cross-modal plasticity and come to process visual information. Thus the behavioral enhancements in motion processing in the deaf that were observed in Neville and Lawson (1987b) are associated not only with greater activation in typical visual-motion-sensitive areas, but also with additional activation in parietal and temporal auditory brain regions with which they are reciprocally connected. Finally, as with the earlier ERP findings (Neville and Lawson,

440

neural plasticity in development

1987c), these population differences were related to deafness and not to the use of a visuospatial language. Hearing signers produced activation similar to hearing nonsigners in area MT when attending to peripheral visual motion, and did not produce activation in STS and PPC (Bavelier et al., 2001). An important question is whether cross-modal plasticity in deafness is specific to some visual processes and not others. To examine this issue, normally hearing adults and congenitally, genetically deaf adults were presented with apparent motion and color change stimuli in an ERP paradigm (Armstrong et al., 2002). The motion stimulus was a lowspatial-frequency grayscale grating in which the bars made occasional rightward movements; ERPs were time locked to this rightward movement of the bars. The color stimulus was a high-spatial-frequency grating of blue and green bars, and ERPs to this stimulus were time locked to a fast color change of the green bars to red. Subjects responded when a stimulus was replaced by a black square. Color elicited similar amplitudes and latencies of the N1 component in deaf and hearing adults. Motion, however, elicited reliably larger N1 amplitudes in deaf than in hearing adults (see figure 26.1). Furthermore, the N1 elicited by motion was larger in deaf than in hearing participants in medial and anterior sites. The results of this ERP study show that, even when participants are not required to attend to the visual motion itself, it elicits greater neural activity in deaf than in hearing adults. These results also demonstrate that cross-modal plasticity in deafness selectively involves motion processing but not color processing. The development of these population differences is the product of long-term auditory deprivation and, in some instances, the use of a visuospatial language. The age at which reliable differences between deaf and hearing individuals are observed varies across the few studies that include young deaf children. Studies investigating higher-order attentional differences between deaf and hearing children report mixed results. One study suggested that enhanced attention to visual features does not emerge until adulthood (Rettenbach, Diller, and Sireteanu, 1999). Two studies employing continuous performance tasks, however, reported poorer performance in deaf as compared to hearing children, with deaf children more attentive to irrelevant peripheral stimuli than hearing children (Quittner et al., 1994; Mitchell and Quittner, 1996; but see Tharpe, Ashmead, and Rothpletz, 2002). Additional research investigated the effects of motion and color as distracters in a visual search task involving deaf and hearing children and adults (Mitchell and Smith, 1996; Mitchell, 1996). This study was designed to test the hypothesis that, on the one hand, if enhanced attention to motion is obligatory in the deaf, then motion should capture attention and suppression of attention to it should be difficult. On the other hand, if attention to color is not obligatory in this population, it should not capture attention and suppression of attention to it should be observed. The developmental

Figure 26.1 Event-related potentials recorded from hearing and deaf adults in response to PVF presentations of motion and color stimuli. Hearing ERPs are shown in solid lines; deaf ERPs are shown in dashed lines. (Reprinted from Armstrong, B., H. Neville,

S. Hillyard, and T. Mitchell, 2002. Auditory deprivation affects processing of motion, but not color. Cogn. Brain Res. 14:422–434. Copyright 2002, with permission from Elsevier.)

prediction was that population differences would not be observed during early school years but would emerge by adulthood. Deaf and hearing children 6–9 years of age and adults performed two visual search tasks that required attention to shape in the presence of both color and motion distracters. In the first task, stimuli were presented in a circle in the center of the visual field. Deaf and hearing children performed similarly, but deaf adults were affected more by both distracter types than hearing adults. In the second task, stimuli were presented across the central and peripheral visual fields. This extension of the stimulus display revealed that both deaf children and adults were more distracted by the task-irrelevant information than hearing children and adults. Thus, across the two tasks, attentional capture by task-irrelevant information is more similar in deaf and hearing children than it is in deaf and hearing adults. Population differences in school-aged children were only observed

when stimuli extended into peripheral visual space, indicating that deafness may affect sensitivity to peripheral space earlier than it affects attentional capture (see the next section). Finally, this study indicated that both color and motion captured attention and were difficult to ignore, a finding which suggests that deafness enhances attention to any change within the visual field. To further investigate the development of effects of auditory deprivation on visual motion processing, data from 20 congenitally, genetically deaf children and 20 hearing children ages 6 through 10 were collected using the same color and motion ERP paradigm employed by Armstrong and colleagues (2002), described earlier (Mitchell and Neville, 2002). Results were similar to those in the adult study. Color stimuli elicited similar N1 amplitudes and latencies across the two groups. By contrast, motion stimuli presented in the CVF elicited larger N1 amplitudes in deaf than hearing

mitchell: cross-modal plasticity and deafness

441

children than motion presented in the PVF, specifically in RH electrodes. Motion presented in the periphery elicited larger overall amplitudes from deaf children than from hearing children. These results, along with those described previously, suggest that some effects of cross-modal plasticity that are documented in adults emerge in the early school years, but others are slowly developing and dependent upon protracted experience with hearing loss or natural signed language.

Effects of deafness and sign language on face processing More recent research has shown that the effects of deafness on visual processing extend to face perception. Individuals with hearing loss must rely heavily on facial expressions for the kind of emotional and social information generally gleaned from tone of voice. Those who communicate with natural signed languages like ASL also extract crucial semantic and syntactic information from the face (Corina, 1989; Reilly, McIntire, and Bellugi, 1990; Snitzer Reilly, McIntire, and Bellugi, 1990). In light of these functional pressures, several studies have investigated whether deafness and sign language use affect the structure and function of face processing. Deaf children and adults are better than hearing nonsigners at recognizing faces when stimuli are upright, but not when they are inverted (Bellugi et al., 1990; Bettger et al., 1997). Hearing adults who are native ASL users are also better than hearing nonsigners, but deaf adults who communicate orally and never learned sign language perform like hearing nonsigners (Parasnis et al., 1996). These findings suggest that the use of a natural signed language can enhance face recognition skills as well. When subjects are required to perform a subtler face-matching task in which two faces might differ only in the identity of a single feature (e.g., eyes, mouth, nose), deaf signers are also better than both hearing native signers and nonsigners (McCullough and Emmorey, 1997), particularly when the mouth differs. Thus this effect is not traceable to experience with ASL, but it may be due to deafness itself or experience with speech-reading. Finally, when processing Mooney faces, which are devoid of individual facial features but maintain the first-order configuration of a face, with two eyes above a mouth, deaf signers are slightly worse than hearing nonsigners at categorizing them by age and gender (there were no data from hearing signers on this task) (McCullough and Emmorey, 1997). The fact that deaf signers excelled in processing individual features but not in processing Mooney faces suggests that deafness may enhance featural or analytic face processing but not holistic/configural processing. Featural/analytic face processing requires attending selectively to individual parts or features of the face (Farah et al., 1998; Mondloch, Le Grand, and Maurer, 2002). This can be contrasted with holistic/con-

442

neural plasticity in development

figural processing, which is characterized by encoding of the face as a whole stimulus on the basis of its overall configuration and is accompanied by difficulties attending selectively to parts of faces (Farah et al., 1998; Mondloch, Le Grand, and Maurer, 2002). Together, this group of studies suggests that deafness may enhance attention to the bottom of the face and may enhance featural/analytic processing but not holistic/configurational processing. Deafness affects not only behavioral responses to faces, but also the neural substrates of face perception. The typical neural substrate of face processing involves greater and faster activity in the right hemisphere (RH) than the left hemisphere (LH; de Schonen et al., 1993; Puce et al., 1996). Behavioral studies of visual field asymmetries demonstrate that deaf children respond similarly to faces presented in the left and right visual fields, in contrast to the leftvisual-field/RH asymmetry observed in hearing children (Szelag and Wasilewski, 1992; Szelag, Wasilewski, and Fersten, 1992). Functional MRI work reports that hearing subjects produce more activation in the RH than the LH in response to faces, while deaf subjects produce more activation in the LH than the RH (McCullough, Emmorey, and Sereno, 2005). This left lateralization in deaf subjects was greater in response to linguistic facial expressions that occur in ASL than in response to canonical emotional expressions. Thus deafness and sign language use reorganize the neural substrates of face processing by recruiting greater activation of LH structures. This LH recruitment may be driven by the kind of heavier reliance on local, featural/analytic processing in the deaf that was described earlier, as this type of processing is associated with greater LH activation in typically hearing individuals (Martinez et al., 1997; Moses et al., 2002). The stronger activation of LH structures may also be driven by the recruitment of language areas in the LH into basic face processing, as a result of the central role of facial expressions in ASL (Corina, 1989; Reilly, McIntire, and Bellugi, 1990; Snitzer Reilly, McIntire, and Bellugi, 1990). Recent work from our laboratory extends the results of these studies by further exploring featural/analytic and holistic/configural processing in the deaf. Severely to profoundly deaf signers and typically hearing children and adults completed two face-processing tasks. The first task investigated effects of face inversion. Inversion affects face processing much more than it affects object processing (Yin, 1969). This effect is considered an index of configural processing; face processing mechanisms are tuned to stimulus orientation and cannot work quickly and accurately when the typical configuration is disrupted (Carey and Diamond, 1977). Face inversion elicited greater decrements in accuracy for deaf children and adults than for hearing children and adults (Mitchell, More, and Van der Heide, 2005). Furthermore, a RH asymmetry was observed in ERP data collected from hearing adults, but similar ERP amplitudes and latencies were

observed across the two hemispheres in deaf adults (Letourneau, Thrasher, and Mitchell, 2006). These findings indicate that deafness increases sensitivity to face orientation and provide further evidence that LH structures are activated as strongly as RH structures in face processing in the deaf. Holistic face processing was also investigated by using composite faces, created by joining the top and bottom halves of two different faces. Subjects made same/different judgments based on only the top or bottom half of the faces. In this task, holistic processing is indexed by the effect of the unattended half on judgments of the attended half. Results revealed that hearing adults were influenced by the unattended half of the faces—that is, they processed them holistically, regardless of which half was attended. Deaf adults, meanwhile, processed faces holistically when they attended to the top half, but were better than hearing adults at restricting their attention when judging the bottom half of the face stimuli (Letourneau and Mitchell, 2006). Finally, ERP results from adults again revealed a RH asymmetry for hearing adults that was absent in deaf adults. The results from these two studies agree with published data reviewed previously and extend them to suggest that deafness and sign language use lead to (1) increased sensitivity to face orientation, (2) increased involvement of LH brain regions in face processing, (3) increased feature-based face processing, and (4) increased attention to information in the bottom half of the face. Therefore, the absence of auditory input across the life span, as well as the use of a visuospatial language, affects the relative importance of particular aspects of the face and reorganizes the neural substrates of face processing. In sum, deafness affects the processing of visual motion and extends to the processing of faces but does not affect color processing. This specificity suggests that the mechanism of this plasticity may be found in the interaction between the functional needs of the organism and intrinsic characteristics of subsystems within vision. One hypothesis is that more slowly developing aspects within vision are those more likely to be affected by chronic atypical experiences such as deafness (Mitchell and Neville, 2002, 2004). Evidence suggests that motion processing continues to develop into late childhood (Mitchell and Neville, 2004), and face processing is known to develop into early adulthood (Taylor et al., 1999; Mondloch, Le Grand, and Maurer, 2002; Itier and Taylor, 2004). Furthermore, motion and face processing are known to be affected by other types of atypical experience (Atkinson et al., 2001; Spencer et al., 2000; Braddick, Atkinson, and Wattam-Bell, 2003; Le Grand et al., 2001; Pollak and Sinha, 2002). The protracted developmental time courses of motion and face processing would provide a large window of time within which auditory deprivation could affect perception and attention within these visual subsystems.

Effects of deafness on processing visual space Deafness affects not only the processing of visual features, but also the processing of visual space. Electrophysiological and fMRI studies reviewed earlier show that the greatest enhancements of visual motion processing in the deaf are observed when attention is directed to the periphery (Neville and Lawson, 1987b; Bavelier et al., 2000, 2001). Results from other studies indicate that this effect may be the product of an enhancement of motion processing as well as an enhancement of visuospatial attention to the periphery. Behavioral studies have reported that deafness enhances the processing of other stimuli in the periphery of the visual field as well. In one study, teenaged and young adult deaf and hearing subjects were presented with a single digit in the center of a computer screen that was followed by an asterisk, which appeared near the preceding digit or 25 degrees to the left or right (Loke and Song, 1991). Subjects were asked to respond upon detection of the asterisk, and then to recall the identity of the digit presented at the beginning of the trial. Deaf subjects were significantly faster than hearing subjects to respond to peripheral asterisks. A second study employed a task that presented stimuli in the center of the visual field that either correctly cued the spatial location of a subsequent peripheral target or cued the location opposite the subsequent target (Parasnis and Samar, 1985). Deaf college students were better than hearing college students at disengaging attention from an incongruently cued location to the correct target location. These studies show that deafness increases attention to a variety of peripheral visual stimuli and suggest that this increased attention may be accompanied by faster disengagement of focused, central attention. Proksch and Bavelier (2002) extended these ideas, showing that the increase in attention to peripheral events observed in deaf individuals may reduce attentional resources available to central events. Subjects were presented with a ring of 6 circles and were told to respond to a target shape that appeared in any of the possible locations. Attentional load was manipulated by presenting irrelevant shape stimuli within either 0 or 5 of the remaining circles. Results showed that attentional load and eccentricity of the distracter stimuli affected the deaf signers and hearing nonsigners. Hearing nonsigners were slower and less accurate when the distracter stimuli were presented in the periphery, while deaf subjects were slower and less accurate when the distracter stimuli were presented in the center. The authors concluded that deaf individuals distribute more attentional resources to the visual periphery but at the cost of reduced attentional resources in the center. The fact that hearing native signers performed like hearing nonsigners in this task suggests that this change in the distribution of attentional resources is due to auditory

mitchell: cross-modal plasticity and deafness

443

deprivation and not to the use of a signed language. These results support the notion that spatial attention is a limited resource such that an increase in its distribution to the periphery results in a decrease in available resources in the center. Deafness affects this balance between peripheral and central events by enhancing the processing of stimuli that appear in the periphery and possibly reducing sustained attention to central events. Why is spatial attention distributed more toward the PVF than the CVF? It may well be because peripheral visual field functions and their neural substrates have a longer sensitive period within which atypical experience can affect its development than central visual field functions (Lewis and Maurer, 2005). Alternatively, it could relate to the fact that, in the primate brain, there are projections from the auditory cortex to portions of the visual cortex that represent peripheral visual space (Falchier et al., 2002). If similar projections exist in the human brain, perhaps regions representing the visual periphery expand their representation in the absence of auditory input, and perhaps this expansion supports greater behavioral sensitivity. Finally, it is possible that sensitivity to events in the CVF may be so well developed in both deaf and hearing populations that a ceiling effect prevents the development of compensatory plasticity. A similar increase in attention to peripheral stimuli is observed in the blind (Röder et al., 1999). A general picture is emerging in which deafness increases attention to information in the PVF, possibly because many initial indications of environmental changes appear in the PVF. This increased attention may serve to alert the individual that something has entered the visual field and needs to be processed further. More work is needed to understand how the system maintains the balance between this alerting function in the periphery and task-related focused attention in the center, how this balance is affected by task demands, and how these effects emerge with development.

Visual plasticity in language: American Sign Language (ASL) This review has thus far described how deafness affects the development of visual functions over the life span, but deafness also clearly impacts the development of language. Deafness significantly impacts the development of spoken language, and deaf individuals often use sign language as their primary means of communication. Children with significant hearing loss can develop speech production and perception skills with intensive practice and training, and some individual deaf children function successfully in the hearing world (Geers, Moog, and Schick, 1984; Geers and Moog, 1992; Moog and Geers, 1999). For others, the visuospatial nature of ASL means that it can be acquired easily, given early and sufficient exposure to a fluently signing model (Singleton et al., 1998). In this section we will explore

444

neural plasticity in development

how the organization and development of ASL are impacted by the visual and manual modalities. The original research documenting that ASL was a full natural language was published in the 1970s (Bellugi and Fischer, 1972; Bellugi and Klima, 1972; Lane, Boyes-Braem, and Bellugi, 1976; Tweney, Heiman, and Hoemann, 1977; see Poizner, Klima, and Bellugi, 1987 for review). This work was an important breakthrough in documenting that ASL was more than just a system of gestures but possessed as full and generative a grammar as spoken and written languages. It was also a breakthrough in demonstrating that speech is not somehow privileged in language development and that the human facility for language is robust enough to enable reliance on the hands and eyes rather than the voice and ears to convey concrete and abstract information and ideas. Important similarities exist between ASL and spoken languages, despite the significant differences in their modes of communication. The classic LH brain regions that are activated by spoken languages are also activated by ASL (Poizner, Klima, and Bellugi, 1987; Bavelier et al., 1998), even when signs are articulated bimanually or with the left hand (Corina et al., 2003). Damage to the LH in adult signers results in aphasia that is highly similar to aphasias observed in talkers in which word order is scrambled and semantics is disordered (Poizner, Klima, and Bellugi, 1987). These LH lesions specifically impact language functions and do not impair the use of simple gestures (Corina, 1992). Furthermore, as in spoken languages, semantic and syntactic processes in ASL are subserved by nonidentical neural pathways. Semantic processing is associated with activity in posterior temporal-parietal regions of the brain, while syntactic processing is associated with activity in right frontal regions, and this observation is made for ASL (Bavelier et al., 1998), as well as for spoken and written languages (Neville, Mills, and Lawson, 1992; Nobre and McCarthy, 1994; Newman et al., 2001). Semantic and syntactic processes are also differently affected by late language acquisition. For both ASL and spoken language, late learners can achieve high levels of semantic processing but typically do not acquire high levels of syntactic processing (Weber-Fox and Neville, 1996; Neville et al., 1997). Thus, despite significant differences in modalities, major principles of the neural organization of language are maintained across signing, speech, and text. The visuospatial nature of ASL does give rise to unique characteristics that are not shared with spoken and written languages. Significant activation of the RH is seen in the perception of ASL (Bavelier et al., 1998; Neville et al., 1998). These RH regions include homologues of Broca’s and Wernicke’s areas, as shown in figure 26.2 and plate 46, as well as posterior temporal and parietal regions. Activation of these areas has been linked to the spatial aspects of ASL grammar and discourse. For example, in ASL actors and objects are first prespecified by establishing a location for

Figure 26.2 Functional MRI activation in Broca’s and Wernicke’s areas in the LH and their homologues in the RH, as elicited by written English in native English speakers (top) and ASL in native signers (bottom). Activation is depicted in bar graphs, as percent signal change as well as spatial extent (mm3). (Reprinted

with permission from Bavelier, D., D. Corina, P. Jezzard, V. Clark, A. Karni, A. Lalwani, J. P. Rauschecker, A. Braun, R. Turner, and H. J. Neville, 1998. Hemispheric specialization for English and ASL: Left invariance–right variability. NeuroReport 9(7):1537–1542.) (See plate 46.)

each within the signing space, and then the action is described by referring back to those locations in specific movements and sequences. Lesions to the RH in signers result in a unique aphasia in which the consistency and accuracy of such spatial relationships are disrupted (Corina and McBurney, 2001). Finally, recruitment of these regions for ASL processing is dependent upon early acquisition of the language; RH activation is not observed in individuals who learn ASL after puberty (Neville et al., 1997; Newman et al., 2002). The recruitment of the RH into the neural substrate of ASL is likely due to its typical role in visual spatial processing. An important piece of evidence that ASL is a natural language is that it is acquired in infancy at the same rate and sequence as spoken language, as long as the crucial element— a fluently signing adult—is present (Bellugi and Klima, 1972). Milestones such as babbling, first words, and first word strings are acquired at a similar rate and sequence in the acquisition of ASL as in the acquisition of speech (Vol-

terra, 1987; Newport, 1988; Meier and Newport, 1990; Petitto and Marentette, 1991). There is a slight advantage in the appearance of first signs as compared to first words. First signs appear earlier in development than first words, because motor control of the hands emerges earlier than motor control of the vocal apparatus, and, by extension, early vocabulary is built at a faster rate (Meier and Newport, 1990; Goodwyn and Acredolo, 1993; Volterra and Iverson, 1995). Early two- and three-sign strings obey appropriate subject-verb-object word orderings in ASL (Newport and Meier, 1985). Thus the visual and manual nature of ASL does not preclude its developing at a rate and sequence similar to those of spoken language. In sum, ASL as a visuospatial language does not differ from spoken language in its overall organization. Both language modalities recruit the same anterior and posterior LH structures, and rely on nonidentical networks to subserve semantic and syntactic processing. Both languages are

mitchell: cross-modal plasticity and deafness

445

acquired in highly similar rates and sequences and are optimally learned early in life. The unique visual and spatial properties of ASL do affect its structure and function. Significant recruitment of the RH is seen in native users of the language, and this recruitment is tied to the unique spatial aspects of ASL grammar. Thus the natural human capacity for language learning is not necessarily limited by the absence of auditory information but can be fully expressed in the visual and manual modalities.

Cross-modal plasticity in speech perception following cochlear implantation The advent of cochlear implants (CIs) has had a dramatic impact on the ability of deaf children to learn spoken language. Implants provide direct stimulation of the auditory nerve through electrical impulses delivered by electrodes placed within the cochlea. This stimulation can restore hearing to those who lost it adventitiously or introduce it to those with congenital hearing loss. Recipients of CIs must adjust to their new auditory input, regardless of whether their deafness occurred prelingually or postlingually. Changes in visual perception and attention take place during the preceding periods of deafness, and these changes turn out to play an important role in the adjustment to input from the CI. Cross-modal plasticity in visual cortex that occurs during periods of postlingual deafness is associated with better outcomes in speech perception after CI. Giraud and colleagues documented that purely auditory speech elicits visual cortex activation in both hearing controls and adult, postlingually, adventitiously deafened CI recipients (Giraud et al., 2001) using positron emission tomography (PET). Implant recipients produced greater activation in calcarine cortex than normally hearing controls, and this activation evolved over the three years that subjects were followed postimplantation; visual cortex activity in the CI subjects elicited specifically by auditory speech increased over time, while activity elicited by noise did not. Finally, the CI recipients with the best speech-reading skills showed the greatest activation in areas V1 and V2 elicited by speech. Implant subjects had at most five years of deafness prior to implantation, and at most three years of CI use. These short periods of deprivation and restoration indicate fast reorganization of cross-modal activity in visual cortex even in adulthood, which stands in contrast to the more slowly evolving plasticity described in the earlier sections of this chapter. The existence of cross-modal activity in hearing subjects suggests that preexisting speechreading skills that are linked to activity in the visual cortex can be enhanced in postlingually deafened adults, and that this enhancement predicts better outcomes after implantation. Further research indicated that the activity in visual cortex becomes tuned to the individuals’ known languages

446

neural plasticity in development

(Giraud and Truy, 2002). Known but not unknown languages activated V1/V2, the left posterior inferior temporal gyrus, and the left superior parietal cortex in both CI recipients and controls. CI recipients showed additional activation in the right perirhinal and fusiform gyri, right temporaloccipital-parietal junction, and Broca’s area. Activation of the fusiform gyrus, a highly face-sensitive area, in response to one’s native language suggests that CI recipients imagine the facial gestures of speech when listening to speech. In sum, these cross-modal effects in postlingually deafened CI recipients represent a reactivation of preexisting speechrelated activity in these areas, as well as the functional tuning of this activity to one’s own language. Data from speech perception tasks postimplantation illustrate that improvements in visual speech perception during periods of deafness have a positive impact on the development of speech perception following implantation. Adult CI users and typically hearing adults were compared on their perception of monosyllabic words in three conditions: auditory only (A), visual only (V), and audiovisual (AV) (Kaiser et al., 2003). Background data provided on the 20 CI users showed that most were deafened in adulthood (three in childhood), the average length of CI use was 39.8 months (seven had experienced deafness for one year or less prior to receiving their implant), and the average duration of deafness was 11 years. Auditory stimuli were presented to hearing adults in noise to avoid ceiling effects; therefore, statistical comparisons were only performed on visual performance. The hearing adults and CI users showed a similar pattern of performance, being best in the AV condition and worst in the V-alone condition. However, CI recipients were better than hearing subjects at perceiving words in the V-alone condition. For each subject, a visual gain score was calculated to index the degree to which visual information in the AV condition increased speech recognition over performance in the A-alone condition. Implant users tended to have higher visual gain scores than hearing subjects. Once again, these findings indicate that postlingually deafened adults can enhance their visual perception of speech even during relatively short periods of auditory deprivation. Most of the CI users were deafened in adulthood and had spent at least 30 years with healthy hearing, establishing strong memory and representations of auditory and visual speech. Thus deafness enhanced visual speech perception, and this enhancement, in turn, facilitated the integration of auditory input introduced by the CI. Data from deaf youngsters who receive CIs shows that the interactions between audition and vision are different after long periods of deafness, as well as after prelingual deafness. Longer periods of deafness prior to implantation are associated with poorer speech perception postimplantation (Nikolopoulos, O’Donoghue, and Archbold, 1999). One

possible explanation for this outcome is that visual takeover of auditory brain regions interferes with the ability of the auditory cortex to integrate auditory input. To test this hypothesis, Lee and colleagues measured glucose metabolism at rest using PET as a rough index of general neural activity in prelingually deaf children who received CIs (Lee et al., 2001, 2005, 2007). Metabolic levels in various cortical regions were then correlated with speech perception skills assessed postimplantation. Longer periods of prelingual deafness were associated in these studies with greater metabolic activity in ventral occipital, frontal, and temporal auditory brain regions, as well as lower scores on speech-perception tasks (Lee et al., 2001, 2007). These data suggest that longer periods of prelingual deafness allow functional takeover of auditory brain regions, with the assumption that the nature of the takeover is visual. The correlations with behavior further suggest that this functional takeover interferes with the ability of the auditory cortex to process the new auditory information provided by the implant. Finally, increased metabolism in frontal and parietal brain regions in prelingually deaf children who receive CIs is associated with better speech-processing outcomes, while increased metabolism in ventral visual regions is associated with poorer outcomes (Lee, Kang et al., 2005). These latter results suggest that enhanced neural activity in visual brain regions and possible takeover of auditory brain regions by visual processing in prelingual deafness may interfere with the incorporation of new auditory input by a CI. Prelingually deafened children, like postlingually deafened adults, develop better auditory perception outcomes postimplantation if they develop visual skills in speech reading while deaf (Bergeson, Pisoni, and Davis, 2005). Profoundly, prelingually deaf children who received their CI by age 9 participated in a five-year study and were divided into groups based on communication mode (oral versus total communication) and age at implantation (by median split). The Common Phrases (CP) test, an open set test of sentence comprehension, was used preimplantation as well as postimplantation to assess speech perception in A-alone, V-alone, and AV conditions. Performance in all conditions of the CP test improved over the five-year period for all subjects, as did measures of visual and auditory gain. Age at implantation affected performance on the CP task. Late-implanted children (average age for OC, 73 months; for TC, 74 months) initially performed better than earlyimplanted children (average age for OC, 36 months; for TC, 38 months), but by four years postimplantation, the two groups showed similar levels of performance. The two groups had similar AV scores, but early-implanted children had higher A-alone and auditory-gain scores, while late-implanted children had higher V-alone and visual-gain scores across the conditions. These results suggest that age at implantation impacts which mode is primarily relied upon for speech

perception: late-implanted children continue to rely on the visual modality, which likely served them well prior to implantation, and early-implanted children come to rely on the auditory modality for speech processing. Mode of communication also affected performance in these tasks. When tested before or soon after implantation, children who relied on oral communication (OC) were better at all speech-perception conditions of the CP test than those who relied on total communication (TC), which combines sign and speech. However, the two groups performed similarly by the end of the study. Children relying on OC also had marginally better visual and auditory gain scores than TC children. The early advantage observed in OC children may be due to the fact that most oralaural education methods strongly emphasize the development of visual speech-perception skills (i.e., speech-reading) and training to capitalize upon any residual hearing. Totalcommunication programs, however, vary to a great degree in the emphasis placed on oral-aural training as compared to training in their chosen manually coded English sign system (note that these TC programs do not employ natural signed languages). Therefore, early visual speech training may have facilitated the OC children’s superior performances on the CP test at the earlier intervals. Finally, performance on the V-alone condition of the CP task preimplantation was the best predictor of outcome measures in this study, including measures of purely auditory speech perception (Bergeson, Pisoni, and Davis, 2005). This finding, as well as those from postlingually deafened adults reviewed earlier, suggests that speech-perception skills in one modality transfer, to some degree, to speech-perception skills in the other modality. Thus visual speech perception may be as important to this population as training in auditory speech perception, and it is likely to produce overall gains in spoken-language comprehension. Recent research with very young CI recipients indicates a sensitive period in the development of the ability to fuse auditory and visual information into a single percept. The fusion of auditory and visual speech information is best illustrated with the McGurk paradigm (McGurk and MacDonald, 1976). The McGurk effect is the finding that when people see a person articulate one phoneme, for example /ga/, but hear the person say another phoneme /ba/, they perceive a third phoneme /da/. This effect illustrates that auditory and visual speech signals are not simply added but are fused to create a new percept. Young typically hearing infants are also susceptible to the effect, although it appears that the influence of visual information on the effect increases with age (McGurk and MacDonald, 1976; Rosenblum, Schmuckler, and Johnson, 1997). Young children between the ages of 5 and 14 years who were born deaf and later received unilateral CIs were tested in a classic McGurk paradigm (McGurk and MacDonald,

mitchell: cross-modal plasticity and deafness

447

1976; Schorr et al., 2005). When presented with incongruous auditory and visual speech input, those subjects who received their implants before 30 months of age were capable of perceiving the fusion. Those who received their implants later than 30 months did not perceive the fusion, and their perception of the event was dominated by the visually presented phoneme. Dominance of the event by visual information in later-implanted children may be due to their reliance on vision for communication, prior to as well as postimplantation. The timing of the sensitive period for auditory/visual speech fusion (30 months) is similar to the 36-month sensitive period observed for recovery of the auditory P1 latency in very young CI recipients (Sharma, Dorman, and Spahr, 2002; Sharma, Dorman, Spahr, and Todd, 2002; Sharma, Dorman, and Kral, 2005). The latency of the auditory P1 ERP component elicited by simple phonemes can be used to index the maturation of the auditory cortex. Prelingually deaf children who receive CIs by 36 months develop typical P1 latencies, those implanted later than 7 years of age do not, and those implanted in the intermediate age range show variable recovery (Sharma, Dorman, and Spahr, 2002; Sharma, Dorman, Spahr, and Todd, 2002; Sharma, Dorman, and Kral, 2005). This auditory sensitive period may contribute to the continued reliance of lateimplanted children on visual information, as described previously (Bergeson, Pisoni, and Davis, 2005). The parallel sensitive periods observed for the McGurk effect and for auditory P1 latencies suggest that typical functioning of the auditory cortex, particularly its speed of processing, is necessary for perceptual fusion of auditory and visual speech channels. The data from CI recipients outlined in this section further illustrate that auditory deprivation increases the reliance on visual information. The enhancements in visual speech perception that develop during periods of deafness form a foundation upon which restored auditory input may be integrated once it is provided by the CIs. Developmental studies indicate that there are limitations to this interaction between vision and audition. Sensitive periods in the development of the auditory cortex impose boundaries upon the time course within which it can learn to process sound in the normal manner. This sensitive period, in turn, provides the opportunity for vision to overtake functional activity in the auditory cortex. This cross-modal plasticity is hypothesized to impose its own limitations on the ability of the auditory cortex to process sound much later in life. This sketch of the complex interplay between vision and audition awaits further research to fill in the details, but it describes an emerging picture of dynamic changes in the brain and behavior that have important clinical implications, as well as theoretical implications for plasticity and development.

448

neural plasticity in development

Conclusion Deafness clearly affects the development of visual functions. This chapter reviewed several examples of effects of deafness on visual attention and perception and their neural substrates. These effects can be considered as evidence of compensation for the absence of auditory input. The absence of audition changes the functional pressures placed on the visual system, which responds by enhancing the processing of visual motion, peripheral stimuli, and certain aspects of faces. These visual phenomena have important consequences in the life of the deaf individual because they enable the balance between focused attention and alerting to changes in the environment and because they carry significant social and emotional information, respectively. While much remains to be learned, the extant results from deaf children suggest that reorganization of vision at this level is slow to develop and continues across many years of deafness. Visual compensation for the absence of auditory input is also observed in speechreading. Even short periods of deafness induce improvements in speechreading, and this reorganization is accelerated when there is previous experience with audiovisual speech. This improvement in speechreading is accompanied by cross-modal plasticity in visual cortex, and this reorganization in turn fuels rapid integration of input from a CI. These effects are even observed in young CI recipients. These types of changes appear to be limited by sensitive periods within the auditory cortex. These sensitive periods establish a time window within which the auditory system can recover if provided auditory input, and beyond which takeover of the auditory cortex by visual input may occur if auditory input is not provided. The plasticity described in this chapter includes examples of both experience-expectant and experience-dependent plasticity (Greenough, Black, and Wallace, 1987). The sensitive period for development of the auditory cortex imposes limitations on the timing of cochlear implantation for optimal development of the auditory P1 (Sharma, Dorman, and Spahr, 2002; Sharma, Dorman, Spahr, and Todd, 2002; Sharma et al., 2005), for fusion of auditory and visual information about speech (McGurk and MacDonald, 1976; Schorr et al., 2005), and for possible takeover of auditory brain regions by visual processing (Lee et al., 2001; Sadato et al., 2004; Lee et al., 2007). These effects are examples of experience-expectant plasticity. By contrast, the protracted periods of development seen in the enhanced processing of motion, of faces, and of speech suggest that these types of plasticity are experience dependent. These atypical outcomes in brain and behavioral development are the product of atypical experience across the life span. Cross-modal compensation for deafness is therefore able to take advantage of multiple mechanisms of plasticity to result in optimal and adaptive functioning.

The cross-modal effects reviewed in this chapter also illustrate that perception depends not only upon the properties of the thing being perceived, but also on what the observer does with the thing being perceived. The visual information that enjoys enhanced processing in the deaf is equally available to hearing individuals; deafness does not enhance visual acuity or sensitivity thresholds (Finney and Dobkins, 2001; Brozinsky and Bavelier, 2004). The functional pressures placed by deafness impose lasting changes on visual processing and its neural substrate and ultimately allow the individual to adapt to its experience. This adaptation can be considered a form of perceptual learning. Perceptual learning involves lasting changes to an organism’s perceptual system that improve its ability to respond to its environment (Goldstone, 1998). Perceptual learning experiments in the laboratory manipulate subjects’ experience with stimuli over time (sometimes as short as an hour, other times as long as weeks) to observe how this experience changes the way these and other stimuli are processed. The effects documented in this chapter illustrate that long-term experiential variables such as deafness and early sign-language acquisition produce perceptual learning effects much like those observed in laboratory experiments. Deafness affects the way attention is distributed and devoted across different kinds of information. Because it is a chronic atypical experience, these attentional effects extend in many instances to the organization of the underlying neural architecture devoted to processing that information. Similar principles of plasticity are observed in studies of blind humans. Blindness enhances attention to peripheral auditory information in the azimuth (Röder et al., 1999) and to tactile information (Hotting and Röder, 2004). Further, the occipital cortex of blind adults is activated by somatosensory (Sadato et al., 1996), auditory (Weeks et al., 2000), and speech stimuli (Röder et al., 2002). The robust activation of visual cortex in the blind, in contrast to the smaller, more task-specific activation of auditory cortex in the deaf, suggests that visual cortex is more susceptible to cross-modal plasticity than auditory cortex. Thus blindness imposes its own functional pressures on audition and somatosensation that result in behavioral and brain changes like those that result from deafness, but unique properties of each modality establish nonidentical patterns of plasticity. Ultimately, the effects of unimodal sensory deprivation on behavior and brain organization illustrate the significant role of sensory experience in development. REFERENCES Anderson, R., G. K. Essick, and R. M. Siegel, 1985. Encoding of spatial location by posterior parietal neurons. Science 230: 456–458. Armstrong, B., H. Neville, S. Hillyard, and T. Mitchell, 2002. Auditory deprivation affects processing of motion, but not color. Cogn. Brain Res. 14:422–434.

Atkinson, J., S. Anker, O. Braddick, L. Nokes, A. Mason, and F. Braddick, 2001. Visual and visuospatial development in young children with Williams syndrome. Dev. Med. Child Neurol. 43(5):330–337. Bavelier, D., C. Brozinsky, A. Tomann, T. Mitchell, H. Neville, and G. Liu, 2001. Impact of early deafness and early exposure to sign language on the cerebral organization for motion processing. J. Neurosci. 21(22):8931–8942. Bavelier, D., D. Corina, P. Jezzard, V. Clark, A. Karni, A. Lalwani, J. P. Rauschecker, A. Braun, R. Turner, and H. J. Neville, 1998. Hemispheric specialization for English and ASL: Left invariance–right variability. NeuroReport 9(7):1537–1542. Bavelier, D., A. Tomann, C. Hutton, T. Mitchell, D. Corina, and H. Neville, 2000. Visual attention to the periphery is enhanced in congenitally deaf individuals. J. Neurosci. 20(RC93): 1–6. Beauchamp, M. S., R. W. Cox, and E. A. DeYoe, 1997. Graded effects of spatial and featural attention on human area MT and associated motion processing areas. J. Neurophysiol. 78:516– 520. Bellugi, U., and S. Fischer, 1972. A comparison of sign language and spoken language. Cognition 1(2–3):173–200. Bellugi, U., and E. S. Klima, 1972. The roots of language in the sign talk of the deaf. Psychology Today 6(1), June 1972: 76. Bellugi, U., L. O’Grady, D. Lillo-Martin, M. O’Grady Hynes, K. van Hoek, and D. Corina, 1990. Enhancement of spatial cognition in deaf children. In V. Volterra and C. J. Erting, eds., From Gesture to Language in Hearing and Deaf Children, 278–298. New York: Springer-Verlag. Bergeson, T. R., D. B. Pisoni, and R. A. Davis, 2005. Development of audiovisual comprehension skills in prelingually deaf children with cochlear implants. Ear Hear. 26(2):149– 164. Bettger, J., K. Emmorey, S. McCullough, and U. Bellugi, 1997. Enhanced facial discrimination: Effects of expertise with American Sign Lanuguage. J. Deaf Stud. Deaf Educ. 2(4): 223–233. Braddick, O., J. Atkinson, and J. Wattam-Bell, 2003. Normal and anomalous development of visual motion processing: Motion coherence and “dorsal-stream vulnerability.” Neuropsychologia 41(13):1769–1784. Brozinsky, C. J., and D. Bavelier, 2004. Motion velocity thresholds in deaf signers: Changes in lateralization but not in overall sensitivity. Cogn. Brain Res. 21(1):1–10. Bushara, K. O., R. A. Weeks, K. Ishii, M. J. Catalan, B. Tian, J. P. Rauschecker, and M. Hallett, 1999. Modality-specific frontal and parietal areas for auditory and visual spatial localization in humans. Nature Neurosci. 2(8):759–766. Carey, S., and R. Diamond, 1977. From piecemeal to configurational representation of faces. Science. 195(4275):312–314. Corina, D. P., 1989. Recognition of affective and noncanonical linguistic facial expressions in hearing and deaf subjects. Brain Cogn. 9(2):227–237. Corina, D. P., and S. L. McBurney, 2001. The neural representation of language in users of American Sign Language. J. Commun. Disord. 34(6):455–471. Corina, D. P., H. Poizner, U. Bellugi, T. Feinberg, D. Dowd, and L. O’Grady-Batch, 1992. Dissociation between linguistic and nonlinguistic gestural systems: A case for compositionality. Brain Lang. 43(3):414–447. Corina, D. P., L. San Jose-Robertson, A. Guillemin, J. High, and A. R. Braun, 2003. Language lateralization in a bimanual language. J. Cogn. Neurosci. 15(5):718–730.

mitchell: cross-modal plasticity and deafness

449

de Schonen, S., C. Deruelle, J. Mancini, and O. Pascalis, 1993. Hemispheric differences in face processing and brain maturation. In B. de Boysson-Bardies, S. de Schonen, P. W. Jusczyk, P. McNeilage, and J. Morton, eds. Developmental Neurocognition, Speech and Face Processing in the First Year of Life. Dordrecht: Kluwer Academic Publishers. Falchier, A., S. Clavagnier, P. Barone, and H. Kennedy, 2002. Anatomical evidence of multimodal integration in primate striate cortex. J. Neurosci. 22(13):5749–5759. Farah, M. J., K. D. Wilson, M. Drain, and J. N. Tanaka, 1998. What is “special” about face perception? Psychol. Rev. 105(3): 482–498. Fine, I., E. M. Finney, G. M. Boynton, and K. R. Dobkins, 2005. Comparing the effects of auditory deprivation and sign language within the auditory and visual cortex. J. Cogn. Neurosci. 17(10): 1621–1637. Finney, E. M., and K. R. Dobkins, 2001. Visual contrast sensitivity in deaf versus hearing populations: Exploring the perceptual consequences of auditory deprivation and experience with a visual language. Cogn. Brain Res. 11(1):171–183. Finney, E. M., I. Fine, and K. R. Dobkins, 2001. Visual stimuli activate auditory cortex in the deaf. Nature Neurosci. 4(12): 1171–1173. Geers, A. E., and J. S. Moog, 1992. Speech perception and production skills of students with impaired hearing from oral and total communication education settings. J. Speech Hear. Res. 35(6): 1384–1393. Geers, A., J. Moog, and B. Schick, 1984. Acquisition of spoken and signed English by profoundly deaf children. J. Speech Hear. Disord. 49(4):378–388. Ghazanfar, A. A., and C. E. Schroeder, 2006. Is neocortex essentially multisensory? Trends Cogn. Sci. 10(6):278–285. Giraud, A. L., C. J. Price, J. M. Graham, E. Truy, and R. S. Frackowiak, 2001. Cross-modal plasticity underpins language recovery after cochlear implantation. Neuron 30(3):657–663. Giraud, A. L., and E. Truy, 2002. The contribution of visual areas to speech comprehension: A PET study in cochlear implants patients and normal-hearing subjects. Neuropsychologia 40(9): 1562–1569. Gizewski, E. R., T. Gasser, A. de Greiff, A. Boehm, and M. Forsting, 2003. Cross-modal plasticity for sensory and motor activation patterns in blind subjects. NeuroImage 19(3): 968–975. Goldstone, R. L., 1998. Perceptual learning. Annu. Rev. Psychol. 49: 585–612. Goodwyn, S. W., and L. P. Acredolo, 1993. Symbolic gesture versus word: Is there a modality advantage for onset of symbol use? Child Dev. 64(3):688–701. Grant, A. C., M. C. Thiagarajah, and K. Sathian, 2000. Tactile perception in blind Braille readers: A psychophysical study of acuity and hyperacuity using gratings and dot patterns. Percept. Psychophys. 62(2):301–312. Greenough, W. T., J. E. Black, and C. S. Wallace, 1987. Experience and brain development. Child Dev. 58:539–559. Heil, P., G. Bronchti, Z. Wollberg, and H. Scheich, 1991. Invasion of visual cortex by the auditory system in the naturally blind mole rat. NeuroReport 2:735–738. Hotting, K., and B. Röder, 2004. Hearing cheats touch, but less in congenitally blind than in sighted individuals. Psychol. Sci. 15(1):60–64. Itier, R. J., and M. J. Taylor, 2004. Face recognition memory and configural processing: A developmental ERP study using

450

neural plasticity in development

upright, inverted, and contrast-reversed faces. J. Cogn. Neurosci. 16(3):487–502. Kaiser, A. R., K. I. Kirk, L. Lachs, and D. B. Pisoni, 2003. Talker and lexical effects on audiovisual word recognition by adults with cochlear implants. J. Speech Lang. Hear. Res. 46(2):390–404. Kujala, T., M. Huotilainen, J. Sinkkonen, A. I. Ahonen, K. Alho, M. S. Hamalainen, R. J. Ilmoniemi, M. Kajola, J. E. T. Knuutila, J. Lavikainen, O. Salonen, J. Simola, C. G. Standertskjold-Nordenstam, H. Tiitinen, S. O. Tissari, and R. Naatanen, 1995. Visual cortex activation in blind humans during sound discrimination. Neurosci. Lett. 183:143–146. Lane, H., P. Boyes-Braem, and U. Bellugi, 1976. Preliminaries to a distinctive feature analysis of handshapes in American sign language. Cogn. Psych. 8(2):263–289. Lee, D. S., J. S. Lee, S. H. Oh, S. K. Kim, J. W. Kim, J. K. Chung, M. C. Lee, and C. S. Kim, 2001. Cross-modal plasticity and cochlear implants. Nature 409(6817):149–50. Lee, H. J., A. L. Giraud, E. Kang, S. H. Oh, H. Kang, C. S. Kim, and D. S. Lee, 2007. Cortical activity at rest predicts cochlear implantation outcome. Cerebral Cortex 17:909–917. Lee, H. J., E. Kang, S. H. Oh, H. Kang, D. S. Lee, M. C. Lee, and C. S. Kim, 2005. Preoperative differences of cerebral metabolism relate to the outcome of cochlear implants in congenitally deaf children. Hear. Res. 203(1–2):2–9. Le Grand, R., C. J. Mondloch, D. Maurer, and H. P. Brent, 2001. Neuroperception: Early visual experience and face processing. Nature. 410(6831):890. Letourneau, S. M., and T. Mitchell, 2006. Effects of Deafness on Electrophysiological Markers of Holistic Face Processing. San Francisco: Cognitive Neuroscience Society. Letourneau, S. M., M. Thrasher, and T. Mitchell, 2006. Effects of deafness on configural and holistic face processing. International Conference on Cognitive and Neural Systems, Boston. Lewis, T. L., and D. Maurer, 2005. Multiple sensitive periods in human visual development: Evidence from visually deprived children. Dev. Psychobiol. 46(3):163–183. Loke, W. H., and S. Song, 1991. Central and peripheral visual processing in hearing and nonhearing individuals. Bul. Psychon. Soc. 29(5):437–440. Martinez, A., P. Moses, L. Frank, R. Buxton, E. Wong, and J. Stiles, 1997. Hemispheric asymmetries in global and local processing: Evidence from fMRI. NeuroReport 8(7):1685– 1689. McCullough, S., and K. Emmorey, 1997. Face processing by deaf ASL signers: Evidence for expertise in distinguished local features. J. Deaf Stud. Deaf Educ. 2:212–222. McCullough, S., K. Emmorey, and M. Sereno, 2005. Neural organization for recognition of grammatical and emotional facial expressions in deaf ASL signers and hearing nonsigners. Cogn. Brain Res. 22(2):193–203. McGurk, H., and J. MacDonald, 1976. Hearing lips and seeing voices. Nature 264:746–748. Meier, R. P., and E. L. Newport, 1990. Out of the hands of babes: On a possible sign language advantage in language acquisition. Language 66(1):1–23. Mitchell, T. V., 1996, How Audition Shapes Visual Attention. Bloomington, IN: Indiana University Department of Psychology. Mitchell, T., S. More, and R. Van der Heide, 2005. Effects of Deafness on the Development of Face Perception. New York: Cognitive Neuroscience Society. Mitchell, T. V., and H. J. Neville, 2002. Effects of age and experience on the development of neurocognitive systems. In A.

Zani and A. M. Proverbio, eds., The Cognitive Physiology of Mind, 225–244. New York: Academic Press. Mitchell, T. V., and H. J. Neville, 2004. Asynchrony in the development of electrophysiological responses to motion and color. J. Cogn. Neurosci. 16(8):1363–1374. Mitchell, T. V., and A. L. Quittner, 1996. A multimethod study of attention and behavior problems in hearing-impaired children. J. Clin. Child Psychol. 25:83–96. Mitchell, T. V., and L. B. Smith, 1996. Deafness drives development of attention to change. In G. Cottrell, ed., Proceedings of the Eighteenth Annual Conference of the Cognitive Science Society, 627–632. Mahwah, NJ: Erlbaum. Mondloch, C. J., R. Le Grand, and D. Maurer, 2002. Configural face processing develops more slowly than featural face processing. Perception 31(5):553–566. Moog, J. S., and A. E. Geers, 1999. Speech and language acquisition in young children after cochlear implantation. Otolaryngol. Clin. North Am. 32(6):1127–1141. Morrell, F., 1972. Visual system’s view of acoustic space. Nature 238(5358):44–46. Moses, P., K. Roe, R. B. Buxton, E. C. Wong, L. R. Frank, and J. Stiles, 2002. Functional MRI of global and local processing in children. NeuroImage 16(2):415–424. Neville, H. J., D. Bavelier, D. Corina, J. Rauschecker, A. Karni, A. Lalwani, A. Braun, V. Clark, P. Jezzard, and R. Turner, 1998. Cerebral organization for language in deaf and hearing subjects: Biological constraints and effects of experience. Proc. Natl. Acad. Sci. USA 95(3):922–929. Neville, H. J., S. A. Coffey, D. S. Lawson, A. Fischer, K. Emmorey, and U. Bellugi, 1997. Neural systems mediating American Sign Language: Effects of sensory experience and age of acquisition. Brain Lang. 57(3):285–308. Neville, H. J., and D. Lawson, 1987a. Attention to central and peripheral visual space in a movement detection task: An eventrelated potential and behavioral study. I. Normal hearing adults. Brain Res. 405(2):253–267. Neville, H. J., and D. Lawson, 1987b. Attention to central and peripheral visual space in a movement detection task: An eventrelated potential and behavioral study. II. Congenitally deaf adults. Brain Res. 405(2):268–283. Neville, H. J., and D. Lawson, 1987c. Attention to central and peripheral visual space in a movement detection task. III. Separate effects of auditory deprivation and acquisition of a visual language. Brain Res. 405(2):283–294. Neville, H. J., D. L. Mills, and D. S. Lawson, 1992. Fractionating language: Different neural subsystems with different sensitive periods. Cerebral Cortex 2(3):244–258. Newman, A. J., D. Bavelier, D. Corina, P. Jezzard, and H. J. Neville, 2002. A critical period for right hemisphere recruitment in American Sign Language processing. Nature Neurosci. 5(1):76–80. Newman, A. J., R. Pancheva, K. Ozawa, H. J. Neville, and M. T. Ullman, 2001. An event-related fMRI study of syntactic and semantic violations. J. Psycholinguist. Res. 30(3):339– 364. Newport, E. L., 1988. Constraints on learning and their role in language acquisition: Studies of the acquisition of American Sign Language. Lang. Sci. 10(1):147–172. Newport, E. L., and R. P. Meier, 1985. The acquisition of American Sign Language. In D. I. Slobin, ed., The Crosslinguistic Study of Language Acquisition, vol. 1: The Data, 881–938. Hillsdale, NJ: Erlbaum.

Nikolopoulos, T. P., G. M. O’Donoghue, and S. Archbold, 1999. Age at implantation: Its importance in pediatric cochlear implantation. Laryngoscope 109(4):595–599. Nobre, A. C., and G. McCarthy, 1994. Language-related ERPs: Scalp distributions and modulation by word type and semantic priming. J. Cogn. Neurosci. 6:233–255. Parasnis, I., and V. J. Samar, 1985. Parafoveal attention in congenitally deaf and hearing young adults. Brain Cogn. 4(3):313–327. Parasnis, I., V. J. Samar, J. G. Bettger, and K. Sathe, 1996. Does deafness lead to enhancement of visual spatial cognition in children? Negative evidence from deaf nonsigners. J. Deaf Stud. Deaf Educ. 1(2):145–152. Petitto, L. A., and P. F. Marentette, 1991. Babbling in the manual mode: Evidence for the ontogeny of language. Science 251(5000):1493–1496. Poizner, H., E. S. Klima, and U. Bellugi, 1987. What the Hands Reveal about the Brain. Cambridge, MA: MIT Press. Pollak, S. D., and P. Sinha, 2002. Effects of early experience on children’s recognition of facial displays of emotion. Dev. Psychol. 38(5):784–791. Proksch, J., and D. Bavelier, 2002. Changes in the spatial distribution of visual attention after early deafness. J. Cogn. Neurosci. 14(5):687–701. Puce, A., T. Allison, M. Asgari, J. C. Gore, and G. McCarthy, 1996. Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance imaging study. J. Neurosci. 16(16):5205–5215. Quittner, A. L., L. B. Smith, M. J. Osberger, T. V. Mitchell, and D. B. Katz, 1994. The impact of audition on the development of visual attention. Psychol. Sci. 5(6):347–353. Rauschecker, J. P., and U. Kniepert, 1994. Auditory localization behaviour in visually deprived cats. Eur. J. Neurosci. 6(1):149– 160. Recanzone, G. H., 1998. Rapidly induced auditory plasticity: The ventriloquism aftereffect. Proc. Natl. Acad. Sci. USA 95(3): 869–875. Recanzone, G. H., 2003. Auditory influences on visual temporal rate perception. J. Neurophysiol. 89(2):1078–1093. Reilly, J. S., M. L. McIntire, and U. Bellugi, 1990. The acquisition of conditionals in American Sign Language: Grammaticized facial expressions. Appl. Psycholinguist. 11(4):369–392. Rettenbach, R., G. Diller, and R. Sireteanu, 1999. Do deaf people see better? Texture segmentation and visual search compensate in adult but not in juvenile subjects. J. Cogn. Neurosci. 11(5):560–583. Rockland, K. S., and H. Ojima, 2003. Multisensory convergence in calcarine visual areas in macaque monkey. Int. J. Psychophysiol. 50(1–2):19–26. Röder, B., O. Stock, S. Bien, H. Neville, and F. Rosler, 2002. Speech processing activates visual cortex in congenitally blind humans. Eur. J. Neurosci. 16(5):930–936. Röder, B., W. Teder-Salejarvi, A. Sterr, F. Rosler, S. A. Hillyard, and H. J. Neville, 1999. Improved auditory spatial tuning in blind humans. Nature 400(6740):162–166. Rosenblum, L. D., M. A. Schmuckler, and J. A. Johnson, 1997. The McGurk effect in infants. Percept. Psychophys. 59(3):347–357. Sadato, N., A. Pascual-Leone, J. Grafman, V. Ibanez, M. Deiber, G. Dold, and M. Hallett, 1996. Activation of the primary visual cortex by Braille reading in blind subjects. Nature 380:526–528. Sadato, N., H. Yamada, T. Okada, M. Yoshida, T. Hasegawa, K. Matsuki, Y. Yonekura, and H. Itoh, 2004. Age-dependent

mitchell: cross-modal plasticity and deafness

451

plasticity in the superior temporal sulcus in deaf humans: A functional MRI study. BMC Neurosci. 5(1):56. Schorr, E. A., N. A. Fox, V. van Wassenhove, and E. I. Knudsen, 2005. Auditory-visual fusion in speech perception in children with cochlear implants. Proc. Natl. Acad. Sci. USA 102(51): 18748–18750. Sharma, A., M. F. Dorman, and A. Kral, 2005. The influence of a sensitive period on central auditory development in children with unilateral and bilateral cochlear implants. Hear. Res. 203(1–2):134–143. Sharma, A., M. F. Dorman, and A. J. Spahr, 2002b. A sensitive period for the development of the central auditory system in children with cochlear implants: Implications for age of implantation. Ear Hear. 23(6):532–539. Sharma, A., M. Dorman, A. Spahr, and N. W. Todd, 2002. Early cochlear implantation in children allows normal development of central auditory pathways. Ann. Otol. Rhinol. Laryngol. Suppl. 189:38–41. Singleton, J. L., S. Supalla, S. Litchfield, and S. Schley, 1998. From sign to word: Considering modality constraints in ASL/English bilingual education. Topics Lang. Disord. 18(4): 16–29. Snitzer Reilly, J., M. L. McIntire, and U. Bellugi, 1990. Faces: The relationship between language and affect. In V. Volterra and C. G. Erting, eds., From Gesture to Language in Hearing and Deaf Children, 128–141. New York: Springer-Verlag. Spencer, J., J. O’Brien, K. Riggs, O. Braddick, J. Atkinson, and J. Wattam-Bell, 2000. Motion processing in autism: Evidence for a dorsal stream deficiency. NeuroReport 11(12): 2765–2767. Szelag, E., and R. Wasilewski, 1992. The effect of congenital deafness on cerebral asymmetry in the perception of emotional and non-emotional faces. Acta Psychologica 79(1):45–57. Szelag, E., R. Wasilewski, and E. Fersten, 1992. Hemispheric differences in the perceptions of words and faces in deaf and hearing children. Scand. J. Psychol. 33(1):1–11.

452

neural plasticity in development

Taylor, M. J., G. McCarthy, E. Saliba, and E. Degiovanni, 1999. ERP evidence of developmental changes in processing of faces. Clin. Neurophysiol. 110(5):910–915. Tharpe, A. M., D. H. Ashmead, and A. M. Rothpletz, 2002. Visual attention in children with normal hearing, children with hearing aids, and children with cochlear implants. J. Speech Lang. Hear. Res. 45:403–413. Tootell, R. B. H., J. B. Reppas, K. K. Kwong, R. Malach, R. T. Born, T. J. Brady, B. R. Rosen, and J. W. Belliveau, 1995. Functional analysis of human MT and related visual cortical areas using magnetic resonance imaging. J. Neurosci. 15(4):3215– 3230. Tweney, R. D., G. W. Heiman, and H. W. Hoemann, 1977. Psychological processing of sign language: Effects of visual disruption on sign intelligibility. J. Exp. Psychol. [Gen.] 106(3):255–268. Van Boven, R. W., R. H. Hamilton, T. Kauffman, J. P. Keenan, and A. Pascual-Leone, 2000. Tactile spatial resolution in blind Braille readers. Neurology 54(12):2230–2236. Volterra, V., 1987. From single communicative signal to linguistic combinations in hearing and deaf children. In J. Montangero, A. Tryphon, and S. Dionnet, eds., Symbolism and Knowledge, 89– 106. Geneva: Jean Piaget Archives Foundation. Volterra, V., and J. M. Iverson, 1995. When do modality factors affect the course of language acquisition? In K. Emmorey and J. Snitzer Reilly, eds., Language, Gesture, and Space, 371–390. Hillsdale, NJ: Lawrence Erlbaum. Weber-Fox, C., and H. J. Neville, 1996. Maturational constraints on functional specializations for language processing: ERP and behavioral evidence in bilingual speakers. J. Cogn. Neurosci. 8(3):231–256. Weeks, R., B. Horwitz, A. Aziz-Sultan, B. Tian, C. M. Wessinger, L. G. Cohen, M. Hallett, and J. P. Rauschecker, 2000. A positron emission tomographic study of auditory localization in the congenitally blind. J. Neurosci. 20(7):2664–2672. Yin, R. K., 1969. Looking at upside-down faces. J. Exp. Psychol. 81(1):141–145.

27

Plasticity of Speech (Animal Model) TERESA A. NICK

Similar to humans, songbirds must learn their vocalizations (Thorpe, 1958; Marler and Peters, 1977). Babies and songbirds first memorize speech or song sounds, respectively, during a sensory phase (Doupe and Kuhl, 1999). Later, both groups use auditory feedback to match the memorized sounds and gradually shape their motor skills during a sensorimotor critical period (Konishi, 1965; Kuhl, 2000). Birdsong and human speech share many other remarkable similarities, such as innate preferences for species-typical sounds and strong effects of social context (for review, see Doupe and Kuhl, 1999). Along with these well-described behavioral similarities, birdsong has provided an excellent model of speech for many other reasons, including the following: (1) the brain areas that produce song are discrete clusters of cells (“nuclei”) that enable precise identification, labeling, lesioning, and recording of known modules in a complex network (Nottebohm, Stokes, and Leonard, 1976); (2) some songbirds learn only a short (∼1.5 sec) phrase or motif and repeat it thousands of times per day, a fact that enables thorough analysis of both behavior (Tchernichovski et al., 2001) and underlying neural activity (Crandall, Aoki, and Nick, 2007); and (3) song learning can be manipulated in a variety of ways, such as rearing in isolation or castration. This chapter focuses on the development of song behavior and underlying neural dynamics. For a review of the major contributions of birdsong to the field of adult neurogenesis, the reader is referred to Nottebohm (2004).

General features of song behavior Why sing? As with human speech, birdsong is important for interspecies communication. Song is used to attract potential mates and to deter territorial intruders. The production of song is affected by social context and by the behavior of listeners. Hearing song affects the recipient. Songs of individual birds can be recognized by other birds and even by a skilled human. Some songbirds have a large repertoire of incredibly complex songs, whereas others, such as the zebra finch, produce long repetitions of a very short and simple “motif,” consisting of approximately one second of learned, stereotyped sounds.

The oscine songbirds do not have an in-born or innate ability to sing their species-typical song. They must learn their song through imitation. Early in life, typically before 100 days of age, songbirds memorize the song of a tutor or tutors. This memorization phase, termed the sensory phase, also occurs in humans. During the sensory phase, elements or even entire songs from individual tutors are memorized. After memorization of a tutor song, the bird uses this explicit tutor song memory to perfect his skill at producing his own song using auditory feedback (Konishi, 1965) during the “sensorimotor phase,” which also has a human analogue. During the sensorimotor phase, the bird may synthesize elements from several tutors’ songs into an entirely new song of his own creation or learn to produce an exact copy of a single tutor’s song. Some songbirds, such as the zebra finch, produce a rather invariant “crystallized” song throughout their adult lives after a developmental period of extreme vocal plasticity. In these “closed-ended” learners, adult song shows very little plasticity except in cases of extreme perturbation, such as deafening (Nordeen and Nordeen, 1992; Leonardo and Konishi, 1998; Brainard and Doupe, 2000; Zevin, Seidenberg, and Bottjer, 2004). In contrast, “open-ended” learners show much more capacity for song plasticity throughout their lives. The classic example of an open-ended learner is the canary, which shows seasonal parallel plasticity in brain and behavior (Brenowitz, 2004). In many songbird species, only the male sings. Song behavior and underlying neural structures are affected by circulating gonadal hormones (Brenowitz, 2004). Estrogens synthesized within the brain appear to play a dominant role in masculinizing the brain during development (Wade and Arnold, 2004). The role of hormones and genetics in the plasticity of song behavior and the song control system have been discussed in a number of excellent recent reviews (Ball et al., 2004; Brenowitz, 2004; Harding, 2004; Wade and Arnold, 2004).

The neural song system There is a clear brain-behavior relationship that underlies vocal learning and production in songbirds. Birds that learn

453

Figure 27.1 Simplified schematic of the neural pathways involved in song learning and production. For clarity, not all connections or brain areas are shown. The brain areas of the classical song system are represented by the gray shaded circles (Reiner, Perkel, et al., 2004). Major subdivisions of the brain are represented by the large ovals. The brain areas involved in song learning and production may be thought of as three major loops, shown here concentrically. The outermost loop includes song behavior, auditory feedback, and the nuclei that receive sound stimuli and produce song. The intermediate motor loop, indicated by the bold arrows, contains the

brain areas that are essential for song production (Nottebohm, Stokes, and Leonard, 1976; Coleman and Vu, 2005). The innermost loop contains the anterior forebrain pathway, which enables song plasticity (Bottjer, Miesner, and Arnold, 1984; Brainard and Doupe, 2000) by inducing variability in behavior (Kao, Doupe, and Brainard, 2005; Olveczky, Andalman, and Fee, 2005; Kao and Brainard, 2006). Notably, the area X-to-DLM connection is inhibitory (as indicated by a closed circle instead of an arrowhead; Luo and Perkel, 1999).

their song possess a set of brain areas collectively known as the song system. In contrast, birds that do not possess a robust song system do not learn their song. This generality holds for comparative studies of songbirds that are oscine (learn song) versus those that are suboscine (do not learn song) (Kroodsma and Konishi, 1991; Gahr, 2000). In addition, in many songbird species, singing is sexually dimorphic: males sing and have a robust song system, whereas females do not sing and possess a relatively small or nonexistent song system. Hormonal manipulation of either sex induces corresponding changes in brain and behavior (Harding, 2004; Wade and Arnold, 2004). The telencephalon of mammals and birds consists of subpallium (basal ganglia) and pallium (which includes hippocampus and cortex). Unlike mammals, the pallium of birds is not layered, but instead consists of a system of nuclei. The clustering of functionally related neurons into nuclei reduces

the length of connections and may have evolved to decrease the weight of the brain as an adaptation for flight. The nuclear structure has enabled the detailed study of microcircuits within the functional modules of the song system. The song system consists of seven nuclei (figure 27.1, gray circles; table 27.1). Four of the song nuclei are pallial or cortical-like (table 27.1, figure 27.1; HVC, LMAN, NIf, and RA), two are thalamic (DLM and Uva), and one is in the basal ganglia (area X). In addition to the song nuclei, seven brain-stem nuclei are known to have roles in song production (for review, see Wild, 2004), and two pallial nuclei are known to have roles in song learning (NCM and CM; Bolhuis and Gahr, 2006). Collectively, these brain areas form three major loops (figure 27.1), providing a framework for understanding the basic structure and function of the song system. All these loops appear to be involved in song learning. The outermost loop enables sensorimotor

454

neural plasticity in development

Table 27.1 Brain areas involved in song learning and production Acronym Proper Name* Brain Region Known or Suspected Function CM Caudal mesopallium Pallium Conspecific song memory (Mello, Vicario, and Clayton, 1992; Mello and Clayton, 1994; Gentner and Margoliash, 2003) Tutor song memory (Bolhuis and Gahr, 2006) DLM Medial nucleus of the Thalamus Plasticity (Bottjer, Miesner, and Arnold, 1984; dorsolateral thalamus Brainard and Doupe, 2000) Inducing behavioral variability (by extrapolation from LMAN results) HVC HVC Pallium Song production (Nottebohm, Stokes, and Leonard, 1976) Source of auditory input to downstream nuclei (Williams and Nottebohm, 1985; Doupe and Konishi, 1991; Vicario and Yohay, 1993) Coding of syllables or motifs (Yu and Margoliash, 1996; Hahnloser, Kozhevnikov, and Fee, 2002) Pattern generation (Mooney and Prather, 2005; Solis and Perkel, 2005) Interconnected Dorsomedial nucleus of Brain stem Song production (Ashmore, Wild, and Schmidt, 2005) medullary and the intercollicular Song timing (Ashmore, Wild, and Schmidt, 2005) midbrain nuclei complex; n. paraambigualis; n. retroambigualis LMAN Lateral magnocellular Pallium Plasticity (Bottjer, Miesner, and Arnold, 1984; nucleus of the Brainard and Doupe, 2000) anterior nidopallium Inducing behavioral variability (Kao, Doupe, and Brainard, 2005; Liu and Nottebohm, 2005; Olveczky, Andalman, and Fee, 2005; Kao and Brainard, 2006) NCM Caudal medial Pallium Conspecific song memory (Mello, Vicario, and nidopallium Clayton, 1992; Mello and Clayton, 1994; Bolhuis et al., 2000; Stripling, Kruse, and Clayton, 2001; Phan, Pytte, and Vicario, 2006) Tutor song memory (Bolhuis and Gahr, 2006; Phan, Pytte, and Vicario, 2006) NIf Interfacial nucleus of the Pallium Source of auditory input to song system (Nottebohm, nidopallium Kelley, and Paton, 1982; Janata and Margoliash, 1999; Coleman and Mooney, 2004; Cardin, Raksin, and Schmidt, 2005) RA Robust nucleus of the Pallium Song production (Nottebohm, Stokes, and Leonard, arcopallium 1976) Coding of notes (Yu and Margoliash, 1996) Uva Nucleus uvaeformis Thalamus Song production (Coleman and Vu, 2005) X Area X within songbird Basal Plasticity (Bottjer, Miesner, and Arnold, 1984; medial striatum ganglia Brainard and Doupe, 2000) Inducing behavioral variability (by extrapolation from LMAN results) *For a detailed discussion of the revised song system nomenclature, see Reiner, Perkel, et al., 2004.

integration (NCM, CM, NIf, HVC, RA, singing, sound). The motor loop (indicated by bold arrows in figure 27.1; Uva, HVC, RA, brain-stem nuclei) consists of the four brain areas that are known to be critical for song production (Nottebohm, Stokes, and Leonard, 1976; Vu, Mazurek, and Kuo, 1994; Ashmore, Wild, and Schmidt, 2005; Coleman and Vu,

2005). The innermost loop, the anterior forebrain pathway (area X, DLM, LMAN), is not necessary for song production (Bottjer, Miesner, and Arnold, 1984) but is crucial for song learning (Bottjer, Miesner, and Arnold, 1984) and adult plasticity (Brainard and Doupe, 2000). These three loops will be considered separately, but the reader should note that all

nick: plasticity of speech (animal model)

455

these loops are interconnected and that not all connections and brain areas are shown in figure 27.1. The Sensorimotor Integration Loop The sensorimotor integration loop consists of the brain areas potentially involved in auditory feedback and in translating that auditory feedback into changes in behavior. There has been no conclusive report of neural auditory feedback in the song system (Konishi, 2004). The study of auditory feedback requires recording from awake, behaving songbirds, which is technically difficult. Further, study of the role of auditory feedback in song learning requires recording from birds in the process of learning. In the most popular electrophysiological model, the zebra finch, there is no evidence that adults can learn new song material (Zevin, Seidenberg, and Bottjer, 2004). Thus adult changes in behavior induced by deafening or perturbation of auditory feedback are not necessarily learning, but may instead result from a maintenance process that is revealed by extreme experimental manipulation. In this case, learning may occur only in juveniles, a fact that adds an additional level of difficulty for electrophysiologists because juveniles are harder to record from because of their weaker skulls, neck musculature, and constitution. Because of this difficulty, most of the published studies of auditory responses in the songbird brain were executed in adult, anesthetized birds. However, technological advances in chronic recording (McCasland and Konishi, 1981; Yu and Margoliash, 1996; Schmidt and Konishi, 1998; Fee and Leonardo, 2001) have begun to make possible reliable recordings from unanesthetized birds during learning. Specifically, improvement in commutator design (Schmidt and Konishi, 1998), miniaturization of the headstage electronics and movable drives (Schmidt and Konishi, 1998; Fee and Leonardo, 2001), and introduction of neural ensemble recording techniques (Crandall, Aoki, and Nick, 2007) have all worked together to make it possible to record the single-neuron activity in awake behaving juveniles as they learn their songs (Olveczky, Andalman, and Fee, 2005; Crandall, Aoki, and Nick, 2007). In the most widely studied songbird model, the zebra finch, it appears that the pattern of song system responses to auditory playback during slow-wave sleep (Nick and Konishi, 2001, 2005b) resembles that which is obtained during anesthesia (Margoliash, 1983; Volman, 1993; Doupe, 1997; Mooney, 2000), with playback of the bird’s own song (BOS; the specific and unique song that the individual finch produces) always inducing the largest response. Perhaps surprisingly, auditory responses in awake adults are weak relative to those obtained during sleep and anesthesia and only slightly greater to BOS compared to other stimuli (Rauske, Shea, and Margoliash, 2003; Nick and Konishi, 2005a). These studies and others that compared waking and anesthesia or waking and presumed sleep indicate that responses

456

neural plasticity in development

to auditory stimuli are gated by the bird’s cognitive state (Dave, Yu, and Margoliash, 1998; Schmidt and Konishi, 1998). A longitudinal study in juveniles found that the response to BOS changed as the BOS itself changed (Nick and Konishi, 2005b), suggesting that the BOS response is tightly linked to sensorimotor learning and may reflect a memory of sensorimotor relationships. In contrast, auditory responses during waking are greatest to the song of the adult male song tutor (usually the father) in juveniles (Nick and Konishi, 2005a), suggesting the presence of a sensory memory trace. Altogether, these data suggest that the song system integrates sensory and motor activities, but the exact function of these responses, if any, is still under investigation. In addition to electrophysiological recordings, the measurement of neural auditory responses during singing and awake passive listening has also been achieved by monitoring activity-dependent gene expression (Mello, Vicario, and Clayton, 1992; Mello and Clayton, 1994). Mello and colleagues played back various auditory stimuli, including conspecific songs (songs from members of the same species), and examined expression of the immediate early gene ZENK (acronym for the conserved gene known as zif-268, egr-1, NGF1-A, or Krox-24). Using this measure, neural regions were identified that were selectively activated by passive playback of conspecific song but not by ethologically irrelevant stimuli such as tone bursts (Mello, Vicavio, and Clayton, 1992; Mello and Clayton, 1994). Strikingly, these selectively activated brain regions were not part of the song system. Instead, the brain areas that responded most to passive playback of conspecific song were part of the ascending auditory system (Theunissen et al., 2004), specifically the auditory processing areas L1 and L3 within Field L and the Field L targets: NCM, CM, the shelf region adjacent to HVC, and the cup region adjacent to RA (Mello, Velho, and Pinaud, 2004). Although activity-dependent gene expression does not report all neural activity, it pinpoints brain regions that have enough electrical activity to activate calciumdependent signaling pathways where they exist. The ZENK pathway is activated in the song system during singing (Jarvis and Nottebohm, 1997), indicating that the appropriate pathways exist but are not strongly activated by auditory stimulation. The area NIf provides the primary auditory input to HVC (Nottebohm, Kelley, and Paton, 1982; Janata and Margoliash, 1999; Coleman and Mooney, 2004) but does not play a critical role in song production (Cardin, Raskin, and Schmidt, 2005). The area NIf receives auditory input from the auditory area CM (Vates et al., 1996; Theunissen et al., 2004) and may also receive connections directly from nucleus ovoidalis (Wild, 2004), the principal auditory relay in the thalamus (Wild, 1993). In addition, NIf receives signals from the motor loop via the thalamic nucleus Uva

(Nottebohm, Kelley, and Paton, 1982). If the tutor-songmatching signal is generated afferent of HVC, then recordings from NIf and/or CM in awake finches may reveal a preference for tutor song during the early sensorimotor phase. Interpretations of these and other future results should consider that accumulating data suggest that auditory stimulus preference and, by extrapolation, the tutor song memory, may be distributed across multiple auditory processing areas (Theunissen et al., 2004; Woolley, Gill, and Theunissen, 2006). For auditory signals to guide the development of song, they must detect similarities or differences between auditory feedback and the tutor-song memory (Konishi, 1965, 2004). Recent data indicate that neurons within HVC respond selectively to tutor song during the early sensorimotor phase, when young finches actively match their vocalizations to the learned tutor song (Nick and Konishi, 2005a). The brain region(s) that originates the tutor-song-matching signal is not known. Recent studies have provided intriguing clues that CM and NCM may store the tutor-song memory: CM is involved in the storage of conspecific song memories (Gentner and Margoliash, 2003), the CM of females responds selectively to playback of their father’s song (Terpstra et al., 2006), activity-dependent gene expression in NCM is positively correlated with the strength of song learning (Terpstra, Bolhuis, and den Boer–Visser, 2004), and NCM neurons habituate faster to novel song than to tutor song, suggesting a long-term memory of tutor song (Phan, Pytte, and Vicario, 2006). In addition, zebra finches that sing (males) have more calbindin-positive neurons in NCM that those that do not (females), suggesting sexually dimorphic perceptual differences that parallel song-learning abilities (Pinaud et al., 2006). Collectively, these data suggest that the memory of the tutor song is stored in CM and/or NCM (Bolhuis and Gahr, 2006) and passed to the motor loop via the CM-NIfHVC pathway. The Motor Loop A memory of a species-typical sound (the tutor-song memory that we have discussed) is formed during the sensory phase and apparently stored in auditory processing areas. It is, in essence, an explicit or declarative memory, since the sequence of memorized tutor-song sounds can be declared. In contrast, the ability to emit speciestypical vocalizations (the memory of the bird’s own song) is an implicit or procedural memory. It is a skill that the bird acquires during the sensorimotor phase after much practice and shaping toward the memorized tutor song. These are different types of memory that are acquired at different times in development. Thus a growing sentiment in the field is that different anatomical regions underlie tutor-song memory as compared to the BOS memory. Specifically, the auditory processing areas discussed previously may store the tutorsong memory, whereas the BOS memory may be stored in

the song system itself (Nick and Konishi, 2005a, 2005b; Bolhuis and Gahr, 2006). Lesioning NIf (Cardin, Raksin, and Schmidt, 2005) or nuclei of the anterior forebrain pathway (Bottjer, Miesner, and Arnold, 1984; Brainard and Doupe, 2000) does not result in immediate degradation of song. Thus the memory of the BOS is most likely contained within the nuclei that are essential for singing: HVC, RA, Uva, and seven brainstem nuclei (three of which are interconnected and able to transmit signals to Uva; Wild, 2004). Bold arrows indicate the motor loop in figure 27.1. Bilateral lesions of HVC and RA obliterate singing (Nottebohm, Stokes, and Leonard, 1976), and stimulation of HVC (Vu, Mazurek, and Kuo, 1994) or RA (Ashmore, Wild, and Schmidt, 2005) during singing resets the song. Lesions of Uva are more difficult because of its size and location, but partial bilateral lesions of this nucleus in three finches resulted in severe impairment of singing (Coleman and Vu, 2005). Although lesions of midbrain and medullary nuclei have not been done, a recent study indicated that stimulation of the respiratory nucleus paraambigualis, one of the three interconnected nuclei that transmit signals to Uva, induced cessation of singing (Ashmore, Wild, and Schmidt, 2005). These data indicate that the song motor control loop consists of Uva, HVC, RA, and three interconnected brain-stem nuclei. Since vocalization must be coordinated with breathing, feedback from the brain-stem nuclei may participate in the control of timing in the thalamic and pallial song control nuclei. Recordings of HVC during development reveal that responses to the BOS change dynamically in parallel with the song produced (Nick and Konishi, 2005b). This finding suggests that the BOS response is related to song production and that HVC stores or receives a copy of the current BOS. The area HVC contains a complex microcircuitry (Mooney, 2000; Mooney and Prather, 2005), which suggests that this brain area is capable of storing aspects of the BOS. In addition, available evidence suggests that HVC may have a role in pattern generation (Mooney and Prather, 2005; Solis and Perkel, 2005). For each behavioral unit known as a “motif” (a learned, stereotyped sequence of sounds), single HVC neurons that project to RA fire a burst of action potentials only once, whereas single RA neurons fire more often (Hahnloser, Kozhevnikov, and Fee, 2002). This finding, combined with previous data (Yu and Margoliash, 1996), suggests that HVC codes vocal behavior at the motif or syllable level (vocalizations lasting hundreds of milliseconds), whereas RA codes at the subsyllable or note level (shorterduration vocalizations). Collectively, these data argue for a distributed memory for BOS encoded in both local circuits within HVC and RA and in HVC-to-RA synapses. Whether aspects of BOS memory are also contained within Uva and the brain-stem nuclei involved in vocalization remains to be determined.

nick: plasticity of speech (animal model)

457

The Anterior Forebrain Pathway The anterior forebrain pathway (AFP; the innermost loop in figure 27.1) is required for song learning (Bottjer, Miesner, and Arnold, 1984) and adult plasticity (Williams and Mehta, 1999; Brainard and Doupe, 2000) but not song production. Bottjer, Miesner, and Arnold (1984) found that lesions of MAN (which includes LMAN) disrupted song learning in juveniles but did not destabilize adult song over a period of several weeks. Later, two separate groups found that adult plasticity, such as degradation of song induced by deafening, was blocked by lesions of the AFP (Williams and Mehta, 1999; Brainard and Doupe, 2000). These data indicate that the AFP has a role in plasticity, in both juveniles and adults. The AFP is a cortical–basal ganglionic–thalamocortical loop that is similar to those found in mammals (Perkel, 2004). The direct pathway of the mammalian basal ganglia has four nodes: cortex, striatum, pallidum (globus pallidus), and thalamus. In contrast, the AFP has only three nodes: LMAN, area X, and DLM. However, evidence from the Perkel laboratory suggests that area X contains both striatal and pallidal cell types (Perkel, 2004). The area X–DLM (basal ganglia–thalamus) connection is inhibitory and thus resembles mammalian pallidothalamic connections (Luo and Perkel, 1999). Moreover, in addition to four classes of striatal-like neurons found in area X, there is a fifth class of aspiny fast-firing neuron that resembles mammalian pallidal neurons and appears to constitute the sole output from area X to DLM (Farries and Perkel, 2002). Consistent with these electrophysiological studies, the subclass of area X neurons that project to DLM express a peptide marker for mammalian pallidal neurons (LANT6; Reiner, Laverghetta, et al., 2004). Recent studies have revealed the role of the AFP in plasticity: the AFP induces variability in song behavior (Kao, Doupe, and Brainard, 2005; Olveczky, Andalman, and Fee, 2005; Kao and Brainard, 2006). Vocal learning is achieved by trial and error, as heard by anyone who has listened to a babbling human baby. Trial-and-error learning requires both an accurate memory of the species-typical vocalization and variability in the behavior, such that vocalizations can be shaped in one way or another. Whether mammalian cortical–basal ganglionic–thalamocortical loops function similarly to the AFP is an exciting question for future study.

Sensorimotor development As described previously, song learning proceeds in two phases that may involve distinct, but interconnected, loops. Auditory learning occurs during the sensory phase of vocal development and may primarily involve brain areas outside the song system. These areas are found in nonsinging animals and thus do not appear to uniquely define vocal learning

458

neural plasticity in development

systems (but they may be tailored at the cellular level for song learning; see Pinaud et al., 2006). Motor learning occurs during the sensorimotor phase and requires the specialized and dedicated song system. Since the song system allows direct brain-behavior comparisons and thus has enabled many detailed analyses of the mechanisms of vocal learning, the remainder of this chapter will focus on the sensorimotor aspects of vocal learning that are based in the song system. Since the link between perceptual learning and behavior is sensorimotor learning, increased understanding of the sensorimotor mechanisms of learning will lay the foundation for future investigations of sensory learning. Five major approaches have been used to study the mechanisms of sensorimotor plasticity: longitudinal analyses of behavior during development, measurement of neuromodulators, characterization of the development of synaptic connections, investigation of the expression of genes related to human speech disorders, and measurement of auditory responses in developing songbirds. Each of these will be considered in turn. Development of Behavior and the Role of Sleep Tchernichovski and colleagues have been successful in monitoring vocal behavior over extended periods and quantifying developmental changes in song (Tchernichovski et al., 2000, 2001; Deregnaucourt et al., 2005). They have found that similar initial prototypes are shaped into dissimilar final sounds and that abrupt changes in learned vocalizations can occur (Tchernichovski et al., 2001). Collectively, these data suggest underlying mechanistic limitations on song learning. One key insight was that sound segments of vocalizations are shaped in the context of the overall vocal pattern (Tchernichovski et al., 2001). That is, notes are not learned separately and then pieced together, but instead notes are carved out of preexisting vocal patterns. The role of sleep in song learning has also been illuminated through longitudinal study of song behavior (Deregnaucourt et al., 2005). Features of juvenile song syllables, such as the variance in entropy, progress toward adult values during the day. At night, the features move back toward juvenile values. The authors were careful to show that the effect was not due to circadian rhythms or lack of singing, but required sleep. These data suggest that aspects of song behavior in juveniles are destabilized during sleep, potentially providing a greater capacity for plasticity the next day. Indeed, juveniles that exhibited the most destabilization during sleep also ultimately produced songs that most resembled the learned tutor song (Deregnaucourt et al., 2005). Neuromodulators in the Song System Multiple neuromodulators are regulated within song nuclei during development. For example, during the sensorimotor phase, norepinephrine turnover dramatically decreases in NIf

(Harding, Barclay, and Waterman, 1998), dopamine levels peak in area X (Harding, Barclay, and Waterman, 1998), and acetylcholine levels peak in HVC (Sakaguchi and Saito, 1989). While much has been learned about the expression patterns of neuromodulators, the enzymes that make and degrade them (for example, Ryan and Arnold, 1981), and their receptors (for example, Casto and Ball, 1994), relatively little is known about how these molecules affect neural activity. Experiments in anesthetized adults suggest that acetylcholine may alter auditory responses in HVC (Shea and Margoliash, 2003). This gating in HVC may be orchestrated through Uva (Akutagawa and Konishi, 2005). Recent data from adults sedated with diazepam indicate that norepinephrine injected into NIf increases the response of HVC to auditory stimuli (Cardin and Schmidt, 2004). How neuromodulators affect song-system activity in awake behaving birds, particularly juveniles, provides an exciting path for future research on vocal learning. Development of Synapses The development of synaptic connections has been investigated on two levels in the song system: assessment of connections between song nuclei and characterization of the developmental expression patterns of glutamatergic N-methyl-D-aspartate receptors (NMDARs), which have a role in some forms of activity-dependent plasticity. Mooney and colleagues (Mooney, 1992; Mooney and Rao, 1994) have found that within RA, synapses from LMAN mature much earlier than those from HVC and that the HVC-to-X connection matures prior to the HVC-to-RA connection. Within RA, LMAN synapses appear to rely primarily on transmission through NMDARs, whereas HVC synapses are primarily non-NMDA-dependent. Stark and Perkel (1999) examined the developmental time frame of these two connections. They found that, with development, the duration of NMDA-induced postsynaptic currents decreased in both HVC-to-RA and LMAN-to-RA synapses and that the NMDA contribution in HVC-to-RA synaptic transmission decreased. However, these synaptic changes were not well correlated with behavioral changes. Nordeen and Nordeen (for review see Nordeen and Nordeen, 2004) have systematically tested the role of NMDARs in song learning. They found that systemic blockade of NMDARs prevented sensory learning. They also found that the expression patterns of NMDAR mRNAs are developmentally modulated in several song nuclei, including LMAN, HVC, and RA (Nordeen and Nordeen, 2004; Scott et al., 2004). However, the time frames of NMDAR expression in LMAN and learning can be dissociated with specific behavioral or hormonal manipulations, suggesting that specific NMDAR expression patterns do not define sensitive periods in vocal learning (Livingston, White, and Mooney, 2000). In summary, the data from NMDAR studies indicate that mechanisms of activity-dependent plasticity are at play

in the song system, but the exact role of NMDARs awaits further study. Genetic Aspects of Vocal Development The FOXP2 gene is the monogenetic locus underlying a human language disorder in the KE family (Lai et al., 2001). Individuals with the FOXP2 mutation have difficulties producing sequences of orofacial movements. In addition, they are significantly impaired on both receptive and grammatical language compared to family members without the mutation. However, family members perform similarly on nonverbal intelligence tests regardless of the mutation, suggesting that the FOXP2 gene is specifically involved in vocal learning and not overall cognitive ability. The White and Scharff laboratories independently discovered the FoxP2 homologue in the brains of songbirds (Scharff and White, 2004). In a songbird that exhibits seasonal changes in behavior, the canary, FoxP2 expression was seasonally up-regulated during periods of plasticity. A related gene, FoxP1, was strongly expressed in HVC, area X, and RA, but not in LMAN. Along with the song system, other regions of the songbird brain expressed these genes. Collectively, these data indicate that a highly conserved gene that has a role in human language and speech may also have a role in vocal learning in the songbird. Sensory Development in a Sensorimotor System Extracellular and intracellular recordings from anesthetized adult songbirds have consistently shown that song nuclei respond selectively to the bird’s own song (Margoliash, 1983; Mooney, 2000). No widely accepted argument explains why BOS is the best song system stimulus during sleep and anesthesia, although one study speculated that song preference for BOS is related to rehearsal of song during sleep (Dave and Margoliash, 2000). This finding in songbirds prompted a complementary study in humans that revealed that sleep improves performance on learning tasks involving a naturalistic spoken language (Fenn, Nusbaum, and Margoliash, 2003). In addition to adults, BOS is also the most effective activating stimulus during anesthesia in developing songbirds (Volman, 1993; Doupe, 1997; Solis and Doupe, 1997, 1999, 2000). The overall picture painted by these songbird studies suggests that song nuclei are not selective for any song in very young anesthetized birds (30 days in zebra finches), but these nuclei become selective for BOS through development. By 60 days, the majority of neurons in nuclei of the anterior forebrain pathway are activated most by BOS under anesthesia, while some respond more to tutor song or equally well to both tutor song and BOS (Solis and Doupe, 1999). If the recorded activity is a template-matching signal, it should be most activated by tutor song, not BOS. However, tutor song was not the overall most effective stimulus in any

nick: plasticity of speech (animal model)

459

developmental study in anesthetized birds. What is the impetus for change? Where is the matching or error signal that indicates whether the auditory input matches or mismatches the template? The anesthetized song system behaves differently than the waking song system. Indeed, in adult finches, the song system exhibits strikingly different activity patterns depending on anesthesia and sleep/wake state (Dave, Yu, and Margoliash, 1998; Schmidt and Konishi, 1998; Nick and Konishi, 2001). Recent studies indicate that the song selectivity measured in juvenile finches is also different during waking compared to anesthesia (Nick and Konishi, 2005a). Specifically, HVC responds most to playback to the tutor song, not BOS, in awake juveniles. However, 12 hours later, during sleep, HVC responds best to BOS. These data indicate that understanding how tutor song shapes singing will require recording from awake juveniles. Techniques for chronic recording are already in use in several laboratories and are dramatically increasing our knowledge of song system development.

Implications for human speech Accumulating data, from behavior to genetics, indicate that direct comparisons to birdsong will inform studies of human speech, and vice versa (Fenn, Nusbaum, and Margoliash, 2003; Kuhl, 2003; Scharff and White, 2004). There are striking similarities between the learned vocal behaviors of human speech and birdsong (Doupe and Kuhl, 1999). Notably, both types of vocal learning proceed through two sensitive periods, the sensory and sensorimotor phases. Both humans and songbirds appear to have innate preferences for species-typical sounds. In addition, both types of vocal learning rely on specialized forebrain areas, and both require auditory feedback during the sensorimotor phase to shape the vocal behavior. Thus recent transitions in our understanding of birdsong suggest hypotheses for human research. Social context affects the production of speech (Goldstein, King, and West, 2003) and song (West and King, 1988) in humans and songbirds, respectively. Study of the song system has revealed that social context affects the rhythm of song behavior, neural activity during song, and the expression of immediate early genes (Jarvis et al., 1998; Hessler and Doupe, 1999; Kao, Doupe, and Brainard, 2005). Collectively, these data suggest that social context impacts vocal development through direct effects on the brain during behavior. Speech and language development in humans may be facilitated by direct, reinforcing actions of caregivers. Further, caregiver interactions may have immediate and profound effects on the neural circuits underlying speech and language. The birdsong field has begun to reassess the physical location of sensory memories (Bolhuis and Gahr, 2006). Previous hypotheses placed the memory of the tutor song within the anterior forebrain pathway (for example, see Troyer and

460

neural plasticity in development

Doupe, 2000). However, studies of activity-dependent gene expression (Mello, Vicario, and Clayton, 1992) and recording from brain areas afferent of the AFP (Nick and Konishi, 2005a) suggest that auditory processing areas that are not part of the song system may store sensory memory. These areas are not obviously specialized for song learning and/or production (but see Pinaud et al., 2006). Assuming that similar mechanisms underlie human speech, a complimentary hypothesis would predict that the brain pathway that underlies sensory or perceptual speech learning is distinct from the pathway that underlies sensorimotor learning. However, the area where sensory and sensorimotor pathways converge should have sensory and sensorimotor capabilities. A human brain area that involves speech perception (Wernicke’s area) is anatomically distinct from Broca’s area, which is involved in both language production and comprehension. The birdsong data suggest two distinct but converging pathways upon which mutations may act. Depending on where they act, specific disorders or injuries may affect sensory acquisition, sensorimotor learning, or both. Another major recent advance in birdsong has revealed a role of the AFP in song plasticity. The AFP induces variability in song behavior, thus enabling trial-and-error learning (Kao, Doupe, and Brainard, 2005; Olveczky, Andalman, and Fee, 2005; Kao and Brainard, 2006). Since clearly human infants also learn speech by trial and error, a complimentary hypothesis would predict that a cortical–basal ganglionic–thalamocortical loop is critical for sensorimotor, but not sensory, speech learning in humans. A general role of such loops may be to induce variation in learned behaviors, such that deficits that involve the basal ganglia impair the ability to learn behaviors and to compensate for changes in the environment or in the peripheral nervous system and effector organs, such as muscle. Longitudinal recordings of brain activity and behavior across the sensorimotor phase of song learning are now possible in zebra finches. By systematically comparing the neural activity associated with learned song, a recent study found that activity in the song nucleus HVC changes with development (Crandall, Aoki, and Nick, 2007). During the sensorimotor sensitive period, a subset of HVC neurons initiates activity seconds before song begins and ends activity seconds after the end of the vocalization. Prolonged bursting also characterizes a purely sensory sensitive period, the critical period for ocular dominance plasticity in the visual system (Fagiolini and Hensch, 2000), which suggests that prolonged bursting may be a defining characteristic of sensitive periods in neural development. Collectively, these data suggest that the sensorimotor sensitive period in songbirds and, perhaps, humans is related to the prolonged activity of neurons in vocal control areas. If so, drugs or treatments that affect neural activity during this sensitive period may severely perturb language acquisition.

The road ahead The birdsong field has achieved a “critical mass” of researchers and expertise, enabling accelerated progress on multiple technological and experimental fronts. For example, a zebra finch genome project is under way (http://songbirdgenome. org/), and chronic recordings lasting weeks to months now enable assessment of the circuit changes occurring in song nuclei online during vocal learning (Nick and Konishi, 2005a). Several research goals that are feasible and hold great promise are described in the following paragraphs. Recent data indicate that the song area HVC, which is the site of convergence for sensory and sensorimotor signals, serves as a pattern generator (Mooney and Prather, 2005; Solis and Perkel, 2005), providing the overall rhythm of vocalization. If so, the mechanisms underlying the shaping of this pattern during development and the limitations thereof are key determinants of vocal development. If the cellular and circuit mechanisms that define the sensorimotor critical period are elucidated, then we will be well on our way to potentially extending or reopening it. Data already indicate that activity within HVC is elevated during the sensorimotor phase (Nick and Konishi, 2005a). Since elevated activity is a hallmark of critical periods in other systems (Hensch, 2004), future experiments will address the development of circuit activity during vocal learning. The area HVC transmits a sparse code to RA (Hahnloser, Kozhevnikov, and Fee, 2002) that appears to define the overall timing of vocalizations. The generative mechanisms underlying such a sparse code are not known, but they appear to at least partially arise within HVC (Coleman and Mooney, 2004). The cellular and circuit mechanisms that shape the pattern of extremely low-frequency action potentials may enable rapid vocal learning (Fiete et al., 2004). Thus investigation of these potentially novel cellular mechanisms is a high priority for the birdsong field. The location(s) where the sensory memory of the tutor song is stored and compared to auditory feedback is not known and remains controversial. The birdsong field will make a concerted effort over the next few years to identify where the signal that shapes learned vocalizations during singing originates. The field has a variety of techniques in its arsenal to attack this problem, including sophisticated song analysis software (Tchernichovski et al., 2000), several methods of chronic recording in awake behaving birds (Schmidt and Konishi, 1998; Fee and Leonardo, 2001), measurement of activity-dependent gene expression (Mello, Vicario, and Clayton, 1992), and laser capture microdissection and expression profiling of single projection neurons (Lombardino et al., 2006). In general, the field can and should move forward using an integrative approach. With continued testing of cellular hypotheses generated from behavioral experiments and vice

versa, the birdsong field is poised to crack the neural code that underlies vocal learning. acknowledgments

I thank Monica Luciana, Stephanie White, Naoya Aoki, Tim Balmer, Vanessa Carels, Shane Crandall, Amanda Kinnischtzke, and Jeff Stott for reviewing preliminary drafts of the manuscript; Melissa Coleman and Mark Konishi for helpful discussions of birdsong; and John Allman for a stimulating discussion of avian brain evolution. REFERENCES

Akutagawa, E., and M. Konishi, 2005. Connections of thalamic modulatory centers to the vocal control system of the zebra finch. Proc. Natl. Acad. Sci. USA 102:14086–14091. Ashmore, R. C., J. M. Wild, and M. F. Schmidt, 2005. Brainstem and forebrain contributions to the generation of learned motor behaviors for song. J. Neurosci. 25:8543–8554. Ball, G. F., C. J. Auger, D. J. Bernard, T. D. Charlier, J. J. Sartor, L. V. Riters, and J. Balthazart, 2004. Seasonal plasticity in the song control system: Multiple brain sites of steroid hormone action and the importance of variation in song behavior. Ann. NY Acad. Sci. 1016:586–610. Bolhuis, J. J., and M. Gahr, 2006. Neural mechanisms of birdsong memory. Nature Rev. Neurosci. 7:347–357. Bolhuis, J. J., G. G. O. Zijlstra, A. M. den Boer–Visser, and E. A. Van der Zee, 2000. Localized neuronal activation in the zebra finch brain is related to the strength of song learning. Proc. Natl. Acad. Sci. USA 97:2282–2285. Bottjer, S., E. Miesner, and A. Arnold, 1984. Forebrain lesions disrupt development but not maintenance of song in passerine birds. Science 224:901–903. Brainard, M. S., and A. J. Doupe, 2000. Interruption of a basal ganglia–forebrain circuit prevents plasticity of learned vocalizations. Nature 404:762–766. Brenowitz, E. A., 2004. Plasticity of the adult avian song control system. Ann. NY Acad. Sci. 1016:560–585. Cardin, J. A., J. N. Raksin, and M. F. Schmidt, 2005. Sensorimotor nucleus NIf is necessary for auditory processing but not vocal motor output in the avian song system. J. Neurophysiol. 93: 2157–2166. Cardin, J. A., and M. F. Schmidt, 2004. Noradrenergic inputs mediate state dependence of auditory responses in the avian song system. J. Neurosci. 24:7745–7753. Casto, J. M., and G. F. Ball, 1994. Characterization and localization of D1 dopamine receptors in the sexually dimorphic vocal control nucleus, area X, and the basal ganglia of European starlings. J. Neurobiol. 7:767–780. Coleman, M. J., and R. Mooney, 2004. Synaptic transformations underlying highly selective auditory representations of learned birdsong. J. Neurosci. 24:7251–7265. Coleman, M. J., and E. T. Vu, 2005. Recovery of impaired songs following unilateral but not bilateral lesions of nucleus uvaeformis of adult zebra finches. J. Neurobiol. 63:70–89. Crandall, S. R., N. Aoki, and T. A. Nick, 2007. Developmental modulation of the temporal relationship between brain and behavior. J. Neurophysiol. 97:806–816. Dave, A. S., and D. Margoliash, 2000. Song replay during sleep and computational rules for sensorimotor vocal learning. Science 290:812–816.

nick: plasticity of speech (animal model)

461

Dave, A. S., A. C. Yu, and D. Margoliash, 1998. Behavioral state modulation of auditory activity in a vocal motor system. Science 282:2250–2254. Deregnaucourt, S., P. P. Mitra, O. Feher, C. L. Pytte, and O. Tchernichovski, 2005. How sleep affects the developmental learning of bird song. Nature 433:710–716. Doupe, A. J., 1997. Song- and order-selective neurons in the songbird anterior forebrain and their emergence during vocal development. J. Neurosci. 17:1147–1167. Doupe, A. J., and M. Konishi, 1991. Song-selective auditory circuits in the vocal control system of the zebra finch. Proc. Natl. Acad. Sci. USA 88:11339–11343. Doupe, A. J., and P. K. Kuhl, 1999. Birdsong and human speech: Common themes and mechanisms. Annu. Rev. Neurosci. 22: 567–631. Fagiolini, M., and T. K. Hensch, 2000. Inhibitory threshold for critical-period activation in primary visual cortex. Nature 404:183–186. Farries, M. A., and D. J. Perkel, 2002. A telencephalic nucleus essential for song learning contains neurons with physiological characteristics of both striatum and globus pallidus. J. Neurosci. 22:3776–3787. Fee, M. S., and A. Leonardo, 2001. Miniature motorized microdrive and commutator system for chronic neural recording in small animals. J. Neurosci. Methods 112:83–94. Fenn, K. M., H. C. Nusbaum, and D. Margoliash, 2003. Consolidation during sleep of perceptual learning of spoken language. Nature 425:614–616. Fiete, I. R., R. H. R. Hahnloser, M. S. Fee, and H. S. Seung, 2004. Temporal sparseness of the premotor drive is important for rapid learning in a neural network model of birdsong. J. Neurophysiol. 92:2274–2282. Gahr, M., 2000. Neural song control system of hummingbirds: Comparison to swifts, vocal learning (songbirds) and nonlearning (suboscines) passerines, and vocal learning (budgerigars) and nonlearning (dove, owl, gull, quail, chicken) nonpasserines. J. Comp. Neurol. 426:182–196. Gentner, T. Q., and D. Margoliash, 2003. Neuronal populations and feature detectors representing learned auditory objects. Nature 424:669–674. Goldstein, M. H., A. P. King, and M. J. West, 2003. Social interaction shapes babbling: Testing parallels between birdsong and speech. Proc. Natl. Acad. Sci. USA 100:8030–8035. Hahnloser, R. H. R., A. A. Kozhevnikov, and M. S. Fee, 2002. An ultra-sparse code underlies the generation of neural sequences in a songbird. Nature 419:65–70. Harding, C. F., 2004. Hormonal modulation of singing: Hormonal modulation of the songbird brain and singing behavior. Ann. NY Acad. Sci. 1016:524–539. Harding, C. F., S. R. Barclay, and S. A. Waterman, 1998. Changes in catecholamine levels and turnover rates in hypothalamic, vocal control, and auditory nuclei in male zebra finches during development. J. Neurobiol. 34:329–346. Hensch, T. K., 2004. Critical period regulation. Annu. Rev. Neurosci. 27:549–579. Hessler, N. A., and A. J. Doupe, 1999. Social context modulates singing-related neural activity in the songbird forebrain. Nature Neurosci. 2:209–211. Janata, P., and D. Margoliash, 1999. Gradual emergence of song selectivity in sensorimotor structures of the male zebra finch song system. J. Neurosci. 19:5108–5118. Jarvis, E. D., and F. Nottebohm, 1997. Motor-driven gene expression. Proc. Natl. Acad. Sci. USA 94:4097–4102.

462

neural plasticity in development

Jarvis, E. D., C. Scharff, M. Grossman, J. A. Ramos, and F. Nottebohm, 1998. For whom the bird sings: Context-dependent gene expression. Neuron 21:775–788. Kao, M. H., and M. S. Brainard, 2006. Lesions of an avian basal ganglia circuit prevent context-dependent changes to song variability. J. Neurophysiol. 96:1441–1455. Kao, M. H., A. J. Doupe, and M. S. Brainard, 2005. Contributions of an avian basal ganglia–forebrain circuit to real-time modulation of song. Nature 433:638–643. Konishi, M., 1965. The role of auditory feedback in the control of vocalization in the white-crowned sparrow. Z. Tierpsychol. 22: 770–783. Konishi, M., 2004. The role of auditory feedback in birdsong. Ann. NY Acad. Sci. 1016:463–475. Kroodsma, D. E., and M. Konishi, 1991. A suboscine bird (easter phoebe, Sayornis phoebe) develops normal song without auditory feedback. Anim. Behav. 42:477–487. Kuhl, P. K., 2000. A new view of language acquisition. Proc. Natl. Acad. Sci. USA 97:11850–11857. Kuhl, P. K., 2003. Human speech and birdsong: Communication and the social brain. Proc. Natl. Acad. Sci. USA 100:9645– 9646. Lai, C. S. L., S. E. Fisher, J. A. Hurst, F. Vargha-Khadem, and A. P. Monaco, 2001. A forkhead-domain gene is mutated in a severe speech and language disorder. Nature 413:519–523. Leonardo, A., and M. Konishi, 1998. Decrystallization of adult birdsong by perturbation of auditory feedback. Nature 399: 466–470. Liu, W.-C., and F. Nottebohm, 2005. Variable rate of singing and variable song duration are associated with high immediate early gene expression in two anterior forebrain song nuclei. Proc. Natl. Acad. Sci. USA 102:10724–10729. Livingston, F. S., S. A. White, and R. Mooney, 2000. Slow NMDA-EPSCs at synapses critical for song development are not required for song learning in zebra finches. Nature Neurosci. 3: 482–488. Lombardino, A. J., M. Hertel, X.-C. Li, B. Haripal, L. MartinHarris, E. Pariser, and F. Nottebohm, 2006. Expression profiling of intermingled long-range projection neurons harvested by laser capture microdissection. J. Neurosci. Methods 157: 195–207. Luo, M., and D. J. Perkel, 1999. A GABAergic, strongly inhibitory projection to a thalamic nucleus in the zebra finch song system. J. Neurosci. 19:6700–6711. Margoliash, D., 1983. Acoustic parameters underlying the responses of song-specific neurons in the white-crowned sparrow. J. Neurosci. 3:1039–1057. Marler, P., and S. Peters, 1977. Selective vocal learning in a sparrow. Science 198:519–521. McCasland, J. S., and M. Konishi, 1981. Interaction between auditory and motor activities in an avian song control nucleus. Proc. Natl. Acad. Sci. USA 78:7815–7819. Mello, C. V., and D. F. Clayton, 1994. Song-induced ZENK gene expression in auditory pathways of songbird brain and its relation to the song control system. J. Neurosci. 14:6652–6666. Mello, C. V., T. A. F. Velho, and R. Pinaud, 2004. Song-induced gene expression: A window on song auditory processing and perception. Ann. NY Acad. Sci. 1016:263–281. Mello, C. V., D. S. Vicario, and D. F. Clayton, 1992. Song presentation induces gene expression in the songbird forebrain. Proc. Natl. Acad. Sci. USA 89:6818–6822. Mooney, R., 1992. Synaptic basis for developmental plasticity in a birdsong nucleus. J. Neurosci. 12:2464–2477.

Mooney, R., 2000. Different subthreshold mechanisms underlie song selectivity in identified HVc neurons of the zebra finch. J. Neurosci. 20:5420–5436. Mooney, R., and J. F. Prather, 2005. The HVC microcircuit: The synaptic basis for interactions between song motor and vocal plasticity pathways. J. Neurosci. 25:1952– 1964. Mooney, R., and M. Rao, 1994. Waiting periods versus early innervation: The development of axonal connections in the zebra finch song system. J. Neurosci. 14:6532–6543. Nick, T. A., and M. Konishi, 2001. Dynamic control of auditory activity during sleep: Correlation between song response and EEG. Proc. Natl. Acad. Sci. USA 98:14012–14016. Nick, T. A., and M. Konishi, 2005a. Neural song preference during vocal learning in the zebra finch depends on age and state. J. Neurobiol. 62:231–242. Nick, T. A., and M. Konishi, 2005b. Neural auditory selectivity develops in parallel with song. J. Neurobiol. 62:469–481. Nordeen, K. W., and E. J. Nordeen, 1992. Auditory-feedback is necessary for the maintenance of stereotyped song in adult zebra finches. Behav. Neural Biol. 57:58–66. Nordeen, K. W., and E. J. Nordeen, 2004. Synaptic and molecular mechanisms regulating plasticity during early learning. Ann. NY Acad. Sci. 1016:416–437. Nottebohm, F., 2004. The road we travelled: Discovery, choreography, and significance of brain replaceable neurons. Ann. NY Acad. Sci. 1016:628–658. Nottebohm, F., D. B. Kelley, and J. A. Paton, 1982. Connections of vocal control nuclei in the canary telencephalon. J. Comp. Neurol. 207:344–357. Nottebohm, F., T. M. Stokes, and C. M. Leonard, 1976. Central control of song in the canary, Serinus canarius. J. Comp. Neurol. 165:457–486. Olveczky, B. P., A. S. Andalman, and M. S. Fee, 2005. Vocal experimentation in the juvenile songbird requires a basal ganglia circuit. PLoS Biol. 3:0902–0909. Perkel, D. J., 2004. Origin of the anterior forebrain pathway. Ann. NY Acad. Sci. 1016:736–748. Phan, M. L., C. L. Pytte, and D. S. Vicario, 2006. Early auditory experience generates long-lasting memories that may subserve vocal learning in songbirds. Proc. Natl. Acad. Sci. USA 103: 1088–1093. Pinaud, R., A. F. Fortes, P. Lovell, and C. V. Mello, 2006. Calbindin-positive neurons reveal a sexual dimorphism within the songbird analogue of the mammalian auditory cortex. J. Neurobiol. 66:182–195. Rauske, P. L., S. D. Shea, and D. Margoliash, 2003. State and neuronal class-dependent reconfiguration in the avian song system. J. Neurophysiol. 89:1688–1701. Reiner, A., A. V. Laverghetta, C. A. Meade, S. L. Cuthbertson, and S. W. Bottjer, 2004. An immunohistochemical and pathway tracing study of the striatopallidal organization of area X in the male zebra finch. J. Comp. Neurol. 469:239–261. Reiner, A., D. J. Perkel, C. V. Mello, and E. D. Jarvis, 2004. Songbirds and the revised avian brain nomenclature. Ann. NY Acad. Sci. 1016:77–108. Ryan, S. M., and A. P. Arnold, 1981. Evidence for cholinergic participation in the control of bird song: Acetylcholinesterase distribution and muscarinic receptor autoradiography in the zebra finch brain. J. Comp. Neurol. 202:211–219. Sakaguchi, H., and N. Saito, 1989. The acetylcholine and catecholamine contents in song control nuclei of zebra finch during song ontogeny. Dev. Brain Res. 47:313–317.

Scharff, C., and S. White, 2004. Genetic components of vocal learning. Ann. NY Acad. Sci. 1016:325–347. Schmidt, M. F., and M. Konishi, 1998. Gating of auditory responses in the vocal control system of awake songbirds. Nature Neurosci. 1:513–518. Scott, L. L., T. D. Singh, E. J. Nordeen, and K. W. Nordeen, 2004. Developmental patterns of NMDAR expression within the song system do not recur during adult vocal plasticity in zebra finches. J. Neurobiol. 58:442–454. Shea, S. D., and D. Margoliash, 2003. Basal forebrain cholinergic modulation of auditory activity in the zebra finch song system. Neuron 40:1213–1226. Solis, M. M., and A. J. Doupe, 1997. Anterior forebrain neurons develop selectivity by an intermediate stage of birdsong learning. J. Neurosci. 17:6447–6462. Solis, M. M., and A. J. Doupe, 1999. Contributions of tutor and bird’s own song experience to neural selectivity in the songbird anterior forebrain. J. Neurosci. 19:4559–4584. Solis, M. M., and A. J. Doupe, 2000. Compromised neural selectivity for song in birds with impaired sensorimotor learning. Neuron 25:109–121. Solis, M. M., and D. J. Perkel, 2005. Rhythmic activity in a forebrain vocal control nucleus in vitro. J. Neurosci. 25: 2811–2822. Stark, L. L., and D. J. Perkel, 1999. Two-stage, input-specific synaptic maturation in a nucleus essential for vocal production in the zebra finch. J. Neurosci. 19:9107–9116. Stripling, R., A. A. Kruse, and D. F. Clayton, 2001. Development of song responses in the zebra finch caudomedial neostriatum: Role of genomic and electrophysiological activities. J. Neurobiol. 48:163–180. Tchernichovski, O., P. P. Mitra, T. Lints, and F. Nottebohm, 2001. Dynamics of the vocal imitation process: How a zebra finch learns its song. Science 291:2564–2569. Tchernichovski, O., F. Nottebohm, C. E. Ho, B. Pesaran, and P. P. Mitra, 2000. A procedure for an automated measurement of song similarity. Anim. Behav. 59:1167–1176. Terpstra, N. J., J. J. Bolhuis, and A. M. den Boer–Visser, 2004. An analysis of the neural representation of bird song memory. J. Neurosci. 24:4971–4977. Terpstra, N. J., J. J. Bolhuis, K. Riebel, J. M. M. van der Burg, and A. M. den Boer–Visser, 2006. Localized brain activation specific to auditory memory in a female songbird. J. Comp. Neurol. 494:784–791. Theunissen, F. E., N. Amin, S. S. Shaevitz, S. M. N. Woolley, T. Fremouw, and M. E. Hauber, 2004. Song selectivity in the song system and in the auditory forebrain. Ann. NY Acad. Sci. 1016:222–245. Thorpe, W. H., 1958. The learning of song patterns by birds with special reference to the song of the chaffinch, Fringilla coelebs. Ibis 100:535–642. Troyer, T. W., and A. J. Doupe, 2000. An associational model of birdsong sensorimotor learning. I. Efference copy and the learning of song syllables. J. Neurophysiol. 84:1204–1223. Vates, G. E., B. M. Broome, C. V. Mello, and F. Nottebohm, 1996. Auditory pathways of caudal telencephalon and their relation to the song system of adult male zebra finches (Taenopygia guttata). J. Comp. Neurol. 366:613–642. Vicario, D. S., and K. H. Yohay, 1993. Song-selective auditory input to a forebrain vocal control nucleus in the zebra finch. J. Neurobiol. 24:488–505. Volman, S. F., 1993. Development of neural selectivity for birdsong during vocal learning. J. Neurosci. 13:4737–4747.

nick: plasticity of speech (animal model)

463

Vu, E. T., M. E. Mazurek, and Y.-C. Kuo, 1994. Identification of a forebrain motor programming network for the learned song of zebra finches. J. Neurosci. 14:6924–6934. Wade, J., and A. P. Arnold, 2004. Sexual differentiation of the zebra finch song system. Ann. NY Acad. Sci. 1016:540–559. West, M. J., and A. P. King, 1988. Female visual displays affect the development of male song in the cowbird. Nature 334: 244–246. Wild, J. M., 1993. Visual and somatosensory projections to the avian song system via nucleus uvaeformis (Uva) and a comparison with the projections of a similar thalamic nucleus in a non-songbird (Columba livia). J. Comp. Neurol. 349:512– 535. Wild, J. M., 2004. Functional neuroanatomy of the sensorimotor control of singing. Ann. NY Acad. Sci. 1016:438–462.

464

neural plasticity in development

Williams, H., and N. Mehta, 1999. Changes in adult zebra finch song require a forebrain nucleus that is not necessary for song production. J. Neurobiol. 39:14–28. Williams, H., and F. Nottebohm, 1985. Auditory responses in avian vocal motor neurons: A motor theory for song perception in birds. Science 229:279–282. Woolley, S. M. N., P. R. Gill, and F. E. Theunissen, 2006. Stimulus-dependent auditory tuning results in synchronous population coding of vocalizations in the songbird midbrain. J. Neurosci. 26:2499–2512. Yu, A. C., and D. Margoliash, 1996. Temporal hierarchical control of singing in birds. Science 273:1871–1875. Zevin, J. D., M. S. Seidenberg, and S. W. Bottjer, 2004. Limits on reacquisition of song in adult zebra finches exposed to white noise. J. Neurosci. 24:5849–5862.

IV COGNITION

28

The Development and Integration of the Dorsal and Ventral Visual Pathways in Object Processing MARK H. JOHNSON, DENIS MARESCHAL, AND GERGELY CSIBRA

Developmental cognitive neuroscience is concerned with understanding the functional development of the human brain by integrating evidence from neural development with that from cognitive and behavioral studies. Whereas it is relatively straightforward to correlate changes at the neural level with changes in behavior, it is more challenging to actually integrate information from these sources into a single cohesive account of developmental change. We believe that recent advances in two areas, one theoretical and one methodological, will allow more rapid progress in integrating neural and behavioral evidence in development. First, the advent of neural network and connectionist modeling will provide the appropriate level of theoretical framework in which evidence from neural development and behavior can be modeled simultaneously (see chapter 22 in this volume). Second, recent advances in neuroimaging now allow the noninvasive measurement of brain activity in healthy infants with reasonable spatial and temporal accuracy. In this chapter we discuss evidence gathered through both of these new techniques in an attempt to advance our knowledge of the functional development of the two main cortical visual pathways in the human brain and their role in object processing. There is now a substantive body of evidence that visual information processing in the primate cortex is divided into two relatively distinct streams (see chapter 32 by Stiles, Paul, and Ark, this volume). However, only in the past decade has the dual-route visual-processing paradigm been applied to the study of infant perceptual and cognitive development (e.g., Atkinson, 1998; Berthenthal, 1996; Mareschal, Plunkett, and Harris, 1999), and many fundamental questions remain. One question is whether it is the dorsal or the ventral route that functionally develops first during infancy, while another concerns whether there is increasing separation or increasing integration between the two pathways with development. Physical objects are an ideal topic for study in relation to the development of the dorsal and ventral pathways, since they often require both visual recognition and manual action. Thus both pathways are likely to be

involved in object processing even though certain objects or tasks may depend more heavily on one pathway or the other. In this chapter we review evidence from neuroimaging, computational modeling, and behavior, which indicates that these questions are unlikely to have simple answers. While there is some evidence that dorsal stream eye movement control is relatively delayed, other aspects of dorsal route function (such as reaching) may be more precocial. Computational and behavioral evidence indicates that even when the dorsal route is functioning, there may be an initial lack of integration between the two pathways resulting in specific patterns of behavioral deficits. We conclude with speculations as to the causes and consequences of different developmental timetables in the dorsal and ventral cortical streams.

The dorsal and ventral routes of visual processing In this section we briefly review the neurocomputational properties of the dorsal and ventral routes (a more detailed review can be found in Milner and Goodale, 1995). The data used to identify anatomically distinct pathways are based largely on primates other than humans. However, analogous structures do exist in humans, and lesion studies suggest that these structures also function in a similar way in humans. The connectivity of these cortical routes is very complex (see van Essen, Anderson, and Fellman, 1992). Although two pathways can be identified, there are interconnections between the two pathways at many different points along the pathway, and information can arrive at several points from subcortical routes. One pathway (the ventral route) extends from the primary visual cortex through to portions of the temporal cortex. The other pathway (the dorsal route) extends from the primary visual cortex to the parietal cortex. The exact point at which the routes separate is still debated. The ventral route is sometimes called the “what” or “perception” pathway, and the dorsal route is often called the “where” or “action” pathway (Ungerleider

467

and Mishkin, 1982; Milner and Goodale, 1995). These names reflect the differing processing objectives attributed to each route by different authors. It is likely that adults often engage both pathways when acting on objects. Ungerleider and Mishkin (1982) initially proposed that the ventral and dorsal streams were completely separate from the retina to the temporal or parietal cortex, and that the ventral stream played a crucial role in object identification whereas the dorsal stream was critical in localizing the object in the visual field. Livingstone and Hubel (1988) went further and suggested that the dorsal stream is concerned with the global spatial organization of objects. That is, it segments and defines individual objects in a scene as well as keeping track of their relative locations in space. In contrast, the ventral stream analyzed the scene in much more detail but at a slower rate. It is sensitive to color, shape, and other surface properties of objects. Processing in the Dorsal Pathway There is evidence of multiple (and parallel) spatial processing systems within the dorsal route. Some cells in the parietal cortex anticipate the retinal consequences of saccadic eye movements and update the cortical (body-centered) representation of visual space to provide continuously accurate coding of the location of objects in space. Other cells have gaze-dependent responses. That is, they mark where the animal is looking with respect to eye-centered or body-centered coordinate systems. These representations are only useful over very short periods of time, since every time the animal moves the coordinates have to be recomputed. In many real-world settings, target objects are moving. It is necessary to anticipate an object’s movement in order to act on it effectively. Some cells in the parietal cortex appear to be involved in the tracking of moving objects. Moreover, many of these cells continue to respond during occluded pursuit after the stimulus has disappeared (Newsome, Wurtz, and Komatsu, 1988), a point that we will return to later. In addition, there is selectivity in other parts of the dorsal stream for relative motion and size changes due to looming. Many cells are also driven by large-scale optical flow fields, suggesting that self-motion is being computed. Cells in the dorsal pathway also code size, shape, and orientation, information that is necessary for the proper reaching and grasping of an object (Jeannerod, 1988). There is also some evidence suggesting that the different spatial-temporal systems are partially segregated into different regions and routes within the dorsal pathway. Thus the dorsal stream could be viewed as a pathway with many parallel computations of different spatial-temporal properties occurring at once. Different streams compute different spatial-temporal analyses in different coordinate systems, possibly with different effector systems as outputs. This point is illustrated by the links between the parietal and frontal

468

cognition

lobes. Milner and Goodale (1995) argue that the cells of the parietal cortex are neither sensory nor motor, but rather sensory-motor cells. They are involved in transforming retinal information (sensory) into motor coordinates (motor) and provide an adaptive medium for transducing perceptual input into motor actions. Processing in the Ventral Pathway The properties of cells in the ventral stream seem to complement those in the dorsal stream. As one progresses along the ventral stream, cells respond to more and more complex clusters of features. At the higher levels, cells show remarkable selectivity in their firing (e.g., face cells). These neurons are selective to the figural and surface properties of objects, and have very large receptive fields. These cells develop spatially invariant representations of objects by responding to the presence of a consistent feature cluster independently of its position within the visual field. Some cells seem to respond maximally to a preferred object orientation (independently of position) thereby computing a “view-centered representation.” Other cells respond equally to an object in any orientation: a “transformation-invariant representation.” Transformationinvariant representations could provide the basic raw material for recognition memory and other long-term representations of the visual world. There is evidence that the responsiveness of cells in the ventral stream can be modulated by the prior occurrence of a stimulus. Moreover, there is evidence suggesting that some cells continue to fire for several seconds after the object has disappeared (Ungerleider, 1985), suggesting that some kind of memory trace remains beyond the immediate visibility of an object. In this chapter, we are less concerned with whether these pathways are best characterized as “what” and “where” (cf. Ungerleider and Mishkin, 1982) or “perception” and “action” (cf. Milner and Goodale, 1995) and more concerned with the fact that they process different types of object information, carry out different computations, and develop different object representations with distinct properties. Nevertheless, one can speculate as to the computational reasons for why two streams of processing may have evolved. If we agree with Milner and Goodale (1995) that the representations in the dorsal stream are closely linked to the functions of the motor system, then it is not surprising that spatial-temporal information is at a premium down this pathway. Motor actions involve localizing targets within a three-dimensional spatial-temporal world. In contrast, recognition or identification of objects requires that spatialtemporal variability be minimized. Early work in machine vision found that view-invariant recognition (i.e., the ability to recognize an object as the same independently of orientation and location) was a very difficult computational problem (Boden, 1988). One of the most efficient ways of achieving view-invariant recognition is to factor out spatial variability.

However, removing spatial information from the object representation is completely at odds with the requirements of the motor system. Hence the need for two distinct streams of object representations. Finally, Jeanerrod (1999) has argued that neuroimaging evidence indicates that the dissociation between the two pathways is less clear than presented by Milner and Goodale (1995). Specifically, structures within the dorsal pathway are often activated following the presentation of objects, without manual responses being required. Leaving aside the question of whether eye movements are elicited or planned in these passive viewing paradigms, it is hard to rule out the possibility that reaching actions are automatically planned, even if not executed, on the presentation of graspable objects. Our view is that the coactivation of the two pathways is entirely consistent with two streams of information processing in which the type of processing that occurs within each pathway is incompatible.

The development of the dorsal and ventral streams Identifying the presence of two visual processing streams raises the question of how these streams develop and how they interact with each other during development. Over the past decade, there has been some speculation about the developmental sequence of the two pathways. For example, Atkinson (1998) argued, on the basis that infants and children are delayed on their judgments of motion coherence compared to their thresholds for perceiving coherent forms, that the dorsal pathway develops later than the ventral. In contrast, studies of developmental neuroanatomy in 1-weekold macaque monkeys led Webster, Bachevalier, and Ungerleider (1995) to the view that the patterns of connectivity of temporal lobe structures on the ventral pathway were still relatively immature, while the connectivity of the parietal lobe (on the dorsal route) was already adultlike. Evidence from developmental neuroanatomy in human infants is not decisive in this regard. For example, in resting PET studies of glucose uptake, virtually identical overall patterns of developmental change are seen in the temporal and parietal cortices (Chugani, Phelps, and Mazziotta, 1987). However, resting blood flow measurements and structural neuroanatomy studies cannot inform us directly about function. In one PET experiment, 2-month-old infants showed activity in structures on the ventral pathway, but the task was one in which dorsal pathway activation would have been unlikely even in adults (passive viewing of faces) (de Schonen, Mancini, and Leigeois, 1998). In nonhuman primates, there is evidence of ventral pathway functioning from as young as 6 weeks of age. Rodman, Gross, and Scalaidhe (1993) established that neurons within the superior temporal sulcus were activated by complex visual stimuli, including faces, from the earliest age at which they could record: 6 weeks. Unfortu-

nately, equivalent data are not available for dorsal pathway functions, and so no comparison is possible. This paucity of data about the dorsal pathway may be due to the great technical difficulty in recording from neurons in young monkeys. Thus there is currently very little evidence that addresses the question of the relative development of the dorsal and ventral pathways during postnatal life. In the next two sections we explore evidence from neuroimaging and behavioral studies suggesting that some aspects of dorsal pathway function develop later than the majority of ventral stream functions. The ERP evidence focuses on markers for the two pathways: face processing and saccade planning. The behavioral evidence focuses on infant abilities to form perceptual compounds from features processed preferentially by one cortical stream or the other, as well as infant responses to temporarily occluded objects. Evidence from High-Density ERP Studies of the Dorsal and Ventral Pathways One of the functions most clearly associated with the ventral pathway is the processing of the surface features of objects. Face processing is commonly regarded as a special class of object processing (see chapter 31 by de Haan, this volume). Specifically, in adults PET, fMRI, ERP, and cellular recording experiments have all implicated regions of the inferior temporal cortex as being important for face processing. For example, Bentin and colleagues (1996) identified a component of the scalprecorded ERP which occurs around 170 milliseconds after the presentation of a face and which, using procedures for estimating the likely underlying brain source that gives rise to scalp surface voltage changes, is localizable to parts of the inferior temporal cortex. In many but not all adults, the specificity of this region for face processing is lateralized, with the right side being more face specific than the left, a finding confirmed with other brain-imaging methods (Kanwisher, McDermott, and Chun, 1997). De Haan, Oliver, and Johnson (1998; see chapter 31 by de Haan, this volume) conducted a study in which they showed pictures of upright and inverted faces to both adults and 6-month-old infants while recording ERPs from the participants’ heads. Adults showed a pattern consistent with several previous studies, indicating face-sensitive cortical processing from structures on the ventral visual pathway. Like adults, the infants also showed a face-sensitive response over temporal channels, albeit at a later time after stimulus presentation. This finding indicates some degree of cortical specialization for face processing from this pathway by 6 months of age. However, infants’ responses also differed in two other ways, indicating that at this age ventral processing may be less specialized for face processing than in adults. First, the response in infants shows less hemispheric specialization than in adults, possibly indicating less localization to the right ventral pathway in infants than in adults. Second, the

johnson, mareschal, and csibra: dorsal and ventral visual pathways

469

response is less selective in infants. The latter point was established by showing that, whereas monkey faces elicit similar responses to human faces in infants, in adults the response is human-face selective (see chapter 31 by de Haan, this volume). In sum, these lines of evidence indicate clear functionality in the ventral pathways by 6 months, even though further tuning is required in order to reach adult levels of competence. In the rest of this section we examine whether a similar or a different pattern of development emerges from studies of dorsal route processing. One function attributed to the dorsal visual pathway is the planning of target-directed saccades by way of the parietal eye-movement centers. The functioning of these cortical regions can be measured by ERPs time locked to the initiation of the eye movement (Balaban and Weinstein, 1985; Csibra, Tucker, and Johnson, 1997). These experiments reveal characteristic presaccadic components recorded over the parietal cortex prior to the execution of saccades. The clearest of these components is the presaccadic spike potential (SP), a sharp positive-going deflection which precedes the saccade by 8–20 msec (Csibra et al., 1997). The spike potential is observed in most saccade tasks in adults and is therefore thought to represent an important stage of cortical processing required to generate a saccade. We investigated whether there are presaccadic potentials recordable over parietal channels in 6-month-old infants (Csibra, Tucker, and Johnson, 1998). Given that the prevailing view is that by this age infants have essentially the same pathways active for saccade planning as do adults, we were surprised to find no evidence of these components in our infant subjects (see also Kurtzberg and Vaughan, 1982; Richards, 2001). This finding suggests that the target-driven saccades performed by 6-month-olds in our study were controlled solely by subcortical routes for visually guided responses mediated by the superior colliculus. Because this result was surprising, we have conducted two follow-up studies. In one of these we tested 12-month-olds with the same procedure (Csibra et al., 2000). The results from this experiment indicated that these older infants did show a spike potential like that observed in adults, though it was smaller in amplitude. The other study explored whether the dorsal pathway could be activated in very young infants through a more demanding saccade task. Specifically, we compared ERPs before reactive (target-elicited) and anticipatory (endogenous) saccades in 4-month-old infants (Csibra, Tucker, and Johnson, 2001). We were not able to record any reliable posterior activity prior to either reactive or anticipatory eye movements. Thus, even when the saccade is generated by cortical computation of the likely location of the next stimulus, as in the case of anticipatory eye movements, the dorsal pathway does not seem to be directly involved in the planning of this action.

470

cognition

As an aside, we should note that the lack of evidence for parietal (dorsal pathway) control over eye movements in our experiments with 6-month-olds does not allow us to conclude that there is no cortical influence over saccades at this age. In all our studies with this age group we have observed effects recorded over frontal leads consistent with frontaleye-field disinhibition of subcortical (collicular) circuits when a central foveated stimulus is removed (Csibra, Tucker, and Johnson, 1998, 2001). In brief, we interpret these findings in terms of the frontal eye fields maintaining fixation onto foveated stimuli by inhibiting collicular circuits. However, when saccades to peripheral stimuli are made, we believe these are largely initiated by collicular circuits, sometimes as a consequence of inhibition being released by the frontal eye fields. In the saccadic ERP data from 6-month-olds, there was strong evidence for a postsaccadic component known as the lambda wave. The lambda wave is a sharp potential appearing over visual cortical areas that is generated when a peripheral target stimulus is foveated (Kurtzberg and Vaughan, 1977). While lambda waves can be observed in adults, they were markedly enhanced in our infant subjects. One conclusion from this work is that while dorsal pathway control of eye movements is not evident at 6 months, early visual cortical responses to foveal stimuli, believed to originate from structures at the gateway to the ventral pathway (V2 and V4), are present. Thus within a single ERP trace there is evidence for the relatively delayed development of the dorsal pathway with respect to the ventral pathway. But why should the lambda wave be enhanced in infants relative to adults? One possibility may lie in the fact that in order to successfully integrate visual input with eye movements, adults “forward map” the expected visual input at the end of their saccade (see Csibra, Tucker, and Johnson, 1998). If infants are unable to integrate information about visual input and eye movements because of a lack of development of the dorsal pathway, then a peripheral target stimulus entering the fovea will be “unexpected” and thus elicit a bigger response. A related aspect of dorsal pathway function involves the use of “body-centered” frames of reference for action, including eye movements. In single-unit recording studies in monkeys, Anderson and colleagues (e.g., Anderson et al., 1993) have found cells in the posterior parietal cortex that appear to integrate visual information with proprioceptive information about eye and body position in order to generate accurate saccades to targets. In a series of experiments, Gilmore and Johnson (1997, 1998) investigated the development of these abilities in infants. In one of these experiments, we examined the patterns of saccade sequences made by 3- and 7-month-old infants performing a two-dimensional version of the double-step task (see figure 28.1). When we examined the second saccade made by the infants in response

Response Patterns

Display Sequence 4°

Fixation

Body-centered

Target

Retinocentric

Target

A

24°

B

Figure 28.1 The two-dimensional version of the double-step saccade task. (Reprinted from R. O. Gilmore and M. H. Johnson, 1997. Body-centred representations for visually-guided action emerge during early infancy. Cognition 65:B1–B9. Copyright 1997, with permission from Elsevier.)

to the two sequentially presented visual targets, we found that the younger infants tended to make this saccade to a “retinocentric” location. That is, they acted as if they were unable to integrate the visual information about where the target is with information about their own (changed) eye and head position. In contrast, the 7-month-olds made around two-thirds of their second saccades to the “correct” spatial location, suggesting that they were able to take into account the eye movement that had already taken place to the first target. The combined results from several such studies support the hypothesis that the representations for planning saccades shift from retinocentric to cranio- or egocentric in accordance with the graded experience-dependent development of cortical, especially parietal, centers where these higher-order representations are instantiated. This behavioral evidence reinforces the conclusion of the ERP studies of saccade planning by indicating that the development of dorsal pathway control over saccades occurs relatively late in infant development. Taken together, the results from the ERP studies described indicate that while the ventral pathway can be activated at 6 months (albeit with some further specialization to take place), the dorsal pathway is still not influencing at least some aspects of eye movement control at that age, suggesting that at least this aspect of dorsal pathway function is somewhat slower to develop than the ventral pathway. Event-Related Oscillations (EEG) and Object Processing A recent technique for studying object processing in the infant brain is the recording of event-related EEG oscillations (see chapter 15 by Csibra, Kushnerenko, and Grossmann, this volume). As mentioned previously, there is evidence from both the ventral and the dorsal pathways of sustained responses to objects even if they temporarily

disappear. In particular, in human adults gamma-band (∼40 Hz) EEG activity has been associated with maintaining an object/location in mind (Tallon-Baudry et al., 1998). We (Kaufman, Csibra, and Johnson, 2003, 2005) measured infants’ electrophysiological responses to occlusion events at the age where reaching behavior does not yet show evidence of understanding “object permanence.” In one experiment, we (Kaufman, Csibra, and Johnson, 2005) created digital video sequences of objects involved in typical occlusion events (involving gradual accretion) and other less usual forms of object disappearance (disintegration). A summary of our main results and analyses is illustrated in figure 28.2 and plate 47. These results demonstrate a sustained period during which gamma power was consistently higher during an event where infants represented an object being occluded (and therefore continuing to exist) as opposed to the object disappearing (and ceasing to exist). We argued that this gamma activity provides a neural basis for maintenance of object representations in infants. A question currently under investigation is whether this gamma activity indexes information relevant to action (e.g., location) or whether it is associated with retaining information about the surface features (identity) of the object in question. The answer to this question will determine whether the gamma activity originates from the dorsal or ventral pathway, or both. Behavioral Evidence of Dorsal and Ventral Dissociation in Infants The second hypothesis we wish to explore in this chapter—that there is a dissociation between processing in the two pathways during infancy—requires us to consider both behavioral evidence and computational considerations. Several authors now invoke the dual-stream hypothesis as one of the most important heuristic frameworks for understanding early human infant–object interactions (e.g., Atkinson, 2000; Kaldy and Sigala, 2004; Xu, Carey, and Welch, 1999; Leslie et al., 1998; Berthenthal, 1996). These authors have appealed to the notion of dual-stream processing in order to explain odd paradoxes of object-directed behaviors in early infancy. In particular, a dissociation between the two streams has been invoked to explain the problems infants have when dealing with briefly occluded objects: across the first year of life, young infants appear unable to retain aspects of both surface-feature (e.g., identity) and spatial-temporal (e.g., location) object information (e.g., Wilcox and Schweinle, 2002; Wilcox and Chapa, 2004; Xu and Carey, 1996; Kaldy and Leslie, 2003; Kaufman, Mareschal, and Johnson, 2003). Indirect evidence of the independent processing of surfacefeature properties (a predominantly ventral function) and spatial-temporal properties (a predominantly dorsal function) of objects can be found in the literature addressing infant cognition (largely dealing with infants’ responses to

johnson, mareschal, and csibra: dorsal and ventral visual pathways

471

Figure 28.2 Gamma-band activity presented as a time-frequency analysis of the average EEG at four electrodes over the right temporal cortex during and after the object’s disappearance behind a visible or invisible occluder. Gray indicates the times of statistical difference. There was significantly greater gamma EEG activity in

the occlusion event than in the disintegration event before and just after the disappearance of the object. (Reprinted from J. Kaufman, G. Csibra, and M. J. Johnson, 2005. Oscillatory activity in the brain reflects object maintenance. Proc. Natl. Acad. Sci. USA 102(42):15271–15274, with permission.) (See plate 47.)

hidden objects) as well as infant perception (dealing with infants’ responses to visible objects). The perceptual evidence is presented first, and the cognitive evidence is presented second. A number of studies relying on preferential looking and dishabituation techniques also provide evidence for a feature/ spatial-temporal dissociation. These studies focus on infants’ abilities to form perceptual compounds involving a surface feature (e.g., color) with a spatial-temporal feature (e.g.,

motion). The role of movement in object perception is complex and has been investigated in great detail (see Burnham, 1987, for a review). Movement can have a number of roles in object perception. Movement can (1) act as a suppresser of feature perception, (2) act as a facilitator of feature perception, (3) be incidental to objects, and (4) act as a feature of objects. Contrary to Bower’s initial claims (Bower and Patterson, 1973), it is now well established that by 2.5 months infants can process the features of a moving object and relate

472

cognition

those features to the features of a stationary object (Hartlep, 1983; Hartlep and Forsyth, 1977; Day and Burnham, 1981). Some studies also suggest that infants can selectively attend to either surface feature or spatial-temporal information independently of the other dimension, suggesting that surface features and spatial-temporal object information can be processed independently. Burnham (1987) reports an experiment that looked at 4- and 6-month-olds’ abilities to develop a feature representation over different motion transformations, and infants’ abilities to develop a motion representation over different feature transformations. In one experiment, these authors habituated infants to an object moving in different ways. They then tested the infant with either the same object moving in a way, or a new object moving in the same way. They found that infants looked reliably longer at the novel object. The implication is that the infants had encoded the shapes of the objects and based their responses on the novel shape rather than the novel movement. Thus infants could encode shape independently of movement. In a second experiment, the authors habituated infants to different objects moving in the same way. They then tested the infants with a novel object either moving the same way or moving in a novel way. They found that infants looked reliably longer at the novel movement. The implication is that the infants had encoded the movements of the objects independently of shape and based their responses on the novel movement rather than the novel shape: infants could encode movement independently of shape. Further behavioral evidence comes from studies of infant responses to temporarily occluded objects. Early competence in preferential looking tasks that claimed to show the coordination of position and feature information in reasoning about hidden objects (e.g., Baillargeon, 1993) have—on close scrutiny—provided evidence only for the use of positional information in conjunction with size or volume information (Mareschal, 1997). Both size and volume are spatial dimensions that are likely to be encoded by the dorsal route. Thus these tasks solely provide evidence of dorsal route processing. Further data that support the suggestion of a dissociation between ventral and dorsal processing have recently become available. Infants fail to use surface-feature information to individuate and enumerate objects that move behind and out from a screen (Simon, Hespos, and Rochat, 1995; Xu and Carey, 1996; Leslie et al., 1998). In these studies, infants watched two different objects move in and out (one at a time) from behind an occluder. The screen was subsequently removed to reveal either 1 or 2 objects. Young infants consistently ignored surface-feature information and relied on spatial-temporal cues when assessing the number of objects behind the occluder as indexed by fixation time. Using a similar paradigm to Xu and Carey (1996), Wilcox (1999) systematically varied (one at a time) the features by which

pairs of objects differed when they appeared from behind the occluding screen. She found that at 4.5 months infants will spontaneously use shape and size information to individuate objects, but only at 7.5 months will they use surfacetexture information, and not until 11.5 months do they use color to individuate objects. More recently, Wilcox and Chapa (2004) found that infants could be primed to retain surface-texture or color information in certain occlusion tasks. They found that when infants were preexposed to the importance of a surface feature in a task, infants as young as 7.5 months would retain color information and those as young as 4.5 months would retain surface-texture but not color information. Thus, while young infants appear able to retain some object surface information, the age at which surface information (e.g., texture and color) is used in conjunction with spatial-temporal information to monitor the number of hidden objects behind an occluder corresponds to the age at which infants begin to succeed at manual retrieval tasks (i.e., 7.5–9.5 months). Mareschal and Johnson (2003) recently explored the conditions under which infants would retain visual information about objects characteristic of dorsal processing and those under which they would retain information characteristic of ventral processing. In this study, 4-month-olds were first familiarized with two objects moving in and out from behind two separate occluders. During testing, the infants first saw both objects move behind the occluders. There was then a 5-second retention interval, following which the occluders were raised to reveal four possible outcomes (see figure 28.3, bottom four panels, and plate 48): (1) the original two objects in their expected locations (baseline condition), (2) the same two familiar objects behind a single location (location violation condition; ST), (3) one familiar object and one novel object (identity violation condition; SF), and (4) the same two familiar objects, but in switched locations (binding violation condition). Mareschal and Johnson (2003) found that when the test objects were images of faces or monochromatic asterisks, the infants responded to an identity violation only and not to a violation of location. In contrast, when images of infant toys were used, the infants responded to violations of location only and not to a change in identity as they had when presented with faces or colored asterisks. In addition, the infants did not respond to the binding violation, whatever stimuli were used. Thus it appears that 4-month-olds can maintain either identity (color or face—ventral features) or location (a dorsal feature) following a brief occlusion, but not both. Moreover, they do not maintain the binding of the dorsal and ventral features (e.g., the link between a particular color and its location when there is more than one object present). Mareschal and Johnson proposed that stimuli that potentially support actions (e.g., images of baby toys) lead to the selective retention of location (dorsal) information, whereas

johnson, mareschal, and csibra: dorsal and ventral visual pathways

473

Familiarization trials

Test trials

Baseline trial

Locations (ST) trial

Figure 28.3 Examples of test and familiarization trials in Mareschal and Johnson (2003). The top four panels illustrate a single familiarization event, whereas the bottom four panels illustrate the four possible test trials. (Reprinted from D. Mareschal and M. H.

stimuli that do not support actions (e.g., monochromatic color stars) lead to the selective retention of identity (ventral) information. Taken together, these studies suggest that infants can process surface-feature information independently from spatial-temporal information and, at times, are unable to process both surface features and spatial-temporal information in conjunction. When viewed in the light of the fact that surface features are processed by the ventral route and that spatial-temporal information is processed (predominantly) by the dorsal route, these findings argue for a relative dissociation between dorsal and ventral route processing in early infancy. A Computational Model of Dorsal and Ventral Route Object Processing Figure 28.4 shows a schematic outline of the dual-route processing model described by Mareschal, Plunkett, and Harris (1999). This model was initially developed to account for a developmental lag in infant object-directed behaviors (Baillargeon, 1993). When tested on their memory of hidden objects with perceptual, surprisebased techniques, infants show a precocious understanding of hidden objects. However, it is not until much later that they are able to retrieve a hidden object, even though they possess the motor ability to do so. In short, there appears to be a lag between infants’ knowledge of hidden objects and their abilities to demonstrate that lag in an active retrieval task. The model uses the dual-route processing hypothesis to account for this lag. The model embodies the basic architectural constraints on visual cortical pathways revealed by contemporary neuroscience: an object-recognition network that develops

474

cognition

Identity (SF) trial

Binding trial

Johnson, 2003. The “what” and “where” of infant object representations in infancy. Cognition 88:259–276. Copyright 2003, with permission from Elsevier.) (See plate 48.)

Response Integration Network

Trajectory Prediction Network

100 outputs

100 outputs 75 hiddens

5 complex cells Object Recognition Network

100 visual memory units

Input retina (4-by-25 grid with 4 feature detector units per grid cell)

Figure 28.4 Schema of object-processing model. (Reprinted from D. Mareschal, K. Plunkett, and P. Harris, 1999. A computational and neuropsychological account of object-oriented behaviours in infancy. Dev. Sci. 2:306–317. Courtesy of Developmented Science.)

spatially invariant feature representations of objects, a trajectory-prediction network that is blind to surface features and computes appropriate spatial-temporal properties even if no actions are undertaken toward the object, and a response module that integrates information from the two latter networks for use in voluntary actions (a function attributed to the prefrontal cortex; Rao, Rainer, and Miller, 1997). Each pathway develops specific task-appropriate representations that persist beyond direct perception. It is successful in demonstrating how the requirement to integrate information across two object representations in a voluntary retrieval task can lead to a developmental lag relative to performance on surprise tasks that only require access to either spatialtemporal information concerning an occluded object or

surface-feature information accessed separately. Moreover, this lag only appears when the network is required to deal with hidden objects, as is the case with infants (von Hofsten and Rosander, 1996). Note that early surprise responses can arise from feature violations, from spatial-temporal violations, and even from both types of violation arising concurrently and independently, but not from a violation involving the integration of feature and spatial-temporal information concerning an occluded object. The model predicts that infants will show a developmental lag not just on manual search tasks but also on surprise tasks that involve such integration. Conversely, the model suggests that infants will show early mastery of response tasks that do not require the integration of information across cortical representations. The developmental lag between “surprise” responses and intentional search responses to occluded objects arises as a natural consequence of the associative learning process. Internal object representations developed over the complex cells and the hidden units persist when the object passes behind the screen, but decay with time. Hence activation levels drop when the object is occluded. The learning algorithm updates network weights in proportion to the sending unit’s activation level. For an identical error signal, the weight updates are smaller when the object is hidden given the lower activation of the sending units. Consequently, it will take longer to arrive at an equivalent level of learning for hidden as compared to visible objects. This outcome is not unique to the learning algorithm used in the current model. It will arise in any learning mechanism that updates weights in proportion to the sending unit activation, providing a clear example of how developmental behaviors are constrained by microlevel mechanisms. Finally, the model also shows different rates of development for the surface-feature and spatial-temporal modules. The spatial-temporal module is the slower of the two modules to develop. However, the precision of the anticipatory localization response is crucial in determining the slow rate of development. Identifying the exact next position of a moving object takes a long time. A corollary of this finding is that tasks that require less spatial-temporal accuracy would not be as delayed. This requirement of precision may account for why dorsal control of saccades is not evident at 6 months (discussed earlier), whereas dorsally controlled (but less accurate) reaching is evident by 6 months. Indeed, when accurate or detour reaching is required, infants continue to make errors until they are much older (Diamond and Lee, 2000). Insofar as the dorsal stream consists of multiple parallel systems closely coupled to motor output systems (Milner and Goodale, 1995), we may expect to find some systems operational by 6 months (e.g., those associated with a less accurate reaching response) while others will not be (e.g., those associated with accurate saccadic responses).

It is also worth noting the similarities between the performance of this model and adult neuropsychological data. There are documented cases of patients with ventral-stream damage but an intact dorsal stream who suffer from a kind of visual form agnosia (they are unable to recognize objects based on shape information alone) and yet are able to reach accurately and even catch objects (Goodale et al., 1991; Milner and Goodale, 1995). After training, the featurerecognition module of the model could be damaged in a way that does not affect its ability to respond with a targeted reach (Mareschal, 1997). As discussed previously, shape (or form) can be encoded down both pathways so that damage to the ventral-stream shape-processing function does not interfere with shape processing in the dorsal stream. A recent extension of the model has looked at using temporal synchrony as a mechanism for binding information that is processed separately in the dorsal and ventral streams (Mareschal and Bremner, 2005). Information in different streams relating to the same object can be temporarily bound together by synchronizing the rate of firing of neurons coding that information (Ward, 2003). The augmented model shows bursts of synchronization following the reappearance of an object whose features have been surreptitiously changed while behind an occluder. As discussed earlier, bursts of synchronized gamma-band activity have been found in 6month-olds following the unexpected disappearance of an object (Kaufman, Csibra, and Johnson, 2003, 2005). In summary, the computational model discussed illustrates possible causes and consequences of a lack of integration between the pathways in early infancy, and demonstrates that the degree of integration could be task dependent (i.e., tasks with hidden objects versus tasks with visible objects).

Discussion and future directions In this chapter we have explored empirical data and computational modeling pertaining to hypotheses about the development of the dorsal and ventral streams of visual processing. With regard to differential timetables of development in the two pathways, though ERP/EEG evidence provided support for the contention that the dorsal pathway is delayed relative to the ventral, behavioral evidence indicates that infants of 6 months and younger are capable of fairly well directed reaching to visible objects, a function attributed to the dorsal pathway. Thus we conclude that different aspects of dorsal-pathway function may emerge at different ages. Similarly, with regard to the ventral pathway, behavioral evidence from Wilcox (1999) indicates that, although infants can individuate objects on the basis of shape and size information at 4.5 months, only at 7.5 months can they use surface texture, and at 11.5 months color. We are thus cautious of general claims about the differential maturation of

johnson, mareschal, and csibra: dorsal and ventral visual pathways

475

the two streams of visual processing, and we suggest that more detailed analyses of different streams and structures within the dorsal and ventral pathways will be required. However, we suggest that there are a number of causes and consequences of differential development timetables in cortical visual processing. It is commonly accepted that the later developing a brain system is, the more scope there is for postnatal environmental influence. This may be particularly important for components of the dorsal pathway that have to integrate information from sensory systems with proprioceptive feedback to generate “body-centered” representations for action. As the body of the infant is developing physically, important factors such as length and weight of limbs continue to change. Thus it may be more important to retain plasticity in the dorsal pathway than in the ventral (see Gilmore and Johnson, 1998). Further, as the infant becomes more mobile toward the end of the first year of life, the nature of its experience of the world changes. From his or her stance as a relatively passive viewer of the world, the infant becomes able to manually explore objects and spatial environments. This new “infant-generated” experience no doubt helps shape appropriate dorsal pathway cortical circuitry. Recent evidence suggests that early dissociations between dorsally and ventrally processed visual information subserving object-directed behaviors may even persist into the third year of life. Deloache, Uttal, and Rosengren (2004) documented a number of striking scale errors in 2-year-olds’ action selection. When presented with miniature replicas of objects such as chairs or cars, 2-year-olds occasionally engaged in behaviors that were appropriate for the life-sized version of the object but not the miniature (e.g., trying to climb into a 6-inch car). For example, Deloache, Uttal, and Rosengren interpret these findings in terms of a dissociation between a visual recognition system and a motor control or motor planning system (see also Glover, 2004). Why might there be a dissociation within the dorsal pathway between eye-movement control and reaching? It is established that there is an effective subcortical (collicular) route for eye-movement control. As we suggested earlier, this route may be adequate for many situations in which saccades are required, and it is sufficient to produce an adaptive response in most circumstances. It is only when more challenging saccade paradigms are examined, such as the double-step paradigm, that maladaptive behavior is revealed. Thus, while there are clearly benefits to dorsalpathway control over eye movements, the gains to be made may be less than those for the cortical control of reaching. By delaying aspects of the development of the dorsal pathway, plasticity can be retained longer. However, there is a potential cost in that later-developing systems tend to be more vulnerable to disruptions in development (see Johnson, 2005). Thus we anticipate that the dorsal visual stream

476

cognition

should show more evidence of plastic changes following differences in experience in otherwise healthy children, but that the dorsal pathway will also show more evidence of deficits in some developmental disorders. There is evidence consistent with both of these predictions. In a series of experiments involving comparisons between congenitally deaf and hearing subjects, Neville and colleagues have found results indicating that the dorsal stream may be more modifiable in response to alterations in afferent input than the ventral pathway (Neville, 1995). For example, while visual responses recorded at the scalp to peripheral and transient signals are greatly enhanced in the congenitally deaf as compared to hearing persons, there are no differences between these groups in measured responses to foveal stimuli. In a parallel series of studies, Maurer and colleagues have investigated the effect of early visual deprivation on subsequent visual acuity (Bowering et al., 1996; see chapter 25 by Maurer, Lewis, and Mondloch, this volume). Children who had cataracts earlier in life subsequently showed deficits in their sensitivity to peripheral visual stimuli several years later. Even children who had unilateral cataracts subsequently showed prolonged deficits in their sensitivity to peripheral visual stimuli, leading the authors to conclude that cortical systems subserving peripheral vision appear to be more sensitive to experience early in life. In this chapter we have presented continuing research into the development of the dorsal and ventral visual pathways. In future studies we anticipate a tighter integration between neuroimaging, modeling, and behavior, which will lead us to a more cohesive account of the development of visually guided behavior during infancy. acknowledgments

We acknowledge financial support from MRC Program grant G97 15587, EC grant 516542 (NEST), and Birkbeck College. REFERENCES

Anderson, R. A., L. H. Synder, C. S. Li, and B. Stricanne, 1993. Coordinate transformations in the representation of spatial information. Curr. Opin. Neurobiol. 3:171–176. Atkinson, J., 1998. The “where and what” or “who and how” of visual development. In F. S. G. Butterworth, ed., The Development of Sensory, Motor and Cognitive Capacities in Early Infancy: From Perception to Cognition, 3–20. Hove, UK: Psychology Press. Atkinson, J., 2000. The Developing Visual Brain. Oxford, UK: Oxford University Press. Baillargeon, R., 1993. The object concept revisited: New directions in the investigation of infant’s physical knowledge. In C. E. Granrud, ed., Visual Perception and Cognition in Infancy, 265–315. Hillsdale, NJ: Lawrence Erlbaum. Balaban, C. D., and J. M. Weinstein, 1985. The human pre-saccadic spike potential: Influences of a visual target, saccade direction, electrode laterality and instruction to perform saccades. Brain Res. 347:49–57.

Bentin, S., T. Allison, A. Puce, E. Perez, and G. McCarthy, 1996. Electrophysiological studies of face perception in humans. J. Cogn. Neurosci. 8(6):551–565. Bertenthal, B. I., 1996. Origins and early development of perception, action and representation. Annu. Rev. Psychol. 47:431– 459. Boden, M. A., 1988. Computer Models of Mind. Cambridge, UK: Cambridge University Press. Bower, T. G. R., and J. G. Patterson, 1973. The separation of place, movement and object in the world of the infant. J. Exp. Child Psychol. 15:161–168. Bowering, E., D. Maurer, T. L. Lewis, H. P. Brent, and P. Riedel, 1996. The Visual Field in Childhood: Normal Development and the Influence of Deprivation. Developmental Cognitive Neuroscience Technical Report No. 96.1. Burnham, D. F., 1987. The role of movement in object perception by infants. In B. E. McKenzie and R. H. Day, eds., Perceptual Development in Early Infancy: Problems and Issues. Hillsdale, NJ: Laurence Erlbaum. Chugani, H. T., M. E. Phelps, and J. C. Mazziotta, 1987. Positron emission tomography study of human brain functional development. Ann. Neurol. 22:487–497. Csibra, G., M. H. Johnson, and L. Tucker, 1997. Attention and oculomotor control: A high density ERP study of the gap effect. Neuropsychologia 35:855–865. Csibra, G., L. A. Tucker, and M. H. Johnson, 1998. Neural correlates of saccade planning in infants: A high-density ERP study. Int. J. Psychophysiol. 29:201–215. Csibra, G., L. A. Tucker, and M. H. Johnson, 2001. Differential frontal cortex activation before anticipatory and reactive saccades in infants. Infancy 2:159–174. Csibra, G., L. A. Tucker, Á. Volein, and M. H. Johnson, 2000. Cortical development and saccade planning: The ontogeny of the spike potential. NeuroReport 11:1069–1073. Day, R. H., and D. K. Burnham, 1981. Infants’ perception of shape and colour in laterally moving patterns. Infant Behav. Dev. 4:341–357. de Haan, M., A. Oliver, and M. H. Johnson, 1998. Electrophysiological correlates of face processing by adults and 6-month-old infants. J. Cogn. Neurosci., Ann. Meeting Suppl. 36. DeLoache, J. S., D. Uttal, and K. S. Rosengren, 2004. Scale errors offer evidence for a perception-action dissociation early in life. Science 304:1027–1029. de Schonen, S., J. Mancini, and F. Leigeois, 1998. About functional cortical specialisation: The development of face recognition. In F. Simion and G. Butterworth, eds., The Development of Sensory, Motor and Cognitive Capacities in Early Infancy: From Perception to Cognition, 103–120. Hove, UK: Psychology Press. Diamond, A., and E.-Y. Lee, 2000. Inability of 5-month-old infants to retrieve a contiguous object: A failure of conceptual understanding or of control of action? Child Dev. 71:1477– 1494. Gilmore, R. O., and M. H. Johnson, 1997. Body-centred representations for visually-guided action emerge during early infancy. Cognition 65:B1–B9. Gilmore, R. O., and M. H. Johnson, 1998. Learning what is where: Oculomotor contributions to the development of spatial cognition. In G. Butterworth and F. Simion, eds., The Development of Sensory, Motor and Cognitive Capacities in Early Infancy: From Sensation to Cognition, 25–47. Hove, UK: Psychology Press. Glover, S., 2004. What causes scale errors in children? Trends Cogn. Sci. 8:440–443.

Goodale, M. A., A. D. Milner, L. S. Jakobson, and D. P. Carey, 1991. A neurological dissociation between perceiving objects and grasping them. Nature 349:154–156. Hartlep, K., 1983. Simultaneous presentation of moving objects in an infant tracking task. Infant Behav. Dev. 6:79–84. Hartlep, K., and G. A. Forsythe, 1977. Infants’ discrimination of moving and stationery objects. Percept. Mot. Skills 45:27–33. Jeannerod, M., 1988. The Neural and Behavioural Organization of Goal-Directed Movements. Oxford, UK: Oxford University Press. Jeannerod, M., 1999. A dichotomous visual brain? Psyche 5(25). http://psyche.cs.monash.edu.au/v5/psyche-5-25-jeannerod. html. Johnson, M. H., 2005. Developmental Cognitive Neuroscience, 2nd ed. Oxford, UK: Blackwell. Kaldy, Z., and A. M. Leslie, 2003. Identification of objects in 9-month-old infants: Integrating “what” and “where” information. Dev. Sci. 6:360–373. Kaldy, Z., and N. Sigala, 2004. The neural mechanisms of object working memory: What is where in the infant brain? Neurosci. Biobehav. Rev. 28:113–121. Kanwisher, N., J. McDermott, and M. M. Chun, 1997. The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17(11):4302–4311. Kaufman, J., G. Csibra, and M. H. Johnson, 2003. Representing occluded objects in the human infant brain. Proc. R. Soc. Lond. [Biol.] 270/S2:140–143. Kaufman, J., G. Csibra, and M. J. Johnson, 2005. Oscillatory activity in the brain reflects object maintenance. Proc. Natl. Acad. Sci. USA 102(42):15271–15274. Kaufman, J., D. Mareschal, and M. H. Johnson, 2003. Graspability and object processing in infants. Infant Behav. Dev. 26:516–528. Kurtzberg, D., and H. G. Vaughan, 1977. Electrophysiolo-gical observation on the visuomotor system and visual neuro-sensorium. In J. E. Desmedt, ed., Visual Evoked Potentials in Man: New Developments, 314–331. Oxford, UK: Clarendon Press. Kurtzberg, D., and H. G. Vaughan, 1982. Topographic analysis of human cortical potentials preceding self-initiated and visually triggered saccades. Brain Res. 243:1–9. Leslie, A. M., F. Xu, P. D. Tremoulet, and B. J. Scholl, 1998. Indexing and the object concept: Developing “what” and “where” systems. Trends Cogn. Sci. 2:10–18. Livingstone, M., and D. Hubel, 1988. Segregation of form, color, movement and depth: Anatomy, physiology and perception. Science 240(4853):740–749. Mareschal, D., 1997. Visual tracking and the development of object permanence: A connectionist enquiry. Unpublished doctoral dissertation, Oxford University, Oxford, UK. Mareschal, D., and A. J. Bremner, 2005. When do 4-month-olds remember the “what” and “where” of hidden objects? In Y. Munakata and M. H. Johnson, eds., Attention and Performance XXI: Processes of Change in Brain and Cognitive Development, 427–448. Oxford, UK: Oxford University Press. Mareschal, D., and M. H. Johnson, 2003. The “what” and “where” of infant object representations in infancy. Cognition 88:259–276. Mareschal, D., K. Plunkett, and P. Harris, 1999. A computational and neuropsychological account of object-oriented behaviours in infancy. Dev. Sci. 2:306–317. Milner, A. D., and M. A. Goodale, 1995. The Visual Brain in Action. Oxford, UK: Oxford University Press.

johnson, mareschal, and csibra: dorsal and ventral visual pathways

477

Neville, H. J., 1995. Developmental specificity in neurocognitive development in humans. In M. S. Gazzaniga, ed., The Cognitive Neurosciences, 219–231. Cambridge, MA: MIT Press. Newsome, W. T., R. H. Wvrtz, and H. Komatsu, 1988. Relation of cortical areas MT and MST to pursuit eye movements. II. Differentiation of retinal from extraretinal inputs. J. Neurophysiol. 60:604–620. Rao, S. C., G. Rainer, and E. Miller, 1997. Integration of “what” and “where” in the primate prefrontal cortex. Science 276:821–824. Richards, J. E., 2001. Cortical indices of saccade planning following covert orienting in 20-week-old infants. Infancy 2:135– 157. Rodman, H. R., C. G. Gross, and S. P. Scalaidhe, 1993. Development of brain substrates for pattern recogniton in primates: Physiological and connectional studies of inferior temporal cortex in infant monkeys. In B. d. Boysson–Bardies, S. d. Schonen, P. Jusczyk, P. MacNeilage, and J. Morton, eds., Developmental Neurocognition: Speech and Face Processing in the First Year of Life, 63–75. Dordrecht, The Netherlands: Kluwer Academic Press. Simon, T. J., S. J. Hespos, and P. Rochat, 1995. Do infants understand simple arithmetic? A replication of Wynn (1992). Cogn. Dev. 10:253–269. Tallon-Baudry, C., O. Bertrand, F. Peronnet, and J. Pernier, 1998. Induced gamma-band activity during the delay of a visual short-term memory task in humans. J. Neurosci. 18:4244– 4254. Ungerleider, L. G., ed. 1985. Two Cortical Pathways for Object Recognition and Spatial Perception. Berlin: Springer-Verlag.

478

cognition

Ungerleider, L. G., and M. Mishkin, 1982. Two cortical visual systems: Separation of appearance and location of objects. In D. L. Ingle and M. A. Ungerleider (1985), The cortical pathways for object recognition and spatial perception. In C. Changas and C. Gross, eds., Pattern Recognition Mechanisms, 21–37. Berlin: Springer-Verlag. Van Essen, D. C., C. H. Anderson, and D. J. Felleman, 1992. Information processing in the primate visual system: An integrated system perspective. Science 255:419–422. von Hofsten, C., and K. Rosander, 1996. The development of gaze control and predictive tracking in young infants. Vis. Res. 36:81–96. Ward, L. M., 2003. Synchronous neural oscillations and cognitive processes. Trends Cogn. Sci. 7:553–559. Webster, M. J., J. Bachevalier, and L. G. Ungerleider, 1995. Development and plasticity of visual memory circuits. In B. Julesz and I. Kovacs, eds., Maturational Windows and Adult Cortical Plasticity, 73–86. Reading, MA: Addison-Wesley. Wilcox, T., 1999. Object individuation: Infants’ use of shape, size, pattern and colour. Cognition 72:125–166. Wilcox, T., and C. Chapa, 2004. Priming infants to use color and pattern information in an individuation task. Cognition 90: 265–302. Wilcox, T., and A. Schweinle, 2002. Object individuation and event mapping: Developmental changes in infants’ use of featural information. Dev. Sci. 5(1):132–150. Xu, F., and S. Carey, 1996. Infants’ metaphysics: The case of numerical identity. Cogn. Psychol. 30:11–153. Xu, F., S. Carey, and J. Welch, 1999. Infants’ ability to use object kind information for object individuation. Cognition 70(2): 137–166.

29

Attention in Young Infants: A Developmental Psychophysiological Perspective JOHN E. RICHARDS

Attention, generally defined, shows dramatic development over the period of infancy. At birth infants attend primarily to salient physical characteristics of their environments or attend with nonspecific orienting (Berg and Richards, 1997). Between birth and two years the development of alert, vigilant sustained attention occurs. At the end of the first two years infants’ executive attention system is beginning to function (Ruff and Rothbart, 1996; Rothbart and Posner, 2001). These dramatic changes in infants are commonly thought to be based predominantly on age-related changes in brain structures responsible for attention control. The present chapter will attempt to accomplish three objectives. First, brain systems that may be involved in attention and that show development in infancy will be reviewed. These systems include a general arousal system that affects many cognitive functions, as well as specific attention systems that are limited in their effects on cognition and attention. Second, psychophysiological measures that have been useful in the study of brain-attention relations in infants will be presented. The use of heart rate as a measure of the general arousal system will be emphasized. Finally, several studies will be examined that used these psychophysiological methods to study the development of infant attention. This review will primarily emphasize the use of heart rate as an index of the development of sustained attention, which is a general arousal system affecting a wide number of behavioral and cognitive functions controlled by the brain. These experiments will be related to changes occurring in the neural systems underlying attention.

Brain systems involved in attention Arousal Attention System One emphasis in the cognitive neuroscience of attention has been on the arousal associated with energized cognitive activity (Posner, 1995). The arousal emphasis has focused upon the increased behavioral performance that occurs when attention is engaged. This increased behavioral performance is associated with shortening of reaction times in the course of detection tasks,

increased focus of performance on specific tasks, and the sustaining of performance over extended periods of time. The arousal emphasis is nonspecific, affecting multiple modalities, cognitive systems, and cognitive processes. Moreover, this arousal emphasis characterizes attention’s energizing effect on cognitive and behavioral performance. Attention also may have a selective effect on specific cognitive processes or behavior without arousal properties (next subsection). In fact, selective attention may serve in some situations to inhibit behavior if such inhibition is appropriate for the goal of the task. Specific locations or systems in the brain control the arousal aspect of attention. The brain systems underlying this attentional component have been detailed in the theoretical and empirical research literature for a number of years. An example of this arousal emphasis is a model of neuroanatomical connections between the mesencephalic reticular activating system and the cortex (Heilman et al., 1987; Mesulam, 1983). Figure 29.1 presents a diagram showing this system. This model presumes that information comes into the brain from visual, auditory, somesthetic, and other afferent pathways. These pathways have ascending connections through the thalamus to the cortex and descending connections to midbrain areas. The mesencephalic ascending reticular activating system influences parts of the thalamus that enhance sensory flow and at the same time stimulates extrinsic neurotransmitters. These effects directly or indirectly influence the limbic system, such as the basolateral nucleus of the amygdala and the subicular portion of the hippocampus, the cingulate cortex, prefrontal areas, and association areas (e.g., parietal area PG). This neuroanatomical system acts in synchrony to “energize” primary sensory areas in the cortex and increase the efficiency of responding in those areas. This system also influences association areas and other attention systems, such as the posterior attention system described by Posner (Posner, 1995; Posner and Petersen, 1990). The nonspecificity of this system is implied by its interconnections with multiple areas that influence cognitive processing. This arousal system

479

Eye Movement (FEF, SC)

Superior Parietal Lobe

Secondary Sensory Cortex

Primary Sensory Cortex

Visual (BA 18, 19)

Visual (BA 17)

Auditory (BA ..)

Auditory (STG)

Somatosensory (BA ..)

Somatosensory (PCG)

Ascending Reticular Activating System

Figure 29.1

cognition

LGN

MGN

VPL

The arousal system of the brain.

“invigorates” or “energizes” cognitive processes leading to increased processing efficiency, shorter reaction times, better detection, and sustaining of cognitive performance for extended periods of time. According to the model presented in figure 29.1, the arousal aspect of attention works through two mechanisms. The first mechanism involves the thalamus. The thalamus is the major sensory connection area between afferent activity and the cortex. The reticular nucleus and internal medullary lamina (figure 29.1, IML) play a role in this effect. The reticular nucleus is enhanced both by the ascending reticular activity and by feedback mechanisms from primary sensory cortex. In turn, its increased activity positively affects the activation of several lamina of the thalamus and thus enhances incoming sensory information. The IML acts as a connection area between midbrain reticular activity and other cortical areas. The second mechanism through which the arousal aspect of attention works is through neurochemical systems. Robbins and Everitt (1995; chapter 6 by Stanwood and Levitt, this volume) distinguish four neurochemical systems that form the basis for the arousal functions of attention: noradrenergic, cholinergic, dopaminergic, and serotoninergic. Figure 29.2 (plate 49) shows the projections from midbrain nuclei for these four brain systems. The nuclei that give rise to these four neurochemical systems are located in brain regions adjacent to the mesencephalic reticular activating system. Robbins and Everitt (1995) review

480

al su y Vi r to di ic et Au th es

m So

Extrinsic Neurotransmitters: Locus Coeruleus (NorE) Basal Forebrain (Ach)

Reticular nucleus

Pulvinar

Lat posterior

Dorsomedial

IML

Anterior Cingulate & Prefrontal Cortex

the evidence linking these neurochemical projection systems to attention and arousal. The noradrenergic and cholinergic systems are thought to be the neurochemical systems that are most closely involved in cortical arousal as it is related to attention. The dopaminergic system affects the motivational and energetic aspects of cognitive processing, and the serotonin system affects the overall control of state. These four neurochemical systems are closely linked so that more than one is likely to be operating during an aroused state. These four neurochemical systems also show changes over infancy that imply that the arousal controlled by these systems develops in that time period. Specific Attention Systems The second manner in which the brain affects the development of attention in infants is through brain systems that are specific to selected functions. These brain areas show enhanced functioning in the course of attention but affect only a single (or few) cognitive functions. Therefore, these systems have only a narrow impact on attention-based cognitive functioning. Two of these are worth mentioning in this respect. First, the enhancement of visual receptive fields during attention to visual stimuli has been widely studied by means of animal models using invasive preparations (Desimone and Duncan, 1995; Maunsell and Ferrera, 1995). This type of attention is selective for particular objects, particular spatial locations, or particular tasks. For example, the responses of visual

Cholinergic

Noradrenergic

Dopaminergic

Serotoninergic

Figure 29.2 The neurochemical systems involved in attention and arousal. Abbreviations: III, oculomotor nucleus; T, thalamus; HC, hippocampal formation; RF, reticular formation; PSG, parasympathetic ganglion cell; X, dorsal motor nucleus of the vagus; H,

hypothalamus; LC, locus ceruleus; C, caudate nucleus; P, putamen; S, septal nuclei; V, ventral striatum. (Reprinted with permission of the publisher from J. Nolte and J. B. Angevine, The Human Brain, pp. 134–137, St. Louis: Mosby. Copyright 1995.) (See plate 49.)

receptive fields are enhanced in tasks requiring focused allocation of attention to that specific visual field or to objects occurring in that visual field. Objects occurring outside of that receptive field provoke no responses when the field is irrelevant to the task, or they may elicit attenuated responses if the object occurring in that location interferes with task performance in the specific visual field. For neurons (or neural areas) that respond in this manner, this type of attentional modulation is specific to a limited number of cognitive

aspects (e.g., a specific stimulus, modality, or task) and typically occurs in a very restricted portion of the brain (e.g., individual neurons or restricted brain areas). Second, a specific attention system of interest to the development of attention is the “posterior attention system” described by Posner (Posner, 1995; Posner and Petersen, 1990; Rothbart and Posner, 2001). This system involves the parietal cortex, pulvinar, superior colliculus, and perhaps, the frontal eye fields. This attentional network has a specific

richards: attention in young infants

481

purpose, that of moving attention (visual attention?) around in space and localizing receptors (eyes?) to targets at specific locations. This attention system is not sensitive to specific targets, is unrelated to attention in stimulus modalities or cognitive functions that do not involve spatial localization, and does not enhance or attenuate other cognitive systems when it operates. The individual brain components of the posterior attention system show changes in infancy that affect the infant’s eye movements during attention or inattention. This system also is involved in “covert orienting,” or “covert attention,” which shows changes in young infants. Covert orienting or covert attention refers to cognitive processes whereby attention can be shifted to a new part of the visual field without making an overt eye movement to that location. These specific brain systems show development in the period of infancy and are related to behavioral indices of infant attention that show development in the same time period. Such considerations may be found in chapter 28 by Johnson, Mareschal, and Csibra and chapter 9 by Iliescu and Dannemiller in this volume, as well as in other sources (e.g., see eye movement–attention model of Johnson, in Johnson, 1990, 1995; Johnson, Gilmore, and Csibra, 1998; Johnson, Posner, and Rothbart, 1991). These specific attention systems will be covered insofar as they are affected by the arousal form of attention.

Psychophysiological measures of infant attention Psychophysiological measures are useful in the study of infant attention and infant brain development. Psychophysiology studies psychological processes using physiological measures and is focused on the psychological processes themselves as well as their relation to the processes affecting the physiological measures (Andreassi, 1989). The physiological measures used in psychophysiology are noninvasive and so may be used with human participants such as infants. Additionally, most of these physiological measures are also practical in psychological experiments. Recording equipment and sensors are nonintrusive, and the sensors do not disrupt the infant’s normal behavior patterns. The use of heart rate and EEG/ERP as psychophysiological measures of attention will be reviewed briefly as exemplars of this approach. Heart Rate The most common measure used by psychophysiologists studying young infants is heart rate (Reynolds and Richards, 2007). The electrocardiogram (ECG) is measured with surface electrodes placed on the infant’s chest, back, arms, or legs. Heart rate is derived from the ECG by measuring the interval between two “R-waves” of the ECG and is defined as the “interbeat interval” (IBI; R-R Interval), or as the inverse of the IBI, heart rate (beats

482

cognition

per minute, BPM). The infant’s heart rate may be measured in response to psychological manipulations as a measure of attention. The infant’s heart rate also may be used to determine whether the infant is attending to a stimulus, and psychological manipulations are then made on the basis of the heart rate change (e.g., Richards, 1987). Heart rate may be used to distinguish general and specific forms of attention. The author in several places (Berg and Richards, 1997; Reynolds and Richards, in press; Richards, 1995, 2001, 2004a; Richards and Casey, 1992; Richards and Hunter, 1998) has presented a model where infants’ heart rate changes during stimulus presentation are used to distinguish four attention phases. These phases are the automatic interrupt, the orienting response, sustained attention, and attention termination. Heart rate and attention levels vary during these phases. Figure 29.3 schematically depicts the heart rate changes occurring during these phases of attention. This figure represents heart rate changes of infants from 3 to 6 months of age presented with a visual stimulus (Richards and Casey, 1991). The figure also has labeled a “pre-attention” and “pre-attention termination” phase. These periods are simply the period of time before the presentation of the stimulus (pre-attention) and before heart rate returns to its prestimulus level but after sustained attention has occurred (pre-attention termination). Sustained attention and attention termination affect a wide range of cognitive functions in infants. The heart rate slows down and remains below prestimulus levels during sustained attention. Cognitively, this phase of attention involves subject-controlled processing of stimulus information. Sustained attention is accompanied behaviorally by maintaining fixation on a focal stimulus in the presence of a peripheral distracting stimulus (Hicks and Richards, 1998; Richards and Hunter, 1997; Lansink and Richards, 1997; Richards, 1987, 1997a), acquiring stimulus information (Richards, 1997b) and exhibiting recognition memory (Reynolds and Richards, 2005; Richards, 2003a; Richards and Casey, 1990), and enhancement of responses in a selected stimulus modality and inhibition of responses in a nonselected stimulus modality (Richards, 1998, 2000a). Alternatively, at the end of sustained attention, the heart rate returns to its prestimulus level, and the phase of attention termination occurs. Attention termination is accompanied by inattentiveness toward the stimulus in the presence of continued fixation on the stimulus—that is, heightened levels of distractibility, lack of acquisition of stimulus information, and lack of selective modality effects. The phases of sustained attention and attention termination are markers of the nonspecific arousal system of the brain (Reynolds and Richards, in press; Richards and Casey, 1992; Richards and Hunter, 1998; Richards, 2001, 2004a). The neural control of this heart rate change originates from

PreStimulus Sustained Attention Orienting Attention

Pre-Attention Attention Termination Termination

2 0 -2 Heart Rate -4 Change (bpm) -6 -8 -10 -5

0

5

10

15

20

25

Seconds following stimulus onset

Figure 29.3 Average heart rate change as a function of the stimulus following stimulus onset for the heart rate defined attention phases for infants from 3 to 6 months of age. (From Richards and Casey, 1991.)

cardioinhibitory centers in the orbitofrontal cortex by way of the vagus nerve (10th cranial nerve). This area has reciprocal connections with the limbic system and through these connections is involved in modulating activity within the mesencephalic reticular formation arousal system (Heilman et al., 1987; Mesulam, 1983) and probably the dopaminergic and cholinergic neurotransmitter systems (Robbins and Everitt, 1995; chapter 6 by Stanwood and Levitt, this volume). The cardioinhibitory centers act through the parasympathetic nervous system to slow heart rate when the arousal system is engaged. The heart rate changes occurring during sustained attention (sustained heart rate slowing) index the onset and continuing presence of this arousal. The heart rate changes during attention termination (return of heart rate to its prestimulus level) index the lack of activation of this arousal system. These two phases of attention therefore reflect the nonspecific arousal that may affect a number of sensory and brain systems. Incidentally, these phases and the “automatic interrupt” and “stimulus orienting” attention phases also may be used to measure specific attentional systems in the young infant (e.g., Berg and Richards, 1997; Balaban, 1996; Richards, 1998, 2000a). Other Psychophysiological Measures There are other psychophysiological measures that have been used in the study of infant attention and its development. Two in particular are worth mentioning: the electroencephalogram (EEG) and scalp-recorded event-related potentials (ERPs). These are reviewed extensively in this volume (chapter 15 by Csibra, Kushnerenko, and Grossman) and will not be reviewed in detail here. Spontaneous electrical activity of

very small magnitude may be recorded from the human scalp. However, EEG activity has been used in adults and infants as a measure of nonspecific arousal (e.g., Ray, 1990; Bell, 1998). This measure is interesting, because it is a more direct measure of neural activity than is heart rate and possibly could be used as a noninvasive measure of neural activity level enhanced by arousal. This chapter will not review the developmental changes occurring in EEG, but the reader should refer to other sources (e.g., Bell, 1998, 1999; Bell and Fox, 1992, 1994; Bell and Wolfe, 2007; Berg and Berg, 1987). Scalp-recorded ERPs are derived from the EEG recording (Csibra, Kushnerenko, and Grossman, this volume). The ERP is thought to reflect specific cognitive processes and therefore may provide a noninvasive and direct measure of functioning within specific brain areas (e.g., see Hillyard et al., 1995). For example, specific components of the ERP change in response to familiar and unfamiliar visual stimuli (Nelson and Collins, 1991, 1992). Nelson and Collins (1991, 1992) demonstrated changes in the amplitudes and latencies of specific ERP components in response to visually presented novel stimuli. Likewise, the ERP also may be used to index specific attentional responses. One such measure is the Nc (negative central) component (Courchesne, 1977, 1978), which is thought to represent a relatively automatic alerting response to the presence of a visual stimulus, especially a novel stimulus (cf., heart-rate-defined “stimulus orienting,” Richards and Casey, 1992; Reynolds and Richards, 2005; Richards, 2003a). The ERP has been used in the study of covert orienting (e.g., Richards, 2000b, 2000c, 2004b, 2005)

richards: attention in young infants

483

and thus might be used in the study of some specific aspects of attention. The ERP has been used extensively in infant participants, and many reviews of this measure are available (e.g., Berg and Berg, 1987; de Haan, 2008; Nelson, 1994; Nelson and Dukette, 1998; Csibra, Kushnerenko, and Grossman, this volume). Psychophysiological Measures as “Marker Tasks” Some comments should be made on the nature of the psychophysiological measures as direct or indirect measures of brain activity. Many psychophysiological measures are indirect measures of brain activity. Heart rate as an index of a general arousal system in the brain should be considered an indirect measure. The connections between the mesencephalic reticular activating system, its associated attention-arousal system (Heilman et al., 1987; Mesulam, 1983), and heart rate control are well known. Also, the connection between the neurochemical arousal systems (Robbins and Everett, 1995; chapter 6 by Stanwood and Levitt, this volume) and cardiac control are known. However, the measurement of such brain systems is indirect when using heart rate as a psychophysiological measure of infant attention. The indirect measure of brain activity with heart rate is similar to the “marker task” concept detailed by Johnson (1997; see Richards and Hunter, 2002). Marker tasks are behavioral tasks that have been studied in animal or invasive preparations and are controlled by specific brain areas or systems. Johnson (1995) proposes that such tasks may be used in infants and children with the understanding that developmental changes in these tasks should reflect developmental changes in the brain areas that control their functioning. In the case of behavioral marker tasks or psychophysiological measures, a solid theoretical or empirical basis for relating the measure to a brain system or controlling brain functions is necessary. The study of attention further requires that these brain systems be related to common attention functions (arousal, selection). Finally, heart rate or behavioral tasks should be used in experimental situations in which relevant psychological processes affect the physiological system (or behavioral marker task). Other physiological processes may affect the dependent variable independently of the psychological process in question (e.g., cardiovascular and energetic demands on heart functioning). It is important to ensure that changes in the physiological system are related to the psychological processes manipulated in the experiment rather than caused by other processes. However, with proper caution, marker tasks allow inferences to be made about brain development and help to inform a developmental cognitive neuroscience approach to attention. Some psychophysiological indices reflect brain activity more directly (Richards and Hunter, 2002). The EEG and ERP in some contexts are direct measures of brain function.

484

cognition

Both are generated by neural activity occurring in cell bodies or extracellular space. They are closely related in time to this neural activity and are generated in specific areas of the brain related to cognitive activity. The identification of the brain area generating the electrical activity cannot be done with the scalp-recorded electrical activity alone. However, cortical ERP measures using high-density EEG recording ( Johnson et al., 2001; Tucker, 1993; Tucker et al., 1994) may be used to hypothesize cortical sources of the electrical activity and thus identify specific brain regions involved in cognitive tasks (Michel et al., 2004; Nunez, 1990; Scherg, 1990; Scherg and Picton, 1991). Functioning of the cortical areas may be inferred from these cortical source localization procedures in a direct fashion. The use of the ERP and cortical source localization procedures as a direct measure in the study of attention is beginning with infant participants ( Johnson et al., 2005; Reynolds and Richards, 2005; Richards, 2005, 2006; see section in this chapter “Recognition of briefly presented visual stimuli”). Such use of the EEG and ERP should lead to a higher quality of information about the relation between the brain and attention in infant psychological development. The rest of the chapter will review studies that show the developmental changes that occur in the arousal form of attention. The first section will review some studies that show the effect of the developing arousal system on eye movements that themselves show development over the first six months of infancy (Hunter and Richards, 2003, submitted; Richards and Holley, 1999). The next section of the reviews will present some studies that show developmental changes in sustained attention that are related to a “higher cognitive function,” infants’ recognition of briefly presented visual stimuli (Frick and Richards, 2001; Reynolds and Richards, 2005; Richards, 1997b, 2003a; Richards and Casey, 1990). These studies will show that familiarization of patterns presented for only a few seconds during sustained attention will result in recognition memory (Frick and Richards, 2001; Richards, 1997b). This section also will present some new data that show that during attentive states, infants will recognize stimuli very quickly and will show appropriate EEG and ERP changes associated with recognition memory. These studies identify the brain origins of this activity (Reynolds and Richards, 2005; Richards, 2003a). These studies should be considered examples of how developmental psychophysiology may contribute to developmental cognitive neuroscience of attention.

Eye movements and attention This section will review the relation between the development of the arousal attention system in young infants and three eye-movement control systems that show development in the same period of time (see also chapter 16 by Karatekin,

this volume). There are three types of eye movements that may be made when tracking visual stimuli. Each eyemovement type is controlled by separate areas of the brain. “Reflexive saccadic” eye movements occur in response to the sudden onset of a peripheral stimulus. These eye movements are controlled by a brain pathway involving the retina, lateral geniculate nucleus, superior colliculus, and perhaps, the primary visual area (Schiller, 1985, 1998). “Voluntary saccadic” eye movements occur under voluntary or planned control. These eye movements often involve attentiondirected targeted eye movements. The voluntary saccadic eye movements are controlled by a brain pathway involving several parts of the cortex including extrastriate occipital areas, the fusiform gyrus, the parietal cortex area PG, and the frontal eye fields (Schiller, 1985, 1998). The third type of eye movements used in tracking visual stimuli are “smooth pursuit” eye movements. These eye movements occur only in the presence of smoothly moving visual stimuli, and they smoothly track visual stimuli over a wide range of visual space. Smooth pursuit eye movements also are controlled by brain pathways involving the cortex, including area MT (medial temporal), areas MST (middle superior temporal), and perhaps the parietal cortex (Schiller, 1985, 1998). The voluntary saccadic and smooth pursuit eye movements are affected by attention, whereas reflexive saccadic eye movements are relatively independent of attention control. The brain areas involved in the control of these three eye-movement systems undergo developmental changes in the first six months. There have been several models of the brain changes affecting eye-movement development, including models by Bronson (1974, 1997), Maurer and Lewis

(1979, 1991, 1998), Johnson and colleagues ( Johnson, 1990, 1995; Johnson, Posner, and Rothbart, 1991; Johnson, Gilmore, and Csibra, 1998; chapter 28 by Johnson, Mareschal, and Csibra, this volume), Hood (Hood, 1995; Hood, Atkinson, and Braddick, 1998), and Richards (Richards and Casey, 1992; Richards and Hunter, 1998). Iliescu and Dannemiller (chapter 9, this volume) review several pertinent “neurodevelopmental” models. Johnson’s model (1990, 1995; Johnson, Posner, and Rothbart, 1991; Johnson, Gilmore, and Csibra, 1998) describes the developmental changes in these three eye-movement systems. This model hypothesizes that layers of the primary visual area develop at different rates and become mature at different ages. The primary visual area layers containing brain pathways that control reflexive eye movement are relatively mature at birth, and therefore reflexive saccadic eye movements dominate the infant’s behavior in the first two postnatal months. The primary visual area layers that contain brain pathways that control voluntary saccadic eye movements develop rapidly from the first to the sixth postnatal month. In conjunction with this development, attention-directed voluntary saccades show developmental changes over the first six months. Finally the primary visual area layers that contain brain pathways that control smooth pursuit eye movements develop more slowly than the other layers. Several parts of the brain pathways that control smooth pursuit eye movements show protracted developmental changes over the first two years (Richards and Hunter, 1998). Thus smooth pursuit eye movements are the latest to begin development and show changes over a longer period than just the first six months of infancy. Figure 29.4 (from Richards and Hunter,

Hypothesized Development of Three Eye Movement Systems 100

Percent Maturity

Reflexive Saccades Targeted Saccades

80 60 40

Smooth Pursuit

20 0 0

1

2

3

4

5

6

7

8

9

10

11

12

Month Figure 29.4 Development of three visual systems involved in visual tracking. The percent maturity is shown as a function of months, from birth through 12 months. There are three lines,

corresponding to the reflexive saccadic eye movements, voluntary saccadic eye movements (“Targeted Saccades”), and smooth pursuit eye movements. (From Richards and Hunter, 1998.)

richards: attention in young infants

485

1998) shows a hypothetical developmental trend for these three eye-movement systems. One study (Richards and Holley, 1999) examined the effect of attention on all three eye-movement types. In this study infants’ tracking behavior was monitored over this age range under conditions of attention and inattention. This study shows how the development of the general arousal system affects the exhibition of eye movements in the first six months of infancy. Infants at 8, 14, 20, and 26 weeks of age were presented with stimuli that moved at varying speeds (8 to 24 degrees per second) on a television monitor. The infants’ heart rate was recorded, and periods of visual

tracking were separated into attentive and inattentive states using the heart-rate-defined attention phases described earlier. The infants’ eye movements were recorded with the “electrooculogram” (EOG) by recording electrical potential changes due to shifts in the eyes. The eye movements were separated into smooth pursuit and saccadic eye movements and related to the attentiveness of the infant. There were two important findings from that study. First, there was an increase in tracking ability over this age. This increase in tracking occurred in both the infants’ use of smooth pursuit eye movements and in saccades. Figure 29.5 shows smooth pursuit and saccadic eye movement results

Weeks Legend 8 & 14 20 & 26

Sustained HR Deceleration 2.00

Saccade Frequency

Smooth Pursuit Gain

1.0 0.8 0.6 0.4 0.2 0.0

1.50

1.00

0.50

0.00 4

9

14

19

4

24

Degrees per second

9

14

19

24

19

24

Degrees per second

Return of HR to Prestimulus Level 2.00

Saccade Frequency

Smooth Pursuit Gain

1.0 0.8 0.6 0.4 0.2

1.00

0.50

0.00

0.0

4

9

14

19

24

Degrees per second Figure 29.5 The smooth pursuit EOG gain and saccade frequency (saccades per second) as a function of stimulus tracking speed and testing age (8 and 14 weeks combined, 20 and 26 weeks combined). The top two plots were taken from the period when

486

1.50

cognition

4

9

14

Degrees per second sustained heart rate deceleration was occurring, and the bottom two plots were taken from the period after heart rate had returned to its prestimulus level. (Adapted from Richards and Holley, 1999, figure 4, and Richards and Hunter, 1998, figure 4.7.)

under conditions of attention and inattention. The lower right part of figure 29.5 shows the saccade frequency occurring during the inattentive periods. These would be most similar to the reflexive saccadic eye movements. The younger and older infants show approximately equal numbers of these eye movements. The upper panels of figure 29.5 show saccade frequency and smooth pursuit gain during sustained attention, corresponding to voluntary saccadic and smooth pursuit eye movements. Both showed improvement from the youngest to the oldest ages. These findings show the expected age changes for these three eye movement systems as might be predicted from figure 29.4. The second important finding is related to the speed of the stimulus. The tracking stimulus was presented at speeds ranging from “very slow” to “very fast” for the capabilities of infants’ smooth pursuit (Richards and Holley, 1999). The reflexive saccadic eye movements were unresponsive to the stimulus speed (figure 29.5, lower right panel), whereas smooth pursuit tracking and saccadic tracking during attention were responsive to stimulus speed (figure 29.5, upper panels). When the stimulus became too fast for smooth pursuit eye movements to follow, the infants shifted from smooth pursuit tracking to saccadic tracking (figure 29.5, cf. left and right panels). Thus the oldest infants during aroused attentive states used the smooth pursuit and voluntary saccadic eye movements to track the visual stimulus and adjusted the parameters of the eye movements according to the speed of the tracking stimulus. The results from this study suggest at least two roles that sustained attention may play in behavior. First, the arousal system of the brain acts to energize specific brain systems involved in cognitive activities. In this study, the general level of increased performance during sustained attention reflects this arousal. The simultaneous development of the eye movement systems (smooth pursuit, voluntary saccadic) and arousal system (sustained attention) resulted in a synchrony between attention and eye-movement control. Second, sustained attention does more than just energize involved systems. Tracking behavior during sustained attention was preserved over increases in tracking speeds by shifting from smooth pursuit tracking to saccadic tracking when smooth tracking failed. The attention-arousal system also affects brain areas that select the appropriate behavior given the feedback being received from the stimulus display and whatever goals the infant has in the situation. A second way in which the relation between eye movements and attention has been assessed has been by examining the physiological characteristics of saccades. One characteristic of eye movements is that the maximum velocity of the saccade and the total saccade amplitude are related—that is, the “main sequence” in eye movements (Bahill, Clark, and Stark, 1975). The main sequence is the direct result of the firing rate and firing duration of

the brain-stem motor neurons that control ocular muscles. This area of the brain is hypothesized to be relatively well developed at birth and therefore should not show many changes. According to the neurodevelopmental models cited earlier, one might expect that, since these structures are well developed, the main sequence should be relatively fixed at very early ages for infants. In several studies, the development of the main sequence has been studied (Hunter and Richards, 2003, submitted; Richards and Hunter, 1997). Richards and Hunter (1997) recorded eye movements using the electrooculogram (EOG). Infants at 14, 20, and 26 weeks of age were presented with visual stimuli in the periphery to which a saccade was made. The main sequence was easily seen in the EOG recording and did not differ across these ages. So, in accord with the neurodevelopmental model, this system seemed to be functioning at similar levels in infants. However, more recently we examined the eye movements of younger infants during “free viewing” of interesting audiovideo stimuli (Hunter and Richards, submitted). In this study in the youngest infants, we found a decrease in the linear relation between maximum saccade velocity and saccade amplitude, the main sequence, from 5 to 14 weeks of age, but no difference from 14 to 26 weeks of age. Figure 29.6 shows plots of the velocity/ amplitude relation for infants from 5 to 26 weeks of age. The slope of the linear component of the main sequence decreases from 5 to 14 weeks of age. It did not change for the 20- and 26-week-old-infants. This finding implies that the low-level system involving the brain-stem eye-movement-control areas, motoneurons, and ocular muscles did show postnatal age changes. The effect of attention on the main-sequence relation also has been studied (Hunter and Richards, submitted; Richards and Hunter, 1997). The times at which the infant made an eye movement to the peripheral stimulus (Richards and Hunter, 1997) or during the free viewing of an interesting audiovisual stimulus (Hunter and Richards, submitted) were separated into those trials where sustained attention was occurring or where the infant was inattentive (i.e., attention termination). When attention was engaged, either to a specific stimulus or generally to the visual display, the older infants in these studies had a slower peak velocity per saccade amplitude than the younger infants. These age changes also have been studied in infants and older children (Hunter, 2001). There seemed to be an increase in the amount of depression of the main-sequence relation from 26 weeks to 1 year and from 1 year to about 7 years of age. However, after 2 years of age the difference between the main sequence under conditions of attention versus inattention were not as great. The results from these studies of eye movements and attention are revealing about the development of the arousal system and how it affects behavior that is controlled by other

richards: attention in young infants

487

488

cognition

20 weeks

14 weeks

26 weeks

11 weeks

Figure 29.6 The main-sequence relation between maximum saccade velocity and saccade amplitude for infants from 5 to 26 weeks of age. The lines are the best-fitting linear and quadratic regression lines. (From Richards and Hunter, 2002.)

8 weeks

5 weeks

brain systems. There may be age changes in the underlying brain areas that are modulated by attention. The saccade system showed development in these studies both in the duration of tracking (Richards and Holley, 1999) and in the main sequence relation (Hunter and Richards, submitted). The effect of attention on the main sequence may be due to an increasing top-down influence of the frontal eye fields (and other cortical systems) on the brain-stem eye-movement areas after the age of about 4 months. In another case, it appeared to act directly to energize the infant’s behavior in the support of a goal such as tracking a smoothly moving object.

Recognition of briefly presented visual stimuli This section will review studies showing the effect of sustained attention on infant recognition memory. Infant recognition memory often is studied with the paired-comparison procedure (Fagan, 1974). In this procedure infants are familiarized with a single stimulus (familiar stimulus) during a familiarization phase. Then, during the recognition-memory test phase, the familiar stimulus is paired with a stimulus not previously seen (novel stimulus). Recognition memory for the familiar stimulus is inferred if the infants show a novelty preference, that is, look longer at the novel stimulus than the familiar stimulus during the paired-comparison test phase (see chapter 30 by Bachevalier and chapter 33 by Richmond and Nelson, this volume). Two studies using heart-rate-defined attention phases have shown that exposure to the familiar stimulus during sustained attention results in recognition memory for stimuli presented for only 5 or 6 seconds (Frick and Richards, 2001; Richards, 1997b). In these studies infants at 14, 20, or 26 weeks of age were presented with a “Sesame Street” movie, Follow That Bird, on a television monitor. This movie is very interesting to young infants and reliably elicits the full range of heart rate changes that are related to the attention phases. On separate trials, at a delay defined by the deceleration of heart rate, a delay defined by the return of heart rate to its prestimulus level, or time-defined delays, a familiarization stimulus was presented for 5 or 6 seconds. One condition with the “Sesame Street” movie alone was provided (no familiarization stimulus, i.e., no exposure control), and one condition with 20 seconds of exposure to the familiar stimulus was presented. The no-exposure control was given to get a baseline level of preference in the absence of any exposure, and the 20-second exposure condition was given to determine how an extended exposure would affect preferences. Following each familiar stimulus presentation, a paired-comparison recognition-memory test was done. The infants’ duration of fixation to the novel and the familiar stimulus during the first 10 seconds of the test phase was recorded.

There were several results that showed that the infants recognized the familiar stimulus and preferred to look at the novel stimulus in the test phase, with only 5 seconds of familiar stimulus exposure. For example, when compared to the no-exposure control trial, infants looked longer at the novel stimulus than at the familiar stimulus. Furthermore, infants looked at the novel stimulus in the test phase for the brief exposure trials (5 or 6 s) for just as long as they did during the traditional 20-second accumulated fixation exposure trial. The most interesting result from these studies is illustrated in figure 29.7 (from Richards, 1997b). This figure shows the duration of the exposure to the familiar stimulus during the familiarization phase, but for different lengths of exposure during the sustained heart rate deceleration. That is, for some trials, the infants’ sustained attention overlapped the familiar stimulus exposure for only a brief period of time (e.g., <1 s), and on other trials the overlap was much greater (e.g., >5 s). This exposure is shown for different trials, and the percent fixation on the novel stimulus in the recognition-memory test phase is plotted. A very brief overlap of sustained attention and the familiar stimulus resulted in novelty preference scores at or below the no-exposure control condition. As the amount of familiar stimulus exposure during sustained attention was increased, there was a corresponding increase in novelty preference. This positive correlation between familiar exposure during sustained attention and later recognition memory level (novelty preference level) implies that incorporation of stimulus information is accomplished when the infant is in a highly aroused (and therefore presumably attentive) state. We also have shown that the distributions of the fixations on the novel and familiar stimuli in the test phase of the paired-comparison recognition memory procedure are affected by the infants’ attention state (Richards and Casey, 1990). In that study, heart rate was recorded and the heartrate-defined attention phases were evaluated during the test phases of the recognition-memory procedure. The infants showed novelty preference, indicating recognition memory, primarily during sustained attention. For example, on the average in these 3- to 6-month-old infants, there were about 11.8 seconds of sustained attention on the recognitionmemory test phase. Within this interval, about 7.3 seconds were spent looking at the novel stimulus, and 4.5 seconds were spent looking at the familiar stimulus. Alternatively, during attention termination (or inattentiveness) the infants spent equal amounts of time looking at the novel and familiar stimuli. And, on no-familiar-stimulus trials (no-exposure control) there were equal amounts of looking at the novel and familiar stimuli during each phase. These results show that the exhibition of recognition memory generally takes place during sustained attention, when heart rate is below

richards: attention in young infants

489

55

Percent Fixation on Novel Stimulus

Immediate

50

Return of HR to Prestimulus + 5-S

20 - Second Accumulation

45 Heart Rate Deceleration No-Exposure Control

40

Return of HR to Prestimulus

35

30 0.0 - 0.5 s

0.6 - 2.5 s

2.6 - 5.0 s

> 5.0 s

Duration of Familiar Stimulus Exposure during Heart Rate Deceleration Figure 29.7 The duration of familiar exposure occurring during heart rate deceleration (sustained attention) and the percent fixation on the novel stimulus in the recognition-memory test phase.

The “no-exposure control” time (40%) should be considered the baseline percent fixation with no exposure to the familiar stimulus. (Adapted from Richards, 1997b, figure 3.)

baseline. That is, the infants recognize the familiar stimulus and move fixation to the novel stimulus. This move to the novel stimulus most likely is to acquire new stimulus information. Thus the exhibition of recognition memory during this paired-comparison procedure, shown as novelty preference, is precisely the infant’s attempt to acquire new information from the previously unseen stimulus during sustained attention! Two recent studies show the effect of attention on individual cognitive processes that may occur in the brain after exposure to familiar and novel stimuli. Nelson and his colleagues (Nelson and Collins, 1991, 1992; Nelson and deRegnier, 1992; Nelson and Salapatek, 1986; also see reviews by Nelson, 1994, Nelson and Dukette, 1998, and de Haan, 2008) and others (Karrer and Ackles, 1987, 1988; Karrer and Monti, 1995; Courchesne, 1977, 1978; Courchesne, Ganz, and Norcia, 1981) have examined infant recognition memory by recording ERPs during stimulus presentations of very brief duration (∼150 ms). These studies use the “oddball” paradigm in which one stimulus is presented relatively frequently and a second stimulus is

presented infrequently. These studies report a large negative ERP component occurring about 400–800 ms after stimulus onset located primarily in the frontal and central EEG leads. This has been labeled the Nc component (Nc is negative central; Courchesne, 1977, 1978). In most studies, the Nc component is larger to the infrequently presented stimuli and is thought to represent a general attentive state or alerting to the presence of a novel stimulus. If the frequently presented and infrequently presented stimuli are already familiar to the infant, the Nc component does not differ (Nelson and Collins, 1991, 1992). This distinction does not occur in 4-month-old infants (e.g., Karrer and Ackles, 1987; Nelson and Collins, 1991, 1992) but occurs in 6-month-old and older infants. Two recent studies examined these ERP measures of brain activity and their relation to attention. In both studies, attention was first elicited by showing a “Sesame Street” movie, Follow That Bird, that elicits the heart rate changes that define sustained attention and inattentiveness. Then, during sustained attention or attention termination, brief visual stimuli were presented overlaid on (replacing) the

490

cognition

attention-eliciting stimulus. The brief stimuli had been previously familiarized (frequent familiar, infrequent familiar) or were novel on each trial (infrequent novel). The ERP responses to these stimuli were recorded and separated into those that occurred when the infant was attending to the stimulus (sustained attention) or showing inattentive visual regard (attention termination). There was a close relation between the size of the Nc ERP component and the infant’s attentiveness. Figure 29.8 (plate 50) shows the Nc response occurring during attention and inattention for the three stimulus types. There was a larger Nc during sustained attention, and this was observed regardless of the familiarity (familiar, novel) or frequency (frequent, infrequent) of the stimulus. There also were age changes in the amplitude of the Nc from 20 to 26 to 32 weeks of age. These age changes occurred predominantly in sustained attention. The close association of the Nc with attention suggests that this component reflects a general process of attention orienting to the stimulus rather than a specific measure of recognition memory (cf., Nelson, 1994; Nelson and Dukette, 1998; Nelson and Monk, 2001). The second study used similar procedures and points to specific brain areas that may generate the Nc ERP component (Reynolds and Richards, 2005). In that study a high-density EEG recording (128 electrodes) was used ( Johnson et al., 2001; Tucker, 1993; Tucker et al., 1994). A presentation procedure similar to that in Richards (2003a) was used, and again it was found that the Nc component was significantly affected by attention. Since the high-density recording was used, the ERP components could be analyzed with cortical source localization procedures. Two cortical sources were found that are of interest. First, a cortical source was located that had dipole locations primarily in the frontal pole of the brain. Second, a cortical source was located that had dipole locations scattered throughout several areas of the prefrontal cortex. Figure 29.9 (plate 51) shows the latter cortical locations. Figure 29.10 shows the activity of these cortical sources with respect to the experimental conditions. Both locations seem to be involved in the generation of the Nc ERP component, and both were affected by attention. The “prefrontal” source showed activity with its maximum occurring at about 500 ms following stimulus onset, which was the same time course as the Nc in that study. On novel stimulus presentations this activity occurred nearly immediately (upper right panel, solid line) and was sustained throughout the time course of the Nc. The “frontal pole” component showed activity later with respect to the stimulus, near the end of the Nc occurrence (figure 29.10, bottom figures). However, this occurred primarily for the attention trials on the novel stimulus presentations. This second brain area might be involved in the latter phases of the Nc and in the upcoming slow waves occurring in this task.

The relation between sustained attention and infants’ recognition of briefly presented visual stimuli shows that the arousal form of attention is related to complex infant cognition. Recognition memory is accomplished by several brain areas and cognitive functions. It requires the acquisition of stimulus information and memory storage over some period of time. The measurement of recognition memory also requires performance on a task exhibiting the existence of the stored memory. The results of these studies show that the arousal aspect of attention may “invigorate” each of these cognitive processes. This process enhances familiarization when information acquistion is occurring, may facilitate memory consolidation during the waiting period, and enhances the processes involved in the exhibition of recognition memory. The effect on recognition memory is true for the overall responses to the stimulus in the pairedcomparison recognition-memory test phase (Richards and Casey, 1990) and for the individual cognitive processes occurring for transient responses to the stimulus (Reynolds and Richards, 2005; Richards, 2003a). The enhancement of the Nc ERP component during attention implies that the general arousal system represented by sustained attention affects specific memory or attention processes that have specified locations in the brain. The facilitative effect of attention on infant recognition memory (e.g., Richards, 1997b; Frick and Richards, 2001) may occur because specific brain areas responsible for information acquisition or recognition are enhanced during attention (Reynolds and Richards, 2005; Richards, 2003a). Some comments will be made about the use of the cortical source analysis in this study. If we assume that the activity of specific neurons or groups of neurons is responsible for the electrical activity recorded in the EEG and ERP (chapter 15 by Csibra, Kushnerenko, and Grossmann, this volume), then this electrical activity might be considered a direct measure of brain activity. Analysis of the sources of this electrical activity provides an estimate of the location in the brain of the activity occurring on the scalp. For ERP, this activity is coordinated with experimental events or cognitive activity, so the source analysis provides a measure of “event-related brain activity.” This process is analogous to the functional neuroimaging provided by the BOLD response of the fMRI (chapter 18 by Thomas and Tseng, this volume) or the NIRS response (this volume, chapter 19). The advantage of using ERP is that it provides a neuroimaging modality that is generated by neural activity rather than vascular activity and therefore has the same timescale as neural responses. The cortical sources inferred with this approach contain some unresolved issues for infant work. The models used in these studies (e.g., Johnson et al., 2005; Reynolds and Richards, 2005; Richards, 2005) are based on parameters

richards: attention in young infants

491

492

cognition Infrequent Familiar

Infrequent Familiar

mV +10

Infrequent Novel

Infrequent Novel

Figure 29.8 The Nc component during attention and inattention. The ERP recording from 100 ms prior to stimulus onset through 1 s following stimulus is shown for the FZ and CZ electrodes for attentive (top figures) and inattentive (bottom figures) periods, combined over the three testing ages. The topographical scalp potential maps show the distribution of this component for the three memory stimulus types in attention and inattention. The topographical maps represent an 80-ms average of the ERP for the Nc component at the maximum point of the ERP response. The data are plotted with a cubic spline interpolation algorithm and represent absolute amplitude of the ERP. (From Richards, 2003a, figure 2.) (See plate 50.)

-10 mV

Frequent Familiar

Nc During Inattention

Frequent Familiar

Nc During Attention

richards: attention in young infants

493

-20mV

ICA Projection

Figure 29.9 The ICA component cluster for the prefrontal component. The topographical map of the average ICA loadings is similar to the topographical map of the grand average ERP of the Nc component. The ECD locations are displayed on several MRI slices, and each location represents an ICA from one individual. (From Reynolds and Richards, 2005, figure 4.) (See plate 51.)

20mV

Grand Average ERP

Prefrontal ICA Cluster: Medial Frontal Gyrus (25), Inferior Frontal Gyrus (47), Anterior Cingulate Cortex (8) (Talairach coordinates: 9.4, 42.9, 16.4)

Figure 29.10 The ICA activations for the frontal clusters for 1 second following stimulus onset. The left panel displays combined responses to frequent familiar and infrequent familiar stimulus presentations separately for periods of attention and inattention. The

right panel displays responses to infrequent novel stimulus presentations separately for periods of attention and inattention. (From Reynolds and Richards, 2005, figure 6.)

derived from adult participants. These include impedance (resistance) values, cortical matter, skull, and scalp of adult participants. Adult values of impedance are higher than those in infants. The use of adult impedance values with infant participants may have the effect of inferring the source of the current on infant participants as being deeper in the cortex than where it actually occurred. Second, the use of these models, even with adults, is preferable when structural MRIs exist that can constrain the dipole locations to realistic topographies derived for each participant (Richards, 2003b, submitted). These concerns are particularly relevant for infant participants where skull irregularities (unseamed sutures, thin skull) and head topography (scalp thickness, lobe location) may differ greatly from adults. Because of these limitations, it is appropriate to make these conclusions about the cortical sources of the Nc response with some caution. Notwithstanding these problems, however, the localization of the cortical sources of these ERP components

is a great advance in the study of the ERP components of infant recognition memory (Richards, 2006). In the first edition of this volume (Richards, 2001), I speculated that the measurement of brain function with cortical source localization methods had a bright future for the study of infant attention. The work reviewed in the current chapter shows that this potential is being realized. This work is still in its infancy, but the use of these techniques should be profitably applied to an understanding of the developmental changes in brain areas that are involved in the development of infant attention.

494

cognition

Summary and conclusion Attention shows dramatic development in the early period of infancy, from birth to 12 months. This chapter has emphasized an attention system that represents a general arousal of cognitive functions. The system in the brain con-

trolling this arousal develops in the first few months of life, and this brain development is responsible for the behavioral/attentional development seen in young infants. This chapter reviewed several studies that showed the effect of this arousal system, indexed by heart rate changes showing sustained attention. There were developmental changes in infant sustained attention that were reflected in developmental changes in specific attentional systems or that corresponded to developments occurring in other brain-based attention-directed infant behavior. There are two ways in which future research and progress in the study of the development of attention arousal in infants could progress. First, this review was limited to studies using heart rate as a measure of the general arousal system in the brain. There are other measures that may be useful in this regard. For example, continuous levels of EEG activity are thought to be influenced by general arousal mechanisms in the brain. Since the EEG represents the summed activity of large groups of neurons, one might expect that the brain areas controlling arousal or the neurochemical systems should have an influence on overall neural activity (extent, duration, and localization). Thus measures of EEG such as spectral power and coherence may give information about arousal. Such measures also may show relatively localized CNS arousal. A second area in which research on the development of the brain systems controlling arousal may benefit is through the use of direct measures of the brain. Such measures in animal models have included invasive chemical manipulations and measurement as well as destruction of the areas controlling arousal through lesions or neurochemical inhibitors. These measures cannot be applied in infant participants because of ethical considerations. However, noninvasive measurements from psychophysiological measures that are tuned to specific neurochemical systems might be found. Perhaps one type of quantitative activity in the EEG may be linked to a specific neurochemical system and another type linked to another system. The simple recording of EEG, ERP, or heart rate cannot be used to distinguish the four arousal systems detailed in Robbins and Everitt (1995). The EEG and heart rate would be expected to respond to any manipulation of an underlying arousal system. Some type of quantitative activity in the EEG would have to be linked to the underlying neurochemical system in order to use psychophysiological measures for this direct evaluation of the brain systems controlling this arousal form of attention. acknowledgments

The writing of this chapter was supported by a grant from the National Institute of Child Health and Human Development, R01-HD19842. Correspondence concerning this chapter should be addressed to John E. Richards, Department of Psychology,

University of South Carolina, Columbia, SC 29208. Electronic mail may be sent to [email protected].

REFERENCES Andreassi, J. L., 1989. Psychophysiology. Hillsdale, NJ: Lawrence Erlbaum. Bahill, A. T., M. R. Clark, and L. Stark, 1975. The main sequence: A tool for studying human eye movements. Math. Biosci. 24:191–204. Balaban, M. T., 1996. Probing basic mechanisms of sensory, attentional, and emotional development: Modulation of the infant blink response. In C. Rovee-Collier and L. P. Lipsitt, eds., Advances in Infancy Research, vol. 10, pp. 219–256. Norwood, N J: Ablex. Bell, M. A., 1998. Frontal lobe function during infancy: Implications for the development of cognition and attention. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 287–316. Hillsdale, N J: Lawrence Erlbaum. Bell, M. A., 1999. The ontogeny of the EEG during infancy and childhood: Implications for cognitive development. In B. Garreay, ed., Neuroimaging in Childhood Psychiatric Disorders. Paris: Springer-Verlag. Bell, M. A., and N. A. Fox, 1992. The relations between frontal brain electrical activity and cognitive development during infancy. Child Dev. 63:1142–1163. Bell, M. A., and N. A. Fox, 1994. Brain development over the first year of life: Relations between electroencephalographic frequency and coherence and cognitive and affective behaviors. In G. Dawson and K. W. Fischer, eds., Human Behavior and the Developing Brain, 314–345. New York: Guilford. Bell, M. A., and C. D. Wolfe, 2007. The use of the electroencephalogram in research on cognitive development. In L. A. Schmidt and S. J. Segalowitz, eds., Developmental Psychophysiology, 150–172. Cambridge, UK: Cambridge University Press. Berg, W. K., and K. M. Berg, 1987. Psychophysiological development in infancy: State, startle and attention. In J. Osofsky, ed., Handbook of Infant Development, vol. 2, pp. 238–317. New York: Wiley. Berg, W. K., and J. E. Richards, 1997. Attention across time in infant development. In P. J. Lang, R. F. Simons, and M. T. Balaban, eds., Attention and Orienting: Sensory and Motivational Processes, 347–368. Mahwah, NJ: Erlbaum. Bronson, G. W., 1974. The postnatal growth of visual capacity. Child Dev. 45:873–890. Bronson, G. W., 1997. The growth of visual capacity: Evidence from infant scanning patterns. In C. Rovee-Collier and L. P. Lipsitt, Advances in Infancy Research, vol. 11, pp. 109–141. Greenwich, CT: Ablex. Courchesne, E., 1977. Event-related brain potentials: Comparison between children and adults. Science 197:589–592. Courchesne, E., 1978. Neurophysiological correlates of cognitive development: Changes in long-latency event-related potentials from childhood to adulthood. Electroencephalogr. Clin. Neurophysiol. 45:468–482. Courchesne, E., L. Ganz, and A. M. Norcia, 1981. Event-related brain potentials to human faces in infants. Child Dev. 52:804– 811. de Haan, M., 2008. Infant EEG and Event-Related Potentials. London: Psychology Press. de Haan, M., and C. A. Nelson, 1997. Recognition of the mother’s face by six-month-old infants: A neurobehavioral study. Child Dev. 68:187–210.

richards: attention in young infants

495

Desimone, R., and J. Duncan, 1995. Neural mechanisms of selective visual attention. Annu. Rev. Neurosci. 18:193–222. Fagan, J. F., 1974. Infant recognition memory: The effects of length of familiarization and type of discrimination task. Child Dev. 59:1198–1210. Frick, J., and J. E. Richards, 2001. Individual differences in recognition of briefly presented visual stimuli. Infancy 2:331– 352. Heilman, K. M., R. T. Watson, E. Valenstein, and M. E. Goldberg, 1987. Attention: Behavior and neural mechanisms. In V. B. Mountcastle, F. Plum, and S. R. Geiger, eds., Handbook of Physiology, 461–481. Bethesda, MD: American Physiological Society. Hicks, J. M., and J. E. Richards, 1998. The effects of stimulus movement and attention on peripheral stimulus localization by 8- to 26-week-old infants. Infant Behav. Dev. 21:571–589. Hillyard, S. A., G. R. Mangun, M. G. Woldroff, and S. J. Luck, 1995. Neural systems mediating selective attention. In M. S. Gazzaniga, ed., Cognitive Neurosciences, 665–682. Cambridge, MA: MIT Press. Hood, B. M., 1995. Shifts of visual attention in the human infant: A neuroscientific approach. Adv. Infancy Res. 10:163–216. Hood, B. M., J. Atkinson, and O. J. Braddick, 1998. Selection-for-action and the development of orienting and visual attention. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 219–250. Hillsdale, NJ: Lawrence Erlbaum. Hunter, S. K., 2001. Changes in saccadic eye movements: Attentional and developmental implications. Doctoral dissertion, University of South Carolina, Columbia, SC. Hunter, S. K., and J. E. Richards, 2003. Peripheral stimulus localization by 5- to 14-week-old infants during phases of attention. Infancy 4:1–25. Hunter, S. K., and J. E. Richards, submitted. Development of eye movements in young infants in response to complex dynamic stimuli. Johnson, M. H., 1990. Cortical maturation and the development of visual attention in early infancy. J. Cogn. Neurosci. 2:81–95. Johnson, M. H., 1995. The development of visual attention: A cognitive neuroscience perspective. In M. S. Gazzaniga, ed., The Cognitive Neurosciences, 735–747. Cambridge, MA: MIT Press. Johnson, M. H., 1997. Developmental Cognitive Neuroscience. London: Blackwell. Johnson, M. H., M. de Haan, A. Oliver, W. Smith, H. Hatzakis, L. A. Tucker, and G. Csibra, 2001. Recording and analyzing high-density event-related potentials with infants using the geodesic sensor net. Dev. Neuropsychol. 19:295–323. Johnson, M. H., R. O. Gilmore, and G. Csibra, 1998. Toward a computational model of the development of saccade planning. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 103–130. Hillsdale, NJ: Lawrence Erlbaum. Johnson, M. H., R. Griffin, G. Csibra, H. Halit, T. Farroni, M. de Haan, S. Baron-Cohen, and J. E. Richards, 2005. The emergence of the social brain network: Evidence from typical and atypical development. Dev. Psychopathol. 17:599– 619. Johnson, M. H., M. I. Posner, and M. K. Rothbart, 1991. Components of visual orienting in early infancy: Contingency learning, anticipatory looking and disengaging. J. Cogn. Neurosci. 3:335–344. Karrer, R., and P. K. Ackles, 1987. Visual event-related potentials of infants during a modified oddball procedure. In R. Johnson, J. W. Rohrbaugh, and R. Parasuramanm, eds., Current

496

cognition

Trends in Event-Related Potential Research, 603–608. Amsterdam: Elsevier Science. Karrer, R., and P. K. Ackles, 1988. Brain organization and perceptual/cognitive development in normal and Down syndrome infants: A research program. In P. Vietze and H. G. Vaughan, Jr., eds., The Early Identification of Infants with Developmental Disabilities, 210–234. Philadelphia: Grune and Stratton. Karrer, R., and L. A. Monti, 1995. Event-related potentials of 4–7-week-old infants in a visual recognition memory task. Electroencephalogr. Clin. Neurophysiol. 94:414–424. Lansink, J. M., and J. E. Richards, 1997. Heart rate and behavioral measures of attention in six-, nine-, and twelve-month-old infants during object exploration. Child Dev. 68:610–620. Maunsell, J. H. R., and V. B. Ferrera, 1995. Attentional mechanisms in visual cortex. In M. S. Gazzaniga, ed., The Cognitive Neurosciences, 451–461. Cambridge, MA: MIT Press. Maurer, D., and T. L. Lewis, 1979. A physiological explanation of infants’ early visual development. Can. J. Psychol. 33:232– 252. Maurer, D., and T. L. Lewis, 1991. The development of peripheral vision and its physiological underpinnings. In M. J. S. Weiss and P. R. Zelazo, eds., Newborn Attention: Biological Constraints and the Influence of Experience, 218–255. Norwood, NJ: Ablex. Maurer, D., and T. L. Lewis, 1998. Overt orienting toward peripheral stimuli: Normal development and underlying mechanisms. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 51–102. Hillsdale, NJ: Lawrence Erlbaum. Mesulam, M. M., 1983. The functional anatomy and hemispheric specialization for directed attention. Trends Neurosci. 6:384– 387. Michel, C. M., M. M. Murray, G. Lantz, S. Gonzalez, L. Spinelli, and R. Grave de Peraltz, 2004. EEG source imaging. Clin. Neurophysiol. 115:2195–2222. Nelson, C. A., 1994. Neural correlates of recognition memory in the first postnatal year. In G. Dawson and K. W. Fischer, eds., Human Behavior and the Developing Brain, 269–313. New York: Guilford Press. Nelson, C. A., and P. F. Collins, 1991. Event-related potential and looking-time analysis of infants’ responses to familiar and novel events: Implications for visual recognition memory. Dev. Psychol. 27:50–58. Nelson, C. A., and P. F. Collins, 1992. Neural and behavioral correlates of visual recognition memory in 4- and 8-month-old infants. Brain Cogn. 19:105–121. Nelson, C. A., and R. A. deRegnier, 1992. Neural correlates of attention and memory in the first year of life. Dev. Neuropsychol. 8:119–134. Nelson, C. A., and D. Dukette, 1998. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 327–362. Hillsdale, NJ: Lawrence Erlbaum. Nelson, C. A., and C. S. Monk, 2001. The use of event-related potentials in the study of cognitive development. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 125–135. Cambridge, MA: MIT Press. Nelson, C. A., and P. Salapatek, 1986. Electrophysiological correlates of infant recognition memory. Child Dev. 57:1483– 1497. Nunez, P. L., 1990. Localization of brain activity with electroencephalography. Adv. Neurol. 54:39–65. Posner, M. I., 1995. Attention in cognitive neuroscience: An overview. In M. S. Gazzaniga, ed., Cognitive Neurosciences, 615– 624. Cambridge, MA: MIT Press.

Posner, M. I., and S. E. Petersen, 1990. The attention system of the human brain. Annu. Rev. Neurosci. 13:25–42. Ray, W. J., 1990. Electrical activity of the brain. In J. T. Cacioppo and L. G. Tassinary, eds., Principles of Psychophysiology: Physical, Social, and Inferential Elements, 385–412. Cambridge, UK: Cambridge University Press. Reynolds, G. D., and J. E. Richards, 2005. Familiarization, attention, and recognition memory in infancy: An ERP and cortical source localization study. Dev. Psychol. 41:598– 615. Reynolds, G. D., and J. E. Richards, 2007. Infant heart rate: A developmental psychophysiological perspective. In L. A. Schmidt and S. J. Segalowitz, eds., Developmental Psychophysiology, 173–210. Cambridge, UK: Cambridge University Press. Richards, J. E., 1987. Infant visual sustained attention and respiratory sinus arrhythmia. Child Dev. 58:488–496. Richards, J. E., 1995. Infant cognitive psychophysiology: Normal development and implications for abnormal developmental outcomes. In T. H. Ollendick and R. J. Prinz, eds., Advances in Clinical Child Psychology, vol 17, pp. 77–107. New York: Plenum Press. Richards, J. E., 1997a. Peripheral stimulus localization by infants: Attention, age and individual differences in heart rate variability. J. Exp. Psychol. [Hum. Percept.] 23:667–680. Richards, J. E., 1997b. Effects of attention on infants’ preference for briefly exposed visual stimuli in the paired-comparison recognition-memory paradigm. Dev. Psychol. 33:22–31. Richards, J. E., 1998. Development of selective attention in young infants. Dev. Sci. 1:45–51. Richards, J. E., 2000a. Development of multimodal attention in young infants: Modification of the startle reflex by attention. Psychophysiology 37:1–11. Richards, J. E., 2000b. Cortical indices of saccade planning following covert orienting in 20-week-old infants. Infancy 2:135– 157. Richards, J. E., 2000c. Localizing the development of covert attention in infants using scalp event-related potentials. Dev. Psychol. 36:91–108. Richards, J. E., 2001. Attention in young infants: A developmental psychophysiological perspective. In C. A. Nelson and M. Luciana, eds., Developmental Cognitive Neuroscience, 321–338. Cambridge, MA: MIT Press. Richards, J. E., 2003a. Attention affects the recognition of briefly presented visual stimuli in infants: An ERP study. Dev. Sci. 6:312–328. Richards, J. E., 2003b. Cortical sources of event-related potentials in the prosaccade and antisaccade task. Psychophysiology 40:878–894. Richards, J. E., 2004a. The development of sustained attention in infants. In M. I. Posner, ed., Cognitive Neuroscience of Attention, chap. 25, pp. 342–356. New York: Guilford Press. Richards, J. E., 2004b. Development of covert orienting in young infants. In L. Itti, G. Rees, and J. Tsotsos, eds., Neurobiology of Attention, chap. 14, pp. 82–88 San Diego, CA: Academic Press/Elsevier. Richards, J. E., 2005. Localizing cortical sources of eventrelated potentials in infants’ covert orienting. Dev. Sci. 8:255– 278.

Richards, J. E., 2006. Realistic head models for cortical source analysis in infant participants. International Conference on Infant Studies, Kyoto, Japan, June. Richards, J. E., submitted. Cortical sources of ERP in the prosaccade and antisaccade task using realistic source models based on individual MRIs. Richards, J. E., and B. J. Casey, 1990. Infant visual recognition memory performance as a function of heart rate defined phases of attention. Infant Behav. Dev. 13:585. Richards, J. E., and B. J. Casey, 1991. Heart rate variability during attention phases in young infants. Psychophysiology 28: 43–53. Richards, J. E., and B. J. Casey, 1992. Development of sustained visual attention in the human infant. In B. A. Campbell, H. Hayne, and R. Richardson, eds., Attention and Information Processing in Infants and Adults, 30–60. Mahway, NJ: Erlbaum. Richards, J. E., and F. B. Holley, 1999. Infant attention and the development of smooth pursuit tracking. Dev. Psychol. 35:856–867. Richards, J. E., and S. K. Hunter, 1997. Peripheral stimulus localization by infants with eye and head movements during visual attention. Vis. Res. 37:3021–3035. Richards, J. E., and S. K. Hunter, 1998. Attention and eye movement in young infants: Neural control and development. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 131–162. Mahway, NJ: Erlbaum. Richards, J. E., and S. K. Hunter, 2002. Testing neural models of the development of infant visual attention. Dev. Psychobiol. 40:226–236. Robbins, T. W., and B. J. Everitt, 1995. Arousal systems and attention. In M. S. Gazzaniga, ed., Cognitive Neurosciences, 703– 720. Cambridge, MA: MIT Press. Rothbart, M. K., and M. I. Posner, 2001. Mechanism and variation in the development of attentional networks. In C. A. Nelson and M. Luciana, eds., Developmental Cognitive Neuroscience, 353–364. Cambridge, MA: MIT Press. Ruff, H. A., and M. K. Rothbart, 1996. Attention in Early Development. New York: Oxford University Press. Scherg, M., 1990. Fundamentals of dipole source potential analysis. In F. Grandori, M. Hoke, and G. L. Romani, eds., Auditory Evoked Magnetic Fields and Potentials, 40–69. Basel, Switzerland: Karger. Scherg, M., and T. W. Picton, 1991. Separation and identification of event-related potential components by brain electrical source analysis. In C. H. M. Brunia, G. Mulder, and M. N. Verbaten, eds., Event-Related Brain Research, 24–37. Amsterdam: Elsevier Science. Schiller, P. H., 1985. A model for the generation of visually guided saccadic eye movements. In D. Rose and V. G. Dobson, eds., Models of the Visual Cortex, 62–70. New York: John Wiley. Schiller, P. H., 1998. The neural control of visually guided eye movements. In J. E. Richards, ed., Cognitive Neuroscience of Attention: A Developmental Perspective, 3–50. Mahway, NJ: Erlbaum. Tucker, D. M., 1993. Spatial sampling of head electrical fields: The geodesic sensor net. Electroencephalogr. Clin. Neurophysiol. 87:154–163. Tucker, D. M., M. Liotti, G. F. Potts, G. S. Russell, and M. I. Posner, 1994. Spatiotemporal analysis of brain electrical fields. Hum. Brain Mapping 1:134–152.

richards: attention in young infants

497

30

Nonhuman Primate Models of Memory Development JOCELYNE BACHEVALIER

This chapter will focus on studies that have provided insights into the neurobiological basis of early memory development in nonhuman primates. Two areas of long-term memory processes have been more particularly investigated: one relates to the development of procedural memory, the other to the development of declarative memory. The distinction between procedural and declarative memory emerged from the study of patients with damage to the medial temporal region and from a growing number of animal models that have refined our understanding of the specific brain regions involved in amnesic syndromes and the critical memory processes mediated by each region. Thus the declarative (Squire, 1992) or propositional (Tulving, 1995) memory systems permit the acquisition of facts and specific events and are indexed by memory tests such as recognition and recall, which require explicit recollection. These memory systems are dependent on the medial temporal lobe/diencephalic structures and their interactions with cortical areas. In turn, the procedural memory systems mediate the acquisition and retention of skilled performance indexed by tasks in which memory is expressed implicitly by changes in performance as a result of prior experience. These procedural memory systems involve different brain circuits, such as the neostriatum for sensorimotor skills learning and the cerebellum for classical conditioning (for review see Eichenbaum, 2003). Critical questions that immediately emerged from this categorization of memory systems in the adults were, When do these memory systems emerge during development, and how do the brain systems that mediate them mature during ontogeny? The questions are important not only for providing information on the different types of memory systems and neural processes available at different time points during maturation, but also to inform us regarding how the different components of memory are assembled to finally result in adult memory functions (for a more complete discussion see Nelson, 1997). To answer these questions, neuropsychological studies in nonhuman primates have taken three complementary approaches. The first one identifies when adult proficiency on behavioral tasks known to be critical for a type of memory in the adult monkeys emerges in infancy. Given

the current knowledge of the neural substrate mediating each of these systems in adults, one could infer the specific neural substrate that becomes available to an infant at different time points during maturation. This is, however, an indirect approach to the development of the neurobiological bases of memory that can generally be substantiated by two other sets of studies. The first ones employ more direct neuroanatomical, neurobiological, and electrophysiological techniques to assess the maturation of specific neural structures in infancy and to correlate the time course of the neural maturation with the time course of emergence of the memory abilities mediated by a specific brain area. Finally, the last approach is to determine whether brain lesions that affect a specific memory system in adult monkeys will preclude the emergence of this system during ontogeny when the lesions are done in early infancy. The basic idea behind this last approach is that a behavioral deficit will occur only at a time when the area normally reaches a certain level of functional maturity (Goldman, 1971). There are also some limitations with this last approach that need to be considered when interpreting the data. The limitations relate to the profound plastic neural changes following early lesions that could, in turn, yield significant functional sparing, or on the contrary, deleterious effects on other neural systems. The neuropsychological studies of memory development in nonhuman primates are still scarce but extremely informative. Thus the main objective of this chapter is to review the most relevant findings, emphasize the numerous gaps in our knowledge, and suggest important areas for future investigations.

Development of procedural memory Procedural memory entails an array of diverse learning and memory processes indexed by tasks measuring the acquisition and retention of conditioned responses, skills, or habits, as well as perceptual responses (Squire and Knowlton, 1995). In monkeys, the development of procedural memory has almost exclusively been studied with visual discrimination tasks, requiring the participation of the neostriatum, although one study has used classically conditioned responses, requiring the participation of the cerebellum.

499

Emergence of Adult Procedural Memory Abilities Classical conditioning in infant monkeys was studied by Harlow (1959). During the first postnatal week, infant monkeys can rapidly learn a conditioned response when an auditory stimulus is paired with a brief and mild electric shock. At this early age, monkeys also learn to associate visual cues to the delivery of food, indicating that acquisition of conditioned responses is present almost immediately after birth. In the simplest version of the discrimination task, the animal is presented with two visual cues (or objects) and must select the sensory cue that is consistently associated with a food reward. These two visual cues are presented over several trials until the animal selects the rewarded cue on almost every trial. Nine- to 15-day-old monkeys rapidly master black/white and left/right discriminations (Harlow, 1959), and, between 15 and 25 days of age, they can discriminate between two patterns or forms (Harlow, 1959; Harlow et al., 1960; Zimmerman, 1961; Zimmerman and Torrey, 1965). Furthermore, the ability to solve more complex discrimination tasks in which the animal has to concurrently learn a series of object discriminations also emerges relatively early in life. Thus by 3 to 4 months of age (the earliest age tested) infant monkeys can learn as efficiently as adult monkeys a short list of object pairs with short intertrial intervals (Mahut and Moss, 1986) and even long lists of object pairs with extremely long intertrial intervals (Bachevalier and Mishkin, 1984). These data suggest that the ability to acquire and retain perceptualmotor associative responses, hence to form some types of procedural memory, is present in the first few postnatal months in monkeys. Consequently, some of the neural components of this memory system are clearly functional early in life to enable adult performance on procedural memory tasks. Maturation of Neural Circuits Mediating Procedural Memory The ability to acquire conditioned responses has been shown to involve the cerebellum and associated brainstem structures (for a review see Woodruff-Pak and Thompson, 1988). In addition, perceptual-motor associations (underlying discrimination tasks) require interactions between sensory and motor cortical areas, with the participation of the neostriatum and cerebellum (Gaffan, 1996; Mishkin, Malamut, and Bachevalier, 1984; Ungerleider, 1995; Wise, 1996). Both metabolic (Bachevalier, Hagger, and Mishkin, 1991) and electrophysiological (Rodman, 1994) studies have indicated that functional maturity of the ventral visual cortical areas, which provide visual inputs to the neostriatum and motor cortical areas, proceeds progressively from the occipital areas to the most anterior temporal areas and appears to be completed around 3–6 months postnatally in monkeys.

500

cognition

For the neostriatum, the cellular components are in place in early gestation (which lasts for 165 days in monkeys), and, from the 69th to the 130th day of gestation, cortical inputs begin to invade the neostriatum (Brand and Rakic, 1979; Goldman-Rakic, 1981). In the caudate nucleus and putamen, synaptic density continues to increase until the end of the first postnatal month (Brand and Rakic, 1984), and changes in neuronal and neuropil morphology are observed until 2–4 months postnatally (Cano, Pasik, and Pasik, 1989; Difiglia, Pasik, and Pasik, 1980). Refinement within the striatum proceeds until the end of the first postnatal year when the striatal neurochemical mosaic attains the adult pattern (Martin, Spicer, and Cork, 1992). Furthermore, histoanatomical studies indicate that motor cortical areas are predominantly connected with subcortical structures at birth and that, thereafter, intracortical connections progressively develop between the motor cortex and other cortical areas (Kemper, Caveness, and Yakovlev, 1973). In fact, the corticospinal projections in the macaque monkey mature gradually over a period of at least 11 months, long after simple measures of dexterity show functional maturity. This result suggests that the later changes may contribute to the improved speed and coordination of skilled motor tasks (Armand et al., 1994; Flament, Hall, and Lemon, 1992; Galea and Darian-Smith, 1995; Olivier et al., 1997). Finally, in the cerebellum, neurogenesis in the cerebellar cortex and deep nuclei occurs during the first two months of gestation, and neuron differentiation continues through midgestation (Kornguth, Anderson, and Scott, 1967). At birth, the morphology of the Purkinje cells as well as their axodendritic synapses are similar in structure to the adult pattern (Kornguth, Anderson, and Scott, 1967). Migration of granule cells continues during the first few postnatal months (Rakic, 1971). Synapses from the Purkinje cells are already present prenatally on the soma and axons of deep nuclear neurons, and synapses from climbing fibers of the inferior olive onto Purkinje cells and those from mossy fibers onto granule cells occur approximately at midgestation (Kornguth, Anderson, and Scott, 1968). Finally, at birth, all Purkinje cells have received synaptic contacts (Levitt et al., 1984). Several neurochemical components of the cerebellum are present at birth, whereas others continue to be modified during the postnatal period (Hayashi, 1987; Schatteman et al., 1988; Yamashita et al., 1990). Thus all neural components necessary to support classical conditioning and visual discriminations appear to be in place, and at almost complete maturity, relatively soon after birth. However, no studies have assessed the effects of early damage to either the cerebellum or the striatum on classical conditioning and discrimination learning in monkeys. The only information available at the present time concerns the effects of damage to the higher-order associative cortical areas in the ventral visual stream. Thus

damage to the inferior temporal cortical area TE in adult monkeys retards the acquisition of concurrent visual discriminations but does not totally abolish this learning ability. The same lesions in infant monkeys during the first month after birth result in a mild and transient retardation in learning concurrent discriminations at 3 months of age (Bachevalier et al., 1990) but no impairment when the same operated animals were retested as adults (Málková, Mishkin, and Bachevalier, 1995). The relative preservation of the ability to learn visual discriminations after early area TE damage suggests that this area is not yet fully functional in infancy and that this ability can be served effectively by visual tissue other than area TE. Indeed, we know that visual inputs to the ventral striatum do not originate uniquely from area TE but also from visual cortical areas located earlier in the ventral visual stream. Thus, in the absence of area TE, these sensory processing outputs from occipitotemporal visual areas to the ventral striatum could support visual discrimination abilities. In addition, since the early lesions yield greater sparing than the lesions in adulthood, these occipitotemporostriatal connections can make a greater contribution to discrimination learning in infants than in adults. To summarize, although much remains to be known about the maturation of the neural structures subserving procedural memory processes, all evidence obtained thus far indicates that the neural circuitry mediating this type of memory appears to reach maturity during the first few months postnatally in monkeys.

Development of declarative memory Declarative memory is thought to support the learning and retention of facts and the recollection of prior events and is indexed by memory tests, such as recall and recognition, requiring the explicit remembering of specific episodes. One important characteristic of this type of memory is that, as opposed to procedural memory, it is very often formed after a single exposure to a situation and is dependent on the integrity of the structures in the medial temporal lobe and medial diencephalon. A considerable emphasis has recently been placed on identifying the specific role played by different structures within the medial temporal region (medial temporal cortical areas and hippocampal formation) on declarative memory processes in monkeys. The findings suggest that each structure participates in a different way, although the details of each structure’s specific participation in declarative memory remain a subject of intense debate. The development of declarative memory in monkeys has mostly been studied using recognition-memory tasks and more recently with relational-memory tasks, which are tasks thought to tax declarative memory processes in animals (Alvarado and Bachevalier, in press).

Emergence of Adult Recognition-Memory Abilities Two paradigms in particular have been used to assess the development of recognition-memory abilities in monkeys: the visual paired comparison (VPC) task (also known as “preferential looking”) and the delayed nonmatching-tosample task (DNMS). Like some tests of human recognition or recall, each task assesses whether subjects can demonstrate that a given stimulus has been previously seen (judgment of prior occurrence; Brown, 1996; see also chapter 33 by Richmond and Nelson, this volume). However, substantial differences in the task requirements may alter both the demands made on memory processes and, potentially, the specific brain regions necessary for successful performance on each (Nemanic, Alvarado, and Bachevalier, 2004). The VPC task measures the distribution of time spent looking at recently presented stimuli as compared to novel ones. Recognition memory is inferred when subjects spent a longer time looking at the new stimuli than at the old ones (Fagan, 1970), a measure of recognition-based incidental memory, since no rules need to be learned by the subjects. Earlier studies (Bachevalier, Brickson, and Hagger, 1993; Gunderson and Sackett, 1984; Gunderson and Swartz, 1985, 1986) have demonstrated the presence of preference for novelty in the first postnatal months in monkeys even when long retention intervals are used. A more recent longitudinal study of recognition-based incidental-memory abilities assessed at 1, 6, and 18 months of age (Zeamer et al., 2006; Bachevalier and Vargha-Khadem, 2005) has shown that, although novelty preference is present as early as 1 month of age (average looking time at novel stimuli: 67% at the shortest delay of 10 s to 64% at the longest delay of 120 s), this ability becomes stronger with age (75% at the shortest delay to 73% at the longest delay at 18 months). To investigate the role of the hippocampal formation on the development of recognition-based incidental memory, the novelty preference of monkeys that had received selective neurotoxic hippocampal lesions between 10 and 12 days of age was also measured at 1, 6, and 18 months. Interestingly, novelty preference in these operated infant monkeys does not differ from that of control animals at the age of 1 and 6 months. However, at 18 months, novelty preference, though not totally abolished, was significantly weaker in the monkeys with neonatal hippocampal damage than in control animals, at delays of 60 seconds or longer. These data suggest that the progressive changes in incidental recognition memory abilities measured by preferential viewing reflect the gradual functional maturation of the hippocampus. They also suggest that this form of recognition memory seen at an early age in both control and operated monkeys could be mediated by the medial temporal cortical areas that have been shown to be critical for familiarity judgments and object memory in adults (Murray, 2000; Brown and

bachevalier: nonhuman primate models of memory development

501

Aggleton, 2001; Yonelinas, 2002; Nemanic, Alvarado, and Bachevalier, 2004). In the DNMS task, the animal must first learn the rule to choose on every trial a novel object paired with another object that had been seen just a few seconds earlier. After learning the DNMS rule with a single object and a short delay, the animal is given a memory test in which its recognition ability is assessed further with longer delays and also with longer lists of to-be-remembered objects. The results showed that, despite the availability of recognition-based memory abilities in the first month of life, it is not before 4 months of age that monkeys begin to learn the DNMS rule and not before 2 years of age that they can learn it with adult proficiency. In addition, as compared to adult monkeys, immature monkeys were deficient in performing the memory test with extended delays and lists (Bachevalier and Mishkin, 1984). Thus the difficulty young normal infant monkeys show in the DNMS task cannot be due to an absence of novelty detection or even to a deficiency in recognizing stimuli, which the DNMS task exploits. Instead the results suggest that the difficulty is due to a deficiency in rule learning, and more particularly difficulty associating the reward in the well with the novel and abstract quality of the object (Bachevalier, 1990; Bachevalier, Brickson, and Hagger, 1993; Diamond, 1990; Málková et al., 2000). Because performance on the DNMS task is not impaired by selective neonatal or late damage to the hippocampal formation (for review see Alvarado and Bachevalier, in press), the data indicate that the difficulty in DNMS rule learning is due to immaturity of a neural circuit other than the hippocampus, that is, a neural circuit involving the interactions between the inferior temporal and prefrontal cortical areas. This view emerged from recent findings in adult monkeys showing that the ability to perform the DNMS task requires interactions between the visual temporal cortical areas, that is, the perirhinal cortex (Alvarez, Zola-Morgan, and Squire, 1995; Eacott, Gaffan, and Murray, 1994; George, Horel, and Cirillo, 1989; Horel et al., 1987; Meunier et al., 1993; Murray, 1992; Suzuki et al., 1993; Zola-Morgan et al., 1989, 1993) and area TE (Buffalo et al., 1998; Buffalo et al., 1999; Mishkin, 1982; Mishkin and Phillips, 1990), and between the inferior prefrontal cortex, that is, the inferior convexity (Kowalska, Bachevalier, and Mishkin, 1991; Weinstein, Saunders, and Mishkin, 1988), and orbitofrontal cortex (Meunier, Bachevalier, and Mishkin, 1997). Similar to the adult lesions, neonatal perirhinal and orbital frontal cortical damage results in a striking impairment in DNMS (Málková et al., 1998; Pixley et al., 1997). By contrast, unlike the adult lesions, neonatal lesions of area TE and inferior prefrontal convexity led to substantial and permanent sparing of the ability to perform on the DNMS task (Bachevalier and Mishkin, 1994; Málková et al., 2000; Málková, Mishkin, and Bachevalier, 1995). This sparing of DNMS performance

502

cognition

after early lesions of area TE–inferior prefrontal convexity suggests instead that it is these cortical areas that are not fully functional at birth and are the limiting factors in the young animal’s ability to achieve adult proficiency on the DNMS task. Emergence of Adult Relational-Memory Abilities Relational memory appears to have a more protracted development than recognition memory. Several object and spatial relational tasks have been used to follow the development of this type of memory in monkeys. In the biconditional discrimination task (Saunders and Weiskrantz, 1989), the animal first learns to discriminate four object pairs (AB, AC, CD, and BD), in which only two of the pairs are rewarded (for example, AC and BD). Both 6-month-old and 1-year-old monkeys performed more poorly on biconditional discrimination learning than adults (Killiany, Rehbein, and Mahut, 2005). In the transverse patterning task (Spence, 1952), the animal concurrently learns three object pairs (A+ versus B−; B+ versus C−, C+ versus A−). Although the ability to solve the transverse patterning problem emerges around 1 year of age, complete adult proficiency in this task is not reached before 2 years (Málková et al., 1999). Similarly, for the oddity task (Harlow, 1959), the animal is presented with new sets of three objects on every trial. Each set comprises two identical objects and one that is different (AAB, CCD, EEF, etc.). The animal must select the odd object of the triad to unveil the reward. Adultlike performance on the oddity task is not attained before 3–4 years of age (Harlow, 1959). Finally, in spatial-memory tasks, the animal must use spatial relationships between cues to find a location in its environment. Using a foraging-memory task, Lavenex and Lavenex (2006) recently showed the presence of spatialrelational memory abilities in 9-month-old monkeys. Given that performance on both the VPC task (Clark, Zola, and Squire, 2000; Pascalis et al., 2000; Zola et al., 2000; Nemanic, Alvarado, and Bachevalier, 2004) and the object and spatial-relational memory tasks (Alvarado and Bachevalier, 2005; Bohbot et al., 1998; Reed and Squire, 1999; Kessels et al., 2001; Hampton, Hampstead, and Murray, 2004; Lavenex, Amaral, and Lavenex, 2006) is impaired in monkeys and humans with selective hippocampal damage, the complete set of data suggests that mnemonic abilities thought to be dependent on the hippocampus do not show a single pattern of development. Some processes (e.g., recognition-based incidental memory) are present in the first months of life, whereas others involving more complex cognitive demands (e.g., object and spatialrelational memory) mature over many years (for more details see Alvarado and Bachevalier, 2000; Bachevalier and Vargha-Khadem, 2005). The findings are consistent with the notion that context-free, incidental mnemonic abilities precede the development of context-rich, explicit memory

(Tulving, 1995), which only gradually emerges during childhood and becomes hippocampally dependent (Mishkin et al., 1997; Mishkin, Vargha-Khadem, and Gadian, 1998). Maturation of Neural Circuits Mediating Declarative Memory One striking difference between rodents and primates is that in the latter species neurogenesis in the hippocampus proper and dentate gyrus occurs almost entirely during prenatal life. However, many morphological and neurochemical changes as well as refinement of synaptic connections within the hippocampus persist until the first postnatal years. In the monkey (for review see Alvarado and Bachevalier, 2000; Seress and Ribak, 1995a, 1995b), genesis of neurons in the dentate gyrus continues throughout gestation and is approximately 80 percent complete at birth, but tapers off between the fourth and sixth postnatal months to a low level that may continue through adult life. The postnatal wave of synaptogenesis in the dentate gyrus peaks at 4 to 5 months of age and is accompanied by a 30 percent increase in spine density and in asymmetrical synapses in addition to a decrease in shaft synapses. In the CA fields, CA3 neurons increase in size and number, and the spines increase in complexity in the second half of the first postnatal year, and new mossy-fiber synapses are formed throughout the first year. Last, myelination of hippocampal afferents and efferents shows substantial postnatal maturation. This postnatal development of the hippocampus is also evidenced by an increase in hippocampal volume as well as changes in the ratio of gray to white matter from birth to 1 year of age as revealed by a recent longitudinal structural MRI study (Machado et al., 2002). Given the recent discovery of the role of the entorhinal and perirhinal cortex in recognition memory and object and spatial-relational memory (for review see Alvarado and Bachevalier, in press), there are surprisingly few studies exploring the anatomical developments of these cortical areas. Within the normal monkey gestational period (165 days), cells in the entorhinal cortex are generated beginning on embryonic day 36 (E36), preceding cell generation in the hippocampal formation by two days, and continuing to E70 (Rakic and Nowakowski, 1981). Cytoarchitectonic analysis revealed that the laminar subdivisions characteristic of the adult entorhinal cortex were identifiable by midgestation with the exception of the lateral entorhinal cortex (that area occupying the medial bank of the rhinal sulcus), which develops fully only in the last quarter of gestation. Aminergic innervation of this region is apparent in the second quarter of gestation (Berger, Alvarez, and Goldman-Rakic, 1993; Berger, de Grissac, and Alvarez, 1999; Berger and Alvarez, 1994, 1996) but increases in density at the end of gestation and shows mature characteristics in the newborn (Berger and Alvarez, 1994). The pattern of labeling suggests that at least some of these connections are extrinsic. For example,

neurotensin-reactive terminals (neurotensin, NT, is a pyramidal cell marker), but not NT-reactive neurons, are present in the entorhinal cortex at birth, suggesting the presence of extrinsic innervation at this early age. One possible source of extrinsic innervation originates in area CA1 of the hippocampus proper, which does show NT-reactive neurons at this age (Berger, Alvarez, and Goldman-Rakic, 1993). Neurogenesis in perirhinal cortical areas (35 and 36) has not been directly examined in the primate. Morphologically, however, development of this area lags behind that of the entorhinal cortex (Berger, Alvarez, and Goldman-Rakic, 1993; Berger and Alvarez, 1994). In fact, by the fourth gestational month, the rhinal sulcus is still only a small indent on the cortical surface (Berger and Alvarez, 1994). Yet, despite this apparent lag, NT-immunoreactive terminals are present in the caudal perirhinal cortex at this time, suggesting that the perirhinal cortex receives extrinsic innervation by the last gestational quarter. Though the source of this input has yet to be determined, CA1 pyramidal cells are possible candidates, as they have been shown to send direct projections to the perirhinal cortex in the adult monkey (Blatt and Rosene, 1998) and are NT immunoreactive at this age (Berger and Alvarez, 1994). At birth, however, the perirhinal cortex can be clearly identified cytoarchitecturally and displays adultlike chemoanatomical characteristics (Berger and Alvarez, 1994). Given all evidence gathered until now, it is likely that, although interactions between the hippocampal formation and subcortical and allocortical areas (perirhinal and entorhinal cortex) may be functional early in life to support some form of recognition memory, the interactions between the hippocampal formation and the neocortical areas on the inferior temporal cortex and prefrontal cortex may develop more progressively during the first year postnatally to account for the protracted appearance of adult proficiency on relational-memory tasks (Alvarado and Bachevalier, in press; see also Nelson, 1995, 1997). To summarize, as for procedural memory, the development of declarative memory appears to comprise several subsystems (see also Mishkin, VarghaKadem, and Gadian, 1998, for a similar conclusion), some of which emerge earlier than others during ontogenesis.

Relationship to human memory development As already reviewed in detail by others (Nelson, 1995, 1997), the development of procedural and declarative memory processes appears to show a comparable developmental time course in monkeys and humans (see table 30.1). Such similarities in the emergence of these multiple memory systems in the two species imply that the basic neural systems mediating each type of memory are likely to follow similar development sequences in both species. For example, the recent data on developmental amnesia in children (for review see

bachevalier: nonhuman primate models of memory development

503

Table 30.1 Functional development of multiple memory systems in monkeys and humans Memory Systems/Tasks Monkeys Humans Procedural memory Conditioned responses 1 week 1–2 daysa Visual discrimination tasks: Two pairs 2–3 weeks 1–1.5 yearsa,b Concurrent multiple pairs 3–4 months Declarative memory Recognition tasks: Visual paired comparison Delayed nonmatching Relational tasks: Spatial Biconditional discrimination Transverse patterning Oddity

2 weeks 2 years

3 daysa 4–5 yearsa,b

4–5 yearsb 9 months 2–3 years 4–5 yearsd 2–3 years 7 yearsb,c 3–4 years Note: For visual functions 1 week of development in monkeys corresponds roughly to 1 month of development in humans. a Nelson, 1995; bOverman, Pate, et al., 1996; cOverman, Bachevalier, et al., 1996; dRudy, Keith, and Georgen, 1993.

Bachevalier and Vargha-Khadem, 2005) are consistent with those reported on young nonhuman primates; namely, despite the early onset of hippocampal pathology, there is little evidence for sparing or recovery of those aspects of explicit memory that are critically dependent on the hippocampus. Thus both infant monkeys and children with early hippocampal damage grow into their memory impairments. Finally, the recent findings that neonatal hippocampal lesions in monkeys do not totally abolish recognition-based incidental learning also indicate that, in the absence of a functional hippocampus in early infancy, the medial temporal cortical areas could provide support for the sparing of some forms of recognition memory abilities still present in developmental amnesia.

Concluding remarks Despite the significant progress made in our understanding of the neural bases of memory development in monkeys, much remains to be learned. For example, the maturation of the neural structures mediating procedural and declarative memory systems has only been studied with anatomical techniques. Although informative, these procedures do not provide knowledge on the functional state of a structure very early in life. Therefore, further analyses of the functional maturation of the neostriatum, cerebellum, and hippocampal formation, as well as cortical areas using electrophysiological or metabolic studies, are clearly needed. In addition, nothing is known about the development of other types of procedural memory, such as emotional memory mediated by the amygdala and orbital frontal cortex.

504

cognition

Such developmental animal studies are critical, since they have important implications for our understanding of the development of memory processes in human infants. First the comparable developmental time course of these multiple memory systems in infant monkeys and human infants (see Overman and Bachevalier, 2002) implies that basic neural systems are likely to follow similar development sequences in both species. If so, then, additional studies in infant monkeys on the development of these multiple neural systems are likely to provide invaluable information on the maturation and neural bases of memory functions in humans. This review also highlights the paucity of data that currently exist on the neural structures and circuits that support the development of memory processes in infancy and childhood, and suggests important directions for future research. Specifically, longitudinal studies using techniques and paradigms applicable to both species (e.g., evoked response potentials, visual paired comparison) covering the period from infancy to later childhood will provide much-needed data on the functional maturation of memory processes. For example, the use of animal memory paradigms in amnesic patients and in neuroimaging studies of normal individuals has strengthened the validity of several animal models of memory and has provided critical knowledge on how different brain areas work together to solve a given task (for review see Alvarado and Bachevalier, in press). Similar cross-species studies are likely not only to guide our search for the neurobiological correlates underlying memory processes early in infancy, but also to shed light on the neurobiological correlates underlying adult memory processes. acknowledgments

Preparation of this chapter was supported in part by grants from the National Institutes of Health (MH58846 and HD35471), the Yerkes Base Grant NIH RR00165, and the CBN grant NSF IBN-9876754. REFERENCES

Alvarado, M. C., and J. Bachevalier, 2000. A protracted maturation of hippocampal memory functions in primates. Learn. Memory 7:244–256. Alvarado, M., and J. Bachevalier, 2005. Selective neurotoxic damage to the hippocampal formation impairs acquisition and performance of the transverse patterning task and location memory in rhesus macaques. Hippocampus 15:118–131. Alvarado, M. C., and J. Bachevalier, in press. Animal models of amnesia. In J. Byrne, ed., Learning and Memory: A Comprehensive Reference. New York: Elsevier. Alvarez, P., S. Zola-Morgan, and L. R. Squire, 1995. Damage limited to the hippocampal region produces long-lasting memory impairment in monkeys. J. Neurosci. 15:3796–3807. Armand, J., S. A. Edgey, R. N. Lemon, and E. Olivier, 1994. Protracted postnatal development of corticospinal projections from the primary motor cortex to hand motoneurones in the macaque monkey. Exp. Brain Res. 101:178–182.

Bachevalier, J., 1990. Ontogenetic development of habit and memory formation in primates. Ann. NY Acad. Sci. 608:457– 477. Bachevalier, J., M. Brickson, and C. Hagger, 1993. Limbicdependent recognition memory in monkeys develops early in infancy. NeuroReport 4:77–80. Bachevalier, J., M. Brickson, C. Hagger, and M. Mishkin, 1990. Age and sex differences in the effects of selective temporal lobe lesion on the formation of visual discrimination habits in rhesus monkeys. Behav. Neurosci. 104:885–899. Bachevalier, J., C. Hagger, and M. Mishkin, 1991. Functional maturation of the occipitotemporal pathway in infant rhesus monkeys. In N. A. Lassen, D. H. Ingvar, M. E. Raichle, and L. Friberg, eds., Alfred Benzon Symposium No. 31, Brain Work and Mental Activity: Quantitative Studies with Radioactive Tracers, 231– 240. Copenhagen: Munksgaard. Bachevalier, J., and M. Mishkin, 1984. An early and a late developing system for learning and retention in infant monkeys. Behav. Neurosci. 98:770–778. Bachevalier, J., and M. Mishkin, 1994. Effects of selective neonatal temporal lobe lesions on visual recognition memory in rhesus monkeys. J. Neurosci. 14:2128–2139. Bachevalier, J., and F. Vargha-Khadem, 2005. The primate hippocampus: Ontogeny, early insult and memory. Curr. Opin. Neurobiol. 15:168–174. Berger, B., and C. Alvarez, 1994. Neurochemical develop-ment of the hippocampal region in the fetal rhesus monkey. II. Immunocytochemistry of peptides, calcium-binding proteins, DARPP-32, and monoamine innervation in the entorhinal cortex by the end of gestation. Hippocampus 4:84–114. Berger, B., and C. Alvarez, 1996. Neurochemical development of the hippocampal region in the fetal rhesus monkey. III. Calbindin-D28K, calretinin and parvalbumin with special mention of Cajal-Retzius cells and the retrosplenial cortex. J. Comp. Neurol. 366:674–699. Berger, B., C. Alvarez, and P. S. Goldman-Rakic, 1993. Neurochemical development of the hippocampal region in the fetal rhesus monkey. I. Early appearance of peptides, calcium-binding proteins, DARPP-32 and the N-monoamine innervation in the entorhinal cortex during the first half of gestation (E47 to 90). Hippocampus 3:279–305. Berger, B., N. de Grissac, and C. Alvarez, 1999. Precocious development of parvalbumin-like immunoreactive interneurons in the hippocampal formation and entorhinal cortex of the fetal cynomolgus monkey. J. Comp. Neurol. 403:309–331. Blatt, G. J., and D. L. Rosene, 1998. Organization of direct hippocampal efferent projections to the cerebral cortex of the rhesus monkey: Projections from CA1, prosubiculum, and subiculum to the temporal lobe. J. Comp. Neurol. 392:92– 114. Bohbot, V. D., M. Kalina, K. Stepankova, N. Spackova, M. Petrides, and L. Nadel, 1998. Spatial memory deficits in patients with lesions to the right hippocampus and to the right parahippocampal cortex. Neuropsychologia 36:1217– 1238. Brand, S., and P. Rakic, 1979. Genesis of the primate neostriatum: [3H]thymidine autoradiographic analysis of the time of neuron origin in the rhesus monkey. Neuroscience 4:767–778. Brand, S., and P. Rakic, 1984. Cytodifferentiation and synaptogenesis in the neostriatum of fetal and neonatal rhesus monkeys. Anat. Embryol. (Berl.) 169:21–34. Brown, M. W., 1996. Neuronal responses and recognition memory. Sem. Neurosci. 8:23–32.

Brown, M. W., and J. P. Aggleton, 2001. Recognition memory: What are the roles of the perirhinal cortex and hippocampus? Nature Rev. Neurosci. 2:51–61. Buffalo, E. A., S. J. Ramus, R. E. Clark, E. Teng, L. R. Squire, and S. M. Zola, 1999. Dissociation between the effects of damage to perirhinal cortex and area TE. Learn. Memory 6: 572–599. Buffalo, E. A., L. Stefanacci, L. R. Squire, and S. Zola, 1998. A reexamination of the concurrent discrimination learning task: The importance of anterior inferotemporal cortex, area TE. Behav. Neurosci. 112:3–14. Cano, J., P. Pasik, and T. Pasik, 1989. Early postnatal development of the monkey globus pallidus: A golgi and electron microscopic study. J. Comp. Neurol. 279:353–367. Clark, R. E., S. M. Zola, and L. R. Squire, 2000. Impaired recognition memory in rats after damage to the hippocampus. J. Neurosci. 20:8853–8860. Diamond, A., 1990. Rate of maturation of the hippocampus and the developmental progression of children’s performance on the delayed nonmatching-to-sample and visual paired comparison tasks. Ann. NY Acad. Sci. 608:394–426. Difiglia, M., P. Pasik, and T. Pasik, 1980. Early postnatal development of the monkey neostriatum: A Golgi and ultrastructural study. J. Comp. Neurol. 190:303–331. Eacott, M. J., D. Gaffan, and E. A., Murray, 1994. Preserved recognition memory for small sets, and impaired stimulus identification for large sets, following rhinal cortex ablations in monkeys. Eur. J. Neurosci. 6:1466–1478. Eichenbaum, H. B., 2003. Learning and memory: Brain systems. In L. R. Squire, F. E. Bloom, S. K. McConnell, J. L. Roberts, N. C. Spitzer, and M. J. Zigmond, eds., Fundamental Neuroscience, 1299–1328. New York: Academic Press. Fagan, J. F., 1970. Memory in the infant. J. Exp. Child Psychol. 9:217–226. Flament, D., E. J. Hall, and R. N. Lemon, 1992. The development of cortico-motoneuronal projections investigated using magnetic brain stimulation in the infant macaque. J. Physiol. (Lond.) 447:755–768. Gaffan, D., 1996. Memory, action, and the corpus striatum: Current developments in the memory-habit distinction. Sem. Neurosci. 8:33–38. Galea, M. P., and I. Darian-Smith, 1995. Postnatal maturation of the direct corticospinal projections in the macaque monkey. Cerebral Cortex 5:518–540. George, P. J., J. A. Horel, and R. A. Cirillo, 1989. Reversible cold lesions of the parahippocampal gyrus in monkeys result in deficits of the delayed matching-to-sample and other visual tasks. Behav. Brain Res. 34:163–178. Goldman, P. S., 1971. Functional development of the prefrontal cortex in early life and the problem of neuronal plasticity. Exp. Neurol. 32:366–387. Goldman-Rakic, P., 1981. Prenatal formation of cortical input and development of cytoarchitectonic compartments in the neostriatum of the rhesus monkey. J. Neurosci. 1:721–735. Gunderson, V. M., and G. P. Sackett, 1984. Development of pattern recognition in infant pigtailed macaques (Macaca nemestrina). Dev. Psychol. 20:418–426. Gunderson, V. M., and K. B. Swartz, 1985. Visual recognition in infant pigtailed macaques after 24-hour delay. Am. J. Primatol. 8:259–264. Gunderson, V. M., and K. B. Swartz, 1986. Effects of familiarization time on visual recognition memory in infant pigtailed macaques (Macaca nemestrina). Dev. Psychol. 22:477–480.

bachevalier: nonhuman primate models of memory development

505

Hampton, R. R., B. M. Hampstead, and E. A. Murray, 2004. Selective hippocampal damage in rhesus monkeys impairs spatial memory in an open-field test. Hippocampus 14:808– 818. Harlow, H. F., 1959. The development of learning in the rhesus monkey. Amer. Sci. Winter: 459–479. Harlow, H. F., M. K. Harlow, R. R. Rueping, and W. A. Mason, 1960. Performance of infant rhesus monkeys on discrimination learning, delayed response, and discrimination learning set. J. Comp. Physiol. Psychol. 53:113–121. Hayashi, M., 1987. Ontogeny of glutamic acid decarboxylase, tyrosine hydroxylase, choline acetyltransferase, somatostatin and substance P in monkey cerebellum. Dev. Brain Res. 32:181– 186. Horel, J. A., D. E. Pytko-Joyner, M. L. Voytko, and K. Salsbury, 1987. The perfomance of visual tasks while segments of the inferotemporal cortex are suppressed by cold. Behav. Brain Res. 23:29–42. Kemper, T. L., W. F. Caveness, and P. I. Yakovlev, 1973. The neuronographic and metric study of the dendritic arbours of neurons in the motor cortex of Macaca mulatta at birth and at 24 months of age. Brain 96:765–782. Kessels, R. P., E. H. de Haan, L. J. Kappelle, and A. Postma, 2001. Varieties of human spatial memory: A meta-analysis on the effects of hippocampal lesions. Brian Res. Brain Res. Rev. 35:295–303. Killiany, R., L. Rehbein, and H. Mahut, 2005. Developmental study of the hippocampal formation in rhesus monkeys (Macaca mulatta). II. Early ablations do not spare the capacity to retrieve conditional object-object associations. Behav. Neurosci. 119:651– 661. Kornguth, S. E., J. W. Anderson, and G. Scott, 1967. Observations on the ultrastructure of the developing cerebellum of the Macaca mulatta. J. Comp. Neurol. 130:1–24. Kornguth, S. E., J. W. Anderson, and G. Scott, 1968. The development of synaptic contacts in the cerebellum of Macaca mulatta. J. Comp. Neurol. 132:531–546. Kowalska, D., J. Bachevalier, and M. Mishkin, 1991. The role of the inferior prefrontal convexity in performance of delayed nonmatching-to-sample. Neuropsychologia 29:583– 600. Lavenex, P. B., D. G. Amaral, and P. Lavenex, 2006. Hippocampal lesion prevents spatial relational learning in adult macaque monkeys. J. Neurosci. 26:4546–4558. Lavenex, P., and P. B. Lavenex, 2006. Spatial relational memory in 9-month-old macaque monkeys. Learn. Memory 13: 84–96. Levitt, P., P. Rakic, P. De Camilli, and P. Greengard, 1984. Emergence of cyclic guanosine 3′ : 5′-monophosphatedependent protein kinase immunoreactivity in developing rhesus monkey cerebellum: Corrrelative immunocytochemical and electron microscopic analysis. J. Neurosci. 4:2553–2564. Machado, C. J., S. L. Babin, E. F. Jackson, and J. Bachevalier, 2002. Use of MRI techniques to investigate the maturation of the nonhuman primate temporal lobes. Soc. Neurosci. Abstracts 28, online. Mahut, H., and M. Moss, 1986. The monkey and the seahorse. In R. L. Isaacson and K. H. Pribram, eds., The Hippocampus, 241– 279. New York: Plenum Press. Málková, L., M. C. Alvarado, E. L. Pixley, A. M. Belcher, M. Mishkin, and J. Bachevalier, 1999. Maturation of relational memory processes in monkeys. Soc. Neurosci. Abstracts 25:88.

506

cognition

Málková, L., J. Bachevalier, M. Webster, and M. Mishkin, 2000. Effects of neonatal inferior prefrontal and medial temporal lesions on learning the rule for delayed nonmatching-to-sample. Dev. Neuropsychol. 18:399–421. Málková, L., M. Mishkin, and J. Bachevalier, 1995. Longterm effects of selective neonatal temporal lobe lesions on learning and memory in monkeys. Behav. Neurosci. 109:212– 226. Málková, L., G. I. Pixley, M. J. Webster, M. Mishkin, and J. Bachevalier, 1998. The effects of early rhinal lesions on visual recognition memory in rhesus monkeys. Soc. Neurosci. Abstracts 24:1906. Martin, L. J., D. M. Spicer, and L. C. Cork, 1992. Postnatal development of the striatal mosaic in rhesus monkey. Soc. Neurosci. Abstracts 18:957. Meunier, M., J. Bachevalier, and M. Mishkin, 1997. Effects of orbital frontal and anterior cingulate lesions on object and spatial memory in rhesus monkeys. Neuropsychologia 35:999– 1015. Meunier, M., J. Bachevalier, M. Mishkin, and E. A. Murray, 1993. Effects on visual recognition of combined and separate ablations of the entorhinal and perirhinal cortex in rhesus monkeys. J. Neurosci. 13:5418–5432. Mishkin, M., 1982. A memory system in the monkey. Philos. Trans. R. Soc. London B298:85–95. Mishkin, M., B. Malamut, and J. Bachevalier, 1984. Memories and habits: Two neural systems. In G. Lynch, J. L. McGaugh, and N. M. Weinberger, eds., Neurobiology of Learning and Memory, 287–296. New York: Guilford Press. Mishkin, M., and R. R., Phillips, 1990. A corticolimbic path revealed through its disconnection. In C. Trevarthen, ed., Brain Circuits and Functions of the Mind: Essays in Honor of Roger W. Sperry, 196–210. Cambridge, UK: Cambridge University Press. Mishkin, M., W. A. Suzuki, D. G. Gadian, and F. VarghaKhadem, 1997. Hierarchical organization of cognitive memory. Philos. Trans. R. Soc. London B352:1461–1467. Mishkin, M., F. Vargha-Khadem, and D. G. Gadian, 1998. Amnesia and the organization of the hippocampal system. Hippocampus 8:212–216. Murray, E. A., 1992. Medial temporal lobe structures contributing to recognition memory: The amygdaloid complex versus the rhinal cortex. In J. P. Aggleton, ed., The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction, 453–470. New York: Wiley. Murray, E. A., 2000. Memory for objects in nonhuman primates. In M. S. Gazzaniga, ed., The New Cognitive Neurosciences, 753–764. Cambridge, MA: MIT Press. Nelson, C. A., 1995. The ontogeny of human memory: A cognitive neuroscience perspective. Dev. Psychol. 31:723–738. Nelson, C. A., 1997. The neurobiological basis of early memory development. In N. Cowan, ed., The Development of Memory in Childhood, 41–82. Hove, UK: Psychology Press. Nemanic, S., M. C. Alvarado, and J. Bachevalier, 2004. The hippocampal/parahippocampal regions and recognition memory: Insights from visual paired comparison versus object delayed nonmatching in monkeys. J. Neurosci. 24:2013–2026. Olivier, E., S. A. Edgley, J. Armand, and R. N. Lemon, 1997. An electrophysiological study of the postnatal development of the corticospinal system in the macaque monkey. J. Neurosci. 17:267–276. Overman, W. H., and J. Bachevalier, 2002. Inferences about the functional development of neural systems in children via the application of animal tests of cognition. In C. A. Nelson

and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 109–124. Cambridge, MA: MIT Press. Overman, W., J. Bachevalier, M. Miller, and K. Moore, 1996. Children’s performance on “animal tests” of oddity: Implications for cognitive processes required for tests of oddity and delayed non-match to sample. J. Exp. Child Psychol. 62:223– 242. Overman, W. H., B. J. Pate, K. Moore, and A. Peuster, 1996. Ontogeny of place learning in children as measured in the radial arm maze, Morris search task, and open field task. Behav. Neurosci. 110:1205–1228. Pascalis, O., N. M. Hunkin, J. S. Holdstock, C. L. Isaac, and A. R. Mayes, 2000. Differential performance on recognition memory tests and the visual paired comparison task in a patient with selective hippocampal lesion. Soc. Neurosci. Abstracts 26:1241. Pixley, G. L., L. Malkova, M. J. Webster, M. Mishkin, and J. Bachevalier, 1997. Early damage to both inferior convexity and orbital prefrontal cortices impairs DNMS learning in infant monkeys. Soc. Neurosci. Abstracts 23:499. Rakic, P., 1971. Neuron-glia relationship during granule cell migration in developing cerebellar cortex: A Golgi and electromicroscopic study in macaque rhesus. J. Comp. Neurol. 141:283– 312. Rakic, P., and R. S. Nowakoski, 1981. The time of origin of neurons in the hippocampal region of the rhesus monkey. J. Comp. Neurol. 196:99–128. Reed, J. M., and L. R. Squire, 1999. Impaired transverse patterning in human amnesia is a special case of impaired memory for two-choice discrimination tasks. Behav. Neurosci. 113:3–9. Rodman, H. R., 1994. Development of inferior temporal cortex in the monkey. Cerebral Cortex 5:484–498. Rudy, J. W., J. Keith, and K. Georgian, 1993. The effect of age on children’s learning of problems that require a configural association solution. Dev. Psychobiol. 26:171–184. Saunders, R. C., and L. Weiskrantz, 1989. The effects of fornix transection and combined fornix transection, mammillary body lesions and hippocampal ablation on object-pair association memory in the rhesus monkey. Behav. Brain Res. 35:85– 94. Schatteman, G. C., L. Gibbs, A. A. Lanahan, P. Claude, and M. Bothwell, 1988. Expression of NGF receptor in the developing and adult primate central nervous system. J. Neurosci. 8:860–873. Seress, L., and C. E. Ribak, 1995a. Postnatal development of CA3 pyramidal neurons and their afferents in the Ammon’s horn of rhesus monkeys. Hippocampus 5:217–231. Seress, L., and C. E. Ribak, 1995b. Postnatal development and synaptic connections of hilar mossy cells in the hippocampal dentate gyrus of rhesus monkeys. J. Comp. Neurol. 355:93– 110. Spence, K. W., 1952. The nature of response in discrimination learning. Psychol. Rev. 59:89–93.

Squire, L. R., 1992. Declarative and nondeclarative memory: Multiple brain systems supporting learning and memory. J. Cogn. Neurosci. 4:232–243. Squire, L. R., and B. J. Knowlton, 1995. Memory, hippocampus, and brain systems. In M. Gazzaniga, ed., The Cognitive Neurosciences, 825–837. Cambridge, MA: MIT Press. Suzuki, W. A., S. Zola-Morgan, L. R. Squire, and D. G. Amaral, 1993. Lesions of the perirhinal and parahippocampal cortices in the monkey produce long lasting memory impairments in the visual and tactual modalities. J. Neurosci. 13:2430– 2451. Tulving, E., 1995. Organization of memory: Quo vadis? In M. Gazzaniga, ed., The Cognitive Neurosciences, 839–847. Cambridge, MA: MIT Press. Ungerleider, L. G., 1995. Functional brain imaging studies of cortical mechanisms for memory. Science 270:769–775. Weinstein, J., R. C. Saunders, and M. Mishkin, 1988. Temporoprefrontal interaction in rule learning by macaques. Soc. Neurosci. Abstracts 14:1230. Wise, S. P., 1996. The role of the basal ganglia in procedural memory. Sem. Neurosci. 8:39–46. Woodruff-Pak, D. S., and R. F. Thompson, 1988. Cerebellar correlates of classical conditioning across the life span. In P. B. Baltes, R. M. Lerner, and D. M. Featherman, eds., Life-Span Development and Behavior, 1–37. Hillsdale, NJ: Erlbaum. Yamashita, A., M. Hayashi, K. Shimizu, and K. Oshima, 1990. Neuropeptide-immunoreactive cells and fibers in the developing primate cerebellum. Dev. Brain Res. 51:19–25. Yonelinas, A. P., 2002. The nature of recollection and familiarity: A review of 30 years of research. J. Mem. Lang. 46:441–517. Zeamer, A. E., M. Resende, E. Heuer, and J. Bachevalier, 2006. The infant monkeys’ recognition memory abilities in the absence of a functional hippocampus. Soc. Neurosci. Abstracts online. Zimmerman, R. R., 1961. Analysis of discrimination learning capacities in the infant rhesus monkey. J. Comp. Physiol. Psychol. 54:1–10. Zimmerman, R. R., and C. C. Torrey, 1965. Ontogeny of learning. In A. M. Schrier, H. F. Harlow, and F. Stollnitz, eds., Behavior of Nonhuman Primates: Modern Research Trends, vol. 2, pp. 405–447. New York: Academic Press. Zola, S. M., L. R. Squire, E. Teng, L. Stefanacci, E. A. Buffalo, and S. K. Clark, 2000. Impaired recognition memory in monkeys after damage limited to the hippocampal region. J. Neurosci. 20:451–463. Zola-Morgan, S., L. R. Squire, D. G. Amaral, and W. A. Suzuki, 1989. Lesions of the perirhinal and parahippocampal cortex that spare the amygdala and hippocampal formation produce severe memory impairment. J. Neurosci. 9:4355–4370. Zola-Morgan, S., L. R. Squire, R. P. Clower, and N. L. Rempel, 1993. Damage to the perirhinal cortex exacerbates memory impairment following lesions of the hippocampal formation. J. Neurosci. 13:251–265.

bachevalier: nonhuman primate models of memory development

507

31

Neurocognitive Mechanisms for the Development of Face Processing MICHELLE DE HAAN

The human face holds special significance as a visual source of social information. Adults can detect faces twice as fast as other objects (Pegna et al., 2004) and can quickly and accurately perceive the many signals they simultaneously display such as a person’s identity, emotional state, and direction of eye gaze. Even newborns are attracted to faces: shortly after birth, they will move their eyes further to keep a moving pattern in sight if its elements are positioned in facelike arrangements (Johnson et al., 1991). These impressive abilities make it easy to forget both the complex mechanisms underlying adults’ face-processing abilities and the lengthy developmental pathway by which the abilities of newborns are transformed into the mature state. This chapter provides a description of developmental changes in face processing together with an overview of neurocognitive mechanisms that may underlie these changes. It is divided into three sections, with the first describing the brain systems and perceptual-cognitive processes underlying face processing in adults, the second providing an overview of development with an emphasis on infancy and on studies that have examined brain correlates, and the third considering atypical development of face processing through the example of prosopagnosia originating in childhood.

How do adults process faces? Adult face processing is mediated by a distributed neural network involving subcortical and cortical areas (see figure 31.1 and plate 52). Visual information about faces is initially passed along two neural pathways: (1) a subcortical system that is involved in detecting faces and directing visual attention to them and (2) a core cortical system that is involved in the detailed visual-perceptual analysis of faces. Both of these components interact with (3) an extended cortical-subcortical system involved in further processing of faces (see figure 31.1; Gobbini and Haxby, 2007; Haxby, Hoffman, and Gobbini, 2000; Johnson, 2005). These three systems will be described in the following subsections.

Subcortical System In the subcortical pathway for face processing, information travels from the retina directly to the superior colliculus, then to the pulvinar and on to the amygdala (de Gelder et al., 2003; Johnson, 2005; see figure 31.1). This route is characterized by rapid, automatic processing, relies primarily on low-spatial-frequency information, and functions to detect faces and to direct visual attention to them (possibly with an enhanced response to faces signaling threat or danger; reviewed in Johnson, 2005). This rapid pathway could allow some degree of face processing to be accomplished before slower, conscious cortical processing is completed, and thereby could mediate nonconscious responses to faces and/or modulation of subsequent cortical processing. While there is compelling evidence in favor of the existence of this pathway (reviewed in Johnson, 2005; Palermo and Rhodes, 2007), the data with respect to its temporal precedence relative to cortical processing are less conclusive. For example, whereas neurophysiological studies are often optimal for testing hypotheses about the timing of brain activity, in this instance they have limitations: scalp recordings may not be particularly effective in detecting subcortical activity, and more direct measurements by intracranial recording in humans are based on medicated, diseased brains. Core Cortical System The core system for visual analysis of faces receives input from the retina by way of the geniculostriate pathway and includes the inferior occipital gyrus (encompassing the lateral occipital area, of which the “occipital face area” is a subregion), fusiform gyrus (including the “fusiform face area”), and posterior superior temporal sulcus/gyrus (see figure 31.1). The inferior occipital gyrus mediates the early perception of faces and passes this information to two areas: (1) the fusiform gyrus, which processes invariant aspects of faces, and (2) the superior temporal sulcus/gyrus, which processes changeable aspects of faces. Greater activation to faces than objects has been observed in all these regions (e.g., Yovel and Kanwisher, 2005), though for the inferior occipital gyrus this faceselective response likely reflects feedback from the fusiform

509

Figure 31.1 The face perception system. The three rectangles with beveled edges show the core system for face perception. Solid lines indicate cortical pathways, and dashed lines show the subcortical route. Areas in yellow represent regions involved in processing identity and associated semantic information. Areas in red represent regions involved in emotion analysis, and areas in blue indicate

those involved in spatial attention as it interacts with the face processing system. (Reprinted form R. Palermo and G. Rhodes, 2007. Are you always on my mind? A review of how face perception and attention interact. Neuropschologia 45:75–92, with permission from Elsevier.) (See plate 52.)

gyrus rather than a primary characteristic of the region itself (Rossion et al., 2003). While the core system exists bilaterally in the brain, there is substantial evidence that there is greater involvement on the right side (e.g., Grossman et al., 2000; Haxby et al., 1999; Wheaton et al., 2004).

is genetically predetermined to become face-specific cortex and argue for the necessity of visual learning. In one of these views, the FFA responds strongly to any stimulus category for which we have acquired perceptual expertise in discriminating category members. The stronger response to faces than to other categories is a reflection of the fact that faces are the only category for which all humans typically acquire such a high level of expertise (Tarr and Gauthier, 2000). The second view emphasises the importance of the timing of visual experience with faces in establishing the FFA, arguing that early operation of the subcortical route ensures that young infants frequently look at faces, thereby providing the necessary input for the developing FFA to become face selective (Johnson, 2005). Evidence showing that individuals deprived of visual input during infancy continue to show impairments in face processing even many years after their vision has been restored supports the view that these early inputs are necessary for the normal development of face processing (Le Grand et al., 2001, 2003). Yet another view

Fusiform face area. A large amount of the research on the core system has focused on the fusiform gyrus, in particular a region called the fusiform face area (FFA) that shows substantially stronger functional magnetic resonance imaging (fMRI) responses to faces than other visual stimuli (Kanwisher, McDermott, and Chun, 1997). The FFA has been the focus of debates as to whether there are cortical areas exclusively devoted to processing faces. According to one view, the FFA responds more strongly to faces because it is a domain-specific cortical module (Kanwisher, McDermott, and Chun, 1997) and genetically specified for the neural computations involved in processing faces (Farah et al., 2000). Other views take issue with the notion that the FFA

510

cognition

takes issue with the idea that the FFA is a homogeneous region specialized for the processing of faces. In this view, the FFA instead is part of a distributed neural system for visual object processing that spans the ventral occipitotemporal cortex (Haxby et al., 2001; Ishai et al., 1999). In support of this view, a recent high-resolution MRI study showed that the FFA consists of face-selective voxels intermixed with voxels highly selective for other objects (Grill-Spector, Sayers, and Ress, 2006) rather than functioning as a homogeneous region of face-selective voxels. The FFA is thought to be involved primarily in encoding invariant aspects of faces. Various types of invariant information can be processed, with one important distinction being between (1) processing of isolated features (e.g., eye color) and (2) processing of the relations among features. Relational processing itself can be further divided into three types: (a) first-order relations (the basic position of the facial features with eyes above nose, both above mouth), (b) holistic processing (the binding together of facial features into a gestalt that makes them hard to process independently), and (c) secondorder relations (subtle variations in the spatial characteristics of the basic facial arrangement, e.g., spacing between eyes; see Maurer, Le Grand, and Mondloch, 2002, for further discussion of these different types of processing). Featural processing is thought to be used for objects as well as faces, whereas relational processing, particularly holistic and second-order relational processing, is believed to be used for faces but not typically for other objects. The inversion effect is often used as an index of this reliance on relational processing: rotation by 180 degrees disrupts relational processing more than featural processing (reviewed in Rossion and Gauthier, 2002) and thus disrupts face processing more than object processing (Yin, 1969; Yovel and Kanwisher, 2004). The behavioral face inversion effect is strongly associated with activity in the FFA but not other components of the core system or other object-responsive regions, confirming that the FFA plays an important role in relational encoding (Yovel and Kanwisher, 2005; see also Schiltz and Rossion, 2006). Superior temporal sulcus/gyrus. The posterior region of the superior temporal sulcus (STS) and surrounding areas are activated when adults view moving faces, as well as other types of body movement. This activation does not simply reflect generic detection of motion, because nonbiological movement (e.g., movement in radial patterns) does not activate the region (Grossman et al., 2000; Pelphrey et al., 2005; Puce et al., 1998) and because activation is also seen in still images with only implied biological motion (Kourtzi and Kanwisher, 2000). There is some evidence that the exact area of activation is related to which body part is moving. For example, mouth movements elicit activity along the midposterior STS, eye movements elicit activity in more superior and posterior portions of the right posterior STS region, and

hand or body movements activate still different regions (Peelen, Wiggett, and Downing, 2006; Pelphrey et al., 2005). The STS’s responsiveness to biological motion makes it particularly important for processing changeable aspects of faces such as direction of eye gaze and emotional expressions. For example, fMRI studies show that the STS is activated when participants must monitor direction of eye gaze (Hooker et al., 2003) and fMRI-adaptation experiments show that activity in the STS, but not the fusiform gyrus, decreases when participants view repeated presentations of the same emotional expressions (Winston et al., 2004). The superior temporal cortex appears necessary for processing directional information from eye gaze, as this ability is lost when its function is disrupted with transcranial magnetic stimulation (Pourtois et al., 2004). However, in the case of emotion, the superior temporal cortex may not play a necessary role, as disrupting its functioning does not affect matching of fearful and happy expressions (Pourtois et al., 2004). Extended Cortical-Subcortical System The extended system receives input from, and in return communicates with, both the subcortical system and core cortical system. It encompasses a variety of regions involved in the further processing of these inputs (see figure 31.1) to allow activities such as conscious emotional appraisal, interpretation of the intentions of others, retrieval of semantic information and other memories about individuals, and direction of spatial attention. Because relatively less is known about the development of the extended system and its communication with the other systems, the next section will focus mainly on the subcortical and core cortical systems.

Development Subcortical System: Face Detection The subcortical system for detecting and orienting to faces appears to function from very early in life: newborn babies will move their eyes, and sometimes their heads, longer to keep a moving facelike pattern in view than several other comparison patterns (Goren, Sarty, and Wu, 1975; Johnson et al., 1991; Maurer and Young, 1983; Valenza et al., 1996). It seems that all that is needed to elicit a response is a very schematic version of the face: a triangular arrangement of three blobs for eyes and a mouth will do (Johnson and Morton, 1991), as will a stimulus that, like a face, has more elements in the upper than the lower half (Cassia, Turati, and Simion, 2004; Turati et al., 2002). There remains a debate as to the best interpretation of these results (for further discussion see de Haan, Humphreys, and Johnson, 2002). Some authors have argued that they are best explained by a social hypothesis. In this view, infants have an innate preference for facelike stimuli that is based on a specific knowledge of the configuration of the face. For

de haan: neurocognitive mechanisms for face processing

511

example, in the Conspec/Conlern hypothesis (Johnson and Morton, 1991), an innate, subcortical mechanism, Conspec, causes newborns to orient specifically to patterns with elements arranged in a facelike pattern in preference to other patterns (Johnson and Morton, 1991). In contrast, other authors account for the same data with a sensory hypothesis. A common theme in the various formulations of this hypothesis is that there is no aspect of the system that is responding specifically to faces. Instead, the preferential orienting of the newborn to faces is just a consequence of more general mechanisms guiding visual attention. For example, newborns prefer a “top-heavy” nonfacelike stimulus with more elements in the upper part over a nonfacelike stimulus with more elements in the lower part, and they do not show a preference for a facelike stimulus over a nonfacelike configuration equated for the number of elements in the upper part of the configuration (Turati et al., 2002). Thus infants’ apparent preference for facelike patterns, which naturally tend to have more elements in the upper half, may simply be a result of this more general orienting tendency (Turati et al., 2002). Whichever explanation for newborns’ preferential orienting to faces is correct, there is consensus that the infant is born with perceptual biases that could serve to guide subsequent experience with faces. Such experience, together with other perceptual and social cues present in real life, likely makes faces a frequent and salient stimulus in the world of the newborn. There is indirect evidence to support the view that infants’ preferential orienting to faces is mediated by a subcortical pathway (reviewed in Johnson, 2005). One line of evidence comes from a change in infants’ orienting to faces between 1 and 3 months of age. Over this time, infants’ tendencies to preferentially orient to faces appearing in their peripheral vision declines and begins to be replaced with a preference for fixating faces in central vision (Johnson and Morton, 1991). Since the orienting response declines over the same period as do other newborn reflexes assumed to be under subcortical control, it may also be mediated subcortically. Another line of evidence comes from studies of visual hemifield differences in infants’ orienting to faces. Because the nasal visual field provides relatively more input to the cortical than subcortical pathway, whereas the temporal visual field does the opposite, it is possible to compare cortical and subcortical processing if infants are tested monocularly and stimuli are presented in only one visual field. The results of such studies show that infants’ preferential orienting to faces occurs only when the stimuli are presented in the temporal visual field, a finding consistent with the idea that this early orienting to faces is mediated subcortically (Simion et al., 1998). Core Cortical System: Fusiform The core cortical system mediates the detailed perceptual analysis of faces

512

cognition

seen in the visual environment. According to the Conspec/ Conlern hypothesis, this system begins to functionally emerge between 1 and 3 months of age and supports the infant’s ability to formulate a mental representation of facedness based on the faces actually seen in the visual environment. This idea is supported by the results of a study in which 1and 3-month-old infants were familiarized to four individual faces, following which their abilities to recognize both the previously unseen morphed average of the four familiar faces and one of the four unaltered familiar faces were tested (de Haan et al., 2001). Babies at both ages were able to recognize the individual faces seen during familiarization, but only the 3-month-olds showed evidence of recognizing, and thus having mentally computed, the average of the four familiar faces. These results suggest that babies’ abilities to form a mental category or “prototype” of the face based on the actual faces they experience emerge sometime between 1 and 3 months of age. Consistent with this finding, studies have shown that 3-month-olds (Bar-Haim et al., 2006; Kelly et al., 2005), but not newborns (Kelly et al., 2005), prefer looking at faces of the same race as they experience in their visual environment rather than at faces of other races. A small number of studies using PET or event-related potentials support the idea that occipitotemporal cortical pathways are involved in infant face processing by 2–3 months of age. In the only infant PET study on this topic, 2-month-olds showed greater activation in the inferior occipital gyrus and the fusiform gyrus to a human face than to a set of three diodes (Tzourio-Mazoyer et al., 2002). These results demonstrate that at least some of the components of the core system are functioning by 2 months of age, although they do not indicate whether these areas are specifically activated by faces or more generally by other visual stimuli. The lack of activation of the superior temporal region might indicate a relative immaturity of this area; however, it is also possible that this occurred because the static, neutral faces were not optimal for activating this region. Event-related potential (ERP) studies support the idea that cortical mechanisms are involved in face processing from at least 3 months of age. However, they also suggest that, when cortical mechanisms do become involved, they are less “tuned in” to faces than in the mature system. These studies have focused on the development of the N170, a negative deflection over occipitotemporal electrodes that peaks approximately 170 milliseconds (ms) after stimulus onset that is thought to reflect the initial stages of the perceptual analysis of faces (Bentin et al., 1996). Components of the core system contribute to the N170, including regions of the fusiform gyrus (Shibata et al., 2002), the posterior inferior temporal gyrus (Bentin et al., 1996; Shibata et al., 2002), the lateral occipitotemporal cortex (Bentin et al., 1996; Schweinberger et al., 2002), and the superior temporal sulcus (Henson et al., 2003; Itier and Taylor, 2004b). The

N170 is typically of larger amplitude and/or longer latency for inverted than upright faces (Bentin et al., 1996; Cassia et al., 2006; de Haan, Pascalis, and Johnson, 2002; Eimer, 2000; Itier and Taylor, 2002, 2004a; Rossion et al., 2000) but does not differ for inverted compared to upright exemplars of nonface object categories (Bentin et al., 1996; Rebai et al., 2001; Rossion et al., 2000), even animal (monkey) faces that share the basic eyes-nose-mouth arrangement with the human face (de Haan, Pascalis, and Johnson, 2002). This N170 inversion effect parallels the behavioral inversion effect described earlier and suggests the N170 is involved in relational encoding. In infants as young as 3 months, two components believed to be precursors to the adult N170 are elicited during viewing of faces, the N290 and the P400 (Cassia et al., 2006; Halit, de Haan, and Johnson, 2003; Halit et al., 2004). Source analyses show that the generators of the infant N290 include regions similar to those identified for the adult N170, including the lateral occipital area bilaterally and the fusiform gyrus and superior temporal sulcus particularly on the right (Johnson et al., 2005). Though these precursors are observable by 3 months, they differ from the N170 in adults in that (1) an inversion effect specific to human faces is not seen until 12 months of age (even though there is evidence of discrimination between upright and inverted faces by at least 3 months of age), (2) response latencies are approximately 100–200 ms slower than in adults, (3) even at 12 months of age, responses are spread over a longer time range compared to adults, and (4) the spatial distribution of both the N290 and P400 is more medial and shifts laterally (thereby becoming more adultlike) between 3 and 12 months of age. Together, these results suggest that components of the core cortical system are functioning by 3 months but become more specific to human faces over the first year of life. They also suggest that processing of faces occurs more quickly and more discretely in time with age, and that the relative involvement or position of neural generators that contribute to these components changes with age (for discussion see de Haan, Johnson, and Halit, 2003). The infant N170 does not appear to be sensitive to familiarity in facial identity (de Haan and Nelson, 1997, 1999). This finding is similar to that observed in adults, suggesting that, as in adults, the N170 reflects core system processing of perceptual features that convey identity rather than recognition of familiar personal identity. Less is known about development of the extended system related to recognition of identity in infants, though studies have shown that more anterior and longer-latency components differentiate familiar from unfamiliar faces by 3 to 6 months of age (de Haan and Nelson, 1997, 1999; Pascalis et al., 1998). Interestingly, at 6 months this recognition effect is lateralized to the right for faces but occurs bilaterally for objects (de Haan and Nelson, 1997, 1999).

The N170 continues to develop until well into adolescence. For example, its latency decreases consistently until approximately 14 years of age, and its amplitude shows a U-shaped function with smaller amplitudes at 11–12 years than older or younger ages (Taylor, Batty, and Itier, 2004). Larger amplitudes over the right hemisphere characteristic of adults are not consistently seen until 11–12 years (Taylor, Batty, and Itier, 2004). A small number of fMRI studies also suggest continued development of the core cortical system through childhood. Children from 10 years show greater activation of the fusiform gyrus while viewing faces than viewing houses (Aylward et al., 2005), natural or manufactured objects (Gathers et al., 2004), scrambled faces (Passarotti et al., 2003), or a fixation point (Lobaugh, Gibson, and Taylor, 2006) and greater activation to direct than averted faces (Garrett et al., 2004). One of these studies included an adult comparison group and found that children showed a more distributed pattern of fusiform activation encompassing the medial and lateral regions while adults showed a more focused pattern of activity (Passarotti et al., 2003). Two studies of children younger than 10 years failed to find greater activation of the fusiform gyrus for faces compared to other stimuli in the classic FFA region, but both found evidence of such activation more posteriorly in the inferior occipital region (Aylward et al., 2005; Gathers et al., 2004). These latter results appear to conflict with those mentioned previously, where 2-month-olds showed activation of both the fusiform gyrus and inferior occipital region while viewing faces. One explanation for this apparent discrepancy is the differing comparison stimuli: the developing fusiform may be more active to faces than diodes, as in the study with 2-month-olds (Tzourio-Mazoyer et al., 2002), but may not yet be sufficiently specialized to be more active for faces than houses or other objects, as found with the children younger than 10 years (Aylward et al., 2005; Gathers et al., 2004). These neuroimaging studies suggest that important developments in face processing occur at about 10 years of age, an idea that is consistent with prior conclusions from behavioral studies. According to the classic “encoding switch” hypothesis, younger children are poorer at encoding and remembering faces because they do so only in terms of featural information. At about 10 years of age, children switch encoding styles and begin to use relational information, resulting in improved face processing (Carey and Diamond, 1977; Diamond and Carey, 1977). More recent studies have demonstrated that young children (Freire and Lee, 2001) and even infants (Cohen and Cashon, 2001) are able to encode relational information and discount the strong version of the hypothesis wherein there is a complete switch. However, there is still evidence that reliance on relational information does increase with age (Mondloch, LeGrand, and Maurer, 2002; but see McKone and Boyer, 2006). In

de haan: neurocognitive mechanisms for face processing

513

the neuroimaging studies, activation in the fusiform region correlates with the size of the behavioral inversion effect for faces in both 12- to 14-year-old children (Aylward et al., 2005) and adults (Yovel and Kanwisher, 2005), possibly suggesting that improvements in relational encoding are linked to development of the fusiform region’s role in face encoding. Core Cortical System: Superior Temporal Regions Emotion. Infants show early sensitivity to emotional information in the face, with a few studies showing that newborns are able to discriminate, and even imitate, facial expressions of emotion (Field et al., 1983, 1982). Certainly, by the end of the first half year of life, infants are able to discriminate at least some of the features of the face that to adults denote different expressions (reviewed in de Haan and Groen, 2006). The results are most extensive and most consistent with respect to the ability to discriminate happy from other expressions: With only one exception (Schwartz, Izard, and Ansul, 1985), the results of several studies are in agreement that during the first few months of life, infants are able to discriminate happy from surprised, angry, and fearful expressions (Barrera and Maurer, 1981; Field et al., 1982, 1983; Kotsoni, de Haan, and Johnson, 2001; LaBarbera et al., 1976; Nelson, Morse, and Leavitt, 1979; Nelson and Dolgin, 1985; Soken and Pick, 1992; Walker, 1982). However, in several studies, infants had difficulty discriminating happy from sad expressions (Young-Browne, Rosenfeld, and Horowitz, 1977; Oster, 1981) even by 7 months of age (Soken and Pick, 1992; but see Caron, Caron, and MacLean, 1988; Walker, 1982). Interestingly, one study has demonstrated that infants as young as 3.5 months of age can discriminate between happy and sad expressions, but only when posed by the mother and not when posed by a stranger (Kahana-Kalman and Walker-Andrews, 2001), suggesting that a familiar facial context might facilitate infants’ abilities to discriminate expressions. Discrimination between other expressions has been less extensively studied in infants, and, in particular, discrimination of disgust from other expressions has not been investigated. Very few studies have examined sensitivity to dynamic, changeable features during emotion processing in infants or children. A small number of studies demonstrate that infants in the first year of life prefer looking at dynamic rather than static faces (Courage, Reynolds, and Richards, 2006), can discriminate among dynamic facial expressions including happy, interested, sad, and angry (although sometimes only if the face is speaking and not when silent; Caron, Caron, and Maclean, 1988; Soken and Pick, 1992), can discriminate facial identity in moving, speaking faces (Spencer et al., 2006), and show better memory for actions involving faces (e.g., brushing teeth) than for the identity of the faces

514

cognition

themselves (Bahrick, Gogate, and Ruiz, 2002). These results confirm that infants are sensitive to moving faces and can tell apart expressions and identity under these conditions, but it remains unclear whether the developmental trajectory based on studies using static faces would be altered if more naturalistic, moving stimuli were employed. There is limited direct information about the involvement of the core cortical system in infants’ or children’s processing of facial expressions. Three ERP studies have examined 7-month-old infants’ responses to fearful compared to happy or angry expressions (de Haan et al., 2004; de Haan and Nelson, 1998; Nelson and de Haan, 1996). These studies demonstrated that ERPs differed for happy compared to fearful faces by approximately 140–260 ms after stimulus onset, but ERPs did not differ from angry compared to fearful faces at any latency up to 1,700 ms after stimulus onset. These results demonstrate that 7-month-olds can rapidly distinguish positive from negative expressions. However, none of these studies examined the infant N290 or P400 responses believed to reflect activity of the core system to determine whether there was any evidence of modulation by emotional expression. Studies of children between 4 and 14 years of age suggest that emotion does not greatly influence the N170 (de Haan et al., 1998; Batty and Taylor, 2006), though one study found a faster latency for fearful than happy or angry expressions in 5-year-olds (de Haan et al., 1998), and a larger amplitude to fearful compared to other expressions is observed by 14 to 15 years (Batty and Taylor, 2006). Most neuroimaging studies of emotion processing in children have focused on the amygdala, but some also show that facial expressions of emotion activate the fusiform gyrus (Wang et al., 2004) and superior temporal gyrus (Lobaugh, Gibson, and Taylor, 2006). However, these findings were based on relatively small groups spanning a wide age range and without adult comparison. Eye gaze. The direction of another person’s eye gaze is an important social cue, as it can provide important information both about that person’s intentions and about significant events in the environment. In fact, it has been argued that one factor contributing to the development of infants’ abilities to process direction of gaze is their learning that monitoring their caregiver’s direction of gaze allows them to predict the locations of interesting objects or events in their environment (Moore and Corkum, 1994). Shifts in eye gaze are such a powerful cue that they can influence the viewer’s attention in an automatic, reflexive manner (Driver et al., 1999; but see Vecera and Rizzo, 2006). Infants are sensitive to direction of eye gaze from the first days of life, preferring to look at faces with direct rather than averted gaze (Farroni et al., 2002). Viewing faces with direct gaze also influences how infants process and react to them.

For example, young infants can only effectively use a gaze shift to guide their own looking if it is preceded by at least a brief period of mutual eye contact (Farroni et al., 2003). In addition, 4-month-olds (Farroni et al., 2007), like adults (Mason, Hood, and Macrae, 2004), are better at recognizing facial identity when they have studied faces with direct gaze than when they have studied faces with averted gaze (see Blass and Camp, 2001, for a similar type of result). Fourmonth-olds also show enhanced processing of information about objects cued by an adult’s gaze relative to objects that were not cued (Reid et al., 2004). The latter effect is similar to ERP studies of adults showing that a reflexive shift of attention following the observation of a dynamic or static eye-gaze cue enhances and speeds up early visual processing of a target presented at the gazed-at location (Schuller and Rossion, 2004). While infants’ abilities to process direction of eye gaze are impressive, development of this skill continues for some years, as 6- and 8-year-olds are still worse than adults at matching faces according to eye-gaze direction, but 10-year-olds are adultlike (Mondloch, Geldart, Maurer, and LeGrand, 2003). Four-month-old infants show a larger infant N170 to faces with direct than to faces with averted gaze (Farroni et al., 2002). Four-month-olds show this response even if the head is turned, but not if the face is inverted (Farroni, Johnson, and Csibra, 2004). It is not clear whether the abolition of the response when the face is inverted is, as for adults, due primarily to inversion of the eyes themselves or due to the inversion of the entire face. Interestingly, the N290 response to direction of eye gaze appears to be generated by the fusiform region rather than the superior temporal region, although both regions are involved at longer latencies (Johnson et al., 2005). Studies of older children have produced mixed results, with one study of 3.5- to 7-year-olds showing, as for adults, no influence of direction of gaze on the N170 (Grice et al., 2005) and another in 12-year-olds showing a larger N170 for direct than for averted gaze (Senju et al., 2005). One study comparing children’s N170 responses to eyes only with responses to full faces found that the response to eyes was much larger and quicker than to full faces (Taylor et al., 2001). The study also reported that the N170 to eyes matures more quickly, by 11 years, than the N170 to full faces, which continues to develop until later in adolescence. In addition, the frontal positivity that in adults is the counterpoint to the posterior negative N170 is present by 11 years only for eyes but not yet for full faces. These results suggest developmental changes in the configuration of generators involved in processing of eyes compared to the full face. A study of event-related fMRI activity in 7- to 10-year-old healthy children confirms that the STS is activated by gaze shifts, and also indicates that the STS, middle temporal gyrus, and inferior parietal lobule are sensitive to the intentions underlying the stimulus character’s eye move-

ments. These findings suggest that the neural circuitry underlying the processing of eye gaze and the detection of intentions conveyed through shifts in eye gaze in children are similar to those found previously in adults, although the study did not make a direct comparison between children and adults (Mosconi et al., 2005). Brief summary: Core cortical system. Neurophysiological and neuroimaging studies show that the components of the core cortical system are involved in face processing from infancy. However, these same studies suggest that there are developmental changes in the system into adolescence. In particular, there is evidence that face sensitivity of the system increases with age and that the system may become more focal and less distributed with age, although further, more direct assessment of these possible developmental changes is warranted, as not all studies have directly compared children of different ages or children with adults. Much work to date has focused on the neural correlates of perception of facedness or facial identity, with fewer studies examining the correlates of perception of emotion or direction of eye gaze.

Developmental impairments in face processing In adults, lesions to the occipitotemporal cortices that are bilateral or limited to the right side (DeRenzi et al.,1994) can result in prosopagnosia, an impairment in the ability to recognize familiar faces. This impairment is selective, in that perception or memory for other types of objects must either be absent or much milder, and general cognitive abilities must be intact. The location of the lesions can vary from posteriorly in regions of the core cortical system (fusiform/ lingual gyri) to anterior temporal regions of the extended system (Damasio, Tranel, and Damasio, 1990). Corecortical-system lesions tend to be associated with a perceptual deficit, and extended-system lesions in anterior temporal regions with a deficit in linking an intact percept to its related memories (Damasio, Tvanel, and Damasio, 1990). Studying developmental impairments in face processing can provide valuable insight into the development of the brain systems involved in face processing. For example, if the core cortical system is domain specific and anatomically localized from early in life, then damage to it should result in persistent, major impairments of face processing. Perinatal Brain Injury Several studies have examined children with known perinatal lesions to document whether any cases of prosopagnosia occur. One set of studies with 5- to 17-year-old children who had sustained perinatal unilateral lesions found impairment in sorting of facial identity and/or emotion relative to comparison children (de Schonen et al., 2005; Mancini et al., 1994). However, (1) face-processing deficits were no more common than

de haan: neurocognitive mechanisms for face processing

515

object-processing deficits following a right-hemisphere lesion, (2) face-processing deficits were not more common after right- than left-sided damage, and (3) a face-processing deficit occurred in the absence of an object-processing deficit in only 1 of 11 patients. The main consistent effect of lesion site was that recognition deficits were more likely to occur following posterior or temporal damage than following more anterior (frontal) damage. The importance of medial temporal regions was also highlighted in another study examining emotion recognition in individuals with earlyonset mesial temporal lobe epilepsy (Meletti et al., 2003). This work demonstrated that deficits in emotion recognition, but not identity matching, were apparent in cases of right mesial temporal lobe epilepsy but not those with left or extratemporal seizure foci. Moreover, within the right group earlier age of onset of epilepsy was associated with poorer emotion recognition. Together, the results of these studies suggest that damage to temporal or posterior cortex early in life does result in persistent impairments in face processing. There is no strong evidence to suggest that perceptual tasks (matching, sorting) are more affected following right-sided than left-sided damage, but there is some evidence that correct labeling of emotional faces is more affected after right-sided than left-sided damage. Acquired Prosopagnosia with Childhood Onset Cases of prosopagnosia that emerge in childhood and can be attributed to a documented brain lesion are very rare. There are a few reported cases of individuals with brain injury acquired early in life who appear to have selective deficits in face processing (Bentin, Deouell, and Soroker, 1999; Farah et al., 2000). These facts would suggest that specific deficits following early damage, though rare, can occur and that there thus may be some degree of early brain specialization and localization of face processing early in life. However, in such cases the specificity of the deficit is often not clearly established, for example, by including careful tests of discrimination and recognition of similar objects and by measuring reaction times. One other caution in interpreting these cases is that it can sometimes be difficult to determine exactly when the damage relevant to the prosopagnosia occurred, and age of injury might be an important factor for determining whether specific deficits emerge. Congenital Prosopagnosia Congenital prosopagnosia has been defined as an impairment in face processing that is apparent from birth in the absence of any brain damage and that occurs in the presence of intact sensory and intellectual functions (Behrmann and Avidan, 2005). A necessary criterion is that face processing was never normal in the lifetime of the individual, although this criterion is difficult to test formally. Recently, increasing numbers of congenital prosopagnosics have been identified, with one

516

cognition

study estimating a prevalance of 2.47 percent (Kennerknecht et al., 2006). It is generally thought that this apparent increase in cases is due to better awareness of the condition and opportunities (e.g., Web sites) for such individuals to come forward, rather than a true increase. As in adults with acquired prosopagnosia, there is a large degree of variation among individual cases with regard to the degree and nature of impairments (Le Grand et al., 2006). An interesting finding in these cases, also observed in some adult acquired cases, is that apparently normal activation of the FFA can be observed (reviewed in Kleinschmidt and Cohen, 2006). What causes congenital prosopagnosia? One hypothesis is that it occurs because of a deficit in the subcortical system for orienting to faces (de Gelder and Stecklenberg, 2005). The defective subcortical system fails to provide the appropriate input to the developing core system, leading to a dysfunction of the core system without any obvious cortical lesion. Another, not mutually exclusive, explanation is that there are genetic factors involved that affect some aspect of the development of brain face-processing systems. A recent study of seven family pedigrees found evidence for a simple autosomal dominant mode of inheritance, suggesting that loss of human face recognition can occur by the mutation of a single gene (Grueter et al., 2007). If such a gene exists, however, it need not necessarily encode directly for face processing. For example, as described in the section “How do adults process faces?” deprivation of visual input early in life in cases of congenital cataracts causes later selective impairment in face processing (Le Grand et al., 2001). As a parallel, a genetic factor that affects a more general aspect of visual-processing ability may in the same way have a secondary effect on face processing.

Conclusions: How does face processing develop? Infants arrive in the world ready to learn about faces, and there is evidence that some of the basic components of the subcortical system and core cortical system are functioning within the first months of life. However, this period is just the beginning of a long journey, as the ability to process facial information and the underlying neural mechanisms continue to develop into adolescence. The existing data suggest that the subcortical system is operational from the first days of life. However, it is less clear whether and how this system continues to exert an influence on infants’ and children’s face processing as cortical systems become increasingly involved. In addition, the existing studies of newborns’ orienting to faces have focused on comparing facelike to nonfacelike patterns and have not investigated whether the system is especially responsive to emotional faces. For example, studies in adults suggest that one way the subcortical system may interact with the core system is by direct feedback connections from the amygdala to the

fusiform gyrus that act to enhance processing of emotional faces (reviewed in Vuilleumier and Pourtois, 2007). The available studies suggest that the core cortical system is functioning from early in life, but they also suggest some differences in the developing compared to the mature system. Results from ERP and fMRI studies suggest that in infants and young children faces activate posterior occipitotemporal areas (inferior occipital area) more strongly than anterior regions (classic FFA and STS) and that these anterior regions may become increasingly involved with age (see also Kylliainen et al., 2006). At least in infancy, the STS may become functionally involved more slowly than the fusiform, as evidenced by ERP and PET studies. However, further investigation is needed to determine whether this is a true difference or whether stimuli employed in these studies have been more optimal for activating the fusiform than the STS. When the classic FFA activation is first seen at about 10 years of age, it is initially more widespread and becomes more focal and category specific with age. Some of these developmental changes, particularly the increasingly focal and category-specific response observed in the FFA, have been likened to the changes that occur in adults during perceptual learning. Certainly there is an increasing body of evidence consistent with the view that visual experience with faces is an important part of the developmental process, both with respect to processing of facial identity (e.g., Bar-Haim et al., 2006; Kelly et al., 2007; Pascalis, de Haan, and Nelson, 2002) and facial emotion (e.g., de Haan et al., 2004; Fries and Pollak, 2004; Parker, Nelson, and the Bucharest Early Intervention Project Core Group, 2005; Pollak and Sinha, 2002). However, the conclusion that the development of the ability to process faces can be equated with the acquisition of perceptual expertise in adults is not yet warranted (e.g., see Robbins and McKone, 2006, and reply by Gauthier and Bukach, 2007; also further discussion in de Haan, Humphreys, and Johnson, 2002). Future studies would benefit from examining in more focus aspects of the extended system as well as interactions among all the systems. For example, primate studies of the development of visual cortex suggest that forward projections develop much more quickly than feedback connections (Barone, Dehay, Berland, and Kennedy, 1998; Burkhalter, 1993). A better understanding of constraints such as these might advance inquiries into how processes such as the increased localization and specialization of face-responsive cortex occur in development. These types of studies are increasingly possible with advances in neuroimaging techniques. Another important direction for future research is to provide better methods for assessing the types of visual experience that occur in children’s lives. To understand how experience will shape the development of face processing and other visual abilities, it will be helpful to be able to measure these inputs in a more precise way.

REFERENCES Aylward, E. H., J. E. Park, K. M. Field, A. C. Parsons, T. L. Richards, S. C. Cramer, and A. N. Meltzoff, 2005. Brain activation during face perception: Evidence of a developmental change. J. Cogn. Neurosci. 17:308–319. Bahrick, L. E., L. J. Gogate, and I. Ruiz, 2002. Attention and memory for faces and actions in infancy: The salience of actions over faces in dynamic events. Child Dev. 73:1629–1643. Bar-Haim, Y., T. Ziv, D. Lamy, and R. M. Hodes, 2006. Nature and nurture in own-race face processing. Psychol. Sci. 17:159–163. Barone, P., C. Dehay, M. Berland, J. Bullier, and H. Kennedy, 1998. Role of directed growth and target selection in the formation of cortical pathways: Prenatal development of the project of area V2 to area V4 in the monkey. J. Comp. Neurol. 374:1–20. Barrera, M. E., and D. Maurer, 1981. The perception of facial expressions by the three-month-old infant. Child Dev. 52: 203–206. Batty, M., and M. J. Taylor, 2006. The development of emotional face processing during childhood. Dev. Sci. 9:207–220. Behrmann, M., and G. Avidan, 2005. Congenital prosopagnosia: Face blind from birth. Trends Cogn. Sci. 9:180–187. Bentin, S., T. Allison, A. Puce, E. Perez, and G. McCarthy, 1996. Electrophysiological studies of face perception in humans. J. Cogn. Neurosci. 8:551–565. Bentin, S., L. Y. Deouell, and N. Soroker, 1999. Selective visual streaming in face recognition: Evidence from developmental prosopagnosia. NeuroReport 10:823–827. Blass, E. M., and C. A. Camp, 2001. The ontogeny of face recognition: Eye contact and sweet taste induce face preference in 9- and 12-week-old human infants. Dev. Psychol. 37:762–774. Burkhalter, A., 1993. Development of feedforward and feedback connections between areas V1 and V2 of human visual cortex. Cerebral Cortex 3:478–487. Carey, S., and R. Diamond, 1977. From piecemeal to configurational representation of faces. Science 195:312–314. Caron, A. J., R. F. Caron, and D. J. MacLean, 1988. Infant discrimination of naturalistic emotional expressions: The role of face and voice. Child Dev. 59:604–616. Cassia, V. M., D. Kuefner, A. Westerlund, and C. A. Nelson, 2006. Modulation of face-sensitive event-related potentials by canonical and distorted human faces: The role of vertical symmetry and up-down featural arrangement. J. Cogn. Neurosci. 18:1343–1358. Cassia, V. M., C. Turati, and F. Simion, 2004. Can a non-specific bias toward top-heavy patterns explain newborns’ face preference? Psychol. Sci. 15:379–383. Cohen, L. B., and C. H. Cashon, 2001. Do 7-month-old infants process independent features or facial configurations? Infant Child Dev. 10:83–92. Courage, M. L., G. D. Reynolds, and J. E. Richards, 2006. Infants’ attention to patterned stimuli: Developmental change from 3 to 12 months of age. Child Dev. 77:680–695. Damasio, A. R., D. Tranel, and H. Damasio, 1990. Face agnosia and the neural substrates of memory. Annu. Rev. Neurosci. 13:89–109. de Gelder, B., I. Frissen, J. Barton, N. Hadjikhari, 2003. A modulatory role for facial expressions in prosopagnosia. Proc. Natl. Acad. Sci. USA 100:13105–13110. de Gelder, B., and J. J. Stecklenberg, 2005. Naso-temporal asymmetry of the N170 for processing faces in normal viewers but not in developmental prosopagnosia. Neurosci. Lett. 376:40–45. de Haan, M., J. Belsky, V. Reid, A. Volein, and M. H. Johnson, 2004. Maternal personality and infants’ neural and visual

de haan: neurocognitive mechanisms for face processing

517

responsivity to facial expressions of emotion. J. Child Psychol. Psychiatry 45:1209–1218. de Haan, M., and M. Groen, 2006. Neural bases of infants’ processing of social information in faces. In P. J. Marshall and N. Fox, eds., The Development of Social Engagement, 46–80. New York: Oxford University Press. de Haan, M., K. Humphreys, and M. H. Johnson, 2002. Developing a brain specialized for face perception: A converging methods approach. Dev. Psychobiol. 40:200–212. de Haan, M., M. H. Johnson, and H. Halit, 2003. Development of face-sensitive event-related potential components during infancy. Int. J. Psychophysiol. 51:45–58. de Haan, M., M. H. Johnson, D. Maurer, and D. I. Perrett, 2001. Recognition of individual faces and average face prototypes by 1- and 3-month-old infants. Cogn. Dev. 16:659–678. de Haan, M. and C. A. Nelson, 1997. Recognition of the mother’s face by six-month-old infants: A neurobehavioral study. Child Dev. 68:187–210. de Haan, M., and C. A. Nelson, 1998. Discrimination and categorization of facial expressions of emotion during infancy. In A. Slater, ed., Perceptual Development: Visual, Auditory, and Language Development in Infancy. London: University College London Press. de Haan, M., and C. A. Nelson, 1999. Brain activity differentiates face and object processing in 6-month-old infants. Dev. Psychol. 35:1113–1121. de Haan, M., C. A. Nelson, M. R. Gunnar, and K. Tout, 1998. Hemispheric differences in brain activity related to the recognition of emotional expressions by 5-year-old children. Dev. Neuropsychol. 14:495–518. de Haan, M., O. Pascalis, and M. H. Johnson, 2002. Specialization of neural mechanisms underlying face recognition in human infants. J. Cogn. Neurosci. 14:199–209. de Renzi, E, D. Perani, G. A. Carlesimo, M. C. Silveri, and F. Fazio, 1994. Prosopagnosia can be associated with damage confined to the right hemisphere—An MRI and PET study and a review of the literature. Neuropsychologia 32:893–902. de Schonen, S., J. Mancini, R. Camps, E. Maes, and A. Laurent, 2005. Early brain lesions and face processing development. Dev. Psychobiol. 46:184–208. Diamond, R., and S. Carey, 1977. Developmental changes in the representation of faces. J. Exp. Child Psychol. 23:1–22. Driver, J., G. Davis, P. Ricciardelli, P. Kidd, E. Maxwell, and S. Baron-Cohen, 1999. Gaze perception triggers reflexive visuospatial orienting. Visual Cogn. 6:509–540. Eimer, M., 2000. Effects of face inversion on the structural encoding and recognition of faces: Evidence from event-related brain potentials. Brain Res. Cogn. Brain Res. 10:145–158. Farah, M. J., C. Rabinowitz, G. E. Quinn, and G. Liu, 2000. Early commitment of neural substrates in face recognition: Evidence from congenital prosopagnosia. Cogn. Neuropsychol. 17:117–123. Farroni, T., G. Csibra, F. Simion, and M. H. Johnson, 2002. Eye contact detection in humans from birth. Proc. Natl. Acad. Sci. USA 99:9602–9605. Farroni, T., M. H. Johnson, and G. Csibra, 2004. Mechanisms of eye gaze perception during infancy. J. Cogn. Neurosci. 16: 1320–1326. Farroni, T., E. Mansfield, C. Lai, and M. H. Johnson, 2003. Infants perceiving and acting on the eyes: Tests of an evolutionary hypothesis. J. Exp. Child Psychol. 85:199–212. Farroni, T., S. Massaccesi, E. Menon, and M. H. Johnson, 2007. Direct gaze modulates face recognition in young infants. Cognition 102:396–404.

518

cognition

Field, T. M., D. Cohen, R. Garcia, and R. Collins, 1983. Discrimination and imitation of facial expressions by term and preterm neonates. Infant Behav. Dev. 6:485–489. Field, T. M., R. Woodson, R. Greenberg, and D. Cohen, 1982. Discrimination and imitation of facial expression by neonates. Science 218:179–181. Friere, A., and K. Lee, 2001. Face recognition in 4- to 7-year-olds: Processing of configural, featural and paraphernalia information. J. Exp. Child Psychol. 80:347–371. Fries, A. B., and S. D. Pollak, 2004. Emotion understanding in postinstitutionalized Eastern European children. Dev. Psychopathol. 16:355–369. Garrett, A. S., V. Menon, K. MacKenzie, and A. L. Reiss, 2004. Here’s looking at you kid: Neural systems underlying face and gaze processing in fragile X syndrome. Arch. Gen. Psychiatry 61:281–288. Gathers, A. D., R. Bhatt, C. R. Corbly, A. B. Farley, and J. E. Joseph, 2004. Developmental shifts in cortical loci for face and object recognition. NeuroReport 15:1549–1553. Gauthier, I., and C. Bukach, 2007. Should we reject the expertise hypothesis? Cognition 103:322–330. Gobbini, M. I., and J. V. Haxby, 2007. Neural systems for recognition of familiar faces. Neuropsychologia 45:32–41. Goren, C., M. Sarty, and P. Wu, 1975. Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics 56:544–549. Grice, S. J., H. Halit, T. Farroni, S. Baron-Cohen, P. Bolton, and M. H. Johnson, 2005. Neural correlates of eye-gaze detection in young children with autism. Cortex 41:342–353. Grill-Spector, K., R. Sayres, and D. Ress, 2006. Highresolution imaging reveals highly selective nonface clusters in the fusiform face area. Nature Neurosci. 9:1177–1185. Grossman, E., M. Donnelly, R. Price, D. Pickens, V. Morgan, G. Neighbor, and R. Blake, 2000. Brain areas involved in perception of biological motion. J. Cogn. Neurosci. 12:711–720. Grueter, M., T. Grueter, V. Bell, J. Horst, W. Laskowski, K. Sperling, P. Halligan, H. D. Ellis, and I. Kennerknecht, 2007. Hereditary prosopagnosia: The first case series. Cortex 43:734–749. Halit, H., G. Csibra, A. Volein, and M. H. Johnson, 2004. Facesensitive cortical processing in early infancy. J. Child Psychol. Psychiatry 45:1228–1234. Halit, H., M. de Haan, and M. H. Johnson, 2003. Cortical specialisation for face processing: Face-sensitive event-related potential components in 3 and 12 month old infants. NeuroImage 19:1180–1193. Haxby, J. V., M. I. Gobbini, M. L. Furey, A. Ishai, J. L. Schouten, and P. Pietrini, 2001. Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science 293:2425–2430. Haxby, J. V., E. A. Hoffman, and M. I. Gobbini, 2000. The distributed human neural system for face perception. Trends Cogn. Sci. 4:223–233. Haxby, J. V., L. G. Ungerleider, V. P. Clark, J. L. Schouten, E. A. Hoffman, and A. Martin, 1999. The effect of face inversion on activity in human neural systems for face and object recognition. Neuron 22:189–199. Henson, R. N., Y. Goshen-Gottstein, T. Ganel, L. J. Otten, A. Quayle, and M. D. Rugg, 2003. Electrophysiological and haemodynamic correlates of face perception, recognition and priming. Cerebral Cortex 13:793–805. Hooker, C. I., K. A. Paller, D. R. Gitelman, T. B. Parrish, M. M. Mesulam, and P. J. Reber, 2003. Brain networks

for analyzing eye gaze. Brain Res. Cogn. Brain Res. 17:406– 418. Ishai, A., L. G. Ungerleider, A. Martin, J. L. Schouten, and J. V. Haxby, 1999. Distributed representation of objects in the human ventral visual pathway. Proc. Natl. Acad. Sci. USA 96: 9379–9384. Itier, R. J., and M. J. Taylor, 2002. Inversion and contrast polarity reversal affect both encoding and recognition processes of familiar faces. NeuroImage 15:353–372. Itier, R. J., and M. J. Taylor, 2004a. Effects of repetition learning on upright, inverted and contrast-reversed face processing using ERPs. NeuroImage 21:1518–1532. Itier, R. J., and M. J. Taylor, 2004b. Source analysis of the N170 to faces and objects. NeuroReport 15:1261–1265. Johnson, M. H., 2005. Subcortical face processing. Nature Rev. Neurosci. 6:766–774. Johnson, M. H., S. Dziurawiec, H. Ellis, and J. Morton, 1991. Newborns’ preferential tracking of face-like stimuli and its subsequent decline. Cognition 40:1–19. Johnson, M. H., R. Griffin, G. Csibra, H. Halit, T. Farroni, M. de Haan, L. A. Tucker, S. Baron-Cohen, and J. Richards, 2005. The emergence of the social brain network: Evidence from typical and atypical development. Dev. Psychopathol. 17: 599–619. Johnson, M. H., and J. Morton, 1991. Biology and Cognitive Development: The Case of Face Recognition. Oxford, UK: Blackwell. Kahana-Kalman, R., and A. S. Walker-Andrews, 2001. The role of person familiarity in young infants’ perception of emotional expressions. Child Dev. 72:352–369. Kanwisher, N., J. McDermott, and M. M. Chun, 1997. The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17:4302–4311. Kelly, D. J., P. C. Quinn, A. M. Slater, K. Lee, L. Ge, and O. Pascalis, 2007. The other-race effect develops during infancy: Evidence of perceptual narrowing. Psychol. Sci. 18:1084–1089. Kelly, D. J., P. C. Quinn, A. M. Slater, K. Lee, A. Gibson, M. Smith, L. Ge, and O. Pascalis, 2005. Three-month-olds, but not newborns, prefer own-race faces. Dev. Sci. 8:F31–36. Kennerknecht, I., T. Grueter, B. Welling, S. Wentzek, J. Horst, S. Edwards, and M. Grueter, 2006. First report of prevalence of non-syndromic hereditary prosopagnosia (HPA). Am. J. Med. Genet. [A] 140:1617–1622. Kleinschmidt, A., and L. Cohen, 2006. The neural bases of prosopagnsia and pure alexia: Recent insights from functional neuroimaging. Curr. Opin. Neurol. 19:386–391. Kotsoni, E., M. de Haan, and M. H. Johnson, 2001. Categorical perception of facial expressions by 7-month-old infants. Perception 30:1115–1125. Kourtzi, Z., and M. Kanwisher, 2000. Activation in human MT/ MST by static images with implied motion. J. Cogn. Neurosci. 12:48–55. Kylliainen, A., S. Braeutigam, J. K. Hietanen, S. J. Swithenby, and A. J. Bailey, 2006. Face and gaze processing in normally developing children: A magnetoencephalographic study. Eur. J. Neurosci. 23:801–810. LaBarbera, J. D., C. E. Izard, P. Vietze, and S. A. Parisi, 1976. Four- and six-month old infants’ visual responses to joy, anger, and neutral expressions. Child Dev. 47:535–538. Le Grand, R., C. J. Mondloch, D. Maurer, and H. P. Brent, 2001. Neuroperception: Early visual experience and face processing. Nature 410:890. Le Grand, R., P. R. Cooper, C. J. Mondloch, T. L. Lewis, N. Sagiv, B. de Gelder, and D. Maurer, 2006. What aspects of

face processing are impaired in developmental prosopagnosia. Brain Cogn. 61:139–158. Le Grand, R., C. J. Mondloch, D. Maurer, and H. P. Brent, 2003. Expert face processing requires visual input to the right hemisphere during infancy. Nature Neurosci. 6:1108–1112. Lobaugh, M., E. Gibson, and M. J. Taylor, 2006. Children recruit distinction neural systems for implicit emotional face processing. NeuroReport 17:215–219. Mancini, J., S. de Schonen, C. Deruelle, and A. Massoulier, 1994. Face recognition in children with early right or left brain damage. Dev. Med. Child Neurol. 36:156–166. Mason, M. F., B. M. Hood, and C. N. Macrae, 2004. Look into my eyes: Gaze direction and person memory. Memory 12:647–643. Maurer, D., R. L. LeGrand, and C. J. Mondloch, 2002. The many faces of configural processing. Trends Cogn. Sci. 6:255– 260. Maurer, D., and R. Young, 1983. Newborns’ following of natural and distorted arrangements of facial features. Infant Behav. Dev. 6:127–131. McKone, E., and B. L. Boyer, 2006. Sensitivity of 4-year-olds to featural and second-order relational changes in face distinctiveness. J. Exp. Child Psychol. 94:134–162. Meletti, S., F. Benuzzi, G. Rubboli, G. Canatlupo, M. Stanzani Maserati, P. Nichelli, and C. A. Tassinari, 2003. Impaired facial emotion recognition in early-onset right mesial temporal lobe epilepsy. Neurology 60:426–431. Mondloch, C. J., S. Geldart, D. Maurer, and R. Le Grand, 2003. Developmental changes in face processing skills. J. Exp. Child Psychol. 86:67–84. Mondloch, C. J., R. LeGrand, and D. Maurer, 2002. Configural processing develops more slowly than featural processing. Perception 31:553–566. Moore, C., and V. Corkum, 1994. Social understanding at the end of the first year of life. Dev. Rev. 14:349–372. Mosconi, M. W., P. B. Mack, G. McCarthy, and K. A. Pelphrey, 2005. Taking an “intentional stance” on eye-gaze shifts: A functional neuroimaging study of social perception in children. NeuroImage 27:247–252. Nelson, C. A., and M. de Haan, 1996. Neural correlates of infants’ visual responsiveness to facial expressions of emotion. Dev. Psychobiol. 29:577–595. Nelson, C. A., and K. Dolgin, 1985. The generalized discrimination of facial expressions by 7-month-old infants. Child Dev. 56:58–61. Nelson, C. A., P. A. Morse, and L. A. Leavitt, 1979. Recognition of facial expressions by seven-month-old infants. Child Dev. 50:1239–1242. Oster, H., 1981. “Recognition” of emotional expression in infancy? In M. E. Lamb and L. R. Sherrod, eds., Infant Social Cognition: Empirical and Theoretical Considerations, 85–125. Hillsdale, NJ: Erlbaum. Palmero, R., and G. Rhodes, 2007. Are you always on my mind? A review of how face perception and attention interact. Neuropsychologia 45:75–92. Parker, S. W., C. A. Nelson, and the Bucharest Early Intervention Project Core Group, 2005. The impact of early institutional rearing on the ability to discriminate facial expressions of emotion: An event-related potential study. Child Dev. 76:54–72. Pascalis, O., M. de Haan, and C. A. Nelson, 2002. Is face processing species-specific during the first year of life? Science 296: 1321–1323. Pascalis, O., M. de Haan, C. A. Nelson, and S. de Schonen, 1998. Long-term recognition memory for faces assessed by visual

de haan: neurocognitive mechanisms for face processing

519

paired comparison in 3- and 6-month-old infants. J. Exp. Psychol. [Learn. Mem. Cogn.] 24:249–260. Passarotti, A. M., B. M. Paul, J. R. Bussiere, R. B. Buxton, E. G. Wong, and J. S. Stiles, 2003. The development of face and location processing: An fMRI study. Dev. Sci. 6: 100–117. Peelen, M. V., A. J. Wigget, and P. E. Downing, 2006. Patterns of fMRI activity dissociate overlapping functional brain areas that respond to biological motion. Neuron 49:815–822. Pegna, A. J., A. Khateb, C. M. Michel, and T. Landis, 2004. Visual recognition of faces, objects and words using degraded stimuli: Where and when it occurs. Hum. Brain Mapping 22:300–311. Pelphrey, K. A., J. P. Morris, C. R. Michelich, T. Allison, and G. McCarthy, 2005. Functional anatomy of biological motion perception in posterior temporal cortex: An FMRI study of eye, mouth and hand movements. Cerebral Cortex 15:1866– 1876. Pollak, S. D., and P. Sinha, 2002. Effects of early experience on children’s recognition of facial displays of emotion. Dev. Psychol. 38:784–791. Pourtois, G., D. Sander, M. Andres, D. Grandjean, L. Reveret, E. Olivier, and P. Vuilleumier, 2004. Dissociable roles of the human somatosentory and superior temporal cortices for processing social face signals. Eur. J. Neurosci. 20:3507–3515. Puce, A., T. Allison, S. Bentin, J. C. Gore, and G. McCarthy, 1998. Temporal cortex activation in humans viewing eye and mouth movements. J. Neurosci. 18:2188–2199. Puce, A., and D. Perrett, 2003. Electrophysiology and brain imaging of biological motion. Philos. Trans. R. Soc. London [Biol.] 358:435–445. Rebai, M., S. Poiroux, C. Bernard, and R. Lalonde, 2001. Eventrelated potentials for category-specific information during passive viewing of faces and objects. Int. J. Neurosci. 10:209–226. Reid, V., T. Striano, J. Kaufman, and M. H. Johnson, 2004. Eye gaze cueing facilitates neural processing of objects in 4-monthold infants. NeuroReport 15:2552–2555. Robbins, R., and E. McKone, 2007. No face-like processing for objects-of-expertise in three behavioural tasks. Cognition 103:34–79. Rossion, B., R. Caldara, M. Seghier, A.-M. Schuller, F. Lazeyras, and E. Mayer, 2003. A network of occipito-temporal face-sensitive areas besides the right middle fusiform gyrus is necessary for normal face processing. Brain 126: 2381–2395. Rossion, B., and I. Gauthier, 2002. How does the brain process upright and inverted faces? Behav. Cogn. Neurosci. Rev. 1:62–74. Rossion, B., I. Gauthier, M. J. Tarr, P. Despland, R. Bruyer, S. Linotte, and M. Crommelink, 2000. The N170 occipitotemporal component is delayed and enhanced to inverted faces but not to inverted objects: An electrophysiological account of face-specific processes in the human brain. NeuroReport 11:69– 74. Schiltz, C., and B. Rossion, 2006. Faces are represented holistically in the human occipto-temporal cortex. NeuroImage 32: 1385–1394. Schuller, A. M., and B. Rossion, 2004. Perception of static eye gaze direction facilitates subsequent early visual processing. Clin. Neurophysiol. 115:1161–1168. Schwartz, G. M., C. E. Izard, and S. E. Ansul, 1985. The 5month-old’s ability to discriminate facial expressions of emotion. Infant Behav. Dev. 8:65–77. Schweinberger, S. R., E. C. Pickering, I. Jentzsch, A. M. Burton, and J. M. Kaufmann, 2002. Event-related brain

520

cognition

potential evidence for a response of inferior temporal cortex to familiar face repetitions. Brain Res. Cogn. Brain Res. 14:398–409. Senju, A., Y. Tojo, K. Yaghuchi, and T. Hasegawa, 2005. Deviant gaze processing in children with autism: An ERP study. Neuropsychologia 43:1297–1306. Shibata, T., H. Nishijo, R. Tamura, K. Miyamoto, S. Eifuku, S. Endo, and T. Ono, 2002. Generators of visual evoked potentials for faces and eyes in the human brain as determined by dipole localization. Brain Topogr. 15:51–63. Simion, F., E. Valenza, C. Umilta, and B. Dalla Barba, 1998. Preferential orienting to faces in newborns: A temporal-nasal asymmetry. J. Exp. Psychol. [Hum. Percept.] 24:1399–1405. Soken, N. H., and A. Pick, 1992. Intermodal perception of happy and angry expressive behaviors by seven-month-old infants. Child Dev. 63:787–795. Spencer, J., J. O’Brien, A. Johnston, and H. Hill, 2006. Infants’ discrimination of faces by using biological motion cues. Perception 35:79–89. Tarr, M. J., and I. Gauthier, 2000. FFA: A flexible fusiform area for subordinate-level visual processing automatized by expertise. Nature Neurosci. 3:764–769. Taylor, M. J., M. Batty, and R. Itier, 2004. The faces of development: A review of early face processing over childhood. J. Cogn. Neurosci. 16:1426–1442. Taylor, M. J., G. E. Edmonds, G. McCarthy, and T. Allison, 2001. Eyes first! Eye processing develops before face processing in children. NeuroReport 12:1671–1676. Turati, C., F. Simion, I. Milani, and C. Umilta, 2002. Newborns’ preference for faces: What is crucial? Dev. Psychol. 38:875–882. Tzourio-Mazoyer, N., S. de Schonen, F. Crivello, B. Reutter, Y. Aujard, and B. Mazoyer, 2002. Neural correlates of woman face processing by 2-month-old infants. NeuroImage 15:454–461. Valenza, E., F. Simion, V. M. Cassia, and C. Umilta, 1996. Face preference at birth. J. Exp. Psychol. [Hum. Percept.] 22:892–903. Vecera, S. P., and M. Rizzo, 2006. Eye gaze does not produce reflexive shifts of attention: Evidence from frontal-lobe damage. Neuropsychologia 44:150–159. Vuilleumier, P., and G. Portois, 2007. Distributed and interactive brain mechanisms during emotion face perception: Evidence from functional neuroimaging. Neuropsychologia 45:174–194. Walker, A. S., 1982. Intermodal perception of expressive behaviors by human infants. J. Exp. Child Psychol. 33:514–535. Wang, A. T., M. Dapretto, A. R. Hariri, M. Sigman, and S. Y. Bookheimer, 2004. Neural correlates of facial affect processing in children and adolescents with autism spectrum disorder. J. Am. Acad. Child Adolesc. Psychiatry 43:481–490. Wheaton, K. J., J. C. Thompson, A. Syngeniotis, D. F. Abbott, and A. Puce, 2004. Viewing the motion of human body parts activates different regions of premotor, temporal and parietal cortex. NeuroImage 22:277–288. Winston, J. S., R. N. Henson, M. R. Fine-Goulden, and R. J. Dolan, 2004. fMRI-adaptation reveals dissociable neural representations of identity and expression in face perception. J. Neurophysiol. 92:1830–1839. Yin, R. K., 1969. Looking at upside-down faces. J. Exp. Psychol. 81:141–145. Young-Browne, G., H. M. Rosenfeld, and F. D. Horowitz, 1977. Infant discrimination of facial expressions. Child Dev. 48: 555–562. Yovel, G., and N. Kanwisher, 2004. Face perception: Domain specific, not process specific. Neuron 44:889–898. Yovel, G., and N. Kanwisher, 2005. The neural basis of the behavioural face-inversion effect. Curr. Biol. 15:2256–2262.

32

The Development of Visuospatial Processing JOAN STILES, BRIANNA PAUL, AND WENDY ARK

Visuospatial processing refers to a wide variety of abilities and skills ranging from tracking a moving object, to localizing and attending to an object or event in the spatial array, to understanding how the parts or features of an object combine to form an organized whole. Neuropsychological studies of visual system architecture have identified dozens of interrelated visual areas in the posterior cortices, each of which contributes to some aspect of visuospatial processing (Van Essen, Anderson, and Felleman, 1992). In the early 1980s, Ungerleider and Mishkin (1982) outlined a useful scheme for understanding the organization of this complex set of cortical areas and functions. Their description of a dual visuospatial processing system derived from the growing body of evidence suggesting distinct functional dissociations within striate and extrastriate visual areas. According to their proposal, the cortical visual system can be functionally and anatomically subdivided into two principal processing streams (see figure 32.1 and plate 53), a dorsal pathway and a ventral pathway. The dorsal stream mediates spatial processing associated with attention to movement and location, while the ventral stream is primarily involved in processing information about patterns and objects. Their proposal introduced an alternative to earlier accounts in which attention to location and movement was associated principally with the tectal visual system, while form processing was linked to the geniculostriate system (Schneider, 1967; Trevarthen, 1968). Ungerleider and Mishkin’s account did not deny the role of the tectal system in, particularly, the oculomotor aspects of spatial processing. Rather, it provided a more complete account of the organization and functioning of higher cortical systems, and of spatial cognitive processing more generally. The dorsal visual pathway begins at the retina and projects via the lateral geniculate nucleus (LGN) of the thalamus to primary visual cortex, area V1. From there the pathway proceeds to areas V2 and V3, then projects dorsally to the medial (MT/V5) and medial superior (MST) regions of the temporal lobe, and then to the ventral inferior-parietal lobe (IP). Input to the dorsal pathway is derived principally, though not exclusively, from the large M-type retinal ganglion cells that project to the magnocellular layers of LGN and then to layer 4C alpha of V1. Cells in this pathway are

maximally sensitive to movement and direction, and are less responsive to color or form. The original functions identified for the dorsal stream involved processing of information about spatial location, optic flow, and motion, as well as allocation and maintenance of spatial attention. It was thus described as the “where” pathway. More recently, work examining the dorsal stream’s role in visually guided movements suggests that the function of this pathway may be more complex (e.g., Goodale and Milner, 1992; Andersen et al., 1997; Rizzolatti and Matelli, 2003). In a recent model, Rizzolatti and Matelli (2003) proposed that the dorsal stream is composed of an inferior parietal system for processing visual perceptual inputs (the “where” system) and a superior parietal system that is involved in sensation and action (the “how” system). The ventral visual pathway also begins at the retina and projects via the LGN of the thalamus to primary visual cortex, area V1. From there the pathway proceeds to areas V2 and V4, and then projects ventrally to the posterior (PIT) and anterior (AIT) regions of the inferior temporal lobe. Input to the ventral pathway is derived principally, though not exclusively, from P-type retinal ganglion cells that project to the parvocellular layers of the LGN and then to layer 4C beta of V1. Parvocellular input to V1 organizes into distinct areas called the blob and interblob regions (Wong-Riley, 1979; Livingstone and Hubel, 1984). Cells in the blob regions are maximally sensitive to form, while cells in the interblob regions respond principally to color. The ventral stream processes information about visual properties of objects and patterns, and it has been described as the “what” pathway. Both the dorsal and ventral pathways project rostrally to both common and adjacent areas of the prefrontal cortex. Finally, there is substantial evidence that the two pathways are richly interconnected and at least partially overlapping in both the mature (e.g., Merigan and Maunsell, 1993; Dobkins and Albright, 1994, 1995, 1998; Marangolo et al., 1998; Thiele, Dobkins, and Albright, 2001; Sincich and Horton, 2005) and the developing visual system (Dobkins and Teller, 1996a, 1996b; Dobkins and Anderson, 2002). The anatomical division of visuospatial functioning into the dorsal and ventral streams provides a convenient

521

Figure 32.1 Visual processing pathways in monkeys. Solid lines indicate connections arising from both central and peripheral visual field representations; dotted lines indicate connections restricted to peripheral field representations. Red boxes indicate ventral stream areas related primarily to object vision; green boxes indicate dorsal stream areas related primarily to spatial vision; and white boxes indicate areas not clearly allied with either stream. The shaded region on the lateral view of the brain represents the extent of the cortex included in the diagram. Abbreviations are as follows: DP, dorsal prelunate area; FST, fundus of superior temporal area; HIPP, hippocampus; LIP, lateral intraparietal area; MSTc, medial superior temporal area, central visual field representation; MSTp, medial superior temporal area, peripheral visual field representation; MT, middle temporal area; MTp, middle temporal area, peripheral visual field representation; PO, parieto-occipital area;

PP, posterior parietal sulcal zone; STP, superior temporal polysensory area; V1, primary visual cortex; V2, visual area 2; V3, visual area 3; V3A, visual area 3, part A; V4, visual area 4; and VIP, ventral intraparietal area. Inferior parietal area 7a; prefrontal areas 8, 11 to 13, 45, and 46; perirhinal areas 35 and 36; and entorhinal area 28 are from Brodmann (1909). Inferior temporal areas TEO and TE, parahippocampal area TF, temporal pole area TG, and inferior parietal area PG are from von Bonin and Bailey (1947). Rostral superior temporal sulcal (STS) areas are from Seltzer and Pandya (1978), and VTF is the visually responsive portion of area TF (Boussaoud, Desimone, and Ungerleider, 1991). (See plate 53.) (Figure and caption reprinted from L. G. Ungerleider, 1995. Functional brain imaging studies of cortical mechanisms for memory. Science 270:770.)

organizational scheme for discussing the neurodevelopmental underpinnings of visuospatial processing. Accordingly, the next two sections reflect this anatomical division. The first section focuses on processes associated with the dorsal stream, and the second section takes up issues related to processes associated with the ventral stream. It should be noted that while a great deal is known about the neuropsychology of dorsal and ventral stream organization and function in adult human and animal populations, the study of the development of the neuropsychological underpinnings of visuospatial processing is much more limited. To date, there have been very few developmental

studies that provide direct evidence mapping specific behavioral changes in spatial processing to specific changes in the neural substrate. Nonetheless, important basic information about the postnatal development of at least some aspects of the visuospatial processing system is beginning to become available. The review that follows focuses on those aspects of visuospatial functioning for which neurodevelopmental data are also available and includes the dorsal stream functions of spatial localization, spatial attention, and mental rotation and the ventral stream functioning associated with part-whole or global-local aspects of visual pattern processing.

522

cognition

The development of spatial processes associated with the dorsal stream: Spatial localization, spatial attention, and mental rotation A variety of spatial processes have been associated with activation of the dorsal visual pathway. Three very basic processes will be considered in this section: (1) spatial localization, (2) spatial attention, and (3) mental rotation. The characterization of these basic dorsal-system processes as independent and distinct is somewhat artificial in that, for example, localization of an object may also require a shift in spatial attention, and mental translation of an object must involve both localization and attention in space. Nonetheless, there is substantial evidence for functional and anatomical independence of key features of each process. Spatial Localization Spatial localization is a complex, multimodal process that engages a wide array of both cortical and subcortical brain systems. The cortical dorsal-stream parietal system is centrally involved in processing location, but other neural systems are also involved. For example, the basal ganglia have been implicated in orienting to the location of reward (Hadj-Bouziane, Meunier, and Boussaoud, 2003; Hikosaka, Nakamura, and Nakahara, 2006), and it is well documented that the hippocampus plays a critical role in both spatial map formation (O’Keefe and Nadel, 1978; D. Smith and Mizumori, 2006) and episodic spatial working memory for locations (Chiba, Kesner, and Jackson, 2002; D. Smith and Mizumori, 2006). Further, localization is not confined to the visual system, but rather is an important aspect of both sensory and motor processing (Colby and Goldberg, 1999; Middlebrooks et al., 2002; Konishi, 2003). Visual localization is mediated principally by the dorsal visual system. Evidence from both human and animal studies has shown that the dorsal stream plays a critical role in perceptual localization (Belger et al., 1998; Chiba, Kesner, and Jackson, 2002), while positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) studies with adult humans have also implicated the parietal region in episodic memory for spatial location (Belger et al., 1998; Wagner et al., 2005). In an early series of studies using PET imaging, Haxby examined profiles of posterior brain activation in adults with tasks requiring them to compare the location of objects in two visually presented arrays (Haxby et al., 1991, 1994). The basic activation findings of Haxby’s experiments on location processing are consistent with animal studies (Colby and Duhamel, 1996; Colby and Goldberg, 1999; Rizzolatti and Matelli, 2003) and have been largely replicated in subsequent fMRI, PET, and transcranial magnetic stimulation (TMS) studies using a variety of perceptual and working-memory tasks (Jonides et al., 1993; E. Smith et al., 1995; E. Smith, Jonides, and Koeppe, 1996;

Belger et al., 1998; Casey et al., 1998; Nelson et al., 2000; Oliveri et al., 2001; Ellison and Cowey, 2006). Haxby’s studies showed that in the most posterior brain regions, areas of extrastriate cortex are activated bilaterally; these include the dorsolateral occipital (area 19), the calcarine, and medial and lateral areas of the occipital lobe. These areas are presumed to be involved in early visual processing and are typically activated by a wide array of visual processing tasks. Within the parietal lobe, activation was observed bilaterally in posterior superior parietal areas, extending rostrally to the intraparietal sulcus (Brodmann’s area 7). These areas are considered important for location processing. In addition, subsequent studies identified the inferior parietal lobe as important in perceptual processing of location (Colby and Duhamel, 1996; Courtney et al., 1996). Finally, a large number of functional neuroimaging studies have demonstrated the importance of frontal regions in spatial working memory for locations. Two regions that appear to be particularly important for spatial working memory in humans include the superior frontal cortex (SFC) Courtney et al., 1998; Haxby et al., 2000; Sala, Rama, and Courtney, 2003; Curtis, 2006) and dorsolateral prefrontal cortex (DLPFC) (Postle et al., 2000; Curtis, 2006). The task of looking or reaching to a spatial location involves a complex network of neural areas within the dorsal pathway, and there is strong evidence that the dorsal frontoparietal system plays a major role in control of visually guided movement, including both eye movement and reach (Colby and Duhamel, 1996; P. Johnson et al., 1996; Wise et al., 1997; Colby and Goldberg, 1999; Rizzolatti and Matelli, 2003; Pierrot-Deseilligny, Milea, and Muri, 2004). Prefrontal motor areas mediate planning and preparation for motor action; activation of these areas typically precedes the actual motor event. There is considerable evidence for superior parietal input to dorsal premotor and motor cortices; activation in frontal areas and activation in superior parietal areas are concordant, suggesting a network of spatial-motor control (Rizzolatti and Matelli, 2003). In addition, recent studies have shown that inferior parietal areas connect to frontal premotor areas and play an important modulatory role in spatial-motor activity (Andersen et al., 1997). Cells in this region respond to modifications of motor actions. Further, different cells are selective for movement type such that different groups of cells respond to eye movement and reach. Consistent with these findings, Goodale and Milner (1992) have reported the case of a woman with bilateral parietal lesions who showed specific difficulty in reaching for objects. These data suggest that the “where” system may also serve as a “how” system mediating action in the spatial world. In a recent review of studies of perceptual and motor-based activation within the dorsal stream, Rizzolatti and Matelli (2003) suggested that the dorsal system may comprise two separate but interrelated systems, an inferior

stiles, paul, and ark: development of visuospatial processing

523

parietal system dominated by visual perceptual inputs and a superior parietal system governed by somatosensory information that is used to guide action. Location processing during infancy. One of the largest bodies of data on early visuospatial processing comes from a simple, spatial hiding task, originally introduced by Piaget (1952). In the task, children watch as a toy is hidden under one of two screens (A and B), and then they are encouraged to retrieve it. In a typical test sequence, the object is first hidden at A for two to three trials, but on the next trial the hiding location is changed to B. Eight-month-olds easily retrieve the object hidden under A (but also see L. Smith et al., 1999), but when the object is then hidden under B, they continue to search at A. This error has been termed the A not B error (AB error), and it has been widely conceptualized as an index of object permanence, that is, of the infant’s knowledge that objects exist independently over space and time. However, although there is agreement that the behavior in some way taps infants’ representation of objects, the nature and content of that representation has been the source of controversy for decades (see Harris, 1987; Wellman, Cross, and Bartsch, 1987; L. Smith et al., 1999). The dispute centers on the wide range of factors that affect task performance. The AB error is observed on the standard task between 8 and 12 months. However, children’s experience outside the task setting can influence the likelihood of making the error (see Hauser, 1999). The experience of self-locomotion, either naturally occurring or introduced by the use of an infant walker, significantly reduces the likelihood of AB error (Horobin and Acredolo, 1986; Kermoian and Campos, 1988; Bertenthal and Campos, 1990). Further, healthy preterm infants are more advanced on the AB search task than full-term peers matched for conceptional age (Matthews, Ellis, and Nelson, 1996), suggesting that the extra experience in the world offers the healthy preterms a developmental advantage. Altering task demands also affects AB task performance. Some factors, such as the use of salient landmarks, distinctive screens, or increased distance between the screens, improve performance (Butterworth, Jarrett, and Hicks, 1982; Wellman, Cross, and Bartsch, 1987). Similarly, when children need only look rather than reach to the hiding location, error is reduced (Baillargeon and Graber, 1988; Baillargeon and DeVos, 1991; Hofstadter and Reznick, 1996; Ahmed and Ruffman, 1998). Further, change in the infant’s posture from sitting on A-trials to standing on the B-hiding trials reduces error (L. Smith et al., 1999). However, increasing task demands negatively impacts performance. Introduction of a delay between hiding and retrieval increases error frequency among children as old as 12 months (Diamond, 1985). Similarly, creating very difficult task conditions elicits the AB error in children well beyond 12 months (Spencer, Smith, and Thelen, 2001). The varied behavioral

524

cognition

results from infants on this basic search task suggest that many factors contribute to the level of performance. Although data from neuropsychological studies of performance on the AB search task are very limited, several lines of evidence have begun to emerge that may provide clues about change in the specific neural systems critical for successful search performance. Adult rhesus monkeys can be trained to ceiling performance on the standard AB task, and they perform well with delays up to 15 seconds (Diamond, 1991; Diamond, Werker, and Lalonde, 1994). However, bilateral lesions of dorsolateral prefrontal cortex disrupt performance in the memory load, but not the no-load, condition. By contrast, bilateral lesions to parietal cortex (Diamond, 1991; Diamond, Werker, and Lalonde, 1994) or hippocampus (Diamond, ZolaMorgan, and Squire, 1989) have little effect on AB search performance. Diamond suggests that the dorsolateral prefrontal lesion data implicate two important functions necessary for the successful AB search performance: explicit spatial memory and inhibitory control. It is well documented that prefrontal lesions impair spatial memory (GoldmanRakic, 1987). Rats with lesions to dorsomedial prefrontal cortex have specific memory impairments for the temporal order of spatial location, which is an important component of the AB task (Chiba, Kesner, and Gibson, 1997). Further, the effects of lesions to dorsolateral prefrontal cortex are consistent with the human adult activation studies reported earlier. Although right dorsolateral prefrontal activation was observed in both the perception and memory for location tasks, it was enhanced, particularly under conditions of memory load. Frontal lesions have also been associated with failures of inhibitory control (Welsh and Pennington, 1988). The AB search task requires children to inhibit a prepotent response (reach to A) in order to make the new, correct response (reach to B). Diamond (1991) and Diamond, Werker, and Lalonde (1994) compared performance on the AB search task and an object retrieval task that required inhibition of a prepotent response, but had no memory component, and found that children showed similar patterns of developmental change on both tasks. The results of the behavioral studies with infants, in conjunction with data from lesion studies, led Diamond (1991) to conclude that maturational change in the frontal cortices underlies improved performance on the AB task. In support of this view, she noted that studies of regional change in brain metabolic activation over the first year of life in human children indicate that increases in frontal metabolic activity begin at about 8 months of age (Chugani, Phelps, and Mazziotta, 1987; Diamond, Werker, and Lalonde, 1994; Jacobs et al., 1995). In addition, systematic increases in electroencephalography (EEG) spectral power over frontal, but not parietal or occipital, brain regions have been reported for the critical period of developmental change on the

AB task (Fox and Bell, 1990). More recently, studies using near–infrared spectroscopy (NIS) to measure localized brain activation in infants have provided converging evidence for the association between frontal lobe change and successful search performance (Baird et al., 2002). Although the evidence for the role of frontal lobe maturation in successful search performance is compelling, other data suggest that the full developmental story may be more complex. For example, in addition to their work on frontal EEG spectral power, M. Bell and Fox (1992) have also examined patterns of long-range EEG coherence. They reported that an increase in anterior-posterior coherence was associated with improved performance on the AB error task, suggesting that stabilization of long-range axonal connections may contribute importantly to the observed change in performance. As discussed earlier, these pathways may be critical for control of spatial reaching. Thus maturation of the anterior-posterior system may account for the widely reported differences in performance on reaching and looking tasks. Moreover, while related to frontal lobe development, the coherence data suggest more widespread changes in the dorsal system that may well affect many different aspects of performance on this task. Finally, EEG data have also been used to examine potential markers of object representation. Gamma-band activity has been associated with maintenance of mental representations of objects among adults (TallonBaudry et al., 1998). Recent studies measuring gamma-band activity in the EEGs of 6-month-old infants during objectprocessing and object-occlusion tasks suggest that by the middle of the first year of life, the neural signature of object representation can be detected (Csibra et al., 2000; Kaufman, Csibra, and Johnson, 2003, 2005). Although still limited, the body of neuropsychological data on the AB task has begun to define the neural changes that contribute to the changes in task performance. Together they suggest that a complex network of neural systems emerges across the first year to support performance on this seemingly simple task. The data point to changes in both frontal and parietal regions within the dorsal stream and, likely, to comparable changes within temporal and frontal regions of the ventral stream. As M. H. Johnson has noted, the changes within these different neural regions are unlikely to be unitary events; rather, neural development is likely to reflect a more gradual “coming online” of the different components of the complex neural system that progressively comes to support the range of behaviors involved in the visual search task (M. Johnson, Mareschal, and Csibra, 2001). Location processing among older children. Although studies of location coding in toddlers suggest that they are able to make use of fine-grained distance information when searching for hidden objects, the tendency to subdivide space (“hierarchical” coding) in order to facilitate remembering an object’s

location does not emerge until approximately age 4 (J. Huttenlocher, Newcombe, and Sandberg, 1994). Further, it is not until 10 years that children show reliable, adultlike spatial coding of fine-grained, multidimensional categorical information (Sandberg, Huttenlocher, and Newcombe, 1996). This finding is consistent with other studies demonstrating improvements in location memory through mid to late childhood (Orsini et al., 1987; Zald and Iacono, 1998; S. Bell, 2002; Luciana et al., 2005). Increasing task demands, for example, by requiring that multiple spatial positions be recalled in a certain order, extends the period of immature performance into early adolescence (Gathercole et al., 2004; Luciana et al., 2005; Farrell Pagulayan et al., 2006). Finetuning of location information encoded in memory is reported to extend through late adolescence (Luna et al., 2004). Pediatric neuroimaging studies of spatial working memory have begun to document changes in the neural substrate that may parallel observed behavioral changes. Early studies of spatial working memory showed that children and adults recruit largely similar brain regions (Thomas et al., 1999; Nelson et al., 2000). Subsequent studies further characterized the response pattern in activated regions across childhood and adolescence by correlating it with developmental indices. Both age (Kwon, Reiss, and Menon, 2002) and location memory performance (Klingberg, Forssberg, and Westerberg, 2002; Olesen et al., 2003) were shown to be associated with increased blood oxygen–level dependent (BOLD) activation in regions of the prefrontal and posterior parietal cortices. More recent work with a large sample of typically developing adolescents (Schweinsburg, Nagel, and Tapert, 2005) has also revealed subtle changes in the topography of the brain response within these regions, noting an agerelated superior-to-inferior shift in posterior parietal regions hypothesized to reflect more efficient strategy use with age. Using diffusion tensor imaging (DTI), Klingberg (2006), Nagy, Westerberg, and Klingberg (2004), and Olesen and colleagues (2003) extended these findings by incorporating indices of white matter development. Correlations of BOLD response with fractional anisotropy (FA) within regions recruited by spatial working-memory tasks revealed a frontalparietal network within the dorsal stream that may represent the neural counterpart of change in behavioral performance extending from childhood through adolescence. Spatial Attention A closely related line of investigation focuses on the neural systems associated with attention to different locations in space. In contrast to work that examines profiles of brain activity when subjects are required to directly perceive or remember the location of an object, spatial attention tasks investigate the brain systems activated when subjects are asked to shift their attention to different locations. The ability to shift attention to different locations

stiles, paul, and ark: development of visuospatial processing

525

is one part of the larger attentional system involving widely distributed brain areas. Review of the full literature on attention is beyond the scope of this chapter. This section will focus more narrowly on those aspects of the attention system involved in shifting of attention in space. There is considerable clinical and experimental evidence that the posterior parietal lobes play a crucial role in the ability to shift attention to different spatial locations (Posner, 1980; Posner et al., 1984; Robertson, 1992; Heilman and Valenstein, 1993; Rafal and Robertson, 1995; Hillyard and Anllo-Vento, 1998; Ivry and Robertson, 1998). Posner (1980) and Posner and Petersen (1990) have presented one influential model of the attention system that involves an interconnected network of structures that modulate and control different aspects of attention. According to Posner, the posterior parietal network plays an essential role in disengaging attention from one location and allowing a shift of attention to another location. In the standard task used to test covert shifts of attention (Posner and Cohen, 1980), subjects are seated in front of a computer and instructed to continuously fixate on a point located centrally between two identical, flanking squares. After a fixed period, a visual cue is presented either centrally (e.g., an arrow) or peripherally (e.g., one box brightens), and soon after a target appears briefly in one box. The subject responds as soon as the target is detected. The critical variable is the validity of the cue. On most trials (75 to 80%) the cue is “valid,” and the target appears in the cued box. On the remaining trials, the cue is “invalid,” and the target appears in the opposite box. If cuing serves to covertly shift attention, it should take less time to detect the target when the cue is valid than when it is invalid. One additional, well-established finding concerns response differences associated with the length of the interval between the valid cue and target, or stimulus onset asychrony (SOA). With short SOAs (<200 msec), the classic facilitation of response time is observed. However, at longer SOAs (300–1,300 msec) responses to the cued target are slowed (e.g., Posner et al., 1985). This phenomenon, which has been called inhibition of return (IOR), is thought to reflect an evolutionarily im-portant suppression of a response to an already attended location. To examine patterns of brain activation associated with shifting attention, Corbetta used a variant of the attentional cuing task (Corbetta et al., 1993; Corbetta, 1998). The results of this study confirmed earlier reports from both human and animal work on the role of the parietal lobes in shifting spatial attention. Significant foci of brain activation were observed in both left and right superior parietal regions. However, the patterns of activation to stimuli presented to the right and left visual fields (RVF, LVF) were not symmetrical across the hemispheres. Presentation of targets to the LVF produced significantly more activation in the RH than the LH, while presentation of targets to the RVF produced significant levels

526

cognition

of activation in both the RH and LH. Furthermore, distinct activation sites for RVF and LVF targets were identified within the right superior parietal region, suggesting that different brain regions within the right hemisphere (RH) are responsible for processing information from the two sides of space. Further, adult patient data are consistent with these findings. Patients with injury to posterior parietal areas, including the cortex, superior colliculus, and pulvinar, have difficulty shifting attention; specifically they show exaggerated slowing of reaction time (RT) following presentation of invalid cues (Posner et al., 1984). By contrast, patients with supranuclear palsy, a disorder associated with damage to the superior colliculus, fail to show IOR responses. Finally, patients with thalamic damage are slow to detect targets presented anywhere in the visual field. There is a small but growing literature on infants’ ability to shift attention in the visual field (also see Colombo, 2001). A number of studies have shown that by 6 months, infants show both facilitation and IOR (Clohessy et al., 1991; Hood, 1993; Harman et al., 1994; M. Johnson, Posner, and Rothbart, 1994; M. Johnson and Tucker, 1996). Attempts to evoke these responses from younger children have been mixed. However, control of factors such as SOA duration and cue/target eccentricity appears to be critical for eliciting the responses. Using 200 and 700 SOAs, M. Johnson and Tucker (1996) demonstrated reliable facilitation and IOR among 4-month-olds but not among 6-month-olds. However, when a 133-msec SOA was introduced, 6-month-olds showed strong facilitation. This finding suggests that while the basic attentional responses may be robust as early as 4 months, the timing parameters that elicit the response may change with development. Similarly, Harman and colleagues (1994) found no IOR response among 3-month-old children when stimuli were presented at 30 degrees eccentricity, but a strong response at 10 degrees. Thus distribution of attention across the visual field may also change with development. Few studies have examined facilitation and IOR in children under 2 months. M. H. Johnson reported only weak facilitation effects and no IOR effects among 2-month-old infants. However, Valenza, Simion, and Umilta’s (1994) study of newborns suggests that IOR may be present in the first days of life. Finally, the child’s prior experience in the world, as indexed by familiarity responses, has been shown to affect components of the EEG responses associated with attention. Such findings in children as young as 4.5 months suggest that memory may have a modulatory effect on attention from very early in life (Reynolds and Richards, 2005). Further, recent data have shown that social cues can direct covert shifts of attention as early as 4 months (Reid et al., 2004). Direction of eye gaze is a potent cue for shared attention. When infants observed an adult shift gaze toward (cue) or away from (uncued) a target object, EEG responses to subsequent presentations of the cued or uncued object differed

in frontotemporal brain regions. These findings suggest that a social cue can induce covert shifts in the infants’ attention. Studies of children with early brain injury provide some insight into the neural mechanisms that mediate these basic attentional processes in development. Craft and colleagues (1994) examined attentional shifts in children with perinatal brain injury, using the shift-attention standard task with 100and 800-msec SOAs. Children with posterior lesions did not show any of the specific deficits noted for adults with injury to the parietal complex. However, children with anterior lesions failed to show the facilitation effect to valid cues presented in the right, but not the left, visual field. This finding suggests an effect of left anterior injury not observed among adults. M. H. Johnson reported a similar finding for infants with unilateral left anterior perinatal focal brain injury ( Johnson et al., 1998). These data suggest that the systems crucial to the control of attentional processes may change with development, such that, early on, frontal regions, in particular left frontal regions, may be crucial to efficient allocation of attention across the spatial field. Mental Rotation Mental rotation is an important spatial operation that contributes to processes such as shape perception, spatial reasoning, and problem solving. The typical mental rotation task requires a host of visuospatial skills including visual pattern processing, visuospatial attention, and visuospatial working memory. A common method used to study mental rotation is to present two objects, one upright and one rotated off vertical, and ask participants if the objects are the same or mirror images. The robust result is that response times vary as a linear, monotonically increasing function of angular disparity between the two objects. This linear response time has become the hallmark characteristic of mental rotation (see Shepard and Cooper, 1986). Over the past decade, many neuroimaging studies have examined the neural systems involved with mental rotation in adults. Since there are a number of visuospatial skills involved with mental rotation, it is not surprising that brain activation obtained during mental rotation tasks reflects a network of interacting brain regions rather than a central “mental rotation area.” The most commonly reported neural areas associated with mental rotation are the parietal lobes, particularly the right superior parietal lobe, higher-order visual areas (such as MT), and the premotor area. The functions of the parietal lobe include the encoding of spatial relations and the allocation of visual attention (M. Cohen et al., 1996), the right parietal lobe being dominant in spatial processing (Andersen, 1988). Richter and colleagues (1997) suggested that the superior parietal lobe (SPL) is involved with the actual execution of mental rotation because the duration of the fMRI signal in the SPL is equal to the response time when performing a mental rotation task. The MT area

is important for processing visual motion (Riecansky, 2004). The activation of MT during mental rotation suggests that participants are actually imagining the rotation of the object and provides neural evidence to support the idea that mental rotation is the analogue of perceived motion. One method to mentally rotate an object could involve imagining one’s hand grasping and rotating the object. There are rich connections between the posterior parietal cortex and the supplementary motor area (SMA 6) that allow a match between visual and motor inputs (Sakata and Taira, 1994), and SMA 6 is associated with grasping behavior (Rizzolatti et al., 1988); therefore, SMA 6 activation during a mental rotation task could be an important element in understanding the method by which mental rotation is accomplished (M. Cohen et al., 1996; Kosslyn et al., 1998). Developmental studies of mental rotation. Mental rotation has been documented in children as young as age 5 (Marmor, 1975; Kosslyn et al., 1990). In an early study, Marmor (1975) found that the RT regression slopes of 5-year-olds were similar to adults and concluded that children are able to perform mental spatial transformations. Subsequent studies have replicated this basic finding (Kosslyn et al., 1990) and confirmed through verbal reports that children use mental rotation to make the judgments even under conditions where no explicit instructions to mentally rotate are given (Estes, 1998). Although young children appear to engage in mental rotation, developmental differences are observed in speed and efficiency of processing. Marmor’s original study noted differences in overall RT. Several studies found a developmental decrease in the rate of mental rotation, until about 13 years (Kail, Pellegrino, and Carter, 1980; Merriman, Keating, and List, 1985; Hale, 1990; Snow, 1990). Further, a wide range of factors has been shown to affect mental rotation performance. In addition to age, factors such as IQ, gender, socioeconomic status, videogame playing, stimulus type employed, and practice on the MR task all impact performance (Waber, Carlson, and Mann, 1982; Willis and Schaie, 1988; Okagaki and Frensch, 1994; Cai and Chen, 2000; De Lisi and Wolford, 2002). Thus the observed developmental changes could reflect improvement in initially rudimentary mental rotation skills, or the data could reflect changing strategies for solving the matching problems presented within the context of the standard mental rotation tasks. In short, there is evidence that children perform mental rotation from a young age, but their speed of processing and accuracy improve well into adolescence. Given the number of visuospatial skills that mental rotation relies on and the host of factors that could affect the development of this skill, it is not surprising that the development of this ability is protracted. Only a few studies have reported data on the neural systems that underlie mental rotation during development.

stiles, paul, and ark: development of visuospatial processing

527

In general, children show patterns of parietal activation similar to adults, but children’s activation appears to be more diffuse (Booth et al., 1999, 2000; Roberts and Bell, 2002). The more distinctive patterns of activation, including greater superior parietal activation, reported for adults compared to 9- to 10-year-olds, may be an index of increasing functional specialization (Booth et al., 2000). Further, an EEG study of 8-year-old children reported bilateral parietal activation (right < left), which could reflect emerging hemispheric specialization for the mental rotation task (Roberts and Bell, 2002). More recently, Ark, Haist, and Stiles (in preparation), tested 9- to 10-year-old children using behavioral and fMRI measures of mental rotation with challenging, 3-D stimuli. The major finding from the behavioral data was the presence of two distinct subgroups for children, but not adults. The child group could be equally divided into high performers (HP) and low performers (LP) defined by accuracy on displacements over 40 degrees. Overall, the LP children were faster to respond than HP children, and their RTs failed to reflect the usual slowing with increased angular disparity. Accuracy differences between the LP and HP groups were most evident on trials with the greatest degrees of angular disparity (80 degrees or higher), where performance of the LP group was at chance. These findings suggest that with high angular disparity, the LP group could not perform the mental rotation task. The major findings from the fMRI study were the differences in activation between adults and children in two important brain areas related to mental rotation: the parietal area and MT. Consistent with earlier studies, children had more bilateral and widespread activation in the parietal lobes than adults. In addition, adults had more activation in MT than the LP, but not the HP, group. It is thought that MT plays a role in imagining the movement of the figures in the mental rotation task. Based on the behavioral data, the LP children did not activate MT because they were not efficiently performing mental rotation. In summary, the behavioral and neuroimaging data suggest that the period of development for mental rotation is protracted. Though there is evidence that young children are able to mentally rotate, a close inspection of the behavioral data revealed that only half the children’s data displayed the characteristic mental rotation slope. Also, the children who displayed the mental rotation slope had patterns of neural activation more similar to those of adults than the children who did not display the mental rotation slope.

Development of visuospatial processes associated with the ventral stream: Understanding parts and wholes and how they go together A major function of the ventral visual stream is the analysis of pattern information. Behaviorally, visuospatial analysis is defined as the ability to specify the parts and the

528

cognition

overall configuration of a visually presented pattern, and to understand how the parts are related to form an organized whole (e.g., Vurpillot, 1976; Palmer, 1977; L. Smith and Kemler, 1977; Palmer, 1980; Palmer and Bucher, 1981; Delis, Robertson, and Efron, 1986; Robertson and Delis, 1986; Delis, Kiefner, and Fridlund, 1988). It thus involves the ability both to segment a pattern into a set of constituent parts (referred to as featural or local level processing), and to integrate those parts into a coherent whole (configural or global level processing). Studies of visuospatial processing within the ventral stream have shown systematic differences in the distribution of global and local level processing across the cerebral hemispheres. Specifically, right posterior temporal regions are dominant for global processing, and left posterior temporal regions dominate local processing. Studies of adult patients have shown that left posterior or right posterior focal brain injury results in dissociable disorders of spatial analytic functioning (e.g., McFie and Zangwill, 1960; Gainotti and Tiacci, 1970; Arena and Gainotti, 1978; Delis, Robertson, and Efron, 1986; Robertson and Delis, 1986; Delis, Kiefner, and Fridlund, 1988; Robertson, Lamb, and Knight, 1988; Swindell et al., 1988; Lamb, Robertson, and Knight, 1989, 1990; Ivry and Robertson, 1998). Left posterior injury disrupts local level processing and results in disorders involving difficulty defining the parts of a spatial array. For example, in drawing, patients tend to oversimplify spatial patterns and omit details, while on perceptual judgment tasks, they rely upon overall configural cues and ignore specific elements. By contrast, patients with right posterior lesions have difficulty with global level processing that disrupts the configural aspects of spatial analysis. In drawing, they include details but fail to maintain a coherent organization among the elements. In perceptual judgment tasks, they focus on the parts of the pattern without attending to the overall form. A large number of RT studies with normal adults have confirmed the lateralization of local level processing to the left hemisphere (LH), and global level processing to the right hemisphere (RH), particularly when stimulus and task parameters are well controlled (e.g., Martin, 1979; Sergent, 1982; Martinez, Moses, et al., 1997; Yovel, Levy, and Yovel, 2001; Han et al., 2002; Volberg and Hubner, 2004). Sergent has suggested that hemispheres differ in terms of higherorder perceptual processes that differentially emphasized processing of lower spatial frequencies in the RH and higher spatial frequencies in the LH. In accordance with this view, a number of experiments have presented sinusoidal gratings containing a single spatial frequency presented to the RVF or LVF. Low spatial frequencies elicit faster responses when presented to the LVF-RH than the RVF-LH, while high spatial frequencies elicit the opposite pattern (Kitterle, Christman, and Hellige, 1990; Kitterle and Selig, 1991;

Kitterle, Hellige, and Christman, 1992). This pattern of hemispheric lateralization as a function of attention to spatial frequency has also been confirmed in event-related potential (ERP) studies (Martinez, Anllo-Vento, and Hillyard, 1997, but also see Boeschoten et al., 2005). Lateralized differences in inferior temporal lobe activation during global and local processing have been examined using a variety of functional brain-imaging techniques including PET, fMRI, and ERP (e.g., Fink et al., 1997; Martinez, Moses, et al., 1997; Heinze et al., 1998). The different experimental methods have yielded generally consistent evidence of lateralized differences in brain activation during global versus local level processing. An fMRI study by Martinez, Moses, and colleagues (1997) illustrates the characteristic pattern of lateralized activation within inferior temporal lobes on tasks using a standard type of stimulus, the hierarchical stimulus, originally developed by Navon (1977). Hierarachical stimuli are large forms composed of smaller elements (e.g., a large global square made of small local circles). Martinez used RT and fMRI to measure performance when subjects attended selectively to either the global or the local level of the hierarchical stimuli. On both measures, a clear advantage for global targets was associated with RH processing, and a significant, but less robust, LH advantage for local targets was observed. These findings are consistent with other studies examining global-local processing within the ventral stream. Studies of typically developing children have shown that even young infants are capable of spatial pattern analysis. Evidence for both configural preferences and rudimentary part-whole processing has been reported in studies of newborns (Slater et al., 1991; Quinn, Burke, and Rush, 1993; Farroni et al., 2000; Cassia et al., 2002). Across the first year of life, systematic change is seen in the complexity of visual pattern processing that reflects the engagement and development of global and local level processing (L. Cohen and Younger, 1984). Further, these patterns of change appear to reflect early hemispheric differences in processing. Deruelle and de Schonen (1991, 1995) have shown that infants as young as 4 months show a profile of lateralized processing differences on global and local processing tasks that is similar to that observed in adult neuroimaging studies. Change in the complexity and sophistication of spatial analytic processing has also been documented across the preschool and school-age periods. Data from a large series of studies using different measures and testing children ranging in age from 3 to 12 years show that initially children segment out well-formed, independent parts and use simple combinatorial rules to integrate the parts into the overall configuration (Akshoomoff and Stiles, 1995a, 1995b; Feeney and Stiles, 1996; Tada and Stiles, 1996; Stiles and Stern, 2001). With development, change is observed in both the nature of the parts and the relations children use to organize

the parts. Studies using the standard hierarchical form stimuli have consistently documented a protracted period of developmental change in global-local processing that extends well into adolescence (Dukette and Stiles, 1996, 2001; Moses et al., 2002; Mondloch et al., 2003; Porporino et al., 2004). However, while there is strong evidence that both global and local level processing change with development, evidence documenting the relative rates of developmental change for these two aspects of visual pattern processing is less clear. Most adult studies report an RH advantage for processing global level stimuli. Further, within the adult literature there is strong evidence of a global-local processing asymmetry. Specifically, inconsistent or competing information at the global level interferes with local processing, but inconsistent local information does not affect global processing. These two findings led to the postulation of a global precedence effect in visual pattern processing (Navon, 1977), suggesting that global level information is processed prior to local level information. While many factors have been shown to mitigate the global precedence effect in adults, it remains a robust finding within the standard task. An advantage for global processing has also been consistently reported in developmental studies across the period from infancy to adolescence (Cassia et al., 2002; Moses et al., 2002; Mondloch et al., 2003; Porporino et al., 2004). However, the findings from the child data do not map simply onto the adult data, and findings about the relative dominance of global versus local processing are somewhat mixed. For example, Mondloch and colleagues (2003) has reported a stronger global processing advantage among school-age children than adults; those differences gradually decline as children reach adolescence, but the global advantage is never completely lost. Moses and colleagues (2002) reported similar findings of a global level advantage for their “less mature” child group in both their RT and fMRI data. Evidence of an enhanced global advantage among children suggests an earlier developmental trajectory for global than for local level processing. By contrast, studies that manipulate filtering and interference during global-local processing tasks report greater effects on global than local level processing for children, but not adults (Porporino et al., 2004). Such findings are inconsistent with the global precedence effects usually observed among adults and suggest a more protracted developmental trajectory for global level processing. Further, selective degradation of the global level stimulus induces a shift in processing bias from the global to the local level in preschool-age children, but not in older children or adults (Dukette and Stiles, 1996). The combined data from the studies of hierarchical form processing show that children are clearly able to engage in both global and local level processing from a very early age, and there is evidence that under standard task conditions, children may have an

stiles, paul, and ark: development of visuospatial processing

529

exaggerated global processing advantage compared to adults. However, it is also clear that stable and mature levels of visuospatial processing emerge very slowly, and that for a protracted period of development, variations in stimulus and task demands play a significantly greater role in modulating visual pattern processing than is observed later in life. Thus the functional role of a global or local processing bias or advantage may be very different during development than it is later in life. Imaging studies of typical children confirm the behavioral findings and suggest that the neural systems associated with spatial analytic processing also undergo a protracted period of development. Using fMRI, Moses and associates (2002) tested children between 11 and 15 years old using a hemifield RT task and fMRI protocols identical to those used by Martinez, Moses, and colleagues (1997) with adult subjects. There were two major findings of this study. First, the pattern of RT data obtained from children across the age span tested in this study differed from that of adults. Similar to the findings from the Mondloch and colleagues (2003) study, across all age groups, children were faster with global than with local targets, and children, as a group, did not manifest the kinds of hemifield RT differences observed among adults. Second, in the fMRI study, children’s profiles of activation differed from the patterns observed among adults. For both the global and local tasks, children showed statistically greater activation in the RH than in the LH. However, overall activation among children was greater than for adults, and children showed considerably more bilateral activation, particularly on the local processing tasks, than did adults. These studies suggest that, at least for these perceptually demanding tasks, children show both a global processing advantage and overall RH dominance. An extensive series of studies of children with pre- and perinatal focal brain injury have provided detailed profiles of deficit and recovery for spatial functioning associated with early lateralized brain injury (Stiles-Davis et al., 1988; Stiles et al., 1997, 1998; Vicari et al., 1998; Stiles et al., 2005; Stiles, Paul, and Hesselink, 2006; Stiles et al., 2008). These studies have shown that on construction and perception tasks, children with RH injury have difficulty with spatial integration. While they are able to segment a spatial form into its elements, they have difficulty organizing those elements to form a coherent whole. Children with LH injury show a very different pattern of deficit in that they oversimplify complex spatial forms and fail to encode the details or elements of these forms. These two profiles are consistent with patterns of deficit reported for adults with injury to comparable brain regions and suggest that the basic lateralized dissociation for global and local processing is robust from very early in development. However, it is also important to note that while basic patterns of deficit are evident even among very young children, the severity of deficit is

530

cognition

Figure 32.2 Children with either LH or RH injury were asked to reproduce the model hierarchical forms from memory. LH injury results in difficulty reproducing the parts, or local elements, of the forms; while RH injury results in greater difficulty reproducing the larger, or global, configuration.

greatly attenuated compared to adults, and children appear to compensate in ways that adults cannot. Two studies illustrate the systematic effects of early lateralized injury on visual pattern processing. A memory reproduction version of the hierarchical forms task was used to test for global-local processing deficits in school-age children with early brain injury. In this task, which was adapted for children from a study of adult stroke patients (Delis, Kiefner, and Fridlund, 1988), children were shown hierarchical stimuli, one at a time, and were told that they would be asked to reproduce them from memory. The accuracy of their global and local level reproductions was assessed. Among children with RH injury, accuracy scores for the local level were comparable to those of ageand IQ-matched control children, but their global level accuracy was significantly lower. The children with LH injury showed the reverse pattern, thus providing evidence for a dissociation in global-local processing deficits associated with early lateralized brain injury (see figure 32.2). Further, data from both the large cross-sectional sample of 5- to 12-year-olds and a smaller longitudinal cohort confirmed the persistence of subtle processing deficits lasting into early adolescence (Stiles et al., 2008). A parallel set of findings was observed among school-age children using a more difficult reconstruction task, the Rey Osterrieth Complex Figure (ROCF). The ROCF is a complex geometric pattern that has been used for many years to evaluate spatial planning in adult neurological patients. This figure is organized around a central rectangle that is symmetrically divided by vertical, horizontal, and diagonal bisecting lines; additional pattern details are positioned both within and surrounding the core rectangle. The most efficient and advanced strategy for copying the ROCF

is to begin with the core rectangle and bisectors and then add pattern details. However, this strategy also places great demands on spatial analytic processing. Akshoomoff and Stiles (1995a, 1995b) have shown that normally developing children do not regularly use this advanced copying strategy until quite late in development. Six-year-olds typically use a very piecemeal strategy, drawing each small subdivision separately; in the middle-school age period children produce progressively larger subunits (quadrants, halves); and by about 10 to 12 years, children’s organization strategy centers around the core rectangle. Akshoomoff and associates (2002) and Akshoomoff and Stiles (2003) examined data collected longitudinally from 6- to 13-year-old children with prenatal brain injury. Two versions of the task were administered—a copy task, in which the children were asked to reproduce the figure with the model present, and an immediate memory task. Deficits on the copying task were particularly evident among the youngest children (6–7 years of age) in that their drawings were sparser and less accurate than normal controls. However, there were no differences between LH and RH injury groups. With development, performance improved considerably, such that by 9 to 10 years of age these children were able to produce accurate copies of the ROCF. However, across the 6- to 12-year age period, children in both lesion groups continued to use the most immature and piecemeal production strategy. Thus, while accuracy improved, continuing deficit was evident in their construction procedures. The failure to find differences between the RH and LH injury groups on the ROCF task could reflect underlying task demands that place equal emphasis on the segmentation and integrative processes, and thus equally disrupt performance on the task. Interestingly, while the copy task failed to differentiate the lesion groups, the memory task from the oldest children in the sample (ages 11 to 14 years) did. Among RH children, performance on the copy and memory tasks did not differ; for both tasks children used a piecemeal strategy to produce relatively accurate reproductions. Although a similar pattern of data was observed among LH children on the copy task, their memory task performance diverged significantly. Specifically, the children organized their memory reproductions around the core rectangle but were able to produce relatively few additional pattern details. In the absence of a model, the children adopted a very different and more advanced construction strategy focused on the global configuration that was apparently encoded in their memory representation. But the accuracy of their memory reproductions, in particular their failure to produce the features of the pattern, also reflected their underlying local processing deficit. In summary, the profiles of visual pattern processing deficit for children with LH and RH injury are quite distinct and consistent with profiles of deficit observed among adults.

Longitudinal data indicate that with development, children with both LH and RH injury show considerable behavioral improvement, eventually achieving near ceiling-level performance on most spatial construction tasks. However, the time course over which this improvement occurs is protracted, and in cases where we have been able to examine the underlying processes associated with recovery, anomalous processing profiles have emerged (Stiles, Paul, and Hesselink, 2006). These data suggest that the neural system for visual pattern processing is specified early in life, but it is immature and underspecified, and it undergoes considerable change with development.

Trajectories of dorsal and ventral stream development While the dorsal and ventral visual pathways have been extensively studied in adults, developmental inquiries have yet to produce an adequate response to the need for an account of how development proceeds toward the adult state of the two dissociated systems. Although there is considerable work examining development of ventral-stream functions or dorsal-stream functions separately, there has been very little systematic work comparing developmental trajectories for these systems. Infant studies exploring the visual tracking of objects in the local environment provide some of the earliest evidence for a dissociation of ventral- and dorsal-stream processing. Prior to 9 to 10 months, infants appear to use spatiotemporal information (e.g., motion or location) rather than information about the features (e.g., pattern or color) to differentiate objects (Xu and Carey, 1996; Van de Walle, Carey, and Prevor, 2000). Although some evidence exists for earlier utilization of certain types of featural information in the context of a less demanding task (Wilcox and Baillargeon, 1998; Wilcox, 1999), these studies have not made direct comparisons with spatiotemporal information (Krojgaard, 2004). Direct comparisons have converged on the primacy of spatiotemporal characteristics (Krojgaard, 2004), implicating the dorsal stream as the source of the information with which infants first discriminate objects (Wilcox, 1999). It is only toward the end of the first year of life that ventral stream information, such as color, is incorporated into object processing (Leslie et al., 1998; Kaldy and Leslie, 2003). Only a very small number of studies have explicitly compared dorsal- and ventral-stream processing in older children. A small corpus of psychophysical investigations suggests that even basic visual functions supported by the parvocellular (P) and magnocellular (M) pathways (which can be mapped onto the ventral and dorsal streams, respectively) show evidence of immaturity in school-age children (e.g., Parrish et al., 2005). However, the results of these studies remain equivocal with regard to the progression to

stiles, paul, and ark: development of visuospatial processing

531

the adultlike state. Reports of form versus motion coherence (Gunn et al., 2002; Atkinson et al., 2003) or color versus motion processing (Mitchell and Neville, 2004; Coch et al., 2005) suggest that the dorsal stream may be slower to reach maturity. By contrast, visual evoked potential (VEP) studies varying in spatial frequency (Gordon and McCulloch, 1999) and chromaticity (Madrid and Crognale, 2000) have found evidence of a lag in ventral stream functioning. Behavioral studies using the dual-stream framework, comparing “what” versus “where” (Ungerleider and Mishkin, 1982) or “what” versus “how” (Milner and Goodale, 1995), have rarely been conducted with older children. Atkinson (1998) has gathered a small amount of data from children ages 4 to 7 years using Milner and Goodale’s (1995) “postbox” task, which requires manual posting of a letter into a slot at a particular angle (dorsal) or visual matching of the perceived angle of the slot (ventral). Although her participants performed more accurately on the visual matching task, adult performance was not available for comparison. A limited amount of additional, relevant data from school-age children have come from developmental studies of visuospatial working memory. It has been proposed that the “visuospatial sketchpad” (Baddeley, 1986), the limitedcapacity storage mechanism for visuospatial (versus verbal) information, can be segregated into separate components dedicated to visual (a “visual cache”) and spatial (an “inner scribe”) material (Logie, 1995). Child studies have looked for a dissociation in the development of these two components of visuospatial working memory. Although it can be assumed that these two components of the visuospatial sketchpad are at least partly related to the dorsal and ventral visual streams, there is unlikely to be a one-to-one mapping, as the relationship between developmental age and working-memory ability is undoubtedly influenced by other factors such as the ability to verbally recode visual stimuli (Hitch et al., 1988), use of active rehearsal (Gathercole, Adams, and Hitch, 1994), and general executive processes (Gathercole et al., 2004). Consistent with this suggestion, existing studies of pattern/object versus spatial memory have produced conflicting results. Whereas several groups have provided evidence suggesting that object-based memory undergoes more rapid development (Siemens, Guttentag, and McIntyre, 1989; Logie and Pearson, 1997; Gulya et al., 2002; Hamilton, Coates, and Heffernan, 2003; Lorsbach and Reimer, 2005), others have found that children encode spatial information more efficiently (Finkel, 1973) and that memory for this type of information is less susceptible to interference (Lange-Küttner and Friederici, 2000). Comparisons of the maturational time courses of the dorsal and ventral streams have been hindered by methodological limitations, since many studies have not controlled

532

cognition

for cross-task difficulty or have utilized stimuli known to elicit verbal strategies (e.g., common objects). In a recent study, Paul, Passaroti, and Stiles (2001) attempted to address these issues. They conducted a large-scale RT study using ventral and dorsal stream tasks that were matched a priori for stimuli, procedures, and required response (also see E. Smith et al., 1995). Critically the tasks were designed to yield comparable RT and accuracy performance among adults. Tasks required visual matching of identity or location for a series of three face stimuli (two reference stimuli followed by one test stimulus). Results from children ages 6 to 12 years and adults showed that performance on the two tasks was comparable at each age, while performance improved continuously across the school-age period. Thus the findings from this study document a protracted period of change for relatively basic perceptual matching abilities that rely on the dorsal and ventral visual streams, but no strong dissociation for the developmental trajectories of the two systems. Although no clear evidence for a ventral-dorsal dissociation was found with typically developing children on the identity- and location-matching tasks, data from clinical populations on the same matching task has yielded very strong dissociations. The task was used with a group of 8- to 24-year-olds with pre- or perinatal focal brain injury (Paul et al., in preparation). Overall, the RH and LH groups were less accurate and slower to respond than age- and IQmatched controls. However, the performance of the patient groups also varied according to the task. On the identitymatching task, RH-injured participants were significantly less accurate and slower than controls, while those with LH injury were not significantly affected. On the locationmatching task, both patient groups performed more poorly, although the effects were not as striking as with identity matching. These results imply that the functioning of both streams can be affected by early injury, but that the nature of these effects may depend on differential involvement of the cerebral hemispheres. The most conspicuous effects in this study seemed to be that of RH injury on ventral-stream function. This finding is consistent with imaging studies in adults showing a general RH dominance for facial identity processing, and it suggests that these biases may emerge very early in life. Studying children and adults with neurodevelopmental disorders may also shed light on questions of differential vulnerability and plasticity of the two visual streams. In particular, studies of individuals with Williams syndrome (WS, a rare genetic disorder characterized by distinctive dysmorphologic facial features, mild to moderate mental retardation, distinctive personality characteristics, and strikingly poor visuocognitive ability) have provided valuable information in this regard. This is demonstrated in a third study of identity- and location-matching (Paul et al., 2002).

Performance of adult WS patients was compared to both age-matched adults and mental-age-matched children. Results suggested that while both typical adults and children performed comparably on the two tasks, individuals with WS were significantly impaired on location- but not identitymatching. In light of the observed “hypersociability” in WS (Jones et al., 2001), an additional location-matching control condition was included, in which the intact face stimuli were replaced with scrambled face stimuli. Results from this condition provided evidence that the WS impairment was not due to a failure to engage in the task secondary to being distracted by the faces. Instead, the deficit appears to be consistent with other rapidly emerging neurobiological, neuroanatomical, and cognitive evidence of greater dorsal-stream involvement in this disorder (Atkinson et al., 1997; Galaburda et al., 2002; Meyer-Lindenberg et al., 2004; Eckert et al., 2006) and in other developmental conditions (Spencer et al., 2000; Stein, Talcott, and Walsh, 2000).

Conclusions This chapter reviewed findings from a range of studies focused on different aspects of visuospatial processing. The specific goal was to examine patterns of association between behavioral and neural development. The review was not intended to be a comprehensive survey of all facets of spatial information processing. Rather, it was designed to focus on several major visuospatial processing systems within both the dorsal and ventral visual pathways, and to use data on the anatomical organization and behavioral functioning of those systems at different points in development to try to understand how basic neural systems come to mediate specific functions. Several general points emerge from these data. First, the basic neural systems mediating spatial functions appear to be specified quite early in development. Infants are able to track and retrieve hidden objects by the middle of the first year of life. Basic markers for control of spatial attention can be documented by 4 months and may be available earlier. Dissociable patterns of spatial analytic deficit can be documented in children with pre- or perinatal brain injury. All these data are indicative of neural systems that are specified for processing certain types of information. However, the nature and degree of that specification appears to be constrained. It is these constraints that lead to discussion of the remaining points. Early specification of the neural system does not imply full or optimal functioning early in development. The idea of partial cortical functioning is a crucial one here, and it should be coupled with the idea of partial behavioral functioning (also see Haith and Benson, 1998). Data from the spatial localization tasks make it clear that the neural and

behavioral systems for spatial localization are not discretely absent or present at a given point in early development. For example, children who fail the standard hiding task solve it easily when a landmark is provided. Thus, at some level, the behavioral and neural systems for spatial localization must be available, but they are not functioning effectively enough to support the demands of the full range of task variation. A second example of limited functioning comes from the data on the early development of spatial attention. Data from normal infants suggest that, while the basic behavioral abilities to shift attention may be available early on, they are only detectable under certain temporal and spatial conditions, and those conditions change with development. Further, data from the early-lesion population suggest that neural mediation of apparently adultlike attentional processes may be mediated by different neural systems. Thus the key to the notion of partial behavioral and neural mediation is to understand what the specific task demands are and how they interact with and engage the available neural processing resources. Finally, early specification may not determine final organization. There are large animal and human literatures on the concept of early brain plasticity (Stiles et al., 1998; Nelson, 1999; Stiles, 2000; Huttenlocher, 2002) which demonstrate quite clearly that (1) normal profiles of brain development require specific kinds of input, (2) variation in input can affect the pattern of neural organization, and (3) injury to the neural substrate can induce alternative patterns of neural organization. Within the spatial domain it is clear that early injury results in persistent deficits. Findings from both the spatial attention and spatial analysis studies demonstrate that years after insult, specific deficits can be detected in a population of children with pre- and perinatal brain injury. However, the deficits are subtler than those observed in adults with comparable injury. These findings are consistent with a view of early brain development in which both early specification and plastic adaptation play prominent roles. Early injury constitutes a perturbation of normal development. Specific neural resources are lost, and there should be consequent impairment of the system, and that is precisely what is observed. However, it is also a developing system and, therefore, a system with an exuberance of resources, the fate of which is determined in large measure by input. Thus the magnitude and duration of the initial impairment may well depend on a range of factors such as the timing of insult, extent and location of injury, and specificity of the neural substrate. acknowledgments

This work was supported by the National Institute of Child Health and Human Development Grants R01-HD25077 and R01-HD041581, National Institute of Neurological Disorders and Stroke Grant P50-NS22343, and National Institute of Deafness and

stiles, paul, and ark: development of visuospatial processing

533

Communicative Disorders Grant P50-DC01289. The authors wish to thank the parents and children for their participation in the studies presented in this article. REFERENCES Ahmed, A., and T. Ruffman, 1998. Why do infants make A not B errors in a search task, yet show memory for the location of hidden objects in a nonsearch task? Dev. Psychol. 34:441– 453. Akshoomoff, N. A., C. C. Feroleto, R. E. Doyle, and J. Stiles, 2002. The impact of early unilateral brain injury on perceptual organization and visual memory. Neuropsychologia 40:539–561. Akshoomoff, N. A., and J. Stiles, 1995a. Developmental trends in visuospatial analysis and planning. I. Copying a complex figure. Neuropsychology 9:364–377. Akshoomoff, N. A., and J. Stiles, 1995b. Developmental trends in visuospatial analysis and planning. II. Memory for a complex figure. Neuropsychology 9:378–389. Akshoomoff, N. A., and J. Stiles, 2003. Children’s performance on the ROCF and the development of spatial analysis. In J. A. Knight and E. Kaplan, eds., The Handbook of Rey-Osterrieth Complex Figure Usage: Clinical and Research Applications, 393–409. Lutz, FL: Psychological Assessment Resources. Andersen, R. A., 1988. The neurobiological basis of spatial cognition: Role of the parietal lobe. In J. Stiles-Davis, M. Kritchevsky, and U. Bellugi, eds., Spatial Cognition: Brain Bases and Development, 57–80. Hillsdale, NJ: Lawrence Erlbaum. Andersen, R. A., L. H. Snyder, D. C. Bradley, and J. Xing, 1997. Multimodal representation of space in the posterior parietal cortex and its use in planning movements. Annu. Rev. Neurosci. 20:303–330. Arena, R., and G. Gainotti, 1978. Constructional apraxia and visuoperceptive disabilities in relation to laterality of cerebral lesions. Cortex 14:463–473. Ark, W., F. Haist, and J. Stiles, in preparation. Developmental Performance of Mental Rotation: An FMRI Study. Atkinson, J., 1998. The “where and what” or “who and how” of visual development. In F. Simion and G. Butterworth, eds., The Development of Sensory, Motor and Cognitive Capacities in Early Infancy: From Perception to Cognition, 3–24. Hove, UK: Psychology Press/ Erlbaum (UK) Taylor & Francis. Atkinson, J., O. Braddick, S. Anker, W. Curran, R. Andrew, J. Wattam-Bell, and F. Braddick, 2003. Neurobiological models of visuospatial cognition in children with Williams syndrome: Measures of dorsal-stream and frontal function. Dev. Neuropsychol. (Special Issue: Williams Syndrome) 23:139–172. Atkinson, J., J. King, O. Braddick, L. Nokes, S. Anker, and F. Braddick, 1997. A specific deficit of dorsal stream function in Williams’ syndrome. NeuroReport 8:1919–1922. Baddeley, A., 1986. Working Memory. New York: Clarendon Press/ Oxford University Press. Baillargeon, R., and J. DeVos, 1991. Object permanence in young infants: Further evidence. Child Dev. 62:1227–1246. Baillargeon, R., and M. Graber, 1988. Evidence of location memory in 8-month-old infants in a nonsearch AB task. Dev. Psychol. 24:502–511. Baird, A. A., J. Kagan, T. Gaudette, K. A. Walz, N. Hershlag, and D. A. Boas, 2002. Frontal lobe activation during object permanence: Data from near-infrared spectroscopy. NeuroImage 16:1120–1125.

534

cognition

Belger, A., A. Puce, J. H. Krystal, J. C. Gore, P. GoldmanRakic, and G. McCarthy, 1998. Dissociation of mnemonic and perceptual processes during spatial and nonspatial working memory using fMRI. Hum. Brain Mapping 6:14– 32. Bell, M. A., and N. A. Fox, 1992. The relations between frontal brain electrical activity and cognitive development during infancy. Child Dev. 63:1142–1163. Bell, S., 2002. Spatial cognition and scale: A child’s perspective. J. Environ. Psychol. (Special Issue: Children and the Environment) 22:9–27. Bertenthal, B. I., and J. J. Campos, 1990. A systems approach to the organizing effects of self-produced locomotion during infancy. Adv. Infancy Res. 6:1–60. Boeschoten, M. A., C. Kemner, T. L. Kenenmans, and H. Engeland, 2005. The relationship between local and global processing of high and low spatial frequencies studied by eventrelated potentials and source modeling. Brain Res. Cogn. Brain Res. 24:228–336. Booth, J. R., B. Macwhinney, K. R. Thulborn, K. Sacco, J. Voyvodic, and H. M. Feldman, 1999. Functional organization of activation patterns in children: Whole brain fMRI imaging during three different cognitive tasks. Prog. Neuropsychopharmacol. Biol. Psychiatry 23:669–682. Booth, J. R., B. MacWhinney, K. R. Thulborn, K. Sacco, J. T. Voyvodic, and H. M. Feldman, 2000. Developmental and lesion effects in brain activation during sentence comprehension and mental rotation. Dev. Neuropsychol. 18:139– 169. Boussaoud, D., R. Desimone, and L. G. Ungerleider, 1991. Visual topography of area TEO in the macaque. J. Comp. Neurol. 306:554–575. Brodmann, K., 1909. Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien Dargestellt auf Grund des Zellenbaues. Leipzig, Germany: J. A. Barth. Butterworth, G., N. Jarrett, and L. Hicks, 1982. Spatiotemporal identity in infancy: Perceptual competence or conceptual deficit? Dev. Psychol. 18:435–449. Cai, H., and C. Chen, 2000. The development of mental rotation ability and its correlationship with intelligence. Psychol. Sci. (China) 23:363–365. Casey, B. J., J. D. Cohen, K. O'Craven, R. J. Davidson, W. Irwin, C. A. Nelson, D. C. Noll, X. Hu, M. J. Lowe, B. R. Rosen, C. L. Truwitt, and P. A. Turski, 1998. Reproducibility of fMRI results across four institutions using a spatial working memory task. NeuroImage 8:249–261. Cassia, V. M., F. Simion, I. Milani, and C. Umilta, 2002. Dominance of global visual properties at birth. J. Exp. Psychol. [Gen.] 131:398–411. Chiba, A. A., R. P. Kesner, and C. J. Gibson, 1997. Memory for temporal order of new and familiar spatial location sequences: Role of the medial prefrontal cortex. Learn. Memory 4:311– 317. Chiba, A. A., R. P. Kesner, and P. A. Jackson, 2002. Two forms of spatial memory: A double dissociation between the parietal cortex and the hippocampus in the rat. Behav. Neurosci. 116:874–883. Chugani, H. T., M. E. Phelps, and J. C. Mazziotta, 1987. Positron emission tomography study of human brain functional development. Ann. Neurol. 22:487–497. Clohessy, A. B., M. I. Posner, M. K. Rothbart, and S. P. Vecera, 1991. The development of inhibition of return in early infancy. J. Cogn. Neurosci. 3:345–350.

Coch, D., W. Skendzel, G. Grossi, and H. Neville, 2005. Motion and color processing in school-age children and adults: An ERP study. Dev. Sci. 8:372–386. Cohen, L. B., and B. A. Younger, 1984. Infant perception of angular relations. Infant Behav. Dev. 7:37–47. Cohen, M. S., S. M. Kosslyn, H. C. Breiter, G. J. DiGirolamo, W. L. Thompson, A. K. Anderson, S. Y. Brookheimer, B. R. Rosen, and J. W. Belliveau, 1996. Changes in cortical activity during mental rotation: A mapping study using functional MRI. Brain 119(Pt. 1):89–100. Colby, C. L., and J. R. Duhamel, 1996. Spatial representations for action in parietal cortex. Brain Res. Cogn. Brain Res. 5:105– 115. Colby, C. L., and M. E. Goldberg, 1999. Space and attention in parietal cortex. Annu. Rev. Neurosci. 22:319–349. Colombo, J., 2001. The development of visual attention in infancy. Annu. Rev. Psychol. 52:337–367. Corbetta, M., 1998. Frontoparietal cortical networks for directing attention and the eye to visual locations: Identical, independent, or overlapping neural systems? Proc. Natl. Acad. Sci. USA 95:831– 838. Corbetta, M., F. M. Miezin, G. L. Shulman, and S. E. Petersen, 1993. A PET study of visuospatial attention. J. Neurosci. 13:1202–1226. Courtney, S. M., L. Petit, J. M. Maisog, L. G. Ungerleider, and J. V. Haxby, 1998. An area specialized for spatial working memory in human frontal cortex. Science 279:1347–1351. Courtney, S. M., L. G. Ungerleider, K. Keil, and J. V. Haxby, 1996. Object and spatial visual working memory activate separate neural systems in human cortex. Cerebral Cortex 6:39–49. Craft, S., D. A. White, T. S. Park, and G. Figiel, 1994. Visual attention in children with perinatal brain injury: Asymmetric effects of bilateral lesions. J. Cogn. Neurosci. 6:165–173. Csibra, G., G. Davis, M. W. Spratling, and M. H. Johnson, 2000. Gamma oscillations and object processing in the infant brain. Science 290:1582–1585. Curtis, C. E., 2006. Prefrontal and parietal contributions to spatial working memory. Neuroscience 139:173–180. De Lisi, R., and J. L. Wolford, 2002. Improving children’s mental rotation accuracy with computer game playing. J. Genet. Psychol. 163:272–282. Delis, D. C., M. G. Kiefner, and A. J. Fridlund, 1988. Visuospatial dysfunction following unilateral brain damage: Dissociations in hierarchical hemispatial analysis. J. Clin. Exp. Neuropsychol. 10:421–431. Delis, D. C., L. C. Robertson, and R. Efron, 1986. Hemispheric specialization of memory for visual hierarchical stimuli. Neuropsychologia 24:205–214. Deruelle, C., and S. de Schonen, 1991. Hemispheric asymmetries in visual pattern processing in infancy. Brain Cogn. 16:151– 179. Deruelle, C., and S. de Schonen, 1995. Pattern processing in infancy: Hemispheric differences in the processing of shape and location of visual components. Infant Behav. Dev. 18:123–132. Diamond, A., 1985. Development of the ability to use recall to guide action, as indicated by infants’ performance on AB. Child Dev. 56:868–883. Diamond, A., 1991. Frontal lobe involvement in cognitive changes during the first year of life. In K. R. Gibson, A. C. Petersen, et al., eds., Brain Maturation and Cognitive Development: Comparative and Cross-Cultural Perspectives, 127–180. New York: Aldine de Gruyter.

Diamond, A., J. F. Werker, and C. Lalonde, 1994. Toward understanding commonalities in the development of object search, detour navigation, categorization, and speech perception. In G. Dawson, K. W. Fischer, et al., eds., Human Behavior and the Developing Brain, 380–426. New York: Guilford Press. Diamond, A., S. Zola-Morgan, and L. R. Squire, 1989. Successful performance by monkeys with lesions of the hippocampal formation on AB and object retrieval, two tasks that mark developmental changes in human infants. Behav. Neurosci. 103:526– 537. Dobkins, K. R., and T. D. Albright, 1994. What happens if it changes color when it moves? The nature of chromatic input to macaque visual area MT. J. Neurosci. 14:4854–4870. Dobkins, K. R., and T. D. Albright, 1995. Behavioral and neural effects of chromatic isoluminance in the primate visual motion system. Visual Neurosci. 12:321–332. Dobkins, K. R., and T. D. Albright, 1998. The influence of chromatic information on visual motion processing in the primate visual system. In T. Watanabe et al., ed., High-Level Motion Processing: Computational, Neurobiological, and Psychophysical Perspectives, 53–94. Cambridge, MA: MIT Press. Dobkins, K. R., and C. M. Anderson, 2002. Color-based motion processing is stronger in infants than in adults. Psychol. Sci. 13:76–80. Dobkins, K. R., and D. Y. Teller, 1996a. Infant contrast detectors are selective for direction of motion. Vis. Res. 36:281–294. Dobkins, K. R., and D. Y. Teller, 1996b. Infant motion: Detection (M:D) ratios for chromatically defined and luminancedefined moving stimuli. Vis. Res. 36:3293–3310. Dukette, D., and J. Stiles, 1996. Children’s analysis of hierarchical patterns: Evidence from a similarity judgment task. J. Exp. Child Psychol. 63:103–140. Dukette, D., and J. Stiles, 2001. The effects of stimulus density on children’s analysis of hierarchical patterns. Dev. Sci. 4:233– 251. Eckert, M. A., A. M. Galaburda, D. L. Mills, U. Bellugi, J. R. Korenberg, and A. L. Reiss, 2006. The neurobiology of Williams syndrome: Cascading influences of visual system impairment? Cell. Mol. Life Sci. 63:1867–1875. Ellison, A., and A. Cowey, 2006. TMS can reveal contrasting functions of the dorsal and ventral visual processing streams. Exp. Brain Res. 175:618–625. Estes, D., 1998. Young children’s awareness of their mental activity: The case of mental rotation. Child Dev. 69:1345–1360. Farrell Pagulayan, K., R. M. Busch, K. L. Medina, J. A. Bartok, and R. Krikorian, 2006. Developmental normative data for the Corsi block-tapping task. J. Clin. Exp. Neuropsychol. 28:1043–1052. Farroni, T., E. Valenza, F. Simion, and C. Umilta, 2000. Configural processing at birth: Evidence for perceptual organisation. Perception 29:355–372. Feeney, S. M., and J. Stiles, 1996. Spatial analysis: An examination of preschoolers’ perception and construction of geometric patterns. Dev. Psychol. 32:933–941. Fink, G. R., P. W. Halligan, J. C. Marshall, C. D. Frith, R. S. J. Frackowiak, and R. J. Dolan, 1997. Neural mechanisms involved in the processing of global and local aspects of hierarchically organized visual stimuli. Brain 120:1779–1791. Finkel, D. L., 1973. A developmental comparison of the processing of two types of visual information. J. Exp. Child Psychol. 16: 250–266. Fox, N. A., and M. A. Bell, 1990. Electrophysiological indices of frontal lobe development: Relations to cognitive and affective

stiles, paul, and ark: development of visuospatial processing

535

behavior in human infants over the first year of life. Ann. NY Acad. Sci. 608:677–704. Gainotti, G., and C. Tiacci, 1970. Patterns of drawing disability in right and left hemispheric patients. Neuropsychologia 8:379– 384. Galaburda, A. M., D. P. Holinger, U. Bellugi, and G. F. Sherman, 2002. Williams syndrome: Neuronal size and neuronal-packing density in primary visual cortex. Arch. Neurol. 59: 1461–1467. Gathercole, S. E., A.-M. Adams, and G. J. Hitch, 1994. Do young children rehearse? An individual-differences analysis. Mem. Cogn. 22:201–207. Gathercole, S. E., S. J. Pickering, B. Ambridge, and H. Wearing, 2004. The structure of working memory from 4 to 15 years of age. Dev. Psychol. 40:177–190. Goldman-Rakic, P. S., 1987. Development of cortical circuitry and cognitive function. Child Dev. 58:601–622. Goodale, M. A., and A. D. Milner, 1992. Separate visual pathways for perception and action. Trends Neurosci. 15:20– 25. Gordon, G. E., and D. L. McCulloch, 1999. A VEP investigation of parallel visual pathway development in primary school age children. Doc. Ophthalmol. 99:1–10. Gulya, M., A. Rossi-George, K. Hartshorn, A. Vieira, C. Rovee-Collier, M. K. Johnson, and B. L. Chalfonte, 2002. The development of explicit memory for basic perceptual features. J. Exp. Child Psychol. 81:276–297. Gunn, A., E. Cory, J. Atkinson, O. Braddick, J. Wattam-Bell, A. Guzzetta, and G. Cioni, 2002. Dorsal and ventral stream sensitivity in normal development and hemiplegia. NeuroReport 13:843–847. Hadj-Bouziane, F., M. Meunier, and D. Boussaoud, 2003. Conditional visuo-motor learning in primates: A key role for the basal ganglia. J. Physiol. (Paris) 97:567–579. Haith, M. M., and J. B. Benson, 1998. Infant cognition. In D. Kuhn and R. S. Siegler, eds., Handbook of Child Psychology: Cognition, Perception, and Language, 199–254. New York: J. Wiley. Hale, S., 1990. A global developmental trend in cognitive processing speed. Child Dev. 61:653–663. Hamilton, C. J., R. O. Coates, and T. Heffernan, 2003. What develops in visuo-spatial working memory development? Eur. J. Cogn. Psychol. 15:43–69. Han, S., J. A. Weaver, S. O. Murray, X. Kang, E. W. Yund, and D. L. Woods, 2002. Hemispheric asymmetry in global/local processing: Effects of stimulus position and spatial frequency. NeuroImage 17:1290–1299. Harman, C., M. I. Posner, M. K. Rothbart, and L. ThomasThrapp, 1994. Development of orienting to locations and objects in human infants. Can. J. Exp. Psychol. 48:301–318. Harris, P. L., 1987. Bringing order to the A-not-B error. Monogr. Soc. Res. Child Dev. 51:52–61. Hauser, M. D., 1999. Perseveration, inhibition and the prefrontal cortex: A new look. Curr. Opin. Neurobiol. 9:214–222. Haxby, J. V., C. L. Grady, B. Horwitz, L. G. Ungerleider, M. Mishkin, R. E. Carson, P. Herscovitch, M. B. Schapiro, and S. I. Rapoport, 1991. Dissociation of object and spatial visual processing pathways in human extrastriate cortex. Proc. Natl. Acad. Sci. USA 88:1621–1625. Haxby, J. V., B. Horwitz, L. G. Ungerleider, J. M. Maisog, P. Pietrini, and C. L. Grady, 1994. The functional organization of human extrastriate cortex: A PET-rCBF study of selective attention to faces and locations. J. Neurosci. 14:6336– 6353.

536

cognition

Haxby, J. V., L. Petit, L. G. Ungerleider, and S. M. Courtney, 2000. Distinguishing the functional roles of multiple regions in distributed neural systems for visual working memory. NeuroImage 11:145–156. Heilman, K. M., and E. Valenstein, eds., 1993. Clinical Neuropsychology, 3rd ed. New York: Oxford University Press. Heinze, H. J., H. Hinrichs, M. Scholz, W. Burchert, and G. R. Mangun, 1998. Neural mechanisms of global and local processing: A combined PET and ERP study. J. Cogn. Neurosci. 10:485– 498. Hikosaka, O., K. Nakamura, and H. Nakahara, 2006. Basal ganglia orient eyes to reward. J. Neurophysiol. 95:567–584. Hillyard, S. A., and L. Anllo-Vento, 1998. Event-related brain potentials in the study of visual selective attention. Proc. Natl. Acad. Sci. USA 95:781–787. Hitch, G. J., S. Halliday, A. M. Schaafstal, and J. M. Schraagen, 1988. Visual working memory in young children. Mem. Cogn. 16:120–132. Hofstadter, M., and J. S. Reznick, 1996. Response modality affects human infant delayed-response performance. Child Dev. 67:646–658. Hood, B. M., 1993. Inhibition of return produced by covert shifts of visual attention in 6-month-old infants. Infant Behav. Dev. 16:245–254. Horobin, K., and L. Acredolo, 1986. The role of attentiveness, mobility history, and separation of hiding sites on Stage IV search behavior. J. Exp. Child Psychol. 41:114–127. Huttenlocher, J., N. Newcombe, and E. H. Sandberg, 1994. The coding of spatial location in young children. Cogn. Psych. 27:115–148. Huttenlocher, P. R., 2002. Neural Plasticity: The Effects of Environment on the Development of the Cerebral Cortex. Cambridge, MA: Harvard University Press. Ivry, R. B., and L. C. Robertson, 1998. The Two Sides of Perception. Cambridge, MA: MIT Press. Jacobs, B., H. T. Chugani, V. Allada, S. Chen, M. E. Phelps, D. B. Pollack, and M. J. Raleigh, 1995. Developmental changes in brain metabolism in sedated rhesus macaques and vervet monkeys revealed by positron emission tomography. Cerebral Cortex 5:222–233. Johnson, M. H., D. Mareschal, and G. Csibra, 2001. The functional development and integration of the dorsal and ventral visual pathways: A neurocomputational approach. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 339–351. Cambridge, MA: MIT Press. Johnson, M. H., M. I. Posner, and M. K. Rothbart, 1994. Facilitation of saccades toward a covertly attended location in early infancy. Psychol. Sci. 5:90–93. Johnson, M. H., and L. A. Tucker, 1996. The development and temporal dynamics of spatial orienting in infants. J. Exp. Child Psychol. 63:171–188. Johnson, M. H., L. A. Tucker, J. Stiles, and D. Trauner, 1998. Visual attention in infants with perinatal brain damage: Evidence of the importance of anterior lesions. Dev. Sci. 1:53–58. Johnson, P. B., S. Farraina, L. Bianchi, and R. Caminiti, 1996. Cortical networks for visual reaching: Physiological and anatomical organization of frontal and parietal lobe arm regions. Cerebral Cortex 6:102–119. Jones, W., U. Bellugi, Z. Lai, M. Chiles, J. Reilly, A. Lincoln, and R. Adolphs, 2001. Hypersociability: The social and affective phenotype of Williams syndrome. In U. Bellugi and M. S. George, eds., Journey from Cognition to Brain to Gene: Perspectives from Williams Syndrome, 43–71. Cambridge, MA: MIT Press.

Jonides, J., E. E. Smith, R. A. Koeppe, E. Awh, S. Minoshima, and M. A. Mintun, 1993. Spatial working memory in humans as revealed by PET. Nature 363:623–625. Kail, R., J. Pellegrino, and P. Carter, 1980. Developmental changes in mental rotation. J. Exp. Child Psychol. 29:102–116. Kaldy, Z., and A. M. Leslie, 2003. Identification of objects in 9month-old infants: Integrating “what” and “where” information. Dev. Sci. 6:360–373. Kaufman, J., G. Csibra, and M. H. Johnson, 2003. Representing occluded objects in the human infant brain. Proc. R. Soc. Lond. B. Biol. Sci. 270 (Suppl. 2):S140–143. Kaufman, J., G. Csibra, and M. H. Johnson, 2005. Oscillatory activity in the infant brain reflects object maintenance. Proc. Natl. Acad. Sci. USA 102:15271–15274. Kermoian, R., and J. J. Campos, 1988. Locomotor experience: A facilitator of spatial cognitive development. Child Dev. 59: 908–917. Kitterle, F. L., S. Christman, and J. B. Hellige, 1990. Hemispheric differences are found in the identification, but not the detection, of low versus high spatial frequencies. Percept. Psychophys. 48:297–306. Kitterle, F. L., J. B. Hellige, and S. Christman, 1992. Visual hemispheric asymmetries depend on which spatial frequencies are task relevant. Brain Cogn. 20:308–314. Kitterle, F. L., and L. M. Selig, 1991. Visual field effects in the discrimination of sine-wave gratings. Percept. Psychophys. 50: 15–18. Klingberg, T., 2006. Development of a superior frontalintraparietal network for visuo-spatial working memory. Neuropsychologia 44:2171–2177. Klingberg, T., H. Forssberg, and H. Westerberg, 2002. Increased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14:1–10. Konishi, M., 2003. Coding of auditory space. Annu. Rev. Neurosci. 26:31–55. Kosslyn, S. M., G. J. Digirolamo, W. L. Thompson, and N. M. Alpert, 1998. Mental rotation of objects versus hands: Neural mechanisms revealed by positron emission tomography. Psychophysiology 35:151–161. Kosslyn, S. M., J. A. Margolis, A. M. Barrett, E. J. Goldknopf, et al., 1990. Age differences in imagery abilities. Child Dev. 61:995–1010. Krojgaard, P., 2004. A review of object individuation in infancy. Br. J. Dev. Psychol. 22:159–183. Kwon, H., A. L. Reiss, and V. Menon, 2002. Neural basis of protracted developmental changes in visuo-spatial working memory. Proc. Natl. Acad. Sci. USA 99:13336–13341. Lamb, M. R., L. C. Robertson, and R. T. Knight, 1989. Attention and interference in the processing of global and local information: Effects of unilateral temporal-parietal junction lesions. Neuropsychologia 27:471–483. Lamb, M. R., L. C. Robertson, and R. T. Knight, 1990. Component mechanisms underlying the processing of hierarchically organized patterns: Inferences from patients with unilateral cortical lesions. J. Exp. Psychol. [Learn. Mem. Cogn.] 16:471– 483. Lange-Küttner, C., and A. D. Friederici, 2000. Modularity of object and place memory in children. Brain Cogn. 43:302– 305. Leslie, A. M., F. Xu, P. D. Tremoulet, and B. J. Scholl, 1998. Indexing and the object concept: Developing “what” and “where” systems. Trends Cogn. Sci. 2:10–18.

Livingstone, M. S., and D. H. Hubel, 1984. Anatomy and physiology of a color system in the primate visual cortex. J. Neurosci. 4:309–356. Logie, R. H., 1995. Visuo-spatial Working Memory. Hillsdale, NJ: Lawrence Erlbaum. Logie, R. H., and D. G. Pearson, 1997. The inner eye and the inner scribe of visuo-spatial working memory: Evidence from developmental fractionation. Eur. J. Cogn. Psychol. 9:241– 257. Lorsbach, T. C., and J. F. Reimer, 2005. Feature binding in children and young adults. J. Genet. Psychol. 166:313–327. Luciana, M., H. M. Conklin, C. J. Hooper, and R. S. Yarger, 2005. The development of nonverbal working memory and executive control processes in adolescents. Child Dev. 76: 697–712. Luna, B., K. E. Garver, T. A. Urban, N. A. Lazar, and J. A. Sweeney, 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75:1357–1372. Madrid, M., and M. A. Crognale, 2000. Long-term maturation of visual pathways. Visual Neurosci. 17:831–837. Marangolo, P., E. Di Pace, R. Rafal, and D. Scabini, 1998. Effects of parietal lesions in humans on color and location priming. J. Cogn. Neurosci. 10:704–716. Marmor, G. S., 1975. Development of kinetic images: When does the child first represent movement in mental images? Cogn. Psych. 7:548–559. Martin, M., 1979. Hemispheric specialization for local and global processing. Neuropsychologia 17:33–40. Martinez, A., L. Anllo-Vento, and S. A. Hillyard, 1997. Brain electrical activity during selective attention to spatial frequency. Soc. Neurosci. Abstracts. Martinez, A., P. Moses, L. Frank, R. Buxton, E. Wong, and J. Stiles, 1997. Hemispheric asymmetries in global and local processing: Evidence from fMRI. NeuroReport 8:1685– 1689. Matthews, A., A. E. Ellis, and C. A. Nelson, 1996. Development of preterm and full-term infant ability on AB, recall memory, transparent barrier detour, and means-end tasks. Child Dev. 67:2658–2676. McFie, J., and O. L. Zangwill, 1960. Visual-constructive disabilities associated with lesions of the left cerebral hemisphere. Brain 83:243–259. Merigan, W. H., and J. H. Maunsell, 1993. How parallel are the primate visual pathways? Annu. Rev. Neurosci. 16:369–402. Merriman, W. E., D. P. Keating, and J. A. List, 1985. Mental rotation of facial profiles: Age-, sex-, and ability-related differences. Dev. Psychol. 21:888–900. Meyer-Lindenberg, A., P. Kohn, C. B. Mervis, J. S. Kippenhan, R. K. Olsen, C. A. Morris, and K. F. Berman, 2004. Neural basis of genetically determined visuospatial construction deficit in Williams syndrome. Neuron 43:623–631. Middlebrooks, J. C., L. Xu, S. Furukawa, and E. A. Macpherson, 2002. Cortical neurons that localize sounds. Neuroscientist 8:73– 83. Milner, A. D., and M. A. Goodale, 1995. The Visual Brain in Action. New York: Oxford University Press. Mitchell, T. V., and H. J. Neville, 2004. Asynchronies in the development of electrophysiological responses to motion and color. J. Cogn. Neurosci. 16:1363–1374. Mondloch, C. J., S. Geldart, D. Maurer, and S. de Schonen, 2003. Developmental changes in the processing of hierarchical shapes continue into adolescence. J. Exp. Child Psychol. 84:20– 40.

stiles, paul, and ark: development of visuospatial processing

537

Moses, P., K. Roe, R. B. Buxton, E. C. Wong, L. R. Frank, and J. Stiles, 2002. Functional MRI of global and local processing in children. NeuroImage 16:415–424. Nagy, Z., H. Westerberg, and T. Klingberg, 2004. Maturation of white matter is associated with the development of cognitive functions during childhood. J. Cogn. Neurosci. 16:1227–1233. Navon, D., 1977. Forest before trees: The precedence of global features in visual perception. Cogn. Psych. 9:353–383. Nelson, C. A., 1999. Neural plasticity and human development. Curr. Dir. Psychol. Sci. 8:42–45. Nelson, C. A., C. S. Monk, J. Lin, L. J. Carver, K. M. Thomas, and C. L. Truwit, 2000. Functional neuroanatomy of spatial working memory in children. Dev. Psychol. 36:109–116. Okagaki, L., and P. A. Frensch, 1994. Effects of video game playing on measures of spatial performance: Gender effects in late adolescence. J. Appl. Dev. Psychol. (Special Issue: Effects of Interactive Entertainment Technologies on Development) 15:33–58. O’Keefe, J., and L. Nadel, 1978. The Hippocampus as a Cognitive Map. Oxford, UK: Oxford University Press. Olesen, P. J., Z. Nagy, H. Westerberg, and T. Klingberg, 2003. Combined analysis of DTI and fMRI data reveals a joint maturation of white and grey matter in a fronto-parietal network. Brain Res. Cogn. Brain Res. 18:48–57. Oliveri, M., P. Turriziani, G. A. Carlesimo, G. Koch, F. Tomaiuolo, M. Panella, and C. Caltagirone, 2001. Parietofrontal interactions in visual-object and visual-spatial working memory: Evidence from transcranial magnetic stimulation. Cerebral Cortex 11:606–618. Orsini, A., D. Grossi, E. Capitani, M. Laiacona, C. Papagno, and G. Vallar, 1987. Verbal and spatial immediate memory span: Normative data from 1355 adults and 1112 children. Ital. J. Neurol. Sci. 8:539–548. Palmer, S. E., 1977. Hierarchical structure in perceptual representation. Cogn. Psych. 9:441–474. Palmer, S. E., 1980. What makes triangles point: Local and global effects in configurations of ambiguous triangles. Cogn. Psych. 12:285–305. Palmer, S. E., and N. M. Bucher, 1981. Configural effects in perceived pointing of ambiguous triangles. J. Exp. Psychol. [Hum. Percept.] 7:88–114. Parrish, E. E., D. E. Giaschi, C. Boden, and R. Dougherty, 2005. The maturation of form and motion perception in school age children. Vis. Res. 45:827–837. Paul, B., S. Carapetian, J. Hesselink, R. Nass, D. Trauner, and J. Stiles, in preparation. Face and location processing in children with early unilateral brain injury. Paul, B. M., A. M. Passarotti, and J. Stiles, 2001. Face and spatial location processing: A developmental perspective. Annual Meeting of the Cognitive Neuroscience Society, New York. Paul, B. M., J. Stiles, A. Passarotti, N. Bavar, and U. Bellugi, 2002. Face and place processing in Williams syndrome: Evidence for a dorsal-ventral dissociation. NeuroReport 13:1115–1119. Piaget, J., 1952. The Origins of Intelligence in Children. New York: International Universities Press. Pierrot-Deseilligny, C., D. Milea, and R. M. Muri, 2004. Eye movement control by the cerebral cortex. Curr. Opin. Neurol. 17:17–25. Porporino, M., D. I. Shore, G. larocci, and J. A. Burack, 2004. A developmental change in selective attention and global form perception. Int. J. Behav. Dev. 28:358–364. Posner, M. I., 1980. Orienting of attention. Q. J. Exp. Psychol. 32:3–25.

538

cognition

Posner, M. I., and Y. Cohen, 1980. Attention and the control of movements. In G. E. Stelmach and J. Requin, eds., Tutorials in Motor Behavior, 243–258. Amsterdam: North Holland. Posner, M. I., and S. E. Petersen, 1990. The attention system of the human brain. Annu. Rev. Neurosci. 13:25–42. Posner, M. I., R. D. Rafal, L. S. Choate, and J. Vaughan, 1985. Inhibition of return: Neural basis and function. Cogn. Neuropsychol. 2:211–228. Posner, M. I., J. A. Walker, F. J. Friedrich, and R. D. Rafal, 1984. Effects of parietal injury on covert orienting of attention. J. Neurosci. 4:1863–1874. Postle, B. R., J. S. Berger, A. M. Taich, and M. D’Esposito, 2000. Activity in human frontal cortex associated with spatial working memory and saccadic behavior. J. Cogn. Neurosci. 12 (Suppl. 2):2–14. Quinn, P. C., S. Burke, and A. Rush, 1993. Part-whole perception in early infancy: Evidence for perceptual grouping produced by lightness similarity. Infant Behav. Dev. 16:19–42. Rafal, R., and L. Robertson, 1995. The neurology of visual attention. In M. S. Gazzaniga et al., eds., The Cognitive Neurosciences, 625–648. Cambridge, MA: MIT Press. Reid, V. M., T. Striano, J. Kaufman, and M. H. Johnson, 2004. Eye gaze cueing facilitates neural processing of objects in 4month-old infants. NeuroReport 15:2553–2555. Reynolds, G. D., and J. E. Richards, 2005. Familiarization, attention, and recognition memory in infancy: An event-related potential and cortical source localization study. Dev. Psychol. 41:598–615. Richter, W., K. Ugurbil, A. Georgopoulos, and S. G. Kim, 1997. Time-resolved fMRI of mental rotation. NeuroReport 8: 3697–3702. Riecansky, I., 2004. Extrastriate area V5 (MT) and its role in the processing of visual motion. Cesk. Fysiol. 53:17–22. Rizzolatti, G., R. Camarda, L. Fogassi, M. Gentilucci, G. Luppino, and M. Matelli, 1988. Functional organization of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal movements. Exp. Brain Res. 71:491–507. Rizzolatti, G., and M. Matelli, 2003. Two different streams form the dorsal visual system: Anatomy and functions. Exp. Brain Res. 153:146–157. Roberts, J. E., and M. A. Bell, 2002. The effects of age and sex on mental rotation performance, verbal performance, and brain electrical activity. Dev. Psychobiol. 40:391–407. Robertson, L. C., 1992. Perceptual organization and attentional search in cognitive deficits. In D. I. Margolin et al., eds., Cognitive Neuropsychology in Clinical Practice, 70–95. New York: Oxford University Press. Robertson, L. C., and D. C. Delis, 1986. “Part-whole” processing in unilateral brain-damaged patients: Dysfunction of hierarchical organization. Neuropsychologia 24:363–370. Robertson, L. C., M. R. Lamb, and R. T. Knight, 1988. Effects of lesions of temporal-parietal junction on perceptual and attentional processing in humans. J. Neurosci. 8:3757–3769. Sakata, H., and M. Taira, 1994. Parietal control of hand action. Curr. Opin. Neurobiol. 4:847–856. Sala, J. B., P. Rama, and S. M. Courtney, 2003. Functional topography of a distributed neural system for spatial and nonspatial information maintenance in working memory. Neuropsychologia (Special Issue: Functional Neuroimaging of Memory) 41:341–356. Sandberg, E. H., J. Huttenlocher, and N. Newcombe, 1996. The development of hierarchical representation of two-dimensional space. Child Dev. 67:721–739.

Schneider, G. E., 1967. Contrasting visuomotor functions of tectum and cortex in the golden hamster. Psychologische Forschung 31:52–62. Schweinsburg, A. D., B. J. Nagel, and S. F. Tapert, 2005. fMRI reveals alteration of spatial working memory networks across adolescence. J. Int. Neuropsychol. Soc. 11:631–644. Seltzer, B., and D. N. Pandya, 1978. Afferent cortical connections and architectonics of the superior temporal sulcus and surrounding cortex in the rhesus monkey. Brain Res. 149:1–24. Sergent, J., 1982. The cerebral balance of power: Confrontation or cooperation? J. Exp. Psychol. [Hum. Percept.] 8:253–272. Shepard, R. N., and L. A. Cooper, 1986. Mental Images and Their Transformations. Cambridge, MA: MIT Press. Siemens, L., R. E. Guttentag, and M. McIntyre, 1989. Age differences in memory for item-identity and occupied-location information. Am. J. Psycho. 102:53–68. Sincich, L. C., and J. C. Horton, 2005. The circuitry of V1 and V2: Integration of color, form, and motion. Annu. Rev. Neurosci. 28:303–326. Slater, A., A. Mattock, E. Brown, and J. G. Bremner, 1991. Form perception at birth: Cohen and Younger (1984) revisited. J. Exp. Child Psychol. 51:395–406. Smith, D. M., and S. J. Mizumori, 2006. Hippocampal place cells, context, and episodic memory. Hippocampus 16:716–729. Smith, E. E., J. Jonides, and R. A. Koeppe, 1996. Dissociating verbal and spatial working memory using PET. Cerebral Cortex 6:11–20. Smith, E. E., J. Jonides, R. A. Koeppe, E. Awh, E. H. Schumacher, and S. Minoshima, 1995. Spatial versus object working memory: PET investigations. J. Cogn. Neurosci. 7:337–356. Smith, L. B., and D. G. Kemler, 1977. Developmental trends in free classification: Evidence for a new conceptualization of perceptual development. J. Exp. Child Psychol. 24:279–298. Smith, L. B., E. Thelen, R. Titzer, and D. McLin, 1999. Knowing in the context of acting: The task dynamics of the A-not-B error. Psychol. Rev. 106:235–260. Snow, J. H., 1990. Investigation of a mental rotation task with school-age children. J. Psychoeduc. Assess. 8:538–549. Spencer, J., J. O'Brien, K. Riggs, O. Braddick, J. Atkinson, and J. Wattam-Bell, 2000. Motion processing in autism: Evidence for a dorsal stream deficiency. NeuroReport 11:2765– 2767. Spencer, J. P., L. B. Smith, and E. Thelen, 2001. Tests of a dynamic systems account of the A-not-B error: The influence of prior experience on the spatial memory abilities of two-year-olds. Child Dev. 72:1327–1346. Stein, J., J. Talcott, and V. V. Walsh, 2000. Controversy about the visual magnocellular deficit in developmental dyslexics. Trends Cogn. Sci. 4:209–211. Stiles-Davis, J., J. Janowsky, M. Engel, and R. D. Nass, 1988. Drawing ability in four young children with congenital unilateral brain lesions. Neuropsychologia 26:359–371. Stiles, J., 2000. Neural plasticity and cognitive development. Dev. Neuropsychol. 18:237–272. Stiles, J., E. A. Bates, D. Thal, D. Trauner, and J. Reilly, 1998. Linguistic, cognitive, and affective development in children with pre- and perinatal focal brain injury: A ten-year overview from the San Diego Longitudinal Project. In C. Rovee-Collier, L. P. Lipsitt, and H. Hayne, eds., Advances in Infancy Research, 131–163. Stamford, CT: Ablex. Stiles, J., B. Paul, and J. Hesselink, 2006. Spatial cognitive development following early focal brain injury: Evidence for adaptive

change in brain and cognition. In Y. Munakata and M. H. Johnson, eds., Process of Change in Brain and Cognitive Development: Attention and Performance XXI, 535–561. Oxford, UK: Oxford University Press. Stiles, J., J. Reilly, B. Paul, and P. Moses, 2005. Cognitive development following early brain injury: Evidence for neural adaptation. Trends Cogn. Sci. 9:136–143. Stiles, J., and C. Stern, 2001. Developmental change in spatial cognitive processing: Complexity effects and block construction performance in preschool children. J. Cogn. Dev. 2:157– 187. Stiles, J., C. Stern, M. Appelbaum, R. Nass, D. Trauner, and J. Hesselink, 2008. Effects of early focal brain injury on memory for visuospatial patterns: Selective deficits of global-local processing. Neuropsychology 22:61–73. Stiles, J., D. Trauner, M. Engel, and R. Nass, 1997. The development of drawing in children with congenital focal brain injury: Evidence for limited functional recovery. Neuropsychologia 35:299– 312. Swindell, C. S., A. L. Holland, D. Fromm, and J. B. Greenhouse, 1988. Characteristics of recovery of drawing ability in left and right brain-damaged patients. Brain Cogn. 7:16–30. Tada, W. L., and J. Stiles, 1996. Developmental change in children’s analysis of spatial patterns. Dev. Psychol. 32:951–970. Tallon-Baudry, C., O. Bertrand, F. Peronnet, and J. Pernier, 1998. Induced gamma-band activity during the delay of a visual short-term memory task in humans. J. Neurosci. 18: 4244–4254. Thiele, A., K. R. Dobkins, and T. D. Albright, 2001. Neural correlates of chromatic motion perception. Neuron 32:351– 358. Thomas, K. M., S. W. King, P. L. Franzen, T. F. Welsh, A. L. Berkowitz, D. C. Noll, V. Birmaher, and B. J. Casey, 1999. A developmental functional MRI study of spatial working memory. NeuroImage 10:327–338. Trevarthen, C. B., 1968. Two mechanisms of vision in primates. Psychologische Forschung 31:299–337. Ungerleider, L. G., 1995. Functional brain imaging studies of cortical mechanisms for memory. Science 270:769–775. Ungerleider, L. G., and M. Mishkin, 1982. Two cortical visual systems. In D. J. Ingle, M. A. Goodale, and R. J. W. Mansfield, eds., Analysis of Visual Behavior, 549–586. Cambridge, MA: MIT Press. Valenza, E., F. Simion, and C. Umilta, 1994. Inhibition of return in newborn infants. Infant Behav. Dev. 17:293–302. Van de Walle, G. A., S. Carey, and M. Prevor, 2000. Bases for object individuation in infancy: Evidence from manual search. J. Cogn. Dev. 1:249–280. Van Essen, D. C., C. H. Anderson, and D. J. Felleman, 1992. Information processing in the primate visual system: An integrated systems perspective. Science 255:419–423. Vicari, S., J. Stiles, C. Stern, and A. Resca, 1998. Spatial grouping activity in children with early cortical and subcortical lesions. Dev. Med. Child Neurol. 40:90–99. Volberg, G., and R. Hubner, 2004. On the role of response conflicts and stimulus position for hemispheric differences in global/local processing: An ERP study. Neuropsychologia 42:1805– 1813. von Bonin, G., and P. Bailey, 1947. The Neocortex of Macaca mulatta (Illinois Monogr. Med. Sci., 5, No. 4). Urbana: University of Illinois Press. Vurpillot, E., 1976. The Visual World of the Child. New York: International Universities Press.

stiles, paul, and ark: development of visuospatial processing

539

Waber, D. P., D. Carlson, and M. Mann, 1982. Developmental and differential aspects of mental rotation in early adolescence. Child Dev. 53:1614–1621. Wagner, A. D., B. J. Shannon, I. Kahn, and R. L. Buckner, 2005. Parietal lobe contributions to episodic memory retrieval. Trends Cogn. Sci. 9:445–453. Wellman, H. M., D. Cross, and K. Bartsch, 1987. Infant search and object permanence: A meta-analysis of the A-not-B error. Monogr. Soc. Res. Child Dev. 51:1–51. Welsh, M. C., and B. F. Pennington, 1988. Assessing frontal lobe functioning in children: Views from developmental psychology. Dev. Neuropsychol. 4:199–230. Wilcox, T., 1999. Object individuation: Infants’ use of shape, size, pattern, and color. Cognition 72:125–166. Wilcox, T., and R. Baillargeon, 1998. Object individuation in infancy: The use of featural information in reasoning about occlusion events. Cogn. Psych. 37:97–155.

540

cognition

Willis, S. L., and K. W. Schaie, 1988. Gender differences in spatial ability in old age: Longitudinal and intervention findings. Sex Roles 18:189–203. Wise, S. P., D. Boussaoud, P. B. Johnson, and R. Caminiti, 1997. Premotor and parietal cortex: Corticocortical connectivity and combinatorial computations. Annu. Rev. Neurosci. 20: 25–42. Wong-Riley, M., 1979. Changes in the visual system of monocularly sutured or enucleated cats demonstrable with cytochrome oxidase histochemistry. Brain Res. 171:11–28. Xu, F., and S. Carey, 1996. Infants’ metaphysics: The case of numerical identity. Cogn. Psych. 30:111–153. Yovel, G., J. Levy, and I. Yovel, 2001. Hemispheric asymmetries for global and local visual perception: Effects of stimulus and task factors. J. Exp. Psychol. [Hum. Percept.] 27:1369–1385. Zald, D. H., and W. G. Iacono, 1998. The development of spatial working memory abilities. Dev. Neuropsychol. 14:563–578.

33

Mechanisms of Change: A Cognitive Neuroscience Approach to Declarative Memory Development JENNY RICHMOND AND CHARLES A. NELSON

It is now generally accepted that there are at least two distinct brain systems that are responsible for different kinds of memory function. Structures in the medial temporal lobe, including the hippocampus and parahippocampal cortex, underlie the conscious recollection of facts and events (i.e., explicit or declarative memory). In contrast, parts of the striatum, cerebellum, and brain stem are responsible for the implicit or procedural learning that is evident in priming, conditioning, and skill-learning tasks (for review, see Eichenbaum, 2002). The idea that there may be multiple memory systems leads to obvious questions among developmental psychologists studying memory: How do memory systems develop? How does the maturation of the brain regions that underlie these memory systems contribute to their development? Schacter and Moscovitch (1984) were the first to argue that the memory systems that are dissociated in amnesia are also dissociated during the course of typical development. According to their view, implicit memory, or the unconscious learning that is expressed by changes in task performance as the result of experience, is controlled by an early-developing system, which may be present at birth. In contrast, the development of explicit memory is quite protracted, emerging when a late-developing neural system matures around 8 to 10 months of age. In a similar theoretical vein, Nelson (1995) proposed that the striatum, cerebellum, and brain stem, which are functional at birth, allow even very young infants to demonstrate learning on visual-expectation, operant-conditioning, and classical-conditioning tasks. In contrast, tasks such as the visual paired-comparison (VPC) procedure, which depend on the hippocampus, are controlled by a “preexplicit” memory system, which begins to give way to adultlike explicit memory capabilities around 8 months of age. Although the early-developing hippocampus can sustain performance on novelty preference tasks over short delays,

further maturation of limbic areas (e.g., dentate gyrus) and cortical areas (e.g., inferior temporal cortex) is required for performance on traditional versions of the delayed nonmatching to sample (DNMS) task, deferred imitation, and cross-modal recognition. Nelson (1995) suggests that the development of successful performance on these tasks between 8 and 12 months of age is the result of a transition from reliance on a preexplicit system to use of an adultlike explicit memory system. In contrast, some researchers argue that there is no evidence for a qualitative change in the nature of infants’ memory capabilities late in the first year of life. RoveeCollier and colleagues have shown that when infants who span the proposed critical age period are tested on the same procedure, there is little evidence of a sudden improvement in performance around 8 months of age (Hartshorn et al., 1998). In addition, performance on the mobile conjugate reinforcement task is subject to all the same variables that affect adults’ performance on measures of declarative memory (Rovee-Collier, 1997). In this operant-conditioning paradigm, infants learn to kick their foot to produce movement in an overhanging mobile, which is attached to their foot by a ribbon. The term “conjugate reinforcement” refers to the fact that the rate and vigor of reinforcement (i.e., movement in the mobile) is directly proportional to the rate and vigor of the infants’ responding (i.e., kicking). Despite the similarity between this paradigm and other operant measures of implicit memory, Rovee-Collier and colleagues have shown that infants’ performance on the task is affected by changes in study time, retention interval, and context; these variables typically influence adults’ performance on declarative or explicit memory tasks but do not affect performance on procedural or implicit memory tasks. RoveeCollier, Hayne, and Colombo (2001) suggest that explicit and implicit memory systems develop in parallel gradually and continuously from early in life.

541

Overall, there is little debate about the development of implicit or procedural memory. Areas of the striatum, cerebellum, and brain stem are functionally mature very early in life, and evidence of simple conditioning can be seen in newborn infants (DeCasper and Fifer, 1980). In contrast, there is much controversy regarding the relative development of declarative memory. Studying the development of declarative memory is inherently difficult, because traditional definitions do not apply to preverbal infants. It is difficult to know whether or not infants experience consciousness, and therefore it is unclear whether they are capable of conscious recollection. Infants are unable to express their memories explicitly; researchers must indirectly infer memory from changes in infants’ behaviors as a function of their experiences. As a consequence, much research has been focused on first determining whether a given infant memory task qualifies as a measure of declarative memory, and second determining the earliest age at which infants can perform the task. This largely paradigm-driven approach has focused the field more on determining when performance on a given task develops, rather than considering the processes underlying developmental change (Ornstein and Haden, 2001). In this chapter, we will explore the merit of taking a cognitive neuroscience approach to understanding the development of declarative memory. We will first consider what we know about the development of performance on declarative memory tasks in the context of current knowledge about the brain systems responsible for task performance. We will then review the fundamental changes in memory processes that occur during infancy and consider the development of brain systems involved in encoding, retention, and retrieval. By considering the development of memory processes that underlie memory tasks and integrating current knowledge of brain development, we hope to provide a comprehensive account of age-related changes in declarative memory.

Declarative memory in infancy As discussed previously, a major stumbling block in the study of declarative memory in infants is the difficulty in determining whether infant memory tasks tap declarative or procedural memory. There are at least two strategies that are typically used to assign memory tasks to memory systems. In adults, declarative and procedural memory tasks can be dissociated both in the degree to which patients with temporal lobe amnesia are impaired on the task (i.e., the amnesia filter) and in the effects of certain independent variables, including study time, retention interval, and context change, on the performance of normal adults (i.e., the parameter filter) (for review, see Rovee-Collier, Hayne, and Colombo, 2001). For the purpose of this chapter, we will restrict our

542

cognition

discussion to infant memory tasks for which studies applying both the amnesia filter and the parameter filter have been conducted. Visual Paired-Comparison Task Although research was originally considered to measure early-developing implicit memory (Schacter and Moscovitch, 1984), recent research has shown that novelty preferences are dependent on the hippocampus and may reflect a form of preexplicit memory (Nelson, 1995). The visual paired-comparison (VPC) task typically involves two phases, a familiarization phase and a test phase. During the familiarization phase, infants are exposed to a pair of stimuli for a fixed amount of time or until a fixed amount of looking time is accumulated. Following a delay, infants are shown another pair of stimuli; this time one is the same as the familiarization stimulus and one is novel. Memory is inferred if infants exhibit a novelty preference, spending a greater proportion of time fixating the novel stimulus than the familiar stimulus during the test. The VPC task passes both the amnesia filter and the parameter filter as a measure of declarative memory. Patients with damage to the medial temporal lobe exhibit null preferences when tested on the VPC task after a delay (McKee and Squire, 1993; Pascalis et al., 2004). Similarly, studies with infants and adults have demonstrated that performance on the VPC task is affected by study time, retention interval, and context change in the same way that these variables affect adults’ performance on other measures of declarative memory (Bahrick and Pickens, 1995; Richmond et al., 2004; Robinson and Pascalis, 2004; Rose et al., 1982). There is also considerable evidence from primate lesion studies that the medial temporal lobe memory system generally, and the hippocampus specifically, are critically involved in VPC performance (see also chapter 30 by Bachevalier, this volume). Monkeys with lesions of the hippocampus are impaired on the VPC task in a delay-dependent fashion. Consistent with data from adult amnesics, lesioned monkeys are generally able to exhibit novelty preferences when tested immediately; however, their performance relative to control animals is impaired when they are tested after a delay (Bachevalier, Brickson, and Hagger, 1993; Nemanic, Alvarado, and Bachevalier, 2004; Pascalis and Bachevalier, 1999; Zola et al., 2000). It is important to note that several primate studies have now reported dissociations between performance on the VPC task and other recognition tasks such as delayed nonmatching to sample (DNMS) (for review, see chapter 30 by Bachevalier, this volume). Patients with damage to the medial temporal lobe have also shown impaired VPC performance but relatively intact recognition (McKee and Squire, 1993; Pascalis et al., 2004). These results suggest that the hippo-campus may be critical for the expression of novelty preferences specifically, rather than recognition

memory per se. Recognition memory in a general sense is likely mediated by extrahippocampal structures (Nelson, de Haan, and Thomas, 2006). Taken together, there is substantial evidence that the VPC task depends on the medial temporal lobe memory system. In addition, performance on the VPC task in infants, children, and adults is susceptible to changes in study time, retention interval, and context change in a manner that is consistent with a measure of declarative memory. Deferred Imitation The deferred imitation task assesses infants’ abilities to encode, retain, and reproduce a sequence of novel actions following a delay. Although the deferredimitation task has long been considered a hallmark of sophisticated mental representation (Piaget, 1962), recent studies have shown that it passes both the amnesia filter and the parameter filter and can be considered to measure declarative memory. McDonough and colleagues (1995) tested adult amnesics with bilateral medial temporal lobe damage on an ageappropriate version of the deferred imitation task. Although controls produced more target actions spontaneously and during cued testing than during the baseline phase, amnesic patients were no more likely to use the objects to produce target actions than were inexperienced controls, who had not seen the target actions demonstrated the previous day. Patients with developmental amnesia, who incurred hippocampal damage early in life, also exhibited impairments on the deferred imitation task, although their impairment was restricted to the recall of the temporal order of actions (Adlam et al., 2005). In addition, infants who were at risk for damage to the medial temporal lobe memory system because of adverse fetal environments (e.g., chronic hypoxia) also exhibit impairments in recalling the temporal order of imitation sequences (DeBoer et al., 2005). Infants’ performance on the deferred imitation task is affected by the same variables that influence adults’ performance on declarative memory tasks. Using a deferred imitation task in which a set of actions are demonstrated with an animal hand puppet, Hayne and colleagues have shown that infants’ imitation is enhanced by increasing the length of the demonstration; performance declines as a function of increasing retention interval; and retrieval is disrupted by a change in context (Barr and Hayne, 2000; Hayne, Boniface, and Barr, 2000; Herbert and Hayne, 2000a). Taken together, data from adults and children with amnesia and from typically developing infants converge to suggest that performance on the deferred imitation task depends on the medial temporal lobe memory system, and may reflect declarative memory. In the following section, we will consider what is known about the development of brain systems that subserve performance on declarative memory tasks in infancy.

Brain development The hippocampus is critically involved in declarative memory performance, as measured by both the VPC task and the deferred imitation task. Studies of human brain development have shown that while much of the hippocampal formation is formed early in gestation, development of the dentate gyrus lags behind the development of other hippocampal subfields (chapter 12 by Seress and Ábrahám, this volume). The granule cell layer of the dentate gyrus is present by the 12th week of gestation (Humphrey, 1967); however, cell formation in this area continues until at least 28 weeks. Cell migration is also prolonged; the cytoarchitecture of the dentate gyrus does not appear adultlike until the end of the first postnatal year. Seress and Ábrahám (chapter 12, this volume) also point to the protracted development of inhibitory interneurons within the hippocampal formation. These cells appear in Ammon’s horn earlier than in the dentate gyrus and continue to mature throughout infancy; however, they do not show adultlike morphology until sometime between 2 and 8 years of age. The late development of GABA-ergic interneurons may play an important role in memory and cognitive development, because these cells have been implicated in gamma oscillations that originate in hippocampal cell networks. Gamma-band oscillatory activity (30–80 Hz) has been shown to increase during recognition performance (Tallon-Baudry et al., 1998; Tallon-Baudry, Kreiter, and Bertrand, 1999) and has been implicated in attention and memory processes (for review, see Kahana, 2006).

The development of declarative memory While much of the hippocampus is formed prior to birth, postnatal development of the dentate gyrus, along with maturation of inhibitory interneurons, prolongs the functional maturity of the medial temporal lobe memory system until at least 2 years of age (chapter 12 by Seress and Ábrahám, this volume). These data are consistent with the extant behavioral database, which suggests that while rudimentary declarative memory is evident early, prolonged development of encoding, retention, and retrieval processes contributes to age-related change in declarative memory performance throughout infancy and early childhood. In the following subsections, we will discuss the key characteristics of infant memory development and suggest how brain development may contribute to age-related changes in memory processes during infancy. Encoding As a rule, older infants encode information faster than younger infants (Hayne, 2004). When tested on

richmond and nelson: mechanisms of change

543

the VPC task, the length of the familiarization phase required for infants to display novelty preferences decreases as a function of age (see Rose et al., 1982), so much so that researchers working with different age groups often use infant-controlled familiarization procedures (Diamond, 1995; Pascalis et al., 1998) or design studies allowing younger infants longer familiarization times than older infants (Colombo, Mitchell, and Horowitz, 1988; Jacobs, 2000; Rose, Feldman, and Jankowski, 2001). Similarly, when tested on a deferred imitation task, 6-month-olds require a demonstration period that is twice the length of that used with 12-month-old infants, in order to exhibit equivalent levels of imitation after a delay (Barr, Dowden, and Hayne, 1996). Age-related differences in encoding become less apparent after 12 months of age. Studies using the VPC task with 1- to 4-year-olds have shown that children in different age groups will exhibit equivalent novelty preferences given the same familiarization period (Hayne, 2004). Similarly, whereas some studies using the deferred imitation paradigm report equivalent levels of encoding in 14- and 24-month-olds (Meltzoff, 1985), others report age-related changes in immediate performance until at least 30 months of age (Hayne, Herbert, and Simcock, 2003; Herbert and Hayne, 2000a). Taken together, these data suggest that while declarative memory emerges early in the first year of life, age-related changes in performance on declarative memory tasks may be related to the development of encoding processes. How can we account for age-related changes in encoding during infancy? As discussed earlier, lesions to the hippocampus typically do not impair performance on the declarative memory tasks when the test occurs immediately. For this reason, it is unlikely that the maturation of the hippocampus proper contributes greatly to age-related differences in infants’ encoding. In relation to VPC performance at least, Rose, Feldman, and Jankowski (2004) suggest that developmental changes in encoding may be related to changes in the speed of information processing. Longitudinal event-related potential (ERP) studies lend weight to this idea, demonstrating that the latency to peak amplitude of ERP components decreases as a function of age (Webb, Long, and Nelson, 2005). In this study, ERPs were recorded in response to brief presentations of the infants’ mother and a stranger and presentations of a favorite toy and a novel toy when the same infants were tested at 4, 6, 8, 10, and 12 months of age. Age-related decreases in latency were evident both in early occipital components that are related to visual processing (e.g., Pb) and in the midlatency frontocentral component that is involved in attention (Nc). This result suggests that a domaingeneral mechanism, such as myelination, may be responsible for age-related changes in processing speed.

544

cognition

Myelination of axons within the central nervous system allows for the efficient transmission of electrical impulses. Myelination begins in the fifth fetal month and continues throughout the first two decades of life; however, the most rapid changes in myelination occur during the first postnatal year (Paus et al., 2001). Myelination can be studied noninvasively by looking at the signal intensities of gray and white matter on T1- and T2-weighted magnetic resonance (MR) images. Signal intensities at birth and during the first 6 months of life are the reverse of those seen in the adult (Ballesteros, Hansen, and Soila, 1993; Gilmore et al., 2004). Unmyelinated white matter has a lower signal intensity and thus appears darker than the gray matter in T1-weighted images. An isotense transition period occurs between 8 and 10 months of age, during which time white and gray matter are difficult to differentiate on MR images. The “early adult pattern” in which white matter has a higher signal intensity than gray matter in T1-weighted images becomes evident by 12 months of age (Ballesteros, Hansen, and Soila, 1993). Myelination occurs in a predictable temporal sequence beginning with the brain stem structures responsible for basic functioning and moving in a deep-to-superficial, posterior-to-anterior pattern (Ballesteros, Hansen, and Soila, 1993; Barkovich, 2005). At birth, myelin is evident in the pons and the cerebellar peduncles. Between 1 and 3 months of age, the posterior limb of the internal capsule, optic radiation, and the splenium of the corpus callosum also myelinate. While the anterior limb of the internal capsule and the genu of the corpus callosum are myelinated by 6 months of age, white matter of the cerebral cortex is the last to myelinate. Myelination of the occipital cortex occurs first, closely followed by the white matter of the frontal and parietal lobes around 8 months of age and the temporal lobe by 12 months (Ballesteros, Hansen, and Soila, 1993). Recent studies using diffusion tensor imaging (DTI) have confirmed that the most rapid changes in myelination occur within the first 6 months of life, with slower change between 6 and 24 months and relative stability thereafter (Hermoye et al., 2006; Miller et al., 2003; Prayer and Prayer, 2003). We suggest that the rapid myelination that occurs during the first year may account for age-related changes in processing speed, decreases in the latency of infant ERP components, and ultimately changes in encoding that support declarative memory. Webb, Long, and Nelson (2005) report that age-related decreases in ERP latency plateau between 8 and 10 months of age. Age-related differences in encoding are also less apparent late in the first year. Thus both behavioral and electrophysiological measures point to rapid myelination during the first 12 months followed by slower developmental change thereafter. Retention Given equivalent levels of encoding, older infants will remember for a longer period of time than

younger infants (Hayne, 2004). Using the VPC task, Rose (1981) has shown that when tested after delays ranging from 90 to 160 seconds, whereas both 6- and 9-month-olds displayed novelty preferences when tested immediately, only 9-month-olds exhibited evidence of retention when they were tested after a delay. Age-related changes in retention on the VPC task continue into early childhood. When tested on the VPC task, 1-year-olds will only exhibit retention when tested immediately, 2-year-olds will exhibit novelty preferences for 1 day, 3-year-olds for 1 week, and 4-yearolds for 1 month (Hayne, 2004). Similarly, studies using the deferred imitation paradigm have shown that when levels of immediate imitation are held constant, older infants are able to remember the actions for longer than younger infants. Barr and Hayne (2000) showed that despite equivalent levels of immediate imitation, 6-month-olds will only remember the puppet task for 24 hours, while 12-month-olds will remember the task for 1 week. Like performance on the VPC task, age-related changes in retention continue into the second year of life. Whereas 18- and 24-month-olds will exhibit equivalent imitation when tested immediately, 18-month-olds will imitate for a maximum of 2 weeks, and 24-month-olds will remember the task for a maximum of 12 weeks (Herbert and Hayne, 2000b). Similarly, Bauer (2005) has shown that during the second year of life, younger infants are more susceptible to forgetting after long delays than are older infants, even when levels of initial learning are matched. Bauer (2005) suggests that the fact that older infants exhibit greater savings upon relearning than do younger infants is evidence for age-related differences in storage rather than retrieval processes. Support for the storage account comes from studies combining event-related potentials and imitation tasks over long delays (Bauer et al., 2006, 2003; Carver, Bauer, and Nelson, 2000). In one such study, Bauer and colleagues (2003) showed three elicited imitation sequences to 9-month-old infants during three demonstration sessions. Recognition was assessed using ERPs immediately following the final demonstration and again after a 1-week delay. Potentials recorded during recognition testing were analyzed as a function of whether infants subsequently exhibited ordered recall at the 1-month behavioral test. The results showed that the Nc component differentiated between pictures of novel and familiar sequences during the delayed recognition test, but only for infants who subsequently exhibited ordered recall. Event-related potentials recorded during the immediaterecognition test differentiated between novel and familiar sequences for both groups of infants, ruling out the possibility that subsequent imitation differences were due to differences in initial encoding. Bauer (2005) suggests that the continued development of the dentate gyrus late in the first year of life may account for changes in the infant’s ability to consolidate memories into long-term storage.

This hypothesis is consistent with the data from adult neuroimaging studies, which show greater activation of the hippocampus and medial temporal lobe structures during the encoding of items that are subsequently recalled relative to items that are subsequently forgotten (for review see Henson, 2005). Within the hippocampus, high-resolution fMRI studies have recently shown that CA2/3 subfields and the dentate gyrus may be disproportionately activated during the consolidation of new associative memories, whereas the subiculum may be disproportionately involved in retrieval (Zeinah et al., 2005). In summary, age-related changes in retention occur throughout infancy and into early childhood. Electrophysiological studies suggest that failures in storage may account for age-related changes in retention and point to the prolonged development of the dentate gyrus as a possible mechanism. This hypothesis is consistent with adult neuroimaging studies implicating the hippocampus generally and dentate gyrus specifically in learning and consolidation. Retrieval Infant memories are extremely specific; retrieval will only occur if the cues available at encoding are identical to those available during the test (for review see Hayne, 2004; Rovee-Collier, 1997). Studies using the deferred imitation task have shown that at least until the middle of the second year of life, infants’ memory retrieval is disrupted if they are tested with different, but functionally equivalent, props. Using the puppet task, Hayne, Boniface, and Barr (2000) showed that 6- and 12-month-olds fail to reproduce the target actions after a 24-hour delay if either the color or the form of the puppet is changed prior to the test. It is not until 18 months of age that infants will exhibit retention if the actions are demonstrated with a mouse and tested with a rabbit, or vice versa. This is not to say that by 18 months of age memory is no longer constrained by specific retrieval cues. If the similarity of the puppets is manipulated (i.e., black cow to yellow duck), 18-month-olds are also unable to retrieve their memory when tested with a new puppet stimulus (Hayne, MacDonald, and Barr, 1997). It is not until 24 months of age that infants can generalize the actions they learned with the cow to the duck, or vice versa. Infants’ memory for the puppet task is also specific to the retrieval cues present in the learning environment. Using the same deferred imitation task, Learmonth, Lamberth, and Rovee-Collier (2004) demonstrated the puppet actions to 6and 9-month-old infants while they sat on a distinctive colored mat and then changed either the mat, the room in the house, or both the mat and the room prior to the test. Both 6- and 9-month-olds were able to imitate the actions after a 24-hour delay if the test occurred either on a different mat or in a different room of their house. Only 9-month-olds exhibited retention if both the mat and the room in their

richmond and nelson: mechanisms of change

545

house were changed at the time of the test (Learmonth, Lamberth, and Rovee-Collier, 2004). By 12 months of age, infants tested on the puppet task are able to cope with large changes in the retrieval cues available at the test (Hayne, Boniface, and Barr, 2000). Infants who learn about the puppet in their home and are tested 24 hours later in the lab (or vice versa) are able to reproduce the target actions at a higher rate than controls. Similarly, until at least 12 months of age, infants’ performance on the VPC task is also constrained by retrieval cues available at the test. Robinson and Pascalis (2004) familiarized 6-, 12-, 18-, and 24-month-old infants with objects that were presented on a colored background. When tested immediately with the same proximal context, infants in all age groups exhibited significant novelty preferences. When the color of the background was changed between the familiarization and the test, however, only 18- and 24-month-olds exhibited evidence of retention. In summary, infants’ memories are highly specific. Relatively small changes to either the central stimulus cue or the proximal context disrupt retrieval and cause memory to fail. With age, however, infants’ memory retrieval becomes less constrained, and they develop the ability to retrieve memories in increasingly novel situations. This age-related change in infants’ ability to use memory flexibly is a hallmark of declarative memory development. How can we account for age-related changes in the infants’ memory retrieval? According to Eichenbaum’s (1992) relational-memory account, declarative memory can be applied to nonverbal organisms if it is considered in terms of two fundamental features, relational representations and flexible expression (Eichenbaum, 1992; Eichenbaum, Otto, and Cohen, 1992). By this definition, the hippocampus is critically involved in forming memories that are made up of networks of items that are linked together by causal, logical, or temporal relationships (Eichenbaum, 1999). Flexibility of expression is a natural consequence and advantage of this type of representational coding; relational networks can be used to make inferences about items that are only indirectly connected. In contrast, the regions outside the hippocampus proper, including the entorhinal, perirhinal, and parahippocampal cortices, can support representations of individual items in memory; however, these representations are “hyperspecific.” Hippocampal-independent memories can only be expressed when the specific conditions that were present at learning are reinstated (Eichenbaum, 2002). There is now considerable support for the relational memory hypothesis in both animals and humans. Eichenbaum and colleagues have shown that rats may exhibit hippocampal-dependent memory that is analogous to declarative memory in humans (Bunsey and Eichenbaum, 1995, 1996; Dusek and Eichenbaum, 1997). Rats have demonstrated transitive inference (TI), a classic example of relational rep-

546

cognition

resentation and flexible use of knowledge (Dusek and Eichenbaum, 1997). In this task, rats are trained on a series of odor-based discriminations in which they dig in a pair of scented chow dishes to retrieve a cereal reward. The rewarded odor of any given pair is determined by an arbitrary hierarchical series. For example, when odors A and B are paired together, A is reinforced; when odors B and C are paired together, B is reinforced; and so on. A series of five odors is used (A > B > C > D > E), and a B–D probe test is applied to test for transitive inference. Odors B and D are never presented together and are equally often rewarded (B–C, D–E) and punished (A–B, C–D) during the course of training. If rats are able to use their memory flexibly, they should infer from the series that odor B is the correct choice. When tested on the transitive-inference task, whereas rats with hippocampal disconnection will learn the sequence of odor pairs at the same rate as controls, they are unable to use their knowledge flexibly and thus fail on the B–D probe test (Dusek and Eichenbaum, 1997). Neuroimaging studies of human adults have also shown hippocampal activation during a transitive inference task (Heckers et al., 2004). In this study, adults learned pairs of visual stimuli that did not overlap (i.e., A > B; C > D; E > F) and pairs that formed an overlapping hierarchical sequence (A > B > C > D > E). During scanning, recognition of novel pairs from the overlapping sequence (including the critical B–D pair) was associated with activation in the right anterior hippocampus. In contrast, recognition of nonoverlapping paired associates produced activation in bilateral anterior parahippocampal cortex. Consistent with Dusek and Eichenbaum (1997), these results suggest that the parahippocampal cortex supports simple representations of individual items; however, the flexible use of such knowledge is critically dependent on the hippocampus. Several other fMRI studies have also shown that the hippocampus is activated during relational encoding (Davachi and Wagner, 2002; Henke et al., 1997) and flexible retrieval of memory (Giovanello, Schnyer, and Verfaellie, 2004; Heckers et al., 2004; Preston et al., 2004). By this account, we may attribute infants’ failures to use their memories flexibly to an inability to form relational representations. We now know that, although much of the hippocampus is formed early, the protracted development of the dentate gyrus and inhibitory interneurons precludes adultlike function until at least 2 years of age (chapter 12 by Seress and Ábrahám, this volume). Perhaps in the absence of mature hippocampal function, young infants bind details of the central cue and the peripheral context into a unitary rather than a relational representation, precluding flexible use of memory if aspects of the cue and/or context are changed (Jones and Herbert, 2006). How do we determine whether infants’ memory representations are unitary or relational in nature? The approach

that Eichenbaum (2002) has taken to testing declarative memory in animals is equally applicable to studies of infants and may be useful in answering these questions. By training participants on a set of associated experiences and testing whether acquired knowledge can be used inferentially to solve new problems, we are able to test whether the experiences have been linked in a relational manner. There are now several examples in the infant literature that take just this approach (Barr, Marrott, and Rovee-Collier, 2003; Boller, 1997; Cuevas, Rovee-Collier, and Learmonth, 2006). Using a sensory preconditioning paradigm, Barr, Marrott, and Rovee-Collier (2003) showed that infants associate stimuli that co-occur in their environment and can use their knowledge about the relations among stimuli to support flexible memory. Six-month-old infants were preexposed to two animal hand puppets (a black cow and a yellow duck) that were paired together for 1 hour per day (paired preexposure) or presented for 30 minutes each at separate times of the day (unpaired preexposure). Following 7 days of preexposure to the puppets, infants saw a set of novel actions that were demonstrated using puppet A. Twentyfour hours later, infants were tested for imitation of the target actions using puppet B. Infants in the paired preexposure group reproduced the target actions when they were tested with a different puppet. Infants who had the same amount of preexposure to each puppet but who did not have the opportunity to learn about the relation between the puppets (unpaired preexposure) did not exhibit retention. As discussed earlier, 6-month-olds’ memory for the puppet task is highly specific to the cues present during learning; changes in either the form or the color of the puppet will disrupt retrieval (Hayne, Boniface, and Barr, 2000). Given paired preexposure, however, 6-month-olds learned about the relation between the puppets and were subsequently able to use this knowledge to generalize their learning about the actions from one puppet to the other. While it is tempting to attribute inflexibility in memory retrieval during infancy to immature hippocampal circuitry, these results suggest that under certain circumstances infants as young as 6 months are able to encode associations among items in memory in a relational manner and use these associations to support flexible retrieval (Barr, Marrott, and Rovee-Collier, 2003; Cuevas, Rovee-Collier, and Learmonth, 2006). It is possible that, even without adultlike functioning of the hippocampus, the medial temporal lobe circuitry is sufficiently mature by 6 months of age to support simple relational associations between co-occurring stimuli and contexts, given sufficient experience. Eichenbaum (2002) suggests that declarative memories are encoded in networks of representations that allow new memories to be linked to prior knowledge. Given the limited experience that infants have

with the world, however, the nature of the knowledge networks to which memories can be associated is very sparse. This fact may explain why infants are able to demonstrate relational memory when the relational network is explicitly provided by the experimental situation; however, they are unable to exhibit relational memory in situations where there is less representational support. Barr, Marrott, and Rovee-Collier (2003) have shown that 6-month-olds can imitate the puppet actions when the test occurs with a novel puppet if they are given experiences prior to the demonstration that establish a relational “puppet knowledge” network. In the absence of such experience, however, infants do not possess a sufficiently rich network of knowledge into which the memory for the puppet can be integrated; thus the memory is isolated and inflexible. Evidence for this idea comes from studies demonstrating that locomotor experience influences infants’ ability to use memories flexibly. Herbert, Gross, and Hayne (2007) have shown that 9-month-olds who are crawling are better able to use their memory in a flexible manner than are their noncrawling counterparts. Perhaps the greater experiences that infants encounter during independent locomotion allow them to form richer representational networks into which new memories can be integrated.

Declarative memory beyond infancy Declarative memory performance continues to develop throughout early and middle childhood. Children become increasingly able to use strategies to efficiently encode, retain, and retrieve memories, and episodic memory improves dramatically. Although the use of magnetic resonance imaging to study structural brain development during childhood is not new (Giedd et al., 1995; Lenroot and Giedd, 2006), researchers are only beginning to capitalize on the ability to measure the functional activation of medial temporal lobe memory systems in the developing brain (Chiu et al., 2006; Curtis et al., 2006; Menon, Boyett-Anderson, and Reiss, 2005). Early studies using MRI to study structural brain development during childhood showed gender-specific volumetric change in the hippocampus as a function of age (Giedd et al., 1996; Lenroot and Giedd, 2006). Recent longitudinal analyses have not found such gender differences, but rather have shown that subregions of the hippocampus may have different developmental trajectories (Gogtay et al., 2006). Gogtay and colleagues (2006) scanned 31 children every two years for a 6- to 10-year period. Overall, the volume of the hippocampus remained constant between 4 and 25 years of age; however, the development of subregions within the hippocampus differed considerably. While the posterior portion of the hippocampus gradually increased in volume with age, most prominently in the left hemisphere, the anterior half

richmond and nelson: mechanisms of change

547

of the hippocampus decreased in volume, most prominently on the right. Given this pattern of increasing posterior and decreasing anterior volumes, it is perhaps not surprising that overall hippocampal volume did not change with age (Gogtay et al., 2006). Continued maturation of the connections between the medial temporal lobe and cortical areas, particularly prefrontal cortex, have been implicated in children’s increasing use of memory strategies (Nelson, de Haan, and Thomas, 2006). Recent neuroimaging data confirm that while medial temporal lobe (MTL) activation predicts subsequent story recall in 7- to 8-year-olds, activation in both MTL and prefrontal regions is associated with successful recall in 10- to 18-year-olds (Chiu et al., 2006). Chiu and colleagues (2006) suggest that greater involvement of prefrontal cortex in older children may explain age-related improvements in rehearsal and other encoding strategies across middle childhood. Menon, Boyett-Anderson, and Reiss (2005) have also implicated the connectivity of the MTL and prefrontal cortex in memory development during childhood. In this study, 11- to 19-year-old children were scanned while encoding pictures of outdoor scenes. Encoding produced activation in the striate and extrastriate cortex, along with the hippocampus, entorhinal cortex, and parahippocampal gyrus; however, the extent of encoding-related activation in the left posterior hippocampus and entorhinal cortex decreased with age. In addition, analyses revealed age-related increases in the functional connectivity of the left entorhinal cortex and left dorsolateral prefrontal cortex. Menon, Boyett-Anderson, and Reiss (2005) suggest that decreased involvement of MTL structures, along with increased interaction between MTL and prefrontal structures, may be related to improvements in strategy use, source memory, and awareness during childhood. The development of relational memory abilities during childhood may be related to improvements in episodic and autobiographical memory. Sluzenski, Newcombe, and Kovacs (2006) have shown that between 4 and 6 years of age there is improvement in children’s ability to bind items together in memory. In this study 4- and 6-year-olds and adults studied pictures of animals in an environment and were tested for recognition of the animal, the environment, or the animal/environment combination. While 4- and 6year-olds did not differ in their recognition of the individual items (i.e., animals or environments), 4-year-olds remembered fewer animal/environment combinations than did 6year-olds. Six-year-olds’ performance did not differ from adults. In addition, 4- and 5-year-olds’ delayed recognition of animal/environment combinations was correlated with their ability to recall a story after a delay. Sluzenski, Newcombe, and Kovacs (2006) suggest that the continuing development of children’s ability to bind items together in memory in a relational manner may be related to improvements in

548

cognition

episodic recall. Future neuroimaging studies will determine whether changes in relational memory during childhood can be attributed to continued functional maturation of medial temporal lobe memory systems.

Early damage to hippocampus Damage to the medial temporal lobe results in profound impairments in declarative memory; in adults, memory impairments include the recall of both facts (i.e., semantic memory) and events (i.e., episodic memory). Recent cases of developmental amnesia have highlighted the consequences of early damage to the hippocampus. Vargha-Khadem and colleagues have now documented several cases of children who, as the result of early hypoxic insult, display patterns of memory impairment that are similar to, although perhaps more specific than, those exhibited by adults (for review see de Haan et al., 2006). In the original case report, Vargha-Khadem and colleagues (1997) described cases of three children (Jon, Beth, and Kate) who had suffered hypoxic insult during infancy or early childhood and who presented with severe impairments in everyday memory functioning. Volumetric MRI showed bilateral hippocampal atrophy in all three cases, with volumes ranging from 39 to 57 percent below normal. Neuropsychological evaluation revealed that while these children had IQs within the normal range, their scores on standardized memory tests were significantly lower than would be predicted by their intelligence. The pattern of impairment exhibited by these children closely resembled that of adult amnesic patients. Performance on tests of immediate memory were normal; however, the children exhibited severe deficits when they were asked to recall information after a delay. The children found it difficult to navigate even familiar surroundings, often forgot where they had left objects, were unable to remember appointments or messages, and could not provide accounts of everyday events that they had participated in. Despite profound amnesia for their everyday lives, the children attended mainstream school and had learned to read, write, and spell at a level that was consistent with their IQs. The children had also acquired considerable factual knowledge, as evidenced by normal performance on vocabulary, information, and comprehension subtests. Vargha-Khadem and colleagues (1997) suggest that children with developmental amnesia exhibit severe impairments in episodic memory but relatively spared semantic memory abilities. Since the initial report by Vargha-Khadem and colleagues (1997), several additional cases of developmental amnesia have been identified, and researchers have further characterized the nature of the syndrome (Adlam et al., 2005; Baddeley, Vargha-Khadem, and Mishkin, 2001; Gadian et al., 2000; Isaacs et al., 2003; King et al., 2004; Vargha-

Khadem, Gadian, and Mishkin, 2001; Vargha-Khadem et al., 2003). There is some evidence that, along with spared semantic memory, recognition abilities may be less affected in developmental amnesia than in cases of adult amnesia (Baddeley, Vargha-Khadem, and Mishkin, 2001; Duzel et al., 2001). Subsequent testing of patient Jon showed that, despite severe recall impairments, both behavioral recognition (Baddeley, Vargha-Khadem, and Mishkin, 2001) and ERP components that are associated with familiarity (Duzel et al., 2001) are intact. It is unclear whether this dissociation between recall and recognition impairments is present in all cases of developmental amnesia. Consistent with Eichenbaum’s (1992) relational memory account, it seems that at least part of Jon’s impairment may be related to the ability to bind items together in memory. Spiers and colleagues (2001) tested Jon on a virtual reality paradigm in which he was asked to navigate through a virtual town, collecting objects from people he met along the way. Following the navigation test, Jon was tested using a forced-choice task in which he had to recognize the specific objects he had collected, the people who gave the objects to him, the places where he received the objects, and the order in which they were received. Jon performed as well as controls on the recognition of individual objects; however, his memory for contextual aspects of the virtual environment (i.e., who, where, and when questions) was impaired. Spiers and colleagues (2001) suggest that Jon may perform well on individual items but poorly on these contextual episodic memory questions because he is unable to learn relations between objects and people, objects and places, or objects in time. It seems that the selective pattern of impairment seen in developmental amnesia (i.e., impaired episodic memory but intact semantic memory) is not necessarily restricted to cases in which the injury occurs very early in life. Rather, VarghaKhadem and colleagues (2003) have shown that children who incurred hypoxic-ischemic insult during middle childhood present with a similar pattern of neuropathology and neuropsychological profile as do children who have experienced perinatal hypoxic-ischemic insult. Vargha-Khadem and colleagues (2003) suggest that the syndrome of developmental amnesia may be one that is associated with hypoxicischemic insult incurred between infancy and puberty, rather than one that is specifically associated with perinatal hypoxicischemic insult. It is important to note that while the consequences of perinatal hypoxic-ischemic injury are often not noticed until the children enter school, impairments following insult during middle childhood are evident immediately following the injury (Vargha-Khadem et al., 2003). There are two possible explanations for this difference in impairment onset. First, some researchers have claimed that the functional consequences of very early hippocampal damage may

initially lie “dormant.” Bachevalier and Vargha-Khadem (2005) suggest that children who have incurred hypoxicischemic insult early in life seem to “grow into their memory impairment.” It is possible that such children possess the precursors to explicit memory but fail to develop the full extent of hippocampally dependent memory functions. Because such functions develop gradually across childhood, it is not until the hippocampal memory system approaches maturity that impairments become evident (Bachevalier and Vargha-Khadem, 2005). In contrast, it is possible that memory impairments in infants who have experienced early hypoxic-ischemic insult may be present from the time of the insult, but may go unnoticed until the children enter school simply because it is not until this time that children are held responsible for their own recall and expected to provide accounts of their everyday lives. Future prospective studies of infants who have experienced hypoxic-ischemic insult will resolve this issue. In summary, early hippocampal damage tends to result in impairments that are restricted to the recall of episodic memories, a pattern that is less severe than the memory impairments produced by similar damage incurred during adulthood. To date it is unclear whether differences in the severity of impairments characterizing developmental amnesia and adult amnesia are related to differences in the extent of hippocampal pathology or due to the role of plasticity and compensatory mechanisms in the developing brain (Bachevalier and VarghaKhadem, 2005).

Concluding remarks Behavioral studies suggest that while rudimentary declarative memory abilities develop early in infancy, the fundamental processes that underlie task performance undergo considerable development during infancy and early childhood. This review highlights the potential of a cognitive neuroscience approach in enhancing our understanding of the relation between age-related changes in encoding, retention, and retrieval and the maturation of brain systems that may underlie these processes. By addressing how changes in brain are related to changes in memory, we may begin to move the field beyond simply describing development to understanding the mechanisms that drive it. In summary, there are at least three fundamental characteristics of memory development that are most apparent during infancy; each is independent of the task that is used to measure memory. First, older infants learn faster than younger infants. Second, older infants remember for longer than younger infants. Third, older infants are better able to exploit retrieval cues in the service of memory than are younger infants (Hayne, 2004). The critical question of

richmond and nelson: mechanisms of change

549

interest is: How does brain development drive these agerelated changes in memory processes? Declarative memory depends critically on brain structures in the medial temporal lobe, and age-related changes in retention and retrieval may, at least in part, be attributed to the prolonged maturation of hippocampal circuitry. We now know that the hippocampus and surrounding cortex can support performance on the VPC task and deferred imitation during the first half of the first year, although Seress and Ábrahám (chapter 12, this volume) suggest that adultlike function of the hippocampal region cannot be expected until at least 2 years of age. Prolonged cell formation and migration in the dentate gyrus may explain the continued development of infants’ long-term retention throughout infancy, as the hippocampus has been implicated in consolidation of memory into long-term storage. Continued development of the dentate may also explain age-related changes in infants’ ability to retrieve memories in new situations, as some theories suggest that the hippocampus is critically involved in relational encoding and flexible retrieval (Eichenbaum, 2002). Age-related changes in encoding are likely attributable to rapid myelination during the first year of life, which allows for efficient transmission of electrical signals and faster information processing. It is important to note that identifying a correlation between the relative maturation of brain systems underlying memory and age-related changes in behavior is only the first step to improving our understanding of the mechanisms of development. For example, the coincidental timing of dentate maturation and improvements in memory flexibility only allow us to speculate about a possible association. Studies that combine multiple measures of brain and behavior are critical if we are to determine a causal link between brain and behavior development. Our understanding of the brain mechanisms underlying memory development will benefit from improvements in the armamentarium of neuroimaging tools that allow us to look inside the developing brain. There is much progress being made in customizing ERP source-localization parameters for studies of infants and children (Reynolds and Richards, 2005; Richards, 2005). In addition, technological advances in MRI will allow younger children, although perhaps not infants, to participate in functional imaging studies. The field of developmental cognitive neuroscience, although in its infancy, is certainly beginning an exciting era. acknowledgments

Writing of this paper was made possible, in part, by grants to the second author from the National Institutes of Health (NS32755, NS32976, and MH078829-10A1) and to the first and second authors from the McDonnell Foundation.

550

cognition

REFERENCES Adlam, A., F. Vargha-Khadem, M. Mishkin, and M. de Haan, 2005. Deferred imitation of action sequences in developmental amnesia. J. Cogn. Neurosci. 17:240–248. Bachevalier, J., M. Brickson, and C. Hagger, 1993. Limbicdependent recognition memory in monkeys develops early in infancy. NeuroReport 4:77–80. Bachevalier, J., and F. Vargha-Khadem, 2005. The primate hippocampus: Ontogeny, early insult, and memory. Curr. Opin. Neurobiol. 15:168–174. Baddeley, A., F. Vargha-Khadem, and M. Mishkin, 2001. Preserved recognition in a case of developmental amnesia: Implications for the acquisition of semantic memory? J. Cogn. Neurosci. 13:357–369. Bahrick, L., and J. Pickens, 1995. Infant memory for object motion across a period of three months: Implications for a four-phase attention function. J. Exp. Child Psychol. 59:343– 371. Ballesteros, M., P. Hansen, and K. Soila, 1993. MR imaging of the developing human brain: Part 2. Postnatal development. Radiographics 13:611–622. Barkovich, A. J., 2005. Pediatric Neuroimaging, 4th ed. San Francisco: Lippincott Williams & Wilkins. Barr, R., A. Dowden, and H. Hayne, 1996. Developmental changes in deferred imitation by 6- to 24-month-old infants. Infant Behav. Dev. 19:159–170. Barr, R., and H. Hayne, 2000. Age-related changes in imitation: Implications for memory development. In C. Rovee-Collier, L. Lipsitt, and H. Hayne, eds., Progress in Infancy Research, vol. 1, pp. 21–67. Hillsdale, NJ: Lawrence Erlbaum. Barr, R., H. Marrott, and C. Rovee-Collier, 2003. The role of sensory preconditioning in memory retrieval by preverbal infants. Learn. Behav. 31:111–123. Bauer, P., 2005. Developments in declarative memory: Decreasing susceptibility to storage failure over the second year of life. Psychol. Sci. 16:41–47. Bauer, P., S. Wiebe, L. Carver, A. Lukowski, J. Haight, and C. Nelson, 2006. Electrophysiological indexes of encoding and behavioral indexes of recall: Examining relations and developmental change later in the first year of life. Dev. Neuropsychol. 29:293–320. Bauer, P., S. Wiebe, L. Carver, J. Waters, and C. Nelson, 2003. Developments in long-term explicit memory late in the first year of life: Behavioral and electrophysiological indices. Psychol. Sci. 14:629–635. Boller, K., 1997. Preexposure effects on infant learning and memory. Dev. Psychobiol. 31:93–105. Bunsey, M., and H. Eichenbaum, 1995. Selective damage to the hippocampal region blocks long-term retention of a natural and nonspatial stimulus-stimulus association. Hippocampus 5:546– 556. Bunsey, M., and H. Eichenbaum, 1996. Conservation of hippocampal memory function in rats and humans. Nature 379: 255–257. Carver, L., P. Bauer, and C. Nelson, 2000. Associations between infant brain activity and recall memory. Dev. Sci. 2:234–246. Chiu, P., V. Schmithorst, R. Douglas Brown, S. Holland, and S. Dunn, 2006. Making memories: A cross-sectional investigation of episodic memory encoding in childhood using fMRI. Dev. Neuropsychol. 29:321–340. Colombo, J., D. Mitchell, and F. Horowitz, 1988. Infant visual behavior in the paired-comparison paradigm: Test-retest

and attention-performance relations. Child Dev. 58:1198– 1210. Cuevas, K., C. Rovee-Collier, and A. Learmonth, 2006. Infants form associations between memory representations of stimuli that are absent. Psychol. Sci. 17:543–549. Curtis, W. J., J. Zhuang, E. Townsend, X. Hu, and C. Nelson, 2006. Memory in early adolescents born prematurely: A functional magnetic resonance imaging investigation. Dev. Neuropsychol. 29:341–377. Davachi, L., and A. Wagner, 2002. Hippocampal contributions to episodic encoding: Insights from relational and item-based learning. J. Neurophysiol. 88:982–990. DeBoer, T., S. Wewerka, P. Bauer, M. Georgieff, and C. Nelson, 2005. Explicit memory performance in infants of diabetic mothers at 1 year of age. Dev. Med. Child Neurol. 47: 525–531. DeCasper, A., and W. Fifer, 1980. Of human bonding: Newborns prefer their mothers’ voices. Science 208:1174–1176. de Haan, M., J. Wyatt, S. Roth, F. Vargha-Khadem, D. Gadian, and M. Mishkin, 2006. Brain and cognitive-behavioural development after asphyxia at term birth. Dev. Sci. 9:350–358. Diamond, A., 1995. Evidence of robust recognition memory early in life even when assessed by reaching behavior. J. Exp. Child Psychol. 59:419–456. Dusek, J., and H. Eichenbaum, 1997. The hippocampus and memory for orderly stimulus relations. Proc. Natl. Acad. Sci. USA 94:7109–7114. Duzel, E., F. Vargha-Khadem, H. Heinze, and M. Mishkin, 2001. Brain activity evidence for recognition without recollection after early hippocampal damage. Proc. Natl. Acad. Sci. USA 98:8101–8106. Eichenbaum, H., 1992. The hippocampal system and declarative memory in animals. J. Cogn. Neurosci. 4:217–231. Eichenbaum, H., 1999. The hippocampus and mechanisms of declarative memory. Behav. Brain Res. 103:123–133. Eichenbaum, H., 2002. The Cognitive Neuroscience of Memory: An Introduction. New York: Oxford University Press. Eichenbaum, H., T. Otto, and N. Cohen, 1992. The hippocampus—What does it do? Behav. Neural Biol. 57:2–36. Gadian, D., J. Aicardi, K. Watkins, D. Porter, M. Mishkin, and F. Vargha-Khadem, 2000. Developmental amnesia associated with early hypoxic-ischaemic injury. Brain 123:499–507. Giedd, J., F. Castellanos, J. Rajapakse, D. Kaysen, A. Vaituzis, Y. Vauss, et al., 1995. Cerebral MRI of human brain development: Ages 4–18. Biol. Psychiatry 37:657. Giedd, J., A. Vaituzis, S. Hamburger, N. Lange, J. Rajapakse, D. Kaysen, et al., 1996. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal development: Ages 4–18 years. J. Comp. Neurol. 366:223–230. Gilmore, J., G. Zhai, K. Wilber, J. Smith, W. Lin, and G. Gerig, 2004. 3 Tesla magnetic resonance imaging of the brain in newborns. Psychiatry Res. Neuroimaging 132:81–85. Giovanello, K., D. Schnyer, and M. Verfaellie, 2004. A critical role for the anterior hippocampus in relational memory: Evidence from an fMRI study comparing associative and item recognition. Hippocampus 14:5–8. Gogtay, N., T. Nugent, D. Herman, A. Ordonez, D. Greenstein, K. Hayashi, et al., 2006. Dynamic mapping of normal human hippocampal development. Hippocampus 16:664–672. Hartshorn, K., C. Rovee-Collier, P. Gerhardstein, R. Bhatt, T. Wondoloski, P. Klein, et al., 1998. The ontogeny of longterm memory over the first year-and-a-half of life. Dev. Psychobiol. 32:69–89.

Hayne, H., 2004. Infant memory development: Implications for childhood amnesia. Dev. Rev. 24:33–73. Hayne, H., J. Boniface, and R. Barr, 2000. The development of declarative memory in human infants: Age-related changes in deferred imitation. Behav. Neurosci. 114:77–83. Hayne, H., J. Herbert, and G. Simcock, 2003. Imitation from television by 24- and 30-month-olds. Dev. Sci. 6:254–261. Hayne, H., S. MacDonald, and R. Barr, 1997. Developmental changes in the specificity of memory over the second year of life. Infant Behav. Dev. 20:233–245. Heckers, S., M. Zalesak, A. Weiss, T. Ditman, and D. Titone, 2004. Hippocampal activation during transitive inference in humans. Hippocampus 14:153–162. Henke, K., A. Buck, B. Weber, and H. Weiser, 1997. Human hippocampus establishes associations in memory. Hippocampus 7:249–256. Henson, R., 2005. A mini-review of fMRI studies of human medial temporal lobe activity associated with recognition memory. Q. J. Exp. Psychol. 58B:340–360. Herbert, J., J. Gross, and H. Hayne, 2007. Crawling is associated with more flexible retrieval by 9-month-old infants. Dev. Sci. 10:183–189. Herbert, J., and H. Hayne, 2000a. Memory retrieval by 18–30month-olds: Age-related changes in representational flexibility. Dev. Psychol. 36:473–484. Herbert, J., and H. Hayne, 2000b. The ontogeny of long-term retention during the second year of life. Dev. Sci. 3:50–56. Hermoye, L., C. Saint-Martin, G. Cosnard, S. Lee, J. Kim, M. Nassogne, et al., 2006. Pediatric diffusion tensor imaging: Normal database and observation of the white matter maturation in early childhood. NeuroImage 29:493–504. Humphrey, T., 1967. The development of the human hippocampal fissure. J. Anat. 101:655–676. Isaacs, E., F. Vargha-Khadem, K. Watkins, A. Lucas, M. Mishkin, and D. Gadian, 2003. Developmental amnesia and its relationship to degree of hippocampal atrophy. Proc. Natl. Acad Sci. USA 100:13060–13063. Jacobs, S. E., 2000. Visual recognition memory for objects in six- and nine-month-old infants. Occup. Ther. J. Res. 20:271– 285. Jones, E., and J. Herbert, 2006. Exploring memory in infancy: Deferred imitation and the development of declarative memory. Infant Child Dev. 15:195–205. Kahana, M., 2006. The cognitive correlates of human brain oscillations. J. Neurosci. 26:1669–1672. King, J., I. Trinkler, T. Hartley, F. Vargha-Khadem, and N. Burgess, 2004. The hippocampal role in spatial memory and the familiarity-recollection distinction: A case study. Neuropsychology 18:405–417. Learmonth, A., R. Lamberth, and C. Rovee-Collier, 2004. Generalization of deferred imitation during the first year of life. J. Exp. Child Psychol. 88:297–318. Lenroot, R., and J. Giedd, 2006. Brain development in children and adolescents: Insights from anatomical magnetic resonance imaging. Neurosci. Biobehav. Rev. 30:718–729. McDonough, L., J. Mandler, R. McKee, and L. Squire, 1995. The deferred imitation task as a nonverbal measure of declarative memory. Proc. Natl. Acad. Sci. USA 92:7580–7584. McKee, R., and L. Squire, 1993. On the development of declarative memory. J. Exp. Psychol. [Learn. Mem. Cogn.] 19:397– 404. Meltzoff, A., 1985. Immediate and deferred imitation in fourteenand twenty-four-month-old infants. Child Dev. 56:62–72.

richmond and nelson: mechanisms of change

551

Menon, V., J. Boyett-Anderson, and A. Reiss, 2005. Maturation of medial temporal lobe response and connectivity during memory encoding. Cogn. Brain Res. 25:379–385. Miller, J., R. Mckinstry, J. Philip, P. Mukherjee, and J. Neil, 2003. Diffusion-tensor MR imaging of normal brain maturation: A guide of structural development and myelination. AJR Am. J. Roentgenol. 180:851–859. Nelson, C., 1995. The ontogeny of human memory: A cognitive neuroscience perspective. Dev. Psychol. 31:723–738. Nelson, C., M. de Haan, and K. Thomas, 2006. Neural bases of cognitive development. In W. Damon, R. Lerner, D. Kuhn, and R. Siegler, eds., Handbook of Child Psychology, 6th ed., vol. 2: Cognitive, Perception and Language, 3–57. Hoboken, NJ: John Wiley. Nemanic, S., M. Alvarado, and J. Bachevalier, 2004. The hippocampal/parahippocampal regions and recognition memory: Insights from visual paired comparison versus object-delayed nonmatching in monkeys. J. Neurosci. 24:2013–2026. Ornstein, P., and C. Haden, 2001. Memory development or the development of memory? Curr. Dir. Psychol. Sci. 10:202–205. Pascalis, O., and J. Bachevalier, 1999. Neonatal aspiration lesions of the hippocampal formation impair visual recognition memory when assessed by paired-comparison task but not by delayed nonmatching-to-sample task. Hippocampus 9:609–616. Pascalis, O., M. de Haan, C. Nelson, and S. de Schonen, 1998. Long-term recognition memory for faces assessed by visual paired comparison in 3- and 6-month-old infants. J. Exp. Psychol. [Learn. Mem. Cogn.] 24:249–260. Pascalis, O., N. Hunkin, J. Holdstock, C. Isaac, and A. Mayes, 2004. Visual paired comparison performance is impaired in a patient with selective hippocampal lesions and relatively intact item recognition. Neuropsychologia 42:1293–1300. Paus, T., D. Collins, A. Evans, G. Leonard, B. Pike, and A. Zijdenbos, 2001. Maturation of white matter in the human brain: A review of magnetic resonance studies. Brain Res. Bull. 54:255– 266. Piaget, J., 1962. Play, Dreams, and Imitation in Children. New York: Norton. Prayer, D., and L. Prayer, 2003. Diffusion-weighted magnetic resonance imaging of cerebral white matter development. Eur. J. Radiol. 45:235–243. Preston, A., Y. Shrager, N. Dudukovic, and J. Gabrieli, 2004. Hippocampal contribution to the novel use of relational information in declarative memory. Hippocampus 14:148–152. Reynolds, G., and J. Richards, 2005. Familiarization, attention, and recognition memory in infancy: An event-related potential and cortical source localization study. Dev. Psychol. 41:598–615. Richards, J., 2005. Localizing cortical sources of event-related potentials in infants’ covert orienting. Dev. Sci. 8:255–278. Richmond, J., P. Sowerby, M. Colombo, and H. Hayne, 2004. The effect of familiarization time, retention interval, and context change on adults’ performance on the visual paired-comparison task. Dev. Psychobiol. 42:146–155. Robinson, A., and O. Pascalis, 2004. Development of flexible visual recognition memory in human infants. Dev. Sci. 7:527– 533.

552

cognition

Rose, S., 1981. Developmental changes in infants’ retention of visual stimuli. Child Dev. 52:227–233. Rose, S., J. Feldman, and J. Jankowski, 2001. Attention and recognition memory in the 1st year of life: A longitundinal study of preterm and full-term infants. Dev. Psychol. 37:135–151. Rose, S., J. Feldman, and J. Jankowski, 2004. Infant visual recognition memory. Dev. Rev. 24:74–100. Rose, S., A. Gottfried, P. Melloy-Carminar, and W. Bridger, 1982. Familiarity and novelty preferences in infant recognition memory: Implications for information processing. Dev. Psychol. 18:704–713. Rovee-Collier, C., 1997. Dissociations in infant memory: Rethinking the development of implicit and explicit memory. Psychol. Rev. 104:467–498. Rovee-Collier, C., H. Hayne, and M. Colombo, 2001. The Development of Implicit and Explicit Memory. Amsterdam: John Benjamins. Schacter, D., and M. Moscovitch, 1984. Infants, amnesics, and dissociable memory systems. In M. Moscovitch, ed., Advances in the Study of Communication and Affect, vol. 9, pp. 173–216. New York: Plenum. Sluzenski, J., N. Newcombe, and S. Kovacs, 2006. Binding, relational memory, and recall of naturalistic events: A developmental perspective. J. Exp. Psychol. [Learn. Mem. Cogn.] 32:89–100. Spiers, H., N. Burgess, T. Hartley, F. Vargha-Khadem, and J. O'Keefe, 2001. Bilateral hippocampal pathology impairs topographical and episodic memory but not visual pattern matching. Hippocampus 11:715–725. Tallon-Baudry, C., O. Bertrand, F. Peronnet, and J. Pernier, 1998. Induced gamma-band activity during the delay of a visual short-term memory task in humans. J. Neurosci. 18:4244– 4254. Tallon-Baudry, C., A. Kreiter, and O. Bertrand, 1999. Sustained and transient oscillatory responses in the gamma and beta bands in a visual short-term memory task in humans. Visual Neurosci. 16:449–459. Vargha-Khadem, F., D. Gadian, and M. Mishkin, 2001. Dissociations in cognitive memory: The syndrome of developmental amnesia. Philos. Trans. R. Soc. London [Biol.] 356:1435–1440. Vargha-Khadem, F., D. Gadian, K. Watkins, A. Connelly, W. Van Paesschen, and M. Mishkin, 1997. Differential effects of early hippocampal pathology on episodic and semantic memory. Science 277:376–380. Vargha-Khadem, F., C. Salmond, K. Watkins, K. Friston, D. Gadian, and M. Mishkin, 2003. Developmental amnesia: Effect of age at injury. Proc. Natl. Acad. Sci. USA 100:10055–10060. Webb, S., J. Long, and C. Nelson, 2005. A longitudinal investigation of visual event-related potentials in the first year of life. Dev. Sci. 8:605–616. Zeinah, M., S. Engel, P. Thompson, and S. Bookheimer, 2005. Dynamics of the hippocampus during encoding and retrieval of face-name pairs. Science 299:577–580. Zola, S., L. Squire, E. Teng, L. Stefanacci, E. Buffalo, and R. Clark, 2000. Impaired recognition memory in monkeys after damage limited to the hippocampal region. J. Neurosci. 20: 451–463.

34

The Development of Executive Function in Childhood PHILIP DAVID ZELAZO, STEPHANIE M. CARLSON, AND AMANDA KESEK

Although informed by cybernetic theory (Weiner, 1948) and hierarchical models of action control (e.g., Miller, Galanter, and Pribram, 1960), the construct of executive function (EF) has its origins in analyses of the consequences of damage to prefrontal cortex (PFC; e.g., Goldstein and Scheerer, 1941; Luria, 1966). Early studies on patients with prefrontal damage, such as the famous case of Phineas Gage (Harlow, 1848, 1868), revealed a peculiar pattern of deficits despite preservation of basic cognitive functions (e.g., Hebb, 1945). These deficits include (but are not limited to) failures to make wise judgments, cognitive inflexibility, poor planning of future actions, and difficulty inhibiting inappropriate responses (e.g., Stuss and Benson, 1986; Tranel, Anderson, and Benton, 1994; Wise, Murray, and Gerfen, 1996). The construct of EF is intended to capture the diverse set of psychological abilities whose impairment is presumed to underlie these manifest deficits. Together, these various abilities allow for conscious, goal-directed problem solving. A first step in understanding EF is to describe it in functional terms (e.g., Goldberg and Bilder, 1987; Luria, 1973; Zelazo and Müller, 2002). Figure 34.1 illustrates how EF unfolds as an iterative process, as well as how different aspects of EF typically work together to contribute to the eventual outcome of conscious problem solving. From this perspective, EF is a higher-order function, with numerous subfunctions, consistent with evidence that EF is characterized by functional unity as well as functional diversity (e.g., Miyake et al., 2000). Consider how this functional characterization applies to the Wisconsin Card Sorting Test (WCST; Grant and Berg, 1948), widely regarded as “the prototypical EF task in neuropsychology” (Pennington and Ozonoff, 1996, 55). In the WCST, participants are given test cards that vary on three dimensions (shape, color, and number) and they must discover the rule for matching them to target cards that also vary on these dimensions. After the participants have sorted correctly by this rule, the rule changes, and they must infer the new rule. Patients with lesions to PFC often persist in sorting by the initial rule despite repeated feedback and despite knowing that they are about to err (e.g., Milner, 1963). To perform correctly, participants must formulate a plan, keep that plan in mind (intending), use the plan to

guide behavior (rule use), and evaluate the outcome of the behavior. Inflexibility may occur at any of these stages.

Development of prefrontal cortex in childhood As shown in figure 34.2 and plate 54, PFC is a heterogeneous region that comprises several distinct subregions, including orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), and ventrolateral (VL-PFC), dorsolateral (DL-PFC), and rostrolateral prefrontal cortices (RL-PFC; figure 34.2). Although PFC function is not synonymous with EF, PFC plays a crucial role in the neural systems that support EF, and there is growing evidence that improvements in EF in childhood are associated with the development of PFC. Indeed, although it was sometimes supposed that PFC was not functional at all during childhood (e.g., Golden, 1981), it is now clear that PFC function emerges early—probably toward the end of the first year of life (e.g., Chugani and Phelps, 1986; Diamond and Goldman-Rakic, 1989)—and continues to develop throughout childhood and adolescence (e.g., Giedd et al., 1999; Gogtay et al., 2004). The growth of PFC beyond infancy has been documented using a variety of measures, and a number of consistent patterns have been noted. First, myelination starts postnatally in PFC and then increases monotonically over the course of childhood (e.g., Klingberg et al., 1999; Yakovlev and Lecours, 1967). Second, the corpus callosum shows peak growth rates between 3 and 6 years in the anterior regions that connect the two frontal lobes (Thompson et al., 2000). Third, in contrast to these age-related increases in white matter, gray matter volume in PFC shows a pattern of early increases followed by gradual decreases that start in late childhood and continue into adulthood (e.g., Gogtay et al., 2004; Huttenlocher, 1990; O’Donnell et al., 2005; see figure 34.3 and plate 55). This inverted U-shaped pattern of growth may reflect the overproduction and subsequent pruning of synapses—a process that would allow the brain to be shaped by experience (e.g., Casey, Giedd, and Thomas, 2000; Durston et al., 2001). Fourth, there are sex differences in PFC development. Throughout childhood and into adolescence, the developmental increases in white matter volume and decreases in gray matter volume are more pronounced

553

for boys than they are for girls (e.g., de Bellis et al., 2001). Moreover, OFC function seems to develop more rapidly in males (e.g., Overman et al., 1996) and seems to be under hormonal control (Clark and Goldman-Rakic, 1989). Finally, for both boys and girls, different regions of PFC seem to mature at different rates. In particular, gray matter volume reaches adult levels earliest in OFC, followed by VL-PFC and then by DL-PFC (Giedd et al., 1999). Measures of cortical thickness suggest that DL-PFC and RL-PFC exhibit similar, slow rates of structural change (O’Donnell et al., 2005).

Problem Representation

Planning

Execution: Intending/Rule Use

Evaluation: Error Detection/Correction

Figure 34.1 A functional characterization of executive function. Dashed lines indicate optional recursive feedback loops. (Adapted with permission from P. D. Zelazo, A. S. Carter, J. S. Reznick, and D. Frye, 1997. Early development of executive function: A problem solving approach. Review of General Psychology 1:198–226.)

In addition to these structural changes, there are at least two general changes in the patterns of neural activation that occur during performance on measures of EF. First, several researchers have noted a shift from more diffuse to more focal activation of PFC—more activation in areas related to EF in adults, and less activation in areas unrelated to EF (e.g., Bunge et al., 2002; Durston et al., 2006; Luna et al., 2001). For example, Durston and colleagues (2006) used fMRI to examine neural activation in children and adults as they performed a go-no-go task, and observed an age-related decrease in activation in brain regions unimportant for task performance. The only increase occurred in an area of VLPFC, and this increase was associated with improved task performance. This focalization as a function of age and EF suggests that the functional development of PFC may be associated with more specialized and efficient processing. Second, several researchers have observed an increasing reliance on more anterior regions of PFC with age and EF development (i.e., frontalization; Lamm, Zelazo, and Lewis, 2006; Rubia et al., 2000). For example, Lamm, Zelazo, and Lewis (2006) used high-density (128-channel) electroencephalography (EEG) to measure event-related potentials (ERPs) on the scalp as children and adolescents performed a go-no-go task, and collected a number of behavioral measures of EF. The N2 component of the ERP, an index of cognitive control, was source localized to the cingulate cortex and to OFC. However, the source of the N2 in older children and in children who performed better on the EF tasks (regardless of age) was more anterior than that of younger children and children who performed poorly.

Theories of executive function development The functional characterization of EF introduced earlier provides a framework within which one can understand the hierarchical structure of EF and consider the way in which

DL-PFC

VL-PFC

RL-PFC

Figure 34.2 The human brain, showing various regions of prefrontal cortex on the lateral surface (left) and the medial surface (right). OFC, orbitofrontal cortex; ACC, anterior cingulate cortex;

554

cognition

ACC

OFC

DL-PFC, dorsolateral prefrontal cortex; VL-PFC, ventrolateral prefrontal cortex; RL-PFC, rostrolateral prefrontal cortex. (See plate 54.)

Figure 34.3 Right lateral and top views of the human brain, showing age-related declines in gray matter volume. (Reprinted with permission from N. Gogtay, J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, et al., 2004. Dynamic

mapping of human cortical development during childhood through early adulthood. Proceedings of the National Academy of Sciences of the USA 101:8174–8179.) (See plate 55.)

more basic cognitive functions (e.g., working memory, WM) contribute to particular aspects of EF (e.g., the role of WM in intending). There are several influential theoretical approaches to EF and its development—approaches that are de facto theories of PFC functional development. One approach in fact emphasizes the importance of WM, the ability to hold information in mind and use it to guide responding (e.g., Goldman-Rakic, 1987; Roberts and Pennington, 1996). Working memory clearly improves with age during childhood (e.g., Gathercole, 1998; Luciana and Nelson, 1998), and several authors have suggested that increases in WM capacity (i.e., the number of items one can hold in mind) may underlie increases in EF more generally (e.g., Case, 1992; Pascual-Leone, 1970). One recent example comes from Morton and Munakata (2002), who distinguish between active memory representations (i.e., WM) and latent memory traces. Active memory representations take the form of sustained, trial-specific activity in PFC, whereas latent memory traces are formed in more posterior cortex. Development of EF is hypothesized to occur as a result of increases in the strength of active memory representations, which, in turn, allow children to override prepotent tendencies mediated by latent memory traces. On the WCST, for example, older children may be better able to keep a new hypothesis in mind despite a tendency to keep sorting by the initial dimension.

Another common approach to EF and its development is to invoke the maturation of an inhibition mechanism— either alone or in combination with the development of WM. Following Luria (e.g., 1961), these accounts hold that children often perform poorly on measures of EF not because they have difficulty remembering an appropriate rule but rather because of a weak or inefficient system for inhibiting prepotent tendencies to respond in a particular way or attend to particular aspects of a stimulus (e.g., Barkley, 1997; Carlson, Moses, and Hix, 1998; Dempster, 1992; Diamond and Gilbert, 1989; Harnishfeger and Bjorklund, 1993; Kirkham, Cruess, and Diamond, 2003; White, 1965). In contrast to the WM accounts just described, these accounts imply that children who fail to switch on the WCST, for example, do in fact keep the new hypothesis in mind but simply cannot suppress the tendency to keep sorting by the initial dimension. Some authors have emphasized, however, that difficulties with inhibitory control interact with WM demands, and that it is the joint requirement that one keep information in mind while inhibiting prepotent tendencies that poses particular problems for children (e.g., Diamond, 2002; Carlson, Moses, and Breton, 2002; Roberts and Pennington, 1996). A third approach to the development of EF is to suggest that it depends on conceptual development—in particular, acquisition of concepts of intention or perspectives. For

zelazo, carlson, and kesek: executive function in childhood

555

example, Kloo and Perner (2003) suggested that children exhibit cognitive inflexibility because they do not understand that a single stimulus can be seen from two different perspectives (or “redescribed”). This account has difficulty explaining changes in EF in later childhood (because it is implausible that older children lack the requisite conceptual understanding), but it does address changes in EF during the preschool period as well as their close link to theory of mind (Perner and Lang, 1999; see later section “Correlates of executive function” for discussion). According to a fourth approach, the cognitive complexity and control theory–revised (CCC-r; Zelazo et al., 2003), age-related changes in EF are the result of increases in the hierarchical complexity of the rules that children can formulate, maintain in WM, and use when solving problems. These age-related increases in rule complexity are, in turn, made possible by age-related increases in the ease with which children can reflect on the rule systems that they represent—the highest level of consciousness that children are able to muster (Zelazo, 2004). In terms of the WCST, this account implies that children have difficulty reflecting on the multidimensional nature of the stimuli and the task, and formulating a rule such as “If I’m sorting by shape, then the yellow crosses go here; if I’m sorting by color, then they go there.” Research in developmental cognitive neuroscience has provided support for aspects of all four of these approaches, and indeed, it now seems likely that EF involves the orchestration of a variety of processes, including inhibitory control, WM, and reflection and rule complexity. It may also depend in part on conceptual changes, at least early in development. We will review this research in terms of a neurocognitive model of EF, one that attempts to capture the many aspects of EF, explain how they may be related, and show how they follow from the hierarchical structure of neural regions within PFC.

Age-related changes in executive function Traditionally, conceptualizations of EF have focused on its relatively “cool,” cognitive aspects, often associated with lateral PFC and elicited by relatively abstract, decontextualized problems (e.g., the WCST). Recently, however, there has been growing interest in the development of relatively “hot” aspects of EF, seen in situations that are emotionally and motivationally significant because they involve meaningful rewards or punishers (e.g., Happaney, Zelazo, and Stuss, 2004; Zelazo and Müller, 2002; cf. Metcalfe and Mischel, 1999; Miller and Cohen, 2001). Zelazo and Cunningham (2007) proposed a neural model of EF that positions hot and cool EF on a continuum of reflective processing and emphasizes the way in which relatively hot and relatively cool aspects of EF interact when one is solving

556

cognition

a problem. This model follows the course of information processing from the thalamus through PFC, beginning with the rapid emotional responses generated in the amygdala. These initial responses are then fed into OFC, which furnishes simple approach-avoidance (stimulus-reward) rules and is also involved in learning to reverse these rules. In many situations, these simple rules will suffice to produce an adequate response. If they fail to do so, however, the ACC acts as a performance monitor and signals the need for further, higher-level reprocessing of action-oriented rules in lateral PFC. Figure 34.4 (plate 56) illustrates the way in which regions of lateral PFC are hypothesized to correspond to rule use at different levels of complexity (Bunge and Zelazo, 2006). According to the model, lateral PFC-mediated reprocessing allows one to reflect on relatively simple rules (i.e., at a higher level of consciousness; Zelazo, 2004) and formulate higher-order rules that control the application of these simpler rules. As individuals engage in reflective reprocessing, ascend through levels of consciousness, and formulate more complex rule systems, they recruit an increasingly complex hierarchical network of PFC regions. This process can be viewed as a shift from hotter to cooler aspects of EF. In addition, however, this reprocessing fulfills the functions of WM (i.e., keeping rules in mind) and inhibitory control. (That is, more complex rules allow children to select relevant rules for guiding their behavior and avoid relying on outmoded rules, although of course doing so may be more or less difficult depending on the prepotency of the outmoded rules.) Moreover, the ability to engage in several degrees of reflection, which typically emerges during the preschool years, is required to construe stimuli in terms of multiple dimensions. In the following subsections, we will examine research at each level of the proposed information-processing pathway. As will become clear, there are now numerous tasks for assessing EF in children at various ages (e.g., Carlson, 2005). More will be said about measurement in a subsequent section, but for now we note simply that there are no “pure” behavioral measures of particular neural regions. Nonetheless, Carlson and Moses (2001) make a useful distinction, based on a factor analysis of EF tasks, between “delay tasks,” which require children to delay a prepotent response, and “conflict tasks,” which require children to make a novel response while inhibiting a conflicting, prepotent response. Delay tasks, which typically involve deciding to avoid making a prepotent response (i.e., to delay responding instead of immediately approaching a reward), provide a relatively straightforward assessment of OFC function (e.g., Berlin, Rolls, and Kischka, 2004; Monterosso et al., 2001), whereas conflict tasks also assess the function of lateral areas of PFC.

Figure 34.4 A hierarchical model of rule representation in PFC. A lateral view of the human brain is depicted at the top of the figure, with regions of PFC identified by the Brodmann areas (BA) that comprise them: orbitofrontal cortex (BA 11), ventrolateral PFC (BA 44, 45, 47), dorsolateral PFC (BA 9, 46), and rostrolateral PFC (BA 10). The PFC regions are shown in various shades of gray, indicating which types of rules they represent. Rule structures are depicted below, with darker shades of gray indicating increasing levels of rule complexity. The formulation and maintenance in working memory of more complex rules depends on the reprocess-

ing of information through a series of levels of consciousness, which in turn depends on the recruitment of additional regions of PFC into an increasingly complex hierarchy of PFC activation. Abbreviations: S, stimulus; check mark, reward; cross, nonreward; R, response; C, context, or task set. Brackets indicate a bivalent rule that is currently being ignored. (Reprinted with permission from S. Bunge and P. D. Zelazo, 2006. A brain-based account of the development of rule use in childhood. Current Directions in Psychological Science 15:118–121.) (See plate 56.)

Orbitofrontal Cortex: Approach-Avoidance Decisions Orbitofrontal cortex function has been assessed using a variety of measures, although perhaps the most common is object reversal, in which one learns a simple discrimination between two objects and then the discrimination is reversed (the previously unrewarded object is rewarded and vice versa). Both nonhuman animals (e.g., Butter, 1969; Dias, Robbins, and Roberts, 1996; Iversen and Mishkin, 1970) and human patients (Fellows and Farah, 2003; Rolls et al., 1994) with OFC damage show deficits in reversal learning— they perseverate on the initial discrimination. Standard object-reversal tasks have proven sensitive to developmental changes in OFC function in very young children, with more complex versions of the task sensitive to changes across the life span (Overman, 2004). Findings of this sort suggest that OFC is involved in the reappraisal of the affective or motivational significance of stimuli (e.g., Rolls, 1999, 2004). According to this view, although the amygdala is primarily

involved in the initial learning of stimulus-reward associations (e.g., Killcross, Robbins, and Everitt, 1997; LeDoux, 1996), reprocessing these relations is the provenance of OFC. Delay of Gratification Development of OFC function has been assessed using a variety of delay-of-gratification paradigms, which measure children’s ability to forgo an immediate reward in favor of a larger reward later. In one classic version (Mischel, Ebbesen, and Zeiss, 1972), children are seated in front of an enticing treat while an experimenter leaves the room. If they wait for the experimenter to return, they get two treats; otherwise they get one. Early studies generally failed to find age differences within the preschool range, but did find differences within the school-age years (see Mischel, Shoda, and Rodriguez, 1989, for review). However, a number of variations of the standard delay-ofgratification paradigm have proven sensitive to developmental changes (and individual differences) throughout the

zelazo, carlson, and kesek: executive function in childhood

557

preschool years. For example, in one task, Kochanska and colleagues (1996) asked children to hold a candy on their tongue without eating it until told to eat it by the experimenter. In another task, they instructed children not to peek as the experimenter noisily wrapped a gift for them. Children’s ability to delay increased significantly from 3 to 4 years. Several recent studies have also examined delay of gratification in young children using a modified choice task. When given the option to choose between a small reward now or a larger reward later (prudence) and between a reward for self now or a reward for self and other later (altruism), 4- to 5-year-old children demonstrated significantly more prudence and altruism than 3-year-old children (Thompson, Barresi, and Moore, 1997). Prencipe and Zelazo (2005) used a version of this choice task to examine children’s delay of gratification for self and other (the experimenter). Three-year-olds typically chose the immediate reward for themselves and the delayed reward for the experimenter, suggesting that they are capable of adaptive decision making but have particular difficulty regulating approach behavior in motivationally significant situations. Rather than being governed by OFC under these circumstances, their behavior seems to be driven by the relatively automatic evaluations generated by the limbic system. Overriding these evaluations may be facilitated in part by the adoption of a third-person perspective on one’s own behavior (Barresi and Moore, 1996). Decision Making and Reward Learning Delay-ofgratification paradigms typically require relatively simple decision making: a larger reward is guaranteed if one is willing to wait. However, more complex affective decision making often requires individuals to make approachavoidance decisions in the face of uncertainty. One of the most commonly used measures of complex decision making in adults is the Iowa Gambling Task (Bechara et al., 1994). Kerr and Zelazo (2004) created a version of this task for children that included only two decks of cards, one advantageous and one disadvantageous, and presented reward and loss information in the form of happy and sad faces. Cards in the disadvantageous deck offered more rewards (M&Ms) on every trial but were associated with occasional (unpredictable) large losses. Three-year-olds performed poorly on this task, failing to develop a preference for the advantageous deck. Four- and 5-year-olds, however, were able to make advantageous decisions (see also Garon and Moore, 2004, 2007). Affective decision making evidently increases dramatically in the preschool years, although it continues to improve throughout childhood (Crone and van der Molen, 2004). Anterior Cingulate Cortex: Performance Monitoring A crucial element of EF is the ability to recognize when

558

cognition

actions do not have the intended effect, necessitating further processing. The relatively simple rule systems generated by OFC may not be able to manage complex or ambiguous situations and stimuli. Under these circumstances, a performance-monitoring system mediated by the ACC (Ridderinkhof et al., 2004) signals the need for more elaborate processing dependent on lateral PFC. Typically, younger children make more errors on tasks requiring EF than adults and may be less aware of them. One way to assess performance monitoring is by looking at error detection and correction. In some situations, children’s ability to detect errors seems to develop in advance of their ability to correct them (e.g., Gelman and Meck, 1983). For example, Bullock and Lütkenhaus (1984) found that 18- and 24-month-old children could reliably distinguish between correctly and incorrectly built towers, even when they themselves failed to build the towers correctly. In other situations, however, error detection and correction appear to be more closely linked. Jacques and colleagues (1999) presented 3year-old children with the Dimensional Change Card Sort (DCCS; see figure 34.5 and plate 57), in which children are shown two target cards (e.g., a blue rabbit and a red boat) and asked to sort a series of bivalent test cards (e.g., red rabbits and blue boats) first according to one dimension (e.g., color), and then according to the other (e.g., shape). Most 3-year-olds perseverate during the postswitch phase, continuing to sort test cards by the first dimension (e.g., Zelazo et al., 2003). In order to assess error detection, Jacques and colleagues (1999) also asked children to evaluate the sorting of a puppet. When 3-year-olds watched the puppet perseverate, they judged the puppet to be correct. When they saw the puppet sort correctly, they judged the puppet to be wrong, which suggests that performance and error detection are closely aligned in this task. Another way to assess performance monitoring is to examine whether children, after making an error, respond more slowly on subsequent trials. Backen Jones, Rothbart, and Posner (2003) found that as 3- to 4-year-old children’s performance on a Simon Says–type task improved, they also tended to show posterror slowing. Using a more fast-paced task-switching paradigm, however, Davidson and colleagues (2006) failed to find evidence of posterror slowing in 4- to 5-year-old children. In adults, performance monitoring may be studied by measuring the ERN (error-related negativity), a negativegoing ERP component that probably originates in the ACC (Falkenstein et al., 1991; Gehring et al., 1993). However, athough ERNs are consistently observed in adults when an error is committed, they are not typically seen in young children, even when children seem to be aware that they are making errors (Segalowitz and Davies, 2004). Rather, the ERN seems to emerge in middle childhood and develop into adolescence (e.g., Santesso, Segalowitz, and Schmidt, 2006).

Figure 34.5 Sample target and test cards in the standard version of the Dimensional Change Card Sort (DCCS). (Reprinted with permission from P. D. Zelazo, 2006. The Dimensional Change

Card Sort [DCCS]: A method of assessing executive function in children. Nature Protocols 1:297–301.) (See plate 57.)

zelazo, carlson, and kesek: executive function in childhood

559

The protracted developmental course of the ERN suggests that the ACC continues to develop beyond childhood, and this may correspond to the slow development of performance monitoring. Nonetheless, further work involving a wider range of paradigms and measures (e.g., electrodermal activity) may be required to determine whether precursors of the ERN can be detected in early childhood. Ventrolateral and Dorsolateral PFC: Rule Use at Various Levels of Complexity The activation of ACC, signaling that the simple approach-avoidance tendencies generated by the OFC have not solved the present problem, may lead to the reprocessing of rules in lateral areas of PFC. This reprocessing corresponds to reflection and permits the formulation of increasingly complex rules (maintained in WM) for guiding behavior. Both VL-PFC and DL-PFC have been consistently implicated in the retrieval, maintenance, and use of conditional stimulus-response rules in both lesion and fMRI studies (Bunge, 2004). They are also sensitive to rule complexity, showing more activation for bivalent rules than for univalent rules (Crone, Wendelken, et al., 2006). Dorsolateral PFC appears to be especially important, however, when participants must use one set of bivalent rules while ignoring competing alternatives (e.g., MacDonald et al., 2000; see also Diamond, 2002). With age, children are able to use increasingly complex representations to guide their actions, and the development of this ability is often assessed using conflict tasks that require goal-directed behavior in the face of interference from prepotent tendencies. When studying infants, researchers have relied on search tasks, such as delayed response (Hunter, 1917) and A-not-B (Piaget, 1954). In a typical A-not-B task, infants watch as an object is placed at one of two or more hiding locations (i.e., at location A versus location B), a delay is imposed, and then infants are allowed to search. After a number of trials at location A, infants watch as the object is hidden at location B. Nine-month-old infants often search perseveratively at location A—evidently failing to keep a representation of the object at its current location (i.e., the goal) in mind and use it to guide a search. Instead, their behavior seems to be determined by the stimulus-reward association established during the A trials. Marcovitch and Zelazo (1999) conducted a meta-analysis of research on this task that revealed (inter alia) that errors were more likely to occur after longer delays and that they increased as a function related to the number of A trials. More recent work by Marcovitch, Zelazo, and Schmuckler (2002) suggests that the probability of making an A-not-B error is actually an inverted-U-shaped function of the number of A trials, perhaps reflecting the combined influences of habit strength and the cumulative probability of engaging in lateral-PFC-mediated reflection on the task. Although the A-not-B task is typically used with infants, versions of this paradigm appear sensitive

560

cognition

to developmental changes in toddlers and preschool-age children (Espy et al., 1999; Marcovitch and Zelazo, 2006; Sophian and Wellman, 1983; Spencer, Smith, and Thelen, 2001; Zelazo, Reznick, and Spinazzola, 1998). Extensive research has implicated PFC in the ability to search for hidden objects (e.g., Baird et al., 2002; Bell and Fox, 1992; Diamond and Goldman-Rakic, 1989). In an influential early study, Diamond and Goldman-Rakic (1989) found that adult monkeys with bilateral DL-PFC lesions (versus controls) showed poor A-not-B performance, similar to young human infants. More recently, Baird and colleagues (2002) used near-infrared spectroscopy (NIRS) to compare changes in cerebral blood flow in infants (5 to 12 months) who reliably searched for hidden objects and infants who did not. Infants who searched showed an increase in blood flow (total hemoglobin) in PFC, whereas those who did not search showed a decrease from baseline. Methods of studying lateral PFC function beyond infancy often involve providing children with verbal instructions and examining the circumstances in which they can follow these instructions—the so-called rule use paradigm pioneered by Luria. For example, Luria (e.g., 1959) reported that 2-yearolds often failed to obey a single, conditional rule (e.g., “When the light flashes, you will press the ball”). Younger 2-year-olds often simply ignored the conditional prerequisite of the rule and acted immediately. Older 2-year-olds, in contrast, refrained from responding until the first presentation of the light, although many of them then proceeded to respond indiscriminately. Following a single rule involves keeping in mind a representation of a relation between a stimulus and a response, and considering this relation relative to a goal (e.g., the goal of pleasing the experimenter). Whereas infants may be able to keep a simple goal in mind, as required in the A-not-B task, 2-year-olds may also come to represent a conditionally specified means for obtaining that goal. Following Luria’s seminal work on the subject (see Zelazo and Jacques, 1997, for review), Zelazo and colleagues studied the development of rule use in preschoolers using a number of tasks that vary in complexity. One task is the DCCS, described earlier (see figure 34.5). Three-year-olds perseverate on this task despite being told the new rules on every trial and despite correctly answering questions about the postswitch rules (e.g., “Where do the rabbits go in the shape game?”). They also perseverate despite being able, at this age, to keep four ad hoc rules in mind. In contrast, by 5 years of age, most children switch immediately on the DCCS when instructed to do so. Like adults, they seem to recognize that they know two ways of sorting the cards: “If I’m playing the color game, then if it’s a red rabbit, then it goes here ...”. Variants of this task have been used extensively to test hypotheses about EF and its development in early childhood (e.g., Bialystok, 1999; Bohlmann and Fenson, 2005;

Diamond, Carlson, and Beck, 2005; Dick, Overton, and Kovacs, 2005; Kirkham, Cruess, and Diamond, 2003; Kloo and Perner, 2005; Munakata and Yerys, 2001). For example, by changing aspects of the test and target cards between pre- and postswitch phases, it has been possible to demonstrate that 3-year-olds’ difficulties on the postswitch phase are a function of both the requirement that they inhibit attention to particular stimulus features that were relevant during the preswitch phase (e.g., rabbits and boats) and the requirement that they disinhibit attention to stimulus features that were actively ignored (e.g., red and blue; Müller et al., 2006). The DCCS is one of the simplest paradigms to assess task switching. The use of more complex paradigms across the life span suggests that the ability to switch rapidly between bivalent pairs of rules continues to improve beyond 5 years of age, throughout adolescence into early adulthood (Cepeda, Kramer, and Gonzalez de Sather, 2001; Crone et al., 2006; Davidson et al., 2006; Frye, Zelazo, and Palfai, 1995; Zelazo, Craik, and Booth, 2004). Cepeda, Kramer, and Gonzalez de Sather (2001) looked at task-switching performance over the life span (ages 7–82). Switch costs (above and beyond perceptual speed, WM, and nonswitch reaction time) decreased from childhood into young adulthood and stayed fairly constant until about 60 years of age. As the difficulty of a task increases, adults tend to slow response speed in order to maintain accuracy. Children, however, do not demonstrate this speed/accuracy trade-off; they continue to respond quickly at the expense of accuracy (Davidson et al., 2006). Another task that has been used to study the development of EF in preschool children is the Day-Night Stroop task (Gerstadt, Hong, and Diamond, 1994). Based on the Stroop color-word task (Stroop, 1935), this task requires young children to say “night” in response to picture of a sun and “day” in response to a picture of a moon (i.e., to use one pair of rules despite a tendency to use an incompatible pair). A similar version requires children to point to a green card when the experimenter says “snow” and to a white card when the experimenter says “grass” (Carlson and Moses, 2001). Gerstadt, Hong, and Diamond (1994) found that the Day-Night Stroop was difficult for young children; 3- and 4-year-old children made more errors than older children and took longer to respond. However, even young children performed well when asked to say “night” and “day” in response to two abstract designs—in the absence of interference from prepotent stimulus-response associations. Children also perform well when asked to say “dog” and “pig” in response to the pictures of day and night (Diamond, Kirkham, and Amso, 2002; see also Simpson and Riggs, 2005). Thus, although even young children are able to use a single pair of rules to guide responding when there is no conflicting response, the need to select such rules from among compet-

ing alternatives, keep them in mind, and follow them makes this task difficult. Other measures of EF suitable for use with preschoolers include variations of Luria’s (1966) tapping task, the Bear/Dragon task (Backen Jones, Rothbart, and Posner, 2003; Reed, Pien, and Rothbart, 1984), and the Whisper Task (Kochanska et al., 1996). Working Memory The hypothesis that DL-PFC plays a key role in following bivalent rules—using one pair of rules while ignoring a competing alternative—is consistent with its well-documented role in WM. That is, WM involves working on some information (e.g., trial-unique information) while ignoring other information (e.g., information from previous trials or canonical associations). Working memory has been linked to DL-PFC function using a number of different methods, including fMRI and lesion studies (Braver et al., 1997; Diamond and Goldman-Rakic, 1989; Smith and Jonides, 1999). Although there is little direct work on the neural basis of WM development with very young children, work with older children (9–11 years) provides evidence for the involvement of ACC and DL-PFC in WM (Casey et al., 1995). There is also evidence that maturation of white matter in PFC is associated with increased WM capacity between 8 and 18 years (Nagy, Westerberg, and Klingberg, 2004). A number of tasks, thought to rely heavily on DL-PFC function, have been used to examine WM across the life span, including delayed response (Diamond and Doar, 1989; Jacobsen, 1936; Goldman-Rakic, 1987), Delayed Alternation (Espy et al., 2001), items from the Cambridge Neuropsychological Test Automated Battery (CANTAB; Luciana and Nelson, 1998), and the Self-Ordered Pointing task (SOP; Petrides and Milner, 1982). On each trial of the SOP task, participants must point to a different picture in a constantly changing spatial arrangement of the same set of pictures. Participants must maintain a running tally of the pictures they have already pointed to and use this information to make a correct response (while avoiding interference from repetition priming). Hongwanishkul et al. (2005) administered an SOP task to young children and found improvements in performance between the ages of 3 and 5 years. Selective Attention Dorsolateral PFC also plays a key role in selective attention, which is typically assessed using measures of rule use. Posner and colleagues (e.g., Fan et al., 2002; Posner, 1994; Posner and Petersen, 1990) described an executive attention system involving ACC and lateral PFC, and proposed that this system is involved in the topdown direction and maintenance of attention, often when there is conflict among possible responses. There are clear age-related improvements in selective attention during early childhood (e.g., Davidson et al., 2006;

zelazo, carlson, and kesek: executive function in childhood

561

Enns and Cameron, 1987). For example, Gerardi-Caulton (2000) used a spatial-conflict (Simon) task with children ages 24, 30, and 36 months. Children were presented with a shape (e.g., a star) on a screen and instructed to press a response button that matched the picture. On some trials, the location of the response button and the location of the picture on the screen were congruent, and on other trials they were incongruent. All children performed more slowly and less accurately on incongruent trials, and this effect was more pronounced for younger children. Rueda and colleagues (2004) studied the development of selective attention in children between the ages of 6 and 10 years using a flanker task (Attention Network Test). In this task children were presented with a row of five fish and told to press an arrow that corresponded to the left-right orientation of the middle fish. In congruent trials the fish all faced the same way, and on incongruent trials the middle fish faced the opposite direction. Results showed improvement in reaction time and in accuracy on incongruent trials from ages 6 to 7. Rueda, Posner, and colleagues (2004) also completed an ERP study with 4-year-old children and adults using this same paradigm. Differences in brain activity between the congruent and incongruent condition were delayed in children relative to adults, suggesting that children may be slower to respond in situations requiring selective attention in part because they are slower to recognize conflict (i.e., because of limitations in ACC-mediated performance monitoring). Rostrolateral PFC: Task Set Selection Whereas VLPFC and DL-PFC are thought to be important for the representation and maintenance of rules and DL-PFC is thought to play a special role in implementing rules in the face of interference, RL-PFC is hypothesized to be involved in reflection on task sets, as when switching between two abstract rules (Bunge, Wallie, et al., 2005; Crone et al., 2006) or coordinating hierarchically embedded goals (Koechlin et al., 1999). There is evidence that RL-PFC may interact with different parts of prefrontal cortex (i.e., VL-PFC or DL-PFC) depending on the type of task involved (Sakai and Passing-ham, 2003, 2006) and, we would argue, depending on the complexity of the rule systems involved. Rostrolateral PFC activity has been observed across a number of neuropsychological studies employing tasks requiring complex reasoning and planning, including the WCST (Berman et al., 1995; Nagahama et al., 1996), the Raven Progressive Matrices test (Prabhakaran et al., 1997), tests of analogical reasoning (Bunge, Wendelken et al., 2005), and the Tower of London (Baker et al., 1996). For example, the Tower of London requires participants to arrange a number of colored balls, using a minimum number of moves, to match a specified arrangement. To perform well, one must coordinate a complex set of abstract rules

562

cognition

specifying the way the balls can be moved, and plan several moves ahead. Baker and colleagues (1996) observed increased activation in RL-PFC as adult participants attempted to reason through Tower of London tasks of increasing difficulty. Although each of these tasks involves complex reasoning, variations have been administered to children. For example, versions of the Tower of London have proven sensitive to developmental changes throughout the preschool years and through middle childhood (Carlson, Moses, and Claxton, 2004; Luciana and Nelson, 1998; Welsh, 1991). Tasks requiring analogical reasoning, or the ability to integrate and transfer knowledge across contexts, have also been developed for use with children. Holyoak, Junn, and Billman (1984) found that although even preschool-age children were capable of transferring solutions contained in a story to a problem-solving task, older children were able to generate more subtle analogies. Although there is limited research exploring the neural basis of these tasks in children, our view is that they are likely to be associated with the development of RL-PFC. Another task that taps the higher-order reflection on task sets hypothesized to depend on RL-PFC function is the Flexible Item Selection Task (FIST; Jacques and Zelazo, 2001), a measure of EF adapted from Feldman and Drasgow’s (1951) Visual-Verbal Test. On each trial of the FIST, children are shown a three-card set (e.g., a purple fish, a pink fish, and a pink telephone) and required to select two cards that match each other on one dimension (e.g., shape: the purple fish and the pink fish), and then select a different pair of cards that match each other on a different dimension (e.g., color: the pink fish and the pink telephone). Four- and 5year-olds did not differ on the first selection. On the second selection, however, 4-year-olds were less accurate than 5year-olds, suggesting that although they understood the basic task instructions and could select the first pair correctly, they had trouble re-representing the pivot item (e.g., the pink fish) in a different way.

Correlates of executive function We have so far discussed a hierarchical model of PFC development that is likely to govern the observed age-related changes in EF task performance in early childhood. Next, we turn to investigations of the correlates of individual differences in EF that may be apparent at any age using developmentally sensitive measures. Examining the correlates of EF helps to contextualize it as a critically important aspect of cognitive and social development, and, furthermore, it has the potential to inform theories of direct and indirect influences upon EF and the neurocognitive mechanisms involved. Known correlates of EF include cognitive, socioemotional, and demographic variables.

Cognitive Correlates To the extent that EF refers to processes essential for purposeful, goal-directed problem solving, it ought to be implicated in all manner of conceptual problems, especially novel ones that call for the inhibition of automatic or established thoughts and responses. An area of conceptual development that has been studied extensively in relation to EF is theory of mind (ToM), which refers to the ability to attribute mental states to oneself and others and to recognize that mental representations (e.g., beliefs) can conflict with reality, change over time, and differ across individuals (e.g., Astington, 1993; Perner, 1991; Premack and Woodruff, 1978; Wellman, 1990). In the paradigmatic false-belief problem developed for assessing children’s ToM, young preschoolers tend to claim that a story character will search for an occluded object where the child him- or herself knows it to be in reality, instead of where the character had last seen the object; that is, they fail to appreciate that the character would hold a false belief. By 4 to 5 years of age, children typically overcome this bias to respond according to their own perspective (Wimmer and Perner, 1983; for a meta-analysis, see Wellman, Cross, and Watson, 2001). Understanding conflicting mental states is considered a core achievement in cognitive development that appears to be unique to human beings (e.g., Povinelli and O’Neill, 2000; Premack and Premack, 1996; Tomasello, Kruger, and Ratner, 1993). In addition to sharing a developmental timetable, ToM and EF are linked in more fundamental ways (Moses and Carlson, 2004). First, imaging studies indicate that PFC is also implicated in thinking about mental states (e.g., Amodio and Frith, 2006; Frith and Frith, 2003; Sabbagh and Taylor, 2000; Saxe, Carey, and Kanwisher, 2004; Siegal and Varley, 2002). Second, individuals with autism have deficits in both ToM and EF (e.g., McEvoy, Rogers, and Pennington, 1993; Russell, 1997; Zelazo et al., 2002). Third, Moses and Carlson (2004) posited that EF might influence ToM by placing constraints on both its expression and its emergence. On the one hand, preschool children might have rudimentary concepts of mind but have difficulty making effective use of these concepts because—in the absence of sufficiently developed EF—they cannot flexibly direct their attention to appropriate aspects of the situation (e.g., set aside what they think and put themselves in someone else’s shoes). On the other hand, a certain level of EF development may be critical for the very emergence of mental-state concepts (Russell, 1996). Children would need to have some ability to disengage attention from salient stimuli in order to entertain the existence of abstract mental representations. In contrast, however, Perner and Lang (1999; Perner, Lang, and Kloo, 2002) explained the correlation between EF and ToM differently; they proposed that ToM (or more specifically a concept of one’s own intention) must be acquired before children can reflect on and subsequently monitor and control their own behavior.

Numerous studies have established a robust association between individual differences in EF and ToM in Western samples (e.g., Carlson and Moses, 2001; Frye, Zelazo, and Palfai, 1995; Hala, Hug, and Henderson, 2003; Hughes, 1998a, 1998b; Perner, Lang, and Kloo, 2002; see Perner and Lang, 1999, for a meta-analysis) and in a Chinese sample (Sabbagh et al., 2006). Carlson and colleagues conducted a series of correlational investigations establishing the magnitude of the relation in 3- and 4-year-olds (e.g., r = .66; Carlson and Moses, 2001), and the specificity of the relation to conflict EF tasks (Carlson, Moses, and Breton, 2002; Carlson, Moses, and Claxton, 2004). They concluded that acquiring and expressing mental-state concepts requires both holding in mind the relevant perspectives (WM) and suppressing the irrelevant ones (inhibition)—putative DLPFC functions. Frye, Zelazo, and colleagues (e.g., Frye, Zelazo, and Palfai, 1995; Frye, Zelazo, and Burack, 1998) emphasized the RL-PFC function of selecting among conflicting perspectives. Furthermore, longitudinal studies have shown that the relation persists over time (Carlson, Mandell, and Williams, 2004; Hughes, 1998b) and that early individual differences in EF predict later-developing ToM better than the reverse, suggesting that EF is an important influence on the emergence of some aspects of mental-state understanding. Socioemotional Correlates Executive function is also implicated in the development of compliance, emotion regulation, and social competence. Effortful control, defined by Rothbart and Bates (1998) as the ability to voluntarily inhibit a dominant response to activate a subdominant response, is a dimension of temperament thought to be closely related to EF (e.g., Carlson and Moses, 2001; Kochanska et al., 1996). Kochanska and colleagues have conducted several studies examining the sociomoral correlates of effortful control. Both compliance with social demands and moral conscience (internalization of rules) were related to individual differences in effortful control, both concurrently and over time in the preschool years (e.g., Kochanska, Coy, and Murray, 2001; Kochanska, Murray, and Harlan, 2000). Several studies have also demonstrated a relation between high effortful control and low negative emotionality using a parent report of one or both constructs (e.g., Eisenberg et al., 1997; Kochanska et al., 1998), but surprisingly little research has examined the overlap between EF and control of emotional expression. However, Carlson and Wang (2007) found a significant correlation using behavioral measures of both constructs. This finding helps to illustrate how similar aspects of EF can be used in the service of controlling attention, action, and emotional expression (e.g., LeDoux, 1996; Zelazo and Cunningham, 2007). Interestingly, however, the study also uncovered a nonlinear relation between these

zelazo, carlson, and kesek: executive function in childhood

563

constructs, in which both low and high levels of temperamental inhibition (which may be associated with anxiety) were deleterious for the regulation of emotion in a social context. This research suggests that EF and temperament likely make interactive contributions to socioemotional development. Explaining individual differences in emotion regulation that appear in early childhood is an important undertaking because children who have difficulty managing emotions (e.g., anger) are at risk for developing behavioral disorders (Cole, Michel, and Teti, 1994; Denham et al., 2002; Dodge and Garber, 1991; Hughes, Dunn, and White, 1998), and indeed, deficits in attention and emotion regulation tend to co-occur in certain atypical and at-risk populations, such as children with ADHD (Barkley, 1997). For these reasons, it is perhaps not surprising that impairments in EF have been found in several studies of preschool-aged children with conduct problems (e.g., Hughes et al., 2000; Seguin and Zelazo, 2005; Speltz et al., 1999). Even within typically developing children, however, there is evidence that individual differences in EF are correlated with parent report of social competence and predictive of social skills in elementary school (Carlson, Mallory, and Beck, 2005; Hughes, Dunn, and White, 1998; Riggs, Blair, and Greenberg, 2003). Although very few long-term studies have been done, the stability and predictive value of at least one aspect of EF, delay of gratification, is particularly impressive. Mischel and colleagues found that choosing to wait for a larger reward at age 4 years positively predicted cognitive and social competence, effective coping skills, and performance on the Scholastic Aptitude Test in adolescence, independent of earlier intelligence scores (Mischel, Shoda, and Peake, 1988). Preschool delay of gratification also predicted more efficient attentional control on a go-no-go test at age 18 years (Eigsti et al., 2006), as well as social understanding, goal setting, and self-regulatory abilities at age 30 (Ayduk et al., 2000). In contrast, self-control deficiencies that persist into adulthood have been associated with criminal behavior, poor physical health, and interpersonal difficulties, in addition to being a central component of many psychiatric disorders (Strayhorn, 2002). Demographic Correlates Although the longitudinal evidence has been taken to suggest that individual differences in EF skills reflect stable personality traits (e.g., ego control and ego resiliency; Funder, Block, and Block, 1983), some investigators have found that the relation between EF and social functioning in childhood is moderated by distal environmental factors, such as neighborhood (e.g., Caspi et al., 2000), and by associated proximal risk factors such as maternal education and socioeconomic status (e.g., Ardila et al., 2005; Blair and Razza, 2007; Hughes and Ensor, 2005; Lengua, 2006; Mezzacappa 2004; Noble, Norman,

564

cognition

and Farah, 2005). For example, Noble, Norman, and Farah (2005) reported that parent education level among African American families affected kindergartners’ language and EF skills disproportionately compared to other major aspects of neurocognitive development. Sex differences also are sometimes found on tests of EF in young children. Differences most often favor females (e.g., Bjorklund and Kipp, 1996; Carlson and Moses, 2001; Isquith, Gioia, and Espy, 2004; Kochanska et al., 1996), although these findings are inconsistent with those showing poorer performance in girls on reversal learning and gambling tasks (Overman, 2004; Overman et al., 1996), as well as with sex differences in PFC development (De Bellis et al., 2001). This discrepancy suggests that there might be some sex-linked components of EF development, such as cool EF (abstract, decontextualized, strong reliance on verbal ability), which is better in girls, versus hot EF (flexible representation of the reinforcement value of stimuli), which is better in boys. In fact, in one study including several EF tasks, when verbal ability was held constant, the sex difference disappeared (Carlson, Mandell, et al., 2004). Girls also tend to perform better than boys on compliance measures (Kochanska, Murray, and Harlan, 2000), implicating socialization as a factor that might override biologically based differences in certain circumstances. Findings of interactions among demographic variables and EF in predicting child outcomes suggest that complex nature-nurture interactions must be emphasized for a full understanding of EF development (e.g., chapter 5 by Tarullo, Quevedo, and Gunnar, this volume).

Influences on executive function Studies of the correlates of EF point not only to the consequences of EF development, but also to potential influences on it. Influences, in turn, help point to the potential mechanisms by which environmental inputs contribute to PFC and EF development. Verbal labeling, bilingualism, symbolism, and specific training have each been shown to facilitate EF in childhood. Labeling Luria (1959, 1961) assessed the effects of labeling on the go-no-go task described earlier. When 3-year-olds were asked to accompany their manual responses (i.e., pressing on go trials) with self-directed commands such as “Press,” they were better able to regulate their responses. By contrast, when 3-year-olds were asked to accompany their nonresponses (i.e., withholding responding on nogo trials) with self-directed commands such as “Don’t press,” their performance deteriorated. Older children’s performance improved when they labeled both go and nogo trials. Luria argued that younger children can regulate their behavior using the expressive aspect of labels, but they have difficulty using semantic aspects when these aspects

conflict with the expressive aspects or with children’s prepotent tendencies. Müller and colleagues (2004) examined the effect of labeling on a rule-use task in which children were presented with colored Smarties candies on large cards whose colors mismatched the colors of the Smarties (e.g., a green Smartie on a yellow card). Children were then asked to turn around and retrieve a small card that matched the color of the card on which the Smartie was placed. Whereas 3-year-olds performed poorly (often retrieving small cards that matched the color of the Smartie rather than the color of the large card), 4- to 6-year-olds had no difficulty. When 3-year-old children were specifically asked to label the color of the card before making a selection, however, they performed near ceiling. A number of researchers have examined the effects of labeling on the DCCS (Kirkham, Cruess, and Diamond, 2003; Towse et al., 2000; Yerys and Munakata, 2006). For example, Kirkham, Cruess, and Diamond (2003) found that asking 3-year-old children to label the test card improved their performance. Similarly, using the FIST, Jacques and colleagues (2007) found that asking 4-year-olds to label their basis for selection 1 (e.g., “Why do those two pictures go together?”) improved their performance on selection 2. This improvement occurred whether children provided the label themselves or the experimenter generated it for them. These results suggest that labeling does not simply change the relative salience of stimuli and redirect children’s attention to the postswitch dimension, but instead may facilitate reflection on their initial construal of the stimuli. Bilingualism In addition to the effects of verbal labeling on EF, some research has suggested beneficial effects of duallanguage learning. Bialystok (2001) posited that a key aspect of EF, namely, inhibitory control over attentional resources, develops more rapidly in children with extensive experience shifting between languages. Several studies by Bialystok and colleagues, using a number of EF tasks (e.g., the DCCS), have supported this hypothesis (Bialystok, 1999; Bialystok and Martin, 2004). The developmental advantage was found to be approximately one year, but this advantage held up only in task versions that call for “conceptual inhibition,” for example, resisting attention to the previously relevant feature (e.g., color) in order to represent the newly relevant feature (e.g., shape). Subsequent studies with bilingual children from heterogeneous language backgrounds corroborated these results (e.g., Bialystok, Martin, and Viswanathan, 2005; Bialystok and Shapero, 2005). In each case, bilinguals were better than monolinguals at selectively attending to a stimulus in the presence of distracting information. Extending these findings, Carlson and Meltzoff (in press) found that kindergarten children who were native speakers of both English

and Spanish performed significantly better on a battery of EF tasks than monolingual (English) speakers in traditional schools and those enrolled in second-language immersion kindergarten. The bilingual advantage was significant for conflict EF tasks, which specifically call for managing conflicting attentional demands, whereas there was no advantage on delay EF tasks. Further research will be needed to examine early versus late exposure, including studies of the manner in which initial dual-language input influences brain development (Vaid, 2002). Symbolism Language is, at its core, an abstract symbol system for representing objects, events, and ideas for the purpose of shared communication about them. The evidence described thus far on the role of verbal labeling and bilingualism in EF raises the question whether there might be a more general function of symbolism influencing EF development. Symbols decontextualize concrete stimuli and can do so along a gradient from highly realistic (e.g., a highresolution photograph) to highly abstract (e.g., words). This decontextualization is akin to psychological distancing (Dewey, 1931/1985; Werner and Kaplan, 1963). To the extent that EF refers to goal-directed activity in the face of cognitive interference, symbols may mediate between stimulus and response and permit top-down control over impulses (Carlson and Zelazo, in press). There is some evidence that inducing a more abstract, symbolic mind-set enhances children’s delay of gratification and problem-solving abilities. Singer (1961) interviewed 6to 9-year-old children about their engagement in fantasy and administered a delay task in which they were required to sit or stand still for 15 minutes. The high-fantasy children were able to wait twice as long as the low-fantasy children (8 versus 4 minutes on average). Mischel and colleagues also found that children who fixed attention toward the appetitive aspects of the reward during delay of gratification (e.g., looking at or smelling treats) had shorter delay times (Mischel, Ebbesen, and Zeiss, 1972; Peake, Hebl, and Mischel, 2002; Rodriguez, Mischel, and Shoda, 1989; Sethi et al., 2000; see also Carlson and Beck, in press). In contrast, an experimental manipulation in which preschoolers were asked to pretend that the marshmallows in the experiment were “white fluffy clouds” helped to extend delay times substantially (Mischel, Ebbesen, and Zeiss, 1972). Metcalfe and Mischel (1999) concluded that a symbolic orientation diverts attention away from the “hot” (reward-focus) aspects of the situation. More recently, Carlson, Davis, and Leach (2005) developed a task to investigate the hypothesis that symbols provide degrees of distance from reality, which then might make it possible to reflect on the self and govern one’s responses more effectively. In the Less is More task, children are presented with two piles of candy, one large and one small, and must point to the small pile in order to obtain the large pile.

zelazo, carlson, and kesek: executive function in childhood

565

Three-year-olds, compared to 4-year-olds, have difficulty inhibiting their tendency to point to the preferred, larger reward; they have difficulty reversing the simple stimulusreward rules for the purposes of the task. Carlson, Davis, and Leach trained 3-year-olds on symbolic representations for the quantities of treats, in increasing degree of separation from reality before giving them the task (e.g., one-to-one correspondence with rocks versus a mouse and an elephant to stand for small and large amounts, respectively). Children in the symbol conditions performed better than children presented with real treats and improved as a function of the degree of symbolic distancing from the real rewards. More research is needed on the effects of children’s spontaneous symbolic activities on EF, but there is some evidence that individual differences in preschoolers’ pretense skills are significantly related to EF (Carlson and Davis-Unger, 2007). Furthermore, these findings, like the labeling effects, point to inner speech as a potential explanatory mechanism. A promising direction for future research is to investigate “virtual dialogues” in toddlers and their relevance to EF and related developments such as pretend play and the ability to assume multiple perspectives (Mead, 1934; Rochat, 2001). Training Recent work on neural plasticity (e.g., Huttenlocher, 2002; Neville, 1993) supports the hypothesis that improvements in EF may be a function of experiencedependent maturation of PFC, and that experience with EF tasks may lead to changes in neural activity. Although many studies of EF manipulate some aspect of the task (e.g., labeling, symbols) to elicit improved performance, only a few studies have attempted to improve performance through training or practice. In one, Dowsett and Livesey (2000) found that children who received training on two tasks requiring EF showed improvement in a go-no-go paradigm, suggesting that experience can improve EF in a relatively general way. Kloo and Perner (2003) provided further evidence that training in a specific EF task can lead to more generalized improvements in EF. In that study, children were trained on either the DCCS or on a standard false-belief task, receiving explanations and feedback in two training sessions. Children who were trained on the DCCS performed significantly better on this task after training, but they also improved their performance in the false-belief task. There is also evidence that training on behavioral paradigms can lead to changes in patterns of neural activity. For example, Rueda and colleagues (2005) administered a fish flanker task, described earlier, to 4- and 6-year olds, and they also measured EEG in these children. The N2 component, often associated with conflict-related activity in adults, was of special interest. After completing an initial session, children in the experimental group received five additional training sessions. At baseline, 6-year-olds showed evidence of a larger N2 during incongruent trials, whereas 4-year-olds

566

cognition

did not. After training, performance on the task improved, and 6-year-olds showed N2 activity similar to that observed in adults, while 4-year-olds began to show N2 activity similar to that observed in untrained 6-year-olds. The possibility that EF can be trained has clear clinical implications. Developmental disorders (e.g., conduct disorder, attention-deficit/hyperactivity disorder) can often be characterized as impairments in EF. However, if EF skills can be trained, and even generalized to other contexts, children who have difficulty controlling emotion and behavior may be helped. Preliminary evidence comes from Klingberg and colleagues (2005), who trained children diagnosed with ADHD in a WM task and observed significant improvement in the task, as well as more generalized improvements across symptoms.

Measurement issues A cornerstone of scientific inquiry into EF development is the measurement of EF. Advances in reliable and valid measurement of EF in early childhood have kept pace with—and contributed to—theoretical advances in the field. Nevertheless, several important issues remain at the frontier of the study of EF, including age-appropriate task innovation, coherence of measures, and EF as a componential versus unitary construct. Recent years have seen a proliferation of measurement tools that are based on normative cognitive development in children, in addition to those derived from models of brain damage in adults. Carlson (2005) provided an overview of many of these measures used with typically developing preschoolers, finding support for the notion of rapid progress in EF in the preschool period and identifying the most developmentally sensitive measures. Very few of the preschool measures have been subjected to test/retest reliability analyses. However, a study of the psychometric properties of 11 impulsivity measures (including Stroop and delay tasks) in typical and behaviorally disordered children ages 6 to 16 showed that eight (73%) of the tasks had at least one dependent variable with adequate temporal stability (Kindlon, Mezzacappa, and Earls, 1995). Developmental change, in and of itself, does not necessarily imply coherence of the EF measures that are available. When multiple EF tasks have been administered to preschool children, task intercorrelations are often moderate to strong, although not universally so. Construct validity varies according to (1) child age, (2) the number of tasks included, and (3) the types of tasks included. With respect to age, Cronbach’s alpha levels for batteries of EF tasks tend to be greater than .70 in 3- to 5-year-olds but relatively low (around .50) in toddlers (Carlson, Mandell, et al., 2004; Hughes and Ensor, 2005; Kochanska, Coy, and Murray, 2001). This difference is likely due to the paucity of adequate

validated measurement tools thus far available for toddlers, rather than a genuine lack of EF coherence, judging by the impressively strong stability of individual differences in EF from ages 2 to 4 years (Carlson et al., 2004; Kochanska, Coy, and Murray, 2001). A related source of low coherence is having too few measures of the construct. Carlson (2003) argued that because there are no “process pure” measures of EF, multimethod approaches are valuable for triangulating the processes involved (see Rushton, Brainerd, and Pressley, 1983). Indeed, studies of individual differences in EF in preschoolers have found consistently more robust correlations between the variables of interest and composite EF scores as opposed to individual EF tasks (e.g., Carlson and Meltzoff, in press; Carlson and Moses, 2001; Hughes, 1998; Kochanska et al., 1996; Perner and Lang, 1999). This finding is especially a concern with EF because it is so multifaceted: Low coherence can result not only from too few measures, but also from inclusion of too many tasks reflecting distinct aspects of EF. One dimension that appears to be important in distinguishing multiple measures of EF is the extent to which a task places demands upon WM. For example, Carlson, Moses, and Breton (2002) found that conflict tasks were significantly correlated with WM capacity whereas delay was not, and they suggested that conflict tasks might involve a combination of inhibition and WM whereas delay involves inhibition but relatively low WM demands (see also Espy and Bull, 2005; Gerstadt, Hong, and Diamond, 1994; Welsh, Pennington, and Groisser, 1991). A second dimension hypothesized to be important in distinguishing major aspects of EF is the extent to which the task is motivationally significant, that is, the hot/cool distinction (Zelazo and Müller, 2002; Zelazo and Cunningham, 2007; see also Metcalfe and Mischel, 1999). Kindlon, Mezzacappa, and Earls (1995), for example, conducted an age-corrected factor analysis of impulsivity measures in older children and found evidence for a distinction between cognitive inhibitory control versus a motivational component relating to insensitivity to punishment. This distinction has been less thoroughly investigated to date in preschoolers, although recent task innovations designed to measure hot EF (e.g., gambling tasks) have made it easier to do so (e.g., Carlson, Davis, and Leach, 2005; Garon and Moore, 2004; Kerr and Zelazo, 2004). Both hot and cool EF develop over the course of the preschool period, and there is some evidence that they overlap at the level of individual differences (e.g., Carlson et al., 2005; Hongwanishkul et al., 2005). It remains possible, however, that EF is a relatively undifferentiated construct in young children but becomes more modular with age. This possibility would be in keeping with the shift from diffuse to more focal and efficient frontal networks associated with EF as children develop (e.g., Durston et al., 2006). This line of research is still in its infancy and will benefit from further development of tightly

controlled tasks that tap into the core aspects of EF and can be administered in a variety of settings across the preschool period, to provide age norms and to tell us more about the early organization and differentiation of EF.

Summary Although the development of EF has long been associated with PFC development, recent advances in developmental cognitive neuroscience have begun to elucidate the nature of this relation. Zelazo and Cunningham (2007) proposed an integrated, hierarchical model of PFC development that attempts to account for this new research. This model suggests that goal-directed problem solving begins with quick, unreflective limbic responses that feed into OFC. If the relatively simple approach-avoidance rules generated by OFC prove inadequate, ACC signals the need to reprocess these rules by way of thalamocortical loops involving regions of lateral PFC. With the development of lateral areas of PFC, children are able to formulate and use increasingly complex rule hierarchies (Bunge and Zelazo, 2006). This experiencedependent maturation of PFC allows children to employ top-down control to overcome the bottom-up response tendencies that often guide the behavior of very young children. The development of EF must be considered in the context of the wider range of changes that occur in early childhood. Although it is often difficult to determine the direction of influence, EF has implications for many aspects of social and emotional development, including the emergence of ToM, emotion regulation, and a wide range of adjustment outcomes. Several influences have been identified that may contribute to the development of individual differences in EF, such as labeling, bilingualism, symbolism, and explicit training. The far-reaching consequences of EF highlight the importance of a complete understanding of its developmental course. Numerous developmental disorders are characterized by deficits in EF, a fact which suggests that it is fragile and easily disrupted. The continued development of methods and theory promises a more comprehensive, integrated understanding of the development of EF and its myriad implications. acknowledgments

The preparation of this chapter was supported in part by grants from NSERC of Canada (to PDZ) and NIH R01 HD51495 (to SMC). We thank the editors for providing helpful comments on an earlier draft of this manuscript. REFERENCES

Amodio, D., and C. Frith, 2006. Meeting of minds: The medial frontal cortex and social cognition. Nature Rev. Neurosci. 7:268– 277.

zelazo, carlson, and kesek: executive function in childhood

567

Ardila, A., M. Rosselli, E. Matute, and S. Guajardo, 2005. The influence of the parents’ educational level on the development of executive functions. Dev. Neuropsychol. 28:539–560. Astington, J. W., 1993. The Child’s Discovery of the Mind. Cambridge, MA: Harvard University Press. Ayduk, O., R. Mendoza-Denton, W. Mischel, G. Downey, P. K. Peake, and M. Rodriguez, 2000. Regulating the interpersonal self: Strategic self-regulation for coping with rejection sensitivity. J. Pers. Soc. Psychol. 79:776–792. Backen Jones, L., M. Rothbart, and M. Posner, 2003. Development of executive attention in preschool children. Dev. Sci. 6: 498–504. Baird, A. A., J. Kagan, T. Gaudette, K. Walz, N. Hershlag, and D. Boas, 2002. Frontal lobe activation during object permanence: Evidence from near infrared spectroscopy. NeuroImage 16:1120–1126. Baker, S. C., R. D. Rogers, A. M Owen, C. D. Frith, R. J. Dolan, R. S. J. Frackowiak, and T. W. Robbins, 1996. Neural systems engaged by planning: A PET study of the Tower of London task. Neuropsychologia 34:515–526. Barkley, R. A., 1997. Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD. Psychol. Bull. 121:65–94. Barresi, J., and C. Moore, 1996. Intentional relations and social understanding. Behav. Brain Sci. 19:107–154. Bechara, A., A. R. Damasio, H. Damasio, and S. W. Anderson, 1994. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 50:7–15. Bell, M. A., and N. A. Fox, 1992. The relations between frontal brain electrical activity and cognitive development during infancy. Child Dev. 63:1142–1163. Berlin, H. A., E. T. Rolls, and U. Kischka, 2004. Impulsivity, time perception, emotion and reinforcement sensitivity in patients with orbitofrontal cortex lesions. Brain 127:1108–1126. Berman, K. F., J. L. Ostrem, C. Randolph, J. Gold, T. E. Goldberg, R. Coppola, et al., 1995. Physiological activation of a cortical network during performance of the Wisconsin Card Sorting Test: A positron emission tomography study. Neuropsychologia 33:1027–1046. Bialystok, E., 1999. Cognitive complexity and attentional control in the bilingual mind. Child Dev. 70:636–644. Bialystok, E., 2001. Bilingualism in Development: Language, Literacy, and Cognition. New York: Cambridge University Press. Bialystok, E., and M. M. Martin, 2004. Attention and inhibition in bilingual children: Evidence from the dimensional change card sort task. Dev. Sci. 7:325–339. Bialystok, E., M. M. Martin, and M. Viswanathan, 2005. Bilingualism across the lifespan: The rise and fall of inhibitory control. Int. J. Bilingualism 9:103–119. Bialystok, E., and D. Shapero, 2005. Ambiguous benefits: The effect of bilingualism on reversing ambiguous figures. Dev. Sci. 8:595–604. Bjorklund, D. F., and K. Kipp, 1996. Parental investment theory and gender differences in the evolution of inhibition mechanisms. Psychol. Bull. 120:163–188. Blair, C., and R. P. Razza, 2007. Relating effortful control, executive function, and false-belief understanding to emerging math and literacy ability in kindergarten. Child Dev. 78:647–663. Bohlmann, N. L., and L. Fenson, 2005. The effects of feedback on perseverative errors in preschool aged children. J. Cogn. Dev. 6:119–131. Braver T. S., J. D. Cohen, L. E. Nystron, J. Jonides, E. E. Smith, and D. C. Noll, 1997. A parametric study of prefrontal

568

cognition

cortex involvement in human working memory. NeuroImage 5: 49–62. Bullock, M., and P. Lütkenhaus, 1988. The development of volitional behavior in the toddler years. Child Dev. 59:664– 674. Bunge, S. A. 2004. How we use rules to select actions: A review of evidence from cognitive neuroscience. Cogn. Affective Behav. Neurosci. 4:564–579. Bunge, S. A., N. M. Dudukovic, M. E. Thomason, C. J. Vaidya, and J. D. E. Gabrieli, 2002. Development of frontal lobe contributions to cognitive control in children: Evidence from fMRI. Neuron 33:301–311. Bunge, S. A., J. D. Wallis, A. Parker, M. Brass, E. A. Crone, E. Hoshi, and K. Sakai, 2005. Neural circuitry underlying rule use in humans and non-human primates. J. Neurosci. 9:10347– 10350. Bunge, S. A., C. Wendelken, D. Badre, and A. D. Wagner, 2005. Analogical reasoning and prefrontal cortex: Evidence for separable retrieval and integration mechanisms. Cerebral Cortex 15:239–249. Bunge, S., and P. D. Zelazo, 2006. A brain-based account of the development of rule use in childhood. Curr. Dir. Psychol. Sci. 15:118–121. Butter, C., 1969. Perseveration in extinction and in discrimination reversal tasks following selective frontal ablations in Macaca mulatta. Physiol. Behav. 4:163–171. Carlson, S. M., 2003. Executive function in context: Development, measurement, theory, and experience. Monogr. Soc. Res. Child Dev. 68(3, Serial No. 274):138–151. Carlson, S. M., 2005. Developmentally sensitive measures of executive function in preschool children. Dev. Neuropsychol. 28: 595–616. Carlson, S. M., and D. M. Beck, in press. Symbols as tools in the development of executive function. In A. Winsler, C. Fernyhough, and I. Montero, eds., Private Speech, Executive Functioning, and the Development of Verbal Self-regulation. New York: Cambridge University Press. Carlson, S. M., and A. C. Davis-Unger, 2007. Executive function and pretense in preschool children. Unpublished manuscript. University of Washington. Carlson, S. M., A. C. Davis, and J. G. Leach, 2005. Less is more: Executive function and symbolic representation in preschool children. Psychol. Sci. 16:609–616. Carlson, S. M., J. Mallory, and D. M. Beck, 2005. Executive functioning and social competence. In D. Lieberman, organizer, The Relationship between Executive Functioning and Motivation, Emotional Regulation, and Everyday Behaviors. Paper symposium presented at the annual meeting of the Jean Piaget Society, Vancouver, Canada, June. Carlson, S. M., D. Mandell, and L. Williams, 2004. Executive function and theory of mind: Stability and prediction from ages 2 to 3. Dev. Psychol. 40:1105–1122. Carlson, S. M., and A. N. Meltzoff, in press. Bilingual experience and executive functioning in young children. Dev. Sci. Carlson, S. M., and L. J. Moses, 2001. Individual differences in inhibitory control and theory of mind. Child Dev. 72:1032– 1053. Carlson, S. M., L. J. Moses, and C. Breton, 2002. How specific is the relation between executive function and theory of mind? Contributions of inhibitory control and working memory. Infant Child Dev. 11:73–92. Carlson, S. M., L. J. Moses, and L. J. Claxton, 2004. Individual differences in executive functioning and theory of mind: An

investigation of inhibitory control and planning ability. J. Exp. Child Psychol. 87:299–319. Carlson, S. M., L. J. Moses, and H. R. Hix, 1998. The role of inhibitory processes in young children’s difficulties with deception and false belief. Child Dev. 69:672–691. Carlson, S. M., and T. Wang, 2007. Inhibitory control and emotion regulation in preschool children. Cogn. Dev. 22:489– 510. Carlson, S. M., and P. D. Zelazo, in press. Symbolic thought. In M. Haith and J. Benson, eds., Encyclopedia of Infant and Early Childhood Development. Amsterdam: Elsevier. Case, R., 1992. The role of the frontal lobes in the regulation of cognitive behavior. Brain Cogn. 20:51–73. Casey, B. J., J. D. Cohen, P. Jezzard, R. Turner, D. C. Noll, R. J. Trainor, et al., 1995. Activation of prefrontal cortex in children during a nonspatial working memory task with functional MRI. NeuroImage 2:221–229. Casey, B. J., J. N. Giedd, and K. M. Thomas, 2000. Structural and functional brain development and its relation to cognitive development. Biol. Psychol. 54:241–257. Caspi, A., A. Taylor, T. E. Moffitt, and R. Plomin, 2000. Neighborhood deprivation affects children’s mental health: Environmental risks identified in a genetic design. Psychol. Sci. 11: 338–342. Cepeda, N. J., A. F. Kramer, and J. C. M. Gonzalez de Sather, 2001. Changes in executive control across the life span: Examination of task-switching performance. Dev. Psychol. 37:715– 730. Chugani, H., and M. Phelps, 1986. Maturational changes in cerebral function in infants determined by 18FDG positron emission tomography. Science 231(February 21):840–843. Clark, A. S., and P. S. Goldman-Rakic, 1989. Gonadal hormones influence the emergence of cortical function in nonhuman primates. Behav. Neurosci. 103:1287–1295. Cole, P. M., M. K. Michel, and L. O. Teti, 1994. The development of emotion regulation and dysregulation: A clinical perspective. In N. A. Fox, ed., The development of emotion regulation: Biological and behavioral considerations. Monogr. Soc. Res. Child Dev. 59(2–3, Serial No. 240):73–100. Crone, E. A., S. A. Bunge, M. W. van der Molen, and K. R. Ridderinkhof, 2006. Switching between tasks and responses: A developmental study. Dev. Sci. 9:278–287. Crone, E. A, and M. W. van der Molen, 2004. Developmental changes in real life decision making: Performance on the gambling task previously shown to depend on the ventromedial prefrontal cortex. Dev. Neuropsychol. 25:251–279. Crone, E. A., C. Wendelken, S. E. Donohue, and S. A. Bunge, 2006. Neural evidence for dissociable components of task switching. Cerebral Cortex 16:475–486. Davidson, M. C., D. Amso, L. Cruess-Anderson, and A. Diamond, 2006. Development of cognitive control and executive functions from 4–13 years: Evidence from manipulations of memory, inhibition and task switching. Neuropsychologia 44:2037–2078. de Bellis, M. D., M. S. Keshavan, S. R. Beers, J. Hall, K. Frustaci, A. Masalehdan, et al., 2001. Sex differences in brain maturation during childhood and adolescence. Cerebral Cortex 11:552–557. Dempster, F. N., 1992. The rise and fall of the inhibitory mechanism: Toward a unified theory of cognitive development and aging. Dev. Rev. 12:45–75. Denham, S. A., K. A. Blair, M. Schmidt, and E. DeMulder, 2002. Compromised emotional competence: Seeds of violence sown early? Am. J. Orthopsychiatry 72:70–82.

Dewey, J., 1931/1985. Context and thought. In J. A. Boydston and A. Sharpe, eds., John Dewey: The Later Works, 1925–1953, vol. 6, pp. 3–21. Carbondale: Southern Illinois University Press. Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: Cognitive functions, anatomy, and biochemistry. In D. Stuss and R. Knight, eds., Principles of Frontal Lobe Function, 466–503. New York: Oxford University Press. Diamond, A., and K. Boyer, 1989. A version of the Wisconsin Card Sort Test for use with preschool children, and an exploration of their sources of error. J. Clin. Exp. Neuropsychol. 11:83. Diamond, A., S. M. Carlson, and D. M. Beck, 2005. Preschool children’s performance in task switching on the dimensional change card sort task: Separating the dimensions aids the ability to switch. Dev. Neuropsychol. 28:689–729. Diamond, A., and B. Doar, 1989. The performance of human infants on a measure of frontal cortex function, the delayed response task. Dev. Psychobiol. 22:271–294. Diamond, A., and J. Gilbert, 1989. Development as progressive inhibitory control of action: Retrieval of a contiguous object. Cogn. Dev. 4:223–249. Diamond, A., and P. S. Goldman-Rakic, 1989. Comparison of human infants and rhesus monkeys on Piaget’s AB task: Evidence for dependence on dorsolateral prefrontal cortex. Exp. Brain Res. 74:24–40. Diamond, A., N. Z. Kirkham, and D. Amso, 2002. Conditions under which young children can hold two rules in mind and inhibit a prepotent response. Dev. Psychol. 38:352–362. Dias, R., T. W. Robbins, and A. C. Roberts, 1996. Dissociation in prefrontal cortex of affective and attentional shifting. Nature 380:69–72. Dick, A. S., W. F. Overton, and S. L. Kovacs, 2005. The development of symbolic coordination: Representation of imagined objects, executive function, and theory of mind. J. Cogn. Dev. 6:133–161. Dodge, K. A., and J. Garber, 1991. Domains of emotion regulation. In J. Garber and K. A. Dodge, eds., The Development of Emotion Regulation and Dysregulation, 159–181. Cambridge, UK: Cambridge University Press. Dowsett, S., and D. J. Livesey, 2000. The development of inhibitory control in pre-school children: Effects of “executive skills” training. Dev. Psychobiol. 36:161–174. Durston, S., M. C. Davidson, N. Tottenham, A. Galvan, J. Spicer, J. A. Fossella, and B. J. Casey, 2006. A shift from diffuse to focal cortical activity with development. Dev. Sci. 9:1–20. Durston, S., H. E. Hulshoff Pol, B. J. Casey, J. N. Giedd, J. K. Buitelaar, and H. van Engeland, 2001. Anatomical MRI of the developing human brain: What have we learned? J. Am. Acad. Child Adolesc. Psychiatry 40:1012–1020. Eigsti, I. M., V. Zayas, W. Mischel, Y. Shoda, O. Ayduk, M. Dadlani, et al., 2006. Predicting cognitive control from preschool to late adolescence and young adulthood. Psychol. Sci. 17:478–484. Eisenberg, N., R. A. Fabes, S. A. Shepard, B. C. Murphy, I. K. Guthrie, and S., Jones, et al., 1997. Contemporaneous and longitudinal prediction of children’s social functioning from regulation and emotionality. Child Dev. 68:642–664. Enns, J. T., and S. Cameron, 1987. Selective attention in young children: The relations between visual search, filtering in, and priming. J. Exp. Child Psychol. 44:38–63. Espy, K. A., and R. Bull, 2005. Inhibitory processes in young children and individual variation in short-term memory. Dev. Neuropsychol. 28:669–688.

zelazo, carlson, and kesek: executive function in childhood

569

Espy, K. A., P. M. Kaufmann, M. L. Glisky, and M. D. McDiarmid, 2001. New procedures to assess executive functions in preschool children. Clin. Neuropsychologist 15:46–58. Espy, K. A., P. M. Kaufmann, M. McDiarmid, and M. L. Glisky, 1999. Executive functioning in preschool children: A-not-B and other delayed response format task performance. Brain Cogn. 41:178–199. Falkenstein, M., J. Hohnsbein, J. Hoormann, and L. Blanke, 1991. Effects of crossmodal divided attention on late ERP components. II. Error processing in choice reaction tasks. Electroencephalogr. Clin. Neurophysiol. 78:447–455. Fan, J., B. D. McCandliss, T. Sommer, A. Raz, and M. I. Posner, 2002. Testing the efficiency and independence of attentional networks. J. Cogn. Neurosci. 14:340–347. Feldman, M. J., and J. Drasgow, 1951. A visual-verbal test for schizophrenia. Psychiatr. Q. 25(Suppl.):55–64. Fellows, L. K., and M. J. Farah, 2003. Ventromedial frontal cortex mediates affective shifting in humans: Evidence from a reversal learning paradigm. Brain 126:1830–1837. Frith, U., and C. D. Frith, 2003. Development and neurophysiology of mentalizing. Philos. Trans. R. Soc. London [Biol.] 358:459– 473. Frye, D., P. D. Zelazo, and J. A. Burack, 1998. I. Cognitive complexity and control: Implications for theory of mind in typical and atypical development. Curr. Dir. Psychol. Sci. 7:116–121. Frye, D., P. D. Zelazo, and T. Palfai, 1995. Theory of mind and rule-based reasoning. Cogn. Dev. 10:483–527. Funder, D. C., J. H. Block, and J. Block, 1983. Delay of gratification: Some longitudinal personality correlates. J. Pers. Soc. Psychol. 44:1198–1213. Garon, N., and C. Moore, 2004. Complex decision-making in early childhood. Brain Cogn. 55:158–170. Garon, N., and C. Moore, 2007. Awareness and symbol use improves future-oriented decision making in preschoolers. Dev. Neuropsychol. 31:39–59. Gathercole, S. E., 1998. The development of memory. J. Child Psychol. Psychiatry 39:3–27. Gehring, W. J., B. Goss, M. G. H. Coles, D. E. Meyer, and E. Donchin, 1993. A neural system for error detection and compensation. Psychol. Sci. 4:385–390. Gelman, R., and E. Meck, 1983. Preschoolers’ counting: Principles before skill. Cognition 13:343–359. Gerardi-Caulton, G., 2000. Sensitivity to spatial conflict and the development of self-regulation in children 24–36 months of age. Dev. Sci. 3/4:397–404. Gerstadt, C., Y. Hong, and A. Diamond, 1994. The relationship between cognition and action: Performance on children 3½–7 years old on a Stroop-like day-night test. Cognition 53:129–153. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, L. Hong, A. Zijdenbos, et al., 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2:861–863. Gogtay, N., J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101:8174–8179. Goldberg, E., and R. M. Bilder, 1987. The frontal lobes and hierarchical organization of cognitive control. In E. Perecman, ed., The Frontal Lobes Revisited, 159–187. New York: IRBN Press. Golden, C. J., 1981. The Luria-Nebraska children’s battery: Theory and formulation. In G. W. Hynd and J. E. Obrzut, eds., Neuropsychological Assessment and the School-Aged Child, 277–302. New York: Grune & Stratton.

570

cognition

Goldman-Rakic, P. S., 1987. Development of cortical circuitry and cognitive function. Child Dev. 58:601–622. Goldstein, K., and M. Scheerer, 1941. Abstract and concrete behavior: An experimental study with special tests. Psychol. Monogr. 53:1–151. Grant, D. A., and E. A. Berg, 1948. A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. J. Exp. Psychol. 38:404–411. Hala, S., S. Hug, and A. Henderson, 2003. Executive functioning and false belief understanding in preschool children: Two tasks are harder than one. J. Cogn. Dev. 4:275–298. Happaney, K., P. D. Zelazo, and D. T. Stuss, 2004. Development of orbitofrontal function: Current themes and future directions. Brain Cogn. 55:1–10. Harlow, J. M., 1848. Passage of an iron rod through the head. Boston Med. Surg. J. 39:389–393. Harlow, J. M., 1868. Recovery from the passage of an iron through the head. Publ. Mass. Med. Soc. 2:327–347. Harnishfeger, K. K., and D. F. Bjorklund, 1993. The ontogeny of inhibition mechanisms: A renewed approach to cognitive development. In M. L. Howe and R. Pasnak, eds., Emerging Themes in Cognitive Development, vol. l, Foundations, 28–49. New York: Springer-Verlag. Hebb, D. O., 1945. Man’s frontal lobes: A critical review. Arch. Neurol. 54:431–438. Holyoak, K. J., E. N. Junn, and D. Billman, 1984. Development of analogical problem-solving skill. Child Dev. 55:2042–2055. Hongwanishkul, D., K. R. Happaney, W. Lee, and P. D. Zelazo, 2005. Hot and cool executive function: Age-related changes and individual differences. Dev. Neuropsychol. 28:617–644. Hughes, C., 1998a. Executive function in preschoolers: Links with theory of mind and verbal ability. Br. J. Dev. Psychol. 16: 233–253. Hughes, C., 1998b. Finding your marbles: Does preschoolers’ strategic behavior predict later understanding of mind? Dev. Psychol. 34:1326–1339. Hughes, C., J. Dunn, and A. White, 1998. Trick or treat? Uneven understanding of mind and emotion and executive function among “hard to manage” preschoolers. J. Child Psychol. Psychiatry 39:981–994. Hughes, C., and R. Ensor, 2005. Executive function and theory of mind in 2-year-olds: A family affair? Dev. Neuropsychol. 28: 645–668. Hughes, C., A. White, J. Sharpen, and J. Dunn, 2000. Antisocial, angry, and unsympathetic: “Hard to manage” preschoolers’ peer problems, and possible social and cognitive influences. J. Child Psychol. Psychiatry 41:169–179. Hunter, W. S., 1917. The delayed reaction in a child. Psychol. Rev. 24:74–87. Huttenlocher, P. R., 1990. Morphometric study of human cerebral cortex development. Neuropsychologia 28:517–527. Huttenlocher, P. R., 2002. Neural Plasticity: The Effects of Environment on the Development of the Cerebral Cortex. Cambridge, MA: Harvard University Press. Isquith, P. K., G. A. Gioia, and K. A. Espy, 2004. Executive function in preschool children: Examination through everyday behavior. Dev. Neuropsychol. 26:403–422. Iversen, S. D., and M. Mishkin, 1970. Perseverative interference in monkeys following selective lesions of the inferior prefrontal convexity. Exp. Brain Res. 11:376–386. Jacobsen, C. F., 1936. Studies of cerebral function in primates. I. The functions of the frontal association areas in monkeys. Comp. Psychol. Monogr. 13:1–68.

Jacques, S., and P. D. Zelazo, 2001. The Flexible Item Selection Task (FIST): A measure of executive function in preschoolers. Dev. Neuropsychol. 20:573–591. Jacques, S., P. D. Zelazo, N. Z. Kirkham, and T. K. Semcesen, 1999. Rule selection versus rule execution in preschoolers: An error-detection approach. Dev. Psychol. 35:770–780. Jacques, S., P. D. Zelazo, S. F. Lourenco, and A. E. Sutherland, 2007. The role of labeling in the development of cognitive flexibility. Manuscript under review. Kerr, A., and P. D. Zelazo, 2004. Development of “hot” executive function: The Children’s Gambling Task. Brain Cogn. 55: 148–157. Killcross, S., T. W. Robbins, and B. J. Everitt, 1997. Different types of fear-conditioned behavior mediated by different nuclei within amygdala. Nature 388(July 24):377–380. Kindlon, D., E. Mezzacappa, and F. Earls, 1995. Psychometric properties of impulsivity measures: Temporal stability, validity and factor structure. J. Child Psychol. Psychiatry 36:645–661. Kirkham, N. Z., L. M. Cruess, and A. Diamond, 2003. Helping children apply their knowledge to their behavior on a dimensionswitching task. Dev. Sci. 6:449–467. Klingberg, T., E. Fernell, P. Olesen, M. Johnson, P. Gustafsson, K. Dahlström, et al., 2005. Computerized training of working memory in children with ADHD—A randomized, controlled trial. J. Am. Acad. Child Adolesc. Psychiatry 44:177–186. Klingberg, T., C. J. Vaidya, J. D. E. Gabrieli, M. E. Moseley, and M. Hedehus, 1999. Myelination and organization of the frontal white matter in children: A diffusion tensor MRI study. NeuroReport 10:2817–2821. Kloo, D., and J. Perner, 2003. Training transfer between card sorting and false belief understanding: Helping children apply conflicting descriptions. Child Dev. 74:1823–1839. Kloo, D., and J. Perner, 2005. Disentangling dimensions in the dimensional change card sorting task. Dev. Sci. 8:44–56. Kochanska, G., K. C. Coy, and K. T. Murray, 2001. The development of self-regulation in the first four years of life. Child Dev. 72:1091–1111. Kochanska, G., K. C. Coy, T. L. Tjebkes, and S. J. Husarek, 1998. Individual differences in emotionality in infancy. Child Dev. 69:375–390. Kochanska, G., K. T. Murray, and E. T. Harlan, 2000. Effortful control in early childhood: Continuity and change, antecedents, and implications for social development. Dev. Psychol. 36:220– 232. Kochanska, G., K. Murray, T. Y. Jacques, A. Koenig, and K. A. Vandegeest, 1996. Inhibitory control in young children and its role in emerging internalization. Child Dev. 67:490– 507. Koechlin, E., G. Basso, P. Pietrini, S. Panzer, and J. Grafman, 1999. The role of the anterior prefrontal cortex in human cognition. Nature 399:148–151. Lamm, C., P. D. Zelazo, and M. D. Lewis, 2006. Neural correlates of cognitive control in childhood and adolescence: Disentangling the contributions of age and executive function. Neuropsychologia 44:2139–2148. LeDoux, J. E., 1996. The Emotional Brain: The Mysterious Underpinnings of Emotional Life. New York: Simon and Schuster. Lengua, L. J., 2006. Growth in temperament and parenting as predictors of adjustment during children’s transition to adolescence. Dev. Psychol. 42:819–832. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36:273–293.

Luna, B., K. R. Thulborn, D. P. Munoz, E. P. Merriam, K. E. Garver, N. J. Minshew, et al., 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 13:786–793. Luria, A. R., 1959. The directive function of speech in development and dissolution. Word 15:341–352. Luria, A. R., 1961. The Role of Speech in the Regulation of Normal and Abnormal Behavior, ed. J. Tizard. New York: Liveright. Luria, A. R., 1966. Higher Cortical Functions in Man, 2nd ed. New York: Basic Books. (Original work published in 1962.) Luria, A. R., 1973. The Working Brain: An Introduction to Neuropsychology, trans. B. Haigh. New York: Basic Books. MacDonald, A. W., J. D. Cohen, V. A. Stenger, and C. S. Carter, 2000. Dissociating the role of dorsolateral prefrontal cortex and anterior cingulate cortex in cognitive control. Science 288(June 19):1835–1837. Marcovitch, S., and P. D. Zelazo, 1999. The A-not-B error: Results from a logistic meta-analysis. Child Dev. 70:1297– 1313. Marcovitch, S., and P. D. Zelazo, 2006. Non-monotonic influence of number of A trials on 2-year-olds’ perseverative search. J. Cogn. Dev. 7:477–501. Marcovitch, S., P. D. Zelazo, and M. A. Schmuckler, 2002. The effect of number of A trials on performance on the A-not-B task. Infancy 3:519–529. McEvoy, R., S. J. Rogers, and B. F. Pennington, 1993. Executive function and social communication deficits in young autistic children. J. Child Psychol. Psychiatry 34:563–578. Mead, G. H., 1934. Mind, Self, and Society from the Standpoint of a Social Behaviorist, ed. C. W. Morris. Chicago: University of Chicago Press. Metcalfe, J., and W. Mischel,1999. A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychol. Rev. 106:3–19. Mezzacappa, E., 2004. Alerting, orienting, and executive attention: Developmental properties and sociodemographic correlates in an epidemiological sample of young, urban children. Child Dev. 75:1373–1386. Miller E. K., and J. D. Cohen, 2001. An integrative theory of prefrontal function. Annu. Rev. Neurosci. 24:167–202. Miller, G., E. Galanter, and K. H. Pribram, 1960. Plans and the Structure of Behavior. New York: Henry Holt. Milner, B., 1963. Effects of different brain lesions on card sorting. Arch. Neurol. 9:90–100. Mischel, W., E. B. Ebbesen, and A. R. Zeiss, 1972. Cognitive and attentional mechanisms in delay of gratification. J. Pers. Soc. Psychol. 21:204–218. Mischel, W., Y. Shoda, and P. K. Peake, 1988. The nature of adolescent competencies by preschool delay of gratification. J. Pers. Soc. Psychol. 54:687–696. Mischel, W., Y. Shoda, and M. L. Rodriguez, 1989. Delay of gratification in children. Science 244(May 22):933–938. Miyake, A., N. P. Friedman, M. J. Emerson, A. H. Witzki, and A. Howerter, 2000. The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cogn. Psych. 41:49–100. Monterosso, J., R. Ehrman, K. L. Napier, C. P. O’Brien, and A. R. Childress, 2001. Three decision-making tasks in cocainedependent patients: Do they measure the same construct? Addiction 96:1825–1837. Morton, J. B., and Y. Munakata, 2002. Active vs. latent representations: A neural network model of perseveration, dissociation, and decalage. Dev. Psychobiol. 40:255–265.

zelazo, carlson, and kesek: executive function in childhood

571

Moses, L. J., and S. M. Carlson, 2004. Self-regulation and children’s theories of mind. In M. Chandler, C. Lalonde, and C. Lightfoot, eds., Changing Conceptions of Psychological Life, 127–146. Mahwah, NJ: Erlbaum. Müller, U., A. S. Dick, K. Gela, W. F. Overton, and P. D. Zelazo, 2006. The role of negative priming in preschoolers’ flexible rule use on the dimensional change card sort task. Child Dev. 77:395–412. Müller, U., P. D. Zelazo, T. Leone, and S. Hood, 2004. Interference control in a new rule use task: Age-related changes, labeling, and attention. Child Dev. 75:1594–1609. Munakata, Y., and B. E. Yerys, 2001. All together now: When dissociations between knowledge and action disappear. Psychol. Sci. 12:335–337. Nagahama, Y., H. Fukuyama, H. Yamauchi, S. Matsuzaki, J. Konishi, H. Shibasaki, et al., 1996. Cerebral activation during performance of a card sorting test. Brain 119:1667– 1675. Nagy, Z., H. Westerberg, and T. Klingberg, 2004. Maturation of white matter is associated with the development of cognitive functions during childhood. J. Cogn. Neurosci. 16:1227– 1233. Neville, H. J., 1993. Neurobiology of cognitive and language processing: Effects of early experience. In M. H. Johnson, ed., Brain Development and Cognition: A Reader, 424–448. Oxford: Blackwell. Noble, K. G., M. F. Norman, and M. J. Farah, 2005. Neurocognitive correlates of socioeconomic status in kindergarten children. Dev. Sci. 8:74–87. O’Donnell, S., M. Noseworthy, B. Levine, M. Brandt, and M. Dennis, 2005. Cortical thickness of the frontopolar area in typically developing children and adolescents. NeuroImage 24:948– 954. Overman, W. H., 2004. Sex differences in early childhood, adolescence and adulthood on cognitive tasks that rely on orbital prefrontal cortex. Brain Cogn. 55:134–147. Overman, W. H., J. Bachevalier, E. Schumann, and P. Ryan, 1996. Cognitive gender differences in very young children parallel biologically based cognitive gender differences in monkeys. Behav. Neurosci. 110:673–684. Pascual-Leone, J., 1970. A mathematical model for the transition rule in Piaget’s developmental stages. Acta Psychol. 32:301– 345. Peake, P., M. Hebl, and W. Mischel, 2002. Strategic attention deployment for delay of gratification in working and waiting situations. Dev. Psychol. 38:313–326. Pennington, B. F., and S. Ozonoff, 1996. Executive functions and developmental psychopathologies. J. Child Psychol. Psychiatry Annu. Res. Rev. 37:51–87. Perner, J., 1991. Understanding the Representational Mind. Cambridge, MA: MIT Press. Perner, J., and B. Lang, 1999. Development of theory of mind and cognitive control. Trends Cogn. Sci. 3:337–344. Perner, J., B. Lang, and D. Kloo, 2002. Theory of mind and self-control: More than a common problem of inhibition. Child Dev. 73:752–767. Petrides, M., and B. Milner, 1982. Deficits on subject-ordered tasks after frontal- and temporal-lobe lesions in man. Neuropsychologia 20:249–262. Piaget, J., 1954. The Construction of Reality in the Child. New York: Basic. Posner, M. I., 1994. Attention: The mechanism of consciousness. Proc. Natl. Acad. Sci. USA 91:7398–7402.

572

cognition

Posner, M. I., and S. E. Petersen, 1990. The attention system of the human brain. Annu. Rev. Neurosci. 13:25–42. Povinelli, D. J., and D. K. O’Neill, 2000. Do chimpanzees use their gestures to instruct each other? In S. Baron-Cohen, H. Tager-Flusberg, and D. J. Cohen, eds., Understanding Other Minds: Perspectives from Autism, 2nd ed., pp. 459–487. Oxford, UK: Oxford University Press. Prabhakaran, V., J. A. L. Smith, J. E. Desmond, G. H. Glover, and J. D. E. Gabrieli, 1997. Neural substrates of fluid reasoning: An fMRI study of neocortical activation during performance of the Raven’s Progressive Matrices Test. Cogn. Psych. 33:43– 63. Premack, D., and A. J. Premack, 1996. Why animals lack pedagogy and some cultures have more of it than others. In D. R. Olson and N. Torrance, eds., The Handbook of Human Development and Education, 302–344. Oxford, UK: Blackwell. Premack, D., and G. Woodruff, 1978. Does the chimpanzee have a theory of mind? Behav. Brain Sci. 1:516–526. Prencipe, A., and P. D. Zelazo, 2005. Development of affective decision-making for self and other: Evidence for the integration of first- and third-person perspectives. Psychol. Sci. 16:501– 505. Reed, M. A., D. L. Pien, and M. K. Rothbart, 1984. Inhibitory self-control in children. Merrill-Palmer Q. 30:131–147. Ridderinkhof, K. R., M. Ullsperger, E. A. Crone, and S. Nieuwenhuis, 2004. The role of the medial frontal cortex in cognitive control. Science 306(October 15):443–447. Riggs, N., C. Blair, and M. Greenberg, 2003. Concurrent and 2-year longitudinal relations between executive function and the behavior of 1st and 2nd grade children. Child Neuropsychol. 9:267–276. Roberts, R. J., and B. F. Pennington, 1996. An interactive framework for examining prefrontal cognitive processes. Dev. Neuropsychol. 12:105–126. Rochat, P., 2001. The dialogical nature of cognition. Monogr. Soc. Res. Child Dev. 66(2, Serial No. 265):133–143. Rodriguez, M. L., W. Mischel, and Y. Shoda, 1989. Cognitive person variables in the delay of gratification of older children at risk. J. Pers. Soc. Psychol. 57:358–367. Rolls, E. T., 1999. The Brain and Emotion. Oxford, UK: Oxford University Press. Rolls, E. T., 2004. The functions of the orbitofrontal cortex. Brain Cogn. 55:11–29. Rolls, E. T., J. Hornak, D. Wade, and J. McGrath, 1994. Emotion related learning in patients with social and emotional changes associated with frontal lobe damage. J. Neurol. Neurosurg. Psychiatry 57:1518–1524. Rothbart, M. K., and J. E. Bates, 1998. Temperament. In N. Eisenberg, ed., Handbook of Child Psychology: Social, Emotional, and Personality Development, vol. 3, pp. 105–176. New York: John Wiley. Rubia, K., S. Overmeyer, E. Taylor, M. Brammer, S. C. R. Williams, and A., Simmons, et al., 2000. Functional frontalisation with age: Mapping neurodevelopmental trajectories with fMRI. Neurosci. Biobehav. Rev. 24:13–19. Rueda, M. R., M. I. Posner, M. K. Rothbart, and C. P. DavisStober, 2004. Development of the time course for processing conflict: An event-related potentials study with 4 year olds and adults. BMC Neurosci. 5:39. Rueda, M. R., M. K. Rothbart, L. Saccamanno, and M. I. Posner, 2005. Training, maturation and genetic influences on the development of executive attention. Proc. Natl. Acad. Sci. USA 102:14931–14936.

Rushton, J. P., C. J. Brainerd, and M. Pressley, 1983. Behavioral development and construct validity: The principle of aggregation. Psychol. Bull. 94:18–38. Russell, J., 1996. Agency: Its Role in Mental Development. Mahwah, NJ: Erlbaum. Russell, J., ed. 1997. Autism as an Executive Disorder. New York: Oxford University Press. Sabbagh, M. A., and M. Taylor, 2000. Neural correlates of theory-of-mind reasoning: An event-related potential study. Psychol. Sci. 11:46–50. Sabbagh, M. A., F. Xu, S. M. Carlson, L. J. Moses, and K. Lee, 2006. The development of executive functioning and theory of mind: A comparison of Chinese and U.S. preschoolers. Psychol. Sci. 17:74–81. Sakai, K., and R. E. Passingham, 2003. Prefrontal interactions reflect future task operations. Nature Neurosci. 6:75–81. Sakai, K., and R. E. Passingham, 2006. Prefrontal set activity predicts rule-specific neural processing during subsequent cognitive performance. J. Neurosci. 26:1211–1218. Santesso, D. L., S. J. Segalowitz, and L. A. Schmidt, 2006. Error-related electrocortical responses are enhanced in children with obsessive-compulsive behaviors. Dev. Neuropsychol. 29:431– 445. Saxe, R., S. Carey, and N. Kanwisher, 2004. Understanding other minds: Linking developmental psychology and functional neuroimaging. Annu. Rev. Psychol. 55:87–124. Segalowitz, S. J., and P. L. Davies, 2004. Charting the maturation of the frontal lobe: An electrophysiological strategy. Brain Cogn. 55:116–133. Seguin, J. R., and P. D. Zelazo, 2005. Executive function in early physical aggression. In R. E. Tremblay, W. W. Hartup, and J. Archer, eds., Developmental Origins of Aggression, 307–329. New York: Guilford. Sethi, A., W. Mischel, L. Aber, Y. Shoda, and M. Rodriguez, 2000. The role of strategic attention deployment in development of self-regulation: Predicting preschoolers’ delay of gratification from mother-toddler interactions. Dev. Psychol. 36:767– 777. Siegal, M., and R. Varley, 2002. Neural systems involved in “theory of mind.” Nature Rev. Neurosci. 3:463–471. Simpson, A., and K. J. Riggs, 2005. Factors responsible for performance on the day-night task: Response set or semantics? Dev. Sci. 8:360–371. Singer, J. L. 1961. Imagination and waiting ability in young children. J. Pers. 29:396–413. Smith, E. E., and J. Jonides, 1999. Storage and executive processes in the frontal lobe. Science 283(March 12):1657– 1661. Sophian, C., and H. M. Wellman, 1983. Selective information use and perseveration in the search behavior of infants and young children. J. Exp. Psychol. 35:369–390. Speltz, M., M. Deklyen, R. Calderon, M. Greenberg, and P. Fisher, 1999. Neuropsychological characteristics and test behaviors of boys with early onset conduct problems. J. Abnorm. Psychol. 108:315–325. Spencer, J. P., L. B. Smith, and E. Thelen, 2001. Tests of a dynamic systems account of the A-not-B error: The influence of prior experience on the spatial memory abilities of two-year-olds. Child Dev. 72:1327–1346. Strayhorn, J. M., Jr., 2002. Self-control: Theory and research. J. Am. Acad. Child Adolesc. Psychiatry 41:7–16. Stroop, J. R., 1935. Studies of interference in serial verbal reactions. J. Exp. Psychol. 18:643–662.

Stuss, D. T., and D. F. Benson, 1986. The Frontal Lobes. New York: Raven Press. Thompson, C., J. Barresi, and C. Moore, 1997. The development of future-oriented prudence and altruism in preschoolers. Cogn. Dev. 12:199–212. Thompson, P. M., J. N. Giedd, R. P. Woods, D. MacDonald, A. C. Evans, and A. W. Toga, 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 404:190–193. Tomasello, M., A. C. Kruger, and H. H. Ratner, 1993. Cultural learning. Behav. Brain Sci. 16:495–552. Towse, J. N., J. Redbond, C. M. T. Houston-Price, and S. Cook, 2000. Understanding the Dimensional Change Card Sort: Perspectives from task success and failure. Cogn. Dev. 15:347– 365. Tranel, D., S. W. Anderson, and A. Benton, 1994. Development of the concept of “executive function” and its relationship to the frontal lobes. In F. Boller and J. Grafman, eds., Handbook of Neuropsychology, vol. 9, pp. 125–148. Amsterdam: Elsevier. Vaid, J. 2002. Bilingualism. In V. S. Ramachandran, ed., Encyclopedia of the Human Brain, vol. 1., pp. 417–434. San Diego: Elsevier. Weiner, N., 1948. Cybernetics. Cambridge, MA: MIT Press. Wellman, H. M., 1990. The Child’s Theory of Mind. Cambridge, MA: MIT Press. Wellman, H. M., D. Cross, and J. Watson, 2001. Meta-analysis of theory of mind development: The truth about false belief. Child Dev. 72:655–684. Welsh, M. C., 1991. Rule-guided behavior and self-monitoring on the Tower of Hanoi disk-transfer task. Cogn. Dev. 6:59– 76. Welsh, M. C., B. F. Pennington, and D. B. Groisser, 1991. A normative-developmental study of EF: A window on prefrontal function in children. Dev. Neuropsychol. 7:131–149. Werner, H., and B. Kaplan, 1963. Symbol Formation: An OrganismicDevelopmental Approach to Language and the Expression of Thought. New York: Wiley. White, S. H., 1965. Evidence for a hierarchical arrangement of learning processes. In L. P. Lipsitt and C. C. Spiker, eds., Advances in Child Development and Behavior, vol. 2, pp. 187–220. New York: Academic Press. Wimmer, H., and J. Perner, 1983. Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition 13:103– 128. Wise, S. P., E. A. Murray, and C. R. Gerfen, 1996. The frontal cortex-basal ganglia system in primates. Crit. Rev. Neurobiol. 10:317–356. Yakovlev, P. I., and A. R. Lecours, 1967. The myelogenetic cycles of regional maturation of the brain. In A. Minkowski, ed., Regional Development of the Brain in Early Life, 3–70. Oxford, UK: Blackwell. Yerys, B. E., and Y. Munakata, 2006. When labels hurt but novelty helps: Children’s perseveration and flexibility in a cardsorting task. Child Dev. 77:1589–1607. Zelazo, P. D., 2004. The development of conscious control in childhood. Trends Cogn. Sci. 8:12–17. Zelazo, P. D., F. I. M. Craik, and L. Booth, 2004. Executive function across the lifespan. Acta Psychol. 115:167–183. Zelazo, P. D., and W. Cunningham, 2007. Executive function: Mechanisms underlying emotion regulation. In J. Gross, ed., Handbook of Emotion Regulation. New York: Guilford.

zelazo, carlson, and kesek: executive function in childhood

573

Zelazo, P. D., and S. Jacques, 1997. Children’s rule use: Representation, reflection, and cognitive control. Ann. Child Dev. 12:119–176. Zelazo, P. D., S. Jacques, J. Burack, and D. Frye, 2002. The relation between theory of mind and rule use: Evidence from persons with autism-spectrum disorders. Infant Child Dev. 11:171–195. Zelazo, P. D., and U. Müller, 2002. Executive functions in typical and atypical development. In U. Goswami, ed., Handbook

574

cognition

of Childhood Cognitive Development, 445–469. Oxford, UK: Blackwell. Zelazo, P. D., U. Müller, D. Frye, and S. Marcovitch, 2003. The development of executive function in early childhood. Monogr. Soc. Res. Child Dev. 68(3):vii–137. Zelazo, P. D., J. S. Reznick, and J. Spinazzola, 1998. Representational flexibility and response control in a multi-step, multilocation search task. Dev. Psychol. 34:203–214.

35

The Development of Prefrontal Cortex Functions in Adolescence: Theoretical Models and a Possible Dissociation of Dorsal versus Ventral Subregions ELIZABETH A. OLSON AND MONICA LUCIANA

The prefrontal cortex (PFC) orchestrates the most complex aspects of human behavior, so an understanding of how this region develops to facilitate adult levels of functional competence is an important area of inquiry. Tasks that are supported by the prefrontal cortex include the maintenance and manipulation of information in working memory, foresight and planning for future goals, strategic self-organization, and updating behavior in response to changing reward contingencies. Collectively, these processes have been referred to as executive functions (EFs). Although dramatic and grossly observable changes in brain development occur during infancy and early childhood, leading to a great deal of focus on neurobehavioral development during those stages, emerging evidence indicates that some structural changes in prefrontal cortex and some functional changes in EFs have a very protracted developmental course, with continuing maturation into the third decade of life. This chapter reviews the evidence regarding maturation of the prefrontal cortex during adolescence and considers the contributions that a developmental perspective may make toward understanding the functional organization of the PFC. While some theorists have argued that the PFC contains subsystems specialized for managing different functional processes, others have articulated the view that the PFC contains subsystems specialized for managing different content domains. Examination of developmental processes may help to clarify, albeit indirectly, whether the PFC appears to be primarily organized by the processes it conducts or by the content it manages. For instance, on the one hand, demonstration of differing time courses for maturation of an identical process performed on two different types of information would indicate that development proceeds by content domain and would lend support to the proposition

that different PFC regions (which develop at somewhat different rates) are specialized for management of different content domains. On the other hand, demonstration of identical developmental time courses for maturation of an identical process performed on two different types of information would indicate that development is organized by process; in that case, either the same PFC region may be performing the function across different content domains, or the two domains may be managed by different PFC regions that happen to develop at identical rates. Consideration of developmental time courses for different processes may also help to illuminate the functional organization of the PFC. Functions that develop at different rates may be presumed to rely on different prefrontal areas or on interactions between areas. Functions that develop at identical rates may be supported by the same region or by regions that happen to have very similar maturational patterns. Prior to discussing the functional development of putative PFC functions, the overall structure and connectivity of this region will be reviewed. Please also see chapter 34 by Zelazo, Carlson, and Kesek for an expanded discussion of this topic.

Prefrontal cortex: Structure and connectivity The terms prefrontal cortex and frontal cortex are not synonymous. The PFC is but one part of the frontal lobe and includes all of the cortex in the frontal lobe that is anterior to the supplementary motor area. In humans, the PFC is vast in size, comprising between one-quarter and one-third of the entire cerebral cortex. The PFC is not viewed as a unitary structure. A number of prefrontal regions have been identified, including the dorsolateral PFC (DL-PFC; Brodmann’s area 46), ventrolateral PFC (VL-PFC; Brodmann’s

575

areas 12 and 45), the medial PFC, and the orbitofrontal cortex (OFC) (areas 10, 11, 13, and 14). These divisions can be seen in figure 35.1 (although some orbitofrontal regions are not apparent because of their location on the medial underside of the frontal lobe, the area that would be visible if we lifted the frontal cortex from the temporal cortex and unfolded the brain accordingly). From a functional perspective, the medial PFC and the OFC are often considered together and comprise the ventromedial PFC (VM-PFC) (Happaney, Zelazo, and Stuss, 2004; Fuster, 2002). There is also some suggestion that prefrontal regions of the left versus right hemisphere might mediate different processes (Happaney, Zelazo, and Stuss, 2004; Bechara, 2004; Rolls et al., 1994; Tranel, Bechara, and Denburg, 2002). All these prefrontal regions have reciprocal connections with sensory cortices and with each other. For instance, the DL-PFC has connections with dorsal (“where”) and ventral (“what”) visual streams as well as with the somatosensory cortex in the caudal parietal lobe, with the auditory cortex

Figure 35.1 Brodmann’s cytoarchitectural map of the human brain. (Reprinted with permission from Pandya and Yeterian, 1990, p. 64.)

576

cognition

in the superior temporal gyrus, and with multimodal sensory areas in the rostral superior temporal sulcus (reviewed in Miller and Cohen, 2001). In a landmark series of studies, Goldman-Rakic (1987) demonstrated that different regions of the macaque parietal association cortex map preferentially to different subregions of the DL-PFC, raising the possibility that different areas of the DL-PFC may have a functional organization that parallels that of the corresponding parietal areas. In addition to these connections with other cortical areas, anterograde and retrograde tracing studies in nonhuman animals indicate that prefrontal regions are richly interconnected with subcortical and limbic areas. The PFC receives input from the thalamus, with thalamic relay nuclei transmitting information from the basal ganglia to specific regions of the frontal cortex (for a review, see Parent and Hazrati, 1995); for instance, the mediodorsal nucleus of the thalamus projects to dorsolateral and orbital PFC regions in the macaque (McFarland and Haber, 2002), and indeed, the prefrontal cortex was originally defined as the projection field of the mediodorsal thalamus (Rose and Woolsey, 1948; Nauta, 1962). There are both reciprocal and nonreciprocal projections back from the frontal cortex to the thalamus (McFarland and Haber, 2002). Frontal cortical regions also send outputs to the basal ganglia, and the projections from cortical areas are believed to impose a functional organization upon areas of the basal ganglia including the striatum (Parent and Hazrati, 1995). The PFC receives projections from the hippocampus, and different patterns of hippocampal innervation have been noted in medial versus orbital versus lateral PFC (Barbas and Blatt, 1995). There are projections from medial prefrontal areas to the hypothalamus (Ongur, An, and Price, 1998). The PFC projects to the anterior and posterior cingulate cortex (Brodmann’s areas 24 and 31; see figure 35.1) (Selemon and Goldman-Rakic, 1988). The PFC also sends and receives projections to and from the amygdala (McDonald, 1991; McDonald, Mascagni, and Guo, 1996). As suggested by this brief review, the PFC shares connections with a large number of cortical and subcortical structures. This intricate pattern of connectivity between the PFC and other regions has important implications for our understanding of prefrontal function. The interconnectivity of the prefrontal regions with one another suggests that prefrontally guided behaviors may be mediated by activity in multiple PFC areas. The reciprocal connections with sensory areas may be related to the PFC’s role not only in receiving sensory information but also in biasing sensory processing in response to top-down attentional demands. Connectivity with limbic areas implicates areas such as the VM-PFC in processing information regarding affective valence such as reward seeking, and connectivity with the anterior cingulate cortex suggests a role in monitoring feed-

back based on environmental contingencies. Given these extensive connections, adult levels of performance on prefrontally mediated tasks depend on the functional maturation of these other brain regions, some of which may mature before the PFC and some of which may show concurrent maturational changes throughout adolescence and young adulthood. In addition to supporting the PFC’s role in guiding behavior, these other regions may play a role in organizing the development of the prefrontal regions with which they interact. Similarly, maturation of a given functional area of the PFC may help to organize the development of other connected prefrontal regions. Thus, with respect to the PFC’s role in guiding behavior, it is important to keep in mind that the PFC represents but one node in a series of complex parallel-distributed networks. Any structure-function relations or theorized associations must be viewed accordingly.

Structural development of the PFC during adolescence What, then, do we know about the maturation of the PFC during childhood and throughout adolescence? It has been suggested for a number of years, based on animal and human autopsy data, that the PFC continues to mature beyond middle childhood (e.g., Yakovlev and Lecours, 1967; see Lewis, 1997, for a review). More recently, brain-imaging data have validated this idea and indicate continuing brain development throughout adolescence. The most protracted developmental time courses occur in prefrontal areas, including the DL-PFC and VM-PFC. Overall, the brain reaches 90 percent of its adult size in humans by age 6 years; after this point, head size continues to change, but increments are attributable to changes in skull thickness, not brain volume (Reiss et al., 1996; Giedd, 2004). Recently, an ongoing pediatric brain-imaging study of over 100 medically and psychiatrically healthy adolescents at the National Institute of Mental Health collected a series of repeated magnetic resonance imaging (MRI) scans over the course of development, producing a large body of information about brain development (e.g., Giedd et al., 1996, 1999; Giedd, 2004; Gogtay et al., 2004). The results indicate that significant changes in gray and white matter volumes occur throughout the teenage years (Giedd, 2004). Rather than reflecting a change in overall brain size (or number of neurons), these changes are considered to reflect processes that influence the number and quality of connections among neurons. These processes include myelination (which increases white matter conduction speed), arborization (in which the number of branching connections between neurons increases), and pruning (in which neuronal connections are destroyed). White matter volume in the brain increases in an apparently linear fashion throughout childhood and adolescence

(ages 5–23 years); the rate of increase is similar in the frontal, parietal, temporal, and occipital lobes (Giedd, 2004; Giedd et al., 1999). An earlier, cross-sectional study of infants to older adults found that white matter volume peaked at about age 20 years (Pfefferbaum et al., 1994), and Reiss and colleagues (1996) found increasing white matter volume from age 5 to age 17 years. Unlike white matter volume, gray matter volume increased to a peak in early adolescence and then decreased, resulting in an “inverted-U” pattern over the course of development (Gogtay et al., 2004). This pattern of gray matter increase followed by a decrease may be attributable to arborization followed by subsequent pruning of unused synapses (Giedd, 2004). Because of methodological issues related to how tissue types are parcellated as MRI data are analyzed (see Paus, 2005; Sowell et al., 1999), the decrease in gray matter volume during later adolescence might also reflect increased myelination. That is, an increase in white matter volume, rather than pruning of gray matter, may account for the apparent decrease (Sowell et al., 1999). In contrast to white matter development, which appears to be concordant across brain regions, the rate of change in gray matter volume is different in different lobes. For example, it has been reported that frontal cortical gray matter volume reaches a maximum early in adolescence (at age 11.0 for girls and 12.1 for boys), while temporal cortical gray matter volume reaches a maximum somewhat later, at 16.7 years for girls and 16.2 for boys (Giedd, 2004). An earlier peak for cortical gray matter volume was seen in an earlier cross-sectional study, which found peak gray matter occurring at age 4 years in a sample of infants to older adults (Pfefferbaum et al., 1994); in contrast, Reiss and colleagues (1996) found a linear decrease from age 5 to age 17. Overall, the development of the cortex (at least with respect to gray matter volume) proceeds from the back (parietal cortex) to the front (frontal cortex), with a few sensory areas completing gray matter development first, including the primary somatosensory cortices, occipital pole, and frontal pole (which included medial and posterior aspects of the olfactory cortex) (Gogtay et al., 2004). This overall back-to-front pattern of gray matter development is also seen within the frontal lobe. Similarly, white matter development (i.e., myelination) follows a rostral to caudal pattern overall, as well as proceeding rostrally to caudally within a given region (Sampaio and Truwit, 2001). Gray matter development begins in the primary motor cortex and spreads anteriorly, with the prefrontal cortex developing last (Gogtay et al., 2004). The DL-PFC loses gray matter only at the end of adolescence, and the VM-PFC was still undergoing developmental changes in gray matter at the latest ages studied (age 23) (Gogtay et al., 2004); voxel-byvoxel analysis of structural MRI data has confirmed that

olson and luciana: prefrontal cortex functions in adolescence

577

DL-PFC development does not reach adult levels of cortical gray matter until the 20s (Giedd, 2004). There may be hemispheric differences as well. According to one study, the left side of the PFC appears to mature earlier than corresponding regions on the right side; this pattern may be a general trend or could reflect earlier maturation of the dominant hemisphere in this particular sample of right-handed individuals (Gogtay et al., 2004). More sophisticated techniques have been used to examine more subtle changes in structural brain organization. For instance, diffusion tensor imaging (DTI) has been used to differentiate between white matter that is projectional (connecting the brain to spinal cord), associational (connecting one type of brain part to another), and commissural (connecting the same part in the right and left hemispheres) (Giedd, 2004; see also chapter 17 by Wozniak, Mueller, and Lim, this volume). Diffusion tensor imaging measures the rates of diffusion of water molecules and can therefore be used to assess not merely white matter volume, but also white matter organization. One major outcome variable from DTI studies is fractional anisotropy (FA), which reflects the extent to which diffusion of water molecules is anisotropic (not equal in all directions). Higher FA values are believed to reflect increasing organization of white matter tracts, which forces water molecules to travel in parallel with them. Some white matter changes may affect the measurements of both FA and white matter density (such as increasing myelination), whereas other changes affect FA only (such as changes in fiber orientation) (Barnea-Goraly et al., 2005). A DTI study comparing children (8- to 12-year-olds) to adults (21- to 30-year-olds) found higher frontal FA (reflecting increased white matter organization) in the adults (Klingberg et al., 1999). Another DTI study of 6- to 19year-olds found significant increases in FA in prefrontal regions, the internal capsule, pathways within the basal ganglia, pathways between the basal ganglia and thalamus, corticothalamic and corticospinal tracts extending from sensorimotor regions, ventral visual streams, and intrahemispheric tracts (including the arcuate fasciculus and the corpus callosum) (Barnea-Goraly et al., 2005). Schmithorst and colleagues (2002) found a negative correlation between age (in a sample of 5- to 18-year-olds) and trace diffusion, another DTI measure that reflects the total water molecule diffusion within a given area; less diffusion may indicate increased white matter organization. In addition, the same group has found an association between FA in frontal areas (in 5- to 18-year-olds) and global indices of cognitive performance (i.e., IQ score) (Schmithorst et al., 2005). Most recently, Liston and colleagues (2006) found that radial diffusion decreased with age in a sample of 7- to 31-yearolds; in addition, performance on a measure of cognitive control (a go-no-go task) was associated with radial diffu-

578

cognition

sion in frontostriatal areas, independent of the effects of age. Thus the PFC and associated structures are undergoing a great deal of synaptic refinement during the adolescent period. Whether these changes in gray matter volume, white matter volume, and white matter connectivity parallel behavioral development, particularly the development of EFs, has not been investigated within subjects. Rather, we infer that these structural refinements support behavioral development through evidence that prefrontally guided behaviors are maturing during the same time period. Prior to discussing behavioral development, an overview of functions attributed to the PFC will be provided.

Functions of the prefrontal cortex The role of the prefrontal cortex has intrigued behavioral neuroscientists for decades, since individuals with extensive prefrontal damage often appear (at least superficially) to be neurologically and intellectually intact. However, with increased scrutiny, more sophisticated neurological testing and behavioral observations have revealed a number of deficits associated with prefrontal damage. Damage to this area may result in a diverse pattern of deficits. Functions that could be affected include attention, reasoning, planning, the regulation of behavior based on past experience or verbal instructions, initiative, spontaneity, verbal fluency, behavioral restraint, social affect, and/or global personality features (Goldman-Rakic, 1987). Although not all of these functions will strike the reader as being “executive” in nature, there have been attempts to develop unified accounts of behaviors that are controlled by the PFC. One popular heuristic for understanding the organization of executive functions that was derived from cognitive psychology proposed a hierarchical system in which one central processor managed the operations of two subordinate systems. Thus Baddeley and Hitch’s (1974) model of working-memory function proposed the existence of a functional central executive system that coordinates that operation of two subsidiary systems, one managing visual content (the “visuo-spatial sketch pad”) and the other managing auditory-verbal content (the “phonological loop”). Similarly, Norman and Shallice (1986) proposed the existence of a central supervisory attention system that had hierarchical control over other processes such as contention scheduling. In this model, contention scheduling is the process through which particular behavioral routines are selected or inhibited when elicited by stimuli; the supervisory attention system is activated when the task requires an unknown solution, when the activation of a given schema is weak, when specific selection of a given schema is necessary, and when inappropriate schemata need to be inhibited (Stuss et al., 1995). In contrast to Baddeley and Hitch’s model, Shallice emphasizes that dif-

ferent subsystems implement different processes rather than managing different types of content (Shallice and Burgess, 1998). Miller and Cohen (2001) have also proposed a hierarchical model of cognitive control in which the PFC is responsible for the top-down guidance of behavior through the active maintenance of activation patterns that represent goals as well as strategies that must be implemented to achieve them. In this model, feedback from the PFC is responsible for directing attention (via feedback to sensory systems), response selection and inhibitory control (via feedback to motor systems), and working memory (via feedback to “intermediate” systems). Miller and Cohen argue that it is unlikely that different PFC regions are responsible for manipulating different types of information, since complex behavior requires the integration of relationships across content domains. In contrast to these hierarchical models, the proposition that the PFC is responsible for managing a single key process was suggested on the basis of animal research. GoldmanRakic’s (1987) lesion studies in primates identified the crucial importance of the DL-PFC in mediating completion of spatial delayed-response tasks. In these tasks, individuals must remember a piece of information (such as the location of a hidden reward) over a delay and then respond appropriately. Performance on delayed-response tasks requires the ability to access the relevant information, to hold that information in mind over the delay, and to execute a motor response; therefore, it indexes the ability to use short-term representational memory (i.e., internalized knowledge) to guide behavior in the absence of informative and immediate external cues (Goldman-Rakic, 1987). Based on the finding that lesions to the DL-PFC impair primates’ ability to remember the spatial location of an object during a delay, Goldman-Rakic articulated a primary role for the DL-PFC in actively representing spatial information over time so that this information could be used to guide a future response, a process known as spatial working memory. Goldman-Rakic’s empirical observations that lesions in the nearby ventrolateral PFC (inferior convexity) produced deficits in working memory for nonspatial visual features (such as form) (Goldman-Rakic, 1995) led her to propose that working memory is the primary function of the prefrontal cortex as a whole (Goldman-Rakic, 1998). However, because different types of working memory appeared to be mediated by dorsolateral versus inferior convexity regions of the PFC, she argued that rather than having different functions, different areas of the prefrontal cortex perform identical functions but on different content domains (Goldman-Rakic, 1988). Thus Goldman-Rakic’s influential view was that the PFC is responsible for maintaining information in working memory and initiating subsequent responses; moreover, differences between areas of the PFC arise because the areas are performing the same operations

on different content domains (Goldman-Rakic, 1988). Based on measurements of neuronal activity in monkeys performing delayed-response tasks, Goldman-Rakic (1988) argued that parallel cortical networks are responsible for spatial information processing in dorsal areas of the PFC and for nonspatial information processing in ventral areas of the PFC. One rationale for this theory was that the dorsalventral distinction in the PFC paralleled and extended that of the visual association cortices in the dorsal (“where”) versus ventral (“what”) systems. Analogous to Goldman-Rakic’s proposal that the PFC collectively manages working memory is Fuster’s contention (1995, 2002) that the PFC plays an overarching role in the temporal integration of action. Although Fuster describes somewhat differing functional roles for different areas of the prefrontal cortex, stressing the importance of the dorsal PFC in maintaining preparatory-set and short-term working memory and the importance of the orbital PFC in suppression of influences and inhibitory control, he emphasizes that, overall, these different functions contribute to a primary role of the prefrontal cortex in the formation of temporal structures that guide goal-directed behavior (linking past behavior with future actions, or “memory for the future”). While Goldman-Rakic’s model of prefrontal function was based on animal data, newer conceptualizations have been guided by human neuroimaging studies. Several authors have argued that different prefrontal areas are functionally differentiated and control different processes, with no overarching function attributable to the prefrontal cortex as a whole. In contrast to Goldman-Rakic’s proposal that the DL-PFC and VL-PFC manage different content, Petrides (1995, 1998) formulated a two-level model proposing that these regions of the lateral prefrontal cortex manage different processes. In his model, the VL-PFC is involved in lowerlevel active (explicit) retrieval of information and comparing stimuli held in short-term memory, while the DL-PFC is involved in the manipulation of stimuli held in working memory as well as self-monitoring of behavior. Petrides (1998) points out that performance on self-ordered search tasks requiring the DL-PFC requires not merely working memory but also self-monitoring of activity, suggesting that when deficits are observed on such tasks, these might reflect an executive contribution to performance that extends above and beyond the mnemonic requirements of the task. Petrides (1995) argues that different PFC regions perform these operations across content domains, for instance, that the DLPFC is responsible for working-memory operations on verbal in addition to spatial information. Within this model, the dorsolateral and ventrolateral regions of the PFC differ according to the level of processing that is demanded by the task. Lower levels of processing recruit the ventrolateral PFC, whereas higher, more complex, levels of processing recruit the dorsolateral PFC.

olson and luciana: prefrontal cortex functions in adolescence

579

Although beyond the scope of this review, it should be noted that the process and content models of PFC organization might not be completely contradictory, particularly since each model was derived from data from different species and from different methodologies. It could be that for some species (or at different points in development in humans) a given content domain (spatial versus verbal, for example) might require more in the way of informationprocessing resources than another content domain. Thus one question to be addressed in the neurobehavioral literature is whether behavior matures in a manner that is more consistent with a process model of prefrontal organization or with a content model. Notably, these models largely describe functions associated with the lateral PFC. They fall short of explaining how other areas of the PFC are functionally organized and how these regions can be incorporated into the same theoretical frameworks. In this vein, several authors have ascribed functions to the orbital or ventromedial PFC that differ markedly from functions typically ascribed to the DL-PFC. Based on lesion studies in macaques that demonstrate impairment in rewardbased learning and changing behavior in response to changing contingencies, Rolls (2004) has argued that the OFC enables very rapid behavioral reversals. That is, behavioral responses can be readily elicited by stimulus-reinforcement associations that allow an organism to determine which of two (or more) responses will reliably lead to a positive outcome. Once this type of association has been learned, it is not always stable within a naturalistic environment. A response that has been positively reinforced for some period of time might become maladaptive such that it is necessary to relearn which stimuli will be reinforced. The role of the OFC in managing the updating of stimulus-response contingencies as they change over time has been verified using reversal-learning tasks, in which subjects are required to learn and then reverse associations between stimuli and reward (Cools et al., 2001; Clark, Cools, and Robbins, 2004). Rolls (1998) argues that the OFC is involved in the representation of the reward value of stimuli such as taste; similarly, neurons in the OFC that represent the reward/ nonreward value of facial expressions may play a role in the social reinforcement of behavior (Rolls, 2004). Rolls argues that the OFC not only supports unique functions not conducted by other PFC areas, but also may use different processes to encode the relevant information than are used in other areas such as the DL-PFC. For instance, he indicates that synaptic modification occurs in the OFC based on stimulus-response association learning to encode the ongoing and frequently changing representation of reward value; in contrast, working-memory functions such as those supported by the DL-PFC do not use synaptic modification but rather require continuing neuronal firing to hold the representation (Rolls, 1998).

580

cognition

Damasio (1994) also has argued that the VM-PFC performs unique functions not managed by other areas such as the DL-PFC. Rather than being involved in the execution of behavior related to reinforcement contingencies, Damasio argues that the VM-PFC is responsible for associating events with representations of body states, or somatic markers, that are used to guide decision making (Damasio, Tranel, and Damasio, 1991; Damasio, 1994, 1999). Thus, although some theorists have emphasized a single overarching prefrontally mediated function (such as working memory or temporal integration), suggesting that different prefrontal regions differ not in function but in the content they manage, others have emphasized that different prefrontal regions perform different functions. It is not difficult to imagine that these perspectives might both be true, to some extent. Adjacent regions in the lateral PFC might perform identical working-memory operations on different content domains, such as nonspatial and spatial visual information, whereas the VM-PFC might perform functions not related to working memory, such as formation and updating of reward-related associations. Or, the functions of the VMPFC might be viewed as a type of “affective working memory” where the information held online is reward based. In support of the suggestion that working-memory operations for different content may be regionally segregated, there is evidence that spatial and nonspatial visual working memory may be differentially disrupted by changes in neurotransmitter levels. Luciana and Collins (1997) demonstrated that administration of a dopamine receptor antagonist impaired performance on a spatial delayed-response task, but not on a measure of delayed memory for object features (nonspatial visual delayed response task), suggesting that dopamine may play a role in facilitating memory for spatial but not nonspatial visual cues. Questions that have not been reconciled within the literature include the following: Is the prefrontal cortex functionally organized according to the content of the information to be processed, or is it organized by the level of processing that must be recruited to accomplish a given task? Is it possible that within nonhuman species or at different points in human development there might be a process-bycontent interaction? Are functions subserved by the lateral and medial regions of the prefrontal cortex distinct in the sense that the lateral PFC functions in a way that is more “cognitive” and the medial PFC functions to subserve functions that are more “affective”? Or is there really one overarching function? The examination of adolescent development and the course of maturation of prefrontally mediated behaviors may help to clarify some of these questions about the

functional organization of the PFC. Unequal maturation of spatial and nonspatial visual processing, for instance, would lend support to the suggestion that these different content domains are processed by different prefrontal regions or hemispheres (which, presumably, are developing at slightly different rates). Similarly, differences in the rate of maturation on tasks thought to rely heavily on a single prefrontal region (such as the Iowa Gambling Task and reversal learning tasks for the VM-PFC) would lend support to the suggestion that successful task performance may rely on the recruitment of additional areas outside the VM-PFC during development.

Development of attention and working memory: Process versus content There are relatively few published reports available describing adolescent maturation on prefrontally mediated tasks involving the implementation of comparable processes but for different content. Next, we will summarize work from our laboratory for visual and verbal recognition memory, visual and verbal span tasks (which measure attentional control for forward sequences and working memory for backward sequences), and visual versus verbal self-ordered search tasks. Recognition Memory Luciana and colleagues (2005) found that 9- to 10-year-olds performed at adult levels on a delayed match-to-sample task assessing the ability to make forced-choice recognition decisions for pictures of faces following a delay. Conklin and colleagues (2007), using a partially overlapping data set, found that 9- to 10-year-olds also performed at adult levels on a delayed match-to-sample task assessing the ability to make forced-choice recognition decisions for words following a delay. In addition, they found no interaction between domain (spatial versus verbal information) and age group (from age 9 through age 17), suggesting that the recognition-memory processes develop at the same rate for verbal and spatial information irrespective of content domain. Although these tasks are believed to be primarily mediated by more posterior cortical areas in the temporal lobe, the suggestion that cognitive development is organized by process rather than by content is still relevant, given that developmental processes in different but related and connected brain regions may unfold similarly. Span Tasks The spatial span test, based on the Corsi block task (Milner, 1971), requires subjects to tap blocks in a specified sequence by copying a prearranged pattern (for forward sequences) or by reversing the pattern (for backward sequences). Performance on the task is believed to be mediated by the right ventrolateral PFC (see Luciana and Nelson, 1998, for a review). On a forward spatial span task,

Luciana and Nelson (1998) found that 7- and 8-year-olds had shorter spans than young adults. Expansion of their sample to include children up to 12 years old revealed that span length still had not matured by age 12 (Luciana and Nelson, 2002). Using a different sample, Luciana and colleagues (2005) found that adult levels of performance on the task were reached by age 13–15, with no significant difference between 13- to 15-year-olds and adults on the task. Similarly, DeLuca and colleagues (2003) found that performance on spatial span tasks improved up to ages 15 to 19. On digit span tasks, patients are required to verbally reproduce a series of numbers spoken by the examiner, either in the same sequence (for digits forward) or in the reversed sequence (for digits backward). Digits forward is considered a measure of attentional efficiency or freedom from distractibility, while digits backward requires both retaining the digits and reversing operations and is therefore considered to reflect working memory (Lezak, Howieson, and Loring, 2004, 351–359). Hooper and colleagues (2004) found significant improvement in digit span performance from ages 9–10 to ages 14–17 in a healthy adolescent sample. Notably, however, the discrepancy between digit span forward and backward performance did not vary for the different age groups. In contrast, Lamm, Zelazo, and Lewis (2005) found decreasing discrepancies between digit span backward and digit span forward scores from age 7 to age 16. Normative data from the Wechsler Intelligence Scale for Children–Third Edition (WISC-III, Wechsler, 1991) indicate mildly decreasing discrepancies between forward and backward span with age (e.g., from 1.95 digits for 9-year-olds to 1.49 digits for 16-year-olds). As in Hooper and colleagues (2004), the WISC-III normative data indicate improving digit span scores until at least age 15 (Wechsler, 1991). Thus spatial span performance improves until the midteenage years (approximately age 15), and digit span performance also improves until at least age 15. These findings suggest that the processes underlying retention and manipulation of simple chunks of information have comparable developmental time courses during adolescence, regardless of whether that information is visual spatial (as in the spatial span task) or auditory (as in the digit span task). Moreover, Conklin and colleagues (2007) found no significant interaction between span task domain (verbal or spatial) and age group, confirming comparable developmental trajectories for attention and working memory irrespective of content domain. Although this finding of parallel functional maturation could be produced if the same neural systems mediate both tasks, it could also occur if the systems underlying task performance have similar developmental time courses. At a minimum, this comparison does not provide support for the claim that different areas perform the same processes on different content domains.

olson and luciana: prefrontal cortex functions in adolescence

581

Span task data suggest that similar developmental trajectories are seen when an essentially identical process is performed on different content domains. Self-Ordered Search Tasks Comparable developmental trends also have been seen on self-ordered search tasks that require more complex executive decision making. Conklin and colleagues (2007) found no interaction between domain (verbal versus spatial information) and age group on analogous self-ordered search tasks believed to involve similar processing demands, but they did find that as both tasks became more complex, the age of mastery increased. Neuropsychological studies of maturation of executive processes during adolescence have found that development proceeds at comparable rates for a given process irrespective of the informational domain involved. Thus there is no evidence that separate neural systems develop to perform similar processes on verbal versus visual information, for instance. Of course, the usual caveat holds: absence of evidence does not constitute evidence of absence. The fact that no differences are seen in the maturational pattern does not exclude the possibility that different brain regions with comparable developmental patterns are performing the same processes on different types of content. Although the evidence is far from conclusive, it is more consistent with the proposition that a single brain system is performing these processes across informational domains.

Different functions, different time courses? Consideration of the adolescent maturation of functions believed to be subserved by different prefrontal areas may, again, help to clarify the organization of the prefrontal cortex. In this section, maturational patterns that have been observed on tasks believed to be mediated by the dorsolateral or ventromedial PFC are compared. In structuring this review, we stress that we have grouped functions as primarily “dorsolateral” versus “ventromedial” based on the adult and animal literatures. It may be that these distinctions are not as pronounced during development, a point that we will return to in our concluding section. Dorsolateral PFC Development in Adolescence Spatial delayed-response task (DRT). The earliest neuropsychological studies of DL-PFC functioning employed the spatial delayed-response task. In this task, subjects are required to remember the spatial location of a stimulus during a brief delay and then respond by indicating its earlier location. Task performance on manual and oculomotor variants is disrupted in primates with lesions to the dorsolateral PFC (Goldman-Rakic, 1987). Luciana and colleagues (2005) found that on a DRT task requiring manual responses, the ability to remember one piece of spatial information

582

cognition

across a brief delay appears to stabilize by age 11 to 12, with younger participants (ages 9 to 10) showing impaired performance relative to older adolescents and adults. In contrast, Luna and colleagues (2004) found that performance on an oculomotor version of the spatial delayed-response task (in which subjects had to direct a saccade toward a remembered spatial location) improved until age 19 years in a sample ranging from 8- to 30-year-olds, suggesting the possibility of a somewhat more protracted time course for maturation of visuospatial working memory. Functional imaging studies also support these findings. In a functional neuroimaging study of 9- to 18-year-olds performing a spatial delayed response task, Klingberg, Forssberg, and Westberg (2002) found increasing activity in superior and middle frontal regions and in the parietal lobe with increasing age, demonstrating that age-related improvements in performance are associated with changes in frontal and parietal activity levels. Similarly, Kwon, Reiss, and Menon (2002) reported age-related increases in prefrontal cortical activation associated with visuospatial two-back performance, a putative measure of spatial working memory, in a small group of twenty-three 7- to 22-year-olds. Further examination of developmental maturation of behavior on spatial delayed-response tasks is needed, preferably with direct comparison of tasks requiring manual versus saccadic responses, in order to determine whether task performance matures at different rates for different response modes within the same individuals. Planning ability. The Tower of London test (Shallice, 1982) is a measure of spatial-planning ability and behavioral inhibition. From a given starting arrangement of three rings in three columns, subjects are required to move the rings in a self-determined sequence in order to reach a specified goal arrangement. The dorsolateral PFC, as well as several other cortical, subcortical, and striatal regions, is implicated in task performance (Luciana and Nelson, 1998); clinical data suggest that patients with left anterior lesions are generally more impaired on the task than those with right anterior lesions (Lezak, Howieson, and Loring, 2004, 618). Luciana and Nelson (1998) found that performance matures early (by age 5) for the problems with the easiest difficulty level, but in completing the more difficult four- and five-move problems, 8-yearolds continue to make more excess moves than adults. By age 12, performance still has not reached adult levels for five-move problems (Luciana and Nelson, 2002). Recent findings (Luciana et al., submitted) from our lab suggest that accuracy of performance stabilizes by ages 15 to 17. Similarly, Asato, Sweeney, and Luna (2006) found that children (age 8 to 13 years) showed poorer performance on four- and five-move problems than adolescents (age 14 to 17) or adults (age 18 to 30). By about

age 15, however, performance has matured; 15- to 19year-olds outperform younger adolescents (especially on more difficult problems) and achieve levels of performance on the task comparable to those of young adults (De Luca et al., 2003). However, Huizinga, Dolan, and van der Moley (2006) found significantly worse performance in 15-year-olds than in 21-year-olds, suggesting the possibility of even later maturation. Thus, although children and younger adolescents are able to understand the task demands and execute appropriate responses, as evidenced by their adult-level performance on easier problems, full maturation on this visual-spatial planning task does not occur until later adolescence. Although the task is primarily considered to reflect planning ability, performance on the task is correlated with performance on measures of response inhibition such as the antisaccade task, suggesting that maturation of response inhibition may play a role in age-related changes on the Tower of London (Asato, Sweeney, and Luna, 2006). In fact, suppressing prepotent but inappropriate responses may be a necessary component of spatial planning. Working memory and processing speed also mildly correlate with Tower of London performance (Asato, Sweeney, and Luna, 2006), suggesting that, as is the case with many measures of executive function, successful task performance relies on a heterogeneous set of abilities. Self-organized behavior. The self-ordered search task (Petrides and Milner, 1982) requires subjects to continually monitor previous selections from a set of stimuli. Patients with lesions to the lateral PFC perform poorly on the task, and task performance is considered to reflect DL-PFC activity (Petrides, 1998). The self-ordered search task requires the ability to organize a behavioral strategy and to hold spatial information in working memory to complete a goal. In the spatial variant of this task, subjects are presented with an array of scattered boxes. Tokens are hidden behind the boxes, and on each search the subject must retrieve them. Subjects are informed that the token is never hidden behind the same box more than once; therefore, during a given search, subjects must remember the previously searched boxes to avoid searching them again. Performance can be scored both in terms of the number of times that the subject searches through previously searched boxes and in terms of a strategy score reflecting the extent to which the subject adopts a sequential strategy for performing the search. For the more difficult (six- and eight-item) searches, Luciana and Nelson (1998) found that 8-year-olds made more forgetting errors than adults, with performance maturing earlier for less difficult searches. Continued improvement on the strategy score was seen through age 7, but performance on this measure had matured by age 8. Twelve-year-olds continue to show poorer performance

than adults on more difficult searches (Luciana and Nelson, 2002). Luciana and colleagues (2005) found that the number of forgetting errors continues to decline until age 13 to 15, at which point adult levels of performance are attained; performance on the strategy score matured later, with adult levels being reached at age 16 to 17. De Luca and colleagues (2003) analyzed data on this task separately for “within-search errors,” considered to reflect short-term memory capacity, and forgetting errors, reflecting working memory; they found that both short-term memory capacity and working memory improved until age 15–19, whereas performance on the strategy score did not reach adult levels until age 20–29. Thus performance on the measure most strongly related to spatial memory matures around age 15, with performance on the measure most strongly related to planning maturing in the late teens or early 20s. Adolescent development on selected dorsolateral tasks: A summary. Although most studies of adolescent performance on tasks mediated by the DL-PFC have demonstrated continuing improvement during adolescence, the time course of that improvement is somewhat variable depending on the specific task. Spatial DRT task performance has been demonstrated to improve until the late teen years (with continued improvement at least until approximately age 20) (Luna et al., 2004; Zald and Iacono, 1998) (but see Luciana et al., 2005, for evidence that performance may mature earlier, i.e., by age 11 or 12). These somewhat discrepant findings may be attributable to differences in subject characteristics (e.g., gender composition of the samples) and/or response modality (oculomotor versus manual tasks). In terms of accuracy, performance on the Tower of London matures around age 15 for more difficult items (although mature performance for easier items is seen at earlier ages) (Luciana and Nelson, 2002; De Luca et al., 2003). Problem deliberation, which supports accuracy of performance, continues until at least ages 15 to 17 (Luciana et al., submitted). Similarly, self-ordered search performance matures around age 15 for the memory-related components of the task, but it matures later (late teens to early 20s) for the strategy-related components of the task (Luciana and Nelson, 1998, 2002; Luciana et al., 2005; De Luca et al., 2003). Recently a functional neuroimaging study confirmed that performance on tasks requiring the manipulation of information over a delay is worse in younger children (ages 8–12) than in adolescents (ages 13–17) or adults (ages 18–25) and that poorer performance is related to a failure to activate the DL-PFC during the delay period (Crone et al., 2006). Though all these tasks require the maintenance and manipulation of spatial information over a delay period, there are important differences in task demands. For instance,

olson and luciana: prefrontal cortex functions in adolescence

583

the strategy score on the self-ordered search task reflects the extent to which an individual has spontaneously developed and used a (sequential) strategy to solve the problems—a skill that is never explicitly demanded by the task instructions. In this way, it differs from the spatial DRT and Tower of London tasks in which task demands are made explicit. The spatial DRT also requires the maintenance of only one piece of information at a time, whereas the self-ordered search and, to some extent, the Tower of London require maintaining multiple pieces of information. Although all these tasks are considered to be primarily mediated by the dorsolateral PFC, a universal developmental time course is not seen. There are a number of possible explanations for this lack. One explanation is that different developmental time courses are seen for these tasks because they reflect different levels of difficulty. However, this is more a restatement of the problem than an explanation for it. Why is strategic self-organization on the self-ordered search more difficult for 15-year-olds than it is for 20-yearolds, although strategic planning on the Tower of London is fully developed by age 15? The tasks may differentially recruit additional areas, outside the DL-PFC, which develop at differing rates. Moreover, tasks that require higher levels of processing (more in the way of cognitive resources or “multitasking” on an information-processing level) are those that show more extended maturational trajectories (see Luciana et al., 2005, for discussion). One type of task demand that may or may not require a greater level of information processing concerns motivational pulls on behavior. Most of the tasks attributed to the lateral prefrontal cortex tap so-called cold cognitive functions (see chapter 34 by Zelazo, Carlson, and Kesek, this volume). Though their cognitive demands are rigorous, these tasks do not require participants to make complex cognitive decisions while weighing personally relevant motivational consequences. Tasks that add this type of affective demand are typically those that have been conceptualized as more “ventromedial” in nature. In the next section, we will review what is known about the development of the ventromedial PFC in adolescence. This development is largely inferred through the use of behavioral tasks that recruit this brain region. Ventromedial PFC Development in Adolescence The Iowa Gambling Task (IGT) (Bechara et al., 1994) is thought to reflect the functioning of the ventromedial prefrontal cortex. In this task, participants choose from four decks of cards with different reward and punishment contingencies. Two decks are “advantageous,” involving small gains and small cumulative losses over time (resulting in a net gain), while two decks are “disadvantageous,” involving large gains and large cumulative losses (resulting in a net loss). In addition, one advantageous deck and one disadvantageous

584

cognition

deck are “frequent” punishment decks, in which the punishment occurs more frequently (but is smaller), while the remaining two are “infrequent” punishment decks, in which the punishment occurs less frequently (but is larger). Healthy adult participants typically learn, over the course of 100 trials, to choose more frequently from advantageous than disadvantageous decks; data are typically analyzed in terms of performance on five blocks of 20 trials (see also chapter 54 by Crone and van der Molen, this volume). Hooper and colleagues (2004) found continuing development on this task until at least age 14 to 17, with 14- to 17-year-olds making more advantageous choices overall than 9- to 10-year-olds. These age effects first appeared in block 4 of the task, in which the oldest group outperformed both 9- to 10-year-olds and 11- to 13-year-olds; by block 5, there was no significant difference between the middle group and the oldest group, suggesting that younger participants may “catch up” to older participants if given more trials in which to acquire an advantageous response set. Hooper and colleagues provided data to suggest that 14- to 17-year-olds had not reached adult levels of performance on the task, since their overall scores were descriptively lower than those of adult controls reported upon by other studies. Similarly, Lamm, Zelazo, and Lewis (2005) found improving performance on the task from age 7 to age 16, and Crone, Vendel, and van der Molen (2003) demonstrated that performance in 15- to 16-year-olds had not yet reached adult levels. Although Overman and colleagues (2004) found mature performance in 9th-graders, an extended version of the task (four blocks of 50 trials each) was used in this study; this finding is consistent with the suggestion that younger adolescents can adequately acquire an appropriate response set on the task if given enough time. Crone and van der Molen (2004), using an analogous task that presents the IGT as a “pro-social” game involving feeding a “hungry donkey” (with 200 trials in the task), found that 18- to 25-year-olds made more advantageous choices than 13- to 15-year-olds, suggesting that performance continues to improve past age 15. The oldest group shifted their preference to the advantageous decks by the third trial block (40th trial), whereas the 13- to 15-year-olds did not begin to shift until the 80th trial, and the younger groups (6- to 9-year-olds and 10- to 12-year-olds) never showed a reliable difference in preference for advantageous versus disadvantageous decks. Although all groups were more likely to switch decks following a loss than they were following a gain, the difference between the switching rates was higher for adults than for any of the other age groups, indicating that adults were more likely to stay with a deck following a net win and switch from a deck following a net loss than the other age groups. Performance on a task in which the reward and punishment contingencies were reversed indicated that

the younger groups’ impairment on the task was attributable to insensitivity to future consequences rather than sensitivity to reward. The developmental improvement in decision making on the task was not attributable to improvements in inductive reasoning ability or to working-memory capacity, suggesting that the impairment on this task is not attributable to inability to reason inductively or maintain information about gains and losses in memory (functions that are typically attributed to the dorsolateral PFC). Although other measures of VM-PFC/OFC functioning have been developed, including reversal learning tasks (Cools et al., 2001; Clark, Cools, and Robbins, 2004), there are no published studies examining the development of reversallearning task performance over the course of adolescence. For that reason, we limit our discussion of VM-PFC development to discussion of the IGT and other gambling task variants. At this point, the developmental time course of behavior on the IGT is somewhat difficult to characterize, since available studies either (1) lacked direct comparison with adults, (2) identified deficits in teenagers versus adults but did not sample across the full age range to determine the point at which adult performance is reached, and/or (3) used extended versions of the task that, by providing additional trials on which to learn the contingencies, erased the deficit typically seen in older adolescents completing the task. That said, studies using the standard task length (100 trials) have shown performance improving to the late teenage years; comparison with normative data suggests that performance probably improves into the 20s (Hooper et al., 2004). Behavioral Inhibition This section focuses on developmental changes in the ability to inhibit an inappropriate motor response; this ability is sometimes termed behavioral response inhibition or motoric response inhibition (in contrast to cognitive inhibition, which reflects [non]impulsive decision making). For instance, the ability to delay gratification or to delay obtaining an immediate but small reward in favor of a delayed but larger reward is considered to reflect (non)impulsive decision making, and the ability to inhibit a prepotent motor response is considered to reflect (non)impulsive action. Although a number of tasks have been used to examine response inhibition, including most prominently the Stroop task (Stroop, 1935) and the Eriksen flanker task (Eriksen and Eriksen, 1974), this section focuses primarily on the go-no-go task as a measure of behavioral inhibition. In a typical go-no-go paradigm, letters are presented one at a time on a screen. Participants must quickly press a response button for every letter except one (e.g., except X). This task establishes a prepotent tendency to respond that must be inhibited when the relatively infrequent X appears. A number of dependent measures may be examined, including overall error rate, false-alarm

rate (assessing errors of commission), hit rate (assessing errors of omission), and reaction time. Go-no-go tasks are difficult to conceptualize in terms of PFC mediation, because imaging studies suggest that they recruit numerous regions of the PFC. Orbital, mesial, inferior, and dorsolateral prefrontal regions all have been implicated in task performance, as have the anterior cingulate and other cortical and subcortical areas (Casey et al., 1997; Rubia et al., 2001). Because task performance is considered to involve both dorsolateral and ventromedial activation, examination of the time course of development of mature performance can help to illuminate the development not only of these separate regions but also of the regions working together as a functional system. The ability to inhibit a prepotent response improves over the course of adolescence. Levin and colleagues (1991) found that the greatest improvement in both commissions and omissions occurred from age 7–8 to age 9–12, with no further improvement seen from age 9–12 to age 13–15. However, Hooper and colleagues (2004) found decreasing false-alarm rates across all age groups in a study comparing 9–10-year-olds, 11–13-year-olds, and 14–17-year-olds. Although no differences were seen between these age groups in the hit rate, there was a significant effect of gender, with females performing better. Overall, results indicate improving inhibitory control with age and better sustained vigilance in females. Interestingly, while the go-no-go hit rate was not associated with the number of advantageous choices on the Iowa Gambling Task (IGT), it was related to the preference for the infrequent punishment decks, even after controlling for the effects of age and gender. A possible interpretation of this finding is that individuals who are predisposed to avoid frequent punishment also show better sustained vigilance over time. The stop-signal task is a similar measure of inhibitory control that requires participants to respond on each trial, except when the stimulus is followed by a cue indicating that the response should be withheld. The lag between the stimulus and the cue is adjusted based on performance; when the response is successfully withheld, the lag to the cue is increased in order to make it more difficult to withhold the prepotent response. Children (ages 6–12) show significantly lower response execution accuracy than adolescents (ages 13–17) and adults (ages 18–59) (Bedard et al., 2002). Although Bedard and colleagues (2002) do not report all possible contrasts between younger age groups, examination of their graph reveals significant improvement in reaction time to go trials until age 13–17. On the measure of stopsignal reaction time, performance stabilized by age 9–12. There were no significant differences in accuracy (on the go trials; performance on the no-go trials is calibrated to achieve comparable accuracy across subjects).

olson and luciana: prefrontal cortex functions in adolescence

585

Using a version of the go-no-go task that adjusts the presentation speed to produce equal error rates across age groups, Lamm, Zelazo, and Lewis (2005) found that 11- to 17-year-olds were able to perform the task more quickly than 7- to 10-year-olds. In addition, the amplitude of the N2 event-related-potential (ERP) component produced during the go-no-go task was related to other measures of executive function, including the IGT and the Stroop interference task, above and beyond any association attributable to age. Analysis of the event-related-potential data indicated that the N2 component was attributable to two neural generators, one in the cingulate cortex and one in the orbitofrontal cortex. Using a wider age range, Segalowitz and Davies (2004) also found associations between ERP components in the go-no-go paradigm and other measures of executive functions. A slightly more complex version of the task was used in which a first stimulus precedes a second stimulus and indicates whether or not to respond to the second stimulus. Older adolescents (age 19–25 years) had faster reaction times on go trials than younger adolescents (ages 7–17 years), although there was no age difference in error rates on the go trials. Examination of the contingent negative variation (CNV) ERP component, produced in response to the warning stimulus, demonstrated that the CNV was associated with performance on other measures of executive function (including a two-back task and a Stroop interference task) above and beyond the effects of age, with stronger associations seen in the younger age groups. Although a number of functional imaging studies have examined changes in patterns of neural activation over the course of adolescence on go-no-go tasks, the results have been quite inconsistent, with some demonstrating increasing activation in a number of areas with age and others showing decreasing activation (for a summary, see Rubia et al., 2006). For instance, Booth and colleagues (2003) found that children (ages 9–12) made significantly more errors of commission and had slower reaction times on a go-no-go task than adults. (In this version of the task, errors of commission could occur in the no-go block only). In this study, there were no brain areas with greater activation in adults than in children, although there were a number of regions that showed the reverse pattern. In contrast, Rubia and colleagues (2006) compared adolescents (ages 10–17) to adults (ages 20–43) and found that adolescents had a higher speed-accuracy trade-off and tended to favor speed over accuracy. In fact, in this study adolescents had faster mean reaction times than adults, but they were less likely to successfully inhibit behavior on no-go trials; there was a significant positive correlation between the probability of inhibiting an inappropriate response and age across the age range. Compared to ado-

586

cognition

lescents, adults showed increased activation in mesial PFC, in an area bordering the right OFC, and in the caudate nucleus. Dorsolateral versus Ventromedial PFC Development in Adolescence Longitudinal studies of structural brain development suggest that both the VM-PFC (Gogtay et al., 2004) and the DL-PFC (Giedd, 2004) continue to undergo gray matter development into the 20s. Although direct comparisons of functional development on tasks believed to recruit the VM-PFC and DL-PFC are not yet available, the existing evidence suggests that, in parallel to this protracted structural developmental time course, functional maturation in both these regions occurs into the 20s. Performance on a wide variety of tasks thought to be mediated by the DL-PFC is not fully mature until the middle to late teenage years (age 15–19 for tasks such as the self-ordered search and Tower of London) (Luciana et al., 2005; De Luca et al., 2003; Luna et al., 2004) and possibly into the 20s (De Luca et al., 2003). Similarly, performance on tasks mediated by the VM-PFC, especially the Iowa Gambling Task, also continues to mature until the middle to late teens (Hooper et al., 2004; Crone, Vendel, and van der Molen, 2003; Lamm, Zelazo, and Lewis, 2005) and possibly into the 20s (Hooper et al., 2004). A similar pattern of midteenage-year maturation is typically seen on tasks of behavioral inhibition believed to recruit both dorsolateral and ventromedial areas.

Conclusions: Developmental contributions to understanding PFC organization The adolescent maturation of performance on tasks assessing attention, memory, and working memory occurs at comparable rates for verbal and visual recognition memory, memory span, and self-ordered search. When each task is considered separately, one interpretation of this pattern is that each function may be performed by a single brain region that manages multiple content domains. It is also possible that similar developmental curves could be produced if different regions that happened to develop at comparable rates performed this same process on different types of information. Examination of adolescent maturation of performance on other tasks believed to recruit the DL-PFC indicates a nonuniform pattern of development on these tasks, with performance on tasks with explicitly delineated planning demands (such as the Tower of London) maturing earlier than performance on tasks that require spontaneous self-organization of strategic behavior (exemplified by the strategy score on self-ordered search tasks). This nonuniform pattern of development on tasks believed to rely on the DL-PFC suggests that the later-maturing functions may require coordination with other prefrontal areas that have a more

protracted developmental time course (such as, possibly, the VM-PFC). Comparison of patterns of DL-PFC and VM-PFC development is somewhat limited at this time, given that there are few published studies on adolescent maturation on tasks believed to recruit the VM-PFC besides the Iowa Gambling Task. Other tasks, such as reversal learning tasks, may provide a “cleaner” measure of VM-PFC function (Clark, Cools, and Robbins, 2004). The lack of published reports directly comparing teenage and adult performance on the standard-length task is another limitation. The available evidence suggests that maturation on both dorsolateral and ventromedial tasks continues into the 20s, with spontaneous self-organization (as in the selfordered search) and updating of behavior in response to changing consistencies (as in the IGT) maturing relatively late. There has been some speculation that both these tasks may involve coordinated activity of VM-PFC and DL-PFC; this coordination of functions performed by different regions may be one of the latest-developing aspects of behavior. Studies examining the development of executive functions during adolescence help to illuminate the functional organization of the PFC. Functions including attention (as measured by span tasks) and self-organization (as measured by self-ordered search tasks) have similar developmental trajectories, irrespective of whether the process is being performed on visual or verbal information. This similarity suggests that development of executive functioning is organized by process rather than by content domain. These findings are more consistent with models of prefrontal organization that propose that the PFC is functionally segregated according to process than models suggesting that different regions perform identical processes on different content domains. There are several ways in which our conclusions could be less tentative regarding the adolescent development of PFC functions. Nearly all published reports on development of prefrontal functions during adolescence have involved crosssectional designs; accurate delineation of patterns of prefrontal development requires longitudinal designs in order to observe developmental changes within individuals. Additionally, published reports tend to describe the age after which no significant improvement on a given task is seen, but comparisons between tasks generally are described in qualitative rather than quantitative terms. It is unclear whether discrepancies in developmental trends on different tasks represent significant differences. Finally, recent structural imaging data (Shaw et al., 2006) indicate different patterns of brain development during adolescence depending on global intelligence, with more protracted developmental courses in individuals with higher IQs. Effects of this type of individual-difference factor on the time course of

development of prefrontal functions have not been extensively investigated but might explain discrepancies across studies. Finally, there are few reports of longitudinal studies that directly examine the development of behavior on neuropsychological measures of executive function in conjunction with neuroimaging data assessing the development of prefrontal structures. By assessing whether developmental improvements in prefrontal function can be attributed to structural refinements that we know are occurring during adolescence (i.e., decreases in gray matter volume, increases in white matter volume plus organization), it should be possible to unravel how maturation of prefrontal regions relates to the continuing emergence of the executive functions over the course of adolescence and into adulthood. acknowledgments The writing of this chapter was sup-

ported by a National Science Foundation graduate fellowship awarded to E. Olson and by grant 5R01DA017843–03 awarded by the National Institutes of Health to M. Luciana.

REFERENCES Asato, M. R., J. A. Sweeney, and B. Luna, 2006. Cognitive processes in the development of TOL performance. Neuropsychologia 44:2259–2269. Baddeley, A. D., and G. Hitch, 1974. Working memory. In I. G. A. Bower, ed., The Psychology of Learning and Motivation, 47– 90. New York: Academic Press. Barbas, H., and G. J. Blatt, 1995. Topographically specific hippocampal projections target functionally distinct prefrontal areas in the rhesus monkey. Hippocampus 5:511–533. Barnea-Goraly, N., V. Menon, M. Eckert, L. Tamm, R. Bammer, A. Karchemskiy, C. C. Dant, and A. L. Reiss, 2005. White matter development during childhood and adolescence: A crosssectional diffusion tensor imaging study. Cerebral Cortex 15: 1848–1854. Bechara, A., 2004. The role of emotion in decision-making: Evidence from neurological patients with orbitofrontal damage. Brain Cogn. 55:30–40. Bechara, A., A. R. Damasio, H. Damasio, and S. W. Anderson, 1994. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 50:7–15. Bedard, A. C., S. Nichols, J. A. Barbosa, R. Schachar, G. D. Logan, and R. Tannock, 2002. The development of selective inhibitory control across the life span. Dev. Neuropsychol. 21:93– 111. Booth, J. R., D. D. Burman, J. R. Meyer, Z. Lei, B. L. Trommer, N. D. Davenport, W. Li, T. B. Parrish, D. R. Gitelman, and M. M. Mesulam, 2003. Neural development of selective attention and response inhibition. NeuroImage 20:737–751. Casey, B. J., R. J. Trainor, J. L. Orendi, A. B. Schubert, L. E. Nystrom, J. N. Giedd, F. X. Castellanos, J. V. Haxby, D. C. Noll, J. D. Cohen, S. D. Forman, R. E. Dahl, and J. L. Rapoport, 1997. A developmental functional MRI study of prefrontal activation during performance of a go-no-go task. J. Cogn. Neurosci. 9:835–847.

olson and luciana: prefrontal cortex functions in adolescence

587

Clark, L., R. Cools, and T. W. Robbins, 2004. The neuropsychology of ventral prefrontal cortex: Decision-making and reversal learning. Brain Cogn. 55:41–53. Conklin, H. M., M. L. Luciana, C. J. Hooper, and R. S. Yarger, 2007. Working memory performance in typically developing children and adolescents: Behavioral evidence of protracted frontal lobe development. Dev. Neuropsychol. 31:103–128. Cools, R., R. Barker, B. Sahakian, and T. Robbins, 2001. Enhanced or impaired cognitive function in Parkinson’s disease as a function of dopaminergic medication and task demands. Cerebral Cortex 11:1136–1143. Crone, E. A., and M. W. van der Molen, 2004. Developmental changes in real life decision making: Performance on a gambling task previously shown to depend on the ventromedial prefrontal cortex. Dev. Neuropsychol. 25:251–279. Crone, E., I. Vendel, and M. van der Molen, 2003. Decisionmaking in disinhibited adolescents and adults: Insensitivity to future consequences or driven by immediate reward? Personality and Individual Differences 35:1625–1641. Crone, E. A., C. Wendelken, S. Donohue, L. van Leijenhorst, and S. A. Bunge, 2006. Neurocognitive development of the ability to manipulate information in working memory. Proc. Natl. Acad. Sci. USA 103:9315–9320. Damasio, A. R., 1994. Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Grosset/Putnam. Damasio, A. R., 1999. The Feeling of What Happens. New York: Harcourt-Brace. Damasio, A. R., D. Tranel. and H. Damasio, 1991. Somatic markers and the guidance of behaviour: Theory and preliminary testing. In H. S. Levin, H. M. Eisenberg, and A. L. Benton, eds., Frontal Lobe Function and Dysfunction, 217–229. New York: Oxford University Press. De Luca, C. R., S. J. Wood, V. Anderson, J. A. Buchanan, T. M. Proffitt, K. Mahony, and C. Pantelis, 2003. Normative data from the CANTAB. I. Development of executive function over the lifespan. J. Clin. Exp. Neuropsychol. 25:242– 254. Eriksen, B. A., and C. W. Eriksen, 1974. Effects of noise letters upon the identification of a target letter in a non-search task. Percept. and Psychophys. 16:143–149. Fuster, J. M., 1995. Temporal processing. Ann. NY Acad. Sci. 769:173–181. Fuster, J. M., 2002. Frontal lobe and cognitive development. J. Neurocytol. 31:373–385. Giedd, J. N., 2004. Structural magnetic resonance imaging of the adolescent brain. Ann. NY Acad. Sci. 1021:77–85. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, T. Paus, A. C. Evans, and J. L. Rapoport, 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 2:861–863. Giedd, J. N., J. W. Snell, N. Lange, J. C. Rajapakse, B. J. Casey, P. L. Kozuch, A. C. Vaituzis, Y. C. Vauss, S. D. Hamburger, D. Kaysen, and J. L. Rapoport, 1996. Quantitative magnetic resonance imaging of human brain development: Ages 4–18. Cerebral Cortex 6:551–560. Gogtay, N., J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, T. F. Nugent, D. H. Herman, L. S. Clasen, A. W. Toga, J. L. Rapoport, and P. M. Thompson, 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101: 8174–8179. Goldman-Rakic, P. S., 1987. Circuitry of primate prefrontal cortex and regulation of behavior by representational memory.

588

cognition

In F. Plus and V. Mountcastle, eds., Handbook of Physiology, 373– 417. Bethesda, MD: American Physiological Society. Goldman-Rakic, P. S., 1988. Topography of cognition: Parallel distributed networks in primate association cortex. Annu. Rev. Neurosci. 11:137–156. Goldman-Rakic, P. S., 1995. Architecture of the prefrontal cortex and the central executive. Ann. NY Acad. Sci. 769:71–84. Goldman-Rakic, P. S., 1998. The prefrontal landscape: Implications of functional architecture for understanding human mentation and the central executive. In A. C. Roberts, T. W. Robbins, and L. Weiskrantz, eds., The Prefrontal Cortex: Executive and Cognitive Functions, 87–102. New York: Oxford University Press. Happaney, K., P. D. Zelazo, and D. T. Stuss, 2004. Development of orbitofrontal function: Current themes and future directions. Brain Cogn. 55:1–10. Hooper, C. J., M. Luciana, H. M. Conklin, and R. S. Yarger, 2004. Adolescents’ performance on the Iowa Gambling Task: Implications for the development of decision making and ventromedial prefrontal cortex. Dev. Psychol. 40:1148–1158. Huizinga, M., C. V. Dolan, and M. W. van der Molen, 2006. Age-related change in executive function: Developmental trends and a latent variable analysis. Neuropsychologia 44:2017– 2036. Klingberg, T., H. Forssberg, and H. Westerberg, 2002. Increased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14:1–10. Klingberg, T., C. J. Vaidya, J. D. Gabrieli, M. E. Moseley, and M. Hedehus, 1999. Myelination and organization of the frontal white matter in children: A diffusion tensor MRI study. NeuroReport 10:2817–2821. Kwon, H., A. L. Reiss, and V. Menon, 2002. Neural basis of protracted developmental changes in visuospatial working memory. Proc. Natl. Acad. Sci. USA 99(20):13336–13341. Lamm, C., P. D. Zelazo, and M. D. Lewis, 2005. Neural correlates of cognitive control in childhood and adolescence: Disentangling the contributions of age and executive function. Neuropsychologia 44:2139–2148. Levin, H. S., K. A. Culhane, J. Hartmann, K. Evankovich, A. J. Mattson, H. Harward, G. Ringholz, L. Ewing-Cobbs, and J. M. Fletcher, 1991. Developmental changes in performance on tests of purported frontal lobe functioning. Dev. Neuropsychol. 7:377–395. Lewis, D. A., 1997. Development of the prefrontal cortex during adolescence: Insights into vulnerable neural circuits in schizophrenia. Neuropsychopharmacology 16:385–398. Lezak, M. D., D. B. Howieson, and D. W. Loring, 2004. Neuropsychological Assessment, 4th ed. New York: Oxford University Press. Liston, C., R. Watts, N. Tottenham, M. Davidson, S. Niogi, A. M. Ulug, and B. J. Casey, 2006. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cerebral Cortex 16:553–560. Luciana, M., and P. F. Collins, 1997. Dopaminergic modulation of working memory for spatial but not object cues in normal humans. J. Cogn. Neurosci. 9:330–347. Luciana, M., H. M. Conklin, C. J. Hooper, and R. S. Yarger, 2005. The development of nonverbal working memory and executive control processes in adolescents. Child Dev. 76:697– 712. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36:273–293.

Luciana, M., and C. A. Nelson, 2002. Assessment of neuropsychological function through use of Cambridge Neuropsychological Testing Automated Battery: Performance in 4- to 12-year-old children. Dev. Neuropsychol. 22:595–624. Luciana, M., A. M. Schissel, P. F. Collins, and E. A. Olson, submitted. Tower of London performance in healthy adolescents: The development of planning skills and associations with self-reported impulsivity. Luna, B., K. E. Garver, T. A. Urban, N. A. Lazar, and J. A. Sweeney, 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75:1357–1372. McDonald, A. J., 1991. Organization of amygdaloid projections to the prefrontal cortex and associated striatum in the rat. Neuroscience 44:1–14. McDonald, A. J., F. Mascagni, and L. Guo, 1996. Projections of the medial and lateral prefrontal cortices to the amygdala: A Phaseolus vulgaris leucoagglutinin study in the rat. Neuroscience 71:55–75. McFarland, N. R., and S. N. Haber, 2002. Thalamic relay nuclei of the basal ganglia from both reciprocal and nonreciprocal cortical connections, linking multiple frontal cortical areas. J. Neurosci. 22:8117–8132. Miller, E. K., and J. D. Cohen, 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24:167–202. Milner, B., 1971. Interhemispheric differences and psychological processes. Br. Med. Bull. 27:272–277. Nauta, W. J. H., 1962. Neural associations of the amygdaloid complex in the monkey. Brain 85:505–520. Norman, D. A., and T. Shallice, 1986. Attention to action: Willed and automatic control of behavior. In R. J. Davidson, G. E. Schwartz, and D. Shapiro, eds., Consciousness and Self Regulation: Advances in Research, vol. 4, pp. 1–18. New York: Plenum Press. Ongur, D., X. An, and J. L. Price, 1998. Prefrontal cortical projections to the hypothalamus in macaque monkeys. J. Comp. Neurol. 401:480–505. Overman, W. H., K. Frassrand, S. Ansel, S. Trawalter, B. Bies, and A. Redmond, 2004. Performance on the IOWA card task by adolescents and adults. Neuropsychologia 42:1838–1851. Pandya, D. N., and E. H. Yeterian, 1990. Prefrontal cortex in relation to other cortical areas in rhesus monkey: Architecture and connections. In H. B. M. Uylings, C. G. Van Eden, J. P. C. DeBruin, M. A. Corner, and M. G. P. Feestra, eds., The Prefrontal Cortex: Its Structure, Function, and Pathology. Progress in Brain Research, vol. 85, pp. 63–94. Amsterdam: Elsevier Press. Parent, A., and L.-N. Hazrati, 1995. Functional anatomy of the basal ganglia. I. The cortico-basal ganglia-thalamo-cortical loop. Brain Res. Rev. 20:91–127. Paus, T., 2005. Mapping brain maturation and cognitive development during adolescence. Trends Cogn. Sci. 9:60–68. Petrides, M., 1995. Functional organization of the human frontal cortex for mnemonic processing. Ann. NY Acad. Sci. 769:85– 96. Petrides, M., 1998. Specialized systems for the processing of mnemonic information within the primate frontal cortex. In A. C. Roberts, T. W. Robbins, and L. Weiskrantz, eds., The Prefrontal Cortex: Executive and Cognitive Functions, 103–116. New York: Oxford University Press. Petrides, M., and B. Milner, 1982. Deficits on subject-ordered tasks after frontal-and temporal-lobe lesions in man. Neuropsychologia 20:249–262. Pfefferbaum, A., D. H. Mathalon, E. V. Sullivan, J. M. Rawles, R. B. Zipursky, and K. O. Lim, 1994. A quantitative magnetic

resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch. Neurol. 9:874–887. Reiss, A. L. M. T. Abrams, H. S. Singer, J. L. Ross, and M. B. Denckla, 1996. Brain development, gender and IQ in children: A volumetric imaging study. Brain 119:1763–1774. Rolls, E. T., 1998. The orbitofrontal cortex. In A. C. Roberts, T. W. Robbins, and L. Weiskrantz, eds., The Prefrontal Cortex: Executive and Cognitive Functions, 67–86. New York: Oxford University Press. Rolls, E. T., 2004. The functions of the orbitofrontal cortex. Brain Cogn. 55:11–29. Rolls, E. T., J. Hornak, D. Wade, and J. McGrath, 1994. Emotion-related learning in patients with social and emotional changes associated with frontal lobe damage. J. Neurol. Neurosurg. Psychiatry 57:1518–1524. Rose, J. E., and C. N. Woolsey, 1948. The orbitofrontal cortex and its connections with the mediodorsal nucleus in rabbit, sheep and cat. Res. Publ. Assoc. Res. Nerv. Ment. Dis. 27:210–232. Rubia, K., T. Russell, S. Overmeyer, M. J. Brammer, E. T. Bullmore, T. Sharma, A. Simmons, S. C. R. Williams, V. Giampietro, C. M. Andrew, and E. Taylor, 2001. Mapping motor inhibition: Conjunctive brain activations across different versions of go/no-go and stop tasks. NeuroImage 13:250–261. Rubia, K., A. B. Smith, J. Woolley, C. Nosarti, I. Heyman, E. Taylor, and M. Brammer, 2006. Progressive increase of frontostriatal brain activation from childhood to adulthood during event-related tasks of cognitive control. Hum. Brain Mapping 27:973–993. Sampaio, R. C., and C. L. Truwit, 2001. Myelination in the developing human brain. In C. A. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 35–44. Cambridge, MA: MIT Press. Schmithorst, V. J., M. Wilke, B. J. Dardzinski, and S. K. Holland, 2002. Correlation of white matter diffusivity and anisotropy with age during childhood and adolescence: A crosssectional structural MR imaging study. Radiology 222:212–218. Schmithorst, J. V., M. Wilke, B. J. Dardzinski, and S. K. Holland, 2005. Cognitive functions correlate with white matter architecture in a normal pediatric population: A diffusion tensor MRI study. Hum. Brain Mapping 26:139–147. Segalowitz, S. J., and P. L. Davies, 2004. Charting the maturation of the frontal lobe: An electrophysiological strategy. Brain Cogn. 55:116–133. Selemon, L. D., and P. S. Goldman-Rakic, 1988. Common cortical and subcortical targets of the dorsolateral prefrontal and posterior parietal cortices in the rhesus monkey: Evidence for a distributed neural network subserving spatially guided behavior. J. Neurosci. 8:4049–4066. Shallice, T., 1982. Specific impairments in planning. Philos. Trans. R. Soc. London [Biol.] 298:199–209. Shallice, T., and P. Burgess, 1998. The domain of supervisory processes and the temporal organization of behavior. In A. C. Roberts, T. W. Robbins, and L. Weiskrantz, eds., The Prefrontal Cortex: Executive and Cognitive Functions, 22–35. New York: Oxford University Press. Shaw, P., D. Greenstein, J. Lerch, L. Clasen, R. Lenroot, N. Gogtay, A. Evans, J. Rapoport, and J. Giedd, 2006. Intellectual ability and cortical development in children and adolescents. Nature 440:676–679. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 1999. In vivo evidence for postadolescent brain maturation in frontal and striatal regions. Nature Neurosci. 2: 859–861.

olson and luciana: prefrontal cortex functions in adolescence

589

Stroop, J. R., 1935. Studies of interference in serial verbal reactions. J. Exp. Psychol. 18:643–662. Stuss, D. T., T. Shallice, M. P. Alexander, and T. W. Picton, 1995. A multidisciplinary approach to anterior attentional functions. Ann. NY Acad. Sci. 769:191–211. Tranel, D., A. Bechara, and N. L. Denburg, 2002. Asymmetric functional roles of right and left ventromedial prefrontal cortices in social conduct, decision-making, and emotional processing. Cortex 38:589–612.

590

cognition

Wechsler, D., 1991. The Wechsler Intelligence Scale for Children—Third Edition. San Antonio, TX: Psychological Corporation. Yakovlev, P. I., and A. R. Lecours, 1967. The myelogenetic cycles of regional maturation of the brain. In A. Minkowski, ed., Regional Development of the Brain in Early Life, 3–70. Oxford, UK: Blackwell Scientific Publications. Zald, D. H., and W. G. Iacono, 1998. The development of spatial working memory abilities. Dev Neuropsychol. 14:563– 578.

36

Cognition and Aging: Typical Development JONAS PERSSON AND PATRICIA A. REUTER-LORENZ

Declining cognitive functions are a normal and inevitable part of healthy aging. Some changes may stem from global alterations in brain functions, including metabolic changes; others may result from localized decline within specific neural circuits. Current research on cognitive aging aims not only to identify the mechanisms that underlie cognitive change, but also to understand and harness the genetic, experiential, and environmental factors that promote the preservation of cognitive abilities. Recent technological advances are leading to new breakthroughs in cognitive aging research, while also posing new challenges to understand the relation between genetic, anatomical, physiological, and cognitive factors and to integrate these levels of analysis. Moreover, the availability of high-resolution neuroimaging methods is revising our perspective on aging and giving way to new ideas about the aging mind and brain. We now know that the aging brain is not simply a depleted and reduced version of the younger brain. Instead, recent cognitive neuroscience evidence points to patterns of preservation and decline, along with functional reorganization and plasticity (see Buckner, 2004; for recent reviews see Burke and Barnes, 2006; Reuter-Lorenz and Lustig, 2005). In this chapter, we review current knowledge about the effects of normal aging and its neural correlates as revealed by functional neuroimaging. Although the definition of young and older adults might vary slightly, most studies identify young adults as being between the ages of 18 and 35, and older adults as 60 years of age and up. We begin with an overview of structural changes that occur with increasing age, such as brain atrophy and white matter disruption. We then review the literature on neuroimaging of cognitive functions, with a particular focus on memory and executive functions. We have also included a discussion on how to interpret the findings of reduced and increased activation in older adults, and the possibilities for cognitive reserve capacities to reduce the effects of aging.

Age-related structural changes in the brain Up until about three decades ago, the dominant view was that aging was associated with severe changes in the brain. This was partially based on results showing massive cell loss

(Brody, 1955) and dendritic reduction (Scheibel et al., 1975, 1976) in several cortical regions in the aging brain. More recently, the long-standing belief that vast numbers of brain cells are lost as we grow older has been replaced by the knowledge that most age-related alterations in brain volume stem from rather small changes in dendritic branching and spine density, rather than neuronal loss (Burke and Barnes, 2006). Also, age-related reduction seems to be region specific rather than widespread, and the aging brain appears to be far more resilient than was previously believed. Several aging studies have shown that both gray and white matter change in volume over the life span at different rates (Bartzokis et al., 2001; Jernigan et al., 2001). Gray matter volume seems to decrease at roughly a linear rate throughout adulthood and into old age, while white matter volume shows a marginal increase through adulthood, peaking in the 40–50-year range (Bartzokis et al., 2001; Sowell et al., 2003) and decreasing thereafter. Furthermore, neither gray nor white matter volumes change uniformly throughout the brain. Some regions are subject to greater atrophy than others, and linking such volumetric decreases to changes in mental function is not always straightforward. Recent behavioral and neurobiological evidence suggests that multiple, separable factors contribute to memory decline in aging. One prominent distinction has been proposed between disruptions of executive processes that influence memory and decline in long-term (declarative) memory (Hedden and Gabrieli, 2005). This distinction, although crude, is based on behavioral observations and can be used for separating normal and pathological aging. For example, it has been shown that older adults without dementia often have problems on tasks that involve executive functions (Balota, Dolan, and Duchek, 2000; Hasher and Zacks, 1988; Jennings and Jacoby, 1993; Moscovitch and Winocour, 1995; West, 1996). Early stages of Alzheimer’s disease, in contrast, are associated with impairments in declarative memory, such as difficulty remembering lists of objects and words (Huppert, 1994), although impairments in executive function are also found (Balota and Faust, 2001). Given the relationship between executive functions and the prefrontal cortex (PFC) on one hand, and medial temporal regions and declarative memory on the other, this distinction may also

591

apply to the neural consequences of aging. For example, several large-scale cross-sectional studies of healthy individuals found no significant association between hippocampal volume and age (Bigler et al., 1997; Sullivan et al., 1995). In line with the distinction outlined previously, this finding has led some researchers to believe that normal age-related memory loss may be more related to changes in other structures, such as the prefrontal cortex, rather than medial temporal lobe (MTL) dysfunction (see, e.g., Glisky, Polster, and Routhieaux, 1995, for work distinguishing fron-tal from temporal aging patterns). Longitudinal measures of MTL volume in older adults indicate a somewhat larger decline compared to cross-sectional estimates suggesting a rate of decline in hippocampal volume of about 0.79–0.86 percent per year (Raz et al., 2005; Raz, Rodrigue, et al., 2004). With regards to cognitive decline in aging, some studies have found correlations between atrophy of the medial temporal lobe, including substructures within the medial temporal lobe, and memory performance (Golomb et al., 1996; Persson, Nybery, et al., 2006; Rodrigue and Raz, 2004), while others have not (see Van Petten, 2004, for a review). One possible explanation for these mixed results is that some studies of presumed healthy aging, especially when longitudinal progression is not taken into account, may actually include individuals in the early stages of Alzheimer’s disease. Also, there is large between-study variability in the definition of MTL measures, with some studies including substructures of the hippocampus and others not. In contrast to the rather small changes in MTL volume with age, healthy aging is characterized by more marked changes in frontal circuits. Whether frontal atrophy is due to white or gray matter change is still under debate. For example, in a study including participants across the adult age range, a negative correlation between frontal white matter volume and age was found, but it was small compared to gray matter reductions within the frontal lobes (Raz, Gunning-Dixon, et al., 2004). In other studies, the opposite pattern has been found (Jernigan et al., 2001). The disproportionate reduction in volume in prefrontal regions compared to medial temporal regions suggests that healthy aging and Alzheimer’s disease (AD) may have different neural trajectories given that the neuropathology in AD is most prominent in the MTL (e.g., Buckner, 2004; Head et al., 2005). The prominence of prefrontal declines in healthy aging supports a frontal basis for cognitive aging (West, 1996). According to this hypothesis, and as discussed earlier, the frontal lobes and their associated executive functions may be disproportionately compromised in aging, thereby influencing performance on a wide range of cognitive tasks that draw on executive processing abilities. Furthermore, behavioral performance is related to decline of prefrontal volume as measured from structural MRI. For example,

592

cognition

prefrontal volume is negatively correlated with errors caused by an inability to change strategies on the Wisconsin Card Sorting Test (Raz et al., 1998) and positively correlated with measures of fluid intelligence (Schretlen et al., 2000). One possibility is that changes in frontal lobe functions in aging may be related to disruption in corpus callosal white matter integrity, thereby affecting frontal networks involved in cognitive processes. Recently, advances in neuroimaging techniques have made it possible to assess the microstructural properties of white matter in the corpus callosum using diffusion tensor imaging (DTI; see chapter 17 by Wozniak, Mueller, and Lim, this volume). In contrast to volumetric studies of white matter, DTI is based on the molecular diffusion of water, which is influenced by microstructural factors including myelin density and other white matter fiber components (Beaulieu, 2002). In line with the distinction discussed earlier, several recent papers have noted that while changes in white matter in AD mostly affect posterior parts of the corpus callosum (Head et al., 2004), healthy aging is associated with changes in anterior parts of the corpus callosum (Head et al., 2004; Persson, Nyberg, et al., 2006). There is also recent evidence directly linking anterior white matter disruption to memory impairment (Persson, Nyberg, et al., 2006). Persson and colleagues used longitudinal memory performance over a decade to divide older participants into those that had either stable or declining memory performance over time. It was found that participants with declining memory performance had reduced white matter integrity in the anterior corpus callosum, compared to participants with stable memory performance over time (figure 36.1). Taken together, frontal disruption seems to be more prominent than MTL atrophy in healthy aging, suggesting that frontal deficits may play a more important role in age-related cognitive decline than MTL dysfunction. This finding is also evidence for microstructural changes in white matter, where disruption is most severe in the anterior parts of the corpus callosum in healthy aging, whereas posterior parts are more affected in subtypes of dementia, such as Alzheimer’s disease.

Working memory and aging Working memory (WM) is widely viewed as a cognitive system that maintains a small amount of information (approximately 1–6 items) in an active state for a short period, thereby permitting its use in comprehension, problem solving, decision making, and other higher cognitive abilities. The working-memory system is therefore central to cognition and is strongly related to variability in performance on diverse measures of mental aptitude (see for reviews Ackerman, Beier, and Boyler, 2005; Jonides, 1995). According to one prevailing view (Baddeley and Hitch,

Figure 36.1 Overview of findings from Persson, Nyberg, and colleagues (2006). (A) Stable (N = 20), declining high (N = 13), and declining low (N = 7) longitudinal memory performance in older adults (composite score from three episodic tests). (B) Results from the DTI analyses show the anterior corpus callosum (genu) outlined on a transverse slice of a fractional anisotropy (FA) image. High signal intensity (brightness) reflects higher FA. (C) Mean FA in anterior corpus callosum as a function of longitudinal memory performance.

1974), WM includes several storage and rehearsal components that are specialized for maintaining specific types of information. Verbal material is mediated by the phonological loop, whereas visual-spatial material is managed by the

visual sketch pad (see Repovs and Baddeley, 2006, for a recent view on this architecture). Working memory also includes a central executive, which mediates the host of operations needed to manipulate, recode, select, and manage the contents of the storage buffers. The working memory literature distinguishes between tasks that emphasize the rote maintenance or storage of information and tasks that also require manipulation of the stored contents. Tasks that emphasize maintenance are thought to reflect the operation of storage mechanisms and to rely minimally on executive processes. Some researchers have referred to these tasks as measures of short-term memory rather than working memory. ReuterLorenz and colleagues (Reuter-Lorenz and Sylvester, 2005) have argued that tasks that emphasize rote maintenance can recruit executive processes depending on the memory load, amount of proactive interference, and individual difference variables such as the age of the participant. According to this view, the contribution of executive processing to a task is best described by a continuum of executive demand, such that a task’s position on the continuum is influenced not only by the explicit demands of the task, but also by the recent task history (practice or interference) and the individual performing the task. This view, which was largely inspired by results from neuroimaging measures of working memory, will be considered in more detail in relation to specific experimental results. Age-related performance differences on rote maintenance tasks are substantially smaller than differences on tasks that make explicit demands on the manipulation of information in working memory (Craik and Jennings, 1992). The tendency to equate explicit task demands with underlying neural processes led to the inference that normal aging spared the processes associated with rote maintenance, thereby leaving tasks that emphasize storage intact, while compromising executive processing and thereby leading to disproportional impairment on tasks that require item manipulation, set shifting, and other processes that recruit the central executive. Neuroimaging studies, however, have revealed a very different picture. Despite high performance in seniors that differs minimally, if at all, from their younger counterparts, the neural circuitry engaged during tasks that emphasize storage differs substantially in the younger and older brain. Two of the initial neuroimaging studies of WM and aging (Grady et al., 1998; Reuter-Lorenz et al., 2000a) used PET to investigate age-related differences in WM. Grady and colleagues (1998) used faces in a match-to-sample task with a variable delay that ranged from 1 to 21 seconds. Accuracy was comparable across age groups at most delays, but older adults were generally slower than young adults. Although both young and older adults activated bilateral PFC,

persson and reuter-lorenz: cognition and aging

593

the locus of peak activations within PFC varied with age. Relative to older adults, younger adults showed greater activation in ventrolateral PFC (Brodmann area [BA] 45) regions associated with storage-related processes, whereas older adults had more activation in dorsolateral PFC regions (BA 9) that mediate executive processes. Reuter-Lorenz and colleagues (2000b) examined the effects of age on the neural correlates of verbal and spatial working memory. In their tasks, different groups of subjects viewed small sets of letters or spatial locations marked by dots on a screen and retained these memoranda in mind for approximately 3 seconds. A probe appeared, and they judged whether it matched either the letter or the location they held in memory. Both young and older adults performed the working-memory tasks with high accuracy, although the older participants were significantly slower. The important result from this study was that younger adults showed primarily left-lateralized activation for the letter-memory task and right-lateralized activation for the spatial-memory task. In contrast, older adults showed bilateral activation for both tasks. The increased activation in the older group was primarily in prefrontal sites, including both ventrolateral and dorsolateral PFC regions. This tendency for older adults to show greater bilateral activation is an important trend observed across a wide range of neuroimaging studies of cognitive aging (see for reviews Cabeza, 2002; Reuter-Lorenz, 2002) and it has been speculated that this pattern may reflect functional compensation for agerelated declines in neural efficiency, an interpretation that is supported by the positive correlation between verbal working-memory performance and right dorsolateral activation in older adults (Reuter-Lorenz et al., 2000a, 2001). In two event-related experiments using a similar paradigm, Rypma and D’Esposito (2000, 2001) also found significant age differences in frontal regions. Both experiments included a manipulation of memory load (six items versus two items) with a 12-second retention interval. The event-related design made it possible to separately assess brain activation related to the different components of encoding, maintenance, and retrieval. In both these experiments it was found that younger individuals showed greater activation in right DL-PFC compared to older individuals. This difference was restricted to the retrieval phase and was more pronounced under high-memory-load conditions. Even though these results deviate considerably from the findings by Reuter-Lorenz and colleagues (2000b, 2001), Rypma and D’Esposito (2000, 2001) also found a correlation between right DL-PFC activation and performance. For older adults, greater activation in the right DL-PFC was associated with shorter response times, which is similar to the findings by Reuter-Lorenz and colleagues (2001). Rypma and D’Esposito (2000) also reported an inverse relationship for younger adults. The finding of age differences in activation–response-time relationships is consistent with behavioral data suggesting that

594

cognition

processing speed is an important variable in age-related performance variation (Salthouse, 1996). This finding has also been observed in different brain regions and across task domains (Persson et al., 2004; Rypma et al., 2005). Taken together, regardless of behavioral performance, young and older adults recruit different brain regions in order to perform the same task, suggesting that they are engaging different neural circuitry. Also, differences in brain activation are most prominent in prefrontal and parietal regions. Furthermore, the evidence for increased bilaterality in older adults, together with the observations of performance/ activation correlations, is in line with the idea that the recruitment of brain regions in older adults is compensatory. Reuter-Lorenz and colleagues (Reuter-Lorenz et al., 2000a; Reuter-Lorenz and Mikels, 2006; Reuter-Lorenz, Stanczak, and Miller, 1999) have proposed that compensatory recruitment may result in more rapid utilization of resource reserve in seniors, which can lead to performance decrements with increasing task demands. Reuter-Lorenz and Mikels (2006) have referred to this as the compensation-related utlization of neural circuits hypothesis or CRUNCH. Evidence in support of CRUNCH has emerged from a recent study in our lab (Cappell, Gmeindl, and Reuter-Lorenz, 2006) in which memory load was varied between three and seven items in a letter-recognition task. Younger and older adults were equally accurate for lower memory loads; however, older adults recruited both left and right DL-PFC for these loads, whereas younger adults activated only left DL-PFC. At the highest load, performance for both groups decreased, but older adults dropped substantially more than their younger counterparts. Importantly, at the highest load, younger adults now recruited left and right DL-PFC, whereas seniors had reached an apparent asymptote and showed no change in activation levels in right DL-PFC compared to the pattern they showed at the next lower load. In accordance with the CRUNCH hypothesis, the findings of bilateral activation in older adults in the low-demanding condition, with no further right DL-PFC activation in the high-load condition, suggest that older adults utilize compensatory mechanisms when the demand for executive functions is low. With higher executive demands, however, older adults may have reached an asymptote, which may limit the use of compensatory mechanisms. Another recent study using a more challenging workingmemory task found a similar pattern (Mattay et al., 2006). The task, known as n-back, is a continuous performance task in which single items (i.e., letters) are presented sequentially and must be matched with a previous item that is “n” back in the sequence, where n typically equals one, two, or three items. This task requires the continuous updating, storage, comparison, and inhibition of items, and therefore places far greater demand on executive processes than item-recognition tasks. Age-related differences in behavioral performance

were found for two- and three-back loads, but they were nonsignificant for one-back. Also, although older adults had slower RTs at all levels of task difficulty, young and older adults activated a similar network of brain regions. A direct age comparison, however, showed that in the one-back task, where performance was equal across age, older adults had greater PFC (BA 9) activity bilaterally. At higher WM loads, when older adults performed worse than young adults, reduced PFC activity was found for older adults. These results suggest that as cognitive demand increases, older adults may exceed the limit beyond which compensation can be maintained. Hence a decline in performance occurs. Interference from prior information, also known as proactive interference (PI), is also greater in older adults than in younger adults (Lustig, May, and Hasher, 2001). A PET study from our lab links increased PI in older adults to frontal inhibitory processes that are compromised with age (Jonides et al., 2000). In a working-memory task, PI was introduced by presenting items that were familiar to subjects because they appeared on a previous trial but that required a “no” response because they were not part of the current memory set. Young adults are slower at rejecting such familiar items compared to less familiar ones, and this effect is associated with activation in the inferior frontal gyrus (IFG). Older adults, who show greater interference effects behaviorally, also demonstrate weaker activation in left IFG, suggesting that they are unable to recruit the frontally mediated process involved in resolving familiarity-based interference in favor of selecting the appropriate response. Recently, Gazzaley and colleagues (2005) provided evidence for age-related loss of inhibition of task-irrelevant information and the effect of this impairment on working-memory processes. In an fMRI design that separated inhibition of taskirrelevant information from enhancement of task-relevant information, they showed age-related deficits in inhibiting irrelevant (unattended) scenes or faces but preserved enhancement of relevant (attended) ones. Older participants showing the larger inhibition deficits also showed the lowest working-memory performance. This finding suggests that age-related impairment in inhibitory processes, which has been linked to executive control functions of ventrolateral regions 45 and 47 in left PFC, may lead to declines in working memory in older adults. In sum, the most common pattern is for older adults to show greater PFC activation than young adults, especially at lower task demands. This age difference is not surprising given the strong involvement of prefrontal functions in executive control. The findings of behavioral and neural age differences in working-memory tasks with high demands on the central executive suggest that these tasks might be vulnerable to age-related decline. One possibility is that the need for compensation is greater in tasks with higher

demands on executive processes, and that the lack of additional resources in some cases might constrain the success of compensation on behavioral performance.

Executive functions and aging Although working-memory tasks can function as a platform for investigating the central executive, age-related differences in executive processes are by no means limited to the working-memory domain. Given the similarities between the behavioral impairments that are shown by older adults and those demonstrated by patients with frontal lobe lesions, as well as their failure to engage in executive control processes, such as inhibition, monitoring, interference resolution, and task switching, neuropsychologists have hypothesized that prefrontal impairments are the underlying cause of cognitive aging (e.g., Moscovitch and Winocour, 1995; West, 1996). Since many cognitive tasks involve executive control to some degree, a decline in these executive components may have a negative impact on performance across a wide range of tasks. For example, Hasher and Zacks (1988) have proposed a model in which age-related deficits in inhibition underlie changes in memory function. They suggested that an ineffective inhibitory system may allow irrelevant information into working memory, thereby decreasing the specificity of memory traces. Also, deficits in the inhibitory system could reduce the ability to suppress irrelevant pathways during memory retrieval and may be a central mechanism underlying both distractibility and memory deficits in older adults. Next we discuss tasks that examine age-related differences in various executive components across task domains. Age-related difficulties in selecting task-appropriate information and inhibiting task-irrelevant responses have been demonstrated in a variety of experimental paradigms. Typically, in a conflict task, two stimulus dimensions are associated with two different responses, only one of which is correct. One such task is the Stroop task (Stroop, 1935), where color words printed in colored ink are used, and the task is to name the colors that the words are printed in. When the printed color and the color name are in conflict (incongruent condition; e.g., “red” printed in green ink), subjects are slower and make more errors because of increasing interference caused by the processing of the irrelevant dimension. Older adults’ reaction times in the Stroop task are typically slower than those of young adults. Although some studies show that the interference effect increases as well (Kwong See and Ryan, 1995), others argue that the age sensitivity of the Stroop interference effect is primarily related to general slowing with aging (Verhaegen and De Meersman, 1998). In an fMRI study on age differences in the Stroop task, Milham and colleagues (2002) found activation differences

persson and reuter-lorenz: cognition and aging

595

between young and older adults predominantly in frontal regions. Age-related differences in activation for incongruent trials were found; young adults had greater activation in left middle frontal gyrus (BA 9), the superior parietal lobule, and the anterior cingulate gyrus, whereas older adults had greater activation in bilateral IFG (BA 45/47) during high-conflict trials. Based on these results, together with the findings of marginal age difference in the interference effect, the authors argued that aging compromises the ability to implement strategies during high-conflict conditions, as reflected by less activation in frontal and parietal regions. A reduced ability to inhibit the processing of task-irrelevant information, coupled with a reduced brain response, may result in greater activation in irrelevant and non-task-specific regions. One possibility is that increased activation for older adults in the IFG reflects an attempt to compensate for age-related impairments in executive functions associated with the task. The finding of additional recruitment of frontal regions in older adults for incongruent trials was also observed by Langenecker, Nielson, and Rao (2004), suggesting that the recruitment of multiple frontal regions in Stroop conflict processing is reliable across studies. The lack of a direct behavioral link between frontal activation and measures of interference in the Stroop task makes it unclear whether this additional recruitment reflects nonspecific and taskirrelevant processing, or whether it is related to some kind of compensation. Recent data indicate that it may be possible to compensate for prefrontal dysfunction that accompanies aging. For example, a study from our lab examined interference in a verb-generation task in which subjects were presented with a noun and responded with a single verb associated with that noun (Persson et al., 2004). Some nouns had one highly associated response (e.g., chair—sit), and other nouns had several verb associates (e.g., ball—throw, catch, hit). The increasing demand for selecting between several associates is associated with increased RTs, compared to generating a verb in response to a noun with one dominant associate. Also, the increased selection demands of the “many”-associate condition activates left IFG, in a region that overlaps with the area activated in response to the need to resolve PI in working memory (Jonides et al., 2000; see working-memory section). However, the higher selection demand does not disadvantage older adults more than young ones, indicating an absence of an age-related decline on the performance of this task. Data from fMRI studies showed that older adults had reduced activation in left IFG while showing increased activation of the homologous IFG site in the right hemisphere, a pattern consistent with the possibility that the bilateral activation pattern is compensatory. Other tasks that have been used to assess age-related differences in executive processes include go-no-go tasks, where responses must be withheld unexpectedly (Nielson,

596

cognition

Langenecker, and Garavan, 2002), and task-switching tasks, where the subjects must shift between different sets of rules that govern the association between the stimulus and the response (DiGirolamo et al., 2001). Additional studies have used the flanker task (Colcombe et al., 2005), in which participants are asked to respond to a central arrow cue embedded in an array of five arrows that point either to the left or to the right. In some trials the flanking arrows point in the same direction as the cue (congruent trials), and in some trials the flanking arrows point in the opposite direction (incongruent trials). In each of these cases, age differences in performance have been associated with age differences in prefrontal activation patterns. These neuroimaging results converge with earlier neuropsychological evidence from behavioral tests of frontal lobe function indicating that prefrontal functions are especially compromised with age (Moscovitch and Winocour, 1995; West, 1996). Additional insights about executive functions and aging come from recent investigations of altered patterns of deactivations in aging. It has been suggested that in fMRI experiments, participants may be engaged in spontaneous and organized mental activity during the rest baseline. The deactivations found in many neuroimaging studies may reflect processes engaged during the resting state or “default mode” and could result from a diversion of default-mode processes to an active task (e.g., Raichle et al., 2001). Thus regions involved in mental activity during the rest baseline are deactivated (i.e., task < baseline) during the active task. A network of regions, including medial prefrontal, posterior cingulate cortex, and lateral ventral parietal regions, have been found to be deactivated by experimental tasks relative to a passive baseline (e.g., Mazoyer et al., 2001; Shulman et al., 1997). Since the diversion of resources away from mental activity during the rest baseline to the active task may involve executive processes, age differences in deactivation might indicate a failure to engage these executive processes. It has been suggested that aging is associated with altered patterns of deactivation, which may be related to declining resources, difficulties with resource allocation, or both. These possibilities predict that greater task demand, which increases deactivation levels in younger adults, should exacerbate age-related declines in allocating resources away from the default mode. In a recent study, deactivation in young and older adults was investigated (Persson et al., 2007). Two versions of the verb-generation task that varied in their demand for selection among competing alternatives (discussed earlier) were compared to word reading and a fixation baseline condition. Consistent with previous results (McKiernan et al., 2003), deactivation increased with increasing demand in both young and older adults, and both groups showed equivalent deactivations when the need for selection was minimal. However, age differences in the magnitude of deactivation

increased with higher selection demand, with older adults showing less increase in deactivation with higher task demands compared to young adults (figure 36.2). Demandrelated changes in deactivation magnitude correlated with the amount of interference, suggesting that individual and group differences in deactivation have functional significance. The findings of an inverse relationship between deactivation in specific regions and behavior suggests that deactivation may be related to the degree of cognitive efficiency. The observation of reduced deactivation for older

adults in high-control conditions may indicate a reduction in cognitive efficiency stemming from difficulty disengaging from or inhibiting internal processes in older adults to reallocate attention to the task at hand. Taken together, results from studies using diverse experimental tasks related to executive function suggest that, in many cases, the effects of increased executive demands are disproportionably larger in older adults than in young adults. This age-related impairment is often associated with altered patterns of prefrontal activation, suggesting that prefrontal deficits are common in old age. In some studies, although older adults show a general reduction in performance, the effect of increased executive demands is similar to the effect demonstrated by young adults. One possibility is that under certain conditions older adults may be able to compensate for impairments in executive processes, even if this ability may be limited during high task demands.

Long-term memory and aging

Figure 36.2 Overview of findings from Persson and colleagues (2007). Transverse sections depict the location of the areas used for the region-of-interest analyses. Bar graphs represent the average percent signal change for young and senior participants for each of the conditions (Read, Few, Many) compared to a rest baseline. Error bars represent standard error of the mean.

Memory problems are a common complaint of older people, and cognitive studies show that they are right: memory declines with age. Yet, 20th-century psychology taught us that memory is not a single, unitary entity. There are several distinct types of memory and memory components that are mediated by different neural structures and subsystems. Moreover, certain types of memory show no obvious decline during normal aging. For example, normal elderly often have rich and elaborate autobiographical memories (see, e.g., Maguire and Frith, 2004), nor do they forget how to write, ride a bicycle, or get to the nearest grocery store, and vocabulary actually increases throughout life (Park, 2002; Schaie, 1996). Here we focus on age differences in episodic memory, which is usually defined as a neurocognitive system that enables humans to remember past experiences. Episodic memory involves several dissociable subcomponents, including the learning of new information (encoding) and the storage and subsequent retrieval of information. In a typical episodic-memory experiment, participants are presented with a number of items to keep in memory and are instructed to keep these items in memory for a subsequent memory test. Following this encoding phase, participants are asked to recall as many of the previously studied items as possible (test of free recall), or they are presented with the previously studied items together with a number of nonstudied items and asked to indicate whether a particular item was previously studied or not (recognition test). Because memory relies on both successful encoding and retrieval, these components cannot be dissociated using behavioral measures alone. Therefore, neuroimaging methods provide an important complementary tool that allows for the separate assessment of brain regions that are active during episodic encoding and retrieval.

persson and reuter-lorenz: cognition and aging

597

Episodic Encoding Episodic-memory deficits in older adults seem to originate from difficulties in encoding to-beremembered information, although processes at retrieval are also thought to play a role (Craik and Byrd, 1982; Light, 1991). Behavioral findings suggest that older adults might show encoding impairments because they are less likely to spontaneously engage in effective encoding strategies that are important for remembering, and as a result, older adults’ memory performance is greatly influenced by the amount of support available from the environment (Craik, 1986; Craik and Byrd, 1982). In particular, the more support offered by the environment, the less important spontaneous strategies become, since environmental cues, such as the context in which an item was encoded, can drive memory performance. Next we review neuroimaging studies that have investigated age effects on memory encoding. In one of the first reports on episodic memory, Grady and colleagues (1995) used PET to investigate age-related differences during intentional encoding of faces. Young participants showed increased activity in left IFG (BA 45) and hippocampal regions, but older participants did not show reliable activation in either of these two regions. These findings led Grady and colleagues (1995) to conclude that encoding in old age was accompanied by reduced brain activation in prefrontal and hippocampal regions. The finding of reduced PFC and MTL activation in older adults during episodic encoding has been observed in several subsequent studies with PET and fMRI and using both verbal and nonverbal information (Anderson et al., 2000; Cabeza et al., 1997; Daselaar et al., 2003b; Grady et al., 1999; Logan et al., 2002; Stebbins, Carrillo, and Dorman, 2002). As previously discussed, reduced memory performance in older adults may be related to less spontaneous engagement of encoding strategies compared to young adults. It has been proposed that a reduction in processing resources with age may underlie insufficient encoding processes. One possibility is that reduced left frontal activation may reflect inadequate encoding strategies, reduced processing resources, or both. If so, then the requirement to perform two tasks simultaneously (and thus divert attentional resources to a second task) should be particularly disadvantageous for older individuals. Anderson and colleagues (2000) used a pairedassociate task to examine verbal memory under conditions of full and divided attention. The study was motivated by observations that young adults’ performance under conditions of divided attention closely resembles that associated with aging (Anderson, Craik, and Naveh-Benjamin, 1998). The results showed that both aging and divided attention were related to similar reductions in encoding-related activity in left ventrolateral PFC. Thus aging may adversely affect memory function because it increases the attentional load of the task in a manner parallel to that in which divided

598

cognition

attention affects task performance and reduces the leftPFC-mediated ability to engage in elaborative encoding. The view that aging is associated with a reduction in elaborative encoding is also supported by findings of age-invariant left prefrontal activation under conditions of specific encoding strategies. Logan and colleagues (2002) investigated encoding of words under three conditions: intentional memory, deep incidental processing (abstract/ concrete categorization), and shallow incidental encoding (temporal order of a letter). A similar increase in memory performance for young and old adults was found as a function of depth of encoding. In line with results from several other studies, older adults had overall less activation across prefrontal regions during shallow incidental encoding and also during intentional encoding. Regions showing agerelated underactivation included Brodmann areas 45 and 47, which are associated with semantic processing (Demb et al., 1995; Poldrack et al., 1999). A notable finding was that the age-related underactivation of PFC was not observed during deep incidental encoding. Since deep incidental encoding is a more efficient strategy than intentional memory for subsequent memory performance, the finding of reduced PFC age differences in this condition supports the idea that insufficient encoding strategies in old age can be reversed when environmental support is provided. Thus age-related differences in memory processing might be reduced by providing older adults with effective encoding strategies, such as semantic categorization. A common finding from neuroimaging studies in young adults is that left PFC regions are more engaged during encoding and that right PFC regions show greater activation than left PFC during retrieval of information from memory. This pattern of process-related asymmetry in episodic memory has been referred to as the HERA (hemispheric encoding/retrieval asymmetry) model (Nyberg, Cabeza, and Tulving, 1996; Tulving et al., 1994). Older adults often deviate from this pattern by showing a bilateral pattern of activation during encoding tasks that typically engage only left prefrontal regions in young adults. For example, using a task in which young and older adults made semantic (deep encoding) and nonsemantic (shallow encoding) judgments about words, Stebbins, Carillo, and Dorman (2002) found a bilateral activation pattern in frontal regions in older adults compared to their younger counterparts. Several subsequent studies have found a similar pattern of activation, suggesting that bilateral activation in older adults is a robust finding across studies and task materials (see Rajah and D’Esposito, 2005, for a recent review). This increased bilaterality in older adults has been conceptualized as the HAROLD (hemispheric asymmetry reduction in older adults) model (Cabeza, 2002), and the compensation hypothesis has emerged from data showing bilateral activation in older adults. This view suggests that the observed bilaterality

in older adults could be caused by compensatory recruitment to counteract the neurocognitive decline associated with aging. Note that this pattern has also been found in studies on working memory (e.g., Reuter-Lorenz et al., 2000a, discussed previously). Recent advances in fMRI experimental design have made it possible to measure the activation related to the encoding of a specific item, and then sort the items based on whether the item was remembered or not in a subsequent memory test. This advance has made it possible to address questions about activation in relation to successful encoding of a particular item versus activation for unsuccessful encoding attempts. By investigating activation-performance correlations, several brain-imaging studies have shown that there is a strong overlap between brain responses related to encoding processes and subsequent successful memory for specific information (Brewer et al., 1998; Otten, Henson, and Rugg, 2001; Wagner et al., 1998). This kind of approach is particularly valuable in providing information about the relationship between brain activation, episodic encoding, and compensation in old age. In a study by Morcom and colleagues (2003), verbal encoding was examined using an incidental semantic categorization task. When remembered items were compared to forgotten items, similar left inferior PFC activation was found for young and old adults. These results resemble the findings of Logan and colleagues (2002) showing equivalent left frontal activation during deep encoding. Old individuals also showed more bilateral anterior PFC activation for remembered items compared to their younger counterparts, suggesting that older adults engage much of the same neural circuitry as younger adults when provided with environmental support. The behavioral significance of this additional recruitment is difficult to determine from these findings. One possibility is that additional recruitment may indicate a reduction or alteration in functional specialization of brain circuits (i.e., dedifferentiation). This interpretation was supported in a recent study by Tisserand and colleagues (2005) in which encoding and subsequent recognition of words were tested in young and older adults. Younger adults performed better than older adults at the recognition task, but there were no group differences in performance during encoding. For the neuroimaging data, they found age-related differences in brain networks related to performance such that young adults activated two distinct networks for encoding (including premotor and parietal brain regions) and subsequent recognition (including middle frontal, and lateral and medial temporal regions), while older adults activated overlapping regions for both encoding and recognition (including prefrontal, premotor, and medial temporal regions). The authors suggest that the reorganization of brain networks in older adults may be related to reduced performance in subsequent recognition.

Additional recruitment of the left dorsolateral PFC (BA 8), together with a reduced activation in bilateral MTL regions, was found in the study by Gutchess and colleagues (2005) using a nonverbal (pictures) incidental deep encoding task. Given the bilateral activation often found when young subjects encode pictures (Kelley et al., 1998), increased engagement of left PFC for older adults above the level of young adults indicates that this selective activation may be compensatory. Although this additional prefrontal activation also may represent dedifferentiation, Gutchess and colleagues (2005) interpret this activation as reflecting a productive response to less efficient MTL processes. Reduced MTL recruitment in old adults was also found in the study by Daselaar and colleagues (2003a), but in contrast to the findings by Gutchess and colleagues (2005), no age-related differences were found in prefrontal regions. Together, age-related differences in prefrontal activation using the subsequent memory paradigm have been quite variable, and clearly more studies are needed to specify the functional role of prefrontal activation in aging. Reduced activation in MTL regions in older adults seems to be a consistent finding, and it parallels the observations from several other studies using different task paradigms. Episodic Retrieval Although both encoding and retrieval processes appear to contribute to age-related declines in episodic memory, behavioral observations suggest that episodic retrieval is generally less affected by age than encoding. For example, in divided-attention studies, a disproportionate episodic memory deficit in older adults has been found when attention is divided at encoding compared to retrieval (Park et al., 1989). In general, older adults show greater impairments in more effortful retrieval tasks, a conclusion reached from the consistent observation that age deficits in episodic memory are more marked for free recall than recognition tests (Schonfield and Robertson, 1966). In an early study using PET, Cabeza and colleagues (1997) investigated regional cerebral blood flow (rCBF) activation during encoding, recall, and recognition of word pairs. Consistent with the HERA model, young adults’ brain activation in prefrontal cortex was left lateralized during encoding and right lateralized during retrieval. The older adults exhibited reduced left PFC activation during encoding and a more bilateral pattern of PFC activity during retrieval. In an influential study on retrieval of context information, Cabeza and colleagues (2002) presented participants with words that were spoken aloud or presented visually and measured brain activation at recognition for item and source information. Young adults and low-performing older adults showed right lateralized PFC activations for source information. High-performing older adults, however, had bilateral activity in PFC regions, a finding consistent with compensatory activations in older individuals.

persson and reuter-lorenz: cognition and aging

599

There is also evidence for differential effects of aging on the retrieval processes of familiarity and recollection, where familiarity refers to the feeling that an item is old in the absence of contextual details and recollection to the recovery of specific contextual details of the episode during which an item was initially encountered. There is substantial behavioral evidence that recollection processes are impaired by aging, while familiarity shows minimal or no age-related decline (Bastin and Van der Linden, 2003; Mantyla, 1993; Parkin and Walter, 1992). In a recent study, the effects of aging on recollection- and familiarity-based retrieval processes were investigated using event-related fMRI (Daselaar et al., 2005). Daselaar and colleagues (2005) investigated the neural correlates of this distinction, assessing brain activity associated with recollection and familiarity using recognition confidence ratings in young and older adults. The main finding was a double dissociation within the MTL between recollection-related responses in the hippocampus, which showed reduced activity with age, and familiarity-related activity in the rhinal cortex, which showed increased activity with age. These findings suggest that the differential effects of aging on recollection versus familiarity are related to different effects of aging on hippocampal versus rhinal brain responses. Daselaar and colleagues (2005) speculate that the increase in rhinal activity in older adults might reflect compensation for reduced hippocampal-related recollection by means of reliance on rhinal-mediated familiarity. In general, brain activity patterns associated with episodic retrieval show less pronounced age differences than those associated with encoding. In several studies, older adults showed bilateral PFC activation during episodic retrieval tasks, which deviates from the pattern of typically rightlateralized activation found in young adults. Also, in several studies, frontal decreases with age were observed, predominantly in the right PFC. Age-related hippocampal differences are less pronounced for episodic retrieval than for episodic encoding, although recent findings suggest age-related dissociations within the MTL depending on whether recognition is based on recollection or familiarity responses.

Cognitive reserve, underactivation, and compensation As this review makes evident, several different patterns of age-related changes in activation have emerged in the recent neuroimaging literature. One pattern apparent in a number of studies is that younger adults activate certain brain regions to a greater extent than older adults. Since most studies on aging have found either reduced accuracy or increased response time for older adults, it is possible that such relative underactivation of older adults is related to age-related reductions in cognitive performance. Given the changes observed in brain volume and white matter integrity

600

cognition

discussed earlier in this chapter, performance and activation reductions in older age may in part be a reflection of agerelated changes in brain structure. Another possibility is that age-related underactivation may be process based, rather than indicating structural changes. For example, there are many findings from behavioral studies suggesting that older adults are less likely to spontaneously engage in elaborated cognitive processes that typically promote good behavioral performance (Craik, 1986). When given appropriate instructions and support, however, they are fully capable of using more efficient means to approach a task, thereby enhancing their performance levels. The previously mentioned study by Logan and colleagues (2002) adds neural support to this proposal by showing that underactivation of frontal regions by older adults during memory encoding can be reversed by providing of explicit encoding strategies. This elegant demonstration challenges a simple structural cause for underactivation, because age-equivalent activation was achieved by a cognitive intervention. Although age-related underactivation may be explained in terms of diminished neural resources with age, the finding of regional overactivation in older adults is more difficult to understand. How can we interpret these findings? One possibility is that overactivation reflects recruitment of additional resources to cope with task demands (i.e., extra effort). Specifically, when task demand increases, an increase in prefrontal activity is likely to occur. Such prefrontal increases have been found in working memory (Braver et al., 1997; Rypma et al., 1999) and episodic memory (Nolde, Johnson, and Raye, 1998) in studies of young adults. Overactivation has also led to speculations that the aging brain can reorganize to handle cognitive and other challenges (Cabeza et al., 2002; Gutchess et al., 2005; Reuter-Lorenz et al., 2000b). However, a recent study of structural and functional correlates of cognitive decline in aging (Persson, Nyberg, et al., 2006) provided evidence that additional regional activation may be related to cognitive decline rather than stability, and also showed that older adults who showed overactivation in prefrontal regions had reductions of hippocampal volume and white matter integrity in the anterior corpus callosum. Together with findings of increased recruitment in individuals with dementia (Bäckman et al., 1999; Grady et al., 2003) and following stroke (Buckner et al., 1996), these observations suggest that the “best case” for older adults would be to show patterns of functional brain activity that are similar to those of younger adults (i.e., neither overactivation nor underactivation). These findings, however, still leave room for a compensatory interpretation of overactivation by older adults. First, a compensation view of overactivation is supported by findings of positive correlations between behavioral performance and prefrontal activation in older adults (Cabeza et al., 2002; Reuter-Lorenz et al., 2000b, 2001; Rosen et al., 2002;

Rypma and D’Esposito, 2000). Second, although Grady and colleagues (2003) found that individuals in early stages of Alzheimer’s disease showed higher frontal recruitment than healthy older adults, they also noted that when the dementia group was considered in isolation, performance was positively correlated with frontal activation. It is therefore important to note that when the whole sample is considered, frontal recruitment may reflect detrimental performance, but within a group of individuals with declining performance, additional frontal recruitment may still show positive correlations with performance (see, e.g., Buckner, 2005; Persson, Lind, et al., 2006, for a discussion). One of the clear challenges for future research is to characterize the functional implications of age-related changes in activation. One idea that may prove to be especially important in this endeavor is the notion of “cognitive reserve” (e.g., Corral et al., 2006; Richards and Deary, 2005; Staff et al., 2004). Although the relation between reserve and compensation has yet to be worked out, it is evident that factors such as education, literacy, IQ, and lifestyle variables help reduce the susceptibility to age-related changes. Other factors include more “passive” variables of brain structure, which often include anatomical features such as brain size, synaptic connectivity, and neural density. In a recent PET study (Stern et al., 2005), a factor score that summarized years of education and intelligence was used as an index of cognitive reserve to examine differential activation in young and older adults. In young individuals, a set of brain regions whose activation increased as a function of task demands was increasingly recruited as a function of level of reserve, suggesting that these regions represent the neural substrate of cognitive reserve. In older adults, the network of regions was activated to a lesser degree in individuals with higher cognitive reserve, suggesting that instead of increasing activation in task-specific regions, older adults use an altered, compensatory network to preserve function in response to age-related physiological changes.

Conclusions and future directions It is evident from this review that many questions regarding cognitive aging remain unanswered. In addition to the need to more fully investigate questions derived from behavioral findings, cognitive neuroscience has introduced an agenda of its own. Although functional brain-imaging studies have converged on the interesting pattern of not only age-related decreases in functional brain activity but also age-related increases, the functional consequences of these age changes need to be more clearly defined. There is evidence that overactivation may be compensatory, and if this is so, we might ask what operations are mediated by this activity. Even if compensation can, to some extent, reduce age effects,

there seems to be a limit to compensation. One important goal for future research is to define these boundaries. Important issues that should be investigated also include how functional alterations can be related to structural changes that occur with increasing age. There are also reasons to examine the extent to which specific aspects of genetics and environment influence the degree of age-related changes in behavioral performance, as well as neural function and structure. Such factors may include cognitive reserve or spared neurocognitive capacity that can be drawn upon to meet mental challenges (e.g., Stern et al., 2005). The ability to compensate for the negative changes that occur with increasing age may be influenced by numerous lifestyle factors including physical fitness, exercise, diet, education, and social and intellectual engagement throughout adulthood and later life. While genetics plays an unquestionable role, environmental, cultural, and self-initiated factors clearly contribute to successful aging (for reviews see Reuter-Lorenz and Lustig, 2005; Reuter-Lorenz and Mikels, 2006) and may influence an enduring capacity for cortical plasticity that in turn permits successful adaptation to neurocognitive aging. Recent results from the cognitive neuroscience of aging suggest the possibility for adaptive reorganization and benefits from physical (Colcombe et al., 2004) and cognitive (Nyberg et al., 2003) training, and reduced age-related decline in cognition through social participation (Lövdén, Ghisletta, and Lindenberger, 2005). Together, these findings hold promise for revealing positive sides to aging that may have the potential to promote higher quality of life in the later years.

REFERENCES Ackerman, P. L., M. E. Beier, and M. O. Boyle, 2005. Working memory and intelligence: The same or different constructs? Psychol. Bull. 131(1):30–60. Anderson, N. D., F. I. M. Craik, and M. Naveh-Benjamin, 1998. The attentional demands of encoding and retrieval in younger and older adults. 1. Evidence from divided attention costs. Psychol. Aging, 13:405–423. Anderson, N. D., T. Iidaka, R. Cabeza, S. Kapur, A. R. McIntosh, and F. I. M. Craik, 2000. The effects of divided attention on encoding and retrieval related brain activity: A pet study of younger and older adults. J. Cogn. Neurosci. 12(5):775–792. Bäckman, L., J. L. R. Andersson, L. Nyberg, B. Winblad, A. Nordberg, and O. Almkvist, 1999. Brain regions associated with episodic retrieval in normal aging and Alzheimer’s disease. Neurology 52(9):1861–1870. Baddeley, A. D., and G. J. Hitch, 1974. Working memory. In G. A. Bower, ed., The Psychology of Learning and Motivation, 47–89. New York: Academic Press. Balota, D. A., P. O. Dolan, and J. M. Duchek, 2000. Memory changes in healthy older adults. In E. Tulving and F. I. M. Craik, eds., Handbook of Memory, 395–410. New York: Oxford University Press.

persson and reuter-lorenz: cognition and aging

601

Balota, D. A., and M. Faust, 2001. Attention in dementia of the Alzheimer’s type. In F. Boller and S. Cappa, eds., Handbook of Neuropsychology, vol. 6, pp. 51–80. Amsterdam: Elsevier. Bartzokis, G., M. Beckson, P. H. Lu, K. H. Nuechterlein, N. Edwards, and J. Mintz, 2001. Age-related changes in frontal and temporal lobe volumes in men: A magnetic resonance imaging study. Arch. Gen. Psychiatry 58(5):461– 465. Bastin, C., and M. Van der Linden, 2003. The contribution of recollection and familiarity to recognition memory: A study of the effects of test format and aging. Neuropsychology 17(1):14– 24. Beaulieu, C., 2002. The basis of anisotropic water diffusion in the nervous system—A technical review. NMR Biomed. 15(7–8):435– 455. Bigler, E. D., D. D. Blatter, C. V. Anderson, S. C. Johnson, S. D. Gale, R. O. Hopkins, et al., 1997. Hippocampal volume in normal aging and traumatic brain injury. AJNR Am. J. Neuroradiol. 18(1):11–23. Braver, T. S., J. D. Cohen, L. E. Nystrom, J. Jonides, E. E. Smith, and D. C. Noll, 1997. A parametric study of prefrontal cortex involvement in human working memory. NeuroImage 5:49–62. Brewer, J. B., Z. Zhao, J. E. Desmond, G. H. Glover, and J. D. E. Gabrieli, 1998. Making memories: Brain activity that predicts how well visual experience will be remembered. Science 281:1185–1187. Brody, H., 1955. Organization of the cerebral cortex. 3. A study of aging in the human cerebral cortex. J. Comp. Neurol. 102:511– 556. Buckner, R. L., 2004. Memory and executive function in aging and AD: Multiple factors that cause decline and reserve factors that compensate. Neuron 44:195–208. Buckner, R. L., 2005. Three principles for cognitive aging research: Multiple causes and sequelae, varience in expression and response, and the need for integrative theory. In R. Cabeza, L. Nyberg, and D. Park, eds., Cognitive Neuroscience of Aging, 267–286. New York: Oxford University Press. Buckner, R. L., M. Corbetta, J. Schatz, M. E. Raichle, and S. E. Petersen, 1996. Preserved speech abilities and compensation following prefrontal damage. Proc. Natl. Acad. Sci. USA 93: 1249–1253. Burke, S. N., and C. A. Barnes, 2006. Neural plasticity in the aging brain. Nature Rev. Neurosci. 7(1):30–40. Cabeza, R., 2002. Hemispheric asymmetry reduction in older adults: The HAROLD model. Psychol. Aging 17:85–100. Cabeza, R., N. D. Anderson, J. K. Locantore, and A. R. McIntosh, 2002. Aging gracefully: Compensatory brain activity in high-performing older adults. NeuroImage 17:1394– 1402. Cabeza, R., C. L. Grady, L. Nyberg, A. R. McIntosh, E. Tulving, S. Kapur, et al., 1997. Age-related differences in neural activity during memory encoding and retrieval: A positron emission tomography study. J. Neurosci. 17(1):391– 400. Cappell, K. A., L. Gmeindl, and P. A. Reuter-Lorenz, 2006. Age differences in DLPFC recruitment during verbal working memory maintenance depend on memory load. Paper presented at the Society for Neuroscience, Atlanta. Colcombe, S. J., A. F. Kramer, K. I. Erickson, and P. Scalf, 2005. The implications of cortical recruitment and brain morphology for individual differences in inhibitory function in aging humans. Psychol. Aging 20(3):363–375.

602

cognition

Colcombe, S. J., A. F. Kramer, K. I. Erickson, P. Scalf, E. McAuley, N. J. Cohen, et al., 2004. Cardiovascular fitness, cortical plasticity, and aging. Proc. Natl. Acad. Sci. USA 101(9): 3316–3321. Corral, M., M. Rodriguez, E. Amenedo, J. L. Sanchez, and F. Diaz, 2006. Cognitive reserve, age, and neuropsychological performance in healthy participants. Dev. Neuropsychol. 29(3):479– 491. Craik, F. I. M., 1986. A functional account of age differences in memory. In F. Klix and H. Hagendorf, eds., Human Memory and Cognitive Abilities, Mechanisms, and Performances. North Holland, The Netherlands: Elsevier. Craik, F. I. M., and M. Byrd, 1982. Aging and cognitive deficits: The role of attentional resources. In F. I. M. Craik and D. Tehub, eds., Aging and Cognitive Processes, 191–211. New York: Plenum Press. Craik, F. I. M., and J. M. Jennings, 1992. Human memory. In F. I. M. Craik and T. A. Salthouse, eds., Handbook of Aging and Cognition, 51–109. Hillsdale, NJ: Erlbaum. Daselaar, S. M., M. S. Fleck, I. G. Dobbins, D. J. Madden, and R. Cabeza, 2005. Effects of healthy aging on hippocampal and rhinal memory functions: An event-related fMRI study. Cerebral Cortex 16(12):1771–1782. Daselaar, S. M., D. J. Veltman, S. A. Rombouts, R. H. Lazeron, J. G. Raaijmakers, and C. Jonker, 2003a. Neuroanatomical correlates of episodic encoding and retrieval in young and elderly subjects. Brain 126:43–56. Daselaar, S. M., D. J. Veltman, S. A. Rombouts, J. G. Raaijmakers, and C. Jonker, 2003b. Deep processing activates the medial temporal lobe in young but not elderly adults. Neurobiol. Aging 24:1005–1011. Demb, J. B., J. E. Desmond, A. D. Wagner, C. J. Vaidya, G. H. Glover, and J. D. E. Gabrieli, 1995. Semantic encoding and retrieval in the left inferior prefrontal cortex: A functional MRI study of task difficulty and process specificity. J. Neurosci. 15:5870–5878. DiGirolamo, G. J., A. F. Kramer, V. Barad, N. J. Cepeda, D. H. Weissman, M. P. Milham, et al., 2001. General and taskspecific frontal lobe recruitment in older adults during executive processes: A fMRI investigation of task-switching. NeuroReport 12:2065–2071. Gazzaley, A., J. W. Cooney, J. Rissman, and M. D’Esposito, 2005. Top-down suppression deficit underlies working memory impairment in normal aging. Nature Neuosci. 8(10):1298–1300. Glisky, E. L., M. L. Polster, and B. C. Routhieaux, 1995. Double dissociation between item and source memory. Neuropsychology 9(2):229–235. Golomb, J., A. Kluger, M. J. deLeon, S. H. Ferris, M. Mittelman, J. Cohen, et al., 1996. Hippocampal formation size predicts declining memory performance in normal aging. Neurology 47(3):810–813. Grady, C. L., A. R. McIntosh, S. Beig, M. L. Keightley, H. Burian, and S. E. Black, 2003. Evidence from functional neuroimaging of a compensatory prefrontal network in Alzheimer’s disease. J. Neurosci. 23:986–993. Grady, C. L., A. R. McIntosh, F. Bookstein, B. Horwitz, S. I. Rapoport, and J. V. Haxby, 1998. Age-related changes in regional cerebral blood flow during working memory for faces. NeuroImage. 8:409–425. Grady, C. L., A. R. McIntosh, B. Horwitz, J. M. Maisog, L. G. Ungerleider, M. J. Mentis, et al., 1995. Age-related reductions in human recognition memory due to impaired encoding. Science 269:218–221.

Grady, C. L., A. R. McIntosh, M. N. Rajah, S. Beig, and F. I. M. Craik, 1999. The effects of age on the neural correlates of episodic encoding. Cerebral Cortex 9(8):805–814. Gutchess, A. H., R. C. Welsh, T. Hedden, A. Bangert, M. Minear, L. L. Liu, et al., 2005. Aging and the neural correlates of successful picture encoding: Frontal activations compensate for decreased medial-temporal activity. J. Cogn. Neurosci. 17(1): 84–96. Hasher, L., and R. T. Zacks, 1988. Working memory, comprehension, and aging: A review and a new view. In G. H. Bower, ed., The Psychology of Learning and Motivation, vol. 22, pp. 193–225. New York: Academic Press. Head, D., R. L. Buckner, J. S. Shimony, L. E. Williams, E. Akbudak, T. E. Conturo, et al., 2004. Differential vulnerability of anterior white matter in nondemented aging with minimal acceleration in dementia of the Alzheimer type: Evidence from diffusion tensor imaging. Cerebral Cortex 14(4):410–423. Head, D., A. Z. Snyder, L. E. Girton, J. C. Morris, and R. L. Buckner, 2005. Frontal-hippocampal double dissociation between normal aging and Alzheimer’s disease. Cerebral Cortex 15:732–739. Hedden, T., and J. D. Gabrieli, 2005. Healthy and pathological processes in adult development: New evidence from neuroimaging of the aging brain. Curr. Opin. Neurol. 18(6):740–747. Huppert, F. A., 1994. Memory function in dementia and normal aging—Dimension or dichotomy? In F. A. Huppert, C. Brayne, and D. W. O’Connor, eds., Dementia and Normal Aging, 291–330. Cambridge, UK: Cambridge University Press. Jennings, J. M., and L. L. Jacoby, 1993. Automatic versus intentional uses of memory: Aging, attention, and control. Psychol. Aging 8(2):283–293. Jernigan, T. L., S. L. Archibald, C. Fennema-Notestine, A. C. Gamst, J. C. Stout, J. Bonner, et al., 2001. Effects of age on tissues and regions of the cerebrum and cerebellum. Neurobiol. Aging 22:581–594. Jonides, J., 1995. Working memory and thinking. In D. Osherson and E. E. Smith, eds., An Invitation to Cognitive Science, 215–265. Cambridge, MA: MIT Press. Jonides, J., C. Marshuetz, E. E. Smith, P. A. Reuter-Lorenz, R. A. Koeppe, and A. Hartley, 2000. Age differences in behavior and PET activation reveal differences in interference resolution in verbal working memory. J. Cogn. Neurosci. 12(1): 188–196. Kelley, W. M., F. M. Miezin, K. B. McDermott, R. L. Buckner, M. Raichle, N. J. Cohen, et al., 1998. Hemispheric specialization in human dorsal frontal cortex and medial temporal lobe for verbal and nonverbal memory encoding. Neuron 20:927–936. Kwong See, S. T., and E. B. Ryan, 1995. Cognitive mediation of adult age differences in language performance. Psychol. Aging 10(3):458–468. Langenecker, S. A., K. A. Nielson, and S. M. Rao, 2004. f MRI of healthy older adults during Stroop interference. NeuroImage 21:192–200. Light, L. L., 1991. Memory and aging: Four hypotheses in search of data. Annu. Rev. Psychol. 42:333–376. Logan, J. M., A. L. Sanders, A. Z. Snyder, J. C. Morris, and R. L. Buckner, 2002. Underrecruitment and nonselective recruitment: Dissociable neural mechanisms associated with aging. Neuron 33:827–840. Lövdén, M., P. Ghisletta, and U. Lindenberger, 2005. Social participation attenuates decline in perceptual speed in old and very old age. Psychol. Aging 20(3):423–434.

Lustig, C., C. P. May, and L. Hasher, 2001. Working memory span and the role of proactive interference. J. Exp. Psychol. [Gen.] 130(2):199–207. Maguire, E. A., and C. D. Frith, 2004. The brain network associated with acquiring semantic knowledge. NeuroImage 22(1): 171–178. Mantyla, T. 1993. Knowing but not remembering: Adult age differences in recollective experience. Mem. Cogni. 21(3):379– 388. Mattay, V. S., F. Fera, A. Tessitore, A. R. Hariri, K. F. Berman, S. Das, et al., 2006. Neurophysiological correlates of age-related changes in working memory capacity. Neurosci. Lett. 392(1–2): 32–37. Mazoyer, B., L. Zago, E. Mellet, S. Bricogne, O. Etard, O. Houdé, et al., 2001. Cortical networks for working memory and executive functions sustain the conscious resting state in man. Brain Res. Bull. 54(3):287–298. McKiernan, K. A., J. N. Kaufman, J. Kucera-Thompson, and J. R. Binder, 2003. A parametric modulation of factors affecting task-induced deactivation in functional neuroimaging. J. Cogn. Neurosci. 15(3):394–408. Milham, M. P., K. I. Erickson, M. T. Banich, A. F. Kramer, A. Webb, T. Wszalek, et al., 2002. Attentional control in the aging brain: Insights from an f MRI study of the stroop task. Brain Cogn. 49:277–296. Morcom, A. M., C. D. Good, R. S. J. Frackowiak, and M. D. Rugg, 2003. Age effects on the neural correlates of successful memory encoding. Brain 126:213–229. Moscovitch, M., and G. Winocour, 1995. Frontal lobes, memory, and aging. Ann. NY Acad. Sci. 769:119–150. Nielson, K. A., S. A. Langenecker, and H. Garavan, 2002. Differences in the functional neuroanatomy of inhibitory control across the adult life span. Psychol. Aging 17(1):56–71. Nolde, S. F., M. K. Johnson, and C. I. Raye, 1998. The role of prefrontal cortex during tests of episodic memory. Trends Cogn. Sci. 2(10):399–406. Nyberg, L., R. Cabeza, and E. Tulving, 1996. PET studies of encoding and retrieval: The HERA model. Psychon. Bull. Rev. 3(2):135–148. Nyberg, L., J. Sandblom, S. Jones, A. Stigsdotter Neely, K. M. Petersson, M. Ingvar, et al., 2003. Neural correlates of training-related memory improvement in adulthood and aging. Proc. Natl. Acad. Sci. USA 100:13728–13733. Otten, L. J., R. Henson, and M. D. Rugg, 2001. Depth of processing effects on neural correlates of memory encoding: Relationship between findings from across- and within-task comparisons. Brain 124:399–412. Park, D. C., 2002. Models of visuospatial and verbal memory across the adult life span. Psychol. Aging 17:299–320. Park, D. C., A. D. Smith, W. N. Dudley, and V. N. Lafronza, 1989. Effects of age and a divided attention task presented during encoding and retrieval on memory. J. Exp. Psychol. [Learn. Mem. Cogn.] 15:1185–1191. Parkin, A. J., and B. M. Walter, 1992. Recollective experience, normal aging, and frontal dysfunction. Psychol. Aging 7(2):290– 298. Persson, J., J. Lind, A. Larsson, C. M. Ingvar, M. Cruts, C. Van Broeckhoven, et al., 2006. Altered brain white matter integrity in healthy carriers of the apoe e4 allele: A risk for AD? Neurology 66:1029–1033. Persson, J., C. Lustig, J. K. Nelson, and P. A. Reuter-Lorenz, 2007. Age differences in deactivation: A link to cognitive control? J. Cogn. Neurosci. 19(6):1021–1032.

persson and reuter-lorenz: cognition and aging

603

Persson, J., L. Nyberg, J. Lind, A. Larsson, L. G. Nilsson, M. Ingvar, et al., 2006. Structure-function correlates of cognitive decline in aging. Cerebral Cortex 16(7):907–915. Persson, J., C.-Y. C. Sylvester, J. K. Nelson, K. M. Welsh, J. Jonides, and P. A. Reuter-Lorenz, 2004. Selection requirements during verb generation: Differential recruitment in older and younger adults. NeuroImage 23:1382–1390. Poldrack, R. A., A. D. Wagner, M. W. Prull, J. E. Desmond, G. H. Glover, and J. D. E. Gabrieli, 1999. Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. NeuroImage 10:15–35. Raichle, M. E., A. M. MacLeod, A. Z. Snyder, W. J. Powers, D. A. Gusnard, and G. L. Shulman, 2001. A default mode of brain function. Proc. Natl. Acad. Sci. USA 98(2):676–682. Rajah, M. N., and M. D’Esposito, 2005. Region-specific changes in prefrontal function with age: A review of PET and fMRI studies on working and episodic memory. Brain 128:1964–1983. Raz, N., F. Gunning-Dixon, D. Head, K. M. Rodrigue, A. Williamson, and J. D. Acker, 2004. Aging, sexual dimorphism, and hemispheric asymmetry of the cerebral cortex: Replicability of regional differences in volume. Neurobiol. Aging 25:377–396. Raz, N., F. M. Gunning-Dixon, D. Head, J. H. Dupuis, and J. D. Acker, 1998. Neuroanatomical correlates of cognitive aging: Evidence from structural magnetic resonance imaging. Neuropsychology 12:95–114. Raz, N., U. Lindenberger, K. M. Rodrigue, K. M. Kennedy, D. Head, A. Williamson, et al., 2005. Regional brain changes in aging healthy adults: General trends, individual differences and modifiers. Cerebral Cortex 15(11):1676–1689. Raz, N., K. M. Rodrigue, D. Head, K. M. Kennedy, and J. D. Acker, 2004. Differential aging of the medial temporal lobe: A study of a five-year change. Neurology 62:433–438. Repovs, G., and A. Baddeley, 2006. The multi-component model of working memory: Explorations in experimental cognitive psychology. Neuroscience 139(1):5–21. Reuter-Lorenz, P. A., 2002. New visions for the aging mind and brain. Trends Cogn. Sci. 6(9):394–400. Reuter-Lorenz, P. A., J. Jonides, E. E. Smith, A. Hartley, A. Miller, C. Marschuetz, et al., 2000a. Age differences in the frontal lateralization of verbal and spatial working memory revealed by PET. J. Cogn. Neurosci. 12:174–187. Reuter-Lorenz, P. A., J. Jonides, E. E. Smith, A. Hartley, A. Miller, C. Marshuetz, et al., 2000b. Age differences in the frontal lateralization of verbal and spatial working memory revealed by PET. J. Cogn. Neurosci. 12:174–187. Reuter-Lorenz, P. A., and C. Lustig, 2005. Brain aging: Reorganizing discoveries about the aging mind. Curr. Opin. Neurobiol. 15(2):245–251. Reuter-Lorenz, P. A., C. Marshuetz, J. Jonides, E. E. Smith, A. Hartley, and R. Koeppe, 2001. Neurocognitive aging of storage and executive processes. Eur. J. Cogn. Psychol. 13(1–2): 257–278. Reuter-Lorenz, P. A., and J. A. Mikels, 2006. The aging mind and brain: Implications of enduring plasticity for behavioral and cultural change. In P. B. Baltes, P. A. Reuter-Lorenz, and F. Rosler, eds., Lifespan Development and the Brain: The Perspective of Biocultural Co-constructivism, 255–278. New York: Cambridge University Press. Reuter-Lorenz, P. A., L. Stanczak, and A. C. Miller, 1999. Neural recruitment and cognitive aging: Two hemispheres are better than one, especially as you age. Psychol. Sci. 10(6):494– 500.

604

cognition

Reuter-Lorenz, P. A., and C. -Y. C. Sylvester, 2005. The cognitive neuroscience of working memory and aging. In R. Cabeza, L. Nyberg, and D. Park, eds., Cognitive Neuroscience of Aging, 186– 218. New York: Oxford University Press. Richards, M., and I. J. Deary, 2005. A life course approach to cognitive reserve: A model for cognitive aging and development? Ann. Neurol. 58(4):617–622. Rodrigue, K. M., and N. Raz, 2004. Shrinkage of the entorhinal cortex over five years predicts memory performace in healthy adults. J. Neurosci. 24(4):956–963. Rosen, A. C., M. W. Prull, R. O’Hara, E. A. Race, J. E. Desmond, G. H. Glover, et al., 2002. Variable effects of aging on frontal lobe contributions to memory. NeuroReport 13:2425– 2428. Rypma, B., J. S. Berger, H. M. Genova, D. Rebbechi, and M. D’Esposito, 2005. Dissociating age-related changes in cognitive strategy and neural efficiency using event-related f MRI. Cortex 41:582–594. Rypma, B., and M. D’Esposito, 2000. Isolating the neural mechanisms of age-related changes in human working memory. Nature Neurosci. 3(5):509–515. Rypma, B., and M. D’Esposito, 2001. Age-related changes in brain-behavior relationships: Evidence from event-related functional MRI studies. Eur. J. Cogn. Psychol. 13:235– 256. Rypma, B., V. Prabhakaran, J. E. Desmond, G. H. Glover, and J. D. E. Gabrieli, 1999. Load-dependent roles of frontal brain regions in the maintenance of working memory. NeuroImage 9(2):216–226. Salthouse, T. A., 1996. The processing-speed theory of adult age differences in cognition. Psychol. Rev. 103(3):403–428. Schaie, K. W., 1996. Intellectual Development in Adulthood: The Seattle Longitudinal Study. Cambridge, UK: Cambridge University Press. Scheibel, M. E., R. D. Lindsay, U. Tomiyasu, and A. B. Scheibel, 1975. Progressive dendritic changes in aging human cortex. Exp. Neurol. 47(3):392–403. Scheibel, M. E., R. D. Lindsay, U. Tomiyasu, and A. B. Scheibel, 1976. Progressive dendritic changes in the aging human limbic system. Exp. Neurol. 53(2):420–430. Schonfield, D., and B. A. Robertson, 1966. Memory storage and aging. Can. J. Psychol. 20:228–236. Schretlen, D., G. D. Pearlson, J. C. Anthony, E. H. Aylward, A. M. Augustine, A. Davis, et al., 2000. Elucidating the contributions of processing speed, executive ability, and frontal lobe volume to normal age-related differences in fluid intelligence. J. Int. Neuropsychol. Soc. 6:52–61. Shulman, G. L., J. A. Fiez, M. Corbetta, R. L. Buckner, F. M. Miezin, M. E. Raichle, et al., 1997. Common blood flow changes across visual tasks. 2. Decreases in cerebral cortex. J. Cogn. Neurosci. 9:648–663. Sowell, E. R., B. S. Peterson, P. M. Thompson, S. E. Welcome, A. L. Henkenius, and A. W. Toga, 2003. Mapping cortical change across the human life span. Nature Neuosci. 6: 309–315. Staff, R. T., A. D. Murray, I. J. Deary, and L. J. Whalley, 2004. What provides cerebral reserve? Brain 127:1191– 1199. Stebbins, G. T., M. C. Carrillo, and J. Dorman, 2002. Aging effects on memory encoding in the frontal lobes. Psychol. Aging 17:44–55. Stern, Y., C. Habeck, J. Moeller, N. Scarmeas, K. E. Anderson, H. J. Hilton, et al., 2005. Brain networks associated with

cognitive reserve in healthy young and old adults. Cerebral Cortex 15(4):394–402. Stroop, J. R., 1935. Studies of interference in serial verbal reactions. J. Exp. Psychol. 28:643–662. Sullivan, E. V., L. Marsh, D. H. Mathalon, K. O. Lim, and A. Pfefferbaum, 1995. Age-related decline in MRI volumes of temporal lobe gray matter but not hippocampus. Neurobiol. Aging 16(4):591–606. Tisserand, D., A. R. McIntosh, F. M. Van der Veen, W. H. Backes, and J. Jolles, 2005. Age-related reorganization of encoding networks directly influences subsequent memory performance. Cogn. Brain Res. 25:8–18. Tulving, E., S. Kapur, F. Craik, M. Moscovitch, and S. Houle, 1994. Hemispheric encoding/retrieval asymmetry in episodic

memory: Positron emission tomography findings. Proc. Natl. Acad. Sci. USA 91(6):2016–2020. Van Petten, C., 2004. Relationship between hippocampal volume and memory ability in healthy individuals across the lifespan: Review and meta-analysis. Neuropsychologia 42:1394– 1413. Verhaegen, P., and L. De Meersman, 1998. Aging and the Stroop effect: A meta-analysis. Psychol. Aging 13(1):120–126. Wagner, A. D., D. Schacter, M. Rotte, W. Koustaal, A. Maril, A. M. Dale, et al., 1998. Building memories: Remembering and forgetting of verbal experiences as predicted by brain activity. Science 281:1188–1191. West, R. L., 1996. An application of prefrontal cortex function theory to cognitive aging. Psychol. Bull. 120:272–292.

persson and reuter-lorenz: cognition and aging

605

37

Cognition and Aging: Dementia MISCHA DE ROVER, SHARON MOREIN-ZAMIR, ANDREW D. BLACKWELL, AND BARBARA J. SAHAKIAN

The demographic aging of populations characterizes both economically developed and developing countries worldwide. Aging has long been associated with cognitive decline in many domains, influencing speed of processing, long-term memory, and other skills such as problem solving (e.g., Craik and Salthouse, 2000; see also chapter 36 by Persson and Reuter-Lorenz, this volume). Among age-related cognitive disorders, dementia is at the forefront, in terms of personal as well as social costs. Dementia, as opposed to normal aging, is characterized not only by a much more rapid decline of cognitive function, but also by the accompanying loss of behavioral functionality, changes in personality, and eventual death. It is estimated that dementia currently affects approximately 28 million people worldwide, with an estimated 775,200 cases in the United Kingdom alone (Knapp and Prince, 2007). As populations age worldwide, these prevalence rates can be expected to increase substantially (Hebert et al., 2003; Petersen et al., 1999). The most common dementia, Alzheimer’s disease (AD), afflicts 10 percent of individuals over 65 years of age in the United States, and more than 50 percent of individuals over 80 years of age (Brookmeyer, Gray, and Kawas, 1998). Not only can a diagnosis of dementia have a devastating personal impact upon the lives of patients and their caregivers, but it is also associated with substantial financial costs. To illustrate, the worldwide costs for dementia care are currently US$248 billion (Knapp and Prince, 2007). The economic impact of dementia is expected to worsen in oncoming years with the demographic transitions worldwide. Current criteria for the diagnosis of probable AD as put forth in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) include (1) memory impairment as well as (2) aphasia and/or apraxia, agnosia, or impairment in executive function. The deficits must also include a significant impairment in social or occupational functioning and must constitute a change from a previous level of performance. The neuropathologic hallmarks of AD include amyloid-rich senile plaques, neurofibrillary tangles of phosphorylated tau protein, neuronal degeneration, and synaptic loss. Disruptions of multiple major neurotransmitters also characterize AD, with cholinergic abnormalities being the most prominent. The changes in the brain occur initially in selective regions, with changes first seen in the transentorhinal cortex,

later spreading to the entorhinal cortex and hippocampus proper (Braak and Braak, 1991). Accordingly, AD is characterized by reduced cholinergic neurons in the basal forebrain and selective loss of nicotinic receptors in the hippocampus and cortex (Desai and Grossberg, 2005). Ultimately the effects of dementia impact the entire brain, with regional emphasis (Fox, Warrington, and Rosser, 1999). From a clinical perspective, probable AD patients often present with difficulty in new learning and memory (Sahakian et al., 1988). In particular, AD is characterized by deficits in episodic memory (i.e., memory for specific events or experiences that can be defined in terms of time and space) (Welsh et al., 1992). Critically, however, substantial neuropathological change may have occurred before the appearance of clinically significant symptoms (Jack et al., 1999; Killiany et al., 2002; Visser et al., 2002). As a consequence, treatment of AD at the time of clinical diagnosis may be suboptimal or even ineffective because of the advanced stage of neurodegeneration at that time. Identifying cognitive tests that are sensitive to early pathological changes would facilitate the diagnosis of patients in whom the pathological process is present but whose symptoms are currently subclinical. Additional forms of dementia include vascular dementia (VaD), dementia with Lewy bodies (DLB), and frontotemporal dementias (FTDs), with VaD being the second most common form of dementia, comprising approximately 20 percent of dementia cases in the United Kingdom (Knapp and Prince, 2007). It is increasingly apparent that each form of dementia has a distinct cognitive profile, at least in the mild stages of the disease, and that this reflects the pattern of underlying neuropathological change (Hodges, 1998; Lee et al., 2003; Rahman, Robbins, and Sahakian, 1999; Rahman et al., 1999). With the development of potential therapies, there is a need for reliable, sensitive, and specific in vivo markers for each form of dementia. Thus significant recent advances in our understanding of the cognitive neuropsychology of the dementias have emphasized the importance of neuropsychological assessment in the early stages of dementia as well as differential diagnosis of the various dementias. Moreover, advances in cognitive neuropsychology enable the objective monitoring of changes in cognitive function across the disease course or as a function of treatment, and allow the estimation of functional

607

status. Hence, knowledge from cognitive assessments can be important in aiding diagnosis; informing management, care, and planning; evaluating the efficacy of a treatment; and providing information about competency for legal matters as well as for future research. This chapter will focus on AD, because this is the most prevalent and probably the best understood dementia from a neuropsychological perspective. Consideration of additional dementias will emphasize the differences between their neurocognitive profiles and that of AD.

Cognition in the early stages of Alzheimer’s disease The most common initial complaint reported by patients or their family members is that of memory difficulties. Specifically, anterograde episodic memory is impaired, leading to a compromised ability to remember new information over a period of time. Patients with AD are also impaired on tests of semantic memory and knowledge, such as naming and fluency tests (Hodges, 2006). Semantic memory refers to conceptual and factual knowledge as well as the meaning of words. Although not all patients demonstrate deficits in attention at the early stages, many patients diagnosed with AD will suffer from deficits in selective or inhibitory attention as well as divided attention (Lezak, 1995). Attentional dysfunction in AD has been linked to the involvement of the basal forebrain cholinergic system (Lawrence and Sahakian, 1995). Clear deficits in sustained attention, or vigilance, as measured by continuous-performance tasks, manifest themselves only at later stages of the disease (Sahakian et al., 1988). Visuospatial and perceptual deficits as well as clear impairments in executive function may be present, but these are not often identified in the early stages of the disease. Yet as the disease progresses, these along with all facets of cognition demonstrate impairment, eventually manifesting in the inability to function independently in daily life. Until recent years, the early detection of dementia was neglected. This neglect resulted in part from the absence of effective treatments, but it was also a result of the unavailability of sufficiently sensitive neuropsychological tests. The emergence of drugs to treat the symptoms of Alzheimer’s disease (e.g., rivastigmine, donepezil, galantamine) has placed much emphasis on the need to detect the onset of cognitive deterioration. Moreover, it is hoped that several drugs currently in development will modify disease progression (Geerts, 2007; Masters and Beyreuther, 2006; Vellas et al., 2007). Nevertheless, if neuroprotective agents that modify the disease process are to be effective, or indeed if their efficacy is to be evaluated, then it is vital that dementia be detected early and accurately, before the emergence of global cognitive impairment and substantial and irreversible atrophic damage (Fox, Warrington, and Rosser, 1999; Blackwell et al., 2004). Early detection of dementias can also

608

cognition

provide patients and their families with time to come to terms with the diagnosis of dementia, to make the necessary personal and financial arrangements, and to reduce the anxiety patients may feel when they are unsure of their diagnosis (Morgan and Baade, 1997; see also Holroyd, Snustad, and Chalifoux, 1996; Holroyd, Turnbull, and Wolf, 2002; Geldmacher et al., 2003). Thus early detection would serve to maximize the potential therapeutic benefit of treatment, enhance patient quality of life, and, in so doing, reduce the burden on residential and nursing care services. Consequently, a very high therapeutic and economic premium is placed on the early detection and diagnosis of AD. The focus to identify individuals in the prodrome of AD and other dementias has primarily been on patients who report memory problems but who do not fulfill criteria for clinically probable AD, as their nonmemory cognitive faculties and daily living activities are preserved. Such patients typically suffer from deficits in episodic memory with poor encoding and rapid forgetting of new material (Petersen, 2004). This condition has been given several labels, including “questionable dementia” (see, e.g., Swainson et al., 2001; Blackwell et al., 2004), and is commonly known as mild cognitive impairment (MCI) (Petersen et al., 1999; Petersen, 2004). The conceptualization of MCI has focused on the earliest point in the cognitive decline of individuals who are destined to develop AD (Petersen, 2004). Cognitive function in MCI patients lies along the continuum between normal ageappropriate function and mild AD (Petersen, 2004). The most prominent type of MCI is amnestic MCI. Formally, amnestic MCI is defined by a memory complaint by the patient in addition to objective memory impairment relative to appropriate age and educational norms. At the same time the patient must have normal general cognition and demonstrate preserved activities of daily living (i.e., does not meet criteria for dementia). In addition to the amnestic MCI subtype, other subtypes include multiple domain MCI and single nonmemory domain MCI. It has been proposed that these latter types may characterize dementias other than AD such as DLB and FTD (Petersen, 2004). At this time, however, further research is needed to evaluate the validity and utility of these subtypes. Approximately 7–15 percent per annum of a sample meeting criteria for MCI will “convert” to meet criteria for probable Alzheimer’s disease, several times the conversion rate expected in a general population (Celsis, 2000; Bennett et al., 2002; Petersen, 2004). There is, however, controversy regarding the prognostic utility of MCI as it has been currently defined. In particular, the utility of the subjective memory complaint (Jorm et al., 1997; Busse et al., 2003), the stability of MCI diagnosis over time, and the lack of consensus as to a particular test and criteria to deem a deficit clinically significant have all been

questioned (Larrieu et al., 2002; Ritchie, Artero, and Touchon, 2001; see also Collie, Maruff, and Currie, 2002; Petersen, 2004). The identification of appropriate tests with which to objectively quantify memory impairment is likely to improve the prognostic utility of MCI as a nosological entity. Moreover, such tests would facilitate the crucial step of identifying patients who may benefit from disease-modifying treatments while avoiding exposure of those whose condition is unlikely to progress to AD to potential side effects of these drugs, as well as unnecessary costs. In the search to identify neuropsychological tests that are sensitive to the cognitive markers of MCI and prodromal AD, it is critical to ensure that performance on such tests is not negatively affected by other neuropsychiatric complaints that may confuse diagnosis. For example, patients suffering from depression may perform similarly to MCI patients on various cognitive tasks (Swainson et al., 2001). Furthermore, in order for maximum diagnostic sensitivity to be achieved, it is important to test cognitive functions that are subserved by brain areas directly implicated in AD neuropathogenesis. The pathological markers seen in autopsy studies of AD are neurofibrillary tangles and neuropil threads. In a seminal review, Braak and Braak (1991) charted the neuropathological staging of these pathological markers, determining that the tangles and threads are initially seen in the transentorhinal cortex and then spread to the entorhinal cortex and hippocampus before destroying all cortical association areas. Converging evidence from lesion studies in humans (Smith and Milner, 1981) and experimental animals (McDonald and White, 1993; Miyashita et al., 1998) and functional neuro-imaging studies in normal volunteers (Maguire et al., 1998; Owen et al., 1996) suggests that these brain areas are necessary for visuospatial associative learning. Accordingly, it is likely that a decline in visuospatial associative learning ability may be a good candidate marker of early neuropathological abnormality. Several longitudinal studies have shown some promising results regarding the relative sensitivity of various neuropsychological tests of memory for Alzheimer’s disease in its prodromal phase. Fox and colleagues (1998) followed 63 asymptomatic individuals at risk of autosomal-dominant Alzheimer’s disease over a six-year period. Post hoc analysis showed that the ten subjects who developed dementia during this time already demonstrated cognitive impairments at first assessment, when they were ostensibly unaffected. As a group the patients who went on to receive a probable AD diagnosis had significantly reduced verbal-memory scores and, to a lesser degree, reduced performance IQ scores. These subjects initially did not differ in terms of age, family history, and initial Mini–Mental State Examination scores. First assessment was typically 2–3 years before symptoms were manifest and 4–5 years before a diagnosis of probable AD was made. Such a result clearly illustrates the potential

sensitivity of cognitive testing, possibly even more sensitive than structural brain imaging data (discussed later). Thus the earliest cognitive deficits seen in AD may include objective episodic-memory impairments that may even precede the onset of subjective-memory complaints. Other longitudinal studies in elderly subjects generally dovetail with these results. In another study, “preclinical” deficits in verbal recall preceded clinical diagnosis of AD in some cases by more than six years (Elias et al., 2000; Linn et al., 1995). Similarly, in the Bronx aging study, two tests of verbal memory (delayed recall from the Buschke Selective Reminding Test and recall from the Fuld Object Memory Evaluation) were found to predict a subgroup of subjects who would go on to develop AD (Masur et al., 1994; see also Albert et al., 2001; Artero et al., 2003). Several studies have utilized the Cambridge Neuropsychological Test Automated Battery (CANTAB) to ascertain whether early Alzheimer’s disease could be detected in a group of individuals with “questionable dementia” (QD; Fowler et al., 1995, 1997, 2002). People with questionable dementia present with subjective memory complaints, and may or may not show some degree of impairment on standard neuropsychological tests. The latter consideration is important because it differentiates questionable dementia from MCI. Nevertheless, these patients do not fulfill NINCDS-ADRDA (National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer’s Disease and Related Disorders Association) criteria for dementia. The NINCDS-ADRDA criteria are practically identical to the DSM criteria, but they are research diagnostic criteria rather than clinical in nature (McKhann et al., 1984). The CANTAB Paired Associates Learning (PAL) scores of the questionable dementia group suggested that visuospatial associative learning may hold predictive value (Fowler et al., 1997). The CANTAB PAL requires subjects to learn and remember abstract visual patterns associated with various locations on a touch-sensitive computer screen. Patterns are presented in one of six or eight boxes around the edge of the screen (figure 37.1). After a brief delay, the same patterns are presented in the middle of the screen, and subjects are required to touch the box in which they saw that pattern appear (figure 37.2). If this step is not completed correctly, the subject is reminded where each pattern belonged and tested again. This process continues until the task is satisfactorily completed or ten trials have been attempted (www. cantab.com). The CANTAB PAL scores of the questionable dementia group fell into two clusters (Fowler et al., 1997). A follow-up longitudinal study revealed that individuals in one cluster, characterized by declining PAL performance, had a poor prognosis and an increased likelihood of a diagnosis of AD. However, individuals in another cluster, characterized by stable PAL performance, had a good prognosis and remained unimpaired (Fowler et al., 1997, 2002). A later

de rover et al.: cognition and aging: dementia

609

Figure 37.1 Encoding phase of the CANTAB paired-associates learning task. One, two, three, six, or eight patterns are displayed sequentially in the available boxes in random order (in the eightpattern stage two more boxes are added).

Figure 37.2 Recall phase of the CANTAB paired-associates learning task. Each pattern presented in the encoding phase is presented in the center of the display in random order, and the subject is required to touch the box in which the pattern was previously seen. If all the responses are correct, the test moves on to the next stage; an incorrect response results in all the patterns being redisplayed in their original locations, followed by another recall phase. The task terminates after ten presentations and recall phases if all patterns have not been placed correctly.

longitudinal study found that assessing stability of PAL performance was not even necessary. Performance on the PAL as assessed at one time point alone could already be used to accurately predict incident probable AD diagnosis in a questionable dementia group (Swainson et al., 2001; Blackwell et al., 2004). The utility of the CANTAB PAL in early and differential diagnosis of AD on a case-by-case basis has been confirmed by further studies. The CANTAB PAL performance of patients with mild Alzheimer’s disease was impaired relative to both demographically matched healthy controls (Sahakian et al., 1988) and to individuals with frontal variant frontotemporal dementia (Lee et al., 2003). The PAL six-pattern-errors measure was the most useful measure in classifying individual patients into appropriate diagnostic categories. Specifically, this measure produced only 7 percent

610

cognition

overlap between AD patients and a combined nondemented group, including depressed and healthy controls (Swainson et al., 2001). This result suggests (see figure 37.3) that the PAL task is of utility in the differential diagnosis of early AD and depression (unlike word recall tests—see O’Carroll et al., 1997). In contrast to the Alzheimer’s Disease Assessment Scale–Cognitive Scale (ADAS-cog), performance on the PAL task was also found to correlate significantly with subsequent deterioration in global cognitive function. Furthermore, in a group of individuals with questionable dementia, baseline PAL results revealed an apparent subgroup of patients who performed like AD patients (Swainson et al., 2001). In a follow-up study on the same patients, Blackwell and colleagues (2004) showed that performance on the PAL task was still found to be the most predictive measure, with correct classification of 82 percent of the patients who converted to probable AD (NINCDS-ADRDA criteria) and 97 percent of patients who did not convert to probable AD. By taking into account age and performance on one other neuropsychological test (the Graded Naming Test [GNT]; McKenna and Warrington, 1980), the CANTAB PAL yielded a 100 percent distinction between the converters and nonconverters 32 months after baseline testing (see also de Jager, Milwain, and Budge, 2002). The authors hypothesized that the PAL is sensitive to early neurodegeneration in the transentorhinal region while the GNT is sensitive to neurodegeneration in the temporal neocortex proper. They proposed that the combination of the PAL and GNT tests provided diagnostic sensitivity while accommodating the heterogeneity in the locus and spread of early AD pathology within the medial temporal regions. These studies also revealed that the sensitivity of CANTAB PAL in detecting prodromal AD in a QD group and its specificity in differentiating AD from depression were considerably better than those of all other frequently used tests, including the ADAS-cog and Wechsler Logical Memory Delayed Passage Recall. In sum, accumulating evidence demonstrates the sensitivity and specificity of CANTAB PAL for the early and differential diagnosis of AD. This evidence suggests that the test represents a potential tool for operationalizing the criteria for objective-memory impairment in MCI along with other candidate tests (see Petersen, 2003). Additional promising results suggest that fast decay of iconic memory may also characterize MCI patients (Lu et al., 2005). However, CANTAB PAL has the additional advantage that it is very brief to administer, yielding a lot of useful information in an efficient way. Further, CANTAB PAL exists in a form that can be used in experimental animals, and indeed a version of the test carried out by monkeys has been shown to be sensitive to the effects of systemic scopolamine and ketamine (Taffe et al., 2002; see also Taffe et al., 2004). Such a translational medicine approach could facilitate the discovery of

Figure 37.3 Scores of individual subjects on PAL, ADAS-cog, pattern recognition, and logical memory recall (i.e., the four tests best able to classify AD subjects and depression (De) or control (Co) subjects). Notice that PAL and ADAS-cog are scored in terms of errors, so a high score indicates poor performance, whereas pattern recognition and logical memory recall are scored in terms of items correct, so a low score indicates poor performance. “F” indicates

that the subjects failed to reach the six-pattern stage. In the “questionable dementia” (QD) group, baseline PAL results revealed an apparent subgroup of patients who performed like AD patients. PAL performance of individuals with depression did not differ significantly from that of matched control subjects. (Swainson et al., 2001; reprinted with permission from S. Karger AG, Basel.)

agents that may arrest the progress of AD. Further research and consensus into the utility and operationalization of each of the MCI criteria will doubtless improve their prognostic utility. Ultimately however, the results from neuropsychological tests that are found to be sensitive and specific for early AD should be incorporated into AD diagnostic criteria, rather than invoking amnestic MCI as an intermediary state.

Moreover, from a research perspective, mapping the rate of decline of patient subgroups may yield a more precise characterization of each subtype. When investigating cognitive decline over time, test sensitivity and practice effects should be taken into account. At present the efficacy of tests developed for the detection of MCI or prodromal AD for the assessment of cognitive decline over time is uncertain. Several CANTAB tests (Sahakian et al., 1989, 1993), as well as ADAS-cog (Rosen, Mohs, and Davis, 1984; Mohs et al., 1997), were shown to be useful in the assessment of the progression of dementia. However, as discussed previously, other objective measures of cognitive function have been shown to be more sensitive and specific in the early stages of Alzheimer’s disease and therefore may provide more appropriate indices of change at these early stages.

Measuring change over time in dementia In order to measure the rate of progression of disease, an identical assessment can be administered, usually in parallel forms, at a number of different time intervals. This procedure is important because it allows clinical assessment of the efficacy of any treatment such as pharmacological intervention.

de rover et al.: cognition and aging: dementia

611

Differential cognitive profiles of the dementias In parallel to the requirement for early and accurate detection of Alzheimer’s disease, there is also a need to accurately differentiate the different forms of dementia. Over the last two decades our understanding of the neurocognitive profiles of various dementias has advanced significantly. This knowledge can be used to facilitate accurate differential diagnoses, which are often difficult to make (Bradshaw et al., 2006). An accurate differential diagnosis is important because it allows clinicians to make appropriate treatment decisions and to provide appropriate advice to patients and relatives regarding the likely nature and course of decline. Frontotemporal Dementia versus AD Frontotemporal dementia has a prevalence similar to that of AD in individuals aged 45 to 64 (Ratnavalli et al., 2002). Frontotemporal dementia refers to a group of non-Alzheimer’s conditions most frequently characterized by a predominant pathology of either the frontal lobes (frontal variant) or of the temporal lobes (temporal variant). The frontal variant is often referred to as frontotemporal dementia (FvFTD), whereas the temporal variant is often referred to as semantic dementia (SD). Frontotemporal dementia is usually associated with impaired performance on tests sensitive to frontal lobe damage in the absence of severe amnesia, aphasia, or perceptual or spatial disorders. However, such cognitive dysfunction is frequently accompanied by marked behavioral changes including lack of insight, disinhibition, loss of personal and social awareness, mental rigidity and inflexibility, perseverative behavior, and emotional lability (Gregory and Hodges, 1996; Orrell and Sahakian, 1991). In semantic dementia there is a selective impairment of semantic memory, which can be measured across tasks, suggesting disruption to an amodal semantic system (Rogers et al., 2006). This causes severe anomia, single-word comprehension, reduced exemplar production in category fluency tests, and impaired general knowledge. Other components of speech production, perceptual and nonverbal problem-solving abilities, and episodic memory are relatively spared (Hodges et al., 1992; Breedin, Saffran, and Coslett, 1994; Snowden, Goulding, and Neary, 1989; Saffran and Schwartz, 1994). Many studies have investigated whether cognitive performance could yield differential profiles for the subtypes of frontotemporal dementia. Studies from our laboratory (Rahman, Robbins, and Sahakian, 1999; Rahman, Sahakian, et al. 1999) indicate that performance on neuropsychological tests of decision making that are sensitive to orbitofrontal/ventromedial cortex could be used to differentiate mild FvFTD patients from IQ-matched control subjects. By contrast, patients and controls did not show performance differences on tasks thought to be sensitive to dorsolateral prefrontal cortex function. A different study

612

cognition

directly compared the neuropsychological profiles of FTD (both frontal and temporal variants) and early Alzheimer’s disease (Hodges et al., 1999). Alzheimer’s patients showed a substantial deficit in episodic memory. Namely, performance on the logical memory subtest of Wechsler Memory Scale–Revised (Wechsler, 1989) and a 45-minute recall of the Rey complex figure drawing (Rey, 1941) are impaired. Likewise, AD patients also revealed more modest deficits in semantic memory (memory for meaning and facts) and visuospatial skills. In contrast, semantic dementia was associated with major isolated deficits in semantic memory and anomia. The FvFTD group was overall the least impaired group, showing mild deficits in episodic memory and verbal fluency with relatively persevered semantic memory. A follow-up study with the same diagnostic groups that examined attention, executive function, and semantic memory has recently complemented these findings. Specifically, SD patients showed a preservation of attention and executive function with severe deficits in semantic memory, whereas FvFTD patients showed the opposite pattern (Perry and Hodges, 2000). Further reinforcement for the value of CANTAB PAL has been suggested by a study that examined the differentiation of SD, FvFTD, and AD (Lee et al., 2003). Performance on the PAL, as indexed by stages completed and errors at the six-pattern stage of patients with FvFTD and SD, was generally spared when compared to patients with AD. By contrast, the PAL “memory score” measure was impaired in both SD and AD but not FvFTD. Hence, the cognitive performance provides converging evidence that different pathologies can be measured reliably as different cognitive profiles. Vascular Dementia versus AD While practically all dementias are characterized by heterogeneous presentations, VaD is perhaps necessarily the most variable of all dementias. There are numerous subtypes of VaD, primarily distinguished by the general location (cortical versus subcortical) or exact cause for neuronal damage (hypoperfusion versus intracerebral hemorrhaging) (O’Brien, 2006). Cortical VaD predominantly results from large cortical infarcts, whereas subcortical VaD predominantly results from subcortical lacunae. The most prominent neuropathological features of VaD include microinfarction, diffuse white matter disease, and perivascular changes. The cognitive changes associated with VaD are, not surprisingly, variable. Nevertheless, some generalizations have emerged, suggesting deficits in attention, executive function, and speed of information processing (Desmond, 2004; Stephens et al., 2004). The cross-study comparison between VaD and AD has been impeded by problems such as differing diagnostic criteria and failure to match adequately for demographic factors and disease severity, but some consensus is now forming. Generally, when compared to patients with probable AD, VaD patients

appear to show relatively preserved episodic memory but impaired executive and attentional function. Both AD and VaD show deficits of a comparable severity in the domains of language, constructional abilities, and working memory (Looi and Sachdev, 1999). A recent meta-analysis examined 16 studies comparing various Wechsler subtests in VaD and AD groups (Oosterman and Scherder, 2006). As expected, the AD groups outperformed VaD groups in tests sensitive to executive function including Object Assembly and Digit Span backward. In contrast, VaD groups performed better than AD groups in semantic memory tests including the Information subscale, while no difference in Vocabulary was noted. Comparisons between AD and subcortical VaD indicated that the latter group performed worse on various subtests sensitive to executive functions. In addition to the tests mentioned, these subtests included Picture Arrangement, Block Design, and Picture Completion, thus suggesting more pronounced executive dysfunction in the subcortical VaD subgroup. Nevertheless, the authors of this meta-analysis caution that in any given study the small effect sizes and small group differences may lead to nonsignificant results (Oosterman and Scherder, 2006). Hence, tests developed specifically to gauge executive function and various memory types are recommended in order to distinguish VaD and AD, as well as dividing the VaD into meaningful subgroups. Semantic and episodic memory were explored in detail in a study by Graham, Emery, and Hodges (2004). This study examined neurocognitive performance in groups of patients with subcortical vascular dementia and probable Alzheimer’s disease while matching the groups for age, education, global cognitive function, and everyday function. In addition to several memory tests, the study also included assessment of executive, attentional, and visuospatial function in the testing battery. In accordance with previous literature (see Looi and Sachdev, 1999), AD was associated with more impaired episodic memory, whereas VaD was associated with more deficient executive, attentional, and visuospatial function, including dual performance task and verbal fluency. Surprisingly, semantic memory was also found to be more impaired in the VaD group. However, the relative preserved semantic memory in the AD group in this study could possibly be attributed to the mild stage of this particular sample. Alternatively, differences in test difficulty may account for the unexpected results. Logistic regression analysis indicated that the most useful discriminator of VaD and AD in this study was a combination of WAIS logical memory–delayed recall and performance of a test of naming silhouetted objects at various rotations, which is a test that can be considered to measure both visuospatial processing and semantic processing. In sum, although some consensus is now forming attributing larger executive-function deficits primarily to VaD and

larger episodic-memory deficits primarily to AD, further studies are required. Specifically it would be beneficial to target patients at earlier stages of VaD and to focus on particular subtypes of this dementia. Dementia with Lewy Bodies versus AD It is estimated that dementia with Lewy bodies is the third most prevalent form of dementia, representing approximately 15 percent of all dementia cases (Knapp and Prince, 2007). This condition is characterized by fluctuating but progressive cognitive decline, as well as recurrent visual hallucinations, motor features of Parkinsonism, and heightened sensitivity to neuroleptic drugs (McKeith et al., 1994, 2000). The leading neuropathological finding consists of Lewy body formations, which are mostly localized in the cerebral cortex, brain-stem nuclei (substantia nigra and locus coeruleus), and basal forebrain (Calderon et al., 2001). The neurocognitive profile of DLB is less well understood than that of AD and VaD, possibly because it incorporates both some facets of cognition mediated by cortical substrates and others mediated by subcortical substrates. The neurocognitive differences between DLB and AD may be particularly important, as clinicopathological studies have found that many cases of DLB are misdiagnosed as probable or possible AD during life (Bradshaw et al., 2006). Matching-to-sample visual recognition memory was one of the first aspects of cognition to be compared between DLB and AD patients (Sahgal et al., 1992). The results revealed that although the performance of both DLB and AD patients was impaired relative to controls, patients with DLB performed most poorly, particularly when a delay was introduced between stimulus presentation and recognition. In a follow-up study of mnemonic processing (Sahgal et al., 1995), DLB and AD patients were contrasted on a selfordered test of spatial working memory (CANTAB Spatial Working Memory: SWM). The results revealed that SWM was more impaired in the DLB group than in the AD group, but the authors suggested that, since the two groups did not differ from each other in the performance of a different test of spatial span, the SWM deficit may not reflect a specific impairment per se, but rather reflect the failure of a nonmnemonic cognitive process subserved by frontostriatal circuitry (i.e., attention or planning). Shimomura and colleagues (1998) compared the performance of age-, MMSE-, sex-, and education-matched DLB and AD patients on standard neuropsychological tests from the ADAS-cog, Raven’s Progressive Matrices (RPM; Raven, 1965), and the Wechsler Adult Intelligence Scale–Revised (Wechsler, 1981). Relative to AD, DLB was associated with a greater degree of impairment on the ADAS construction score, WAIS-R picture arrangement, Block Design, Object Assembly, and Picture Substitution scales. Also, RPM score was significantly worse in the DLB group. By contrast,

de rover et al.: cognition and aging: dementia

613

ADAS word-recall performance was generally spared in the DLB relative to the AD group. More recently, DLB was found to be associated with deficits in visual object and space perception relative to AD and controls (Calderon et al., 2001). Patients with DLB also demonstrated deficits in tests of attention and executive function that were greater than those seen in AD at a similar stage of global cognitive decline. At the same time the DLB patients were significantly less impaired on tests of episodic memory. The latter result has been replicated several times and appears to be very reliable (e.g., Guidi et al., 2006; Johnson, Morris, and Galvin, 2005). Additionally, DLB, but not AD, patients showed impaired performance in divided attention (Calderon et al., 2001). Thus it would appear that attention deficits are more widespread in DLB, affecting all aspects of attention. Attentional deficits were not only found to be generalized in nature, but also increased in magnitude as greater demands were placed on attentional selectivity (Bradshaw et al., 2006). Thus attentional deficits in DLB, compared to AD, appear most pronounced under conditions that require more active recruitment of executive control and visuospatial processing. In the future, a direct comparison of the cognitive profiles of VaD subtypes and DLB may elucidate potential similarities and differences between them. The two dementias have also been compared in terms of the rate of decline of cognitive function, as fluctuating cognition has proven difficult to identify reliably (McKeith et al., 2000). Generally, longitudinal studies have indicated that the rate of decline has been similar between DLB and AD (Ballard et al., 1998; Helmes et al., 2003; Johnson, Morris, and Galvin, 2005). One notable difference has been that recognition memory was found to decline more rapidly in patients with AD than in patients with DLB, converging with the increased memory deficits in AD (Stavitsky et al., 2006). This study suggested that recognition memory is more likely to be similar in the two patient groups during early stages and to diverge at later stages, with AD patients progressing more rapidly. In summary, although more research is needed, studies to date suggest that DLB is characterized by a relative sparing of episodic memory with better performance on verbal recall tasks. At the same time, DLB patients demonstrate more pronounced attentional, executive, visuospatial, visuoconstructive, and visuoperceptual deficits when compared to AD patients, particularly in the early course of the dementia.

Functional and structural brain abnormalities in dementia A major part of the challenge in the search to identify neuropsychological tests that are sensitive to the cognitive markers of MCI and prodromal AD may be the plasticity of

614

cognition

the human brain. Several plasticity-related mechanisms are known to compensate for pathology-induced brain damage. For instance, if a certain brain area is damaged, as is the substantia nigra in Parkinson’s disease, the remaining intact neurons in this brain region will “work harder” to compensate for the damage. Therefore, the pathological process may be present in the human brain with no accompanying behavioral symptoms. In the case of Parkinson’s disease, 80 percent of the dopaminergic neurons in the substantia nigra are lost before the symptoms become apparent (Hornykiewicz, 1963). Although no such data are available yet for dementia, it seems plausible that similar plasticityrelated mechanisms cause the pathological process to go unnoticed for quite a while. Brain-imaging techniques have been used to directly identify the pathological processes in the brain, independent from the symptoms. Structural magnetic resonance imaging (MRI) analyzed with volumetric tracing or voxel-based morphometry was used to show right medial temporal lobe atrophy in MCI patients (Pennanen et al., 2005) and volume reductions in the entorhinal cortex and hippocampus in MCI and AD patients (Du et al., 2001). Entorhinal cortex atrophy was suggested to precede hippocampal atrophy in early AD (Pennanen et al., 2004). This result is in line with earlier findings, mentioned previously, that neurofibrillary tangles and neuropil threads are found in the entorhinal cortex earlier in the pathological process and spread to the hippocampus in a later stage (Braak and Braak, 1991). However, medial temporal atrophy is not an AD-specific feature, thus hampering its use as a diagnostic tool (Laakso, 2002). While neuropsychological assessment seems to be more reliable in predicting subsequent progression of patients to AD, structural MRI as a diagnostic tool remains a controversial issue (Laakso, 2002). One potential approach to circumvent this problem could be mapping a number of regions and defining a pattern of atrophy, in order to increase the specificity of the method (Laakso, 2002). Another approach to increase the specificity of brainimaging techniques as a diagnostic tool is to involve functional measurements. Given that a main symptom in AD is difficulty in new learning and memory (Sahakian et al., 1988), it is not surprising that this has been the functional measurement of choice. Notably, brain activity during episodic memory has been studied, because this is most affected in AD (Welsh et al., 1992). Independent from any symptoms, functional brain-imaging techniques can be used to show plasticity-related mechanisms compensating for pathologyinduced brain damage during new learning and episodic memory. In an associative encoding task of novel pictureword pairs, MCI patients showed increased functional MRI (fMRI) responses in the posterior hippocampal, parahippocampal, and fusiform regions (Hamalainen et al., 2007). Hippocampal volume was decreased in MCI compared to

controls, and hippocampal volume was negatively correlated with parahippocampal activation in MCI patients only (Hamalainen et al., 2007). Thus it was concluded that the increased fMRI responses in MCI patients were compensatory because of the hippocampal atrophy (Hamalainen et al., 2007). Further, fMRI was used to show changes in brain function as the pathological process progressed. During encoding of novel face-name pairs, mildly impaired MCI subjects showed hyperactivation of the hippocampus compared with controls, whereas more severely impaired MCI subjects demonstrated significant hypoactivation, similar to the levels observed in mild AD subjects (Celone et al., 2006). These data indicate that in less impaired MCI patients hyperactivation of the hippocampus during new learning and memory compensates for pathological damage, whereas in more impaired MCI patients, compensatory mechanisms break down, possibly because the damage has become too extensive, leading to the first AD-related symptoms. In contrast to picture-word pairs and the face-name pairs, the CANTAB PAL task relies on nonverbal stimuli, thus reducing intercultural biases and the use of nonverbal strategies, and facilitating comparison with animal models (Taffe et al., 2004). Further, as described earlier, the PAL task was shown to be very useful in the early detection and differential diagnosis of prodromal AD. Together these developments led to the use of the PAL task in fMRI experiments in order to directly link these different areas of research in a translational approach. With a variant of the PAL task, Gould and colleagues (2003) found that increasing cognitive demands or task difficulty involves the same, rather than an additional, network of brain regions being relatively more active. Use of the same paradigm while controlling for confounds of varying task difficulty and subsequent performance yielded remarkably similar brain activations during successful paired-associate learning in AD patients and healthy controls (Gould et al., 2005). Task-induced deactivations in brain activity are often found using fMRI. One interpretation of task-induced deactivation has been that deactivations arise as a consequence of the suspension of task-independent processing that occurs when the brain is at rest or in a passive state (e.g., Binder et al., 1999; Hutchinson et al., 1999). Alzheimer’s patients do not show significant differences in PAL-task-induced deactivation or in the type of relationship exhibited between deactivation and task difficulty (Gould et al., 2006b). On the one hand, the process of shifting attentional focus from monitoring of the self and the environment to external, goal-directed behavior, which is considered to depend on task-induced deactivation, may therefore be intact in AD. On the other hand, the matching of task performance and difficulty may have obscured a potential effect in this study (Gould et al., 2006b). A different study found greater deactivation in less-impaired MCI patients and loss of deactivation

in more-impaired MCI patients and mild AD subjects in medial and lateral parietal regions (Celone et al., 2006). When task difficulty was not matched, the pattern of brain activity in patients with AD performing an easy version of the task was indistinguishable from that of controls performing a harder version of the task (Gould et al., 2006a). These data support greater recruitment of the same brain regions in the patients as in age-matched controls, as a means of compensating for neuropathology and associated cognitive impairment in AD (Gould et al., 2006a). It remains to be investigated whether this is also the case in MCI patients and whether this approach may prove to be a useful additional diagnostic tool.

Conclusions and perspectives For understanding, assessment, and differential diagnosis of dementia, different methods, including structured clinical interview, neuroimaging, and genetics, may each provide valuable contributions. Therefore, the most accurate diagnosis and treatment recommendations will always be made by combining converging forms of clinical, neuropsychological, and neuroimaging evidence (Perry and Hodges, 1996). Neuropsychological assessment can make an important contribution to the early and differential diagnosis of the dementias and is also vital for the objective monitoring of changes in cognitive function across disease course or as a function of treatment. Combining neuropsychological assessment with brain imaging enables us to identify the underlying pathology. Although structural neuroimaging and genetics, in particular genotyping of apolipoprotein E (ApoE) allele 4, have been combined (for review see Lehtovirta et al., 2000), the combination of functional neuroimaging and genetics is the goal of the next decade (Hodges, 2006). Other key challenges for neuropsychologists over the next decade include decomposing the cognitive and neural substrates of task performance within each form of dementia in order to more accurately characterize the endophenotypes (objective markers that mediate between the direct vulnerability mechanisms of dementia and the clinical diagnosis and symptomatology; Gottesman and Gould, 2003; Hasler et al., 2004) that characterize each disorder. It will be particularly important to examine the relationship between neuropsychological function and specific neuropathological features using novel in vivo neurochemical techniques (Klunk et al., 2004; Shoghi-Jadid et al., 2002). Improved neuropsychological characterization will be translated into more sensitive and specific diagnostic tools. This process in turn should lead to earlier interventions with pharmacological and nonpharmacological therapies. Considering pharmacological therapies, it remains to be investigated whether drugs currently used to treat the symptoms of Alzheimer’s disease could be useful for earlier interventions. It is, for

de rover et al.: cognition and aging: dementia

615

instance, as yet unknown whether these drugs improve performance of MCI patients on the CANTAB PAL task. Nonpharmacological therapies, such as education or neurocognitive activation, may be considered as a way to enhance cognitive reserve (Barnett et al., 2006). Such interventions could potentially be introduced even prior to the onset of disabling cognitive impairments, or they may be used in combination with pharmacological therapies to further improve daily living. acknowledgments

MdR is funded by a Marie Curie IntraEuropean Fellowship within the 6th European Community Framework Program (grant 025062), and ADB and SM are funded by a Wellcome Trust Program Grant (RN 019407). Portions of the work described in this chapter were supported by an MRC LINK grant and carried out within the Behavioural and Clinical Neuroscience Institute. BJS thanks an anonymous donor through Cambridge in America. REFERENCES

Albert, M. S., M. B. Moss, R. Tanzi, and K. Jones, 2001. Preclinical prediction of AD using neuropsychological tests. J. Int. Neurospsychol. Soc. 7:631–639. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington DC: American Psychiatric Association. Artero, S., M. C. Tierney, J. Touchon, and K. Ritchie, 2003. Prediction of transition from cognitive impairment to senile dementia: A prospective longitudinal study. Acta Psychiatr. Scand. 107(5):390–393. Ballard, C. G., J. O’Brien, K. Lowery, G. A. Ayre, R. Harrison, R. Perry, et al., 1998. A prospective study of dementia with Lewy bodies. Age Ageing 27(5):631–636. Barnett, J. H., C. H. Salmond, P. B. Jones, and B. J. Sahakian, 2006. Cognitive reserve in neuropsychiatry. Psychol. Med. 36(8):1053–1064. Bennett, D. A., R. S. Wilson, J. A. Schneider, D. A. Evans, L. A. Beckett, N. T. Aggarwal, et al., 2002. Natural history of mild cognitive impairment in older persons. Neurology 59(2): 198–205. Binder, J. R., J. A. Frost, T. A. Hammeke, P. S. Bellgowan, S. M. Rao, and R. W. Cox, 1999. Conceptual processing during the conscious resting state: A functional MRI study. J. Cogn. Neurosci. 11(1):80–93. Blackwell, A. D., B. J. Sahakian, R. Vesey, J. M. Semple, T. W. Robbins, and J. R. Hodges, 2004. Detecting dementia: Novel neuropsychological markers of preclinical Alzheimer’s disease. Dement. Geriatr. Cogn. Disord. 17(1–2):42– 48. Braak, H., and E. Braak, 1991. Neuropathological staging of Alzheimer-related changes. Acta Neuropathol. (Berl.) 82(4):239– 259. Bradshaw, J. M., M. Saling, V. Anderson, M. Hopwood, and A. Brodtmann, 2006. Higher cortical deficits influence attentional processing in dementia with Lewy bodies relative to patients with dementia of the Alzheimer’s type and controls. J. Neurol. Neurosurg. Psychiatry 77(10):1129–1135.

616

cognition

Breedin, S. D., E. M. Saffran, and H. B. Coslett, 1994. Reversal of the concreteness effect in a patient with semantic dementia. Cogn. Neuropsychol. 11:617–661. Brookmeyer, R., S. Gray, and C. Kawas, 1998. Projections of Alzheimer’s disease in the United States and the public health impact of delaying disease onset. Am. J. Public Health 88(9): 1337–1342. Busse, A., J. Bischkopf, S. G. Riedel-Heller, and M. C. Angermeyer, 2003. Subclassifications for mild cognitive impairment: Prevalence and predictive validity. Psychol. Med. 33(6):1029– 1038. Calderon, J., R. J. Perry, S. W. Erzinclioglu, G. E. Berrios, T. R. Dening, and J. R. Hodges, 2001. Perception attention and working memory are disproportionately impaired in dementia with Lewy bodies compared with Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry 70(2):157–164. Celone, K. A., V. D. Calhoun, B. C. Dickerson, A. Atri, E. F. Chua, S. L. Miller, et al., 2006. Alterations in memory networks in mild cognitive impairment and Alzheimer’s disease: An independent component analysis. J. Neurosci. 26(40):10222– 10231. Celsis, P., 2000. Age-related cognitive decline mild cognitive impairment or preclinical Alzheimer’s disease? Ann. Med. 32(1): 6–14. Collie, A., P. Maruff, and J. Currie, 2002. Behavioural characterization of mild cognitive impairment. J. Clin. Exp. Neuropsychol. 24(6):720–733. Craik, F. I. M., and T. A. Salthouse, 2000. Handbook of Aging and Cognition, 2nd ed. Hillsdale, NJ: Lawrence Erlbaum. de Jager, C. A., E. Milwain, and M. Budge, 2002. Early detection of isolated memory deficits in the elderly: The need for more sensitive neuropsychological tests. Psychol. Med. 32(3):483– 491. Desai, A. K., and G. T. Grossberg, 2005. Diagnosis and treatment of Alzheimer’s disease. Neurology 64(12 Suppl. 3): S34–39. Desmond, D. W., 2004. The neuropsychology of vascular cognitive impairment: Is there a specific cognitive deficit? J. Neurol. Sci. 226(1–2):3–7. Du, A. T., N. Schuff, D. Amend, M. P. Laakso, Y. Y. Hsu, W. J. Jagust, et al., 2001. Magnetic resonance imaging of the entorhinal cortex and hippocampus in mild cognitive impairment and Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry 71(4):441–447. Elias, M. F., A. Beiser, P. A. Wolf, R. Au, R. F. White, and R. B. D’Agostino, 2000. The preclinical phase of Alzheimer disease: A 22-year prospective study of the Framingham cohort. Arch. Neurol. 57(6):808–813. Fowler, K. S., M. M. Saling, E. L. Conway, J. M. Semple, and W. J. Louis, 1995. Computerized delayed matching to sample and paired associate performance in the early detection of dementia. Appl. Neuropsychol. 2(2):72–78. Fowler, K. S., M. M. Saling, E. L. Conway, J. M. Semple, and W. J. Louis, 1997. Computerized neuropsychological tests in the early detection of dementia: Prospective findings. J. Int. Neuropsychol. Soc. 3(2):139–146. Fowler, K. S., M. M. Saling, E. L. Conway, J. M. Semple, and W. J. Louis, 2002. Paired associate performance in the early detection of DAT. J. Int. Neuropsychol. Soc. 8(1):58– 71. Fox, N. C., E. K. Warrington, and M. N. Rossor, 1999. Serial magnetic resonance imaging of cerebral atrophy in preclinical Alzheimer’s disease. Lancet 353(9170):2125.

Fox, N. C., E. K. Warrington, A. L. Seiffer, S. K. Agnew, and M. N. Rossor, 1998. Presymptomatic cognitive deficits in individuals at risk of familial Alzheimer’s disease: A longitudinal prospective study. Brain 121(Part 9):1631–1639. Geerts, H., 2007. (R)-flurbiprofen—An enantiomer of flurbiprofen for the treatment of Alzheimer’s disease. IDrugs 10(2):121– 133. Geldmacher, D. S., G. Provenzano, T. McRae, V. Mastey, and J. R. Ieni, 2003. Donepezil is associated with delayed nursing home placement in patients with Alzheimer’s disease. J. Am. Geriatr. Soc. 51(7):937–944. Gottesman, I. I., and T. D. Gould, 2003. The endophenotype concept in psychiatry: Etymology and strategic intentions. Am. J. Psychiatry 160(4):636–645. Gould, R. L., B. Arroyo, R. G. Brown, A. M. Owen, E. T. Bullmore, and R. J. Howard, 2006a. Brain mechanisms of successful compensation during learning in Alzheimer disease. Neurology 67(6):1011–1017. Gould, R. L., R. G. Brown, A. M. Owen, E. T. Bullmore, and R. J. Howard, 2006b. Task-induced deactivations during successful paired associates learning: An effect of age but not Alzheimer’s disease. NeuroImage 31(2):818–831. Gould, R. L., R. G. Brown, A. M. Owen, E. T. Bullmore, S. C. Williams, and R. J. Howard, 2005. Functional neuroanatomy of successful paired associate learning in Alzheimer’s disease. Am. J. Psychiatry 162(11):2049–2060. Gould, R. L., R. G. Brown, A. M. Owen, D. H. Fytche, and R. J. Howard, 2003. fMRI BOLD response to increasing task difficulty during successful paired associates learning. NeuroImage 20(2):1006–1019. Graham, N. L., T. Emery, and J. R. Hodges, 2004. Distinctive cognitive profiles in Alzheimer’s disease and subcortical vascular dementia. J. Neurol. Neurosurg. Psychiatry 75(1):61–71. Gregory, C. A., and J. R. Hodges, 1996. Clinical features of frontal lobe dementia in comparison to Alzheimer’s disease. J. Neural Transm. Suppl. 47:103–123. Guidi, M., L. Paciaroni, S. Paolini, S. De Padova, and O. Scarpino, 2006. Differences and similarities in the neuropsychological profile of dementia with Lewy bodies and Alzheimer’s disease in the early stage. J. Neurol. Sci. 248(1–2):120–123. Hamalainen, A., M. Pihlajamaki, H. Tanila, T. Hanninen, E. Niskanen, S. Tervo, et al., 2007. Increased fMRI responses during encoding in mild cognitive impairment. Neurobiol. Aging 28(12):1859–1903. Hasler, G., W. C. Drevets, H. K. Manji, and D. S. Charney, 2004. Discovering endophenotypes for major depression. Neuropsychopharmacology 29(10):1765–1781. Hebert, L. E., P. A. Scherr, J. L. Bienias, D. A. Bennett, and D. A. Evans, 2003. Alzheimer disease in the US population: Prevalence estimates using the 2000 census. Arch. Neurol. 60(8): 1119–1122. Helmes, E., J. V. Bowler, H. Merskey, D. G. Munoz, and V. C. Hachinski, 2003. Rates of cognitive decline in Alzheimer’s disease and dementia with Lewy bodies. Dement. Geriatr. Cogn. Disord. 15(2):67–71. Hodges, J., 1998. The amnestic prodrome of Alzheimer’s disease. Brain 121(Part 9):1601–1602. Hodges, J. R., 2006. Alzheimer’s centennial legacy: Origins, landmarks, and the current status of knowledge concerning cognitive aspects. Brain 129(Part 11):2811–2822. Hodges, J. R., K. Patterson, S. Oxbury, and E. Funnell, 1992. Semantic dementia: Progressive fluent aphasia with temporal lobe atrophy. Brain 115(Part 6):1783–1806.

Hodges, J. R., K. Patterson, R. Ward, P. Garrard, T. Bak, R. Perry, et al., 1999. The differentiation of semantic dementia and frontal lobe dementia (temporal and frontal variants of frontotemporal dementia) from early Alzheimer’s disease: A comparative neuropsychological study. Neuropsychology 13(1):31– 40. Holroyd, S., D. G. Snustad, and Z. L. Chalifoux, 1996. Attitudes of older adults on being told the diagnosis of Alzheimer’s disease. J. Am. Geriatr. Soc. 44(4):400–403. Holroyd, S., Q. Turnbull, and A. M. Wolf, 2002. What are patients and their families told about the diagnosis of dementia? Results of a family survey. Int. J. Geriatr. Psychiatry 17(3):218– 221. Hornykiewicz, O., 1963. The tropical localization and content of noradrenalin and dopamine (3-hydroxytyramine) in the substantia nigra of normal persons and patients with Parkinson’s disease. Wien Klin. Wochenschr. 75:309–312. Hutchinson, M., W. Schiffer, S. Joseffer, R. Schlosser, S. Dikshit, E. Goldberg, and J. D. Brodie, 1999. Task-specific deactivation patterns in functional magnetic resonance imaging. Magn. Reson. Imaging 17(10):1427–1436. Jack, C. R., Jr., R. C. Petersen, Y. C. Xu, P. C. O’Brien, G. E. Smith, R. J. Ivnik, et al., 1999. Prediction of AD with MRI-based hippocampal volume in mild cognitive impairment. Neurology 52(7):1397–1403. Johnson, D. K., J. C. Morris, and J. E. Galvin, 2005. Verbal and visuospatial deficits in dementia with Lewy bodies. Neurology 65(8):1232–1238. Jorm, A. F., H. Christensen, A. E. Korten, A. S. Henderson, P. A. Jacomb, and A. Mackinnon, 1997. Do cognitive complaints either predict future cognitive decline or reflect past cognitive decline? A longitudinal study of an elderly community sample. Psychol. Med. 27(1):91–98. Killiany, R. J., B. T. Hyman, T. Gomez-Isla, M. B. Moss, R. Kikinis, F. Jolesz, et al., 2002. MRI measures of entorhinal cortex vs. hippocampus in preclinical AD. Neurology 58(8):1188– 1196. Klunk, W. E., H. Engler, A. Nordberg, Y. Wang, G. Blomqvist, D. P. Holt, et al., 2004. Imaging brain amyloid in Alzheimer’s disease with Pittsburgh Compound-B. Ann. Neurol. 55(3):306–319. Knapp, M., M. Prince, et al., 2007. Dementia UK: A Report to the Alzheimer’s Society on the Prevalence and Economic Cost of Dementia in the UK, Produced by King’s College London and the London School of Economics. London: Alzheimer’s Society. Laakso, M. P., 2002. Structural imaging in cognitive impairment and the dementias: An update. Curr. Opin. Neurol. 15(4):415– 421. Larrieu, S., L. Letenneur, J. M. Orgogozo, C. Fabrigoule, H. Amieva, N. Le Carret, et al., 2002. Incidence and outcome of mild cognitive impairment in a population-based prospective cohort. Neurology 59(10):1594–1599. Lawrence, A. D., and B. J. Sahakian, 1995. Alzheimer disease, attention, and the cholinergic system. Alzheimer Dis. Assoc. Disord. 9(Suppl. 2):43–49. Lee, A. C., S. Rahman, J. R. Hodges, B. J. Sahakian, and K. S. Graham, 2003. Associative and recognition memory for novel objects in dementia: Implications for diagnosis. Eur. J. Neurosci. 18(6):1660–1670. Lehtovirta, M., M. P. Laakso, G. B. Frisoni, and H. Soininen, 2000. How does the apolipoprotein E genotype modulate the brain in aging and in Alzheimer’s disease?

de rover et al.: cognition and aging: dementia

617

A review of neuroimaging studies. Neurobiol. Aging 21(2):293– 300. Lezak, M. D., 1995. Neuropsychological Assessment, 3rd ed. New York: Oxford University Press. Linn, R. T., P. A. Wolf, D. L. Bachman, J. E. Knoefel, J. L. Cobb, A. J. Belanger, et al., 1995. The “preclinical phase” of probable Alzheimer’s disease: A 13-year prospective study of the Framingham cohort. Arch. Neurol. 52(5):485–490. Looi, J. C., and P. S. Sachdev, 1999. Differentiation of vascular dementia from AD on neuropsychological tests. Neurology 53(4): 670–678. Lu, Z. L., J. Neuse, S. Madigan, and B. A. Dosher, 2005. Fast decay of iconic memory in observers with mild cognitive impairments. Proc. Natl. Acad. Sci. USA 102(5):1797–1802. Maguire, E. A., C. D. Frith, N. Burgess, J. G. Donnett, and J. O’Keefe, 1998. Knowing where things are: Parahippocampal involvement in encoding object locations in virtual large-scale space. J. Cogn. Neurosci. 10(1):61–76. Masters, C. L., and K. Beyreuther, 2006. Alzheimer’s centennial legacy: Prospects for rational therapeutic intervention targeting the Abeta amyloid pathway. Brain 129(Part 11):2823– 2839. Masur, D. M., M. Sliwinski, R. B. Lipton, A. D. Blau, and H. A. Crystal, 1994. Neuropsychological prediction of dementia and the absence of dementia in healthy elderly persons. Neurology 44(8):1427–1432. McDonald, R. J., and N. M. White, 1993. A triple dissociation of memory systems: Hippocampus, amygdala, and dorsal striatum. Behav. Neurosci. 107:3–22. McKeith, I. G., C. G. Ballard, R. H. Perry, P. G. Ince, J. T. O’Brien, D. Neill, et al., 2000. Prospective validation of consensus criteria for the diagnosis of dementia with Lewy bodies. Neurology 54(5):1050–1058. McKeith, I. G., A. F. Fairbairn, R. A. Bothwell, P. B. Moore, I. N. Ferrier, P. Thompson, et al., 1994. An evaluation of the predictive validity and inter-rater reliability of clinical diagnostic criteria for senile dementia of Lewy body type. Neurology 44(5):872–877. McKenna, P., and E. K. Warrington, 1980. Testing for nominal dysphasia. J. Neurol. Neurosurg. Psychiatry 43(9):781–788. McKhann, G., D. Drachman, M. Folstein, R. Katzman, D. Price, and E. M. Stadlan, 1984. Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34(7): 939–944. Miyashita, Y., M. Kameyama, I. Hasegawa, and T. Fukushima, 1998. Consolidation of visual associative long-term memory in the temporal cortex of primates. Neurobiol. Learn. Mem. 70(1–2):197–211. Mohs, R. C., D. Knopman, R. C. Petersen, S. H. Ferris, C. Ernesto, M. Grundman, et al., 1997. Development of cognitive instruments for use in clinical trials of antidementia drugs: Additions to the Alzheimer’s Disease Assessment Scale that broaden its scope. The Alzheimer’s Disease Cooperative Study. Alzheimer Dis. Assoc. Disord. 11 (Suppl. 2):S13–21. Morgan, C. D., and L. E. Baade, 1997. Neuropsychological testing and assessment scales for dementia of the Alzheimer’s type. Psychiatr. Clin. North Am. 20(1):25–43. O’Brien, J. T., 2006. Vascular cognitive impairment. Am. J. Geriatr. Psychiatry 14(9):724–733. O’Carroll, R. E., S. Conway, A. Ryman, and N. Prentice, 1997. Performance on the delayed word recall test (DWR) fails to

618

cognition

differentiate clearly between depression and Alzheimer’s disease in the elderly. Psychol. Med. 27(4):967–971. Oosterman, J. M., and E. J. Scherder, 2006. Distinguishing between vascular dementia and Alzheimer’s disease by means of the WAIS: A meta-analysis. J. Clin. Exp. Neuropsychol. 28(7):1158– 1175. Orrell, M. W., and B. J. Sahakian, 1991. Dementia of frontal lobe type. Psychol. Med. 21(3):553–556. Owen, A. M., B. Milner, M. Petrides, and A. Evans, 1996. A specific role for the right parahippocampal region in the retrieval of object-location: A positron emission tomography study. J. Cogn. Neurosci. 8:588–602. Pennanen, C., M. Kivipelto, S. Tuomainen, P. Hartikainen, T. Hanninen, M. P. Laakso, et al., 2004. Hippocampus and entorhinal cortex in mild cognitive impairment and early AD. Neurobiol. Aging 25(3):303–310. Pennanen, C., C. Testa, M. P. Laakso, M. Hallikainen, E. L. Helkala, T. Hanninen, et al., 2005. A voxel-based morphometry study on mild cognitive impairment. J. Neurol. Neurosurg. Psychiatry 76(1):11–14. Perry, R. J., and J. R. Hodges, 1996. Spectrum of memory dysfunction in degenerative disease. Curr. Opin. Neurol. 9(4):281–285. Perry, R. J., and J. R. Hodges, 2000. Differentiating frontal and temporal variant frontotemporal dementia from Alzheimer’s disease. Neurology 54(12):2277–2284. Petersen, R. C., 2003. Mild cognitive impairment clinical trials. Nature Rev. Drug Discov. 2(8):646–653. Petersen, R. C., 2004. Mild cognitive impairment as a diagnostic entity. J. Intern. Med. 256(3):183–194. Petersen, R. C., G. E. Smith, S. C. Waring, R. J. Ivnik, E. G. Tangalos, and E. Kokmen, 1999. Mild cognitive impairment: Clinical characterization and outcome. Arch. Neurol. 56(3): 303–308. Rahman, S., T. W. Robbins, and B. J. Sahakian, 1999. Comparative cognitive neuropsychological studies of frontal lobe function: Implications for therapeutic strategies in frontal variant frontotemporal dementia. Dement. Geriatr. Cogn. Disord. 10(Suppl. 1):15–28. Rahman, S., B. J. Sahakian, J. R. Hodges, R. D. Rogers, and T. W. Robbins, 1999. Specific cognitive deficits in mild frontal variant frontotemporal dementia. Brain 122(Part 8):1469– 1493. Ratnavalli, E., C. Brayne, K. Dawson, and J. R. Hodges, 2002. The prevalence of frontotemporal dementia. Neurology 58(11): 1615–1621. Raven, J. C., 1965. RCPM: Guide to Using the Colored Progressive Matrices. New York: Psychological Corporation. Rey, A., 1941. L’examen psychologique dans les cas d’encephalopathie traumatique [The psychological examination of traumatic encephalopathy]. Archives de Psychologie 28:268–340. Ritchie, K., S. Artero, and J. Touchon, 2001. Classification criteria for mild cognitive impairment: A population-based validation study. Neurology 56(1):37–42. Rogers, T. T., A. Ivanoiu, K. Patterson, and J. R. Hodges, 2006. Semantic memory in Alzheimer’s disease and the frontotemporal dementias: A longitudinal study of 236 patients. Neuropsychology 20(3):319–335. Rosen, W. G., R. C. Mohs, and K. L. Davis, 1984. A new rating scale for Alzheimer’s disease. Am. J. Psychiatry 141(11):1356– 1364. Saffran, F. M., and M. F. Schwartz, 1994. Of cabbages and things: Semantic memory from a neuropsychological perspective—A tutorial review. In C. Umita and M. Moscovitch, eds.,

Attention and Performance XV, 507–535. Cambridge, MA: MIT Press. Sahakian, B., G. Jones, R. Levy, J. Gray, and D. Warburton, 1989. The effects of nicotine on attention information processing and short-term memory in patients with dementia of the Alzheimer type. Br. J. Psychiatry 154:797–800. Sahakian, B. J., R. G. Morris, J. L. Evenden, A. Heald, R. Levy, M. Philpot, et al., 1988. A comparative study of visuospatial memory and learning in Alzheimer-type dementia and Parkinson’s disease. Brain 111(Part 3):695–718. Sahakian, B. J., A. M. Owen, N. J. Morant, S. A. Eagger, S. Boddington, L. Crayton, et al., 1993. Further analysis of the cognitive effects of tetrahydroaminoacridine (THA) in Alzheimer’s disease: Assessment of attentional and mnemonic function using CANTAB. Psychopharmacology (Berl.) 110(4):395– 401. Sahgal, A., P. H. Galloway, I. G. McKeith, S. Lloyd, J. H. Cook, I. N. Ferrier, et al., 1992. Matching-to-sample deficits in patients with senile dementias of the Alzheimer and Lewy body types. Arch. Neurol. 49(10):1043–1046. Sahgal, A., I. G. McKeith, P. M. Galloway, N. Tasker, and T. Stechler, 1995. Do differences in visuospatial ability between senile dementias of the Alzheimer and Lewy body type reflect differences solely in mnemonic function? J. Clin. Exp. Neuropsychol. 17(1):35–43. Shimomura, T., E. Mori, H. Yamashita, T. Imamura, N. Hirono, M. Hashimoto, et al., 1998. Cognitive loss in dementia with Lewy bodies and Alzheimer disease. Arch. Neurol. 55(12):1547– 1552. Shoghi-Jadid, K., G. W. Small, E. D. Agdeppa, V. Kepe, L. M. Ercoli, P. Siddarth, et al., 2002. Localization of neurofibrillary tangles and beta-amyloid plaques in the brains of living patients with Alzheimer disease. Am. J. Geriatr. Psychiatry 10(1): 24–35. Smith, M. L., and B. Milner, 1981. The role of the right hippocampus in the recall of spatial location. Neuropsychologia 19(6):781–793. Snowden, J. S., P. J. Goulding, and D. Neary, 1989. Semantic dementia: A form of circumscribed cerebral atrophy. Behav. Neurol. 2:167–182.

Stavitsky, K., A. M. Brickman, N. Scarmeas, R. L. Torgan, M. X. Tang, M. Albert, et al., 2006. The progression of cognition, psychiatric symptoms, and functional abilities in dementia with Lewy bodies and Alzheimer disease. Arch. Neurol. 63(10): 1450–1456. Stephens, S., R. A. Kenny, E. Rowan, L. Allan, R. N. Kalaria, M. Bradbury, et al., 2004. Neuropsychological characteristics of mild vascular cognitive impairment and dementia after stroke. Int. J. Geriatr. Psychiatry 19(11):1053–1057. Swainson, R., J. R. Hodges, C. J. Galton, J. Semple, A. Michael, B. D. Dunn, et al., 2001. Early detection and differential diagnosis of Alzheimer’s disease and depression with neuropsychological tasks. Dement. Geriatr. Cogn. Disord. 12(4):265– 280. Taffe, M. A., M. R. Weed, T. Gutierrez, S. A. Davis, and L. H. Gold, 2002. Differential muscarinic and NMDA contributions to visuo-spatial paired-associate learning in rhesus monkeys. Psychopharmacology (Berl.) 160(3) 253–262. Taffe, M. A., M. R. Weed, T. Gutierrez, S. A. Davis, and L. H. Gold, 2004. Modelling a task that is sensitive to dementia of the Alzheimer’s type: Individual differences in acquisition of a visuo-spatial paired-associate learning task in rhesus monkeys. Behav. Brain Res. 149(2):123–133. Vellas, B., S. Andrieu, C. Sampaio, and G. Wilcock, European Task Force group, 2007. Disease-modifying trials in Alzheimer’s disease: A European Task Force consensus. Lancet Neurol. 6(1) 56–62. Visser, P. J., F. R. Verhey, P. A. Hofman, P. Scheltens, and J. Jolles, 2002. Medial temporal lobe atrophy predicts Alzheimer’s disease in patients with minor cognitive impairment. J. Neurol. Neurosurg. Psychiatry 72(4):491–497. Wechsler, D., 1981. Wechsler Adult Intelligence Scale–Revised. San Antonio TX: Psychological Corporation. Wechsler, D., 1989. Wechsler Memory Scale–Revised. San Antonio TX: Psychological Corporation. Welsh, K. A., N. Butters, J. P. Hughes, R. C. Mohs, and A. Heyman, 1992. Detection and staging of dementia in Alzheimer’s disease: Use of the neuropsychological measures developed for the Consortium to Establish a Registry for Alzheimer’s Disease. Arch. Neurol. 49(5):448–452.

de rover et al.: cognition and aging: dementia

619

V NEURODEVELOPMENTAL ASPECTS OF CLINICAL DISORDERS

38

The Role of Nutrition in Cognitive Development ANITA J. FUGLESTAD, RAGHAVENDRA RAO, AND MICHAEL K. GEORGIEFF

Nutrient categories Macronutrients Protein, fat, and carbohydrates are considered the three macronutrients. Protein is typically used by the body for somatic (tissue) protein and serum protein synthesis. Proteins also include all enzymes found within organs. Energy in the form of fat or carbohydrates is the metabolic fuel on which the body depends. The body preferentially uses carbohydrates (typically glucose) as the substrate for cellular metabolism to generate adenosine triphosphate (ATP). Neuronal metabolism is particularly dependent on glucose availability and is very sensitive to periods of carbohydrate deprivation (Volpe, 1995). The body maintains a small storage pool of carbohydrates in the form of glycogen that supplies glucose equivalents for a short period of time when glucose supply is limited. Hypoglycemia (low blood sugar concentration) has a particularly profound negative effect on the developing hippocampus (Kim et al., 2005; Yamada et al., 2004). Fat, with its higher caloric value, provides energy for storage (as adipose tissue), but can also be utilized more slowly than glucose to provide energy to the entire body, including the brain (e.g., as ketone bodies). Certain fats and lipoproteins are important for normal neuronal cell membrane integrity and myelination. For example, cholesterol, phosphatidyl choline, and certain fatty acids are essential for cell membrane synthesis and integrity. Linolenic acid, linoleic acid, arachidonic acid, and docosohexaenoic acid are essential for normal brain membrane formation and myelination. During starvation or periods of illness, protein can also be used as an energy substrate. When protein is utilized in such a manner, it is not available for structural tissue (including brain) synthesis. There is an extensive literature in both humans and animal models on the isolated and combined adverse effects of protein and energy malnutrition on the developing brain (Pollitt, 1996; Pollitt and Gorman, 1994; Pollitt, Watkins, and Husaini, 1997; Pollitt et al., 1993, 1995; Winick and Nobel, 1966; Winick and Rosso, 1969a, 1969b). Pollitt and Gorman (1994) have pointed out that other nutrient deficiencies usually coexist with protein-energy malnutrition (PEM) in free-living populations.

Micronutrients Minerals, trace elements, and vitamins are collectively grouped as micronutrients. Minerals. The major minerals of the body are sodium, potassium (with their usual accompanying dietary anion, chloride), calcium, and phosphorus. Minerals are not classically considered essential for brain development, but deficiencies in these nutrients will lead to abnormal brain function, mostly through altering neuronal electrical function. Trace elements. This class of nutrients contains elements that are required in trace quantities by the body and are used, for the most part, in intermediary cellular metabolism. Members of this category include magnesium, manganese, iodine, zinc, copper, molybdenum, cobalt, selenium, fluoride, and iron. As with the major minerals, these elements are not classically considered to be uniquely important for normal brain development except as their deficiencies affect cellular (including neuronal) function. Some elements in this group, however, are exceptional in their particularly profound effect on cognitive development: iron, iodine, and zinc. Iron is required for enzymes that regulate central nervous system cell division (ribonucleotide reductase), monoamine synthesis (e.g., tyrosine hydroxylase), myelination (delta-9 desaturase), and oxidative metabolism (cytochromes). The effects of iron deficiency on the growing and mature brain are well documented. Iodine is essential for normal thyroid hormone synthesis. Brain development is severely compromised by hypothyroidism (cretinism) with particularly profound cognitive effects (Hetzel and Mano, 1989; Kretchmer, Beard, and Carlson, 1996). Similarly, zinc, through its role in nucleic acid synthesis, plays a significant role in neurodevelopment. Both neuroanatomic and neurochemical changes have been described in zinc deficiency. Other than the extensive literature on these trace elements, little developmental work has been performed to assess the roles of the remaining micronutrients on cognitive development. Vitamins. Vitamins are categorized as water- and fatsoluble. They generally are cofactors in intermediary metabolism, although some, like vitamin A, bind promoter regions

623

of genes that regulate cell differentiation and neuronal growth (Mangelsdorf, 1994). As with the trace elements, vitamin deficiencies can potentially affect total body metabolism and consequently brain growth and development. Nevertheless, certain vitamins (e.g., vitamin A, folic acid, pyridoxine [B6]) appear to be more critical during certain periods of central nervous system (CNS) development, and their deficiencies present a greater risk to neurodevelopment (for review see Pollitt, 1996). Folic acid deficiency during early pregnancy has been closely linked, both epidemiologically and in animal models, to neural tube defects such as meningomyelocele and encephalocele (Copp and Bernfield, 1994). Vitamin A deficiency is an important neuroteratogenic risk factor during the periconceptional period, but it is also associated with retinal and neuronal degeneration in the postnatal period. Pyridoxine is critical for N-methyl d-aspartate (NMDA) receptor synthesis and function (Guilarte, 1993).

The role of nutrition within the context of cognitive development It is critical to appreciate that insufficient nutrition does not determine poor cognitive outcome. Likewise, adequate nutrient delivery alone does not ensure normal brain growth and development. Although adequate nutrition is essential for normal development, the role of nutrition in cognitive development must be considered with regard to other biological and environmental factors (figure 38.1). The effect of

Growth factors Inborn Errors of Metabolism

Genetic Factors

Brain microarchitecture

Nutrients

Developing Brain

MaternalInfant bonding

Environmental Enrichment Figure 38.1 The role of nutrition in the developing brain. Genetic factors, nutrients, and environmental factors play major roles in neurodevelopment. In addition to their direct involvements (thin arrows) in neurological processes such as myelination, cell proliferation, synaptogenesis, and neurotransmission, these factors interact with each other (thick arrows) in shaping the developing brain. Curved arrows indicate interaction within a major group (e.g., nutrient-nutrient interaction).

624

poor nutrition on development may be influenced by environmental factors. For instance, cognitive effects (of nutritional deficiencies as measured by the mental development index of Bayley Scales of Infant Development [BSID]) are more severe for children living in homes where there is less stimulation compared to homes with higher levels of stimulation (Grantham-McGregor et al., 1998). Just as nutrient deficiency does not determine cognitive development, simply providing adequate nutrition may not support normal brain growth and development. Along with adequate nutrition, inborn errors of metabolism, growth factors, and stress need to be considered. An example of an inborn error of metabolism is phenylketonuria (PKU). Individuals with PKU lack the enzyme phenylalanine hydroxylase, which converts phenylalanine to tyrosine, and therefore phenylalanine accumulates. If this condition is not treated by limiting phenylalanine in the diet (particularly early in life), brain development is altered, and individuals develop mental retardation. Growth Factors It is important to recognize that nutritional status is not determined simply by nutrient availability. Once nutrients have been ingested, they must be absorbed and translated into useful metabolic products that promote growth and development. This process occurs through the action of growth factors and anabolic hormones, which translate potential nutritional value into tissue synthesis and function. Nutrients for the body are analogous to fuel for a car. Growth factors act as the transmission that “puts the car into drive” and makes it progress forward. Growth factors (and therefore nutrient handling) are profoundly affected by the organism’s physiologic state. For example, starvation has a different effect than illness on somatic and cerebral metabolism even though both are characterized by low nutrient intakes. Typically, during starvation the body lowers its metabolic set point, thereby requiring fewer calories for maintenance of vital functions. Although insulin levels are low, counterregulatory hormone concentrations that promote tissue breakdown (cortisol and glucagon) are not typically elevated. Provision of protein and energy at a level slightly higher than that needed for maintenance will result in some, albeit suboptimal, growth. In the young child, the fascinating phenomenon of “head sparing” occurs, where somatic growth will suffer at the expense of brain growth during periods of marginal malnutrition. The mechanism of this regional growth effect is unknown. In contrast, illness activates cortisol and glucagon secretion to provide a ready source of glucose, and the body is relatively insulin resistant. Thus brain growth is severely impaired during illness, whereas it is spared during simple starvation. The brain is dependent on a host of growth factors for normal neurogenesis, synaptogenesis, dendritic arborization,

neurodevelopmental aspects of clinical disorders

and myelination. A complete review of these factors and their functions is beyond the scope of this chapter (for review see Pleasure and Pleasure, 2003). By “growth factors” in this context, we mean small proteins that enhance proliferation of target cells either by encouraging cell division or preventing cell death. The brain contains growth factors that are found throughout the body, including insulin-like growth factor–I (IGF-I), epidermal growth factor (EGF), and fibroblast growth factor (FGF), as well as some that are specific to brain (e.g., brainderived neurotrophic factor [BDNF] or glial growth factor [GGF]). The expression and function of these growth factors are influenced by nutritional status. Malnutrition during fetal life (i.e., intrauterine growth restriction) down-regulates IGF-I and IGF-I binding protein expression (Nishijima, 1986). The growth factor IGF-I has a mitogenic and posttranscriptional effect on oligodendrocytes, stimulates neurite outgrowth, and promotes neuronal differentiation (Fellows, 1987). Reduction in IGF-I levels can thus influence myelin production (McMorris and Dubois-Dalcq, 1988; Saneto et al., 1988), as well as neuronal number and complexity. The reduction in IGF-I concentrations noted in growthrestricted fetuses may account for the high rate of microcephaly seen in this population. A study of transgenic mice underscores the importance of the interaction between IGF-I and malnutrition as they influence brain growth (Lee et al., 1999). Whole brain and regional brain growth were assessed in well-nourished and malnourished suckling transgenic mice that overexpress IGF-I. Transgenic overexpression of IGF-I in well-nourished mice increased whole brain weight by 20 percent over well-nourished controls, predominantly because of increased myelination, a modest increase in DNA content, and increased neuronal survival resulting from decreased apoptosis. Whereas generalized malnutrition reduced brain weight by 10 percent in control mice, the IGF-Ioverexpressing mice had brain weights comparable to wellnourished controls, implying sparing of undernutrition effects by the IGF-I. More importantly, the malnutrition in the control group and the ameliorating effect of IGF-I in the transgenic group were regional. The hippocampus and cerebellum were more affected by malnutrition than the cortex, diencephalon, or brain stem. The brain-sparing effects of IGF-I overexpression were more prominent in the undernourished hippocampus, cortex, and diencephalon, but were not seen in the cerebellum. This regionalization of brain effects from malnutrition may be related to the metabolic demand of the areas at this time of development as well as variations in the expression of IGF-I. A similar effect in terms of iron accretion has been described in young rats (Erikson et al., 1997; deUngria et al., 2000).

Selected nutrients and their effect on brain development Table 38.1 categorizes the components of human nutrition and their likely relationships to central nervous system development and function. All nutrients are needed for normal somatic development, but some play a greater role in neurodevelopment than others. The following subsections present the evidence for the nutrients that are most important to the developing cognitive systems of the brain. Protein-Energy Status The effect of protein-energy nutritional status on brain growth and neurodevelopment is one of the most extensively studied subjects in nutrition. The importance of providing protein and energy to developing and mature brains has been assessed almost exclusively through studies of protein-energy insufficiency (as opposed to evaluating whether there is a beneficial effect of supplementing a replete organism). Even large-scale epidemiological studies demonstrating beneficial effects of protein and energy supplementation have been conducted in populations where protein-energy malnutrition (PEM) is endemic (Gorman, 1995). Pollitt and Gorman (1994) suggest that chronic energy rather than protein deficiency is responsible for most of the neurobehavioral changes associated with PEM. Animal models, however, strongly implicate independent roles for protein and energy in brain structural changes (Bass, Netsky, and Young, 1970; Cragg, 1972; Jones and Dyson, 1981; Wiggins, Fuller, and Enna, 1984). Protein-energy malnutrition can occur throughout the life span from fetal life through adulthood. The likely neuropathology underlying the significant changes in brain development or function seen with PEM has been elucidated using animal models. Effects of prenatal PEM. As with most nutritional deficiencies, the most significant neurodevelopmental effects appear to occur when severe PEM is imposed on a rapidly growing brain. The brain grows most rapidly during fetal and early postnatal life. Restriction of macronutrients during fetal life results in deceleration of growth and bears the term “intrauterine growth restriction” (IUGR). Maternal hypertension during gestation accounts for 75 percent of all cases of IUGR, although maternal malnutrition, intrauterine infections, and chromosomal anomalies can alter fetal growth as well (Low and Galbraith, 1974). In the latter two instances, the adverse neurodevelopmental outcome associated with small brain size is not due to PEM but to the underlying pathology of the precipitating condition (e.g., viral invasion of CNS cells). In the case of maternal hypertension, high blood pressure causes atheromatous (plaquelike) changes in the vessels of the placenta, which in turn restricts blood and

fuglestad, rao, and georgieff: nutrition in cognitive development

625

Table 38.1 Selected nutrient effects on cognitive development at the (1) molecular, (2) biochemical, (3) structural, and (4) behavioral levels a Nutrient Effect of Deficit Targeted Brain Region Protein-energy Whole brain (1) ↓ Total DNA, RNA, protein content Cerebellar cortex (2) ↓ mRNA for neuronal, glial proteins Cerebral cortex ↓ Neurotransmitter production Hippocampus ↓ Altered fatty acid profile ↓ Growth factor synthesis (3) ↓ Synapse number ↓ Myelin ↓ Brain weight (4) ↓ IQ ↓ Bender-Gestalt scores Weak visual recognition memory Weak verbal ability/reduced vocabulary Decreased speed of processing Iron

(1) ↓ Brain DNA, RNA protein content ↓ Ribonucleotide reductase activity (2) ↓ Tyrosine hydroxylase activity ↓ Cytochrome c and c oxidase activity ↓ Delta-9 desaturase activity (3) ↓ Dopamine activity ↓ Neuronal oxidative metabolism ↓ Myelination (4) ↓ Bayley Mental Development Index ↓ Bayley Psychomotor Development Index ↓ Spontaneous movement Delayed latency on evoked responses ↓ Spatial working memory ↓ Memory and learning

Variable, based on age of insult (see text for details) Hippocampus targeted at young ages

Zinc

(1) ↓ DNA, RNA, and protein content ↓ Cell replication (2) ↓ IGF-I activity ↓ Synaptic Zn release Altered neurotransmitter receptor binding (3) Truncated dendritic arborization Reduced regional brain mass ↓ Inhibition of GABA ↓ Binding to receptors (4) ↓ Spontaneous motor activity ↓ Short-term visual memory ↓ Concept formation and abstract reasoning

Cerebellum, limbic system, cerebral cortex Many effects are neurochemical/ neurophysiological, given their reversibility with treatment

Iodineb

(1) ↓ Brain DNA, stable protein:DNA ratio ↓ Membrane signaling proteins ↓ mRNA for microtubule proteins ↓ Binding of gene promoter regions for stem cell differentiation (2) Abnormal fatty acid synthesis ↓ Axonal and dendritic microtubule protein ↓ Neuronal oxidative metabolism (3) ↓ Brain weight ↓ Dendritic arborization Migration defects ↓ Neuropil Hypomyelination (4) ↓ Verbal IQ ↓ Subset coding ability (WISC-R) Motor impairment (reaction time) Spastic diplegia Mental retardation

626

neurodevelopmental aspects of clinical disorders

Table 38.1—Continued Effect of Deficit

Nutrient c

Selenium

(1) Down-regulation of myelin genes in oligodendrocytes (2) Biochemical findings of hypothyroidism (see iodine deficiency) (3) Increased dopamine turnover Hypomyelination (↓ myelin basic protein) (4) ↓ Thermoregulation Impaired motor ability

Targeted Brain Region All areas, but primarily cerebellar cortex, prefrontal cortex, and hippocampus

Hippocampus (1) ↑ Apoptosis ↓ Transmethylation reactions (2) ↓ Decreased acetylcholine concentrations (4) Impaired memory and attention a The model portrayed is based on integrated human and animal behavioral, histological, and cell-culture data. b Iodine’s effect on the CNS is strictly through its role in thyroid hormone, not elemental iodine deficiency. c Selenium’s primary effects are through interaction with iodine and thyroid status.

Choline

nutrient flow to a smaller, more calcified placenta (DeWolf, Robertson, and Brosen, 1975). Evidence of PEM resides in the smaller weight, length, and fat and muscle content of the growth-restricted fetus and newborn. Brain growth is compromised if the hypertensive insult is early (second trimester), prolonged, or severe. The neurodevelopmental effects of restricted fetal nutritional delivery and delayed brain growth have been a subject of extensive clinical and laboratory investigation. Most larger trials measured outcomes that are relatively nonspecific neurologically, such as IQ (Strauss and Dietz, 1998). Smaller studies have looked more specifically at what areas of the brain may be preferentially affected by fetal PEM. These studies demonstrate a fivefold higher prevalence of mild neurodevelopmental abnormalities at 2 years of age (Spinillo et al., 1993), weak novelty preference on visual recognitionmemory tasks (Gottleib, Biasini, and Bray, 1988), and reduced verbal ability (Pollitt and Gorman, 1994) in the IUGR infants compared with appropriate-sized controls. It is important to recognize that these effects may be due to PEM alone (based on animal models), but may also be due to other micronutrient deficiencies (iron, zinc, selenium) or chronic intrauterine hypoxia. A significant postnatal confounder has been the consistent finding that fetal growth restriction occurs more frequently in women of lower socioeconomic status who receive less prenatal care and in whom there is an increased incidence of smoking (Strauss and Dietz, 1998). Epidemiological studies show that poor prenatal head growth presages poor developmental outcome (Gottleib, Biasini, and Bray, 1988; Harvey et al., 1982; Low et al., 1982; Strauss and Dietz, 1998; Winer and Tejani, 1994). Employing a clever study design, Strauss and Dietz (1998) matched term IUGR infants to sibling controls who were of appropriate weight for term gestation, thus controlling for genetic and postnatal environment; they showed a signifi-

cant decrement in broad measures such as Wechsler Intelligence Scale for Children IQ and Bender-Gestalt scores at 7 years only if head growth had been compromised in the fetal period, but found no effect if intrauterine head growth had been spared. The biochemical and neuroanatomic bases of neurodevelopmental impairment from early PEM can be found in human autopsy studies and in investigations using animal models. The human studies show significant reductions in brain DNA, RNA, and protein content (Winick and Nobel, 1966; Winick and Rosso, 1969a, 1969b). Infants with IUGR infants have lower brain cell number, smaller cell size, and smaller head circumferences. Certain areas (e.g., the cerebellum, cerebral cortex, hippocampus) demonstrate more profound effects than others, suggesting that the developing brain in some way prioritizes protein and energy during deficiency states. Animal models of IUGR support the human findings of lighter brain weights and reduced neuronal DNA and RNA content, as well as reductions in mRNA for neuronal and glial structural proteins, synapse number, synaptic structures, and neurotransmitter peptide production (Bass, Netsky, and Young, 1970; Cragg, 1972; Jones and Dyson, 1981; Wiggins, Fuller and Enna, 1984). Fatty acid profiles are profoundly altered, with subsequent reductions in myelination, brain lipid composition, and learning ability in fat-restricted rats (Yamamoto et al., 1987). Malnutrition also down-regulates CNS growth factors critical for normal brain development (Nishijima, 1986). The study by Lee and colleagues (1999) in transgenic mice emphasizes the importance of maintaining IGF-I levels during periods of malnutrition in order to spare regions of the brain important for cognition. It will be important in future research in IUGR infants to link specific regional neuropathologic or neurochemical findings (elucidated from controlled animal models of PEM) with deficits in behaviors known to be based in those regions.

fuglestad, rao, and georgieff: nutrition in cognitive development

627

Thus one should be able to relate the reduction in visual recognition-memory processing in IUGR infants (Gottleib, Biasini, and Bray, 1988) to either reduced hippocampal volume or metabolism on MRI/NMR spectroscopy. Several studies have begun to examine such relationships between brain and behavior in IUGR infants. For instance, IUGR infants, when compared with matched-for-gestational-age controls, had reduced cerebral cortical gray matter, and this reduction in gray matter was correlated with attention interaction capacity (as measured by the Assessment of Preterm Infants’ Behavior) (Tolsa et al., 2004). Another recent study (Black et al., 2004) assessed electrophysiological correlates of recognition-memory processing using event-related potentials (ERP); IUGR newborns, compared with matched-forage controls, had ERP patterns indicative of accelerated maturation of auditory recognition memory. Continuation of such lines of research will be important to clarify the relationship between specific aspects of brain development and cognition in IUGR infants. Effects of postnatal PEM. The postnatal brain grows rapidly during the first six postnatal months and then slows considerably during the last six months of the first year. Important regional changes that occur during this year include myelination of motor tract fibers, rapid cerebellar development, and establishment of hippocampal-prefrontal connections. Similar to intrauterine malnutrition, extrauterine malnutrition bodes poorly for the developing brain. Preterm infants or those with severe illnesses during the neonatal period are particularly vulnerable. The protein and energy needs of these infants are high, and it is often difficult to meet the nutritional requirements of these infants because of their illnesses. Preterm infants whose head circumference (which reflects brain growth during typical development) fails to catch up to the gestation-specific norm within 4 weeks have suboptimal head growth and lower developmental quotient at 12 months corrected age (Georgieff et al., 1985). In preterm infants with IUGR, these adverse effects are seen after only 2 weeks of poor postnatal head growth. Similarly, a recent prospective study found that IUGR infants had lower IQs and performed worse on measures of frontal lobe functioning at age 9 years compared with carefully matched controls (Geva et al., 2006). These findings were attenuated by postnatal catch-up growth, such that IUGR children with incomplete catch-up growth had more cognitive difficulties than those with complete catch-up growth. Once infants are weaned from human milk or formula, their risk of PEM increases because they are now dependent on the same foods as their adult food providers. A child growing up in an area of the world where PEM is endemic will be placed at a risk equivalent to or greater than the adult population of that area. The effects of PEM have been extensively studied in children from age 6 months through

628

the early teenage years in these endemic areas. Although the studies show mixed results owing to differences in degree, timing, type, and duration of malnutrition, it is safe to state that PEM of sufficient degree and duration at a critical time during a growing child’s life will affect cognitive abilities. In some cases, the negative effect on the brain is reversible with nutritional rehabilitation, while in other cases the effects appear more “permanent.” However, as we will discuss, the assignment of nutritional causality to poorer cognitive outcomes in free-living societies has been very difficult, given concurrent confounding variables such as maternal mental health, maternal socioeconomic status, lack of infant mental stimulation, and malnutrition-induced lack of infant motivation (Stein and Susser, 1985). Until the mid-1990s, studies assessing the role of postnatal malnutrition in cognitive development were dominated by two important theoretical concepts. The first centered on the idea of “critical periods.” Winick and Noble’s histopathologic studies introduced the idea that PEM during the period of rapid brain growth in the perinatal period would cause more significant and hence more permanent perturbations in brain structure and, ultimately, function (Winick and Nobel, 1966). The vulnerable period was subsequently extended until at least 3 years of age as it became apparent that synaptogenesis, dendritic pruning, programmed cell death, and myelinogenesis continued well into the postnatal period. Dobbings’ work codified the concept of vulnerable or critical periods in the field of nutritional neuroscience (Dobbing, 1990). Based on these developmental neuroanatomic considerations, it was logical to expect that field studies of PEM in 6-month-old to 3-year-old children should demonstrate significant cognitive effects. The other important theoretical concept centered on the role of covariates such as maternal mental status, maternalchild interactions, and infant motivation in determining the cognitive abilities of the malnourished infant (Lozoff et al., 1998; Pollitt and Gorman, 1994). Theoretically, these covariates could be highly dependent on either maternal or infant nutritional status, or they could be completely separate issues. Another example concerns the effects of individual differences in infant motivation. An infant could perform poorly on a set of cognitive measures as a function of what is perceived by the investigator as “reduced motivation.” Reduced motivation could be due to depression, decreased energy due to marginal malnutrition, or poor maternal-child stimulation based on the mother’s (and not the infant’s) nutritional status. Lozoff has published a diagram of these complex interactions with respect to iron deficiency (Lozoff et al., 1998; Lozoff and Georgieff, 2006), but the tenets are clearly applicable to PEM as well. Finally, some of the impairments in developmental tests could be due to the effect of malnutrition on the musculoskeletal system. As an example, preterm infants with bronchopulmonary dysplasia,

neurodevelopmental aspects of clinical disorders

a condition that is associated with malnutrition (deRegnier et al., 1996) perform poorly on developmental tests, such as the psychomotor development index component of BSID during infancy and toddlerhood (Raman, Georgieff, and Rao, 2006). With nutritional rehabilitation and improvement in physical growth, these impairments also tend to improve. Prior to the mid-1990s, studies in humans that attempted to associate mild to moderate PEM with poorer cognitive outcome were so hopelessly flawed that researchers in the field would reach diametrically opposite conclusions about the studies (see Stein and Susser, 1985). It may not be ethically possible to separate all of the nonnutritional confounding variables in a free-living human society, but studies published by Pollitt’s group in the mid-1990s have given credence to the hypothesis that nutritional factors (specifically, PEM) do influence postnatal cognitive development (Pollitt and Gorman, 1994; Pollitt et al., 1995; Pollitt, Watkins, and Husaini, 1997). In these studies, long-term cognitive effects of early supplemental feeding in children from birth to 7 years of age were examined in four Guatemalan villages from 1969 to 1977. Seventy percent of the children were examined again in 1988 at ages 11–26 years. The villages were in an area marked by mild to moderate PEM. Participants in two of the villages received a high-calorie/high-protein supplement that was administered to the mothers, infants, and young children. Participants in the other two villages received a supplement with 40 percent fewer calories and 20 percent less protein. Test scores on knowledge, numeracy, reading, processing time, and vocabulary at follow-up (adolescence) were significantly higher in the group receiving more robust protein-energy supplementation. Within that group, no effect of socioeconomic status was observed. Interestingly, the group that received less supplementation not only had lower scores on cognitive tasks, but demonstrated a profound socioeconomic effect as well. This result suggests that marginal nutritional status may potentiate the effects of other risk factors on neurodevelopment. A similar effect has been described in animal models of iron deficiency (Rao et al., 1999). Pollitt and colleagues (1995) very carefully considered potential alternative explanations—including differences between the villages, intervening factors between the initial intervention and the follow-up 10 years later, compliance with the dietary supplements, and the 30 percent dropout rate at follow-up—and concluded that none accounted for the observed differences in cognition. They do not claim that the study provided definitive proof of the effect of early mild PEM on later cognitive ability, but felt that the evidence is compelling enough to justify nutritional intervention as sound public health policy (Pollitt and Gorman, 1994). From a mechanistic standpoint, it would be most useful for researchers studying postnatal PEM to assess whether the

neurodevelopmental abnormalities that are observed fit the expected pattern of vulnerable brain areas and processes predicted from human and animal histopathologic, neurochemical, or neurophysiologic models. Although it is laudable to demonstrate effects of mild PEM on relatively broad cognitive function, it would be interesting to know whether neurologic circuits involved in the behaviors where differences are observed between groups are particularly at risk specifically for protein or energy malnutrition. For example, poorer performance on broad cognitive tasks might be due to nutritionally induced hypomyelination. Thus reduced speed of processing would explain the effects of PEM on cognitive function. Alternatively, the same finding at the behavioral level could be due to reduced motivation to perform the task and may be related to damage or dysfunction of limbic structures such as the amygdala. If no data support amygdaloid vulnerability to PEM, the latter may be more effectively ruled out. Newer tools that assess the structure and functions underlying certain cognitive capabilities (e.g., f MRI, event-related potential) now allow for more precise delineation of cause and effect of nutritional deficiencies such as early PEM. Effects of specific macronutrients. Infants in the United States and Canada are typically fed either human milk or infant formula derived from cow’s milk or soy plant products during the postnatal period. The significant, positive relationship between breast-feeding and cognitive development is intriguing, since it is unclear whether the positive effect is related to nutritional factor(s) present in human milk but not in formula, to positive maternal-infant interactions (including the propensity of higher-IQ mothers to choose breastfeeding), or both. There is good reason to believe that multiple factors found in breast milk promote normal CNS development and that deficiencies of these nutrients in cow’s milk or soy-based formulas are responsible for slower rates of cognitive development (Lucas, 1997; Morrow-Tlucak, Haude, and Ernhart, 1988; Wang and Wu, 1996). These factors include compounds such as nucleotides (DeLucchi, Pita, and Faus, 1987), oligosaccharides, and long-chain polyunsaturated fatty acids (LCPUFAs) (Innis, 1992) that are simply not synthesized by cows and soybean plants. Because of their potential neurologic effects, within the past few years, LCPUFAs have been added to some cow’s milk and soy-based formulas. Other neurotrophic compounds (e.g., growth factors) have not been added because they are destroyed in formula processing or storage (MacLean and Benson, 1989). The oligosaccharides and lactose that are present in human milk may play a significant role in brain development. Oligosaccharides are present only in the milk of humans and elephants, both species in whom the central nervous system develops predominantly after birth

fuglestad, rao, and georgieff: nutrition in cognitive development

629

(Kunz et al., 2000). Galactose- and sialic acid–containing oligosaccharides appear be important for myelination, synaptogenesis, and learning (Kunz et al., 2000; Wang and Brand-Miller, 2003; Wang et al., 2003). Oral and parenteral supplementation of sialic acid results in greater accumulation of sialic acid–containing glycolipid and glycoprotein in the cerebrum and cerebellum in rats (Carlson and House, 1986). The role of oligosaccharide supplementation has yet to be studied in human infants. LCPUFAs are one type of macronutrient that have been given specific attention in neurodevelopment. These LCPUFAs include docosahexaenoic acid (DHA; 22:6n-3) and arachidonic acid (AA; 20:4n-6). Both are transported by the placenta to the fetus (Carlson, 1997), and both accumulate in the CNS during the third trimester (Clandinin et al., 1980). These compounds are provided transplacentally by the mother during the last trimester. Accordingly, preterm infants are at greater risk of deficiency as a result of shortened nutrient placental transfer during late gestation. Furthermore, DHA and AA can be synthesized from the precursors alpha-linolenic acid and linoleic acid, respectively; however, this synthetic pathway is immature in neonates, with limited synthetic capacity as early as 33 weeks gestation. However, full synthetic capacity is not thought to develop until 2 months postterm. Because they are essential in all cell membranes, LCPUFAs potentially affect neurodevelopment. They are involved in intercellular communication, signal transduction (Carlson, 1997), and monoamine metabolism (Innis, 2003). They are critical for visual development in primates and likely in humans (Neuringer et al., 1986; Uauy et al., 1990). Evidence from animal studies shows that gestational DHA is necessary for normal brain development. Gestational DHA deficiency decreases neurogenesis in the rat brain (Coti Bertrand, O’Kusky, and Innis, 2006) and affects dopaminergic function (Levant, Radel, and Carlson, 2004). There is evidence that prenatal LCPUFA supplementation affects cognitive development in humans. In a double-blinded randomized supplementation study, children born to mothers supplemented with DHA during pregnancy and three months of lactation performed better on cognitive tasks (mental processing composite of the Kaufman Assessment Battery for Children) at 4 years of age. Furthermore, the extent of DHA supplementation was correlated with the children’s scores (Helland et al., 2003). In another maternal supplementation study, maternal DHA levels at birth were found to be correlated with the children’s performance on attentional tasks at 1 year and 2 years of age (Colombo et al., 2004). During the first 18 months postnatally LCPUFAs continue to accumulate (Clandinin et al., 1980); however, postnatal supplementation of LCPUFAs seems to have different effects in preterm and term infants. Randomized trials of LCPUFA supplementation in preterm infants show

630

transiently improved visual acuity, faster processing time on electroretinogram, better visual recognition memory, and higher scores on the Mental Developmental Index of the Bayley Scales of Infant Development at 12 months of age (Birch et al., 1992; Carlson, Werkman, and Tolley, 1996). The retinal results with respect to rod electroretinogram threshold at 36 weeks gestation and visual acuity by forcedchoice preferential looking and by visual evoked potential in preterm infants fed supplemented formula approach the findings in breast-fed infants, and are significantly better than those of infants fed unsupplemented formula (UauyDagach and Mena, 1995). Nevertheless, most of the effects appear transient in that they do not persist after the supplementation is discontinued. Furthermore, the transience of the findings is not due to withdrawal of the supplementation with subsequent lack of synthesis by the infants, since the trials typically lasted well beyond the age (6 months) when infants can synthesize LCPUFAs de novo. Term infants supplemented with LCPUFAs do not demonstrate nearly the beneficial effect that preterm infants show, perhaps because of an increased capacity to synthesize de novo AA and DHA at an earlier postnatal age (Jensen et al., 1997). In spite of the mixed long-term results, the data are quite convincing that LCPUFAs (administered without the remainder of trophic factors and nutrients found in breast milk) independently alter neurologic function and perhaps development, and these findings provide clear evidence that nutrients can affect brain development and function. Although they appear to be supplementation studies, they are better considered studies of correction of deficiencies of essential nutrients in nonhuman milk formulas. Nucleotides are nitrogenous compounds derived from the combination of a nucleic acid base (adenine, guanine, cytosine, thymine, or uracil) with a phosphorylated pentose sugar. Nucleotides and their precursors (nucleosides and nucleic acids) are critical for DNA and RNA synthesis. Thus a steady intracellular pool of nucleotides is required to ensure cell division and protein synthesis. This de novo synthetic pathway is immature in all newborns, but particularly in premature infants, raising the still unresolved issue of whether dietary nucleotides are semiessential in the newborn period. Nucleotide supplementation of infant formula has been postulated to have an impact on the developing brain by increasing the levels of LCPUFAs such as arachidonic and docosahexaenoic acid (DeLucchi, Pita, and Faus, 1987; Gil et al., 1986). Iron Iron is the second most commonly studied nutrient in relation to brain development, after protein-energy. Iron deficiency (ID) is common worldwide; 30 percent of the developing world’s population and 9 percent of U.S. toddlers are iron-deficient as a result of a low-iron diet combined (especially in the developing world) with a high

neurodevelopmental aspects of clinical disorders

rate of intestinal blood loss. Iron deficiency is the world’s most common nutritional cause of anemia. It is not unreasonable to study the relationship of iron to neurodevelopment and neurologic function, given the role of iron in many brain cellular metabolic processes (Cammack, Wrigglesworth, and Baum, 1990; Larkin and Rao, 1990; Thelander, 1990; Youdim, Ben-Sachar, and Yehuda, 1989). Iron-containing enzymes and hemoproteins are involved in numerous neurodevelopmental processes including myelination, energy metabolism, dendritogenesis, synaptogenesis, and monoamine metabolism (Rao et al., 2003). As with PEM, a convincing relationship can be drawn between this nutrient and brain development based on deficiency states. There are three time periods when children are at particular risk for iron deficiency: the fetal/neonatal period, infancy and early toddlerhood (6–24 months of age), and following the onset of menarche in girls (Bruner et al., 1996). Brain growth and development are relatively rapid during two of these periods, but they are nearly complete during the third. This distinction allows for studying the comparative effects of a single-nutrient deficiency on a growing versus a mature brain. Because of its intricate involvement in cell cycle kinetics and myelination (Larkin and Rao, 1990; Thelander, 1990), one would expect profound neuroanatomic changes in a brain that is still growing, but perhaps no structural effect in a relatively mature brain. Teenage iron deficiency would not be expected to affect myelination, since it is largely completed by that time (with the exception of myelination in the prefrontal cortex); however, iron deficiency prior to 3 years of age would likely result in profound and possibly permanent myelin changes (Algarin et al., 2003). Furthermore, areas that are growing particularly rapidly might be expected to be most affected. Since the brain does not develop homogeneously (i.e., not all parts mature simultaneously), iron deficiency during one growth period (e.g., fetal life) may result in very different neuroanatomic and neurobehavioral deficits than iron deficiency during another growth period (e.g., infancy). Iron also has important effects on neurochemistry and neurometabolism through its effects on monoamine metabolism and oxidative phosphorylation (Cammack, Wrigglesworth, and Baum, 1990; Youdim, Ben-Sachar, and Yehuda, 1989). Iron deficiency may affect these chemical and metabolic aspects of brain function similarly in developing and mature brains. Therefore, neurotransmitters such as dopamine and glutamate (Rao et al., 2003) would be vulnerable to iron deficiency at any age. Iron deficiency most commonly occurs during infancy, between 6 and 24 months of age, resulting from low dietary iron intake (through consumption of either low-iron formula or cow’s milk and delayed introduction of iron-containing solid foods). It is not surprising that this age group has been more extensively studied than any other iron-deficient group,

and animal models have focused on ID during this time. Dietary iron deficiency has been induced in rats from weaning (21 days) to 35 days of age in order to model iron deficiency of human infancy. Multiple well-controlled clinical studies in this age group demonstrate significant decrements in motor and mental achievement (Lozoff, 1990; Lozoff et al., 1982, 1987; Nokes, van den Bosch, and Bundy, 1998; Walter, Kowalskys, and Stekel, 1983). While the motor findings are less robust and generally reversible with iron therapy, the significant cognitive deficits that have been documented are more resistant to reversal, and in some studies are very long-term (Lozoff, 1990). The cognitive deficits include a 10- to 12-point reduction in the Mental Developmental Index of the Bayley Scales of Infant Development, significantly associated with the degree of anemia (Lozoff, 1990). These cognitive deficits appear to be irreversible with treatment. The findings are not thought to be due to anemia itself, since rapid correction of the anemia does not affect the neurodevelopmental test scores. The apparent permanence of these neurobehavioral findings has three important implications. First, there appear to be differential regional effects of iron on the growing brain, perhaps mediated by differences in regional blood-brain iron-transport regulation. This regionalization concept is supported by the nonhomogeneous (e.g., cognitive greater than motor) deficits induced by iron deficiency. Second, iron deficiency during this period of brain development likely changes neuroanatomy (see next paragraph), since the deficits remain extant over much longer periods of time than would be expected for neurochemical or neurophysiological alterations. Third, and perhaps of greatest concern, the results demonstrate that certain brain lesions induced by nutritional deficiencies are beyond the reach of CNS reparative or compensatory processes that may be termed “plasticity.” The presence of iron in the brain is critical for myelination (Larkin, Jarratt, and Rao, 1986). Iron deficiency in immature animal models results in a predictable loss of enzyme activity and in hypomyelination (Larkin and Rao, 1990). In animal models, early ID affects myelin lipid synthesis and permanently alters brain lipid composition despite repletion (Kwik-Uribe et al., 2000; Ortiz et al., 2004; Rao et al., 2003). If the same process occurs in infants, one could predict a specific neuroanatomic effect (potentially visible with highresolution MRI) or a neurophysiologic effect (potentially detected by electrophysiologic assessment). Although the former has not been assessed, Roncagliolo and colleagues (1998) have reported delayed latencies on auditory brainstem-evoked responses in iron-deficient 6-month-old infants. They attributed this delay to the effects of iron-deficiencyinduced hypomyelination. These effects appear to be longlasting despite iron treatment. At ages 3 and 4, delayed

fuglestad, rao, and georgieff: nutrition in cognitive development

631

latencies in auditory brain-stem responses and visual evoked potentials have been observed in children with irondeficiency anemia despite treatment during infancy (Algarin et al., 2003). Iron also has important effects on neurochemistry and neurometabolism through its effects on monoamine metabolism and oxidative phosphorylation (Cammack, Wrigglesworth, and Baum, 1990; Youdim, Ben-Sachar, and Yehuda, 1989). Monoamine metabolism involves iron-dependent enzymes (tyrosine hydroxylase and tryptophan hydroxylase), and animal studies that model ID in infancy show long-term monoamine alterations (Beard and Connor, 2003; Lozoff et al., 2006). For instance, extracellular dopamine and epinephrine increase, while monoamine transporters and dopamine receptors D1 and D2 decrease, and such alterations occur even before noticeable reductions in brain iron concentrations (Beard and Connor, 2003; Beard et al., 2006). Decreased exploration and increased hesitancy have been observed as a result of early ID in rodents, likely a result of altered dopaminergic function (Felt and Lozoff, 1996; Pinero, Jones, and Beard, 2001). There is evidence for similar behaviors in infants with ID anemia. They were observed to be more wary and hesitant, easily tired, and less playful, and to make fewer attempts at test items (Lozoff et al., 1998). Accordingly, reduced activity and inhibited exploration may lead to hampered cognitive abilities. The selectivity of iron deficiency for certain areas of the brain can be similarly assessed. Striatal development is also rapid during late gestation and early infancy and is altered by ID during this time. The striatum contains high concentrations of both iron and dopamine (Beard and Connor, 2003), and striatally mediated behaviors in the rodent are affected by ID (Felt et al., 2006). There is evidence that striatal development is altered by ID during infancy and has long-term cognitive effects. Children with severe ID during infancy performed worse on tasks that rely on striatal-frontal connections. For instance, at 11–14 years of age, children who had severe, chronic ID during infancy had poorer performance on tasks of spatial memory and selective attention than children without severe ID during infancy (Lozoff et al., 2000). Furthermore, at the age of 19, poor performance on tasks of executive function, primarily those of inhibition and planning, was associated with ID during infancy (Burden, Koss, and Lozoff, 2004). The second group of developing humans at risk for iron deficiency is late-gestation fetuses and newborns. Infants born to mothers with iron deficiency, gestational conditions such as diabetes mellitus, IUGR, and maternal smoking are at risk of fetal and neonatal iron deficiency. The adverse effects of fetal and neonatal iron deficiency have been studied in infants of mothers with diabetes mellitus during pregnancy and infants who suffered IUGR. Each group has a 50 percent prevalence of low iron stores at birth, 25 percent of

632

which are at risk for brain iron deficiency. Autopsy studies have documented up to a 40 percent decrease in brain iron in the most severely affected infants. Both groups of infants are at increased risk for neurobehavioral abnormalities. Animal models of perinatal iron deficiency support the concept of regional loss of iron-dependent brain metabolic function, with the hippocampus and its prefrontal projections demonstrating particular vulnerability (deUngria et al., 2000). Regions that are developing rapidly during a state of ID are likely to be most affected. During late fetal and neonatal development, the hippocampus is developing rapidly (see chapter 12 by Seress and Ábrahám, this volume), and animal models show alterations in hippocampal neurochemistry, structure, and electrophysiology (Jorgenson et al., 2005; Rao et al., 2003; Jorgenson, Wobken, and Georgieff, 2003). Poor performance on hippocampally mediated memory performance (e.g., spatial-memory tasks) is associated with ID in the rodent model of early ID (Felt and Lozoff, 1996). Thus a putative link between perinatal iron deficiency and newborn and long-term neurobehavioral sequelae can be proposed. Unfortunately, like all other clinical population studies, each group has other significant neurologic risk factors (e.g., hypoglycemia, hypoxia, PEM) that may affect long-term outcome and confound any attempt at causally relating perinatal iron deficiency with neurobehavioral deficits. However, in a study of infants born to mothers with diabetes mellitus, those with iron deficiency at birth showed delays in auditory recognition memory when tested at birth compared with iron-sufficient infants of diabetic mothers (Siddappa et al., 2004). This finding suggests that the effects in recognition memory are due to iron rather than to other risk factors associated with gestational diabetes. The neurobehavioral study of iron-deficient teenage girls provides an interesting contrast to iron-deficient infants, since the brain is relatively mature neuroanatomically and myelination is complete at this age. Iron supplementation of young women with iron deficiency anemia improves memory and learning but has no effect on attention (Groner et al., 1986). These findings can be contrasted with the long-term (and perhaps permanent) deficits found in toddlers who became iron deficient and were subsequently treated. The reversibility of the neurobehavioral deficits with iron therapy in the teenage population argue for an effect of iron on neurochemistry (e.g., dopamine) or neurophysiology (oxidative metabolism) as opposed to potential structural changes in neuroanatomy that may have occurred in the toddlers. Zinc A strong case can be made for the essentiality of zinc for normal brain development and function. Zinc deficiency affects neuroanatomy, neurochemistry, and neurophysiology through a variety of mechanisms. The global effects of zinc are in part due to the essential role that zinc plays in basic protein biochemistry and in cell replication. Zinc has a direct

neurodevelopmental aspects of clinical disorders

effect on brain growth and morphology through its role in enzymes that mediate protein and nucleic acid synthesis (Sandstead, 1985; Terhune and Sandstead, 1972). Profound zinc deficiency in the growing animal results in decreased brain DNA, RNA, and protein content (Duncan and Hurley, 1978; Sandstead, 1985). Zinc has an indirect effect on brain growth because its presence is needed for normal insulin-like growth factor–I (IGF-I) activity (McNall, Etherton, and Fosmire, 1995). Zinc influences brain neurochemistry (and presumably function) by inhibiting binding of opioids to μ receptors and of magnesium to μ and δ receptors in the cerebral cortex (Tejwani and Hanissian, 1990). Zinc also inhibits γ-aminobutyric acid (GABA) stimulated chloride influx into hippocampal neurons (Li, Rosenberg, and Chiu, 1994). Zinc is released into the interneuronal space from presynaptic boutons (Frederickson and Danscher, 1990). Zinc’s effect on brain neurophysiology is evident in abnormal electroencephalographic tracings found in zinc-deficient rats (Hesse, 1979). The cerebellum, limbic system, and cerebral cortex are particularly rich in zinc and demonstrate the most profound effects of zinc deprivation (Frederickson and Danscher, 1990). These effects include truncated dendritic arborization and reduced regional brain mass in rats. As with iron deficiency, the neuroanatomic sequelae of zinc deficiency in young rats persist into adulthood with attendant persistent behavioral sequelae (Frederickson and Danscher, 1990). Two-year-old rhesus monkeys fed a zinc-deficient diet have reduced spontaneous motor activity and poorer short-term memory compared with the predeficiency baseline period (Golub et al., 1994). Similar results have been found in mice and rats. Zinc is essential for growth, as it is involved in cell replication and nucleic-acid and protein synthesis. Accordingly, zinc is especially critical for infants, children, adolescents, and pregnant women (FAO/WHO, 2004). It is likely that during times of rapid growth and/or rapid brain growth, individuals might be vulnerable to zinc deficiency. The rate of inadequate dietary intakes of zinc in infants and toddlers in developing countries suggests that the rate of zinc deficiency is fairly high (FAO/WHO, 2004). Furthermore, inadequate dietary zinc intakes among middle-class and upper-class infants and toddlers in the United States also suggest that the rate of zinc deficiency is common among American children (Skinner et al., 1997), and according to the latest national nutrition survey (NHANES III), young children ages 1–3 and adolescent females are at risk for inadequate zinc intakes (Briefel et al., 2000). Despite the evidence in animal models, zinc supplementation trials in children have yielded inconsistent results. During infancy, there is some implication that zinc supplementation improves motor development and promotes activity in the most severe cases of zinc deficiency (for review,

see Black, 2003). However, other findings show that zinc deficiency in humans results in significant changes in neuropsychological performance on tests that tap the anatomical areas shown to be vulnerable in the animal models (Sandstead et al., 1998). Six- to nine-year-old first-graders with low zinc status were assessed biochemically and neuropsychologically before and after treatment. The tasks included design matching to assess visual perception, delayed design matching to assess short-term visual memory, a spatial orientation memory test, and Pollitt’s oddity task to assess concept formation and abstract reasoning. Zinc supplementation for 10 weeks resulted in significantly better zinc status and improvement in these particular neuropsychological assessments. Iodine There is overwhelming evidence that iodine sufficiency is critical for normal early CNS development. An analysis of available epidemiological studies has helped elucidate the effect of severity, timing, and duration of iodine deficiency on brain development and neurologic outcome. Iodine deficiency can range from severe to mild based on the availability of the iodine in the food supply (Hetzel and Mano, 1989). Endemic areas of severe dietary iodine deficiency include parts of China, Zaire, Iran, and India (Hetzel and Mano, 1989; Kretchmer, Beard, and Carlson, 1996), with 35 percent of the world’s population having insufficient iodine intake (de Benoist et al., 2004). Classic older studies from these areas helped characterize the syndrome of endemic cretinism that, in its severest neurologic manifestation, includes mental retardation, spastic diplegia, and deaf-mutism. More recently, investigators have concentrated on describing the neuropsychological and motor effects of moderate or mild iodine deficiency and the effect of iodine prophylaxis or treatment in high-risk groups (Aghini-Lombardi et al., 1995; Azizi et al., 1993). A clear dose-response effect can be appreciated across the studies, with moderate iodine deficiency resulting in reduced verbal IQ and subset-coding ability on the WISC-R as well as motor impairments on simple reaction-time tests in children of elementary school age (Fenzi, Giusti, and AghiniLombardi, 1990). Mild iodine deficiency resulted in reduced motor ability without any apparent effect on cognition when children of the same age were assessed with the same tools (Azizi et al., 1993). It is unclear whether the latter group had early cognitive findings that were “reversed” with postnatal iodine treatment or whether they were never affected in the first place. As with any nutrient deficiency, timing and duration are of critical importance. Cretinism results from severe iodine deficiency during early pregnancy, rather than late pregnancy. Moreover, developmental delays associated with prenatal iodine deficiency appear to be specific to deficiency during the first two trimesters, but not the third (Cao et al.,

fuglestad, rao, and georgieff: nutrition in cognitive development

633

1994; O’Donnell et al., 2002). Interventional studies with iodine treatment clearly demonstrate that prophylaxis is effective. Iodine-deficient women living in endemic areas who were injected with iodinated oil prior to pregnancy did not produce infants with cretinism (Hetzel, 1987). In a Chinese population at risk for iodine deficiency, iodine supplementation before the third trimester predicted higher BSID scores at age 2 (Cao et al., 1994) and higher psychomotor test scores around the age of 6 (O’Donnell et al., 2002) compared with those who received supplementation during the third trimester or postnatally. In animal models, treatment later in pregnancy had a much less dramatic effect, and postnatal treatment of humans and animals appeared to have little effect. These observations lend credence to the hypothesis that the critical window for brain responsiveness to iodine is during early fetal life. Nevertheless, it should be noted that children and adolescents living in areas of endemic iodine deficiency will demonstrate alterations in psychomotor development with reduced IQ, even with normal physical growth (an indicator of less severe iodine deficiency) (Azizi et al., 1993). Studies that assess the effect of repletion of iodine status on psychomotor performance in children with mild to moderate iodine deficiency remain to be performed. However, a recent randomized, placebo-controlled intervention found that iodine supplementation improved performance on cognitive tasks of information processing and problem solving in children between the ages of 9 and 10 who had moderate to severe deficiency (Zimmerman et al., 2006). There is reason to believe that late-onset iodine deficiency (hypothyroidism) affects brain function, but not anatomy. Smith and Ain (1995), using 31P magnetic resonance spectroscopy, demonstrated reduced oxidative metabolism in the hypothyroid brain (Pleasure and Pleasure, 2003) that was reversible with thyroid replacement therapy. Selenium Selenium is a micronutrient whose role in brain development is only now being elucidated. Selenium deficiency is seen in geographical regions where there are low selenium levels in the soil. Food crops and pasture grasses grown in these soils will have lower selenium content; therefore, populations who depend on local food crops and animal products are at risk for selenium deficiency. The highest rates of selenium deficiency are in communities living in several regions throughout China with low soil content (FAO/WHO, 2004). Several other regions with low soil content (e.g., Finland, New Zealand, United Kingdom) have reduced the rates of selenium deficiency by importing crops and animal products raised in high-content regions. Finally, another population at relatively high risk of selenium deficiency is preterm infants, because of their lower body stores and generally poorer antioxidant status. Although direct evidence of selenium’s effect on cognitive development in humans is lacking, its role in brain thyroid

634

and iodine metabolism, as well as its interaction with other micronutrients (e.g., iron, copper, zinc, lead) that affect brain development, is important to note. Selenium is required for the synthesis of proteins (selenoproteins) that are involved in thyroid metabolism. Thus, as with iodine, selenium deficiency can lead to hypothyroidism and cretinism. Studies linking selenium status with behavioral development in animal models have been published (Mitchell et al., 1998; Watanabe and Satoh, 1994). Thus it is likely that research on the role of this nutrient in human brain development will be forthcoming. Choline Developing fetuses and postnatally breast-fed infants may be at risk for choline deficiency if the mother has poor choline body stores and/or poor dietary choline intake. Formula-fed infants may also be at risk for choline deficiency depending on the formula used, as the choline concentrations range between 50 and 140 percent of that found in human milk. Human milk is also abundant in several cholinecontaining compounds, and formulas do not contain the same composition of these compounds as human milk (Blusztajn, 1998; Meck and Williams, 1999). Additionally, during the early postnatal period, there is a brief spike in choline concentrations in human milk that is not mirrored in infant formulas (Meck and Williams, 1999). Choline serves several metabolic functions essential for neurodevelopment. The majority of choline in the body is present in two phospholipids that are essential for both plasma-membrane and myelin synthesis (Blusztajn, 1998; Colombo et al., 2003). Choline is also involved in cholinergic neurotransmission, as it is the precursor for acetylcholine, serves as a methyl source in single-carbon metabolism of protein synthesis and transmethylation reactions (Blusztajn, 1998), and is involved in transmembrane signaling during neurogenesis and synaptogenesis (Meck and Williams, 1999). During fetal development, choline is essential for stem-cell proliferation and apoptosis (Blusztajn, 1998). Choline has permanent effects on rodent memory and attention development. Prenatal choline supplementation results in better spatial-memory performance in the adult rat compared with control rats and impairs memory performance in prenatally choline-deficient rats (Meck and Williams, 1999). Prenatal choline deficiency is associated with long-term effects on attention that may be related to the long-term effects on memory. Gestationally deficient rats perform poorly on attention tasks compared with controls and choline-supplemented rats, showing difficulty selectively attending to stimuli and showing an increased rate of agerelated decline in attention (Meck and Williams, 2003). Several mechanisms for the effects of choline supplementation on memory development have been identified (Meck and Williams, 2003). Prenatal choline deficiency increases the rate of apoptosis in the hippocampus, thus potentially

neurodevelopmental aspects of clinical disorders

altering hippocampal structure and function (HolmesMcNary et al., 1997). Another potential mechanism involves adaptations in cholinergic transmission (Meck and Williams, 2003); prenatal choline has long-term effects on cholinergic transmission, altering hippocampal acetylcholine function despite repletion (Blusztajn et al., 1998). Yet another mechanism linking choline to memory is through hippocampal long-term potentiation (LTP); the stimulus threshold for LTP is inversely related to prenatal choline status (Pyapali et al., 1998). Because choline supplementation of older animals does not seem to improve memory as it does following supplementation during fetal development, it is likely that choline during early development alters the developmental trajectory of neural circuits that support memory, leading to long-term alterations in memory and attention (Meck and Williams, 2003). Despite the mounting evidence from animal research for the role of choline in neurodevelopment and memory and attention performance, questions still exist whether these findings translate to humans. However, the evidence from animal models highly supports the hypothesis that dietary choline during pregnancy and the early postnatal period is necessary for normal cognitive development.

The role of timing in nutrient deprivation and subsequent repletion A common theme in the discussion of each of the nutrients has been the roles of timing, duration, and severity of nutrient administration or deficiency. Clear evidence exists that brain growth and development is a nonlinear process and that various “circuits” come on line at different times between conception and adolescence. Nutrient requirements are increased during a period of rapid growth for any organ, including the brain. Human and animal model experiments also support the concept that nutritional deprivations during a period of rapid growth result in more profound structural, chemical, and physiological changes than if the same degree of deprivation is imposed during a more quiescent period. What remains of great interest is whether these early changes are reversible. Reversibility (or amenability to nutritional rehabilitation) can be looked at in two ways—through a large-scale population-based approach, as Pollitt has done (Pollitt and Gorman, 1994), and through further understanding of the critical nutrients required for normal ontogeny of brain development. Both approaches beg the question of “plasticity” within the developing system. Pollitt’s studies clearly demonstrate that developmental benefits of nutrient supplementation are a function of the timing of that intervention (Pollitt et al., 1993). The earlier the timing, the greater the benefits. Yet, it is difficult to argue that the “critical window”

is ever “closed” during childhood. Studies of children suffering from prenatal malnutrition (IUGR; Morgane et al., 1993; Strauss and Dietz, 1998) or postnatal malnutrition (Pollitt et al., 1993) clearly demonstrate recoverability of function well beyond the period of rapid brain growth. A similar effect is seen in the growth and development of preterm infants. More than 50 percent of preterm infants become microcephalic during their hospital stay as a result of inadequate nutrient delivery, severe illness resulting in catabolism, and a very rapid expected rate of brain growth (Georgieff et al., 1985). Although nutritional management of preterm infants has improved substantially (Georgieff et al., 1989), preterm infants leave the NICU with significant postnatal growth restriction and altered body composition that remain extant until more than 1 year of life. Catch-up growth and development have been described well into the teenage years. The developmental outcome of these infants far exceeds the expectations based on the number of neurologic risk factors (including malnutrition) they encounter. A reasonable hypothesis is that the expected developmental sequelae of this “extrauterine growth restriction” are reversed postnatally with catch-up growth. Although it would be preferable never to have had a nutritional deficit in the first place, it is clear that recovery is entirely possible and, in fact, likely. Consequently, periods of rapid growth may not only confer a risk factor when nutrients are deprived, but may also offer a window of opportunity for rapid and complete repair when nutrients are subsequently provided. Pollitt (1996) notes that there is evidence for recovery following PEM in all age groups, including fetuses that receive adequate postnatal nutrition (IUGR), infants malnourished before 2 years of age who were not supplemented until after 3 years, and infants malnourished postnatally but also treated before 2 years of age. A review of the nutrients that affect brain growth and development reveals striking differences with respect to vulnerability and timing. Deficiencies of certain nutrients, such as selenium, folate, and vitamin A, exert their effects in a very narrow postconceptional window. Later supplementation of these nutrients will do little to alter damage that occurred in the first 12 weeks postconception. These nutrients play important roles in neuronal differentiation, cell division, protein synthesis, and neuronal migration. Other nutrients (e.g., protein, energy, iron, zinc, iodine) play a role throughout development. Deficiencies of these nutrients at different ages result in variable neuroanatomic, neurochemical, and neurobehavioral effects. These differences may be regionalized within the brain, with certain areas being spared at one age and not at another (deUngria et al., 2000; Erikson et al., 1997). In humans, nutrient deficiencies tend to cluster and are frequently prolonged, thus exposing vulnerable brain regions to multiple nutrient deficiencies over time. Only through careful modeling of each nutrient’s effect on

fuglestad, rao, and georgieff: nutrition in cognitive development

635

brain development combined with utilization of nutrientspecific assessments (e.g., if hypomyelination is a distinguishing hallmark of iron deficiency, the effect in a population could be studied electrophysiologically) (Roncagliolo et al., 1998) will it become possible to unravel the precise effects of nutrient deficiencies at the various stages of neurodevelopment.

Future Directions Epigenetic Processes Epigenetic processes are those that alter the mammalian genome by covalently adding a methyl group to specific genomic sites (CpG sites—cytosine followed by guanosine). This methylation of DNA affects numerous events including DNA repair, gene stability, and gene transcription (Robertson and Jones, 2000). For example, addition of methyl within promoter regions inhibits genetic transcription. Several nutrients, such as choline, folate, vitamin B12, methionine, and betaine, are involved in the metabolic pathways that supply methyl groups for all methylation reactions (Mason, 2003), and reductions in dietary methyl sources can change genetic expression and subsequent phenotypes (Cooney, Dave, and Wolff, 2002; Waterland and Jirtle, 2003). Choline is a key dietary source of methyl groups, and there is a growing body of evidence for choline’s role in epigenetic processes and related brain development (Zeisel, 2004). Choline deficiency in human neuroblast cell cultures decreases methylation of the promoter region of a gene responsible for inhibition of cell proliferation. Hypomethylation of this gene’s promoter leads to its overexpression and subsequent reduction in cell proliferation (Niculescu, Yamamuro, and Zeisel, 2004). Choline has also been shown to alter expression of numerous genes of neuronal precursor cells, several of which are known to be regulated by the methylation of promoter or intron regions (Niculescu, Craciunescu, and Zeisel, 2005). Furthermore, epigenetic process (genomic methylation) is a mechanism by which early maternal care can affect offspring throughout life. In the rat, maternal care affects the development of the hypothalamic-pituitary-adrenal system through epigenetic processes (Weaver et al., 2004). Although nutrition has not been examined in the context of such epigenetic processes, which link maternal behavior to offspring development, it will be interesting in future research to examine the potential role of nutrients required for methylation reactions in these and other epigenetic processes involved in neurodevelopment. Fetal Origins Hypothesis According to the fetal origins hypothesis, organisms adapt to their in utero nutritional environment, and these adaptations have permanent effects

636

on physiology and metabolism (Barker, 1997). The in utero environment acts as a “forecast” of the nutritional environment into which the fetus will be born; therefore, structural, physiological, and metabolic adaptations occur during fetal development to prepare for this “forecasted” environment (Hales and Barker, 2001). For instance, newborns born small for gestational age (IUGR) may develop a thrifty phenotype. Fetal undernutrition leads to physiological and structural adaptations in various tissues and organs (such as pancreas, liver, and blood vessels) in preparation for a suboptimal postnatal nutritional environment. The initial findings have found associations between IUGR and coronary heart disease, type 2 diabetes, and hypertension; however, it will be important to consider brain development as well. Physiological and metabolic adaptations to undernutrition involve regulatory hormones of fetal growth, including insulin and growth hormones (Hales and Barker, 2001). These adaptations may alter the availability and utilization of nutrients. Accordingly, depending on the early environment, differential programming would lead to different nutrient metabolism throughout life and, therefore, nutrient availability to the CNS. Thus it will be important to consider the consequences of fetal programming of nutrient metabolism on brain development. Complementary and Supplementary Measures The apparent permanence of some cognitive impairments in nutrient deficiencies suggests the need for additional measures that complement the specific nutrient supplementation. Nutritional and nonnutritional measures that influence disparate pathways of a specific neurological process appear attractive for this purpose. For example, simultaneous supplementation of LCPUFAs may correct hypomyelination that is due to iron deficiency, since both nutrients are involved in fatty acid metabolism (Rioux, Lindmark, and Hernell, 2006). Similarly, environmental enrichment may have additive beneficial effects in nutritional deficiency conditions (figure 38.1). However, a thorough understanding of their mechanistic aspects is necessary for these measures to be successful.

Conclusions Adequate intake of all nutrients throughout life is important for brain health and function. The specific nutrients reviewed in this chapter (e.g., macronutrients, iron, zinc, iodine, selenium, and choline) have a more profound effect on cognitive development than others. However, nutrient intake alone is not sufficient to support brain health and function, although nutrient assimilation influenced by growth factors is important. Furthermore, the effects of nutrients on cognitive development are dependent on both environmental and biological contexts. The timing and extent of nutrient

neurodevelopmental aspects of clinical disorders

supplementation or deprivation have an important effect on brain development and function. Although some of the cognitive impairments may be amenable to nutritional rehabilitation, long-term effects remote from the period of nutritional deficiency or overload may occur.

REFERENCES Aghini-lombardi, F. A., A. Pinchera, L. Antonangeli, T. Rago, L. Chiovato, S. Bargagna, B. Bertucelli, G. Ferretti, B. Sbrana, and M. Marcheschi, 1995. Mild iodine deficiency during fetal/neonatal life and neuropsychological impairment in Tuscany. J. Endocrinol. Invest. 18:57–72. Algarin, C., P. Peirano, M. Garrido, F. Pizarro, and B. Lozoff, 2003. Iron deficiency anemia in infancy: Long-lasting effects on auditory and visual system functioning. Pediatr. Res. 5:217–223. Azizi, F., A. Sarshar, M. Nafarabadi, A. Ghazi, M. Kimiagar, S. Noohi, N. Rahbar, A. Bahrami, and S. Kalantari, 1993. Impairment of neuromotor and cognitive development in iodinedeficient schoolchildren with normal physical growth. Acta Endocrinol. (Copenh.) 129:497–500. Barker, D. J. P., 1997. The long-term outcome of retarded fetal growth. Clin. Obstet. Gynecol. 40:853–863. Bass, N. H., M. G. Netsky, and E. Young, 1970. Effects of neonatal malnutrition on developing cerebrum. I. Microchemical and histologic study of cellular differentiation in the rat. Arch. Neurol. 23:289–302. Beard, J. L., and J. R. Connor, 2003. Iron status and neural functioning. Annu. Rev. Nutr. 23:41–58. Beard, J. L., B. Felt, T. Schallert, M. Burhans, J. R. Connor, and M. K. Georgieff, 2006. Moderate iron deficiency in infancy: Biology and behavior in young rats. Behav. Brain Res. 170:224–232. Birch, E. E., D. G. Birch, D. R. Hoffman, and R. Uauy, 1992. Dietary essential fatty acid supply and visual acuity development. Invest. Ophthalmol. Vis. Sci. 33:3242–3253. Black, L. S., R. A. DeRegnier, J. Long, M. K. Georgieff, and C. A. Nelson, 2004. Electrographic imaging of recognition memory in 34–38 week gestation intrauterine growth restricted newborns. Exp. Neurol. 190:72–83. Black, M. M., 2003. The evidence linking zinc deficiency with children’s cognitive and motor functioning. J. Nutr. 133: 1473S–1476S. Blusztajn, J. K., 1998. Choline, a vital amine. Science 281: 794–795. Blusztajn, J. K., J. M. Cermak, T. Holler, and D. A. Jackson, 1998. Imprinting of hippocampal metabolism of choline by its availability during gestation: Implications for cholinergic transmission. J. Physiol. 92:199–203. Briefel, R. R., K. Bialostosky, J. Kennedy-Stephenson, M. A. McDowell, R. B. Ervin, and J. D. Wright, 2000. Zinc intake of the U.S. population: Findings from the third National Health and Nutrition Examination Survey, 1988–1994. J. Nutr. 130: 1367S–1373S. Bruner, A. B., A. Joffe, A. K. Duggan, J. F. Casella, and J. Brandt, 1996. Randomized study of cognitive effects of iron supplementation in non-anemic iron-deficient adolescent girls. Lancet 348:992–996. Burden, M., M. Koss, and B. Lozoff, 2004. Neurocognitive differences in 19-year-olds treated for iron deficiency in infancy. Pediatr. Res. 55:279A.

Cammack, R., J. M. Wrigglesworth, and H. Baum, 1990. Irondependent enzymes in mammalian systems. In Iron Transport and Storage, ed. P. Ponka, H. M. Schulman, and R. C. Woodworth, 18–39. Boca Raton, FL: CRC Press. Cao, X.-Y., X. M. Jiang, Z. H. Dou, M. A. Rakeman, M. L. Zhang, K. O’Donnell, T. Ma, K. Amette, N. DeLong, and G. R. DeLong, 1994. Timing of vulnerability of the brain to iodine deficiency in endemic cretinism. N. Engl. J. Med. 331:1739–1744. Carlson, S. E., 1997. Long-chain polyunsaturated fatty acid supplementation of preterm infants. In J. Dobbing, ed., Developing Brain and Behaviour, 41–78. London: Academic Press. Carlson, S., and S. House, 1986. Oral and intraperitoneal administration of N-acetylneuraminic acid: Effect on rat cerebral and cerebellar N-acetylneuraminic acid. J. Nutr. 116:881– 886. Carlson, S. E., S. H. Werkman, and E. A. Tolley, 1996. The effect of long chain n-3 fatty acid supplementation on visual acuity and growth of preterm infants with and without bronchopulmonary dysplasia. Am. J. Clin. Nutr. 63:687–697. Clandinin, M. T., J. E. Chappell, S. Leong, T. Heim, P. R. Swyer, and G. W. Chance, 1980. Intrauterine fatty acid accretion rates in human brain: Implications for fatty acid requirements. Early Hum. Deve. 4:121–129. Colombo, J. P., C. Garcia-Rodenas, P. R. Guesry, and J. Rey, 2003. Potential effects of supplementation with amino acids, choline or sialic acid on cognitive development in young infants. Acta Paediatr. 92:42–46. Colombo, J., K. N. Kannass, D. J. Shaddy, S. Kundurthi, J. M. Maikranz, C. J. Anderson, O. M. Blaga, and S. E. Carlson, 2004. Maternal DHA and the development of attention in infancy and toddlerhood. Child Dev. 75:1254–1267. Cooney, C. A., A. A. Dave, and G. L. Wolff, 2002. Maternal methyl supplements in mice affect epigenetic variation and DNA methylation of offspring. J. Nutr. 132:2393S–2400S. Copp, A. J., and M. Bernfield, 1994. Etiology and pathogenesis of human neural tube defects. Curr. Opin. Pediatr. 6:626–631. Coti Bertrand, P., J. R. O’Kusky, and S. M. Innis, 2006. Maternal dietary (n-3) fatty acid deficiency alters neurogenesis in the embryonic rat brain. J. Nutr. 136:1570–1575. Cragg, B. G., 1972. The development of cortical synapses during starvation in the rat. Brain 95(1):143–150. de Benoist, B., M. Andersson, I. Egli, B. Takkouche, and H. Allen, eds., 2004. Iodine Status Worldwide: WHO Global Database on Iodine Deficiency. Geneva: World Health Organization. DeLucchi, C., M. L. Pita, and M. J. Faus, 1987. Effects of dietary nucleotides on the fatty acid composition of erythrocyte membrane lipids in term infants. J. Pediatr. Gastroenterol. Nutr. 6:568– 574. deRegnier, R. A., T. W. Guilbert, M. M. Mills, and M. K. Georgieff, 1996. Growth failure and altered body composition are established by one month of age in infants with bronchopulmonary dysplasia. J. Nut. 126:168–175. deUngria, M., R. Rao, J. D. Wobken, M. Luciana, C. A. Nelson, and M. K. Georgieff, 2000. Perinatal iron deficiency decreases cytochrome c oxidase activity in selected regions of neonatal rat brain. Pediatr. Res. 48:169–176. DeWolf, F., W. B. Robertson, and I. Brosen, 1975. The ultrastructure of acute atherosis in hypertensive pregnancy. Am. J. Obstet. Gynecol. 123:164–174. Dobbing, J., 1990. Vulnerable periods in developing brain. In J. Dobbing, ed., Brain, Behavior and Iron in the Infant Diet, 1–25. London: Springer-Verlag.

fuglestad, rao, and georgieff: nutrition in cognitive development

637

Duncan, J., and L. Hurley, 1978. Thymidine kinase and DNA polymerase activity in normal and zinc deficient developing rat embryos. Proc. Soc. Exp. Biol. Med. 159:39–43. Erikson, K. M., D. J. Pinero, J. R. Connor, and J. L. Beard, 1997. Regional brain iron, ferritin and transferring concentrations during iron deficiency and iron repletion in developing rats. J. Nutr. 127:2030–2038. FAO/WHO, 2004. Vitamin and Mineral Requirements in Human Nutrition. Geneva: WHO Press. Fellows, R., 1987. 1GF-1 supports and differentiation of fetal rat brain neurons in serum-free, hormone-free defined medium. Soc. Neurosci. Abstracts 13:1615. Felt, B., J. Beard, T. Schallert, J. Shao, J. W. Aldridge, J. R. Connor, M. K. Georgieff, and B. Lozoff, 2006. Persistent neurochemical and behavioral abnormalities in adulthood despite early iron supplementation for perinatal iron deficiency anemia in rats. Behav. Brain Res. 171(2):261–270. Felt, B. T., and B. Lozoff, 1996. Brain iron and behavior of rats are not normalized by treatment of iron deficiency anemia during early development. J. Nutr. 126:693–701. Fenzi, G. F., L. F. Giusti, and F. Aghini-Lombardi, 1990. Neuropsychological assessment of schoolchildren from an area of moderate iodine deficiency. J. Endocrinol. Invest. 13:427–432. Frederickson, C., and G. Danscher, 1990. Zinc-containing neurons in hippocampus and related CNS structures. Prog. Brain Res. 83:71–84. Georgieff, M. K., J. S. Hoffman, G. R. Pereira, J. Bernbaum, and M. Hoffman-Williamson, 1985. Effect of neonatal caloric deprivation on head growth and 1-year developmental status in preterm infants. J. Pediatr. 107:581–587. Georgieff, M. K., M. M. Mills, L. L. Lindeke, S. Iverson, D. E. Johnson, and T. R. Thompson, 1989. Changes in nutritional management and outcome of very-low-birth-weight infants. Am. J. Dis. Child. 143:82–85. Geva, R., R. Eshel, Y. Leitner, A. F. Valevski, and S. Harel, 2006. Neuropsychological outcome of children with intrauterine growth restriction: A 9-year prospective study. Pediatrics 118:91–100. Gil, A., M. Pita, A. Martinez, J. A. Molina, and F. Sanchez Medina, 1986. Effect of dietary nucleotides on the plasma fatty acids in at-term neonates. Hum. Nutr. Clin. Nutr. 40:185– 195. Golub, M. S., P. T. Takeuchi, C. L. Keen, M. E. Gershwin, A. G. Hendricks, and B. Lonnerdal, 1994. Modulation of behavioral performance of prepubertal monkeys by moderate dietary zinc deprivation. Am. J. Clin. Nutr. 60:238–243. Gorman, K. S., 1995. Malnutrition and cognitive development: Evidence from experimental/quasi-experimental studies among the mild-moderately malnourished. J. Nutr. 125:S2239–S2244. Gottleib, S. J., F. J. Biasini, and N. W. Bray, 1988. Visual recognition memory in IUGR and normal birthweight infants. Infant Behav. Dev. 11:223–228. Grantham-McGregor, S. M., P. I. C. Lira, A. Ashworth, S. S. Morris, and A. M. S. Assuncao, 1998. The development of low birth weight term infants and the effects of the environment in Northeast Brazil. J. Pediatr. 132:661–666. Groner, J. A., N. A. Holtzman, E. Charney, and E. D. Mellits, 1986. A randomized trial of oral iron on tests of short-term memory and attention span in young pregnant women. J. Adolesc. Health Care 7:44–48. Guilarte, T. R., 1993. Vitamin B6 and cognitive development: Recent research findings from human and animal studies. Nutr. Rev. 51:193–198.

638

Hales, C. N., and D. J. P. Barker, 2001. The thrifty phenotype hypothesis. Br. Med. Bull. 60:5–20. Harvey, D., J. Prince, J. Bunton, C. Parkinson, and S. Campbell, 1982. Abilities of children who were small-for-gestational-age babies. Pediatrics 69:296–300. Helland, I. B., L. Smith, K. Saarem, O. D. Saugstad, and C. A. Drevon, 2003. Maternal supplementation with very-long-chain n-3 fatty acids during pregnancy and lactation augments children’s IQ at 4 years of age. Pediatrics 111:e39–44. Hesse, G., 1979. Chronic zinc deficiency alters neuronal function of hippocampal mossy fibers. Science 205:1005–1007. Hetzel, B. S., 1987. An overview of the prevention and control of iodine deficiency disorders. In B. S. Hetzel, J. T. Dunn, and J. B. Standbury, eds., The Prevention and Control of Iodine Deficiency Disorders, 7–34. Amsterdam: Elsevier. Hetzel, B. S., and M. T. Mano, 1989. A review of experimental studies of iodine deficiency during fetal development. J. Nutr. 119:145–151. Holmes-McNary, M. Q., R. Loy, M. H. Mar, C. D. Albright, and S. H. Zeisel, 1997. Apoptosis is induced by choline deficiency in fetal brain and in PC12 cells. Dev. Brain Res. 101: 9–16. Innis, S. M., 1992. Human milk and formula fatty acids. J. Pediatr. 120:S55–61. Innis, S. M., 2003. Perinatal biochemistry and physiology of longchain polyunsaturated fatty acids. J. Pediatr. 143:S1–8. Jensen, C. L., T. C. Prager, J. K. Fraley, H. Chen, R. E. Anderson, and W. C. Heird, 1997. Effects of dietary linoleic/alphalinolenic acid ratio on growth and visual function of term infants. J. Pediatr. 131:173–175. Jones, D. G., and S. E. Dyson, 1981. The influence of protein restriction, rehabilitation and changing nutritional status on synaptic development: A quantitative study in rat brain. Brain Res. 208:97–111. Jorgenson, L. A., M. Sun, M. O’Connor, and M. K. Georgieff, 2005. Fetal iron deficiency disrupts the maturation of synaptic function and efficacy in area CA1 of the developing rat hippocampus. Hippocampus 15:1094–1102. Jorgenson, L. A., J. D. Wobken, and M. K. Georgieff, 2003. Perinatal iron deficiency alters apical dendritic growth in hippocampal CA-1 pyramidal neurons. Dev. Neurosci. 25:412– 420. Kim, M., Z. X. Yu, B. B. Fredholm, and S. A. Rivkees, 2005. Susceptibility of the developing brain to acute hypoglycemia involving A1 adenosine receptor activation. Am. J. Physiol. Endocrinol. Metab. 289:E562–569. Kretchmer, N., J. L. Beard, and S. Carlson, 1996. The role of nutrition in the development of normal cognition. Am. J. Clin. Nutr. 63:997S–1001S. Kunz, C., S. Rudolf, W. Baier, N. Klein, and S. Strobe, 2000. Oligosaccharides in human milk: Structural, functional and metabolic aspects. Annu. Rev. Nutr. 20:699–722. Kwik-Uribe, C. L., D. Gietzen, J. B. German, M. S. Golub, and C. L. Keen, 2000. Chronic marginal iron intakes during early development in mice result in persistent changes in dopamine metabolism and myelin composition. J. Nutr. 130:2821–2830. Larkin, E. C., B. A. Jarratt, and G. A. Rao, 1986. Reduction of relative levels of nervonic to lignoceric acid in the brain of rat pups due to iron deficiency. Nutr. Res. 6:309–314. Larkin, E. C., and G. A. Rao, 1990. Importance of fetal and neonatal iron: Adequacy for normal development of central nervous system. In J. Dobbing, ed., Brain, Behavior and Iron in the Infant Diet, 43–62. London: Springer-Verlag.

neurodevelopmental aspects of clinical disorders

Lee, K.-H., A. Kalikoglu, P. Ye, and A. J. D’Ercole, 1999. Insulin-like growth factor–1 (1GF-1) ameliorates and 1GF binding protein–1 (1GFBP-1) exacerbates the effects of undernutrition on brain growth during early postnatal life: Studies in 1GF-1 and 1GFBP-1 transgenic mice. Pediatr. Res. 45:331– 336. Levant, B., J. D. Radel, and S. E. Carlson, 2004. Decreased brain docosahexaenoic acid during development alters dopamine-related behaviors in adult rats that are differentially affected by dietary remediation. Behav. Brain Res. 152: 49–57. Li, M., H. Rosenberg, and T. Chiu, 1994. Zinc inhibition of GABA-stimulated Cl− flux in rat brain regions is unaffected by acute or chronic benzodiazepine. Pharmacol. Biochem. Behav. 49:477–482. Low, J. A., and R. S. Galbraith, 1974. Pregnancy characteristics of intrauterine growth retardation. Obstet. Gynecol. 44:122–126. Low, J. A., R. S. Galbraith, D. Muir, H. Killen, B. Pater, and J. Karchmar, 1982. Intrauterine growth retardation: A study of long-term morbidity. Am. J. Obstet. Gynecol. 142:670–677. Lozoff, B., 1990. Has iron deficiency been shown to cause altered behavior in infants? In J. Dobbing, ed., Brain, Behavior and Iron in the Infant Diet, 107–131. London: Springer-Verlag. Lozoff, B., J. L. Beard, J. R. Connor, B. T. Felt, M. K. Georgieff, and T. Schallert, 2006. Long-lasting neural and behavioral effects of iron deficiency in infancy. Nutr. Rev. 64: S34–S43. Lozoff, B., G. M. Brittenham, F. E. Viteri, A. W. Wolf, and J. J. Urrutia, 1982. The effects of short-term oral iron therapy on developmental deficits in iron-deficient anemic infants. J. Pediatr. 100:351–357. Lozoff, B., G. M. Britteham, A. W. Wolf, D. K. McClish, P. M. Kuhnert, E. Jimenez, L. A. Mora, I. Gomez, and D. Krauskoph, 1987. Iron-deficiency anemia and iron therapy: Effects on infant developmental test performance. Pediatrics 79:981–995. Lozoff, B., and M. K. Georgieff, 2006. Iron deficiency and brain development. Sem. Pediatr. Neurol. 13(3):158–165. Lozoff, B., E. Jimenez, J. Hagen, E. Mollen, and A. W. Wolf, 2000. Poorer behavioral and developmental outcome more than 10 years after treatment for iron deficiency in infancy. Pediatrics 105:e51. Lozoff, B., N. Klein, E. Nelson, D. McClish, M. Manuel, and M. Chacon, 1998. Behavior of infants with iron-deficiency anemia. Child Dev. 69:24–36. Lucas, A., 1997. Long-chain polyunsaturated fatty acids, infant feeding and cognitive development. In J. Dobbing, ed., Developing Brain and Behaviour, 3–27. London: Academic Press. MacLean, W. C., Jr., and J. D. Benson, 1989. Theory into practice: The incorporation of new knowledge into infant formula. Sem. Perinatol. 13:104–111. Mangelsdorf, D. J., 1994. Vitamin A receptors. Nutr. Rev. 52: S32–S37. Mason, J. B., 2003. Biomarkers of nutrient exposure and status in one-carbon (methyl) metabolism. J. Nutr. 133:941S–947S. McMorris, F. A., and M. Dubois-Dalcq, 1988. Insulin-like growth factor 1 promotes cell proliferation and oligodendroglial commitment in rat glial progenitor cells developing in vitro. J. Neurosci. Res. 21:199–209. McNall, A., T. Etherton, and G. Fosmire, 1995. The impaired growth induced by zinc deficiency in rats is associated with decreased expression of hepatic insulin-like growth factor 1 and growth hormone receptor genes. J. Nutr. 124:874–879.

Meck, W. H., and C. L. Williams, 1999. Choline supplementation during prenatal development reduces proactive interference in spatial memory. Dev. Brain Res. 118:51–59. Meck, W. H., and C. L. Williams, 2003. Metabolic imprinting of choline by its availability during gestation: Implications for memory and attentional processing across the lifespan. Neurosci. Biobehav. Rev. 27:385–399. Mitchell, J. H., F. Nicol, G. J. Beckett, and J. R. Arthur, 1998. Selenoprotein expression and brain development in preweanling selenium and iodine deficient rats. J. Mol. Endocrinol. 20: 203–210. Morgane, P. J., R. Austin-LaFrance, J. Bronzino, J. Tonkiss, S. Diaz-Cintra, L. Cintra, T. Kemper, and J. R. Galler, 1993. Prenatal malnutrition and development of the brain. Neurosci. Biobehav. Rev. 17:91–128. Morrow-Tlucak, M., R. H. Haude, and C. B. Ernhart, 1988. Breastfeeding and cognitive development in the first 2 years of life. Soc. Sci. Med. 26:625–639. Neuringer, M., W. E. Connor, D. S. Lin, L. Barstad, and S. Luck, 1986. Biochemical and functional effects of prenatal and postnatal omega-3-fatty acid deficiency on retina and brain in rhesus monkeys. Proc. Natl. Acad. Sci. USA 83:4021– 4025. Niculescu, M. D., C. N. Craciunescu, and S. H. Zeisel, 2005. Gene expression profiling of choline-deprived neural precursor cells isolated from mouse brain. Mol. Brain Res. 134: 309–322. Niculescu, M. D., Y. Yamamuro, and S. H. Zeisel, 2004. Choline availability modulates human neuroblastoma cell proliferation and alters the methylation of the promoter region of the cyclindependent kinase inhibitor 3 gene. J. Neurochem. 89:1252– 1259. Nishijima, M., 1986. Somatomedin-C as a fetal growth promoting factor and amino acid composition of cord blood in Japanese neonates. J. Perinatal. Med. 14:163–166. Nokes, C., C. van den Bosch, and D. A. P. Bundy, 1998. Infants and young children (6–24 months). In A Report of the International Anemia Consultative Group: The Effects of Iron Deficiency and Anemia on Mental and Motor Performance, Educational Achievement and Behavior in Children: An Annotated Bibliography, 7–22. Washington, DC. O’Donnell, K., M. A. Rakeman, D. Zhi-Hong, C. Xue-Yi, Z. Y. Mei, N. DeLong, G. Brenner, M. Tai, W. Dong and G. R. DeLong, 2002. Effects of iodine supplementation during pregnancy on child growth and development at school age. Dev. Med. Child Neurol. 44:76–81. Ortiz, E., J. M. Pasquini, K. Thompson, B. Felt, G. Butkus, J. Beard, and J. R. Connor, 2004. Effect of manipulation of iron storage, transport, or availability on myelin composition and brain iron content in three different animal models. J. Neurosci. Res. 77:681–689. Pinero, D. J., B. C. Jones, and J. L. Beard, 2001. Variations in dietary iron alter behavior in developing rats. J. Nutr. 131: 311–318. Pleasure, J., and D. Pleasure, 2003. Trophic factor and nutritional and hormonal regulation of brain development. In R. A. Polin, W. W. Fox, and S. H. Abman, eds., Fetal and Neonatal Physiology, 1785–1792. Philadelphia: W. B. Saunders. Pollitt, E., 1996. Timing and vulnerability in research on malnutrition and cognition. Nutr. Rev. 54:S49–S55. Pollitt, E., and K. S. Gorman, 1994. Nutritional deficiencies as developmental risk factors. In C. A. Nelson, ed., Threats to Optimal Development: The Minnesota Symposia on Child Psychology, vol. 27, pp. 121–144. Hillsdale, NJ: Erlbaum.

fuglestad, rao, and georgieff: nutrition in cognitive development

639

Pollitt, E., K. S. Gorman, P. L. Engle, R. Martorell, and J. Rivera, 1993. Early supplementary feeding and cognition: Effects over two decades. Monogr. Soc. Res. Child Dev. 58: 1–99. Pollitt, E., K. S. Gorman, P. L. Engle, J. A. Rivera, and R. Martorell, 1995. Nutrition in early life and the fulfillment of intellectual potential. J. Nutr. 125:S1111–S1118. Pollitt, E., W. E. Watkins, and M. A. Husaini, 1997. Threemonth nutritional supplementation in Indonesian infants and toddlers benefits memory function 8 Y later. Am. J. Clin. Nutr. 66:1357–1363. Pyapali, G. K., D. A. Turner, C. L. Williams, W. H. Meck, and H. S. Swartzwelder, 1998. Prenatal dietary choline supplementation decreases the threshold for induction of long-term potentiation in young adult rats. J. Neurophysiol. 79:1790– 1796. Raman, L., M. K. Georgieff, and R. Rao, 2006. The role of chronic hypoxia in the development of neurocognitive abnormalities in preterm infants with bronchopulmonary dysplasia. Dev. Sci. 9:359–367. Rao, R., M. DeUngria, D. Sullivan, P. Wu, J. D. Wobken, C. A. Nelson, and M. K. Georgieff, 1999. Perinatal brain iron deficiency increases the vulnerability of rat hippocampus to hypoxic ischemic insult. J. Nutr. 129:199–206. Rao, R., I. Tkac, E. L. Townsend, R. Gruetter, and M. K. Georgieff, 2003. Perinatal iron deficiency alters the neurochemical profile of the developing rat hippocampus. J. Nutr. 133:3215–3221. Rioux, F. M., G. Lindmark, and O. Hernell, 2006. Does inadequate maternal iron or DHA status have a negative impact on an infant’s functional outcomes? Acta Paediatr. 95: 137–144. Robertson, K. D., and P. A. Jones, 2000. DNA methylation: Past, present and future directions. Carcinogenesis 21:461–467. Roncagliolo, M., M. Garrido, T. Walter, P. Peirano, and B. Lozoff, 1998. Evidence of altered central nervous system development in infants with iron deficiency anemia at 6 mo: Delayed maturation of auditory brainstem responses. Am. J. Clin. Nutr. 68:683–690. Sandstead, H. H., 1985. Zinc: Essentiality for brain development and function. Nutr. Rev. 43:129–137. Sandstead, H. H., J. G. Penland, N. W. Alcock, and H. H. Dayal, 1998. Effects of repletion with zinc and other micronutrients on neuropsychologic performance and growth of Chinese children. Am. J. Clin. Nutr. 68:470S–475S. Saneto, R. P., K. G. Low, M. H. Melner, and J. De Vellis, 1988. Insulin/insulin-like growth factor 1 and other epigenetic modulators of myelin basic protein expression in isolated oligodendrocyte progenitor cells. J. Neurosci. Res. 21:210–219. Siddappa, A. M., M. K. Georgieff, S. Wewerka, C. Worwa, C. A. Nelson, and R. A. DeRegnier, 2004. Iron deficiency alters auditory recognition memory in newborn infants of diabetic mothers. Pediatr. Res. 55:1034–1041. Skinner, J. D., B. R. Carruth, K. S. Houck, F. Coletta, R. Cotter, D. Ott, and M. McLeod, 1997. Longitudinal study of nutrient and food intakes of infants aged 2 to 24 months. J. Am. Diet. Assoc. 97:496–504. Smith, C. D., and K. B. Ain, 1995. Brain metabolism in hypothyroidism studied with 31P magnetic-resonance spectroscopy. Lancet 345:619–620. Spinillo, A., M. Stronati, A. Ometto, E. Fazzi, G. Lanzi, and S. Guaschino, 1993. Infant neurodevelopmental outcome in pregnancies complicated by gestational hypertension and

640

intrauterine growth retardation. J. Perinatol. Med. 21:195– 203. Stein, Z., and M. Susser, 1985. Effects of early nutrition on neurological and mental competence in human beings. Psychol. Med. 15:717–726. Strauss, R. S., and W. H. Dietz, 1998. Growth and development of term children born with low birth weight: Effects of genetic and environmental factors. J. Pediatr. 133:67–72. Tejwani, G., and S. Hanissian, 1990. Modulation of mu, delta and kappa opioid receptors in rat brain by metal ions and histidine. Neuropharmacology 29:445–452. Terhune, M. W., and H. H. Sandstead, 1972. Decreased RNA polymerase activity in mammalian zinc deficiency. Science 177: 68–69. Thelander, L., 1990. Ribonucleotide reductase. In P. Ponka, H. M. Schulman, and R. C. Woodworth, eds., Iron Transport and Storage, 193–200. Boca Raton, FL: CRC Press. Tolsa, C. B., S. Zimine, S. K. Warfield, M. Freschi, A. Sancho Rossignol, F. Lazeyras, S. Hanquinet, M. Pfizenmaier, and P. S. Huppi, 2004. Early alteration of structural and functional brain development in premature infants born with intrauterine growth restriction. Pediatr. Res. 56:132–138. Uauy, R. D., D. G. Birch, E. E. Birch, J. E. Tyson, and D. R. Hoffman, 1990. Effect of dietary omega-3 fatty acids on retinal function of very-low-birth-weight neonates. Pediatr. Res. 28: 485–492. Uauy-Dagach, R., and P. Mena, 1995. Nutritional role of omega3 fatty acids during the perinatal period. Clin. Perinatol. 22: 157–176. Volpe, J. J., 1995. Hypoglycemia and brain injury. In J. J. Volpe, ed., Neurology of the Newborn, 3rd ed., 467–489. Philadelphia: W. B. Saunders. Walter, T., J. Kowalskys, and A. Stekel, 1983. Effect of mild iron deficiency on infant mental developmental scores. J. Pediatr. 102:519–522. Wang, B., and J. Brand-Miller, 2003. The role and potential of sialic acid in human nutrition. Eur. J. Clin. Nutr. 57:1351–1369. Wang, B., P. McVeagh, P. Petocz, and J. Brand-Miller, 2003. Brain ganglioside and glycoprotein sialic acid in breastfed compared with formula-fed infants. Am. J. Clin. Nutr. 78: 1024–1029. Wang, Y. S. and S. Y. Wu, 1996. The effect of exclusive breastfeeding on development and incidence of infection in infants. J. Hum. Lact. 12:27–30. Watanabe, C., and H. Satoh, 1994. Brain selenium status and behavioral development in selenium-deficient preweanling mice. Physiol. Behav. 56:927–932. Waterland, R. A., and R. L. Jirtle, 2003. Transposable elements: Targets for early nutritional effects on epigenetic gene regulation. Mol. Cell. Biol. 23:5293–5300. Weaver, I. C., N. Cervoni, F. A. Champagne, A. C. D. Alessio, S. Sharma, J. R. Seckl, S. Dymov, M. Szyf, and M. J. Meaney, (2004). Epigenetic programming by maternal behavior. Nature Neurosci. 7:847–854. Wiggins, R. C., G. Fuller, and S. J. Enna, 1984. Undernutrition and the development of brain neurotransmitter systems. Life Sci. 35:2085–2094. Winer, E. K., and N. Tejani, 1994. Four- to seven-year evaluation in two groups of small-for-gestational-age infants. In N. Tejani, ed., Obstetrical Events and Developmental Sequelae, 2nd ed., 77–94. Boca Raton, FL: CRC Press. Winick, M., and A. Nobel, 1966. Cellular responses in rats during malnutrition at various ages. J. Nutr. 89:300–306.

neurodevelopmental aspects of clinical disorders

Winick, M., and P. Rosso, 1969a. Head circumference and cellular growth of the brain in normal and marasmic children. J. Pediatr. 74:774–778. Winick, M., and P. Rosso, 1969b. The effect of severe early malnutrition on cellular growth of the human brain. Pediatr. Res. 3:181–184. Yamada, K. A., N. Rensing, Y. Izumi, G. A. De Erausquin, V. Gazit, D. A. Dorsey, and D. G. Herrera, 2004. Repetitive hypoglycemia in young rats impairs hippocampal long-term potentiation. Pediatr. Res. 55:372–379. Yamamoto, N., M. Saitoh, A. Moriuchi, M. Nomura, and H. Okuyama, 1987. Effect of dietary alpha-linolenate/linoleate

balance on brain lipid compositions and learning ability of rats. J. Lipid Res. 28:144–151. Youdim, M. B. H., D. Ben-Sachar, and S. Yehuda, 1989. Putative biological mechanisms of the effect of iron deficiency on brain biochemistry and behavior. Am. J. Clin. Nutr. 50:607–612. Zeisel, S. H., 2004. Nutritional importance of choline for brain development. J. Am. Coll. Nutr. 23:621S–626S. Zimmermann, M. B., K. Connolly, M. Bozo, J. Bridson, F. Rohner, and L. Grimci, 2006. Iodine supplementation improves cognition in iodine-deficient schoolchildren in Albania: A randomized, controlled, double-blind study. Am. J. Clin. Nutr. 83:108–114.

fuglestad, rao, and georgieff: nutrition in cognitive development

641

39

Fetal Alcohol Syndrome SARAH N. MATTSON, SUSANNA L. FRYER, CHRISTIE L. MCGEE, AND EDWARD P. RILEY

The effects of heavy prenatal alcohol exposure on the developing fetus are devastating and permanent, yet preventable. Despite anecdotal descriptions throughout history associating maternal alcohol consumption during pregnancy with harmful consequences to the offspring, scientific recognition of the fetal alcohol syndrome (FAS) occurred only about 35 years ago (Jones and Smith, 1973; Jones et al., 1973; Lemoine et al., 1968). Estimates of the incidence rate of FAS in the United States range from 0.5 to 2.0 cases per 1,000 live births, with 1,000–6,000 new cases each year (Bertrand, Floyd, and Weber, 2005). International rates are often higher, with estimates in some locations as high as 7.4 percent of school-age children (May et al., 2000, 2006; Viljoen et al., 2005). Moreover, FAS is a major public health concern in the United States, given the personal and financial costs to the individual, the family, and society (Lupton, Burd, and Harwood, 2004). The diagnosis of FAS requires the presence of three distinct criteria: (1) a specific pattern of facial anomalies, (2) pre- and/or postnatal growth deficiency, and (3) some evidence of a central nervous system dysfunction. However, more individuals are adversely affected by prenatal alcohol exposure than those that meet the formal criteria for a diagnosis of FAS, and behavioral or cognitive symptoms may be present in the absence of the facial characteristics or growth deficiency seen in FAS (Mattson, Riley, Gramling, et al., 1998). The umbrella term “fetal alcohol spectrum disorders” (FASD) was adopted to describe the wide range of effects resulting from prenatal alcohol exposure, from FAS to subtle neurobehavioral, growth, or physical deficits (Bertrand et al., 2004). Using this classification, FAS is considered dysmorphic FASD (i.e., all three of the diagnostic criteria are met, including the characteristic facial pattern). Individuals who lack all the distinct characteristics of FAS but have a documented history of significant prenatal alcohol exposure and some physical or neurocognitive/ behavioral alteration thought to be related to that exposure are considered to be within the range of nondysmorphic FASD. It must be stressed that the term “FASD” is not a diagnostic term, but simply an acronym used to describe the range of outcomes following prenatal alcohol exposure.

Methodological considerations Convergent evidence from animal models of perinatal ethanol exposure and from observational human studies suggests a general dose-response relationship to alcohol’s teratogenic effects (Meyer and Riley, 1986). For example, FAS, which is associated with maternal alcohol abuse or dependence, falls at the severe end of the continuum of teratogenic effects. However, as discussed in the chapter’s introduction, neurobehavioral dysfunction can result from prenatal alcohol exposure independently of other FASrelated signs. The reason for this variation in outcome is unknown and may, in some cases, be attributable to aspects of exposure timing and dose. It is important to note that there are methodological barriers to precise examination of dose-response relationships in human studies of alcohol teratogenesis. Retrospective measures of exposure based on maternal report of alcohol intake generally have poor reliability (Jacobson et al., 2002). Postnatal identification of fetal alcohol exposure is not feasible because alcohol consumption does not produce long-lasting metabolites. By collecting information about maternal alcohol consumption at or near the time of birth, prospectively designed studies offer some mitigation of these methodological difficulties, but such studies still typically rely on subjective measures of fetal alcohol exposure, such as maternal report. Thus, in human studies of FASD, exact timing and dose of gestational alcohol exposure are usually not known. Another methodological barrier to consider is that attempts to establish a causal relationship between alcohol exposure and behavioral outcome must account for potentially confounding influences such as demographic characteristics of the fetus and mother, other teratogenic exposures, and quality of postnatal environment. Although animal models have clearly shown a causal relationship between alcohol and subsequent brain and behavioral changes, in humans such confounding influences may contribute to the brain-behavior profile observed in FASD.

Autopsy studies Initially the devastating effects of alcohol on brain structure were seen on autopsy of a small number of cases. These reports indicated that developing brains exposed to

643

alcohol suffer from a variety of structural insults including overall decreases in brain size and anomalies of certain brain structures, such as the corpus callosum, ventricles, and cerebellum (Clarren, 1986; Jones and Smith, 1973). A review of available autopsy examinations indicates extensive and diffuse damage. The variable and nonspecific autopsy findings led to the reasonable conclusion that no particular pattern of behavioral or intellectual functioning would be characteristic of individuals with FAS (Clarren, 1986). Yet neuropsychological studies of individuals with FAS suggested that there might be a characteristic pattern of cognitive and behavioral deficits (for review, see Mattson and Riley, 1998). One possible explanation for this discrepancy is that the autopsies were conducted primarily on the most severe cases of FAS, those that had abnormalities not compatible with life. Thus these cases may not have been representative of the larger population of individuals with FAS.

Neuroimaging studies Imaging technologies, such as magnetic resonance imaging (MRI), have enabled the examination of brains of living individuals exposed prenatally to alcohol. MRI studies provide a more representative sampling of exposed individuals than previously possible and suggest that although the overall brain size appears to be reduced, this decrease is not uniform. Examination of the lobes of the brain in individuals with FASD suggests that parietal regions (Archibald et al., 2001; Sowell, Thompson, et al., 2001; Sowell, Thompson, Mattson, et al., 2002; Sowell, Thompson, Peterson, et al., 2002) and portions of the frontal lobes (Sowell, Thompson, Mattson, et al., 2002) are disproportionately reduced compared to other lobes of the brain. Changes in brain shape and tissue distributions have also been reported in this population. Morphological studies have demonstrated narrowing of the parietal regions (Sowell, Thompson, Mattson, et al., 2002) and reduced cerebral asymmetry (Sowell, Thompson, Peterson, et al., 2002). Changes in the distribution of white and gray matter have been described in two studies from the same cohort. One study indicated that, within the cerebrum, white matter decreases are greater than those of gray matter (Archibald et al., 2001). Similarly, Sowell, Thompson, and colleagues (2001) found a relative increase in gray matter and a decrease in white matter in the perisylvian cortices, particularly in the left hemisphere. Because typical development is associated with a decrease in gray matter density, the authors speculated that increased gray matter density in the alcoholexposed brain may occur as a result of abnormalities in the myelination process (Sowell, Thompson, Mattson, et al., 2002). These results suggest that brain development contin-

644

ues to be affected long after the prenatal insult of alcohol exposure. Cerebellum Prenatal alcohol exposure is also associated with reductions in overall cerebellar volume (Archibald et al., 2001; Mattson, Jernigan, and Riley, 1994; Mattson and Riley, 1996), as well as alterations in the cerebellar vermis, or midline. For instance, using MRI techniques, Autti-Rämö and colleagues (2002) reported a size reduction in the vermis and found that this was a sensitive neuroanatomical marker of prenatal alcohol exposure (as documented by maternal report and measured blood and urine alcohol concentrations), occurring even in cases exposed only in the first trimester. Two other studies indicated that this reduction in the vermis was not uniform. Specifically, disproportionate decreases occurred in the anterior and posterior portions of the cerebellar vermis (O’Hare et al., 2005; Sowell et al., 1996). Abnormal spatial displacement was also noted in these regions of the vermis, with the anterior vermis being shifted inferiorly and posteriorly in the alcoholexposed group compared to unexposed controls (O’Hare et al., 2005). More recently, Bookstein and colleagues reported that a single multivariate summary score of cerebellar size and shape could distinguish between individuals with FASD and controls with about 75 percent accuracy (Bookstein et al., 2006). Cerebellar irregularities are of interest in this population because of the structure’s association with balance and motor coordination. Alcoholexposed children show impaired motor development, including disruptions of fine and gross motor skills and poor balance (Autti-Rämö and Granström, 1991; Barr et al., 1990; Roebuck, Mattson, and Riley, 1998), which may persist into adulthood in heavily exposed individuals (Connor et al., 2006). In addition, recent literature suggests a role for the cerebellum beyond motor coordination (Cabeza and Nyberg, 2000; Middleton and Strick, 2000), including involvement in attention (Akshoomoff, Courchesne, and Townsend, 1997; Allen et al., 1997) and classical conditioning (Woodruff-Pak et al., 2000; Woodruff-Pak, Papka, and Ivry, 1996), both of which are affected in FASD (e.g., Coffin et al., 2005; Coles et al., 2002). Interestingly, displacement of the anterior vermis correlates with verbal-learning and memory deficits in children and adolescents with FASD (O’Hare et al., 2005). Corpus Callosum Irregularities of the corpus callosum are among the most commonly reported alcohol-induced alterations in the developing brain. Perhaps the most striking finding is partial or complete agenesis of the corpus callosum, which has been reported by a number of groups studying alcohol-exposed individuals (Bhatara et al., 2002; Clark et al., 2000; V. Johnson et al., 1996; Riikonen et al., 1999; Riley et al., 1995; Swayze et al., 1997). In fact, it has been

neurodevelopmental aspects of clinical disorders

suggested that prenatal alcohol exposure is the most common cause of callosal agenesis (Jeret and Serur, 1991), although this anomaly remains a rare condition in FASD, even within the heavily exposed population. However, in the absence of agenesis, callosal size is reduced in alcohol-exposed individuals even after controlling for overall brain size. Specifically, midsagittal measurements of the corpus callosum taken from MRI images indicate that anterior and posterior callosal regions are reduced in children and adolescents prenatally exposed to alcohol (Riley et al., 1995; Sowell, Mattson, et al., 2001). Interestingly, a similar pattern of results has been reported in children with attention-deficit/hyperactivity disorder (ADHD) (Hynd et al., 1991), prompting some to speculate that callosal structural changes may be associated with the attention deficits noted in individuals with prenatal alcohol exposure. In addition to size decreases, morphological analyses indicate significant spatial displacement of the corpus callosum in alcohol-exposed individuals; posterior regions of this structure are located more anterior and inferior in alcohol-exposed participants than in controls (Sowell, Mattson, et al., 2001). Interestingly, this displacement was associated with verbal but not nonverbal learning ability. Other studies have related corpus callosum size or morphology to neuropsychological function in alcohol-exposed individuals, including inter-hemispheric transfer of somatosensory information (Roebuck, Mattson, and Riley, 2002) and patterns of executive and motor function (Bookstein et al., 2002). Finally, one recent study using diffusion tensor imaging in young adults revealed reduced integrity of white matter tracts in the corpus callosum, suggesting that this region is affected at the microstructural level (Ma et al., 2005). Basal Ganglia MRI studies of children and adolescents with heavy alcohol exposure also show a disproportionate reduction in the size of the basal ganglia (Mattson, Riley, Jernigan, et al., 1992; Mattson, Riley, Jernigan, et al., 1994; Mattson, Riley, Sowell, et al., 1996). When the size of individual structures that make up the basal ganglia were examined in relation to overall brain size, only the caudate nucleus was disproportionately reduced (Mattson, Riley, Sowell, et al., 1996). This result was replicated in a larger sample (Archibald et al., 2001). The caudate nucleus is known to have extensive connections to the frontal lobes, and it has been speculated that caudate volume reductions may lead to a disruption in frontal-subcortical circuitry. This disruption may underlie a number of neuropsychological deficits related to executive functions such as planning ability, concept formation, and fluency that are exhibited by individuals with histories of heavy prenatal alcohol exposure. Brain Metabolism Studies Researchers have only recently begun to assess functional brain changes in individuals with FASD. Three studies involving single-photon emission

computerized tomography (SPECT) or positron emission tomography (PET) have been conducted in individuals with FASD. The SPECT results indicated decreased cerebral blood flow in the left parieto-occipital region, abnormal functional symmetry with similar blood flow in the right and left frontal lobes (Riikonen et al., 1999), reduced serotonin in the medial frontal cortex, and an increase in striatal dopamine transporter binding (Riikonen et al., 2005). The PET study indicated subtle reductions in glucose metabolism in medial subcortical areas, including caudate and thalamic nuclei (Clark et al., 2000). Importantly, these brain regions, implicated by the PET and SPECT studies, have also been implicated in structural studies, as described previously, and these data suggest that structural differences may relate to metabolic changes in the brains of individuals with FASD. Recently, brain metabolism in 10 adolescents and young adults prenatally exposed to alcohol was compared to matched controls using magnetic resonance spectroscopy (MRS). Metabolite ratios NAA/Cho, NAA/Cr, and Cho/ Cr and absolute metabolite concentrations were calculated for several anatomic regions. Prenatal alcohol exposure was associated with altered brain metabolism in several brain areas. Relative to controls, participants with FASD had lower NAA/Cho and/or NAA/Cr ratios in parietal and lateral frontal cortices, as well as within the frontal white matter, corpus callosum, thalamus, and cerebellar dentate nucleus. Importantly, these changes are in accordance with previous findings discussed earlier. It appears these brain metabolite ratio values were driven by Cr and Cho levels, rather than NAA, as there was an increase in the absolute intensity of the glial markers Cho and Cr but no change in the neuronal marker NAA. Therefore, the authors speculated that these results may reflect changes in the glial cell pool rather than in neurons (Fagerlund et al., in press). Another MRS study, focused on the caudate nucleus, reported preliminary findings on a small sample of children with and without prenatal alcohol exposure (Cortese et al., 2006). In addition to bilateral caudate nucleus volume reductions, alcohol-exposed children had higher NAA/Cr ratios in these regions. Though the precise role of these metabolites is not fully understood, the NAA/Cr ratio is thought to reflect the integrity of neuronal functioning. Interestingly, in this small sample, the increase in the NAA/Cr ratio was driven by increased amounts of absolute NAA concentrations. Thus these findings, though preliminary, suggest that neuronal functioning may be compromised in striatal areas of individuals with FASD. Functional MRI Studies Three recent studies have utilized functional magnetic resonance imaging (fMRI) to examine the hemodynamic brain response in adults and children with FASD. The first study involved performance on low- and high-load working-memory tasks. Within-group

mattson et al.: fetal alcohol syndrome

645

activation maps indicated increased activation for low-level working-memory tasks in the inferior and middle frontal cortices in alcohol-exposed participants but not in controls, although no between-group comparisons were conducted (Malisza et al., 2005). An accompanying commentary (Bookheimer and Sowell, 2005) suggested that these findings are difficult to interpret in the context of the participants’ ability to perform the task, as the alcohol-exposed participants showed increased activation but were also less accurate and/ or had slower reaction times than controls. Another fMRI study of alcohol-exposed children examined activation during a response-inhibition task and found that individuals with FASD had alterations in the frontal and subcortical regions during trials requiring inhibition of action in comparison to controls. These differences were exhibited despite comparable behavioral performance between the two groups. Specifically, increased frontal and decreased caudate activation were noted in the alcohol-exposed group (Fryer et al., 2006). These areas are consistent with frontalstriatal anatomical connections thought to be involved in response inhibition (Middleton and Strick, 2000). Because task performance was comparable between the groups, the accompanying activation differences were likely taskspecific. The authors speculate that these results may indicate that the frontal-subcortical circuitry thought to mediate inhibitory control is sensitive to prenatal alcohol exposure. Finally, the third fMRI study examined activation during a verbal paired-associate task. Results indicated less activation in the left medial and posterior temporal regions and more activation in the right dorsal frontal cortex in the FASD group relative to controls during task performance. The authors speculated that the increased activation in the frontal systems may have been a compensatory mechanism for dysfunctional medial temporal memory systems in the FASD group (Sowell et al., 2007).

Neuropsychological studies Neuropsychological studies have attempted to identify the specific strengths and weaknesses of children exposed prenatally to alcohol, with the goal of defining a profile of core cognitive or behavioral deficits and strengths. This important task has the potential to lead to improvement in the diagnosis and treatment of individuals exposed to alcohol prenatally, regardless of facial dysmorphology. The neuropsychological literature on prenatal alcohol exposure indicates impairment of general cognitive functioning, along with specific deficits of learning, memory, language, attention, executive functioning, and visual-spatial and motor skills. Intellectual Function Generally, prenatal alcohol exposure leads to decreases in IQ, with the mean IQ score

646

of FAS children falling in the low 70s (Mattson and Riley, 1998). Fetal alcohol syndrome is even cited as the most common preventable cause of mental retardation in the Western world (Abel and Sokol, 1987). However, it should be noted that the majority of individuals with FAS are not mentally retarded, as only about 25 percent of individuals with FAS have IQ scores that fall below 70 (Streissguth et al., 1996). Moreover, IQ scores among alcohol-exposed children are extremely variable; for instance, a review of available reports cited a range of scores from 20 to 120 (Mattson and Riley, 1998). As stated earlier, neurocognitive deficits, including mental retardation, may be present in alcohol-exposed individuals with or without the presence of the facial abnormalities required for a diagnosis of FAS. However, the severity of these facial anomalies appears to relate to intellectual functioning, with those showing the clearest physical signs of FAS having the lowest IQ scores (e.g., Mattson, Riley, Gramling, et al., 1997). Learning and Memory Given the evidence of intellectual impairments, it is not surprising that children prenatally exposed to alcohol have decreased academic achievement and an increased rate of learning problems (Howell et al., 2006; Streissguth, Barr, and Sampson, 1990). These deficits may relate to impairments in verbal and nonverbal learning and memory that have been reported repeatedly, although deficits in memory may not be as pervasive as originally thought. There is evidence that long-term retention may be preserved in alcohol-exposed individuals, suggesting that once information is learned, it will likely be retained, at least in the verbal domain (Mattson, Riley, Delis, et al., 1996; Mattson, Riley, Gramling, et al., 1998; Mattson and Roebuck, 2002; Willford et al., 2004). These memory and learning deficits may result from problems with response inhibition and the initial encoding of verbal information. Specifically, in one study, children with FAS showed impaired ability to learn a word list, and although retention was intact, they made more errors of intrusion, perseveration, and false-positive responding (Mattson, Riley, Delis, et al., 1996). In contrast to verbal learning and memory, nonverbal performance is relatively less studied, and results have been conflicting. Some studies suggest that, as in the verbal domain, alcohol-exposed children benefit from repeated exposure to nonverbal information, but demonstrate both encoding and retrieval deficits (Mattson and Roebuck, 2002; Roebuck-Spencer and Mattson, 2004). In contrast, other studies of visual learning suggest a pattern of spared retention, as seen in the verbal domain (Kaemingk, Mulvaney, and Tanner Halverson, 2003). Similar inconsistencies are reported within the nonverbal memory domain, with some studies demonstrating intact object recall but impaired spatial location recall (Uecker and Nadel, 1996, 1998) and

neurodevelopmental aspects of clinical disorders

others suggesting that spatial recall is comparable to controls once basic perceptual and verbal memory skills are taken into account (Kaemingk and Halverson, 2000). Finally, a recent small study indicated that place learning, a type of spatial learning associated with the hippocampus, is impaired in children with FAS (Hamilton et al., 2003). This finding is consistent with previous animal studies (e.g., T. Johnson and Goodlett, 2002) and was recently replicated in a large international study of children with FASD (Mattson et al., 2006). Visual-Spatial Function In addition to possible alterations in nonverbal memory functioning, individuals prenatally exposed to alcohol display deficits in visual-spatial processing. When presented with hierarchical stimuli consisting of large, global figures made up of smaller, local figures, children with FAS showed local but not global processing deficits (Mattson, Gramling, Delis, et al., 1996). Local processing difficulties were independent of memory function and construction ability. This impairment in local feature processing may be a result of poor attention shifting, a disruption in frontal-subcortical pathways, and/or a pattern of brain abnormality involving the left hemisphere to a greater degree than the right hemisphere, all of which are supported by studies of brain structure in this population (e.g., Archibald et al., 2001; Sowell, Thompson, et al., 2001). In addition to the higher-order visual-spatial processing dysfunction tapped by the global-local task, prenatal alcohol exposure is associated with more basic visual-spatial disabilities, as evidenced by impairment on simple drawing tasks (Conry, 1990; Janzen, Nanson, and Block, 1995; Mattson, Riley, Gramling, et al., 1998). Such impairment may relate to observed disproportionate parietal lobe volume reductions, as discussed previously. Attention Hyperactive disorders and attention-deficit/ hyperactivity disorder (ADHD) are among the most frequent psychiatric diagnoses in individuals with FASD (Fryer, McGee, et al., in press; Steinhausen, Willms, and Spohr, 1993). One recent study reviewed the charts of 2,231 youth referred for FASD, divided into four groups by risk for gestational alcohol exposure. Those at the highest risk for exposure had the highest rates of ADHD, whereas those with no risk of exposure had the lowest rates (Bhatara, Loudenberg, and Ellis, 2006). Attention deficits and/or hyperactivity are associated with prenatal alcohol exposure regardless of the degree of cognitive impairment. One study indicated that commonly used measures of attention can be used to accurately distinguish children with heavy prenatal alcohol exposure from controls with a high degree of accuracy, even after controlling for IQ (Lee, Mattson, and Riley, 2004). A more recent, although preliminary, study indicated that alcohol-exposed children can also be

distinguished from individuals with ADHD using the same measures (Mattson, Vaurio, and Riley, 2007). Additional studies directly comparing children with FASD to children with ADHD are sparse. However, although there are some inconsistencies (Coles et al., 1997; Nanson and Hiscock, 1990), it appears that these two groups can be distinguished based on measures of vigilance and visual-spatial processing (Coles et al., 1997), and eyeblink conditioning (Coffin et al., 2005). Additional, albeit preliminary, studies suggest that these two groups can be distinguished on the basis of verbal learning and memory, executive function, and adaptive behavior (Crocker, Riley, and Mattson, 2007; Mattson, Lee, et al., 2005; Mattson, Vaurio, and Riley, 2006). Two studies have examined visual and auditory attention in children with prenatal alcohol exposure (Coles et al., 2002; Mattson, Calarco, and Lang, 2006). In both studies, visual attention deficits were more pervasive than auditory attention deficits, and in one study (Mattson, Calarco, and Lang, 2006), auditory attention deficits were only evident when task demands increased. These findings suggest that some attentional systems may be more vulnerable to the effects of prenatal alcohol exposure than others. Executive Function Individuals with heavy prenatal alcohol exposure have been shown to have executive function deficits (Kodituwakku et al., 1995; Mattson and Riley, 1999; Rasmussen, 2005; Schonfeld et al., 2001). Specific deficits have been noted in cognitive flexibility, response inhibition, planning and concept formation, reasoning, and verbal and nonverbal fluency. In most cases, executive function deficits were beyond what were expected based on deficiencies of more basic component skills (Mattson and Riley, 1999). One aspect of executive function that has not been well studied in FASD is working memory (for review see Rasmussen, 2005). The studies that have been conducted have conflicting results (Burden et al., 2005; Kodituwakku et al., 1995) or suggest that basic perceptual skills may account for apparent working-memory deficits (Vaurio et al., 2006). The relationship between executive function and IQ in this population is unclear. One study of adults found that executive function deficits were worse than predicted by IQ alone, suggesting that executive function deficits are independent of diminished intellectual ability (Connor et al., 2000). In contrast, other studies find that executive function ability in alcohol-exposed participants is either not significantly associated with IQ (Mattson, Goodman, et al., 1999) or is better than would be predicted by IQ (Kane et al., 2003; Mattson, Roebuck, and Riley, 1996; Mattson, Vaurio, and Riley, 2006; McGee, Schonfeld, submitted), although this finding may be limited to samples with lower IQ scores. These changes in executive function may relate to alcohol-induced brain changes. For example, as discussed

mattson et al.: fetal alcohol syndrome

647

previously, caudate nucleus abnormalities and anterior and orbital frontal cortical shape changes occur following such exposure, and these brain areas are linked to executive function. More specifically, a preliminary study suggested that caudate volume reductions noted in alcohol-exposed individuals may relate to executive dysfunction. In this study, caudate volume predicted performance of alcohol-exposed children on several tests of executive function, including cognitive flexibility as measured by perseverative errors on the Wisconsin Card Sorting Test (Mattson, Riley, Archibald, and Jernigan, 2001). Speech and Language Language functioning is another relatively unstudied functional domain in FASD. A small number of studies demonstrate both speech and languageprocessing dysfunction in heavily exposed populations (Church et al., 1997; Conry, 1990; Janzen, Nanson, and Block, 1995). One recent study indicated that children with heavy prenatal alcohol exposure had poorer language skills than controls, although both groups had better receptive than expressive language abilities. However, the observed deficits were no greater than predicted by overall IQ scores (McGee, Bjorkquist, et al., 2007). Motor Function Early reports of motor dysfunction in FAS described delayed motor development and fine and gross motor deficits such as tremors, weak grasp, and poor hand/eye coordination (Jones et al., 1973). The majority of studies investigating motor abilities have supported these early descriptions (reviewed in Mattson and Riley, 1998). Several studies have supported effects of prenatal alcohol exposure on gait and balance (Barr et al., 1990; Roebuck, Simmons, Mattson, and Riley, 1998; Roebuck, Simmons, Richardson, et al., 1998). One study, comparing postural balance in alcohol-exposed children and matched nonexposed controls under conditions of systematic variations of visual and somatosensory information, indicated that alcohol-exposed individuals were overly reliant on the somatosensory system in maintaining balance (Roebuck, Simmons, Richardson, et al., 1998). In addition, motor deficits appear to persist to adulthood in heavily exposed populations (Connor et al., 2006). These results are consistent with studies involving animal models, which show an association of motor deficits with prenatal alcohol exposure and suggest, as mentioned previously, that these deficits may be associated with cerebellar damage (e.g., Goodlett, Thomas, and West, 1991; Meyer, Kotch, and Riley, 1990).

Future research directions Although 35 years of research have documented numerous impairments of brain structure and function in individuals with histories of prenatal alcohol exposure, much remains

648

to be done. For example, individual studies documenting specific strengths or weaknesses could be examined to determine a neurobehavioral profile or phenotype that will improve both identification and treatment of alcohol-affected individuals. Research involving multiple measures and multiple sites is under way to further this aim. Such multisite collaborations, including international or cross-cultural samples, are a promising avenue of FASD research (Riley et al., 2003) and should serve to strengthen the generalizability of any hypothesized neurobehavioral phenotype. Additional future research directions include increased use of functional imaging techniques to further specify brain changes in alcohol-exposed individuals and to discern the importance of noted neuroanatomical changes. Also, computer-assisted diagnosis of FAS utilizing novel facial dysmorphology technology and examination of risk factors including genetic susceptibility and alcohol dose and exposure pattern are expanding areas of FASD research.

Conclusion Prenatal alcohol exposure can lead to permanent changes in the brain that impact the behavior and cognition of exposed individuals. The consequences for affected individuals, their families, and society are grave, yet this spectrum of disorders is entirely preventable. Currently, research is focusing on defining the neurobehavioral phenotype associated with heavy prenatal alcohol exposure. This phenotype, if it exists, is likely to be continuous rather than categorical in nature, reflecting the continuum of effects seen following prenatal alcohol exposure. Further delineation of the brain-behavior relationship specific to individuals exposed to alcohol prenatally can inform the diagnosis and treatment strategies that attempt to mitigate the teratogenic effects of alcohol. acknowledgments

Preparation of this chapter was supported in part by National Institute on Alcohol Abuse and Alcoholism Grant numbers R01 AA010417, R01 AA010820, T32 AA013525, F31 AA016051, and F31 AA016047. We gratefully acknowledge the assistance and support of the Center for Behavioral Teratology, San Diego State University.

REFERENCES Abel, E. L., and R. J. Sokol, 1987. Incidence of fetal alcohol syndrome and economic impact of FAS-related anomalies. Drug Alcohol Depend. 19:51–70. Akshoomoff, N. A., E. Courchesne, and J. P. Townsend, 1997. Attention coordination and anticipatory control. Int. Rev. Neurobiol. 41:575–598. Allen, G., R. B. Buxton, E. C. Wong, and E. Courchesne, 1997. Attentional activation of the cerebellum independent of motor involvement. Science 275:1940–1943.

neurodevelopmental aspects of clinical disorders

Archibald, S. L., C. Fennema-Notestine, A. Gamst, E. P. Riley, S. N. Mattson, and T. L. Jernigan, 2001. Brain dysmorphology in individuals with severe prenatal alcohol exposure. Dev. Med. Child Neurol. 43:148–154. Autti-Rämö, I., T. Autti, M. Korkman, S. Kettunen, O. Salonen, and L. Valanne, 2002. MRI findings in children with school problems who had been exposed prenatally to alcohol. Dev. Med. Child Neurol. 44:98–106. Autti-Rämö, I., and M.-L. Granström, 1991. The psychomotor development during the first year of life of infants exposed to intrauterine alcohol of various duration. Neuropediatrics 22:59–64. Barr, H. M., A. P. Streissguth, B. L. Darby, and P. D. Sampson, 1990. Prenatal exposure to alcohol, caffeine, tobacco, and aspirin: Effects on fine and gross motor performance in 4-yearold children. Dev. Psychol. 26:339–348. Bertrand, J., R. L. Floyd, and M. K. Weber, 2005. Guidelines for identifying and referring persons with fetal alcohol syndrome. MMWR Morb. Mortal. Wkly. Rep. Recommendations and Reports 54:1–14. Bertrand, J., R. L. Floyd, M. K. Weber, M. O’Connor, E. P. Riley, K. A. Johnson, and D. E. Cohen, 2004. National Task Force on FAS/FAE: Guidelines for Referral and Diagnosis. Atlanta: Centers for Disease Control and Prevention. Bhatara, V., R. Loudenberg, and R. Ellis, 2006. Association of attention deficit hyperactivity disorder and gestational alcohol exposure: An exploratory study. J. Attention Disord. 9:515– 522. Bhatara, V. S., F. Lovrein, J. Kirkeby, V. Swayze II, E. Unruh, and V. Johnson, 2002. Brain function in fetal alcohol syndrome assessed by single photon emission computed tomography. SD J. Med. 55:59–62. Bookheimer, S. Y., and E. R. Sowell, 2005. Brain imaging in FAS: Commentary on the article by Malisza et al. Pediatr. Res. 58:1148–1149. Bookstein, F. L., A. P. Streissguth, P. D. Connor, and P. D. Sampson, 2006. Damage to the human cerebellum from prenatal alcohol exposure: The anatomy of a simple biometrical explanation. Anat. Rec. B New Anatomist 289B:195– 209. Bookstein, F. L., A. P. Streissguth, P. D. Sampson, P. D. Connor, and H. M. Barr, 2002. Corpus callosum shape and neuropsychological deficits in adult males with heavy fetal alcohol exposure. NeuroImage 15:233–251. Burden, M. J., S. W. Jacobson, R. J. Sokol, and J. L. Jacobson, 2005. Effects of prenatal alcohol exposure on attention and working memory at 7.5 years of age. Alcohol. Clin. Exp. Res. 29:443–452. Cabeza, R., and L. Nyberg, 2000. Imaging cognition. II. An empirical review of 275 PET and fMRI studies. J. Cogn. Neurosci. 12:1–47. Church, M. W., F. Eldis, B. W. Blakley, and E. V. Bawle, 1997. Hearing, language, speech, vestibular, and dentofacial disorders in fetal alcohol syndrome. Alcohol. Clin. Exp. Res. 21: 227–237. Clark, C. M., D. Li, J. Conry, R. Conry, and C. Loock, 2000. Structural and functional brain integrity of fetal alcohol syndrome in nonretarded cases. Pediatrics 105:1096–1099. Clarren, S. K., 1986. Neuropathology in fetal alcohol syndrome. In J. R. West, ed., Alcohol and Brain Development, 158–166. New York: Oxford University Press. Coffin, J. M., S. Baroody, K. Schneider, and J. O’Neill, 2005. Impaired cerebellar learning in children with prenatal alcohol

exposure: A comparative study of eyeblink conditioning in children with ADHD and dyslexia. Cortex 41:389–398. Coles, C. D., K. A. Platzman, M. E. Lynch, and D. Freides, 2002. Auditory and visual sustained attention in adolescents prenatally exposed to alcohol. Alcohol. Clin. Exp. Res. 26: 263–271. Coles, C. D., K. A. Platzman, C. L. Raskind-Hood, R. T. Brown, A. Falek, and I. E. Smith, 1997. A comparison of children affected by prenatal alcohol exposure and attention deficit, hyperactivity disorder. Alcohol. Clin. Exp. Res. 21:150–161. Connor, P. D., P. D. Sampson, F. L. Bookstein, H. M. Barr, and A. P. Streissguth, 2000. Direct and indirect effects of prenatal alcohol damage on executive function. Dev. Neuropsychol. 18:331–354. Connor, P. D., P. D. Sampson, A. P. Streissguth, F. L. Bookstein, and H. M. Barr, 2006. Effects of prenatal alcohol exposure on fine motor coordination and balance: A study of two adult samples. Neuropsychologia 44:744–751. Conry, J., 1990. Neuropsychological deficits in fetal alcohol syndrome and fetal alcohol effects. Alcohol. Clin. Exp. Res. 14:650–655. Cortese, B. M., G. J. Moore, B. A. Bailey, S. W. Jacobson, V. Delaney-Black, and J. H. Hannigan, 2006. Magnetic resonance and spectroscopic imaging in prenatal alcohol-exposed children: Preliminary findings in the caudate nucleus. Neurotoxicol. Teratol. 28:597–606. Crocker, N. A., E. P. Riley, and S. N. Mattson, 2007. Comparison of verbal learning and memory in children with ADHD with or without heavy prenatal alcohol exposure. Presented at International Neuropsychological Society meeting, Portland, February. Fagerlund, Å., S. Heikkinen, I. Autti-Rämö, M. Korkman, M. Timonen, T. Kuusi, E. P. Riley, and N. Lundbom, 2006. Brain metabolic alterations in adolescents and young adults with fetal alcohol spectrum disorders. Alcohol. Clin. Exp. Res. 30:2097– 2104. Fryer, S. L., C. L. McGee, G. E. Matt, E. P. Riley, and S. N. Mattson, 2007. Evaluation of psychopathological conditions in children with heavy prenatal alcohol exposure. Pediatrics 119: e733–e741. Fryer, S. L., S. F. Tapert, S. N. Mattson, M. P. Paulus, A. D. Spadoni, and E. P. Riley, 2007. Prenatal alcohol exposure affects frontal-striatal BOLD response during inhibitory control. Alcohol. Clin. Exp. Res. 31:1415–1424. Goodlett, C. R., J. D. Thomas, and J. R. West, 1991. Long-term deficits in cerebellar growth and rotarod performance of rats following “binge-like” alcohol exposure during the neonatal brain growth spurt. Neurotoxicol. Teratol. 13:69–74. Hamilton, D. A., P. Kodituwakku, R. J. Sutherland, and D. D. Savage, 2003. Children with fetal alcohol syndrome are impaired at place learning but not cued-navigation in a virtual Morris water task. Behav. Brain Res. 143:85–94. Howell, K. K., M. E. Lynch, K. A. Platzman, G. H. Smith, and C. D. Coles, 2006. Prenatal alcohol exposure and ability, academic achievement, and school functioning in adolescence: A longitudinal follow-up. J. Pediatr. Psychol. 31:116–126. Hynd, G. W., M. Semrud-Clikeman, A. R. Lorys, E. S. Novey, D. Eliopulos, and H. Lyytinen, 1991. Corpus callosum morphology in attention deficit-hyperactivity disorder: Morphometric analysis of MRI. J. Learn. Disabil. 24:141– 146. Jacobson, S. W., L. M. Chiodo, R. J. Sokol, and J. L. Jacobson, 2002. Validity of maternal report of prenatal alcohol, cocaine,

mattson et al.: fetal alcohol syndrome

649

and smoking in relation to neurobehavioral outcome. Pediatrics 109:815–825. Janzen, L. A., J. L. Nanson, and G. W. Block, 1995. Neuropsychological evaluation of preschoolers with fetal alcohol syndrome. Neurotoxicol. Teratol. 17:273–279. Jeret, J. S., and D. Serur, 1991. Fetal alcohol syndrome in adolescents and adults. JAMA 266:1077. Johnson, T. B., and C. R. Goodlett, 2002. Selective and enduring deficits in spatial learning after limited neonatal binge alcohol exposure in male rats. Alcohol. Clin. Exp. Res. 26: 83–93. Johnson, V. P., V. W. Swayze II , Y. Sato, and N. C. Andreasen, 1996. Fetal alcohol syndrome: Craniofacial and central nervous system manifestations. Am. J. Med. Genet. 61:329– 339. Jones, K. L., and D. W. Smith, 1973. Recognition of the fetal alcohol syndrome in early infancy. Lancet 2:999–1001. Jones, K. L., D. W. Smith, C. N. Ulleland, and A. P. Streissguth, 1973. Pattern of malformation in offspring of chronic alcoholic mothers. Lancet 1:1267–1271. Kaemingk, K. L., and P. T. Halverson, 2000. Spatial memory following prenatal alcohol exposure: More than a material specific memory deficit. Child Neuropsychol. 6:115–128. Kaemingk, K. L., S. Mulvaney, and P. T. Halverson, 2003. Learning following prenatal alcohol exposure: Performance on verbal and visual multitrial tasks. Arch. Clin. Neuropsychol. 18: 33–47. Kane, A., S. N. Mattson, R. Klopp, C. D. Coles, G. Marintcheva, and E. P. Riley, 2003. Relationship of IQ and WCST performance in children with fetal alcohol syndrome. J. Int. Neuropsychol. Soc. 9:202. Kodituwakku, P. W., N. S. Handmaker, S. K. Cutler, E. K. Weathersby, and S. D. Handmaker, 1995. Specific impairments in self-regulation in children exposed to alcohol prenatally. Alcohol. Clin. Exp. Res. 19:1558–1564. Lee, K. T., S. N. Mattson, and E. P. Riley, 2004. Classifying children with heavy prenatal alcohol exposure using measures of attention. J. Int. Neuropsychol. Soc. 10:271–277. Lemoine, P., H. Harousseau, J.-P. Borteyru, and J.-C. Menuet, 1968. Les enfants de parents alcooliques: Anomalies observées, à propos de 127 cas [Children of alcoholic parents: Abnormalities observed in 127 cases]. Ouest Med. 21: 476–482. Lupton, C., L. Burd, and R. Harwood, 2004. Cost of fetal alcohol spectrum disorders. Am. J. Med. Genet. (Sem. Med. Genet.) 127: 42–50. Ma, X., C. D. Coles, M. E. Lynch, S. M. LaConte, O. Zurkiya, D. Wang, and X. Hu, 2005. Evaluation of corpus callosum anisotropy in young adults with fetal alcohol syndrome according to diffusion tensor imaging. Alcohol. Clin. Exp. Res. 29:1214– 1222. Malisza, K. L., A.-A. Allman, D. Shiloff, L. Jakobson, S. Longstaffe, and A. E. Chudley, 2005. Evaluation of spatial working memory function in children and adults with fetal alcohol spectrum disorders: A functional magnetic resonance imaging study. Pediatr. Res. 58:1150–1157. Mattson, S. N., K. E. Calarco, and A. R. Lang, 2006. Focused and shifting attention in children with heavy prenatal alcohol exposure. Neuropsychology 20:361–369. Mattson, S. N., A. M. Goodman, C. Caine, D. C. Delis, and E. P. Riley, 1999. Executive functioning in children with heavy prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 23:1808– 1815.

650

Mattson, S. N., L. Gramling, D. Delis, K. L. Jones, and E. P. Riley, 1996. Global-local processing in children prenatally exposed to alcohol. Child Neuropsychol. 2:165–175. Mattson, S. N., T. L. Jernigan, and E. P. Riley, 1994. MRI and prenatal alcohol exposure: Images provide insight into FAS. Alcohol Health Res. World 18:49–52. Mattson, S. N., K. T. Lee, K. Hayden, and E. P. Riley, 2005. Comparison of adaptive behavior deficits in children with fetal alcohol spectrum disorders and/or ADHD. Alcohol. Clin. Exp. Res. 29:46A. Mattson, S. N., and E. P. Riley, 1996. Brain anomalies in fetal alcohol syndrome. In E. L. Abel, ed., Fetal Alcohol Syndrome: From Mechanism to Prevention, 51–68. Boca Raton, FL: CRC Press. Mattson, S. N., and E. P. Riley, 1998. A review of the neurobehavioral deficits in children with fetal alcohol syndrome or prenatal exposure to alcohol. Alcohol. Clin. Exp. Res. 22: 279–294. Mattson, S. N., and E. P. Riley, 1999. Implicit and explicit memory functioning in children with heavy prenatal alcohol exposure. J. Int. Neuropsychol. Soc. 5:462–471. Mattson, S. N., E. P. Riley, S. A. Archibald, and T. L. Jernigan, 2001. Caudate volume predicts CVLT-C and WCST performance in children with heavy prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 25:74A. Mattson, S. N., E. P. Riley, I. Autti-Rämö, P. A. May, V. Konovalova, K. L. Jones, et al., submitted. Spatial learning and navigation deficits in an international sample of children with heavy prenatal alcohol exposure. Mattson, S. N., E. P. Riley, D. C. Delis, C. Stern, and K. L. Jones, 1996. Verbal learning and memory in children with fetal alcohol syndrome. Alcohol. Clin. Exp. Res. 20:810–816. Mattson, S. N., E. P. Riley, L. J. Gramling, D. C. Delis, and K. L. Jones, 1997. Heavy prenatal alcohol exposure with or without physical features of fetal alcohol syndrome leads to IQ deficits. J. Pediatr. 131:718–721. Mattson, S. N., E. P. Riley, L. J. Gramling, D. C. Delis, and K. L. Jones, 1998. Neuropsychological comparison of alcoholexposed children with or without physical features of fetal alcohol syndrome. Neuropsychology 12:146–153. Mattson, S. N., E. P. Riley, T. L. Jernigan, C. L. Ehlers, D. C. Delis, K. L. Jones, et al., 1992. Fetal alcohol syndrome: A case report of neuropsychological, MRI and EEG assessment of two children. Alcohol. Clin. Exp. Res. 16:1001–1003. Mattson, S. N., E. P. Riley, T. L. Jernigan, A. Garcia, W. M. Kaneko, C. L. Ehlers, and K. L. Jones, 1994. A decrease in the size of the basal ganglia following prenatal alcohol exposure: A preliminary report. Neurotoxicol. Teratol. 16:283– 289. Mattson, S. N., E. P. Riley, E. R. Sowell, T. L. Jernigan, D. F. Sobel, and K. L. Jones, 1996. A decrease in the size of the basal ganglia in children with fetal alcohol syndrome. Alcohol. Clin. Exp. Res. 20:1088–1093. Mattson, S. N., and T. M. Roebuck, 2002. Acquisition and retention of verbal and nonverbal information in children with heavy prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 26: 875–882. Mattson, S. N., T. M. Roebuck, and E. P. Riley, 1996. Wisconsin card sorting performance in children prenatally exposed to alcohol. Alcohol. Clin. Exp. Res. 20:74A. Mattson, S. N., L. Vaurio, and E. P. Riley, 2006. Differences in executive functioning in children with fetal alcohol spectrum disorders and/or ADHD. Presented at the International Neuropsychological Society meeting, Boston, February.

neurodevelopmental aspects of clinical disorders

Mattson, S. N., L. Vaurio, and E. P. Riley, 2007. Can children with FASD be distinguished from children with ADHD using measures of attention? Presented at International Neuropsychological Society meeting, Portland, February. May, P. A., L. Brooke, J. P. Gossage, J. Croxford, C. Adnams, K. L. Jones, et al., 2000. Epidemiology of fetal alcohol syndrome in a South African community in the Western Cape Province. Am. J. Public Health 90:1905–1912. May, P. A., D. Fiorentino, J. P. Gossage, W. O. Kalberg, H. E. Hoyme, L. K. Robinson, et al., 2006. Epidemiology of FASD in a province in Italy: Prevalence and characteristics of children in a random sample of schools. Alcohol. Clin. Exp. Res. 30: 1562–1575. McGee, C. L., O. A. Bjorkquist, E. P. Riley, and S. N. Mattson, 2007. Language performance in young children with heavy prenatal alcohol exposure. Presented at International Neuropsychological Society meeting, Portland, February. McGee, C. L., A. M. Schonfeld, T. M. Roebuck, E. P. Riley, and S. N. Mattson, submitted. Problem solving in children with heavy prenatal alcohol exposure. Meyer, L. S., L. E. Kotch, and E. P. Riley, 1990. Alterations in gait following ethanol exposure during the brain growth spurt in rats. Alcohol. Clin. Exp. Res. 14:23–27. Meyer, L. S., and E. P. Riley, 1986. Behavioral teratology of alcohol. In E. P. Riley and C. V. Vorhees, eds., Handbook of Behavioral Teratology, 101–140. New York: Plenum Press. Middleton, F. A., and P. L. Strick, 2000. Basal ganglia and cerebellar loops: Motor and cognitive circuits. Brain Res. Rev. 31:236–250. Nanson, J. L., and M. Hiscock, 1990. Attention deficits in children exposed to alcohol prenatally. Alcohol. Clin. Exp. Res. 14:656– 661. O’Hare, E. D., E. Kan, J. Yoshii, S. N. Mattson, E. P. Riley, P. M. Thompson, et al., 2005. Mapping cerebellar vermal morphology and cognitive correlates in prenatal alcohol exposure. NeuroReport 16:1285–1290. Rasmussen, C., 2005. Executive functioning and working memory in fetal alcohol spectrum disorder. Alcohol. Clin. Exp. Res. 29: 1359–1367. Riikonen, R., I. Salonen, K. Partanen, and S. Verho, 1999. Brain perfusion SPECT and MRI in foetal alcohol syndrome. Dev. Med. Child Neurol. 41:652–659. Riikonen, R. S., P. Nokelainen, K. Valkonen, A. I. Kolehmainen, K. I. Kumpulainen, M. Kononen, et al., 2005. Deep serotonergic and dopaminergic structures in fetal alcoholic syndrome: A study with nor-β-CIT-single-photon emission computed tomography and magnetic resonance imaging volumetry. Biol. Psychiatry 57:1565–1572. Riley, E. P., C. Guerri, F. Calhoun, M. E. Charness, T. M. Foroud, T.-K. Li, et al., 2003. Prenatal alcohol exposure: Advancing knowledge through international collaborations. Alcohol. Clin. Exp. Res. 27:118–135. Riley, E. P., S. N. Mattson, E. R. Sowell, T. L. Jernigan, D. F. Sobel, and K. L. Jones, 1995. Abnormalities of the corpus callosum in children prenatally exposed to alcohol. Alcohol. Clin. Exp. Res. 19:1198–1202. Roebuck, T. M., S. N. Mattson, and E. P. Riley, 1998. A review of the neuroanatomical findings in children with fetal alcohol syndrome or prenatal exposure to alcohol. Alcohol. Clin. Exp. Res. 22:339–344. Roebuck, T. M., S. N. Mattson, and E. P. Riley, 2002. Interhemispheric transfer in children with heavy prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 26:1863–1871.

Roebuck, T. M., R. W. Simmons, S. N. Mattson, and E. P. Riley, 1998. Prenatal exposure to alcohol affects the ability to maintain postural balance. Alcohol. Clin. Exp. Res. 22:252–258. Roebuck, T. M., R. W. Simmons, C. Richardson, S. N. Mattson, and E. P. Riley, 1998. Neuromuscular responses to disturbance of balance in children with prenatal exposure to alcohol. Alcohol. Clin. Exp. Res. 22:1992–1997. Roebuck-Spencer, T. M., and S. N. Mattson, 2004. Implicit strategy affects learning in children with heavy prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 28:1424–1431. Schonfeld, A. M., S. N. Mattson, A. R. Lang, D. C. Delis, and E. P. Riley, 2001. Verbal and nonverbal fluency in children with heavy prenatal alcohol exposure. J. Stud. Alcohol 62: 239–246. Sowell, E. R., T. L. Jernigan, S. N. Mattson, E. P. Riley, D. F. Sobel, and K. L. Jones, 1996. Abnormal development of the cerebellar vermis in children prenatally exposed to alcohol: Size reduction in lobules I–V. Alcohol. Clin. Exp. Res. 20:31–34. Sowell, E. R., L. H. Lu, E. D. O’Hare, S. T. McCourt, S. N. Mattson, M. J. O’Connor, and S. Y. Bookheimer, 2007. Functional magnetic resonance imaging of verbal learning in children with heavy prenatal alcohol exposure. NeuroReport 18(7):635–639. Sowell, E. R., S. N. Mattson, P. M. Thompson, T. L. Jernigan, E. P. Riley, and A. W. Toga, 2001. Mapping callosal morphology and cognitive correlates: Effects of heavy prenatal alcohol exposure. Neurology 57:235–244. Sowell, E. R., P. M. Thompson, S. N. Mattson, K. D. Tessner, T. L. Jernigan, E. P. Riley, and A. W. Toga, 2001. Voxel-based morphometric analyses of the brain in children and adolescents prenatally exposed to alcohol. NeuroReport 12:515–523. Sowell, E. R., P. M. Thompson, S. N. Mattson, K. D. Tessner, T. L. Jernigan, E. P. Riley, and A. W. Toga, 2002. Regional brain shape abnormalities persist into adolescence after heavy prenatal alcohol exposure. Cerebral Cortex 12:856–865. Sowell, E. R., P. M. Thompson, B. S. Peterson, S. N. Mattson, S. E. Welcome, A. L. Henkenius, et al., 2002. Mapping cortical gray matter asymmetry patterns in adolescents with heavy prenatal alcohol exposure. NeuroImage 17:1807–1819. Steinhausen, H.-C., J. Willms, and H.-L. Spohr, 1993. Long-term psychopathological and cognitive outcome of children with fetal alcohol syndrome. J. Am. Acad. Child Adolesc. Psychiatry 32:990– 994. Streissguth, A. P., H. M. Barr, J. Kogan, and F. L. Bookstein, 1996. Final Report: Understanding the Occurrence of Secondary Disabilities in Clients with Fetal Alcohol Syndrome (FAS) and Fetal Alcohol Effects (FAE). Seattle: University of Washington Publication Services. Streissguth, A. P., H. M. Barr, and P. D. Sampson, 1990. Moderate prenatal alcohol exposure: Effects on child IQ and learning problems at age 7½ years. Alcohol. Clin. Exp. Res. 14:662–669. Swayze, V. W., II, V. P. Johnson, J. W. Hanson, J. Piven, Y. Sato, J. N. Giedd, et al., 1997. Magnetic resonance imaging of brain anomalies in fetal alcohol syndrome. Pediatrics 99:232–240. Uecker, A., and L. Nadel, 1996. Spatial locations gone awry: Object and spatial memory deficits in children with fetal alcohol syndrome. Neuropsychologia 34:209–223. Uecker, A., and L. Nadel, 1998. Spatial but not object memory impairments in children with fetal alcohol syndrome. Am. J. Ment. Retard. 103:12–18. Vaurio, L., A. Repp, C. L. McGee, E. P. Riley, and S. N. Mattson, 2006. Visual working memory deficits in children with heavy prenatal alcohol exposure are secondary to lower-order

mattson et al.: fetal alcohol syndrome

651

deficits. Presented at the 13th congress of the International Society for Biomedical Research on Alcoholism, Sydney, Australia, September. Alcohol. Clin. Exp. Res. 30:178A. Viljoen, D. L., J. P. Gossage, L. Brooke, C. M. Adnams, K. L. Jones, L. K. Robinson, et al., 2005. Fetal alcohol syndrome epidemiology in a South African community: A second study of a very high prevalence area. J. Stud. Alcohol 66:593–604. Willford, J. A., G. A. Richardson, S. L. Leech, and N. L. Day, 2004. Verbal and visuospatial learning and memory function in

652

children with moderate prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 28:497–507. Woodruff-Pak, D. S., G. Goldenberg, M. M. Downey-Lamb, O. B. Boyko, and S. K. Lemieux, 2000. Cerebellar volume in humans related to magnitude of classical conditioning. NeuroReport 11:609–615. Woodruff-Pak, D. S., M. Papka, and R. B. Ivry, 1996. Cerebellar involvement in eyeblink classical conditioning in humans. Neuropsychology 10:443–458.

neurodevelopmental aspects of clinical disorders

40

Impact of Prenatal Cocaine Exposure on the Developing Nervous System ERIC M. LANGLOIS AND LINDA C. MAYES

Conservative estimates have reported that prenatal exposure to cocaine occurs in roughly 45,000 infants each year in the United States (National Institute on Drug Abuse, 1995). These estimates have triggered serious concerns with respect to the possible teratologic consequences that cocaine may have on human fetal and postnatal development. Despite nearly two decades of extensive research by numerous investigators, the potential physical, neurodevelopmental, and neuropsychological impact of prenatal cocaine exposure on the postnatal development of the neonate continues to be debated. This disagreement has been sustained by numerous methodological problems, inconsistencies in study design, and confounding factors such as inadequate pre- and postnatal nutrition and care, socioeconomic status, and exposure to other drugs (for a review of central methodological issues in studies of prenatal cocaine exposure see Brooks-Gunn, McCarton, and Hawley, 1994a, 1994b; Frank et al., 2001; Frank, Augustyn, and Zuckerman, 1998; Lester, Freier, and LaGasse, 1995; Mayes and Fahy, 2001; Neuspiel, 1995). Some researchers argue that children exposed in utero to cocaine do not display noticeable aberrations in neurological functioning (Eyler et al., 2001; Frank et al., 2001; Haasen and Krausz, 2001; Rose-Jacobs et al., 2002). Others have concluded that although still scant, inconsistent, or inconclusive on many crucial issues and marked by a number of methodological problems, published studies to date nonetheless reveal the beginnings of a profile of possible cocainerelated effects on neuropsychological functions subserving arousal and attention regulation and reactivity to stressful conditions (Alessandri et al., 1993; Azuma and Chasnoff, 1993; Bendersky et al., 2003; Bendersky and Lewis, 1998; Griffith, Azuma, and Chasnoff, 1994; Hansen, Struthers, and Gospe, 1993; Heffelfinger et al., 2002; Jacobson, Jacobson, and Sokol, 1994; Jacobson et al., 1996; Mayes and Fahy, 2001; Mayes et al., 1998; Metosky and Vondra, 1995; Richardson, Conroy, and Day, 1996; Singer et al., 2000, 2004; Struthers and Hansen, 1992). That profile is further elaborated by findings from several animal models in which crucial factors such as duration and type of exposure as

well as environmental conditions may be more adequately controlled. In this chapter we review the state of knowledge regarding the neurobiological effects of prenatal cocaine exposure on the developing nervous system as elaborated through the use of animal models. Studies in this area represent an important interface between basic studies of normal neural ontogeny and those of neuroteratology. Because cocaine is a potent central nervous system stimulant, understanding its effects on the developing brain not only explicates the possible neurodevelopmental impairments resulting from intrauterine exposure to a potential neurotoxin, but also sheds light on mechanisms of normal neural ontogeny. This review focuses on neurobiological mechanisms purported to account for the neurotoxicity of cocaine in developing vertebrates. We examine the mechanisms for the effect of cocaine exposure on brain and cellular structure, neurotransmitter function, and genetic regulatory mechanisms. Wherever data are available, we point to possible neurobehavioral correlates for these alterations in structure and function.

Candidate mechanisms for cocaine-related neurobehavioral teratologic effects There are several candidate mechanisms proposed for the effect of prenatal cocaine exposure on the ontogeny of neural systems and, by extension, for the cocaine-related neurobehavioral profiles. These are effects on monoaminergic system development; changes in neural growth factors; alteration of ion channel and monoamine transport development; effects on other neurotransmitter systems (such as the cholinergic system) and on neuropeptides (including substance P), dynorphin, gamma aminobutyric acid (GABA), and glutamate sites; or alteration of immediate early gene expression (Kosofsky and Wilkins, 1998; Mayes, 1999; Nestler, 1994). Cocaine has potent vasoconstrictive effects that may also have a more generalized and less system-specific

653

hypoxia-related deleterious effect on neural growth. The effect on central and peripheral noradrenergic systems accounts for the cardiovascular effects of cocaine, including tachycardia, hypertension, and increased risks for strokes and myocardial infarctions and arrhythmias. Thus in utero cocaine exposure may compromise fetal brain development through an effect on placental vascular function with decreased placental blood flow and consequent fetal hypoxemia and possibly ischemic injury. Similarly, cocaine-related noradrenergic effects on developing fetal vasculature may also compromise blood flow to the developing brain (Koegler et al., 1991; Lipton et al., 2002; Woods, Plessinger, and Clark, 1987). This more general level of effect through fetal hypoxemia or relative ischemia may not be as specifically directed to one region of the brain as to another, and the severity of effect, if any, will depend particularly on the timing of the exposure and the level of relative hypoxemia. Another group of mechanisms of in utero cocaine effect on neurodevelopmental outcome is also more general but is unique to the human model. Substance-abusing adults rarely use cocaine in isolation. Most often, cocaine abuse is combined with other potential neuroteratogens including alcohol, marijuana, and tobacco. While there are no studies modeling various combinations of polydrug exposure in animals, it is reasonable to hypothesize that interactions among different drugs may produce both potentially neurotoxic metabolites and directly additive effects on neural development. Related to human substance abuse are other more general health and environmental factors that have effects on fetal health and development. These include poor maternal nutrition (in part reflecting the appetite suppression effect of cocaine) and, more often than not, maternal poverty and environmental chaos that may contribute to postnatal neglect and poor parenting, factors that are themselves developmentally compromising. These levels of effect unique to humans cannot be adequately modeled in animals but do confound efforts to attribute any neurobehavioral impairments found in infants exposed to cocaine prenatally and young children exposed to cocaine alone. The methodologic dilemmas in human studies of prenatal cocaine exposure are reviewed in detail elsewhere (Brooks-Gunn, McCarton, and Hawley, 1994b; Chiriboga, 1998; Frank et al., 2001; Frank, Augustyn, and Zuckerman, 1998; Lester, Freier, and LaGasse, 1995; Mayes, 1999; Mayes and Fahy, 2001; Neuspiel, 1995; Olson and Toth, 1999; Singer, 1999; Tronick and Beeghly, 1999). Importantly, the nonspecific but frequent mixture of acute and chronic stress characterizing the postnatal caregiving environments for many prenatally exposed children may also have enduring effects on attention-andarousal-system ontogeny, perhaps through some of the same monoamine-related mechanisms.

654

Preclinical models of prenatal cocaine exposure Cortical Morphology Growing evidence is available for cocaine-related effects on the earliest phases of neurogenesis, cell proliferation, and neuronal migration and resulting alterations in cortical morphology. Investigators have discovered that differentiation of neuroblastoma cells by nerve growth factor (NGF) and cell proliferation stimulated by insulin-like growth factor I (IGF-I) are both inhibited by cocaine (Zachor et al., 1994, 1996). This interruption is likely mediated through cocaine’s effect on D1 receptor activation, since similar inhibitory effects on neuronal differentiation of NGF-treated cells occur with dopamine agonists selective to the D1 subtype receptors (Zachor et al., 2000). Similarly, at the stage of neuronal formation and proliferation in cultured preparations of glioblastoma and neocortical cells, cocaine exposure delays uridine and thymidine incorporation (Garg, Turndorf, and Bansinath, 1993). Cocaine-induced suppression of cell proliferative processes has also been demonstrated in neonatal rats (Anderson-Brown, Slotkin, and Seidler, 1990). The neurogenesis of cultured noradrenergic neurons of the locus coeruleus is significantly disrupted by cocaine administration, with a reduction in cell adhesion and neurite elongation (Snow et al., 2001). Prenatal exposure in the intact pregnancy of rats early in gestation interferes with radial gliogenesis and thus disrupts neuronal migration and resulting cortical architecture (Akbari, Whitaker-Azmitia, and Azmitia, 1994; Gressens, Gofflot, et al., 1992; Gressens, Kosofsky, and Evrard, 1992; Yablonsky-Alter et al., 1992). Intermittent prenatal cocaine exposure in rhesus monkeys further yields cerebral cortices with highly abnormal structural characteristics comprising disrupted cortical laminar architecture with an increased number of cells in the underlying white matter, consistent with markedly impaired neuronal migration (He and Lidow, 2004; Lidow, 1995; Lidow and Song, 2001a). There is also a decrease in the volume and density of cortical glial (connective) elements with numerous cortical cells failing to reach their proper destination. This alteration in cytoarchitecture persists into adulthood, suggesting that the disruption may be permanent with little chance of recovery following gestation (Lidow and Song, 2001b). Although an increase in cell death may be a source of altered cortical migration and cortical neuronal distribution in the rhesus model—and indeed a significant increase in cell death throughout the fetal cerebral wall has been reported during neocortical neurogenesis in primates (He, Song, and Lidow, 1999)—exposure to cocaine either prior to neocortical neurogenesis or after neurogenesis does not result in decreased neocortical neurons (Lidow, Bozian, and Song, 2001). Cocaine-induced cell death may be restricted in primates to the period of neurogenesis, or cocaine-related cell death may have an overall minimal

neurodevelopmental aspects of clinical disorders

impact on the developing structure of the cortex (Lidow, 2003). A significant reduction in neuronal migration from the ganglionic eminence to the cerebral wall, as well as a decrease in overall GABAergic cell density in those regions, has also been associated with prenatal cocaine exposure (Stanwood et al., 2001). Examination of the primary somatosensory cortex of the prenatally cocaine-exposed mouse reveals depletions in infragranular neuron density (Ren et al., 2004). Prenatal cocaine exposure has also been shown to reduce the number of inhibitory axo-axonic interneuron structures in rats (Morrow, Elsworth, and Roth, 2003). Mouse embryos exposed to cocaine very early in gestation show a decrease in glial formation and glial density, and a disorganization of axon-dendrite bundles (Gressens, Gofflot, et al., 1992). Findings such as these may reflect disrupted monoaminergic-system-regulated processes that control the genesis of radial glial cells necessary for proper neuronal migration to cortical layers and regulate the formation of connections among neurons (Lidow and Rakic, 1994, 1995; Seidler et al., 1995; Slotkin and Seidler, 1992). Cocaine may also have a direct toxic effect on early stages of neuronal proliferation and early connectivity, as suggested by cell culture preparations of both glia and neurons. A decrease in the number of neurons and a degeneration of the connective processes through which communication occurs has been demonstrated with cocaine exposure to growing neurons in culture (Nassogne et al., 1995; Nassogne, Evrard, and Courtoy, 1998). There is no apparent reduction in actual glial numbers, but there is a change in cell structure, suggesting again a cocaine-related effect on the morphology of cells involved in neural connectivity. Closer examination of cocaine-exposed neurons shows morphologic changes in nuclei indicative of apoptosis or neuronal death after approximately two days of exposure, suggesting that cocaine exposure initiates apoptotic processes within the neuron by yet-undefined mechanisms (Nassogne et al., 1997). One possibility may be through the action of monoamines, given that exposure of neuronal cell cultures to catecholamines can lead to neuronal death through action as a glutamate agonist (Nassogne, Evrard, and Courtoy, 1998). The occurrence of neuronal death, as measured by an increase in the number of TUNEL-positive nuclei, has also been demonstrated in the cerebral wall of cocaine-exposed primates (He, Song, and Lidow, 1999). Specific studies examining the relation between cocaine exposure and neurogenesis in the cerebral wall of the rhesus have revealed cocaine-induced alterations in the proliferative process (Lidow and Song, 2001b). More specifically, cocaine first triggers an initial suppression in cell growth, which is then followed by a significant boost in compensatory proliferation that occurs when the concentration of cocaine deteriorates.

Prenatal cocaine exposure has also been associated with structural abnormalities in the dendrites of corticolimbic pyramidal neurons of cocaine-exposed animals. Using a rabbit model, investigators have found that dendrites of pyramidal neurons in layers III and V of the anterior cingulate cortex are 30 to 50 percent longer in exposed compared to nonexposed, saline-treated animals (L. Jones et al., 2000; Levitt et al., 1997; Stanwood et al., 2001). Confocal analysis shows that these dendrites course abnormally through the cortex. Rather than the expected straight distribution, the dendrites course in and out of the plane of a section. Since the exposed animals show normal lamination and thickness of cortical layers in the anterior cingulate, the trajectory pattern of the dendrites suggests less-controlled growth with the extended dendritic projections undulating to fit within the limits of the cortical layers. Similar observations are made in the prefrontal cortex, another cortical region targeted by dense dopaminergic afferents, but these changes in dendritic growth are not found in regions that receive sparse dopaminergic input such as the primary sensory cortex (L. Jones et al., 2000; Levitt et al., 1997; Lloyd et al., 2003; Stanwood et al., 2001). Analyses of the effects of prenatal cocaine exposure on the histotypic layers of the frontal cortex in the mouse reveal comparable findings. Similarly, only regions receiving dense monoaminergic input showed cocaine-related effects in the development of neuronal dendritic trees. Dendrite lengths are increased in layer III; however, unlike findings in the anterior cingulate of the rabbit, dendrite lengths are decreased in layer IV (Lloyd et al., 2003). This dichotomy may be explained by the differential expression of dopamine receptor subtypes in neurons of different histotypical layers of the cortex. These morphologic changes appear rapidly and persist into adulthood. In vitro studies show that after only two weeks of exposure to cocaine, fetal neurons from the anterior cingulate plated in culture without further cocaine exposure show a marked increased in dendritic growth, a finding suggesting that the cocaine-related effect on growth is not simply a direct exposure effect but rather an effect that is brought about through an alteration in growth-regulating mechanisms. Direct Effects on Neurotransmitter Function Several neurochemical effects have been reported in both in vivo and in vitro models, each pointing to prenatal cocaineexposure-related effects on neurotransmitter function and, more specifically, monoaminergic system development at the level of transmitter synthesis, activity, or receptor formation-function. Findings from preclinical rodent models point to specific effects on circuitry mediated by dopamine (DA) D1/D2 that is critical to arousal regulation of the prefrontal cortex (Seamans, Floresco, and Phillips, 1998) and to up-regulation of norepinephrine (NE) and serotonergic

langlois and mayes: prenatal cocaine exposure

655

systems at a time in early development that could permanently distort the excitatory/inhibitory balance between arousal systems necessary for optimal cortical function. Dopamine. Adult rats exposed prenatally to cocaine show a reduction in basal dopamine release and decreased activity in dopaminergic cells in the substantia nigra and ventral tegmentum (Minabe et al., 1992). These decreases in dopamine release and activity (Glatt et al., 2004), as well as reductions in glucose metabolic activity, are particularly marked in prenatally exposed male rats in the hypothalamus, nigrostriatal pathway, and structures within the limbic system (Dow-Edwards, 1993). Juvenile and adult rats exposed in utero to cocaine also exhibit higher levels of presynaptic dopamine reuptake (Glatt et al., 2004). Studies examining the effects of cocaine exposure on dopamine neurotransmission in vivo in mature offspring report contradictory findings. Phillips and colleagues (2003) employed fast cyclic voltammetry recordings using carbon fiber microelectrodes to study presynaptic dopaminergic function and found no differences between cocaine- and saline-exposed adult rats in the dopaminergic response to single-pulse electrical stimulation in either the caudate or the nucleus accumbens. Using a different cocaine dosage and analysis method, other investigators found a significant reduction in potassiumevoked dopamine release and transporter (DAT) protein levels (Salvatore et al., 2004). Prenatally exposed rabbits show reduced striatal dopamine levels at birth (Weese-Mayer et al., 1993), and rhesus at 60 days prenatally show reduced tyrosine hydroxylase mRNA (Ronnekleiv and Naylor, 1995) in the substantia nigra and ventral tegmental area. A reduction in dopamine synthesis may lead to increased receptor sensitivity and proliferation, as is the case in both rhesus and rat that are prenatally exposed (Dow-Edwards et al., 1993; Ronnekleiv et al., 1998; Scalzo et al., 1990). Other studies examining receptor sensitivity and density following prenatal exposure have examined postnatal brains following early or late exposure and have reported conflicting results with either a decrease, no change, or an increase in the expression of D1 and D2 receptors and ligand binding to those receptors (Booze et al., 2006; Henderson, McConnaughey, and McMillen, 1991; Leslie et al., 1994; Scalzo et al., 1990; H.-Y. Wang, Runyan, et al., 1995; H.-Y. Wang, Yeung, and Friedman, 1995). These conflicting results may reflect compensatory changes in dopamine neural transmission following withdrawal from cocaine or cessation of exposure, a suggestion that has been confirmed in adult animals (Bonci and Williams, 1996). When cocaine is administered postnatally to rat pups in either the first or second week postnatally (equivalent to later gestation in the human and during a period of rapid synaptogenesis), there is widespread increase in glucose metabolic activity across several different regions of the brain including

656

cortex (Dow-Edwards et al., 1993) but with particular increases in highly dopaminergic-innervated regions. These effects are most dramatic in the female. For the male, there are no substantial areas of increased metabolism, and the nucleus accumbens continues to show decreased activity, as was true prenatally. When ligands specific for dopamine D1 and D2 receptors are injected into postnatally treated animals, areas showing marked increase in glucose uptake are those with greater dopaminergic activity (Dow-Edwards, 1989). Earlier postnatal exposure (days 1 to 10) affects the mesolimbic dopamine system and other structures within the limbic system differentially more than later postnatal exposure (days 11 to 20), resulting in more general activation (Dow-Edwards, Freed, and Milhorat, 1988). Such general activation consistent with the relation between monoamines and synaptic formation suggests that exposure to a monoaminergic agonist such as cocaine during these periods of cortical maturation results in an overall increase in monoaminergic (and specifically dopaminergic) activity. More specific studies of receptor functioning have shown that prenatal cocaine exposure is associated with a functional decoupling of the D1 receptor from G protein (Friedman and Wang, 1998; H.-Y. Wang, Runyan, et al., 1995), the second messenger within the cell that begins a cascade of intracelluar events. Prenatal cocaine exposure impairs coupling in the frontal and cingulate areas of the dopaminergic D1 receptor system to one of the second messengers, GαS protein, a functional decrease that persists into adulthood while coupling between D2 and Gi protein remains normal (Friedman, Yadin, and Wang, 1996; H.-Y. Wang, Runyan, et al., 1995; X.-H. Wang et al., 1995). This decoupling of D1 to GαS protein may be secondary to increased phosphorylation of the receptor that in turn renders the receptor desensitized to G protein (Svenningsson et al., 2000; Zhen et al., 2001). Also, mRNA for the dopamine transporter is reduced in these same regions following both prenatal and postnatal exposure (Dow-Edwards, Yin, and Hurd, 1997; Leslie et al., 1994). Taken as a group, these findings suggest a selective or at least a more specific attenuating effect of cocaine on the development of the D1 system. In the normal individual, D1 and D2 receptor systems function interactively and in some ways reciprocally to regulate thalamic and cortical inputs. In particular, a reciprocal relation between D1 and D2 activity coupled also with a balance between noradrenergic α2 and α1 activity regulates and protects prefrontal cortical activity during increasing arousal or stressful states (Arnsten, 1997, 1998; Arnsten et al., 1998). Increased D2 (and α1) (Arnsten and Jentsch, 1997) activity may in effect take certain prefrontal cortical activities off-line and, in so doing, interfere with aspects of attention and other higher order reasoning and discriminative abilities (functions that typically also deteriorate in hyperaroused states). An increase in D2

neurodevelopmental aspects of clinical disorders

activity might be expected to parallel attenuation in D1 systems; and indeed, prenatal cocaine exposure does appear to be associated with an increase in D2 binding affinity, receptor density, and mRNA for the D2 receptor in the striatum (Leslie et al., 1994; Moody, Frambes, and Spear, 1992; Stadlin et al., 1995). Postnatal exposure (days 1–9) also results in an increased density of D2 in striatum (Howard, Fisher, and Landry, 1997). Additionally, challenges to the D2 system using a D2 antagonist, haloperidol, result in an exaggerated response in postnatally exposed male rats (Dow-Edwards, 1998). More recently, investigators have begun to explore the possible impact of prenatal cocaine exposure on D3 and sigma receptors (Silvers et al., 2006). Intravenous administration of cocaine to adult rats produces alterations in D3 receptors (Peris et al., 1990; Wallace, Mactutus, and Booze, 1996), and evidence exists of a modulatory relationship between the sigma and dopaminergic systems (Bastianetto et al., 1995; McCracken, Bowen, and Matsumoto, 1999; Ritz and George, 1997; Santolaria-Fernandez et al., 1995). Sigma receptor protein levels and gene expression significantly escalate in adult male rats following acute cocaine exposure, and sigma1-specific receptor antagonists effectively reduce neurological and behavioral responses to cocaine as well as cocaine-induced toxicity (Lason, 2001; R. Matsumoto et al., 2001, 2004; Maurice and Romieu, 2004; Romieu, Martin-Fardon, and Maurice, 2000; Romieu et al., 2004; Ujike, Kuroda, and Otsuki, 1996). Studies examining the effects of prenatal cocaine exposure reveal a significant increase in D3 and sigma receptor binding in the striatum and nucleus accumbens (Silvers et al., 2006), suggesting another possible mechanism for the cocaine-induced alteration in monoaminergic system development. Gamma aminobutyric acid (GABA). An abundance of recent data is emerging supporting the essential role of gamma aminobutyric acid (GABA) in the development and plasticity of the cortex. The GABAergic system is one of the earliest neurotransmitter systems to develop and may have an important, early neurotrophic role during the development of cortical neurons (Lauder, 1988; Meier et al., 1991; Parnavelas, 1992; Van Eden et al., 1989). In addition to being one of the earliest neurotransmitter systems formed, the GABAergic system is portrayed by a long period of plasticity during the neonatal period, underscoring the potential contributions of the GABAergic system in synaptic plasticity (Micheva, 1997). GABA may play an important role in the development and maintenance of behavioral sensitization (Kalivas and Duffy, 1997; Karler et al., 1995; Pierce and Kalivas, 1997; Wolf, 1998). Virtually every GABAergic cortical neuron expresses only one of the three major calcium-binding proteins, and in many cortical areas

these proteins are expressed only in GABAergic neurons (Blaustein, 1988; Du et al., 1995; Nitsch et al., 1989, 1994; Nitsch, Scotti, and Nitsch, 1995). Parvalbumin is one of these calcium-binding proteins, which is characteristically observed in interneurons that have a powerful inhibitory influence on their target neurons. The onset of parvalbumin immunoreactivity coincides with synaptic formation, innervation by cortical afferents, and the growth of dendritic arborizations (Hendrickson et al., 1991; Soriano et al., 1992; Sven and Marco, 1991). An increase in GABA-immunoreactive cells (H.-Y. Wang, Runyan, et al., 1995; X.-H. Wang et al., 1995) and immunoreactivity of parvalbumin in primary, secondary, and tertiary dendrites of interneurons in the anterior cingulate cortex is significantly increased in cocaine-exposed rabbits compared to controls (Murphy and Segal, 1997; Stanwood et al., 2001; X.-H. Wang et al., 1996). This increase in GABA-immunoreactive cells seems evident when cocaine exposure occurs during a period of peak neuronal differentiation (embryonic days E16–E29) and is detectable in both postnatal days 10 and 20 (Stanwood et al., 2001). Contradictorily, Morrow, Elsworth, and Roth (2005) found a significant reduction in the number of primary, secondary, and tertiary dendrites in parvalbumin-immunoreactive local-circuit neurons in the frontal cortex of rats, possibly reflecting methodological and species differences. There are primarily two types of GABAergic local-circuit neurons that exhibit parvalbumin immunoreactivity and have the potential to be important in the cognitive function of the frontal cortex: axosomatic basket cells and axo-axonic chandelier cells (Gabbott and Bacon, 1996). The axons of basket cells target the pyramidal soma, while the axon terminals of the chandelier cells form the inhibitory axo-axonic interneuron structures. Prenatal cocaine exposure produces a 50 percent reduction in the number of spindle-shaped parvalbuminimmunoreactive cells in the medial prefrontal cortex of rats, suggesting either a loss of chandelier cells or an alteration in their shape (Morrow, Elsworth, and Roth, 2005). These chandelier cells are anatomically positioned to intercept the action potential of pyramidal neurons, suggesting that a decreased influence of these cells, mediated by either a diminution or change in shape, may result in a hyperresponsiveness of pyramidal neurons (Morrow, Elsworth, and Roth, 2005). Following prenatal cocaine exposure, the expression of select GABAA receptor subunit mRNAs is differentially regulated in which α1 mRNA levels are significantly higher in layer III and β2 mRNA levels are significantly lower in layer II of the anterior cingulate of cocaine-exposed rabbits compared to controls (Shumsky et al., 2002). These differences are transient, appearing at postnatal day 20, during a time of peak plasticity in the formation, refinement, and stabilization of synaptic connections, and diminishing by postnatal day 60 (Shumsky et al., 2002).

langlois and mayes: prenatal cocaine exposure

657

The direct mechanism explaining cocaine’s effect on the developing GABAergic system remains to be discovered. However, many investigators hint at the dopaminergic system’s role in the cocaine-provoked developmental abnormalities observed in the GABAergic system. The function of the central nervous system depends on complex interactions between a variety of neurotransmitters and neuromodulators in which alterations in one system may result in compensatory changes in another. There is little evidence that cocaine acts directly on the GABAergic system; however, compelling support of an interaction between the dopaminergic and GABAergic systems is beginning to emerge (Aceves et al., 1992; Beauregard and Ferron, 1991; Bergson et al., 1995; Bernath and Zigmond, 1989; Cameron and Williams, 1993, 1994; Campbell et al., 1993; Goeders et al., 1990; Koob, 1992; Lindefors, 1993; Retaux, Besson, and Penit-Soria, 1991; Santiago, Machado, and Cano, 1993; Verney et al., 1991). Dopamine D1 receptors may play a role in modulating the release of GABA (Aceves et al., 1992; Beauregard and Ferron, 1991; Cameron and Williams, 1994; Campbell et al., 1993). The altered development of the GABAergic system in animals prenatally exposed to cocaine provides further evidence of a cocaine-induced disruption in the inhibitory/excitatory balance of neurotransmitter systems. Serotonin. Cocaine appears to have a greater affinity for the serotonin transporter in developing brain, in contrast to the adult brain in which cocaine binds predominantly to dopamine transporters (Meijer et al., 2000; Shearman and Collins, 1996; Whitaker-Azmitia, 1998). These effects have led several to suggest that the relationship between cocaine and serotonin systems may be more important to understanding mechanisms for teratogenic effects in developing brain, whereas dopamine systems may be more influential in mature, and perhaps postnatal, brain (Henderson et al., 1991). Though less work has been done to date on serotonin systems and cocaine exposure, the findings are nonetheless consistent with hypotheses based on cocaine-related effects on serotonin metabolism and serotonin’s role as an early trophic factor. During development, serotonin neurons use a negative feedback mechanism for regulating neuronal and cortical growth. Both tissue culture and whole animal studies show that, early in development, increased extracellular serotonin causes a decrease in serotonin neuronal outgrowth. It has been demonstrated that, early on, cocaine impedes the development of reuptake mechanisms (DeGeorge et al., 1989). Thus, based on cocaine’s inhibiting effect on synaptic reuptake, more serotonin is present in the extracellular synaptic cleft, which should limit the growth of serotonin terminals (Whitaker-Azmitia, 1998). These predictions have been supported: studies of prenatally exposed rats reveal a

658

decrease in serotonin terminal density and an accompanying reduction in the normal growth of serotonin fibers in the hippocampus (Akbari et al., 1992; Yan, 2002). However, Akbari and colleagues (1992) found this effect to be transient, and by a month postnatally no differences between exposed and nonexposed animals are apparent. However, others have shown that by a month postnatally late-onset increased growth of serotonin fibers is observed in the striatum (Snyder-Keller and Keller, 1993). These discrepancies may reflect differences in dosage (Yan, 2002), or it may be that if growth is inhibited or delayed in one region, it is increased elsewhere (Whitaker-Azmitia, 1998). The ability of serotonin fiber density in cortex and hippocampus to return to normal levels by a month postnatally introduces an interesting yet speculative hypothesis for a cocaine-induced alteration in serotonergic systems. Rather than suggesting transient to no effect of cocaine, even early, brief disruption of one aspect of serotonin system ontogeny may disrupt other areas of serotonin system development. Delayed serotonin terminal development may lead to decreased serotonin levels in later phases even after exposure has stopped. Decreased serotonin levels in the later prenatal and early postnatal period may have marked consequences for neuronal connectivity. For example, depletion of serotonin by selective synthesis inhibitors results in delayed neuronal differentiation and decreased synaptic connectivity in both cortex and hippocampus in the chick model (Cheng et al., 1993). In rats, depletion of serotonin on the third postnatal day by a selective neurotoxin results in decreased density of granule cells dendritic spines or decreased connectivity that actually continues to worsen as the animals mature (Genkova-Papazova et al., 1997). Depletion at postnatal day 10 to 20 causes marked loss in dendrites, with reductions still apparent up to a year of age (Mazer et al., 1997). Moreover, delaying serotonin terminal development has been associated with behavioral deficits in the adult animal (Shemer, Azmitia, and Whitaker-Azmitia, 1991). By altering serotonin terminal development, prenatal exposure alters events weeks downstream from the exposure. In the first few weeks postnatally in the rat, serotonin is required for astroglial maturation, and in the second to third postnatal week, serotonin plays a critical role in neuronal maturation and synaptogenesis (Chubakov, Tsyganova, and Sarkisova, 1993). Postnatal synaptogenesis and neuronal connectivity depend on the maturation of a number of cell types including astroglial cells. Astroglial cells that play a role early in gestation as the radial glial cells for neuronal migration are also crucial for later connectivity and the production of neural growth factors. Serotonin has a regulatory role in the maturation of the astroglial cell (Lauder and Liu, 1994; Whitaker-Azmitia and Azmitia, 1994); and the rapid increase of serotonin 1A receptors (5-HT1A) early in gestation in

neurodevelopmental aspects of clinical disorders

mammalian brain is largely through serotonin receptors on astroglial cells (Hellendall et al., 1992). One aspect of that maturation is a morphologic change from the long processes needed for guiding neuronal migration to shorter but more processes with greater connectivity among cells. A second aspect of serotonin-mediated astroglial maturation is the release of growth factor S-100β (Whitaker-Azmitia, Murphy, and Azmitia, 1990). This particular growth factor causes neuronal outgrowth and neuronal plasticity in response to experience (such as activity, handling, enriched stimulation environments) (Muller, Akhavan, and Bette, 1993; Whitaker-Azmitia and Azmitia, 1994). Thus S-100β may be one factor involved in potentiation of learning and experience-dependent synaptogenesis. Depletion or decrease in S-100β release from astroglial cells should result in decreased dendritic development and delayed neural outgrowth. Prenatal cocaine exposure may interfere with astroglial cell maturation and with the release of S-100β. Prenatally exposed rat pups examined on postnatal day 6 show a delay in the morphologic change in astroglial cells from their radial glial function to more mature cells (Clarke et al., 1996). There are also fewer astroglial cells in the cortex and hippocampal regions, with overall cortical thinning. (Similar reductions in cortical astroglial cells have also been reported for the rhesus model at 2 months postnatal age (Lidow, 1995). Immunochemical staining for S-100β revealed significant reductions (Akbari, Whitaker-Azmitia, and Azmitia, 1994; Clarke et al., 1996). It is possible then that even if serotonin levels returned to normal in the subsequent postnatal period, the reduction of S-100β during that critical period may have more enduring effects. Conversely, treatment of prenatally exposed rat pups with 5-HT1A agonists on postnatal days 1 through 5 showed increases in S-100β in both cortex and hippocampus and a significant increase in brain growth (Akbari, Whitaker-Azmitia, and Azmitia, 1994). Prenatal cocaine-exposure-associated delay in serotonin terminal growth, delay in astroglial maturation, and delay in release of S-100β may lead to loss of synapses in the adult animal and thus to deficits in learning and memory (Whitaker-Azmitia, 1998). Postnatal or preweaning exposure appears to increase serotonin levels. Postnatal exposure to serotonin agonists in the rat from birth to postnatal day 20 increases brain activity during active synaptogenesis leading to accelerated brain maturation with attendant behavioral changes, including increased activity, and to behavioral changes in the mature animal, including increased anxiety and activity (Borella, Bindra, and Whitaker-Azmitia, 1997; Whitaker-Azmitia, 1998). In addition, preweaning administration of cocaine decreases 5-HT1A receptor density in the raphe nuclei (Dow-Edwards et al., 1998; Shi et al., 1998) where the action of the 5-HT1A is inhibitory. Thus a

decrease in receptor density is, in effect, an increase in afferent or ascending serotonergic activity (Pennington, Kelly, and Fox, 1991). Serotonin neurons innervating the forebrain region also send projections to the hypothalamus and mediate changes in plasma ACTH, corticosterone, renin, and prolactin (Van de Kar, Richardson-Morton, and Rittenhouse, 1991). Disruptions in serotonin system ontogeny in both receptor distribution and density of serotonin projections may also alter postsynaptic serotonin receptor function and hypothalamic response to neuroendocrine challenge tests. Longterm consequences of prenatal cocaine exposure on mature serotonergic systems have been somewhat contradictory. Recently, investigators have reported that when prenatal cocaine-exposed rats reached adulthood, the 5-HT1A (Chen et al., 2005) and 5-HT2A (Chen et al., 2004) receptor– mediated neuroendocrine responses were indistinguishable from control rats (Chen et al., 2005). Others demonstrate that adult rats treated prenatally between day 13 and day 20 of gestation and studied at postnatal day 30 show increased ACTH and renin response to 5-HT1A (Battaglia and Cabrera, 1994; Battaglia et al., 1998) and 5-HT2A (Battaglia et al., 2000) receptor agonists. In contrast, ACTH and renin response is attenuated by a serotonin releaser acting at the presynapse, suggesting the potentiated receptor responses may represent a compensatory response to the reported presynaptic deficit. Some studies report that these changes are not paralleled by alterations in 5-HT1A receptor density in hypothalamus or midbrain or by any alteration in serotonin uptake sites at the presynapse or in levels of serotonin (Battaglia and Cabrera, 1994; Battaglia et al., 1998, 2000). Others report that by postnatal day 70, 5-HT1A receptor densities are increased in cortex and midbrain but not hypothalamus (Cabrera et al., 1993; Cabrera-Vera et al., 2000). Still others report that consequences of prenatal cocaine exposure on 5-HT1A receptor densities are present at birth ( Johns et al., 2002) in a dynamic manner in male rat offspring that is dissimilar to that of females ( Johns et al., 2002). In male progeny, receptors show an initial upsurge neonatally, followed by a permanent reduction in adulthood. Females, on the contrary, undergo an opposite effect in which a neonatal reduction in receptor density returns to control levels by postnatal day 60, suggesting that prenatal receptor development may be hindered by cocaine exposure in males whereas the process is accelerated in females ( Johns et al., 2002). Taken together, these findings hint at an ongoing, pervasive disruption to monoamine-related systems after exposure has ceased and the continued alteration of effects downstream following early changes in system function. Furthermore, the evidence also implies that certain alterations in monoaminergic system function related to prenatal cocaine exposure may surface only when the system is challenged.

langlois and mayes: prenatal cocaine exposure

659

Norepinepherine. Far fewer studies have been done on the relation between prenatal cocaine exposure and the developing noradrenergic system, and the majority have been in the postnatal but preweaning period. Preweaning cocaine administration appears to increase activity within the overall noradrenergic system (Seidler and Slotkin, 1992). Also demonstrated has been increased fiber density in norepinephrinerich regions, increased receptor binding (β1 in cortex at 30 days postnatal but not later; α1 and α2 in cerebellum and forebrain), and increased presynaptic synthesis of norepinephrine (Akbari, Whitaker-Azmitia, and Azmitia, 1992; Henderson, McConnaughey, and McMillen, 1991; Seidler and Slotkin, 1992). Prenatal exposure studies have also found increased adrenergic-α2 receptor density in the hippocampus, parietal cortex, and amygdala (Booze et al., 2006). Rat pups exposed to cocaine on postnatal days 4 through 9 and then exposed to a noradrenergic agonist (clonidine) showed enhanced response to the locomotor stimulating effects of the agonist (Barron et al., 1998). Noradrenergic neurons arising from the locus coeruleus innervate the entire forebrain, with many projections to the posterior and prefrontal cortex. The locus coeruleus appears critical to vigilance or the response to a salient target (Delagrange et al., 1993), possibly through projections to the prefrontal cortex. Of special significance is the afferent input from the prefrontal cortex to the locus coeruleus that suggests a feedback, modulating link between cortical activity and ongoing noradrenergic activity (Arnsten and Goldman-Rakic, 1984). Excessive stimulation of adrenergic α1 receptors, in effect, initiates the action of taking the prefrontal cortex off-line (Arnsten and Jentsch, 1997). This in turn disrupts the regulatory loop between the locus coeruleus and the prefrontal cortex and may disrupt the capacity for vigilance and the ability to distinguish salient from nonsalient information (Rajkowsi et al., 1998), a behavioral finding demonstrated in cocaine-exposed animal models (Gabriel and Taylor, 1998; Romano and Harvey, 1998) and suggested in human studies (Mayes, Bornstein, et al., 1996; Mayes, Grillon, et al., 1998). Prenatal cocaine exposure has been demonstrated to have a direct effect on the development of noradrenergic neurons in the locus coeruleus in which total neurite length, neurite length per cell, and overall percentage of cells with neurites is significantly reduced (Snow et al., 2001, 2004). These findings underscore the importance for more preclinical work to examine cocainerelated effects on noradrenergic system ontogeny.

In utero cocaine exposure and immediate early genes In the mature brain exposed chronically to cocaine (and other drugs of abuse), neuronal functions gradually change (e.g., tolerance) and these changes may endure even after cessation of drug use. Such delayed, gradually developing,

660

and persistent changes in function suggest that a candidate mechanism for addictive phenomena may be long-term changes in neuronal gene expression (Nestler et al., 1993). Similarly, in the developing brain, changes that persist beyond the acute prenatal or preweaning exposure as described earlier (e.g., changes in receptor density, expressions of neural growth factors, and connecting fiber density) suggest alterations in the genetic programs for ongoing neural function and development. In the mature brain, genes regulate the synthesis of neuropeptides, monoamines, their receptors, transporters, second messengers, and so on. The expression of specific genes in the central nervous system appears regulated by a class of DNA-binding proteins termed transcription factors. These transcription factors bind to the regulatory regions of certain genes and affect the rate at which these genes are transcribed (and thus the synthesis of the relevant synaptic components). These transcription factors, called as a group immediate early genes (IEGs; Sheng and Greenberg, 1990) appear rapidly (within minutes) in response to neuronal stimulation. The mRNAs transcribed from IEGs often have a very short half-life (i.e., the induction of these genes is short-lived). For example, the IEG c-fos can be undetectable within 0.5 hour of stimulus induction (Sheng and Greenberg, 1990). For the induction of c-fos, c-jun, and zif/268 in response to novel stimulation, mRNA transcription occurs within 5 minutes, with peak steady-state levels of mRNA at 30–45 minutes and peak protein synthesis within 2 hours of stimulation (Kosofsky and Hyman, 1993). Immediate early genes may activate or repress the rate at which other genes are transcribed, though the mechanism for this regulation of transcription rate is not yet clear (Plashne and Gann, 1990). The transient nature of the mRNAs from induction of IEGs suggests a complicated stimulus-dependent mechanism for regulating such events as the synthesis of receptors and transporters. The IEG response can also be used as a marker of neuronal activation (Sheng and Greenberg, 1990) and provides information complementary to that of regional glucose metabolism, which predominantly labels nerve terminals activated by a stimulus or drug. Drugs of abuse affect the expression of IEGs in the mature and the developing animal, and the effect is apparent during and after treatment. This relation may be one mechanism to explain how substances of abuse alter genetic programs in the different phases of neural ontogeny. Drug-induced alterations in IEG expression can modify neural gene expression during development and thereby alter, perhaps permanently, cellular identity and neuronal repertoire (Kosofsky and Hyman, 1993). In the adult animal, exposure to cocaine enhances the expression of several IEGs including c-fos and zif/268, as well as genes encoding for the neuropeptides substance P and dynorphin, which are involved in the

neurodevelopmental aspects of clinical disorders

predominantly D1 striatonigral dopaminergic pathways (Snyder-Keller and Keller, 1995; Steiner and Gerfen, 1993, 1995) and in limbic areas such as the amygdala (SnyderKeller and Keller, 2001). Similar induction of these same IEGs is seen with D1 agonists (Robertson, Vincent, and Fibiger, 1992). The cocaine-related induction can be blocked with D1 antagonists (Steiner and Gerfen, 1995) and does not occur in a D1-deficient mouse (Drago et al., 1996; Moratella et al., 1996). The cocaine-related effect on the serotonin system (but not the norepinephrine system) also seems related to induction of c-fos and zif/268 (Bhat and Baraban, 1993). Similarly, with chronic exposure there may be long-term perturbations in IEG expression. For example, during withdrawal from cocaine treatment adult rats showed reduction in c-fos expression for several days (Ennulat, Babb, and Cohen, 1994). In the preweaning exposure model, beginning lines of work point to cocaine-related induction of IEG expression following neural stimulation, particularly in the striatum D1rich regions (Kosofsky, Genova, and Hyman, 1995a, 1995b; Kosofsky and Hyman, 1993). On postnatal day 8, induction of c-fos and zif/268 occurs primarily in the striatum, but by day 15 induction of both occurs in striatum and cortex at the same time that corticostriatal connections are rapidly forming (Kosofsky, Genova, and Hyman, 1995a, 1995b). Induction of fos and cellular apoptosis significantly increase in the paraventricular hypothalamus, amygdala, and medial thalamus following prenatal cocaine exposure in rats, mainly as a result of D1 dopaminergic signaling rather than asphyxia (Mitchell and Snyder-Keller, 2003a, 2003b). Expression of c-fos was also induced in the striatum and frontal cortex of rabbits prenatally exposed to cocaine (Tilakaratne, Cai, and Friedman, 2001). The IEG c-fos may play a pivotal role in developmental gene regulation that may be disrupted by excessive induction (Arnauld et al., 1995), and also shows increased induction by other phenomena such as stress, seizures, and ethanol administration (Hiscock et al., 2001; Martinez, Calvo-Torrent, and Herbert, 2002; Mitchell and Snyder-Keller, 2003a). The candidate target genes whose expression may be altered by cocaine-related induction of these IEGs include those involved in growth-factor synthesis, neuropeptide precursors, and monoamine synthesis; and the present challenge is to identify the target genes for the IEGs in the developing brain (Nestler, Gao, and Tamminga, 1997). Mistimed induction of IEGs with the subsequent alteration of the expression of a cascade of related genes may have continued effects on the genetically coordinated development of neural circuitry. Administration of cocaine in vitro and in cultured cells inhibits activation of cAMP response elementbinding protein (CREB) and suppresses gene expression of brain-derived neurotrophic factor (BDNF; Feng, Yan, and Yan, 2006). The growth factor BDNF plays an important

role in the development and function of dopamine and serotonin neurons, acting as a neurotrophic factor during neurogenesis (Zhou, Bradford, and Stern, 1994; Zhou and Iacovitti, 2000), whereas CREB is an essential regulator of BDNF-induced gene expression (Finkbeiner et al., 1997). Gene expression mediated by CREB has been shown to be necessary for survival of numerous neuronal subtypes. The proliferation of newborn cells in the hippocampus of adult mice is increased following activation of CREB and is reduced in mice that express a negative mutant of CREB (Nakagawa et al., 2002). Glycogen synthase kinase-3β (GSK3β) is abundant in brain neurons and is proposed to regulate neuronal differentiation and neurite growth, mainly through the phosphorylation of IEGs (Goold and Gordon-Weeks, 2001; Leroy and Brion, 1999; Spear, Kirstein, and Frambes, 1989). Prenatal cocaine exposure selectively reduces basal GSK3β activity in the frontal cortex and striatum of 20-day-old rabbits (Gil, Zhen, and Friedman, 2003), possibly suggesting an ensuing delay in the phosphorylation of transcription factors necessary for gene expression. Chronic cocaine exposure is capable of affecting the expression of multiple gene-encoding proteins belonging to the Wnt and cadherin systems in the fetal cerebral wall (Novikova et al., 2005). The Wnt-b-catenin-induced gene transcription is involved in regulation of cell proliferation and differentiation of progenitor cells in the fetal brain (Chenn and Walsh, 2002; Galceran et al., 2000; Hirabayashi et al., 2004; Megason and McMahon, 2002; Zechner et al., 2003), and several reports have described abnormalities in the cortical architecture of humans and animals with altered gene expressions in the Wnt/cadherin network (Campos, Du, and Li, 2004; Cotter et al., 1999; Ligon et al., 2003; Machon et al., 2003). These morphologic disruptions are notably similar to those previously described in prenatally cocaine-exposed animals (Gressens, Kosofsky, and Evrard, 1992; Kosofsky, Genova, and Hyman, 1995a; Ren et al., 2004). Mistimed induction of IEGs with the subsequent alteration of the expression of a cascade of related genes may have continued effects on the genetically coordinated development of neural circuitry. It also may be that long-term exposure, as is the case in the human model, may alter the relation between stimulation and IEG expression such that IEGs are either over- or underexpressed with stimulation long after exposure has ceased. Such a possibility would have profound implications for ongoing neural development during the synaptogenic phase. These are lines of work yet to be extensively pursued.

Behavioral correlates in preclinical models The neurobehavioral areas most frequently studied are (1) locomotor behavior in open-field exploration and

langlois and mayes: prenatal cocaine exposure

661

responses to monoaminergic agonists and antagonists; (2) attention and conditioned learning; and (3) stress reactivity. In each of these areas, a pattern of neurobehavioral correlates has emerged. Several laboratories studying cocaine and monoaminergic function in the rat, mouse, rabbit, or rhesus monkey are also examining related behavioral alterations in those domains of behavior most likely related to alterations of mesolimbic or nigrostriatal systems. These include changes in locomotor behavior and in attention. The findings regarding locomotion are mixed depending on the assessment paradigm, the exposure period, and the time of assessment. Prenatal exposure appears to increase spontaneous locomotor activity in open-field exploration during the first 3 weeks of life but not by 60 days of age (Henderson and McMillen, 1990). Others have reported the same lack of relation between prenatal exposure and locomotor activity after the first weeks of life (Peris, Coleman, and Millard, 1992; Riley and Foss, 1991a, 1991b). Female rats treated on postnatal days 1 through 10 and studied between 60 and 65 days of age show decreased levels of activity (Dow-Edwards, 1989). Prenatally exposed animals also inhibit approaching novel conditions and open-field exploration ( Johns et al., 1992); and wall-climbing behavior, a dopaminergic regulated behavior, is attenuated with foot shock in prenatally exposed animals (Spear, Kirstein, Frambes, and Moody, 1989). Exposure on postnatal days 11–20 reduced the locomotor response to amphetamine challenge in rats (Higgins et al., 1991). Similarly, amphetamine challenge in prenatally exposed rabbits failed to elicit a stereotypic head bobbing, usually associated with activation of the D1 receptor, and this effect persisted through adulthood (Levitt et al., 1997). The aforementioned hypothesis regarding the uncoupling of the D1 receptor rendering it dysfunctional has support in related behavioral studies. For one, D1 receptor activation regulates the increased locomotor response to cocaine (Cabib et al., 1991). Reduced D1 activity would be expected to reduce behaviors usually stimulated by dopamine agonists, especially selective D1 agonists. Mice that are deficient in D1 do not show the usual increase in locomotor behavior on cocaine challenge (Drago et al., 1996), an observation perhaps consistent with the reduced response of pre- and postnatal cocaine-exposed rabbit and rat, respectively, to amphetamine challenge (Hughes et al., 1991). Also, D1deficient mice show reduced rearing (or vertical exploratory) behavior, a complex motor activity that is an essential part of an exploratory repertoire (Drago et al., 1996) and the reduction of which is similar to behavioral patterns seen in prenatally exposed animals. While other motor behaviors are variously altered in D1-deficient mice (including overall activity level, grooming, and sniffing; Xu et al., 1994), rearing activity during exploration may be a more sensitive behavioral marker of D1 receptor dysfunction (Drago et al., 1996).

662

Also, studies of locomotor response to novelty suggest that high-responding animals show greater D1 binding in the nucleus accumbens and less D2 binding in both accumbens and striatum (Hooks et al., 1994), again a possible parallel with the reduced open-field exploration behaviors of cocaine-exposed animals. Second, in rodents, cocaine or stimulant administration in adulthood elicits a number of characteristic responses, including head bobbing, increased sniffing, and activity, behaviors mediated in large part through the D1 system. Selective D1 antagonists abolish this characteristic behavioral response to cocaine, as does prenatal cocaine exposure, presumably at least in part through the decoupling effect on the D1-GαS system. Similarly, behavioral responses to agonists selective for the D1 receptor are reduced following both prenatal and postnatal exposure. There may also be evidence suggestive of a role for the D1 uncoupling in the learning deficits. Prenatally and/or preweaning exposed animals inhibit approaching novel conditions and open-field exploration ( Johns et al., 1992). Deficits in learning are suggested by impairments in classical conditioning (Brunzell et al., 2002; Heyser et al., 1990; Spear, Kirstein, Bell, et al., 1989), deficiencies in active and passive avoidance tasks (Heyser et al., 1990), poor short-term memory, increased response perseveration, diminished proximal cue learning (Vorhees et al., 1995), impaired learning on serial reversal tasks (Garavan et al., 2000), impaired habituation (Heyser et al., 1994), impaired reversal learning (Chelonis, Gillam, and Paule, 2003), increased susceptibility to distractors in a visual attention task (Gendle et al., 2004), deficits in spontaneous alternation behavior (Thompson, Levitt, and Stanwood, 2005), and less efficient error detection (Morgan et al., 2002). Prenatally exposed animals appear unable to attend preferentially to less salient but relevant stimuli in the context of more salient but distracting and nonrelevant background stimuli (Gabriel, Taylor, and Burhans, 2003; He et al., 2004; Romano and Harvey, 1998) and have difficulty changing previously learned conditions (Heyser, Spear, and Spear, 1992; Kosofsky and Wilkins, 1998). This collection of findings cuts across prenatal and preweaning exposure. Prenatally exposed rabbits show pronounced anatomical changes in the anterior cingulate cortex (a richly dopaminergic region). Specifically, neurons in this region show marked increase in dendrite outgrowth (Levitt et al., 1997; Stanwood et al., 2001), a deficit possibly connected to the decoupling of the D1 receptor inasmuch as D1 receptor activation in normal animals decreases dendritic outgrowth (Reinoso, Undie, and Levitt, 1996). Indeed, D1 receptor–G protein coupling has been shown to be reduced in the cingulate cortex in prenatally exposed animals (Friedman, Yadin, and Wang, 1996; H.-Y. Wang, Yeung, and Friedman, 1995).

neurodevelopmental aspects of clinical disorders

Also, the anterior cingulate cortex plays a central role in processes of learning and memory and is especially crucial to situations that demand attention preferentially to less salient but relevant stimuli when more salient, but not necessarily relevant, stimuli occur in the same context. Lesions in the anterior cingulate cortex impair attentional processes, and in intact animals attentional tasks lead to activation of this region (Vogt and Gabriel, 1984). Uncoupling the D1 receptor from its second-messenger systems dampens D1-mediated responses within striatal neurons and within the cingulate-cortex-regulated functions and impairs learning (Gabriel and Taylor, 1998). Prenatally cocaine-exposed rabbits with changes in the anterior cingulate region show reduced learning ability when the positive conditioning stimulus is less salient than the negative stimulus (e.g., a soft tone versus a loud tone) (Levitt et al., 1997; Romano and Harvey, 1998). In other words, these animals cannot attend preferentially to less salient but relevant stimuli in the context of more salient background (and irrelevant) stimuli. These deficits persisted into adulthood (Romano and Harvey, 1998). Other studies have also suggested an association between prenatal cocaine exposure and impaired attentional processes that impede the animal’s ability to change previously learned conditions or to screen out distracting but nonrelevant information (Heyser, Spear, and Spear, 1992; Kosofsky and Wilkins, 1998). Findings of this nature parallel reports from studies of prenatally cocaine-exposed preschool and school-aged children suggesting deficits in selective attention and in information processing. Altered stress responsivity is suggested by several behavioral observations and is the domain most relevant to the earlier suggestion that early compensatory mechanisms that “normalize” basal conditions may nonetheless leave the adult animal with a vulnerability to novel, challenging, or stressful contexts (Spear et al., 1998). While neural “stress circuitry” is complex and involves many interactive systems, including the corticosteroid-based HPA axis, dopaminergic and noradrenergic pathways in the amygdala and other limbic and corticothalamic systems are central (Dunn and Kramarcy, 1984). Furthermore, stress selectively stimulates many of the same dopaminergic regions as does cocaine. Prenatally exposed rats, while initially showing less motor response to open-field exploration, engage in exploration after a delay and then do so with markedly increased activity suggesting overarousal (Henderson and Fuller, 1992; Hutchings, Fico, and Dow-Edwards, 1989; Spear, Kirstein, and Frambes, 1989). Similarly, adult rats exposed to cocaine prenatally demonstrate less expected stress-induced immobility during a forced swim or intermittent foot shock, or when placed in a novel condition following foot shock (Bilitzke and Church, 1992; McMillen et al., 1991; Molina, Wagner, and Spear, 1994; Smith et al., 1989; Spear et al.,

1998; Spear, Kirstein, and Frambes, 1989). This alteration in stress responsivity, as manifested by immobility, changes with age. Prepubertal or preweaning animals exposed prenatally to cocaine show greater immobility in such stressful conditions as forced swim or foot shock (Goodwin et al., 1997; Molina, Wagner, and Spear, 1994; Wood et al., 1995). Other behaviors suggest increased arousal and altered stress responsivity. Prenatally exposed animals are more sensitive to environmental demands and show more aggression than nontreated animals in conditions of water deprivation and competition for water (Spear et al., 1998; Wood and Spear, 1998). Increased aggression is not apparent, however, if the animals are not stressed (Estelles et al., 2005; Spear et al., 1998). Far more work is required to link these behavioral observations of altered stress responsivity with the alterations in neurochemical function related to prenatal cocaine exposure. Finally, there is some suggestion that prenatally cocaineexposed animals may be more sensitive to the effects of early experience (stressful and nonstressful). With chronic stress (such as repeated foot shock over time), preadolescent cocaine-exposed animals “reverse” their acute stressor pattern of reduced immobility and show more immobility, as is expected in nonexposed animals (Goodwin et al., 1997). Nonexposed but repeatedly stressed animals show less significant changes in the pattern of their stress response. Findings such as these about the effects of “early experience” point to potentially productive lines of work using more traditional early experiences, such as increased handling, that have been studied in relation to adrenocortical function and stress (Meaney et al., 1989; Spear et al., 1998).

Human model of prenatal exposure Outcomes of prenatally cocaine-exposed children seen through adolescence range from assessments of physical growth and motoric maturation to standardized assessments of cognition and language, neurophysiological indices of arousal and sensory response, emotional regulation, maladaptive behavior, and a range of executive control functions. However, in both extensive descriptive reviews and meta-analyses, appropriate concerns have been raised about singularly attributing any one disruption in developmental capacities to the prenatal effects of cocaine on emerging neural systems (Bland-Stewart et al., 1998; Brooks-Gunn, McCarton, and Hawley, 1994a; Chiriboga, 1998; Frank et al., 2001; Lester, Freier, and LaGasse, 1995; Malakoff et al., 1999; Mayes and Fahy, 2001; Neuspiel, 1995; Olson and Toth, 1999; Singer, 1999; Tronick and Beeghly, 1999). Prenatal exposure to cocaine is often paired with a simultaneous exposure to other drugs and to severe postnatal environmental deprivation (Frank et al., 2001; Held, Riggs, and Dorman, 1999; Lagasse, Seifer, and Lester, 1999; Lutiger

langlois and mayes: prenatal cocaine exposure

663

et al., 1991; Tronick and Beeghly, 1999; Vogel, 1997). Only a few studies have addressed apparent cocaine-related effects on children’s development and behavior that are moderated or mediated (Baron and Kenny, 1986) by way of pathways of maternal health or quality of caregiving (Black, Schuler, and Nair, 1993; Singer et al., 1997), or have taken into consideration parental attentional or learning difficulties that may serve as markers of possible genetic contributions to infant functioning (Mayes and Bornstein, 1995). Nonetheless, bearing in mind these cautions, a number of findings in human studies have appeared across laboratories and are consistent with those in preclinical models. Neurochemical Findings in Infants and Children In human infants, neurochemical studies following prenatal exposure are scant, but two areas of work point to effects on the neurochemical systems that are most closely related to arousal regulation and altered response to stress. Three studies have examined metabolites, precursors, or levels of norepinephrine or dopamine in the serum or cerebrospinal fluid of newborns exposed to cocaine prenatally and have found a significant increase in plasma norepinephrine (Davidson Ward et al., 1991) and catecholamine precursors in CSF (Mirochnick et al., 1991) and a decrease in the metabolites of dopamine (Needlman et al., 1993), findings suggesting an alteration in monoaminergic system function at least neonatally. Further, for the cocaine-exposed infants, norepinephrine levels were inversely related to features of the infant’s neurobehavioral profile (Mirochnick et al., 1991). Second, cocaine-exposed infants exhibit an attenuated cortisol response to noninvasive (neurobehavioral examination) and invasive (heel-stick) manipulations despite showing no differences in baseline cortisol levels (Mangano, Gardner, and Karmel, 1992), suggesting that glucocorticoid-mediated arousal regulatory systems are altered by in utero cocaine exposure or by the chronically stressful conditions associated with cocaine exposure (Karmel, Gardner, and Magnano, 1991). Neurobehavioral and Developmental Findings There are at least two pathways by which disruptions in monoaminergic system ontogeny might affect neurocognitive functioning in prenatally exposed children (Mayes et al., 2003). One relates to arousal-modulated attention and related cognitive functioning (Mayes and Fahy, 2001; Mayes et al., 1998). In this model, neurocognitive impairments in a range of prefrontal cortical functions are expected especially as children are more emotionally aroused. The second model is based on a direct effect of cocaine on cortical morphology and specialization (He and Lidow, 2004; Lidow, 1995; Lidow, Bozian, and Song, 2001; Lidow and Song, 2001a). In this model, neurocognitive deficits might be expected in even nonstressed conditions in a range of

664

functions including impaired inhibition and slowed reaction times (Lidow, 2003). Findings from several groups are consistent with both these models. A number of investigators have reported difficulties among prenatally cocaine-exposed children in arousal-regulatory capacities ranging from increased excitability with poor state regulation and rapid changes in arousal with stimulation (measured by changes in heart rate), to increased arousal from sleep, increased EEG asymmetry, and greater physiological lability (Alessandri, Bendersky, and Lewis, 1998; Brown et al., 1998; DiPietro et al., 1995; Gingras et al., 1995; N. Jones, et al., 2004; Karmel, Gardner, and Freedland, 1996; Mayes et al., 1995, 1996; Regaldo et al., 1995, 1996), with some findings persisting at least through one year of follow-up (Bendersky and Lewis, 1998; Kaplan, 1995; Scher, Richardson, and Day, 2000). These differences in behavioral and physiological arousal are expanded upon by a few studies examining the relation between arousal-modulated attention and impaired information processing in infants and toddlers, including diminished responsiveness to novel stimuli and recognition memory, poor impulse control and task persistence, diminished sustained attention, and greater emotional lability and/or behavioral disorganization (Alessandri et al., 1993; Alessandri, Bendersky, and Lewis, 1998; Azuma and Chasnoff, 1993; Coles et al., 1999; Gaultney, Gingras, and Martin, 2005; Griffith, Azuma, and Chasnoff, 1994; Hansen, Struthers, and Gospe, 1993; Heffelfinger et al., 2002; S. Jacobson, Bihun, and Chiodo, 1999; S. Jacobson et al., 1996; Lewis et al., 2004; Mayes et al., 1998; Metosky and Vondra, 1995; Singer et al., 1999; Struthers and Hansen, 1992), as well as other aspects of the stress response system, including the studies cited earlier of baseline and peak cortisol response to stressors in which cocaine-exposed infants show a depressed cortisol response to challenge (Mangano, Gardner, and Karmel, 1992) or lower baseline levels (S. Jacobson, Bihun, and Chiodo, 1999). Thus evidence suggests that cocaine-exposed children may be more easily aroused and hence vulnerable to difficulties in information processing and learning in novel conditions consistent with one model of the developmental impact of prenatal cocaine exposure. At the same time, a number of labs are also reporting neurocognitive deficits under even optimal testing conditions using narrow-band assessments to test specific functions such as task switching and inhibition. Collected findings include slower reaction times to visual stimulus presentation (Heffelfinger, Craft, and Shyken, 1997), slowed reaction times in continuous performance tasks (Eghbalieh, Crinella, Hunt, and Swanson, 2000), greater perseveration, diminished response inhibition (Bendersky et al., 2003; Bendersky and Lewis, 1998), increased errors of commission or omission on A-not-B or continuous performance tasks (Eghbalieh, Crinella, and Swanson, 2000; Espy,

neurodevelopmental aspects of clinical disorders

Kaufmann, and Glisky, 1999; Mayes and Fisher, 1998; Richardson, Conroy, and Day, 1996), diminished capacity for sustained attention (Bandstra and Burkett, 1991; Schiller and Allen, 2005), and deficits in spatial learning (Schroder et al., 2004). The range of attentional/executive-control-function effects that seem related to prenatal cocaine exposure may be consistent with preclinical work suggesting a cocainerelated effect on the development of the anterior cingulate cortex (Stanwood et al., 2001) and, as discussed in earlier sections, on dopaminergic D1/D2-mediated circuitry critical to regulation of prefrontal cortex (Seamans, Floresco, and Phillips, 1998) and to up-regulation of noradrenergic (and serotonergic) systems at a time in early development that could permanently distort the excitatory/inhibitory balance between regulatory systems necessary for optimal cortical function. Consistent with findings discussed earlier regarding the decoupling of the D1 system, in preclinical models, behavioral responses to agonists selective for the D1 receptor are reduced following both prenatal and postnatal exposure (Gilde, 1994). Similarly, differential reduction in cingulate activity in cocaine-exposed animals is associated with an impaired ability to discriminate among salient and/ or relevant stimuli (Gabriel and Taylor, 1998; Romano and Harvey, 1998), a finding consistent with emerging data cited previously in human studies. Also, because, as cited before, cocaine-associated disruptions in neuronal migration with attendant disruptions in cortical lamination (He et al., 2004; Lidow, 1995, 2003; Lidow and Rakic, 1995; Lidow and Song, 2001b) might interfere with regional specialization, performance on certain cortically related tasks might be slowed (Lidow, 2003). This suggestion is supported by findings from prenatally cocaine-exposed monkeys followed through six years of age who were tested on their ability to learn a new set of responses following a change in the rules of reinforcement for a simple visual discrimination task (Chelonis, Gillam, and Paule, 2003). The task required inhibition of a previously learned and rewarding response. The prenatally cocaine-exposed animals performed much more poorly; that is, they took more sessions to attain or never attained pre-reversal type responding when compared to the drug-naive control animals. These data are also consistent with findings from human studies using eventrelated-potential methods in which prenatally cocaineexposed 7- to 9-year-old children used more diffuse regions of the cortex to complete a simple response-inhibition task (Mayes et al., 2003). Even though the children were able to respond correctly in an untimed testing situation, their responses were slower, and they required more time to complete the task, as might be considered consonant with more diffuse activation of the cortex. Although the findings in overall cognitive development in preschool and school-aged children are inconsistent (Hurt

et al., 1997; Phelps and Cottone, 1999; Richardson, Conroy, and Day, 1996), the evidence increasingly documents somewhat greater cognitive impairments in cocaine-exposed children after controlling for other drug exposures (Alessandri, Bendersky, and Lewis, 1998; Arendt et al., 2004; Koren et al., 1998; Richardson, 1998; Singer et al., 2004). While the effects on IQ may be subtle (Arendt et al., 2004; Lester, LaGasse, and Seifer, 1998; Singer et al., 2004), moderate to severe delays and impairments in both expressive and receptive language development are more consistently reported (Angelilli et al., 1994; Beeghly et al., 2006; Bland-Stewart et al., 1998; Delaney-Black et al., 1998, 2000; Johnson et al., 1997; Lewis et al., 2004; Madison et al., 1998; Malakoff et al., 1999; Malakoff, Mayes, and Schottenfeld, 1994; Morrow, George, and Roth, 2004; Morrow, Elsworth, and Roth, 2003; van Baar, 1990; van Baar and de Graaff, 1994). Some of these investigations include controls for environmental stimulation. Studies of developmental language differences among cocaine-exposed children are also beginning to focus on more specific outcome measures, such as differences in phonological patterns (Madison et al., 1998), that are less confounded by environmental deprivation. In terms of behavior and school performance, a few studies document a suggested increase in behavioral and academic problems among cocaine-exposed children, though these findings are confounded by reporting bias of often nonblinded observers (Delaney-Black et al., 2004, 1998). Importantly, several investigators have reported apparent cocaine-related effects on children’s development and behavior that are moderated or mediated (Baron and Kenny, 1986) by means of pathways of maternal health or quality of caregiving (Black, Schuler, and Nair, 1993; Singer et al., 1997). Conversely, others have examined infants exposed prenatally who were adopted into stable environments and have documented apparent direct effects of the prenatal exposure on neurocognitive outcomes (Koren et al., 1998; Nulman et al., 1994). Also, several studies have differentiated between heavy and moderate to light exposure and found dose-response outcomes. Several studies have also documented dose-response effects on attention control, executive function, and general cognitive deficits (Alessandri, Bendersky, and Lewis, 1998; Bendersky and Lewis, 1998; Coles et al., 1992; Delaney-Black et al., 2004; S. Jacobson et al., 1999, 1996; Morrow, George, and Roth, 2004; Swanson et al., 1999), as well as an interactive effect with the level of environmental risk (Bendersky et al., 2003). Thus, despite a number of methodological concerns and still scant data for school-aged children, the profile emerging for children prenatally exposed to cocaine suggests a range of neurocognitive, neuropsychological dysfunction that may be a combination of effects, mediated partially through adverse environment, partially through the perinatal impact

langlois and mayes: prenatal cocaine exposure

665

of the exposure, and partially through a direct effect on neural structure-function relationships.

Summary While far more work is needed at the preclinical level, findings are accumulating to suggest a targeted and enduring effect of prenatal cocaine exposure on key aspects of neural ontogeny and function. Early gestation exposure to cocaine may affect the most basic processes of neuronal proliferation and migration, whereas later exposure influences neuronal maturation and synaptogenesis. It is probable that there are differential effects on monoaminergic and GABA systems and that impairment in one system may be compensated for through up- or down-regulation of another. This interactive and compensatory mechanism may account in part for variations in findings depending on the age of the animal at the time of study. The preclinical models are extremely useful for suggesting hypotheses and potentially fruitful avenues of study in the human. Combining prenatal and preweaning exposure models will be important to model the more common human condition of exposure throughout pregnancy. In addition, promising work suggests that with maturation some of the baseline behavioral effects of prenatal exposure may be less apparent but that deficits continue to be manifest under stressful or novel conditions. acknowledgments

The work has been supported by National Institute on Drug Abuse grants ROI-DA-06025 (LCM), KO2-, KO5-DA20091 (LCM), and RO1-DA017863 (LCM). This work also draws extensively on other reviews completed by the second author, including Mayes and Ward (2003) and Mayes and Pajulo (2006).

REFERENCES Aceves, J., B. Floran, D. Martinez-Fong, J. Benitez, A. Sierra, and G. Flores, 1992. Activation of D1 receptors stimulates accumulation of [gamma]-aminobutyric acid in slices of the pars reticulata of 6-hydroxydopamine-lesioned rats. Neurosci. Lett. 145(1):40–42. Akbari, H. M., H. K. Kramer, P. M. Whitaker-Azmitia, L. P. Spear, and E. C. Azmitia, 1992. Prenatal cocaine exposure disrupts the development of the serotonergic system. Brain Res. 572(1–2):57–63. Akbari, H. M., P. M. Whitaker-Azmitia, and E. C. Azmitia, 1994. Prenatal cocaine decreases the trophic factor S-100 beta and induced microcephaly: Reversal by postnatal 5-HT1A receptor agonist [published erratum appears in Neurosci. Lett. 1994 July 4; 175(1–2): 176]. Neurosci. Lett. 170(1):141–144. Alessandri, S. M., M. Bendersky, and M. Lewis, 1998. Cognitive functioning in 8- to 18- month old drug-exposed infants. Dev. Psychol. 34:565–573. Alessandri, S., M. Sullivan, S. Imaizumi, and M. Lewis, 1993. Learning and emotional responsivity in cocaine-exposed infants. Dev. Psychol. 29(6):989–997.

666

Anderson-Brown, T., T. A. Slotkin, and F. J. Seidler, 1990. Cocaine acutely inhibits DNA synthesis in developing rat brain regions: Evidence for direct actions. Brain Res. 537(1–2):197– 202. Angelilli, M. L., H. Fischer, V. Delaney-Black, M. Rubinstein, J. W. Ager, and R. J. Sokol, 1994. History of in utero cocaine exposure in language-delayed children. Clin. Pediatr. (Phila.) 33(9):514–516. Arendt, R. E., E. J. Short, L. T. Singer, S. Minnes, J. Hewitt, S. Flynn, et al., 2004. Children prenatally exposed to cocaine: Developmental outcomes and environmental risks at seven years of age. J. Dev. Behav. Pediatr. 25(2):83–90. Arnauld, E., J. Arsaut, J. A. Tafani, and J. Demotes-Mainard, 1995. Dopaminergic control of gene transcription during striatal ontogeny: c-fos induction by D1 receptor activation in the developing striosomes. Mol. Brain Res. 30(2):223–232. Arnsten, A. F. T., 1997. Catecholamine regulation of the prefrontal cortex. J. Psychopharmacol. 11(2):151–162. Arnsten, A. F., 1998. Catecholamine modulation of prefrontal cortical cognitive function. Trends Cogn. Sci. 280:436–447. Arnsten, A. F., and P. Goldman-Rakic, 1984. Selective prefrontal cortical projections to the region of the locus coeruleus and raphe nuclei in the rhesus monkey. Brain Res. 306:9–18. Arnsten, A. F. T., and J. D. Jentsch, 1997. The alpha-1 adrenergic agonist, cirazoline, impairs spatial working memory performance in aged monkeys. Pharmacol. Biochem. Behav. 58:55– 59. Arnsten, A. F. T., J. C. Steere, D. J. Jentsch, and B. M. Li, 1998. Noradrenergic influences on prefrontal cortical cognitive function: Opposing actions at portjunctional alpha 1 versus alpha 2adrenergic receptors. Adv. Pharmacol. 42:764–767. Azuma, S. D., and I. J. Chasnoff, 1993. Outcome of children prenatally exposed to cocaine and other drugs: A path analysis of three-year data. Pediatrics 92(3):396–402. Bandstra, E. S., and G. Burkett, 1991. Maternal-fetal and neonatal effects of in utero cocaine exposure. Sem. Perinatol. 15(4): 288–301. Baron, R. M., and D. A. Kenny, 1986. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. J. Pers. Soc. Psychol. 51(6): 1173–1182. Barron, S., C. M. Staton, J. A. Willford, and L. S. Hansen-Trench, 1998. The effects of clonidine on rat pups neonatally exposed to cocaine. Neurotoxicol. Teratol. 20:401–410. Bastianetto, S., L. Rouquier, G. Perrault, and D. J. Sanger, 1995. DTG-induced circling behaviour in rats may involve the interaction between [sigma] sites and nigro-striatal dopaminergic pathways. Neuropharmacology 34(3):281–287. Battaglia, G., and T. M. Cabrera, 1994. Potentiation of 5-HT1A receptor-mediated neuroendocrine responses in male but not female rat progeny after prenatal cocaine: Evidence for gender differences. J. Pharmacol. Exp. Ther. 271(3):1453–1461. Battaglia, G., T. M. Cabrera-Vera, L. D. Van de Kar, F. Garcia, A. Vicentic, and W. Pinto, 1998. Prenatal cocaine exposure produces long-term impairments in brain serotonin function in rat offspring. Ann. NY Acad. Sci. 846:355–357. Battaglia, G., T. M. Cabrera-Vera, and L. D. Van de Kar, 2000. Prenatal cocaine exposure potentiates 5-HT1a receptor function in male and female rat offspring. Synapse 35(3):163– 172. Beauregard, M., and A. Ferron, 1991. Dopamine modulates the inhibition induced by GABA in rat cerebral cortex: An iontophoretic study. Eur. J. Pharmacol. 205(3):225–231.

neurodevelopmental aspects of clinical disorders

Beeghly, M., B. Martin, R. Rose-Jacobs, H. Cabral, T. Heeren, M. Augustyn, et al., 2006. Prenatal cocaine exposure and children’s language functioning at 6 and 9.5 years: Moderating effects of child age, birthweight, and gender. J. Pediatr. Psychol. 31(1):98–115. Bendersky, M., G. Gambini, A. Lastella, D. Bennett, and M. Lewis, 2003. Inhibitory motor control at five years as a function of prenatal cocaine exposure. J. Dev. Behav. Pediatr. 24(5):345–351. Bendersky, M., and M. Lewis, 1998. Prenatal cocaine exposure and impulse control at two years. Ann. NY Acad. Sci. 846:365– 367. Bergson, C., L. Mrzljak, J. F. Smiley, M. Pappy, R. Levenson, and P. S. Goldman-Rakic, 1995. Regional, cellular, and subcellular variations in the distribution of D1 and D5 dopamine receptors in primate brain. J. Neurosci. 15(12):7821–7836. Bernath, S., and M. J. Zigmond, 1989. Dopamine may influence striatal GABA release via three separate mechanisms. Brain Res. 476(2):373–376. Bhat, R. V., and J. M. Baraban, 1993. Activation of transcription factor genes in striatum by cocaine: Role of both serotonin and dopamine systems. J. Pharmacol. Exp. Ther. 267:496–505. Bilitzke, P. J., and M. W. Church, 1992. Prenatal cocaine and alcohol exposures affect rat behavior in a stress test (the Porsolt swim test). Neurotoxicol. Teratol. 14(5):359–364. Black, M., M. Schuler, and P. Nair, 1993. Prenatal drug exposure: Neurodevelopmental outcome and parenting environment. J. Pediatr. Psychol. 18(5):605–620. Bland-Stewart, L. M., H. N. Seymour, M. Beeghley, and D. A. Frank, 1998. Semantic development of African-American children prenatally exposed to cocaine. Sem. Speech Lang. 19:167– 187. Blaustein, M. P., 1988. Calcium transport and buffering in neurons. Trends Neurosci. 11(10):438–443. Bonci, A., and J. T. Williams, 1996. A common mechanism mediates long-term changes in synaptic transmission after chronic cocaine and morphine. Neuron 16:631–639. Booze, R., D. Wallace, J. Silvers, B. Strupp, D. Snow, and C. Mactutus, 2006. Prenatal cocaine exposure alters alpha2 receptor expression in adolescent rats. BMC Neurosci. 7(1):33. Borella, A., M. Bindra, and P. M. Whitaker-Azmitia, 1997. Role of the 5-HTIA receptor in development of the neonatal rat brain: Preliminary behavioral studies. Neuropharmacology 36:445– 450. Brooks-Gunn, J., C. McCarton, and T. Hawley, 1994a. Effects of in utero drug exposure on children’s development: Review and recommendations. Arch. Pediatr. Adolesc. Med. 148(1): 33–39. Brooks-Gunn, J., C. McCarton, and T. Hawley, 1994b. Effects of in utero drug exposure on children’s development. Arch. Pediatr. Adolesc. Med. 148:33–39. Brown, J. V., R. Bakeman, C. D. Coles, W. R. Sexson, and A. S. Demi, 1998. Maternal drug use during pregnancy: Are preterm and full-term infants affected differently? Dev. Psychol. 34(3):540–554. Brunzell, D. H., A. E. Coy, J. J. B. Ayres, and J. S. Meyer, 2002. Prenatal cocaine effects on fear conditioning: Exaggeration of sex-dependent context extinction. Neurotoxicol. Teratol. 24(2): 161–172. Cabib, S., C. Castellano, V. Cestari, U. Filibeck, and S. Puglisi-Allegra, 1991. D1 and D2 receptor antagonists differently affect cocaine-induced locomotor hyperactivity in the mouse. Psychopharmacology 105:335–339.

Cabrera, R., A. Diaz, A. Pinter, and J. Belmar, 1993. In vitro progesterone effects on 3H-dopamine release from rat corpus striatum slices obtained under different endocrine conditions. Life Sci. 53(23):1767–1777. Cabrera-Vera, T. M., F. Garcia, W. Pinto, and G. Battaglia, 2000. Neurochemical changes in brain serotonin neurons in immature and adult offspring prenatally exposed to cocaine. Brain Res. 870(1–2):1–9. Cameron, D. L., and J. T. Williams, 1993. Dopamine D1 receptors facilitate transmitter release. Nature 366(6453):344–347. Cameron, D. L., and J. T. Williams, 1994. Cocaine inhibits GABA release in the VTA through endogenous 5-HT. J. Neurosci. 14(11 Part 1):6763–6767. Campbell, K., P. Kalen, K. Wictorin, C. Lundberg, R. J. Mandel, and A. Bjorklund, 1993. Characterization of GABA release from intrastriatal striatal transplants: Dependence on host-derived afferents. Neuroscience 53(2):403–415. Campos, V. E., M. Du, and Y. Li, 2004. Increased seizure susceptibility and cortical malformation in [beta]-catenin mutant mice. Biochem. Biophys. Res. Commun. 320(2):606–614. Chelonis, J. J., M. P. Gillam, and M. G. Paule, 2003. The effects of prenatal cocaine exposure on reversal learning using a simple visual discrimination task in rhesus monkeys. Neurotoxicol. Teratol. 25(4):437–446. Chen, Z., J. Tetzlaff, K. Sripathirathan, G. A. Carrasco, M. Shankaran, L. D. Van de Kar, et al., 2005. Paroxetine is effective in desensitizing 5-HT1A receptor function in adult offspring exposed prenatally to cocaine. Psychopharmacology V180(2):316–326. Chen, Z., K. Waimey, L. D. Van de Kar, G. A. Carrasco, M. Landry, and G. Battaglia, 2004. Prenatal cocaine exposure potentiates paroxetine-induced desensitization of 5-HT2A receptor function in adult male rat offspring. Neuropharmacology 46(7):942–953. Cheng, L., K. Hamaguchi, M. Ogawa, S. Hamada, and N. Okado, 1993. PCPA reduces both monoaminergic afferents and nonmonoaminergic synapses in the cerebral cortex. Neurosci. Res. 19:111–116. Chenn, A., and C. A. Walsh, 2002. Regulation of cerebral cortical size by control of cell cycle exit in neural precursors. Science 297(5580):365–369. Chiriboga, C. A., 1998. Neurological correlates of fetal cocaine. Ann. NY Acad. Sci. 846:109–125. Chubakov, A. R., V. G. Tsyganova, and E. F. Sarkisova, 1993. The stimulating influence of the raphe nuclei on the morphofunctional development of the hippocampus during their combined cultivation. Neurosci. Behav. Physiol. 23:271–276. Clarke, C., K. Clarke, J. Muneyyirci, E. Azmitia, and P. M. Whitaker-Azmitia, 1996. Prenatal cocaine delays astroglial maturation: Immunodensitometry shows increased markers of immaturity (vimentin and GAP-43) and decreased proliferation and production of growth factor S-100. Dev. Brain Res. 91:268– 273. Coles, C. D., K. A. Bard, K. A. Platzman, and M. E. Lynch, 1999. Attentional responses at eight weeks in prenatally drugexposed and preterm infants. Neurotoxicol. Teratol. 21:527–537. Coles, C. D., K. A. Platzman, I. Smith, M. E. James, and A. Falek, 1992. Effects of cocaine and alcohol use in pregnancy on neonatal growth and neurobehavioral status. Neurotoxicol. Teratol. 14(1):23–33. Cotter, D., H. Mrinalini, L. Simon, R. Lucy, K. Robert, A. Brian, et al., 1999. Disturbance of Notch-1 and Wnt signalling proteins in neuroglial balloon cells and abnormal large neurons

langlois and mayes: prenatal cocaine exposure

667

in focal cortical dysplasia in human cortex. Acta Neuropathol. (Berl.) V98(5):465–472. Davidson Ward, S., S. Schuetz, L. Wachsman, X. Bean, D. Bautista, S. Buckley, et al., 1991. Elevated plasma norepinephrine levels in infants of substance-abusing mothers. Am. J. Dis. Child. 145:44–48. DeGeorge, G., S. Hochberg, R. Murphy, and E. Azmitia, 1989. Serotonin neurons may be a major target for the action of cocaine: Evidence from in vitro and tissue culture studies. Soc. Neurosci. Abstracts 15:418. Delagrange, P., M. H. Canu, A. Rougeul, P. Buser, and J. J. Bouyer, 1993. Effects of locus coeruleus lesions on vigilance and attentive behavior in cat. Behav. Brain Res. 53:155–165. Delaney-Black, V., C. Covington, B. Nordstrom, J. Ager, J. Janisse, J. H. Hannigan, et al., 2004. Prenatal cocaine: Quantity of exposure and gender moderation. J. Dev. Behav. Pediatr. 25(4):254–264. Delaney-Black, V., C. Covington, T. Templin, J. Ager, S. Martier, and R. Sokol, 1998. Prenatal cocaine exposure and child behavior. Pediatrics 102:945–950. Delaney-Black, V., C. Covington, T. Templin, T. Kershaw, B. Nordstrom-Klee, J. Ager, et al., 2000. Expressive language development of children exposed to cocaine prenatally: Literature review and report of a prospective cohort study. J. Communi. Disord. 33(6):463–481. DiPietro, J. A., P. E. Suess, J. S. Wheeler, P. H. Smouse, and D. B. Newlin, 1995. Reactivity and regulation in cocaineexposed neonates. Infant Behav. Dev. 18:407–414. Ditza, A., J. F. Zachor, T. M. Smoot, and A. K. Percy, 1996. Inhibitory effects of cocaine on NGF-induced neuronal differentiation: Incomplete reversibility after a critical time period. Brain Res. 729(2):270–272. Dow-Edwards, D. L., 1989. Long-term neurochemical and neurobehavioral consequences of cocaine use during pregnancy. Ann. NY Acad. Sci. 562(June):280–289. Dow-Edwards, D., 1993. The puzzle of cocaine’s effects following maternal use during pregnancy: Still unsolved [comment]. Neurotoxicol. Teratol. 15(5):295–296; discussion 311–292. Dow-Edwards, D. L., 1998. Preweaning cocaine administration alters the adult response to quipazine: Comparison with fluoxetine. Neurotoxicol. Teratol. 20:133–142. Dow-Edwards, D. L., L. A. Freed, and T. H. Milhorat, 1988. Stimulation of brain metabolism by perinatal cocaine exposure. Dev. Brain Res. 42:137–141. Dow-Edwards, D. L., L. A. Freed-Malen, and H. E. Hughes, 1993. Long-term alterations in brain function following cocaine administration during the preweanling period. Dev. Brain Res. 72(2):309–313. Dow-Edwards, D. L., X. Shi, R. Yin, and G. Frick, 1998. DA receptors in adult rats exposed to preweaning cocaine or GBR: Effects of chronic haloperidol. CPDD abstract, 81. Dow-Edwards, D. L., R. Yin, and Y. L. Hurd, 1997. Alterations in dopamine transporter mRNA in adult rats following developmental exposure to cocaine. Soc. Neurosci. Abstracts 432.4. Drago, J., C. R. Gerfen, H. Westphal, and H. Steiner, 1996. D1 dopamine receptor–deficient mouse: Cocaine-induced regulation of immediate-early gene and substance P expression in the striatum. Neuroscience 74:813–823. Du, F., T. Eid, E. W. Lothman, C. Kohler, and R. Schwarcz, 1995. Preferential neuronal loss in layer III of the medial entorhinal cortex in rat models of temporal lobe epilepsy. J. Neurosci. 15(10):6301–6313.

668

Dunn, A. J., and N. R. Kramarcy, eds. 1984. Neurochemical Responses in Stress: Relationship between the Hypothalamic-Pituitary-Adrenal and Catecholamine Systems. New York: Plenum. Eghbalieh, B., F. M. Crinella, L. E. Hunt, and J. M. Swanson, 2000. Prenatal stimulant exposure affects alerting and executive control in six- and seven-year-olds. J. Attention Disord. 4(1): 5–13. Eghbalieh, B., L. E. H. Crinella, and J. M. Swanson, 2000. Prenatal stimulant exposure affects alerting and executive control in six- and seven-year-olds. J. Attention Disord. 4:5–13. Ennulat, D. J., S. M. Babb, and B. M. Cohen, 1994. Persistent reduction of immediate early gene mRNA in rat forebrain following single or multiple doses of cocaine. Brain Res. 26:106–112. Espy, K. A., P. M. Kaufmann, and M. L. Glisky, 1999. Neuropsychologic function in toddler exposed to cocaine in utero: A preliminary study. Dev. Neuropsychol. 15:447–460. Estelles, J., M. Rodrigues-Arias, C. Maldonado, M. Aguilar, and J. Minarro, 2005. Prenatal cocaine exposure alters spontaneous and cocaine-induced motor and social behaviors. Neurotoxicol. Teratol. 27:449–457. Eyler, F. D., M. Behnke, C. W. Garvan, N. S. Woods, K. Wobie, and M. Conlon, 2001. Newborn evaluations of toxicity and withdrawal related to prenatal cocaine exposure. Neurotoxicol. Teratol. 23:399–411. Feng, M. J., S. E. Yan, and Q. S. Yan, 2006. Cocaine exposure at a sublethal concentration downregulates CREB functions in cultured neuroblastoma cells. Brain Res. 1077:59–66. Finkbeiner, S., S. F. Tavazoie, A. Maloratsky, K. M. Jacobs, K. M. Harris, and M. E. Greenberg, 1997. CREB: A major mediator of neuronal neurotrophin responses. Neuron 19:1031– 1047. Frank, D. A., M. Augustyn, W. G. Knight, T. Pell, and B. Zuckerman, 2001. Growth, development, and behavior in early childhood following prenatal cocaine exposure: A systematic review. JAMA 285(12):1613–1625. Frank, D. A., M. Augustyn, and B. S. Zuckerman, 1998. Neonatal neurobehavioral and neuroanatomic correlates of prenatal cocaine exposure: Problems of dose and confounding. Ann. NY Acad. Sci. 846(1):40–50. Friedman, E., and H.-Y. Wang, 1998. Prenatal cocaine exposure alters signal transduction in brain D1 dopamine receptor system. Ann. NY Acad. Sci. 846:238–247. Friedman, E., E. Yadin, and H.-Y. Wang, 1996. Effect of prenatal cocaine on dopamine-receptor-G protein coupling in mesocortical regions of the rabbit brain. Neuroscience 70:739–747. Gabbott, P. L. A., and S. J. Bacon, 1996. Local circuit neurons in the medial prefrontal cortex (areas 24a,b,c, 25 and 32) in the monkey: II. Quantitative areal and laminar distributions. J. Comp. Neurol. 364(4):609–636. Gabriel, M., and C. Taylor, 1998. Prenatal exposure to cocaine impairs neuronal coding of attention and discrimination learning. Ann. NY Acad. Sci. 846:194–212. Gabriel, M., C. Taylor, and L. Burhans, 2003. In utero–cocaine, discriminative avoidance learning with low-salient stimuli and learning related neuronal activity in rabbits Oryctolagus cuniculus. Behav. Neurosci. 117(5):912–926. Galceran, J., E. M. Miyashita-Lin, E. Devaney, J. L. Rubenstein, and R. Grosschedl, 2000. Hippocampus development and generation of dentate gyrus granule cells is regulated by LEF1. Development 127(3):469–482. Garavan, H., R. E. Morgan, C. F. Mactutus, D. A. Levitsky, and R. M. Booze, 2000. Prenatal cocaine exposure

neurodevelopmental aspects of clinical disorders

impairs selective attention: Evidence from serial reversal and extradimensional shift tasks. Behav. Neurosci. 114:725–738. Garg, U. C., H. Turndorf, and M. Bansinath, 1993. Effect of cocaine on macromolecular syntheses and cell proliferation in cultured glial cells. Neuroscience 57:467–472. Gaultney, J. F., J. L. Gingras, and M. Martin, 2005. Prenatal cocaine exposure and infants’ preference for novelty and distractibility. J. Genet. Psychol. 166(4):385–406. Gendle, M. H., T. L. White, M. Strawderman, C. F. Mactutus, R. M. Booze, D. A. Levitsky, et al., 2004. Enduring effects of prenatal cocaine exposure on selective attention and reactivity to errors: Evidence from an animal model. Behav. Neurosci. 118(2):290–297. Genkova-Papazova, M., B. P. Petkova, M. Lazarova-Bakarova, A. Boyanova, and D. Staneva-Stoytcheva, 1997. Effects of flunarizine and nitrendipine on electroconvulsive shock- and clonidine-induced amnesia. Pharmacol. Biochem. Behav. 56:583– 587. Gil, M., X. Zhen, and E. Friedman, 2003. Prenatal cocaine exposure alters glycogen synthase kinase-3beta (GSK3beta) pathway in select rabbit brain areas. Neurosci. Lett. 349:143– 146. Gilde, A. B., 1994. Prenatal cocaine exposure affects motor activity following acute quinpirole injection. Dissertation Abstracts International. B. The Sciences and Engineering [0419-4217], 55 iss:4-B, 1702. Gingras, J. L., J. Feibel, B. Dalley, and A. Muelenaer, 1995. Maternal polydrug use including cocaine and postnatal infant sleep architecture: Preliminary observations and implications for respiratory control and behavior. Early Hum. Dev. 43:197– 204. Glatt, S. J., G. H. Trksak, O. S. Cohen, B. P. Simeone, and D. Jackson, 2004. Prenatal cocaine exposure decreases nigrostriatal dopamine release in vitro: Effects of age and sex. Synapse 53:74–89. Goeders, N., V. Bell, A. Guidroz, and M. McNulty, 1990. Dopaminergic involvement in the cocaine-induced up-regulation of benzodiazepine receptors in the rat caudate nucleus. Brain Res. 515(1–2):1–8. Goodwin, G. A., T. Bliven, C. Kuhn, R. Francis, and L. P. Spear, 1997. Immediate early gene expression to examine neuronal activity following acute and chronic stressors in rat pups: Examination of the neurophysiological alterations underlying behavioral consequences of prenatal cocaine exposure. Physiol. Behav. 61:895–902. Goold, R. G., and P. R. Gordon-Weeks, 2001. Microtubuleassociated protein 1B phosphorylation by glycogen synthase kinase 3{beta} is induced during PC12 cell differentiation. J. Cell Sci. 114(23):4273–4284. Gressens, P., F. Gofflot, G. Van Maele-Fabry, J. P. Misson, J. F. Gadisseux, P. Evard, et al., 1992. Early neurogenesis and teratogenesis in whole mouse embryo cultures: Histochemical, immunocytological, and ultrastructural study of the premigratory neuronal-glial units in mouse embryo influenced by cocaine and retinoic acid. J. Neuropathol. Exp. Neurol. 51:206– 219. Gressens, P., B. E. Kosofsky, and P. Evrard, 1992. Cocaineinduced disturbances of corticogenesis in the developing murine brain. Neurosci. Lett. 140(1):113–116. Griffith, D. R., S. D. Azuma, and I. J. Chasnoff, 1994. Threeyear outcome of children exposed prenatally to drugs. Special section: Cocaine babies. J. Am. Acad. Child Adolesc. Psychiatry 33(1):20–27.

Haasen, C., and M. Krausz, 2001. Myths versus evidence with respect to cocaine and crack: Learning from the US experience. Eur. Addict. Res. 7(4):159–161. Hansen, R. L., J. M. Struthers, and S. M. Gospe, Jr., 1993. Visual evoked potentials and visual processing in stimulant drugexposed infants. Dev. Med. Child Neurol. 35(9):798–805. He, N., J. Bai, M. Champoux, S. J. Suomi, and M. S. Lidow, 2004. Neurobehavioral deficits in neonatal rhesus monkeys exposed to cocaine in utero. Neurotoxicol. Teratol. 26:13–26. He, N., and M. S. Lidow, 2004. Cerebral cortical abnormalities seen in a nonhuman primate model of prenatal cocaine exposure are not related to vasoconstriction. Neurotoxicology 25:419– 432. He, N., Z.-M. Song, and M. S. Lidow, 1999. Cocaine induces cell death within the primate fetal cerebral wall. Neuropathol. Appl. Neurobiol. 25:504–512. Heffelfinger, A., S. Craft, and J. Shyken, 1997. Visual attention in children with prenatal cocaine exposure. J. Int. Neuropsychol. Soc. 1997:237–245. Heffelfinger, A. K., S. Craft, D. A. White, and J. Shyken, 2002. Visual attention in preschool children prenatally exposed to cocaine: Implications for behavioral regulation. J. Int. Neuropsychol. Soc. 8:12–21. Held, J., M. Riggs, and C. Dorman, 1999. The effect of prenatal cocaine exposure on neurobehavioral outcome: A meta-analysis. Neurotoxicol. Teratol. 21(6):619–625. Hellendall, R. P., U. Schambra, J. Liu, G. R. Breese, D. E. Millhorn, and J. M. Lauder, 1992. Detection of serotonin receptor transcripts in the developing nervous system. J. Chem. Neuroanat. 5:299–310. Henderson, M., and R. Fuller, 1992. Comparative brain levels of phenylpropanolamine and amphetamine in rats. Res. Commun. Chem. Pathol. Pharmacol. 75(2):247–250. Henderson, M. G., M. M. McConnaughey, and B. A. McMillen, 1991. Long-term consequences of prenatal exposure to cocaine or related drugs: Effects on rat brain monoaminergic receptors. Brain Res. Bull. 26(6):941–945. Henderson, M. G., and B. A. McMillen, 1990. Effects of prenatal exposure to cocaine or related drugs on rat development and neurological indices. Brain Res. Bull. 24:207–212. Hendrickson, A., J. Van Brederode, K. Mulligan, and M. Celio, 1991. Development of the calcium-binding protein parvalbumin and calbindin in monkey striate cortex. J. Comp. Neurol. 307(4):626–646. Heyser, C., W. Chen, J. Miller, N. Spear, and L. Spear, 1990. Prenatal cocaine exposure induces deficits in Pavlovian conditioning and sensory preconditioning among infant rat pups. Behav. Neurosci. 104:955–963. Heyser, C. J., L. Rajachandran, N. E. Spear, and L. P. Spear, 1994. Responsiveness to cocaine challenge in adult rats following prenatal exposure to cocaine. Psychopharmacology (Berl.) 116(1):45– 55. Heyser, C., N. Spear, and L. Spear, 1992. The effects of prenatal exposure to cocaine on conditional discrimination learning in adult rats. Behav. Neurosci. 106:841–849. Higgins, S. T., D. D. Delaney, A. J. Budney, W. K. Bickel, J. R. Hughes, F. E. Foerg, et al., 1991. A behavioral approach to achieving initial cocaine abstinence. Am. J. Psychiatry 148(9): 1218–1224. Hirabayashi, Y., Y. Itoh, H. Tabata, K. Nakajima, T. Akiyama, N. Masuyama, et al., 2004. The Wnt/{beta}-catenin pathway directs neuronal differentiation of cortical neural precursor cells. Development 131(12):2791–2801.

langlois and mayes: prenatal cocaine exposure

669

Hiscock, J. J., L. Mackenzie, A. Medvedev, and J. O. Willoughby, 2001. Kainic acid and seizure-induced fos in subtypes of cerebrocortical neurons. J. Neurosci. Res. 66(6):1094– 1100. Hooks, M. S., J. L. Juncos, J. B. Justice, S. M. Meiergerd, S. L. Povlock, J. O. Schenk, et al., 1994. Individual locomotor response to novelty predicts selective alterations in D1 and D2 receptors and mRNA. J. Neurosci. 14:6144–6152. Howard, S. G., R. Fisher, and C. F. Landry, 1997. Alterations in the spontaneous release of dopamine and the density of the DA D2 receptor mRNA after chronic postnatal exposure to cocaine. Brain Res. Bull. 43:101–106. Hughes, H. E., G. F. Pringle, L. A. Scribani, D. L. DowEdwards, et al., 1991. Cocaine treatment in neonatal rats affects the adult behavioral response to amphetamine. Neurotoxicol. Teratol. 13(3):335–339. Hurt, H., E. Malmud, L. Betancourt, L. E. Braitman, N. L. Brodsky, and J. Giannetta, 1997. Children with in utero cocaine exposure do not differ from control subjects on intelligence testing. Arch. Pediatr. Adolesc. Med. 151:1237– 1241. Hutchings, D. E., T. A. Fico, and D. L. Dow-Edwards, 1989. Prenatal cocaine: Maternal toxicity, fetal effects and locomotor activity in rat offspring. Neurotoxicol. Teratol. 11(1):65–69. Jacobson, J. L., S. W. Jacobson, and R. J. Sokol, 1994. Effects of prenatal exposure to alcohol, smoking, and illicit drugs on postpartum somatic growth. Alcohol Clin. Exp. Res. 18(2):317– 323. Jacobson, S. W., J. T. Bihun, and L. M. Chiodo, 1999. Effects of prenatal alcohol and cocaine exposure on infant cortisol levels. Dev. Psychopathol. 11:195–208. Jacobson, S. W., J. L. Jacobson, R. J. Sokol, S. S. Martier, and L. M. Chiodo, 1996. New evidence for neurobehavioral effects of in utero cocaine exposure. J. Pediatr. 129:581–590. Johns, J. M., D. A. Lubin, J. A. Lieberman, and J. M. Lauder, 2002. Developmental effects of prenatal cocaine exposure on 5-HT 1A receptors in male and female rat offspring. Dev. Neurosci. 24(6):522–530. Johns, J. M., M. J. Means, D. R. Anderson, L. W. Means, and B. A. McMillen, 1992. Prenatal exposure to cocaine. II. Effects on open-field activity and cognitive behavior in Sprague-Dawley rats. Neurotoxicol. Teratol. 14(5):343–349. Johnson, J. M., A. Seikel, C. L. Madison, S. M. Foose, and K. D. Rinard, 1997. Standardized test performance of children with a history of prenatal exposure to multiple drugs/cocaine. J. Commun. Disord. 30:45–73. Jones, L. B., G. D. Stanwood, B. S. Reinoso, R. A. Washington, H.-Y. Wang, E. Friedman, et al., 2000. In utero cocaine-induced dysfunction of dopamine D1 receptor signaling and abnormal differentiation of cerebral cortical neurons. J. Neurosci. 20(12):4606–4614. Jones, N. A., T. Field, M. Davalos, and S. Hart, 2004. Greater right frontal EEG asymmetry and nonemphatic behavior are observed in children prenatally exposed to cocaine. Int. J. Neurosci. 114:459–480. Kalivas, P. W., and P. Duffy, 1997. Dopamine regulation of extracellular glutamate in the nucleus accumbens. Brain Res. 761(1):173–177. Kaplan, D., 1995. Statistical power in structure equation modeling. In R. H. Hoyle, ed., Structure Equation Modeling, Concepts, Issues, and Applications, 100–117. Newbury Park, CA: Sage Publications. Karler, R., L. D. Calder, L. H. Thai, and J. B. Bedingfield, 1995. The dopaminergic, glutamatergic, GABAergic bases for

670

the action of amphetamine and cocaine. Brain Res. 671(1):100– 104. Karmel, B. Z., J. M. Gardner, and R. L. Freedland, 1996. Arousal modulated attention at four months as a function of intrauterine cocaine exposure and central nervous system injury. J. Pediatr. Psychol. 21:821–832. Karmel, B. Z., J. M. Gardner, and C. L. Magnano, 1991. Neurofunctional consequences of in utero cocaine exposure. NIDA Res. Monogr. 105:535–536. Koegler, S. M., F. J. Seidler, et al., 1991. Ischemia contributes to adverse effects of cocaine on brain development: Suppression of ornithine decarboxylase activity in neonatal rat. Brain Res. Bull. 27:829–931. Koob, G. F., 1992. Drugs of abuse: Anatomy, pharmacology and function of reward pathways. Trends Pharmacol. Sci. 13(5):177– 184. Koren, G., I. Nulman, J. Rovet, R. Greenbaum, M. Loebstein, and T. Einarson, 1998. Long-term neurodevelopmental risks in children exposed in utero to cocaine. The Toronto adoption study. Ann. NY Acad. Sci. 846:306–313. Kosofsky, B. E., L. M. Genova, and S. E. Hyman, 1995a. Postnatal age defines specificity of immediate early gene induction by cocaine in developing rat brain. J. Comp. Neurol. 351: 27–40. Kosofsky, B. E., L. M. Genova, and S. E. Hyman, 1995b. Substance P phenotype defines specificity of c-fos induction by cocaine in developing rat striatum. J Comp. Neurol. 351:41–50. Kosofsky, B. E., and S. E. Hyman, 1993. The ontogeny of immediate early gene response to cocaine: A molecular analysis of the effects of cocaine on developing rat brain. NIDA Res. Monogr. 125:161–171. Kosofsky, B. E., and A. S. Wilkins, 1998. A mouse model of transplacental cocaine exposure: Clinical implications for exposed infants and children. Ann. NY Acad. Sci. 846:248–261. Lagasse, L. L., R. Seifer, and B. Lester, 1999. Interpreting research on prenatal substance exposure in the context of multiple confounding factors. Clin. Perinatol. 26(1):39–54. Lason, W., 2001. Neurochemical and pharmacological aspects of cocaine-induced seizures. Pol. J. Pharmacol. Pharm. 53(1):57–60. Lauder, J., 1988. Neurotransmitters as morphogens. Prog. Brain Res. 73:365–387. Lauder, J. M., and J. Liu, 1994. Glial heterogeneity and developing neurotransmitter systems. Perspect. Dev. Neurobiol. 2:239–250. Leroy, K., and J.-P. Brion, 1999. Developmental expression and localization of glycogen synthase kinase-3[beta] in rat brain. J. Chem. Neuroanat. 16(4):279–293. Leslie, C. A., M. W. Robertson, A. B. Jung, J. Liebermann, and J. P. Bennett, 1994. Effects of prenatal cocaine exposure upon postnatal development of neostriatal dopaminergic function. Synapse 17:210–215. Lester, B. M., K. Freier, and L. LaGasse, 1995. Prenatal cocaine exposure and child outcome: What do we really know? In M. Lewis and M. Bendersky, eds., Mothers, Babies, and Cocaine: The Role of Toxins in Development, 19–40. Hillsdale, NJ: Lawrence Erlbaum. Lester, B. M., L. L. LaGasse, and R. Seifer, 1998. Cocaine exposure and children: The meaning of subtle effects. Science 282:633–634. Levitt, P., J. A. Harvey, E. Friedman, K. Simansky, and E. H. Murphy, 1997. New evidence for neurotransmitter influences on brain development. Trends Neurosci. 20:269–274. Lewis, B. A., L. T. Singer, E. J. Short, S. Minnesa, R. Arendt, P. Weishampel, et al., 2004. Four-year language outcomes

neurodevelopmental aspects of clinical disorders

of children exposed to cocaine in utero. Neurotoxicol. Teratol. 26:617–627. Lidow, M., 1995. Prenatal cocaine exposure adversely affects development of the primate cerebral cortex. Synapse 21:332– 341. Lidow, M. S., 2003. Consequences of prenatal cocaine exposure in nonhuman primates. Dev. Brain Res. 147(1–2):23–36. Lidow, M. S., D. Bozian, and Z.-M. Song, 2001. Cocaine affects cerebral neocortical cytoarchitecture in primates only if administered during neocortical neuronogenesis. Dev. Brain Res. 128:45–52. Lidow, M., and P. Rakic, 1994. Unique deployment of adrenergic receptor subtypes during development of the primate cerebral cortex including transient subplate and proliferative zones. J. Neuoscience. 14:4064–4078. Lidow, M. S., and P. Rakic, 1995. Neurotransmitter receptors in the proliferative zones of the developing primate occipital lobe. J. Comp. Neurol. 360:393–402. Lidow, M. S., and Z.-M. Song, 2001a. Effect of cocaine on cell proliferation in the cerebral wall of monkey fetuses. Cerebral Cortex 11:545–551. Lidow, M. S., and Z.-M. Song, 2001b. Primates exposed to cocaine in utero display reduced density and number of cerebral cortical neurons. J. Comp. Neurol. 435:263–275. Ligon, K. L., Y. Echelard, S. Assimacopoulos, P. S. Danielian, S. Kaing, E. A. Grove, et al., 2003. Loss of Emx2 function leads to ectopic expression of Wnt1 in the developing telencephalon and cortical dysplasia. Development 130(10): 2275–2287. Lindefors, N., 1993. Dopaminergic regulation of glutamic acid decarboxylase mRNA expression and GABA release in the striatum: A review. Prog. Neuropsychopharmacol. Biol. Psychiatry 17(6): 887–903. Lipton, J. W., T. Q. Vu, Z. Ling, S. Gyawali, J. R. Mayer, and P. M. Carvey, 2002. Prenatal cocaine exposure induces an attenuation of uterine blood flow in the rat. Neurotoxicol. Teratol. 24:143–148. Lloyd, S. A., B. Wensley, C. J. Faherty, and R. J. Smeyne, 2003. Regional differences in cortical dendrite morphology following in utero exposure to cocaine. Dev. Brain Res. 147:59– 66. Lutiger, B., K. Graham, T. R. Einarson, and G. Koren, 1991. Relationship between gestational cocaine use and pregnancy outcome: A meta-analysis. Teratology 44(4):405–414. Machon, O., C. J. van den Bout, M. Backman, R. Kemler, and S. Krauss, 2003. Role of [beta]-catenin in the developing cortical and hippocampal neuroepithelium. Neuroscience 122(1):129– 143. Madison, C. L., J. M. Johnson, A. Seikel, M. Arnold, and L. Schultheis, 1998. Comparative study of the phonology of preschool children prenatally exposed to cocaine and multiple drugs and nonexposed children. J. Commun. Disord. 31:231– 244. Malakoff, M. E., L. C. Mayes, and R. S. Schottenfeld, 1994. Language abilities of preschool-age children living with cocaineusing mothers. Am. J. Addict. 3(4):346–354. Malakoff, M. E., L. C. Mayes, R. Schottenfeld, and S. Howell, 1999. Language production in 24-month-old inner-city children of cocaine-and-other-drug-using mothers. J. Appl. Dev. Psychol. 20:159–180. Mangano, C., J. Gardner, and B. Karmel, 1992. Differences in salivary cortisol levels in cocaine-exposed and noncocaineexposed NICU infants. Dev. Psychobiol. 25:93–103.

Martinez, M., A. Calvo-Torrent, and J. Herbert, 2002. Mapping brain response to social stress in rodents with c-fos expression: A review. Stress 5(1):3–13. Matsumoto, R. R., D. L. Gilmore, B. Pouw, W. D. Bowen, W. Williams, A. Kausar, et al., 2004. Novel analogs of the sigma receptor ligand BD1008 attenuate cocaine-induced toxicity in mice. Eur. J. Pharmacol. 492(1):21–26. Matsumoto, R., K. Hewett, B. Pouw, W. Bowen, S. Husbands, J. Cao, et al., 2001. Rimcazole analogs attenuate the convulsive effects of cocaine: Correlation with binding to sigma receptors rather than dopamine transporters. Neuropharmacology 41(7):878–886. Maurice, T., and P. Romieu, 2004. Involvement of the sigma1 receptor in the appetitive effects of cocaine. Pharmacopsychiatry 37:198–207. Mayes, L. C., 1999. Developing brain and in-utero cocaine exposure: Effects on neural ontogeny. Dev. Psychopathol. 11:685–714. Mayes, L. C., and M. H. Bornstein, 1995. Developmental Dilemmas for Cocaine-Abusing Parents and Their Children. Hillsdale, NJ: Erlbaum. Mayes, L. C., M. H. Bornstein, K. Chawarska, and R. H. Granger, 1995. Information processing and developmental assessments in 3-month-old infants exposed prenatally to cocaine. Pediatrics 95(4):539–545. Mayes, L. C., M. H. Bornstein, K. Chawarska, O. M. Haynes, and R. H. Granger, 1996. Impaired regulation of arousal in three-month-old infants exposed prenatally to cocaine and other drugs. Dev. Psychopathol. 8:29–42. Mayes, L. C., D. Cicchetti, S. Acharyya, and H. Zhang, 2003. Developmental trajectories of cocaine-and-other-drugexposed and non-cocaine-exposed children. J. Dev. Behav. Pediatr. 24(5):323–335. Mayes, L. C., and T. Fahy, 2001. Prenatal drug exposure and cognitive development. In R. J. Sternberg and E. L. Grigorenko, eds., Environmental Effects on Cognitive Abilities, 189–220. Mahwah, NJ: Lawrence Erlbaum. Mayes, L. C., and S. Fisher, 1998. A computerized attention task for preschool and school aged children. Programmed in Visual Basic. Copyright 1998. Mayes, L. C., C. Grillon, R. Granger, and R. Schottenfeld, 1998. Regulation of arousal and attention in preschool children exposed to cocaine prenatally. Ann. NY Acad. Sci. 846:126–143. Mayes, L. C., and M. Pajulo, 2006. Neurodevelopmental sequelae of prenatal cocaine exposure. In D. C., Bellinger, ed., Human Developmental Neurotoxicology, 192–229. New York: Taylor and Francis. Mayes, L. C., and A. Ward, 2003. Principles of neurobehavioral teratology. In D. Cicchetti and E. Walker, eds., Neurodevelopmental Mechanisms in Psychopathology, 3–33. New York: Cambridge University Press. Mazer, C., J. Muneyyirci, K. Taheny, N. Raio, A. Borella, and P. M. Whitaker-Azmitia, 1997. Serotonin depletion during synaptogenesis leads to decreased synaptic density and learning deficits in the adult rat: A model of neurodevelopmental disorders with cognitive deficits. Brain Res. 760:68–74. McCracken, K., W. Bowen, and R. Matsumoto, 1999. Novel sigma receptor ligands attenuate the locomotor stimulatory effects of cocaine. Eur. J. Pharmacol. 365(1):35–38. McMillen, B., J. Johns, E. Bass, and L. Means, 1991. Learning and behavior of rats exposed to cocaine throughout gestation. Teratology 43:495. Meaney, M. J., D. H. Aitken, V. Viau, S. Sharma, and S. Sarrieau, 1989. Neonatal handling alters adrenocortical

langlois and mayes: prenatal cocaine exposure

671

negative feedback sensitivity and hippocampal type II glucocorticoid receptor binding in the rat. Neuroendocrinology 50:597–604. Megason, S. G., and A. P. McMahon, 2002. A mitogen gradient of dorsal midline Wnts organizes growth in the CNS. Development 129(9):2087–2098. Meier, C. A., D. Fabbro, I. Meyhack, B. Hemmings, U. Olbrecht, A. Jakob, et al., 1991. Effect of hypothyroidism and thyroid hormone replacement on the level of protein kinase C and protein kinase A in rat liver. FEBS Lett. 282(2):397–400. Meijer, O. C., A. Williamson, M. F. Dallman, and D. Pearce, 2000. Transcriptional repression of the 5-HT1A receptor promoter by corticosterone via mineralocorticoid receptors depends on the cellular context. J. Neuroendocrinol. 12(3):245–254. Metosky, P., and J. Vondra, 1995. Prenatal drug exposure and play and coping in toddlers: A comparative study. Infant Behav. Dev. 18:15–25. Micheva, K. D., and C. Beaulieu, 1997. Development and plasticity of the inhibitory neocortical circuitry with an emphasis on the rodent barrel field cortex: A review. Can. J. Physiol. Pharmacol. 75(5):470–478. Minabe, Y., C. R. Ashby, Jr., C. Heyser, L. P. Spear, and R. Y. Wang, 1992. The effects of prenatal cocaine exposure on spontaneously active midbrain dopamine neurons in adult male offspring: An electrophysiological study. Brain Res. 586(1):152– 156. Mirochnick, M., J. Meyer, J. Cole, T. Herren, and B. Zuckerman, 1991. Circulating catecholamine concentrations in cocaine-exposed neonates: A pilot study. Pediatrics 88(3):481– 485. Mitchell, E. S., and A. Snyder-Keller, 2003a. Blockade of D1 dopaminergic transmission alleviates c-fos induction and cleaved caspase-3 expression in the brains of rat pups exposed to prenatal cocaine or perinatal asphyxia. Exp. Neurol. 182:64–74. Mitchell, E. S., and A. Snyder-Keller, 2003b. c-fos and cleaved caspase-3 expression after perinatal exposure to ethanol, cocaine, or the combination of both drugs. Dev. Brain Res. 147:107– 117. Molina, V. A., J. M. Wagner, and L. P. Spear, 1994. The behavioral response to stress is altered in adult rats exposed prenatally to cocaine. Physiol. Behav. 55(5):941–945. Moody, C. A., N. A. Frambes, and L. P. Spear, 1992. Psychopharmacological responsiveness to the dopamine agonist quinpirole in normal weanlings and in weanling offspring exposed gestationally to cocaine. Psychopharmacology (Berl.) 109:256–262. Moratella, R., M. Xu, S. Tonegawa, and A. M. Graybiel, 1996. Cellular responses to psychomotor stimulant and neuroleptic drugs are abnormal in mice lacking the D1 receptor. Proc. Natl. Acad. Sci. USA 93:14928–14933. Morgan, D., K. Grant, H. Gage, R. Mach, J. Kaplan, O. Prileau, et al., 2002. Social dominance in monkeys: Dopamine D2 receptors and cocaine self-administration. Nature Neurosci. 5:169–174. Morrow, B. A., J. D. Elsworth, and R. H. Roth, 2003. Axo-axonic structures in the medial prefrontal cortex of the rat: Reduction by prenatal exposure to cocaine. J. Neurosci. 23(12): 5227–5234. Morrow, B. A., J. D. Elsworth, and R. H. Roth, 2005. Prenatal exposure to cocaine selectively disrupts the development of parvalbumin containing local circuit neurons in the medial prefrontal cortex of the rat. Synapse 56(1):1–11. Morrow, B., T. George, and R. Roth, 2004. Noradrenergic alpha-2 agonists have anxiolytic-like actions on stress-related

672

behavior and mesoprefrontal dopamine biochemistry. Brain Res. 1027:173–178. Muller, C. M., A. C. Akhavan, and M. Bette, 1993. Possible role of S-100 in glia-neuronal signalling involved in activity dependent plasticity in the developing mammalian cortex. J. Chem. Neuroanat. 6:215–220. Murphy, D. D., and M. Segal, 1997. Morphological plasticity of dendritic spines in central neurons is mediated by activation of cAMP response element binding protein. Proc. Natl. Acad. Sci. USA 94(4):1482–1487. Nakagawa, S., J.-E. Kim, R. Lee, J. E. Malberg, J. Chen, C. Steffen, et al., 2002. Regulation of neurogenesis in adult mouse hippocampus by cAMP and the cAMP response elementbinding protein. J. Neurosci. 22(9):3673–3682. Nassogne, M. C., P. Evrard, et al., 1995. Selective neuronal toxicity of cocaine in embryonic mouse brain cocultures. Proc. Natl. Acad. Sci. USA 92:11029–11033. Nassogne, M. C., P. Evrard, and P. Courtoy, 1998. Selective direct toxicity of cocaine on fetal mouse neurons. Ann. NY Acad. Sci. 846:51–68. Nassogne, M. C., J. Louahed, P. Evrard, and P. J. Courtoy, 1997. Cocaine induces apoptosis in cortical neurons of fetal mice. J. Neurochem. 68:2442–2450. National Institute on Drug Abuse, 1996. National Pregnancy and Health Survey—Drug Use Among Women Delivering Live Births, 1992. Rockville, MD: U.S. Department of Health and Human Services. Needlman, R., B. Zuckerman, G. M. Anderson, M. Mirochnick, and D. J. Cohen, 1993. Cerebrospinal fluid monoamine precursors and metabolites in human neonates following in utero cocaine exposure: A preliminary study. Pediatrics 92(1):55–60. Nestler, E. J., 1994. Molecular neurobiology of drug addiction. Neuroscientist 11:77–87. Nestler, E. J., C. M. Bergson, X. Gultart, and B. T. Hope, 1993. Regulation of neural gene expression in opiate and cocaine addiction. NIDA Res. Monogr. 125:92–116. Nestler, E. J., X. M. Gao, and C. A. Tamminga, 1997. Molecular biology. V. Immediate early genes. Am. J. Psychiatry 15:312. Neuspiel, D. R., 1995. The problem of confounding in research on prenatal cocaine effects on behavior and development. In M. Lewis and M. Bendersky, eds., Mothers, Babies, and Cocaine: The Role of Toxins in Development, 95–110. Hillsdale, NJ: Lawrence Erlbaum. Nitsch, C., I. P. Maly, D. Mori, and A. L. Scotti, 1994. Evidence for the colocalization of parvalbumin and glutamate, but not GABA, in the perforant path of the gerbil hippocampal formation: A combined immunocytochemical and microquantitative analysis. J. Neurochem. 62(4):1276–1284. Nitsch, C., A. L. Scotti, and F. M. Nitsch, 1995. Distribution of parvalbumin-containing interneurons in the hippocampus of the gerbil—a qualitative and quantitative statistical analysis. J. Chem. Neuroanat. 9(2):135–147. Nitsch, C., A. Scotti, A. Sommacal, and G. Kalt, 1989. GABAergic hippocampal neurons resistant to ischemia-induced neuronal death contain the Ca2+-binding protein parvalbumin. Neurosci. Lett. 105(3):263–268. Novikova, S. I., F. He, J. Bai, and M. S. Lidow, 2005. Neuropathology of the cerebral cortex observed in a range of animal models of prenatal cocaine exposure may reflect alterations in genes involved in the Wnt and cadherin systems. Synapse 56:105–116.

neurodevelopmental aspects of clinical disorders

Nulman, I., J. Rovet, D. Altmann, C. Bradley, T. Einarson, and G. Koren, 1994. Neurodevelopment of adopted children exposed in utero to cocaine [see comments]. Can. Med. Assoc. J. 151(11):1591–1597. Olson, H. C., and S. K. Toth, 1999. Samples in research on prenatal cocaine exposure: Vexing problems and practical solutions. J. Drug Issues 29:237–253. Parnavelas, J. G., 1992. Development of GABA-containing neurons in the visual cortex. Prog. Brain Res. 90:523–537. Pennington, N. J., J. S. Kelly, and A. P. Fox, 1991. A study of the mechanism of Ca2+ current inhibition produced by serotonin in rat dorsal raphe neurons. J. Neurosci. 11:3594–3609. Peris, J., S. J. Boyson, W. A. Cass, P. Curella, L. P. Dwoskin, G. Larson, et al., 1990. Persistence of neurochemical changes in dopamine systems after repeated cocaine administration. J. Pharmacol. Exp. Ther. 253(1):38–44. Peris, J., H. M. Coleman, and W. J. Millard, 1992. Cocaine inutero enhances the behavioral response to cocaine in adult rats. Pharmacol. Biochem. Behav. 42:509–515. Phelps, L., and J. W. Cottone, 1999. Long-term developmental outcome of prenatal cocaine exposure. J. Psychoeduc. Assess. 17: 343–353. Phillips, P. E., J. M. Johns, D. A. Lubin, E. A. Budygin, R. R. Gainetdinov, J. A. Lieberman, et al., 2003. Presynaptic dopaminergic function is largely unaltered in mesolimbic and mesostriatal terminals of adult rats that were prenatally exposed to cocaine. Brain Res. 961:63–72. Pierce, R. C., and P. W. Kalivas, 1997. A circuitry model of the expression of behavioral sensitization to amphetamine-like psychostimulants. Brain Res. Rev. 25(2):192–216. Plashne, M., and A. F. Gann, 1990. Activators and targets. Nature 346:329–331. Rajkowsi, J., P. Kubiak, S. Ivanova, and G. Aston-Jones, 1998. State-related activity and reactivity of the locus coeruleus neurons in behaving monkeys. Adv. Pharmacol. (NY) 42:740–744. Regaldo, M., V. Schechtman, A. Del Angel, and X. Bean, 1995. Sleep disorganization in cocaine-exposed neonates. Infant Behav. Dev. 18:319–327. Regaldo, M. G., V. L. Schechtman, A. P. Del Angel, and X. D. Bean, 1996. Cardiac and respiratory patterns during sleep in cocaine-exposed neonates. Early Hum. Dev. 44:187–200. Reinoso, B. S., A. S. Undie, and P. Levitt, 1996. Dopamine receptors mediate differential morphological effects on cerebral cortical neurons in vitro. J. Neurosci. Res. 43:439–453. Ren, J.-Q., C. J. Malanga, E. Tabit, and B. E. Kosofsky, 2004. Neuropathological consequences of prenatal cocaine exposure in the mouse. Int. J. Dev. Neurosci. 22:309–320. Retaux, S., M. J. Besson, and J. Penit-Soria, 1991. Synergism between D1 and D2 dopamine receptors in the inhibition of the evoked release of [3H]GABA in the rat prefrontal cortex. Neuroscience 43(2–3):323–329. Richardson, G., 1998. Prenatal cocaine exposure: A longitudinal study of development. Ann. NY Acad. Sci. 846:144–152. Richardson, G., M. Conroy, and N. Day, 1996. Prenatal cocaine exposure: Effects on the development of school-age children. Neurotoxicol. Teratol. 18:627–634. Riley, E., and J. Foss, 1991a. The acquisition of passive avoidance, active avoidance, and spatial navigation tasks by animals prenatally exposed to cocaine. Neurotoxicol. Teratol. 13:559–564. Riley, E. P., and J. A. Foss, 1991b. Exploratory behavior and locomotor activity: A failure to find effects in animals prenatally exposed to cocaine. Neurotoxicol. Teratol. 13(5):553–558.

Ritz, M. C., and F. R. George, 1997. Cocaine-induced convulsions: Pharmacological antagonism at serotonergic, muscarinic and sigma receptors. Psychopharmacology 129(4):299–310. Robertson, G. S., S. R. Vincent, and H. C. Fibiger, 1992. D1 and D2 dopamine receptors differentially regulate c-fos expression in striatonigral and striatopallidal neurons. Neuroscience 49:285–296. Romano, A. G., and J. A. Harvey, 1998. Prenatal cocaine exposure: Long term deficits in learning and motor performance. Ann. NY Acad. Sci. 846:89–108. Romieu, P., R. Martin-Fardon, and T. Maurice, 2000. Involvement of the sigma1 receptor in the cocaine-induced conditioned place preference. NeuroReport 11:2885–2888. Romieu, P., J. Meunier, D. Garcia, N. Zozime, R. Martin-Fardon, W. Bowen, et al., 2004. The sigma1 receptor activation is a key step for the reactivation of cocaine conditioned place preference by drug priming. Psychopharmacology 175:154–162. Ronnekleiv, O. K., Y. Fang, W. S. Choi, and L. Chai, 1998. Changes in the midbrain-rostral forebrain dopamine circuitry in the cocaine-exposed primate fetal brain. Ann. NY Acad. Sci. 846:165–181. Ronnekleiv, O. K., and B. R. Naylor, 1995. Chronic cocaine exposure in fetal rhesus monkey: Consequences for early development of dopamine neurons. J. Neurosci. 15:7330– 7343. Rose-Jacobs, R., H. Cabral, M. A. Posner, J. Epstein, and D. A. Frank, 2002. Do “we just know”? Masked assessors’ ability to accurately identify children with prenatal cocaine exposure. J. Dev. Behav. Pediatr. 23(5):340–346. Salvatore, M. F., O. Hudspeth, L. E. Arnold, P. E. Wilson, J. A. Stanford, C. F. MacTutus, et al., 2004. Prenatal cocaine exposure alters potassium-evoked dopamine release dynamics in rat striatum. Neuroscience 123:481–490. Santiago, M., A. Machado, and J. Cano, 1993. Regulation of the prefrontal cortical dopamine release by GABAA and GABAB receptor agonists and antagonists. Brain Res. 630(1–2): 28–31. Santolaria-Fernandez, F. J., J. L. Gomez-Sirvent, C. E. Gonzalez-Reimers, J. N. Batista-Lopez, J. A. JorgeHernandez, F. Rodriguez-Moreno, et al., 1995. Nutritional assessment of drug addicts. Drug Alcohol Depend. 38(1):11–18. Scalzo, F. M., S. F. Ali, N. A. Frambes, and L. P. Spear, 1990. Weanling rats exposed prenatally to cocaine exhibit an increase in striatal D2 dopamine binding associated with an increase in ligand affinity. Pharmacol. Biochem. Behav. 37(2):371–373. Scher, M. S., G. A. Richardson, and N. L. Day, 2000. Effects of prenatal cocaine/crack and other drug exposure on electroencephalographic sleep studies at birth and one year. Pediatrics 105:39–48. Schiller, C., and P. Allen, 2005. Follow-up of infants prenatally exposed to cocaine. Pediatr. Nurs. 31(5):427–436. Schroder, M., P. Snyder, I. Sielski, and L. Mayes, 2004. Impaired performance of children exposed in utero to cocaine on a novel test of visospatial working memory. Brain Cogn. 55:409–412. Seamans, J. K., S. B. Floresco, and A. G. Phillips, 1998. D1 receptor modulation of hippocampal-prefrontal cortical circuits integrating spatial memory with executive functions in the rat. J. Neurosci. 18:1613–1621. Seidler, F. J., and T. A. Slotkin, 1992. Fetal cocaine exposure causes persistent noradrenergic hyperactivity in rat brain regions: Effects on neurotransmitter turnover and receptors. J. Pharmacol. Exp. Ther. 263(2):413–421.

langlois and mayes: prenatal cocaine exposure

673

Seidler, F. J., S. W. Temple, E. C. McCook, and T. A. Slotkin, 1995. Cocaine inhibits central noradrenergic and dopaminergic activity during the critical developmental period in which catecholamines influence cell development. Brain Res. Dev. Brain Res. 85(1):48–53. Shearman, L. P., and L. M. Collins, 1996. Characterization and localization of 125-I-RTI-55-labeled cocaine binding sites in fetal and adult rat brain. J. Pharmacol. Exp. Ther. 277:1770– 1783. Shemer, A. V., E. C. Azmitia, and P. M. Whitaker-Azmitia, 1991. Dose-related effects of prenatal 5-methoxytryptamine (5-MT) on development of serotonin terminal density and behavior. Brain Res. 59:59–63. Sheng, M., and M. E. Greenberg, 1990. The regulation and function of c-fos and other immediate early genes in the nervous system. Neuron 4:477–485. Shi, X., N. Goeders, B. Irby, and D. Dow-Edwards, 1998. Gender-specific effects of inhibition of the serotonin or dopamine transporters on 8-OH-DPAT binding in developing rat brain: Abstract. Soc. Neurosci. Abstr. 538.533. Shumsky, J. S., Y. Wu, E. H. Murphy, J. Nissanov, A. O’Brien-Jenkins, and D. R. Grayson, 2002. Differential effects of prenatal cocaine exposure on selected subunit mRNAs of the GABAA receptor in rabbit anterior cingulate cortex. J. Chem. Neuroanat. 24(4):243–255. Silvers, J. M., D. R. Wallace, S. B. Harrod, C. F. Mactutus, and R. M. Booze, 2006. Prenatal cocaine alters dopamine and sigma receptor binding in nucleus accumbens and striatum in dams and adolescent offspring. Neurotoxicol. Teratol. 28:173–180. Singer, L., 1999. Advances and redirections in understanding effects of fetal drug exposure. J. Drug Issues 29:253–262. Singer, L. T., R. Arendt, J. Fagan, S. Minnes, A. Salvator, T. Bolek, et al., 1999. Neonatal visual information processing in cocaine-exposed and nonexposed infants. Infant Behav. Dev. 22:1–15. Singer, L., R. Arendt, S. Minnes, K. Farkas, and A. Salvator, 2000. Neurobehavioral outcomes of cocaine-exposed infants. Neurotoxicol. Teratol. 22:653–666. Singer, L. T., R. Arendt, K. Farkas, S. Minnes, J. Huang, and T. Yamashita, 1997. Relationship of prenatal cocaine exposure and maternal postpartum psychological distress to child developmental outcome. Dev. Psychopathol. 9(3):473– 489. Singer, L. T., S. Minnes, E. Short, R. Arendt, K. Farkas, B. Lewis, et al., 2004. Cognitive outcomes of preschool children with prenatal cocaine exposure. JAMA 291(20): 2448–2456. Smith, R. F., K. M. Mattran, M. F. Kurkjian, and S. L. Kurtz, 1989. Alterations in offspring behavior induced by chronic prenatal cocaine dosing. Neurotoxicol. Teratol. 11(1):35– 38. Snow, D. M., H. M. Carman, J. D. Smith, R. M. Booze, M. A. Welch, and C. F. Mactutus, 2004. Cocaine-induced inhibition of process outgrowth in locus coerleus neurons: Role of gestational exposure period and offspring sex. Int. J. Dev. Neurosci. 22:297–308. Snow, D. M., J. D. Smith, R. M. Booze, M. A. Welch, and C. F. Mactutus, 2001. Cocaine decreases cell survival and inhibits neurite extension of rat locus coeruleus neurons. Neurotoxicol. Teratol. 23:223–234. Snyder-Keller, A. M., and R. W. Keller, Jr., 1993. Prenatal cocaine increases striatal serotonin innervation without altering

674

the patch/matrix organization of intrinsic cell types. Dev. Brain Res. 74(2):261–267. Snyder-Keller, A. M., and R. W. Keller, Jr., 1995. Prenatal cocaine alters later sensitivity to cocaine-induced seizures. Neurosci. Lett. 191(3):149–152. Snyder-Keller, A., and R. W. Keller, 2001. Spatiotemporal analysis of fos expression associated with cocaine- and PTZinduced seizures in prenatally cocaine-treated rats. Exp. Neurol. 170(1):109–120. Soriano, E., J. A. Del Rio, I. Ferrer, C. Auladell, L. De Lecea, and S. Alcantara, 1992. Late appearance of parvalbuminimmunoreactive neurons in the rodent cerebral cortex does not follow an “inside-out” sequence. Neurosci. Lett. 142(2):147– 150. Spear, L. P., J. Campbell, K. Snyder, M. Silveri, and N. Katovic, 1998. Animal behavior models: Increased sensitivity to stressors and other environmental experiences after prenatal cocaine exposure. Ann. NY Acad. Sci. 846:76–88. Spear, L. P., C. L. Kirstein, J. Bell, V. Yoottanasumpun, R. Greenbaum, J. O’Shea, H. Hoffman, and N. E. Spear, 1989. Effects of prenatal cocaine exposure on behavior during the early postnatal period. Neurotoxicol. Teratol. 11(1):57–63. Spear, L. P., C. L. Kirstein, and N. A. Frambes, 1989. Cocaine effects on the developing central nervous system: Behavioral, psychopharmacological, and neurochemical studies. Ann. NY Acad. Sci. 562:290–307. Spear, L. P., C. L. Kirstein, N. A. Frambes, and C. A. Moody, 1989. Neurobehavioral teratogenicity of gestational cocaine exposure. NIDA Res. Monogr. 95:232–238. Stadlin, A., H. L. Choi, K. W. Tsim, and D. Tsang, 1995. Prenatal cocaine exposure revealed minimal postnatal changes in rat striatal dopamine D2 receptor sites and mRNA levels in the offspring. Mol. Neurobiol. 11:67–76. Stanwood, G. D., R. A. Washington, J. S. Shumsky, and P. Levitt, 2001. Prenatal cocaine exposure produces consistent developmental alterations in dopamine-rich regions of the cerebral cortex. Neuroscience 106(1):5–14. Steiner, H., and C. R. Gerfen, 1993. Cocaine-induced c-fos messenger RNA is inversely related to dynorphin expression in striatum. J. Neurosci. 13:5066–5081. Steiner, H., and C. R. Gerfen, 1995. Dynorphin opioid inhibition of cocaine-induced, D1 dopamine receptor-mediated immediate-early gene expression in the striatum. J. Comp. Neurol. 353:200–212. Struthers, J. M., and R. L. Hansen, 1992. Visual recognition memory in drug-exposed infants. J. Dev. Behav. Pediatr. 13(2): 108–111. Sven, S., and R. C. Marco, 1991. Ontogeny of the calcium binding protein parvalbumin in the rat nervous system. Anat. Embryol. (Berl.) V184(2):103–124. Svenningsson, P., M. Lindskog, C. Ledent, M. Parmentier, P. Greengard, B. B. Fredholm, et al., 2000. Regulation of the phosphorylation of the dopamine- and cAMP-regulated phosphoprotein of 32 kDa in vivo by dopamine D1, dopamine D2, and adenosine A2A receptors. Proc. Natl. Acad. Sci. USA 97:1856–1860. Swanson, M. W., A. P. Streissguth, P. D. Sampson, and H. Carmichael Olson, 1999. Prenatal cocaine and neuromotor outcome at four months: Effect of duration of exposure. J. Dev. Behav. Pediatr. 20(5):325–334. Thompson, B. L., P. Levitt, and G. D. Stanwood, 2005. Prenatal cocaine exposure specifically alters spontaneous alternation behavior. Behav. Brain Res. 164(1):107–116.

neurodevelopmental aspects of clinical disorders

Tilakaratne, N., G. Cai, and E. Friedman, 2001. Attenuation of cocaine-induced genomic and functional responses in prenatal cocaine-exposed rabbits. Pharmacol. Biochem. Behav. 69:225–232. Tronick, E., and M. Beeghly, 1999. Prenatal cocaine exposure, child development, and the compromising effects of cumulative risk. Clin. Perinatol. 26(1):151–171. Ujike, H., S. Kuroda, and S. Otsuki, 1996. Sigma receptor antagonists block the development of sensitization to cocaine. Eur. J. Pharmacol. 296:123–126. van Baar, A., 1990. Development of infants of drug dependent mothers. J. Child Psychol. Psychiatry 31:911–920. van Baar, A., and B. M. de Graaff, 1994. Cognitive development at preschool-age of infants of drug-dependent mothers. Dev. Med. Child Neurol. 36(12):1063–1075. Van de Kar, L. D., K. D. Richardson-Morton, and P. A. Rittenhouse, 1991. Stress: Neuroendocrine and pharmacological mechanisms. In J. G. Cantin, ed., Stress Revisited. 1. Neuroendocrinology of Stress: Methods and Achievements in Experimental Pathology, vol. 14. Basel: Karger. Van Eden, C. G., L. Mrzljak, P. Voorn, and H. B. Uylings, 1989. Prenatal development of GABA-ergic neurons in the neocortex of the rat. J. Comp. Neurol. 289(2):213–227. Verney, C., N. Zecevic, B. Nikolic, C. Alvarez, and B. Berger, 1991. Early evidence of catecholaminergic cell groups in 5 and 6 week old human embryos using tyrosine hydroxylase and dopamine-b-hydroxylase immunochemistry. Neurosci. Lett. 131: 121–124. Vogel, G., 1997. Scientists probe feelings behind decision-making [news; comment]. Science 275(5304):1269. Vogt, B. A., and M. Gabriel, eds., 1984. Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Handbook, Boston: Birkhauser. Vorhees, C. V., T. M. Reed, K. D. Acuff-Smith, M. A. Schilling, G. D. Cappon, J. E. Fisher, et al., 1995. Long-term learning deficits and changes in unlearned behaviors following in utero exposure to multiple daily doses of cocaine during different exposure periods and maternal plasma cocaine concentrations. Neurotoxicol. Teratol. 17(3):253–264. Wallace, D. R., C. F. Mactutus, and R. M. Booze, 1996. Repeated intravenous cocaine administration: Locomotor activity and dopamine D/D receptors. Synapse 23:152–163. Wang, H.-Y., S. Runyan, E. Yadin, and E. Friedman, 1995. Prenatal exposure to cocaine selectively reduces D1 dopamine receptor-mediated activation of striatal G proteins. J. Pharmacol. Exp. Ther. 273:492–498. Wang, H.-Y., J. M. Yeung, and E. Friedman, 1995. Prenatal cocaine exposure selectively reduces mesocortical dopamine release. J. Pharmacol. Exp. Ther. 273(3):1211–1215. Wang, X.-H., A. O. B. Jenkins, L. Choi, and E. H. Murphy, 1996. Altered neuronal distribution of parvalbumin in anterior cingulate cortex of rabbits exposed in utero to cocaine. Exp. Brain Res. V112(3):359–371. Wang, X.-H., P. Levitt, D. R. Grayson, and E. H. Murphy, 1995. Intrauterine cocaine exposure of rabbits: Persistent elevation of GABA-immunoreactive neurons in anterior cingulate cortex but not visual cortex. Brain Res. 689(1):32–46. Weese-Mayer, D. E., J. M. Silvestri, D. Lin, C. M. Buhrfiend, E. S. Lo, and P. M. Carvey, 1993. Effect of cocaine in early gestation on striatal dopamine and neurotrophic activity. Pediatr. Res. 34(3):389–392.

Whitaker-Azmitia, P. M., 1998. Role of the neurotrophic properties of serotonin in the delay of brain maturation induced by cocaine. Ann. NY Acad. Sci. 846:158–164. Whitaker-Azmitia, P. M., and E. C. Azmitia, 1994. Astroglial 5HT1a receptors and S-100 beta in development and plasticity. Perspect. Dev. Neurobiol. 2:233–238. Whitaker-Azmitia, P. M., R. Murphy, and E. C. Azmitia, 1990. S-100 protein is released from astroglial cells by stimulation of 5-HT1a receptors and regulates development of serotonin neurons. Brain Res. 528:155–160. Wolf, M. E., 1998. The role of excitatory amino acids in behavioral sensitization to psychomotor stimulants. Prog. Neurobiol. 54(6):679–720. Wood, R. D., V. A. Molina, J. M. Wagner, and L. P. Spear, 1995. Play behavior and stress responsivity in periadolescent offspring exposed prenatally to cocaine. Pharmacol. Biochem. Behav. 52:367–374. Wood, R. D., and L. P. Spear, 1998. Prenatal cocaine alters social competition of infant, adolescent, and adult rats. Behav. Neurosci. 112:419–431. Woods, J. R., Jr., M. A. Plessinger, and K. E. Clark, 1987. Effect of cocaine on uterine blood flow and fetal oxygenation. JAMA 257(7):957–961. Xu, M., J. Xiu-Ti, D. C. Cooper, R. Moratella, A. M. Graybiel, F. J. White, et al., 1994. Elimination of cocaineinduced hyperactivity and dopamine-mediated neurophysiological effects in dopamine D1 receptor mutant mice. Cell 79:945–955. Yablonsky-Alter, E., I. Gleser, C. Carter, and M. Juvan, 1992. Effects of prenatal cocaine treatment on postnatal development of neocortex in white mice: Immunocytochemistry of calbindinand paralbumin-positive populations of gabaergic neurons. Soc. Neurosci. Abstracts 18:367. Yan, Q.-S., 2002. Reduced serotonin release and serotonin uptake sites in the rat nucleus accumbens and striatum after prenatal cocaine exposure. Brain Res. 929(1):59–69. Zachor, D., J. Cherkes, C. Fay, and I. Ocrant, 1994. Cocaine differentially inhibits neuronal differentiation and proliferation in vitro. J. Clin. Invest. 93:1179–1185. Zachor, D. A., J. F. Moore, C. M. Brezausek, A. B. Theibert, and A. K. Percy, 2000. Cocaine inhibition of neuronal differentiation in NGF-induced PC12 cells is independent of ras signaling. Int. J. Dev. Neurosci. 18:765–772. Zechner, D., Y. Fujita, J. Hulsken, T. Muller, I. Walther, M. M. Taketo, et al., 2003. [beta]-Catenin signals regulate cell growth and the balance between progenitor cell expansion and differentiation in the nervous system. Dev. Biol. 258(2):406– 418. Zhen, X., C. Torres, H.-Y. Wang, and E. Friedman, 2001. Prenatal exposure to cocaine disrupts D1A dopamine receptor function via selective inhibition of protein phosphatase 1 pathway in rabbit frontal cortex. J. Neurosci. 21(23): 9160–9167. Zhou, J., H. F. Bradford, and G. M. Stern, 1994. The response of human and rat fetal ventral mesencephalon in culture to the brain-derived neurotrophic factor treatment. Brain Res. 656(1): 147–156. Zhou, J., and L. Iacovitti, 2000. Mechanisms governing the differentiation of a serotonergic phenotype in culture. Brain Res. 877(1):37–46.

langlois and mayes: prenatal cocaine exposure

675

41

Neurocognitive Models of Early-Treated Phenylketonuria: Insights from Meta-analysis and New Molecular Genetic Findings MARILYN WELSH, KATHRYN DEROCHE, AND DAVID GILLIAM

In the early part of the 20th century, a Norwegian mother of two developmentally delayed children made a seemingly mundane discovery, setting in motion decades of research in such disparate fields as molecular genetics, biochemistry, neuroscience, nutrition, and psychology that would ultimately lead to the discovery of the single most common, treatable cause of mental retardation, phenylketonuria (PKU). Mrs. Egeland noticed a musty odor surrounding her children, despite frequent bathing, and the strong odor was finally identified as emanating from their urine (Centerwall and Centerwall, 2000). This case was referred to Asbjorn Folling, a Norwegian medical doctor and biochemist who systematically tested the children’s urine. This investigation resulted in the identification of phenylpyruvic acid, the source of the musty odor and, he suspected, a clue to the chain of events that had led to the children’s mental retardation and severe behavioral disorder. Folling’s suspicion was confirmed when his own examination of hundreds of local mentally retarded patients led to the discovery of eight patients, including one pair of siblings, who excreted phenylpyruvic acid in their urine (Christ, 2003). In the decades following Folling’s original discovery, an understanding of the genetics of the disorder emerged, with Jervis and colleagues suggesting that the inheritance of a single autosomal recessive gene was related to a metabolic error in which the ability to hydroxylate phenylalanine (Phe) to tyrosine is disrupted. In PKU individuals, phenylalanine hydroxylase is inactive in the liver, and Phe is converted instead into the phenylpyruvic acid that was excreted in the urine of mentally retarded individuals ( Jervis et al., 1939, 1947, 1953; as cited in Knox, 1972). Given this understanding of the metabolic deficit, an effective preventive treatment was devised in which the dietary levels of Phe were restricted (Armstrong and Tyler, 1955; Bickle, Gerrard, and Hickmans, 1954; as cited in Knox, 1972). A Phe tolerance test was developed to identify

heterozygous carriers of PKU (Hsia et al., 1956, as cited in Knox, 1972). The tolerance test was soon followed by an effective neonatal screening procedure that detected elevated plasma Phe and, thus, identified newborns who were homozygous for the disorder (Guthrie and Susi, 1963; as cited in Knox, 1972). Millions of newborns have been screened for PKU by the Guthrie test since federal mandates were put into place in the early 1960s (Scriver and Clow, 1980), averting severe mental retardation and psychiatric disturbance in individuals who would have otherwise been institutionalized. The story of PKU is one that is recounted in nearly every developmental psychology text published today because it represents a classic model of “nature-nurture interaction.” That is, it is the story of a genetically based disorder, the negative consequences of which can be substantially ameliorated through environmental intervention in the form of a Phe-restricted diet (e.g., Williamson et al., 1981). Phenylketonuria is a disorder in which the links between gene, enzyme, biochemistry/neurochemistry, brain function, and behavior are clearer than in most other neurodevelopmental disorders (Welsh and Pennington, 2000), and the systematic neuropsychological investigation of this condition has led to fascinating insights regarding brain development in general, and the function of specific brain regions, such as the prefrontal cortex, in developing children (e.g., Diamond et al., 1997; Welsh et al., 1990; Welsh, 1996; White et al., 2001). After providing a brief overview of PKU, we will discuss two models that have been proposed to understand the particular profile of neurocognitive characteristics of early-treated PKU, one implicating the prefrontal cortex and the other focusing on white matter integrity. This discussion will be followed by a meta-analysis of 24 years of cognitive and neuropsychological research that is designed to examine the reported differences between early-treated PKU individuals and controls on tests of prefrontal cortical functions (i.e.,

677

executive functions) and measures of white matter integrity (i.e., speed of cognitive processing). Next, we will discuss recent, cutting-edge discoveries in molecular genetics, specifically the identification of the coexistence of the BH4 cofactor with PKU, which may ultimately explain much of the individual variation seen in the behavioral outcomes of PKU today. Finally, based on this review, we will develop some tentative conclusions and generate suggestions for future research.

Overview of phenylketonuria Phenylketonuria is a well-known genetic cause of mental retardation if left untreated, and it affects one in 10,000 to 20,000 live births (Stanbury, Wyngaarden, and Friedrickson, 1983). The disorder is the consequence of mutations in the gene that codes for the enzyme phenylalanine hydroxylase (PAH), which is essential for hydroxylation of dietary Phe to tyrosine in the liver (e.g., Guttler and Lou, 1986). Given normal intake of the amino acid Phe, severe mental handicap results ( Jervis, 1939; as cited in Knox, 1972); however, for nearly five decades newborn screening programs have identified individuals with PKU so that a diet low in Phe may be initiated early in development. If this treatment is begun early and controlled consistently during childhood, mental retardation is averted; however, specific areas of neuropsychological impairment may still exist, even with early dietary treatment. Mutations in the PAH gene can result in benign hyperphenylalaninemia, commonly referred to as mild hyperphenylanlainemia (mHPA), which results in blood Phe levels between 120 and 600 μmol/l. Mutations in the PAH gene resulting in blood Phe levels higher than 600 μmol/l are collectively referred to as phenylketonuria (PKU) (Kayaalp et al., 1997). Phenylketonuria is further defined by two categories: classical PKU and atypical PKU. Classical PKU is considered to be a mutation that results in blood Phe levels over 1,200 μmol/l (Guldberg et al., 1998), whereas atypical PKU is characterized by blood Phe levels ranging from 600 to 1,200 μmol/l (Guldberg et al., 1998). The diagnostic category of PKU most frequently described in the published empirical literature is classical PKU, and this particular classification is the target of our current review and metaanalysis examining the characteristic neurocognitive outcomes. As genetic technologies continue to advance, leading to a more comprehensive understanding of the range of mutations that may be involved in the overall group of hyperphenylalaninemias, as well as within the subcategory of classical PKU, this knowledge base will undoubtedly contribute to our understanding of the individual differences that have been observed in the neuropsychological sequelae of the disorder. The objective of this meta-analysis is to identify whether a consistent profile of deficits has emerged

678

in empirical research published over 24 years of study, from 1980 through 2004. Even with early dietary treatment of PKU, research has suggested that impairments in intellectual, cognitive, and behavioral domains exist in this population of children and adults. The results of the comprehensive, longitudinal PKU Collaborative Study (Williamson, Dobson, and Koch, 1977), conducted in the 1970s, indicated that early dietary treatment is related to intellectual levels in the normal range; however, the IQ scores of children with PKU often do not achieve the levels predicted from the scores of unaffected parents and siblings (e.g., Williamson, Dobson, and Koch, 1977). In addition, studies reporting declines in IQ with age during the school years suggest that such declines can be linked to elevations in Phe caused by poor or absent dietary control (e.g., Waisbren, Schnell, and Levy, 1980). More recently, Burgard (2000) examined the effect sizes for the studies cited by Waisbren, Schnell, and Levy (1980) with a specific focus on longitudinal studies that examined IQ change after diet discontinuation. He found effect sizes ranging from −1.0, reflecting a large drop in IQ after diet discontinuation, to 0, suggesting no change in measured intelligence; thus the negative impact of high Phe on IQ scores is not universal. Although most studies focus on fullscale IQ scores of early-treated PKU children, generally there is the expectation that performance IQ should be more negatively impacted than verbal IQ based on the profile of neurocognitive outcomes observed in decades of research. That is, studies exploring the cognitive skills of early-treated PKU children conducted in the 1970s and 1980s indicate a pattern of cognitive function characterized by intact speech and language skills but impaired visual-spatial, perceptualmotor, and problem-solving abilities. With regard to behavior, early-treated children do not present with a consistent clinical profile; however, when behavioral problems are exhibited they tend to cluster in the areas of hyperactivity, impulsivity, poor planning, and less task persistence (see Welsh and Pennington, 2000, for a review).

Two models of the neuropathology of PKU: Prefrontal cortical function and white matter integrity Although the specific neuropathology of PKU has not been explicated fully, the two most common models discussed currently involve (1) prefrontal dysfunction, secondary to decreased functional dopamine, and (2) abnormalities in white matter as a consequence of demyelination and/or hypomyelination. The neuroanatomical and neurophysiological evidence for the two models, as well as the behavioral sequelae consistent with each, will be discussed in this section. Although the prefrontal and white matter hypotheses tend to be discussed separately as explanations of the neurocognitive consequences of PKU, it is not entirely clear that these

neurodevelopmental aspects of clinical disorders

two models are mutually exclusive, and evidence germane to this issue will be discussed in a later section. In fact, White and colleagues (2001) suggested that the neurocognitive impairments typically observed in early-treated PKU children and adolescents may, in fact, be a consequence of some combination of neurological abnormalities in both the prefrontal cortex and cerebral white matter. Several groups of researchers have proposed that PKU produces a prefrontal dysfunction (e.g., Chamove and Molinaro, 1978; Diamond et al., 1994; Luciana, Hanson, and Whitley, 2004; Pennington et al., 1985; Welsh, 1996), and various lines of biochemical and behavioral evidence are consistent with this hypothesis. First, the biochemical alterations resulting from the genetic mutation cause a disruption in catecholamine biosynthesis, particularly that of the neurotransmitter dopamine, which is essential for prefrontal cortical function. The genetic mutation characteristic of classical PKU disrupts the hydroxylation of Phe to tyrosine (Choo et al., 1979; Guttler and Lou, 1986), the amino acid necessary for the rate-limiting step in the synthesis of dopamine and norepinephrine ( Cooper, Bloom, and Roth, 1991). Tyrosine in the brain is also limited by high levels of Phe, which compete with tyrosine for passage across the blood-brain barrier, further reducing the amount of functional dopamine in the brain (Curtius, Völlmin, and Baerlocher, 1972; McKean, 1972). The dopaminergic neurons in prefrontal cortex are extremely sensitive to even small decreases in tyrosine, given that the neurons of this region lack the synthesis-modulating autoreceptors present in other brain areas (Chiodo et al., 1984). Animal models of classical PKU, both the untreated condition (Chamove and Molinaro, 1978) and the early- and continuously treated condition (Diamond et al., 1994), have demonstrated behavioral deficits characteristic of prefrontal damage. Additionally, reduced CSF levels of dopamine have been observed in untreated PKU individuals, and an inverse association between Phe and dopamine metabolites has been found in early-treated subjects (Burlina et al., 2000; Krause et al., 1986; Salvo, 1985). In a recent study, Luciana, Hanson, and Whitley (2004) demonstrated that haloperidol, a dopamine antagonist, produced deficits on executive-function tasks in both unaffected individuals and those with early-treated PKU; however, the impairments were more severe in the diagnosed group. Second, the hypothesized impairment in prefrontal cortical function, resulting from diminished levels of dopamine, is consistent with decades of cognitive and behavioral research demonstrating a particular profile of deficits in early-treated PKU: lower nonverbal intelligence, impairments in novel problem solving, and impulsive behavior lacking in planning and goal orientation (Welsh and Pennington, 2000). Moreover, studies specifically designed to explore the prefrontal-dysfunction model of PKU have

found evidence of specific impairments in the cognitive domain of executive function, which is characterized by cognitive skills that have been linked to this brain region (Stuss and Benson, 1984). Despite some disagreement regarding definition, there is general consensus that executive function refers to cognitive processes that are necessary for purposeful, future-oriented behavior. These include, but are not limited to, regulation of attention, inhibition of inappropriate responses, coordination of information in working memory, and capacities to organize, sequence, and plan adaptive behavior (Blair, 2002; Eslinger, 1996; Shonkoff and Phillips, 2000; Welsh, 2002; Zelazo and Frye, 1998; Zelazo et al., 2003). Over the past two decades, there has been mounting evidence that children diagnosed with classical PKU exhibit deficits in executive functions such as planning, working memory, and inhibition, even in the context of normal intelligence, and in some cases these deficits have been linked to blood levels of Phe at or near the time of testing (e.g., Diamond et al., 1997; Welsh et al., 1990). In recent studies, executive-function impairments have been observed in children and adolescents with PKU in the form of inhibition and flexibility (e.g., Huijbregts et al., 2002), working memory (e.g., White et al., 2002), and metacognition and strategy use (e.g., Antshel and Waisbren, 2003; White et al., 2001). The white-matter-integrity model of PKU derives from consistent evidence of neuropathology in white matter regions of the brains of individuals diagnosed with PKU. In early postmortem examinations (presumably of persons with untreated PKU), substantial white matter abnormalities were found that varied in degree of damage, as well as location throughout the brain (Alford et al., 1950; Poser and van Bogaert, 1959; Malamud, 1966; as cited in Christ, 2003). More recent neuroimaging studies of individuals with PKU who benefited from early dietary treatment have, likewise, revealed abnormalities in white matter tracts primarily in the parietal-occipital, posterior periventricular, and frontal regions, with damage extending to subcortical white matter in more severe cases of PKU (Bick et al., 1993; Dyer, 1999; Lou et al., 1992; Pietz, Meyding-Lamade, and Schmidt, 1996; Ulrich et al., 1994). Hasselbalch and colleagues (1996) found evidence of structural abnormalities in both frontal and occipital white matter in the brains of early-treated adults with PKU, as well as evidence of reduced glucose metabolism in anterior periventricular white matter. Although the causal mechanism underlying the white matter abnormalities is currently unclear, there is evidence that pathological processes interfering with myelin formation may occur early in neurodevelopment (Pueschel, 1996). These prenatal events result in reduced white matter throughout the brain and a susceptibility to demyelination later in development due to the presence of defective myelin (Christ, 2003). Studies in which hyperphenylalaninemia is

welsh, deroche, and gilliam: early-treated phenylketonuria

679

induced in mice have demonstrated increases in myelin turnover, disrupted synaptogenesis, and neuronal death (Hommes and Moss, 1992; Shedlovsky et al., 1993), suggesting that high blood levels of Phe in humans with PKU may be associated with demyelination of axon tracts. Consistent with this hypothesis, it has been observed that lowering blood Phe levels can partially reverse these abnormalities in cerebral white matter (Cleary et al., 1995). Neurocognitive findings are also supportive of the whitematter-integrity model of early-treated PKU, as cognitive functions presumed to be mediated by white matter tracts have been found to be impaired in diagnosed individuals. In general, the white matter tracts, particularly those connecting cortical regions, are presumed to subserve informationprocessing speed (Banich, 2004). Slowed information processing has been observed in early-treated PKU individuals in several recent studies (e.g., Gourovitch et al., 1994; Griffiths et al., 2005; Waisbren et al., 1994; White et al., 2001). Other deficits consistent with the white matter model that have been documented include speeded motor responses (Huijbregts et al., 2003; Pietz et al., 1998) and slowed interhemispheric transfer time reflecting potentially abnormal myelination of the corpus callosum (Banich et al., 2000; Gourovitch et al., 1994). It is important to note that many of the tasks used to assess information-processing speed (e.g., Stroop, Continuous Performance Task, etc.) also involve demands for executive-function skills, such as controlled and flexible attention, working memory, and inhibition. In light of this fact, it is difficult to know the degree to which speed-of-processing deficits actually reflect executive impairment or, conversely, the extent to which deficits on certain executive-function measures are better explained as processing-speed difficulties. To address this question, studies by Channon and colleagues (Channon et al., 2004; Channon, Mockler, and Lee, 2005) have attempted to disentangle these explanations by independently manipulating the executive-function and processing-speed demands in a working-memory task. Using this methodology, Channon, Mockler, and Lee (2005) found that the performance of adults with early-treated PKU was more disrupted by processing-speed manipulations than by the executivefunction manipulations, providing support for the white matter hypothesis. It is important to note that this cognitive slowing account of the cognitive deficits in early-treated PKU was identified in a sample of adults, and the metaanalysis described in the following section is based on studies almost exclusively conducted on children and adolescents. In summary, there is substantial evidence for both prefrontal cortical dysfunction due to depleted functional dopamine and abnormalities in the myelination of white matter tracts in the brain. These two models of the neuropathology underlying PKU may not be mutually exclusive; both types of abnormalities may coexist in the brains of

680

some or all individuals with early-treated PKU. The two models do predict somewhat different patterns of neuro cognitive sequelae, with a prefrontal dysfunction resulting in executive-function deficits and abnormalities in white matter leading to specific impairments in information-processing speed. In what follows, the published literature on earlytreated PKU between the years of 1980 and 2004 has been examined with meta-analytic techniques to compare the two accounts of the neurocognitive outcomes of early-treated PKU—that is, the prefrontal-dysfunction/executive-function model and the white-matter/speed-of-processing model.

Neurocognitive outcomes in early-treated PKU: A meta-analytic approach Research studies investigating the neuropsychological abilities of early-treated PKU individuals tend to be characterized by methodological limitations, including low statistical power resulting from small sample sizes, low generalizability due to predominate use of nonexperimental designs, heterogeneous inclusion requirements for participants, and a diverse array of measurement tools. The methodological limitations result in discrepancies among the conclusions drawn from the primary studies (Antshel and Waisbren, 2003; Huijbregts et al., 2002; Luciana, Sullivan, and Nelson, 2001; Stemerdink et al., 1999; White et al., 2001), as reflected in the empirical support for the two models of cognitive deficits in early-treated PKU. The lack of consensus among findings precludes the identification of a “clear profile” of neuropsychological functioning. To examine the presence of consistent themes in the neurocognitive research conducted over 24 years (1980–2004), Welsh and colleagues (DeRoche and Welsh, in press; Welsh, Kelley, and Tran, 2000) conducted meta-analyses of the research involving the performance of early-treated PKU individuals on a variety of cognitive domains. Meta-analysis represents one reasonable approach to characterizing general findings of cognitive function in early-treated PKU by allowing the collective analysis of primary studies that involve different participants, sample sizes, and methodological conditions, while adjusting for sample size. In the process of combining primary studies, statistical power is increased, and the researcher is afforded the opportunity to test competing theoretical models of the behavior in question (H. Cooper and Hedges, 1994). Two analyses were conducted to establish overall measures of effect size (Hedges’ g and r, respectively) for sets of studies that examined (1) group differences between earlytreated PKU individuals and control individuals on various measures of executive function, processing speed, and intelligence, and (2) the correlations between concurrent blood Phe levels in diagnosed individuals and scores on relevant measures of the three cognitive domains. The prefrontaldysfunction model was examined by calculating the average

neurodevelopmental aspects of clinical disorders

effect sizes for executive-function measures as a whole, as well as for the subdomains of working memory, inhibition, flexibility (i.e., set shifting), planning, and “other” (e.g., self-monitoring, organization). The white-matter-integrity model was explored with regard to the average effect sizes for processing-speed tasks overall, as well as the subdomains of pure speed of processing (e.g., choice reaction time) and speeded attention (e.g., Continuous Performance Test). Effect sizes for both domains were compared to effect sizes for intelligence (full scale, verbal, performance, and “other”), as a baseline for general cognitive function that is not directly related to either neurocognitive model. The meta-analysis synthesized data from 71 primary studies and 438 outcomes, collected from selected databases that commonly included studies of neuropsychological outcomes in early-treated PKU from 1980 to 2004. Based on the low prevalence rates of PKU, participants may have contributed to outcomes of more than one of the dependent variables; therefore, independence among domain scores and among studies cannot be assumed. All primary studies were coded into an Excel file by two trained researchers. The file comprised 28 items of information falling into three categories: (1) study characteristics, (2) subject characteristics, and (3) study outcomes. In general, the outcomes were coded as executive function, processing speed, or intelligence (and specific subdomains) in accordance with the operational definitions provided by the author(s) of the primary study. Effect sizes were calculated for all outcomes within each domain and pooled to produce the overall domains of intelligence, executive function, and processing speed. With regard to the group difference analysis, table 41.1 includes the domains and subdomains of interest, the effect

Domain Pure speed of processing Speeded attention Speed-of-processing total Executive function–flexibility Executive function–inhibition Executive function–other Executive function–planning Executive function–working memory Executive function–total

sizes for the group differences, confidence intervals, z-scores, and significance levels. We followed Cohen’s (1992) index for the interpretation of effect sizes for group differences (d) (i.e., small .20, moderate .50, and larger .80), with a positive effect size indicative of a “more impaired” performance by the PKU group for the behavioral domain. The domain of executive function contained 121 outcomes and produced the largest effect size (.79, p < .01), suggesting that PKU individuals had a large deficit in their performance on executive-function tasks, with the most impaired functioning displayed for flexibility (1.15, p < .01) and inhibition (.78, p < .01). Analysis of executive function did produce a larger effect size than the total IQ domains (.42, p < .01), indicating relatively greater executive function impairments observed in PKU individuals than in controls, a finding that is consistent with the prefrontal-dysfunction model of early-treated PKU. The domain of processing-speed produced a moderate effect size (.58, p < .01), suggesting that PKU individuals have a moderate deficit in processing-speed ability compared to controls, providing some support for the white matter hypothesis. However, it is important to note that the average of effect size for processing speed was not much larger than the effect size identified for intelligence. Importantly, the two subdomains of processing speed exhibited quite different effect sizes, with pure speed of processing tasks (e.g., choice reaction time) associated with a small effect (.23, p < .01) and speeded attention tasks with a large effect size (.87, p < .01). As will be discussed later in the chapter, it could be argued that many of these speeded attention tasks also involved demands for executive-function skills. In summary, the group-difference analysis provided fairly strong support for the prefrontal model of early-treated

Table 41.1 Effect sizes for group difference analysis: PKU versus controls Confidence Intervals N Effect Size (Hedge’s g) Lower Upper

Z-score

p-value

42 23 63

.23 .87 .58

.10 .74 .49

.37 .99 .67

3.37 13.89 12.55

.00* .00* .00*

37 29 15 19 21 121

1.15 .78 .63 .51 .59 .79

1.05 .67 .50 .39 .45 .73

1.25 .90 .76 .64 .73 .84

22.54 13.24 9.40 7.92 8.25 28.64

.00* .00* .00* .00* .00* .00*

.49 .36 .44 .43 .47

10.42 2.67 4.20 3.19 14.46

.00* .01* .00* .00* .00*

IQ–full scale 40 .42 IQ–other 4 .20 IQ–performance 14 .30 IQ–verbal 13 .29 IQ–total 80 .42 Note: The asterisk indicates that a significant p-value was obtained.

.34 .051 .16 .14 .36

welsh, deroche, and gilliam: early-treated phenylketonuria

681

PKU, as reflected by large effect sizes, and some support for the white-matter-integrity model, as seen in the moderate effect sizes identified. The second analysis examined the association between the measured blood Phe level, at or near the time of testing (i.e., “concurrent level”), and the neuropsychological outcomes of interest within samples of early-treated PKU individuals. A predicted negative correlation can be interpreted as higher blood Phe levels associated with “lower” (i.e., more impaired) scores on the measures of intelligence, executive function, and processing speed. With regard to this second analysis, effect sizes were interpreted according to Cohen’s (1992) index for correlations (r) (i.e., small .10, moderate .30, and larger .50), with higher numbers indicating more shared variance between Phe level and performance scores. Table 41.2 reports the number of outcomes, effect size (r), confidence intervals, z-scores, and significance values for the relationship between concurrent blood Phe level and executive function, speed of processing, and intelligence. In contrast to the previous analysis, executive function as a whole produced a low negative relationship (−.15, p < .01), implying that concurrent blood Phe levels of participants with PKU have a weak negative association with performance on executive tasks. In addition, executivefunction abilities displayed a smaller association with Phe level than did intelligence (−19, p < .01). However, upon closer examination of the subdomains of executive function, one finds a moderate to large relationship between concurrent blood Phe level and performance on tests of working memory (−.50, p < .01), planning (−.38, p < .01), and “other”

Domain Pure speed of processing Speeded attention Speed-of-processing

executive function skills (−.40, p < .01). Ironically, the domains of flexibility and inhibition, which had the largest effect sizes in the group-difference analysis, were correlated with concurrent blood Phe levels at a very low level (.04 for flexibility and −.15 for inhibition). Because the flexibility and inhibition subdomains contributed the most outcomes to the analysis of executive function, this result served to lower the overall correlation between concurrent Phe and this domain. The analysis of the processing-speed domain produced a moderate negative relationship (−.25, p < .01), but it was highest among all three cognitive domains, implying that PKU individuals with lower blood Phe levels performed better (i.e., faster) on speed-of-processing tasks. The pure speed-of-processing tasks produce a low relationship between blood Phe levels and performance scores (−.10, p < .01), while performance on speeded attention tasks displayed a moderate to large relationship between Phe levels and performance scores (−.40, p < .01). Interestingly, the effect size (i.e., correlation with Phe level) for performance on speeded-attention tasks was at a level exhibited by the executive-function subdomains of planning and working memory. These meta-analytic results provide support for both models. The group difference analysis is consistent with the prefrontal-dysfunction model, with larger effect sizes (i.e., PKU deficits) revealed for performance on executivefunction tasks. Processing-speed correlations with concurrent blood Phe levels, in the expected direction, provide support for the white-matter-integrity hypothesis. That this particular analysis would lend support to the white matter hypothesis is interesting, given that higher Phe levels should be

Table 41.2 Relationship between concurrent blood Phe level and functional ability Confidence Intervals N Effect Size (r) Lower Upper 11 −.10 −.30 −.09 6 −.40 −.54 −.24 17 −.25 −.33 −.16

Z-score −3.56 −4.63 −5.43

p-value .00* .00* .00*

.04 −.15 −.40 −.38 −.50 −.15

−.05 −.26 −.52 −.57 −.63 −.21

.12 −.04 −.26 −.15 −.34 −.10

.85 −2.74 −5.40 −3.14 −5.55 −5.24

.40 .01* .00* .00* .00* .00*

IQ–full scale 18 −.35 10 .01 IQ–other IQ–performance 2 −.41 IQ–verbal 2 −.32 IQ–total 106 −.19 Note: The asterisk indicates that a significant p-value was obtained.

−.43 −.11 −.61 −.54 −.22

−.27 .13 −.16 −.07 −.15

−7.99 .14 −3.16 −2.44 −9.99

.00* .89 .00* .02* .00*

Executive function–flexibility Executive function–inhibition Executive function–other Executive function–planning Executive function–working memory Executive function–total

682

17 13 10 6 3 49

neurodevelopmental aspects of clinical disorders

associated with lower availability of dopamine in the prefrontal cortex, which should result in executive-function impairments. However, it is important to point out that the correlations for speeded-attention measures of processing speed were at the same level as the correlations for the planning, working-memory, and “other” (i.e., organizational skills) measures of executive function, consistent with the speculation that these tasks may share demands for common cognitive processes. Moreover, an additional explanation for the weak association between some subdomains of executive function and concurrent Phe level may reflect a “history effect.” That is, from the late 1990s through 2004, studies of executive function took center stage, while at the same time most PKU children were treated under current recommendations for tighter dietary control. That is, these later studies of executive function tended to involve earlytreated PKU individuals with lower and less variable Phe levels, potentially contributing to the correlations of lower magnitude. One way to reconcile the support in this meta-analysis for both neurocognitive models of early-treated PKU is to examine evidence for an association between the depleted dopamine and the white matter hypomyelination and demyelination observed in treated PKU. Joseph and Dyer (2003) describe an alternative theory that has been developed to understand the neurobiological mechanisms underlying PKU. The most common perspective, the tyrosine/dopamine theory, was described earlier as the tendency toward low tyrosine levels in the central nervous system resulting from tyrosine competing for transport systems with abnormally high concentrations of Phe. The relative lack of tyrosine, the rate-limiting step in DA synthesis, leads to decreased functional DA in the brain. An alternative is the myelin/dopamine theory in which myelin/axonal interactions are assumed to convert signals relating to the up-regulation of the key enzyme in the biosynthetic pathway of DA, tyrosine hydroxylase. Therefore, this perspective suggests that myelin/axonal contact is a contributor to the synthesis of DA, either increasing or decreasing functional DA in concert with the existence of myelinated axonal tracts. Using a genetic mouse model of PKU involving a PAH gene mutation, Joseph and Dyer (2003) designed a study to examine the evidence for each of these two theories. The authors found that exposing their PKU mice to a low-Phe diet for a period of four weeks was followed by a rebound of DA levels to near-normal levels in the frontal cortex and striatum regions, with a corresponding increase in myelination in these same brain structures. In contrast, central levels of tyrosine did not rebound to normal levels after the fourweek diet; these levels were at 65 percent of normal in the striatum and 68 percent of normal in the frontal cortex at the conclusion of the diet. The authors’ conclusions from these data were that myelin production, and not tyrosine

levels, was the more effective regulator of DA synthesis. One mechanism proposed is that contact between myelin and axons stimulates the phosphorylation of existing tyrosine hydroxylase. Therefore, if tyrosine and corresponding enzymes are deficient, as they continued to be in the brains of the mice on low-Phe diets, the existence of myelinated axons in frontal cortex and striatum enhances the activity level of the tyrosine hydroxylase that is available, thus synthesizing more DA. The myelin/axonal theory of the neurobiological underpinnings of PKU is compelling, particularly given the results of our meta-analysis. This perspective clearly highlights the possibility that the prefrontal-dysfunction/executivefunction model and the white-matter-integrity/processingspeed models are not mutually exclusive. First, the findings of Joseph and Dyer (2003) clearly show that demyelination or aberrant myelination of axon tracts in the frontal cortex and/or striatum would be expected to result in disrupted DA synthesis. According to the prefrontal-dysfunction model, the lower levels of functional dopamine in the prefrontal cortex, where DA receptors are particularly sensitive to small fluctuations, are the cause of the executive function deficits commonly observed in children with early-treated PKU. Second, the hypo- and/or demyelination of axonal tracts in the frontal cortex and striatum would be expected to lead to impairments in processing-speed tasks, commonly observed in this population. Therefore, the white matter abnormalities, which have been documented for decades in neuroimaging studies of early-treated PKU, would be predicted to manifest in processing-speed deficits; however, the dysregulation of DA synthesis and the decreased levels of functional DA in the prefrontal cortex also would be expected to result in executive-function impairments. This expectation dovetails nicely with the findings of our meta-analysis. In the following section, we discuss new findings in molecular genetics that, in the coming decades, will undoubtedly provide one of the key explanations for the individual differences in cognitive functioning found in early-treated PKU individuals. Discoveries of different genetic mutations currently are being linked to differential responsiveness to treatment and to an alternative to the low-Phe diet, as in the case of BH4 genotype. The goals of such research will be to map the genotypes to protein synthesis, enzymatic activity, neurodevelopment and neural function, and finally to cognition and behavior; however, the state of knowledge is in relatively early stages of this exploration. As stated by Scriver and Kaufman (2001, 1667–1724), All studies of genotype-phenotype correlations reveal reasonable correlations at the proximal (enzyme) level; but, at intermediate (metabolic) and distal (cognitive) levels, phenotypes have emergent properties and behave as complex traits in which the effects of PAH, the major locus, [are] modulated by “modifiers.”

welsh, deroche, and gilliam: early-treated phenylketonuria

683

Genetics and individual differences in outcome: The tetrahydrobiopterin-responsive form of phenylalanine hydroxylase deficiency As we have discussed, the deficiency in phenylalanine hydroxylase due to mutations in the PAH gene (also called a locus) leads to accumulation of phenylalanine in the blood and brain, which results in PKU. More than 500 diseasecausing mutations are found in the PAH gene. A Web site (http://www.pahdb.mcgill.ca/) is maintained that contains access to up-to-date information about mutations at the phenylalanine hydroxylase locus. Tetrahydrobiopterin, or BH4, is an essential cofactor that is used by the PAH enzyme to convert phenylalanine to tyrosine. Since people inherit two copies of the PAH gene, one from their mother and one from their father, each copy of the gene may contain a different mutation. A complete listing of known BH4-responsive genotypes can be found at http://www.bh4.org/BH4DatabasesBiopku.asp. Most genotypes that are BH4-responsive show some PAH enzyme activity, and recent research has found that some PKU patients, typically with the milder form of the disorder, benefit from BH4 administration because it results in decreased phenylalanine levels in the blood (Blau and Erlandsen, 2004). Patients who are sensitive to BH4 are identified by a BH4 loading test, but how this test is conducted and how it is interpreted remain open for further study (Baldellou Vazquez et al., 2006). We now know that some PKU patients are BH4responsive and show reduced phenylalanine levels without Phe-restricted diets (Kure et al., 1999). Oral administration of BH4, in pill form appears to stabilize the PAH enzyme and prevent its degradation. This process is referred to as the chaperone-like activity of BH4 on the PAH gene (other mechanisms of BH4-responsiveness are discussed in N. Blau and Erlandsen, 2004). The stabilized PAH enzyme is able to convert phenylalanine to tyrosine, an essential amino acid involved in the synthesis of several neurotransmitters (Bernegger and Blau, 2002). Deficiency in BH4 may lead not only to PKU, but also to impaired synthesis of l-dopa and dopamine, serotonin, and nitric oxide (Lee et al., 2006), which can manifest as neurological symptoms. Given the role these neurotransmitters play in movement, reinforcement, memory, and problem solving (Luciana, Hanson, and Whitley, 2004), BH4 therapy or therapies augmenting neurotransmitter functioning could be beneficial for many PKU patients. As mentioned, some PKU patients benefit from BH4 treatment because they are no longer required to maintain a diet low in phenylalanine. One of the first PKU patients treated with BH4 for longer than three years (5 mg BH4/kg/ day) showed plasma Phe levels averaging 130 μmol/L without Phe restriction (Spaapen and Rubio-Gozalbo, 2003).

684

Levels of Phe in other mild PKU patients receiving BH4 therapy decreased from 934 μmol/L to values ranging from 84 to 222 μmol/L (Trefz, Aulela-Scholz, and Blau, 2001). Other researchers report treating PKU patients with BH4 doses ranging from 3.5 mg BH4/kg/day to 50 mg BH4/day, resulting in significant reduction in Phe levels (Bonafé et al., 2001; Lindner et al., 2001; Steinfeld et al., 2002). These levels of circulating Phe with BH4 therapy are well within the range that characterizes mild hyperphenylalaninemia (mHPA; Kayaalp et al., 1997). Adult patients with PKU also appear to benefit from BH4 therapy. One 25-year-old female patient who received BH4 supplementation (100 mg/day) showed a 33 percent decrease in her Phe level and improved psychological affect. Amelioration of depression and panic attacks led to the discontinuation of psychotropic medication (Koch, Guttler, and Blau, 2002). There is a complex cascade of events that may be associated with BH4 deficiency, and this condition can be caused by defects in the enzymes that convert precursors to BH4 (see figure 41.1). One enzyme (6-pyruvoyl-tetrahydropterin synthase or PTPS) is associated with 60 percent of all BH4 deficiencies (Blau and Diazani, 2003). Other enzymes involved in the biosynthesis or regeneration of BH4 include DHPR (dihydropteridine reductase), GTPCH (guanosine triphosphate cyclohydrolase I), SR (sepiapterin reductase), and PCD (pterin-4α-carbinolamine dehydratase). An overview of the biosynthetic pathways of BH4 is given in Lee and colleagues (2006) (or at http://www.bh4.org/BH4_ Deficiency_Biochemistry.asp). Blau and Burgard (2006) provide recommendations for treating each type of enzyme deficiency at different ages of development. These recommendations include protein requirements, Phe tolerance, target blood Phe, and Phe-free amino acid mixture. In one of the few long-term studies that used BH4 therapy, Lee and colleagues (2006) reported on the beneficial effects of treating PTPS deficiency with BH4 and neurotransmitter precursors over a 15-year period. Ten PKU patients showed improved intelligence (increase IQ ≈ 20 points), decreased seizures, and improved eye contact. Other neurological symptoms, which included limb spasticity, truncal hypotonia, dysphagia, and hypersalivation, were lessened over a 6–12 month period of BH4 treatment. The absence of a positive correlation between predicted severity of PAH mutation and cognitive development in untreated PKU patients, either within or between families, implies that PKU is a complex disease (see comprehensive review by Scriver and Kaufman, 2001). Furthermore, blood phenylalanine levels cannot be predicted reliably from mutant PAH genotypes, suggesting that PKU is a genetically heterogeneous trait under the control of multiple interacting mechanisms (Leuzzi et al., 2006). According to Scriver and Kaufman (2001), in addition to individual differences of BH4 generation or regeneration, other allelic variations, such as

neurodevelopmental aspects of clinical disorders

GTP GTPCH DHNP

Biopterin

PTPS 6PTS SR BH4

Phenylalanine

Tyrosine

Tryptophan

TH

TPH

Arginine

DHPR Neopterin

qBH2

PAH

NOS

PCD 4 α OHBH4

Tyrosine

L-DOPA

5-OH-tryptophan

NO

Dopamine

Serotonin

Citrulline

+

Figure 41.1 The biochemical pathway of tetrahydrobiopterin (BH4) metabolism. Abbreviations: 4αOHBH4, pterin-4αcarbinolamine; 6PTS, 6-pyruvoyl-tetrahydropterin; BH4, tetrahydropterin; DHNP, dihydroneopterin triphosphate; DHPR, dihydropteridine reductase; GTP, guanosine triphosphate; GTPCH, guanosine triphosphate cyclohydrolase I; NO, nitric

oxide; NOS, nitric oxide synthase; PAH, phenylalanine hydroxylase; PCD, pterin-4α-carbinolamine dehydratase; PTPS, 6-pyruvoyl-tetrahydropterin synthase; qBH2, quininoid dihydrobiopterin; SR, sepiapterin reductase; TH, tyrosine hydroxylase; TPH, tryptophan hydroxylase. (This figure is modified from Lee et al., 2006.)

the brain phenylalanine transporter, may modify PKU phenotypes and modulate phenylalanine homeostasis. Since concentration of free phenylalanine in brain tissue is the major determinant of brain phenotype in PKU, understanding the genetic (and nongenetic) mechanisms responsible for differences in transport of phenylalanine into brain cells will help explain variations in cognitive development in PKU individuals and in patients with similar mutant PAH genotypes (Scriver and Kaufman, 2001).

meta-analysis of studies that reported associations between concurrent Phe level and test performance indicated somewhat higher correlations for processing-speed tasks than for, at least some, executive-function tasks. It is important to point out that these two sets of analyses (i.e., group comparisons and correlations) only can be compared in an indirect manner, given that, in most cases, different studies contributed to each analysis. That is, one can point to many potential factors (e.g., sample demographics, dietary control, assessments utilized, methods of data collection) that might contribute to these apparently contradictory findings. Nevertheless, we draw the following tentative conclusion from both sets of meta-analyses: there is convincing evidence for deficits in executive-function skills and processing speed, and these are frequently linked to circulating blood levels of Phe at or near the time of testing. Conducting such a quantitative review of the literature serves to highlight ways in which a research literature can be improved. One suggestion we will offer is that both group differences and Phe-level– task-performance correlations should be included routinely within the same study (e.g., Diamond et al., 1997; Welsh et al., 1990). In this way, we can begin to examine the degree to which evidence of neurocognitive patterns of strengths and weaknesses found in group-difference studies (i.e., specific neurocognitive deficits relative to controls) converges with that reported for within-group studies (i.e., specific neurocognitive performances that are associated with blood Phe level at the time of testing) while controlling for various differences in study methodology.

Conclusions and future directions for research This chapter explored the reported patterns of neurocognitive deficit in early-treated PKU and the recent molecular genetics findings, specifically related to the BH4 cofactor, as one potential explanation for cognitive and behavioral individual differences that have been observed in this population. We will conclude with three issues that have been highlighted by the analyses and evidence discussed, as well as suggestions for future research that directly bear on these issues. First, we presented a quantitative review of the evidence over a 24-year period for neurocognitive deficits in earlytreated PKU that putatively reflect either prefrontal cortical dysfunction, secondary to dopamine depletion, or white matter abnormalities. The meta-analysis of group differences suggested that children with early-treated PKU exhibit larger deficits relative to control children on executivefunction tasks than on processing-speed tasks. However, the

welsh, deroche, and gilliam: early-treated phenylketonuria

685

Second, these meta-analyses, as well as preliminary evidence of a myelin-dopamine theory of early-treated PKU as reported by Joseph and Dyer (2003), suggest that the prefrontal/executive-function model and white-matter/ processing-speed model should not be considered to be competing, mutually exclusive models of neurocognitive deficit for this disorder. There is evidence for deficits in both cognitive domains, when examining the group difference and correlational studies; however, again we are typically comparing studies that may involve somewhat different samples, assessments, and methods. Therefore, a second suggestion for future research is to examine both executive-function skills and processing-speed efficiency within the same study, as demonstrated in recent studies by Channon and colleagues (2004; Channon, Mockler, and Lee, 2005) and Luciana, Hanson, and Whitley (2004). Only when researchers investigate both domains of cognition (as well as discriminant domains such as language), while holding constant important variables such as sample demographics, dietary treatment, and the like, can we begin to compare the nature of the deficits in executive function and processing speed directly. If the myelin-dopamine model of the neuropathology of PKU proves to be a valid explanation for even a subgroup of affected individuals, then one would expect to see evidence of both executive-function impairments and slowed processing speed, reflecting both decreased functional dopamine in the prefrontal cortex and white matter abnormalities. It is important to point out that the suggestion to include measures of executive function and processing speed within the same study is complicated by the fact that researchers in the field are still debating how to measure each of these constructs “cleanly” (Anderson, 2002; Espy and Kaufman, 2002; Welsh, 2002; Welsh, Friedman, and Spieker, 2006). With the exception of “pure” speed-of-processing tasks, such as simple and choice reaction-time tests, many processingspeed tasks demand aspects of executive function (e.g., working memory, inhibition). For that matter, it is likely that many executive-function tasks, even if untimed and devoid of specific latency of response measures (e.g., Tower of Hanoi, Wisconsin Card Sorting Test), may nevertheless benefit from faster information processing as one attempts to plan sequences of actions, inhibit responses, and flexibly modify problem-solving strategies (DeRoche and Welsh, in press). Channon and colleagues propose that it is underlying processing speed that contributes to deficits that have been observed on executive-function tasks; however, future research may reveal that it is some core executive process, such as working memory or inhibitory control, that is responsible for the impairments we observe on processingspeed measures. Third, the new findings with regard to the BH4-responsive form of PKU, as well as the 500 or more different mutations

686

evident on the PAH gene, open the door to an exciting avenue of research that will begin to document genotypephenotype correlations. Whereas there is evidence that genotypes are linked to phenotypic variability with regard to Phe level (e.g., Kayaalp et al., 1997), there has been a dearth of research associating genotypes to phenotypic variability in terms of neurocognitive function. In fact, as described in an earlier section, there is only one study we know of that has examined the consequences of BH4 therapy on cognitive function, and this study reported intelligence scores only (Lee et al., 2006). The meta-analysis findings reported here, as well as the conclusions of several qualitative reviews (e.g., Welsh and Pennington, 2000), suggest that general cognitive function, as is measured by intelligence tests, may not be the most fruitful approach for exploring the specific cognitive sequelae that may be associated with the different genotypes involved in PKU. For example, we have discussed the importance of the BH4 cofactor for the conversion of Phe to tyrosine and the negative implications this holds for production of the dopamine necessary for effective prefrontal function. Is it possible that PKU individuals who also prove to be BH4-responsive will exhibit a different pattern of neurocognitive deficits— perhaps more specifically in the area of executive functions— than PKU individuals who are not BH4-responsive? In any case, the fast-moving field of molecular genetics is demonstrating that PKU is clearly a heterogeneous disorder that will undoubtedly be associated with a range of cognitive profiles. From the first systematic investigations of this genetic disorder by Folling in the 1930s, through the comprehensive, multisite studies of treatment efficacy in the 1970s and 1980s, to the current cutting-edge investigations utilizing the best tools that genetic science, neuroscience, and cognitive science have to offer, each era reveals important insights into a complex genetic disorder that is amenable to environmental intervention. The insights we will continue to gain on the mechanisms and outcomes of PKU not only will be useful to developing more effective treatments of this disorder, but will also undoubtedly provide a window on both typical and atypical brain and behavior relationships. REFERENCES Anderson, V., 2002. Executive function in children: Introduction. Child Neuropsychol. 8:69–70. Antshel, K. M., and S. E. Waisbren, 2003. Timing is everything: Executive functions in children exposed to elevated levels of phenylalanine. Neuropsychology 17(3):458–468. Baldellou Vazquez, A., M. I. Salazar Garcia-Blanco, M. P. Ruiz-Echarri Zalaya, C. Campos Calleja, L. Ruiz Desviat, and M. Ugarte Perez, 2006. Tetrahydrobiopterin therapy for hyperphenylalaninemia due to phenylalanine hydroxylase deficiency. When and how? An. Pediatr. (Barc.) 64(2):146–152.

neurodevelopmental aspects of clinical disorders

Banich, M. T., 2004. Cognitive Neuroscience and Neuropsychology, 2nd ed. Boston: Houghton Mifflin. Banich, M. T., A. M. Passarotti, D. A. White, M. J. Nortz, and R. D. Steiner, 2000. Interhemispheric interaction during childhood. II. Children with early-treated phenylketonuria. Dev. Neuropsychol. 18(1):53–71. Bernegger, C., and N. Blau, 2002. High frequency of tetrahydrobiopterin-responsiveness among hyperphenylalaninemias: A study of 1,919 patients observed from 1988 to 2002. Mol. Gen. Metab. 77(4):304. Bick, U., K. Ullrich, U. Stober, H. Moller, G. Schuierer, A. C. Ludolph, et al., 1993. White matter abnormalities in patients with treated hyperphenylalaninaemia: Magnetic resonance relaxometry and proton spectroscopy findings. Eur. J. Pediatr. 152(12):1012–1020. Blair, C., 2002. School readiness: Integrating cognition and emotion in a neurobiological conceptualization of children’s functioning at school entry. Am. Psychol. 57(2):111–127. Blau, N., and P. Burgard, 2006. Disorders of phenylalanine and tetrahydrobiopterin. In N. Blau, G. F. Hoffmann, J. Leonard, and J. T. R. Clarke, eds., Physicians Guide to the Treatment and Follow-up of Metabolic Diseases, 25–34. Heidelberg: Springer. Blau, N., and H. Erlandsen, 2004. The metabolic and molecular bases of tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency. Mol. Gen. Metab. 82:101–111. Blau, T. B., and N. Diazani, 2003. BIOMDB: Database of mutations causing tetrahydrobiopterin deficiency. Retrieved July 15, 2006, from http://www.bh4.org/BH4DatabasesBiomdb.asp. Bonafé, L., N. Blau, A. P. Burlina, A. Romstad, F. Güttler, and A. B. Burlina, 2001. Treatable neurotransmitter deficiency in mild phynylketonuria. Neurology 57:908–911. Burgard, P., 2000. Development of intelligence in early treated phenylketonuria. Eur. J. Pediatr. 159(Suppl. 2):S74–S79. Burlina, A. B., L. Bonafe, V. Ferrari, A. Suppiej, F. Zacchello, and A. P. Burlina, 2000. Measurement of neurotransmitter metabolites in the cerebrospinal fluid of phenylketonuric patients under dietary treatment. J. Inherit. Metab. Dis. 23(4):313–316. Centerwall, S. A., and W. R. Centerwall, 2000. The discovery of phenylketonuria: The story of a young couple, two retarded children, and a scientist. Pediatrics 105:89–103. Chamove, A. S., and T. J. Molinaro, 1978. Monkey retarded learning analysis. J. Ment. Defic. Res. 22:223. Channon, S., E. German, C. Cassina, and P. Lee, 2004. Executive functioning, memory, and learning in phenylketonuria. Neuropsychology 18(4):613–620. Channon, S., C. Mockler, and P. Lee, 2005. Executive functioning and speed of processing in phenylketonuria. Neuropsychology 19(5):679–686. Chiodo, L. A., M. J. Bannon, A. A. Grace, R. H. Roth, and B. S. Bunney, 1984. Evidence for the absence of impulseregulating somatodendritic and synthesis-modulating nerve terminal autoreceptors on subpopulations of mesocortical dopamine neurons. Neuroscience 12(1):1–16. Choo, K. H., R. G. Cotton, D. M. Danks, and I. G. Jennings, 1979. Genetics of the mammalian phenylalanine hydroxylase system: Studies of human liver phenylalanine hydroxylase subunit structure and of mutations in phenylketonuria. Biochem. J. 181(2):285–294. Christ, S. W., 2003. Asbjorn folling and the discovery of phenylketonuria. J. Hist. Neurosci. 12:44–54. Cleary, M. A., J. H. Walter, J. E. Wraith, F. White, K. Tyler, and J. P. Jenkins, 1995. Magnetic resonance imaging in phenyl-

ketonuria: Reversal of cerebral white matter change. J. Pediatr. 127(2):251–255. Cohen, J., 1992. A power primer. Quant. Meth. Psychol. 112:155– 159. Cooper, H., and L. V. Hedges, 1994. The Handbook of Research Synthesis. New York: Russell Sage Foundation. Cooper, J. R., F. E. Bloom, and R. H. Roth, 1991. The Biochemical Basis of Neuropharmacology. New York: Oxford University Press. Curtius, H. C., J. A. Völlmin, and K. Baerlocher, 1972. The use of deuterated phenylalanine for the elucidation of the phenylalanine-tyrosine metabolism. Clin. Chim. Acta 37:277– 285. DeRoche, K., and M. C. Welsh, in press. Twenty-four years of research on the neurocognitive outcomes in early-treated phenylketonuria: Intelligence and executive function. Dev. Neuropsychol. Diamond, A., V. Ciaramitaro, E. Donner, S. Djali, and M. B. Robinson, 1994. An animal model of early-treated PKU. J. Neurosci. 14(5 Part 2):3072–3082. Diamond, A., M. B. Prevor, G. Callender, and D. P. Druin, 1997. Prefrontal cortex cognitive deficits in children treated early and continuously for PKU. Monogr. Soc. Res. Child Dev. 62(4):1– 208. Dyer, C. A., 1999. Pathophysiology of phenylketonuria. Ment. Retard. Dev. Disabil. Res. Rev. 5(2):104–112. Eslinger, P. J., 1996. Conceptualizing, describing, and measuring components of executive function: A summary. In G. R. Lyon and N. A. Krasnegor, eds., Attention, Memory, and Executive Function, 367–395. Baltimore: Paul H. Brookes. Espy, K. A., and P. M. Kaufmann, 2002. Individual differences in the development of executive function in children: Lessons from the delayed response and A-not-B tasks. In D. L. Molfese and V. J. Molfese, eds., Developmental Variations in Learning: Applications to Social, Executive Function, Language, and Reading Skills, 113–137. Mahwah, NJ: LEA. Gourovitch, M. L., S. Craft, S. B. Dowton, P. Ambrose, and S. Sparta, 1994. Interhemispheric transfer in children with early-treated phenylketonuria. J. Clin. Exp. Neuropsychol. 16: 393–404. Griffiths, P., P. Robinson, R. Davies, K. Hayward, K. Lewis, K. Livingstone, et al., 2005. Speed of decision-making and set-switching: Subtle executive deficits in children with treated phenylketonuria. Educ. Child Psychol. 22(2):81– 89. Guldberg, P., F. Rey, J. Zschocke, V. Romano, B. Francois, L. Michiels, et al., 1998. A European multicenter study of phenylalanine hydroxylase deficiency: Classification of 105 mutations and a general system for genotype-based prediction of metabolic phenotype. Am. J. Hum. Genet. 63(1):71– 79. Guttler, F., and H. Lou, 1986. Dietary problems of phenylketonuria: Effect on CNS transmitters and their possible role in behaviour and neuropsychological function. J. Inherit. Metab. Dis. 9(Suppl. 2):169–177. Hasselbalch, S., G. M. Knudsen, P. B. Toft, P. Hogh, E. Tedeschi, S. Holm, et al., 1996. Cerebral glucose metabolism is decreased in white matter changes in patients with phenylketonuria. Pediatr. Res. 40(1):21–24. Hommes, F. A., and L. Moss, 1992. Myelin turnover in hyperphenylalaninaemia: A re-evaluation with the HPH-5 mouse. J. Inherit. Metab. Dis. 15(2):243–251. Huijbregts, S., L. de Sonneville, R. Licht, J. Sergeant, and F. van Spronsen, 2002. Inhibition of prepotent responding

welsh, deroche, and gilliam: early-treated phenylketonuria

687

and attentional flexibility in treated phenylketonuria. Dev. Neuropsychol. 22(2):481–499. Huijbregts, S. C. J., L. M. J. De Sonneville, F. J. Van Spronsen, I. E. Berends, R. Licht, P. H. Verkerk, et al., 2003. Motor function under lower and higher controlled processing demands in early and continuously treated phenylketonuria. Neuropsychology 17(3):369–379. Joseph, B., and C. A. Dyer, 2003. Relationship between myelin production and dopamine synthesis in the PKU mouse brain. J. Neurochem. 86(3):615–626. Kayaalp, E., E. Treacy, P. J. Waters, S. Byck, P. Nowacki, and C. R. Scriver, 1997. Human phenylalanine hydroxylase mutations and hyperphenylalaninemia phenotypes: A metanalysis of genotype-phenotype correlations. Am. J. Hum. Genet. 61(6): 1309–1317. Knox, W. E., 1972. Phenylketonuria. In J. B. Stanbury, J. B. Wyngaarden, and D. S. Fredrickson, eds., The Metabolic Basis of Inherited Disease, 266–295. New York: McGraw-Hill. Koch, R., F. Guttler, and N. Blau, 2002. Mental illness in mild PKU responds to biopterin. Mol. Gen. Metab. 75:284–286. Krause, W., C. Epstein, A. Averbook, P. Dembure, and L. Elsas, 1986. Phenylalanine alters the mean power frequency of electroencephalograms and plasma L-dopa in treated patients with phenylketonuria. Pediatr. Res. 20(11):1112–1116. Krause, W., M. Halminski, L. McDonald, P. Dembure, R. Salvo, D. Freides, et al., 1985. Biochemical and neuropsychological effects of elevated plasma phenylalanine in patients with treated phenylketonuria: A model for the study of phenylalanine and brain function in man. J. Clin. Invest. 75(1):40– 48. Kure, S., D.-C. Hou, T. Ohura, H. Iwamoto, S. Suzuki, N. Sugiyama, et al., 1999. Tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency. J. Pediatr. 135:375– 378. Lee, N.-C., L.-Y. Cheng, T.-T. Liu, K.-J. Hsiao, P.-C. Chiu, and D.-M. Niu, 2006. Long-term follow-up of Chinese patients who received delayed treatment for 6-pyruvoyl-tetrahydropterin synthase deficiency. Mol. Gen. Metab. 87:128–134. Leuzzi, V., C. Carducci, C. Carducci, F. Chiarotti, C. Artiola, T. Giovanniello, et al., 2006. The spectrum of phenylalanine variations under tetrahydrobiopterin load in subjects affected by phenylalanine hydroxylase deficiency. J. Inherit. Metab. Dis. 29(1):38–46. Lindner, M., D. Haas, J. Zschocke, and P. Burgard, 2001. Tetrahydrobiopterin responsiveness in phenylketonuria differs between patients with the same genotype. Mol. Gen. Metab. 73(1):104–106. Lou, H. C., P. B. Toft, J. Andreson, I. Nikkelsen, B. Olsen, and F. Guttler, 1992. An occipito-temporal syndrome in adolescents with optimally controlled hyperphenylalaninemia. J. Inherit. Metab. Dis. 15:687–695. Luciana, M., K. L. Hanson, and C. B. Whitley, 2004. A preliminary report on dopamine system reactivity in PKU: Acute effects of haloperidol on neuropsychological, physiological, and neuroendocrine functions. Psychopharmacology 175(1):18–25. Luciana, M., J. Sullivan, and C. A. Nelson, 2001. Associations between phenylalanine-to-tyrosine ratios and performance on tests of neuropsychological function in adolescents treated early and continuously for phenylketonuria. Child Dev. 72(6): 1637. McKean, C. M., 1972. The effects of high phenylalanine concentrations on serotonin and catecholamine metabolism in the human brain. Brain Res. 47(2):469–476.

688

Pennington, B. F., W. J. van Doorninck, L. L. McCabe, and E. R. McCabe, 1985. Neuropsychological deficits in early treated phenylketonuric children. Am. J. Ment. Defic. 89(5):467– 474. Pietz, J., R. Dunckelmann, A. Rupp, D. Rating, H. M. Meinck, H. Schmidt, et al., 1998. Neurological outcome in adult patients with early-treated phenylketonuria. Eur. J. Pediatr. 157(10):824– 830. Pietz, J., U. K. Meyding-Lamade, and H. Schmidt, 1996. Magnetic resonance imaging of the brain in adolescents with phenylketonuria and in one case of 6-pyruvoyl tetrahydropteridine synthase deficiency. Eur. J. Pediatr. 155(Suppl. 1):S69– 73. Pueschel, S. M., 1996. Central nervous system effects in individuals with phenylketonuria. Dev. Brain Dysfunc. 90:165–179. Scriver, C. R., and C. L. Clow, 1980. Phenylketonuria: Epitome of human biochemical genetics (1st of 2 parts). N. Engl. J. Med. 303:1336–1400. Scriver, C. R., and S. Kaufman, 2001. Hyperphenylalaninemia: Phenylalanine hydroxylase deficiency. In C. R. Scriver, A. L. Beaudet, W. S. Sly, and D. Valle, eds., The Metabolic and Molecular Bases of Inherited Disease, 1667–1724. New York: McGraw-Hill. Shedlovsky, A., J. D. McDonald, D. Symula, and W. F. Dove, 1993. Mouse models of human phenylketonuria. Genetics 134(4):1205–1210. Shonkoff, J. P., and D. A. Phillips, 2000. From Neurons to Neighborhoods: The Science of Early Childhood Development. Washington, DC: National Academy Press. Spaapen, L. J. M., and M. E. Rubio-Gozalbo, 2003. Tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency, state of the art. Mol. Gen. Metab. 78(2):93. Stanbury, J. B., J. B. Wyngaarden, and D. S. Fredrickson, 1983. The Metabolic Basis of Inherited Disease, 2nd ed. New York: McGraw-Hill. Steinfeld, R., A. Kohlschutter, J. Zschocke, M. Lindner, K. Ullrich, and Z. Lukacs, 2002. Tetrahydrobiopterin monotherapy for phenylketonuria patients with common mild mutations. Eur. J. Pediatr. 161(7):403. Stemerdink, N., M. van der Molen, A. Kalverboer, A. J. van der Meere, J. Huisman, L. de Jong, F. Slijper, P. Verkerk, and F. Spronsen, 1999. Prefrontal dysfunction in early and continuously treated phenylketonuria. Dev. Neuropsychol. 16:29–57. Stuss, D. T., and D. F. Benson, 1984. Neuropsychological studies of the frontal lobes. Psychol. Bull. 95(1):3–28. Trefz, F. K., C. Aulela-Scholz, and N. Blau, 2001. Successful treatment of phenylketonuria with tetrahydrobiopterin. Eur. J. Pediatr. 160(5):315. Ullrich, K., H. Moller, J. Weglage, G. Schuierer, U. Bick, A. Ludolph, et al., 1994. White matter abnormalities in phenylketonuria: Results of magnetic resonance measurements. Acta Paediatr.: 1992 (Supplement)407:78–82. Waisbren, S. E., M. J. Brown, L. M. de Sonneville, and H. L. Levy, 1994. Review of neuropsychological functioning in treated phenylketonuria: An information processing approach. Acta Paediatr.: 1992 (Supplement)407:98–103. Waisbren, S. E., R. R. Schnell, and H. L. Levy, 1980. Diet termination in children with phenylketonuria: A review of psychological assessments used to determine outcome. J. Inherit. Metab. Dis. 3(4):149–153. Welsh, M. C., 1996. A prefrontal dysfunction model of earlytreated phenylketonuria. Eur. J. Pediatr. 155(Suppl. 1):S87–89. Welsh, M. C., 2002. Developmental and clinical variations in executive functions. In D. L. Molfese and V. J. Molfese, eds.,

neurodevelopmental aspects of clinical disorders

Developmental Variations in Learning: Applications to Social, Executive Function, Language, And Reading Skills, 139–185. Mahwah, NJ: Lawrence Erlbaum. Welsh, M. C., S. L. Friedman, and S. J. Spieker, 2006. Executive functions in developing children: Current conceptualizations and questions for the future. In K. McCartney and D. Phillips, eds., Blackwell Handbook of Early Childhood Development, 167–187. Malden, MA: Blackwell Publishing. Welsh, M. C., G. Kelley, and Z. V. Tran, 2000. Treatment regimens and their effectiveness: A meta-analytic review. Technical report prepared for the NIH Consensus Development Conference on Early-Treated Phenylketonuria, October 16–18, 2000, Washington, DC. Welsh, M. C., and B. F. Pennington, 2000. Phenylketonuria. In K. O. Yeates, M. D. Ris, and H. G. Taylor, eds., Pediatric Neuropsychology: Research, Theory and Practice, 275–299. New York: Guilford. Welsh, M. C., B. F. Pennington, S. Ozonoff, B. Rouse, and E. R. McCabe, 1990. Neuropsychology of early-treated phenylketonuria: Specific executive function deficits. Child Dev. 61(6): 1697–1713.

White, D. A., M. J. Nortz, T. Mandernach, K. Huntington, and R. D. Steiner, 2001. Deficits in memory strategy use related to prefrontal dysfunction during early development: Evidence from children with phenylketonuria. Neuropsychology 15(2):221–229. White, D. A., M. J. Nortz, T. Mandernach, K. Huntington, and R. D. Steiner, 2002. Age-related working memory impairments in children with prefrontal dysfunction associated with phenylketonuria. J. Int. Neuropsychol. Soc. 8:1–11. Williamson, M., J. C. Dobson, and R. Koch, 1977. Collaborative study of children treated for phenylketonuria: Study design. Pediatrics 60(6):815. Williamson, M. L., R. Koch, C. Azen, and C. Chang, 1981. Correlates of intelligence test results in treated phenylketonuric children. Pediatrics 68(2):161–167. Zelazo, P. D., and D. Frye, 1998. Cognitive complexity and control. II. The development of executive function in childhood. Curr. Dir. Psychol. Sci. 7(4):121–126. Zelazo, P. D., U. Muller, D. Frye, and S. Marcovitch, 2003. The development of executive function. Monogr. Soc. Res. Child Dev. 68:1–27.

welsh, deroche, and gilliam: early-treated phenylketonuria

689

42

Research into Williams Syndrome: The State of the Art ANNETTE KARMILOFF-SMITH

For instance, children with Williams syndrome have a barely measurable general intelligence and require constant parental care, yet they have an exquisite mastery of syntax and vocabulary. They are, however, unable to understand even the most immediate implications of their admirably constructed sentences. (PiattelliPalmarini, 2001, 887) In sum, brain volume, brain anatomy, brain chemistry, hemispheric asymmetry, and the temporal patterns of brain activity are all atypical in people with Williams syndrome. How could the resulting system be described in terms of a normal brain with parts intact and parts impaired, as the popular view holds? Rather, the brains of infants with WS develop differently from the outset, which has subtle, widespread repercussions at the cognitive level. (KarmiloffSmith, 1998, 393)

The striking difference between these two quotations not only encapsulates early research into the neurodevelopmental disorder Williams syndrome (WS), but also continues to illustrate the theoretical differences guiding current research into this fascinating syndrome. And there is no doubt that debates will continue to rage over the extent to which WS is a window on the nature/nurture debate. Williams syndrome was first described by two cardiology groups (Williams, Barrett-Boyes, and Lowe, 1961; Beuren, Apitz, and Harmjanz, 1962), both identifying the association of several clinical features in affected individuals: narrowing of the aorta (supravalvular aortic stenosis: SVAS), distinctive facial dysmorphology, slow physical growth, and learning difficulties. It took another couple of decades before the syndrome started to be extensively investigated by cognitive psychologists and neuroscientists, hoping thereby to gather data that might contribute directly to the question of whether or not the human mind/brain starts out with innately specified, cognitive-level modules (like grammar, face processing, spatial cognition, number, and the like) that operate independently of one another and of general intelligence. The widely held view was and still is to some extent that WS involves a juxtaposition of intact and impaired modules— that is, fluent language, excellent face-processing abilities, and a very friendly social disposition, alongside very poor spatial, problem-solving, and numerical abilities. But, as we shall see later, the real picture of the WS profile is far more complex.

At the level of the brain, WS is characterized by structural abnormalities, with total brain volume being about 80 percent of normal brains (Neville, Mills, and Bellugi, 1994). Once overall brain volume is controlled for, structural imaging studies point to specific reductions in parietal cortex (Eckert et al., 2005) as well as in the brain stem (Reiss et al., 2000). The corpus callosum is particularly small (Schmitt et al., 2002), but primary auditory cortex is proportionally large, with atypical cytoarchitecture (Holinger et al., 2005). While the cerebellum is also large in WS brains ( Jones et al., 2002), this is an area in which we identified abnormal brain biochemistry (Rae et al., 1998). Atypical neuronal size, orientation, and density have also been identified (Galaburda et al., 2002), as well as increased cortical gyrification in the right parietal and occipital regions, as well as the left frontal areas (Schmitt et al., 2002). Gray matter volume is close to that found in normal controls, but white matter volume is reduced (Reiss et al., 2000). Finally, the hippocampus in WS brains appears to have an atypical shape (Meyer-Lindenberg et al., 2005). Apart from these structural abnormalities, other studies using functional magnetic resonance imaging (MeyerLindenberg et al., 2005) and high-density event-related potentials (Grice et al., 2001, 2003; Mills et al., 2000) point to atypical functioning in the WS brain. It is worth noting, however, that most of the studies of the WS brain structure and function have been based on adult brains. We have as yet little understanding of the developmental mechanisms that yield these brain abnormalities. Paradoxically, and despite agreement on the abnormalities of the WS brain, some researchers continue to portray the WS brain in terms of a juxtaposition of intact and impaired cognitive modules, while others refute the possibility of intact modules in such a generally atypically developing brain. So, let’s first describe the agreed-upon facts about the genotype and noncognitive features of WS and then examine the more controversial cognitive-level phenotype.

The Williams syndrome genotype It is now known that in the vast majority of individuals with WS at least 28 genes are deleted on the long arm of one copy of chromosome 7q11.23 (Donnai and KarmiloffSmith, 2000; Tassabehji et al., 2005). In most cases the

691

deletion is sporadic, although a few cases of parent-to-child transmission and concordant monozygotic twins have been reported (Morris, Thomas, and Greenberg, 1993; Pankau et al., 1993). No indication of parental age effects has emerged, and deletions occur with equal frequency on both maternally and paternally inherited chromosomes. For a long time, the prevalence of WS was estimated to be between 1 in 20,000 and 1 in 50,000 (Morris et al., 1998), but more recently a Norwegian study has situated the prevalence much higher, at closer to 1 in 7,500 (Stromme, Bjornstad, and Ramstad, 2002). The deletion occurs at the outset of pregnancy during meiosis and is due to the existence of identical repeats on the regions flanking the deletion, thereby allowing for the chance misalignment of segments (Peoples et al., 2000; Urban et al., 1996). The genetic basis of the syndrome started to be identified in the early 1990s when it was found that the elastin gene (ELN) at 7q11.23 was disrupted by a translocation associated with SVAS (Curran et al., 1993), leading to the hypothesis that hemizygosity at the elastin locus might help situate the extent of the full deletion in individuals with WS (Ewart et al., 1993), who suffer from SVAS (Hallidie-Smith and Karas, 1988). The discovery of the ELN deletion in WS supported the view that haploinsufficiency for ELN causes the vascular abnormalities of WS such as SVAS (Ewart et al., 1993) and perhaps prematurely aging skin, but subsequent research has challenged the role of ELN in the WS facial dysmorphology (Hammond et al., 2005; Tassabehji et al., 2005) and in other physical or cognitive anomalies of the syndrome (Li et al., 1997; Tassabehji et al., 1999, 2005). Since then, a great deal of progress has been made in identifying several of the other genes responsible for the fullblown WS phenotype (for reviews see Franke, 1999; Donnai and Karmiloff-Smith, 2000; Meyer-Lindenberg et al., 2005). The microdeletion is approximately 1.5 Mb, with ELN being midway between the two break points (Perez-Jurado et al., 1996). Adjacent to ELN, the limkinase-1 gene (LIMK1), which is expressed in the brain, was found to be deleted in all patients with full-blown WS (Tassabehji et al., 1996). Notable is the fact that defects in the expression of LIMK1 may well affect axonal guidance during central nervous system development (Arber et al., 1998). Moreover, changes in the expression levels of the genes neighboring the deletion in WS may also play a role (Merla et al., 2006). Other genes are currently under investigation, often helped by the identification of patients with partial deletions in the WS critical region (Frangiskakis et al., 1996; Gray et al., 2006; Karmiloff-Smith, Grant, et al., 2003; Tassabehji et al., 1996, 2005). But, because of the complexities of both gene expression and human ontogeny, no genotype-phenotype relations, apart from the role of ELN in SVAS, have yet been unequivocally identified. Nonetheless, studies of

692

partial deletion (PD) patients do point to testable hypotheses about possible genotype/phenotype relations. For example, Frangiskakis and colleagues identified a family, some of whose members had ELN and LIMK1 deleted and others who did not. It turned out that PD family members displayed spatial deficits similar to those encountered in individuals with WS, whereas those without the PD showed no spatial deficits (Frangiskakis et al., 1996). Based also on findings from a LIMK1 knockout mouse model, which resulted in spatial deficits in the Morris maze (Meng et al., 2002), Frangiskakis and colleagues concluded that the LIMK1 gene deletion in individuals with WS caused their spatial impairment. However, subsequent studies of PD patients, using first the same tasks as in the Frangiskakis study (Karmiloff-Smith, Grant, et al., 2003; Tassabehji et al., 1999) and then a very wide range of 16 different tasks to detect subtle spatial impairments (Gray et al., 2006), demonstrated that the deletion of LIMK1 could not alone explain the spatial deficits, since our ELN/LIMK1 PD patients were completely unimpaired across all these studies. How does this finding sit with the fact that the mouse LIMK1 knockout does show spatial deficits? It is important to note that the LIMK1 mouse knockout model was tested in the Morris maze, in which the mouse must represent the changing positions of its body in space, whereas the human tests had targeted tabletop constructions during which participants remain seated and stationary. It thus remained crucial to test PD and WS patients in navigational tasks in which they also had to mentally represent their changing positions in space. Karmiloff-Smith and collaborators (submitted) therefore devised child and adult human versions of the Morris maze measuring navigational abilities. Our ELN-LIMK1 PD patients continued to show no spatial deficits, whereas the WS controls were very impaired. However, a patient with a much larger PD (24 of the 28 WS genes) turned out to be just as impaired as the WS controls on the navigational task, highlighting the importance of two genes (CYLYN2 and GTF2IRD1) at the telomeric end of the WS critical region as the most likely contributors to spatial cognition. Interestingly, this patient had a much milder social phenotype compared to the typical overly friendly disposition of those with full-blown WS, suggesting that one or more of her four nondeleted genes contribute, probably with many other genes, to the development of normal social cognition (Karmiloff-Smith et al., submitted). It is obvious from the previous discussion that while PD patients help narrow hypotheses about the contribution of particular genes to the resulting WS phenotype, these are rare cases and the story always turns out to be extremely complex at both the genotypic and phenotypic levels. Given the pleiotropic nature of most genes, it is highly unlikely that scientists will end up with neat one-to-one mappings between mutated genes and phenotypic outcomes at the cognitive or social level.

neurodevelopmental aspects of clinical disorders

The Williams syndrome phenotype First let us note a general point. To understand any developmental syndrome, it is essential to distinguish between the behavioral phenotype (based on scores from standardized tests of overt behavior) and the cognitive phenotype (based on in-depth analyses of the mental processes underlying the overt behavior) (Karmiloff-Smith, 1998). For instance, it can happen that equivalent behavioral scores camouflage very different cognitive processes, as we shall see when we examine in detail some aspects of the WS phenotypic profile. Apart from the physical abnormalities in WS (facial dysmorphology, SVAS, small stature, and hyperacusis), the personality profile of people with WS is very specific to the syndrome. Even in early infancy, children with WS fixate the faces of others and smile very frequently. As they reach middle childhood and onward, individuals with WS tend to be overly friendly with strangers and to lack social judgment skills (Einfeld, Tonge, and Florio, 1997; Gosch and Pankau, 1997). While tending not to be shy in new surroundings, they also display extreme anxiety where unexpected things happen. This clinical population also shows empathy toward others’ emotions, although they have difficulty in interpreting some more complex facial expressions, such as anger and fear, compared to controls (Tsirempolou et al., 2006). They are also less skilled at understanding human intentionality than originally thought. In other words, their socialaffective understanding seems relatively proficient, but their social-cognitive understanding of other minds is clearly impaired (Tager-Flusberg, Boshart, and Baron-Cohen, 1998). Most studies suggest that individuals with WS exhibit a relatively uneven cognitive-linguistic profile (although see meta-analysis in Brock, 2007, arguing for a more even profile in this syndrome), together with mild to severe mental retardation. Their WS full intelligence quotient is estimated at 51–70, with a mean of 56 (Mervis et al., 1999; Udwin and Yule, 1991). To be noted, however, is the fact that the full IQ score camouflages marked differences in specific cognitive abilities. The pioneering work of Bellugi and her collaborators suggested some clear-cut dissociations in the cognitive architecture of WS. Language and face processing appeared to be preserved in the face of both general retardation and particularly serious problems with visuospatial cognition (Bellugi, Wang, and Jernigan, 1994). However, the notion that abilities in developmental disorders are “preserved” or “intact” takes overt behavior in the adult as if it were a direct index of underlying cognitive processes, which it clearly is not (Karmiloff-Smith, 1998). In-depth analyses of the language and face processing in adults with WS—two areas purported to be “intact”—strongly suggest that their behavioral proficiencies are supported by different cognitive processes compared with normal controls. Moreover,

analyses of the WS infant cognitive profile demonstrate that the latter differs from the adult phenotypic outcome. In other words, it is critical to examine full developmental trajectories from infancy to adulthood at the level of cognitive processes rather than merely recording overt behavior (Karmiloff-Smith, 1998; Annaz, Karmiloff-Smith, and Thomas, 2008). Let us start historically with what is known about the adult cognitive phenotype in WS, since studies of infants with the syndrome are only recently being undertaken. All researchers agree that WS presents with serious deficits in spatial cognition and number (Ansari et al., 2003; Ansari and KarmiloffSmith, 2002; Bellugi, Wang, and Jernigan, 1994; Mervis et al., 1999). It is therefore of interest to focus mainly on face processing and language, the two areas on which much of the debate continues to hinge.

The phenotypic outcome in adults: Face processing There is no doubt that people with WS are very proficient at recognizing faces. They score in the normal range on at least two standardized face-processing tasks (Bellugi, Wang, and Jernigan, 1994; Udwin and Yule, 1991). As we shall see with respect to the case of WS language, initial hypotheses about face processing in WS suggested an intact, innately specified face-processing module (Bellugi, Wang, and Jernigan, 1994; Rossen et al., 1996). However, as mentioned earlier, it is vital to distinguish between the behavioral phenotype and the cognitive phenotype (Karmiloff-Smith, 1998). Several studies have replicated Bellugi’s earlier work revealing normal or near normal WS scores on standardized face-processing tasks (Deruelle et al., 1999; KarmiloffSmith, 1998; Karmiloff-Smith et al., 2004). However, this work has also seriously challenged the notion that the behavioral success displayed in WS face processing is normal. Rather, it has been shown that, whereas normal controls tend to use configural processes to recognize faces, people with WS use predominantly featural or holistic processes (Deruelle, et al., 1999; Karmiloff-Smith, 1998; Karmiloff-Smith et al., 2004; Nakamura et al., 2006). Moreover, the tendency to use greater featural than configural processing in WS, as compared to, say, patients with Down syndrome and normal controls, is seen not only with respect to faces, but also in other visuospatial tasks such as drawing and construction (Bellugi, Wang, and Jernigan, 1994; Deruelle et al., 1999; Mervis et al., 1999). Imaging studies focusing on the electrophysiology of face processing in WS also support the notion that different cognitive processes sustain the WS behavioral proficiency. Thus, using high-density event-related potentials (ERPs), we discovered abnormalities in the early waveforms of WS patients (Grice et al., 2003) not found in any of our healthy controls.

karmiloff-smith: research into williams syndrome

693

For example, whereas normal adult brains show differences in both amplitude and latency when processing faces compared to cars, the individuals with WS displayed the same brain signatures for both faces and cars. Moreover, in contrast to normal controls, there was less right-hemisphere lateralization in WS brains when processing faces. Other ERP laboratories have found similar differences in WS adults compared to controls (Mills et al., 2000), particularly with respect to the normally different brain signatures for upright and inverted faces, which are less differentiated in participants with WS. These various data refute the idea of an “intact” face-processing module and instead suggest that people with WS may use a general “object processor” to process all visual stimuli. Furthermore, a comparison of 40hertz gamma-band activity in WS, autism, and healthy controls yielded interesting cross-syndrome differences (Grice et al., 2001). The adolescents and adults with autism displayed bursts of activity similar to those of the controls, except that they failed to differentiate upright and inverted faces in the group with autism. By contrast, the adolescents and adults with WS displayed almost no gamma bursts, with their brain patterns resembling those of normal 2-month-old infants (Grice et al., 2001; see also Farran, 2005, regarding deficits in perceptual integration in WS). This finding yet again points to a lack of a normal developmental trajectory over time in the WS brain.

The phenotypic outcome in adults: Language Despite claims to the contrary (Bellugi, Wang, and Jernigan, 1994; Piatelli-Palmarini, 2001; Pinker, 1994; Rossen et al., 1996; Smith and Tsimpli, 1995), it is questionable whether any aspect of language—syntax, semantics, phonology, or pragmatics—is intact in WS. Indeed, an abundance of empirical studies from numerous laboratories across the world now challenge intactness claims with respect to all aspects of WS language (e.g., the lexicon: Jarrold et al., 2000; Temple, Almazan, and Sherwood, 2002; morphosyntax: Grant, Valian, and Karmiloff-Smith, 2002; KarmiloffSmith et al., 1997; Thomas et al., 2001; Volterra, Capirci, and Caselli, 2001; phonology: Grant et al., 1997; Laing et al., 2001; pragmatics: Laws and Bishop, 2004). In a report by Clahsen and Almazan (1998), they tried to retain the innateness hypothesis but for an aspect of morphosyntax rather than a whole cognitive module. These researchers argued for a dissociation of innate mechanisms, on the basis of their claim that in WS memory for vocabulary is impaired but grammar is intact. However, their arguments were based on a very small sample of children with WS (N = 2 for MA = 5 years, and N = 2 for MA = 7 years), with considerable individual variation between the few participants. By contrast, a much broader, in-depth study using the same tasks (Thomas et al., 2001) compared the performance of

694

21 patients with WS with that of four typically developing control groups at ages 6, 8, 10, and adult. Thomas and colleagues argued that it is not sufficient to demonstrate that one aspect of language is relatively poorer than another, because this also obtains at younger stages in normal development. One cannot take a relative comparison and make an absolute claim. Rather, it is necessary to demonstrate that the level of a specific aspect of language is poorer than would be expected in WS given the subjects’ overall level of language development. The Thomas and colleagues study showed that when verbal mental age was controlled for, the WS group not only was generally impaired, but also displayed no selective deficit across vocabulary and morphosyntax. Results of a subsequent study of rapid naming in WS (Thomas et al., in press) also pointed to atypicality in WS: slower and less accurate naming in the clinical group compared with both chronologically age-matched and receptive-vocabulary age-matched controls. In fact, Mervis and her collaborators (e.g., Klein and Mervis, 1999) have concluded that the best way to characterize WS language is that it is delayed, revealing patterns typical of considerably younger children. Indeed, a meta-analysis of a large number of language studies reveals that WS language is neither intact nor at the level expected for chronological age; rather, it is often no better than would be expected for nonverbal mental age (Brock, in press). A number of other empirical findings suggest that the WS language system not only is delayed, but actually develops along a different trajectory compared to healthy controls (Karmiloff-Smith et al., 1997; Klein and Mervis, 1999; Mervis et al., 1999; Stevens and Karmiloff-Smith, 1997; Thomas and Karmiloff-Smith, 2003; Thomas et al., 2001, in press). For example, unlike healthy controls, young children with WS use pointing after the appearance of naming. Exhaustive sorting follows the vocabulary spurt in WS, rather than preceding it as in typical development (Mervis et al., 1999). Sensitivity to the sound patterns of the language may act as a greater constraint in WS language than sensitivity to meaning (Grant et al., 1997; Klein and Mervis, 1999; Laing et al., 2001). Finally, several studies across different languages (e.g., Karmiloff-Smith et al., 1997; Klein and Mervis, 1999; Levy and Bechar, 2003; Lucas, Pleh, and Racsmany, 2004; Vicari et al., 1996; Volterra et al., 1996, 2003) now suggest that the subtle problems that people with WS have with semantics, morphosyntax, and pragmatics are often camouflaged by their good verbal memory, challenging the popular belief that their language is intact. Taken together, these and many other studies point to the fact that when learning language as children, as well as when processing language as adults, individuals with WS follow a deviant developmental trajectory. Behaviorally, WS language may appear to be relatively proficient, but cognitively,

neurodevelopmental aspects of clinical disorders

it seems increasingly likely to involve different cognitive processes from the language of normally developing controls.

The infant start state versus the phenotypic outcome Let us now turn briefly to the infant start state and how it relates to the phenotypic end state in WS. Early assumptions, made on the basis of the adult behavioral phenotype, held that the pattern of abilities and deficits found in the end state would also characterize the infant start state, leading to claims about the innate specification of certain WS abilities. We challenged these assumptions and addressed directly the relationship between the end state and the start state in, for example, a study of two cognitive domains, one of relative proficiency in the phenotypic end state— language—and one of serious impairment—number. Infants, toddlers, children, and adults with WS were compared with infants, toddlers, children, and adults from another syndrome, Down syndrome (DS) (Paterson et al., 1999, 2006), matched for mental age and chronological age. The findings were very clear. For adults, the WS and DS groups had significantly different scores on a vocabulary test, with the WS adults outstripping the DS adults. For number, the pattern of performance of the DS adults, although delayed, resembled that of the normal controls. By contrast, the WS adults performed far more poorly on number tasks than the DS adults and showed signs of a deviant trajectory (Paterson et al., 2006; see also Ansari et al., 2003). So the phenotype in the adult end state was as follows: DS significantly worse than WS on vocabulary, WS significantly worse than DS on number. If the start state could be directly derived from the end state and used to make claims about innateness and genotype/phenotype relations, then atypical infants should show the same profile of cognitive abilities and impairments as the atypical adults. But this was not the case. The infancy studies revealed a very different pattern from the adult studies when comparing WS to DS. For the vocabulary task, the WS and DS infants were equally impaired (at approximately half their chronological age), despite the fact that WS adults are significantly better than DS adults. By contrast, for number, although the WS adults are more impaired than DS adults, in infancy the group with WS was unimpaired on a numerosity judgment task. The WS infants performed like the chronological-age controls, whereas the DS infants were seriously impaired and did not even reach the level of the mental-age controls. Again, the pattern in infancy differed considerably from that observed in adulthood across the two syndromes. The infancy/adult data suggest that the learning trajectories of the two syndromes, WS and DS, differ across developmental time. This finding highlights the need to consider the process of development itself when studying

developmental disorders, by building full developmental trajectories. It also underlines the fact that claims about innateness and genotype/phenotype relations cannot be based on patterns found in phenotypic outcomes in adults (Annaz, Karmiloff-Smith, and Thomas, 2008; Karmiloff-Smith, 1998; Paterson et al., 1999).

Building full developmental trajectories within domains and across domains The healthy infant cortex starts out as highly interconnected, and it is only with time that areas of the brain become increasingly specialized (progressively restricting the inputs that a particular circuit processes) and increasingly localized (Huttenlocher and de Courten, 1987; Huttenlocher and Dabholkar, 1997; Johnson, 2001; Neville, 2006). Timing plays a crucial role in human ontogeny, and it is possible that in individuals with WS this progressive modularization of function does not occur (Karmiloff-Smith, 1992). We have seen in an earlier section how face processing in WS, despite its superficial proficiency, involves brain processes that do not differentiate between, say, cars and faces, or between upright and inverted faces, and that the typical right-hemispheric specialization for faces does not obtain in WS. This finding indicates that, despite the behavioral proficiency on some standardized tasks, the gradual process of modularization may fail to occur in this syndrome. This observation suggests that the atypical brain may remain more interconnected over time with less progressive modularization than in the normal case. Preliminary analyses of our recent data comparing the symbolic-distance effect and the semantic-distance effect in WS and controls yielded interesting results in this respect (Scerif et al., submittted). The symbolic distance effect measures participants’ reaction time when they compare two numbers and click on the larger of the two: the closer the two numbers, the slower the reaction time. The semantic distance effect measures participants’ reaction time when they compare two words/ images and judge whether they are the same or different: the closer the words/images are semantically, the slower the reaction time. Our initial analyses point to correlations across the numerical and lexicosemantic domains in the adults with WS, whereas healthy controls showed no such correlations. This result suggests that in the normal case domains like number and vocabulary become progressively modularized (i.e., localized and specialized), whereas in our adults with WS more commonality of processing remains across domains. Of course, it will next be essential to examine these questions in younger children (for whom such tasks are suitable) in order to build full normal and atypical developmental trajectories, and to pinpoint when and how in normal development these two domains become increasingly separated and specialized.

karmiloff-smith: research into williams syndrome

695

Multiple developmental factors contributing to lexical development in Williams syndrome An attempt to understand the early interplay across domains is often missing from developmental studies, and yet this issue is crucial. Our current research program aims to address this issue further in the case of early lexical development in WS infants and toddlers. How does our developmental approach sit with earlier claims about an intact language module in WS (Bellugi, Wang, and Jernigan, 1994; Pinker, 1994, 1999; Smith and Tsimpli, 1995)? Interestingly, language onset in WS children is very late. Indeed, despite the superficially fluent language peppered with eruditesounding words of adolescents and adults with WS, language production in this clinical population often does not occur until the 5th or 6th year (Singer-Harris et al., 1997). Why is language so delayed in this population? Is the delay simply due to a late-maturing language module? Or is there a developmental explanation? In my view, the roots of the delay reside in deficits in multiple interacting earlier processes. For example, infants and toddlers with WS are extremely delayed in hand movements and babbling (Masataka, 2001). They are also seriously delayed in segmenting the speech stream (Nazzi, Paterson, and Karmiloff-Smith, 2003), a capacity seen as early as 8 months in typically developing infants but still impaired at 20 months in WS toddlers. Second, unlike typical controls, toddlers and young children with WS rely more on perceptual cues than on linguistic labels when identifying new objects (Nazzi, Gopnik, and KarmiloffSmith, 2005). Furthermore, early categorization abilities in WS are impaired (Nazzi and Karmiloff-Smith, 2002), and exhaustive sorting follows word onset rather than preceding it as in the normal case (Mervis and Bertrand, 1997). Pointing is also atypical in WS toddlers. Whereas in typical development, referential pointing precedes the onset of language, in WS this order is unusually reversed (Mervis and Bertrand, 1997). Moreover, our recent studies revealed that WS toddlers do not use or follow eye gaze for referential communication and do not properly understand the referential function of pointing (Laing et al., 2002). They are good at dyadic attention but poor at triadic attention—shared attention to an external object or event. Finally, in normal language acquisition, young children’s comprehension outstrips their levels of production. This clear-cut asymmetry does not hold for WS (Paterson et al., 1999). In sum, many different aspects of communication show an early, unusual pattern in WS, jointly contributing in complex ways to the explanation of the late onset of their language. However, an even earlier deficit outside the domain of language may offer a compelling explanation of some of these early deficits: atypical eye movement planning. In a

696

study of saccadic planning in infants and toddlers with WS and DS compared to mental-age and chronological-age controls, we found that although children with DS resembled controls, apart from being somewhat slower, the infants and toddlers with WS displayed a range of impairments in planning saccadic eye movements (Brown et al., 2003). Some stayed fixated on one stimulus without moving their eyes to the second stimulus at all, whereas others made one saccade to a new stimulus, but failed to make the second saccade that all the controls and young children with DS made. For the infants with WS who did make a double saccadic movement, two errors appeared: either they failed to update their retinal image after the first eye movement and ended in the wrong location after their second saccade (the retinocentric error); or they summated the two saccades before moving their eyes, thus making the vector summation error, typical of normal 2-month-olds. In other words, planning visual attention and making saccadic eye movements to explore the environment, as well as to follow another’s eye gaze and pointing gestures, turns out to be atypical in infants and toddlers with WS. Recall that in normal interaction, many vocabulary items are learned through triadic interaction with caregivers, which involve saccadic eye movements to follow the partner’s focus of attention. Thus early visuospatial deficits in the WS developmental trajectory outside the domain of language can have cascading developmental effects over time on several emerging higher-level linguistic and cognitive domains. Moreover, infants and toddlers with WS show specific impairments in visual search tasks when looking for targets among distractors. Their errors differ from the errors of young children with fragile X syndrome, for instance (Scerif et al., 2004). Moreover, the fact that domains (like visual cortex and auditory cortex) are highly interrelated in early cortical development (Huttenlocher and de Courten, 1987; Huttenlocher and Dabholkar, 1997; Neville, 2006) turns out to play a critical role in the formation of more general, albeit sometimes subtle, deficits in later development.

Concluding thoughts It has become increasingly clear that the uneven profile of abilities and impairments at the behavioral level in WS must be reanalyzed at the cognitive level if we are to begin to adequately relate genotype to phenotype and gain a deeper understanding of the WS cognitive phenotype. Debates are likely to continue to rage. Notable is the fact that the choice of control group can influence the way in which data from individuals with WS are interpreted. Often, when reporting levels of WS performance consistent with mental-age controls, researchers then tend to conclude that the ability is “intact,” despite being several years behind the typical child. We must not dismiss delay as irrelevant or count a “relative”

neurodevelopmental aspects of clinical disorders

advantage of one system over another as an “absolute” one, leading to claims of intactness (see discussion in KarmiloffSmith, 1998; Karmiloff-Smith, Scerif, and Ansari, 2003). Moreover, infancy studies have highlighted the fact that we cannot use the phenotypic outcome in adults to simply assume the pattern of abilities and impairments in the start state. In other words, we should not directly relate the effects of deleted genes to cognitive-level outcomes in adults. In fact, genetic mutations are more likely to affect low-level cognitive processes that will have differing, cascading effects on different domains as development proceeds over time. Indeed, timing plays a critical role in normal development, and its effects on atypical development must be center stage when we endeavor to build a comprehensive theory of Williams syndrome in particular, and of developmental disorders in general. acknowledgments

The studies by A.K.-S. and colleagues were funded by Program and Projects Grants from the UK Medical Research Council, Fogarty/National Institute of Health, USA (Grant No. R21TW06761-01), and by PhD studentships from the Medical Research Council, the Down Syndrome Association, and the Williams Syndrome Foundation, UK. I particularly wish to acknowledge the full support of the WSF and of the families who took part in our studies.

REFERENCES Annaz, D., A. Karmiloff-Smith, and M. S. C. Thomas, 2008. The importance of tracing developmental trajectories for clinical child neuropsychology. In J. Reed and J. Warner Rogers, eds., Child Neuropsychology: Concepts, Theory and Practice. Oxford: Blackwell. Ansari, D., C. Donlan, M. Thomas, S. Ewing, T. Peen, and A. Karmiloff-Smith, 2003. What makes counting count? Verbal and visuo-spatial contributions to typical and atypical number development. J. Exp. Child Psychol. 85(1):50–62. Ansari, D., and A. Karmiloff-Smith, 2002. Atypical trajectories of number development. Trends Cogn. Sci. 6(12):511–516. Arber, S., F. A. Barbayannis, H. Hanser, C. Schneider, C. A. Stanyon, O. Bernard, and P. Caroni, 1998. Regulation of actin dynamics through phosphorylation of cofilin by LIMkinase. Nature 393:805–809. Bellugi, U., P. P. Wang, and T. Jernigan, 1994. Williams syndrome: An unusual neuropsychological profile. In S. H. Broman and J. Grafman, eds., Atypical Cognitive Deficit in Developmental Disorders: Implications for Brain Function, 23–56. Hillsdale, NJ: Lawrence Erlbaum. Beuren, A. J., J. Apitz, and D. Harmjanz, 1962. Supravalvular aortic stenosis in association with mental retardation and a certain facial appearance. Circulation 26:1235–1240. Brock, J., 2007. Language abilities in Williams syndrome: A critical review. Dev. Psychopathol. 19:97–127. Brown, J., M. H. Johnson, S. Paterson, R. Gilmore, M. Gsödl, E. Longhi, and A. Karmiloff-Smith, 2003. Spatial representation and attention in toddlers with Williams syndrome and Down syndrome. Neuropsychologia 41:1037–1046.

Clahsen, H., and M. Almazan, 1998. Syntax and morphology in Williams syndrome. Cognition 68:167–198. Curran, M. E., D. L. Atkinson, A. K. Ewart, C. A. Morris, M. F. Leppert, and M. T. Keating, 1993. The elastin gene is disrupted by a translocation associated with supravalvular aortic stenosis. Cell 73:159–168. Deruelle, C., J. Mancini, M. O. Livet, C. Casse-Perrot, and S. de Schonen, 1999. Configural and local processing of faces in children with Williams syndrome. Brain Cogn. 41:276–298. Donnai, D., and A. Karmiloff-Smith, 2000. Williams syndrome: From genotype through to the cognitive phenotype. Am. J. Med. Genet. 97(2):164–171. Eckert, M. A., D. Hu, S. Eliez, U. Bellugi, A. Galaburda, J. Korenberg, D. Mills, and A. L. Reiss, 2005. Evidence for superior parietal impairment in Williams syndrome. Neurology 64:152–153. Einfeld, S. L., B. J. Tonge, and T. Florio, 1997. Behavioral and emotional disturbance in individuals with Williams syndrome. Am. J. Ment. Retard. 102:45–53. Ewart, A. K., C. A. Morris, D. Atkinson, J. Weishan, K. Sternes, P. Spallone, D. A. Stock, M. Leppert, and M. T. Keating, 1993. Hemizygosity at the elastin locus in a developmental disorder, Williams syndrome. Nature Genet. 5:11–16. Farran, E., 2005. Perceptual grouping ability in Williams syndrome: Evidence for deviant patterns of performance. Neuropsychologia 43:815–822. Frangiskakis, J. M., A. Ewart, C. A. Morris, C. B. Mervis, J. Bertrand, B. F. Robinson, B. P. Klein, G. Ensing, L. A. Everett, E. D. Green, C. Proschel, N. J. Gutowski, M. Noble, D. L. Atkinson, S. J. Odelberg, and M. T. Keating, 1996. LIM-kinase1 hemizygosity implicated in impaired visuospatial constructive cognition. Cell 86:59–69. Franke, U., 1999. Williams-Beuren syndrome: Genes and mechanisms. Hum. Mol. Gen. 8:1947–1954. Galaburda, A. M., D. P. Holinger, U. Bellugi, and G. F. Sherman, 2002. Williams syndrome: Neuronal size and neuronal packing density in primary visual cortex. Arch. Neurol. 59:1461–1467. Gosch, A., and R. Pankau, 1997. Personality characteristics and behaviour problems in individuals of different ages with Williams syndrome. Dev. Med. Child Neurol. 39:527– 533. Grant, J., A. Karmiloff-Smith, S. E. Gathercole, S. Paterson, P. Howlin, M. Davies, and O. Udwin, 1997. Phonological short-term memory and its relationship to language in Williams syndrome. J. Cogn. Neuropsychol. 2:81–99. Grant, J., V. V. Valian, and A. Karmiloff-Smith, 2002. A study of relative clauses in Williams syndrome. J. Child Lang. 29: 403–416. Gray, V., A. Karmiloff-Smith, E. Funnell, and M. Tassabehji, 2006. In-depth analysis of spatial cognition in Williams syndrome: A critical assessment of the role of the LIMK1 gene. Neuropsychologia 44(5):679–685. Grice, S., M. W. Spratling, A. Karmiloff-Smith, H. Halit, G. Csibra, M. de Haan, and M. H. Johnson, 2001. Disordered visual processing and oscillatory brain activity in autism and Williams syndrome. NeuroReport 12:2697–2700. Grice, S. J., M. de Haan, H. Halit, M. H. Johnson, G. Csibra, J. Grant, and A. Karmiloff-Smith, 2003. ERP abnormalities of visual perception in Williams syndrome. NeuroReport 14:1773– 1777. Hallidie-Smith, K. A., and S. Karas, 1988. Cardiac anomalies in Williams-Beuren syndrome. Arch. Dis. Child 63:809–813.

karmiloff-smith: research into williams syndrome

697

Hammond, P., T. J. Hutton, J. E. Allanson, B. Buxton, L. E. Campbell, J. Clayton-Smith, D. Donnai, A. Karmiloff-Smith, K. Metcalfe, K. C. Murphy, M. Patton, B. Pober, K. Prescott, P. Scambler, A. Shaw, A. C. M. Smith, A. F. Stevens, I. K. Temple, R. Hennekam, and M. Tassabehji, 2005. Discriminating power of localized three-dimensional facial morphology. Am. J. Hum. Genet. 77:999–1010. Holinger, D. P., U. Bellugi, D. L. Mills, J. R. Korenberg, A. L. Reiss, G. F. Sherman, and A. M. Galaburda, 2005. Relative sparing of primary auditory cortex in Williams syndrome. Brain Res. 1037:35–42. Huttenlocher, P. R., and C. de Courten, 1987. The development of synapses in striate cortex of man. Hum. Neurobiol. 6:1–9. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387:167–178. Jarrold, C., S. J. Harley, C. Phillips, and A. D. Baddeley, 2000. Word fluency in Williams syndrome: Evidence for unusual semantic organization. Cogn. Neuropsychiatry 5:293– 319. Johnson, M. H., 2001. Functional brain development in humans. Nature Rev. Neurosci. 2:475–483. Jones, W., J. Hesselink, E. Courchesne, T. Duncan, K. Matsuda, and U. Bellugi, 2005. Cerebellar abnormalities in infants and toddlers with Williams syndrome. Dev. Med. Child Neurol. 44: 688–694. Karmiloff-Smith, A., 1992. Beyond Modularity. Cambridge, MA: MIT Press. Karmiloff-Smith, A., 1998. Development itself is the key to understanding developmental disorders. Trends Cogn. Sci. 2:389– 398. Karmiloff-Smith, A., J. Grant, I. Berthoud, M. Davies, P. Howlin, and O. Udwin, 1997. Language and Williams syndrome: How intact is “intact”? Child Dev. 68:246–262. Karmiloff-Smith, A., J. Grant, S. Ewing, M. J. Carette, K. Metcalfe, D. Donnai, A. P. Read, and M. Tassabehji, 2003. Using case study comparisons to explore genotype-phenotype correlations in Williams-Beuren syndrome. J. Med. Genet. 40(2):136–140. Karmiloff-Smith, A., G. Scerif, and D. Ansari, 2003. Double dissociations in developmental disorders? Theoretically misconceived, empirically dubious. Cortex 39:161–163. Karmiloff-Smith, A., A. D. Smith, E. Chin, A. Annaz, M. Elsabbagh, I. D. Gilchrist, P. Hammond, A. M. Hogan, B. Hood, and M. Tassabehji, submitted. Of mice and men: Genetic contributions to spatial cognition. Karmiloff-Smith, A., and M. S. C. Thomas, 2003. What can developmental disorders tell us about the neurocomputational constraints that shape development? The case of Williams syndrome. Dev. Psychopathol. 15:969–990. Karmiloff-Smith, A., M. Thomas, D. Annaz, K. Humphreys, S. Ewing, N. Brace, M. van Duuren, G. Pike, S. Grice, and R. Campbell, 2004. Exploring the Williams syndrome face processing debate: The importance of building developmental trajectories. J. Child Psychol. Psychiatry 45(7):1258– 1274. Klein, B. P., and C. B. Mervis, 1999. Contrasting patterns of cognitive abilities of 9- and 10-year-olds with Williams syndrome or Down syndrome. Dev. Neuropsychol. 16:177–196. Laing, E., G. Butterworth, D. Ansari, M. Gsödl, E. Longhi, G. Panagiotaki, S. Paterson, and A. Karmiloff-Smith, 2002. Atypical development of language and social communica-

698

tion in toddlers with Williams syndrome. Dev. Sci. 5:233– 246. Laing, E., C. Hulme, J. Grant, and A. Karmiloff-Smith, 2001. Learning to read in Williams syndrome: Looking beneath the surface of atypical reading development. J. Child Psychol. Psychiatry 42(6):729–739. Laws, G., and D. V. M. Bishop, 2004. Pragmatic language impairment and social deficits in Williams syndrome: A comparison with Down’s syndrome and specific language impairment. Int. J. Lang. Commun. Disord. 39(1):45–64. Levy, Y., and T. Bechar, 2003. Cognitive, lexical and morphosyntactic profiles of Israeli children with Williams syndrome. Cortex 39:255–271. Li, D. Y., A. E. Toland, B. B. Boak, D. L. Atkinson, G. J. Ensing, C. A. Morris, and M. T. Keating, 1997. Elastin point mutations cause an obstructive vascular disease, supravalvular aortic stenosis. Hum. Mol. Gen. 6:1021–1028. Lukas, A., C. Pleh, and M. Racsmany, 2004. Language in Hungarian children with Williams syndrome. In S. Bartke and J. Siegmuller, eds., Williams Syndrome across Languages, 187–220. Amsterdam: John Benjamins. Masataka, N., 2001. Why early linguistic milestones are delayed in children with Williams syndrome: Late onset of hand banging as a possible rate-limiting constraint on the emergence of canonical babbling. Dev. Sci. 4:158–164. Meng, Y., Y. Zhang, V. Tregoubov, C. Janus, L. Cruz, M. Jackson, W.-Y. Lu, J. F. MacDonald, J. Y. Wang, D. L. Falls, and Z. Jia, 2002. Abnormal spine morphology and enhanced LTP in LIMK-1 knockout mice. Neuron 35:121– 133. Merla, G., and R. Lyle, et al., 2006. Submicroscopic deletion in patients with Williams-Beuren syndrome influences expression levels of the non-hemizygous flanking genes. Am. J. Hum. Genet. 79(2):332–341. Mervis, C., and J. Bertrand, 1997. Developmental relations between cognition and language: Evidence from Williams syndrome. In L. B. Adamson and M. A. Romski, eds., Research on Communication and Language Disorders: Contributions to Theories of Language Development, 75–106. New York: Brookes. Mervis, C. B., C. A. Morris, J. Bertrand, and B. F. Robinson, 1999. Williams syndrome: Findings from an integrated program of research. In H. Tager-Flusberg, ed., Neurodevelopmental Disorders, 65–110. Cambridge, MA: MIT Press. Meyer-Lindenberg, A., C. B. Mervis, D. Sarpal, P. Koch, S. Steele, P. Kohn, S. Marenco, C. A. Morris, S. Das, S. Kippenham, V. S. Mattay, D. R. Weinbergerm, and K. F. Berman, 2005. Functional, structural and metabolic abnormalities of the hippocampal formation in Williams syndrome. J. Clin. Invest. 115:1888–1895. Mills, D. L., T. D. Alvarez, M. St. George, L. G. Appelbaum, U. Bellugi, and H. Neville, 2000. Electrophysiological studies of face processing in Williams syndrome. J. Cogn. Neurosci. 12:47–64. Morris, C. A., I. T. Thomas, and F. Greenberg, 1993. Williams syndrome: Autosomal dominant inheritance. Am. J. Med. Genet. 47:478–481. Morris, C. A., S. A. Demsey, C. O. Leonard, C. Dilts, and B. L. Blackburn, 1998. Natural history of Williams syndrome: Physical characteristics. J. Pediatr. 113:318–326. Nakamura, M., S. Watanabe, A. Gunji, and R. Kakigi, 2006. The magnetoencephalographic response to upright and inverted face stimuli in a patient with Williams syndrome. Pediatr. Neurol. 34(5):412–414.

neurodevelopmental aspects of clinical disorders

Nazzi, T., A. Gopnik, and A. Karmiloff-Smith, 2005. Asynchrony in the cognitive and lexical development of young children with Williams syndrome. J. Child Lang. 32:427–438. Nazzi, T., and A. Karmiloff-Smith, 2002. Early categorization abilities in young children with Williams syndrome. NeuroReport 13(10):1259–1262. Nazzi, T., S. Paterson, and A. Karmiloff-Smith, 2003. Early word segmentation by infants and toddlers with Williams syndrome. Infancy 4(2):251–271. Neville, H. J., 2006. Flexibility and plasticity in cortical development. In Y. Munakata and M. J. Johnson, eds., Attention and Performance, XXI: Processes of Change in Brain and Cognitive Development, 287–314. Oxford, UK: Oxford University Press. Neville, H. J., D. L. Mills, and U. Bellugi, 1994. Effects of altered auditory sensitivity and age of language acquisition on the development of language-relevant neural systems: Preliminary studies of Williams syndrome. In S. Broman and J. Grafman, eds., Atypical Cognitive Deficits in Developmental Disorders: Implications for Brain Function, 67–83. Hillsdale, NJ: Erlbaum. Pankau, R., A. Gosch, E. Simeoni, and A. Wessel, 1993. WilliamsBeuren syndrome in monozygotic twins with variable expression. Am. J. Med. Genet. 47:475–477. Paterson, S. J., J. H. Brown, M. K. Gsodl, M. H. Johnson, and A. Karmiloff-Smith, 1999. Cognitive modularity and genetic disorders. Science 286:2355–2358. Paterson, S. J., L. Girelli, B. Butterworth, and A. KarmiloffSmith, 2006. Are numerical impairments syndrome specific? Evidence from Williams syndrome and Down’s syndrome. J. Child Psychol. Psychiatry 47(2):190–204. Peoples, R., Y. Franke, Y.-K. Wang, L. Perez-Jurado, T. Paperna, M. Cisco, and U. Franke, 2000. A physical map, including a BAC/PAC clone contig of the Williams-Beuren syndrome deletion region at 7qll.23. Am. J. Hum. Genet. 66:47– 67. Perez-Jurado, L. A., R. Peoples, P. Kaplan, B. C. J. Hamel, and U. Franke, 1996. Molecular definition of the chromosome 7 deletion in Williams syndrome and parent-of-origin effects on growth. Am. J. Hum. Genet. 59:781–792. Piatelli-Palmerini, M., 2001. Speaking of learning. Nature 411: 887–888. Pinker, S., 1994. The Language Instinct. London: Penguin. Rae, C., A. Karmiloff-Smith, M. A. Lee, R. M. Dixon, et al., 1998. Brain biochemistry in Williams syndrome: Evidence for a role of the cerebellum in cognition? Neurology 51:33–40. Reiss, A. L., S. Eliez, J. E. Schmitt, E. Straus, Z. Lai, W. Jones, and U. Bellugi, 2000. Neuroanatomy of Williams syndrome: A high-resolution MRI study. J. Cogn. Neurosci. 12(1):65– 73. Rossen, M., E. S. Klima, U. Bellugi, A. Bihrle, and W. Jones 1996. Interaction between language and cognition: Evidence from Williams syndrome. In J. H. Beitchman, N. J. Cohen, M. M. Konstantareas, and R. Tannock, eds., Language, Learning, and Behavior Disorders: Developmental, Biological, and Clinical Perspectives, 367–392. New York: Cambridge University Press. Scerif, G., K. Cornish, J. Wilding, J. Driver, and A. Karmiloff-Smith, 2004. Visual selective attention in typically developing toddlers and toddlers with fragile X and Williams syndrome. Dev. Sci. 7:116–130. Scerif, G., A. Karmiloff-Smith, D. Ansari, and L. K. Tyler, submitted. Is cognitive processing modular or domain-general? A comparison of lexical and numerical representations in Williams syndrome and healthy controls.

Schmitt, J. E., K. Watts, S. Eliez, U. Bellugi, A. M. Galaburda, and A. L. Reiss, 2002. Increased gyrification in Williams syndrome: Evidence using 3D MRI methods. Dev. Med. Child Neurol. 44:292–295. Singer-Harris, N. G., U. Bellugi, E. Bates, W. Jones, and M. Rossen, 1997. Contrasting profiles of language development in children with Williams and Down syndromes. Dev. Neuropsychol. 13:345–370. Smith, N., and I.-M. Tsimpli, 1995. The Mind of a Savant: Language, Learning and Modularity. Oxford, UK: Blackwell. Stevens, T., and A. Karmiloff-Smith, 1997. Word learning in a special population: Do individuals with Williams syndrome obey lexical constraints? J. Child Lang. 2:737–765. Stromme, P., P. G. Bjornstad, and K. Ramstad, 2002. Prevalence estimation of Williams syndrome. J. Child Neurol. 17:269– 271. Tager-Flusberg, H., J. Boshart, and S. Baron-Cohen, 1998. Reading the windows to the soul: Evidence of domain-specific sparing in Williams syndrome. J. Cogn. Neurosci. 10:631– 639. Tager-Flusberg, H., and K. Sullivan, 2000. A componential view of theory of mind: Evidence from Williams syndrome. Cognition 76:59–90. Tassabehji, M., K. Metcalfe, W. D. Fergusson, M. J. Carette, J. K. Dore, D. Donnai, A. P. Read, C. Proschel, N. J. Gutowski, X. Mao, and D. Sheer, 1996. LIM-kinase deleted in Williams syndrome. Nature Genet. 13:272–273. Tassabehji, M., K. Metcalfe, A. Karmiloff-Smith, M. J. Carette, J. Grant, N. Dennis, W. Reardon, M. Splitt, A. P. Read, and D. Donnai, 1999. Williams syndrome: Use of chromosomal microdeletions as a tool to dissect cognitive and physical phenotypes. Am. J. Hum. Genet. 64:118–125. Tassabehji, M., P. Hammond, A. Karmiloff-Smith, P. Thompson, M. E. Durkin, S. Thorgeirsson, K. Metcalfe, A. Rucka, T. Hutton, A. Hogan, H. Stewart, A. P. Read, M. Maconochie, and D. Donnai, 2005. GTF2IRD1 in craniofacial development of humans and mice. Science 310(5751):1184– 1187. Temple, C. M., M. Almazan, and S. Sherwood, 2002. Lexical skills in Williams syndrome: A cognitive neuropsychological analysis. J. Neurolinguistics 15(1):463–495. Thomas, M. S. C., D. Annaz, D. Ansari, G. Scerif, C. Jarrold, and A. Karmiloff-Smith, in press. The case for using developmental trajectories in the study of developmental disorders: Methodological considerations. J. Speech Lang. Hear. Res. Thomas, M. S. C., J Grant, Z. Barham, M. Gsödl, E. Laing, L Lakusta, L. K. Tyler, S. Grice, S. Paterson, and A. Karmiloff-Smith, 2001. Past tense formation in Williams syndrome. Lang. Cogn. Processes 2(16):143–176. Thomas, M. S. C., and A. Karmiloff-Smith, 2003. Connectionist models of cognitive development, atypical development and individual differences. In R. J. Sternberg, J. Lautrey, and T. Lubart, eds., Models of Intelligence for the Next Millennium, 133–150. American Psychological Association. Tsirempolou, E., K. Lawrence, K. Lee, S. Ewing, and A. Karmiloff-Smith, 2006. Understanding the social meaning of the eyes: Is Williams syndrome so different from autism? World J. Pediatr. 2:288–296. Udwin, O., and W. Yule, 1991. A cognitive and behavioural phenotype in Williams syndrome. J. Clin. Exp. Neuropsychol. 13:232–244. Urban, Z., C. Helms, G. Fekete, K. Csiszar, D. Bonnet, A. Munnich, H. Donis-Keller, and C. D. Boyd, 1996. 7q11.23

karmiloff-smith: research into williams syndrome

699

deletions in Williams syndrome arise as a consequence of unequal meiotic crossover. Am. J. Hum. Genet. 59:958–962. Vicari, S., G. Carlesimo, D. Brizzolara, and G. Pezzini, 1996. Short-term memory in children with Williams syndrome: A reduced contribution of lexical-semantic knowledge to word span. Neuropsychologia 34(9):919–992. Volterra, V., O. Capirci, and M. C. Caselli, 2001. What atypical populations can reveal about language development: The contrast between deafness and Williams syndrome. Lang. Cogn. Processes 16:19–239.

700

Volterra, V., O. Capirci, G. Pezzini, L. Sabbadini, and S. Vicari, 1996. Linguistic abilities in Italian children with Williams syndrome. Cortex 32:663–677. Volterra, V., M. C. Caselli, O. Capirci, F. Tonucci, and S. Vicari, 2003. Early linguistic abilities of Italian children with Williams syndrome. Dev. Neuropsychol. 23:33– 58. Williams, J. C. P., B. G. Barratt-Boyes, and J. B. Lowe, 1961. Supravalvular aortic stenosis. Circulation 24:1311– 1318.

neurodevelopmental aspects of clinical disorders

43

Neurocognitive Development in Autism MIKLE SOUTH, SALLY OZONOFF, AND ROBERT T. SCHULTZ

Background Diagnosis Autism is part of a spectrum of disorders characterized by deficits in social relatedness, communication, and a variety of behavioral problems such as restricted interests, sensory sensitivities, and repetitive behaviors. Onset is by the age of three. In the social domain, symptoms include impaired use of nonverbal behaviors (e.g., eye contact, facial expression, gestures) to regulate social interaction, failure to develop age-appropriate peer relationships, little seeking to share enjoyment or interests with other people, and limited social-emotional reciprocity. Communication deficits include delays in or absence of spoken language, difficulty with conversational reciprocity, idiosyncratic or repetitive language, and imitation and pretend play deficits. In the behaviors and interests domain, there are often encompassing, unusual interests, inflexible adherence to nonfunctional routines, stereotyped body movements, and preoccupation with parts or sensory qualities of objects (American Psychiatric Association, 2000). Associated features that frequently co-occur with autism include mental retardation, seizures, and dysregulation of both eating and sleeping patterns. Autism and related disorders can co-occur with a variety of additional psychiatric and behavioral disturbances: the most common comorbid conditions in the autism spectrum population are anxiety disorders and depressed mood, with attention problems and tic disorders also elevated (Lainhart, 1999). Autism is recognized as a heterogeneous disorder and is part of a continuum of disability referred to as autism spectrum disorders (ASDs). For instance, within this spectrum, Asperger syndrome is defined by the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV) as sharing the social disabilities and restricted behaviors and interests of autism, but not significant delays in the onset or early course of language; the recognition of other symptoms may occur somewhat later than for autism. Evidence is mixed for the external validity of Asperger syndrome as distinct from autistic disorder when it is accompanied by normal intellectual functioning (“high-functioning autism”). Pervasive developmental disorder not otherwise specified (PDDNOS) is a label used for children who experience difficulties in at

least two of the three autism-related symptom clusters but who do not meet criteria for autism or Asperger syndrome. The diagnosis is often misused, with substantial proportions of children carrying this label either meeting full criteria for autism or not meeting criteria for any autism spectrum disorder (Willemsen-Swinkels and Buitelaar, 2002). A major theme of this chapter is the heterogeneity of the autism spectrum with regard to developmental symptoms, neuropsychological profiles, and degree and quality of social impairment. For example, heterogeneity may occur across diagnostic categories (e.g., autism versus Asperger syndrome), across levels of functioning (e.g., level of language development and/or cognitive impairment), or with regard to genetic and/or psychiatric comorbidity. This heterogeneity almost certainly reflects multiple etiological pathways, although the ASDs may collectively share some underlying neurobiological bases. Developmental Course The majority of children diagnosed with autism (approximately two-thirds) display developmental abnormalities within the first two years of life. A smaller group display a period of normal or mostly normal development, followed by a loss of communication and social skills and onset of autism (Lord, Shulman, and DiLavore, 2004). Currently, the average age of diagnosis of autism is about 36 months (Mandell, Novak, and Zubritsky, 2005), although symptoms often appear much earlier (Zwaigenbaum et al., 2005). Parents often begin to be concerned when language fails to develop as expected. However, several other behavioral differences, particularly social ones, appear to predate the language abnormalities that parents report at the time of recognition. These include less looking at faces, decreased responses to hearing one’s name, decreased pointing, and lack of sharing enjoyment and interests with others (see Zwaigenbaum et al., 2005). The onset of Asperger syndrome and PDDNOS are less well understood. Children with these disorders usually present at older ages (Mandell, Novak, and Zubritsky, 2005), and parent report of early development may not be as accurate when more time has passed. Additionally, the symptoms of Asperger syndrome and PDDNOS can be more subtle in preschool than those of autism and may be difficult to detect even by professionals.

701

Epidemiology The prevalence of autism has increased substantially in the past decade, leading some to speculate about potential causal contributions of environmental factors. However, the best evidence to date indicates that the rise in autism cases does not reflect an actual increase in incidence; rather it is due to a combination of factors including improved ascertainment and significant broadening of the diagnostic criteria (Rutter, 2005b). One theory of environmental insult that has had enormous influence in the public and the media is the notion that the combined measles-mumps-rubella vaccine was causally linked to increased prevalence of autism. Subsequent epidemiological studies have discounted this theory, and the original report on the matter has been questioned on numerous grounds (Parker et al., 2004; Taylor et al., 2006). Nonetheless, the possibility of a major CNS insult through early viral infection or other external factors cannot be ruled out at this time, and further research in this area is clearly needed (Libbey et al., 2005; Rutter, 2005b).

Etiology Genetic Basis of Autism In his original (1943) description of autism, Leo Kanner suggested that children with autism were born with “an innate inability to form the usual, biologically provided affective contacts with people” (p. 250). Subsequent research has made it clear that biological mechanisms produce brain changes that lead to the symptoms of autism. The heritability for autism has been estimated to lie above .90, and there are no viable socialenvironmental hypotheses of etiology. Nevertheless, it is clear that there are nongenetic sources of influence; this finding is clearest in the case of identical twins where one sibling has full-blown autism and the other either has a milder variant or, in rare cases, does not present on the spectrum at all. The term “broader autism phenotype” has been used to describe milder manifestations of autistic-like behaviors seen in some relatives of autism probands. Numerous studies of first-degree relatives of people with autism have documented increased frequency, relative to a variety of comparison groups, of behaviors that are qualitatively similar to autism, including few friendships, limited conversational reciprocity, delayed speech onset, and narrow focused interests (Bishop et al., 2006; Constantino et al., 2006; Kolevzon, Matthewson, and Hollander, 2006; Le Couteur et al., 1996; Piven et al., 1997). Increases in secondary characteristics including mood disorders (Bolton et al., 2001; Ghaziuddin, 2005) and atypical cognitive profiles (Bolte and Poustka, 2006; Hughes, Plumet, and Leboyer, 1999) have also been documented. The mode of transmission for ASDs (and the broader manifestations seen in family members) is thought to be extremely complex, involving multiple genes and pathways

702

and probable epistasis (modification of one gene by another) and locus heterogeneity (Sutcliffe et al., 2005). As we will discuss, genome scans of autism have produced the most interest in candidate gene regions on chromosomes 7q, 2q, and 17q, but to date no clear and consistent genetic markers have been identified. Comprehensive reviews of the genetics of autism can be found elsewhere (e.g., Rutter, 2005a); this chapter highlights genetic findings that relate specifically to brain development in autism. Endophenotypes. The heterogeneity of symptom expression in autism poses significant challenges for traditional genetic approaches. Endophenotypes are measurable characteristics that lie on the pathway between genes and symptom expression (Gottesman and Gould, 2003). For example, rather than focus on the genes underlying “autism” as a disorder, genetics research is more concerned with identifying the genetic basis for intermediate, observable traits such as abnormal head growth, serotonin levels, eye gaze behavior, or language development. Such endophenotypes may serve to create more genetically homogeneous subgroups as well as to indicate shared traits in relatives with subclinical symptom expression, as indicators of susceptibility genes (Coon, 2006). Candidate genes. There have been a number of candidate gene models that incorporate reasonable functional links to potential causal brain mechanisms in autism (reviewed in Bachelli and Maestrani, 2006). Some evidence suggests that the RELN gene on chromosome 7q22 (related to neuronal migration) and several X-linked neuroligin genes (related to synaptogenesis) could theoretically be involved in the brain bases of autism. In addition, there is a longstanding interest in the role of the serotonin transporter gene in the area of 17q11-12, SLC6A4. While much of the evidence is mixed (Veenstra-VanderWeele, Christian, and Cook, 2004), there is newer evidence for a collection of multiple, often rare alleles at this site that bestow increased risk for autism (Sutcliffe et al., 2005). This gene could be an important contributor to findings of serotonergic abnormalities in autism (reviewed in Anderson and Hoshino, 2005). Millonig and colleagues (Benayed et al., 2005) have found significant associations between autism and the homeobox transcription factor gene, Engrailed 2, in three samples. In mice, mutations in this gene lead to atypical cerebellar development, including hypoplasticity and a decrease in the number of Purkinje cells. Similarities to abnormalities that have been found in autism (e.g., Kemper and Bauman, 2002) led Benayed and colleagues (2005) to propose that Engrailed 2 might act as a susceptibility locus for autism. At this time there is little solid evidence for any other candidate genes. Although the literature is replete with

neurodevelopmental aspects of clinical disorders

significant association studies, failures to replicate with independent samples are the norm. New techniques offer realistic potential for the eventual resolution of these and many other genetic puzzles in autism, however, and the development of large consortiums that increase study sample size provides a sense of great optimism in the field of autism genetics, despite the genetic complexities that appear for both individuals and extended families (Coon, 2006). While large sample sizes are certainly key for finding common variants, a different but complementary approach stresses finding rare variants, which have a large impact on the developing brain and thus are important for pointing to specific neurodevelopmental pathways. It now seems clear that autism is not a single entity at either the phenotypic or genetic level; the hope is that heterogeneity will be less at the level of the endophenotype and that the unifying principals will emerge from a focus on observable abnormalities in brain structure and function. Brain Development Prenatal course. Arndt, Stodgell, and Rodier (2005) conclude from several lines of evidence that at least some forms of autism are consistent with early prenatal insults to the developing brain, perhaps as early as the first eight weeks. The critical period for exposure to a variety of known teratogens with possible relation to autism (e.g., maternal rubella infection and thalidomide) has been shown to occur between 3 and 8 weeks postconception (see also Miller, 2005). Nelson and colleagues (2001) found atypical concentrations of certain neuropeptides and neurotrophins during newborn blood screening in babies later diagnosed with autism or mental retardation. Patterns of congenital dysmorphic features in idiopathic cases of autism suggest early origins (Miles and Hillman, 2000). Histological studies of postmortem autism brains (e.g., Bailey et al., 1998) have found reductions in fibers and other abnormalities that are very consistent with early developmental atypicality. There may be an increased risk of autism in children carried by pregnant women taking valproic acid (VPA), a drug used to treat seizure disorders and migraine headaches and the mood instabilities found in conditions like bipolar disorder and borderline personality disorder. This drug or its metabolites may pass through the placenta and/or breast milk. Valproic acid given to rats at certain periods of prenatal development causes morphological and behavioral abnormalities consistent with some aspects of autism, especially with regard to cerebellar development and function (see Stodgell et al., 2001). Despite these and other proposed links (Miyazaki, Narita, and Narita, 2005) between early teratogens and autism, pre- or perinatal difficulties appear to play a significant role in only a small number (∼5%) of autism cases (Rutter, 2000).

Head and brain growth. Kanner’s original description of autism noted that many of the cases had enlarged heads (Kanner, 1943). Although this observation was confirmed subsequently by others (e.g., Steg and Rapoport, 1975; Walker, 1977), it did not receive much attention until postmortem and MRI studies began to confirm that the brain was enlarged in autism (e.g., Aylward et al., 2002; Bailey et al., 1995; Courchesne et al., 2001; Piven et al., 1995). Enlargement of the brain (by about 3–10%) is now one of the most consistently observed biological correlates of autism in the entire literature. Postnatally, some children with autism demonstrate accelerated head growth (reviewed in Lainhart, 2006). Macrocephaly (i.e., head circumference above the 97th percentile) reported in the second half of the first year is associated with an up to five times greater risk for autism (Bolton et al., 2001). Increased head growth, which reflects increased brain growth (Courchesne, Carper, and Akshoomoff, 2003; Rice et al., 2005), is reported for 36–73 percent of children with autism during the first year of life (Lainhart, 2006). Differences in head size and/or brain size between autism and typical development appear to be smaller in adolescents and adults than in individuals within the earliest stages of childhood (Aylward et al., 2002; Courchesne et al., 2001; Piven et al., 1995; Schultz et al., 2005). In general, macrocephaly is associated with an increased rate of neurodevelopmental disorders and genetic syndromes. Benign macrocephaly is defined as large head size that is not associated with any major developmental problems, including normal intelligence and neurological examinations. Nonetheless, children with benign macrocephaly do have an increased rate of subtle developmental differences, including mild articulation problems, mild speech delays, and/or mild attention-deficit/hyperactivity disorder (ADHD) (personal communication from Janet Lainhart). Magnetic resonance imaging (MRI) studies of the brain allow for much more precise investigation of contributions to enlarged head size, as well as characterization of specific neural systems. Recently, several studies have found that the white matter is disproportionately enlarged in autism compared to gray matter (Courchesne, Carper, and Akshoomoff, 2001; Herbert et al., 2004; Schultz et al., 2005). At the same time, there is good evidence that not all white matter tracts are enlarged. One of the most consistent structural MRI findings in the autism literature is decreased cross-sectional width of the corpus callosum, which contains long association fibers (e.g., Hardan et al., 2000; Piven et al., 1997). Thus, although overall white matter is increased in volume, it must be disproportionately distributed among certain fiber tracks. These data are suggestive of fundamental difficulties with systems-level neuroconnectivity in autism. In fact, reductions in connectivity, especially long-range connectivity, are predicted on the basis of theoretical models of increased brain size (Ringo, 1991). There is a physical limit

south, ozonoff, and schultz: neurocognitive development in autism

703

on the degree of interconnectedness with increasing gray matter, such that the degree of systems-level connectivity must be decreased with increasing brain size (Schultz, Romanski, and Tsatsanis, 2000). Numerous fMRI studies (reviewed in Schultz and Robins, 2005) have shown reduced functional connectivity between anterior and posterior brain regions, abnormal temporal synchronization, reflecting in part reduced frontal lobe activity during language and working-memory tasks, and a variety of perceptual processing abnormalities (for example, face perception deficits reviewed later). One report of abnormally large amygdala volume in young children with autism, but no difference from controls in older children and teenagers (Schumann et al., 2004), has been interpreted to indicate the possibility that hyperanxiety in early childhood leads to slowed amygdala development (Dalton et al., 2005). Diffusion tensor imaging (DTI) is a relatively new MRIbased tool that examines variability in water diffusion to trace the integrity and anatomy of white matter tracts and bundles (see chapter 17 by Wozniak, Mueller, and Lim, this volume). Several research programs are currently investigating the pattern of white matter abnormalities in autism. One published study (Barnea-Goraly et al., 2004) demonstrated widespread and substantial decreases in white matter, especially in prefrontal and temporal-parietal regions, in a small group of children and adolescents diagnosed with autism and with IQ scores in the normal range. Neurochemistry. Central nervous system (CNS) function is clearly altered in autism (reviewed in Anderson and Hoshino, 2005). Nonetheless, there are few replicated differences between individuals with autism and controls in the levels of neurotransmitters, neuropeptides, and other hormonal and metabolic measures. The most consistent finding in autism is that of elevated peripheral serotonin levels (especially in blood platelets). Serotonin is an important neurotransmitter contributing to body regulatory function and mood. It has a rich theoretical potential for links to autism (Scott and Deneris, 2005), and there is evidence for partial efficacy of serotonin-based medicines for improving anxiety and repetitive behavior in autism (reviewed in Kolevzon, Matthewson, and Hollander, 2006). It appears that there is a bimodal distribution of serotonin levels along the autism spectrum, with approximately half of ASD individuals showing hyperserotonemia measured in blood platelets (Mulder et al., 2004). However, studies of urine metabolites and other measures of serotonin systems suggest these elevated levels may arise as a result of the way in which platelets process serotonin, rather than from increased serotonin production in the gut. Peripheral serotonin levels from platelets can be independent of CNS levels (Anderson and Hoshino, 2005), so it is not clear whether the findings of elevated peripheral serotonin reflect abnormal brain

704

development or functioning. The few studies of CNS serotonin levels in autism (see Narayan et al., 1993) have not demonstrated pervasive changes. The GABA system may also play a role in autism. GABAergic interneurons are important for establishing the computational architecture of cortical columns, and they may be particularly susceptible to developmental errors (Belmonte et al., 2004). There is evidence for low GABA receptor binding in the hippocampus as well as low platelet GABA levels in autism (see Belmonte et al., 2004, for a thorough review of possible GABA contributions to autism). Cognitive Models Four cognitive models have created intense interest among autism researchers over the past two decades: the executive dysfunction, weak central coherence, theory of mind, and complex information processing models. Although each of these theories has provided a reasonable explanation of specific deficits that are commonly seen in autism, none has been able to fully account for either the range of symptoms (social interaction, communication, and repetitive behaviors) or the variability across different levels of cognitive ability and adaptive skills seen within the autism spectrum. Executive function. Executive functions (EFs) are the cognitive abilities that allow an individual to maintain an appropriate focus in order to attain a desired goal (e.g., planning, set shifting, inhibiting, dealing with novelty and unpredictability). Four decades of research have demonstrated that individuals with ASD have clear deficits in this neuropsychological domain (see Ozonoff, South, and Provencal, 2005, for a review). A recent neuroimaging study (Schmitz et al., 2006) found that individuals with ASD show abnormalities on functional MRI when completing EF tasks, suggesting that the brain systems underlying these cognitive abilities are dysfunctional. The EF model of autism posited that deficits in EF gave rise to other symptoms of autism, such as perseveration, repetitive behaviors, and social dysfunction. Early criticisms of the EF model highlighted its limited specificity for autism, as executive function difficulties are found in a range of disorders (Pennington and Ozonoff, 1996). One of the strengths of recent research in this area has been a “component process approach” which seeks to decompose the multidimensional construct of EF into more elementary and specific operations. For example, it appears that inhibitory control and working memory are relatively spared executive functions in autism but that several aspects of mental flexibility (e.g., set shifting and attention shifting) are impaired. A second weakness of the EF account of autism is that executive deficits are not present, relative to children with typical or delayed development, until after the preschool period (Griffith et al., 1999), and it is difficult to explain how a deficit that only emerges over time can

neurodevelopmental aspects of clinical disorders

account for symptoms that precede it by several years. A third issue, as we will discuss later, is the unclear picture regarding links between EF performance and core diagnostic features of autism such as repetitive behaviors. Weak central coherence. A second major theory has focused on what has been labeled weak central coherence (WCC), or a limited capacity to integrate information meaningfully into a coherent whole (Frith, 1989). This theory emphasizes a cognitive style that is biased toward local rather than global information processing. The theory explains the social deficits in autism as stemming from difficulties integrating local-level information with global-level information, such as the need to appreciate social context in order to modulate social behavior. Shah and Frith (1993) first suggested that superior performance on the Block Design IQ subtest by ASD individuals arises from their ability to segment the design “gestalt” even in the absence of visual segmentation cues. Many other studies have supported findings of weak central coherence (reviewed in Happe and Frith, 2006). More recently, Mottron and colleagues (e.g., Caron et al., 2006) suggest that this Block Design strength is due to enhanced visual processing abilities arising out of primary visual cortex. However, there have also been a number of failed replications of WCC tasks in ASDs (see Happe and Frith, 2006). Happe and Frith (2006) suggest that these negative findings help to demarcate the boundaries of weak central coherence in autism, in that ASD individuals comprehend object properties and meanings but have difficulty connecting objects together (including visual objects as well as words). Other negative findings (e.g., Plaisted, Swettenham, and Rees, 1999) have highlighted the importance of the modulation of attention on central coherence tasks. Others have suggested that the problems in central coherence in autism arise from the social deficits rather than the reverse; for example, lacking an early drive for social mastery, children with autism lack the important social “templates” that guide later development (Caron et al., 1997). Thus the WCC theory, like the EF theory, is not at the present time accepted as a unifying cognitive account of autism symptomatology. Theory of mind. The theory of mind (ToM) account has probably been the most dominant cognitive model of autism, suggesting that social and other dysfunctions of autism arise from impaired capacity for thinking about thoughts, beliefs, intentions, and feelings of both the self and others (BaronCohen, 1995). Seminal work by Baron-Cohen and colleagues (Baron-Cohen, 1989; Baron-Cohen et al., 1997; Baron-Cohen, Leslie, and Frith, 1985) found deficits in individuals with autism in both simple and more complex social perspective taking, as well as in perceiving the emotional states of others. This account has stimulated important

advances in the field’s thinking about the primary developmental deficits in autism. Klin, Volkmar, and Sparrow (1992) have described a number of difficulties with this theory, however. First, language and IQ are highly correlated with theory-of-mind performance and explain a great deal of variance in these abilities. In fact, many higher-functioning individuals with autism or Asperger syndrome may be able to solve lab-based theory-of-mind tasks quite nicely, and yet they remain significantly impaired socially. In addition, the social difficulties in autism arise in very young infants before theory-of-mind abilities (as typically defined) are first evident. Complex information processing. A more recent cognitive account of autism highlights difficulties with complex information processing (Minshew, Goldstein, and Siegel, 1997). This model states that individuals with autism demonstrate spared performance on neuropsychological tasks of any sort (perceptual, spatial, verbal, or motor) when informationprocessing demands are low, but are impaired on tasks that require higher-order information processing. For example, simple language processes (e.g., phonology, syntax), simple executive functions (e.g., inhibition), simple attentional processes (e.g., selective and focused attention), and simple memory functions (e.g., rote and recognition) may all be spared in autism, while the same domains are impaired at more complex levels (e.g., language pragmatics, cognitive flexibility, abstraction, attention shifting, working memory). Lowered efficiency of information processing may reflect disrupted neural networks arising from enlarged brains (Courchesne, Carper, and Akshoomoff, 2003) or an increased number of cortical columns of reduced size and density (Casanova, Buxhoeveden, and Brown, 2002). Conceptually, however, the definition of “complexity” remains elusive. Is complexity the same as difficulty? Cognitive complexity needs to be defined independently of the empirical results in autism, yet be able to explain the specific pattern of cognitive strengths and weaknesses found in autism (see the general discussion of external validity in Pennington, 2002). Chapman and Chapman (1973) argued a long time ago that many demonstrations of apparently specific cognitive deficits in schizophrenia did not equate the experimental and comparison tasks for difficulty. A similar pattern could be expected for any extreme neurodevelopmental disorder, including mental retardation syndromes and autism. It will be important to examine whether discrepancies in simple versus complex information processing are specific to autism or exist also in other complex neurocognitive disorders. As with the theory-of-mind account, the complexinformation-processing theory of autism does not adequately explain the very early symptoms of autism that are often apparent before the first birthday (Osterling and Dawson,

south, ozonoff, and schultz: neurocognitive development in autism

705

1994). Autism is particularly interesting because impairments in very basic interpersonal processes coexist with spared abilities in certain cognitive domains, such as memory and visual-spatial processes. The complex-information-processing theory cannot explain why a 2-year-old with autism can put together puzzles at a 3-year mental-age level, but is not able to point, a developmentally simpler skill normally apparent by 1 year of age. Summary of cognitive models. Each of the four models discussed in this section is supported by evidence from a number of domains, but none is sufficient to fully describe the complexity of symptom expression across development that is part and parcel of the autism spectrum disorders. Data in support of these models can be viewed as symptoms of underlying brain-based abnormalities (such as disorganized white matter circuitry), which may then lead to the creation or exacerbation of diagnostic symptoms (such as poor social development arising from the inability to quickly process numerous, rapid social cues). In this sense, the cognitive patterns described by these models are intermediate manifestations of atypical, lower-level neural mechanisms, rather than causes of autism per se. Animal Models Animal models of autism and other complex psychiatric conditions are complicated by the difficulty in probing multifaceted behaviors using simple techniques, often with little ethological relevance, and are also complicated by the consequences of environmental factors such as laboratory rearing conditions and social context for brain-behavior relationships (Insel and Fernald, 2004). Nonetheless, animal models are clearly useful for identifying and subtyping many neurochemical, neuroanatomic, and neurobehavioral pathways in autism. For instance, although mice lack language, theory of mind, and other important behaviors seen in humans and dysfunctional in autism, they may be quite useful for studying simpler behaviors impaired in autism, such as motor stereotypies or atypical social proximity seeking. Mice with knockout genes, mutated genes, or altered neurotransmitter expression are providing important information on gene functionality, developmental patterns, and so forth (Crawley, 2004). Crawley (2004) postulates that using a variety of inbred strains and mutant lines, it will be possible to find genes that are linked to phenotypic extremes. Amaral and colleagues (e.g., Amaral and Corbett, 2003) have found that early amygdala lesions in rhesus monkeys do not impair a variety of social behaviors or the monkeys’ ability to recognize species-specific social cues. They suggest that the amygdala is not essential for reciprocal social interaction, although it may be an important contributor to the development and maintenance of anxiety, which is often a comorbid complication of autism.

706

Scientists from other areas who are approaching the problem of autism for the first time need to understand the complexity and heterogeneity of the autism spectrum and carefully consider the adequacy of symptom matching between animals and humans.

Autism and the social brain Because the triad of symptoms that define autism is rather broad in nature, one would presume that the syndrome affects a widely distributed set of neural systems (Schultz and Robins, 2005). However, there are also areas of spared function in autism, including basic perceptual skills and certain cognitive functions. Thus significant social disability in autism may be accompanied by average or above average general intelligence, implying that social intelligence and the cognitive intelligence measured by traditional IQ and memory tests arise from distinct neural systems. Leo Kanner’s original report highlighted social dysfunction (the term “autism” comes from the Greek root autos, describing the preoccupation with the self), and this emphasis has continued as a fundamental principle for research into the neuropathophysiology of autism. Although some social skills develop over time in individuals with ASD, by definition social development in autism is both delayed and deviant; these social deficits appear to be more unique to autism and less shared with other neuropsychiatric disabilities than other core autism deficits (Schultz, 2005). This section reviews findings regarding a number of areas of socialcommunicative development, including face perception, social cognition, and motor imitation. Social Perception Facial identity processing. Impaired face-to-face social engagement is a hallmark symptom of autism: paucity of eye contact, poor pragmatic communication, and difficulties in face perception have been described repeatedly. In contrast to “face expertise,” which typically develops naturally and vigorously, individuals with ASD have difficulty with recognizing face identity despite vision that is intact generally and for other complex visual objects (reviewed in Grelotti, Gauthier, and Schultz, 2002). Dawson, Webb, McPartland, (2005) have found face-recognition difficulties in behavioral and event-related potential (ERP) studies of parents of children with autism, strongly suggesting that face processing is a functional trait marker (i.e., a valid endophenotype) for autism. Ongoing work suggests that the magnitude of the deficit in recognizing facial identity is as large as or larger than that in any other neuropsychological domain. This deficit is marked by an overemphasis on individual features and a deemphasis on the overall holistic gestalt in the face.

neurodevelopmental aspects of clinical disorders

Figure 43.1 Functional MRI abnormalities observed in autism spectrum disorders (ASD). A, These coronal MRI images show the cerebral hemispheres above, the cerebellum below, and a circle over the fusiform gyrus of the temporal lobe. The examples illustrate the frequent finding of hypoactivation of the fusiform gyrus to faces in an adolescent male with ASD (right) compared with an age- and IQ-matched healthy control male (left). The red/yellow signal shows brain areas that are significantly more active during perception of faces; signals in blue show areas more active during perception of nonface objects. Note the lack of face activation in the boy with ASD but average levels of nonface object activation. B, Schematic diagrams of the brain from lateral and medial orien-

tations illustrating the broader array of brain areas found to be hypoactive in ASD during a variety of cognitive and perceptual tasks that are explicitly social in nature. Some evidence suggests that these areas are linked to form a “social brain” network. IFG, Inferior frontal gyrus (hypoactive during facial expression imitation); pSTS, posterior superior temporal sulcus (hypoactive during perception of facial expression and eye gaze tasks); SFG, superior frontal gyrus (hypoactive during theory of mind tasks, i.e. when taking another person’s perspective); A, amygdala (hypoactive during a variety of social tasks); FFA, fusiform face area (hypoactive during perception of personal identity). (From DiCicco-Bloom et al., 2006.) (See plate 58.)

For example, ASD individuals show less performance degradation with inverted faces than do controls (e.g., Teunisse and de Gelder, 2003). Typically developing individuals demonstrate facilitated face recognition for “lowspatial-frequency” information (i.e., the position of features in space), while children with ASD do better on behavioral tasks that emphasize high-spatial-frequency information (i.e., sharp changes in brightness around the edges; Deruelle et al., 2004). Schultz, Ganthier, et al. (2000) published the first functional neuroimaging study of face perception among indi-

viduals with ASD, demonstrating that the fusiform face area (FFA; a small region in the lateral extent of the middle portion of the temporal lobe fusiform gyrus) was hypoactive in a group of 14 persons with autism or Asperger syndrome compared to two independent samples of 14 control participants (see figure 43.1 and plate 58 for an example). In typical development, the FFA shows selectivity (i.e., enhanced activation) for faces compared to other complex objects (Kanwisher, McDermott, and Chun, 1997). Findings of hypoactivity of the FFA in autism have been replicated by nine other labs and now represent the single best-

south, ozonoff, and schultz: neurocognitive development in autism

707

replicated functional MRI marker of ASD in the literature (see Schultz, 2005, for more detailed reviews of these studies, as well as two failed replications, and a theoretical model for this line of research). Preliminary data show that individual differences in autism symptom severity strongly predict individual differences in degree of FFA activation, such that less FFA activation is associated with greater social disability. Accuracy on face identity perception tests also correlates with degree of FFA activation. Although deficits in face perception should theoretically be a hindrance to social functioning, individuals who are blind are not necessarily affected by autism or autistic features, although they do appear to be at an increased risk for developing the condition (e.g., Hobson and Bishop, 2003; Hobson, Lee, and Brown, 1999). In addition, lesions in adulthood that result in prosopagnosia, a disorder of face recognition, generally do not also cause social dysfunction reminiscent of autism (Marlene Behrmann, personal communication). Thus it is still not clear whether developmental deficits in face perception might play a causal role in producing the social deficits seen in autism. Facial expression processing. Children and adults with ASD demonstrate deficits in perceiving, labeling, and understanding facial expressions, which may underlie more complex difficulties in empathy. While the FFA, as described in the preceding subsection, appears to be specific for perceiving face identity, recognizing facial expressions occurs in other distinct brain areas, including the superior temporal sulcus (STS). The posterior STS includes groups of neurons that are specialized for interpreting dynamic social signals, such as direction of eye gaze, gestures, facial expression, and other dynamic aspects of the face and body. Pelphrey and colleagues (2005) have recently shown differential activation of the posterior STS depending on the type of body movement: perception of mouth movements elicited activity along the midposterior STS, whereas eye movements elicited activity that was somewhat more posterior. These different functional subdivisions of the STS may have unique relationships to symptom origin and/or expression in autism. Schultz and colleagues have found preliminary data to support abnormalities in several aspects of STS functioning, including perceptual discrimination of facial expressions and integration of auditory prosody with visual information about facial expressions, in adolescents with ASD (Robins, Hunyadi, and Schultz, 2005). There are now also at least two reports of the morphology of the STS being altered in adolescents and young adults with autism (Boddaert et al., 2004; Waiter et al., 2004). As studies of other brain regions (e.g., the amygdala) and cognitive theories (e.g., executive-function and theory-of-mind accounts) have shown that younger children with autism show a different pattern of development than adolescents and adults,

708

there is a clear need to create techniques and tasks for measuring specific aspects of STS function in younger children with autism. Eye-tracking studies of eye gaze. Some researchers have focused on the eye region as the most critical part of the face for understanding another’s state of mind (see Adolphs et al., 2005). Notably, Baron-Cohen and colleagues (2000) have shown that individuals with ASD have significant difficulty extracting the “language” of complex emotional states from the eyes. Using high-density event-related potentials, Dawson and colleagues have shown atypical viewing of fearful faces in children with autism as young as 3 years old, and have shown that children with autism have difficulty encoding and remembering face-related information (reviewed in Dawson, Webb, and McPartland, 2005). Klin and colleagues (2002) have used infrared eye-tracking technology to measure visual scan paths in persons with autism as they try to make sense of interpersonal social interactions. He found that persons with autism focused much more than typical viewers on the mouth while focusing much less on the rest of the face, particularly the eye region. The distribution of percent viewing time on the eye region for each group did not overlap at all, showing that this one behavioral variable could classify participants with 100 percent sensitivity and specificity. Studies of infant siblings of children with autism have found that later diagnosis of autism spectrum disorders in infants is associated with impaired visual tracking (Zwaigenbaum et al., 2005). Some 6-month-old siblings at familial risk for autism also demonstrate decreased eye gaze and increased gaze at the mouth during eye-tracking studies of a “still face” paradigm (Merin et al., 2007), although most of these infants do not go on to develop autism. The ability to quantify primary social functions such as eye gaze in a variety of naturalistic and experimental settings, across a range of ages and developmental levels, promises to provide valuable insight into the neural mechanisms underlying autism. Mirror neurons. Impaired social-communicative behaviors are the earliest observable symptoms of autism in young children (Toth et al., 2006). These include joint attention (sharing attention to objects or events with another person; see chapter 50 by Mundy and Van Hecke, this volume), the use of communicative gestures (pointing to objects, reaching for objects to request them), and imitation (immediate and deferred imitation of actions on objects; imitation of facial and body movements). These behaviors appear to form essential building blocks to later communicative and social development, in particular the development of language and a theory of mind. Neuroimaging studies of potential early-onset markers have been recently energized by paradigms used to investi-

neurodevelopmental aspects of clinical disorders

gate motor mirror neurons, named for highly specific neural firing in monkey ventral premotor cortex (area F5), not just while executing an action with an object (e.g., reaching for a glass) but also while viewing somebody else do the same (“mirror”) action. It has been suggested that this system is critical for understanding others’ intentions and may be an essential building block for developing a theory of mind (reviewed in Williams et al., 2006). Williams and colleagues found decreased fMRI blood oxygen level–dependent (BOLD) activation in right parietal areas for adolescents with ASD compared to controls during motor mirroring activities; they also report atypical activation for the ASD group in the right temporoparietal junction thought to subserve theory of mind, as well as decreased left amygdala activity. Iacoboni and colleagues (Dapretto et al., 2006) reported reduced fMRI activity for children with highfunctioning autism, relative to controls, in the inferior frontal gyrus during a task comparison of observing versus imitating facial expressions. A challenge to the mirror neuron account is to provide neural models that effectively link motor-system function with the affective dysfunction that is a hallmark of autism. Auditory perception. Although most previous research on social perception in autism has dealt with visual stimuli, the human voice contributes rich social information similar to faces. It is known that autism is associated with a number of deficits in auditory and voice processing such as prosody comprehension, auditory filtering, and verbal content comprehension (reviewed in Tager-Flusberg, Paul, and Lord, 2005). Samson and colleagues (2006) review a number of studies by their group showing deficits in auditory processing in response to nonvoice stimuli (e.g., tones, environmental sounds) across multiple behavioral and psychophysiological methods. A consistent theme of their work is that dysfunction in autism corresponds to the level of stimulus and/or task complexity, in line with the complex-informationprocessing model (Minshew, Goldstein, and Siegel, 1997). Affiliation. Insel and Fernald (2004) have described a framework in which the evolutionary significance of affiliative and reproductive behaviors means that their neural and hormonal correlates are likely to be highly conserved within species. Insel and Fernald suggest that social learning and social motivation are hardwired, rapid, and strong because of their evolutionary importance, and that selective neuropeptides in the mesolimbic dopamine system, especially vasopressin and oxytocin, form the essential link between perception and evaluation of the reward value of a stimulus in species-specific behaviors. Oxytocin knockout mice show profound deficits in social recognition, and pair bonding in prairie voles is affected by manipulations of the vasopressin and/or oxytocin system (Lim, Bielsky, and Young, 2005).

One published study has shown a possible association between autism and markers of the vasopressin V1a receptor (Kim et al., 2002). Social motivation and the amygdala. A failure of appropriate levels of social motivation, when deficient from birth, may derail a whole host of normal developmental processes (reviewed in Baron-Cohen et al., 2000; Schultz, 2005). One example of the developmental consequences of congenitally low levels of social motivation would be a failure to develop normal perceptual expertise for faces (Grelotti, Gauthier, and Schultz, 2002; Rolls, 1999). More broadly, insufficient motivation for social engagement would be expected to lead to less social experience, which would impact social perceptual and social cognitive skill development. From the perspective of this model, congenital deficits in social motivation are a linchpin to the development of autism (see figure 43.2). Amygdala-fusiform connections. Clearly both the STS and the FFA support critical social perceptual functions that are abnormal in autism. However, neither the STS nor the fusiform gyrus (FG) functions as an encapsulated module that can perform functions in an autonomous manner; rather they operate as part of larger neural systems, and information processing subserving social perception likely reflects the interactions between multiple nodes within these systems. The precise interactions between nodes have not yet been well specified, but it appears that the FFA has a special relationship with the amygdala to form the core of one functional neural network (Schultz, 2005). Damage to the amygdala can cause reduced levels of FFA activity to faces (Vuilleumier et al., 2004), suggesting that there are direct and active inputs from the amygdala to the FFA that support or prime its computational activities. In addition to its primary role in social perception, the FFA also appears to be involved in select aspects of social cognition. Studies employing tasks of the visual theory-of-mind type have now shown the FFA to be active during social judgments in the absence of any presentation of a face or a facelike object (e.g., Schultz et al., 2003). One interpretation of the FFA’s activity during social cognitive tasks is that it is reactivated in service of the semantic system and that the coding and storage of social semantic knowledge normally entail reactivation of perceptual representations. In this context, the FFA’s low activity level during face perception in individuals with ASD might, in part, reflect a paucity of social ideation in response to a face, as well as deficits in face perception. Amygdala-frontal connections. Beyond its role as modulator and interpreter of the emotional significance of data processed in the perceptual cortices, the amygdala has dense

south, ozonoff, and schultz: neurocognitive development in autism

709

Figure 43.2 A heuristic model of the relationship between the development of face perceptual skills and social skills as mediated by a system involving the amygdala and select cortical regions such as the fusiform gyrus and the superior temporal sulcus. The amygdala is hypothesized to have a critical role early in development,

while the cortical areas are believed to have a protracted development across childhood so as to allow the advanced computations characteristic of adult levels of face and facial expression recognition. (From Schultz, 2005.)

reciprocal connections to orbital and medial prefrontal cortices (PFC; see chapter 11 by Bauman and Amaral, this volume). These areas are known to subserve the processing of feelings and intentions, and the amygdala is thought to provide integration of emotion and cognition for decision making and action in the frontal cortices. The amygdala is known to be necessary for learning to associate sensory perceptions with reinforcers; contemporary models of its functions suggest that it adds an emotional tag to incoming sensory stimuli, which is important for triaging information in terms of motivational salience and linking the sensory information with prior knowledge and experience to formulate thoughts and actions. In autism samples, orbital and medial PFC areas have shown reduced fMRI activity during theory-of-mind tasks (Castelli et al., 2002), as well as reduced dopaminergic activity using positron emission topography (Ernst et al., 1997). Damage to the amygdala causes impairment in recognizing facial expressions, detecting social faux pas, judging trustworthiness, and attributing social intentions (see Adolphs et al., 2005).

and that its activation reflects stimulus intensity. Thus hypoactivation of the amygdala in autism may reflect nonspecific task effects. In this regard, reduced activity on fMRI in individuals with ASD might best be characterized as reflecting generalizable low levels of arousal to a broad range of stimuli, with the common thread being the stimuli’s importance to social perception. From this perspective, amygdala hypoarousal may be a principal source of low social motivation in autism, which in turn leads to a broad array of social perceptual and social deficits across the course of development. Although this model is likely an oversimplification, it does provide a structure for future studies, which are clearly needed for testing the precise role of these and other neural substrates of social motivation deficits in autism. One fruitful area of study will be to define the specificity of amygdala dysfunction with regard to social information processing. For example, a recent behavioral study of adolescents and adults with autism by South and colleagues (2008) found intact facilitation of emotion on a series of learning and memory tasks that did not include overtly social elements. South and colleagues suggested that amygdala function is disrupted in autism only for social information, perhaps as a result of atypical feedback from social perception areas such as the FFA. It will also be important to characterize the specificity of the “amygdala phenotype” for autism vis-à-vis other developmental psychopathologies: very similar models of patho-

Questions about the amygdala’s role in autism. It is not clear whether amygdala activation reflects the nature of the sensory inputs to this structure, whether the activation is reflective of computations by component nuclei, or whether it is a combination of these processes. One perspective is that the amygdala’s role in social-cognitive and perceptual processes might largely be one of mediating physiological arousal

710

neurodevelopmental aspects of clinical disorders

physiology have been suggested for anxiety, depression, psychopathy, bipolar disorder, substance abuse, schizophrenia, and posttraumatic stress disorder (PTSD). Identification of focused contributions to behavior by specific amygdala regions and pathways in autism may benefit from the use of multiple techniques, including neuroimaging, as well as from techniques used to study other disorders (such as startle paradigms used in PTSD and psychopathy). This in turn may be compatible with research to identify endophenotypes and associated genetic markers that correspond directly to subnuclei of the amygdala (Zirlinger and Anderson, 2003). Further development of animal models (e.g., Amaral and Corbett, 2003) will continue to provide valuable insight regarding the role of the amygdala in social development.

Repetitive behavior Although stereotyped, repetitive behavior has been identified as a hallmark symptom of the autism spectrum through all iterations of the major diagnostic systems, research in this area has historically been neglected (South, Ozonoff, and McMahon, 2005). There is currently little understanding of even the most basic issues of causality, function, maintenance, and treatment of such behaviors; nonetheless, progress in both imaging and genetics has opened the door for potentially significant contributions from expanded work in this area (Bodfish et al., 2000). A fundamental question related to possible endophenotypes of repetitive behavior is whether the disparate variety of stereotyped behaviors grouped together in autism spectrum disorders (including stereotyped motor movements, nonfunctional use of objects, rigid resistance to change, and the development of narrow, intense “circumscribed interests,” in addition to many obsessive-compulsive-like behaviors seen in autism) are part of a unitary factor or whether further subtyping would be useful (reviewed in South, Ozonoff, and McMahon, 2005). Because nonautistic forms of developmental disabilities, such as mental retardation, also entail stereotyped behaviors, it is important to determine similarities to and differences from autism in the developmental course of these stereotyped behaviors (Greaves et al., 2006). A third question is whether repetitive behaviors and impaired reciprocal social interaction share an underlying genetic basis or whether these are orthogonal factors with regard to heredity. With current data, the answer to this question hinges on whether one studies an autism population or the general population. In a study of children with an autism spectrum disorder, Constantino and colleagues (2004) presented cluster analyses from maternal-report instruments that suggest a “singular underlying factor of major effect” for autism symptoms (p. 724). In contrast, Ronald, Happe, and Plomin (2006), using parent and teacher data from a very large community sample

of twin pairs, not selected for autistic symptomatology, report only modest associations between deficits in social behavior and levels of stereotyped, repetitive behavior; Ronald, Happe, and Plomin suggest that both classes of behavior are highly heritable but that they share limited genetic overlap in the general population and therefore should be studied separately. It is possible that genetic mechanisms that normally moderate the development of social behavior and repetitive behaviors are distinct and separate, but that the genes that cause autism also create linkages between these phenotypic elements. Various repetitive-behavior subtypes have been proposed (e.g., “higher-order” versus “lower-order” behaviors related to degree of cognitive ability; Turner, 1999; CarcaniRathwell, Rabe-Hasketh, and Santosh, 2006), but the consistency of definition for the repetitive-behaviors category throughout all revisions of autism diagnostic criteria suggests that at least the clinical manifestations of these behaviors run together. For a group of high-functioning individuals with ASD, parent-report data demonstrated an internal consistency coefficient of .84 for a combined factor taken from the four DSM-IV repetitive-behavior symptoms (South, Ozonoff, and McMahon, 2005). This study also demonstrated a convergence of DSM-IV-based repetitive-behavior patterns between high-functioning autism and Asperger syndrome as children moved from childhood into adolescence. Prospective and longitudinal studies will help determine the developmental course of repetitive behavior symptoms with the goal of more closely linking behavior to the underlying brain mechanisms. Brain Mechanisms There are few published studies to date that have investigated brain models of repetitive behaviors in autism. Collins and colleagues (2006) have reported several studies linking repetitive behaviors in autism with GABA receptor genes on chromosome 4, perhaps mediated by the presence or absence of a family history of seizure activity; there are also candidate genes in GABArelated genes on chromosomes 15, 7, and 2. Collins and colleagues suggest a link between repetitive behavior and reduced inhibitory activity in the brain resulting from atypical GABAergic function; GABA appears in mouse models to also be related to hypersensitivities as well as motor stereotypies (reviewed in Belmonte et al., 2004). A recent study utilizing symptom-rating scales in conjunction with MRI found a significant negative correlation between narrow interest patterns and amygdala volume in adults with Asperger syndrome (Dziobek et al., 2006). Studies of both drug and environmental mediation of stereotyped motor movements in deer mice have suggested altered corticobasal ganglia circuitry similar to models of Tourette syndrome and obsessive compulsive disorders (Lewis, 2004), in specific nigrostriatal dopaminergic pathways. Grossman and

south, ozonoff, and schultz: neurocognitive development in autism

711

Verobyev (1998) conclude that multiple neurochemical systems necessarily contribute to stereotyped behavior. Progress in understanding the phenotype of repetitive behaviors in autism will be essential to improved characterization of the neurobiology, but the reverse may also hold true.

Conclusions Cognitive neuroscience research regarding the etiological bases of autism has expanded exponentially since the last edition of this Handbook. The recognition that the behavioral symptoms of autism emerge through numerous genotypic pathways has advanced the field from a focus on categorical definitions to a search for observable endophenotypes that link behaviors with underlying biological mechanisms. Whereas the previous Handbook chapter described the neural basis of autism in terms of a few major theories with origins in higher-order cognitive functions (e.g., executive functions and theory of mind), this chapter summarizes a wide variety of techniques and theories that are in large part dedicated to the search for endophenotypes at increasingly narrow levels (e.g., in vivo neurochemical spectroscopy; Friedman et al., 2006) and in increasingly narrower samples (e.g., the rare variant genetics approach to developmental psychopathology; Abelson et al., 2005). Yet this chapter is not by any means comprehensive, and it seems the more we learn, the more questions we encounter. For instance, although the finding of hypoactivation of the fusiform face area while viewing faces is the most replicated marker of neurophysiological atypicality in autism, new questions have arisen regarding task-specific effects, cross-modal connectivity, and the role of anxiety in gaze avoidance. Conceptually, it is essential to consider how any number of etiological pathways may converge in neurodevelopment to comprise the behaviorally defined autism spectrum disorders. What are the neural organizing principles and shared pathways that lead to a group of behaviors that can be defined by the common term “autism”? Such questions ensure a rich prospect for research on autism for the foreseeable future. REFERENCES Abelson, J. F., K. Y. Kwan, B. J. O’Roak, D. Y. Baek, A. A. Stillman, T. M. Morgan, C. A. Matthews, D. L. Pauls, M.-R. Rasin, M. Gunel, et al., 2005. Sequence variants in SLITRK1 are associated with Tourette’s syndrome. Science 310:317– 320. Adolphs, R., F. Gosselin, T. W. Buchanan, D. Tranel, P. Schyns, and A. R. Damasio, 2005. A mechanism for impaired fear recognition after amygdala damage. Nature 433:68–72. Amaral, D. G., and B. A. Corbett, 2003. The amygdala, autism and anxiety. Novartis Foundation Symposium, 251:177.

712

American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders, 4th ed.—text revision. Washington, DC: Author. Anderson, G. M., and Y. Hoshino, 2005. Neurochemical studies of autism. In F. R. Volkmar, R. Paul, A. Klin, and D. Cohen, eds., Handbook of Autism and Developmental Disorders, 3rd ed., 453– 472. New York: Wiley. Arndt, T. L., C. J. Stodgell, and P. M. Rodier, 2005. The teratology of autism. Int. J. Dev. Neurosci. 23 (2–3):189–199. Aylward, E. H., N. J. Minshew, K. Field, B. F. Sparks, and N. Singh, 2002. Effects of age on brain volume and head circumference in autism. Neurology 59:175–183. Bailey, A., A. Le Couteur, I. Gottesman, P. Bolton, E. Simonoff, E. Yuzda, et al., 1995. Autism as a strongly genetic disorder: Evidence from a British twin study. Psychol. Med. 25(1):63–77. Bailey, A., P. Luther, A. Dean, B. Harding, I. Janota, M. Montgomery, M. Rutter, and P. Lantos, 1998. A clinicopathological study of autism. Brain 121:889–905. Barnea-Goraly, N., H. Kwon, V. Menon, S. Eliez, L. Lotspeich, and A. L. Reiss, 2004. White matter structure in autism: Preliminary evidence from diffusion tensor imaging. Biol. Psychiatry 55:323–326. Baron-Cohen, S., 1989. The autistic child’s theory of mind: A case of specific developmental delay. J. Child Psychol. Psychiatry 30(2):285–297. Baron-Cohen, S., 1995. Mind Blindness. Cambridge, MA: MIT Press. Baron-Cohen, S., T. Jolliffe, C. Mortimore, and M. Robertson, 1997. Another advanced test of theory of mind: Evidence from very high functioning adults with autism or Asperger syndrome. J. Child Psychol. Psychiatry 38(7):813–822. Baron-Cohen, S., A. M. Leslie, and U. Frith, 1985. Does the autistic child have a “theory of mind”? Cognition 21:37–46. Baron-Cohen, S., H. A. Ring, E. T. Bullmore, S. Wheelwright, C. Ashwin, and S. C. R. Williams, 2000. The amygdala theory of autism. Neurosci. Biobehav. Rev. 24:355–364. Belmonte, M. K., E. H. Cook, Jr., G. M. Anderson, J. L. R. Rubenstein, W. T. Greenough, A. Beckel-Mitchener, E. Courchesne, L. M. Boulanger, S. B. Powell, P. R. Levitt, et al., 2004. Autism as a disorder of neural information processing: Directions for research and targets for therapy. Mol. Psychiatry 9:646–663. Benayed, R., N. Gharani, I. Rossman, V. Mancuso, G. Lazar, S. Kamdar, S. E. Bruse, S. Tishfield, B. J. Smith, R. A. Zimmerman, et al., 2005. Support for the homeobox transcription factor gene Engrailed 2 as an autism spectrum disorder susceptibility locus. Am. J. Hum. Genet. 77:851–868. Bishop, D. V., M. Mayberry, D. Wong, A. Mayley, and J. Hallmayer, 2006. Characteristics of the broader phenotype in autism: A study of siblings using the children’s communication checklist-2. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 141:117–122. Boddaert, N., N. Chabane, H. Gervais, C. D. Good, M. Bourgeois, M. H. Plumet, C. Barthelemy, M. C. Mouren, E. Artiges, Y. Samson, F. Brunelle, R. S. Frackowiak, and M. Zilbovicius, 2004. Superior temporal sulcus anatomical abnormalities in childhood autism: A voxel-based morphometry MRI study. Neuroimage 23(1):364–369. Bodfish, J. W., F. J. Symons, D. E. Parker, and M. H. Lewis, 2000. Varieties of repetitive behavior in autism: Comparisons to mental retardation. J. Autism Dev. Disord. 30:237–243. Bölte, S., and F. Poustka, 2006. The broader cognitive phenotype of autism in parents: How specific is the tendency for local

neurodevelopmental aspects of clinical disorders

processing and executive dysfunction? J. Child Psychol. Psychiatry 47(6):639–645. Bolton, P. F., M. Roobol, L. Allsopp, and A. Pickles, 2001. Association between idiopathic infantile macrocephaly and autism spectrum disorders. Lancet 358(9283):726–727. Carcani-Rathwell, I., S. Rabe-Hasketh, and P. J. Santosh, 2006. Repetitive and stereotyped behaviours in pervasive developmental disorders. J. Child Psychol. Psychiatry 47:573–581. Caron, R. F., A. J. Caron, J. Roberts, and R. Brooks, 1997. Infant sensitivity to deviation in dynamic facial-vocal displays: The role of eye regard. Dev. Psychol. 33:802–813. Casanova, M. F., D. P. Buxhoeveden, and C. Brown, 2002. Clinical and macroscopic correlates of minicolumnar pathology in autism. J. Child Neurol. 17(9):692–695. Castelli, F., C. Frith, F. Happe, and U. Frith, 2002. Autism, Asperger syndrome and brain mechanisms for the attribution of mental states to animated shapes. Brain 125:1839– 1849. Chapman, L. J., and J. P. Chapman, 1973. Disordered Thought in Schizophrenia. New York: Appleton Century Crofts. Collins, A. L., D. Ma, P. L. Whitehead, E. R. Martin, H. H. Wright, R. K. Abramson, J. P. Hussman, J. L. Haines, M. L. Cuccaro, J. R. Gilbert, et al., 2006. Investigation of autism and GABA receptor subunit genes in multiple ethnic groups. Neurogenetics 7:167–174. Constantino, J. N., C. P. Gruber, S. Davis, S. Hayes, N. Passanante, and T. Przybeck, 2004. The factor structure of autistic traits. J. Child Psychol. Psychiatry 45:719–726. Constantino, J. N., C. Lajonchere, M. Lutz, T. Gray, A. Abbacchi, K. McKenna, D. Singh, and R. D. Todd, 2006. Autistic social impairment in the siblings of children with pervasive developmental disorders. Am. J. Psychiatry 163:294–296. Coon, H., 2006. Current perspectives on the genetic analysis of autism. Am. J. Med. Genet.(Sem. Med. Genet.) 142:24–32. Courchesne, E., 2002. Abnormal early brain development in autism. Mol. Psychiatry 7:S21–23. Courchesne, E., R. Carper, and N. Akshoomoff, 2003. Evidence of brain overgrowth in the first year of life in autism. JAMA 290(3):337–344. Courchesne, E., C. M. Karns, H. R. Davis, R. Ziccardi, R. A. Carper, Z. D. Tigue, H. J. Chisum, P. Moses, K. Pierce, C. Lord, et al., 2001. Unusual brain growth patterns in early life in patients with autistic disorder: An MRI study. Neurology 57:245–254. Crawley, J. N., 2004. Designing mouse behavioral tasks relevant to autistic-like behaviors. Ment. Retard. Dev. Disabil. Res. Rev. 10:248–258. Dalton, K. M., B. M. Nacewicz, T. Johnstone, H. S. Schaefer, M. A. Gernsbacher, H. H. Goldsmith, et al., 2005. Gaze fixation and the neural circuitry of face processing in autism. Nature Neurosci. 8:519–526. Dapretto, M., M. S. Davies, J. H. Pfeifer, A. A. Scott, M. Sigman, S. Y. Bookheimer, and M. Iacoboni, 2006. Understanding emotions in others: Mirror neuron dysfunction in children with autism spectrum disorders. Nature Neurosci. 9:28–30. Dawson, G., S. J. Webb, and J. McPartland, 2005. Understanding the nature of face processing impairment in autism: Insights from behavioral and electrophysiological studies. Dev. Psychopathol. 27:403–424. Dawson, G., S. J. Webb, E. Wijsman, G. Schellenberg, A. Estes, I. Munson, and S. Faja, 2005. Neurocognitive and electrophysiological evidence of altered face processing in parents of children

with autism: Implications for a model of abnormal development of social brain circuitry in autism. Dev. Psychopathol. 17: 679–697. Deruelle, C., C. Rondan, B. Gepner, and C. Tardif, 2004. Spatial frequency and face processing in children with autism and Asperger syndrome. J. Autism Dev. Disord. 34:199–210. DiCicco-Bloom, E., C. Lord, L. Zwaigenbaum, E. Courchesne, S. R. Dager, C. Schmitz, R. T. Schultz, J. Crawley, and L. J. Young, 2006. The developmental neurobiology of autism spectrum disorder. J. Neurosci. 26:6897–6906. Dziobek, I., S. Fleck, K. Rogers, O. T. Wolf, and A. Convit, 2006. The “amygdala theory of autism” revisited: Linking structure to behavior. Neuropsychologia 44:1891–1899. Ernst, M., A. J. Zametkin, J. A. Matochik, D. Pascualvaca, and R. M. Cohen, 1997. Reduced medial prefrontal dopaminergic activity in autistic children. Lancet 350:638. Frith, U., 1989. Autism: Explaining the Enigma. London: Basil Blackwell. Friedman, S. D., D. W. Shaw, A. A. Artruu, G. Dawson, H. Petropolous, and S. R. Dager, 2006. Gray and white matter chemistry in young children with autism. Arch. Gen. Psychiatry 63:786–794. Ghaziuddin, M., 2005. A family history study of Asperger Syndrome. J. Autism Dev. Disord. 35(2):177–182. Gottesman, I. I., and T. G. Gould, 2003. The endophenotypes concept in psychiatry: Etymology and strategic intentions. Am. J. Psychiatry 160:636–645. Greaves, N., E. Prince, D. W. Evans, and T. Charman, 2006. Repetitive and ritualistic behaviour in children with Prader-Willi syndrome and children with autism. J. Intellectual Disabil. Res. 50:92–100. Grelotti, D., I. Gauthier, and R. T. Schultz, 2002. Social interest and the development of cortical face specialization: What autism teaches us about face processing. Dev. Psychobiol. 40: 213–225. Griffith, E. M., B. F. Pennington, E. A. Wehner, and S. J. Rogers, 1999. Executive functions in young children with autism. Child Dev. 70:817–832. Grossman, R., and L. Verobyev, 1998. The neurobiology of stereotypic movement behaviors and SMD. Psychiatr. Ann. 28: 317–323. Hardan, A. Y., N. J. Minshew, M. S. Keshavan, 2000. Corpus callosum size in autism. Neurology 55(7):1033–1036. Herbert, M. R., D. A. Ziegler, N. Makris, P. A. Filipek, T. L. Kemper, J. J. Normandin, et al., 2004. Localization of white matter volume increase in autism and developmental language disorder. Ann. Neurol. 55(4):530–540. Hobson, R. P., and M. Bishop, 2003. The pathogenesis of autism: Insights from congenital blindness. Philos. Trans. R. Soc. Lond. B [Biol.] 358(1430):335–344. Hobson, R. P., A. Lee, and R. Brown, 1999. Autism and congenital blindness. J. Autism Dev. Disord. 29(1):45–56. Hughes, C., M. H. Plumet, and M. Leboyer, 1999. Towards a cognitive phenotype for autism: Increased prevalence of executive dysfunction and superior spatial span amongst siblings of children with autism. J. Child Psychol. Psychiatry 40(5):705–718. Insel, T. R., and R. D. Fernald, 2004. How the brain processes social information: Searching for the social brain. Ann. Rev. Neurosci. 27:697–722. Kanner, L., 1943. Autistic disturbances of affective contact. Nervous Child 2:217–253. Kanwisher, N., J. McDermott, and M. M. Chun, 1997. The fusiform face area: A module in human extrastriate cortex

south, ozonoff, and schultz: neurocognitive development in autism

713

specialized for the perception of faces. J. Neurosci. 17:4302– 4311. Kemper, T. L., and M. L. Bauman, 2002. Neuropathology of infantile autism. Mol. Psychiatry 7(Suppl 2):S12–13. Kim, S. J., L. J. Young, D. Gonen, J. Veendra-Vanderweele, R. Courchesne, E. Courchesne, C. Lord, B. L. Leventhal, E. H. Cook, Jr., and T. R. Insel, 2002. Transmission disequilibrium testing of arginine surpressor receptor 1A (AVPR1A) polymorphisms in autism. Mol. Psychiatry 7:503–507. Klin, A., W. Jones, R. Schultz, F. Volkmar, and D. Cohen, 2002. Visual fixation patterns during viewing of naturalistic social situations as predictors of social competence in individuals with autism. Arch. Gen. Psychiatry 59:809–816. Klin, A., F. R. Volkmar, and S. S. Sparrow, 1992. Autistic social dysfunction: Some limitations of the theory of mind hypothesis. J. Child Psychol. Psychiatry 33:861–876. Kolevzon, A., K. A. Matthewson, and E. Hollander, 2006. Selective serotonin reuptake inhibitors in autism: A review of efficacy and tolerability. J. Clin. Psychiatry 67:407–414. Lainhart, J. E., 1999. Psychiatric problems in individuals with autism, their parents and siblings. Int. Rev. Psychiatry 11: 278–298. Lainhart, J. E., 2006. Advances in autism neuroimaging research for the clinician and geneticist. Am. J. Med. Genet. (Sem. Med. Genet.) 142:33–39. Le Couteur, A. Bailey, S. Goode, A. Pickles, A. Robertson, I. Gottesman, and M. Rutter, 2006. A broader phenotype of autism: The clinical spectrum in twins. J. Child Psychol. Psychiatry 37:785–801. Lewis, M. H., 2004. Environmental complexity and central nervous system development and function. Ment. Retard. Dev. Disabil. Res. 10:91–95. Libbey, J. E., T. L. Sweeten, W. M. McMahon, and R. S. Fujinami, 2005. Autistic disorder and viral infections. J. Neurovirol. 11:1–10. Lim, M. M., I. F. Bielsky, and L. J. Young, 2005. Neuropeptides and the social brain: Potential rodent models of autism. Int. J. Dev. Neurosci. 23:235–243. Lord, C., C. Shulman, and P. DiLavore, 2004. Regression and word loss in autistic spectrum disorders. J. Child Psychol. Psychiatry 45(5):936–955. Mandell, D. S., M. M. Novak, and C. D. Zubritsky, 2005. Factors associated with age of diagnosis among children with autism spectrum disorders. Pediatrics 116:1480–1486. Merin, N., G. S. Young, S. Ozonoff, and S. J. Rogers, 2007. Visual fixation patterns during reciprocal social interaction distinguish a subgroup of 6-month-old infants at risk for autism from comparison infants. J. Autism Dev. Disord. 37(1):108– 121. Miles, J. H., and R. E. Hillman, 2000. Value of a clinical morphology examination in autism. Am. J. Med. Genet. 91: 245–253. Miller, M. T., K. Stromland, L. Ventura, M. Johansson, J. M. Bandim, and C. Gillberg, 2005. Autism associated with conditions characterized by developmental errors in early embryogenesis. Int. J. Dev. Neurosci. 23:201–219. Minshew, N. J., G. Goldstein, and D. J. Siegel, 1997. Neuropsychologic functioning in autism: Profile of a complex information processing disorder. J. Int. Neuropsychol. Soc. 3:303–316. Miyazaki, K., M. Narita, and N. Narita, 2005. Maternal administration of thalidomide or valproic acid causes abnormal serotonergic neurons in the offspring: Implication for pathogenesis of autism. Int. J. Dev. Neurosci. 23:287–297.

714

Mulder, E. J., G. M. Anderson, I. P. Kema, A. de Bildt, N. D. van Lang, J. A. den Boer, and R. B. Minderaa, 2004. Platelet serotonin levels in pervasive developmental disorders and mental retardation: Diagnostic group differences, within-group distribution, and behavioral correlates. J. Am. Acad. Child Adolesc. Psychiatry 43:491–499. Narayan, M., S. Srinath, G. M. Anderson, and D. B. Meundi, 1993. Cerebrospinal fluid levels of homovanillic acid and 5-hydroxyindoleacetic acid in autism. Biol. Psychiatry 33:630– 635. Nelson, K. B., J. K. Grether, L. A. Croen, J. M. Dambrosia, B. F. Dickens, L. L. Jelliffe, R. L. Hansen, and T. M. Phillips, 2001. Neuropeptides and neurotrophins in neonatal blood of children with autism or mental retardation. Ann. Neurol. 49:597–606. Osterling, J., and G. Dawson, 1994. Early recognition of children with autism: A study of first birthday home videotapes. J. Autism Dev. Disord. 24(3):247–257. Ozonoff, S., M. South, and S. Provencal, 2005. Executive functions. In F. R. Volkmar, R. Paul, A. Klin, and D. Cohen, eds., Handbook of Autism and Developmental Disorders, 3rd ed., 606–627. New York: Wiley. Parker, S. K., B. Schwartz, J. Todd, and L. K. Pickering, 2004. Thimerosal-containing vaccines and autistic spectrum disorder: A critical review of published original data. Pediatrics 114(3): 793–804. Pelphrey, K. A., J. P. Morris, C. R. Michelich, T. Allison, and G. McCarthy, 2005. Functional anatomy of biological motion perception in posterior temporal cortex: An fMRI study of eye, mouth, and hand movements. Cerebral Cortex 15(12):1866– 1876. Pennington, B. F., 2002. The Development of Psychopathology: Nature and Nurture. New York: Guilford. Pennington, B. F., and S. Ozonoff, 1996. Executive functions in developmental psychopathology. J. Child Psychol. Psychiatry 37:51–87. Piven, J., S. Arndt, J. Bailey, S. Havercamp, N. Andreasen, and P. Palmer, 1995. An MRI study of brain size in autism. Am. J. Psychiatry 152:1145–1149. Piven, J., J. Bailey, B. J. Ranson, and S. Arndt, 1997. An MRI study of the corpus callosum in autism. Am. J. Psychiatry 154(8):1051–1056. Piven, J., P. Palmer, D. Jacobi, D. Childress, and S. Arndt, 1997. Broader autism phenotype: Evidence from a family history study of multiple-incidence autism families. Am. J. Psychiatry 154(2):185–190. Rice, S. A., E. D. Bigler, H. B. Cleavinger, D. F. Tate, J. Sayer, W. McMahon, et al., 2005. Macrocephaly, corpus callosum morphology, and autism. J. Child Neurol. 20(1):34– 41. Ringo, J. L., 1991. Neuronal interconnection as a function of brain size. Brain Behav. Evol. 38(1):1–6. Robins, D. L., E. T. Hunyadi, and R. T. Schultz, 2005. Perception of emotion through facial expression and tone of voice in autism spectrum disorders: An fMRI study. Paper presented at the International Meeting for Autism Research, Boston, May. Rolls, E. T., 1999. The Brain and Emotion. Oxford, UK: Oxford University Press. Ronald, A., F. Happe, and R. Plomin, 2006. Genetic research into autism. Science 311:952. Rutter, M., 2000. Genetic studies of autism: From the 1970s into the millennium. J. Abnorm. Child Psychol. 28(1):3–14.

neurodevelopmental aspects of clinical disorders

Rutter, M., 2005a. Genetic influences of autism. In F. R. Volkmar, R. Paul, A. Klin, and D. Cohen, eds., Handbook of Autism and Developmental Disorders, 3rd ed., 425–452. New York: Wiley. Rutter, M., 2005b. Incidence of autism spectrum disorders: Changes over time and their meaning. Acta Paediatr. 94:2–15. Samson, F., L. Mottron, B. Jemel, P. Belin, and V. Ciocca, 2006. Can spectro-temporal complexity explain the autistic pattern of performance on auditory tasks? J. Autism Dev. Disord. 36:65– 76. Schmitz, N., K. Rubia, E. Daly, A. Smith, S. Williams, and D. G. Murphy, 2006. Neural correlates of executive function in autistic spectrum disorders. Biol. Psychiatry 59:7–16. Schultz, R. T., 2005. Developmental deficits in social perception in autism: The role of the amygdala and fusiform face area. Int. J. Dev. Neurosci. 23:125–141. Schultz, R. T., I. Gauthier, A. Klin, R. K. Fulbright, A. W. Anderson, F. R. Volkmar, P. Skudlarski, C. Lacadie, D. J. Cohen, and J. C. Gore, 2000. Abnormal ventral temporal cortical activity during face discrimination among individuals with autism and Asperger syndrome. Arch. Gen. Psychiatry 57:331– 340. Schultz, R. T., D. J. Grelotti, A. Klin, J. Kleinman, C. Van der Gaag, R. Marois, et al., 2003. The role of the fusiform face area in social cognition: Implications for the pathobiology of autism. Philos. Trans. R. Soc. Lond. B [Biol.] 358(1430):415–427. Schultz, R.T., and D. L. Robins, 2005. Functional neuroimaging studies of autism spectrum disorders. In F. R. Volkmar, A. Klin, and R. Paul, eds. Handbook of Autism and Pervasive Developmental Disorders, 3rd ed., 515–533. New York: Wiley. Schultz, R. T., L. M. Romanski, and K. D. Tsatsanis, 2000. Neurofunctional models of autistic disorder and Asperger syndrome: Clues from neuroimaging. In A. Klin, F. R. Volkmar, and S. S. Sparrow, eds., Asperger Syndrome, 172–209. New York: Guilford. Schultz, R. T., L. Win, A. Jackowski, A. Klin, L. Staib, X. Papademetris, T. Babitz, E. Carter, C. Klaiman, A. Feiler, and F. Volkmar, 2005. Brain morphology in autism spectrum disorders: An MRI study. Presentation at the annual International Meeting for Autism Research (IMFAR), Boston, May 5–7. Schumann, C. M., J. Hamstra, B. L. Goodlin-Jones, L. J. Lotspeich, H. Kwon, M. H. Buonocore, et al., 2004. The amygdala is enlarged in children but not adolescents with autism; the hippocampus is enlarged at all ages. J. Neurosci. 24:6392– 6401. Scott, M. M., and E. S. Deneris, 2005. Making and breaking serotonin neurons and autism. Int. J. Dev. Neurosci. 23:277– 285. Shah, A., and U. Frith, 1993. Why do autistic individuals show superior performance on the block design task? J. Child Psychol. Psychiatry 34:1351–1364. South, M., S. Ozonoff, and W. M. McMahon, 2005. Repetitive behavior profiles in Asperger syndrome and high-functioning autism. J. Autism Dev. Disord. 35:145–153. South, M., S. Ozonoff, Y. Suchy, R. P. Kesner, W. M. McMahon, and J. E. Lainhart, 2008. Intact emotion facilitation for nonsocial stimuli in autism: Is amygdala impairment in

autism specific for social information? J. Int. Neuropsychol. Soc. 14(1):42–54. Steg, J. P., and J. L. Rapoport, 1975. Minor physical anomalies in normal, neurotic, learning disabled and severely neurotic children. J. Autism Dev. Disord. 5:299–307. Stodgell, C. J., J. Ingram, S. Gnall, and P. M. Rodier, 2001. In utero valproic acid exposure alters Hoxa1 expression in the rat embryo: Mechanism of teratogenicity and relationship to autism. Teratology 63:247. Sutcliffe, J. S., R. J. Delahanty, H. C. Prasad, J. L. McCauley, Q. Han, L. Jiang, C. Li, S. E. Folstein, and R. D. Blakely, 2005. Allelic heterogeneity at the serotonin transporter locus (SLC6A4) confers susceptibility to autism and rigid-compulsive behaviors. Am. J. Hum. Genet. 77:265–269. Tager-Flusberg, H., R. Paul, and C. Lord, 2005. Language and communication in autism. In F. R. Volkmar, R. Paul, A. Klin, and D. Cohen, eds., Handbook of Autism and Developmental Disorders, 3rd ed., 335–364. New York: Wiley. Taylor, B., 2006. Vaccines and the changing epidemiology of autism. Child Care Health Dev. 32(5):511-519. Teunisse, J., and B. de Gelder, 2003. Face processing in adolescents with autistic disorder: The inversion and composite effects. Brain Cogn. 52:285–294. Toth, K., J. N. Munson, A. Meltzoff, and G. Dawson, 2006. Early predictors of communication development in young children with autism spectrum disorder: Joint attention, imitation, and toy play. J. Autism Dev. Disord. 36(8):993– 1005. Turner, M. A., 1999. Annotation: Repetitive behavior in autism: A review of psychological research. J. Child Psychol. Psychiatry 40:839–849. Veenstra-VanderWeele, J., S. L. Christian, and E. H. Cook, 2004. Autism as a paradigmatic complex genetic disorder. Annu. Rev. Genomics Hum. Genet. 5:379–405. Vuilleumier, P., M. P. Richardson, J. L. Armony, J. Driver, and R. J. Dolan, 2004. Distant influences of amygdala lesion on visual cortical activation during emotional face processing. Nature Neurosci. 7(11):1271–1278. Waiter, G. D., J. H. Williams, A. D. Murray, A. Gilchrist, D. I. Perrett, and A. Whiten, 2004. A voxel-based investigation of brain structure in male adolescents with autistic spectrum disorder. Neuroimage 22:619–625. Walker, H. A., 1977. Incidence of minor physical anomaly in autism. J. Autism Childhood Schizophr. 7(2):165–176. Willemsen-Swinkels, S. H., and J. K. Buitelaar, 2002. The autistic spectrum: Subgroups, boundaries, and treatment. Psychiatr. Clin. North Am. 25:811–836. Williams, J. H., G. D. Waiter, A. Gilchrist, D. I. Perrett, A. D. Murray, and A. Whiten, 2006. Neural mechanisms of imitation and “mirror neuron” functioning in autistic spectrum disorder. Neuropsychologia 44:610–621. Zirlinger, M., and D. Anderson, 2003. Molecular dissection of the amygdala and its relevance to autism. Genes Brain Behav. 2:282–294. Zwaigenbaum, L., S. Bryson, T. Rogers, W. Roberts, J. Brian, and P. Szatmari, 2005. Behavioral manifestations of autism in the first year of life. Int. J. Dev. Neurosci. 23:143– 152.

south, ozonoff, and schultz: neurocognitive development in autism

715

44

Tics and Compulsions: Disturbances of Self-Regulatory Control in the Development of Habitual Behaviors RACHEL MARSH, JAMES F. LECKMAN, MICHAEL H. BLOCH, YANKI YAZGAN, AND BRADLEY S. PETERSON

Tics and compulsions lie on a continuum of semiinvoluntary or “habitual” behaviors that are released from self-regulatory control. The conceptual construct of selfregulatory control encompasses many aspects of cognitive and psychosocial functioning. Individuals must engage selfregulatory control to organize their many perceptions, feelings, memories, and thoughts during the planning, execution, and monitoring of any goal-directed behavior in the context of competing urges, desires, or situational demands (Peterson, 2005; Posner and Rothbart, 2000; Tucker, Luu, and Pribram, 1995). Self-regulatory control is present in virtually every action we perform, in that choosing to execute one action always necessitates not choosing or inhibiting another. Learning to control or inhibit behaviors that violate personal or societal norms is one of the centrally defining characteristics of normal child development. Disturbances in the maturation of neural processes that subserve self-regulatory control likely contribute to the development of a variety of psychiatric disorders in which children have difficulty controlling their cognitive, affective, and motoric behaviors. These disturbances may release from regulatory control an underlying impulse to move or perform some kind of compulsory behavior, or to act on various appetitive or aggressive drives. Children with impaired regulation or control of impulses may begin performing some of these behaviors at a young age. The continued presence of self-regulatory impairments may allow these impulsive behaviors to crystallize into habits and, eventually, into a full-fledged disorder. Understanding the neurobiological origins of these disturbances in development can enhance our understanding of normal brain functioning and maturation. Tourette’s syndrome (TS) is a neuropsychiatric disorder of childhood onset that is characterized by the presence of motor and vocal tics that wax and wane in severity (American Psychiatric Association, 1987). Tics are typically brief,

nonpurposeful or semipurposeful fragments of behaviors that are often responses to stimuli or environmental cues either from within the body or from the outside world ( J. Leckman and Riddle, 2000). Sensitivity to these cues is usually experienced as a compulsory urge that is only relieved by the performance of a tic (Leckman and Riddle, 2000; Leckman, 2002; Peterson and Klein, 1997). These urges, and the patient’s preoccupation with them, bear a phenomenological resemblance to the obsessional urges to act that typically precede compulsive behaviors. In fact, patients with TS are often affected with comorbid obsessive-compulsive disorder (OCD). Obsessive-compulsive disorder is a neuropsychiatric disorder of either childhood or adult onset. It is characterized by recurrent, distressing, and intrusive thoughts, ideas, or images (obsessions) together with their repetitive behavioral counterparts (compulsions) (American Psychiatric Association, 1987). Several large factor-analytic studies have confirmed the presence of at least four components to OCD symptoms: (1) aggressive, sexual, religious, and somatic obsessions, and checking compulsions; (2) symmetry and ordering; (3) cleanliness and washing; and (4) hoarding (Baer, 1994; Leckman, Grice, et al., 1997). Evidence from familygenetic and twin studies indicates that TS and OCD are genetically related (Pauls, Towbin, et al., 1986; Peterson and Klein, 1997), and neuroimaging studies suggest that the neural bases of TS and OCD are related as well (Leckman, Peterson, Anderson, et al., 1997). Both tics and compulsions can be suppressed voluntarily (Peterson, Skudlarski, et al., 1998), but the internal struggle to control the urges to tic or to perform a compulsion is often as debilitating as the tic or obsession itself (Evans, Lewis, and Iobst, 2004; Leckman, 2003). The phenomenological similarities between tics and compulsions, and their common genetic and neural basis in childhood, suggest that TS and pediatric OCD may be manifestations of the same underlying disease process. Both

717

tics and compulsions clinically present as “habits” that have escaped regulatory control. Neuroimaging studies increasingly suggest that the neural basis for both TS and OCD resides in anatomical and functional disturbances of cortical-striatal-thalamic-cortical (CSTC) circuits. These circuits comprise multiple, partially overlapping but largely “parallel” pathways that direct information from the cerebral cortex to the subcortex, and then back again to specific regions of the cortex (Alexander, Crutcher, and DeLong, 1990; Goldman-Rakic and Selemon, 1990; Parent and Hazrati, 1995). Although the number of anatomically and functionally discrete pathways is still controversial, current consensus holds that CSTC circuitry has at least four components—those initiating from and projecting back to sensorimotor cortex, orbitofrontal cortex (OFC), limbic and associated anterior cingulate cortices, or association cortices. The motor portions of CSTC circuits are believed to be primarily involved in the pathophysiology of TS; and OFC portions, in the pathophysiology of OCD (Peterson and Klein, 1997; Peterson, 2000; Peterson, Leckman, et al., 1999; Peterson and Thomas, 2000). Abundant clinical and preclinical evidence suggests that CSTC circuits subserve self-regulatory functions (E. Miller and Cohen, 2001; Pasupathy and Miller, 2005). The prefrontal portions of these circuits are thought to regulate or engage top-down control over memory and learning functions that are based with subcortical structures (Wise, Murray, and Gerfen, 1996). The learning of procedures, stimulus-response (S-R) associations, or “habits” is mediated by the striatum (Knowlton, Mangels, and Squire, 1996; Knowlton, Squire, and Gluck, 1994; Packard and Knowlton, 2002; Squire and Kandel, 1999; Squire and Zola, 1996). Morphological studies suggest reduced volume of the caudate nucleus in children and adults with TS (Peterson, Riddle, et al., 1993; Peterson, Staib, et al., 2001; Peterson, Thomas, et al., 2003) and in adults with OCD (Rosenberg, Keshavan, O’Hearn, et al., 1997). The cellular composition of the caudate nucleus also appears to be abnormal in some adults with severe TS (Kalanithi et al., 2005). We presume that dysfunction of frontostriatal regulatory systems exemplifies the core pathophysiologic disturbances in TS and OCD. These disturbances may then interact with habit-learning systems within the striatum, contributing to the habitual nature of tics and compulsions. This review will focus on disturbances in frontostriatal neural systems that may contribute to the development and perpetuation of tics and compulsions. First, we will present a brief overview of the natural history of TS and OCD, highlighting their comorbidity in childhood and adolescence. We will then review neuroimaging and cognitive neuroscience studies of self-regulatory control processes and habit learning in these disorders. We will attempt to synthesize findings that suggest the presence of functional dis-

718

turbances in these systems in both children and adults with TS or OCD. Finally, we will speculate on how disturbances in the maturation of frontostriatal systems that subserve self-regulatory control and habit learning likely contribute to the development, persistence, and severity of tics and compulsions.

Natural histories of TS and OCD The modal age of onset of tics is 6 years. Tics affect 10–20 percent of children at some time in their life, with a ratio of boys to girls of approximately 3 or 4 to 1 (Achenbach and Edelbrock, 1981; Costello et al., 1996; Nomoto and Machiyama, 1990; Peterson, Pine, et al., 2001; Verhulst, Akkerhuis, and Althaus, 1985). Tics most commonly begin at a low frequency and with minimal forcefulness, so that parents often regard them as habits or unusual mannerisms. In the majority of children, tics disappear in a matter of weeks to months. Roughly 1 percent of all children will have “chronic” tics that, by definition, persist for more than a year. No phenomenological, natural history, or neurobiological evidence exists to suggest that transient tics, chronic tics, or TS differ from each other in any way other than their duration or the declaration that TS comprises both motor and vocal tics (Peterson, Pine, et al., 2001). Family-genetic and twin studies, in fact, suggest that TS and chronic tic disorders represent continua of the same underlying genetic diathesis (Pauls and Leckman, 1986; Price et al., 1985). Chronic tics gradually increase in number, frequency, and forcefulness until they peak in severity at age 11 or 12. They then decline gradually in severity throughout adolescence (Leckman, Zhang, et al., 1998). At all ages, however, tic severity tends to fluctuate unpredictably over minutes, hours, days, and weeks throughout childhood and adolescence. By age 18 years, tic symptoms are substantially reduced in roughly 90 percent of patients with TS, and more than 40 percent are symptom free (Bloch et al., 2005; Burd et al., 2001; Leckman, Zhang, et al., 1998). For many individuals with TS, obsessive thoughts and compulsive rituals emerge during late childhood or early adolescence, several years after the early onset of motor tics (Leckman, Walker, et al., 1994). The age of onset of OCD in the general community is likely bimodal, with one mode of onset at 10–12 years of age and the other in early adulthood (Berg et al., 1989; Rasmussen and Tsuang, 1986; Valleni-Basile et al., 1996). The childhood-onset form of OCD most commonly occurs in the context of a personal history of a tic disorder, and it occurs even more commonly in the context of a personal or family history of tic disorder (together these constitute the “tic-related” form of OCD). The adult-onset form of OCD, in contrast, is much less likely to occur in the context of a personal or family history of tics (Black et al., 1992; Pauls

neurodevelopmental aspects of clinical disorders

et al., 1995). The early-onset form is thought to be more strongly familial than the adult-onset form (Pauls et al., 1995). Symptoms of the tic-related form of OCD are significantly more likely to include aggressive, sexual, checking, and cleaning compulsions, whereas the non-tic-related form is more likely to involve symmetry and ordering concerns (Leckman, Grice, et al., 1997). When present together, the severities of OCD and tic symptoms have been shown to covary with one another, suggesting an underlying common modulator of severity over the short term (Lin et al., 2002). In contrast to tics, childhood-onset OCD symptoms over the long term tend more often to persist into late adolescence and adulthood, and they are usually more functionally debilitating than are tics alone (Leonard et al., 1990; Swedo, Rapoport, et al., 1989). In addition to the increased risk of having tics and ticrelated disorders in children with OCD, the incidence of OCD is also increased in children with TS (Cohen and Leckman, 1994; Leckman, Peterson, Anderson, et al., 1997; Leckman, Walker, et al., 1994), suggesting a common neurobiology between TS and pediatric OCD, and indicating that the child- and adult-onset forms of OCD may compose distinct subtypes of OCD that have differing etiologies (Eichstedt and Arnold, 2001). However, the high rate of tics and TS in children with OCD, as well as in the families of individuals with childhood-onset OCD, indicates that tic disorders and childhood-onset OCD are related and may in fact be manifestations of a single underlying pathology. Evidence suggests increasingly that the common neurobiological origin of the two disorders resides in the frontostriatal circuitry that mediates self-regulatory control. We will review evidence suggesting that frontostriatal abnormalities in children with TS or OCD may interfere with their exerting control over motoric behaviors or thoughts, thereby contributing to tics and compulsions.

Self-regulatory systems in TS and OCD Self-regulatory control is a higher level of CNS organization that is also referred to as executive functioning (GoldmanRakic, 1996; Robbins, 1996), attentional processing (Posner, 1994), supervisory processing (Norman and Shallice, 1986), willed action (Libet, 1985), or top-down processing (Bar, 2003; Frith and Dolan, 1997; Pessoa, Kastner, and Ungerleider, 2003; Przybyszewski, 1998; Sarter, Givens, and Bruno, 2001). In healthy individuals, self-regulatory control processes mature gradually over childhood and early adolescence (Diamond, 1988, 2002; Luna et al., 2004; Luna and Sweeney, 2001, 2004). Experimental paradigms that study self-regulatory processes typically require subjects to inhibit a more automatic behavior in favor of a less automatic one. They are therefore regarded as experimental models for studying the resolution of behavioral conflict. Findings from

developmental studies of self-regulatory control show that performance on Stroop, flanker, go-no-go, and stop-signal reaction time tasks continues to improve over childhood and does not reach full maturity until at least 12 years of age (Bunge et al., 2002; Carver, Livesey, and Charles, 2001; Casey et al., 2001; Casey, Trainor, et al., 1997; Ridderinkhof et al., 1997; Rubia et al., 2000). The Stroop task is one of the most commonly studied of these paradigms (Stroop, 1935). It requires subjects to inhibit word reading in favor of a less automatic behavior, naming the color of letters. When the color that a written word denotes matches the color of the ink in which the letters are printed (e.g., “R-E-D” written in red ink), subjects perform the task easily, as indexed by their rapid responses and infrequent errors. However, when the color that a word denotes does not match the color of the printed letters (e.g., “R-E-D” written in blue ink), the task is more difficult, as indicated by slower responses and more frequent errors. Inhibiting the prepotent reading response in this condition requires the mobilization of attentional resources, the resolution of cognitive conflict, the modulation of the automatic response tendency, and, ultimately, the engagement of selfregulatory control. Imaging studies of brain activity during color naming of the mismatching compared with the matching stimuli have demonstrated activation in large expanses of anterior cingulate and prefrontal cortices, as well as the striatum, in both adults (Carter et al., 2000; Leung et al., 2000; MacDonald et al., 2000; Milham et al., 2002; Pardo et al., 1990; Peterson, Skudlarski, et al., 1999) and children (Adleman et al., 2002; Blumberg et al., 2003). Findings from our recent developmental fMRI study of the Stroop task (Marsh et al., 2006) suggest that the maturation of frontostriatal systems underlies the normal development of selfregulatory processes, consistent with findings from previous developmental imaging studies of response inhibition (Adleman et al., 2002; Bunge et al., 2002; Casey, Castellanos, et al., 1997; Casey, Trainor, et al., 1997; Luna and Sweeney, 2004; Luna et al., 2001; Tamm, Menon, and Reiss, 2002). A major limitation of the Stroop task as a paradigm to study self-regulatory control within the scanning environment is that it requires a verbal response, precluding the online monitoring of performance during scanning. The “Simon Spatial Incompatibility Task” overcomes this limitation. In the Simon task, subjects are instructed to indicate, by using one of two possible button responses, the direction in which a white arrow is pointing (left or right). The individual arrows appear on one or the other side of a screen. When the side on which the arrow appears matches the direction in which the arrow points (e.g., a leftward-pointing arrow on the left side of the screen), subjects perform the task rapidly with few errors. When the side on which the arrow appears does not match the direction of the arrow, however (e.g., a leftward-pointing arrow on the right side of

marsh et al.: tics and compulsions

719

the screen), subjects perform the task more slowly and make more errors. The Simon task has been shown to produce a pattern of brain activation nearly identical to that produced by the Stroop task (figure 44.1 and plate 59; Peterson, Kane, et al., 2002). It can therefore be considered a nonverbal analogue of the Stroop task with the same underlying task demands and processing functions. Both tasks require the

Figure 44.1 Simon versus Stroop activations. Axial slices are labeled with Talairach Z-coordinates at far left. “Contrast”: comparison of Simon versus Stroop activity with p < 0.05, cluster filter 9 pixels. The Simon tends to activate superior parietal cortices slightly more than does the Stroop, and the Stroop tends to activate posterior temporal areas more than does the Simon, consistent with their greater demands on spatial and receptive language functions, respectively. (See plate 59.)

720

engagement of self-regulatory control to inhibit a more automatic behavior in favor of a less automatic one. Tourette’s Syndrome Recent findings from our laboratory suggest that children with TS perform similarly to their healthy counterparts on both the Stroop and Simon tasks, and that interference scores decline normally with age, likely reflecting maturation of neural systems that subserve response inhibition (Comalli, Wapner, and Werner, 1962; Dash and Dash, 1982; Schiller, 1966). Compared to healthy individuals, however, and despite their normal behavioral performance, those with TS seem to rely more on exaggerated activation of frontostriatal systems in order to maintain adequate Stroop task performance, perhaps compensating for underlying disturbances in the efficiency of these neuroregulatory systems (Marsh et al., 2007), a dysfunction that is likely to affect many aspects of self-regulatory control, including their relative inability to regulate tic behaviors (figure 44.2 and plate 60). In addition, the continued presence of disturbances in frontostriatal circuits in adults with persistent TS likely contributed to their inability to suspend task-related activity in ventral prefrontal and posterior cingulate regions during performance of the baseline task. Unlike healthy adults, adults with TS did not demonstrate activity in these “default mode” processing regions during the relative “resting” state (Greicius et al., 2003; Greicius and Menon, 2004; Gusnard et al., 2001; Gusnard and Raichle, 2001; McKiernan et al., 2006, 2003; Shulman et al., 1997). Impairments in self-regulatory control likely required adults with TS to allocate a greater proportion of their attentional resources to maintain task performance, thus precluding them from letting their minds wander and from monitoring their emotions or internal states during the baseline task (Marsh et al., 2007). Other findings from our laboratory have indicated the presence of exaggerated activation of frontostriatal circuits during performance of the Simon task in adults who have persistent TS, also likely reflecting a compensatory response that enhances self-regulatory control within dysfunctional circuits, in the service of maintaining normal behavioral performance on the task (Raz et al., in press). In a previous fMRI study, the willful suppression of tics also produced changes in activity of frontostriatal systems, with the magnitude of change in those regions correlating with the severity of tic symptoms (Peterson, Skudlarsky, et al., 1998). Those findings suggested that a greater ability of the cortex to suppress basal ganglia activity might be linked with decreased tic severity, consistent with findings from PET and SPECT studies that have suggested involvement of the basal ganglia in TS (Braun et al., 1995; Butler, Stern, and Silbersweig, 2006). Recent neurophysiological findings also indicate that the frontal cortex plays an important compensatory role in tic suppression (Serrien et al., 2005). In this study, the

neurodevelopmental aspects of clinical disorders

Figure 44.2 Main effects of Stroop task performance in (A) Tourette’s syndrome and (B) normal control groups. Diagnosis-byperformance interactions (C) were detected in frontostriatal and posterior cingulate regions in which interactions of diagnosis with age were observed. The TS subjects likely needed to engage frontostriatal regions more to maintain task performance. (See plate 60.)

involvement of sensorimotor-frontal connections in tic suppression was evidenced by increased EEG coherence during the suppression of voluntary movements in individuals with TS compared with healthy subjects during a go-no-go task. Taken together, these findings suggest that when frontostriatal control systems fail, tics and possibly other behaviors are more likely to be released from inhibitory influences of

the frontal cortex. Individuals with TS must rely on frontostriatal systems to both suppress their tics and perform cognitive tasks that require the engagement of self-regulatory control. Anatomical findings have revealed larger dorsal prefrontal cortices, as well as smaller overall corpus callosum size, in children compared to adults with TS. In addition, the magnitudes of these anatomical changes were associated with the severity of tic symptoms, with larger prefrontal volumes and smaller corpus collosum size accompanying less severe tic symptoms (Peterson, Staib, et al., 2001; Plessen et al., 2004; Spessot, Plessen, and Peterson, 2004). Larger cortical volumes in children with TS likely represent a compensatory or adaptive process that serves to attenuate tics. The continuous need to suppress tics at school, on the playground, or in other social settings will activate prefrontal regions, thereby stimulating activity-dependent neural plasticity in frontal regions that provide increased reserves of the neural resources that subserve self-regulatory control. Likewise, neural plasticity, particularly in the form of axonal pruning, may contribute to the smaller size of the corpus callosum detected in children with TS by limiting neuronal trafficking across this interhemispheric commissure and reducing input to cortical inhibitory interneurons within the prefrontal cortices. Reduced inhibitory input may in turn enhance prefrontal excitatory activity, thus helping children with TS to engage self-regulatory control over their tics (Plessen et al., 2004). In contrast, smaller prefrontal volumes in adults with TS may represent an inability to produce this activity-dependent plastic response to the need to engage self-regulatory control over tics and other behaviors. Greater activation of prefrontal cortices and basal ganglia nuclei was associated with poorer performance on the Stroop task in both children and adults with TS, and activation of these regions was generally exaggerated in adults with TS during performance of the Simon task (Marsh et al., 2007; Raz et al., in press). These findings suggest that the continued presence of abnormalities in frontostriatal functioning may contribute to the continued presence of impaired self-regulatory control in TS adults. Smaller prefrontal volumes in adults with TS may provide insufficient neural and functional reserve to control the urges to tic that originate from abnormalities within the striatum, thereby contributing to the release of tic symptoms from top-down control and the persistence of TS into adulthood (Peterson, Staib, et al., 2001). The extensive reliance on frontostriatal circuits to maintain performance on the Stroop task in children and adults with TS, as well as the exaggerated reliance on these circuits during performance of the Simon task in TS adults, may represent a co-opting of the developmental circuits that subserve normative improvement in selfregulatory control during human development, thereby accounting for the similar age-related improvements in

marsh et al.: tics and compulsions

721

behavioral performance on these tasks across diagnostic groups. Obsessive-Compulsive Disorder Findings from numerous neuropsychological studies suggest that children and adults with OCD perform poorly on tasks that require the engagement of self-regulatory control over behaviors and cognitions (Bannon et al., 2002; Rankins, Bradshaw, and Georgiou-Karistianis, 2006; Rosenberg, Averbach, et al., 1997; Rosenberg, Dick, et al., 1997; Van der Linden et al., 2005). Evidence indicates that the degree of selfregulatory impairments on Stroop, go-no-go (Bannon et al., 2002), and oculomotor suppression tasks (Rosenberg, Averbach, et al., 1997) is associated with the severity of OCD symptoms, suggesting that functional abnormalities in self-regulatory systems may underlie the symptoms of this disorder. Symptoms of OCD may in fact be characterized by failures to self-regulate or shift attention from the intrusive, troubling thoughts that precede repetitive, compulsive behaviors or rituals (Chamberlain et al., 2005). However, OCD symptoms may also be considered failures to regulate the behaviors themselves (e.g., ritualistic check-ing behaviors), similar to the tic symptoms of individuals with TS. To our knowledge, no studies thus far have investigated age-related, functional changes in the frontostriatal systems that underlie self-regulatory control processes in individuals with OCD. Some anatomical findings suggest that abnormal maturation of frontostriatal circuitry may contribute to the clinical phenotype of OCD (Behar et al., 1984; Bolton et al., 2001; Fitzgerald et al., 2000; Giedd et al., 2000; Gilbert et al., 2000; Luxenberg et al., 1988; MacMaster et al., 1999; Moore et al., 1998; Rosenberg, Benazon, et al., 2000; Rosenberg, Keshavan, Dick, et al., 1997; Rosenberg, Keshavan, O’Hearn, et al., 1997; Rosenberg, MacMaster, et al., 2000; Russell et al., 2003). Findings from the few studies that have investigated brain structure and function in children with OCD, or in adults with childhood-onset illness, indicate abnormalities of the striatum (Benkelfat et al., 1990; Luxenberg et al., 1988; Rosenberg, Keshavan, O’Hearn, et al., 1997). However, unpublished findings suggest that striatal abnormalities in children with OCD may be attributable to the presence of comorbid tic disorders (Peterson, Zhu, et al., submitted). Evidence from two studies of brain structure in adult-onset OCD suggest the presence of smaller caudate volumes (Luxenberg et al., 1988; Robinson et al., 1995), whereas another study reported normal caudate volumes (Aylward et al., 1996). The inconsistencies in these findings may be attributable to methodological inconsistencies, or they may indicate that striatal abnormalities are unique to pediatric, tic-related OCD. Although further research is needed, the presence of striatal abnormalities in children with TS and childhood-onset OCD points to a common etiology in these pediatric disorders,

722

perhaps contributing to shared deficits in self-regulation of behaviors and cognitions. Morphological studies of the frontal cortex in OCD have also produced inconsistent findings, with some reporting decreases (Luxenberg et al., 1988; Szeszko et al., 1999) and some reporting no differences (Riffkin et al., 2005) in OFC volumes relative to healthy adults. One investigation of prefrontal volumes in children and adolescents with OCD reported increased volumes of ventral prefrontal cortices relative to those of control children (Rosenberg and Keshavan, 1998). A recent study, however, reported smaller volumes of dorsal prefrontal cortex in children with OCD, and volumes in these regions correlated inversely with the severity of OCD symptoms (Peterson, Zhu, et al., submitted). In contrast, children with chronic tic disorders but no OCD had larger volumes of the dorsal prefrontal cortex (DPFC), consistent with previous findings in children with TS (Peterson, Staib, et al., 2001). These findings suggest that pediatric-onset OCD may differ from chronic tic disorders in their neuroanatomical bases: whereas TS likely arises from dysfunction in the striatum, dysfunction in OCD may be based more primarily within prefrontal cortices. This interpretation is consistent with findings that individuals with OCD are impaired at tasks that require self-regulatory control (Bannon et al., 2002; Rosenberg, Averbach, et al., 1997), whereas individuals with TS perform as well as their healthy counterparts. Children with TS likely engage DPFC regions to compensate for their self-regulatory impairments, an adaptive process that is thought to contribute to the enlarged DPFC volumes observed in children with TS (Peterson, Staib et al., 2001; Peterson, Zhu, et al., submitted). A preliminary PET study reported that during a continuous performance task, tic-related OCD symptoms (e.g., checking compulsions and religious, aggressive, and sexual obsessions) were associated with regional blood flow to the striatum, whereas non-tic-related symptoms (e.g., washing and cleaning) correlated with blood flow to dorsolateral prefrontal cortex (Rauch et al., 1998). Thus tic-related symptoms were associated with the functioning of the striatum and non-tic-related symptoms with prefrontal functioning. These findings further suggest that the dysfunction in pediatric OCD may be based in prefrontal regions, whereas chronic tic disorders are based more in the striatum. Functional imaging studies of children or adolescents with OCD are scarce. However, resting-state studies employing PET or SPECT in OCD adults have reported hyperactivity of the lateral orbitofrontal cortex, caudate nucleus, anterior cingulate cortex, and thalamus (Baxter, 1990; Baxter et al., 1987, 1988, 1992; Machlin et al., 1991; Rubin et al., 1992; Swedo, Schapiro, et al., 1989). These associations are also presumed to characterize children and adolescents with OCD (Bradshaw and Sheppard, 2000; Rosenberg and Keshavan, 1998).

neurodevelopmental aspects of clinical disorders

In addition, functional imaging studies of OCD patients have reported with reasonable consistency increased activity within these regions during the experimental provocation of symptoms (Breiter et al., 1996; McGuire et al., 1994; Rauch et al., 1994; Saxena et al., 1998), suggesting perhaps a causal relationship between frontostriatal hyperactivity and OCD symptoms (Evans, Lewis, and Iobst, 2004) and possibly indicating that frontostriatal hyperactivation represents a less efficient regulatory system, consistent with findings from neuropsychological studies of impaired regulatory control in individuals with OCD (Evans, Lewis, and Iobst, 2004). Few studies have investigated the functioning of frontostriatal regulatory systems in pediatric OCD, and none have assessed volumetric or functional abnormalities in prefrontal and striatal regions in both children and adults with childhood-onset OCD. Identifying the neurodevelopmental correlates of OCD symptoms is therefore difficult. Although further studies are needed, we speculate that OCD may arise from dysfunction in the prefrontal cortices, contributing to performance deficits on neuropsychological tasks of selfregulatory control that are spared in TS, as well as to the nonshared clinical symptoms of the two disorders. Self-regulatory control over behaviors and the urges to perform them is impaired in individuals with TS and OCD. The phenomenological similarities of tics and compulsions, along with their common genetic and neural basis, suggest that they lie on a continuum of semi-involuntary or habitual behaviors (Leckman, Yeh, and Cohen, 2001). Parents of children with tics and OCD symptoms in fact often describe the behaviors as habits or mannerisms.

Habit-learning systems Habits are assembled routines that link sensory cues with motor action. They allow us to act without thinking, as we do when riding a bicycle or driving a car. As such, they are enormously adaptive and part of a common evolutionary heritage that we share with other vertebrates, and probably with all other species capable of goal-directed behavior. Repetitive sensory-cognitive-emotive-motor ensembles, as occur when a gifted pianist plays a sonata or a surgeon sutures a wound, have habitual elements. When we do things over and over again, we get better at it. We think less about the action, and we can respond in a more nuanced manner to other internal cues or to perturbations in the external world. When we do something over and over, our brains compress the relevant information into “chunks” coordinating bits of behavior into action sequences (Graybiel, 2005; G. Miller, 1956). These events seem to involve the same CSTC circuits that underlie self-regulatory functions. The frontostriatal portions of these circuits, including the projections of ventral prefrontal cortex (VPFC) and anterior cingulate (AC) cortex to the basal ganglia (Ferry et al.,

2000), have been most clearly identified as subserving self-regulation of motor functions (Baumeister and Vohs, 2004; Bronson, 2000; Peterson, Skudlarsky, et al., 1998; Peterson, Staib, et al., 2001; Peterson, Thomas, et al., 2003; Posner and Rothbart, 1998; Tucker, Luu, and Pribram, 1995). Similarly, dense monosynaptic projections in these circuits from VPFC to the amygdala (Amaral and Price, 1984) and hippocampus (Ongur, Drevets, and Price, 1998) (directly to the hippocampus and indirectly via the parahippocampus) (Stefanacci, Suzuki, and Amaral, 1996; Suzuki and Amaral, 1994) have been increasingly implicated in the regulation of emotions (Beauregard, Levesque, and Bourgouin, 2001; Davidson et al., 2002; Hariri et al., 2003; Levesque et al., 2004; Ochsner et al., 2002; Phillips et al., 2003). Prefrontal portions of these circuits generally are viewed as the central “executor” of the regulatory functions that these circuits subserve. Although the aspects of self-regulation that the striatum and mesial temporal lobe structures subserve are still controversial, a vast recent literature from animal and human lesion studies suggests that these structures together participate centrally in the affective and mnemonic components of self-regulation. Mesial temporal lobe structures (the hippocampus and amygdala) mediate memories for declarative and emotional memories (Eichenbaum, 2000). Structures in the striatum (the caudate nucleus and putamen), in contrast, mediate memories for stimulus-response associations (Jog et al., 1999; Packard and Knowlton, 2002). Disturbances in frontostriatal systems may interfere with self-regulatory functions either by impairing the learning of S-R associations, thereby reducing overall flexibility of the behavioral program, whereas disturbances in frontotemporal systems may interfere with emotional learning and higher-order cognitive functions that are required to organize behavior appropriately in space and time, within an emotionally salient social environment (figure 44.3). Thus the striatum and mesial temporal lobe structures mediate different forms of learning and memory that are needed for self-regulation. Whereas mesial temporal lobe structures mediate memories for conscious facts, previous experiences, and semantics (collectively termed a “declarative” or “episodic” memory system) (Eichenbaum, 2000), structures in the striatum (the caudate nucleus and putamen) mediate memories for motor skills, procedures, and habits (collectively termed procedural, stimulus-response, or habit learning) (Jog et al., 1999; Packard and Knowlton, 2002). Both animal (Packard, Hirsh, and White, 1989; Packard and Knowlton, 2002; Packard and McGaugh, 1992; Packard and Teather, 1997) and human studies (Knowlton, Mangels, and Squire, 1996; Knowlton, Squire, and Gluck, 1994; Reber, Knowlton, and Squire, 1996) have demonstrated the mediation of habit learning by the basal ganglia, and they have shown the independence of habit learning from

marsh et al.: tics and compulsions

723

Figure 44.3 Pathophysiological model of the shared clinical phenotype of TS and OCD. Frontal abnormalities may have downstream effects on procedural memory and habit-forming systems

within the striatum and on declarative and emotional memory systems within the amygdala and hippocampus.

declarative learning and memory functions that are based within medial temporal lobe structures. Studies of memory systems in rodents have shown that electrolytic or neurochemical lesions of the dorsal striatum impair performance on habit-learning tasks but not on declarative-memory tasks, whereas lesions of the hippocampal system impair performance of declarative-memory tasks but not on habitlearning tasks (McDonald and White, 1993; Packard, Hirsh, and White, 1989). Similar studies in rodents have also highlighted the importance of a certain class of interneurons in habit learning (Berke et al., 2004). Tasks of probabilistic classification learning (PCL) are commonly used to assess habit learning in human subjects (Keri et al., 2002; Knowlton, Squire, and Gluck, 1994; Poldrack et al., 1999; Witt, Nuhsman, and Deuschl, 2002). This form of habit learning circumvents the use of declarative memory by probabilistically associating cues with specific outcomes. One version of a PCL task is a weather-prediction game that requires the gradual learning of S-R associations. Declarative memory of the previous trial is not as useful in improving performance as is information gleaned across many trials. Subjects try to predict rain or sunshine based on the presentation of a varying combination of a set of cards on a computer screen (Knowlton, Mangels, and Squire, 1996; Knowlton, Squire, and Gluck, 1994). Each card is independently and probabilistically related to the outcomes (rain or shine), which occur equally often. Because of the probabilistic nature of the

task, subjects usually believe that they are simply guessing at the outcome. Normal subjects do, however, exhibit learning on this task, in that they gradually improve in their ability to predict the correct weather outcome, even though it is outside of their conscious awareness. Patients with diseases affecting the striatum, such as Huntington’s disease and Parkinson’s disease, exhibit impaired learning on this task, even though they are able to answer explicit factual questions about the task (Knowlton, Mangels, and Squire, 1996; Knowlton, Squire, and Gluck, 1994). This pattern of findings in human diseases involving the basal ganglia is consistent with earlier studies in lower animals indicating that the dorsal striatum subserves habit learning (Packard, Hirsh, and White, 1989; Packard and Knowlton, 2002; Packard and McGaugh, 1992; Packard and Teather, 1997). Conversely, patients with temporal lobe lesions that affect declarative memory systems are impaired at answering explicit factual questions about the task (Knowlton, Mangels, and Squire, 1996; Knowlton, Squire, and Gluck, 1994; Packard and Knowlton, 2002; Reber, Knowlton, and Squire, 1996), whereas their learning on the probabilistic features of the task is intact. Consistent with the implications from these behavioral findings that habit learning is based within the dorsal striatum, human functional imaging studies have demonstrated increased neuronal activity in the striatum and reduced activity in the hippocampus during habit learning in the weather-prediction task (Poldrack et al., 1999; Seger and Cincotta, 2005).

724

neurodevelopmental aspects of clinical disorders

Although a common distinction made in the study of multiple memory systems is between declarative and procedural habit-learning systems, these broad classes of memory functions are themselves likely composed of several anatomically and functionally distinct subsystems. Procedural memory, for example, is not a unitary construct, but rather a heterogeneous collection of learning capacities or skills (Squire, 1987; Squire and Kandel, 1999; Squire and Zola, 1996). Human research that has focused on the acquisition and functional anatomy of procedural motor skills has concluded that several brain structures are of critical importance for acquiring these skills (Doyon, Penhune, and Ungerleider, 2003; Georgopoulos, 2000; Graybiel, 1995; Laforce and Doyon, 2001; Sanes, 2003; Ungerleider, Doyon, and Karni, 2002). These structures include the neostriatum as well as the cerebellum and motor cortex of the frontal lobe. Studies that have used the pursuit-rotor task to assess motor-skill habit learning in patients with striatal dysfunction (Huntington’s disease, Parkinson’s disease, or TS) have reported variable findings of either impaired performance (Gabrieli, 1997; Heindel, Butters, and Salmon, 1988; Sarazin et al., 2002; Stebbins et al., 1995) or normal performance (Bondi and Kaszniak, 1991; Haaland et al., 1997; Harrington et al., 1990; Heindel et al., 1989). One human imaging study of the pursuit-rotor task emphasized the role of motor areas during skill acquisition, showing an increase in activity in corticocerebellar circuits rather than in corticostriatal pathways (Grafton, Woods, and Mike, 1994). Another study reported intact mirror-tracing skill learning in patients with Huntington’s disease (Gabrieli et al., 1997), and no studies had previously used this task to assess procedural skill learning in patients with TS. Tourette’s Syndrome The phenomenological similarity of tics with habits, together with both the documented abnormalities of the striatum in persons with TS and the role of the striatum in habit learning, prompted us as well as other investigators to hypothesize that tics could represent habit learning gone awry (Peterson, 2000; Peterson, Anderson, et al., 2003; Peterson and Klein, 1997; Peterson, Staib et al., 2001; Peterson, Thomas, et al., 2003). Findings from our laboratory indicate that relative to their healthy counterparts, both children and adults with TS were impaired at habit learning on the weather-prediction task (Marsh et al., 2004). However, the same individuals were not impaired at learning on the pursuit-rotor or mirror-tracing tasks (Marsh et al., 2005). We previously interpreted the normal learning of perceptual-motor skills in persons with TS in the context of the previously documented impairments in the probabilistic classification form of habit learning (i.e., in the weather task) to suggest that these forms of learning are indeed dissociable and that the deficits in habit learning in TS are not merely attributable to motor limitations

imposed by the presence of TS. In addition, Keri and colleagues (2002) observed similar habit-learning deficits in TS subjects despite evidence of their explicit memory system being intact (Keri et al., 2002). Together, the findings from these three studies of habitbased learning in persons with TS indicate that their deficient habit learning likely contributes to their habitlike, stereotyped behaviors. The tics in TS may be the product of core disturbances in the structure and function of the striatum that predispose an individual to impairments in habit learning and to the expression of fragmented motor and vocal behaviors. These predispositions to habitual tic behaviors may then be released from regulatory influences of the prefrontal cortex (Gerard and Peterson, 2003; Spessot, Plessen, and Peterson, 2004). Thus disturbances in frontostriatal systems in patients with TS may interfere with self-regulatory functions, thereby altering habit learning and releasing habitual tic behaviors from inhibitory control. This phenomenon likely extends to other psychopathologies in which patients are also plagued by recurrent habitual behaviors that are associated with a loss of control. Obsessive-Compulsive Disorder Evidence suggests that despite the purported striatal abnormalities in both TS and OCD, and despite the deficient habit learning in individuals with TS, adults with OCD are not impaired at tasks that are mediated by the striatum. Findings from a study that employed a serial reaction time (SRT) task during PET imaging revealed that, compared with healthy participants, adults with OCD exhibited increased metabolism in the hippocampus rather than in the striatum, although the behavioral performances of the two groups on the task was equivalent (Rauch, Savage, et al., 1997). These findings suggest that the intact SRT performance of adults with OCD may be achieved by their compensatory engagement of declarative learning systems based within the mesial temporal lobe (Rauch, Savage, et al., 1997; Rauch, Whalen, et al., 1997). When required to perform a concurrent explicit task, however, individuals with OCD were impaired at implicit sequence learning (Deckersbach et al., 2002). Based on these findings, the compensatory involvement of hippocampal structures that mediate explicit, conscious processing has been suggested to contribute to the clinical presentation of OCD, particularly the repetitive, intrusive thoughts that linger in conscious awareness (Rauch, Wedig, et al., 2007). Findings of aberrant hippocampal recruitment during performance of the SRT in adults with OCD were replicated in a recent fMRI study, which further indicated that striatal activation was inversely associated with the specific, non-tic-related symptoms of symmetry, ordering, washing, and contamination (Rauch et al., 2007). Additionally, ticrelated symptoms of aggressive and sexual obsessions and

marsh et al.: tics and compulsions

725

checking were associated with activation of the OFC (Chamberlain et al., 2005; Evans, Lewis, and Iobst, 2004; Saxena et al., 1998). In other words, greater activation of the striatum accompanied less severe tic-related symptoms, and greater activation of the prefrontal cortex accompanied more severe non-tic-related OCD symptoms in this sample of adults with OCD. These findings further indicate that OCD may be a disorder of prefrontal cortices, whereas TS is primarily based within the striatum, giving rise to the selective impairments on tasks mediated by the striatum in persons with TS. To our knowledge, no previous studies have employed cognitive-neuroscience-based tasks to investigate the functioning of the neostriatal habit-learning system in children with OCD, and it is unclear whether the existing studies in adults with OCD were studying adults with child- or adultonset illness. Thus the extant data do not allow us to make direct inferences regarding the role of striatal-based learning systems in determining the neurodevelopmental trajectory of OCD. However, given the more central role of prefrontal cortices in OCD and the striatum in TS, we suspect that children with OCD may not show similar impairments on tasks that are mediated by the striatum.

CSTC circuits that subserve tics and compulsions The neural systems that subserve both tics and compulsions, as well as self-regulatory control processes and habit learning, loop between cortical and subcortical brain regions (Leckman, Knorr, et al., 1991; Peterson and Thomas,

CSTC Component Cortical afferents

Striatum

GP/SNpr

2000). Although the number of anatomically and functionally discrete loops is still controversial (Alexander, Crutcher, and DeLong, 1990; Goldman-Rakic and Selemon, 1990; Parent and Hazrati, 1995), current consensus holds that CSTC circuitry has at least four components—those initiating from and projecting back to sensorimotor cortex, orbitofrontal cortex, association cortices, and limbic and anterior cingulate cortices (table 44.1). These cortical regions map excitatory input onto specific regions of the striatum, which then project inhibitory impulses onto specific regions of the pallidum, then through the thalamus, and back again to the cortex. Information flows predominantly through either a direct or an indirect pathway, depending on whether information from the striatum passes primarily through the internal or external globus pallidus, respectively (figure 44.4). Striatal neurons contain the neurotransmitter gamma-aminobutyric acid (GABA) and are inhibitory to their targets. The projections from the thalamus to the cortex are glutamatergic and excitatory. The direct outflow pathway maps inhibitory signals from the striatum onto the internal segment of globus pallidus and substantia nigra pars reticulata (GPi/SNpr), and then directly onto the thalamus before projecting back to the cortex. The indirect pathway projects inhibitory signals from the striatum to the globus pallidus pars externa (GPe) and a second inhibitory connection to the subthalamic nucleus (STN). The STN neurons send excitatory projections to the GPi/SNpr before sending inhibitory connections to the thalamus and then back to the cortex (Saint-Cyr, 2003; Wichmann and DeLong, 1996).

Table 44.1 Cortical-striatal-thalamic-cortical (CSTC) circuit pathways Sensorimotor Orbitofrontal Pathways Pathways Association Pathways Somatosensory Orbitofrontal DLPFC Primary motor STG Posterior SMA ITG Parietal AC Arcuate Premotor Dorsolateral putamen Ventral caudate Dorsolateral caudate Dorsolateral caudate Ventral putamen Ventrolateral GPi Caudolateral SNpr

Dorsomedial GPi Rostromedial SNpr

Thalamic nuclei

Dorsomedial GPi Rostrolateral SNpr

Limbic-System Pathways AC Hippocampus Entorhinal cortex STG, ITG Ventral caudate Ventral putamen NAcc Rostrolateral GPi Ventral pallidum Rostrodorsal SNpr Medial dorsal

Ventrolateral Medial dorsal Ventral anterior Centromedial Intralaminal Cortical projections Supplementary Orbitofrontal DLPFC AC motor Abbreviations: STG, superior temporal gyrus; ITG, inferior temporal gyrus; GPi, globus pallidus (internal segment); SNpr, substantia nigra pars reticulata; DLPFC, dorsolateral prefrontal cortex; NAcc, nucleus accumbens; AC, anterior cingulate; SMA, supplementary motor area.

726

neurodevelopmental aspects of clinical disorders

Figure 44.4 Normal cortical-striatal-thalamic-cortical circuits and the hyperinnervation of the striatum by dopaminergic neurons. Excitatory (glutamatergic) projections are indicated by solid lines, and inhibitory (GABAergic) projection by dotted lines. GPe, exter-

nal segment of the globus pallidus; STN, subthalamic nucleus; GPi/SNpr, internal segment of the globus pallidus or the substantia nigra pars reticulata.

The presence of a markedly higher total neuron number in the GPi of TS in contrast to a lower neuron number and density in the GPe and caudate nucleus of TS subjects is consistent with our findings of deficits in habit learning (Kalanithi et al., 2005; Leckman, Vaccarino, et al., 2006). Specifically, the observation that there was a deceased number and proportion of the caudate neurons that were positive for the calcium-binding protein parvalbumin in tissue from TS subjects suggests that these cells are the fastspiking GABAergic interneurons that are crucial to habit formation (Berke et al., 2004). The neurotransmitter dopamine selectively modulates both the direct and indirect pathways by exciting the striatal neurons that send inhibitory, GABAergic projections to the output nuclei of the direct pathway, and by inhibiting the striatal neurons that send GABAergic projections to the GPe in the indirect pathway (figure 44.4) (Graybiel, 1990). In the direct pathway, dopaminergic hyperinnervation stimulates D1 receptors that are located on GABAergic postsynaptic neurons (Gerfen et al., 1990). These striatal neurons, projecting from the GPi/SNpr to the thalamus, then block inhibitory interneurons within the thalamus, thereby enhancing glutamatergic excitation of the cortex and increasing motor output. Another population of striatal neurons within the indirect pathway express D2 dopamine receptors, which are

thought to decrease the effect of cortical input to striatal neurons (Surmeier et al., 1993). The striatal neurons in the indirect pathway project to the GPe and then subsequently to the STN (Gerfen et al., 1990). In this pathway, the excitatory output from the STN to the GPi/SNpr likely enhances rather than blocks inhibitory projections of these basal ganglia output neurons, thereby reducing thalamocortical excitation. The presence of abnormalities in dopamine transmission in persons with TS (Heinz et al., 1998; Wolf et al., 1996) is consistent with findings that the administration of dopamine antagonists often reduces the severity of tic symptoms (Leckman, Peterson, Pauls, and Cohen, 1997). We and others suspect that hyperinnervation of the striatum by dopaminergic neurons may excessively stimulate the direct pathways and inhibit the indirect pathways in individuals with TS, thereby contributing to their difficulty inhibiting or engaging self-regulatory control over motor behaviors (Leckman, Peterson, Anderson, et al., 1997; Peterson and Thomas, 2000). Based on computational models of normal basal ganglia functioning, a pathophysiological model of tic disorders has been proposed to explain the genesis of tics (Albin and Mink, 2006; Mink, 1996, 2001). In normal striatal functioning, the inhibitory output of the basal ganglia is said to act as a “brake” on the activity of motor patterns that

marsh et al.: tics and compulsions

727

are generated in the cortex and brain stem. To initiate a movement, striatal neurons in the direct pathway are activated and inhibit basal ganglia output neurons in the GPi/ SNpr. The removal of this inhibition may release the brake on motor pattern activity, thereby allowing execution of the intended motor behavior. Simultaneously, excitatory projections from STN to the GPi/SNpr then project to the thalamus, inhibiting and engaging the brake on competing motor patterns, further facilitating performance of the intended movement (Albin and Mink, 2006). Based on this model of normal basal ganglia functioning, tics are hypothesized to be a product of striatal neurons becoming inappropriately active, thereby inhibiting basal ganglia output neurons and, in turn, disinhibiting competing motor patterns (Leckman, Vaccarino, et al., 2006). This abnormal striatal activation may then produce an involuntary movement (figure 44.5) (Albin and Mink, 2006; Mink, 1996, 2001). The repeated and stereotyped nature of these unwanted movements in the form of tics therefore may be produced by excess activity in these striatal neurons. The

basic scheme of this model, the facilitation and inhibition of competing movements, likely applies to more complex behaviors and thoughts as well as to tics, thereby accounting for the similarities between various forms of tic behaviors and compulsions. Both tics and compulsions may arise from the failure to inhibit unwanted behaviors and thoughts that result from abnormal patterns of activity in basal ganglia output neurons (Mink, 2001). The segregation of the striatum based on the outputs to differing cortical areas (see table 44.1) through the thalamus may provide the neuroanatomical basis for the differential production of simple tics, complex tics, and compulsions (Mink, 2001). Both simple and complex tics likely arise from abnormal function within the sensorimotor pathway. Simple motor patterns (simple tics) may result from abnormal activation of the supplementary motor area through basal ganglia–thalamocortical connections in this pathway, whereas complex tics may arise from abnormal activation of supplementary motor, premotor, and cingulate cortices within the same pathway. The premonitory “urges”

Figure 44.5 Dopaminergic hyperinnervation could produce multiple effects in the direct and indirect pathways through the basal ganglia, which predispose to disinhibition of thalamocortical excitatory projections and concomitant disinhibition of behavior in TS-related conditions. Excitatory (glutamatergic) projections are

indicated by solid lines, and inhibitory (GABAergic) projections by dotted lines. GPe, external segment of the globus pallidus; STN, subthalamic nucleus; GPi/SNpr, internal segment of the globus pallidus or the substantia nigra pars reticulata.

728

neurodevelopmental aspects of clinical disorders

that typically precede tics are similar to the urges to move described by patients who have undergone electrical stimulation of the sensorimotor cortex (Lim et al., 1994). Recently, a brain network of paralimbic areas including the anterior cingulate and insular cortex, supplementary motor area, and parietal operculum was found to be associated with these premonitory urges preceding tics (Bohlhalter et al., 2006). The anterior cingulate is also involved in the pathophysiology of OCD (Adler et al., 2000; Rauch et al., 1994), and some compulsions are sometimes difficult to differentiate from complex tics. Compulsions, however, may also be a consequence of abnormal activation in the OFC that derives from activity in basal ganglia–thalamocortical connections within the orbitofrontal pathway, consistent with findings of increased metabolism or blood flow to the OFC in OCD imaging studies (Alptekin et al., 2001; Baxter et al., 1987; Nordahl et al., 1989; Rubin et al., 1995; Swedo, Schapiro, et al., 1989) and with recent findings

GABA

CSTC circuits that subserve perceptually cued learning Cortical neurons projecting to the striatum outnumber striatal spiny neurons (MSs) by about a factor of ten (Zheng and Wilson, 2002). These convergent cortical efferent neurons project to the dendrites of MSs within the striosomes and matrix of the striatum, two structurally similar but neurochemically distinct, histologically defined compartments in the basal ganglia (figure 44.6 and plate 61). These two compartments differ in their cortical inputs, with the striosomal MS projection mainly receiving convergent limbic and prelimbic inputs, and neurons in the matrix mainly receiving convergent input from ipsilateral primary motor and sensorimotor cortices, and from contralateral primary motor cortices.

S

Striosomal medium spiny neuron

M

Medium spiny neuron in the matrix

Glutamate Acetylcholine Dopamine

that acute deep-brain stimulation of the ventral striatum in patients with OCD modulates activity in this pathway (Rauch et al., 2006).

FS

Tan

Midline thalamic nuclei

Inputs to striosomes

Fast spiking aspiny neuron

Tonically active neuron

Cortical inputs to the matrix

Hippocampus Prefrontal cortex

Striosome

M FS

S Anterior cingulate cortex

M

Tan Orbital frontal cortex

S

Motor and premotor cortex

FS M Sensorimotor cortex

Amygdala

Matrix

Substantia nigra, pars compacta

Figure 44.6 A schematic diagram of the major inputs into the medium spiny GABAergic projection neurons of the striatum. Tonically active neurons (TANs) synapse on fast-spiking neurons (FSNs), which play a key role in modulating the medium spiny neurons (M). (See plate 61.)

marsh et al.: tics and compulsions

729

Several other less abundant striatal cell types probably have a key role in habit learning, including cholinergic tonically active neurons (TANs) and fast-spiking GABAergic interneurons (FSNs) (Gonzalez-Burgos et al., 2005; Jog et al., 1999). Tonically active neurons are very sensitive to salient perceptual cues because they signal the networks within the cortical–basal ganglia learning circuits when these cues arise (Raz et al., 1996). Specifically, they are responsive to dopaminergic inputs from the substantia nigra, and these signals likely participate in calculation of the perceived salience (reward value) of perceptual cues along with excitatory inputs from midline thalamic nuclei (figure 44.6). The FSNs of the striatum receive direct cortical inputs predominantly from lateral cortical regions, including the primary motor and somatosensory cortex, and they are highly sensitive to cortical activity in these regions (McGeorge and Faull, 1989). They are also known to be electrically coupled through gap junctions that connect adjacent dendrites. Once activated, these FSNs can inhibit many nearby striatal projection neurons synchronously by way of synapses on cell bodies and proximal dendrites (Koos and Tepper, 1999). These FSNs are also very sensitive to cholinergic drugs, suggesting that they are functionally related to the TANs (Koos and Tepper, 2002). The response of particular medium spiny neurons in the striatum seems to depend on a given set of perceptual cues and environmental conditions, suggesting that the coordinated striatal response is acquired through learning and experience (Aosaki, Graybiel, and Kimura, 1994). Inputs from ascending dopamine pathways originating in the substantia nigra, pars compacta (SNpc) play a crucial role in this learning process (Aosaki, Graybiel, and Kimura, 1994). Ensemble recordings, in which the activity of multiple striatal neurons is recorded simultaneously, further clarify the role of the striatum and related brain circuits in the learning and production of habitual or “automatic” behavioral responses (Jog et al., 1996). When ensembles of electrodes in the sensorimotor areas of the rat striatum have been recorded during cued learning tasks, a large array of neurons initially fired rapidly in response to the cue, but only a few neurons fired once the rats learned the task. Thus the cells that do fire may represent a chunk of behavior. In addition, distinct, individually coordinated, motor and cognitive action chunks may be combined and implemented as goaldirected behavioral performance sets. When performed repeatedly, these ensembles become habits. Other electrophysiological studies have shown that neural oscillations in basal ganglia circuits play a role in the emergence of these goal-directed behaviors and in the formation of habits (Courtemanche, Fujii, and Graybiel, 2003; Fujii and Graybiel, 2005; Graybiel, 2005). Fast-spiking neurons seem to be instrumental in orchestrating the oscillatory activity of the MSs in the dorsal lateral striatum (Berke

730

et al., 2004), and this orchestration is likely relevant to the pathogenesis of TS. A recent quantitative examination of the brains of three severely affected individuals with TS revealed altered numbers of FSNs in the striatum (Kalanithi et al., 2005). It is hypothesized that the loss of these FSN cells would allow clusters of MSs within the somatotopic areas that are associated with tics to become disengaged from their usually strong entrainment to synchronized oscillatory activity, thereby giving rise to tics (Leckman, Vaccarino, et al., 2006). Over time, tic-related oscillations may become fixed patterns that both resemble habits and are likely to interfere with the normal functioning of these basal ganglia circuits (e.g., habit learning; Marsh et al., 2004). Other investigators have focused on the role of the STN as a regulator of striatal output. Given the excitatory nature of the STN neurons, their threshold and peak firing rates, a simple model of neuron responses has revealed that large regions of this highly interconnected nucleus respond to excitatory input in the form of a widespread uniform pulse (Gillies and Willshaw, 1998). These pulses of activity may act as a braking signal that resets the basal ganglia output nuclei, thus interrupting habits and setting the stage for the emergence of new goal-directed emotive-cognitivebehavioral sets. If habits are indeed coordinated chunks of thought and action, then tics may be best conceptualized as those prewired bits of behavior that are available to be chunked together to produce habits (Leckman, 2002; Leckman, Vaccarino, et al., 2006). Like habits, tic action sequences often arise from a heightened and selective sensitivity to environmental cues from within the body or from the outside world (Bliss, 1980; Leckman, Walker, and Cohen, 1993; Leckman, Walker, et al., 1994). These perceptual cues include faint premonitory feelings or urges that are relieved with the performance of tics and a need to perform tics or compulsions until they are felt ineffably to be “just right.” Although the neural mechanisms that conspire to produce tics and compulsions have yet to be elucidated, preliminary evidence suggests that they involve the same structures that underlie habit formation.

Conclusions Disturbances in frontostriatal systems likely release the brakes that inhibit underlying impulses to move or to perform more complex, seemingly goal-directed behaviors, thereby contributing to the genesis of tics and compulsions. In typically developing children, the age-related maturation of these systems improves self-regulatory control and the ability to inhibit these unwanted behaviors. Findings from neuroimaging and cognitive neuroscience studies suggest that the neural basis for both TS and OCD resides in anatomical and functional disturbances of frontostriatal systems. In

neurodevelopmental aspects of clinical disorders

addition, the phenomenological similarities between tics and compulsions, along with their common genetic and neural bases, suggest that these childhood neuropsychiatric disorders may be manifestations of the same disease process. The growing number of brain-imaging modalities that can be applied safely in children, as well as the growing number of cognitive neuroscience probes that are based on animal models of behavior, have enormous potential for furthering our study of cognitive developmental neuroscience in both health and illness. Future studies of brain structure-function relationships in children should employ functional imaging techniques to compare the developmental trajectory of brain activity during performance of tasks that require the engagement of self-regulatory control across healthy children and those with TS or OCD. These studies would further our understanding of the role of frontostriatal regulatory systems in the development of these disorders. In addition, research incorporating cognitive neuroscience tasks that probe the function of the striatum in TS and OCD would allow us to determine whether tics and compulsions represent habit learning gone awry. The potential availability of both animal models and human paradigms for studying cognitive processes, such as self-regulatory control and habit learning, offers the exciting promise not only of improving our knowledge of the neurobiological origins of TS and OCD, but also of developing novel therapeutics through bona fide translational research programs and methods that are unavailable to human clinical studies alone. Increasing our knowledge of the maturation of self-regulatory control and habit learning, and of the neuroanatomical systems that subserve these processes, will allow us to determine the ways in which these systems can be derailed during development and how these neurodevelopmental disruptions contribute to the development of tics and compulsions. acknowledgments

This work was supported in part by NIMH grants MH01232, MH59139, MH068318, K0274677 (BSP), and K01-MH077652 (RM); the Suzanne Crosby Murphy Endowment at Columbia University College of Physicians and Surgeons; the Thomas D. Klingenstein and Nancy D. Perlman Family Fund; and in part by NIH grants P01 MH49351, P01 HD03008, T32 MH18268, M01 RR06022, P30 MH30929, and K05 MH076273 (JFL). REFERENCES

Achenbach, T., and C. Edelbrock, 1981. Behavioral problems and competencies reported by parents of normal and disturbed children aged 4–16. Monogr. Soc. Res. Child Dev. 46:1–86. Adleman, N. E., V. Menon, C. M. Blasey, C. D. White, I. S. Warsofsky, G. H. Glover, et al., 2002. A developmental fMRI study of the Stroop color-word task. NeuroImage 16(1):61–75.

Adler, C. M., P. Mcdonough-ryan, K. W. Sax, S. K. Holland, S. Arndt, and S. M. Strakowski, 2000. fMRI of neuronal activation with symptom provocation in unmedicated patients with obsessive compulsive disorder. J. Psychiatr. Res. 34:317– 324. Albin, R. L., and J. W. Mink, 2006. Recent advances in Tourette syndrome research. Trends Neurosci. 29(3):175–182. Alexander, G., M. Crutcher, and M. Delong, 1990. Basal ganglia–thalamocortical circuits: Parallel substrates for motor, oculomotor, “prefrontal,” and “limbic” functions. Prog. Brain Res. 85:119–146. Alptekin, K., B. Degirmenci, B. Kivircik, H. Durak, B. Yemez, E. Derebek, et al., 2001. Tc-99m HMPAO brain perfusion SPECT in drug-free obsessive-compulsive patients without depression. Psychiatry Res. Neuroimaging 107:51–56. Amaral, D. G., and J. L. Price, 1984. Amygdalo-cortical projections in the monkey (Macaca fascicularis). J. Comp. Neurol. 230: 465–496. American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., rev. Washington, DC: American Psychiatric Association. Aosaki, T., A. M. Graybiel, and M. Kimura, 1994. Effect of the nigrostriatal dopamine system on acquired neural responses in the striatum of behaving monkeys. Science 265(5170):412–415. Aylward, E. H., G. J. Harris, R. Hoehn-Saric, P. E. Barta, S. R. Machlin, and G. D. Pearlson, 1996. Normal caudate nucleus in obsessive-compulsive disorder assessed by quantitative neuroimaging. Arch. Gen. Psychiatry 53:577–584. Baer, L., 1994. Factor analysis of symptom subtypes of obsessive compulsive disorder and their relation to personality and tic disorders. J. Clin. Psychiatry 55:18–23. Bannon, S., C. J. Gonsalvez, R. J. Croft, and P. M. Boyce, 2002. Response inhibition deficits in obsessive-compulsive disorder. Psychiatry Res. 110(2):165–174. Bar, M., 2003. A cortical mechanism for triggering top-down facilitation in visual object recognition. J. Cogn. Neurosci. 15(4): 600–609. Baumeister, R. F., and K. D. Vohs, 2004. Handbook of SelfRegulation. New York: Guilford Press. Baxter, L. R., 1990. Brain imaging as a tool in establishing a theory of brain pathology in obsessive compulsive disorder. J. Clin. Psychiatry 51(Suppl.):22–25; discussion, 26. Baxter, L. R. J., J. M. Phelps, J. C. Mazziotta, B. H. Guze, and J. M. Schwartz, 1987. Local cerebral glucose metabolic rates in obsessive-compulsive disorder: A comparison with rates in unipolar depression and normal controls. Arch. Gen. Psychiatry 44:211–218. Baxter, L., Jr., J. M. Schwartz, K. S. Bergman, M. P. Szuba, B. H. Guze, J. C. Mazziotta, et al., 1992. Caudate glucose metabolic rate changes with both drug and behavior therapy for obsessive-compulsive disorder. Arch. Gen. Psychiatry 49(9): 681–689. Baxter, L., J. Schwartz, J. Mazziotta, M. E. Phelps, J. J. Pahl, B. E. Buze, et al., 1988. Cerebral glucose metabolic rates in nondepressed patients with obsessive-compulsive disorder. Am. J. Psychiatry 145:1560–1563. Beauregard, M., J. Levesque, and P. Bourgouin, 2001. Neural correlates of conscious self-regulation of emotion. J. Neurosci. 21(18):RC165. Behar, D., J. L. Rapoport, C. J. Berg, M. B. Denckla, L. Mann, C. Cox, et al., 1984. Computerized tomography and neuropsychological test measures in adolescents with obsessivecompulsive disorder. Am. J. Psychiatry 141(3):363–369.

marsh et al.: tics and compulsions

731

Benkelfat, C., T. E. Nordahl, W. E. Semple, A. C. King, D. L. Murphy, and R. M. Cohen, 1990. Local cerebral glucose metabolic rates in obsessive-compulsive disorder: Patients treated with clomipramine. Arch. Gen. Psychiatry 47(9):840–848. Berg, C. Z., J. L. Rapoport, A. Whitaker, M. Davies, H. Leonard, S. E. Swedo, et al., 1989. Childhood obsessive compulsive disorder: A two-year prospective follow-up of a community sample. J. Am. Acad. Child Adolesc. Psychiatry 28(4):528–533. Berke, J. D., M. Okatan, J. Skurski, and H. B. Eichenbaum, 2004. Oscillatory entrainment of striatal neurons in freely moving rats. Neuron 43(6):883–896. Black, D. W., R. Noyes, Jr., R. B. Goldstein, and N. Blum, 1992. A family study of obsessive-compulsive disorder. Arch. Gen. Psychiatry 49(5):362–368. Bliss, J., 1980. Sensory experiences of Gilles de la Tourette syndrome. Arch. Gen. Psychiatry 37(12):1343–1347. Bloch, M. H., J. F. Leckman, H. Zhu, and B. S. Peterson, 2005. Caudate volumes in childhood predict symptom severity in adults with Tourette syndrome. Neurology 65(8):1253–1258. Blumberg, H. P., A. Martin, J. Kaufman, H. C. Leung, P. Skudlarski, C. Lacadie, et al., 2003. Frontostriatal abnormalities in adolescents with bipolar disorder: Preliminary observations from functional MRI. Am. J. Psychiatry 160(7):1345–1347. Bohlhalter, S., A. Goldfine, S. Matteson, G. Garraux, T. Hanakawa, K. Kansaku, et al., 2006. Neural correlates of tic generation in Tourette syndrome: An event-related functional MRI study. Brain 129(8):2029–2037. Bolton, J., G. J. Moore, S. Macmillan, C. M. Stewart, and D. R. Rosenberg, 2001. Case study: Caudate glutamatergic changes with paroxetine persist after medication discontinuation in pediatric OCD. J. Am. Acad. Child Adolesc. Psychiatry 40(8):903–906. Bondi, M. W., and A. W. Kaszniak, 1991. Implicit and explicit memory in Alzheimer’s disease and Parkinson’s disease. J. Clin. Exp. Neuropsychol. 13(2):339–358. Bradshaw, J. L., and D. M. Sheppard, 2000. The neurodevelopmental frontostriatal disorders: Evolutionary adaptiveness and anomalous lateralization. Brain Lang. 73(2):297–320. Braun, A. R., C. Randolph, B. Stoetter, E. Mohr, C. Cox, K. Vladar, et al., 1995. The functional neuroanatomy of Tourette’s syndrome: An FDG-PET study. II. Relationships between regional cerebral metabolism and associated behavioral and cognitive features of the illness. Neuropsychopharmacology 13(2): 151–168. Breiter, H. C., S. L. Rauch, K. K. Kwong, J. R. Baker, R. M. Weisskoff, D. N. Kennedy, et al., 1996. Functional magnetic resonance imaging of symptom provocation in obsessivecompulsive disorder. Arch. Gen. Psychiatry 53:595–606. Bronson, M. B., 2000. Self-Regulation in Early Childhood: Nature and Nurture. New York: Guilford Press. Bunge, S. A., N. M. Dudukovic, M. E. Thomason, C. J. Vaidya, and J. D. Gabrieli, 2002. Immature frontal lobe contributions to cognitive control in children: Evidence from fMRI. Neuron 33(2):301–311. Burd, L., P. J. Kerbeshian, A. Barth, M. G. Klug, P. K. Avery, and B. Benz, 2001. Long-term follow-up of an epidemiologically defined cohort of patients with Tourette syndrome. J. Child Neurol. 16:431–437. Butler, T., E. Stern, and D. Silbersweig, 2006. Functional neuroimaging of Tourette syndrome: Advances and future directions. Adv. Neurol. 99:115–129. Carter, C. S., A. M. Macdonald, M. Botvinick, L. L. Ross, V. A. Stenger, D. Noll, et al., 2000. Parsing executive proces-

732

ses: Strategic vs. evaluative functions of the anterior cingulate cortex. Proc. Natl. Acad. Sci. USA 97(4):1944–1948. Carver, A. C., D. J. Livesey, and M. Charles, 2001. Age related changes in inhibitory control as measured by stop signal task performance. Int. J. Neurosci. 107(1–2):43–61. Casey, B. J., F. X. Castellanos, J. N. Giedd, W. L. Marsh, S. D. Hamburger, A. B. Schubert, et al., 1997. Implication of right frontostriatal circuitry in response inhibition and attentiondeficit/hyperactivity disorder. J. Am. Acad. Child Adolesc. Psychiatry 36(3):374–383. Casey, B. J., S. D. Forman, P. Franzen, A. Berkowitz, T. S. Braver, L. E. Nystrom, et al., 2001. Sensitivity of prefrontal cortex to changes in target probability: A functional MRI study. Hum. Brain Mapping 13(1):26–33. Casey, B. J., R. Trainor, J. L. Orendi, A. B. Schubert, L. E. Nystrom, J. N. Giedd, et al., 1997. A developmental functional MRI study of prefrontal activation during performance of a go/no-go task. J. Cogn. Neurosci. 9:835–847. Chamberlain, S. R., A. D. Blackwell, N. A. Fineberg, T. W. Robbins, and B. J. Sahakian, 2005. The neuropsychology of obsessive compulsive disorder: The importance of failures in cognitive and behavioural inhibition as candidate endophenotypic markers. Neurosci. Biobehav. Rev. 29(3):399–419. Cohen, D. J., and J. F. Leckman, 1994. Developmental psychopathology and neurobiology of Tourette’s syndrome [review]. J. Am. Acad. Child Adolesc. Psychiatry 33(1):2–15. Comalli, P. E., Jr., S. Wapner, and H. Werner, 1962. Interference effects of Stroop color-word test in childhood, adulthood, and aging. J. Genet. Psychol. 100:47–53. Costello, E. J., A. Angold, B. J. Burns, D. K. Stangl, D. L. Tweed, A. Erkanli, et al., 1996. The Great Smoky Mountains study of youth: Goals, design, methods, and the prevalence of DSM-III-R disorders. Arch. Gen. Psychiatry 53:1129–1136. Courtemanche, R., N. Fujii, and A. M. Graybiel, 2003. Synchronous, focally modulated beta-band oscillations characterize local field potential activity in the striatum of awake behaving monkeys. J. Neurosci. 23(37):11741–11752. Dash, X., and A. S. Dash, 1982. Cognitive developmental studies of the Stroop phenomena: Cross-sectional and longitudinal data. Indian Psychol. 1:24–33. Davidson, R. J., D. A. Lewis, L. B. Alloy, D. G. Amaral, G. Bush, J. D. Cohen, et al., 2002. Neural and behavioral substrates of mood and mood regulation. Biol. Psychiatry 52(6):478–502. Deckersbach, T., C. R. Savage, T. Curran, A. Bohne, S. Wilhelm, L. Baer, et al., 2002. A study of parallel implicit and explicit information processing in patients with obsessivecompulsive disorder. Am. J. Psychiatry 159(10):1780–1782. Diamond, A., 1988. Abilities and neural mechanisms underlying AB performance. Child Dev. 59(2):523–527. Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: Cognitive functions, anatomy, and biochemistry. In D. T. Stuss and R. T. Knight, eds., Principles of Frontal Lobe Function, 466–503. London: Oxford University Press. Doyon, J., V. Penhune, and L. G. Ungerleider, 2003. Distinct contribution of the cortico-striatal and cortico-cerebellar systems to motor skill learning. Neuropsychologia 41(3):252–262. Eichenbaum, H., 2000. A cortical-hippocampal system for declarative memory. Nature Rev. Neurosci. 1(1):41–50. Eichstedt, J. A., and S. L. Arnold, 2001. Childhood-onset obsessive-compulsive disorder: A tic-related subtype of OCD? Clin. Psychol. Rev. 21(1):137–157. Evans, D. W., M. D. Lewis, and E. Iobst, 2004. The role of the orbitofrontal cortex in normally developing compulsive-like

neurodevelopmental aspects of clinical disorders

behaviors and obsessive-compulsive disorder. Brain Cogn. 55(1): 220–234. Ferry, A. T., D. Ongur, X. An, and J. L. Price, 2000. Prefrontal cortical projections to the striatum in macaque monkeys: Evidence for an organization related to prefrontal networks. J. Comp. Neurol. 425(3):447–470. Fitzgerald, K. D., G. J. Moore, L. A. Paulson, C. M. Stewart, and D. R. Rosenberg, 2000. Proton spectroscopic imaging of the thalamus in treatment-naive pediatric obsessive-compulsive disorder. Biol. Psychiatry 47(3):174–182. Frith, C., and R. J. Dolan, 1997. Brain mechanisms associated with top-down processes in perception. Philos. Trans. R. Soc. London [Biol.] 352(1358):1221–1230. Fujii, N., and A. M. Graybiel, 2005. Time-varying covariance of neural activities recorded in striatum and frontal cortex as monkeys perform sequential-saccade tasks. Proc. Natl. Acad. Sci. USA 102(25):9032–9037. Gabrieli, J. E., G. T. Stebbins, J. Singh, D. B. Willingham, et al., 1997. Intact mirror-tracing and impaired rotary-pursuit skill learning in patients with Huntington’s disease: Evidence for dissociable memory systems in skill learning. Neuropsychology 11(2):272–281. Georgopoulos, A. P., 2000. Neural aspects of cognitive motor control. Curr. Opin. Neurobiol. 10(2):238–241. Gerard, E., and B. S. Peterson, 2003. Developmental processes and brain imaging studies in Tourette syndrome. J. Psychosom. Res. 55:13–22. Gerfen, C. R., T. M. Engber, L. C. Mahan, Z. Susel, T. N. Chase, F. J. Monsma, Jr., et al., 1990. D1 and D2 dopamine receptor–regulated gene expression of striatonigral and striatopallidal neurons. Science 250(4986):1429–1432. Giedd, J. N., J. L. Rapoport, M. A. Garvey, S. Perlmutter, and S. E. Swedo, 2000. MRI assessment of children with obsessivecompulsive disorder or tics associated with streptococcal infection. Am. J. Psychiatry 157(2):281–283. Gilbert, A. R., G. J. Moore, M. S. Keshavan, L. A. Paulson, V. Narula, F. P. Mac Master, et al., 2000. Decrease in thalamic volumes of pediatric patients with obsessive-compulsive disorder who are taking paroxetine. Arch. Gen. Psychiatry 57(5):449– 456. Gillies, A. J., and D. J. Willshaw, 1998. A massively connected subthalamic nucleus leads to the generation of widespread pulses. Proc. Biol. Sci. 265(1410):2101–2109. Goldman-Rakic, P. S., 1996. The prefrontal landscape: Implications of functional architecture for understanding human mentation and the central executive. Philos. Trans. R. Soc. London 351: 1445–1453. Goldman-Rakic, P., and L. Selemon, 1990. New frontiers in basal ganglia research. Trends Neurosci. 13:241–243. Gonzalez-Burgos, G., L. S. Krimer, N. V. Povysheva, G. Barrionuevo, and D. A. Lewis, 2005. Functional properties of fast spiking interneurons and their synaptic connections with pyramidal cells in primate dorsolateral prefrontal cortex. J. Neurophysiol. 93(2):942–953. Grafton, S. T., R. P. Woods, and T. Mike, 1994. Functional imaging of procedural motor learning: Relating cerebral blood flow with individual subject performance. Hum. Brain Mapping 1:221–234. Graybiel, A. M., 1990. Neurotransmitters and neuromodulators in the basal ganglia. Trends Neurosci. 13:244–254. Graybiel, A. M., 1995. Building action repertoires: Memory and learning functions of the basal ganglia. Curr. Opin. Neurobiol. 5(6): 733–741.

Graybiel, A. M., 2005. The basal ganglia: Learning new tricks and loving it. Curr. Opin. Neurobiol. 15(6):638–644. Greicius, M. D., B. Krasnow, A. L. Reiss, and V. Menon, 2003. Functional connectivity in the resting brain: A network analysis of the default mode hypothesis. Proc. Natl. Acad. Sci. USA 100(1): 253–258. Greicius, M. D., and V. Menon, 2004. Default-mode activity during a passive sensory task: Uncoupled from deactivation but impacting activation. J. Cogn. Neurosci. 16(9):1484–1492. Gusnard, D. A., E. Akbudak, G. L. Shulman, and M. E. Raichle, 2001. Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. Proc. Natl. Acad. Sci. USA 98(7):4259–4264. Gusnard, D. A., and M. E. Raichle, 2001. Searching for a baseline: Functional imaging and the resting human brain. Nature Rev. Neurosci. 2(10):685–694. Haaland, K. Y., D. L. Harrington, S. O’Brien, and N. Hermanowicz, 1997. Cognitive-motor learning in Parkinson’s disease. Neuropsychology 11(2):180–186. Hariri, A. R., V. S. Mattay, A. Tessitore, F. Fera, and D. R. Weinberger, 2003. Neocortical modulation of the amygdala response to fearful stimuli. Biol. Psychiatry 53(6):494–501. Harrington, D. L., K. Y. Haaland, R. A. Yeo, and E. Marder, 1990. Procedural memory in Parkinson’s disease: Impaired motor but not visuoperceptual learning. J. Clin. Exp. Neuropsychol. 12(2):323–339. Heindel, W. C., N. Butters, and D. P. Salmon, 1988. Impaired learning of a motor skill in patients with Huntington’s disease. Behav. Neurosci. 102(1):141–147. Heindel, W. C., D. P. Salmon, C. W. Shults, P. A. Walicke, and N. Butters, 1989. Neuropsychological evidence for multiple implicit memory systems: A comparison of Alzheimer’s, Huntington’s, and Parkinson’s disease patients. J. Neurosci. 9(2):582–587. Heinz, A., M. B. Knable, S. S. Wolf, D. W. Jones, J. G. Gorey, T. W. Hyde, et al., 1998. Tourette’s syndrome. [I-123]b-CIT SPECT correlates of vocal tic severity. Neurology 51:1069– 1074. Jog, M. S., C. I. Connolly, V. Hillegaart, M. Wilson, and A. M. Graybiel, 1996. Ensemble recordings from striatum of freely-behaving rats. Soc. Neurosci. Abstracts 22:1086. Jog, M. S., Y. Kubota, C. I. Connolly, V. Hillegaart, and A. M. Graybiel, 1999. Building neural representations of habits. Science 286(5445):1745–1749. Kalanithi, P. S., W. Zheng, Y. Kataoka, M. Difiglia, H. Grantz, C. B. Saper, et al., 2005. Altered parvalbumin-positive neuron distribution in basal ganglia of individuals with Tourette syndrome. Proc. Natl. Acad. Sci. USA 102(37):13307–13312. Keri, S., C. Szlobodnyik, G. Benedek, Z. Janka, and J. Gadoros, 2002. Probabilistic classification learning in Tourette syndrome. Neuropsychologia 40(8):1356–1362. Knowlton, B. J., J. A. Mangels, and L. R. Squire, 1996. A neostriatal habit learning system in humans. Science 273:1399– 1402. Knowlton, B. J., L. R. Squire, and M. A. Gluck, 1994. Probabilistic classification learning in amnesia. Learn. Memory 1:106– 120. Koos, T., and J. M. Tepper, 1999. Inhibitory control of neostriatal projection neurons by GABAergic interneurons. Nature Neurosci. 2(5):467–472. Koos, T., and J. M. Tepper, 2002. Dual cholinergic control of fast-spiking interneurons in the neostriatum. J. Neurosci. 22(2): 529–535.

marsh et al.: tics and compulsions

733

Laforce, R., Jr., and J. Doyon, 2001. Distinct contribution of the striatum and cerebellum to motor learning. Brain Cogn. 45(2): 189–211. Leckman, J. F., 2002. Tourette’s syndrome. Lancet 360(9345): 1577–1586. Leckman, J. F., 2003. Phenomenology of tics and natural history of tic disorders. Brain Dev. 25(Suppl. 1):S24–28. Leckman, J. F., D. E. Grice, J. Boardman, H. Zhang, A. Vitale, C. Bondi, et al., 1997. Symptoms of obsessive-compulsive disorder. Am. J. Psychiatry 154:911–917. Leckman, J. F., A. M. Knorr, A. M. Rasmusson, and D. J. Cohen, 1991. Basal ganglia research and Tourette’s syndromes [letter]. Trends Neurosci. 14(3):94. Leckman, J. F., B. S. Peterson, G. M. Anderson, A. F. Arnsten, D. L. Pauls, and D. J. Cohen, 1997a. Pathogenesis of Tourette’s syndrome. J. Child Psychol. Psychiatry 38(1):119–142. Leckman, J. F., B. S. Peterson, D. L. Pauls, and D. J. Cohen, 1997. Tic disorders. Psychiatr. Clin. North Am. 20(4):839–861. Leckman, J., and M. Riddle, 2000. Tourette’s syndrome: When habit-forming systems form habits of their own? Neuron 28: 349–354. Leckman, J. F., F. M. Vaccarino, P. S. Kalanithi, and A. Rothenberger, 2006. Annotation: Tourette syndrome: A relentless drumbeat—driven by misguided brain oscillations. J. Child Psychol. Psychiatry 47(6):537–550. Leckman, J. F., D. E. Walker, and D. J. Cohen, 1993. Premonitory urges in Tourette’s syndrome. Am. J. Psychiatry 150(1):98– 102. Leckman, J. F., D. E. Walker, W. K. Goodman, D. L. Pauls, and D. J. Cohen, 1994. “Just right” perceptions associated with compulsive behavior in Tourette’s syndrome. Am. J. Psychiatry 151(5):675–680. Leckman, J. F., C. B. Yeh, and D. J. Cohen, 2001. Tic disorders: When habit forming neural systems form habits of their own? Zhonghua Yi Xue Za Zhi (Taipei) 64(12):669–692. Leckman, J. F., H. Zhang, A. Vitale, F. Lahnin, K. Lynch, C. Bondi, et al., 1998. Course of tic severity in Tourette’s syndrome: The first two decades. Pediatrics 102:14–19. Leonard, H. L., E. L. Goldberger, J. L. Rapoport, D. L. Cheslow, and S. E. Swedo, 1990. Childhood rituals: Normal development or obsessive-compulsive symptoms? J. Am. Acad. Child Adolesc. Psychiatry 29(1):17–23. Leung, H. C., P. Skudlarski, J. C. Gatenby, B. S. Peterson, and J. C. Gore, 2000. An event-related functional MRI study of the stroop color word interference task. Cerebral Cortex 10(6): 552–560. Levesque, J., Y. Joanette, B. Mensour, G. Beaudoin, J. M. Leroux, P. Bourgouin, et al., 2004. Neural basis of emotional self-regulation in childhood. Neuroscience 129(2):361–369. Libet, B., 1985. Unconscious cerebral initiative and the role of conscious will in voluntary action. Behav. Brain Sci. 8:529–566. Lim, S. H., D. S. Dinner, P. K. Pillay, H. Luders, H. H. Morris, G. Klem, et al., 1994. Functional anatomy of the human supplementary sensorimotor area: Results of extraoperative electrical stimulation. Electroencephalogr. Clin. Neurophysiol. 91:179–193. Lin, H., C. B. Yeh, B. S. Peterson, L. Scahill, H. Grantz, D. B. Findley, et al., 2002. Assessment of symptom exacerbations in a longitudinal study of children with Tourette’s syndrome or obsessive-compulsive disorder. J. Am. Acad. Child Adolesc. Psychiatry 41(9):1070–1077. Luna, B., K. E. Garver, T. A. Urban, N. A. Lazar, and J. A. Sweeney, 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75(5):1357–1372.

734

Luna, B., and J. A. Sweeney, 2001. Studies of brain and cognitive maturation through childhood and adolescence: A strategy for testing neurodevelopmental hypotheses. Schizophr. Bull. 27(3): 443–455. Luna, B., and J. A. Sweeney, 2004. The emergence of collaborative brain function: fMRI studies of the development of response inhibition. Ann. NY Acad. Sci. 1021:296–309. Luna, B., K. R. Thulborn, D. P. Munoz, E. P. Merriam, K. E. Garver, N. J. Minshew, et al., 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 13(5):786–793. Luxenberg, J. S., S. E. Swedo, M. F. Flament, R. P. Friedland, J. Rapoport, and S. I. Rapoport, 1988. Neuroanatomical abnormalities in obsessive-compulsive disorder detected with quantitative X-ray computed tomography. Am. J. Psychiatry 145(9):1089–1093. MacDonald, A. W., 3rd, J. D. Cohen, V. A. Stenger, and C. S. Carter, 2000. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science 288(5472):1835–1838. Machlin, S. R., G. J. Harris, G. D. Pearlson, R. Hoehn-Saric, P. Jeffery, and E. E. Camargo, 1991. Elevated medial-frontal cerebral blood flow in obsessive-compulsive patients: A SPECT study. Am. J. Psychiatry 148:1240–1242. MacMaster, F. P., E. L. Dick, M. S. Keshavan, and D. R. Rosenberg, 1999. Corpus callosal signal intensity in treatment naive pediatric obsessive compulsive disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 23:601–612. Marsh, R., G. M. Alexander, M. G. Packard, H.-T. Zhu, and B. S. Peterson, 2005. Perceptual-motor skill learning in Tourette syndrome: Evidence for multiple procedural learning and memory systems. Neuropsychologia 43(10):1456–1465. Marsh, R., G. M. Alexander, M. G. Packard, H. Zhu, J. C. Wingard, G. Quackenbush, et al., 2004. Habit learning in Tourette syndrome: A translational neuroscience approach to a developmental psychopathology. Arch. Gen. Psychiatry 61(12): 1259–1268. Marsh, R., H. Zhu, R. T. Schultz, G. Quackenbush, J. Royal, P. Skudlarski, et al., 2006. A developmental fMRI study of self-regulatory control. Hum. Brain Mapping 27(11):848– 863. Marsh, R., H. Zhu, Z. Wang, P. Skudlarski, and B. S. Peterson, 2007. A developmental fMRI study of self-regulatory control in Tourette syndrome. Am. J. Psychiatry 164(6):955–966. McDonald, R. J., and N. M. White, 1993. A triple dissociation of memory systems: Hippocampus, amygdala, and dorsal striatum. Behav. Neurosci. 107(1):3–22. McGeorge, A. J., and R. L. Faull, 1989. The organization of the projection from the cerebral cortex to the striatum in the rat. Neuroscience 29(3):503–537. McGuire, P. K., C. J. Bench, C. D. Frith, I. M. Marks, R. S. Frackowiak, and R. J. Dolan, 1994. Functional anatomy of obsessive-compulsive phenomena. Br. J. Psychiatry 164(4):459– 468. McKiernan, K. A., B. R. D’Angelo, J. N. Kaufman, and J. R. Binder, 2006. Interrupting the “stream of consciousness”: An fMRI investigation. NeuroImage 29(4):1185–1191. McKiernan, K. A., J. N. Kaufman, J. Kucera-Thompson, and J. R. Binder, 2003. A parametric manipulation of factors affecting task-induced deactivation in functional neuroimaging. J. Cogn. Neurosci. 15(3):394–408. Milham, M. P., K. I. Erickson, M. T. Banich, A. F. Kramer, A. Webb, T. Wszalek, et al., 2002. Attentional control in the aging

neurodevelopmental aspects of clinical disorders

brain: Insights from an fMRI study of the stroop task. Brain Cogn. 49(3):277–296. Miller, E. K., and J. D. Cohen, 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24:167–202. Miller, G. A., 1956. The magic number seven, plus or minus two: Some limits on our capacity for processing information. Psychol. Rev. 63:81–97. Mink, J. W., 1996. The basal ganglia: Focused selection and inhibition of competing motor programs. Prog. Neurobiol. 50(4):381– 425. Mink, J. W., 2001. Basal ganglia dysfunction in Tourette’s syndrome: A new hypothesis. Pediatr. Neurol. 25(3):190–198. Moore, G. J., F. P. MacMaster, C. Stewart, and D. R. Rosenberg, 1998. Case study: Caudate glutamatergic changes with paroxetine therapy for pediatric obsessive-compulsive disorder. J. Am. Acad. Child Adolesc. Psychiatry 37(6):663–667. Nomoto, F., and Y. Machiyama, 1990. An epidemiological study of tics. Jpn. J. Psychiatry Neurol. 44:649–655. Nordahl, T. E., C. Benkelfat, W. E. Semple, M. Gross, A. C. King, and R. M. Cohen, 1989. Cerebral glucose metabolic rates in obsessive compulsive disorder. Neuropsychopharmacology 2:23– 28. Norman, D. A., and T. Shallice, 1986. Attention to action: Willed and automatic control of behaviour. In G. E. Schwartz and D. Shapiro, eds., Consciousness and Self-Regulation, vol. 4, pp. 1–18. New York: Plenum Press. Ochsner, K. N., S. A. Bunge, J. J. Gross, and J. D. Gabrieli, 2002. Rethinking feelings: An FMRI study of the cognitive regulation of emotion. J. Cogn. Neurosci. 14(8):1215–1229. Ongur, D., W. C. Drevets, and J. L. Price, 1998. Glial reduction in the subgenual prefrontal cortex in mood disorders. Proc. Natl. Acad. Sci. USA 95(22):13290–13295. Packard, M. G., R. Hirsh, and N. M. White, 1989. Differential effects of fornix and caudate nucleus lesions on two radial maze tasks: Evidence for multiple memory systems. J. Neurosci. 9(5): 1465–1472. Packard, M. G., and B. J. Knowlton, 2002. Learning and memory functions of the basal ganglia. Annu. Rev. Neurosci. 25:563–593. Packard, M. G., and J. L. McGaugh, 1992. Double dissociation of fornix and caudate nucleus lesions on acquisition of two water maze tasks: Further evidence for multiple memory systems. Behav. Neurosci. 106(3):439–446. Packard, M. G., and L. A. Teather, 1997. Double dissociation of hippocampal and dorsal-striatal memory systems by posttraining intracerebral injections of 2-amino-5-phosphonopentanoic acid. Behav. Neurosci. 111(3):543–551. Pardo, J. V., P. J. Pardo, K. W. Janer, and M. E. Raichle, 1990. The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc. Natl. Acad. Sci. USA 87:256–259. Parent, A., and L. Hazrati, 1995. Functional anatomy of the basal ganglia. I. The cortico–basal ganglia–thalamo–cortical loop. Brain Res. Rev. 20:91–127. Pasupathy, A., and E. K. Miller, 2005. Different time courses of learning-related activity in the prefrontal cortex and striatum. Nature 433(7028):873–876. Pauls, D. L., J. N. Alsobrook, W. Goodman, S. Rasmussen, and J. F. Leckman, 1995. A family study of obsessive-compulsive disorder. Am. J. Psychiatry 152(1):76–84. Pauls, D. L., and J. F. Leckman, 1986. The inheritance of Gilles de la Tourette’s syndrome and associated behaviors: Evidence for autosomal dominant transmission. N. Engl. J. Med. 315(16): 993–997.

Pauls, D. L., K. E. Towbin, J. F. Leckman, G. E. Zahner, and D. J. Cohen, 1986. Gilles de la Tourette’s syndrome and obsessive-compulsive disorder: Evidence supporting a genetic relationship. Arch. Gen. Psychiatry 43(12):1180–1182. Pessoa, L., S. Kastner, and L. G. Ungerleider, 2003. Neuroimaging studies of attention: From modulation of sensory processing to top-down control. J. Neurosci. 23(10):3990– 3998. Peterson, B. S., 2000. Neuroimaging studies of Tourette syndrome: A decade of progress. In D. J. Cohen, C. G. Goetz, and J. Jankovic, eds., Advances in Neurology: Tourette Syndrome and Associated Disorders, 179–196. Philadelphia: Lippincott Williams & Wilkins. Peterson, B. S., 2005. Clinical neuroscience and imaging studies of core psychoanalytic constructs. Clin. Neurosci. Res. 4(5):349– 365. Peterson, B. S., A. W. Anderson, R. Ehrenkranz, L. H. Staib, M. Tageldin, E. Colson, et al., 2003. Regional brain volumes and their later neurodevelopmental correlates in term and preterm infants. Pediatrics 111:938–948. Peterson, B. S., M. J. Kane, G. M. Alexander, C. Lacadie, P. Skudlarski, H. C. Leung, et al., 2002. An event-related functional MRI study comparing interference effects in the Simon and Stroop tasks. Brain Res. Cogn. Brain Res. 13(3):427–440. Peterson, B., and J. Klein, 1997. Neuroimaging of Tourette’s syndrome neurobiologic substrate. In B. S. Peterson, ed., Child Psychiatry Clinics of North America: Neuroimaging, vol. 6 (April), pp. 343–364. Philadelphia: W. B. Saunders. Peterson, B. S., J. F. Leckman, A. Arnsten, G. Anderson, L. Staib, J. Gore, et al., 1999. Neuroanatomical circuitry. In J. F. Leckman, ed., Tourette’s Syndrome—Tics, Obsessions, Compulsions; Developmental Psychopathology and Clinical Care, 230–259. New York: John Wiley & Sons. Peterson, B. S., D. S. Pine, P. Cohen, and J. Brook, 2001. A prospective, longitudinal study of tic, obsessive-compulsive, and attention deficit–hyperactivity disorders in an epidemiological sample. J. Am. Acad. Child Adolesc. Psychiatry 40:685–695. Peterson, B., M. A. Riddle, D. J. Cohen, L. D. Katz, J. C. Smith, M. T. Hardin, et al., 1993. Reduced basal ganglia volumes in Tourette’s syndrome using three-dimensional reconstruction techniques from magnetic resonance images [see comments]. Neurology 43(5):941–949. Peterson, B. S., P. Skudlarski, A. W. Anderson, H. Zhang, J. C. Gatenby, C. M. Lacadie, et al., 1998. A functional magnetic resonance imaging study of tic suppression in Tourette syndrome. Arch. Gen. Psychiatry 55:326–333. Peterson, B. S., P. Skudlarski, H. Zhang, J. C. Gatenby, A. W. Anderson, and J. C. Gore, 1999. An fMRI study of Stroop word-color interference: Evidence for cingulate subregions subserving multiple distributed attentional systems. Biol. Psychiatry 45:1237–1258. Peterson, B. S., L. Staib, L. Scahill, H. Zhang, C. Anderson, J. F. Leckman, et al., 2001. Regional brain and ventricular volumes in Tourette syndrome. Arch. Gen. Psychiatry 58:427– 440. Peterson, B. S., and P. Thomas, 2000. Functional brain imaging in Tourette’s syndrome: What are we really imaging? In M. Ernst and J. Rumsey, eds., Functional Neuroimaging in Child Psychiatry, 242–265. Cambridge, UK: Cambridge University Press. Peterson, B. S., P. Thomas, M. J. Kane, L. Scahill, Z. Zhang, R. Bronen, et al., 2003. Basal ganglia volumes in patients with Gilles de la Tourette syndrome. Arch. Gen. Psychiatry 60:415– 424.

marsh et al.: tics and compulsions

735

Peterson, B. S., H. Zhu, L. Scahill, L. Staib, Z. Zhang, R. Bronen, et al., submitted. Neuroanatomical subtyping of pediatric obsessive-compulsive disorder. Phillips, P. E., J. M. Johns, D. A. Lubin, E. A. Budygin, R. R. Gainetdinov, J. A. Lieberman, et al., 2003. Presynaptic dopaminergic function is largely unaltered in mesolimbic and mesostriatal terminals of adult rats that were prenatally exposed to cocaine. Brain Res. 961(1):63–72. Plessen, K. J., T. Wentzel-Larsen, K. Hugdahl, P. Feineigle, J. Klein, L. H. Staib, et al., 2004. Altered interhemispheric connectivity in individuals with Tourette’s disorder. Am. J. Psychiatry 161(11):2028–2037. Poldrack, R. A., V. Prabhakaran, C. A. Seger, and J. D. Gabrieli, 1999. Striatal activation during acquisition of a cognitive skill. Neuropsychology 13(4):564–574. Posner, M. I., 1994. Attention: The mechanisms of consciousness. Proc. Natl. Acad. Sci. USA 91:7398–7403. Posner, M. I., and M. K. Rothbart, 1998. Attention, selfregulation and consciousness. Philos. Trans. R. Soc. London [Biol.] 353(1377):1915–1927. Posner, M. I., and M. K. Rothbart, 2000. Developing mechanisms of self-regulation. Dev. Psychopathol. 12(3):427–441. Price, R. A., K. K. Kidd, D. J. Cohen, D. L. Pauls, and J. F. Leckman, 1985. A twin study of Tourette syndrome. Arch. Gen. Psychiatry 42(8):815–820. Przybyszewski, A. W., 1998. Vision: Does top-down processing help us to see? Curr. Biol. 8(4):R135–139. Rankins, D., J. L. Bradshaw, and N. Georgiou-Karistianis, 2006. The semantic Simon effect in Tourette’s syndrome and obsessive-compulsive disorder. Brain Cogn. 61(3):225–234. Rasmussen, S. A., and M. T. Tsuang, 1986. Clinical characteristics and family history in DSM-III obsessive-compulsive disorder. Am. J. Psychiatry 143:317–322. Rauch, S. L., D. D. Dougherty, D. Malone, A. Rezai, G. Friehs, A. J. Fischman, et al., 2006. A functional neuroimaging investigation of deep brain stimulation in patients with obsessivecompulsive disorder. J. Neurosurg. 104(4):558–565. Rauch, S. L., D. D. Dougherty, L. M. Shin, N. M. Alpert, P. Manzo, L. Leahy, et al., 1998. Neural correlates of factoranalyzed OCD symptom dimensions: A PET study. CNS Spectrums 3:37–43. Rauch, S. L., M. A. Jenike, N. M. Alpert, L. Baer, H. C. Breiter, C. R. Savage, et al., 1994. Regional cerebral blood flow measured during symptom provocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography. Arch. Gen. Psychiatry 51(1):62–70. Rauch, S. L., C. R. Savage, N. M. Alpert, D. Dougherty, A. Kendrick, T. Curran, et al., 1997. Probing striatal function in obsessive-compulsive disorder: A PET study of implicit sequence learning. J. Neuropsychiatry Clin. Neurosci. 9:568–573. Rauch, S. L., M. M. Wedig, C. I. Wright, B. Martis, K. G. McMullin, L. M. Shin, et al., 2007. Functional magnetic resonance imaging study of regional brain activation during implicit sequence learning in obsessive-compusive disorder. Biol. Psychiatry 61(3):330–336. Rauch, S. L., P. J. Whalen, C. R. Savage, T. Curran, A. Kendrick, H. D. Brown, et al., 1997. Striatal recruitment during an implicit sequence learning task as measured by functional magnetic resonance imaging. Hum. Brain Mapping 5(2):124–132. Raz, A., A. Feingold, V. Zelanskaya, E. Vaadia, and H. Bergman, 1996. Neuronal synchronization of tonically active neurons in the striatum of normal and parkinsonian primates. J. Neurophysiol. 76(3):2083–2088.

736

Raz, A., H. Zhu, S. Yu, R. Bansal, Z. Wang, G. M. Alexander, et al., in press. Neural substrates of self-regulatory control in children and adults with Tourette syndrome. Can. J. Psychiatry. Reber, M. E., B. J. Knowlton, L. R. Squire, 1996. Dissociable properties of memory systems: Differences in the flexibility of declarative and nondeclarative knowledge. Behav. Neurosci. 110(5):861–871. Ridderinkhof, K. R., M. W. van der Molen, G. P. Band, and T. R. Bashore, 1997. Sources of interference from irrelevant information: A developmental study. J. Exp. Child Psychol. 65(3): 315–341. Riffkin, J., M. Yucel, P. Maruff, S. J. Wood, B. Soulsby, J. Olver, et al., 2005. A manual and automated MRI study of anterior cingulate and orbito-frontal cortices, and caudate nucleus in obsessive-compulsive disorder: Comparison with healthy controls and patients with schizophrenia. Psychiatry Res. 138(2):99–113. Robbins, T. W., 1996. Dissociating executive functions of the prefrontal cortex. Philos. Trans. R. Soc. London [Biol.] 351:1463– 1471. Robinson, D., H. Wu, R. A. Munne, M. Ashtari, J. M. J. Alvir, G. Lerner, et al., 1995. Reduced caudate nucleus volume in obsessive-compulsive disorder. Arch. Gen. Psychiatry 52:393– 398. Rosenberg, D. R., D. H. Averbach, K. M. O’Hearn, A. B. Seymour, B. Birmaher, and J. A. Sweeney, 1997. Oculomotor response inhibition abnormalities in pediatric obsessivecompulsive disorder. Arch. Gen. Psychiatry 54(9):831–838. Rosenberg, D. R., N. R. Benazon, A. Gilbert, A. Sullivan, and G. J. Moore, 2000. Thalamic volume in pediatric obsessivecompulsive disorder patients before and after cognitive behavioral therapy. Biol. Psychiatry 48(4):294–300. Rosenberg, D. R., E. L. Dick, K. M. O’Hearn, and J. A. Sweeney, 1997. Response-inhibition deficits in obsessive-compulsive disorder: An indicator of dysfunction in frontostriatal circuits. J. Psychiatry Neurosci. 22(1):29–38. Rosenberg, D. R., and M. S. Keshavan, 1998. Toward a neurodevelopmental model of obsessive-compulsive disorder. Biol. Psychiatry 43(9):623–640. Rosenberg, D. R., M. S. Keshavan, E. L. Dick, W. W. Bagwell, F. P. MacMaster, and B. Birmaher, 1997. Corpus callosal morphology in treatment-naive pediatric obsessive compulsive disorder. Progr. Neuropsychopharmacol. Biol. Psychiatry 21(8):1269– 1283. Rosenberg, D. R., M. S. Keshavan, K. M. O’Hearn, E. L. Dick, W. W. Bagwell, A. B. Seymour, et al., 1997. Frontostriatal measurement in treatment-naive children with obsessivecompulsive disorder. Arch. Gen. Psychiatry 54:824–830. Rosenberg, D. R., F. P. MacMaster, M. S. Keshavan, K. D. Fitzgerald, C. M. Stewart, and G. J. Moore, 2000. Decrease in caudate glutamatergic concentrations in pediatric obsessivecompulsive disorder patients taking paroxetine. J. Am. Acad. Child Adolesc. Psychiatry 39(9):1096–1103. Rubia, K., S. Overmeyer, E. Taylor, M. Brammer, S. C. Williams, A. Simmons, et al., 2000. Functional frontalization with age: Mapping neurodevelopmental trajectories with fMRI. Neurosci. Biobehav. Rev. 24(1):13–19. Rubin, R. T., J. Ananth, J. Villanueva-Meyer, P. G. Trajmar, and I. Mena, 1995. Regionial [133]Xenon cerebral blood flow and cerebral [99m]Tc-HMPAO uptake in patients with obsessive-compulsive disorder before and during treatment. Biol. Psychiatry 38:429–437.

neurodevelopmental aspects of clinical disorders

Rubin, R. T., J. Villanueva-Meyer, J. Ananth, P. G. Trajmar, and I. Mena, 1992. Regional xenon 133 cerebral blood flow and cerebral technetium 99m HMPAO uptake in unmedicated patients with obsessive-compulsive disorder and matched normal control subjects: Determination by high-resolution single-photon emission computed tomography [see comments]. Arch. Gen. Psychiatry 49(9):695–702. Russell, A., B. Cortese, E. Lorch, J. Ivey, S. P. Banerjee, G. J. Moore, et al., 2003. Localized functional neurochemical marker abnormalities in dorsolateral prefrontal cortex in pediatric obsessive-compulsive disorder. J. Child Adolesc. Psychopharmacol. 13(Suppl. 1):S31–38. Saint-Cyr, J. A., 2003. Frontal-striatal circuit functions: Context, sequence, and consequence. J. Int. Neuropsychol. Soc. 9(1):103– 127. Sanes, J. N., 2003. Neocortical mechanisms in motor learning. Curr. Opin. Neurobiol. 13(2):225–231. Sarazin, M., B. Deweer, A. Merkl, N. Von Poser, B. Pillon, and B. Dubois, 2002. Procedural learning and striatofrontal dysfunction in Parkinson’s disease. Mov. Disorders 17(2):265–273. Sarter, M., B. Givens, and J. P. Bruno, 2001. The cognitive neuroscience of sustained attention: Where top-down meets bottom-up. Brain Res. Brain Res. Rev. 35(2):146–160. Saxena, S., A. L. Brody, J. M. Schwartz, and L. R. Baxter, 1998. Neuroimaging and frontal-subcortical circuitry in obsessive-compulsive disorder. Br. J. Psychiatry, Supplement(35):26– 37. Schiller, P. H., 1966. Developmental study of color-word interference. J. Exp. Psychol. 72(1):105–108. Seger, C. A., and C. M. Cincotta, 2005. The roles of the caudate nucleus in human classification learning. J. Neurosci. 25(11): 2941–2951. Serrien, D. J., M. Orth, A. H. Evans, A. J. Lees, and P. Brown, 2005. Motor inhibition in patients with Gilles de la Tourette syndrome: Functional activation patterns as revealed by EEG coherence. Brain 128(Part 1):116–125. Shulman, G. L., J. A. Fiez, M. Corbetta, R. L. Buckner, F. M. Miezin, M. E. Raichle, et al., 1997. Common blood flow changes across visual tasks: II. Decreases in cerebral cortex. J. Cogn. Neurosci. 9:648–663. Spessot, A. L., K. J. Plessen, and B. S. Peterson, 2004. Neuroimaging of developmental psychopathologies: The importance of self-regulatory and neuroplastic processes in adolescence. Ann. NY Acad. Sci. 1021:86–104. Squire, L. R., 1987. Memory and Brain. New York: Oxford University Press. Squire, L. R., and E. R. Kandel, 1999. Memory: From Mind to Molecules. New York: Scientific American Library. Squire, L. R., and S. M. Zola, 1996. Structure and function of declarative and nondeclarative memory systems. Proc. Natl. Acad. Sci. USA 93:13515–13522. Stebbins, G. T., J. Singh, J. Weiner, R. S. Wilson, C. G. Goetz, and J. D. Gabrieli, 1995. Selective impairments in memory functioning in unmedicated adults with Gilles de la Tourette’s syndrome. Neuropsychology 9(3):329–337. Stefanacci, L., W. A. Suzuki, and D. G. Amaral, 1996. Organization of connections between the amygdaloid complex and the

perirhinal and parahippocampal cortices in macaque monkeys. J. Comp. Neurol. 375:552–582. Stroop, J. R., 1935. Studies of interference in serial verbal reactions. J. Exp. Psychol. 18:643–662. Surmeier, D. J., A. Reiner, M. S. Levine, and M. A. Ariano, 1993. Are neostriatal dopamine receptors co-localized? Trends Neurosci. 16(8):299–305. Suzuki, W. A., and D. G. Amaral, 1994. Perirhinal and parahippocampal cortices of the macaque monkey: Cortical afferents. J. Comp. Neurol. 350:497–533. Swedo, S. E., J. L. Rapoport, H. Leonard, M. Lenane, and D. Cheslow, 1989. Obsessive-compulsive disorder in children and adolescents: Clinical phenomenology of 70 consecutive cases. Arch. Gen. Psychiatry 46(4):335–341. Swedo, S. E., M. B. Schapiro, C. L. Grady, D. L. Cheslow, H. L. Leonard, A. Kumar, et al., 1989. Cerebral glucose metabolism in childhood-onset obsessive-compulsive disorder. Arch. Gen. Psychiatry 46(6):518–523. Szeszko, P. R., D. Robinson, J. M. J. Alvir, R. M. Bilder, T. Lencz, M. Ashtari, et al., 1999. Orbital frontal and amygdala volume reductions in obsessive-compulsive disorder. Arch. Gen. Psychiatry 56:913–919. Tamm, L., V. Menon, and A. L. Reiss, 2002. Maturation of brain function associated with response inhibition. J. Am. Acad. Child Adolesc. Psychiatry 41(10):1231–1238. Tucker, D. M., P. Luu, and K. H. Pribram, 1995. Social and emotional self-regulation. Ann. NY Acad. Sci. 769:213–239. Ungerleider, L. G., J. Doyon, and A. Karni, 2002. Imaging brain plasticity during motor skill learning. Neurobiol. Learn. Mem. 78(3):553–564. Valleni-Basile, L. A., C. Z. Garrison, J. L. Waller, C. L. Addy, R. E. Mckeown, K. L. Jackson, et al., 1996. Incidence of obssessive-compulsive disorder in a community sample of young adolescents. J. Am. Acad. Child Adolesc. Psychiatry 35:898–906. Van der Linden, M., G. Ceschi, A. Zermatten, D. Dunker, and A. Perroud, 2005. Investigation of response inhibition in obsessive-compulsive disorder using the Hayling task. J. Int. Neuropsychol. Soc. 11(6):776–783. Verhulst, F. C., G. W. Akkerhuis, and M. Althaus, 1985. Mental health in Dutch children. I. A cross-cultural comparison. Acta Psychiatr. Scand. 323(Suppl. 72):1–108. Wichmann, T., and M. R. Delong, 1996. Functional and pathophysiological models of the basal ganglia. Curr. Opin. Neurobiol. 6(6):751–758. Wise, S. P., E. A. Murray, and C. R. Gerfen, 1996. The frontal cortex-basal ganglia system in primates. Crit. Rev. Neurobiol. 10(3–4):317–356. Witt, K., A. Nuhsman, and G. Deuschl, 2002. Dissociation of habit-learning in Parkinson’s and cerebellar disease. J. Cogn. Neurosci. 14(3):493–499. Wolf, S. S., D. W. Jones, M. B. Knable, J. G. Gorey, K. S. Lee, T. M. Hyde, et al., 1996. Tourette syndrome: Prediction of phenotypic variation in monozygotic twins by caudate nucleus D2 receptor binding. Science 273:1225–1227. Zheng, T., and C. J. Wilson, 2002. Corticostriatal combinatorics: The implications of corticostriatal axonal arborizations. J. Neurophysiol. 87(2):1007–1017.

marsh et al.: tics and compulsions

737

45

Developmental Dyslexia GUINEVERE F. EDEN AND D. LYNN FLOWERS

Developmental dyslexia is a learning disability that manifests during the early years of schooling as unexpected reading problems. Children who demonstrate otherwise typical intellectual skills and who are the recipients of conventional reading instructions struggle to pronounce words accurately and fluently and hence fail in their attempts to extract meaning from text (Lyon, Shaywitz, and Shaywitz, 2003). Developmental dyslexia, or as it is often referred to in the field of education, “reading disability,” affects between 5 and 12 percent of the population (Katusic et al., 2001), with boys more often affected than girls (Rutter et al., 2004), and presents an example where subtle deviations in cognitive development can lead to devastating consequences for a person in a literacy-driven society. Developmental dyslexia has drawn interest from researchers across a diverse range of disciplines, including neurology, child psychiatry, psychology, neuroscience, and education, and has thereby provided a lens by which to examine typical reading acquisition and its impairments from multiple perspectives. In the course of this research, interests have focused on the definition and diagnosis of reading disability, where the boundary should be placed to distinguish normal from abnormal reading, and under what conditions optimal reading acquisition occurs. More recent efforts have questioned to what degree reading skills are genetically rather than environmentally determined, the neurobiological basis of reading and reading disability, and avenues for intervention. As will become apparent, this work has been conducted employing various techniques and applying different theoretical frameworks. For example, while some of these are guided by neurological models of reading, others are informed by behavioral observations of typical reading acquisition; or a combination of these factors may be emphasized. The tendency to include behavioral, cognitive, and biological factors in theoretical frameworks of dyslexia is growing, as is the recognition that linguistic as well as sensorimotor systems must be incorporated during the testing of these frameworks. Ultimately, all these research avenues have a common goal: to better recognize the early signs of dyslexia and employ successful methods for its treatment. These studies shall be summarized in this chapter, framed by the relevant historical, cognitive, clinical, and educational context. The earliest recognition of dyslexia occurred in the late 19th century among members of the medical profession—for example, the ophthalmologist James Hinshelwood,

who described a boy with “word blindness” (Hinshelwood, 1896). Although it may seem intuitive to invoke such a visual-deficit hypothesis to account for dyslexia, more detailed neurological accounts were subsequently offered in the 20th century by the neuropathologist Samuel Orton in the United States and by the neurologist Macdonald Critchley in the United Kingdom, whose work combined sensorimotor and language-based aspects to reading—for example, recognizing that the condition involved “a primary disturbance of the sound-structure of the written words” (Schilder, 1944). The 1968 definition of dyslexia by the World Federation of Neurology influenced educational and research practices for decades (Critchley, 1970). Orton proposed that dyslexia involved a failure to represent print appropriately in the occipital poles (Orton, 1937). Although his hypothesis of ambiguous occipital dominance has not proven to be accurate, today’s brain-imaging studies have shown that numerous brain regions differ between dyslexic and nondyslexic groups, including the occipital cortex, but also temporal, parietal, frontal, and cerebellar regions. Attempts to integrate clinical observations with educational practices date back to Orton’s time, during which instructional approaches first emerged. These approaches used “multisensory-based” techniques to teach the features of spoken language incorporating motor, visual, auditory, and kinesthetic feedback still used today (Birsh, 1999). Although dyslexia has often been considered strictly from the neurobiological standpoint, there is a continuous effort to bridge the gap between clinical approaches, research, and education, so that knowledge about learning and acquisition of complex skills can inform educational practices in the context of reading in general, and reading disability specifically (Eden and Moats, 2002). Notably, research and educational practices have proven to be influential in the determination of the definition of developmental dyslexia.

Definition of dyslexia Dyslexia is a specific learning disability that is neurological in origin. It is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities. These difficulties typically result from a deficit in the phonological component of language that is often unexpected in relation to other cognitive abilities and the provision of effective classroom instruction. Secondary consequences may include problems in reading comprehension and reduced reading experience that can impede

739

growth of vocabulary and background knowledge. (Lyon, Shaywitz, and Shaywitz, 2003).

This most recent definition of dyslexia, which emerged from a working group commissioned by the National Institutes of Health and the International Dyslexia Association, has in common with numerous prior versions that dyslexia is described as a learning disability with features of “distinct” or “specific” reading-related deficits in spite of otherwise adequate health, ability, and access to instruction. Over the last two decades it has been generally accepted that word-decoding and reading-fluency deficiencies observed in dyslexia are the manifestations of an underlying deficit in phonological processes, described in more detail later in this chapter. The current definition also places an emphasis on a neurobiological origin, the nature of which is better understood today than when the definition was formulated in 1968, in which it was stated that dyslexia “is dependent upon fundamental cognitive disabilities which are frequently of constitutional origin” (World Federation of Neurology, 1968). Specifically, in the last three decades, microscopic anomalies have been reported from postmortem studies (Galaburda and Kemper, 1979), as well as structural brain anomalies (for review see Eckert, 2004) and functional anomalies (Brunswick et al., 1999; Eden et al., 1996; Flowers, Wood, and Naylor, 1991; Rumsey, Donohue, et al., 1997) identified from in vivo neuroimaging studies. As will be discussed in more detail, genetic origins have been revealed through linkage studies, and together with the brain-based findings, these provide the motivation for this portion of the definition (for review see McGrath, 2006). For practical purposes of service provision and also for research studies, the earlier definitions of dyslexia had relied on a discrepancy model—that is, a specified difference between expected and actual achievement. Commonly, IQ tests have been the benchmark for “expected” achievement, while a standardized reading achievement test provided the “actual” reading skill, the latter often constituting one or a combination of measures such as paragraph reading comprehension, single-word reading, and paragraph reading fluency. The amount of discrepancy defining specific reading disability has varied among research groups, from state to state, even among school systems, but typically at least a standard deviation (15 standard score points) between the two measures is mandated. However, in recent years researchers and educators have noted numerous problems associated with this discrepancy approach. First, there is a strong correlation between IQ and reading achievement, likely reflecting a decline over time among poorer readers resulting from the fact that verbal portions of the IQ are bolstered through reading experience (Siegel, 1989, 1992). Further, it has been shown that some of the core deficits

740

originally attributed to dyslexia are also observed in poor readers whose reading level is on par with their IQ (“gardenvariety poor readers”) and that both groups benefit from interventions that target these areas of phonological weakness (Stanovich, 1988). From the point of view of educational practice, the discrepancy formula is problematic, because eligibility may be delayed until a time when the discrepancy is large enough to meet the criteria for educational services, a “wait to fail approach.” Further, eligibility may be lost when a child moves to another geographical location or as a result of a positive response to intervention even though the deficit has not been entirely remediated (Fletcher et al., 1992; Meyer, 2000). In longitudinal studies, the IQ-discrepancy model has not been useful in predicting long-term outcome (Flowers et al., 2001), and, importantly, it does not provide clues to the most effective interventions. Consequently, it is now being advocated that an alternative, a “response to intervention” model, be used for addressing reading disability. Although it is too soon to assess its efficacy, this alternative approach tracks children’s responses to intervention throughout a three-tier model (Fletcher et al., 2005) with the level of instruction changing and intensifying with each tier (Tier 1, high-quality instruction in the general education setting; Tier 2, targeted intervention at the group or individual level; Tier 3, intervention in the special education setting). This new model is not without controversy, with some suggesting that response to intervention has simply replaced IQ in the discrepancy model (Gerber, 2005). Furthermore, it begs the question of which instructional approaches are best suited for which children.

Behavioral profile Poorly developed phonological skills are widely accepted to be at the basis of dyslexia because early phonological awareness and decoding skills predict the course of reading acquisition with a high degree of confidence (Schatschneider et al., 2004). As shall be discussed later, the fact that interventions addressing these phonological weaknesses are largely successful in bringing about gains in decoding in individuals with dyslexia (Bradley and Bryant, 1983; Torgesen et al., 2001) has further led to the proposal that a phonological core deficit hypothesis best describes the condition of dyslexia (Stanovich and Siegel, 1994). Specific examples of tests that involve phonological coding include phonemic awareness (sound manipulation at the phoneme level), phonological retrieval (rapid naming of objects, colors, letters, and numbers), and phonological recoding (short-term verbal memory) (Wagner et al., 1993; Wagner and Torgesen, 1987). Dyslexic children as early as kindergarten level are noted to differ from their peers with respect to awareness of the phonemic structure of spoken words (Wagner, 1986) and then

neurodevelopmental aspects of clinical disorders

to the ease with which they acquire grapheme-phoneme correspondence rules affecting their ability to decode words (Stanovich and West, 1989). However, other language-related processes may be impacted in dyslexia. For example, children with early language impairments are likely to go on to develop reading problems. These children may be diagnosed as dyslexic because of their eventual reading difficulties, but an evaluation for the expressive and receptive language skills may result in a diagnosis of specific language impairment (SLI). Many studies predicting reading skill from preschool testing have found that expressive and receptive language (especially syntactic skill), along with phonological awareness (as measured by syllable counting) and orthographic skill (as measured by visual matching), contribute significantly to reading achievement (Badian, 1982, 1998, 2001; Bryant et al., 1989, 1990; Scarborough, 1991; Schatschneider et al., 2004). In a review of the language basis of reading, Catts and Hogan (2003) point out that, whereas phonological processes are impaired in poor readers, numerous studies find that many poor readers also have deficits in other languagerelated processes, specifically vocabulary, morphology, and syntax, and text comprehension (referred to as “nonphonological language problems”). For example, students’ kindergarten scores on tests of nonphonological language measures (e.g., vocabulary and grammar) were strongly associated with reading outcome in second grade (Catts et al., 1999). Indeed, an assessment battery that added sentence imitation to the more typical tasks of phonological awareness, letter identification, and rapid naming was found to be 90 percent accurate in predicting second grade reading achievement (Catts et al., 2001). In addition to other underlying skills necessary for reading acquisition, orthographic skills—that is, recognizing whole words without decoding as well as recognizing the letter patterns that are “legal” in the written form of the language—are important for fluent reading (Badian, 2005). The relative ease of accessing the orthography appears to ultimately determine reading fluency and, therefore, to aid in comprehension (Goswami, 1990). While variance in early reading skills is best predicted by phonemic awareness and decoding, simple word calling and reading comprehension diverge in advancing grades. Indeed, some late-emerging poor readers only come to attention around fourth grade without having displayed the characteristic phonological problems (Leach, Scarborough, and Rescorla, 2003). Compensation in word reading, as well as in comprehension, is frequently observed. Deficits in the underlying mechanisms of phonemic awareness and fluency, however, do not normalize without explicit instruction (Bruck, 1992; Flowers, 1995). Finally, as reflected in the definition given before, dyslexic readers often present with poor spelling abilities, and these are also explained by their difficulties

in dealing with phonological coding (Perfetti, Rieben, and Fayol, 1997). Since the earliest definition of dyslexia, the diagnostic criteria have been largely exclusionary rather than inclusionary, thereby leading to the identification of heterogeneous populations of dyslexics, which in turn has led to some variability in the description of the dyslexic phenotype. In an attempt to characterize the wide-ranging deficits observed within this population, researchers have often attempted to identify subgroups. While these have largely been unfruitful for the purpose of understanding the etiology of dyslexia, identifying subtypes of impaired readers across a broader range of reading disabilities is helpful for treatment (Catts et al., 2005). Although the phonological deficits are prominent and most directly related to the process of reading acquisition, other behavioral manifestations have been observed and have given rise to alternative hypotheses.

Theories of dyslexia The behavioral manifestations, genetic contributions, and neural basis of dyslexia have been studied intensely as a way to better understand its etiology (for a recent thorough review see Vellutino et al., 2004). Data acquired at the behavioral and biological level have proven most useful when combined with theories at the cognitive level (Frith, 2001). In this regard, the position taken to describe dyslexia in terms of a phonological deficit has been fruitful; however, it has not sufficed in explaining the plurality of deficits seen in individuals with dyslexia. Other models have fared better at integrating the diverse symptoms observed in dyslexia and have explained these in terms of temporal integration (Tallal, 1980), magnocellular-based visual perceptual dysfunction (Lovegrove et al., 1980), and parietal (Stein, 2001) and cerebellar (Nicolson et al., 1999) deficits. Further, there exist conceptual frameworks by which all these models can be combined into one unifying theory (Ramus, 2004; Stein, 1993; Eden and Zeffiro, 1999). The work by Tallal and colleagues emerged from a series of studies investigating the relationship between early sensory processing of verbal and nonverbal sounds and its effects on phonological processing in individuals with specific language impairment (SLI). As will be discussed later, children with SLI present with expressive and receptive language problems, whereas children with dyslexia are identified based on their reading deficits. In their work, Tallal and colleagues found that individuals with SLI required longer time intervals between rapidly successive auditory inputs to discriminate or sequence them, suggesting a causal mechanism for impaired phoneme discrimination (Tallal and Piercy, 1973). To remediate this problem, computer-based training exercises were devised to drive processing of rapidly successive acoustic stimuli to faster rates, and to temporarily extend

eden and flowers: developmental dyslexia

741

acoustic speech stimuli to improve speech perception (Tallal, 2004). However, since demonstrations of improved language processing for children with SLI in the laboratory were carried out (Merzenich et al., 1996; Tallal et al., 1996), no independent, randomized, controlled studies have demonstrated long-term benefits to reading for children with SLI or dyslexia. The magnocellular visual system has been demonstrated to be abnormal in dyslexia using behavioral, electrophysiological, neuroanatomical, and functional neuroimaging techniques (Eden et al., 1996; Livingstone et al., 1991; Lovegrove et al., 1980). Tasks that involve perception of visual motion stimuli rely on magnocellular function and include the parietal cortex and area MT/V5 in the dorsal visual stream (Ungerleider, 1982). It has been suggested that the parietal impairments observed in dyslexia are associated with the magnocellular deficit (Eden, Wood, and Stein, 2003; Stein and Walsh, 1997), and one explanation put forward for a dual symptomatology of impaired visual motion perception and phonological processing is that cortical regions that are functionally specialized for these different tasks are colocalized in these regions of the brain. Hence, abnormalities in the vicinity of these areas in dyslexia may result in multiple behavioral manifestations (Eden and Zeffiro, 1999). A similar “twin-level” theoretical framework for dyslexia has been put forward with the cerebellum as the anatomical locus. Nicolson and Fawcett have provided an account for how reading and phonological deficits arise in parallel with poor skill automaticity, the latter being dependent on cerebellar integrity (Nicolson, Fawcett, and Dean, 2001). This theoretical framework can also been extended to the deficits observed in dyslexia during implicit learning (Howard et al., 2006) and motor learning (Nicolson et al., 1999). Automaticity is also the focus of the “double-deficit hypothesis” of dyslexia (Wolf, 1999; Wolf and Bowers, 1999), which proposes that deficits in phonological processing and deficits in visual naming speed (rapid automatized naming, also referred to as phonological retrieval; Wagner et al., 1993; Wagner and Torgesen, 1987) represent two independent causal deficits in dyslexia, and that children with both problems are more impaired and harder to remediate (Wolf and Obregon, 1992). These naming-speed deficits are thought to be brought about by impaired temporal integration of the phonological and visual counterparts of print. Remedial approaches that accompany this theory focus on fluencybased reading intervention at the level of phonological, orthographic, semantic, and lexical retrieval skills (Wolf, Miller, and Donnelly, 2000), the efficacy of which is still under investigation. These approaches have in common that they address sensorimotor as well as cognitive models of reading and offer a parsimonious account of the diverse deficits noted

742

in dyslexia. For more detailed descriptions of the research around these models we refer the reader to a thorough review of this work (Vellutino et al., 2004). However, several issues need to be kept in mind when considering some of the existing models. In many cases the causal relationship between sensorimotor deficits and reading is hypothetical, and empirical data are tenuous. Clearer answers may emerge when these symptoms are studied over a wide age range or longitudinally, in order to determine whether the prevalence of one type of deficit over another represents a developmental process. For example, it has been suggested that perceptual processes may be instrumental in establishing phonological representations and that the failure to do so results in more severe and easily documented deficits in phonological compared to sensory processing (Galaburda et al., 2006). Another explanation for why the presence of sensorimotor impairments is confined to a smaller section of the dyslexia population may reflect the fact that the neural representations of these skills are harder to perturb, with neurobiological anomalies having a relatively greater impact on cognitive over sensorimotor skills. The issue of causality with regard to cortical abnormalities has recently been reevaluated (Ramus, 2004), and seemingly disparate theories, the phonological theory and the magnocellular theory of dyslexia, may be accounted for by brain anomalies at the thalamic level as a result of cortical anomalies, possibly attributed to dyslexia susceptibility genes that are involved in neural migration (Galaburda et al., 2006). Core Cognitive Impairment: Phonological Representation As described previously, a number of theories of dyslexia have been proposed. An element common to most of these, or a prominent claim among them, is the notion that children and adults with dyslexia have problems with phonological processing. Following the definition of Scarborough and Brady, phonological representation (or phonological code) involves mentally represented information about the phonological characteristics of a particular word (Scarborough and Brady, 2002). Phonological awareness refers to a broad class of skills that involve attending to, thinking about, and intentionally manipulating the phonological aspects of spoken, as opposed to written, language. Phonemic/phoneme awareness is a kind of phonological awareness involving individual phonemes within spoken words and syllables. These have been implicated as the best predictor of children’s future reading and spelling levels at most developmental stages (Bradley and Bryant, 1983; Felton and Wood, 1989; Torgesen, Wagner, and Rashotte, 1994; Torgesen et al., 1990; Wagner et al., 1997; Wood and Grigorenko, 2001). Importantly, phonological awareness is the only skill for which strong evidence exists for a causal role in dyslexia.

neurodevelopmental aspects of clinical disorders

That is, children with poor phonological awareness likely go on to become poor readers. Teaching children who are at risk for reading problems with programs that emphasize phonological awareness facilitates reading acquisition. It has been suggested that strengthening phonological awareness skills can bring about reading gains in multiple domains (Stanovich, 1988). Most therapies for the remediation of dyslexia have, therefore, tended to target this domain (Alexander et al., 1991; Eden and Moats, 2002), whereas successful training studies in other domains are still lacking (McCardle, Scarborough, and Catts, 2001). The vast majority of this work has been conducted in alphabetic languages. However, phonological coding also has a place in logographic writing systems, such as Chinese, that do not allow for grapheme-phoneme mapping (Perfetti and Zhang, 1991). A logogram is a single written character that represents a complete word (more precisely, a morpheme). There is a role for both phonological and orthographic processing in Chinese character recognition, suggesting some common requirements for tackling logographic and alphabetic text (Perfetti, Liu, and Tan, 2005). In Chinese, despite such “universal” predictors of reading acquisition that transcend writing systems, rapid naming and motor skills involved in copying characters may be the best predictors of later reading skills (Tan et al., 2005). Interestingly, the neural signature of dyslexia in Chinese is quite distinct from that observed in alphabetic languages, and the hypoactivity observed in frontal systems of poor readers in China is likely to be related to the kinds of skills that also predict reading of Chinese (Siok et al., 2004).

Functional anatomy of typical reading and dyslexia How is the complex interplay of sensorimotor and cognitive systems necessary for reading represented in the brain? Positron emission tomography (PET) and fMRI studies have revealed the neural signature of adult normal letter processing (Flowers et al., 2004), reading (Bookheimer et al., 1995; Fiez and Petersen, 1998; Price and Friston, 1997; Pugh et al., 1996; Turkeltaub et al., 2002), and phonological representation (Gelfand and Bookheimer, 2003; Poldrack et al., 1999; Price, Moore, et al., 1997; Rumsey, Donohue, et al., 1997). Neuroimaging studies of reading in typically developing children have revealed activation of brain areas often associated with reading and language in adults, including left dominant areas of inferotemporal cortex, posterior superior temporal cortex, and inferior frontal cortex (Ahmad et al., 2003; Balsamo et al., 2002). However, adults have consistently exhibited greater activity in the dorsal left inferior frontal gyrus than children (Burgund et al., 2002). These findings may demonstrate the developmental engagement of phonological or semantic processing units for reading (Fiez, 1997; Pugh et al., 1996),

or they may simply reflect a more general maturation of the left inferior frontal cortex (Chugani, 1998; Huttenlocher and Dabholkar, 1997). To address the development of neural mechanisms for reading, a study conducted in our laboratory enrolled 41 healthy, good readers between the ages of 6 and 22 years (Turkeltaub et al., 2003) and asked them to perform an “implicit” reading task requiring detection of a visual feature within words and matched false font strings. This paradigm has been previously shown to activate reading-related areas in the brain, even though subjects are not instructed to read the words (Price, Wise, and Frackowiak, 1996) and simultaneously allows accuracy and reaction times to be matched across age. We found that learning to read was associated with increasing activity in the left middle temporal and inferior frontal gyri (figure 45.1a). To establish brain-behavioral correlates, we regressed brain activity evoked during this implicit reading task with measures of phonological awareness, phonetic retrieval, and phonological working memory. These three types of skills, which independently predict future achievement in reading, were related to three different networks of brain areas (figure 45.1b). Verbal working memory was concentrated in an area of the left inferior parietal sulcus. The measure of phonemic awareness was related to brain activity in the left superior temporal sulcus and the left inferior frontal gyrus. In contrast, phonetic retrieval was related to activity in a distributed network of bilateral frontal and temporal cortical areas. These results suggest that wide-ranging networks are at play in establishing the skills that are necessary for successful reading. However, at the same time, these studies raise the possibility that any of these networks could be perturbed in the case of developmental dyslexia. Brain-imaging studies have also characterized the anomalous patterns of neural activation associated with reading and phonological representation in adults with persistent or compensated developmental dyslexia (Brunswick et al., 1999; Demonet et al., 1992; Eden et al., 2004; Flowers, 1995; Horwitz, Rumsey, and Donohue, 1998; Ingvar et al., 2002; Paulesu et al., 1996; Pugh et al., 2000; Rumsey, Donohue, et al., 1997; Shaywitz et al., 1998). Employing various experimental approaches and paradigms (e.g., the detection or judgment of rhymes, nonword reading, and implicit reading), these studies have converged on findings of differential brain activity between dyslexics and nondyslexics underlying phonological and reading processes in left-hemisphere perisylvian regions. Differences in taskrelated signal change in the left temporoparietal and occipitotemporal cortices have emerged as the most consistent findings in studies of dyslexia in the alphabetic writing system. Specifically, hypoactivity in the parietal cortex is commonly reported in investigations of phonological representation and reading in dyslexic pediatric (Shaywitz

eden and flowers: developmental dyslexia

743

A

B

Figure 45.1 (A) Brain areas of increased (light) and decreased (dark) activity related to maturation of reading skills. (B) Neural correlates of reading and performance of phonemic awareness (light), phonological retrieval (dark), and verbal short-term memory

(dark and hashed). (Reprinted from Turkeltaub, P. E., L. Gareau, D. L. Flowers, T. A. Zeffiro, and G. F. Eden, 2003. Development of neural mechanisms for reading, Nature Neuroscience, 6(7):767–773. Copyright 2003, with permission.)

et al., 2002; Simos et al., 2000; Temple et al., 2001) and adult populations (Brunswick et al., 1999; Eden et al., 2004; Pugh et al., 2000; Rumsey, Donohue, et al., 1997; Shaywitz et al., 1998). The differences observed in dyslexic children are not the result of impoverished reading experience, as they can be observed in comparisons to both chronological- and reading-age-matched controls (Hoeft et al., 2006). In addition, there are reports of underactivity in the left occipitotemporal junction or BA 37 (Brunswick et al., 1999; Paulesu et al., 2001; Rumsey, Donohue, et al., 1997; Shaywitz et al., 1998) and hyperactivation in the left inferior frontal gyrus in adult dyslexics when performing some tasks (Brunswick et al., 1999; Rumsey, Donohue, et al., 1997; Shaywitz et al., 1998), but no difference or hypoactivation when performing other tasks (Brunswick et al., 1999; Eden et al., 2004; Paulesu et al., 1996; Rumsey, Nace, et al., 1997; Rumsey et al., 1994). The distribution of regions across the brain during these reading-related tasks implies that there is not simply a single area that is involved in reading or phonological processes and that depending on which of these areas is affected, there may be a number of different ways by which dyslexia may

arise. Also, whether these functional differences at the level of the cortex represent the primary site of pathology is unknown. Although the finding of reduced fractional anisotropy in left temporal parietal white matter in dyslexia (Klingberg et al., 2000; Niogi and McCandliss, 2006) is consistent with these functional findings and also with neuroanatomical findings (Eckert, 2004), the cause of this white matter pathology remains to be determined. It is interesting to note that these in vivo observations in dyslexia also dovetail with the abnormal cell formations found in posterior cerebral cortex postmortem (Galaburda and Kemper, 1979). As functional and anatomical findings converge, genetic research is playing an important role in shedding light on possible genetic influences on abnormal neuronal migrations.

744

Genetics There is hope that genetic research will eventually result in early screening for the risk factors of dyslexia. Twin studies find that components of reading skill as well as reading composite scores are strongly influenced by genes (Gayan and Olson, 2001; Wadsworth et al., 2000). However, the

neurodevelopmental aspects of clinical disorders

complex relationships among reading phenotypes make it highly unlikely that there is a “dyslexia gene.” Indeed, the evidence to date does not support a one-to-one relationship between an isolated reading component and a specific gene. Replicated linkage studies indicate regions of susceptibility on chromosomes 1, 2, 3, 6, 15, 18, and X. Four candidate genes on chromosomes 3, 6, and 15 have been identified that are associated with aspects of the dyslexia phenotype (Cope et al., 2005; Hannula-Jouppi et al., 2005; Meng et al., 2005; Taipale et al., 2003). These genes code for various aspects of brain development, such as axonal guidance (Hannula-Jouppi et al., 2005) and neuronal migration (Cope et al., 2005; Meng et al., 2005). As already insinuated, this result is consistent with the early findings of migration anomalies found in postmortem tissue by Galaburda and colleagues (1985) as well as the later findings based on animal models (Rosen, Sherman, and Galaburda, 1989; Sherman et al., 1990; Wang et al., 2006) and a human parallel in a Finnish cohort (Taipale et al., 2003). The four candidate genes have been proposed by Galaburda and colleagues to be involved in abnormal axonal growth and neuronal migration and in abnormal patterns of organization in subsequent cortical networks involved in the processes involved in reading acquisition (Galaburda, 2005; Galaburda et al., 2006). However, it is also important to note that the known functions of the candidate genes are global whereas dyslexia is a specific learning disorder (McGrath, 2006). Also, functional mutations have not been reliably identified in the candidate genes, pointing to mutations in regions that regulate those genes. It has been suggested that genetic influences may also provide an explanation for the co-occurrence of dyslexia with other developmental disorders and learning disabilities (Ramus, 2004).

Comorbid disorders Attention-Deficit/Hyperactivity Disorder (ADHD) It is commonly observed, and borne out in several large studies, that attention-deficit/hyperactivity disorder (ADHD) and reading disability often co-occur, as much as 40 percent of the time by some reports (Shaywitz, Fletcher, and Shaywitz, 1994; Shaywitz et al., 1995). Pennington (1991) observed that dyslexia is not more common in children with ADHD, but ADHD is often seen in children with dyslexia. It had been assumed that the two disorders were distinct and independent, since each disorder occurs in isolation or along with other disorders. While this assumption is true, and there is clearly heterogeneity within each disorder, recent genetic evidence points to a common etiology for their co-occurrence (Gayan et al., 2005; Gilger, Pennington, and DeFries, 1992). A report on sibling pairs by Gayan and colleagues suggests that a locus on chromosome 14 affects both ADHD and

reading disability (Gayan et al., 2005). Another study has implicated other loci including reading disorder susceptibility regions on chromosomes 2, 8, and 15 (Loo et al., 2004). Either disorder can impact learning, but the combination is especially forceful. Behaviorally, reading disability and ADHD are sometimes mistaken for one another, with a resulting loss of appropriate treatment (Knivsberg, Reichelt, and Modland, 1999). Attention-deficit/hyperactivity disorder and reading disability share some but not all deficits, suggesting that within the heterogeneity of these disorders, some subtypes may have shared etiology while others do not (deJong et al., 2006). The neurobiological focus within ADHD research has been quite distinct from that of dyslexia, with neuroimaging studies of ADHD examining frontal lobe function during tasks that tap executive control (Vaidya et al., 2005). Accurate diagnosis has important implications for treatment, which in the case of ADHD frequently involves pharmacological approaches. Dyslexia and Specific Language Impairment (SLI) Children with SLI have difficulties with expressive and receptive language. Reading impairment is not one of the criteria for receiving a diagnosis of SLI. However, reading deficits are commonly observed in this population, leading to a suggestion that children with phonological impairments in the presence of normal comprehension have dyslexia while children with deficits on both dimensions are diagnosed as SLI (Bishop and Snowling, 2004). Catts and colleagues directly examined the relationship between dyslexia and SLI by comparing the two disorders in isolation as well as in combination (Catts et al., 2005). They found that on measures of phonological processing skill (phonological awareness and nonword repetition) the SLI-only group had only mild deficits, while the dyslexic subjects, either with or without comorbid SLI, had significantly poorer phonological skills than either the SLI-only or a normally developing group. Thus it appears that although dyslexia and SLI are frequently found in the same individuals, and even though reading acquisition relies on intact language components, dyslexia does not inevitably emerge from developmental language impairment. Nor do all dyslexics have a history of language impairment. The existence of a group with language-comprehension weaknesses yet good wordidentification skills suggests that these reading skills are to some degree independent. Anatomical findings from studies employing brain imaging support the notion that phonological deficits and language comprehension in children with an array of reading and oral language skills have different neural origins (Eckert et al., 2003; Leonard et al., 2006). This has important implications for the choice of intervention, as it cannot be assumed that children with dyslexia and SLI can be remediated using identical strategies.

eden and flowers: developmental dyslexia

745

From the view of SLI, children with developmental language delays are likely to have problems in reading achievement. It is estimated that at least 50 percent of children with a diagnosis of SLI in early childhood are at increased risk for developing reading problems and other learning difficulties (Aram and Hall, 1989). While those with mild problems are at some risk, the severity and duration of language deficits impact the degree of the reading deficits (Bishop and Adams, 1990; Scarborough and Dobrich, 1990; Stothard et al., 1998).

Risk assessment and prediction of dyslexia The long history of reading research has reaped the benefit of revealing which specific skills measured in young children predict eventual reading acquisition. Several studies have now been conducted in a deliberate effort to identify the best combination of predictors of reading. The age at which initial assessment and outcome assessment are undertaken, as well as the choice of measures, influences the nature and strength of any predictions. For example, preschool testing is limited to those constructs that can be assessed in preliterate children, and these early language skills, such as mean length of utterance, phonological awareness, and visual perception, account for significant variance in reading in the early grades (Badian, 1998; Bryant et al., 1989, 1990; Scarborough, 1991). By the time a child has entered school, sentence or story recall, receptive language (sentence comprehension and picture-word association), expressive language (sentence completion and confrontation naming), rapid serial naming, phonological awareness, and letter name and sound knowledge each individually explain significant amounts of variance (on the order of 10% to 30%) in the early acquisition of reading skills (Felton et al., 1987; Scarborough, 1998; Snow, Burns, and Griffin, 1998; Snow, Scarborough, and Burns, 1999; Wolf and Obregon, 1992). In a large, randomly selected cohort, Schatschneider and colleagues used dominance analysis to test hypotheses regarding the efficacy of kindergarten testing at different time points on reading achievement in grades 1 and 2 (Schatschneider et al., 2004). Single-word identification, silent paragraph reading, and rapid word reading were the outcome measures for all of these. The best predictors were phonological awareness, rapid serial letter naming, and letter sound knowledge. Vocabulary, expressive and receptive language, and visual perceptual skills were only very weak predictors, accounting for 1 percent or less of the variance. However, as the authors point out, the sample did not include a large number of children with SLI, as the study by Catts and colleagues had done, and had a limited end point. Others who have followed children for a longer period (grades 4 and higher) and who have used a variety of tests

746

of comprehension have found a stronger relationship between early oral language skills and reading outcome (Catts et al., 1999; Storch and Whitehurst, 2002; Stothard et al., 1998). Because reading is a complex function that undoubtedly involves several skills, some researchers have attempted to formulate multivariate predictor models to identify those at risk for developing reading problems. Regression models are powerful in their ability to distinguish at-risk and good readers (Badian, 1982, 1998, 2001; Felton, 1992; Wood et al., 2005), and several predictive assessments have emerged. For example, Wood and colleagues followed a large randomly selected cohort from first to eighth grade and applied multiple regression analysis predicting broad reading scores from a variety of early skills (Wood et al., 2005). The model that emerged included four core skills, phonemic awareness, word reading, naming fluency, and vocabulary knowledge. These findings were subsequently cross-validated on a sample of 500 geographically diverse late kindergarten through third grade students whose ethnic and racial composition closely approximated the national early elementary school population. In the validation sample, the average Woodcock Johnson III Broad Reading score correlation with the model was .93. Sensitivity/specificity values for predicting the 15 percent and 30 percent cut-off for the Woodcock Johnson III were 91.3/88.0 and 94.1/89.1, respectively, making the model a powerful tool to identify children at risk for underdevelopment of reading skills as well as to predict typical or exceptional reading. Predictive tests such as these hold the promise of flagging children at risk for reading deficits, identifying specific areas of weakness, and using this information for refining remedial instruction.

Intervention Performance on measures of language and phonological coding has been used to build models that predict reading development. The same models may be used to guide appropriate instruction. In its 2000 review of reading instruction, the National Reading Panel (NRP) concluded that good reading instruction combines phonics, phonemic awareness, fluency, vocabulary, and reading comprehension. It also revealed that phonological awareness-based instruction significantly improved the reading performance of poor readers. This finding is consistent with a recent review of studies conducted in clinics and classrooms that have demonstrated that teaching the principles of phonological awareness to children can raise scores on reading and is the most effective approach to treating individuals with dyslexia (Alexander and Slinger-Constant, 2004). Usually these intensive instructions for students with dyslexia involve explicit and systematic teaching of the phonological code,

neurodevelopmental aspects of clinical disorders

repeated practice, small-group delivery, and the use of enhancing techniques (usually multisensory techniques linking listening, speaking, reading, and writing). For example, explicit, structured, systematic instruction targeting phonological awareness, especially when combined with decoding instruction, has been found to improve reading in poor readers (Torgesen, 1999; Torgesen et al., 2001; Wise, Ring, and Olson, 1999). A recent study suggests that early identification leading to kindergarten and first grade intervention may have preventative usefulness, at least through third grade, in at-risk children (Vellutino et al., 2006). Increasingly, with evidence that phonology and decoding alone are insufficient to support fully efficient reading comprehension, training in fluency is also included (Wolf, Miller, and Donnelly, 2000). As previously discussed, where listening comprehension is also weak, instruction aimed at vocabulary and grammatical structure may be necessary. A comprehensive review of the available programs aimed at improving the teaching of reading-related skills to young children is beyond the scope of this chapter. However, there is a vehicle for accessing information that promises to be helpful in determining the research-based effectiveness of available programs for young children. The What Works Clearinghouse (WWC) Intervention Report may be accessed online (http://ies.ed.gov/ncee/wwc). The advantage of using the information available on this site is that, while programs making claims of a positive response to intervention may have good face validity, many fail to provide rigorous research in support of these claims. The studies reported at the WWC are vetted for the use of randomized, controlled trials that meet the Department of Education evidence standards and are examined for conflict of interest, and the results are ranked according to the relative effectiveness of outcome in several areas. The specific studies reviewed, including those not meeting WWC standards, are summarized on this site with respect to the age of the children studied. To date, information is available on only a few programs, yet in the meantime the Florida Center for Reading Web site provides information on the nature of frequently used interventions (http://www.fcrr.org/ FCRRreports/). In addition, it cites related peer-reviewed publications and provides current reports from schools that have adopted these interventions. The latter, however, need to be considered with caution, as they often do not comply with rigorous research practices, such as the use of controlled, randomized experimental designs.

Neurobiological basis of reading intervention Whereas functional neuroimaging studies have identified differences in those areas of the brain subserving reading and phonological coding in dyslexic compared to nondyslexic

subjects, the same technology has been employed more recently to examine pediatric brain function following intensive reading interventions (Aylward et al., 2003; Shaywitz et al., 2004; Simos et al., 2002). In the first such study, children between the ages of 7 and 17 received an average of about 80 hours of intervention, which resulted in significant postintervention increases in brain activity (measured with magnetoencephelography, MEG) in the left posterior superior temporal gyrus, akin to that measured in the nondyslexic sample (Simos et al., 2002). Using a different variant on the paradigm and employing fMRI, Aylward also showed that following 28 hours of reading intervention, dyslexic children’s brain activity resembled that of the controls following the intervention but not prior to it (Aylward et al., 2003). When similar studies were conducted with the inclusion of a dyslexic control group that received a community-based rather than an experimental intervention, the experimental intervention, compared to a standard intervention, led to increases in activity in the left inferior frontal gyrus and posterior portion of the middle temporal gyrus (Shaywitz et al., 2004). However, these changes were also observed in a group of typical readers followed over the same time frame, suggesting that they are not uniquely associated with receiving reading intervention (unique changes in the readingdisabled group were constrained to increased caudate nucleus activity). Taken together, results from studies performed to date are varied, most likely because of methodological differences in subject recruitment and reading intervention. Although more studies are needed, revelation of the neural correlates of treatment can provide important insight into brain mechanisms of behavioral plasticity. Naturally, functional brainimaging studies of reading intervention are plagued by experimental design issues similar to those noted for behavioral studies. For example, there is a need to include a dyslexic control group, so that it can be inferred whether the observed gains are specific to the intervention or not (prepost differences in the dyslexic sample are not sufficient to establish efficacy of the intervention at hand). This requires treatment of a nonintervention dyslexic group to be withheld for the duration of the study, or the group may be provided with an alternative intervention, the results of which are compared to that of the experimental treatment. Studies of children carry their own complexities. Extraneous experience-driven and developmentally driven physiological changes are at work and may interact with the brain variables of interest with regard to intervention. Studies in adults offer a less dynamic situation and circumvent some of these issues. In our own work, we have examined the effect of an average of 120 hours of intensive reading intervention in dyslexic adults with persistent reading problems. Using fMRI, we examined the neural systems subserving phonological processing (using a sound deletion task) prior

eden and flowers: developmental dyslexia

747

LEFT

RIGHT

A Controls

B Dyslexics

C Controls> Dyslexics

D Intervention Dyslexics> Non-Intervention Dyslexics

Figure 45.2 Functional anatomy of phonological manipulation in (A) typical adult readers, (B) dyslexic adult readers, and (C) typical readers greater than dyslexic readers. Activation maps were generated by contrasting simple repetition of an aurally presented word with repeating words after performing sound deletion. Signal increases attributed to phonological manipulation were observed in left occipitotemporal, inferior parietal, and inferior frontal cortex in typical readers (A) and in bilateral inferior parietal, inferior frontal, middle temporal cortex, precuneus, and cerebellum in the dyslexic group (B). A between-group statistical comparison of the control and dyslexic groups revealed less activity in the dyslexic group in left inferior parietal regions (supramarginal and angular

gyri), superior parietal lobule, precuneus, and medial frontal gyrus and right hemisphere occipitotemporal junction (C). (D) Functional anatomy of phonological manipulation following reading remediation (group X session interaction) revealed increases during phonological manipulation in left parietal cortex and fusiform gyrus. Right hemisphere increases included posterior superior temporal sulcus/gyrus and parietal cortex. (Reprinted from Eden, G. F., K. M. Jones, K. Cappell, L. Gareau, F. B. Wood, T. A. Zeffiro, N. A. E. Dietz, J. A. Agnew, and D. L. Flowers, 2004. Neural changes following remediation in adult developmental dyslexia, Neuron, 44(3). Copyright 2004, with permission from Elsevier.)

to and following a phonologically based intervention (see figure 45.2) that resulted in improved performance in reading accuracy. Behavioral gains were associated with increased activity in bilateral parietal and right perisylvian cortex. These results suggest that in adult developmental dyslexia, behavioral plasticity involves two neural mechanisms, the same increases in the left hemisphere as reported for children (Aylward et al., 2003; Simos et al., 2002) and a right-hemisphere increase not unlike that observed in acquired reading disorders (Adair et al., 2000). Future studies will need to directly examine intervention-induced changes in children of different ages, to better understand the role of brain maturation in the context of brain plasticity.

Conclusions and future goals Sensorimotor and cognitive skills follow developmental trajectories that adhere to different time scales and are shaped by the interplay of biological factors (e.g., genetic predisposition, brain architecture) and environmental factors (e.g., instructions). A comprehensive understanding of the relationship that exists between learning, development, and experiential factors is urgently needed for the field of reading. The processing of written language is one of the most complex cognitive skills that we master, as it relies on the integration of numerous sensorimotor and cognitive processes. Their successful coordination, in turn, relies on explicit, staged instructions addressing numerous domains of oral and written language. Reading itself does not develop naturally, and the process of learning to read further promotes a number of skills that alter the brain’s function and anatomy (Castro-Caldas et al., 1998, 1999). How these can best be used to optimize reading skills of dyslexic readers is still a matter of investigation. This chapter serves to summarize the most relevant and prevalent findings in our current understanding of the reading disability developmental dyslexia. From these, a variety of theoretical frameworks have emerged to best explain the multiplicity of difficulties experienced by individuals with dyslexia, and consequently these have led to the advocacy of different types of interventions. A defining deficit in dyslexia is in phonological coding, the inability to understand how words are broken up into their constituent sounds, how these can be manipulated, and how they represent print. This core phonological coding deficit has become the primary target for intervention strategies in dyslexia, and future research will continue to quantify the efficacy of these approaches. Compounding the difficulty in establishing an allencompassing account by which to explain all the symptoms of dyslexia across the genetic, behavioral, and neurological levels is the fact that dyslexia is comorbid with other condi-

tions that also interfere with learning and language. At the same time, the study of the hereditary mechanisms of these co-occurring conditions may help to provide insight into dyslexia. Independent of these complexities, there is general agreement on the cognitive processes that are involved in reading. It is known that if these fail to become established, inaccurate and dysfluent reading follows. One of the more challenging questions revolves around the degree to which interventions are successful in bringing about improved ability across a range of reading skills, that is, not just reading accuracy, but also reading fluency and comprehension, the latter representing the raison d’être for reading. Not only is transfer of learning crucial for specific words (gains should not only be observed on trained exemplar items, but also transfer to untrained items), but dyslexic readers should be able to generalize their newly acquired skills to all domains of reading. Coupled with this realization, investigators who design brain-imaging studies to probe the neural correlates of reading gain will need to increase their repertoire of paradigms to evaluate the brain-based changes that occur beyond single-word reading. acknowledgments

The authors are supported by grants from the National Institute of Child Health and Human Development (HD40095, HD37890, HD21887). We thank Ashley Wall and Eileen Napoliello for preparation of this manuscript. REFERENCES

Adair, J. C., S. E. Nadeau, T. W. Conway, L. J. Gonzalez-Rothi, P. C. Heilman, I. A. Green, et al., 2000. Alterations in the functional anatomy of reading induced by rehabilitation of an alexic patient. Neuropsychiatry Neuropsychol. Behav. Neurol. 13(4): 303–311. Ahmad, Z., L. M. Balsamo, B. C. Sachs, B. Xu, and W. D. Gaillard, 2003. Auditory comprehension of language in young children: Neural networks identified with fMRI. Neurology 60(10): 1598–1605. Alexander, A. W., H. Anderson, P. Heilman, and K. Voelles, 1991. Phonological awareness training and remediation of analytic decoding deficits in a group of severe dyslexics. Ann. Dyslexia 41:193–206. Alexander, A. W., and A. M. Slinger-Constant, 2004. Current status of treatments for dyslexia: Critical review. J. Child Neurol. 19(10):744–758. Aram, D. M., and N. E. Hall, 1989. Longitudinal followup of children with preschool communication disorders: Treatment implications. School Psychol. Rev. 18(4):487– 501. Aylward, E. H., T. L. Richards, V. W. Berninger, W. E. Nagy, K. M. Field, A. C. Grimme, et al., 2003. Instructional treatment associated with changes in brain activation in children with dyslexia. Neurology 61(2):212–219. Badian, N., 1982. The prediction of good and poor reading before kindergarten entry: A 4-year follow-up. J. Special Educ. 16(3): 309–318.

eden and flowers: developmental dyslexia

749

Badian, N., 1998. A validation of the role of preschool phonological and orthographic skill in the prediction of reading. J. Learn. Disabil. 31(5):472–481. Badian, N., 2001. Phonological and orthographic processing: Their roles in reading prediction. Ann. Dyslexia 51:179–202. Badian, N. A., 2005. Does a visual-orthographic deficit contribute to reading disability? Ann. Dyslexia 55(1):28–52. Balsamo, L. M., B. Xu, C. B. Grandin, J. R. Petrella, S. H. Braniecki, T. K. Elliott, et al., 2002. A functional magnetic resonance imaging study of left hemisphere language dominance in children. Arch. Neurol. 59(7):1168–1174. Birsh, J., ed., 1999. Multisensory Teaching of Basic Language Skills. Baltimore, London, Sydney: Paul H. Brookes. Bishop, D. V. M., and C. Adams, 1990. A prospective study of the relationship between specific language impairment phonological disorders and reading retardation. J. Child Psychol. Psychiatry 31: 1027–1050. Bishop, D. V. M., and M. J. Snowling, 2004. Developmental dyslexia and specific language impairment: Same or different? Psychol. Bull. 130(6):858–886. Bookheimer, S. Y., T. A. Zeffiro, T. Blaxton, W. Gaillard, and W. Theodore, 1995. Regional cerebral blood flow during object naming and word reading. Hum. Brain Mapping 3:93– 106. Bradley, L., and P. Bryant, 1983. Categorizing sounds and learning to read—A causal connection. Nature 301:419–421. Bruck, M., 1992. Persistence of dyslexics’ phonological awareness deficits. Dev. Psychol. 28:874–886. Brunswick, N., E. McCrory, C. J. Price, C. D. Frith, and U. Frith, 1999. Explicit and implicit processing of words and pseudowords by adult developmental dyslexics: A search for Wernicke’s Wortschatz? Brain 122(Part 10):1901–1917. Bryant, P. E., L. L. Bradley, M. Camlean, and J. Crossland, 1989. Nursery rhymes, phonological skills and reading. J. Child Lang. 16:407–428. Bryant, P. E., M. MacLean, L. L. Bradley, and J. Crossland, 1990. Rhyme and alliteration, phoneme detection, and learning to read. Dev. Neuropsychol. 26(3):429–438. Burgund, E. D., H. C. Kang, J. E. Kelly, R. L. Buckner, A. Z. Snyder, S. E. Petersen, et al., 2002. The feasibility of a common stereotactic space for children and adults in fMRI studies of development. Neuroimage 17(1):184–200. Castro-Caldas, A., P. C. Miranda, I. Carmo, A. Reis, F. Leote, C. Ribeiro, et al., 1999. Influence of learning to read and write on the morphology of the corpus callosum. Eur. J. Neurol. 6(1): 23–28. Castro-Caldas, A., K. M. Petersson, A. Reis, S. Stone-Elander, and M. Ingvar, 1998. The illiterate brain: Learning to read and write during childhood influences the functional organization of the adult brain. Brain 121(Part 6):1053–1063. Catts, H. W., S. M. Adlof, T. P. Hogan, and S. E. Weismer, 2005. Are specific language impairment and dyslexia distinct disorders? J. Speech Lang. Hear. Res. 48:1378–1396. Catts, H. W., M. E. Fey, J. B. Tomblin, and X. Zhang, 2001. Estimating the risk of future reading difficulties in kindergarten children: A research-based model and its clinical implications. Lang. Speech Hear. Serv. Schools 32:38–50. Catts, H. W., M. E. Fey, X. Zhang, and J. B. Tomblin, 1999. Language basis of reading and reading disabilities. Sci. Stud. Reading 3:331–361. Catts, H. W., and T. P. Hogan, 2003. Language basis of reading disabilities and implications for early identification and remediation. Reading Psychol. 24:223–246.

750

Chugani, H. T., 1998. A critical period of brain development: Studies of cerebral glucose utilization with PET. Prev. Med. 27(2):184–188. Cope, N., D. Harold, G. Hill, V. Moskvina, J. Stevenson, P. Holmans, et al., 2005. Strong evidence that KIAA0319 on chromosome 6p is a susceptibility gene for developmental dyslexia. Am. J. Hum. Genet. 76(4):581–591. Critchley, M., 1970. Developmental dyslexia: A constitutional disorder of symbolic perception. Res. Publ. Assoc. Res. Nerv. Ment. Dis. 48:266–271. de Jong, C .G. W., J. Oosterlaan, and J. A. Sergeant, 2006. The role of double dissociation studies in the search for candidate endophenotypes for the comorbidity of attention deficit hyperactivity disorder and reading disability, Int J. Disability, Develop. Edu. 53(2):177–193. Demonet, J. F., F. Chollet, S. Ramsay, D. Cardebat, J. L. Nespoulous, R. Wise, et al., 1992. The anatomy of phonological and semantic processing in normal subjects. Brain 115(Part 6):1753–1768. Eckert, M., 2004. Neuroanatomical markers for dyslexia: A review of dyslexia structural imaging studies. Neuroscientist 10(4):362– 371. Eckert, M. A., C. M. Leonard, T. L. Richards, E. H. Aylward, J. Thomson, and V. W. Berninger, 2003. Anatomical correlates of dyslexia: Frontal and cerebellar findings. Brain 126(2):482– 494. Eden, G. F., K. M. Jones, K. Cappell, L. Gareau, F. B. Wood, T. A. Zeffiro, et al., 2004. Neural changes following remediation in adult developmental dyslexia. Neuron 44(3):411–422. Eden, G. F., and L. Moats, 2002. The role of neuroscience in the remediation of students with dyslexia. Nature Neurosci. 5(Suppl.):1080–1084. Eden, G. F., J. W. VanMeter, J. M. Rumsey, J. M. Maisog, R. P. Woods, and T. A. Zeffiro, 1996. Abnormal processing of visual motion in dyslexia revealed by functional brain imaging. Nature 382(6586):66–69. Eden, G. F., F. B. Wood, and J. F. Stein, 2003. Clock drawing in developmental dyslexia. J. Learn. Disabil. 36(3):216–228. Eden, G. F., and Zeffiro, T. A., 1998. Neural mechanisms affected in developmental dyslexia revealed by functional neuroimaging. Neuron 21:279–282. Felton, R. H., 1992. Early identification of children at risk for reading disabilities. Topics Early Child. Special Educ. 12:212– 229. Felton, R. H., and F. B. Wood, 1989. Cognitive deficits in reading disability and attention deficit disorder. J. Learn. Disabil. 22(1):3– 13, 22. Felton, R. H., F. B. Wood, I. S. Brown, S. K. Campbell, and M. R. Harter, 1987. Separate verbal memory and naming deficits in attention deficit disorder and reading disability. Brain Lang. 31(1):171–184. Fiez, J. A., 1997. Phonology, semantics, and the role of the left inferior prefrontal cortex. Hum. Brain Mapping 5(2):79–83. Fiez, J. A., and S. E. Petersen, 1998. Neuroimaging studies of word reading. Proc. Natl. Acad. Sci. USA 95(3):914–921. Fletcher, J. M., D. J. Francis, R. D. Morris, and G. R. Lyon, 2005. Evidence-based assessment of learning disabilities in children and adolescents. J. Clin. Child Adolesc. Psychol. 34(3):506– 522. Fletcher, J. M., D. J. Francis, B. P. Rourke, S. E. Shaywitz, and B. A. Shaywitz, 1992. The validity of discrepancy-based definitions of reading disabilities. J. Learn. Disabil. 25(9):555– 561.

neurodevelopmental aspects of clinical disorders

Flowers, D. L., 1995. Neuropsychological profiles of persistent reading disability and reading improvement. In C. K. J. Leong and R. M. Joshi, eds., Developmental and Acquired Dyslexia, 61–77. Netherlands: Kluwer Academic. Flowers, D. L., K. Jones, K. Noble, J. VanMeter, T. A. Zeffiro, F. B. Wood, et al., 2004. Attention to single letters activates left extrastriate cortex. Neuroimage 21(3):829–839. Flowers, D., M. Meyer, J. Lovato, F. Wood, and R. Felton, 2001. Does third grade discrepancy status predict the course of reading development? Ann. Dyslexia 51:49–74. Flowers, D. L., F. B. Wood, and C. E. Naylor, 1991. Regional cerebral blood flow correlates of language processes in reading disability. Arch. Neurol. 48(6):637–643. Frith, U., 2001. What framework should we use for understanding developmental disorders? Dev. Neuropsychol. 20(2):555–563. Galaburda, A. M., 2005. Dyslexia—A molecular disorder of neuronal migration. Ann. Dyslexia 55(2):151–165. Galaburda, A. M., and T. L. Kemper, 1979. Cytoarchitectonic abnormalities in developmental dyslexia: A case study. Ann. Neurol. 6:94–100. Galaburda, A. M., J. LoTurco, F. Ramus, R. H. Fitch, and D. G. Rosen, 2006. From genes to behavior in developmental dyslexia. Nature Neurosci. 9(10):1213–1217. Galaburda, A. M., G. F. Sherman, D. G. Rosen, F. Aboitiz, and N. Geschwind, 1985. Developmental dyslexia: Four consecutive patients with cortical anomalies. Ann. Neurol. 18(2): 222–233. Gayan, J., and R. K. Olson, 2001. Genetic and environmental influences on orthographic and phonological skills in children with reading disabilities. Dev. Neuropsychol. 20:483–507. Gayan, J., E. G. Willcutt, S. E. Fisher, C. Francks, L. R. Cardon, R. K. Olson, et al., 2005. Bivariate linkage scan for reading disability and attention-deficit/hyperactivity disorder localizes pleiotropic loci. J. Child Psychol. Psychiatry 46(10): 1045–1056. Gelfand, J. R., and S. Y. Bookheimer, 2003. Dissociating neural mechanisms of temporal sequencing and processing phonemes. Neuron 38(5):831–842. Gerber, M. M., 2005. Teachers are still the test: Limitations of response to instruction strategies for identifying children with learning disabilities, J. Learn. Disabil. 38(6):516–524. Gilger, J. W., B. F. Pennington, and J. C. DeFries, 1992. A twin study of the etiology and comorbidity: Attention-deficit hyperactivity disorder and dyslexia. J. Am. Acad. Child Adolesc. Psychiatry 31:343–348. Goswami, U., 1990. Phonological priming and orthographic analogies in reading. J. Exp. Child Psychol. 49(2):323–340. Hannula-Jouppi, K., N. Kaminen-Ahola, M. Taipate, R. Eklund, J. Nopola-Hemmi, H. Kaariainen, et al., 2005. The axon guidance receptor on ROBO1 is a candidate gene for developmental dyslexia. PLoS Genet. 1(4):e50. Hinshelwood, J., 1896. A case of dyselxia: A peculiar form of word-blindness. Lancet 2:1451–1454. Hoeft, F., A. Hernandez, G. McMillon, H. Taylor-Hill, J. L. Martindale, A. Meyler, et al., 2006. Neural basis of dyslexia: A comparison between dyslexic and nondyslexic children equated for reading ability. J. Neurosci. 26(42):10700– 10708. Horwitz, B., J. M. Rumsey, and B. C. Donohue, 1998. Functional connectivity of the angular gyrus in normal reading and dyslexia. Proc. Natl. Acad. Sci. USA 95(15):8939–8944. Howard, J. H., Jr., D. V. Howard, K. C. Japikse, and G. F. Eden, 2006. Dyslexics are impaired on implicit higher-order sequence

learning, but not on implicit spatial context learning. Neuropsychologia 44(7):1131–1144. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387(2):167–178. Ingvar, M., P. af Trampe, T. Greitz, L. Eriksson, S. StoneElander, and C. von Euler, 2002. Residual differences in language processing in compensated dyslexics revealed in simple word reading tasks. Brain Lang. 83(2):249–267. Katusic, S. K., R. C. Colligan, W. J. Barbaresi, D. J. Schaid, and S. J. Jacobsen, 2001. Incidence of reading disability in a population-based birth cohort, 1976–1982, Rochester, Minn. Mayo Clin. Proc. 76(11):1081–1092. Klingberg, T., M. Hedehus, E. Temple, T. Salz, J. D. Gabrieli, M. E. Moseley, et al., 2000. Microstructures of temporoparietal white matter as a basis for reading ability: Evidence from diffusion tensor magnetic resonance imaging. Neuron 25(2): 493–500. Knivsberg, A.-M., K.-L. Reichelt, and M. Modland, 1999. Comorbidity, or coexistence, between dyslexia and attention deficit hyperactivity disorder. Br. J. Special Educ. 26(1):42– 47. Leach, J. M., H. S. Scarborough, and L. Rescorla, 2003. Lateemerging reading disabililties. J. Learn. Disabil. 95(2):211–224. Leonard, C. M., M. A. Eckert, B. K. Given, V. W. Berninger, and G. F. Eden, 2006. Individual differences in anatomy predict reading and oral language deficits. Brain 129(Part 12):3329– 3342. Livingstone, M. S., G. D. Rosen, F. W. Drislane, and A. M. Galaburda, 1991. Physiological and anatomical evidence for a magnocellular deficit in developmental dyslexia. Proc. Natl. Acad. Sci. USA 88:7943–7947. Loo, S. K., S. E. Fisher, C. Francks, M. N. Ogdie, I. L. MacPhie, M. Yang, et al., 2004. Genome-wide scan of reading ability in affected sibling pairs with attention-deficit/hyperactivity disorder: Unique and shared genetic effects. Mol. Psychiatry 9(5): 485–493. Lovegrove, W. J., A. Bowling, D. Badcock, and M. Blackwood, 1980. Specific reading disability: Differences in contrast sensitivity as a function of spatial frequency. Science 210(4468):439– 440. Lyon, G. R., S. E. Shaywitz, and B. A. Shaywitz, 2003. A definition of dyslexia. Ann. Dyslexia 53:1–14. McCardle, P., H. S. Scarborough, and H. W. Catts, 2001. Predicting, explaining, and preventing children’s reading difficulties. Learn. Disabil. Res. Pract. 16(4):230–239. McGrath, L. M., 2006. Breakthroughs in the search for dyslexia candidate genes. Trends Mol. Med. 12(7):333–341. Meng, H., S. D. Smith, K. Hager, M. Held, J. Liu, R. K. Olson, et al., 2005. DCDC2 is associated with reading disability and modulates neuronal development in the brain. Proc. Natl. Acad. Sci. USA 102:17053–17058. Merzenich, M. M., W. M. Jenkins, P. Johnston, C. Schreiner, S. L. Miller, and P. Tallal, 1996. Temporal processing deficits of language-learning impaired children ameliorated by training. Science 271(5245):77–81. Meyer, M. S., 2000. The ability-achievement discrepancy: Does it contribute to an understanding of learning disabilities? Educ. Psychol. Rev. 12:315–337. Nicolson, R. I., A. J. Fawcett, E. L. Berry, I. H. Jenkins, P. Dean, and D. J. Brooks, 1999. Association of abnormal cerebellar activation with motor learning difficulties in dyslexic adults. Lancet 353(9165). 1662–1667.

eden and flowers: developmental dyslexia

751

Nicolson, R., A. J. Fawcett, and P. Dean, 2001. Dyslexia, development and the cerebellum. Trends Neurosci. 24(9):515– 516. Niogi, S. N., and B. D. McCandliss, 2006. Left lateralized white matter microstructure accounts for individual differences in reading ability and disability. Neuropsychologia 44(11):2178– 2188. Orton, S. T., 1937. Reading, Writing, and Speech Problems in Children and Selected Papers. Baltimore: International Dyslexia Association. Paulesu, E., J.-F. Demonet, F. Fazio, E. McCrory, V. Chanoine, N. Brunswick, et al., 2001. Dyslexia: Cultural diversity and biological unity. Science 291:2165. Paulesu, E., C. Frith, M. Snowling, A. Gallagher, J. Morton, R. S. J. Frackowiak, et al., 1996. Is developmental dyslexia a disconnection syndrome? Evidence from PET scanning. Brain 119:143–157. Pennington, B. F., 1991. Diagnosing Learning Disorders. New York: Guilford Press. Perfetti, C. A., Y. Liu, and L. H. Tan, 2005. The lexical constituency model: Some implications of research on Chinese for general theories of reading. Psychol. Rev. 112(1):43–59. Perfetti, C. A., L. Rieben, and M. Fayol, eds., 1997. Learning to Spell: Research, Theory, and Practice Across Languages. Mahwah, NJ: Erlbaum. Perfetti, C. A., and S. L. Zhang, 1991. Phonological processes in reading Chinese characters. J. Exp. Psychol. [Learn. Mem. Cogn.] 17(4):633–643. Poldrack, R. A., A. D. Wagner, M. W. Prull, J. E. Desmond, G. H. Glover, and J. D. Gabrieli, 1999. Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. Neuroimage 10(1):15–35. Price, C. J., and K. J. Friston, 1997. The temporal dynamics of reading: A PET study. Proc. R. Soc. Lond. [Biol.] 264(1389): 1785–1791. Price, C. J., C. J. Moore, G. W. Humphreys, and R. J. S. Wise, 1997. Segregating semantic from phonological processes during reading. J. Cogn. Neurosci. 9(6):727–733. Price, C. J., R. J. Wise, and R. S. Frackowiak, 1996. Demonstrating the implicit processing of visually presented words and pseudowords. Cerebral Cortex 6(1):62–70. Pugh, K. R., W. E. Mencl, A. R. Jenner, L. Katz, S. J. Frost, J. R. Lee, et al., 2000. Functional neuroimaging studies of reading and reading disability (developmental dyslexia). Ment. Retard. Dev. Disabil. Res. Rev. 6(3):207–213. Pugh, K. R., B. A. Shaywitz, S. E. Shaywitz, R. T. Constable, P. Skudlarski, R. K. Fulbright, et al., 1996. Cerebral organization of component processes in reading. Brain 119(Part 4): 1221–1238. Ramus, F., 2004. Neurobiology of dyslexia: A reinterpretation of the data. Trends Neurosci. 27(12):720–726. Rosen, D. G., G. F. Sherman, and A. M. Galaburda, 1989. Interhemispheric connections differ between symmetrical and asymmetrical brain regions. Neuroscience 33:525–533. Rumsey, J. M., B. C. Donohue, D. R. Brady, K. Nace, J. N. Giedd, and P. Andreason, 1997. A magnetic resonance imaging study of planum temporale asymmetry in men with developmental dyslexia. Arch. Neurol. 54(12):1481– 1489. Rumsey, J. M., K. Nace, B. Donohue, D. Wise, J. M. Maisog, and P. Andreason, 1997. A positron emission tomographic study of impaired word recognition and phonological processing in dyslexic men. Arch. Neurol. 54(5):562–573.

752

Rumsey, J. M., A. J. Zametkin, P. Andreason, A. P. Hanahan, S. D. Hamburger, T. Aquino, et al., 1994. Normal activation of frontotemporal language cortex in dyslexia, as measured with oxygen 15 positron emission tomography. Arch. Neurol. 51(1): 27–38. Rutter, M., A. Caspi, D. Fergusson, L. J. Horwood, R. Goodman, B. Maughan, et al., 2004. Sex differences in developmental reading disability: New findings from 4 epidemiological studies. JAMA 291(16):2007–2012. Scarborough, H. S., 1991. Early syntactic development of dyslexic children. Ann. Dyslexia 41:207–220. Scarborough, H. S., 1998. Early identification of children at risk for reading disabilities: Phonological awareness and some other promising predictors. In B. K. Shapiro, P. J. Accardo, and A. J. Capute, eds., Specific Reading Disability: A View of the Spectrum, 75–119. Timonium, MD: York Press. Scarborough, H. S., and S. A. Brady, 2002. Toward a common terminology for talking about speech and reading: A glossary of the “phon” words and some related terms. J. Literacy Res. 34(3):299–336. Scarborough, H., and W. Dobrich, 1990. Development of children with early delay. J. Speech Lang. Hear. Res. 33:70– 83. Schatschneider, C., J. M. Fletcher, D. J. Francis, C. D. Carlson, and B. R. Foorman, 2004. Kindergarten predictions of reading skills: A longitudinal comparative analysis. J. Educ. Psychol. 96(2):265–282. Schilder, P., 1944. Congenital alexia and its relation to optic perception. J. Genet. Psychol. 65:67–88. Shaywitz, B. A., S. E. Shaywitz, B. A. Blachman, K. R. Pugh, R. K. Fulbright, P. Skudlarski, et al., 2004. Development of left occipitotemporal systems for skilled reading in children after a phonologically based intervention. Biol. Psychiatry 55(9): 926–933. Shaywitz, B. A., S. E. Shaywitz, K. R. Pugh, W. E. Mencl, R. K. Fulbright, P. Skudlarski, et al., 2002. Disruption of posterior brain systems for reading in children with developmental dyslexia. Biol. Psychiatry 52(2):101–110. Shaywitz, S. E., J. M. Fletcher, and B. A. Shaywitz, 1994. Issues in the definition and classification of attention deficit disorder. Topics Lang. Disord. 14(4):1–25. Shaywitz, S. E., B. A. Shaywitz, K. R. Pugh, R. K. Fulbright, R. T. Constable, W. E. Mencl, et al., 1998. Functional disruption in the organization of the brain for reading in dyslexia. Proc. Natl. Acad. Sci. USA 95(5):2636–2641. Sherman, G. F., L. Morrison, D. G. Rosen, P. O. Behan, and A. M. Galaburda, 1990. Brain abnormalities in immune defective mice. Brain Res. 532:25–33. Siegel, L. S., 1989. IQ is irrelevant to the definition of learning disabilities. J. Learn. Disabil. 22(8):469–478, 486. Siegel, L. S., 1992. An evaluation of the discrepancy definition of dyslexia. J. Learn. Disabil. 25(10):618–629. Simos, P. G., J. I. Breier, J. M. Fletcher, B. R. Foorman, E. Bergman, K. Fishbeck, et al., 2000. Brain activation profiles in dyslexic children during nonword reading: A magnetic source imaging study. Neurosci. Lett. 290(1):61–65. Simos, P. G., J. M. Fletcher, E. Bergman, J. I. Breier, B. R. Foorman, E. M. Castillo, et al., 2002. Dyslexia-specific brain activation profile becomes normal following successful remedial training [see comments]. Neurology 58(8):1203–1213. Siok, W. T., C. A. Perfetti, Z. Jin, and L. H. Tan, 2004. Biological abnormality of impaired reading is constrained by culture. Nature 431(7004):71–76.

neurodevelopmental aspects of clinical disorders

Snow, C. E., M. S. Burns, and P. Griffin, 1998. Preventing Reading Difficulties in Young Children. Washington, DC: National Academy Press. Snow, C. E., H. S. Scarborough, and M. S. Burns, 1999. What speech-language pathologists need to know about early reading. Topics Lang. Disord. 20(1):48–58. Stanovich, K. E., 1988. Explaining the differences between the dyslexic and the garden-variety poor reader: The phonologicalcore variable-difference model. J. Learn. Disabil. 21(10):590– 604. Stanovich, K. E., and L. S. Siegel, 1994. Phenotypic performance profile of children with reading disabilities: A regression-based test of the phonological-core variable-difference model. J. Educ. Psychol. 86:24–53. Stanovich, K. E., and R. F. West, 1989. Exposure to print and orthographic processing. Reading Res. Q. 24:402–433. Stein, J. F., 1993. Dyslexia—Impaired temporal information processing? In P. Tallal, A. M. Galaburda, R. R. Llinas, and C. von Euler, eds., Temporal Information Processing in the Nervous System: Special Reference to Dyslexia and Dysphasia, vol. 14, pp. 83–86. New York: New York Academy of Sciences. Stein, J., 2001. The magnocellular theory of developmental dyslexia. Dyslexia 7(1):12–36. Stein, J., and V. Walsh, 1997. To see but not to read: The magnocellular theory of dyslexia. Trends Neurosci. 20(4):147–152. Storch, S. A., and G. J. Whitehurst, 2002. Oral language and code-related precursors to reading: Evidence from a longitudinal structural model. Dev. Psychol. 38:934–947. Stothard, S. E., M. Snowling, D. V. M. Bishop, B. B. Chipchase, and C. A. Kaplan, 1998. Language-impaired preschoolers: A follow-up into adolescence. J. Speech Lang. Hear. Res. 41(2):407. Taipale, M., N. Kaminen, J. Nopola-Hemmi, T. Haltia, B. Hyllyluoma, H. Lyytinen, et al., 2003. A candidate gene for developmental dyslexia encodes a nuclear tetratricopeptide repeat domain protein dynamically regulated in brain. Proc. Natl. Acad. Sci. USA 100(20):11553–11558. Tallal, P., 1980. Auditory temporal perception, phonics, and reading disabilities in children. Brain Lang. 9:182–198. Tallal, P., 2004. Improving language and literacy is a matter of time. Nature Rev. Neurosci. 5(9):721–728. Tallal, P., S. L. Miller, G. Bedi, G. Byma, X. Wang, S. S. Nagarajan, et al., 1996. Language comprehension in languagelearning impaired children improved with acoustically modified speech. Science 271(5245):81–84. Tallal, P., and M. Piercy, 1973. Defects of nonverbal auditory perception in children with developmental aphasia. Nature 241(5390):468–469. Tan, L., J. Spinks, G. Eden, C. Perfetti, and W. Siok, 2005. Reading depends on writing, in Chinese. Proc. Natl. Acad. Sci. USA 102:8781–8785. Temple, E., R. A. Poldrack, J. Salidis, G. K. Deutsch, P. Tallal, M. M. Merzenich, et al., 2001. Disrupted neural responses to phonological and orthographic processing in dyslexic children: An fMRI study. Neuroreport 12(2):299–307. Torgesen, J. K., 1999. Assessment and instruction for phonemic awareness and word recognition skills. In H. W. Catts and A. G. Kamhi, eds., Language and Reading Disabilities, 428–453. Needham Heights, MA: Allyn & Bacon. Torgesen, J. K., A. W. Alexander, R. K. Wagner, C. A. Rashotte, K. K. S. Voeller, and T. Conay, 2001. Intensive remedial instruction for children with severe reading disabilities: Immediate and long-term outcomes from two instructional approaches. J. Learn. Disabil. 1:33–58.

Torgesen, J. K., R. K. Wagner, and C. A. Rashotte, 1994. Longitudinal studies of phonological processing and reading. J. Learn. Disabil. 27(5):276–286; discussion 287–291. Torgesen, J. K., R. K. Wagner, K. Simmons, and P. Laughon, 1990. Identifying phonological coding problems in disabled readers: Naming, counting, or span measures? Learn. Disabil. Q. 13(Fall):236–243. Turkeltaub, P. E., G. F. Eden, K. M. Jones, and T. A. Zeffiro, 2002. Meta-analysis of the functional neuroanatomy of singleword reading: Method and validation. Neuroimage 16(3 Part 1): 765–780. Turkeltaub, P. E., L. Gareau, D. L. Flowers, T. A. Zeffiro, and G. F. Eden, 2003. Development of neural mechanisms for reading. Nature Neurosci. 6(7):767–773. Ungerleider, L. G., and. M. Mishkin, 1982. Two cortical visual systems. In D. J. Ingle, M. A. Goodale, and R. J. W. Mansfield, eds., Analysis of Visual Behavior, 549–586. Cambridge, MA: MIT Press. Vaidya, C. J., S. A. Bunge, N. M. Dudukovic, C. A. Zalecki, G. R. Elliott, and J. D. Gabrieli, 2005. Altered neural substrates of cognitive control in childhood ADHD: Evidence from functional magnetic resonance imaging. Am. J. Psychiatry 162(9): 1605–1613. Vellutino, F. R., J. M. Fletcher, M. J. Snowling, and D. M. Scanlon, 2004. Specific reading disability (dyslexia): What have we learned in the past four decades? J. Child Psychol. Psychiatry 45(1):2–40. Vellutino, F. R., D. M. Scanlon, S. Small, and D. P. Fanuele, 2006. Response to intervention as a vehicle for distinguishing between children with and without reading disabilities: Evidence for the role of kindergarten and first-grade interventions. J. Learn. Disabil. 39(2):157–169. Wadsworth, S. J., R. K. Olson, B. F. Pennington, and J. C. DeFries, 2000. Differential genetic etiology of reading disability as a function of IQ. J. Learn. Disabil. 33(2):192–199. Wagner, R. K., 1986. Phonological processing abilities and reading: Implications for disabled readers. J. Learn. Disabil. 19(10):623–631. Wagner, R. K., J. K. Torgesen, P. Laughon, K. Simmons, and C. A Rashotte, 1993. Development of young readers’ phonological processing abilities. J. Educ. Psychol. 85(1):83– 103. Wagner, R. K., and J. K. Torgesen, 1987. The nature of phonological processing and its causal role in the acquisition of reading skills. Psychol. Bull. 101(2):192–212. Wagner, R. K., J. K. Torgesen, C. A. Rashotte, S. A. Hecht, T. A. Barker, S. R. Burgess, et al., 1997. Changing relations between phonological processing abilities and word-level reading as children develop from beginning to skilled readers: A 5-year longitudinal study. Dev. Psychol. 33(3):468–479. Wang, Y., M. Paramasivam, A. Thomas, J. Bai, N. KaminenAhola, J. Kere, J. Voskuil, G. D. Rosen, A. M. Galaburad, and J. J. Loturco, 2006. DYX1C1 functions in neuronal migration in developing neocortex. Neuroscience 143(2):515– 522. Wise, B. W., J. Ring, and R. K. Olson, 1999. Training phonological awareness with and without explicit attention to articulation. J. Exper. Child Psychol. 72(4):271–304. Wolf, M., 1999. What time may tell: Towards a new conceptualization of developmental dyslexia. Ann. Dyslexia 49:219–247. Wolf, M., and P. G. Bowers, 1999. The double-deficit hypothesis for the developmental dyslexias. J. Educ. Psychol. 91:415– 438.

eden and flowers: developmental dyslexia

753

Wolf, M., L. E. Miller, and K. Donnelly, 2000. Retrieval, automaticity, vocabulary elaboration, orthography (RAVE-O): A comprehensive, fluency-based reading intervention program. J. Learn. Disabil. 33(4):375–386. Wolf, M., and M. Obregon, 1992. Early naming deficits, developmental dyslexia, and a specific deficit hypothesis. Brain Lang. 42:219–247.

754

Wood, F. B., and E. L. Grigorenko, 2001. Emerging issues in the genetics of dyslexia: A methodological preview. J. Learn. Disabil. 34(6):503–511. Wood, F. B., D. F. Hill, M. Meyer, and D. L. Flowers, 2005. Predictive assessment of reading. Ann. Dyslexia 55(2):193–216. World Federation of Neurology, 1968. Report of research group on dyslexia and world illiteracy. Dallas: W.F.N.

neurodevelopmental aspects of clinical disorders

46

The Development and Cognitive Neuroscience of Anxiety DANIEL S. PINE AND CHRISTOPHER S. MONK

This chapter provides an integrative perspective on pediatric anxiety disorders. The chapter is divided into four sections. The first defines key terms, differentiates normal from pathological anxiety states, and describes three specific anxiety disorders. The second section delineates three major advances that profoundly influence plans for future work on anxiety. Third, an integrative framework is described that draws insights from basic and clinical science. Finally, this framework is applied specifically to research on pediatric anxiety disorders.

Definitions The terms “fear” and “anxiety” define diverse phenomena across distinct research perspectives, including cognitive and affective neuroscience, biological psychiatry, and experimental and clinical psychology. This chapter integrates these approaches; accordingly, a uniform set of definitions is proffered. Specifically, the term “fear” defines a neural, cognitive, and behavioral state elicited in an organism by a dangerous stimulus or circumstance (Pine, 2007). Cues associated with danger are referred to as “threats” (Grillon, 2002). In these instances, threats, or dangerous stimuli and circumstances themselves, represent stimuli that an organism classifies as capable of producing harm and worthy of being avoided. Threats engage diverse processes across many neurocognitive levels (Pine, in press). For example, very briefly presented threats can engage an isolated set of neural, cognitive, and behavioral responses, whereas threats presented for longer time periods can engage a more elaborated set of responses (Davis, 1998). The term “anxiety” refers to a neural, cognitive, and behavioral state that resembles fear in terms of the organism’s response but differs from fear in terms of the associated stimuli. Whereas fear occurs in the immediate presence of threat, anxiety occurs in the absence of threat. Specifically, anxiety can be elicited by cues that suggest a threat may be forthcoming, or anxiety can represent a state that persists in the organism following elicitation of a fear response. Normal versus Pathological Anxiety Fear and anxiety represent normal aspects of childhood. Children and adolescents exhibit a characteristic pattern of fears and

anxieties that show striking similarities across cultures (Ollendick et al., 1996). Thus children consistently exhibit fears of strange situations and people within their first years of life. As these fears abate prior to school entry, they typically are replaced by concerns about separation and fears of animals or other danger, such as darkness or natural threats. Finally, as children approach adolescence, social concerns emerge as a dominant focus of anxiety. In all these areas, anxiety and fear represent aspects of normal development. Indeed, parents might be equally concerned by the absence of prototypical fears as by the presence of extreme fears. The ubiquity of anxiety during normal development raises major questions about delineating normal and pathological anxiety states. Fear and anxiety represent aspects of various psychopathological states, including personality disorders, psychoses, and mood disorders. The term “anxiety disorder” refers to the class of mental illnesses where anxiety represents the predominant focus of symptoms. Fear or anxiety is considered pathological when it causes marked distress or interferes with the child’s ability to function, typically leading to avoidance of situations perceived as dangerous. The fourth, revised edition of the Diagnostic and Statistical Manual (DSM-IV) recognizes more than 10 anxiety disorders as unique conditions, characterized by disorder-specific sets of criteria (American Psychiatric Association, 1994). Some of these disorders, such as panic disorder, occur primarily in adults: panic disorder virtually never manifests before puberty and rarely presents in adolescence (Pine et al., 1998). For the other anxiety disorders, which present more commonly in children and adolescents, considerable debate persists concerning the validity of the current narrow definitions in DSM-IV. As a result, the current chapter reviews prior research on pediatric anxiety disorders both broadly considered as a group of syndromes and considered individually, from the more narrow DSM-IV perspective. When data on individual disorders are discussed, the emphasis is on three specific conditions that are prevalent in childhood and diagnostically related. These comprise social anxiety disorder and generalized anxiety disorder, both of which are classified similarly in children, adolescents, and adults; and separation anxiety disorder, which is only classified in children and adolescents. These three conditions are

755

discussed as a group in the clinical and research literature because of high comorbidity and physiological similarities. A fourth condition, specific phobia, is often also discussed with these other three conditions. The current chapter provides limited discussion of specific phobia, since it shows weak associations with impairment and a distinct course (Shaffer et al., 1996; Pine et al., 1998). Finally, the chapter also briefly discusses data on pediatric posttraumatic stress disorder (PTSD). While most research on pediatric anxiety disorders excludes children with PTSD, interest in this condition has grown in light of data in rodents demonstrating robust effects of early life stress on threat response behavior throughout adulthood. Specific Pediatric Anxiety Disorders Social anxiety disorder involves extreme fear in situations where an individual is scrutinized. The condition typically emerges in late childhood or early adolescence, though subclinical signs typically manifest earlier. Similarly, behavioral inhibition, a temperamental classification in toddlers, also shows an association with social anxiety disorder (Perez-Edgar and Fox, 2005). Generalized anxiety disorder (GAD) involves worries about various situations and circumstances. Classically, these worries reflect concerns about competence or potential harm. Worries about social situations also typically occur in GAD. Technically, GAD is only diagnosed when associated worries extend beyond social situations. However, there are high rates of comorbidity between GAD and social phobia. Furthermore, GAD is highly comorbid with other conditions, such as depression (Angold, Costello, and Erkanli, 1999). Such high rates of comorbidity for GAD with other conditions raise questions about the degree to which GAD represents a valid, distinct pathological condition. Separation anxiety disorder involves concerns about harm to an attachment figure. Such worries lead affected children to show anxiety in anticipation of separation and to avoid situations, such as sleepovers, where separation occurs. Separation anxiety disorder shows a robust inverse relationship with age, representing the most common anxiety disorder before puberty and one of the rarer ones in late adolescence. Some evidence from both family-based and psychophysiological research suggests that separation anxiety disorder during childhood may represent a specific precursor of panic disorder, which emerges during adulthood (Pine et al., 1998; Biederman et al., 2001). Finally, PTSD involves anxiety associated with exposure to a traumatic event that produces marked and immediate fear in a child. Following exposure to trauma, children with PTSD demonstrate a pattern of reexperiencing this reaction as either flashbacks or recurring nightmares, in tandem with avoidance of situations related to the trauma and signs of increased arousal, such as exaggerated startle.

756

Major research advances and clinical implications Current classifications of psychopathological states rely entirely on descriptions of symptom patterns. Unlike other medical conditions, information on physiological indicators of perturbed organ system function is not currently useful in making distinctions between normal and pathological behaviors or emotions. Unfortunately, mental health professionals must wait some time before they will have tools analogous to the electrocardiogram. The practice of basing diagnoses on symptom reports represents a necessary compromise, given the limited understanding of physiologic influences on human behavior. While clearly limited, the current classification system revolutionized clinical approaches to mental disorders in the final decades of the 20th century. Such advances capitalized on the availability of reliable terminology that facilitated communication. Nevertheless, as scientific approaches mature, pathophysiology data will play an increasingly prominent role in classification. Arriving at a framework that allows integration of physiology data with symptom-based approaches represents a major challenge for mental health science (Pine et al., 2002). Integrating developmental approaches creates further complications. Three Major Advances Over the past quarter century, advances in three distinct research areas have made this a particularly opportune time for integrating clinical and basic research to better understand the development of anxiety. These three areas are research in developmental psychopathology, the neural correlates of fear, and cross-species parallels in the threat response. For developmental psychopathology, refinements in psychiatric nosology in the early 1980s produced a classification scheme for mental disorders that provided reliable definitions of various mental syndromes. This classification scheme was originally codified in the third version of the Diagnostic and Statistical Manual (DSM-III). The criteria for individual disorders in this manual have undergone two revisions, and various other classification schemes have been published (American Psychiatric Association, 1994). However, similarities among these schemes are far greater than the differences between any of these schemes and classification schemes used in the earlier parts of the 20th century. Two developments became possible following this change in classification. First, a series of large-scale epidemiological investigations could be initiated that defined the range of normal and abnormal behaviors in representative populations (Costello et al., 2002). This work revealed the tremendous morbidity associated with psychiatric states, including anxiety disorders. Second, studies used these innovations to track large cohorts of children. As these children aged throughout the end of the 20th century, the developmental unfolding of psychiatric states was revealed (Rutter,

neurodevelopmental aspects of clinical disorders

Kim-Cohen, and Maughan, 2006). Of note, these longitudinal studies revealed anxiety disorders to represent more clinically significant conditions than previously recognized, given the strong associations between pediatric anxiety disorders and various forms of adult psychopathology. Moreover, these data also served to stimulate debate concerning the validity of narrow DSM definitions of the anxiety disorders, given that an individual, narrowly defined anxiety disorder present in childhood typically showed equally strong associations with the same and with other anxiety disorders in late adolescence or adulthood. At the same time, the theoretical perspective of developmental psychopathology was becoming more broadly accepted. This school of thought focused on both normal and abnormal fluctuations in behavior across development. By understanding the range of behaviors that typically manifest in normal children, clarification of what is abnormal could be grasped. Moreover, by focusing on trajectories, the school placed behaviors in a developmental context. From this perspective, one symptom or another might raise concerns both for its immediate effect and its association with perturbed development (Gross and Hen, 2004). The importance of the developmental psychopathology perspective appears particularly prescient for anxiety disorders. Pediatric anxiety disorders consistently emerge as the most prevalent form of psychopathology in children and adolescents, with the rates of the conditions covered in this chapter appearing particularly high (Costello, Egger, and Angold, 2004). Moreover, pediatric anxiety disorders represent an early-life manifestation of various pathological states that unfold across the life span (Pine et al., 1998). This association appears particularly strong for anxiety disorders, such that pediatric social anxiety disorder, GAD, and separation anxiety disorder predict high risk for a range of adult anxiety disorders (Pine et al., 2007). Moreover, these three pediatric anxiety disorders also predict risk for other adverse adult outcomes, including mood disorders, substance use, and suicide (Pine et al., 2007). In terms of the second major advance, work in rodents and nonhuman primates demonstrates robust developmental influences on the threat response, involving genetic effects, environmental effects, and interactions between the two. For this chapter, the focus will be on measuring the threat response in immature organisms. This focus reflects the fact that variations on how immature organisms respond to threat may provide insight into why humans with anxiety disorders respond differently to situations and events that are perceived as threatening. The animal work suggests that immature rodent and primate brains show unique levels of plasticity in the way they process threat. The effects of rodent maternal behavior provide the most compelling evidence of the role of the environment on an organism’s response to threats. Stimulated by Levine’s

studies on maternal behavior (Levine, 1957, 1967), Meaney and colleagues defined the manner in which variations in maternal care during the initial weeks of life produce robust alterations in stress responses (Meaney, 2001; Gross and Hen, 2004). Rat pups exposed to intensive stimulation, associated with licking and grooming from their mothers, exhibit relatively low levels of threat response behavior throughout life. Cross-fostering work demonstrates that this effect is mediated by environmental influences. Moreover, a developmental window exists: rats exposed to environmental influences during adulthood do not show these permanent alterations in stress responses. While less extensive work examines environmental influences in nonhuman primates, analogous long-term influences have been demonstrated (Amaral, 2002). Thus this work shows that environmental effects have the capacity to interact with genes to shape an organism’s lifelong pattern of threat responses. Moreover, the environmental effects may have greater impact when the organism is young than when mature. Research in rodents has been vital in delineating developmental plasticity in genetic effects. Here, the most compelling examples emerge in gene-manipulation studies. Mice with a deletion of the serotonin 1A receptor (5-HT1a) exhibit extremely high levels of anxiety-like behaviors. This genetic effect emerges through developmental influences on anxietylike behaviors, as the effect only occurs if a functional gene is absent prior to day 21 of development: mice that are allowed to mature past this point with a fully functional 5HT1a receptor show no changes in anxiety-like behaviors when the gene is deleted later in life (Gross and Hen, 2004). Moreover, work with a related gene, which codes for the 5HT transporter (5HTT) protein, suggests that these genetic effects interact with environmental influences to shape anxiety-related phenotypes (Ansorge et al., 2004). Such work is consistent with studies in both nonhuman primates and humans (Suomi, 2003). Thus both environmental and genetic studies demonstrate unique plasticity in the immature rodent and nonhuman primate. Influences on fear behavior appear more robust and lasting in immature than in mature organisms. Finally, the third major advance involves research in cognitive neuroscience on cross-species parallels in the neural circuitry of fear responses. As reviewed in this chapter, work in this area has delineated a suite of information-processing functions that become engaged when an organism processes threats. These functions can be reliably engaged in rodents, nonhuman primates, and humans, as indexed by comparable measures of behavior (LeDoux, 2000). Moreover, with the advent of modern brain-imaging techniques, parallel measures of brain activity can be derived across these species. Furthermore, the inclusion of participants with brain lesions in conjunction with neuroimaging facilitates further understanding of cross-species similarities and differences (Amaral,

pine and monk: anxiety

757

2002; Adolphs et al., 2005). As a whole, this work demonstrates strong cross-species parallels not only in informationprocessing functions engaged by threats but also in the neural architecture that mediates the engagement of these processes. While such work has focused predominantly on mature organisms, recent studies in nonhuman primates suggest the neural correlates of these information-processing functions develop gradually as organisms mature (Prather et al., 2001; Amaral, 2002). Moreover, methods used to demonstrate neural correlates of these functions in adult humans have recently been successfully applied in research with children (Monk et al., 2006; McClure et al., 2007). These advances set the stage for truly novel research approaches to pediatric anxiety disorders. Clinical Questions on Pediatric Anxiety Realization of the potential afforded by these three advances will require studies to integrate emerging insights from neuroscience to address major questions emerging in clinical domains. Three specific questions have emerged in available research on pediatric anxiety. For advances in cognitive and affective neuroscience to have traction, methods might be brought to bear directly on these questions, thus integrating research on pathophysiology and clinical symptomatology. One major question concerns the validity of current nosology. On the one hand, preliminary data document specific longitudinal trajectories and cross-generational associations for some anxiety disorders (Pine et al., 1998; Merikangas et al., 1999; Biederman et al., 2001). For example, pediatric social phobia appears to show the strongest association with adult social phobia in longitudinal and familybased studies, whereas pediatric separation anxiety disorders appear to show the strongest association with adult panic disorder in such work. Generalized anxiety disorder (GAD), in contrast, shows associations with virtually all of the anxiety disorders, as well as with major depression. This finding has led to greater questions on the validity of GAD than either social phobia or separation anxiety disorder. On the other hand, none of the findings documenting specificity for social phobia and separation anxiety disorder are well replicated, and some studies generate contradictory findings (Costello, Egger, and Angold, 2004). Moreover, because of the high rates of comorbidity among all anxiety disorders, most studies of therapeutics target the three disorders together, rather than as distinct conditions. Thus controlled trials of either anxiolytic medication or psychotherapy typically randomize children, as a group, who present with any combination of social phobia, separation anxiety disorder, and GAD, to experimental and control treatments. The data on treatment response emerging from these studies also document stronger similarities than differences among the three disorders. Thus the weight of the evidence provides limited support for the validity of current classifications.

758

Data on therapeutics give rise to a second major question. Specifically, pediatric anxiety disorders exhibit nonspecificity in treatment outcomes; while both selective serotonin reuptake inhibitors (SSRIs) and cognitive behavioral psychotherapy (CBT) represent effective treatments, the treatments work equally well for all pediatric anxiety disorders (Kendall et al., 1997; RUPP, 2001). Moreover, heterogeneity clearly exists in terms of treatment response, and at present it is unclear what variables predict a favorable response (Birmaher et al., 2003; RUPP, 2003). Finally, the third major question concerns specificity in familial aggregation. Family studies document strong and consistent relationships between anxiety in parents and their offspring, with perhaps the most consistent association emerging between panic disorder in parents and separation anxiety disorder in offspring (Capps et al., 1996; Biederman et al., 2001). Nevertheless, family studies of depression document similarly strong associations: offspring of parents with major depression also show high rates of anxiety disorders in general and pediatric separation anxiety disorder specifically (Beidel and Turner, 1997; Weissman et al., 1997; Biederman et al., 2004). Moreover, genetic studies provide further evidence of nonspecificity in risk: identical genetic factors contribute to anxiety during childhood and major depression during adolescence or adulthood (Silberg, Rutter, and Eaves, 2001). These data raise basic questions concerning the degree to which children manifest narrow as opposed to broad vulnerabilities. Answers to questions on nosology, therapeutics, and risk are unlikely to emerge from current research approaches. The inconsistent pattern of associations in prior research suggests that current classification schemes do not accurately capture underlying variations in pathophysiology. A novel approach might integrate insights from neuroscience.

Explicating a model of clinical anxiety An Integrative Framework Progress in mental health science requires an integration of data on pathophysiology with data on nosology and developmental psychopathology. Ultimately such integration will lead to a nosology based on both symptomatic expression and pathophysiology. Figure 46.1 illustrates a framework for achieving such integration. As shown on the left of the figure, pathophysiological processes in virtually all forms of psychopathology ultimately involve interactions between genetic and environmental influences, with distinct disorders involving different mixtures of these effects and distinct genes or environmental influences. These effects do not directly influence expressions of or risks for psychopathology. Rather, genetic and environmental influences are thought to sculpt functional aspects of neural circuits, with specific circuits playing distinct roles in specific disorders.

neurodevelopmental aspects of clinical disorders

Genetic influences

Information Processing Functions Fear Circuit Medial Temporal Lobe (Amygdala & Hippocampus)

Prefrontal Cortex (PFC)

Clinical Features Anxiety Disorders

Fear Conditioning, Extinction & Other Aspects of Emotional Memory

Temperament

Emotion-Attention Interaction

Exposure to Risks

Response to Innate Threats Sub-Clinical Fears Other Functions (e.g., stimulus-reward association, stimulus classification, etc.)

Environmental influences Figure 46.1 This figure depicts a model that integrates neuroscience research on pathophysiology with clinical research on between-subject variation in anxiety-related constructs. The patho-

physiologic chain of events flows from left to right in explicating causes of between-subject variations in clinical anxiety states.

Current understanding of brain-behavior associations appears much stronger for between-subject variations in cognitive or affective processing than for between-subject variations in psychiatric symptoms. In research on brainbehavior associations, children can be classified based on profiles derived from measures of brain imaging or from information-processing paradigms. These neuroscienceoriented classifications then can be linked to clinically based classification schemes. The term “endophenotype” has been used to refer to neuroscience-based classifications that show consistent relationships with clinical classifications. Thus a crucial step in refining current nosology involves linking measures of brain function and information-processing function to clinical measures of psychiatric symptoms and risk. For research on anxiety disorders, this will require investigators to assess information-processing functions engaged when subjects respond to threats. These information-processing functions can then be mapped both onto clinical symptom patterns and onto underlying neural circuitry engaged by threats.

threat responses, as well as responding to innately threatening stimuli or situations. Perhaps the most extensive literature examines fear conditioning, the process whereby a neutral conditioned stimulus, such as a light or a tone (CS+), is repeatedly paired with an aversive unconditioned stimulus (UCS), such as an electric shock or a loud sound (LeDoux, 2000). As shown in figure 46.2, following repeated pairings, mammals rapidly learn that a CS+ predicts the UCS, leading the organism to treat the CS+ as a threat. Mammals also learn to represent the context in which the UCS–CS+ relationship is established. Such contextual representations code for the place in which the pairing occurred as well as the state of the organism during conditioning. The neural circuit engaged during fear conditioning is centered on the amygdala and distributed throughout various regions, including sensory-cortical pathways and the thalamus (LeDoux, 2000; Phelps and LeDoux, 2005). Though disagreements persist concerning the precise role of specific amygdala subnuclei in fear conditioning, consensus has emerged regarding the role of the amygdala in a broader series of processes. These processes include aspects of reward evaluation, each characterized by rapid evaluations of stimulus salience (Baxter and Murray, 2002). Representation of context involves a series of associated regions, which include

Neural Circuits Engaged by Threats Research in rodents and nonhuman primates precisely elucidates neural circuits engaged by distinct aspects of threat processing. These aspects include formation and extinction of learned

pine and monk: anxiety

759

TRAINING

Exposure to context (2 min)

Onset of sound (CS: 30 s)

Onset of shock (US: 2 s)

TESTING: Context Test at 1 hour and 24 hours

Same context (5 min)

TESTING: Cued Test at 1 hour and 24 hours

Onset of sound (CS: 3 min)

Figure 46.2 This figure depicts methods for a fear-conditioning experiment in rodents. Following exposure of the rodent to the context (i.e., the cage), a series of shock-unconditioned stimulus presentations (UCS) is paired with repeated presentation of a conditioned stimulus (the “sound CS”). At a later time, the rodent shows fear both of the context, when placed in the cage where conditioning occurred, and fear of the sound CS, when placed in

a novel cage and presented with the same sound CS. Following repeated presentation of this sound CS in the novel context, without the shock UCS, the rodent exhibits extinction of fear to the sound CS. (Figure reproduced with permission from Larry R. Squire and Eric R. Kandel, Memory: From Mind to Molecules, 1999, Scientific American Library, Henry Holt and Company, LLC.)

the hippocampus and prefrontal cortex (PFC) (LeDoux, 2000; Miller and Cohen, 2001). Advances in research on fear conditioning during the past decade have stimulated increasing interest in extinction, the processes whereby mammals learn that previously learned CS+ are no longer dangerous. Extinction requires new learning about the CS+–UCS relationship, as opposed to forgetting that a connection exists between CS+ and UCS. This new learning requires reclassification of a CS+ as currently not dangerous, which can be achieved by representing the context of CS+ presentation, instantiated in the PFC (Bouton, 2002; Quirk and Gehlert, 2003). Virtually no research examines development changes in either conditioning or extinction in the periadolescent period, a particularly important time for understanding of clinically significant anxiety. As a result, work in this area represents a particularly important focus for future research efforts. Fear conditioning and extinction are used to examine processes whereby affective value is assigned to initially neutral stimuli. Other stimuli are recognized as dangerous independent of any previous exposure, presumably because of the evolutionary advantage afforded to ancestors who

demonstrated the capacity to immediately recognize such threats in the absence of prior exposure. Such stimuli relate to species-specific dangers. For example, well-lit rooms represent innate threats for nocturnal rodents, whereas dark rooms represent innate threats for diurnal organisms, such as humans (Davis, 1998). The circuitry through which these innate threats are processed shows both parallels and discontinuities with the circuitry of learned fears. For example, whereas both types involve the amygdala, the basolateral nucleus may play a particularly strong role in processing innate threats (Davis, 1998). Moreover, components of the extended amygdala, such as the bed nucleus of the stria terminalis, are more strongly implicated in processing of innate fears. Finally, pharmacological studies demonstrate distinctions among these classes of fears. Thus rodent and nonhuman primate studies elucidate neural structures engaged by various classes of threats. Most of this research examines aspects of neural function that are presumed to operate in all individuals of a given species. Relatively few studies reveal factors that are associated with individual differences. Moreover, to the extent that neural contributions have been identified, available studies from

760

neurodevelopmental aspects of clinical disorders

neuroscience most strongly implicate developmental factors in individual differences (Amaral, 2002; Gross and Hen, 2004). For example, work on the effects of amygdala lesions in monkeys shows that the developmental timing of an injury influences the nature of effects on individual differences in responding to specific threats. When injury occurs in adulthood, amygdala lesions reduce responding to a range of threats. When injury occurs during childhood, in contrast, such lesions reduce responding to snakes but augment responding to social threats. Similarly, as noted previously, studies in rodents and nonhuman primates show that alterations in the rearing environment produce robust and enduring alterations in fear responses. For example, in rodents, handling manipulations produce long-term alterations in threat-response behavior and associated neural circuit functioning that persist through adulthood. Similar manipulations of the social environment in mature rodents produce no such lasting effects. These effects emerge through alterations specifically on the fear circuit. Information-Processing Functions of the Fear Circuit As depicted in figure 46.1, integration of data from neuroscience and clinical science may emerge through research on information processing. Such work benefits from a cross-species perspective that examines how functions manifest in rodents, nonhuman primates, and humans at various developmental stages. Given the precise anatomical understanding of fear conditioning, it should come as no surprise that this process has been examined in many studies. Brain-imaging studies reliably implicate neural circuits in various aspects of fear conditioning among humans (Phelps and LeDoux, 2005). Thus the amygdala is consistently engaged during the acquisition of fears to specific cues, whereas the PFC is engaged during extinction (Kalisch et al., 2006). This research extends to humans as depicted in figure 46.1. Nevertheless, the relationship that between-subject variation in fear conditioning shows with either risk for, or overt expression of, clinical anxiety remains unclear. Thus the tie between informationprocessing functions and clinical states in figure 46.1 appears more tenuous. The only meta-analysis addressing this issue, which included all studies in children as well as adults, found a weak relationship (Lissek et al., 2005). Moreover, the relationship was entirely accounted for by studies using simple fear conditioning, which involves only a CS+ but no CS− stimulus. Without this experimental control, simple fear conditioning experiments cannot differentiate associations with conditioning from sensitization in anxiety disorders. Recent interest focuses on processes beyond the acquisition of conditioned fears. Thus Grillon suggests that clinical anxiety states result more from perturbations in discrimination among various cues associated with a UCS, as opposed to a tendency to show strong conditioning to one or another

CS+ (Grillon, 2002). Alternatively, Bouton suggests that clinical anxiety states result from deficits in extinction as opposed to formation of conditioned fears (Bouton, 2002). Data supporting this position emerge from studies demonstrating the efficacy of cognitive behavioral therapy, which uses extinction to reduce anxiety (Ressler et al., 2004). This position received further support from recent studies demonstrating the clinical benefits of therapy involving extinction learning in combination with medication that enhances extinction (Ressler et al., 2004; Hofmann et al., 2006). Fear conditioning is conceptualized as a form of emotional learning or memory formation; distinct memoryrelated processes associated with fear conditioning require distinct functions that can be classified as either declarative or nondeclarative in nature. Recent interest in clinical states has focused on declarative forms of emotional learning. For example, emotionally arousing stimuli are thought to receive prioritization during cognitive processing, and this prioritization is thought to produce a memory advantage on certain declarative mnemonic tasks (Phelps, 2006). Such a memory advantage is conceptualized as a memory bias to selectively encode certain aspects of emotionally arousing stimuli. Subjects show a consistent memory advantage on declarative tasks for arousing words, pictures, or stories, relative to neutral items. Moreover, neuroimaging data implicate the amygdala in this memory advantage (Cahill, 1999). Nevertheless, clinical research reveals weak and inconsistent relationships between emotional memory and anxiety disorders (Pine et al., 2004). From the clinical perspective, perturbations in attention probably represent the most consistently implicated information-processing function in anxiety disorders. Thus tasks that probe interactions between emotion and attention reveal a strong tie in figure 46.1 between this specific information-processing function and clinical anxiety states (Williams, Mathews, and MacLeod, 1996; Bar-Haim et al., 2007). These associations have been found in both children and adults, both in children with specific anxiety disorders, such as GAD or social phobia, and in children with high ratings on anxiety symptom scales in the absence of a clinical anxiety disorder diagnosis. On these tasks, cues signaling threats have been shown to exert greater disruptions in attention among patients with anxiety disorders than in healthy individuals. Moreover, emerging evidence suggests that these attention biases play a role in the genesis of clinical anxiety disorders, such that between-subject variation in attention predicts response to stress (MacLeod et al., 2002). Such attention disruptions manifest as biased processing of neutral information that appears particularly proximal or distal to threatening information. This capacity to bias attention has been demonstrated using diverse tasks, each designed to tap distinct aspects of attention. For example, spatial orienting has been probed

pine and monk: anxiety

761

using either dot-probe tasks or tasks modeled after the spatial-orienting paradigm of Posner (Williams, Mathews, and MacLeod, 1996; Bar-Haim et al., 2007). These tasks explicitly ask subjects to attend to neutral target cues by performing a motor or oculomotor response to the target. Potentially distracting stimuli of both neutral and emotionally evocative valences are introduced using various methods, and the effects of these distractors are manifested in variations of motor or oculomotor responses. A differential bias in patients is reflected in the effects of emotion on responses to neutral targets. Alternatively, strategic attention has been monitored using variants of the Stroop task, whereby emotional distractors are introduced based on the meaning of color words. In this task, subjects are required to ignore the word meaning as they identify word colors. With this design, threat words produce greater perturbations in color naming among patients with anxiety disorders than among healthy subjects (Williams, Mathews, and MacLeod, 1996). A final set of tasks asks subjects to vary their attention focus as they attend to distinct aspects of stimuli varying in valence. For example, subjects have been asked to alternatively attend to either physical aspects of photographs or their own subjective response to these photographs (Lane et al., 1997; Pine et al., 2005a). Both the emotional content in photographs and the subject’s focus of attention influence task performance in these designs, providing yet another example of an emotion-attention interaction. Both neuroimaging and lesion studies among adult humans implicate the amygdala as a crucial modulator of the emotion-attention interface, as probed by these tasks. In other clinically focused research, anxiety disorders have been characterized based on their relationships with hypersensitivities to innate threats. Probably the most developed line of work examines associations with respiratory threats (Klein, 1996; Pine et al., 2005a, 2005b). Stimuli associated with suffocation represent innate threats for airbreathing organisms such as humans. Adults with panic disorder, which typically afflicts adults but not children, show hypersensitivity to suffocation cues. Moreover, consistent with data on cross-generational transmission, children with separation anxiety disorder also show signs of such hypersensitivity, suggesting that pediatric separation anxiety disorder represents a precursor of adult panic disorder. Data examining the response to other innate threats generate less consistent associations. Risk for anxiety in adolescents has been linked to extreme responses to aversive air puffs (Merikangas et al., 1999). However, adolescents at risk for depression also show this hypersensitivity, but adolescents with ongoing anxiety disorders do not (Grillon et al., 2005). Finally, other studies in patients implicate the amygdala and associated fear circuitry in an array of other information-processing functions. For example, studies in rodents, nonhuman primates, and humans clearly implicate the

762

amygdala in formation of stimulus-reward associations (Baxter and Murray, 2002). During tasks that tap this information-processing function, the amygdala is thought to interact with the PFC to represent in a flexible manner the current valence of a stimulus. Similarly, the amygdala has been implicated in the classification of stimuli that vary in emotional details. For example, both lesion and neuroimaging studies in adults implicate the amygdala in the decoding of face-motion cues (Adolphs et al., 2005). However, sparse evidence implicates these information-processing functions in clinical anxiety states. Specifically, the few available studies in children with social phobia, GAD, or separation anxiety disorder find normal performance on tests of faceemotion classification. Modulators of Fear-Circuit Function Studies in humans demonstrate the robust effects of various factors on the fear circuit. These effects derive from both environmental and genetic influences. Studies that manipulate serotonin (5HT)–related genes in rodents using knockout strategies reliably produce changes in threat response behavior. These effects have been found both for manipulations of the serotonin transporter gene (5HTT) and of the 5HT1a gene (Gross and Hen, 2004). Interestingly, these effects emerge during development, in that manipulations of the 5HTT and 5HT1a gene produce stronger effects when they occur in immature rodents than when they occur in mature ones. Integrating insights from this work with insights from Meaney’s work on early-life environmental manipulations leads to the suggestion that genes interact with the environment as organisms mature to sculpt activity in the fear circuit. Studies in humans have not fully extended this model. Nevertheless, variation in the 5HTT has been reliably shown to predict response in the fear circuit in at least five studies among adults (Hariri et al., 2002; Canli et al., 2005; Hariri et al., 2005; Heinz et al., 2005; Pezawas et al., 2005). This response is reflected in the threshold for engaging both the amygdala and associated expanses of the PFC (Pezawas et al., 2005). Individuals who inherit the low-activity form of the 5HTT exhibit a lower threshold for engaging this circuit. As in studies among rodents and nonhuman primates, this heightened sensitivity is thought to interact with the environment to ultimately shape in tandem neural circuit function, information-processing capacities, and risk for psychopathology (Caspi et al., 2003; Gross and Hen, 2004). Environmental factors also have been implicated in between-subject variations in fear-circuit function. Probably the most extensive line of research examines the degree to which stress exposure relates to clinical compromise and associated fear-circuit dysfunction. A sizable minority of individuals exposed to traumatic events will develop chronic anxiety, manifested as posttraumatic stress disorder (PTSD),

neurodevelopmental aspects of clinical disorders

which involves both subject-specific and environmental factors: humans exhibit robust between-subject variability in response to comparable aversive stressors, but the risk for PTSD clearly increases significantly among all people as a function of stressor severity (Pine, 2003). Research using both information-processing and neuroimaging approaches implicates fear-circuit dysfunction in PTSD, although most work focusing on these issues examines adults. Thus adults with PTSD, relative to traumatized adults without PTSD, show signs of perturbed attention and memory, occurring in tandem with perturbations in amygdala, hippocampal, and PFC function. Nevertheless, consistent variability emerges across studies in terms of the precise nature of the associated abnormalities. Finally, other evidence of fear-circuit dysfunction emerges in a surprisingly wide range of adult clinical conditions, including various anxiety disorders (Rauch, Shin, and Wright, 2003). Fear-circuit dysfunction also occurs in a range of other clinical states. Most prominent among there is major depressive disorder (MDD), where abnormalities in both amygdala and PFC emerge with consistency (Drevets, 2000; Whalen et al., 2002). In summary, adult variations in threat response behavior and fear-circuit function reflect ontological influences of genetic and environmental factors. Clinical research demonstrates that adult anxiety disorders reflect comparable ontogeny, such that the majority of adult disorders have their roots in childhood. Therefore, a full application of the framework depicted in figure 46.1 is predicated on an extension to children and adolescents.

Affective neuroscience and pediatric anxiety Integrative studies on pediatric anxiety must consider two major factors, each related to key constructs illustrated in figure 46.1. First, as noted at the right-hand side, betweensubject clinical features can be conceptualized from various perspectives. Second, various information-processing functions, as appearing in the middle of figure 46.1, can be probed and linked to functional aspects of the fear circuit, as appearing at the left-hand side of figure 46.1. Clinical Classification Four distinct categories of children and adolescents have been studied from integrative research perspectives. These categories involve group classification based on either particular features of the child or the child’s level of risk. First, studies in children have examined individuals affected by specific pediatric anxiety disorders. Typically, this work focuses on children with social phobia, GAD, or separation anxiety disorder to provide vital data concerning the pathophysiology of clinical anxiety states. One would expect to detect perturbations in information-processing

and fear-circuit functions implicated in these conditions. Nevertheless, such work is complicated by the heterogeneity of most clinical conditions. For example, children may manifest syndromes that appear similar, in terms of their overt symptomatic expressions, but emerge from distinct forms of pathophysiology. This pattern is ubiquitous in many chronic medical illnesses, such that various forms of cancers or endocrine-based syndromes might present with similar clinical manifestations, each emerging from diverse pathophysiologies. Second, children and adolescents can be classified meaningfully based on their temperament, which can be observed early in development. Perhaps the strongest data linking temperament and pediatric anxiety come from work on behavioral inhibition: the tendency for toddlers to react with wariness, avoidance, and suppression of motor response when confronted with novel stimuli or situations (PerezEdgar and Fox, 2005). Questions remain concerning the optimal stimuli to elicit inhibitory reactions, because inhibition has been linked to biases in the processing of threats and rewards, as well as neutral novel or discrepant stimuli. Nevertheless a clear association with clinical anxiety states has been demonstrated. Children formerly classified as inhibited exhibit an increased risk for various anxiety states, with some evidence suggesting a particularly strong tie with social anxiety disorder (Perez-Edgar and Fox, 2005). Moreover, behavioral inhibition in children is associated with a high rate of panic disorder in parents (Rosenbaum et al., 2000). However, the degree to which specific associations emerge with risk for anxiety remains unclear. Behavioral inhibition is not associated with behavior disorders, such as attention-deficit/ hyperactivity disorder, but strong associations with risk for MDD emerge in both longitudinal and family-based data (Caspi et al., 1996; Rosenbaum et al., 2000). This pattern of associations is consistent with the data reviewed earlier linking fear-circuit dysfunction to various adult psychopathological states. Third, independent of any temperamental or clinical factors associated directly with the child, risk for anxiety in children is manifest based on aspects of the environment in which they are raised. These risks can be mediated by internal factors related to the child, they may be purely environmental, or they may be combinations of both (Rutter, 2000; Silberg, Rutter, and Eaves, 2001; Costello et al., 2002). Some categorizations involve classifications likely to be influenced by constellations of these factors. For example, as described previously, cross-generational transmission of pathological states has been noted, such that pediatric anxiety is strongly linked to both anxiety and MDD in parents (Beidel and Turner, 1997). Similarly, children exposed to trauma show high risk for mood and anxiety disorders (Pine, 2003). These associations are likely to reflect complex interactions among

pine and monk: anxiety

763

genetic and environmental factors. Other classifications can involve more narrowly conceptualized risk factors. For example, the same 5HTT polymorphism associated with adult fear-circuit dysfunction has been implicated in pediatric anxiety (Fox, 2005). Nevertheless, such factors are likely to interact in complex ways with other factors to increase risk. In the instance of the 5HTT, this polymorphism interacts with experiential factors to increase risk for mood and anxiety disorders (Caspi et al., 2003). Complex interactions involving the 5HTT gene have been demonstrated as predictors of overt clinical states in adults and children as well as temperamental variations in adults and children. Finally, as noted earlier, major questions persist concerning boundaries between normal and pathological manifestations of anxiety in children. Particularly important questions emerge concerning specific fears, which can be classified as normal or as clinically significant, in the form of specific phobias. Children with subclinical fears clearly represent a population at risk for various adult psychopathologies, particularly MDD (Weissman et al., 1997; Pine, Cohen, and Brook, 2001). Such associations emerge in both longitudinal and family-based studies. Information-Processing and Fear-Circuit Functions in Pediatric Anxiety Relatively few studies in children and adolescents use associations between clinical anxiety and information-processing functions and relate them to findings of fear circuitry dysfunction. Virtually no research in children and adolescents specifically examines the relationship between fear conditioning and anxiety disorders, one of the best-studied areas in adults. Clearly, ethical factors complicate attempts to implement such work. For instance, the use of electric shocks, the bestvalidated UCS in adult fear-conditioning studies, raises ethical questions for a vulnerable population. In general, attempts to use less aversive UCS in children generate suggestive but far from definitive associations between various indices of fear conditioning and clinical symptoms or risk (Merikangas et al., 1999; Grillon et al., 2005). Moreover, brain-imaging studies using such UCS in children and adolescents reveal amygdala engagement but in a manner less consistent than in adult studies with more aversive UCS (Monk et al., 2003). As described previously, research on attention provides the best evidence of perturbed information processing in anxiety disorders, and this statement holds true in pediatric anxiety as well. A relatively small but growing constellation of studies demonstrates relationships between biased attention and pediatric anxiety disorders (Bar-Haim et al., 2007). The most comprehensive literature examines biases in attention orienting, and this work raises questions concerning developmental influences on attention-emotion interactions. For example, studies using orienting paradigms,

764

which reveal vigilance for threats in adult anxiety states, have revealed attention avoidance in pediatric anxiety disorders (Monk et al., 2006). Figure 46.3 (plate 62) depicts the task on which this finding emerged: GAD patients studied with this task tended to show avoidance of asterisk targets appearing proximal to angry-threat faces, whereas healthy adolescents tended to show attention toward such asterisk targets. Adults with and without GAD show the opposite patterns of responding. This finding may suggest that the threshold for disengaging attention to avoid threats may change during development. Another measure of attention is the emotional Stroop task. Using these procedures, both children and adults with anxiety disorders, relative to healthy age-mates, exhibit greater disruption in color naming of threat relative to neutral color words (Vasa and Pine, 2004). Finally, one study used an attention task that required subjects to shift attention while viewing evocative facial photographs (Pine, Mogg, et al., 2005). This study found that both ongoing pediatric anxiety disorders and a history of panic disorder in a parent predicted enhanced attention allocation when monitoring internal fear states. Neuroimaging studies extend these data by examining attention: of the three functional magnetic resonance imaging (fMRI) studies on pediatric anxiety disorders, two examine attention. The first study compared 12 pediatric patients with GAD or panic disorder to 12 healthy children and adolescents, acquiring data while subjects passively viewed blocks of face-emotion photos (Thomas et al., 2001). The study generated provocative findings by showing that patients exhibited greater amygdala activation than healthy peers when viewing fearful relative to neutral faces. However, the study failed to demonstrate amygdala engagement to fear faces in the healthy subjects. Since no behavioral data were acquired at imaging, the study left unanswered questions concerning attention influences that might account for the group difference in amygdala activation. Data from other fMRI studies implicate attention in the modulation of these group differences. Monk and colleagues found attention avoidance in 18 adolescents with GAD, relative to 15 healthy adolescents, using a dot-probe task that presented facial threat cues to manipulate orienting (Monk et al., 2006). This study demonstrated enhanced right ventrolateral PFC (VLPFC) engagement in GAD relative to healthy adolescents during threat processing. Based on a negative correlation between VLPFC activation and anxiety severity, PFC engagement in this study was seen as essentially regulatory in nature. Figure 46.3 presents data from this study. Specifically, the figure shows brain regions where GAD patients showed greater activation than healthy adolescents, for events where angry threat faces appear spatially incongruent with asterisk targets, contrasted with events where angry threat faces appear

neurodevelopmental aspects of clinical disorders

Direction of Attention Avoid Anger

Attend to Anger

vs. *

*

Figure 46.3 This figure shows a group contrast of an eventrelated analysis in 18 pediatric GAD patients and 15 age-matched healthy adolescents, drawing data from Monk and colleagues (2006). The specific contrast shown maps brain areas where GAD patients show enhanced activation for a contrast of events where

asterisk targets appear proximal to neutral faces, relative to when they appear proximal to angry faces. As shown in the upper left, a group difference emerges in the prefrontal cortex, at coordinates of x, y, z = 51, 30, −2, in Talairach space, with t = 3.3 (p = .001). (See plate 62.)

spatially congruent with asterisk targets. While the study found consistent between-group differences in VLPFC engagement, the study failed to detect differences in amygdala activation. These data suggest that anxiety-related differences emerge in a distributed circuit encompassing the PFC. The only other fMRI study in pediatric anxiety disorders supported this suggestion. In a study of 15 GAD patients and 20 age-gender-IQ-matched healthy adolescents, McClure and colleagues used a paradigm that required subjects to shift attention while viewing photographs of faces depicting evocative poses (McClure et al., 2007). This study found enhanced activation in a distributed network encompassing the fear circuit. This network included portions of the ventral and medial PFC, as well as the amygdala. Attention modulated group differences in circuitry engagement: patients only differed from healthy adolescents when monitoring their internal fear state but not when viewing faces while adopting different attention states. Taken together, these three studies implicate perturbations in fear-circuit function in pediatric anxiety disorders, particularly GAD. The degree to which these perturbations

reflect state as opposed to trait factors remains unclear. Studies in behavioral inhibition, a major risk factor for pediatric anxiety disorders, generate some insights on this question. The only published study in this area compared amygdala activation in 22 adults formerly classified as either inhibited or noninhibited as toddlers (Schwartz et al., 2003). Examining the response to novel relative to nonnovel faces, the formerly inhibited subjects showed greater amygdala activation, suggesting that hypersensitivity in the brain’s fear network may represent a risk factor for anxiety. These imaging data on the relationship between anxiety and fear-circuit function raise questions on associations among anxiety, brain function, and brain structure. Studies in adults suggest that individual variations in anxiety might reflect individual differences in anatomy of structures encompassing the fear circuit (Drevets, 2000; Whalen et al., 2002). Probably the most consistent findings document reduced volumes of medial PFC structures, which occur in a range of anxiety disorders, most prominently in PTSD, as well as in individuals inheriting genetic risk factors for anxiety disorders (Pine, 2003; Pezawas et al., 2005). Such reductions also occur in depression. Other studies in adults

pine and monk: anxiety

765

find reductions in medial temporal lobe structures, including both the hippocampus and amygdala. Reduced hippocampal volumes emerge most consistently in studies of traumatized individuals. Reduced amygdala volumes emerge in studies of various conditions, albeit with relative inconsistency, including studies of depression, anxiety, and risk for depressive or anxiety disorders. Interestingly, at least for data in depression, a model of pathophysiology implicating glutamate excitotoxicity emerges from this confluence of findings (Siegle et al., 2003). Specifically, this model is consistent with data in the amygdala documenting reductions in brain volume, in tandem with increases in neural activation, as well as negative correlations between volume and activation. As with fMRI data, fewer studies examine morphometry in pediatric anxiety disorders. Two studies specifically examined amygdala volume. Findings from these two studies were inconsistent: one reported increased volume, whereas the other found decreased volume (De Bellis et al., 2000; Millham et al., 2005). However, data documenting increased activation and reduced volumes in the same series of adolescents support excitotoxicity-based models (De Bellis et al., 2000; Millham, 2005). Studies of other brain structures have generated other findings that also need replication. Beyond studies of fear-circuit function, research implicating other aspects of perturbed information processing in pediatric anxiety appears weak or inconsistent. Thus some evidence of perturbed reward-system function is found, but results have not been well replicated. Similar findings emerge in studies of memory. In fact, the only fMRI study of emotional memory found enhanced amygdala activation in adolescents with major depression but not in adolescents with anxiety disorders, relative to healthy adolescents (Roberson-Nay et al., 2006).

Conclusions Research in pediatric anxiety disorders might provide an important guide for investigators attempting to integrate research on clinical conditions with research on underlying brain function. From the clinical perspective, advances in developmental psychopathology have delineated the range of anxiety states exhibited by children and adolescents. While the current chapter provides the most detailed discussions of work on three specific anxiety disorders, major questions remain concerning the validity of nosology for the broad family of anxiety states manifest during development. Research in neuroscience documents the manner in which early-life individual differences in fear-circuit function relate to long-term information-processing functions instantiated in the fear circuit. Studies applying cognitive neuroscience methods hold the hope of using these findings

766

to ground clinical research in current understandings of pathophysiology. Of particular relevance for future advances, this chapter has reviewed recent data from fMRI studies that suggest the potential utility of this integrative, translational approach. While this research raises as many questions as it answers, the work does provide an important road map for possible future research endeavors. Namely, recent work on information processing and fMRI demonstrates the capacity to map in children brain systems and associated informationprocessing functions previously mapped in fine detail among rodents and nonhuman primates. As such, future work on pediatric anxiety disorders might map in both healthy and anxious children the nature of developmental relationships between information processing and brain function associated with threat processing. For example, such work might involve longitudinal analysis examining changes in PFC and amygdala activation on dot-probe threat paradigms in both healthy and anxious children, studied as they mature from adolescence to adulthood. Such work would be expected to show the manner in which changes in attention and fear circuit function relate to developmental changes in anxiety. REFERENCES Adolphs, R., F. Gosselin, T. W. Buchanan, D. Tranel, P. Schyns, and A. R. Damasio, 2005. A mechanism for impaired fear recognition after amygdala damage. Nature 433:68–72. Amaral, D. G., 2002. The primate amygdala and the neurobiology of social behavior: Implications for understanding social anxiety. Biol. Psychiatry 51:11–17. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, fourth edition. Washington, DC: American Psychiatric Association. Angold, A., E. J. Costello, and A. Erkanli, 1999. Comorbidity. J. Child Psychol. Psychiatry 40:57–87. Ansorge, M. S., M. Zhou, A. Lira, R. Hen, and J. A. Gingrich, 2004. Early-life blockade of the 5-HT transporter alters emotional behavior in adult mice. Science 306:879–881. Bar-Haim, Y., D. Lamy, L. Pergamin, M. J. BakermansKranenburg, van M. H. Izendoorn, 2007. Threat-related attentional bias in anxious and non-anxious individuals: A metaanalytic study. Psychol. Bull. 133(1):1–24. Baxter, M. G., and E. A. Murray, 2002. The amygdala and reward. Nature Rev. Neurosci. 3:563–573. Beidel, D. C., and S. M. Turner, 1997. At risk for anxiety. I. Psychopathology in the offspring of anxious parents. J. Am. Acad. Child Adolesc. Psychiatry 36:918–924. Biederman, J., S. V. Faraone, D. R. Hirshfeld-Becker, D. Friedman, J. A. Robin, and J. F. Rosenbaum, 2001. Patterns of psychopathology and dysfunction in high-risk children of parents with panic disorder and major depression. Am. J. Psychiatry 158:49–57. Biederman, J., M. C. Monuteaux, S. V. Faraone, D. R. Hirshfeld-Becker, A. Henin, J. Gilbert, and J. F. Rosenbaum, 2004. Does referral bias impact findings in highrisk offspring for anxiety disorders? A controlled study of

neurodevelopmental aspects of clinical disorders

high-risk children of non-referred parents with panic disorder/agoraphobia and major depression. J. Affect. Disord. 82:209– 216. Birmaher, B., D. A. Axelson, K. Monk, C. Kalas, D. B. Clark, M. Ehmann, J. Bridge, J. Heo, and D. A. Brent, 2003. Fluoxetine for the treatment of childhood anxiety disorders. J. Am. Acad. Child Adolesc. Psychiatry 42:415–423. Bouton, M. E., 2002. Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biol. Psychiatry 52:976– 986. Cahill, L., 1999. A neurobiological perspective on emotionally influenced, long-term memory. Semin. Clin. Neuropsychiatry 4:266– 273. Canli, T., E. Congdon, L. Gutknecht, R. T. Constable, and K. P. Lesch, 2005. Amygdala responsiveness is modulated by tryptophan hydroxylase-2 gene variation. J. Neural Transm. 112: 1479–1485. Capps, L., M. Sigman, R. Sena, B. Henker, and C. Whalen, 1996. Fear, anxiety and perceived control in children of agoraphobic parents. J. Child Psychol. Psychiatry 37:445–452. Caspi, A., T. E. Moffitt, D. L. Newman, and P. A. Silva, 1996. Behavioral observations at age 3 years predict adult psychiatric disorders: Longitudinal evidence from a birth cohort. Arch. Gen. Psychiatry 53:1033–1039. Caspi, A., K. Sugden, T. E. Moffitt, A. Taylor, I. W. Craig, H. Harrington, J. McClay, J. Mill, J. Martin, A. Braithwaite, and R. Poulton, 2003. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301:386–389. Costello, E., D. S. Pine, C. Hammen, J. S. March, P. M. Plotsky, M. M. Weissman, J. Biederman, H. H. Goldsmith, J. Kaufman, P. Lewinsohn, M. Hellander, K. Hoagwood, D. S. Koretz, C. A. Nelson, and J. F. Leckman, 2002. Development and natural history of mood disorders. Biol. Psychiatry 52(6):529–542. Costello, E. J., H. L. Egger, and A. Angold, 2004. Developmental epidemiology of anxiety disorders. In T. H. Ollendick and J. S. March, eds., Phobic and Anxiety Disorders in Children and Adolescents, 61–91. New York: Oxford Press. Davis, M., 1998. Are different parts of the extended amygdala involved in fear versus anxiety? Biol. Psychiatry 44:1239– 1247. De Bellis, M. D., B. J. Casey, R. E. Dahl, B. Birmaher, D. E. Williamson, K. M. Thomas, D. A. Axelson, K. Frustaci, A. M. Boring, J. Hall, and N. D. Ryan, 2000. A pilot study of amygdala volumes in pediatric generalized anxiety disorder. Biol. Psychiatry 48:51–57. Drevets, W. C., 2000. Neuroimaging studies of mood disorders. Biol. Psychiatry 48:813–829. Fox, N. A., K. Nichols, H. Henderson, K. Rubin, C. Schmidt, D. Hamer, M. Ernst, and D. S. Pine, 2005. Evidence for a gene-environment interaction in predicting behavioral inhibition in middle childhood. Psychol. Sci. 16(12):921–926. Grillon, C., 2002. Associative learning deficits increase symptoms of anxiety in humans. Biol. Psychiatry 51:851–858. Grillon, C., V. Warner, J. Hille, K. R. Merikangas, G. E. Bruder, C. E. Tenke, Y. Nomura, P. Leite, and M. M. Weissman, 2005. Families at high and low risk for depression: A three-generation startle study. Biol. Psychiatry 57:953–960. Gross, C., and Hen, R., 2004. The developmental origins of anxiety. Nature Rev. Neurosci. 5:545–552. Hariri, A. R., E. M. Drabant, K. E. Munoz, B. S. Kolachana, V. S. Mattay, M. F. Egan, and D. R. Weinberger,

2005. A susceptibility gene for affective disorders and the response of the human amygdala. Arch. Gen. Psychiatry 62: 146–152. Hariri, A. R., V. S. Mattay, A. Tessitore, B. Kolachana, F. Fera, D. Goldman, M. F. Egan, and D. R. Weinberger, 2002. Serotonin transporter genetic variation and the response of the human amygdala. Science 297:400–403. Heinz, A., D. F. Braus, M. N. Smolka, J. Wrase, I. Puls, D. Hermann, S. Klein, S. M. Grusser, H. Flor, G. Schumann, K. Mann, and C. Buchel, 2005. Amygdala-prefrontal coupling depends on a genetic variation of the serotonin transporter. Nature Neurosci. 8:20–21. Hofmann, S. G., A. E. Meuret, J. A. Smits, N. M. Simon, M. H. Pollack, K. Eisenmenger, M. Shiekh, and M. W. Otto, 2006. Augmentation of exposure therapy with D-cycloserine for social anxiety disorder. Arch. Gen. Psychiatry 63:298– 304. Kalisch, R., E. Korenfeld, K. E. Stephan, N. Weiskopf, B. Seymour, and R. J. Dolan, 2006. Context-dependent human extinction memory is mediated by a ventromedial prefrontal and hippocampal network. J. Neurosci. 26:9503–9511. Kendall, P. C., E. Flannery-Schroeder, S. M. PanichelliMindel, M. Southam-Gerow, A. Henin, and M. Warman, 1997. Therapy for youths with anxiety disorders: A second randomized clinical trial. J. Consult. Clin. Psychol. 65:366– 380. Klein, D. F., 1996. Panic disorder and agoraphobia: Hypothesis hothouse. J. Clin. Psychiatry 57(Suppl. 6.):21–27. Lane, R. D., G. R. Fink, P. M. Chau, and R. J. Dolan, 1997. Neural activation during selective attention to subjective emotional responses. NeuroReport 8:3969–3972. LeDoux, J. E., 2000. Emotion circuits in the brain. Annu. Rev. Neurosci. 23:155–184. Levine, S., 1957. Infantile experience and resistance to stress. Science 126:405. Levine, S., 1967. Maternal and environmental influences on the adrenocortical response to stress in weanling rats. Science 156:258–260. Lissek, S., A. S. Powers, E. B. McClure, E. A. Phelps, G. Woldehawariat, C. Grillon, and D. S. Pine, 2005. Classical fear conditioning in the anxiety disorders: A meta-analysis. Behav. Res. Ther. 43:1391–1424. MacLeod, C., E. Rutherford, L. Campbell, G. Ebsworthy, and L. Holker, 2002. Selective attention and emotional vulnerability: Assessing the causal basis of their association through the experimental manipulation of attentional bias. J. Abnorm. Psychol. 111:107–123. McClure, E. B., C. S. Monk, E. E. Nelson, J. M. Parrish, A. Adler, R. J. R. Blair, S. Fromm, D. S. Charney, E. Leibenluft, E. Ernst, and D. S. Pine, in press. Abnormal attention modulation of fear circuit function in pediatric generalized anxiety disorder. Arch. Gen. Psychiatry 64(1):97–106. Meaney, M. J., 2001. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu. Rev. Neurosci. 24:1161–1192. Merikangas, K. R., S. Avenevoli, L. Dierker, and C. Grillon, 1999. Vulnerability factors among children at risk for anxiety disorders. Biol. Psychiatry 46:1523–1535. Miller, E. K., and J. D. Cohen, 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24:167– 202. Millham, M. P., A. C. Nugent, W. C. Drevets, E. Leibenluft, M. Ernst, D. S. Charney, and D. S. Pine, 2005. Selective

pine and monk: anxiety

767

reduction in amygdala volume in pediatric generalized anxiety disorder: A voxel-based morphometry investigation. Biol. Psychiatry 57:961–966. Monk, C. S., C. Grillon, J. M. Baas, E. B. McClure, E. E. Nelson, E. Zarahn, D. S. Charney, M. Ernst, and D. S. Pine, 2003. A neuroimaging method for the study of threat in adolescents. Dev. Psychobiol. 43:359–366. Monk, C. S., E. E. Nelson, E. B. McClure, K. Mogg, B. P. Bradley, E. Leibenluft, R. J. Blair, G. Chen, D. S. Charney, M. Ernst, and D. S. Pine, 2006. Ventrolateral prefrontal cortex activation and attentional bias in response to angry faces in adolescents with generalized anxiety disorder. Am. J. Psychiatry 163:1091–1097. Ollendick, T. H., B. Yang, N. J. King, Q. Dong, and A. Akande, 1996. Fears in American, Australian, Chinese, and Nigerian children and adolescents: A cross-cultural study. J. Child Psychol. Psychiatry 37:213–220. Perez-Edgar, K., and N. A. Fox, 2005. Temperament and anxiety disorders. Child Adolesc. Psychiatr. Clin. N. Am. 14:681–706, viii. Pezawas, L., A. Meyer-Lindenberg, E. M. Drabant, B. A. Verchinski, K. E. Munoz, B. S. Kolachana, M. F. Egan, V. S. Mattay, A. R. Hariri, and D. R. Weinberger, 2005. 5-HTTLPR polymorphism impacts human cingulate-amygdala interactions: A genetic susceptibility mechanism for depression. Nature Neurosci. 8:828–834. Phelps, E. A., 2006. Emotion and cognition: Insights from studies of the human amygdala. Annu. Rev. Psychol 57:27–53. Phelps, E. A., and J. E. LeDoux, 2005. Contributions of the amygdala to emotion processing: From animal models to human behavior. Neuron 48:175–187. Pine, D. S., 2003. Developmental psychobiology and response to threats: Relevance to trauma in children and adolescents. Biol. Psychiatry 53:796–808. Pine, D. S., 2007. Research review: A neuroscience framework for pediatric anxiety disorders. J. Child Psychol. Psychiatry 48(7): 631–648. Pine, D. S., M. Alegria, E. H. Cook, E. J. Costello, R. E. Dahl, D. Koretz, K. R. Merikangas, A. L. Riess, and B. Vitiello, 2002. Advances in developmental science and DSM-V. In D. J. Kupfer, M. B. First, and D. A. Regier, eds., A Research, Agenda for DSM-V. 85–122. Washington, DC: American Psychiatric Association Press. Pine, D. S., P. Cohen, and J. Brook, 2001. Adolescent fears as predictors of depression. Biol. Psychiatry 50:721–724. Pine, D. S., P. Cohen, D. Gurley, J. Brook, and Y. Ma, 1998. The risk for early-adulthood anxiety and depressive disorders in adolescents with anxiety and depressive disorders. Arch. Gen. Psychiatry 55:56–64. Pine, D. S., R. G. Klein, R. Roberson-Nay, S. Mannuzza, J. L. Moutlon, G. Woldehawariat, and M. Guardino, 2005a. Face emotion processing and risk for panic disorder in youth. J. Am. Acad. Child Adolesc. Psychiatry 44:664–672. Pine, D. S., R. G. Klein, R. Roberson-Nay, S. Mannuzza, J. L. Moulton, 3rd, G. Woldehawariat, and M. Guardino, 2005b. Response to 5% carbon dioxide in children and adolescents: Relationship to panic disorder in parents and anxiety disorders in subjects. Arch. Gen. Psychiatry 62(1):73– 80. Pine, D. S., S. Lissek, R. G. Klein, S. Mannuzza, J. L. Moulton, 3rd, M. Guardino, and G. Woldehawariat, 2004. Facememory and emotion: Associations with major depression in children and adolescents. J. Child Psychol. Psychiatry 45:1199– 1208.

768

Pine, D. S., K. Mogg, B. P. Bradley, et al., 2005. Attention bias to threat in maltreated children: Implications for vulnerability to stress-related psychopathology. Am. J. Psychiatry 162(2): 291–296. Prather, M. D., P. Lavenex, M. L. Mauldin-Jourdain, W. A. Mason, J. P. Capitanio, S. P. Mendoza, and D. G. Amaral, 2001. Increased social fear and decreased fear of objects in monkeys with neonatal amygdala lesions. Neuroscience 106:653– 658. Quirk, G. J., and D. R. Gehlert, 2003. Inhibition of the amygdala: Key to pathological states? Ann. NY Acad. Sci. 985:263– 272. Rauch, S. L., L. M. Shin, and C. I. Wright, 2003. Neuroimaging studies of amygdala function in anxiety disorders. Ann. NY Acad. Sci. 985:389–410. Ressler, K. J., B. O. Rothbaum, L. Tannenbaum, P. Anderson, K. Graap, E. Zimand, L. Hodges, and M. Davis, 2004. Cognitive enhancers as adjuncts to psychotherapy: Use of Dcycloserine in phobic individuals to facilitate extinction of fear. Arch. Gen. Psychiatry 61:1136–1144. Roberson-Nay, R., E. B. McClure, C. S. Monk, E. E. Nelson, A. E. Guyer, S. J. Fromm, D. S. Charney, E. Leibenluft, J. Blair, M. Ernst, and D. S. Pine, 2006. Increased amygdala activity during successful memory encoding in adolescent major depressive disorder: An fMRI study. Biol. Psychiatry 60(9):966–973. Rosenbaum, J. F., J. Biederman, D. R. Hirshfeld-Becker, J. Kagan, N. Snidman, D. Friedman, A. Nineberg, D. J. Gallery, and S. V. Faraone, 2000. A controlled study of behavioral inhibition in children of parents with panic disorder and depression. Am. J. Psychiatry 157:2002– 2010. RUPP, A.S.G., 2001. Research Unit on Pediatric Psychopharmacology Anxiety Study Group. Fluvoxamine for the treatment of anxiety disorders in children and adolescents. N. Engl. J. Med. 344:1279–1285. RUPP, A.S.G., 2003. Research Unit on Pediatric Psychopharmacology Anxiety Study Group. Searching for moderators and mediators of pharmacological treatment effects in children and adolescents with anxiety disorders. J. Am. Acad. Child Adolesc. Psychiatry 42:13–21. Rutter, M., 2000. Psychosocial influences: Critiques, findings, and research needs. Dev. Psychopathol 12:375–405. Rutter, M., J. Kim-Cohen, and B. Maughan, 2006. Continuities and discontinuities in psychopathology between childhood and adult life. J. Child Psychol. Psychiatry 47:276–295. Schwartz, C. E., C. I. Wright, L. M. Shin, J. Kagan, and S. L. Rauch, 2003. Inhibited and uninhibited infants “grown up”: Adult amygdalar response to novelty. Science 300:1952– 1953. Shaffer, D., P. Fisher, M. K. Dulcan, M. Davies, J. Piacentini, M. E. Schwab-Stone, B. B. Lahey, K. Bourdon, P. S. Jensen, H. R. Bird, G. Canino, and D. A. Regier, 1996. The NIMH Diagnostic Interview Schedule for Children Version 2.3 (DISC-2.3): Description, acceptability, prevalence rates, and performance in the MECA Study: Methods for the Epidemiology of Child and Adolescent Mental Disorders Study. J. Am. Acad. Child Adolesc. Psychiatry 35:865–877. Siegle, G. J., R. O. Konecky, M. E. Thase, and C. S. Carter, 2003. Relationships between amygdala volume and activity during emotional information processing tasks in depressed and never-depressed individuals: An fMRI investigation. Ann. NY Acad. Sci. 985:481–484.

neurodevelopmental aspects of clinical disorders

Silberg, J. L., M. Rutter, and L. Eaves, 2001. Genetic and environmental influences on the temporal association between earlier anxiety and later depression in girls. Biol. Psychiatry 49:1040– 1049. Suomi, S. J., 2003. Gene-environment interactions and the neurobiology of social conflict. Ann. NY Acad. Sci. 1008:132– 139. Thomas, K. M., W. C. Drevets, R. E. Dahl, N. D. Ryan, B. Birmaher, C. H. Eccard, D. Axelson, P. J. Whalen, and B. J. Casey, 2001. Amygdala response to fearful faces in anxious and depressed children. Arch. Gen. Psychiatry 58:1057– 1063.

Vasa, R., and D. Pine, 2004. Neurobiology. In T. L. Morris and J. S. March, eds., Anxiety Disorders in Children and Adolescents, 2nd ed., 3–26. New York: Guilford Press. Weissman, M. M., V. Warner, P. Wickramaratne, D. Moreau, and M. Olfson, 1997. Offspring of depressed parents: 10 Years later. Arch. Gen. Psychiatry 54:932–940. Whalen, P. J., L. M. Shin, L. H. Somerville, A. A. McLean, and H. Kim, 2002. Functional neuroimaging studies of the amygdala in depression. Semin. Clin. Neuropsychiatry 7:234–242. Williams, J. M., A. Mathews, and C. MacLeod, 1996. The emotional Stroop task and psychopathology. Psychol. Bull. 120:3– 24.

pine and monk: anxiety

769

47

Developmental Neuropsychology of Unipolar Depressions IAN M. GOODYER AND ZOË KYTE

Over the past two decades research findings have emphasized marked individual differences in the characteristics, treatment response, and outcomes of depressive disorders in young people (Goodyer, 2001a). Longitudinal studies suggest that a depressive illness before 18 years of age may have adverse consequences for the developmental trajectory into adult life (Pine et al., 1999; Fombonne et al., 2001a, 2001b; Dunn and Goodyer, 2006). Though there has been considerable research on the effects of social adversities on onset, course, and outcomes of depressions, it is only in the last decade that psychological functions have begun to be intensively studied (Tavares, Drevets, and Sahakian, 2003; Kyte and Goodyer, 2005). The majority of the work has been in depressed adults using predominantly cross-sectional designs comparing this group with nondepressed control participants. These important proof-of-concept studies have noted that currently ill patients with unipolar depressions show a range of deficits in memory, attention, social cognition, decision making, and planning that are not explained by variations in general intelligence within the normal range or other premorbid factors. Most recently, functional neuroimaging studies have delineated some of the putative neural substrates that subserve these functions (Liotti and Mayberg, 2001; Jack, Sylvester, and Corbetta, 2006). The precise role of these mood-valent cognitive processes in the onset, treatment response, and outcome of unipolar depressions will unfold through prospective and longitudinal research in population- and clinical-based studies. The application of sensitive tests of emotion and cognition to children and adolescents brings the added complexity of maturation of the central nervous system and the need for developmentally sensitive methods and measures (Luciana and Nelson, 2002; Luciana, 2003a; Luciana et al., 2005). At the neural level, we lack a clear understanding of the correlates between brain structure and functions over the first two decades of life, although these are beginning to emerge (Durston et al., 2001; Paus et al., 2001; Paus, 2005; Blakemore and Choudhury, 2006a). As a result, we do not yet possess a clear understanding of the continuities and discontinuities between skills at different ages and stages of development and the role of deficits in these skills for the onset and outcomes of depressions in young people. For example, prospective psychosocial research has noted that

greater than 90 percent of episodes of depression in adolescents arise in populations who have been exposed to longstanding chronic adversities including marital disharmony, negative effects of parental psychiatric disorder, neglect, and abuse (Goodyer, 2001b). Within this exposed adolescent group, however, a significant proportion does not become mentally ill. Exposure to acute highly undesirable life events brings forward the onset of episodes in those exposed to chronic social difficulties, but such exposure alone has minimal effects on the risk for unipolar depression (Goodyer et al., 2000a, 2000b). These marked variations beg the question, what are the affective-cognitive processes that mediate the relationship between social difficulties and adverse behavioral response (Goodyer, 2002). The psychological findings to date represent a first pass at characterizing the neuropsychological components of the intermediate biology of depressions in young people.

Building a brain Primary neurulation is a critical phase of early development in which there is considerable shaping, by means of folding and rolling of neural components giving rise to the brain and spinal cord (Brown, Keynes, and Lumsden, 2001). Secondary neurulation refers to morphogenetic movements of the developing nervous system. The size, shape, and movement of cells are under genetic control, and these changes promote cell migration and the development of specialized portions of the nervous system, such as sensory organs. Whether defects in secondary neurulation can affect neural systems that are required for higher mental functions such as emotion recognition and behavioral inhibition is not yet known. Neurotransmitters and their receptors are expressed on a wide scale before synaptic connections are made, with regulatory mechanisms coordinating the patterning and connectivity of neurotransmitter phenotypes. Early disruptions of the normative trajectory of monoamine pathways may have profound effects on the development of neural systems subserving psychiatric syndromes (Whitaker-Azmitia, 2001). Neurochemicals may also have quite different actions at differing times in development. In addition to acting as switches in serotonin functions, steroids such as cortisol and testosterone exert organizing effects on brain structure during

771

prenatal life but activating effects on functions during postnatal life (Buchanan, Eccles, and Becker, 1992). Similarly, GABA exerts excitatory and inhibitory functions in the prenatal and postnatal brain, respectively. Cross talk is frequent between chemical systems. For example, high central neurosteroid levels may diminish inhibitory GABA activity, and both may play a role in modulating emotional state (Barbaccia et al., 2001). For many aspects of their development and normal physiology, neurons are dependent on a small group of highly specific proteins known as neurotrophic factors (Huang and Reichardt, 2001). The principal roles of these factors are to regulate axonal and dendritic growth, synaptic plasticity, and neuronal survival. Reduced neuronal survival and deficits in morphological differentiation can lead to abnormal formations of brain regions. In the mature nervous system, neurotrophins can regulate short-term synaptic transmission and long-term potentiation, a process that is used as a model for learning and memory. Later in development, they may determine the phenotype of developing sensory and autonomic nerve cells during the critical period of rearrangement and stabilization of nerve cells. Throughout the first few years of postnatal life, there is simultaneous progressive and regressive restructuring of neural networks occurring in a region-specific manner (Durston et al., 2001). A striking feature of this neural development and growth is synaptogenesis (the formation of new synapses), with differential bursts of activity and variations in peak density occurring at different ages in different brain regions (Huttenlocher and Dabholkar, 1997). The prefrontal cortex is the slowest brain area to reach peak synaptogenesis—by about the end of the first year of life. Sensory areas tend to myelinate earlier than motor areas. Magnetic resonance studies of children and adolescents have demonstrated a gray-white matter contrast in a temporal sequence reflecting the time course of myelination (Durston et al., 2001; Paus et al., 2001). White matter increases in overall volume and becomes increasingly myelinated in a region-specific manner. Thus the brain myelinates from caudal to rostral and from inferior to superior, with the prefrontal cortex continuing this process (albeit more slowly than in the first two years) until young adult life (Casey, 1999; Giedd et al., 1999; Sowell et al., 2000). The largest white matter tract, the corpus callosum (CC), also continues to increase in size as a consequence of myelination (Giedd et al., 1999). Contrary to the typical pattern of caudal to rostral development in the brain, the corpus callosum and anterior regions of the cortex, which have been related to primary sensory and motor functions, appear to mature in childhood and the posterior areas in adolescence (Jernigan et al., 1991; Thompson et al., 2000). In contrast to white matter, gray matter generally exhibits a net decrease in volume across the school-age years (Jernigan et al., 1991; Thompson et al., 2000). This may be due

772

to the ongoing processes of pruning and cell death (apoptosis) among both neurons and glial cells into young adult life. This apoptotic effect of programmed cell death is critical for normal brain development (Kuan et al., 2000). Subcortical gray matter typically shows the same pattern, with decreases in basal ganglia volume with increasing age. By contrast, the temporal lobe structures amygdala and hippocampus appear to increase in volume with age (Giedd et al., 1996). Some sex differences have been reported, with the male brain being 10 percent larger than the female and most structures displaying this volume difference (Giedd et al., 1997, 1996). The caudate, and possibly the globus pallidus and hippocampus, are disproportionately larger in female brains, whereas the amygdala is disproportionately smaller. The rate and pace of nonlinear decreases in gray matter and linear increases in white matter over time are correlated with individual differences in general intellectual ability, indicating a broad association between psychological functions and neural maturation (Barnea-Goraly et al., 2005; Paus, 2005; Shaw et al., 2006). Thus there is a developmental correlation between brain and behavioral maturation, the precise details of which remain underspecified. Defining the changing relationships between mental functions and the neural systems that subserve these will form the thrust of developmental neuroscience in the next few years. Such studies will be a key component in mapping typical development and will provide a platform for elucidating the mechanisms and processes that result in atypical development.

Building a mind The maturing brain over the child and adolescent period subserves the emergence of complex emotions and cognitions that will be lifelong functions (Blakemore and Choudhury, 2006a; Shaw et al., 2006). Although the origins of key features of cognitive functions, including executive processes, are embedded early in neurodevelopment (Luciana and Nelson, 1998, 2002; Hughes et al., 2000; Luciana et al., 2005; Blakemore and Choudhury, 2006b), their comprehensive emergence as mature functions unfolds over the first two decades. Indeed neural system reorganization during the adolescent period is itself associated with improving executive capacity to control and coordinate feelings, thoughts, and behaviors. Thus information-processing speed, working memory, and decision making continue to improve throughout the second decade of life (Luciana et al., 2005; Blakemore and Choudhury, 2006b). One striking observation of studies to date is an overall transient dip in performance during puberty (Blakemore and Choudhury, 2006a, 2006b). This has led to the suggestion that there is a perturbation of cognitive performance in early adolescence perhaps associated with synaptic pruning into specialized

neurodevelopmental aspects of clinical disorders

efficient networks. These behavioral findings are consistent with histological studies showing synaptic proliferation in the prefrontal cortex during childhood, with a subsequent elimination and reorganization of these connections after puberty (Huttenlocher, 1979; Huttenlocher and Dabholkar, 1997). Adolescence is also a period associated with the emergence of major mental illnesses including unipolar depressions. During adolescence, clinical depressions are estimated to occur in around 3–5 percent of the population, compared with around 1–2 percent in prepubertal children (Angold and Costello, 2001). Among these, some 50 percent will experience a second episode before their twenties, and around 20 percent will not recover at all (Lewinsohn et al., 1999; Fombonne et al., 2001b; Dunn and Goodyer, 2006). An episode of depression during a period of marked brain reorganization may have long-term effects on subsequent structure and function. Thus neuropsychological processes may be altered as a consequence of depressive disorders as well as being vulnerability factors that may be instrumental in their emergence.

Clinical characteristics of depressive disorders Depressive disorders constitute a serious group of mental disorders with considerable risk for recurrence and subsequent psychosocial impairment with, in some cases, onset in childhood and adolescence and continuity into adult life (Brotman et al., 2006; Dunn and Goodyer, 2006; Fombonne et al., 2001a, 2001b). Use of the American Psychiatric Association DSM-IV clinical criteria (American Psychiatric Association, 1994) successfully identifies the same clinical syndromes in children as young as 4, adolescents, and adults, although with differing clinical presentations, which we will describe in detail. Phenomenological psychopathology remains the core construct guiding the identification of criteria for diagnosis with no dependency on putative causation. This system has proven highly successful in generating a common language for the diagnosis of depressive disorders that makes research findings comparable between investigations despite marked differences in the nature of the populations studied. The categorical approach to classification of depression, however, is not comprehensive. For example, within diagnostic groups the relative importance of different types of symptoms that are present at different ages is underspecified (Cooper and Goodyer, 1993; Goodyer and Cooper, 1993). In addition, many individuals present with a range of symptoms that fall below the inclusion threshold for a diagnosis (too few or insufficient duration) but are of sufficient severity to cause personal impairment and a poor adjustment in the long term (Costello et al., 1996; Fergusson et al., 2005). Some advantage is gained by describing patients as suffering from mild, moderate, or severe disorder depending

Mood Dysphoria or irritability in children

Table 47.1 Symptoms of major depressive disorder Cognitive Physical Anhedonia Weight changes Feelings of (includes a failure worthlessness to make expected Inappropriate guilt weight gains) Diminished ability Fatigue or loss of to think or energy concentrate Psychomotor Recurrent thoughts agitation or of death or retardation suicidal ideation

on the number of symptoms (4–5, 6–7, 8 and above, respectively), thereby incorporating some of the dimensional qualities of depression. Unipolar Major Depression According to DSM-IV (American Psychiatric Association, 1994), a diagnosis of unipolar depression requires establishing first the mandatory presence of lowered mood (dysphoria, or irritability if in children) or loss of interest/pleasure (anhedonia) together with four of seven other possible nonmandatory symptoms from the two broad domains of disordered cognitions and physical changes. These are shown in table 47.1. It is essential that the five symptoms occur concurrently over a minimum two-week period. In children and adolescents, but not in adults, the entry criteria of lowered mood can be met by the presence of irritability or sadness. It is important to establish that these symptoms are not accounted for by the direct effects of substance misuse or a general medical condition, particularly one that involves known brain changes, as these reduce the likelihood of reliable and valid mental state assessments. In addition, the symptoms should not be accounted for by recent bereavement. It is worth noting, however, that these caveats do not imply that such subjects cannot be subsequently clinically depressed as a consequence of these experiences, rather that symptoms essential to the diagnosis may be acute or transient and therefore increase the liability of a false positive diagnosis at the time of presentation. There are no requirements for a particular pattern of cognitions and/or physical symptoms, although such patterns may be developmentally sensitive (Ryan, Puig-Antich, and Ambrosini, 1987; Goodyer and Cooper, 1993; Kovacs, 1996). Depressed children are more likely to complain of physical symptoms such as headache and abdominal pains with low levels of suicidal thinking, hopelessness, and helplessness present at clinical assessment (Kovacs, 1996). In contrast, by midadolescence negative cognitions about the self, the world, and the future become prominent together with higher levels of suicidal thoughts and acts (Ryan,

goodyer and kyte: unipolar depressions

773

Puig-Antich, and Ambrosini, 1987). Anhedonia, a complex state characterized by lack of interest and pleasure and often accompanied by low energy, also appears more common in depressed adolescents than in children. Symptom profiles may also change over the adolescent period. For example, girls over 15 years report higher rates of feeling and looking depressed and having self-deprecatory thoughts than younger depressed females (Goodyer and Cooper, 1993). There are no community studies that have examined age effects on clinical symptoms in adolescent males. In general, psychotic symptoms are less common than in depressed adults. Collectively, evidence suggests that there are significant developmental influences on the nature, characteristics, outcomes, and treatments of depressions (Kaufman et al., 2001). Because of these developmental influences, the precise identification of mood disorders in young people remains problematic. For example, severe mood dysregulation with high levels of irritability, disorganized behavior, and hyperarousal in childhood are frequently diagnosed as suffering from conduct disorder or attention-deficit/hyperactivity disorder (ADHD), yet many of these individuals develop frank clinical depression by early adult life (Brotman et al., 2006). This broad concept of severe mood dysregulation syndrome requires much further investigation than has been provided to date. Children as young as 3 or 4 years may, with careful assessment, show the full major depressive syndrome (Luby et al., 2003). The superficial presentation of externalizing symptoms in childhood (the most common referral characteristic to clinicians) may mask underlying mood and cognitive disturbance. Thus “hard to manage children” who show executive difficulties may do so because of mood as well as behavioral dysregulation. In contrast to symptom profiles, no developmental distinction is made regarding the duration of unipolar depressions which, providing symptoms have been present for at least two weeks, may vary in length for any period of time, up to a number of years. The disorder may also differ in its severity or degree of psychosocial impairment ranging from mild, indicating only a modest deviation from normal behavioral functioning, to being unable to care for oneself and requiring 24-hour intensive psychiatric care. A significant proportion of unipolar depressions across the life span are preceded by the existence of nondepressive clinical disorders that appear to convey an increased risk for a poor outcome of the depressive disorder (Kim-Cohen et al., 2003). There is now general agreement that unipolar depressions are heterogeneous in etiology and clinical presentation. Genetic and environmental (physiological, psychological, and social) processes combine in as yet undetermined ways to increase the liability for these complex and common mental disorders. For some, episodes are sporadic and relatively short-lived (weeks), whereas for others, a first

774

episode heralds the emergence of a protracted and disabling condition with a high risk of recurrence and persistence for many months and years (Goodyer, Herbert, and Tamplin, 2003; Dunn and Goodyer, 2006). Mild disorders with fewer than six symptoms have a better prognosis overall, with some two-thirds likely to undergo spontaneous recovery with no increased risk of relapse in the ensuing 5-year period (Fombonne et al., 2001a, 2001b; Dunn and Goodyer, 2006). In contrast, child and adolescent patients with severe depression (more than seven symptoms at presentation) are at considerable risk of relapse and in some cases chronic illness into adult life (Pine et al., 1999; Fombonne et al., 2001a, 2001b; Dunn and Goodyer, 2006). The majority of clinical neuropsychological research into affective disorders in young people has been focused on moderate to severe unipolar major depressions where there are at least six symptoms and often more. These studies have, in general, been cross-sectional in nature involving participants with a recent or a current episode of unipolar depression. In this regard, studies to date represent the first step toward an understanding of the neurocognitive and neuroaffective systems that may act as vulnerability processes for the clinical phenotype of the unipolar depressions.

Social cognition, unipolar depression, and the brain Social cognition refers to the cognitive processes that subserve the diverse and flexible range of social behaviors. A key issue in investigating the psychology of depressions in recent years has been to compare and contrast cognitive processes in neutral mood (cold cognitions) and dysphoric mood (hot cognitions) (Kyte and Goodyer, 2005). This approach is taken because of the distinctions reported between mood-valent and mood-neutral cognitive processing (Kelvin et al., 1999; Teasdale and Barnard, 1993). Given that unipolar depression appears to be associated with abnormal functioning in both higher cognitive and limbic domains, consensus is now emerging to explain the phenomenology of depression on the basis of a malfunction in the regulation of an entire network of brain regions involved in both emotional behavior and social cognition (Drevets, 2000; Seminowicz et al., 2004). These neural structures are now considered within two critical circuits: a limbic-thalamic-cortical (LTC) circuit, involving the amygdala, medial thalamus, and orbital and medial prefrontal cortices; and a limbic-cortical-striatal-pallidal-thalamic (LCSPT) circuit, involving components of the previous circuit with the addition of related parts of the striatum and pallidum. From the developmental perspective, we have yet to map neural systems that subserve depressive symptoms at different ages. Thus it is not known to what extent some of these brain regions may be developmentally impaired, structurally immature, or functionally inactive prior to the

neurodevelopmental aspects of clinical disorders

onset of a disorder at different ages. Equally, it is unclear to what extent depressive episodes may alter the pathway of neural development, resulting in a vulnerable brain liable to recurrent illness. Thus a number of features of the study population need to be taken into account when researching the underlying cognitive processes that mediate the liability for clinical symptoms. Etiological research needs to study individuals at risk but with no prior history of mood disorder; developmental influences on etiology require at-risk populations at different stages of maturity to determine the constancy and stability of vulnerable brain regions in the risk process; pathophysiological studies require patients with matched clinical characteristics including duration, severity, number of prior episodes, comorbidity, and levels of impairment. Such precision is likely to inform neuro-cognitive systems at differential levels of sensitivity to the disease process. Finally, there is a critical need to distinguish experimentally between “cold” and “hot” cognitive processing at both the psychological and neural systems levels.

Cognitive theories of depression Cognitive theories have been developed in relative ignorance of underlying neural systems. In addition, while all theories of depression postulate a role of adverse early experiences in current thinking style, none have determined precisely what these experiences are or how they result in the formation of negative cognitions. Developmental psychology has, however, extensively examined the effects of social influences on general cognitive development. One focus of this work is to add key information on how individual differences in early maternal experiences and the resultant relational psychology between mother and child could result in negative views of the self (Murray et al., 1999; Gibb and Alloy, 2006; Murray et al., 2006). These distal processes have yet to be fully delineated or incorporated into theoretical frameworks. The majority of investigations to date have focused on elucidating proximal psychosocial mechanisms that lead to clinical depression. Appraisal and Attribution Theories A key component of these theories is the presence of a latent negative cognitive set that is activated by a recent undesirable personal experience and results in clinical depression (Beck, 1967; Abramson, Seligman, and Teasdale, 1978; Seligman et al., 1984). Under these proximal conditions, working memory is fully occupied by negative dysphoric thoughts about the self and/or attributing negative effects of events to the self. There is considerable evidence that such cognitive vulnerabilities exist within adults with a previous episode of depression (Segal and Ingram, 1994; Ingram, Miranda, and Segal, 1998; Just, Abramson, and Alloy, 2001).

The Oregon Adolescent Depression Project (OADP) reported that dysfunctional attitudes acted as a vulnerability factor for both first episode and recurrent adolescent depressive disorder, if combined with conditions of stress (Lewinsohn et al., 1994; Lewinsohn, Rohde, and Seeley, 1998; Lewinsohn, Joiner, and Rohde, 2001). The Cognitive Vulnerability to Depression Project showed that depressogenic cognitive styles conferred specific risk for first-onset depression, regardless of negative life events (Alloy et al., 1999, 2000). Mood-Valent Negative Cognitions There is some evidence that perhaps up to 50 percent of adolescents with “cold” negative cognitions exposed to social adversities develop depression in the short term (Goodyer et al., 2000a, 2000b). Teasdale and Barnard (1993) suggested that vulnerability to intense and persistent depression may be determined by individual differences in the patterns of negative thinking that become activated as a consequence of an event-related dysphoric mood state. During such a state, vulnerable individuals are hypothesized to access qualitatively different and more depressogenic types of negative cognitions. Teasdale and colleagues initially demonstrated dysphoric activation of an attentional bias for negative thinking style within individuals with a prior history of depression (Teasdale and Dent, 1987; Teasdale et al., 1995; Teasdale, Lloyd, and Hutton, 1998). Subsequent experimental work demonstrated the presence of mood-valent negative cognitions in well adolescents with no lifetime history of affective disorders (Kelvin et al., 1999). Kelvin and colleagues further showed that this mood-related attentional bias was particularly salient in those at temperamental risk (defined as persistently high emotionality over 3 years but with no evidence of psychiatric disorder) for mood disorder and provided the first evidence for a potential causal role of cognitive vulnerabilities arising in the first two decades of life. Individual differences in basic attentional processing of social information may be a key variable in determining moodvalent attentional bias. Mood-Valent Cognitive Response Style Moderate to severely depressed patients are frequently severely preoccupied by their negative thoughts, invariably being unable to remove them from consciousness. Mood-valent theories of activating negative cognitions do not explain how or why these become persistent and in many cases immovable from consciousness. A ruminative response style to dysphoric activated negative cognitions has been proposed as another intermediate vulnerability process between underlying latent negative cognitions and the onset of clinical depressions (Just and Alloy, 1997; Nolen-Hoeksema, 1991). Ruminating on the putative causes (“Why am I like this?”) and the salience of depressive symptoms (“I won’t be able to do my work because I feel so bad”) has been proposed as the most

goodyer and kyte: unipolar depressions

775

proximal cognitive mechanism through which other cognitive vulnerability factors affect depression (Spasojevic and Alloy, 2001). Ruminative response to dysphoria and related negative thoughts is associated with prolonged low mood, increased depressive symptoms, onset of clinical depression in young adults, and persistence of an episode in adolescents (Lyubomirsky and Nolen-Hoeksma, 1993; Nolen-Hoeksema, 2000; Just, Abramson, and Alloy, 2001; Spasojevic and Alloy, 2001; Goodyer, Herbert, and Tamplin, 2003). Interestingly, little is known about either the developmental origins or the neuropsychological correlates of this highly dysfunctional process of perseverative negative thinking. Some early evidence argues that ruminating is more likely in those with a childhood history of abuse or neglect, or may be a familial cognitive style with a stronger association reported between mothers’ and daughters’ ruminative style than between that of fathers and sons (Nolen-Hoeksma, 2004).

Executive dysfunctions of unipolar depression Within the domain of human social interaction, performance of any task generally requires that a sequence of operations occurs that includes both mental processes and overt actions. Among others, these consist of developing an awareness of the desired goal or outcome of the event, developing a strategic plan of action in order to achieve this goal, and inhibiting or deferring inappropriate behavioral responses so that the most appropriate response option can be initiated. Each of these individual processes, or “context-specific action selections” (Pennington and Ozonoff, 1996) is thought to be dependent upon the integration of information from a range of abilities, including attention, planning, decision making, inhibition, and memory (Tamminga, 2000). These abilities are referred to as executive functions. Any social interaction or circumstance that demands optimal responding is therefore likely to recruit and be reliant upon, at least in part, the expression of these abilities. Depressed patients’ inabilities to change their negative mental set and their tendencies to remain highly dysphoric and become increasingly inactive and withdrawn suggest dysfunctions in one or more executive processes. Overall, however, studies have been unsuccessful in accumulating evidence to suggest a profile of deficits that is specific to unipolar depression compared to other psychopathologies. Instead, it seems there is an extensive range of neuropsychological deficits measurable during an episode or after recovery that are associated with a globaldiffuse impairment in executive functions that may be dependent on the age of the population and the severity, duration, and characteristics of the depressive episode (Lockwood et al., 2000; Schatzberg et al., 2000; Sweeney, Kmiec, and Kupfer, 2000a; Grant, Thase, and Sweeney, 2001; Weiland-Fiedler et al., 2004; Paelecke-Habermann, Pohl,

776

and Leplow, 2005; Pardo et al., 2006). These and many other studies demonstrate difficulties in the areas of attention, behavioral inhibition, memory, decision making, and planning. Very little is known about healthy individuals who are at risk for depression. Studies are beginning to emerge that document executive skills in child and adolescent populations in general (Hughes et al., 2000; Luciana, 2003b). There are few on depressed young people (Kyte, Goodyer, and Sahakian, 2005; Perez-Edgar et al., 2006). The main areas that have been examined to date have focused on three domains: decision making, behavioral inhibition, and attentional processing. Decision Making in Unipolar Depression Efficient and effective decision making is required by individuals to ensure that correct choices are made between alternative courses of action in order to keep behavior coherent and adaptive. Making decisions involves considering and anticipating desired future outcomes, weighing the probability of an outcome being successful, and judging whether personal resources are available and necessary to achieve that outcome. Central components of the decision-making process involve attentional and working-memory resources, are context sensitive, and are influenced by mood. The development of a computerized decision-making task (Rogers, Everitt, et al., 1999) has improved the measurement precision of decision making. In this task, the individual is required to make a decision, in terms of color, as to where the computer has hidden a yellow token (under a red or under a blue box, with ratios of each color changing across trials from 6 : 4 through 9 : 1). Once participants have made their decision on a given trial, they must then place a bet depending on how confident they are that they have chosen the correct color. Possible bets begin low and increase (ascend condition), or begin high and decrease (descend condition), and in each case, if the correct choice is made, the individual wins the points they were willing to bet. If the wrong decision is made, the individual loses the points she or he was willing to bet. In an attempt to best model real-life decision making, three features of this task are of interest: decision-making behavior across a range of contingencies (manipulation of the ratio of red to blue boxes across trials); individual efficiency (allowing individuals to decide for themselves how many of their points they wish to bet); and impulsivity versus genuine risk-taking behavior (offering bets in both ascending and descending conditions). From this conceptual framework, three principal measures are derived: (1) speed of decision making—how long it takes to decide which color box is hiding the token; (2) quality of decision making—how much of the time the subject chooses the most likely outcome; and (3) risk adjustment—the rate at which a subject increases the percentage of available bets

neurodevelopmental aspects of clinical disorders

100 Controls (n = 49) Cases (n = 30)

90 80 Mean % bets

70 60 50 40 30 20 10 0 6:4

7:3

8:2

9:1

6:4

7:3

8:2

9:1

Ratio (ascend/descend)

Figure 47.1 Decision making characteristics of recently depressed adolescents and never depressed controls. The figure shows formerly depressed adolescents making significantly higher bets on the descend condition on all ratios except 9 : 1 and in the ascend condi-

tion on ratios 6 : 4 and 7 : 3 only. The presence of inappropriate allocation of resources in both conditions is consistent with impulsive rather than risk-taking decision making.

in response to more favorable ratios of red to blue (i.e., 9 : 1 versus 6 : 4). In adult patients with unipolar depression, administration of this task has resulted in reports of suboptimal decision making characterized by a protracted time to make decisions and the employment of less responsive betting strategies indicated by the allocation of an inappropriate number of points to a given decision (Murphy et al., 2001). The ability to make decisions that are likely to produce the desired outcome, however, appears to be less impaired (Murphy et al., 2001), indicating that adults with unipolar depression remain able to effectively encode information about the likelihood of a reward response and make decisions accordingly. Using the same task, adolescents who had recently experienced a first-episode unipolar depression also displayed a preserved ability to make decisions that were likely to produce the desired outcome (Kyte, Goodyer, and Sahakian, 2005) as well as displaying suboptimal decision making in the form of an inappropriate allocation of resources (i.e., betting a large proportion of their points on a decision with an unfavorable outcome), although to a greater extent than seen in previous adult populations (Elliot et al., 1996, 1997). Although this finding may be consistent with the notion that adolescents are more inclined to take risks than older populations, when combined with results that indicated higher bets made overall in descending conditions (where bets begin high and decrease) compared to ascending conditions (where bets begin low and increase), the pattern becomes more consistent with that of a tendency to be impulsive (i.e., a failure to consider, analyze, and reflect prior to engaging in a response, leading adolescents with recent first-episode

unipolar depression to choose bets early in both conditions). This pattern of decision making is shown in figure 47.1. Although increases in impulsivity may serve as some explanation for the patterns of suboptimal decision-making profiles reported in unipolar depression, a number of other neuropsychological processes have also been implicated. These include a hyposensitivity to reward (the prospect of a large immediate gain outweighing any prospect of future loss); insensitivity to punishment (prospect of a large loss not overriding any prospect of gain); or insensitivity to future consequences (behavior always guided by immediate prospects). Recent fMRI studies in children and adolescents have noted hypoactivation in currently depressed patients in reward-related brain areas that involve the prefrontal cortex (Forbes et al., 2006). There are suggestions that the neural systems that subserve reward may be impaired in development, blunting the ability to discern positive emotional states (Forbes and Dahl, 2005; Forbes et al., 2006), and there is some evidence for similar impairments in adult depressed patients (Tremblay et al., 2005). Impairments in reward systems seem very pertinent to the symptom of anhedonia in currently depressed individuals. Developmentally, this finding suggests that reward systems may be less impaired in younger depressed patients, since anhedonia is less common as a depressive symptom in childhood than it is in adolescence (Goodyer and Cooper, 1993). Not all impairments on the decision-making task reflect deficits in reward systems. Depressed individuals may display compromised decision making as a result of an inability to resolve effectively between two competing response options (Rogers, Owen et al., 1999), a loss of the ability to ponder different courses of action (Tranel, Bechara, and Damasio,

goodyer and kyte: unipolar depressions

777

2000), or a degree of cognitive (rather than behavioral) impulsivity (Bechara, Tranel, and Damasio, 2000). A task of further developmental research is to disaggregate the effects of maturation on decision-making processes and the neural systems that subserve these (Crone et al., 2005; Crone and van der Molen, 2004). This would include describing the expected level of impulsivity (for example, by age, pubertal stage, or some physiological index of developmental change; see chapter 54 by Crone and van der Molen, this volume) in order to examine the unique contribution of being at risk for or suffering from illness to task performance at different stages in the life cycle. For example recent neuropsychological findings on participants aged 9 to 17 years suggest that the ventromedial prefrontal cortex or its connections are functionally maturing during adolescence in a manner that can be distinguished from maturation of other prefrontal regions. Development of these functions may continue into young adulthood (Hooper et al., 2004). Determining decision-making skills in adolescents in the population at large would also help us to understand how individual differences in risk taking and impulsivity produce age- and sex-differentiated standards and examine the extent to which variability in decision making acts as a vulnerability factor for mood disorders. Behavioral Inhibition and Biases in Unipolar Depression Inhibitory control is another central cognitive process that is contained within the range of executive functions. Inhibitory processes exert effects on the interrelated components of reasoning, planning, and control of behavior. In addition to using executive control to choose, construct, execute, and maintain optimal strategies in a given situation, an individual must also be able to inhibit strategies that become inappropriate when goals alter or task errors occur. In order to keep behavior coherent, relevant actions need to continue while irrelevant actions are inhibited. Plaisted and Sahakian (1997) proposed the “inhibition hypothesis,” which was developed in response to reports that patients with frontal lobe damage display deficits in behavioral inhibition in relation to salient social scenarios but not when tested on abstract scenarios. This hypothesis was based on observations that damage to the prefrontal cortex (PFC) often results in a loss of inhibitory control over inappropriate responses to a current situation (Dias, Robbins, and Roberts, 1997). Loss of behavioral inhibition can be characterized by deficits of PFC inhibitory mechanisms to suppress inappropriate behaviors elicited by cues from the immediate environment. Such disinhibition would prevent the selection of alternative and more appropriate action plans that are governed by more long-term goals, leaving behavior to become dominated by the immediate emotional evaluation of the environment. The result is that reactions are based purely on the experience of associated emotions and that behavior is unconstrained

778

because of a disregard for social norms, often, therefore, seeming inappropriate. Thus patients with deficits in social cognition respond to immediate but not future social situations. In a sense, such emotionally based, behaviorally disinhibited reactions may be seen as a form of impulsivity: a failure to fully consider, analyze, and reflect before engaging in a particular behavior. Since depressed patients have been shown to exhibit a bias in their processing for negative affective stimuli in both adult (Murphy et al., 1999) and adolescent (Kyte, Goodyer, and Sahakian, 2005) populations, impulsivity is likely to be emotion dependent during episodes of mood disorder. These findings raise the possibility of a primary attentional vulnerability in a subpopulation of adolescents that is activated under specific dysphoric-related circumstances that leads to impulsive behavior. Such vulnerability could underpin the selective encoding and/or recall of negative information. As already noted, mood-congruent processing biases are among the most robust findings in neuropsychological studies of depression and are central to several of the cognitive theories of the disorder discussed previously. The evidence for such a bias among adolescents with no lifetime history of unipolar depression who possess a temperamental vulnerability of high emotionality suggests that such a process may indeed be a true mood-valent cognitive vulnerability for subsequent affective disorders (Kelvin et al., 1999). Functional neuroimaging findings in adults using a moodvalent behavioral-inhibition task have shown that depressed patients show a marked adverse response to negative feedback, which may play a key role in mediating the interaction between mood and cognition in affective disorder (Elliott et al., 2002). It is unclear if this sensitivity to critical comment about performance is a state- or trait-related phenomenon. As yet there are no prospective studies of this process in individuals at risk for depressive disorders. For example, negative feedback sensitivity may be more common in adolescents with underlying latent self-devaluative traits or in those with higher mood-related ruminative response style. Heightened social sensitivity is found in adolescents exposed to postnatal depression and with a history of insecure attachment indicating a developmental ontogeny for an adverse reaction to negative feedback (Elliot et al., 1996; Murray et al., 2006). Furthermore this process when depressed suggests a psychological mechanism that may give rise to poor friendships that are present in depressed adolescents. They may be oversensitive to the ordinary rough and tumble of day-to-day interpersonal communications resulting in high levels of personally disappointing life events resulting from misinterpretation of perturbations in relationships. Interestingly, the relative strengths and weaknesses of the attentional system itself have not been well studied in depressive conditions. As noted previously, determining the relative strength of the attentional system in those at risk for

neurodevelopmental aspects of clinical disorders

depressive disorders is an important research task for the future. Whether individuals who show a mood-valent attentional bias also have subtle deficits in one or more aspects of the attentional processing system that are mood independent is not known. Depression and Impaired Attention A striking clinical symptom of unipolar depressions is difficulty with attention and concentration. For example, clinical interviews with adolescent girls have noted impaired concentration and/or attention in around 9 percent of well adolescents in the community but 70 percent of those with current major depression (Cooper and Goodyer, 1993). Findings from studies on mood-related ruminative styles and behavioralinhibition difficulties both suggest vulnerabilities in one or more aspects of the attentional system. Attention itself is not, however, considered a unitary concept and may consist of an interrelated set of components comprising searching, selection, sustaining, and switching strategies. The precise deficit, if present, within depressed individuals remains unclear. Perseverative problems with mood-valent ruminating on negative thoughts suggest problems in switching attention. Mood-valent attentional biases noted on the behavioralinhibition task suggest that if switching is not the problem, then there may be vulnerability in sustained attention that is activated in dysphoric circumstances. This vulnerability, in turn, evokes the bias. Studies using the Wisconsin Card Sort Task have shown that depressed adults have some difficulties in switching attention (Merriam et al., 1999; Grant, Thase, and Sweeney, 2001). Studies using a computerized analogue of this task, the Intra-Dimensional, Extra-Dimensional Set Shifting task (ID-ED) of the Cambridge Neuropsychological Test Automated Battery (CANTAB), have found somewhat equivocal findings, with attentional switching being deficient in some studies of depressed individuals (Beats, Sahakian, and Levy, 1996; Purcell et al., 1997) but not others (Elliot et al., 1996; Sweeney, Kmiec, and Kupfer, 2000b). Recent findings on recovered depressed adults noted a sustained attention deficit not explained by residual depressive symptoms (WeilandFiedler et al., 2004), suggesting that this may be either a vulnerability marker or a consequence of disorder. There is only one published study using the ID-ED in recently depressed adolescents, which showed no difference compared to healthy controls, suggesting no clear-cut vulnerability present in younger individuals (Kyte, Goodyer, and Sahakian, 2005). As noted earlier, recently depressed adolescents and adults have difficulty in switching attentional set away from a sad focus. The assumption could be made that attentional biases are only relevant in mood-valent contexts. To date, however, there has been little detailed investigation of attentional processing in depressed children and adolescents or in those

young people at risk for mood disorders. The Test of Everyday Attention for Children (TEA-Ch; Manly et al., 2001) was developed and validated to provide a comprehensive test of three components of attentional processing, namely, attentional control/switching (the ability to switch from one way of working to another), selective attention (the ability to filter information to detect relevant and ignore irrelevant stimuli), and sustained attention (the ability to maintain performance in a task that is inherently uninteresting and unrewarding). This noncomputerized task, presented as a board game, may provide a more comprehensive evaluation of attention than has been available to date. The tests are shorter and simpler than computerized tasks (thus minimizing the need to be able to sustain attention); therefore, the attentional switching tasks may be purer measures of shifting attentional set. Utilizing this measure in currently depressed adolescents confirms that depressed adolescents are significantly slower than controls at switching attentional resources from one task to another independent of mood (Wilkinson and Goodyer, 2006). The findings were consistent with depressed adolescents finding the task more difficult rather than simply being more careful on the task. The increased latency in attentional switching remained when processing speed was controlled for and, together with the absence of a deficit in educational ability, suggests that the findings are unlikely to be due to reduced motivation. Differences were greatest in those not taking antidepressants, suggesting this was not a treatment effect. Impairments in selective and sustained attention were also noted, although it is possible that the antidepressants, possibly because of the side effect of tiredness, led to the difficulties with sustained attention. There was no evidence for an association between the degree of rumination and impairments in attentional switching. Thus ruminating on dysphoric-related negative cognition does not appear to be arising from any of the measured components of attentional processing in this sample. Deficits in attentional switching and increases in mood-related ruminations in adolescent depression may, therefore, be independent processes. Since the TEA-Ch attentional switching tasks are emotionally neutral, they suggest that adolescents who develop mood disorders may have a basic deficit in attentional switching that renders them vulnerable to attentional biases during dysphoric-related circumstances. The neuropsychological components of mood-valent rumination remain underspecified. It is possible that such perseverative thinking reflects deficits in other aspects of executive functions such as decision making or planning. Alternatively, one can consider ruminative thinking as a normative process that is activated when depressed as initially an adaptive process. Reflecting and being preoccupied by problems may lead to solution-focused thinking. This appears not to be the case in some depressed individuals whose ruminating is little more

goodyer and kyte: unipolar depressions

779

than brooding and perseverating on depressogenic thoughts. It should be emphasized that this is not a deficit of attention. Rather, the recruitment of a normative and adaptive thinking style has become maladaptive for unknown reasons. This result may indicate a deficit or block in mnemonic retrieval processes, with the perseverative thinking indexing a failure in recall of alternative thoughts, either mood-valent, happy, distracting ones or enabling solution-focused ones. There is some recent experimental evidence in support of this notion. Dysphoric-related ruminative thinking diminishes the precision of autobiographical recall independently of level of depression (Park, Goodyer, and Teasdale, 2004). This recall interference may increase metabolic effort at both the physiological and psychological levels but inhibit emotional processing and block alternate forms of more hedonic or solution-focused thoughts.

Cognitive resilience There has been considerable interest for decades in the notion of resilience in the face of adversity (Connor and Zhang, 2006). Why is it that individuals exposed to the same or similar undesirable life events can show markedly different responses, with some showing no negative effects? Many different studies with varied designs, populations, and measures have demonstrated that key resilience processes are most likely to be found within the person (Yehuda et al., 2006). As yet, however, little is really known about the resilience mechanisms that result from activating neurocognitive processes during exposure to social adversities (Goodyer, 2002). For example, what adaptive cognitive process is required following dysphoric mood change to prevent negative cognitions and depressive outcomes? Perhaps this is the function of executive processes, whose role is to construct, execute, and maintain optimal response patterns. Thus resilience may be dependent on the processes of attention, behavioral inhibition, and decision making working in a cooperative manner to either effect a reduction in negative outcomes or to enhance positive and alternative strategies. Cognitive resilience can be considered as occurring cooperatively at three levels of processing: (1) Selective and sustained attention + Behavioral inhibition = Construction phase; (2) Choice + Action = Execution phase; and (3) Outcome = Maintenance phase (Kyte and Goodyer, 2005). The cooperative implication is that disruption in one or more of these individual phases may result in interference in optimal behavioral responses and diminish resilience and therefore adaptive outcomes. The absence of sufficient resilience therefore increases the likelihood of psychopathology. Impaired resilience may be characterized at any one or more of the following levels: behavioral rigidity, distractibility, inability to maintain goal-directed behavior, prevention of selecting the most appropriate or alternative action plans,

780

impaired shifting of attention and response, suboptimal decision making, and a diminished opportunity for adaptive coping. Thus cognitive resilience is likely to arise from an aggregation of complex processing and will require multiple testing of each component to determine its efficiency and effectiveness. As yet this has not been undertaken in experimental studies of young people evaluating the neuropsychology of mood disorders. Specific psychological processes that are involved in the processing of and responding to adversity and in turn perhaps activating cognitive resilience processes have also been associated with discrete brain regions within a limbic-cortical circuitry. For example, recognizing the emotional tone of a social experience has been associated with activation primarily within the amygdala (Calder, Lawrence, and Young, 2001; Thomas et al., 2001). Processing affect-related meanings of life events appears to be mediated by the medial PFC functioning as the executive component for limbic-cortical activity (Teasdale et al., 1999). The neural basis for evaluating, organizing, and consolidating the meaning of environmental stimuli in declarative memory has been shown to be a function of the hippocampus (Eichenbaum, 1999). And finally, establishing the degree of difficulty of inductive inference from external stimuli appears preferentially associated with activation within the orbitofrontal cortex (Goel and Dolan, 2000).

Conclusions Over the next decade we can expect continuing refinement of the affective-cognitive tasks for use in functional neuroimaging studies to determine the neural correlates that underpin mental processes in well individuals, adults, and those at high risk for and those suffering from unipolar depression (Elliott et al., 2002). The findings to date indicate that there is no single cognitive processing component or related brain area that is the key to psychological vulnerability for unipolar depression. Rather, there is a coherence of neurocognitive systems involved in the recognition, processing, and response formation to emotionally meaningful stimuli from the environment that is required to prevent inefficient mental functions. Exactly how these neural systems operate to process affectively valent information is slowly emerging, but we remain relatively ignorant of the precise cooperative psychological processes that, if not functioning, will result in depressive disorders. It seems highly likely that fully integrated social and performance functions are needed for mental competence and resilience in the face of adversity. The breakdown in one or more components involved in processing environmental stimuli may result in unipolar depression. The characteristics of depression, the severity and duration of the episode in particular, may result in further or new cognitive deficits. Thus

neurodevelopmental aspects of clinical disorders

the neuropsychology of individuals with a single episode of illness may not be the same as that in individuals who experience recurrent episodes. Whether there are developmental differences that result in neurocognitive profiles being distinct in younger versus older individuals, or in those at high versus low risk for psychopathology, remains unclear. Unipolar depression is one of the most serious mental disorders to emerge in the second decade of life. Delineating the neurocognitive phenotypes of this heterogeneous condition is an important and achievable goal for the early part of the 21st century. acknowledgments

This chapter was completed within the Adolescents Mood Disorders Program funded by the Wellcome Trust and sponsored by the Cambridge and Peterborough Mental Health Partnership Trust.

REFERENCES Abramson, L., M. Seligman, and J. Teasdale, 1978. Learned helplesness in humans: Critique and reformulation. J. Abnorm. Psychol. 87:49–74. Alloy, L. B., L. Y. Abramson, M. E. Hogan, W. G. Whitehouse, D. T. Rose, M. S. Robinson, et al., 2000. The TempleWisconsin Cognitive Vulnerability to Depression Project: Lifetime history of axis I psychopathology in individuals at high and low cognitive risk for depression. J. Abnorm. Psychol. 109(3): 403–418. Alloy, L. B., L. Y. Abramson, W. G. Whitehouse, M. E. Hogan, N. A. Tashman, D. L. Steinberg, et al., 1999. Depressogenic cognitive styles: Predictive validity, information processing and personality characteristics, and developmental origins. Behav. Res. Ther. 37(6):503–531. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental and Behavioral Disorders, 4th ed. Washington, DC: American Psychiatric Association. Angold, A., and E. J. Costello, 2001. The epidemiology of depression in children and adolescents. In I. Goodyer, ed., The Depressed Child and Adolescent, 2nd ed., 143–178. Cambridge, UK: Cambridge University Press. Barbaccia, M. L., M. Serra, R. H. Purdy, and G. Biggio, 2001. Stress and neuroactive steroids. Int. Rev. Neurobiol. 46:243–272. Barnea-Goraly, N., V. Menon, M. Eckert, L. Tamm, R. Bammer, A. Karchemskiy, et al., 2005. White matter development during childhood and adolescence: A cross-sectional diffusion tensor imaging study. Cerebral Cortex 15(12):1848–1854. Beats, B. C., B. J. Sahakian, and R. Levy, 1996. Cognitive performance in tests sensitive to frontal lobe dysfunction in the elderly depressed. Psychol. Med. 26(3):591–603. Bechara, A., D. Tranel, and H. Damasio, 2000. Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions. Brain 123(Part 11):2189–2202. Beck, A., 1967. Depression: Clinical Experiments with a Theoretical Aspect. New York: Harper and Row. Blakemore, S. J., and S. Choudhury, 2006a. Brain development during puberty: State of the science. Dev. Sci. 9(1):11–14. Blakemore, S. J., and S. Choudhury, 2006b. Development of the adolescent brain: Implications for executive function and social cognition. J. Child Psychol. Psychiatry 47(3–4):296–312.

Brotman, M. A., M. Schmajuk, B. A. Rich, D. P. Dickstein, A. E. Guyer, E. J. Costello, et al., 2006. Prevalence, clinical correlates, and longitudinal course of severe mood dysregulation in children. Biol. Psychiatry 60(9):991–997. Brown, M., R. Keynes, and A. Lumsden, eds. 2001. The Developing Brain. Oxford, UK: Oxford University Press. Buchanan, C., J. Eccles, and J. Becker, 1992. Are adolescents victims of raging hormones: Evidence for activational effects of hormones on moods and behavior at adolescence. Psychol. Bull. 111(1):62–107. Calder, A., A. D. Lawrence, and A. W. Young, 2001. Neuropsychology of fear and loathing. Nature 2:351–363. Casey, B. J., 1999. Images in neuroscience: Brain development. XII. Maturation in brain activation. Am. J. Psychiatry 156(4): 504. Connor, K. M., and W. Zhang, 2006. Resilience: Determinants, measurement and treatment responsiveness. CNS Spectrums 11: 5–12. Cooper, P. J., and I. M. Goodyer, 1993. A community study of depression in adolescent girls. 1. Estimates of symptom and syndrome prevalence. Br. J. Psychiatry 163:369–374. Costello, E. J., A. Angold, B. J. Burns, A. Erkanli, D. K. Stangl, and D. L. Tweed, 1996. The Great Smoky Mountains Study of Youth: Functional impairment and serious emotional disturbance. Arch. Gen. Psychiatry 53(12):1137–1143. Crone, E. A., S. A. Bunge, H. Latenstein, and M. W. van der Molen, 2005. Characterization of children’s decision making: Sensitivity to punishment frequency, not task complexity. Child Neuropsychol. 11(3):245–263. Crone, E. A., and M. W. van der Molen, 2004. Developmental changes in real life decision making: Performance on a gambling task previously shown to depend on the ventromedial prefrontal cortex. Dev. Neuropsychol. 25(3):251–279 Dias, R., T. W. Robbins, and A. C. Roberts, 1997. Dissociable forms of inhibitory control within prefrontal cortex with an analog of the Wisconsin Card Sort Test: Restriction to novel situations and independence from “on-line” processing. J. Neurosci. 17(23):9285–9297. Drevets, W. C., 2000. Functional anatomical abnormalities in limbic and prefrontal cortical structures in major depression. Prog. Brain Res. 126:413–431. Dunn, V., and I. M. Goodyer, 2006. Longitudinal investigation into childhood- and adolescence-onset depression: Psychiatric outcome in early adulthood. Br. J. Psychiatry 188:216–222. Durston, S., H. Hulshoff, B. J. Casey, J. N. Giedd, J. Buitelaar, and H. van Engeland, 2001. Anatomical MRI of the developing human brain: What have we learned? J. Am. Acad. Child Adolesc. Psychiatry 40(9):10112–11020. Eichenbaum, H., 1999. The hippocampus and mechanisms of declarative memory. Behav. Brain Res. 103(2):123–133. Elliot, R., B. J. Sahakian, J. J. McKay, J. J. Herrod, T. W. Robbins, and E. S., P. 1996. Neuropsychological impairments in unipolar depression: The influence of perceived failure on subsequent performance. Psychol. Med. 26:975– 989. Elliott, R., S. C. Baker, R. D. Rogers, D. A. O’Leary, E. S. Paykel, C. D. Frith, et al., 1997. Prefrontal dysfunction in depressed patients performing a complex planning task: A study using positron emission tomography. Psychol. Med. 27(4):931– 942. Elliott, R., J. S. Rubinsztein, B. J. Sahakian, and R. J. Dolan, 2002. The neural basis of mood-congruent processing biases in depression. Arch. Gen. Psychiatry 59(7):597–604.

goodyer and kyte: unipolar depressions

781

Fergusson, D. M., L. J. Horwood, E. M. Ridder, and A. L. Beautrais, 2005. Subthreshold depression in adolescence and mental health outcomes in adulthood. Arch. Gen. Psychiatry 62(1):66–72. Fombonne, E., G. Wostear, V. Cooper, R. Harrington, and M. Rutter, 2001a. The Maudsley long-term follow-up of child and adolescent depression. 1. Psychiatric outcomes in adulthood. Br. J. Psychiatry 179:210–217. Fombonne, E., G. Wostear, V. Cooper, R. Harrington, and M. Rutter, 2001b. The Maudsley long-term follow-up of child and adolescent depression. 2. Suicidality, criminality and social dysfunction in adulthood. Br. J. Psychiatry 179:218–223. Forbes, E. E., J. Christopher May, G. J. Siegle, C. D. Ladouceur, N. D. Ryan, C. S. Carter, et al., 2006. Rewardrelated decision-making in pediatric major depressive disorder: An fMRI study. J. Child Psychol. Psychiatry 47(10):1031–1040. Forbes, E. E., and R. E. Dahl, 2005. Neural systems of positive affect: Relevance to understanding child and adolescent depression? Dev. Psychopathol. 17(3):827–850. Gibb, B. E., and L. B. Alloy, 2006. A prospective test of the hopelessness theory of depression in children. J. Clin. Child Adolesc. Psychol. 35(2):264–274. Giedd, J. N., J. Blumenthal, N. O. Jeffries, J. C. Rajapakse, A. C. Vaituzis, H. Liu, et al., 1999. Development of the human corpus callosum during childhood and adolescence: A longitudinal MRI study. Prog. Neuropsychopharmacol. Biol. Psychiatry 23(4): 571–588. Giedd, J. N., F. X. Castellanos, J. C. Rajapakse, A. C. Vaituzis, and J. L. Rapoport, 1997. Sexual dimorphism of the developing human brain. Prog. Neuropsychopharmacol. Biol. Psychiatry 21(8): 1185–1201. Giedd, J. N., A. C. Vaituzis, S. D. Hamburger, N. Lange, J. C. Rajapakse, D. Kaysen, et al., 1996. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: Ages 4–18 years. J. Comp. Neurol. 366(2):223– 230. Goel, V., and R. J. Dolan, 2000. Anatomical segregation of component processes in an inductive inference task. J. Cogn. Neurosci. 12(1):110–119. Goodyer, I. M., ed., 2001a. The Depressed Child and Adolescent, 2nd ed. Cambridge, UK: Cambridge University Press. Goodyer, I. M., 2001b. Life events: Their nature and effects. In I. M. Goodyer, ed., The Depressed Child and Adolescent, 2nd ed., pp. 204–233. Cambridge, UK: Cambridge University Press. Goodyer, I. M., 2002. Social adversity and mental functions in adolescents at high risk of psychopathology: Position paper and suggested framework for future research. Br. J. Psychiatry 181: 383–386. Goodyer, I. M., and P. J. Cooper, 1993. A community study of depression in adolescent girls. 2. The clinical features of identified disorder. Br. J. Psychiatry 163:374–380. Goodyer, I. M., J. Herbert, and A. Tamplin, 2003. Psychoendocrine antecedents of persistent first-episode major depression in adolescents: A community-based longitudinal enquiry. Psychol. Med. 33(4):601–610. Goodyer, I. M., J. Herbert, A. Tamplin, and P. M. Altham, 2000a. First-episode major depression in adolescents: Affective, cognitive and endocrine characteristics of risk status and predictors of onset. Br. J. Psychiatry 176:142–149. Goodyer, I. M., J. Herbert, A. Tamplin, and P. M. Altham, 2000b. Recent life events, cortisol, dehydroepiandrosterone and the onset of major depression in high-risk adolescents. Br. J. Psychiatry 177:499–504.

782

Grant, M. M., M. E. Thase, and J. A. Sweeney, 2001. Cognitive disturbance in outpatient depressed younger adults: Evidence of modest impairment. Biol. Psychiatry 50(1):35–43. Hooper, C. J., M. Luciana, H. M. Conklin, and R. S. Yarger, 2004. Adolescents’ performance on the Iowa Gambling Task: Implications for the development of decision making and ventromedial prefrontal cortex. Dev. Psychol. 40(6):1148–1158. Huang, E. J., and L. F. Reichardt, 2001. Neurotrophins: Roles in neuronal development and function. Annu. Rev. Neurosci. 24: 677–736. Hughes, C., A. White, J. Sharpen, and J. Dunn, 2000. Antisocial, angry, and unsympathetic: “Hard-to-manage” preschoolers’ peer problems and possible cognitive influences. J. Child Psychol. Psychiatry 41(2):169–179. Huttenlocher, P. R., 1979. Synaptic density in human frontal cortex—Developmental changes and effects of aging. Brain Res. 163(2):195–205. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387(2):167–178. Ingram, R., J. Miranda, and Z. Segal, 1998. Cognitive Vulnerability to Depression. New York: Guilford Press. Jack, A. I., C. M. Sylvester, and M. Corbetta, 2006. Losing our brainless minds: How neuroimaging informs cognition. Cortex 42(3):418–421. Jernigan, T. L., D. A. Trauner, J. R. Hesselink, and P. A. Tallal, 1991. Maturation of human cerebrum observed in vivo during adolescence. Brain 114(Part 5):2037–2049. Just, N., L. Y. Abramson, and L. B. Alloy, 2001. Remitted depression studies as tests of the cognitive vulnerability hypotheses of depression onset: A critique and conceptual analysis. Clin. Psychol. Rev. 21(1):63–83. Just, N., and L. Alloy, 1997. The response styles theory of depression: Tests and extension of the theory. J. Abnorm. Psychol. 2:221–229. Kaufman, J., A. Martin, R. A. King, and D. Charney, 2001. Are child-, adolescent-, and adult-onset depression one and the same disorder? Biol. Psychiatry 49(12):980–1001. Kelvin, R. G., I. M. Goodyer, J. D. Teasdale, and D. Brechin, 1999. Latent negative self-schema and high emotionality in well adolescents at risk for psychopathology. J. Child Psychol. Psychiatry 40(6):959–968. Kim-Cohen, J., A. Caspi, T. E. Moffitt, H. Harrington, B. J. Milne, and R. Poulton, 2003. Prior juvenile diagnoses in adults with mental disorder: Developmental follow-back of a prospective-longitudinal cohort. Arch. Gen. Psychiatry 60(7):709–717. Kovacs, M., 1996. Presentation and course of major depressive disorder during childhood and later years of the life span. J. Am. Acad. Child Adolesc. Psychiatry 35(6):705–715. Kuan, C. Y., K. A. Roth, R. A. Flavell, and P. Rakic, 2000. Mechanisms of programmed cell death in the developing brain. Trends Neurosci. 23(7):291–297. Kyte, Z. A., and I. M. Goodyer, 2005. Neurobiology of social cognition and its relationship to unipolar depression. In A. Easton and N. Emery, eds., The Cognitive Neuroscience of Social Behaviour, 257–299. London: Psychology Press. Kyte, Z. A., I. M. Goodyer, and B. J. Sahakian, 2005. Selected executive skills in adolescents with recent first episode major depression. J. Child Psychol. Psychiatry 46(9):995–1005. Lewinsohn, P. M., G. N. Clarke, P. Rohde, and J. R. Seeley, 1994. Major depression in community adolescents: Age at onset, episode duration, and time to recurrence. J. Am. Acad. Child Adolesc. Psychiatry 33:809–818.

neurodevelopmental aspects of clinical disorders

Lewinsohn, P. M., T. E. Joiner, Jr., and P. Rohde, 2001. Evaluation of cognitive diathesis-stress models in predicting major depressive disorder in adolescents. J. Abnorm. Psychol. 110(2): 203–215. Lewinsohn, P. M., P. Rohde, D. N. Klein, and J. R. Seeley, 1999. Natural course of adolescent major depressive disorder. I. Continuity into young adulthood. J. Am. Acad. Child Adolesc. Psychiatry 38(1):56–63. Lewinsohn, P. M., P. Rohde, and J. R. Seeley, 1998. Major depressive disorder in older adolescents: Prevalence, risk factors, and clinical implications. Clin. Psychol. Rev. 18(7):765–794. Liotti, M., and H. S. Mayberg, 2001. The role of functional neuroimaging in the neuropsychology of depression. J. Clin. Exp. Psychol. 23(1):121–136. Lockwood, K. A., G. S. Alexopoulos, T. Kakuma, and W. G. Van Gorp, 2000. Subtypes of cognitive impairment in depressed older adults. Am. J. Geriatr. Psychiatry 8(3):201–208. Luby, J. L., A. K. Heffelfinger, C. Mrakotsky, K. M. Brown, M. J. Hessler, J. M. Wallis, et al., 2003. The clinical picture of depression in preschool children. J. Am. Acad. Child Adolesc. Psychiatry 42(3):340–348. Luciana, M., 2003a. Cognitive development in children born preterm: Implications for theories of brain plasticity following early injury. Dev. Psychopathol. 15(4):1017–1047. Luciana, M., 2003b. Practitioner review: Computerized assessment of neuropsychological function in children: Clinical and research applications of the Cambridge Neuropsychological Testing Automated Battery (CANTAB). J. Child Psychol. Psychiatry 44(5):649–663. Luciana, M., H. M. Conklin, C. J. Hooper, and R. S. Yarger, 2005. The development of nonverbal working memory and executive control processes in adolescents. Child Dev. 76(3): 697–712. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36(3):273–293. Luciana, M., and C. A. Nelson, 2002. Assessment of neuropsychological function through use of the Cambridge Neuropsychological Testing Automated Battery: Performance in 4- to 12-year-old children. Dev. Neuropsychol. 22(3):595–624. Lyubomirsky, S., and N. Nolen-Hoeksma, 1993. Self-perpetuating properties of dysphoric rumination. J. Pers. Soc. Psychol. 65:339– 349. Manly, T., V. Anderson, I. Nimmo-Smith, A. Turner, P. Watson, and I. H. Robertson, 2001. The differential assessment of children’s attention: The Test of Everyday Attention for Children (TEA-Ch), normative sample and ADHD performance. J. Child Psychol. Psychiatry 42(8):1065–1081. Merriam, E. P., M. E. Thase, G. L. Haas, M. S. Keshavan, and J. A. Sweeney, 1999. Prefrontal cortical dysfunction in depression determined by Wisconsin Card Sorting Test performance. Am. J. Psychiatry 156(5):780–782. Murphy, F. C., J. S. Rubinsztein, A. Michael, R. D. Rogers, T. W. Robbins, E. S. Paykel, et al., 2001. Decision-making cognition in mania and depression. Psychol. Med. 31(4):679–693. Murphy, F. C., B. J. Sahakian, J. S. Rubinsztein, A. Michael, R. D. Rogers, T. W. Robbins, et al., 1999. Emotional bias and inhibitory control processes in mania and depression. Psychol. Med. 29(6):1307–1321. Murray, L., S. L. Halligan, G. Adams, P. Patterson, and I. M. Goodyer, 2006. Socioemotional development in adolescents at risk for depression: The role of maternal depression and attachment style. Dev. Psychopathol. 18(2):489–516.

Murray, L., D. Sinclair, P. Cooper, P. Ducournau, P. Turner, and A. Stein, 1999. The socioemotional development of 5-yearold children of postnatally depressed mothers. J. Child Psychol. Psychiatry 40(8):1259–1271. Nolen-Hoeksema, S., 1991. Responses to depression and their effects on the duration of depressive episodes. J. Abnorm. Psychol. 100:569–582. Nolen-Hoeksema, S., 2000. The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. J. Abnorm. Psychol. 109(3):504–511. Nolen-Hoeksma, S., 2004. The response styles theory. In C. Papageorgiou and A. Wells, eds., Depressive Ruminations, 107– 124. Chichester, UK: Wiley. Paelecke-Habermann, Y., J. Pohl, and B. Leplow, 2005. Attention and executive functions in remitted major depression patients. J. Affect. Disord. 89(1–3):125–135. Pardo, J. V., P. J. Pardo, S. W. Humes, and I. P. M. Posner, 2006. Neurocognitive dysfunction in antidepressant-free, non-elderly patients with unipolar depression: Alerting and covert orienting of visuospatial attention. J. Affect. Disord. 92(1):71–78. Park, R. J., I. M. Goodyer, and J. D. Teasdale, 2004. Effects of induced rumination and distraction on mood and overgeneral autobiographical memory in adolescent major depressive disorder and controls. J. Child Psychol. Psychiatry 45(5):996– 1006. Paus, T., 2005. Mapping brain maturation and cognitive development during adolescence. Trends Cogn. Sci. 9(2):60–68. Paus, T., D. L. Collins, A. C. Evans, G. Leonard, B. Pike, and A. Zijdenbos, 2001. Maturation of white matter in the human brain: A review of magnetic resonance studies. Brain Res. Bull. 54(3):255–266. Pennington, B. F., and S. Ozonoff, 1996. Executive functions and developmental psychopathoogy. J. Child Psychol. Psychiatry 37(1): 51–89. Perez-Edgar, K., N. A. Fox, J. F. Cohn, and M. Kovacs, 2006. Behavioral and electrophysiological markers of selective attention in children of parents with a history of depression. Biol. Psychiatry 60(10):1131–1138. Pine, D. S., E. Cohen, P. Cohen, and J. Brook, 1999. Adolescent depressive symptoms as predictors of adult depression: Moodiness or mood disorder? Am. J. Psychiatry 156(1):133–135. Plaisted, K. C., and B. Sahakian, 1997. Dementia of the frontal lobe type—Living in the here and now. Aging Ment. Health 1: 293–295. Purcell, R., P. Maruff, M. Kyrios, and C. Pantelis, 1997. Neuropsychological function in young patients with unipolar major depression. Psychol. Med. 27(6):1277–1285. Rogers, R. D., B. J. Everitt, A. Baldacchino, A. J. Blackshaw, R. Swainson, K. Wynne, et al., 1999. Dissociable deficits in the decision-making cognition of chronic amphetamine abusers, opiate abusers, patients with focal damage to prefrontal cortex, and tryptophan-depleted normal volunteers: Evidence for monoaminergic mechanisms. Neuropsychopharmacology 20(4):322–339. Rogers, R. D., A. M. Owen, H. C. Middleton, E. J. Williams, J. D. Pickard, B. J. Sahakian, et al., 1999. Choosing between small, likely rewards and large, unlikely rewards activates inferior and orbital prefrontal cortex. J. Neurosci. 19(20): 9029–9038. Ryan, N. D., J. Puig-Antich, and P. Ambrosini, 1987. The clinical picture of major depression in childhood and adolescence. Arch. Gen. Psychiatry 45:486–495. Schatzberg, A. F., J. A. Posener, C. DeBattista, B. M. Kalehzan, A. J. Rothschild, and P. K. Shear, 2000.

goodyer and kyte: unipolar depressions

783

Neuropsychological deficits in psychotic versus nonpsychotic major depression and no mental illness. Am. J. Psychiatry 157(7): 1095–1100. Segal, Z. V., and R. Ingram, 1994. Mood priming and construct activation in tests of cognitive vulnerability to unipolar depression. Clin. Psychol. Rev. 14(7):663–695. Seligman, M., C. Peterson, N. Kaslow, R. Tanenbaum, L. Alloy, and L. Abramson, 1984. Attributional style and depressive symptoms among children. J. Abnorm. Psychol. 93:235–238. Seminowicz, D. A., H. S. Mayberg, A. R. McIntosh, K. Goldapple, S. Kennedy, Z. Segal, et al., 2004. Limbicfrontal circuitry in major depression: A path modeling metanalysis. NeuroImage 22(1):409–418. Shaw, P., D. Greenstein, J. Lerch, L. Clasen, R. Lenroot, N. Gogtay, et al., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440(7084): 676–679. Sowell, E. R., P. M. Thompson, C. J. Holmes, T. L. Jernigan, and A. W. Toga, 2000. In vivo evidence for postadolescent brain maturation in frontal and striatal regions. Nature Neurosci. 2(10): 859–861. Spasojevic, J., and L. B. Alloy, 2001. Rumination as a common mechanism relating depressive risk factors to depression. Emotion 1(1):25–37. Sweeney, J. A., J. A. Kmiec, and D. J. Kupfer, 2000a. Neuropsychologic impairments in bipolar and unipolar mood disorders on the CANTAB neurocognitive battery. Biol. Psychiatry 48(7): 674–684. Sweeney, J. A., J. A. Kmiec, and D. J. Kupfer, 2000b. Neuropsychologic impairments in bipolar and unipolar mood disorders on the CANTAB neurocognitive battery. Biol. Psychiatry 48(7): 674–684. Tamminga, C., 2000. Images in neuroscience: Executive function. Am. J. Psychiatry 157:3. Tavares, J. V., W. C. Drevets, and B. J. Sahakian, 2003. Cognition in mania and depression. Psychol. Med. 33(6):959–967. Teasdale, J. D., and P. J. Barnard, 1993. Affect, Cognition and Change: Remodelling Depressive Thought. Hillsdale, NJ: Lawrence Erlbaum. Teasdale, J., and J. Dent, 1987. Cognitive vulnerability to depression: An investigation of two hypotheses. Br. J. Clin. Psychol. 26: 113–126.

784

Teasdale, J. D., R. J. Howard, S. G. Cox, Y. Ha, M. J. Brammer, S. C. Williams, et al., 1999. Functional MRI study of the cognitive generation of affect. Am. J. Psychiatry 156(2):209–215. Teasdale, J. D., C. A. Lloyd, and J. M. Hutton, 1998. Depressive thinking and dysfunctional schematic mental models. Br. J. Clin. Psychol. 37(Part 3):247–257. Teasdale, J. D., M. J. Taylor, Z. Cooper, H. Hayhurst, and E. S. Paykel, 1995. Depressive thinking: Shifts in construct accessibility or in schematic mental models? J. Abnorm. Psychol. 104(3):500–507. Thomas, K., W. C. Drevetts, R. E. Dahl, N. D. Ryan, B. Birmaher, H. C. Eccard, et al., 2001. Amygdala response to fearful faces in anxious and depressed children. Arch. Gen. Psychiatry 58:1057–1063. Thompson, P. M., J. N. Giedd, R. P. Woods, D. MacDonald, A. C. Evans, and A. W. Toga, 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 404(6774):190–193. Tranel, D., A. Bechara, and A. R. Damasio, 2000. Decision making and the somatic marker hypothesis. In M. S. Gazzaniga, ed., The Cognitive Neurosciences, 2nd ed., pp. 1046–1106. Cambridge, MA: MIT Press. Tremblay, L. K., C. A. Naranjo, S. J. Graham, N. Herrmann, H. S. Mayberg, S. Hevenor, et al., 2005. Functional neuroanatomical substrates of altered reward processing in major depressive disorder revealed by a dopaminergic probe. Arch. Gen. Psychiatry 62(11):1228–1236. Weiland-Fiedler, P., K. Erickson, T. Waldeck, D. A. Luckenbaugh, D. Pike, O. Bonne, et al., 2004. Evidence for continuing neuropsychological impairments in depression. J. Affect. Disord. 82(2):253–258. Whitaker-Azmitia, P. M., 2001. Serotonin and brain development: Role in human developmental diseases. Brain Res. Bull. 56(5):479–485. Wilkinson, P. O., and I. M. Goodyer, 2006. Attention difficulties and mood-related ruminative response style in adolescents with unipolar depression. J. Child Psychol. Psychiatry 47(12)1284– 1291. Yehuda, R., J. D. Flory, S. Southwick, and D. S. Charney, 2006. Developing an agenda for translational studies of resilience and vulnerability following trauma exposure. Ann. NY Acad. Sci. 1071:379–396.

neurodevelopmental aspects of clinical disorders

VI EMOTION/ COGNITION INTERACTIONS

48

Toward a Neurobiology of Attachment MYRON A. HOFER AND REGINA M. SULLIVAN

The word attachment has come to refer to a broad range of behavioral processes and mental states unified by a single central concept. Attachment theory envisages a unique motivational system with evolutionary survival value on a par with hunger and sex that is organized to maintain physical proximity to the mother/caretaker soon after birth and is responsible for psychological proximity, or sense of closeness, later in development, the “bond” (Bowlby, 1969). Rupture of this “bond” by separation was the concept used to explain the traumatic effects of early loss on behavior and physiology of the child. The special qualities of the bond reflect the nature of the infant’s and the caretaker’s mental representations of the behavioral dynamics and patterns of their many previous interactions, referred to as an “internal waking model” of the relationship. Most of the research on attachment subsequent to Bowlby’s landmark volume defining the field (Bowlby, 1969) has focused on the different patterns of early attachment and their later developmental correlates, including the transmission of patterns of attachment to the next generation. This focus left the earlier observations far less studied: the developmental processes through which attachment is initially formed in altricial (slow-developing) mammals and the behavioral mechanisms underlying the intense behavioral and physiological responses of infants to separation. Meanwhile others interested in processes underlying the early development of motor and sensory systems, perception, attention, learning, memory, communication, motivation, and emotion, within the field of developmental psychobiology, developed new methods (Shair, Barr, and Hofer, 1991) and a knowledge base in each of these areas (Michel and Moore, 1995) without relation to the concept of attachment or its field of research. To these researchers, the behaviors included within the concept of attachment did not appear to be organized as a unitary system, and attachment concepts were too global to generate incisive research questions that could lead to a deeper understanding of the phenomena described. Now that a knowledge base has been established in the various research areas of developmental psychobiology, we can revisit the concept of basic attachment theory and begin to describe the underlying processes responsible for them in terms that can be related to neural structure and function.

When one does these things, as will be described in the body of this chapter, it becomes apparent that the unique features of early attachment phenomena can be shown to be the result of certain unique features of early sensory and motor integration—of early learning, communication, motivation, and the regulation of developing biological and behavioral systems by the mother-infant interaction. The apparently unified nature of attachment and its function as a system are the result of the fact that for the mammalian fetus and infant, its environment virtually begins and ends within the confines of a unified source of interaction, the mother. Thus “attachment,” like “hunger,” is made up of a number of underlying component processes and exists as a useful concept because it describes the output of these subprocesses as they work together within the larger-scale arena of social relationships. The observations that attachment theory was formulated to explain will be used as a framework for this chapter. In each case we will describe recent evidence for the basic behavioral and neurobiological processes that underlie the concepts of attachment theory. First, how does the infant come to know and prefer its own mother, maintain proximity with her, and continue to do so, even despite abuse and neglect at her hands? Second, why does separation from the mother produce such intense and widespread emotional responses in the infant? Third, how do individual differences arise in the characteristic patterns or qualities of the motherinfant interaction, and how do these early interactions become translated into long-term effects on development and ultimately into the transmission of similar maternal behavior to the next generation? In this chapter we will focus on those areas where we have the most detailed knowledge of the underlying behavioral, cognitive, and neurobiological processes involved, using laboratory animal models. It has been a surprise to find how many of the basic processes of attachment can be studied in relatively simple mammalian species, and the results interpreted in relation to similar observations in humans.

Initial formation of a specific attachment Infants of mammalian species that are born in an immature state, such as the human and the rat, face a daunting

787

cognitive task. They must eventually learn to identify, remember, and prefer their own mother, and they must learn to use these new cognitive capacities to reorganize their simple motor repertoires, long adapted to the uterine environment, so as to be able to approach, remain close, and orient themselves to their mothers for the first nursing bout. It has been assumed until recently that these processes were well beyond the capacities of newborn mammals (except in precocial species such as the sheep) and that the relationship initially depended almost entirely upon maternal behavior until well into the nursing period (Bowlby, 1982; Kraemer, 1992). Attachment has thus been supposed to be built up slowly in the weeks or days after birth in human or rat. But the last decade has produced a number of studies revealing earlier and earlier evidence of learning, even extending into the prenatal period, as described in the next subsection. In addition, coordinated motor acts have been demonstrated in fetuses in response to specific stimuli that will not be encountered until after birth. Thus the solutions for the infant’s cognitive tasks appear to be found much earlier than previously thought and to take place through novel developmental processes that had not been imagined until recently. Prenatal Origins The first strong evidence for fetal learning came from studies on early voice recognition in humans, in which it was found that babies recognize and prefer their own mother’s voice, even when tested within hours after birth (DeCasper and Fifer, 1980). William Fifer continued these studies using a device through which newborns can choose between two tape-recorded voices by sucking at different rates on a pacifier rigged to control an audiotape player (reviewed in Fifer and Moon, 1995). He has found that newborn infants, in the first hours after birth, prefer human voices to silence, female voices to males, their native language to another language, and their own mother to another mother reading the same Dr. Seuss story. In order to obtain more direct evidence for the prenatal origins of these preferences (rather than very rapid postnatal learning), Fifer filtered the high-frequency components from the tapes to make the mother’s voice resemble recordings of maternal voice by hydrophone placed within the amniotic space of pregnant women. This altered recording, in which the words were virtually unrecognizable to adults, was preferred to the standard mothers’ voice by newborns in the first hours after birth, a preference that tended to wane in the second and third postnatal days. Furthermore, there is now evidence that newborns prefer familiar rhythmic phrase sequences to which they have been repeatedly exposed prenatally (DeCasper and Spence, 1986). In a striking interspecies similarity, rat pups were subsequently shown to discriminate and prefer their own dams’ amniotic fluid to that of another dam when offered a choice

788

emotion⁄cognition interactions

in a head-turning task (Hepper, 1987). Newborn pups were also shown to require amniotic fluid on a teat in order to find and attach to it for their first nursing attempt (Blass, 1990). Robinson and Smotherman (1995) have directly tested the hypothesis that pups begin to learn about their mothers’ scent in utero. They have been able to demonstrate one-trial taste-aversion learning and classical conditioning in late-term rat fetuses, using intraoral cannula infusions and perioral stimulation. Taste aversions learned in utero were expressed in the free feeding responses of weanling rats nearly 3 weeks later. The researchers went on to determine that aversive responses to vibrissa stimulation were attenuated or blocked by intraoral milk infusion, a prenatal “comfort” effect they found to be mediated by a central kappa-opioid receptor system. Interestingly, when artificial nipple stimulation was presented as a predictive cue (CS) for intraoral milk in a learning paradigm and a reduction in fetal responsiveness to the CS was acquired in this way, the conditioned response was blocked by a mu-opioid antagonist rather than a kappa antagonist. This finding suggests that the opioid system plays an important role in the organization of fetal behavior by redirecting and focusing responsiveness to a subset of available stimuli (Robinson and Smotherman, 1995). These forms of fetal learning, involving maternal voice in humans and amniotic fluid in rodents, appear to play an adaptive role in preparing the infant for its first extrauterine encounter with its mother. They are thus the earliest origins we have yet found for attachment to the mother. The Perinatal Transition In addition to the evidence for fetal learning that we have described, specific adaptive motor-response capabilities have also been found in laterterm rat fetuses, using an exteriorized in vivo preparation (Smotherman and Robinson, 1992). The reflexes of licking, mouthing, and sucking necessary for postnatal nipple grasp and nursing can be elicited in late-term fetuses, and even the characteristic stretch response to oral milk and the facial wiping (rejection) response to unfamiliar tastes that are usually associated with rat pups tested several days postnatally. The emergence of these reflex responses in anticipation of the postnatal environment in which they will be expressed provides the newborn with the components needed for the transition to postnatal life and for the rapid organization of an integrated response repertoire tuned to specific characteristics of the maternal body. The spontaneous motor acts needed for an attachment system also appear to be developing prior to birth. Rat fetuses engage in a number of spontaneous behaviors in utero, including curls, stretches, and trunk and limb movements. These acts were observed to increase markedly in frequency with progressive removal of intrauterine space constraints, as pups were observed first through the uterine

wall, then through the thin amniotic sac, and finally unrestrained in a warm saline bath (Smotherman and Robinson, 1986). When newborn pups are observed prior to their first nursing bout, they resemble exteriorized fetuses, until the mother lowers her ventrum over them. Their behavior then changes rapidly over the first few nursing bouts, into the complex repertoire described in the next subsection. An Attachment System in the Newborn When pups less than a day old are stimulated gently by soft surfaces from above, as when the mother hovers over them, they show a surprisingly vigorous repertoire of behaviors (Polan and Hofer, 1999). These include the curling and stretching seen prenatally, but now also include locomotor movement toward the suspended surface, directed wriggling, audible vocalizations, and, most strikingly, turning upside down toward the surface above them. Evidently these behaviors propel the pup into close contact with the ventrum, maintain it in proximity, and keep it oriented toward the surface. They thus appear to be very early attachment behaviors. In a series of experiments, we found that these are not stereotyped reflex acts, but organized responses that are graded according to the number of maternal modalities present on the surface (e.g., texture, warmth, odor) (Polan and Hofer, 1999). Furthermore, they are enhanced by periods of prior maternal deprivation, suggesting the rapid development of a motivational component. We found that pups discriminate their own mother’s odor in preference to equally familiar nest odors by two days of age (Polan and Hofer, 1998), and Hepper (1986) has shown that pups discriminate and prefer their own mother, father, and siblings to other lactating females, males, or age-mates by the first postnatal week. These results show that a highly specific “behavioral attachment system” (Bowlby, 1969), capable of approach and proximity maintenance to the mother and motivated by brief periods of separation from her, may occur much earlier in development than previously supposed. The remarkable specificity of the approach response of the infant rat to individual family members acquired within the first few postnatal days demonstrates that specificity of attachment does not require long experience or advanced cognitive and emotional capabilities. Olfaction in the rat and vision in the human provide the necessary basis for approach responses that are specific to a single individual. But this remarkable capability can develop independently of the specificity of the rat pup’s contact comfort response. For even by two weeks of age a rat pup will show an equal comfort response to contact with any female that is available. This nonspecificity is limited, however, and a form of “stranger anxiety” develops by the second week of life, well before weaning. Pups will avoid the odor of unfamiliar adult males (but not of familiar or unfamiliar prepubertal males), and they show

immobility and a brisk adrenocortical response when the stranger is too near (Takahashi, 1992). This early fear response, like approach responses to the mother, depends upon olfactory cues (Shair et al., 1997). Recent work in humans, inspired by these findings in lower animals, has shown that human newborns too are capable of slowly locomoting across the bare surface of the mother’s abdomen and locating the breast scented with amniotic fluid in preference to the untreated breast (Varendi, Porter, and Winberg, 1996). Although newborns are attracted to natural breast odors even before the first nursing bout (Makin and Porter, 1989), amniotic fluid can override this effect. Apparently, human newborns are not as helpless as previously thought and possess approach and orienting behaviors that anticipate the recognized onset of maternal attachment at 6–8 months. Postnatal Learning Although specific olfactory and/or auditory predispositions toward the infant’s own mother are acquired prenatally, after birth the newborn mammal enters a new world where contingent events, so important for more advanced forms of learning, are now occurring with great frequency. As noted earlier, the abrupt transition from prenatal to postnatal life appears to be eased for the newborn by the presence of prenatal stimuli continuing into the postnatal environment (i.e., the odor of amniotic fluid; Lecanuet and Schaal, 1996; Mennella, Johnson, and Beauchamp, 1995; Schaal, Marlier, and Soussignan, 1995; Varendi et al., 1998). However, neonatal capacities for stimulus discrimination and preferential approach, orienting, and proximity-maintenance behavior described in the previous subsection seemed likely to have been formed by some type of rapid postnatal learning process, resembling avian imprinting, as hypothesized by Bowlby. Yet until recently, no such process had been discovered. When developmental psychobiologists first began assessing infant rat development, any form of learning appeared beyond the capabilities of the neonatal rat. However, as our understanding of the newborn’s environment began to evolve and experimental conditions became more naturalistic, the surprising learning capabilities of rodent neonates have emerged. Since then, extensive work characterizing early learning has demonstrated that the basic laws of adult learning also apply to infant rats and that learning occurs naturally within the nest (Brunjes and Alberts, 1979; Campbell, 1984; Galef and Kaner, 1980; Leon, 1975; Miller, Jagielo, and Spear, 1989; Pedersen, Williams, and Blass, 1982; Rudy and Cheatle, 1979; Sullivan, Brake, et al., 1986; Sullivan, Hofer, and Brake, 1986; Sullivan, Wilson, et al., 1990; Terry and Johanson, 1996). These early studies revealed that newborn pups were capable of learning to discriminate, prefer, approach, and maintain proximity to an odor that had been associated with

hofer and sullivan: neurobiology of attachment

789

forms of stimulation that naturally occurred within the early mother-infant interaction. Random presentations of the two stimuli (odor and reward) had no such effect and provided the control procedure necessary to identify the change in behavior as due to associative conditioning. Since this learning required only two or three paired presentations and since the preference was retained into adulthood to enhance sexual behavior (Fillion and Blass, 1986; Moore, Jordan, and Wong, 1996), it seemed to qualify as an “imprinting-like process” that is likely to be central to attachment in slowdeveloping mammals. Indeed, a human analogue of this process was found by Sullivan, Tabarsky-Barba, and colleagues (1991), who showed that when human newborns were presented with a novel odor and then rubbed along their torsos to simulate maternal care, the next day they became activated and turned their head preferentially toward that odor. This result suggests that rapid learning of orientation to olfactory cues is an evolutionarily conserved process in mammalian newborns. Somatosensory information is also of importance in mother-infant interactions, and pup mortality rate (caused by disturbance of nipple orientation and grasp) increases markedly when a pup’s facial somatosensory system is disrupted (Hofer, 1981). Somatosensory stimulation evokes specific orientation behaviors in 2- to 3-day-old pups (Polan and Hofer, 1999), and specific contingent stimulation can be shown to alter pups’ responsiveness to somatosensory stimulation. Following a conditioning procedure where whisker stimulation was paired with a reward, pups showed more vigorous responding to whisker stimulation (Landers and Sullivan, 1999a, 1999b; Sullivan, Landers, et al., 2003), resulting in enhanced responding to tactile stimulation from the mother. From an evolutionary perspective, a reliance on learning for early attachment might appear rather risky considering the potential for inappropriate object choice. However, considering the physical constraints of the nest and pups’ immature motor system, the range of possible attachment figures is limited. Additionally, unique characteristics of infant learning appear to greatly enhance the likelihood of pups developing odor preferences necessary for forming maternal attachment. However, while pups can learn to avoid odors, this learning is greatly constrained. For example, while neonatal rats are capable of learning to avoid tastes/odors paired with malaise, this learning is blocked if pups are nursing during conditioning (Gemberling and Domjan, 1982; Haroutunian and Campbell, 1979; Martin and Alberts, 1979; Melcer, Alberts, and Gubernick, 1985; Rudy and Cheatle 1977, 1978; Shionoya et al., 2006). Other types of learning such as fear conditioning, inhibitory conditioning, and passive avoidance do not appear to emerge until after postnatal days 10–11 (Blozovski and Cudennec, 1980; Camp and Rudy, 1988; Collier et al., 1979; Goldman

790

emotion⁄cognition interactions

and Tobach, 1967; Haroutunian and Campbell, 1979; Stehouwer and Campbell, 1978; Sullivan, Stackenwalt, et al., 2000; Myslivecek, 1997), presumably because this learning could interfere with pups’ interaction with the mother. Clinical observations have taught us not only that attachment occurs to supportive caretakers, but also that children tolerate considerable abuse while remaining strongly attached to the abusive caretaker (Helfer, Kempe, and Krugman, 1997). Although it may initially appear to be counterproductive to form and maintain an attachment to an abusive caretaker, from an evolutionary perspective, it may have been better for an altricial infant to have a bad caretaker than no caretaker. We suggest that this aspect of human attachment may be modeled in the infant rat. During the first postnatal week, we and others have found that a surprisingly broad spectrum of stimuli can function as reinforcers to produce an odor preference in rat pups. These stimuli range from apparently rewarding ones such as milk and access to the mother (Alberts and May, 1984; Brake, 1981; Galef and Sherry, 1973; Johanson and Hall, 1979; Johanson and Teicher, 1980; Leon, 1975; McLean et al., 1993; Pedersen, Williams, and Blass, 1982; Sullivan, Brake, et al., 1986; Sullivan, Hofer, and Brake, 1986; Weldon, Travis, and Kennedy, 1991; Wilson and Sullivan, 1994) to apparently aversive stimuli such as moderate shock and tail pinch (Camp and Rudy, 1988; Moriceau and Sullivan, 2006; Roth and Sullivan, 2005; Sullivan, Hofer, and Brake, 1986; Sullivan, Stackenwalt, et al., 2000; Spear, 1978). These aversive stimuli supporting odor-preference learning elicit escape responses from pups during conditioning, and neonatal rat pups feel pain (Barr, 1995; Collier and Bolles, 1980; Emerich et al., 1985; Fitzgerald, 2005; Stehouwer and Campbell, 1978). As pups mature and reach an age when leaving the nest becomes more likely (Bolles and Woods, 1964; pups begin to walk between 9 and 11 days old), olfactory learning comes to resemble learning in adults more closely and effectively ends the enhanced-learning “sensitive period.” Specifically, odor aversions are easily learned by two-week-olds, and acquisition of odor preferences is limited to odors paired with stimuli of positive value (Camp and Rudy, 1988; Haroutunian and Campbell, 1979; Sullivan and Wilson, 1995; Spear, 1978). Thus the odor learning that underlies early attachment appears to take place in response to a very broad range of contingent events while pups are confined to the nest, but becomes more selective at a time in development when pups begin leaving the nest and encountering novel odors not associated with the mother. Neural Basis of Olfactory Preference Learning The development of a specific olfactory-based attachment system in the rat pup during the first week and a half of life is associated with the acquisition of olfactory bulb neural changes. We found that rat pups express this modified

olfactory bulb response to both natural maternal odors and artificial odors experienced in the nest (Sullivan, Wilson, et al., 1990), as well as to odors in controlled learning experiments (Sullivan and Leon, 1986; Johnson et al., 1995; Moriceau and Sullivan, 2004b; Roth and Sullivan, 2005; Yuan et al., 2002; Wilson and Sullivan, 1991; Wilson, Sullivan, and Leon, 1987). The modified olfactory bulb response is characterized by enhanced immediate-early gene activity (c-Fos), optical imaging, and enhanced 2deoxyglucose (2-DG) uptake in focal, odor-specific glomerular regions in response to the conditioned odor. Within the underlying neural substrate, modified single-unit response patterns of mitral/tufted cells near the enhanced glomerular foci were found (Wilson, Sullivan, and Leon, 1987; Wilson and Leon, 1988; Wilson and Sullivan, 1990), and olfactory bulb anatomical changes were reflected in enlarged glomeruli within these foci (Woo, Coopersmith, and Leon, 1987). As with the behavioral changes in attachment, these neural changes are retained into adulthood, but their acquisition is dependent upon experiences during infancy (Woo and Leon, 1988; Pager, 1974). In adults, learning is distributed throughout the brain, and a similar distribution is also likely occurring in pups. Thus far, in addition to the olfactory bulb learning-associated changes found in sensitive-period pups, piriform “olfactory” cortex and anterior olfactory nucleus also appear to encode learning (Roth and Sullivan, 2005; Kucharski and Hall, 1987; Moriceau et al., 2006; Sullivan and Leon, 1986; review, Roth, Wilson, and Sullivan, 2004). These brain areas, as well as other yet unidentified areas, may mediate the development of a close association of emotional states with events within the sphere of attachment processes (see figure 48.1). Many neurotransmitters have a role in early olfactory learning in neonatal rats (5-HT—McLean et al., 1993;

dopamine—Weldon, Travis, and Kennedy, 1991; Barr and Wang, 1992; glutamate—Mickley et al., 1998; Lincoln et al., 1988; opiates—Barr and Rossi, 1992; Kehoe and Blass, 1986; Roth and Sullivan, 2003, 2006; GABA—Okutani et al., 2003), although the action of norepinephrine (NE) appears particularly important in neural plasticity during early development and in the form of olfactory (Brennen and Keverne, 1997; Wilson and Sullivan, 1994) and somatosensory (Landers and Sullivan, 1999b) learning-induced plasticity used in early attachment. Norepinephrine input to the neonatal olfactory bulb is widespread in the granule cell layer (Woo and Leon, 1995; McLean and Shipley, 1991), which is composed of inhibitory interneurons modulating the bulb’s output through the adjacent mitral cells (Brunjes, Smith-Craft, and McCarty, 1985; Lauder and Bloom, 1974; Trombley and Shepherd, 1992; Wilson and Leon, 1988). Wilson has shown that activation of the NE input to the infant rat’s olfactory bulb during an odor presentation maintains mitral cell responsiveness to that odor, preventing the habituation that these cells normally exhibit to repeated odor presentations (Wilson and Sullivan, 1992). Olfactory bulb NE increases by activation of the locus coeruleus (LC) by the reward during infant odor learning (Rangel and Leon, 1995). Manipulation of NE within the bulb during odor learning further supports a critical role for NE in pup learning. Specifically, an odor paired with increasing olfactory bulb NE (intrabulbar infusions or stimulating LC which is the source of NE) supports odor-preference learning (Sullivan, Stackenwalt, et al., 2000). Conversely, odor learning is blocked if olfactory bulb NE is depleted or its receptors are blocked (Langdon, Harley, and McLean, 1997; Sullivan, Zyzak, et al., 1992; Sullivan, Wilson, et al., 1994). Together, all this suggests that the contingent events of increasing

Figure 48.1 The neural circuit for learning changes with development. In early life, during the sensitive period, the olfactory bulb and anterior piriform cortex code odor learning. With maturation (>10 days old) and termination of the sensitive period, the circuit

expands and includes the posterior piriform cortex and amygdala. Other yet-to-be-identified brain areas are likely involved in this circuit.

hofer and sullivan: neurobiology of attachment

791

olfactory bulb NE and odor stimulation support the neural plasticity responsible for the acquisition of olfactory-based attachment behavior. Developmental changes within the noradrenergic pontine nucleus LC (McLean et al., 1989; McLean and Shipley, 1991; Shipley, Halloran, and De la Torre, 1985) may contribute to the termination of the infant rat’s rapid odorpreference learning. The neonatal LC has unique response characteristics that enable it to respond to a broad range of sensory stimuli and release abundant olfactory bulb NE. Specifically, compared to the adult LC, the infant LC responds to a broader range of sensory stimuli and fails to habituate after repeated presentation of stimuli (Kimura and Nakamura, 1985; Nakamura and Sakaguchi, 1990; Foote, Aston-Jones, and Bloom, 1980; Harley and Sara, 1992; Sara, Dyon-Laurent, and Herve, 1995; Vankov, Herve-Minvielle, and Sara, 1995; Kimura and Nakamura, 1985; Nakamura and Sakaguchi, 1990). The most dramatic response difference between the adult and infant LC is the prolonged response time of the infant. Specifically, a 1-second tactile stimulation is likely to cause a response of a few milliseconds in the adult LC, but a 20- to 30-second response in the neonates. The neonatal LC’s prolonged response is probably due to extensive electronic coupling of neurons through gap junction but also through the unique functioning of the infant’s autoreceptors located on the somadendritic membranes of the LC neurons (Marshall et al., 1991; Nakamura and Sakaguchi, 1990; Winzer-Serhan et al., 1997). Specifically, the LC has recurrent axon collaterals that go back into the LC to release NE. When this NE is released in the adult LC, it activates noradrenergic α2 inhibitory autoreceptors that effectively terminate the LC response within milliseconds. The neonatal LC’s α2 autoreceptors, while present, do not appear functional. The neonate LC also has abundant excitatory α1 autoreceptors that maintain the infant response longer than the adult LC response. At about 10 days old, as the sensitive period ends, the infant LC begins to take on adult characteristics with functional emergence of the α2 inhibitory autoreceptors and diminished α2 excitatory autoreceptors. These results suggest that the neonate’s LC is designed to release copious amounts of NE to support the olfactory bulb neural changes required for neonatal pup odor-preference learning. With maturational changes in the LC during the second week of life, NE release into the bulb is reduced and effectively closes the sensitive period for the imprinting-like learning underlying attachment in this species. We have tested this hypothesis by pharmacologically altering the LC autoreceptors in older pups to reinstate the neonatal LC characteristics. Specifically, we implanted bilateral cannulas into the LC of older pups and infused an NE α2 receptor antagonist to block the newly emerging α2 inhibitory autoreceptors that have begun to terminate the LC response and an NE

792

emotion⁄cognition interactions

α1 receptor agonist to further excite the diminishing α1 excitatory autoreceptors, all while stimulating the LC. Indeed, we were able to reinstate the rapid NE-dependent preference learning in pups simply by pharmacologically reinstating the neonatal characteristics of the LC (Moriceau and Sullivan, 2004b). Neural Basis of Attenuated Aversion Learning Another characteristic of attachment identified by Bowlby (1969) was the child’s strong ties to an abusive attachment figure, which others have identified in other species (Harlow and Harlow, 1965; Fisher, 1955, cited in Rajecki, Lamb, and Obmascher, 1978; review, Helfer, Kempe, and Krugman, 1997; Salzen, 1970). We have been attempting to model this aspect of attachment in rats in an attempt to understand this paradoxical learning in infants. In adults, pairing a neutral stimulus with pain results in the neutral stimulus becoming aversive, eliciting fear as well as avoidance, fleeing, or freezing depending on the experimental context. This learning involves the amygdala, although many additional brain areas are involved in this complex neural circuit (Davis, Walker, and Lee, 1997; Fanselow and Gale, 2003; Fanselow and LeDoux, 1999; Herzog and Otto, 1997; Maren, 2003; McGaugh, Roozendaal, and Cahill, 1999; Pape and Stork, 2003; Pare, Quirk, and LeDoux, 2004; Rosenkranz and Grace, 2002; Sananes and Campbell, 1989; Schettino and Otto, 2001; Sevelinges et al., 2004). We have data suggesting that the inability of pups to learn this early-life version of fear conditioning may be due to the amygdala’s failure to participate in early-life learning. During the neonatal sensitive period, when pups readily learn to prefer odors paired with pain, the two major odor inputs to the amygdala (olfactory bulb and piriform “olfactory” cortex) (see figure 48.1) show activation during learning (Roth and Sullivan, 2005; Moriceau et al., 2006), although the amygdala did not appear to participate as indicated by 14C 2-DG or c-Fos (Roth and Sullivan, 2005; Moriceau and Sullivan, 2006; Moriceau et al., 2006). A similar analysis in older, 12-day-old pups that easily learn odor aversions with odor–0.5-mA shock conditioning shows activation of the basolateral, lateral, and cortical amygdala nuclei (Moriceau et al., 2006; Moriceau and Sullivan, 2006; Sullivan, Stackenwalt, et al., 2000) (see figure 48.2). These results suggest the amygdala or its connections mature around 10 days of age to permit its participation in learning and avoidance acquisition (Berdel and Morys, 2000; Berdel, Morys, and Maciejewska, 1997; Bouwmeester, Wolterink, and van Ree, 2002; Cunningham, Bhattacharyya, and Benes, 2002; Nair and Gonzalaz-Lima, 1999; Morys et al., 1999; Schwob, Haberly, and Price, 1984; Wilson, Best, and Sullivan, 2004). However, our ability to manipulate the age at which pups begin to learn the odor-shock avoidance fear conditioning and amygdala activity with

Figure 48.2 Learning an odor preference during early life from odor–0.5-mA shock does not activate any amygdala nuclei (basolateral complex, BLA/LA; and central, CeA; others not shown) but

corticosterone (CORT) suggests amygdala maturation is not the limiting factor in pup learning (see figure 48.3). There was existing literature that strongly suggested CORT might alter the age when pups begin to learn avoidance and fear to unlearned predator odor (Bialik, Pappas, and Roberts, 1984; Blozovski and Cudennec, 1980; Collier et al., 1979; Moriceau and Sullivan, 2004a; Myslivecek, 1997; Takahashi, 1992; Wiedenmayer and Barr, 2001). Additionally, the emergence of odor–0.5-mA shock-learned fear was related to the declining “stress hyporesponsive period” when pups’ low CORT level is not raised by most stressful stimuli (i.e., restraint, shock; Levine, 2001; Rosenfeld, Wetmore, and Levine, 1992). Indeed, as illustrated in figure 48.3A, we can cause the precocious emergence of fear conditioning and amygdala activation (odor–0.5-mA shock) in pups simply by increasing endogenous CORT (Moriceau and Sullivan, 2004a, 2006; Moriceau et al., 2006). Specifically, CORT injected (3mg/kg) 24 hours and 30 minutes before 8-day-olds received odor–0.05-mA shock conditioning causes pups to learn an odor aversion rather than the age-typical odor preference. This learning activated the amygdala. Moreover, as illustrated in figure 48.3B, decreasing CORT in older pups (conditioning at 12 days old with adrenalectomy at 8 days old) prolongs pups’ inability to learn fear and odor-shock conditioning and continues to produce an odor preference without amygdala activation. The amygdala appears to be the site of CORT action, since intra-amygdala infusions of CORT receptor agonists and antagonists during conditioning respectively terminate or prolong the odor-pain-induced odor-preference learning (Moriceau et al., 2006; Moriceau and Sullivan, 2004a; 2006). We suggest that natural fluctuations in pups’ CORT level can occur in nature and can prematurely end or

does alter its input (GCL, olfactory bulb granule cell layer; PIR, piriform; Roth and Sullivan, 2005).

Figure 48.3 Odor-shock conditioning produces an (A) odor preference in young sensitive period pups (<10 days old) (B) without apparent participation of the amygdala, although CORT permits the learning of an odor aversion and amygdala activity. Older postsensitive-period pups (A) learn the odor aversion (B) along with amygdala activity, although depleting pups of CORT causes pups to revert to the earlier odor-preference learning without amygdala activation (Moriceau et al., 2006).

hofer and sullivan: neurobiology of attachment

793

prolong pups’ attachment-learning system. In early life, pups have a “stress hyporesponsive period” (S. Levine, 1962; A. Levine, 1994), which comes to an end as odorshock conditioning begins to produce the odor avoidance characteristic of fear conditioning. Pups CORT level can be increased through mother’s milk (Yeh, 1984) or lack of sensory stimulation from the mother. Indeed, maternal behavior toward pups appears to be critical in maintaining pups’ low CORT level, as the mother provides milk and sensory stimulation (licking and maternal odor; Levine, Stanton, and Gutierrez, 1988; Rosenfeld et al., 1991; Stanton, Gutierrez, and Levine, 1988; Wiedenmayer et al., 2003). Even in postsensitive-period pups, presence of an anesthetized dam attenuates pups’ response to shock and prevents normal novelty and stress-induced increase in CORT (Richardson, Siegel, and Campbell, 1989; Levine, Stanton, and Gutierrez, 1988; Moriceau and Sullivan, 2006; Stanton and Levine, 1990; Wiedenmayer and Barr, 2001). Similar CORT attenuation by social cues has been found in other social situations throughout development (Carter and Keverne, 2002; Carter et al., 1997; Dallman, 2000; Deschamp, Woodside, and Walker, 2003; Devries, 2002: DeVries, Glasper, and Detillion, 2003; Hennessy, 1984; Hennessy et al., 2006; Hofer, 1996a; Hofer and Shair, 1991). Since increased CORT is important in pup odor–0.5-mA shock-aversion learning and the maternal presence can lower CORT in older pups, we assessed whether the maternal presence would cause older pups to revert back to the attachment odor-shock preference learning. Indeed, the mother’s presence during conditioning had a dramatic effect on postsensitive-period 12- to 15-day-old pups’ odor–0.5mA shock learning, reinstating the odor-shock-induced odor preference characteristic of the sensitive period, and pups failed to show amygdala activity (Moriceau and Sullivan, 2006). Again, CORT action on the amygdala seems to be important, since temporary inactivation of the amygdala (muscimol) had no effect on pup odor preference, while intra-amygdala CORT infusion enabled pups to learn an odor aversion even when the mother was present. These data suggest that two circuits for odor-shock conditioning exist in pups, with maternal presence providing the “switch” by lowering pups’ corticosterone level. Since pups must learn the diet-dependent maternal odor for interactions with the mother (i.e., nipple attachment, approach) this dual learning system may ensure that pups still learn to approach maternal odor, yet learn complex contingencies necessary for life outside the nest. It should be noted that even fetal pups can learn to avoid an odor paired with LiCl and that neonatal pups can learn to avoid odors paired with very high shock (1.2 mA), both of which produce malaise in pups as indicated by diarrhea (Martin and Alberts, 1979; Shionoya et al., 2006; Stickrod,

794

emotion⁄cognition interactions

Kimble, and Smotherman, 1982). This learning is evident before the maturation of the amygdala (required for adult odor-LiCl learning; Touzani and Sclafani, 2005). This finding suggests that pups are capable of learning odor avoidance although a nonamygdala learning circuit exists. Learning appears to use the olfactory bulb, with amygdala activation during learning only occurring close to weaning age (Shionoya et al., 2006). Interestingly, when the nursinginduced blocking of odor-LiCl aversion learning occurs, amygdala activity is also suppressed in the weaning pups (Martin and Alberts, 1979; Shionoya et al., 2006). Unique learning capabilities, facilitating infant attachment, appear throughout the animal kingdom and may have evolved to ensure that altricial animals easily form a repertoire of proximity-seeking behaviors, regardless of the specific qualities of the treatment they receive from the primary caretaker. Observations of mother-infant interactions within the rat nest demonstrate that maternal behaviors are sometimes painful to pups. For example, when the rat mother steps on pups entering or leaving the nest or when she retrieves pups by a leg rather than at the nape of the neck, “broadband” (mixed ultrasonic and audible range) vocalizations, the type associated with noxious stimuli, are elicited (White et al., 1992). It is certainly beneficial to pups not to learn an aversion to their mother’s odor or inhibit approach responses to nest odors; instead, pups need to exhibit approach behaviors to procure the mother’s milk, warmth, and protection. A similar phenomenon exists in avian species. Specifically, experiencing an aversive shock during exposure to an imprinting object strengthens the following response. For example, in a series of classic experiments by Hess (1962) and Salzen (1970), recently hatched chicks were shocked (3 mA for 0.5 sec) while presented with a surrogate mother. The next day, chicks that were shocked showed a significantly stronger following response than chicks that were not shocked. With striking similarity to the infant rat, similar pairings in older chicks resulted in a subsequent aversion to the surrogate mother. Additional mammalian species where a similar phenomenon has been documented include nonhuman primates (Harlow and Harlow, 1965; Sanchez, Ladd, and Plotsky, 2001), dogs (Fisher, 1955, cited in Rajecki, Lamb, and Obmascher, 1978), and humans (review, Helfer, Kempe, and Krugman, 1997). The ability of low levels of CORT to limit both inhibitory and avoidance learning of pups (Bialik, Pappas, and Roberts, 1984; Blozovski and Cudennec, 1980; Collier et al., 1979; Moriceau and Sullivan, 2004a, 2006; Moriceau et al., 2006; Myslivecek, 1997) suggests that the neurobiology of neonatal learning is designed to prevent pups from learning to avoid or inhibit the responses to the mother’s odor that are necessary for survival. Thus, as pups mature and CORT increases, the amygdala becomes more important in learning, although other structures important in adult learning, such as the

hippocampus, may still be immature (Crain et al., 1973; Fanselow and Rudy, 1998; Rudy and Morledge, 1994; Rudy, Stadler-Morris, and Alberts, 1987; Sananes and Campbell, 1989; Stanton, 2000). Furthermore, connectivity between the hippocampus, amygdala, and other brain areas may not become functional until sometime between 12 and 17 days old (Nair and Gonzalez-Lima, 1999). Thus, while some features of the adult learning system appear functional in early life, others have a more protracted development, and their incorporation into the learning circuit needs to be explored.

Maternal separation responses, hidden regulators, and the emergence of emotion in attachment Soon after birth, prenatally acquired perceptual biases and motor programs, stimulus-guided tactile responses, and associative learning create a powerful behavioral control system through which the infant maintains close proximity to its mother. But there is another important attribute of attachment, one that has long been viewed as the strongest evidence for an emotional tie of the infant to its mother: the response to separation. In attachment theory, this response is viewed as an integral part of the proximity-maintenance system, one that represents the affective expression of its motivational nature. Thus the degree or strength of attachment is thought to be responsible for the intensity of the response to separation, and the separation response itself is taken to represent a full expression of the attachment behaviors in the absence of their “goal object.” Experiments in our lab led us to a very different view, in which the processes underlying attachment and the responses to separation are seen as separate and distinct early in postnatal life (Hofer, 1975a, 1983). The response of infant rats, and primates, to maternal separation has been found to involve a complex pattern of changes in a number of different behavioral and physiologic systems (Kraemer, 1992; Hofer, 1996a). We found that this pattern was not an integrated psychophysiological response, as had been supposed (the “despair” phase of the separation response), but the result of a novel mechanism. During separation, we found that each of the individual systems of the infant rat responded to the absence of one or another of the components of the infants’ previous interaction with its mother. Providing one of these components to a separated pup—for example, maternal warmth—maintained the level of brain biogenic amine function underlying the pups’ general activity level (Stone, Bonnet, and Hofer, 1976; Hofer, 1980) but had no effect on other systems, for example, the pups’ cardiac rate. Heart rate fell 40 percent after 18 hours of separation, regardless of whether supplemental heat was provided (Hofer, 1971). The heart rate, normally maintained by sympathetic tone, we found was regulated by maternal provision

of milk to neural receptors in the lining of the pup’s stomach (Hofer and Weiner, 1975). By studying a number of additional systems, such as those controlling sleep-wake states (Hofer, 1976), activity level (Hofer, 1975b), sucking pattern (Brake et al., 1982), vocalization (Hofer and Shair, 1980), and blood pressure (Shear, Brunelli, and Hofer, 1983), we concluded that in maternal separation, all these regulatory components of the motherinfant interaction are withdrawn at once. This widespread loss creates a pattern of increases or decreases in level of function of the infant’s systems, depending upon whether the particular system had been up- or down-regulated by the previous mother-infant interaction. We called these “hidden regulators” because they were not evident in simply observing the mother-infant interaction. Other investigators, using this approach, have discovered other maternal regulatory systems of the same sort. For example, removal of the dam from rat pups was found to produce a rapid (30 min) fall in the pup’s growth hormone (GH) levels, and vigorous tactile stroking of maternally separated pups (mimicking maternal licking) prevented the fall in GH (reviewed in Kuhn and Schanberg, 1991). Brain substrates for this effect were then investigated, and it now appears that GH levels are normally maintained by maternal licking, acting through serotonin (5-HT) 2A and 2C receptor modulation of the balance between growth hormone releasing factor (GRH) and somatostatin (SS), which together act on the anterior pituitary release of GH (Katz et al., 1996). The withdrawal of maternal licking by separation allows GRH to fall and SS to rise, resulting in a precipitate fall in GH. There are several biological similarities between this maternal deprivation effect in rats and the growth retardation that occurs in some variants of human reactive attachment disorders of infancy. Applying this new basic-science knowledge about the regulation of GH to low-birth-weight prematurely born babies, Tiffany Field and coworkers joined the Schanberg group. They used a combination of stroking and limb movement, administered three times a day for 15 minutes each time, and continued throughout the babies’ two-week hospitalization. This intervention increased weight gain, head circumference, and behavior development test scores in relation to a randomly chosen control group, with beneficial effects discernible many months later (Field et al., 1986). These results show that maternal-infant interactions are not limited to a transient regulatory role but can function as ongoing regulators of early development. Attachment, Emotion, and the Origins of Vocal Communication One of the best-known responses to maternal separation is the infant’s isolation call, a behavior that occurs in a wide variety of species (Newman, 1988). In the rat, this call is in the ultrasonic range (40 kHz) and

hofer and sullivan: neurobiology of attachment

795

appears in the first or second postnatal day (Noirot, 1968). Pharmacological studies show that the ultrasonic vocalization (USV) response to isolation is attenuated or blocked in a dose-dependent manner by clinically effective anxiolytics that act at benzodiazepine and serotonin receptors and, conversely, that USV rates are increased by compounds known to be anxiogenic in humans, such as benzodiazepine receptor inverse agonists (beta-carboline, FG 1742) and GABAa receptor ligands such as pentylenetetrazol (Miczek, Tornatsky, and Vivian, 1991; Hofer, 1996b). Within serotonin and opioid systems, receptor subtypes known to have opposing effects on experimental anxiety in adult rats also have opposing effects on infant calling rates. Neuroanatomical studies in infant rats show that stimulation of the periaquaductal gray area produces USV and that chemical lesions of this area prevent calling (Goodwin and Barr, 1998). The more distal motor pathway is through nucleus ambiguous and both laryngeal branches of the vagus nerve (Wetzel, Kelly, and Campbell, 1980). Higher centers known to be involved in cats and primates suggest a neural substrate for isolation calls involving primarily the hypothalamus, amygdala, thalamus, and hippocampus (Jurgens, 1994), brain areas known to be involved in adult human and adult animal anxiety and defensive responses. This evidence strongly suggests that sudden isolation produces an early anxiety-like state in rat pups, one that is expressed overtly in the rates of infant calling as well as covertly by the autonomic and adrenocortical systems (A. Levine, 1994). How does this calling behavior, together with its inferred underlying affective state, function as a communication system between mother and pup? Infant rat USVs are a powerful stimulus for the lactating rat, capable of causing her to interrupt an ongoing nursing bout, initiate searching outside the nest, and direct her search toward the source of the calls (Smotherman et al., 1978). The mother’s retrieval response to the pup’s vocal signals then results in renewed contact between pup and mother. This contact in turn quiets the pup. The isolation and comfort responses in attachment theory are described as expressions of interruption and reestablishment of a social bond. Such a formulation would predict that, since the pup recognizes its mother by her scent (as described earlier), pups made acutely anosmic would fail to show a comfort response. But anosmic pups show comfort responses that are virtually unaffected by loss of their capacity to recognize their mother in this way (Hofer and Shair, 1991). Instead, we and others have found multiple regulators of infant USV within the contact between a pup and its mother or littermates: warmth, tactile stimuli, and milk as well as their scent (Hofer, 1996b). Provision of these modalities separately, by experimental design, and then in combination, elicits a graded response, with the maximum isolation

796

emotion⁄cognition interactions

Table 48.1 Maternal regulators of infant systems in two-week-old rat pups Infant Systems Maternal Regulators and Effects Behavioral Activity level Thermal, heat (+) Thermal, cold (−) Tactile stimulation (−) Olfactory cues (−) Sucking Nutritive Gastric distention (−) Perioral, tactile/chemical (−) Nonnutritive Vocalization Thermal, texture, olfactory, milk (−) Ultrasound Autonomic Sympathetic Cardiac (beta-adrenergic) Vasomotor (alpha-adrenergic)

Milk at gastric receptors (+) Milk at gastric receptors (−)

Endocrine Growth hormone ACTH, CRH Corticosterone

Tactile stimulation (+) Tactile stimulation (−) Milk, acting on adrenal response to ACTH (−) Regularly timed tactile and/or Sleep-wake states nutrient stimuli REM duration (+) Awakening frequency (−); duration (−) The direction of regulation is indicated by (+) or (−) for up- or down-regulation during normal ongoing mother-infant interaction. Upon maternal separation, release from regulation produces changes in the opposite direction in each infant system.

calling rates occurring when all are withdrawn at once and with the full comfort response being elicited only when all are presented together in the form of a multimodal surrogate (see table 48.1). We have recently learned that this early affective state and its “social” regulation is not a transitory developmental adaptation that is limited to the preweaning period, but is part of the long-term trajectory of affective and social development. We have been selectively breeding rats according to the level of their isolation vocalization level at 10 days, creating high, low, and random bred lines (Hofer et al., 2001). Recently we have found that this infant behavior is on a developmental trajectory with specific adolescent and adult traits linked to juvenile social play (Brunelli et al., 2006) and to measures linked to anxiety and depression in adulthood (Zimmerberg et al., 2005). Adults that were bred to vocalize at high levels at 10 days of age showed significantly more behavioral indices of anxiety in openfield emergence and novel social-interaction tests, while showing depression-like responses in the Porsolt swim test.

Adolescents in the low line showed low levels of social play overall, engaging in social investigation behavior instead. These studies of early isolation calling suggest that individual differences in early attachment responses and their neurobiological substrates can be found to be developmentally continuous with certain aspects of adult social and emotional behavior, and evidence for this hypothesis will be presented in the next section.

Mother-infant interaction patterns, infant development, and maternal behavior in the next generation One of the major tenets of attachment theory is the idea that the nature of an infant’s early attachment interaction pattern can have long-term developmental importance; not only for their patterns of mothering when they become adults, but for broader aspects of behavior as well, such as levels of anxiety, depressive affect, and social interactions generally. The evidence for this idea in humans is suggestive but not at all conclusive. Our discovery of early mother-infant regulatory interactions suggested that they might represent a novel mechanism for long-term effects of early relationships if these interactions could be shown to regulate the course of development into adulthood. So, by observing naturally occurring variations in a wide variety of mother-infant interactions in strains of genetically identical rats in the first two weeks of postnatal life, we were able to identify three behaviors that were correlated significantly with the severity of hypertension in adults of the genetically hypertensive strain (Myers et al., 1989). We were able to replicate this correlation with blood pressure in the normotensive progenitor strain, and since the animals in each strain were genetically identical, the differences in adulthood could be attributed to the different early mother-infant interaction patterns: the time spent in contact, the amount of maternal licking and grooming, and the time mothers spent in a highly stimulating high-arched resting position. Thus we were able to conclude that early differences in the pattern of early maternal regulating interactions could initiate long-term developmental effects lasting into adulthood. But the possible physiological mechanisms for maternal regulation of offspring blood pressure were not evident, and we had no clue as to how the biological effects of such early experiences were represented in a way that could be maintained into adulthood. The work of Michael Meaney and his colleagues over the past decade has greatly enlarged our understanding of the biological processes at work in these lasting effects of early relationships (reviewed in Cameron et al., 2005). They discovered that normal variation in two of the same mother-infant behaviors found in our studies (maternal licking of pups and high-arched maternal nursing posi-

tion) systematically modifies the development of the adrenocortical stress response and the behavioral fear response to open spaces in the mother’s adult offspring: low levels of those maternal behaviors lead to more intense responses. In a series of studies, Meaney and his colleagues have been able, first, to trace the effects of different levels of maternal behaviors to different levels of hippocampal cell membrane corticosterone receptors that sense the level of adrenocortical hormone and modulate the hormonal response to stress. Next they found that levels of activation of the genes responsible for the synthesis of these receptors in the adult offspring of the high- or low-licking/grooming mothers were differently regulated. These findings link variation within normally occurring levels of mother-infant interactions to molecular processes regulating gene expression in the adult. But how are the behavioral interactions translated into molecular genetic changes in the brains of the infants, and how are the changes maintained into adulthood? Answers to these questions were not forthcoming until very recently. First, Fenoglio, Chen, and Barum (2006) reported that the changes in brain systems known to be primarily involved in the integration of adrenocortical and behavioral stress responses (central nucleus of the amygdala, thalamic paraventricular area, and bed nucleus of the stria terminalis) were specifically activated in pups by high levels of maternal grooming that are known to take place when pups are returned to their mothers after a 15-minute handling period. One grooming bout produced minor changes, but after repeated daily handling during the first nine postnatal days, there was activation in cell signaling pathways known to be involved in the synaptic facilitation underlying learning and the gene expression changes of memory storage. At day 9, only the first steps in the pathway were involved; by 23 postnatal days, areas regulating corticosterone production showed persistent reduction in gene activation in the hypothalamic paraventricular nucleus; and by 45 days, the stressrelated brain regions we have described were found to show enduring changes in the activation levels of key regulatory factors and gene functions that are responsible for the full range of adrenocortical and behavioral fear responses. What these studies have found is that a form of activity-dependent process, utilizing known mechanisms of learning, underlies the long-term developmental regulatory effects of different levels of maternal-infant interaction during the early postnatal period. Second, Meaney’s group (Weaver et al., 2004) had found that differences between the two mothering types were associated with changes in a newly recognized system of gene regulation called epigenetics. The complex molecular structure, chromatin, that supports and surrounds the long thin DNA strands includes several mechanisms for

hofer and sullivan: neurobiology of attachment

797

modifying gene function, silencing some genes and opening others to activation (gene transcription) in response to outside signals. They found that two genes in brain cell regions known to regulate adrenocortical responses showed strong evidence of such modification (methylation and deacetylation) in their epigenetic regulation. Most recently, Weaver, Meaney, and Szf (2006) carried out a massive analysis of all the changes in gene activation in adults of the two early-experience groups. They used a new technique, the gene “chip,” that allows simultaneous assessment of the activation state of all the roughly 30,000 genes present in each cell in a piece of brain tissue, in this case the hippocampus. Of the 300 genes that they found to be differently activated in the two mothering groups, about 100 are known to be involved in cell-to-cell signaling in pathways of brain formation and function. These results give us a first view of an entirely new means of studying the long-term developmental roles of early attachment. Can the cellular memory mechanism of the Fenoglio, Chen, and Barum group be integrated with the epigenetic changes described by the Meaney group? In the process of answering this question in the next few years, we will learn a great deal more about developmental mechanisms underlying the long-range developmental effects of different early mother-infant interaction patterns. What these studies have done for us is to uncover for the first time some of the component processes at the cell/molecular level, through which different patterns of early mother-infant interactions act to regulate the long-term development of physiological and behavioral systems into adulthood. Since the genes involved in this long-term regulation apparently number in the hundreds, it seems likely that the effects of different early mother-infant interaction patterns are more extensive and may well involve more systems than those already identified. We know from human observations that attachment patterns tend to be repeated by daughters in the next generation, an effect generally thought to be mediated by cultural processes and by the formation of a psychological representation of mothering in daughters, based on their experiences with their own mothers. Now we have good evidence for underlying biological processes in this transgenerational process. Meaney’s group (Cameron et al., 2005) has found that mothers with high and low interaction levels pass these different maternal behavior patterns on to their daughters, along with the different levels of adult adrenocortical and fear responses. This transgenerational effect on maternal behavior is beginning to be linked to the effects of different maternal interaction patterns on the one-week-old pups’ level of estrogen-induced oxytocin receptors in the preoptic area of the pups’ developing hypothalamus, a region known to be of central

798

emotion⁄cognition interactions

importance to later maternal behavior (Champagne et al., 2001). Oxytocin has been shown in animal studies to have a “priming” role in the initiation of maternal behavior in the immediate postpartum period (Pederson, 1999), and recent studies in humans have found that maternal oxytocin levels were positively correlated with measures of maternal behavior in the first month postpartum and with measures of maternal preoccupations and mental representations during pregnancy as well as postpartum. Interestingly, maternal cortisol levels sampled simultaneously were negatively correlated with maternal behavior. Taken together, these findings begin to reveal specific developmental pathways that underlie and support, at a biological level, the cultural, social, and subjective mental processes that, together with genetic predisposition, transmit patterns of “attachment” and maternal care from one generation to the next.

Summary and implications for human development In this chapter we have reexamined the major psychological constructs of attachment theory and described a series of component processes, discovered in experiments with laboratory rats, that enable us to begin to understand the behavioral mechanisms involved and to identify their underlying neural substrates. We first described the processes that bring the neonatal mammal into close proximity to the mother, continue to keep the infant close, act to reunite the separated infant with the mother, and cause a complex patterned response to prolonged maternal separation. These several components tell us that an “enduring social bond” (to use Bowlby’s term) has been formed, but we can now understand the bond in terms of separate processes that can be delineated as they work independently, serially, or in parallel to produce the familiar behavioral signs of “attachment.” The discovery of these component processes allows us to begin to understand what makes up the “glue” that holds the infant to the mother. The discovery of regulatory interactions within the motherinfant relationship allows us to escape the circularity of the traditional attachment model, in which the response to separation is attributed to disruption of a social bond, the existence of which is inferred from the presence of that same separation response. Some of the individual processes described allow us to understand how the infant comes to identify and orient toward the mother by different means at different stages in development. Beginning before birth and continuing in the newborn period, novel processes of associative learning have been discovered that allow us for the first time to identify and understand the mysterious “imprinting-like process” that Bowlby envisioned as the altricial mammalian equivalent of avian imprinting. And,

finally, we can begin to see how one of the consequences of these early learning processes, acting within repeated regulatory interactions, is to provide a novel source of experiences for the formation of the mental representation of the infant-mother relationship. With this enlarged view, some processes have been discovered to play an important role in the development of proximity maintenance between infant and mother that were not previously considered to be important (e.g., prenatal learning). And events that were thought to be an integral part of the attachment bond (e.g., separation responses) have been found to be capable of being produced by other, independent mechanisms (e.g., the withdrawal of previously unknown maternal regulators). More broadly, our understanding of the evolutionary survival value of remaining close to the mother has been expanded to include the many pathways available for regulation of the infant’s physiological and behavioral systems by its interactions with the caregiver. The relationship thus provides an opportunity for the mother to shape both the developing physiology and the behavior of her offspring through her patterned interactions with the infant. Behavioral adaptations to environmental events occurring in the life of the mother can thus lead to “anticipatory” changes in offspring brain and behavior development preparing them in advance for different environmental demands—a novel evolutionary mechanism. The discovery of regulatory interactions and the effects of their withdrawal allows us to understand not only the responses to separation in young organisms of limited cognitive-emotional capacity, but also the familiar experienced emotions and memories that can be verbally described to us by older children and adults (Hofer, 1996a). It is not that rat pups respond to loss of regulatory processes, while human infants respond to emotions of love, sadness, anger, and grief. Human infants, as they mature, can respond at the symbolic level as well as at the level of the behavioral and physiological processes of the regulatory interactions. The two levels appear to be organized as parallel and complementary response systems. Even adult humans continue to respond in important ways at the sensorimotorphysiologic level in their social interactions, separations, and losses, continuing a process begun in infancy. A good example of such responses is the mutual regulation of menstrual synchrony among close female friends, an effect that takes place out of conscious awareness and has recently been found to be mediated at least in part by a pheromonal cue (Stern and McClintock, 1998). Other examples may well include the role of social interactions in entraining circadian physiological rhythms, the disorganizing effects of sensory deprivation, and the remarkable therapeutic effects of social support on the course of medical illness (reviewed in Hofer, 1984). In this way, adult love,

grief, and bereavement may well contain elements of the simpler regulatory processes that we can clearly see in the attachment responses of infant animals to separation from their social companions. This is perhaps the most challenging area for future research: to find out how to apply what we have learned in basic brain and behavior studies to the human condition. Studies on other animals cannot be used to define human nature, but many of the principles and new ways to approach the mother-infant interaction described in this chapter can be useful in studies of the human mother-infant relationship. We must take into account obvious differences between species, such as the primacy of olfaction and tactile senses in the newborn rat as contrasted to the wider range of senses available to the human newborn. For example, the learning processes regulating approach and proximity that are based on olfaction and tactile senses in the rat pup are likely to be mediated by visual and auditory systems as well as olfaction and touch in the human newborn. At present, there is widespread use of the concept of regulation as inherent in the mother-infant interaction in humans. This word is generally used in two ways: first, to refer to the graded effects of different patterns of interaction on the emotional responses of the infant, the so-called “regulation of affect” (Schore, 1994); second, to refer to how the behaviors initiated by the infant or mother and/or their responses to each other act to regulate the interaction itself, its tempo, or its rhythm (hence its “quality”) or the distance (both psychological and physical) between the members of the dyad (e.g., Gergeley and Watson, 1999). At a more basic science level, the word “regulation” is used extensively in the literature on molecular genetic, cellular, and electrophysiological brain processes in development. Indeed gene regulation is now recognized as the central process for building an organism and its brain. Thus regulation serves as a useful conceptual link across wide differences in the level of organization at which developmental processes are studied. A role for nonverbal features of the early mother-infant interaction in the specification of lasting mental representations of maternal behavior in the adult is a central hypothesis of clinical attachment theory. It would be difficult to confirm clinically this useful idea with any degree of certainty. But the transgenerational effects on affective responses, maternal behavior, and adrenocortical function described in this chapter can serve as a research model for understanding the psychobiological mechanisms for early nonverbal influences on development. Prospective clinical studies from infancy to childhood would be most interesting and could reveal which residues of particular early interactions can be related to which later characteristics of the stories, play, or social relationships of older children, and eventually the parental behavior of adults.

hofer and sullivan: neurobiology of attachment

799

REFERENCES Alberts, J. R., and B. May, 1984. Nonnutritive, thermotactile induction of filial huddling in rat pups. Dev. Psychobiol. 17:161– 181. Barr, G. A., 1995. Ontogeny of nociception and antinociception. NIDA Res. Monogr. 158:172–201. Barr, G. A., and G. Rossi, 1992. Conditioned place preference from ventral tegmental injections of morphine in neonatal rats. Dev. Brain Res. 66:133–136. Barr, G. A., and S. Wang, 1992. Peripheral and central administration of cocaine produces conditioned odor preferences in the infant rats. Brain Res. 599:181–185. Berdel, B., and J. Morys, 2000. Expression of calbindin-D28k and parvalbumin during development of rat’s basolateral amygdaloid complex. Int. J. Dev. Neurosci. 18(6):501–513. Berdel, B., J. Morys, and B. Maciejewska, 1997. Neuronal changes in the basolateral complex during development of the amygdala of the rat. Int. J. Dev. Neurosci. 15(6):755–765. Bialik, R. J., B. A. Pappas, and D. C. Roberts, 1984. Neonatal 6-hydroxydopamine prevents adaptation to chemical disruption of the pituitary-adrenal system in the rat. Horm. Behav. 18:12– 21. Blass, E. M., 1990. Suckling: Determinants, changes, mechanisms, and lasting impressions. Dev. Psychol. 26(4):520–533. Blozovski, D., and A. Cudennec, 1980. Passive avoidance learning in the young rat. Dev. Psychobiol. 13:513–518. Bolles, R. C., and P. J. Woods, 1964. The ontogeny of behavior in the albino rat. Anim. Behav. 12:427–441. Bouwmeester, H., G. Wolterink, and J. M. van Ree, 2002. Neonatal development of projections from the basolateral amygdala to prefrontal, striatal, and thalamic structures in the rat. J. Comp. Neurol. 442(3):239–249. Bowlby, J., 1969. Attachment and Loss, vol. 1. Attachment. New York: Basic Books. Bowlby, J., 1982. Attachment and loss: Retrospect and prospect. Am. J. Orthopsychiatry 52:664–678. Brake, S., 1981. Suckling infant rats learn a preference for a novel olfactory stimulus paired with milk delivery. Science 211:506– 508. Brake, S., D. J. Sager, R. Sullivan, and M. A. Hofer, 1982. The role of intraoral and gastrointestinal cues in the control of sucking and milk consumption in rat pups. Dev. Psychobiol. 15: 529–541. Brennan, P. A., and E. B. Keverne, 1997. Neural mechanisms of mammalian olfactory learning. Prog. Neurobiol. 51:457– 481. Brunelli, S. A., R. Nie, C. Whipple, V. Winiger, M. A. Hofer, and B. Zimmerberg, 2006. The effects of selective breeding for infant ultrasonic vocalizations on play behavior in juvenile rat. Physiol. Behav. 87(3):527–536. Brunjes, P. C., and J. R. Alberts, 1979. Olfactory stimulation induces filial preferences for huddling in rat pups. J. Comp. Physiol. Psychol. 93:548–555. Brunjes, P. C., L. K. Smith-Craft, and R. McCarty, 1985. Unilateral odor deprivation: Effects on the development of olfactory bulb catecholamines and behavior. Dev. Brain Res. 22: 1–6. Cameron, N. M., F. A. Champagne, P. Carine, E. W. Fish, K. Ozaki-Kuroda, and M. J. Meaney, 2005. The programming of individual differences in defensive responses and reproductive strategies in the rat through variations in maternal care. Neurosci. Biobehav. Rev. 29:843–865.

800

emotion⁄cognition interactions

Camp, L. L., and J. W. Rudy, 1988. Changes in the categorization of appetitive and aversive events during postnatal development of the rat. Dev. Psychobiol. 21:25–42. Campbell, B. A., 1984. Reflections on the ontogeny of learning and memory. In R. Kail and N. E. Spear, eds., Comparative Perspectives on the Development of Memory, 23–35. Hillsdale, NJ: Lawrence Erlbaum. Carter, C. S., and E. B. Keverne, 2002. The neurobiology of social affiliation and pair bonding. In D. W. Pfaff, ed., Hormones, Brain, and Behavior, 299–337. San Diego: Academic Press. Carter, C. S., J. I. Lederhandler, and B. Kirkpatrick, 1997. The Integrative Neurobiology of Affiliation. New York: New York Academy of Sciences. Champagne, F., J. Diorio, S. Sharma, and M. J. Meaney, 2001. Variations in maternal care in the rat are associated with differences in estrogen-rated changes in oxytocin receptor levels. Proc. Natl. Acad. Sci. USA 98:12736–12741. Collier, A. C., and R. C. Bolles, 1980. The ontogenesis of defensive reactions to shock in preweanling rats. Dev. Psychobiol. 13(2): 141–150. Collier, A. C., J. Mast, D. R. Meyer, and C. E. Jacobs, 1979. Approach-avoidance conflict in preweanling rats: A developmental study. Anim. Learn. Behav. 7:514–520. Crain, B., C. Cotman, D. Taylor, and G. Lynch, 1973. A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat. Brain Res. 63:195–204. Cunningham, M. G., S. Bhattacharyya, and F. M. Benes, 2002. Amygdalocortical sprouting continues into early adulthood: Implications for the development of normal and abnormal function during adolescence. J. Comp. Neurol. 453(2):116– 130. Dallman, M. F., 2000. Interaction among social environment, the hypothalamic pituitary-adrenal axis, and behavior. Horm. Behav. 41:405–413. Davis, M., D. L. Walker, and Y. Lee, 1997. Roles of the amygdala and bed nucleus of the stria terminalis in fear and anxiety measured with the acoustic startle reflex: Possible relevance to PTSD. Ann. NY Acad. Sci. 821:305–331. DeCasper, A. J., and W. P. Fifer, 1980. Of human bonding: Newborns prefer their mothers’ voices. Science 208:1174– 1176. DeCasper, A. J., and M. J. Spence, 1986. Prenatal maternal speech influences a newborn’s perception of speech sounds. Infant Behav. Dev. 9:133–150. Deschamps, S., B. Woodside, and C. D. Walker, 2003. Pups presence eliminates the stress hyporesponsiveness of early lactating females to a psychological stress representing a threat to the pups. J. Neuroendocrinol. 15:486–497. DeVries, A. C., 2002. Interaction among social environment, the hypothalamic-pituitary-adrenal axis, and behavior. Horm. Behav. 41:405–413. DeVries, A. C., E. R. Glasper, and C. E. Detillion, 2003. Social modulation of stress responses. Physiol. Behav. 79:399–407. Emerich, D. F., F. M. Scalzo, E. K. Enters, N. Spear, and L. Spear, 1985. Effects of 6-hydroxydopamine-induced catecholamine depletion on shock-precipitated wall climbing of infant rat pups. Dev. Psychobiol. 18:215–227. Fanselow, M. S., and G. D. Gale, 2003. The amygdala, fear, and memory. Ann. NY Acad. Sci. 985:125–134. Fanselow, M. S., and J. E. LeDoux, 1999. Why we think plasticity underlying Pavlovian fear conditioning occurs in the basolateral amygdala. Neuron 23:229–232.

Fanselow, M. S., and J. W. Rudy, 1998. Convergence of experimental and developmental approaches to animal learning and memory processes. In T. J. Carew, R. Menzel, and C. J. Shatz, eds., Mechanistic Relationships Between Development and Learning, 15– 28. New York: J. Wiley & Sons. Fenoglio, K. A., Y. Chen, and T. Z. Baram, 2006. Neuroplasticity of the hypothalamic-pituitary-adrenal axis early in life requires recurrent recruitments of stress-regulating brain regions. J. Neurosci. 26(9):2434–2442. Field, T. M., S. M. Schanberg, F. Scafidid, C. R. Bauer, N. Vega-Lahr, R. Garcia, J. Nystrom, and C. M. Kuhn, 1986. Tactile/kinesthetic stimulation effects on preterm neonates. Pediatrics 77(5):654–658. Fifer, W. P., and C. M. Moon, 1995. The effects of fetal experience with sound. In J. -P. Lecanuet, W. P. Fifer, N. A. Krasnegor, and W. P. Smotherman, eds., Fetal Development—A Psychobiological Perspective, Hillsdale, NJ: Erlbaum. Fillion, T. J., and E. M. Blass, 1986. Infantile experience with suckling odors determines adult sexual behavior in male rats. Science 231:729–731. Fitzgerald, M., 2005. The development of nociceptive circuits. Nature Rev. Neurosci. 6:507–520. Foote, S. L., G. Aston-Jones, and F. E. Bloom, 1980. Impulse activity of locus coeruleus neurons in awake rats and monkey is a function of sensory stimulation and arousal. Proc. Natl. Acad. Sci. USA 77:3033–3037. Galef, G. G., and D. F. Sherry, 1973. Mother’s milk: Medium for transmission of cues reflecting the flavor of mother’s diet. J. Comp. Physiol. Psychol. 83:374–378. Galef, G. G., and H. C. Kaner, 1980. Establishment and maintenance of preference for natural and artificial olfactory stimuli in juvenile rats. J. Comp. Physiol. Psychol. 94:588–595. Gemberling, G. A., and M. Domjan, 1982. Selective associations in one-day-old rats: Taste-toxicosis and texture-shock aversion learning. J. Comp. Physiol. Psychol. 96(1):105–113. Gergeley, G., and J. S. Watson, 1999. Early social-emotional development: Contingency perception and the social biofeedback model. In P. Rochat, ed., Early Social Cognition, 101–136. Hillsdale, NJ: Erlbaum. Goldman, P. S., and E. Tobach, 1967. Behavior modification in infant rats. Anim. Behav. 15:559–562. Goodwin, G. A., and G. A. Barr, 1998. Behavioral and heart rate effects of infusing kainic acid into the dorsal midbrain during early development in the rat. Dev. Brain Res. 107:11–20. Harley, C. W., and S. J. Sara, 1992. Locus coeruleus burst induced by glutamate trigger delayed perforant path spike amplitude potentiation in the dentate gyrus. Exp. Brain Res. 89: 581–587. Harlow, H. F., and M. K. Harlow, 1965. The affectional systems. In A Schrier, H. F. Harlow, and F. Stollnitz, eds., Behavior of Nonhuman Primates, vol. 2. New York: Academic Press. Haroutunian, V., and B. A. Campbell, 1979. Emergence of interoceptive and exteroceptive control of behavior in rats. Science 205:927–929. Helfer, M. E., R. S. Kempe, and R. D. Krugman, 1997. The Battered Child. Chicago: University of Chicago Press. Hennessey, M. B., G. Hornscauh, S. Kaiser, and N. Sachser, 2006. Cortisol response and social buffering: A study throughout the lifespan. Horm. Behav. 49:383–390. Hepper, P. G., 1986. Parental recognition in the rat. Q. J. Exp. Psychol. 38B:151–160. Hepper, P. G., 1987. The amniotic fluid: An important priming role in kin recognition. Anim. Behav. 35(5):1343–1346.

Herzog, C., and T. Otto, 1997. Odor-guided fear conditioning in rats. 2. Lesions of the anterior perirhinal cortex disrupt fear conditioned to the explicit conditioned stimulus but not to the training context. Behav. Neurosci. 111(6):1265–1272. Hess, E. H., 1962. Ethology: An approach to the complete analysis of behavior. In R. Brown, E. Galanter, E. H. Hess, and G. Mendler, eds., New Directions in Psychology, New York: Holt, Rinehart and Winston. Hofer, M. A., 1971. Cardiac rate regulated by nutritional factor in young rats. Science 172:1039–1041. Hofer, M. A., 1975a. Infant separation responses and the maternal role. Biol. Psychiatry 10:149–153. Hofer, M. A., 1975b. Studies on how early maternal separation produces behavioral change in young rats. Psychosom. Med. 37: 245–264. Hofer, M. A., 1976. The organization of sleep and wakefulness after maternal separation in young rats. Dev. Psychobiol. 9:189– 205. Hofer, M. A. 1980. The effects of reserpine and amphetamine on the development of hyperactivity in maternally deprived rat pups. Psychosom. Med. 42:513–520. Hofer, M. A., 1981. Effects of infraorbital nerve section on survival, growth and suckling behaviors of developing rats. J. Comp. Physiol. Psychol. 95:123–133. Hofer, M. A., 1983. On the relationship between attachment and separation processes in infancy. In R. Plutchik and H. Kellerman, eds., Emotions in Early Development, 199–216. New York: Academic Press. Hofer, M. A., 1984. Relationships as regulators: A psychobiological perspective on bereavement. Psychosom. Med. 46:183– 197. Hofer, M. A. 1996a. On the nature and consequences of early loss. Psychosom. Med. 58(6):570–581. Hofer, M. A. 1996b. Multiple regulators of ultrasonic vocalization in the infant rat. Psychoneuroendocrinology 21(2):203–217. Hofer, M. A., J. R. Mesmeda, H. N. Shair, and S. A. Brunelli, 2001. Developmental effects of selective breeding for an infantile trait: The rat pup isolation cell. Dev. Psychobiol. 39: 231–246. Hofer, M. A., and H. N. Shair, 1980. Sensory processes in the control of isolation-induced ultrasonic vocalization by 2 week old rats. J. Comp. Physiol. Psychol. 94:271–279. Hofer, M. A., and Shair, H. N. 1991.Trigeminal and olfactory pathways mediating isolation distress and companion comfort responses in rat pups. Behav. Neurosci. 105:699–706. Hofer, M. A., and H. Weiner, 1975. Physiological mechanisms for cardiac control by nutritional intake after early maternal separation in the young rat. Psychosom. Med. 37:8–24. Johanson, I. B., and W. G. Hall, 1979. Appetitive learning in 1day-old rat pups. Science 205:419–421. Johanson, I. B., and M. Teicher, 1980. Classical conditioning of an odor preference in 3-day-old rats. Behav. Neural Biol. 29:132– 136. Johnson, B. A., C. C. Woo, H. Duong, V. Nguyen, and M. Leon, 1995. A learned odor evokes an enhanced Fos-like glomerular response in the olfactory bulb of young rats. Brain Res. 699: 192–200. Jurgens, U., 1994. Review article: The role of periaqueductal grey in vocal behaviour. Behav. Brain Res. 62:107–117. Katz, L. M., L. Nathan, C. M. Kuhn, and S. M. Schanberg, 1996. Inhibition of GH in maternal separation may be mediated through altered serotonergic activity at 5HT2A and 5HT2C receptors. Psychoneuroendocrinology 21(2):219–235.

hofer and sullivan: neurobiology of attachment

801

Kehoe, P., and E. Blass, 1986. Central nervous system mediation of positive and negative reinforcement in neonatal albino rats. Dev. Brain Res. 27:69–75. Kimura, F., and S. Nakamura, 1985. Locus coeruleus neurons in the neonatal rat: Electrical activity and responses to sensory stimulation. Dev. Brain Res. 23:301–305. Kraemer, G. W., 1992. A psychobiological theory of attachment. Behav. Brain Sci. 15(3):493–511. Kucharski, D., and W. G. Hall, 1987. New routes to early memories. Science 238:786–788. Kuhn, C. M., and S. M. Schanberg, 1991. Stimulation in infancy and brain development. In B. J. Carroll, ed., Psychopathology and the Brain, 97–112. New York: Raven Press. Landers, M., and R. M. Sullivan, 1999a. Vibrissae evoked behavior and conditioning before functional ontogeny of the somatosensory vibrissae cortex. J. Neurosci. 19:5131– 5137. Landers, M., and R. M. Sullivan, 1999b. Norepinephrine and associative conditioning in the neonatal rat somatosensory system. Dev. Brain Res. 114:261–264. Langdon, P. E., C. W. Harley, and J. H. McLean, 1997. Increased ß adrenoceptor activation overcomes conditioned olfactory learning induced by serotonin depletion. Dev. Brain Res. 102: 291–293. Lauder, J. M., and F. E. Bloom, 1974. Ontogeny of monoamine neurons in the locus coeruleus, raphe nuclei and substantia nigra of the rat. J. Comp. Neurol. 155:469–482. Lecanuet, J., and B. Schaal, 1996. Fetal sensory competencies. Eur. J. Obstet. Gynaecol. Reprod. Biol. 68:1–23. Leon, M., 1975. Dietary control of maternal pheromone in the lactating rat. Physiol. Behav. 14:311–319. Levine, S., 1994. The ontogeny of the hypothalamic-pituitaryadrenal axis: The influence of maternal factors. In R. deKloer, E. C. Azmitia, and P. W. Landfield, eds. Ann. NY Acad. Sci. 746: 275–293. Levine, S., 1962. Plasma-free corticosteroid response to electric shock in rats stimulated in infancy. Science 135:795–796. Levine, S., 2001. Primary social relationships influence the development of the hypothalamic-pituitary-adrenal axis in the rat. Physiol. Behav. 73:255–260. Levine, S., M. E. Stanton, and Y. R. Gutierrez, 1988. Maternal modulation of pituitary-adrenal activity during ontogeny. Adv. Exp. Med. Biol. 245:295–310. Lincoln, J., R. Coopersmith, E. W. Harris, C. W. Cotman, and M. Leon, 1988. NMDA receptor activation and early olfactory learning. Brain Res. 467:309–312. Makin, J. W., and R. H. Porter, 1989. Attractiveness of lactating females’ breast odors to neonates. Child Dev. 60:803– 810. Maren, S., 2003. The amygdala, synaptic plasticity, and fear memory. Ann. NY Acad. Sci. 985:106–113. Marshall, K. C., M. M. Christi, P. G. Finlayson, and J. T. Williams, 1991. Developmental aspects of the locus coeruleus-noradrenaline system. Prog. Brain Res. 88:173– 185. Martin, L. T., and J. R. Alberts, 1979. Taste aversion to mother’s milk: The age-related role of nursing in acquisition and expression of a learned association. J. Comp. Physiol. Psychol. 93:430– 445. McGaugh, J. L., B. Roozendaal, and L. Cahill, 1999. Modulation of memory storage by stress hormones and the amygdaloid complex. In M. Gazzaniga, ed., Cognitive Neuroscience, 2nd ed., 1081–1098. Cambridge, MA: MIT Press.

802

emotion⁄cognition interactions

McLean, J. H., A. Darby-King, R. M. Sullivan, and S. R. King, 1993. Serotonergic influences on olfactory learning in the neonatal rat. Behav. Neural Biol. 60:152–162. McLean, J. H., and M. T. Shipley, 1991. Postnatal development of the noradrenergic projection from the locus coeruleus to the olfactory bulb in the rat. J. Comp. Neurol. 304:467–477. McLean, J. H., M. T. Shipley, W. T. Nickell, G. Ashton-Jones, and C. K. Reyher, 1989. Chemoanatomical organization of the noradrenergic input from locus coeruleus to the olfactory bulb of the adult rat. J. Comp. Neurol. 285:339–349. Melcer, T., J. R. Alberts, and D. J. Gubernick, 1985. Early weaning does not accelerate the expression of nursing-related taste aversions. Dev. Psychobiol. 18:375–381. Mennella, J., A. Johnson, and G. Beauchamp, 1995. Garlic ingestion by pregnant women alters the odor of amniotic fluid. Chem. Senses 20:207–209. Michel, G. F., and C. L. Moore, 1995. Developmental Psychobiology: An Interdisciplinary Science. Cambridge, MA: MIT Press. Mickley, G. A., M. A. Schaldach, K. J. Snyder, S. A. Balagh, T. Len, K. Neimanis, P. Gaulis, J. Hug, K. Sauchak, D. R. Remmers-Roeber, C. Walker, and B. K. Yamamoto, 1998. Ketamine blocks a conditioned taste aversion in neonatal rats. Physiol. Behav. 64:381–390. Miczek, K. A., W. Tornatsky, and J. Vivian, 1991. Ethology and neuropharmacology: Rodent ultrasounds. In Advances in Pharmacologic Sciences, 409–429. Basel: Birkhauser. Miller, J. S., J. A. Jagielo, and N. E. Spear, 1989. Age-related differences in short-term retention of separable elements of an odor aversion. J. Exp. Psychol. 15:194–201. Moore, C. L., L. Jordan, and L. Wong, 1996. Early olfactory experience, novelty and choice of sexual partner by male rats. Physiol. Behav. 60:1361–1367. Moriceau, S., T. L. Roth, T. Okotoghaide, and R. M. Sullivan, 2004. Corticosterone controls the developmental emergence of fear and amygdala function to predator odors in infant rat pups. Int. J. Dev. Neurosci. 22:415–422. Moriceau, S., and R. M. Sullivan, 2004a. Corticosterone influences on mammalian neonatal sensitive period learning. Behav. Neurosci. 118:274–281. Moriceau, S., and R. M. Sullivan, 2004b. Unique neural circuitry for neonatal olfactory learning. J. Neurosci. 24(5):1182– 1189. Moriceau, S., and R. M. Sullivan, 2006. Maternal presence serves as a switch between learning fear and attraction in infancy. Nature Neurosci. 9:1004–1006. Moriceau, S., D. A. Wilson, S. Levine, and R. M. Sullivan, 2006. Dual circuitry for odor-shock conditioning during infancy: Corticosterone switches between fear and attraction via amygdala. J. Neurosci. 26:6737–6748. Morys, J., B. Berdel, H. Jagalska-Majewska, and A. Luczynska, 1999. The basolateral amygdaloid complex—Its development, morphology and functions. Folia Morphol. (Warsz.) 58:29– 46. Myers, M. M., S. A. Brunelli, J. M. Squire, R. Shindledecker, and M. A. Hofer, 1989. Maternal behavior of SHR rats in its relationship to offspring blood pressure. Dev. Psychobiol. 22(1): 29–53. Myslivecek, J., 1997. Inhibitory learning and memory in newborn rats. Prog. Neurobiol. 53:399–430. Nair, H. P., and F. Gonzalez-Lima, 1999. Extinction of behavior in infant rats: Development of functional coupling between septal, hippocampal, and ventral tegmental regions. J. Neurosci. 19:8646–8655.

Nakamura, S. T., and T. Sakaguchi, 1990. Development and plasticity of the locus coeruleus: A review of recent physiological and pharmacological experimentation. Prog. Neurobiol. 34:505– 526. Newman, J. D., ed., 1988. The Physiological Control of Mammalian Vocalization. New York: Plenum Press. Noirot, E., 1968. Ultrasounds in young rodents. II. Changes with age in albino rats. Anim. Behav. 16:129–134. Okutani, F., J. J. Zhang, T. Otsuka, F. Yagi, and H. Kaba, 2003. Modulation of olfactory learning in young rats through intrabulbar GABA(B) receptors. Eur. J. Neurosci. 18(7):2031–2036. Pager, J., 1974. A selective modulation of olfactory bulb electrical activity in relation to the learning of palatability in hungry and satiated rats. Physiol. Behav. 12:189–195. Pape, H. C., and O. Stork, 2003. Genes and mechanisms in the amygdala involved in the formation of fear memory. Ann. NY Acad. Sci. 985:92–105. Pare, D., G. J. Quirk, and J. E. LeDoux, 2004. New vistas on amygdala networks in conditioned fear. J. Neurophysiol. 92:1–9. Pedersen, P., C. L. Williams, and E. M. Blass, 1982. Activation and odor conditioning of suckling behavior in 3-day-old albino rats. J. Exp. Psychol. [Anim. Behav.] 8:329–341. Polan, H. J., and M. A. Hofer, 1998. Olfactory preference for mother over home nest shavings by newborn rats. Dev. Psychobiol. 33:5–20. Polan, H. J., and M. A. Hofer, 1999. Maternally directed orienting behaviors of newborn rats. Dev. Psychobiol. 34:269–280. Rajecki, D. W, M. E. Lamb, and P. Obmascher, 1978. Behav. Brain Sci. 3:417–464. Rangel, S., and M. Leon, 1995. Early odor preference training increases olfactory bulb norepinephrine. Dev. Brain Res. 85: 187–191. Richardson, R., M. A. Siegel, and B. A. Campbell, 1989. Effect of maternal presence on the cardiac and behavioral responses to shock in rats as a function of age. Dev. Psychobiol. 22:567–583. Robinson, S. R., and W. P. Smotherman, 1995. Habituation and classical conditioning in the rat fetus: Opioid involvements. In J.-P. Lecanuet, W. P. Fifer, N. A. Krasnegor, and W. P. Smotherman, eds., Fetal Development—A Psychobiolotgical Perspective, 295–314. Hillsdale, NJ: Lawrence Erlbaum. Rosenfeld, P., Y. A. Gutierrez, A. M. Martin, H. A. Mallett, E. Alleva, and S. Levine, 1991. Maternal regulation of the adrenocortical response in preweanling rats. Physiol. Behav. 50:661–671. Rosenfeld, P., J. B. Wetmore, and S. Levine, 1992. Effects of repeated maternal separations on the adrenocortical response to stress of preweanling rats. Physiol. Behav. 52:787–791. Rosenkranz, J. A., and A. A. Grace, 2002. Dopamine-mediated modulation of odour-evoked amygdala potentials during pavlovian conditioning. Nature 417:282–287. Roth, T. L., and R. M. Sullivan, 2003. Consolidation and expression of a shock-induced odor preference in rat pups is facilitated by opioids. Physiol. Behav. 78(1):135–142. Roth, T. L., and R. M. Sullivan, 2005. Memory of early maltreatment: Neonatal behavioral and neural correlates of maternal maltreatment within the context of classical conditioning. Biol. Psychiatry 57(8):823–831. Roth, T. L., and R. M. Sullivan, 2006. Examining the role of endogenous opioids in learned odor—Stroke associations in infant rats. Dev. Psychobiol. 48:71–78. Roth, T. L., D. A. Wilson, and R. M. Sullivan, 2004. Neurobehavioral development of infant learning and memory: Designed

for attachment. In P. Slater, J. Rosenblatt, C. T. Snowdon, and T. J. Ropers, eds., Adv. Study Behav. 31:103–123. Rudy, J. W., and M. D. Cheatle, 1977. Odor-aversion learning in neonatal rats. Science 198(4319):845–846. Rudy, J. W., and M. D. Cheatle, 1978. A role for conditioned stimulus duration in toxiphobia conditioning. J. Exp. Psychol. [Anim. Behav.] 4(4):399–411. Rudy, J. W., and M. D. Cheatle, 1979. Odor aversion learning in neonatal rats. Science 198:845–846. Rudy, J. W., and P. Morledge, 1994. The ontogeny of contextural fear conditioning: Implications for consolidation, infantile amnesia, and hippocampal system function. Behav. Neurosci. 108: 227–234. Rudy, J. W., S. Stadler-Morris, and P. A. Alberts, 1987. Ontogeny of spatial navigation behaviors in the rat: Dissociation of “proximal-” and “distal-cue” based behaviors. Behav. Neurosci. 101:62–73. Salzen, E. A., 1970. Imprinting and environmental learning. In L. R. Aronson, E. Tobach, D. S. Lehrman, and J. Rosensblatt, eds., Development and Evolution of Behavior. San Francisco: W. H. Freeman. Sananes, C. B., and B. A. Campbell, 1989. Role of the central nucleus of the amygdala in olfactory heart rate conditioning. Behav. Neurosci. 103:519–525. Sanchez, M. M., C. O. Ladd, and P. M. Plotsky, 2001. Early adverse experience as a developmental risk factor for later psychopathology: Evidence from rodent and primate models. Dev. Psychopathol. 13:419–449. Sara, S. J., D. Dyon-Laurent, and A. Herve, 1995. Novelty seeking behavior in the rat is dependent upon the integrity of the of the noradrenergic system. Cogn. Brain Res. 2:181– 187. Schaal, B., L. Marlier, and R. Soussignan, 1995. Responsiveness to the odour of amniotic fluid in the human neonate. Biol. Neonate 67:396–406. Schettino, L. F., and T. Otto, 2001. Patterns of Fos expression in the amygdala and ventral perirhinal cortex induced by training in an olfactory fear conditioning paradigm. Behav. Neurosci. 115:1257–1272. Schore, A. N., 1994. Affect Regulation and the Origin of the Self. Hillsdale, NJ: Erlbaum. Schwob, J. E., L. B. Haberly, and J. L. Price, 1984. The development of physiological responses of the piriform cortex in rats to stimulation of the lateral olfactory tract. J. Comp. Neurol. 223(2): 223–237. Sevelinges, Y., R. Gervais, B. Messaoudi, L. Granjon, and A. M. Mouly, 2004. Olfactory fear conditioning induces field potential potentiation in rat olfactory cortex and amygdala. Learn. Memory 11:761–769. Shair, H. N., G. A. Barr, and M. A. Hofer, eds., 1991. Developmental Psychobiology, New Methods and Changing Concepts. New York: Oxford University Press. Shair, H. N., J. R. Masmela, S. A. Brunelli, and M. A. Hofer, 1997. Potentiation and inhibition of ultrasonic vocalization of rat pups: Regulation by social cues. Dev. Psychobiol. 30:195– 200. Shear, M. K., S. A. Brunelli, and M. A. Hofer, 1983. The effects of maternal deprivation and of refeeding on the blood pressure of infant rats. Psychosom. Med. 45:3–9. Shionoya, K., S. Moriceau, L. Lunday, C. Miner, T. L. Roth, and R. M. Sullivan, 2006. Development switch in neural circuitry underlying odor-malaise learning. Learn. Memory 13:801–808.

hofer and sullivan: neurobiology of attachment

803

Shipley, M. T., F. J. Halloran, and J. De la Torre, 1985. Surprisingly rich projection from locus coeruleus to the olfactory bulb in the rat. Brain Res. 239:294–299. Smotherman, W. P., R. W. Bell, W. A. Hershberger, and G. D. Coover, 1978. Orientation to rat pup cues: Effects of maternal experiential history. Animal Behav. 26:265–273. Smotherman, W. P., and S. R. Robinson, 1986. Environmental determinants of behaviour in the rat fetus. Anim. Behav. 34: 1859–1873. Smotherman, W. P., and S. R. Robinson, 1992. Prenatal experience with milk: Fetal behavior and endogenous opioid systems. Neurosci. Biobehav. Rev. 16:351–364. Spear, N. E., 1978. Processing Memories: Forgetting and Retention. Hillsdale, NJ: Erlbaum. Stanton, M. E., 2000. Multiple memory systems, development and conditioning. Behav. Brain Res. 110:25–37. Stanton, M. E., Y. R. Gutierrez, and S. Levine, 1988. Maternal deprivation potentiates pituitary-adrenal stress responses in infant rats. Behav. Neurosci. 102:692–700. Stanton, M. E., and S. Levine, 1990. Inhibition of infant glucocorticoid stress response: Specific role of maternal cues. Dev. Psychobiol. 23:411–426. Stehouwer, D. J., and B. A. Campbell, 1978. Habituation of the forelimb-withdrawal response in neonatal rats. J. Exp. Psychol. [Anim. Behav.] 4:104–119. Stern, K., and M. K. McClintock, 1998. Regulation of ovulation by human pheromones. Nature 392:177–179. Stickrod, G., D. P. Kimble, and W. P. Smotherman, 1982. In utero taste/odor aversion conditioning in the rat. Physiol. Behav. 28:5–7. Stone, E., K. Bonnet, and M. A. Hofer, 1976. Survival and development of maternally deprived rat pups: Role of body temperature. Psychosom. Med. 38:242–249. Sullivan, R. M., S. C. Brake, M. A. Hofer, and C. L. Williams, 1986. Huddling and independent feeding of neonatal rats can be facilitated by a conditioned change in behavioral state. Dev. Psychobiol. 19:625–635. Sullivan, R. M., M. A. Hofer, and S. C. Brake, 1986. Olfactoryguided orientation in neonatal rats is enhanced by a conditioned change in behavioral state. Dev. Psychobiol. 19:615–623. Sullivan, R. M., M. Landers, J. Fleming, T. Young, and J. Polan, 2003. Characterizing the functional significance of neonatal rat vibrissae. Somatosensory Mot. Res. 20(2):157–162. Sullivan, R. M., and M. Leon, 1986. Early olfactory learning induces an enhanced olfactory bulb response in young rats. Dev. Brain Res. 27:278–282. Sullivan, R. M., G. Stackenwalt, F. Nasr, C. Lemon, and D. A. Wilson, 2000. Association of an odor with activation of olfactory bulb noradrenergic ß-receptors or locus coeruleus stimulation is sufficient to produce learned approach response to that odor in neonatal rats. Behav. Neurosci. 114:957–962. Sullivan, R. M., S. Taborsky-Barba, R. Mendoza, A. Itano, M. Leon, C. Cotman, T. F. Payne, and I. Lott, 1991. Olfactory classical conditioning in neonates. Pediatrics 87:511–518. Sullivan, R. M., and D. A. Wilson, 1995. Dissociation of behavioral and neural correlates of early associative learning. Dev. Psychobiol. 28:213–219. Sullivan, R. M., D. A. Wilson, C. Lemon, and G. A. Gerhardt, 1994. Bilateral 6-OHDA lesions of the locus coeruleus impair associative olfactory learning in newborn rats. Brain Res. 643: 306–309. Sullivan, R. M., D. A. Wilson, R. Wong, A. Correa, and M. Leon, 1990. Modified behavioral olfactory bulb responses

804

emotion⁄cognition interactions

to maternal odors in preweanling rats. Dev. Brain Res. 53:243– 247. Sullivan, R. M., D. Zyzak, P. Skierkowski, and D. A. Wilson, 1992. The role of olfactory bulb norepinephrine in early olfactory learning. Dev. Brain Res. 70:279–282. Suomi, S. J., 1997. Early determinants of behaviour: Evidence from primate studies. Br. Med. Bull. 53(1):170–184. Takahashi, L. K., 1992. Developmental expression of defensive responses during exposure to conspecific adults in preweanling rats (Rattus norvegicus). J. Comp. Psychol. 106:69–77. Terry, L. M., and I. B. Johanson, 1996. Effects of altered olfactory experiences on the development of infant rats’ responses to odors. Dev. Psychobiol. 29:353–377. Touzani, K., and A. Sclafani, 2005. Critical role of amygdala in flavor but not taste preference learning in rats. Eur. J. Neurosci. 22:1767–1774. Trombley, P. Q., and G. M. Shepherd, 1992. Noradrenergic inhibition of synaptic transmission between mitral and granule cells in mammalian olfactory bulb cultures. J. Neurosci. 12:3985– 3991. Vankov, A., A. Herve-Minvielle, and S. J. Sara, 1995. Response to novelty and its rapid habituation in locus coeruleus neurons of the freely exploring rat. Eur. J. Neurosci. 7: 1180–1187. Varendi, H., K. Christenson, R. H. Porter, and J. Winberg, 1998. Early Hum. Dev. 51:47–55. Varendi, H., R. H., Porter, and J. Winberg, 1996. Attractiveness of amniotic fluid odor: Evidence of prenatal olfactory learning? Acta Paediatr. 85:1223–1227. Weaver, I. C. G., N. Cervoni, F. A. Champagne, A. C. D’Alessio, S. Sharma, J. R. Seckl, S. Dymov, M. Szf, and M. J. Meaney, 2004. Epigenetic programming by maternal behavior. Nature Neurosci. 7(8):847–854. Weaver, I. C. G., M. J. Meaney, and M. Szf, 2006. Maternal care effects on the hippocampal transcriptome and anxiety-mediated behaviors in the offspring that are reversible in adulthood. Proc. Natl. Acad. Sci. USA 103(9):3480–3486. Weldon, D. A., M. L., Travis, and D. A. Kennedy, 1991. Posttraining D1 receptor blockade impairs odor conditioning in neonatal rats. Behav. Neurosc. 105:450–458. Wetzel, D. M., D. B. Kelley, and B. A. Campbell, 1980. Central control of ultrasonic vocalizations in neonatal rats. I. Brain stem motor nuclei. J. Comp. Physiol. Psychol. 94:596–605. White, N. R., R. Adox, A. Reddy, and R. J. Barfield, 1992. Regulation of rat maternal behavior by broadband pup vocalizations. Behav. Neural Biol. 58:131–137. Wiedenmayer, C. P., and G. A. Barr, 2001. Developmental changes in c-fos expression to an age-specific social stressor in infant rats. Behav. Brain Res. 126:147–157. Wiedenmayer, C. P., A. M. Magarinos, B. S. McEwen, and G. A. Barr, 2003. Mother lowers glucorticoid levels of preweaning rats after acute threat. Ann. NY Acad. Sci. 1008: 304–307. Wilson, D. A., A. R. Best, and R. M. Sullivan, 2004. Plasticity in the olfactory system: Lessons for the neurobiology of memory. Neuroscientist 10(6):513–524. Wilson, D. A., and M. Leon, 1988. Noradrenergic modulation of olfactory bulb excitability in the postnatal rat. Dev. Brain Res. 42: 69–75. Wilson, D. A., and R. M. Sullivan, 1990. Olfactory associative conditioning in infant rats with brain stimulation as reward. I. Neurobehavioral consequences. Dev. Brain Res. 53:215– 221.

Wilson, D. A., and R. M. Sullivan, 1991. Olfactory associative conditioning in infant rats with brain stimulation as reward. II. Norepinephrine mediates a specific component of the bulb response to reward. Behav. Neurosci. 105:843–849. Wilson, D. A., and R. M. Sullivan, 1992. Blockade of mitral/ tufted cell habituation to odors by association with reward: A preliminary note. Brain Res. 594:143–145. Wilson, D. A., and R. M. Sullivan, 1994. Review: Neurobiology of associative learning in the neonate: Early olfactory learning. Behav. Neural Biol. 61:1–18. Wilson, D. A., R. M. Sullivan, and M. Leon, 1987. Single-unit analysis of postnatal olfactory learning: Modified olfactory bulb output response patterns to learned attractive odors. J. Neurosci. 7:3154–3162. Winzer-Serhan, U. H., H. K. Raymond, R. S. Broide, Y. Chen, and F. M. Leslie, 1997. Expression of alpha 2 adrenoceptors during rat brain development. II. Alpha 2C messenger RNA expression and [3H]rauwolscine binding. Neuroscience 76:261– 272.

Woo, C. C., R. Coopersmith, and M. Leon, 1987. Localized changes in olfactory bulb morphology associated with early olfactory learning. J. Comp. Neurol. 263:113–125. Woo, C. C., and M. Leon, 1988. Sensitive period for neural and behavioral responses to learned odors. Dev. Brain Res. 36:309– 313. Woo, C. C., and M. Leon, 1995. Distribution and development of Beta-adrenergic receptors in the rat olfactory bulb. J. Comp. Neurol. 352:1–10. Yeh, K. Y., 1984. Corticosterone concentration in the serum and milk of lactating rats: Parallel changes after induced stress. Endocrinology 115:1364–1370. Yuan, Q., C. Harley, J. H. McLean, and T. Knopfel, 2002. Optical imaging of odor preference memory in the rat olfactory bulb. J. Neurophysiol. 87:3156–3159. Zimmerberg, B., S. A. Brunelli, A. J. Fluty, and C. A. Frye, 2005. Differences in affective behavior and hippocampal allopregnenolone levels in adult rats of lines selectively bred for infantile vocalizations. Behav. Brain Res. 159(2):301–311.

hofer and sullivan: neurobiology of attachment

805

49

Sleep, Cognition, and Emotion: A Developmental View OSKAR G. JENNI AND RONALD E. DAHL

In the course of the past century, developmentalists have learned a great deal about cognitive and emotional development of children from birth to adulthood (Cole, Cole, and Lightfoot, 2005). The vast majority of this research, however, has focused on the child’s development and behavior during daytime when the child is active and awake. This focus on waking behavior is understandable, yet it ignores the fact that children spend more than half of the first 10 years of their lives asleep (Iglowstein et al., 2003). One important aspect to consider in the light of this relative disregard of sleep in the study of child development is the increasing recognition that sleep and wakefulness are intimately intertwined ( Jenni and LeBourgeois, 2006; Borbély and Achermann, 2005; Borbély, 1982). For example, it is clear that sleep during the night plays an important role in cognitive performance and emotional well-being during the following day, and waking experiences during the day subsequently affect nocturnal sleep. Thus, as will be described in this chapter, a strong case can be made for paying greater attention to sleep and sleep-wake interactions in children’s development—and more specifically, for considering the emerging evidence that sleep processes may play an active role in brain maturation, development, and learning. Since the 1990s an exciting and rapidly growing new field has been established—the cognitive neuroscience of sleep—which has contributed substantially to the understanding of sleep in relation to cognitive processes in adults (Stickgold, 2005; Walker and Stickgold, 2006). A large body of research indicates that certain types of learning may depend on specific characteristics of sleep (Smith, 2001; Maquet, 2001; Walker, 2004; Walker and Stickgold, 2004, 2006). However, relatively limited experimental research has focused on the relationship between sleep, cognition, and emotion in children and adolescents. This developmental dimension of understanding the role of sleep in learning processes is emerging as an area of great interest, which may finally bring us closer to answering one of the most enigmatic questions in bioscience: “Why do we sleep?” This chapter will first briefly introduce basic aspects of sleep organization and regulation during childhood, and then review current evidence for the role of sleep in cognitive and emotional development. In the light of the limited

empirical evidence that is currently available, this chapter will also pose some hypotheses that warrant further confirmation by experimental studies and discuss why these questions may be of central importance with respect not only to understanding basic aspects of human development, but also to how these have important implications for clinical and social policy regarding the importance of children’s and adolescents’ sleep.

Sleep organization and regulation during human development It is important to note that the brain does not reside in a single state throughout the 24-hour period, but rather, the brain actively cycles through periods of different neural activity, associated with distinct behavioral states, most obviously separated into those of wakefulness and sleep. Sleep itself is broadly divided into rapid eye movement (REM) sleep and non-REM (NREM) sleep on the basis of polysomnography, which monitors electroencephalographic (EEG) patterns, eye movements, and muscle tone. Non-REM sleep is characterized by low-frequency, high-voltage EEG activity, low muscle tone, and absence of eye movements. Respiration patterns and heart rate are regular. Non-REM sleep can be divided into four stages on the basis of distinct EEG features. Stage-1 NREM sleep occurs at transitions of sleep and wakefulness. Stage 2 is characterized by frequent bursts of rhythmic EEG activity, so-called sleep spindles (first occurring after 4 weeks of age), and high-voltage slow spikes, so-called K-complexes (first appearing after 6 months). In stages 3 and 4 (slow wave sleep, SWS), the EEG pattern comprises more or less continuous high-voltage activity in the slowest (<2 Hz) frequency range. In contrast, REM sleep is characterized by high levels of desynchronized cortical EEG activity (mixed frequencies, relatively low voltage), absence of muscle tone, irregular heart rate and respiratory patterns, and episodic bursts of phasic eye movements, the hallmark of REM sleep. In the first few months of life, infants’ sleep is divided evenly (50 : 50) between NREM and REM sleep (figure 49.1). The proportion of REM sleep decreases throughout early childhood until it reaches the adolescent and adult

807

Figure 49.1 Upper panel: Schematic representation of the three processes underlying sleep regulation. W, waking; S, sleep; N, NREM sleep; R, REM sleep. The two top diagrams indicate the homeostatic and circadian modulation of sleep pressure across the day and at night. The homeostatic sleep-wake dependent

process is an hourglass process, while the circadian process is a clocklike process (original from Achermann and Borbély, 2003). Lower panel: Sleep duration, NREM sleep and REM sleep as a function of age (redrawn from Roffwarg, Muzio, and Dement, 1966). Picture: Six-week-old infant smiling during REM sleep.

levels of about 20 to 25 percent of nocturnal sleep. When young infants fall asleep, the initial sleep episode is typically REM sleep, that is, sleep-onset REM periods. After 3 months, sleep-onset REM periods become replaced by the adult pattern, that is, sleep-onset NREM periods. The distribution of sleep states is unequal across the night; SWS is most abundant in the first half of the night, whereas the proportion of REM sleep increases in the second half of the night (for an overview, see Jenni and Carskadon, 2007). Current theoretical models suggest that there are two interacting but independent regulatory processes that control

the timing, intensity, and duration of sleep (as described in the two-process model of sleep regulation): one is a homeostatic sleep process, and the other is a circadian sleep process (see Borbély, 1982; Borbély and Achermann, 2005; Borbély et al., 1981; and see figure 49.1). The first regulatory process, sometimes called Process S, represents a sleep-wake-dependent homeostatic component of sleep. Process S increases (or “builds up”) as a function of previous wakefulness, and it gradually decreases (or “dissipates”) over the course of a sleep period. This process accounts for an increase of sleep pressure as waking is

808

emotion⁄cognition interactions

extended and appears to underpin some type of recovery or restorative process that occurs during sleep. The time course of the homeostatic process was initially derived from EEG slow-wave activity (SWA, EEG power in the frequency range 0.75–4.5 Hz) during NREM sleep that increases as a function of previous wakefulness and gradually decreases in the course of sleep. Recent research has indicated that the homeostatic sleep drive may be due to some sleep-promoting substances that accumulate during prolonged wakefulness and are depleted during sleep (such as adenosine or other somnogens; Porkka-Heiskanen et al., 2002, 1997; Kong et al., 2002). Even more interesting from the perspective of developmental neuroscience is recent evidence that sleep homeostasis is involved in aspects of learning and neural plasticity (see later section “The synaptic homeostasis hypothesis of sleep” (Tononi and Cirelli, 2003, 2006). The homeostatic process interacts with the other, sleepwake-independent, circadian process. This clocklike aspect of sleep regulation is often called Process C, and under normal conditions it is entrained to the light-dark cycle. The underlying circadian mechanisms have a distinct neuroanatomical locus, and various molecular components have been identified (Carskadon, Acebo, and Jenni, 2004; Albrecht and Eichele, 2003). More recently, neural mechanisms that regulate the ability for sleep and wakefulness have been described (Saper, Cano, and Scammell, 2005; Saper, Scammell, and Lu, 2005). These studies show that the control of wake and sleep emerges from the interaction of cell groups that cause arousal with other nuclei that induce sleep such as the ventrolateral preoptic nucleus in the hypothalamus. From a developmental perspective, both the homeostatic and the circadian process undergo significant maturational changes ( Jenni and LeBourgeois, 2006). For example, the circadian phase marker melatonin correlates with pubertal stage such that more mature adolescents show a delay in the melatonin excretion pattern even under conditions controlling for psychosocial influences on sleep-wake patterns (Carskadon, Acebo, and Jenni, 2004). This correlation appears to contribute to the tendency to stay up later and sleep in later among adolescents. Furthermore, evidence is accumulating that the dynamics of sleep homeostatic processes slow down in the course of childhood (i.e., sleep pressure accumulates more slowly with increasing age) enabling children to be awake for consolidated periods during the day ( Jenni, Achermann, and Carskadon, 2005; Jenni and LeBourgeois, 2006).

The function(s) of sleep Despite the growing understanding of the mechanisms generating and maintaining sleep, the specific function(s) of

sleep are still not clear. Among a number of hypotheses, two theories (not mutually exclusive) have dominated the field: first, sleep is involved in a restorative metabolic function for the brain (Benington and Heller, 1995), and second, sleep plays a fundamental role in memory consolidation and learning (Stickgold, 2005; Walker, 2004; Walker and Stickgold, 2004, 2006; Peigneux et al., 2001; Hobson and Pace-Schott, 2002). The latter hypothesis implies that in some way sleep facilitates brain plasticity—the brain’s ability to modify cerebral structures and functions depending not only upon genetic information but also on the effects of individual experience and environmental inputs. This view of sleep has long been supported by studies showing impaired memory consolidation following episodes with sleep loss (Pilcher and Huffcut, 1996), and recently there have been findings suggesting such a link on the cellular and molecular level (Benington and Frank, 2003; Steriade, 1999; Sejnowski and Destexhe, 2000; Tononi and Cirelli, 2003, 2006; Graves, Pack, and Abel, 2001). The initial studies investigating the relationship between sleep and memory examined the influence of sleep on posttraining consolidation. Recently, data also have been presented on initial memory encoding (Walker and Stickgold, 2006). Most relevant to the focus on developmental dimensions of learning, a recent study reported preliminary evidence for the role of sleep in learning in infants (Gomez, Bootzin, and Nadel, 2006). It is important to note, however, that some researchers have also expressed skepticism (Vertes, 2004; Vertes and Siegel, 2005) about the role of sleep in memory consolidation and learning (but see also reply by Stickgold and Walker, 2005b). In fact, the potential purpose of sleep in the processing of learned facts should be viewed on multiple levels at the same time (Maquet, 2001), because it is possible that sleep functions are not universal for all living organisms and across the life span, but may be limited to particular species or developmental periods (Siegel, 2005). For example, multidimensional levels include that (1) sleep may be needed for brain maturation in early life and for cognitive processes that involve learning and adaptation in adults, (2) different sleep states may contribute in different ways to specific types of learning processes and memory consolidation, and finally, (3) experimental manipulations of sleep (such as sleep deprivation or restriction) may alter the performance of learned tasks, and exposure to a new environment may change subsequent sleep. Over the past decades, a large number of studies have been published on these multidimensional levels leading to distinct theories about the functions of sleep. Relatively few of these have focused specifically on the role of sleep in the developing brain, despite the fact, as described earlier, that a strong case can be made for sleep that may be particularly important during periods of maturation.

jenni and dahl: sleep, cognition, and emotion

809

Differential effects of sleep stages on brain development REM Sleep In 1966, Roffwarg, Muzio, and Dement published their seminal paper “The Ontogenetic Development of the Human Sleep-Dream Cycle” and hypothesized that the high neuronal activity levels during REM sleep may serve as an intrinsic stimulation of the brain at a period when exogenous sensory input is very low. They based their view on the large amount of high-frequency, low-amplitude REM sleep in neonates that sharply declines as a function of age, reflecting less need for intrinsic activation of the brain as extrinsic activation increased during development. The authors concluded that REM sleep plays a primary role in the growth and development of the brain. Over the past 40 years, two kinds of experiments have shown preliminary evidence for the role of REM sleep in neural development, although the exact mechanisms remain unclear (Mirmiran and Ariagno, 2003; Dang-Vu et al., 2006; Mirmiran, 1995; Mirmiran and Van Someren, 1993). Correlative studies have demonstrated that REM sleep duration is associated with the degree of brain maturity at birth (Zepelin, Siegel, and Tobler, 2005; Siegel, 2005). In the rat (an altrical mammal born in a relatively immature state) REM sleep amount is high at birth; in precocial mammals (with a relatively completed neural development at birth, e.g., in lambs, giraffes, baboons, guinea pigs, etc.) REM sleep shows low levels at birth. In humans, the diminution of REM sleep amount is slow in the course of the first few months of life, and low levels are reached only in the preschool period (see figure 49.1) ( Jenni, Borbély, and Achermann, 2004; Louis et al., 1997). In fact, these findings suggest that REM sleep reaches adult levels when the most rapid period of brain maturation is completed. Other investigations in rats have shown that enriched play environments are associated with increased synaptic densities and elevated REM sleep amount (Mirmiran, van den Dungen, and Uylings, 1982). Furthermore, phasic REM sleep activity (such as ponto-geniculooccipital [PGO] waves recorded in the visual system of young cats) is also considered as an intrinsic activation of the visual system during periods without visual waking input (Oksenberg et al., 1996; Pompeiano, Pompeiano, and Corvaja, 1995). Another approach to studying the potential function of REM sleep during development is to selectively deprive infants of REM sleep and then follow their subsequent development (however, this method is challenging because of the difficulties of depriving infants of sleep due to high sleep pressure early in life). Some studies have used pharmacological agents that are known to suppress REM sleep (e.g., clonidine, clomipramine). In brief, neonatal REM sleep deprivation induced substantial changes in behavior,

810

emotion⁄cognition interactions

brain morphology, and neurotransmitter circuitries (Shaffery et al., 1998, 1999, 2002). Notably, these findings should be used with caution because of multiple effects of the drugs and stress on the developing brain. Non-REM Sleep Other researchers have suggested that NREM sleep is equally important for neural development because of its homeostatic regulation (Frank, Morrissette, and Heller, 1998; Jenni, Borbély, and Achermann, 2004; Frank and Stryker, 2003). Endogenous neural activity during NREM sleep is highly synchronized and may facilitate the shaping of connections within the thalamocortical system— the neural circuitry that generates EEG slow waves (Steriade, McCormick, and Sejnowski, 1993). In fact, the increase of SWS during NREM sleep in the first few years after birth parallels the period of intense synaptogenesis in the cerebral cortex (Feinberg et al., 1990). Later, during adolescence, the SWS decrease may be closely associated with the decline in synaptic connectivity that occurs at the transition from childhood to adulthood (Huttenlocher and Dabholkar, 1997; Jenni and Carskadon, 2004). Direct evidence for the association of NREM sleep and neural development comes from the study of Frank, Issa, and Stryker, who showed that experience-dependent cortical plasticity in the visual system of developing kittens is highly correlated with NREM sleep time and intensity (Frank, Issa, and Stryker, 2001). Another study demonstrated that EEG activity during NREM sleep changes as a function of the sensory experience during a late critical period in cats and mice (Miyamoto, Katagiri, and Hensch, 2003). Altogether, these findings may indicate that NREM sleep consolidates waking experiences during critical periods of brain development (Frank and Stryker, 2003).

Sleep patterns and neurocognitive functioning in children: Statelike effects If sleep does play a critical role in cognitive development and learning, then disturbances of sleep, sleep restriction, or even total sleep loss are likely to impair these functions. Evidence for such statelike effects on neurocognitive and behavioral functioning in children and adolescents comes from three different lines of research: (1) from the “naturalistic” correlation between abnormal sleep patterns and cognitive functioning in representative populations, (2) from the study of cognitive performance of children with specific sleep disorders such as sleep-disordered breathing, and (3) from studies examining the effects of experimental manipulation of sleep on subsequent cognitive functioning and behavior. A link between sleep patterns and daytime functioning was provided by Wolfson and Carskadon, who examined 3,000 high school students showing higher grades with more sleep and earlier bedtimes (Wolfson and Carskadon, 1998). These findings were confirmed by several studies that

indicate poorer school performance and more difficulties in concentrating with less sleep, low sleep quality, frequent nighttime awakenings, and earlier school start times (Carskadon et al., 1998; Epstein, Chillag, and Lavie, 1998; Meijer, Habekothe, and Van Den Wittenboer, 2000; Wolfson and Carskadon, 2003). Sadeh, Raviv, and Gruber were the first to use an objective measure for evaluating sleep patterns. They found that sleep fragmentation assessed by actigraphy was significantly correlated to daytime sleepiness, attentional deficits, and learning impairment (Sadeh, Raviv, and Gruber, 2000). Moreover, auditory and visual working memory was affected in children with lower sleep efficiency and longer sleep latency as assessed by actigraphy (Steenari et al., 2003). Sleep quality and sleep fragmentation appear to play a crucial role in performance decrements of sustained attention, concentration, memory, and learning in children. These effects are even more evident in younger children, who may be more vulnerable to insufficient sleep (Sadeh, Gruber, and Raviv, 2002). In sum, daytime sleepiness and poor quality of sleep can substantially impair cognitive functioning and behavioral performance in children. Another line of evidence for the link between sleep and cognition during childhood comes from the study of sleepdisordered breathing, which is known to be related to a wide variety of neurocognitive deficits and behavioral disorders, reduced academic achievement, and impairments of learning and memory (Halbower and Mahone, 2006; Gozal, 1998; O’Brien and Gozal, 2004; Rhodes et al., 1995; Kaemingk et al., 2003). Sleep-disordered breathing is considered a whole spectrum of disorders, which range from primary snoring to severe obstructive sleep apnea syndrome. A recent population-based study showed lower performance in memory processes, executive function, and general intelligence in children with sleep-disordered breathing compared to healthy controls (Gottlieb, Chase, and Vezina, 2004). Other studies even demonstrated deficits in general intelligence, language, memory, and visuospatial skills in children with only primary snoring (Kennedy, Blunden, and Hirte, 2004; O’Brien et al., 2004). These neurocognitive deficits may be a direct consequence of sleep fragmentation and/or disruption of ventilation leading to hypoxemia (O’Brien and Gozal, 2004). It remains unclear, however, whether there are critical periods of neural plasticity in the rapidly developing brain during which sleep-disordered breathing causes long-term and irreversible deficits in cognition and learning. Notably, the previously mentioned research was correlative in nature and did not provide a direct and causal relationship between sleep and cognition during development. Some studies have tried to experimentally manipulate sleep amount in order to evaluate the effects of subsequent cognitive functioning (review in Curcio, Ferrara, and De Gennero,

2006). The first study examining the effects of complete sleep loss (i.e., 38 hours of sustained wakefulness) on cognitive performance was carried out by Carskadon, Harvey, and Dement (1981). They found an increase in sleepiness and a decrease in memory and problem-solving abilities. In contrast, however, the same research group found no significant impairments in neurobehavior and cognition when sleep was transiently restricted to 4 hours of sleep (Carskadon, Harvey, and Dement, 1981). These authors concluded that total sleep deprivation always leads to observable deficits in neurocognition, while sleep restriction does not necessarily do so. This conclusion was challenged by a study of Randazzo and coworkers, who found impaired psychomotor and cognitive performance in a sleeprestriction paradigm (5 hours for a single night; Randazzo et al., 1998). More recently, a particularly interesting report was published by Sadeh, Gruber, and Raviv, who demonstrated that even a small change in sleep duration for three consecutive nights can have a measurable impact on daytime performance in children. They experimentally manipulated bedtime on school nights by one hour (subjects were randomized to either going to bed an hour earlier or an hour later). They found distinct performance patterns with improvement of attention and memory in the group who increased sleep duration compared to those with an hour less sleep on three consecutive nights (Sadeh, Gruber, and Raviv, 2003). It is important to note, however, that despite numerous studies showing an association between sleep and cognition under short-term experimental conditions, some of the most important clinical and social policy questions remain open: What are the long-term effects of impaired sleep on neurocognitive development? Moreover, are there specific developmental time windows (critical periods) that are particularly sensitive to sleep disruption and consequently have important longterm effects on cognitive development?

Sleep patterns and developmental outcome: Traitlike effects We know that sleep patterns show substantial interindividual variability (between subjects) and intraindividual stability (within subjects across time), indicating traitlike individual characteristics (Iglowstein et al., 2003; Buckelmuller et al., 2006; Finelli, Achermann, and Borbély, 2001; Jenni et al., 2007). In fact, a growing body of evidence suggests that sleep varies among individuals because of biological differences in sleep (i.e., in neurological processes; Aeschbach et al., 2003; Saper, Cano, and Scammell, 2005) and/ or genetic mechanisms (Tafti, Maret, and Danvilliers, 2005). Given this traitlike sleep stability and the potential relationship between sleep and cognition, the question arises whether sleep characteristics may be used to predict

jenni and dahl: sleep, cognition, and emotion

811

(cognitive) development. Some authors have suggested sleep as “a window on the developing brain” (Kohyama, 1998; Scher, 1998). In fact, sleep represents an ideal state to study the developing brain because it minimizes possible confounding factors related to waking activities, including changes in the level of attention and distractibility, and issues of motivation or cognitive capacity. Along these lines, several earlier studies have examined the predictive validity of sleepwake patterns during infancy and childhood (see overview, Halpern, MacLean, and Baumeister, 1995). For example, Richards and others studied aspects of sleep-state organization and EEG patterns at term and at age 3 months in terms of developmental outcome at age 5 years and found that specific EEG patterns were predictors of later outcome (Richards, Parmalee, and Beckwith, 1986). Beckwith and Parmelee confirmed these findings in a prospective study including preterm and term infants (Beckwith and Parmelee, 1986). They showed that quantitative EEG patterns of NREM sleep were related to general intelligence at age 8 years. Becker and Thoman used a prospective-longitudinal design and recorded rapid eye movement bursts during REM sleep (Becker and Thoman, 1981). They reported a close relationship between rapid eye movements and later cognitive functioning. Overall, sleep characteristics appear to be closely related to later developmental outcome, supporting the view of a close sleep/cognition relationship. These findings, however, have not been proved to be robust enough for clinical practice. Whether sleep patterns will serve in future as a reliable predictor for identifying infants or children whose development will later be compromised is unsure, because the influence of the child’s environment on final developmental outcome is well recognized.

The synaptic homeostasis hypothesis of sleep: A model for understanding the function of sleep during development? The recently proposed synaptic homeostasis hypothesis of sleep offers an explanation of the close relationship between sleep, cognitive processes, and brain plasticity (Tononi and Cirelli, 2003, 2006) and may also provide a model for the sleep function during development. An appealing feature of the hypothesis is that it reconciles the restorative and homeostatic function of sleep with its beneficial effects on learning and memory. According to this hypothesis, the homeostatic regulation of SWA is directly related to the amount of synaptic potentiation that occurs during wakefulness (Tononi and Cirelli, 2003, 2006). Long-term synaptic potentiation (LTP) is considered the major candidate mechanism underlying learning and cognitive functions (Bliss and Lomo, 1973; Whitlock et al., 2006). The synaptic homeostasis hypothesis predicts that the progressive increase in synaptic strength of neural circuits from daytime activity must be “downscaled” during

812

emotion⁄cognition interactions

sleep to maintain synaptic homeostasis. This downscaling process would be responsible for keeping the overall synaptic weight in balance and may represent a key mechanism for the modulation of cognitive functioning (Turrigiano and Nelson, 2000). Other predictions of the synaptic homeostasis hypothesis of sleep were supported by findings that a visuomotor task involving a circumscribed cortical region produced a local increase in SWA during subsequent sleep (Huber et al., 2004). Conversely, locally decreased SWA was demonstrated following synaptic depression by arm immobilization (Huber et al., 2006). The model may also be useful for understanding the link between sleep and cognition during development. Recent analyses of the Zurich Longitudinal Studies—some of the largest and most complete birth-to-maturity cohorts on child development ever collected (Iglowstein et al., 2003; Jenni, Zinggeler, et al., 2005; Tanner, 1998)—have shown a consistent negative relationship between sleep amount and cognition (measured by psychometric intelligence assessments) in a population of normal healthy children (unpublished observation, Jenni). Individuals with a traitlike short sleep duration showed higher cognitive capacities than those with a long sleep duration (the effect was independent of social class). Are habitual short sleepers in fact more intelligent than habitual long sleepers? Several theories about the neural bases of cognition have suggested that individual differences in information processing and neural functioning may explain individual differences in intelligence (see for example “the neural efficiency hypothesis”—Duncan et al., 2000; Gray and Thompson, 2004). One hypothesis that may fit these data focuses on the concept that the brain of some individuals may function more efficiently during the day (contributing to higher measures of intelligence) and also at night (producing more efficient and therefore “shorter” sleep). Traitlike individual differences of sleep amount are known to be related to differences in sleep-pressure tolerance (Aeschbach et al., 1996). Thus sleep pressure in children with higher cognitive abilities may accumulate and dissipate faster than in children with lower capacities. As a consequence, children with higher mental abilities may potentiate synapses more strongly during wakefulness and downscale them faster during sleep than children do with lower capacities. In other words, traitlike individual differences in habitual sleep duration and individual variation in cognitive capacities may share common neurobiological underpinnings. These predictions, however, warrant further empirical studies. Nevertheless, a traitlike relationship between sleep and cognition on the basis of the synaptic homeostasis hypothesis may offer a compelling explanation for the function of sleep during childhood—to facilitate brain development, learning, and memory. Indeed, the homeostatic aspect of sleep is particularly strong early in human development (e.g., the

amount of SWA, see Jenni, Achermann, and Carskadon, 2005; Jenni and Carskadon, 2004)—at a time period in life when learning processes are at their peak. Recent evidence shows that the maturational changes of sleep homeostasis coincide with periods of heightened cortical plasticity (Frank, Issa, and Stryker, 2001). Thus it may be more than a coincidence that sleep homeostasis is only fully developed after the process of synapse elimination that occurs during adolescence (Feinberg et al., 1990; Jenni, Achermann, and Carskadon, 2005).

Sleep, learning, and the developing brain: Relevance to affective, clinical, and social policy domains? As described previously, advancing knowledge of the neuroscience of sleep-wake regulation is providing a growing understanding of the role of sleep in learning and memory (Hobson and Pace-Schott, 2002; Peigneux et al., 2001; Stickgold, 2005; Stickgold and Walker, 2005a; Walker and Stickgold, 2004). More broadly, emerging data are consistent with the hypothesis that sleep facilitates brain plasticity—the ability to constantly modify cerebral structures and functions depending on traitlike genetic information, specific experience, and statelike environmental inputs. These models of the role of sleep in learning and brain maturation have relevance not only for cognitive processes and explicit learning (in ways that should be of central importance to education policies for youth emphasizing the importance of sleep), but also for understanding implicit learning processes, including emotional learning. As was the case with the cognitive neuroscience of sleep, the vast majority of the existing research has focused on adults. However, there are several reasons to believe that the affective neuroscience of sleep may also be particularly important to consider from a developmental perspective. That is, the same issues of plasticity, learning, and critical periods of maturation have great salience for understanding the role of sleep in early social and emotional development in children (see Dahl, 1996; Dahl and Harvey, in press).

Sleep and emotion regulation Over the past decade there has been increasing interest in understanding the bidirectional relationships between sleep disturbances and problems with emotion regulation in children and adolescents. There is growing evidence that problems with sleep can create and/or exacerbate emotional problems and that emotional difficulties can interfere with sleep. These bidirectional interactions (and the potential for a vicious cycle of negative effects) may have particular clinical relevance in at least two clinical domains, affective disorders (including anxiety, depression, and bipolar disorders in youth—Dahl, 1996; Harvey, Mullin, and Hinshaw, 2006)

and clinical problems with aggression, anger, and impulse control (Haynes et al., 2006). With respect to sleep and mood, evidence from the sleep deprivation literature suggests that one of the strongest adverse effects of sleep deprivation is increased negative mood (Dinges et al., 1997; Pilcher and Huffcut, 1996; Van Dongen et al., 2003). This suggestion makes sense given the evidence suggesting that REM sleep has an emotionalprocessing and mood-regulation function (Stickgold et al., 2001). Perhaps not surprisingly, in adults with bipolar disorder the case for sleep contributing to mood disturbance is fairly compelling: sleep loss is highly correlated with daily manic symptoms (Barbini et al., 1996); among patients with bipolar disorder sleep disturbance is the most common prodrome of mania and the sixth most common prodrome of depression ( Jackson, Cavanagh, and Scott, 2003); and induced sleep deprivation triggers hypomania or mania in a proportion of patients (Colombo et al., 1999; Kasper and Wehr, 1992; Leibenluft et al., 1996; Wehr, Sack, and Rosenthal, 1987; Wu and Bunney, 1990; Zwi, ShaweTaylor, and Murray, 2005). The clinical study of the sleep of children and adolescents with bipolar disorder represents a relatively new domain of investigation. Rates of significant sleep disturbance in youth diagnosed with bipolar disorder range from 35 to 45 percent, and samples of children with a parent with an affective illness (a group at high risk) also display severe and persistent sleep problems (see Harvey, Mullin, and Hinshaw, 2006). However, in these younger samples research and clinical questions relating to the possible link between sleep disturbance and mood are yet to be addressed. With respect to sleep and aggression, there are both human and animal data showing increased aggression and impulsivity following experimental sleep loss. Rats, for example, show increases in aggression and defensive fighting after sleep deprivation. One recent study found that animals who were easy to handle at baseline became irritable and aggressive following modest amounts of sleep deprivation, with evidence of related changes in synaptic plasticity associated with these behavioral changes (Marks and Wayner, 2005). Clinical studies of children and adolescents also have revealed associations between sleep deprivation and irritability/aggression and difficulties with self-regulation in youth (Chervin et al., 2003; Dahl, Pelham, and Weirson, 1991; Ireland and Culpin, 2006). Most relevant is a recent study by Haynes and colleagues (2006) which examined behavioral and emotional changes in adolescents with substancerelated difficulties undergoing a behavioral sleep treatment. This study reported that improvements in sleep time were associated with significant decreases in the reporting of aggressive thoughts and actions. Taken together, these data suggest that inadequate sleep in adolescence may contribute to aggressive thoughts and actions and that increased or

jenni and dahl: sleep, cognition, and emotion

813

improved sleep may reduce problematic aggression at least in some cases.

Caveats, conclusions, and compelling questions On one hand, it is important to acknowledge again the preliminary nature of the developmental data specifically addressing sleep processes in relation to cognitive and affective neuroscience. The majority of animal and human research supporting the key elements of this conceptual model of sleep have focused on studies of adults. Meanwhile, the vast majority of developmental studies of children and adolescents have focused almost exclusively on waking behavior. On the other hand, we have put forth what we believe is a compelling case for research to bridge these gaps. There is a need for developmental researchers in several areas of cognitive and affective neuroscience to devote increased attention and consideration to sleep, as well as a need for sleep researchers to consider the crucial dimension of developmental research. We have also pointed to the importance of conceptual models such as the synaptic homeostasis hypothesis of sleep, including specific testable views regarding the role of sleep and brain development. The questions about sleep, learning, brain plasticity, and their relevance to cognitive and affective neuroscience not only are compelling scientific issues that may shed light on crucial mysteries about the basic function of sleep, but are also highly important issues of relevance to clinical and social policy. Insights into the importance of sleep in learning, memory, emotion regulation, and social development represent a broad range of issues with great impact on education, as well as physical and emotional health in youth. These compelling questions are made even more urgent by recent evidence indicating that many youngsters in modern society may be obtaining insufficient sleep or may suffer from sleep disturbances. Empirical data, together with scientific progress in elucidating the specific roles of sleep in particular types of learning processes and addressing the issues about sensitive periods of development when sleep may be particularly important, are needed to inform clinical and social policies in these important areas. acknowledgments

The research of Dr. Jenni was supported by the Swiss National Science Foundation. Dr. Dahl has been supported by the National Institute of Mental Health grants K02MH00132 and R24MH067346. REFERENCES

Achermann, P., and A. A. Borbély, 2003. Mathematical models of sleep regulation. Front. Biosci. 8:S683–693. Aeschbach, D., C. Cajochen, H. Landolt, and A. A. Borbély, 1996. Homeostatic sleep regulation in habitual short sleepers and long sleepers. Am. J. Physiol. 270(1 Part 2):R41–53.

814

emotion⁄cognition interactions

Aeschbach, D., L. Sher, T. T. Postolache, J. R. Matthews, M. A. Jackson, and T. A. Wehr, 2003. A longer biological night in long sleepers than in short sleepers. J. Clin. Endocrinol. Metab. 88(1):26–30. Albrecht, U., and G. Eichele, 2003. The mammalian circadian clock. Curr. Opin. Genetics Dev. 13(3):271–277. Barbini, B., S. Bertelli, C. Colombo, and E. Smeraldi, 1996. Sleep loss, a possible factor in augmenting manic episode. Psychiatry Res. 65:121–125. Becker, P. T., and E. B. Thoman, 1981. Rapid eye movement storms in infants: Rate of occurrence at 6 months predicts mental development at 1 year. Science 212(4501):1415–1416. Beckwith, L., and A. H. Parmelee, Jr., 1986. EEG patterns of preterm infants, home environment, and later IQ. Child Dev. 57(3):777–789. Benington, J. H., and M. G. Frank, 2003. Cellular and molecular connections between sleep and synaptic plasticity. Prog. Neurobiol. 69(2):71–101. Benington, J. H., and H. C. Heller, 1995. Restoration of brain energy metabolism as the function of sleep. Prog. Neurobiol. 45(4): 347–360. Bliss, T. V., and T. Lomo, 1973. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol. 232(2):331–356. Borbély, A. A., 1982. A two process model of sleep regulation. Hum. Neurobiol. 1(3):195–204. Borbély, A. A., and P. Achermann, 2005. Sleep homeostasis and models of sleep regulation. In M. H. Kryger, T. Roth, and W. C. Dement, eds., Principles and Practice of Sleep Medicine, 3rd ed., 405–417. Philadelphia: Elsevier Saunders. Borbély, A. A., F. Baumann, D. Brandeis, I. Strauch, and D. Lehmann, 1981. Sleep deprivation: Effect on sleep stages and EEG power density in man. Electroencephalogr. Clin. Neurophysiol. 51(5):483–495. Buckelmuller, J., H. P. Landolt, H. H. Stassen, and P. Achermann, 2006. Trait-like individual differences in the human sleep electroencephalogram. Neuroscience 138(1):351–356. Carskadon, M. A., C. Acebo, and O. G. Jenni, 2004. Regulation of adolescent sleep: Implications for behavior. Ann. NY Acad. Sci. 1021:276–291. Carskadon, M. A., K. Harvey, and W. C. Dement, 1981. Sleep loss in young adolescents. Sleep 4(3):299–312. Carskadon, M. A., A. R. Wolfson, C. Acebo, O. Tzischinsky, and R. Seifer, 1998. Adolescent sleep patterns, circadian timing, and sleepiness at a transition to early school days. Sleep 21(8): 871–881. Chervin, R. D., J. E. Dillon, K. H. Archbold, and D. L. Ruzicka, 2003. Conduct problems and symptoms of sleep disorders in children. J. Am. Acad. Child Adolesc. Psychiatry 42(2):201–208. Cole, M., S. R. Cole, and C. Lightfoot, 2005. The Development of Children. New York: Worth. Colombo, C., F. Benedetti, B. Barbini, E. Campori, and E. Smeraldi, 1999. Rate of switch from depression into mania after therapeutic sleep deprivation in bipolar depression. Psychiatry Res. 86:267–270. Curcio, G., M. Ferrara, and L. De Gennaro, 2006. Sleep loss, learning and academic performance. Sleep Med. Rev. 10(5):323– 337. Dahl, R. E., 1996. The regulation of sleep and arousal: Development and psychopathology. Dev. Psychopathol. 8:3–27. Dahl, R. E., and A. G. Harvey, in press. Sleep disorders. Rutter’s Child and Adolescent Psychiatry, 5th ed. New York: Wiley-Blackwell.

Dahl, R. E., W. E. Pelham, and M. Wierson, 1991. The role of sleep disturbances in attention deficit disorder symptoms: A case study. J. Pediatr. Psychol. 16(2):229–239. Dang-Vu, T. T., M. Desseilles, P. Peigneux, and P. Maquet, 2006. A role for sleep in brain plasticity. Pediatr. Rehabil. 9(2):98– 118. Dinges, D. F., F. Pack, K. Williams, K. A. Gillen, J. W. Powell, G. E. Ott, et al., 1997. Cumulative sleepiness, mood disturbance, and psychomotor vigilance performance decrements during a week of sleep restricted to 4–5 hours per night. Sleep 20(4):267–267. Duncan, J., R. J. Seitz, J. Kolodny, D. Bor, H. Herzog, A. Ahmed, et al., 2000. A neural basis for general intelligence. Science 289(5478):457–460. Epstein, R., N. Chillag, and P. Lavie, 1998. Starting times of school: Effects on daytime functioning of fifth-grade children in israel. Sleep 21(3):250–256. Feinberg, I., H. C. Thode, Jr., H. T. Chugani, and J. D. March, 1990. Gamma distribution model describes maturational curves for delta wave amplitude, cortical metabolic rate and synaptic density. J. Theor. Biol. 142(2):149–161. Finelli, L. A., P. Achermann, and A. A. Borbély, 2001. Individual “fingerprints” in human sleep EEG topography. Neuropsychopharmacology 25(5 Suppl.):S57–62. Frank, M. G., N. P. Issa, and M. P. Stryker, 2001. Sleep enhances plasticity in the developing visual cortex. Neuron 30(1):275–287. Frank, M. G., R. Morrissette, and H. C. Heller, 1998. Effects of sleep deprivation in neonatal rats. Am. J. Physiol. 275(1 Part 2):R148–157. Frank, M. G., and M. P. Stryker, 2003. The role of sleep in the development of central visual pathways. In P. Maquet, C. Smith, and R. Stickgold, eds., Sleep and Brain Plasticity, 189–206. Oxford, UK: Oxford University Press. Gomez, R. L., R. R. Bootzin, and L. Nadel, 2006. Naps promote abstraction in language learning infants. Psychol. Sci. 39(2):158– 170. Gottlieb, D. J., C. Chase, and R. M. Vezina, 2004. Sleepdisordered breathing symptoms are associated with poorer cognitve function in 5-year-old children. J. Pediatr. 145:458–464. Gozal, D., 1998. Sleep-disordered breathing and school performance in children. Pediatrics 102(3 Part 1):616–620. Graves, L., A. Pack, and T. Abel, 2001. Sleep and memory: A molecular perspective. Trends Neurosci. 24(4):237–243. Gray, J. R., and P. M. Thompson, 2004. Neurobiology of intelligence: Science and ethics. Nature Rev. Neurosci. 5(6):471– 482. Halbower, A. C., and E. M. Mahone, 2006. Neuropsychological morbidity linked to childhhod sleep-disordered breathing. Sleep Med. Rev. 10:97–107. Halpern, L. F., W. E. MacLean, and A. A. Baumeister, 1995. Infant sleep-wake characteristics: Relation to neurological status and the prediction of developmental outcome. Dev. Rev. 15:255– 291. Harvey, A. G., B. C. Mullin, and S. P. Hinshaw, 2006. Sleep and circadian rhythms in children and adolescents with bipolar disorder. Dev. Psychopathol. 18:1147–1168. Haynes, P. L., R. R. Bootzin, L. Smith, J. Cousins, M. Cameron, and S. Stevens, 2006. Sleep and aggression in substanceabusing adolescents: Results from an integrative behavioral sleep-treatment pilot program. Sleep 29:512–520. Hobson, J. A., and E. F. Pace-Schott, 2002. The cognitive neuroscience of sleep: Neuronal systems, consciousness, and learning. Nature Rev. Neurosci. 3:679–693.

Huber, R., M. F. Ghilardi, M. Massimini, F. Ferrarelli, B. A. Riedner, M. J. Peterson, et al., 2006. Arm immobilization causes cortical plastic changes and locally decreases sleep slow wave activity. Nature Neurosci. 9(9):1169–1176. Huber, R., M. F. Ghilardi, M. Massimini, and G. Tononi, 2004. Local sleep and learning. Nature 430(6995):78–81. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387(2):167–178. Iglowstein, I., O. G. Jenni, L. Molinari, and R. H. Largo, 2003. Sleep duration from infancy to adolescence: Reference values and generational trends. Pediatrics 111(2):302–307. Ireland, J. L., and V. Culpin, 2006. The relationship between sleeping problems and aggression, anger, and impulsivity in a population of juvenile and young offenders. J. Adolesc. Health 38(6):649–655. Jackson, A., J. Cavanagh, and J. Scott, 2003. A systematic review of manic and depressive prodromes. J. Affect. Disord. 74:209– 217. Jenni, O. G., P. Achermann, and M. A. Carskadon, 2005. Homeostatic sleep regulation in adolescents. Sleep 28(11):1446–1454. Jenni, O. G., A. A. Borbély, and P. Achermann, 2004. Development of the nocturnal sleep electroencephalogram in human infants. Am. J. Physiol. 286:R528–538. Jenni, O. G., J. A. Caflisch, L. Molinari, and R. H. Largo, 2007. Sleep duration from age 1 to 10 years: Variability, stability and comparison with growth. Pediatrics 120(4):e769–776. Jenni, O. G., and M. A. Carskadon, 2004. Spectral analysis of the sleep electroencephalogram during adolescence. Sleep 27(4): 774–783. Jenni, O. G., and M. A. Carskadon, 2007. Sleep behavior and sleep regulation from infancy through adolescence: Normative aspects. In O. G. Jenni and M. A. Carskadon, eds., Sleep Medicine Clinics of North America: Children and Adolescents, 2(3):321–329. Philadelphia: Elsevier. Jenni, O. G., and M. K. LeBourgeois, 2006. Understanding sleep-wake regulation and sleep disorders during childhood: The value of a model. Curr. Opin. Psychiatry 19(3):282–287. Jenni, O. G., H. Zinggeler Fuhrer, I. Iglowstein, L. Molinari, and R. H. Largo, 2005. A longitudinal study of bedsharing and sleep problems among Swiss children in the first 10 years of life. Pediatrics 115:233–240. Kaemingk, K. L., A. E. Pasvogel, J. L. Goodwin, S. A. Mulvaney, F. Martinez, P. L. Enright, et al., 2003. Learning in children and sleep disordered breathing: Findings of the Tucson Children’s Assessment of Sleep Apnea (TUCASA) prospective cohort study. J. Int. Neuropsychol. Soc. 9(7):1016– 1026. Kasper, S., and T. A. Wehr, 1992. The role of sleep and wakefulness in the genesis of depression and mania. Encephale 18:45– 50. Kennedy, J. D., S. Blunden, and C. Hirte, 2004. Reduced neurocognition in children who snore. Pediatr. Pulmonol. 37:330– 337. Kohyama, J., 1998. Sleep as a window on the developing brain. Curr. Probl. Pediatr. 28(3):69–92. Kong, J., P. N. Shepel, C. P. Holden, M. Mackiewicz, A. I. Pack, and J. D. Geiger, 2002. Brain glycogen decreases with increased periods of wakefulness: Implications for homeostatic drive to sleep. J. Neurosci. 22(13):5581–5587. Leibenluft, E., P. S. Albert, N. E. Rosenthal, and T. A. Wehr, 1996. Relationship between sleep and mood in patients with rapid-cycling bipolar disorder. Psychiatry Res. 63:161–168.

jenni and dahl: sleep, cognition, and emotion

815

Louis, J., C. Cannard, H. Bastuji, and M. J. Challamel, 1997. Sleep ontogenesis revisited: A longitudinal 24-hour home polygraphic study on 15 normal infants during the first two years of life. Sleep 20(5):323–333. Maquet, P., 2001. The role of sleep in learning and memory. Science 294(5544):1048–1052. Marks, C. A., and M. J. Wayner, 2005. Effects of sleep disruption on rat dentate granule cell LTP in vivo. Brain Res. Bull. 66(2): 114–119. Meijer, A. M., H. T. Habekothe, and G. L. Van Den Wittenboer, 2000. Time in bed, quality of sleep and school functioning of children. J. Sleep Res. 9:145–153. Mirmiran, M., 1995. The function of fetal/neonatal rapid eye movement sleep. Behav. Brain Res. 69(1–2):13–22. Mirmiran, M., and R. L. Ariagno, 2003. Role of REM sleep in brain development and plasticity. In P. Maquet, C. Smith, and R. Stickgold, eds., Sleep and Brain Plasticity, 181–187. Oxford, UK: Oxford University Press. Mirmiran, M., H. van den Dungen, and H. B. Uylings, 1982. Sleep patterns during rearing under different environmental conditions in juvenile rats. Brain Res. 233(2):287–298. Mirmiran, M., and E. Van Someren, 1993. Symposium: Normal and abnormal REM sleep regulation: The importance of REM sleep for brain maturation. J. Sleep Res. 2(4):188–192. Miyamoto, H., H. Katagiri, and T. Hensch, 2003. Experiencedependent slow-wave sleep development. Nature Neurosci. 6(6): 553–554. O’Brien, L. M., and D. Gozal, 2004. Neurocognitive dysfunction and sleep in children: From human to rodent. Pediatr. Clin. North Am. 51(1):187–202. O’Brien, L. M., C. B. Mervis, C. R. Holbrook, J. L. Bruner, N. H. Smith, N. McNally, et al., 2004. Neurobehavioral correlates of sleep-disordered breathing in children. J. Sleep Res. 13(2):165–172. Oksenberg, A., J. P. Shaffery, G. A. Marks, S. G. Speciale, G. Mihailoff, and H. P. Roffwarg, 1996. Rapid eye movement sleep deprivation in kittens amplifies LGN cell-size disparity induced by monocular deprivation. Brain Res. Dev. Brain Res. 97(1):51–61. Peigneux, P., S. Larureys, X. Delbeuck, and P. Maquet, 2001. Sleeping brain, learning brain: The role of sleep for memory systems. NeuroReport 12:A111–124. Pilcher, J. J., and A. I. Huffcut, 1996. Effects of sleep deprivation on performance: A meta-analysis. Sleep 19:318–326. Pompeiano, O., M. Pompeiano, and N. Corvaja, 1995. Effects of sleep deprivation on the postnatal development of visual-deprived cells in the cat’s lateral geniculate nucleus. Arch. Ital. Biol. 134(1):121–140. Porkka-Heiskanen, T., L. Alanko, A. Kalinchuk, and D. Stenberg, 2002. Adenosine and sleep. Sleep Med. Rev. 6(4):321– 332. Porkka-Heiskanen, T., R. E. Strecker, M. Thakkar, A. A. Bjorkum, R. W. Greene, and R. W. McCarley, 1997. Adenosine: A mediator of the sleep-inducing effects of prolonged wakefulness. Science 276(5316):1265–1268. Randazzo, A. C., M. J. Muehlbach, P. K. Schweitzer, and J. K. Walsh, 1998. Cognitive function following acute sleep restriction in children ages 10–14. Sleep 21(8):861– 868. Rhodes, S. K., K. C. Shimoda, L. R. Waid, P. M. O’Neil, M. J. Oexmann, N. A. Collop, et al., 1995. Neurocognitive deficits in morbidly obese children with obstructive sleep apnea. J. Pediatr. 127(5):741–744.

816

emotion⁄cognition interactions

Richards, J. E., A. H. Parmelee, Jr., and L. Beckwith, 1986. Spectral analysis of infant EEG and behavioral outcome at age five. Electroencephalogr. Clin. Neurophysiol. 64(1):1–11. Roffwarg, H. P., J. N. Muzio, and W. C. Dement, 1966. The ontogenetic development of the human sleep-dream cycle. Science 152(3722):604–619. Sadeh, A., R. Gruber, and A. Raviv, 2002. Sleep, neurobehavioral functioning, and behavior problems in school-age children. Child Dev. 73(2):405–417. Sadeh, A., R. Gruber, and A. Raviv, 2003. The effects of sleep restriction and extension on school-age children: What a difference an hour makes. Child Dev. 74(2):444–455. Sadeh, A., A. Raviv, and R. Gruber, 2000. Sleep patterns and sleep disruptions in school-age children. Dev. Psychol. 36:291– 301. Saper, C. B., G. Cano, and T. E. Scammell, 2005. Homeostatic, circadian, and emotional regulation of sleep. J. Comp. Neurol. 493(1):92–98. Saper, C. B., T. E. Scammell, and J. Lu, 2005. Hypothalamic regulation of sleep and circadian rhythms. Nature 437(7063): 1257–1263. Scher, M. S., 1998. Understanding sleep ontogeny to assess brain dysfunction in neonates and infants. J. Child Neurol. 13(10):467– 474. Sejnowski, T. J., and A. Destexhe, 2000. Why do we sleep? Brain Res. 886:208–223. Shaffery, J. P., A. Oksenberg, G. A. Marks, S. G. Speciale, G. Mihailoff, and H. P. Roffwarg, 1998. REM sleep deprivation in monocularly occluded kittens reduces the size of cells in LGN monocular segment. Sleep 21(8):837–845. Shaffery, J. P., H. P. Roffwarg, S. G. Speciale, and G. A. Marks, 1999. Ponto-geniculo-occipital-wave suppression amplifies lateral geniculate nucleus cell-size changes in monocularly deprived kittens. Brain Res. Dev. Brain Res. 114(1):109– 119. Shaffery, J. P., C. M. Sinton, G. Bissette, H. P. Roffwarg, and G. A. Marks, 2002. Rapid eye movement sleep deprivation modifies expression of long-term potentiation in visual cortex of immature rats. Neuroscience 110(3):431–443. Siegel, J. M., 2005. Clues to the functions of mammalian sleep. Nature 437(7063):1264–1271. Smith, C., 2001. Sleep states and memory processes in humans: Procedural versus declarative memory systems. Sleep Med. Rev. 5(6):491–506. Steenari, M. R., V. Vuotela, E. J. Paavonen, S. Carlson, M. Fjallberg, and E. Aronen, 2003. Working memory and sleep in 6- to 13-year-old schoolchildren. J. Am. Acad. Child Adolesc. Psychiatry 42:85–92. Steriade, M., 1999. Coherent oscillations and short-term plasticity in corticothalamic networks. Trends Neurosci. 22(8):337–345. Steriade, M., D. A. McCormick, and T. J. Sejnowski, 1993. Thalamocortical oscillations in the sleeping and aroused brain. Science 262:679–685. Stickgold, R., 2005. Sleep-dependent memory consolidation. Nature 437(7063):1272–1278. Stickgold, R., J. A. Hobson, R. Fosse, and M. Fosse, 2001. Sleep, learning, and dreams: Off-line memory reprocessing. Science 294(5544):1052–1057. Stickgold, R., and M. P. Walker, 2005a. Memory consolidation and reconsolidation: What is the role of sleep? Trends Neurosci. 28(8):408–415. Stickgold, R., and M. P. Walker, 2005b. Sleep and memory: The ongoing debate. Sleep 28(10):1225–1227.

Tafti, M., S. Maret, and Y. Dauvilliers, 2005. Genes for normal sleep and sleep disorders. Ann. Med. 37(8):580–589. Tanner, J. M., 1998. A brief history of the study of human growth. In S. J. Ulijaszek, F. E. Johnston, and M. A. Preece, eds., The Cambridge Encyclopedia of Human Growth and Development, 3–12. Cambridge, UK: Cambridge University Press. Tononi, G., and C. Cirelli, 2003. Sleep and synaptic homeostasis: A hypothesis. Brain Res. Bull. 62:143–150. Tononi, G., and C. Cirelli, 2006. Sleep function and synaptic homeostasis. Sleep Med. Rev. 10:49–62. Turrigiano, G. G., and S. B. Nelson, 2000. Hebb and homeostasis in neuronal plasticity. Curr. Opin. Neurobiol. 10(3):358–364. Van Dongen, H. P., G. Maislin, J. M. Mullington, and D. F. Dinges, 2003. The cumulative cost of additional wakefulness: Dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep 26(2):117–126. Vertes, R. P., 2004. Memory consolidation in sleep: Dream or reality. Neuron 44(1):135–148. Vertes, R. P., and J. M. Siegel, 2005. Time for the sleep community to take a critical look at the purported role of sleep in memory processing. Sleep 28(10):1228–1229; discussion, 1230– 1233. Walker, M. P., 2004. Issues surrounding sleep-dependent memory consolidation and plasticity. Cell Mol. Life Sci. 61(24):3009– 3015.

Walker, M. P., and R. Stickgold, 2004. Sleep-dependent learning and memory consolidation. Neuron 44(1):121–133. Walker, M. P., and R. Stickgold, 2006. Sleep, memory, and plasticity. Annu. Rev. Psychol. 57:139–166. Wehr, T. A., D. A. Sack, and N. E. Rosenthal, 1987. Sleep reduction as a final common pathway in the genesis of mania. Am. J. Psychiatry 144:201–204. Whitlock, J. R., A. J. Heynen, M. G. Shuler, and M. F. Bear, 2006. Learning induces long-term potentiation in the hippocampus. Science 313(5790):1093–1097. Wolfson, A. R., and M. A. Carskadon, 1998. Sleep schedules and daytime functioning in adolescents. Child Dev. 69(4):875– 887. Wolfson, A. R., and M. A. Carskadon, 2003. Understanding adolescents’ sleep patterns and school performance: A critical appraisal. Sleep Med. Rev. 7(6):491–506. Wu, J. C., and W. E. Bunney, 1990. The biological basis of an antidepressant response to sleep deprivation and relapse: Review and hypothesis. Am. J. Psychiatry 147:14–21. Zepelin, H., J. M. Siegel, I. Tobler, 2005. Mammalian sleep. In M. H. Kryger, T. Roth, and W. C. Dement, eds., Principles and Practice of Sleep Medicine, 91–100. Philadelphia: Elsevier Saunders. Zwi, R., M. Shawe-Taylor, and J. Murray, 2005. Cognitive processes in the maintenance of insomnia and comorbid anxiety. Behav. Cogn. Psychother. 33:333–342.

jenni and dahl: sleep, cognition, and emotion

817

50

Neural Systems, Gaze Following, and the Development of Joint Attention PETER MUNDY AND AMY VAN HECKE

In a 1975 paper in Nature, Scaife and Bruner reported that, between the ages of 6 and 18 months, infants increasingly display the ability to follow the direction of gaze of a social partner. This observation was groundbreaking, in part because it was inconsistent with the prevailing Piagetian egocentric view of early cognitive development. To refer to this newly observed domain of development, Bruner (1975) adopted the general term joint attention. This term, though, was applied not only to gaze-following skills but to the development of related abilities as well, such as the tendency of infants to initiate episodes of social-attention coordination with gestures (e.g., pointing, showing) and alternating eye contact. Subsequently, the study of gaze-following and related joint-attention skills has become a major focus of research within the field of developmental psychology. Contemporary research on development of gaze following and joint attention follows several paths. One common perspective or model in the field revolves around the longstanding hypothesis that the development of gaze following and joint attention reflects the understanding of intentionality in others, which provides a necessary social-cognitive foundation for subsequent language development in typically and atypically developing children (e.g., Bretherton, 1991; Carpenter, Nagell, and Tomasello, 1998). Beyond social-cognition theory, researchers have begun to consider the links between joint attention and the early development of executive processes involved in self-monitoring, attention regulation, and social motivation (Mundy, 1995, 2003; Sheinkopf et al., 2004; Vaughan et al., 2003; Williams et al., 2005; Zelazo, Qu, and Muller, 2005). Relative to these lines of research, several groups have been applying developmental science to understand the role that syndrome-specific deficits in gaze following and joint attention may play in the development of autism (e.g., Dawson, Munson, et al., 2002; Filipek et al., 1999; Mundy, Sigman, et al., 1986; Mundy, Sigman, and Kasari, 1994; Sigman and Ruskin, 1999). Each of these lines of research has recently motivated researchers to begin to consider the neural systems and

neurodevelopmental processes that may be involved in the ontogeny of gaze following and joint attention. Our previous work in this regard has primarily focused on the nature of the systems that may be involved in the capacity of infants to initiate joint attention bids with others (Mundy, 1995, 2003; Mundy, Card, and Fox, 2000). In this chapter, however, we address what is known about the neural systems that are associated with gaze following and related abilities. To begin this discussion, we first present a brief overview of our perspective on the nature and importance of gazefollowing and joint-attention development. This is followed by a discussion of the functional neuroanatomy of gaze following, as well as other forms of joint attention. This leads to a consideration of a dual-process theory of attention development (Posner and Peterson, 1990; Rothbart and Posner, 2001) and social-cognitive development (Frith and Frith, 1999, 2001), in order to develop a sense of the possible common, as well as unique, neuropsychological processes that may be involved in gaze following and initiating forms of joint attention. Finally, a preliminary sketch of the important features of an integrative neurodevelopmental model of gaze following and initiating joint attention is provided.

The nature of gaze following and joint attention in early development Joint attention refers to processes and behaviors involved in the capacity of infants to coordinate their attention with a social partner. One type of joint-attention skill involves infants’ ability to follow the direction of another person’s gaze (Butterworth and Jarrett, 1991; Scaife and Bruner, 1975). In studies of this domain, children or adults are presented with a real or analogue social partner (e.g., a picture of a face). The gaze of the social partner is directed to the left, right, behind, or below the observing child/adult, and his/her behavioral or neurophysiological response to the direction of the social partner’s gaze is observed. Many of these studies control for and eliminate behaviors that may naturally occur when people use gaze shifting to redirect the

819

A

B

C

Figure 50.1 Illustrations of RJA following gaze and point (A), IJA pointing (B), and IJA alternating gaze (C) from the Early Social Communication Scales (ESCS; Mundy, Delgado, et al., 2003). (See plate 63.)

attention of a social partner, such as head turning, pointing, and vocalizing. Alternatively, these multimodal stimuli may be more ecologically valid and more powerful elicitors of attention following, at least in infants (see Flom and Pick, 2003). Therefore, several studies (e.g., Mundy, Card, and Fox, 2000; Ulvund and Smith, 1996) have observed infants’ responses to gaze shifting accompanied by head turns, vocalization, and/or pointing in a social interaction (responding to joint attention, RJA: Seibert, Hogan, and Mundy, 1982). Another dimension of joint attention that arises in the first year involves the infant’s use of eye contact and/or deictic gestures (e.g., a pointing or showing) to spontaneously initiate coordinated or shared attention with a social partner. This latter type of protodeclarative act (Bates, 1976) will be referred to as initiating joint attention skill (IJA: Seibert, Hogan, and Mundy, 1982) (see figure 50.1 and plate 63). Both types of joint attention skills are critical milestones of early development and social learning. For example, much of early vocabulary acquisition in the second year takes place in unstructured or incidental social-learning situations where (1) parents provide learning opportunities by referring to new objects or events, but (2) the infant may need to sort through a number of stimuli in order to focus on the correct object/event and to acquire the appropriate new word association. In this situation, infants are confronted with the possibility of “referential mapping errors” (Baldwin, 1995). To deal with this problem, infants utilize the direction of gaze of the parent to limit the number of potential stimuli to which to attend and, therefore, increase the likelihood of a correct word learning experience (Baldwin,

820

emotion⁄cognition interactions

1995). Similarly, when infants initiate bids for joint attention, the responsive caregiver may follow the child’s line of regard and take advantage of the child’s focus of attention to provide a new word in a context that maximizes the opportunity to learn (cf. Tomasello, 1995). Hence, both gaze following and IJA may be regarded as early developing self-organizing facilities that are critical for much of subsequent language development (Baldwin, 1995; Mundy and Burnette, 2005; Mundy and Sigman, 2006). As previously noted, gaze following/RJA and IJA may reflect social-cognitive development, or an awareness that others have perspectives and intentions (e.g., Bretherton, 1991; Carpenter, Nagell, and Tomasello, 1998). However, this may be the case only in the later stages of development of these skills, in the second year of life (Brooks and Meltzoff, 2002). The rudiments of joint attention and especially gaze following may emerge as early as 3 months (D’Entremont, Hains, and Muir, 1997; Farroni, Massaccessi, and Francesca, 2002; Hood, Willen, and Driver, 1998; Striano and Stahl, 2005) and, thus, well before theory suggests that social-cognitive processes affect or organize behavior (Mundy and Sigman, 1989, 2006). Indeed, recent research suggests that, in particular, gaze following/RJA prior to the second year of life is not necessarily complete evidence of socialcognitive development (Brooks and Meltzoff, 2002; Meltzoff and Decety, 2005; Moore, 1996; Woodward, 2003). What other processes may be involved? Candidates here include processes associated with imitation (Meltzoff and Decety, 2003; Pomares et al., 2003), attention-regulation and/or social-motivation processes (Mundy, Card, and Fox, 2000;

Sheinkopf et al., 2004), early operant learning processes (Corkum and Moore, 1998), spatial analytic processes (Butterworth and Jarrett, 1991), and intersensory integration processes (Flom and Pick, 2003). It may well be the case, though, that gaze-following and RJA behavior in the first year of life provides a critical source of social information that contributes to the development of later-emerging socialcognitive facilities (e.g., Baron-Cohen, 1995; Meltzoff and Decety, 2003; Mundy and Neal, 2001). A more complete discussion of this hypothesis will be provided in the concluding section of this chapter. One infrequently considered aspect of theory and research alluded to previously is that gaze following and joint attention may involve executive functions and social-motivation processes. For example, to engage in successful gaze following, infants need to attend to people, process social cues, and then inhibit looking to their social partner and flexibly redirect their attention based on social cues that provide spatial orientation information. Thus gaze following may provide a unique and valid measure of the early development of attention-regulation skills in infancy. Indeed, Morales and colleagues (2000b) have observed that RJA at 6 months is associated with mothers’ ratings of individual differences in attention regulation on the Infant Behavior Questionnaire (IBQ: Rothbart, 1981). These observations may be especially important, since attention regulation is one of the fundamental building blocks for the development of social competence (Masten and Coatsworth, 1998). Therefore, above and beyond its links to language and social cognition, gaze following may reflect aspects of executive development that are important for the subsequent regulation of social behaviors (Mundy and Sigman, 2006; Sheinkopf et al., 2004; Vaughan Van Hecke et al., 2007). Individual differences in the development of gaze following, RJA, and IJA also reflect factors associated with socialemotional and motivation processes (Mundy, 1995). These factors, though, may be more prominent features of IJA than gaze following and RJA. Research suggests that IJA, but not RJA, is associated with temperament measures of emotional reactivity (Vaughan et al., 2003) and with the direct observation of the display of positive affect (e.g., Mundy, Kasari, and Sigman, 1992). Moreover, the selfinitiation of joint attention behaviors (i.e., IJA) may more clearly reflect social-motivation processes than behaviors involving the response to cues, such as gaze following and RJA (Mundy, 1995). Indeed, it may be that these differences in social-emotional and motivation processes contribute to divergent patterns of growth and associations across observations of IJA and RJA (Block et al., 2003; Mundy, Sigman, and Kasari, 1994; Mundy and Gomes, 1998; Mundy, Card, and Fox, 2000). These observations are not consistent with the notion that gaze following/RJA and IJA measure the same processes involved in the early

development of social cognition (Baron-Cohen, 1995; Tomasello, 1995; Carpenter, Nagell, and Tomasello, 1998). Thus there is a need to better understand the degree to which the development of gaze following/RJA and IJA reflect common and/or unique sets of processes. To address this issue, it is useful to compare and contrast the development of gaze following and IJA in terms of behavioral, psychological, and neurophysiological processes. Much of our own work, including this chapter, has adopted this perspective.

The importance of gaze following and joint attention in early development One of the more important tests of theory on the importance of joint attention development involves examining the longitudinal continuity between infant gaze following, or related joint-attention skills, and later childhood developmental outcomes. A consistent observation from this type of research has been that individual differences in infant gaze following and RJA are related to subsequent language development. Robust correlations between RJA measures in the second year and subsequent language acquisition have been observed in both typically and atypically developing children, even after controlling for individual differences in general aspects of cognitive development (Mundy and Gomes, 1998; Mundy, Kasari, et al., 1995). Indeed, recent data indicate that infant RJA measures have long-term predictive validity. Measures of RJA at 12 and 18 months have been observed to significantly correlate with language development at 3 to 7 years in typical and at-risk samples (Acra et al., 2003; Block et al., 2003; Neal et al., 2006). Moreover, there is evidence that individual differences in RJA skills may be observed as early as 6 months of age and that these differences predict language outcomes at 12, 24, and 30 months (Morales, Mundy, and Rojas, 1998; Morales et al., 2000a, 2000b). Initiating-joint-attention skills also appear to be related to language outcomes (Carpenter, Nagell, and Tomasello, 1998; Mundy and Gomes, 1998), but some evidence suggests that RJA measures may be the more consistent or robust correlates of early language outcomes (Mundy, Kasari, et al., 1995; Mundy and Gomes, 1998). However, others have reported that IJA, rather than RJA, appears to be the strongest correlate of 5- to 6-year-old verbal and performance IQ outcomes in at-risk infants (Ulvund and Smith, 1996). Obviously, more research is needed to understand the relative contributions of processes tapped by both these dimensions of joint attention vis-à-vis language development. There have been surprisingly few empirical reports on the longitudinal links between infant joint attention

mundy and van hecke: neural systems, gaze following, and joint attention

821

development and social cognitive development. Indeed, there currently may be only one empirical study on this issue. Charman and colleagues (2000) followed a sample of 13 typically developing infants from 20 to 44 months. At 20 months, observations were recorded of toddlers spontaneously alternating eye contact between a tester and an interesting toy spectacle. This type of “alternating gaze” measure is an early index of IJA (e.g., Mundy, Sigman, et al., 1986; Tomasello, 1995). After controlling for differences in IQ and language development, the 20-month IJA alternating gaze measure was a significant predictor of 44-month theory of mind (ToM) performance. No data, though, were presented on gaze following or RJA in this study. The paucity of information on this topic is a significant gap. Nevertheless, research on the neural systems involved in gaze following/ RJA, IJA, and attention, to be described later in this chapter, provides an intriguing, albeit indirect, empirical bridge to social-cognitive development. As briefly noted earlier, recent theory and research also suggest that there may be important links between gaze following, joint attention, and social-emotional processes that may be associated with the development of social competence (see Mundy and Sigman, 2006, for review). Skill in RJA, assessed with caregivers at 6 months of age, was related to toddlers’ adaptive emotion and attention regulatory behaviors on a delay-of-gratification task at 24 months (Morales et al., 2005). Vaughan Van Hecke and colleagues (2007) have reported that infant-tester interaction measures of both RJA and IJA at 12 months were negatively related to parent reports of externalizing behaviors at 30 months in a sample of typically developing infants. Similarly, Sheinkopf and colleagues (2004) reported that 12-, 15-, and 18-month average IJA and RJA scores were significantly associated with preschool teacher reports of classroom externalizing behavior at 36 months in cocaine-exposed infants. Multiple regression data in this study indicated that both joint-attention measures made unique contributions to the prediction of externalizing behavior. Furthermore, infant RJA was also a significant negative predictor of teacher reports of withdrawn behavior, but a positive predictor of teacher reports of social competence. In a longer-term follow-up of this sample, though, Acra and colleagues (2003) observed that IJA, but not RJA, was negatively related to parent and teacher observations of externalizing, but positively associated with observations of social competence at 7 years of age. These studies are interesting, but much more work needs to be done to delineate and understand these possible lines of developmental continuity. In summary, recent longitudinal data are consistent with theory that suggests that there are important points of biobehavioral continuity between joint-attention processes and later language, cognitive, and social development. One implication of this literature is that research on infant

822

emotion⁄cognition interactions

gaze-following and joint-attention measures have much to contribute to understanding the processes that contribute to problems in social and cognitive outcome among “at-risk” and developmentally disordered children. Nowhere has this potential been more evident than in the study of autism.

Joint-attention impairments in autism Autism is a biologically based disorder that may be more prevalent than once thought, with the spectrum of autismrelated disorders occurring at a rate between 2–6 per thousand (Dawson, Webb, et al., 2002; Fombonne, 2003). Pathognomic features of this syndrome include impaired social and communication development. One important characteristic of these social impairments is a robust early disturbance of joint-attention development (Mundy, Sigman, et al., 1986; Mundy and Sigman, 1989). Observations of joint-attention deficits are extremely well replicated and appear to be universal among young children with autism (Filipek et al., 1999). Moreover, the joint-attention domain appears to be sensitive to early individual differences in the development of children with autism that are related to language and social outcomes (Bono, Daley, and Sigman, 2004; Mundy, Sigman, and Kasari, 1994; Sigman and Ruskin, 1999). Young children with autism display deficits in both IJA and RJA skills. However, with development, there is evidence of dissociation in correlates and course of these joint-attention impairments. For example, although some studies have suggested that children with autism may not use RJA in language development (Baron-Cohen, Baldwin, and Crowson, 1997), other research has suggested that RJA is significantly related to language development (Carpenter, Pennington, and Rogers, 2002; Sigman and Ruskin, 1999) and may even moderate the effectiveness of early intervention on language development with these children (Bono, Daley, and Sigman, 2004). The relation of IJA to language development, though, has been less clear in these studies. Research has also suggested that RJA may be more sensitive to parent-child interaction effects associated with attachment among children with autism (Capps, Sigman, and Mundy, 1994). In contrast to RJA, observations have suggested that IJA deficits may be more highly associated with impairments in positive affective expression in autism (Kasari et al., 1990), parent reports of symptoms of social impairments (Mundy, Sigman, and Kasari, 1994; see also Rogers et al., 2003), and long-term social outcomes (Lord et al., 2003; Sigman and Ruskin, 1999). Of course, these observations were derived from studies with modest sample sizes; hence they may be prone to type II error. However, several studies now indicate that, although delayed in development, children with autism display basic gaze-following ability by 2 years of age (Charwarska, Klin,

and Volkmar, 2003), and problems in RJA may remit among older children with autism or those with higher mental ages (Mundy, Sigman, and Kasari, 1994; Leekam and Moore, 2001; Sigman and Ruskin, 1999). However, IJA deficits remain robust. As alluded to earlier, this difference in course of joint-attention skills development in autism may be explained by the basic difference between gaze-following/ RJA and IJA skills. The latter reflect spontaneously generated social attention coordination behavior, whereas gaze following/RJA involves the perception and response to the social cues of another person. Initiating joint attention, then, may be more affected by executive and social-motivation processes involved in the generation and self-initiation of behavioral goals than RJA (Mundy, 1995; Mundy, Card, and Fox, 2000). In particular, IJA deficits in autism appear to reflect an impairment in the tendency to spontaneously initiate episodes of shared affective experience concerning an object or an event with a social partner, and this process does not appear to involved to as great an extent in RJA (Kasari et al., 1990). To review, gaze-following and related joint-attention skills not only reflect early social cognitive development, but also contribute to self-organizing systems that facilitate early social learning. They also provide developmental markers of aspects of the development of attention regulation, and emotional and motivation processes that play a role in subsequent individual differences in social competence. Furthermore, deficits in gaze-following/RJA and, especially, IJA are cardinal symptoms of autism, a complex developmental disorder that results in a pernicious attenuation of social competencies. Therefore, it seems likely that understanding and comparing the brain systems involved in various types of joint-attention skill development may provide clues with respect to critical aspects of the neurobiology of social development, as well as the neural systems that may be central to autism. In this regard, we first consider the emergent literature on the neurobiology of gaze following and RJA.

Gaze following, RJA, and the ventral “social brain” Social perception appears to be supported by a complex ventromedial “social brain” circuit involving the orbitofrontal cortex, temporal cortical areas, including the superior temporal sulcus (STS) as well as superior temporal gyrus (STG), and ventral subcortical areas such as the amygdala (Adolphs, 2001; Brothers, 1990). With respect to gaze following, research using a variety of methods (electrophysiological methods, imaging, and comparative methods) has converged to indicate that gaze monitoring and following are mediated by part of this social brain circuit contained within the superior temporal sulcus (STS) and gyrus (STG), adjacent parietal areas (e.g., Brodmann’s area 40), and the

amygdala. The temporal and parietal areas of the brain are thought to contain neural networks that respond preferentially to faces, animate movement, and spatial orientation, including head, eye, and body orientation (e.g., Emery, 2000; Calder et al., 2002), while the amygdala, in conjunction with the orbitofrontal cortex, is thought to be involved in evaluating the valence or reward value of stimuli (e.g., Adolphs, 200; Dawson, Webb, et al., 2002). Wicker and colleagues (1999) observed that neural groups in the posterior STS were activated in response to faces with direct or horizontally averted eye gaze, but not to faces with downward eye gaze. Wicker and colleagues, however, did not observe differences between direct and averted eye gaze conditions. Alternatively, Puce and colleagues (1998) reported that videos of face stimuli with gaze moving horizontally from forward to averted elicited greater posterior STS activation than did faces with static forward gaze. Face matching on the basis of direction of gaze also elicited activation of neurons in the left posterior STS, while identitybased face matching elicited bilateral activation from the fusiform and inferior occipital gyri (Hoffman and Haxby, 2000). Similarly, George, Driver, and Dolan (2001) reported that direct gaze stimuli elicited more fusiform activation than averted gaze stimuli. Finally, Kingstone, Friesen, and Gazzaniga (2000) reported data on gaze following in two split-brain patients that were consistent with the notion that parietal, as well as temporal, subsystems specialized for face processing and processing of information relevant to spatial orientation combine to support the development of gaze following. Mundy, Card, and Fox (2000) have also reported observations linking gaze following/RJA to parietal processes. These authors examined the longitudinal relations between baseline EEG at 14 months and RJA at 18 months in 32 typically developing infants as measured on the Early Social Communication Scales (ESCS: Seibert, Hogan, and Mundy, 1982; Mundy, Delgado, et al., 2003). The EEG data were collected with electrodes placed bilaterally at dorsofrontal, central, temporal, parietal, and occipital sites. Responding to joint attention at 18 months was predicted by EEG indexes of left-parietal activation and right-parietal deactivation at 14 months. These data were quite consistent with previous research (Emery, 2000) that suggested that parietal areas specialized for spatial orienting and attention, along with temporal systems specialized for processing gaze, may contribute to gaze-following or related RJA skill development. Dawson and colleagues (2002) have reported a neuropsychological study of joint-attention ability in young children with autism that is also indicative of relations between gaze following and temporal “social brain” activity. Dawson and colleagues observed that performance on a type of delayed nonmatch to sample (DNMS) measure, which has previously

mundy and van hecke: neural systems, gaze following, and joint attention

823

been associated with functions of a temporal-ventromedialfrontal circuit, was significantly associated with performance on a combined measure of RJA/IJA in children with autism and typical development. Although the joint-attention measure reflected both RJA and IJA skills, it was constructed in such a fashion as to likely be biased to the former. The DNMS measure was thought to provide an index of processes associated with rule learning that depends on the orbitofrontal and amygdala-mediated capacity to associate novel stimuli with reward value (see Dawson et al., 2002), as well as selective encoding of stimulus features guided by the hippocampus (Hampson et al., 2004). Hence the authors suggested that this ventral brain function plays a role in the social learning that contributes to gaze-following/RJA development. However, the results of this study did not indicate that processes tapped by the DNMS data played a major role in the nature of autism. The results of these human imaging, electrophysiology, and neuropsychological studies are consistent with earlier reports from comparative research that provide experimental evidence of temporal (i.e., STS) and parietal involvement in gaze following (Emery, 2000). In two studies, presurgical monkeys demonstrated a clear ability to discriminate face stimuli on the basis of direction of gaze. After resection of the STS, however, the gaze discrimination abilities of the monkeys fell to chance (Cambell et al., 1990; Heyward and Cowey, 1992). Eacott and colleagues (1993) compared two groups of monkeys—those with and without STS surgically induced lesions—on a task of discriminating pairs of eyes directed straight ahead or averted 5 or more degrees. The results indicated that the nonlesioned monkeys were capable of discriminating targets involving horizontal eye gaze shifts of greater than 5 degrees, but the STS lesioned animals were not. Eacott and colleagues also reported, though, that the lesioned animals also performed worse than the nonlesioned animals on a nonsocial task involving the discrimination between stimuli formed from ASCII characters. The latter observation reminds us that, although research is beginning to pinpoint the systems involved in gaze following and social perception, the specificity of these systems for social versus nonsocial processing has still to be definitively examined. One recent study, though, has addressed a related issue in imaging research on human gaze following. Hooker (2002) used whole-brain f MRI to compare neural activity in response to (1) horizontal eye movement stimuli that provided directional information about where a visual stimulus would appear, or (2) arrow stimuli that provided equivalent directional information, or (3) eye movements that did not provide directional information. Hooker (2002) observed more activity in the STS in the first condition than in either of the other conditions. Alternatively, Hooker reported more activity in the fusiform gyrus and prefrontal cortex in the eye-motion control condition (condition 3) than in the other

824

emotion⁄cognition interactions

conditions. These data were consistent with the notion that the STS may develop a specialization for processing gazerelated, social-spatial orientation information. Activity in other areas of the “social brain” has also been observed to be involved in gaze following. In one study, Kawashima and colleagues (1999) reported that the left amygdala displayed activation to faces with horizontal gaze aversion, but that right-amygdala activation increased with presentation of direct-gaze pictures of faces. It may be important to note, though, that the subjects in this study were eight righted-handed men from the Imuyama region of Japan, and the stimuli involved digitized cinematic images of an attractive Japanese woman’s face (see figure 1 in Kawashima et al., 1999). The relative attractiveness of face stimuli, however, may affect the pattern of brain activity observed in response to gaze stimuli (Kampe et al., 2001). Specifically, the study by Kampe and colleagues (2001) revealed increased activity in the ventral striatum in response to more attractive faces with direct gaze and decreased activity to less attractive faces with averted gaze. The ventral striatum is part of the input system of the basal ganglia and plays a specific role in the “limbic loop” that receives information from the medial and lateral temporal lobes (e.g., amygdala) and passes it on to thalamic nuclei and, ultimately, to the anterior cingulate and orbitofrontal cortex (Martin, 1996). Based on this finding, Kampe and colleagues suggested that the ventral striatum and associated neural groups may be involved in processing the relative reward value associated with eye contact and gaze aversion. This observation is not only consistent with Dawson and colleagues (2002), but also suggests that variability of stimulus characteristics (e.g., their representation of reward value) may both inform and complicate research on the complex neural systems involved in processing gaze and gaze direction. One puzzling aspect of the imaging research on gaze following has recently been noted by Calder and colleagues (2002). Theoretically, gaze-following behavior reflects aspects of social-cognitive processes (e.g., Baron-Cohen, 1995). In this regard, recent imaging studies have suggested that, in addition to ventral “social brain” activation, social cognition is associated with activation of the dorsal medial frontal cortex (Brodmann’s areas 8/9) and the anterior cingulate (Frith and Frith, 1999, 2001; Mundy, 2003). So why have studies of gaze following not observed a link with frontal, dorsal medial activation? Calder and colleagues (2002) suggested that task difficulty should be considered in this regard. Most studies have examined passive gaze following on tasks that did not require the perception or inference of intentions on the basis of eye gaze. These authors suggested that more complex presentations of sequences of stimuli involving gaze directed toward and away from participants may elicit this type of processing, as well as evidence

of more dorsal contributions to the neural substrate of gaze following. To this end, Calder and colleagues (2002) used positron emission tomography (PET) to examine the neural responses of nine female volunteers to a relatively complex sequence of faces with gaze-averted, gaze-direct, and gazedown orientations. It was not completely clear from the methods described in this paper how these stimuli might have elicited more processing of intentionality than in previous studies. Nevertheless, the results provided evidence of activation in the dorsal medial frontal cortex (BA 8/9) and medial frontal cortex proximal to the anterior cingulate (BA 32), as well as areas of the STS, in response to horizontal gaze aversion. This study, though, unlike most others, employed a sample of women rather than men. So it is unclear whether an as-yet-unrecognized gender effect may have also played in a role in these observations. This caveat notwithstanding, it is clearly the case that when the interpretation of intentions is an overt feature of the demands of a gaze-processing task, activation of the dorsal medial frontal cortex appears to be involved. In an fMRI study, Baron-Cohen and colleagues (1999) presented six individuals with Asperger disorder and 12 typical controls with the “eyes test,” which requires inferring of people’s emotional states or gender from pictures of their eyes. The results of this study indicated that, in addition to activity in the orbitofrontal cortex, amygdala, and STS of the “social brain,” activation of the left and right dorsal medial frontal cortex was also a specific correlate of performance on this task in the typical sample. Interestingly, the autism groups failed to display amygdala activation and may not have shown as much right dorsal medial activation. Russell and colleagues (2000) have also employed the “eyes test” (Baron-Cohen et al., 1999) in an fMRI study of the neural characteristics of individuals affected by schizophrenia. The control sample displayed relatively more activity in the medial-frontal lobe (BA 9 and 45) in association with performance on this task, relative to the individuals with schizophrenia. In addition, more ventral “social brain” components of the left inferior frontal gyrus (BA 44/45/47) and the left middle and superior temporal gyri (BA 21/22) contributed to clinical group differences on performance on this task. In summary, the emerging literature on gaze following indicates that the most consistent correlates of gaze following and RJA appear to involve ventral “social brain” neural clusters in the STS. However, there is also evidence that parietal, orbitofrontal, and amygdala processes may play a role in gaze following. Moreover, when and if tasks involve inference of intentions (i.e., social cognition), more dorsal medial cortical systems may be brought to bear in processing gaze-related stimuli. This pattern of results is interesting in that it is consistent with developmental studies that suggest that gaze following may initially not reflect social cognitive activity per se, but that it comes to do so over time and

experience (Brooks and Meltzoff, 2002; Moore, 1996; Woodward, 2003). We will return to this issue after considering the neural correlates of other forms of joint attention, especially IJA.

IJA, social cognition, and the dorsal “social brain” In addition to the work of Dawson, Munson, and colleagues (2002), two other studies examined the neuropsychological correlates of joint attention in children with autism and controls (Griffith et al., 1999; McEvoy, Rogers, and Pennington, 1993). Unlike the research of Dawson and colleagues, though, these studies used separate IJA and RJA measures from the ESCS. Both of these studies observed that performance on a spatial reversal task, which presumably taps into dorsolateral frontal response inhibition, memory, and planning processes, was related to both IJA and RJA in young typically developing children and children with autism. Hence these studies provided data that were consistent with the notion that common executive functions may play a role in RJA and IJA development. In addition to common processes, though, Mundy (2003) has observed that the input or perception of social behaviors, such as gaze following/RJA, is very different from the output or spontaneous organization and initiation of social behaviors, such as in initiating joint attention bids. Therefore, there may be differences, as well as similarities, in the neural systems that contribute to these types of joint-attention behaviors. One aspect of these differences was observed in a study of the behavioral outcome of 13 infants who underwent hemispherectomies in an attempt to treat their intractable seizure disorders (Caplan et al., 1993). Positron emission tomography data were gathered prior to surgical intervention with the infants, and the ESCS was used to assess the postsurgical development of IJA and RJA, as well as the tendency for infants to initiate behavior requests (IBR). In addition to comparing IJA and RJA, IBR may also be important to consider in research on joint attention. Whereas IJA reflects the social use of attention-directing behavior (e.g., directing attention to show or share interest in an object), IBR involves the instrumental use of attentiondirecting behavior (e.g., directing attention to elicit aid in obtaining an object or event). Hence, by comparing IJA behaviors (e.g., showing a jar containing toys) with IBR behaviors (e.g., giving a jar containing toys to request aid in opening the jar), brain systems associated with the spontaneous initiation of “social” joint-attention bids versus “instrumental” joint-attention bids may be examined. The results indicated that metabolic activity in the frontal hemispheres, especially the left frontal hemisphere, predicted the development of IJA skill in this sample. However, the postsurgical development of the capacity of children to respond to the joint-attention bids (RJA) or initiate requesting bids

mundy and van hecke: neural systems, gaze following, and joint attention

825

(IBR) was not observed to relate to any of the PET indexes of cortical activity. Moreover, metabolic activity recorded from other brain regions, including ventral social brain regions of the orbital, temporal, parietal, and occipital cortex, were not significantly associated with IJA or other social-communication skills in this study. Thus frontal activity appeared to be specifically related to the development of the tendency to spontaneously initiate social joint attention bids to share experiences with others. Mundy, Card, and Fox (2000) followed up on these observations with a study of the links between EEG activity and joint-attention development between 14 and 18 months. As previously noted, one result of this study was the observation that 14-month parietal activity predicted differences in RJA skill development at 18 months. Alternatively, individual differences in 18-month IJA were predicted by a different and complex pattern of 14-month EEG activity at left medial-frontal electrode sites, as well as indices of right central deactivation, left occipital activation, and right occipital deactivation. Although the source location of the EEG data could not be definitively determined in this study, the frontal correlates of IJA reflected activity only from electrodes positioned above a point of confluence of Brodmann’s areas 8 and 9 of the medial-frontal cortex of the left hemisphere (Martin, 1996). This area includes aspects of the frontal eye fields and supplementary motor cortex commonly observed to be involved in an anterior system of attention control (Posner and Petersen, 1990). This study suggested that a dual-process, or multiplesystems, neurodevelopmental model of joint-attention development might be useful to consider (Mundy, Card, and Fox, 2000). In particular, activity of ventral and parietal systems may be a common correlate of gaze following/RJA, while more dorsal activation may be a relatively stronger component of IJA. However, even with the combined data from Caplan and colleagues (1993) and Mundy, Card, and Fox (2000), the evidence for frontal IJA connections was far more tenuous than the evidence for temporal-parietal RJA connections. Thus the recent contribution of Henderson and colleagues (2002) becomes even more important. Henderson and colleagues (2002) also employed the ESCS to examine baseline EEG at 14 months as a predictor of 18-month joint-attention development in 27 typically developing infants. However, to improve the spatial resolution of their data, they used a higher-density array of 64 electrodes. Moreover, they reasoned that, since the total ESCS scores for measures of IJA and other domains used in Mundy, Card, and Fox (2000) were composites of several behaviors, the exact nature of the associations with EEG activity were unclear. Therefore, Henderson and colleagues compared the EEG correlates of only two types of behaviors, selfinitiated pointing to share attention with respect to an active mechanical toy (IJA pointing) and self-initiated pointing to

826

emotion⁄cognition interactions

elicit aid in obtaining an out-of-reach object (IBR pointing). In the ESCS, the former involves pointing to a toy that is within easy reach, and the latter involves pointing to a toy that is out of reach. In this study, no significant correlations were observed between any of the 14-month EEG data and IBR pointing at 18 months. However, in the 3–6 Hz band, 14-month EEG data indicative of greater brain activity over the medial frontal cortex was strongly associated with more IJA pointing at 18 months. These correlations involved electrodes that were placed above cortical regions corresponding to Brodmann’s areas 8, 9, and 6. Henderson and colleagues (2002) also analyzed data from the 6–9-Hz band, which revealed 15 significant correlations between 14-month EEG data and 18-month IJA pointing. Again, higher bilateral activity corresponding to the previously identified medial-frontal sites was a strong predictor of IJA pointing at 18 months. In addition, though, IJA pointing at 18 months was also predicted by activity in this bandwidth from regions of the orbitofrontal, temporal, and dorsolateral frontal cortical regions. Thus this study suggested that that IJA development might reflect an integration of dorsal cortical functions (Mundy, Card, and Fox, 2000) with the ventral “social brain” and dorsolateral functions identified in other studies (Dawson et al., 2002; Griffith et al., 1999). However, there was little evidence for parietal involvement in IJA. The study of Henderson and colleagues (2002) also provided information about the social specificity of the link between IJA and dorsal cortical brain activity. As previously noted, the specific medial frontal cortical areas of involvement suggested by data from Mundy, Card, and Fox (2000) and some of the data from Henderson and colleagues (2002) correspond to aspects of both the frontal eye fields and supplementary motor cortex associated with the control of saccadic eye movement and motor planning (Martin, 1996). Therefore, these associations could simply reflect the motor control of the eye movements and/or gestural behaviors that are intrinsic to joint attention behavior. However, the study of Henderson and colleagues controls for this possible interpretation. The gross motor topography of IJA pointing and IBR pointing is virtually identical on the ESCS. Therefore, a neuromotor explanation of the different cortical correlates of IJA and IBR appears unlikely. Instead, since IJA pointing and IBR pointing appear to serve different socialcommunicative functions, it is reasonable to assume that the difference in EEG correlates of these infant behaviors also reflects differences in the neurodevelopmental substrates of these social-communicative functions. In addition to these studies on infants, two recent imaging studies on adults also provide data that converge on the observation that initiating joint attention and or awareness of engaging in joint attention involves parts of the frontal

eye fields and supplementary motor cortex, especially in the left hemisphere. Astafiev and colleagues (2003) compared cortical activation associated with attention, looking, and pointing to a peripheral target. They found that activation of the frontal eye fields and intraparietal sulcus was common to all activities. However, pointing also recruited activation of the supplementary motor cortex as well as the precuneus, parietal lobes, and posterior superior temporal sulcus. Finally they reported that the patterns of posterior activity were very consistent with related functional mapping of the posterior cortex of other primates (macaques) but not with respect to frontal functional associations noted in this study. In another study Williams and colleagues (2005) examined cortical activation in adults in two conditions: visually tracking a moving dot on a monitor where a simulation of a social partner on the screen also appeared to be visually tracking the dot and visually tracking the dot where the simulated social partner appeared to be visually tracking something other than the dot. The former “joint attention” simulation condition uniquely elicited activation of the medial frontal cortex and particularly the left superior frontal lobe. Of course, even data from five studies (Astafiev et al., 2003; Caplan et al., 1993; Henderson et al., 2002; Mundy, Card, and Fox, 2000; Williams et al., 2005) are not sufficient to draw strong conclusions about frontal cortical associations with IJA. Recall, though, that joint-attention skills, and especially IJA (Swettenham et al., 1998), may reflect an incipient aspect of social-cognitive and, specifically, theory-of-mind development. If this hypothesis is true, then the neural correlates of ToM-related task performance may provide information about the neural systems involved in IJA. Indeed, recent imaging research indicates that brain activity in the dorsal medial cortex (Brodmann’s areas 8/9) and adjacent subcortical areas of the anterior cingulate is the most consistent correlate of ToM task performance (Frith and Frith, 1999, 2001). This finding is true for both verbal and nonverbal measures of social cognition, though in the latter, areas of the right inferior-frontal cortex, right cerebellum, and right and left temporal cortices may also be involved in solving social-cognitive tasks (see Mundy, 2003, for a review). This literature provides evidence of a potentially significant neurofunctional linkage that lends credence to observations of a dorsal medial-frontal contribution to IJA. Thus recent research suggests that the most consistent brain process correlates of IJA, ToM, and social-cognitive performance appear to involve activity in the dorsal medial frontal cortex (DMFC), although more ventral-temporal social brain processes may also be involved (Frith and Frith, 1999, 2001; Mundy, 2003). Alternatively, current research suggests that ventral brain and parietal processes are the most consistent correlates of gaze following and RJA, but some evidence of DMFC involvement has also been

presented. Little evidence of parietal involvement in IJA has been presented. Hence the literature reviewed so far suggests that multiple neural systems contribute to joint attention and social cognitive development. Research and theory on attention development, which describes the initial contribution of parietal systems to less volitional attention control followed by the advent of more medial frontal contributions to volitional attention control (e.g., Rothbart, Posner, and Rosicky, 1994), may assist in understanding the function of these multiple neural systems. Similarly, social cognitive theory that describes the integration of ventral brain processes associated with the perception of others’ social behavior and self-monitoring related to more dorsal medial cortical systems (Frith and Frith, 1999, 2001) might also be useful to consider. In the next section, the implications of these perspectives for a more coherent view of the neuropsychological processes involved in the development of gaze following and other aspects of joint attention are examined.

The posterior attention system and gaze following/RJA Posner, Rothbart, and colleagues (e.g., Posner and Petersen, 1990; Rothbart, Posner, and Rosicky, 1994) have described the development of mechanisms of attention in the first years of life in terms of an early-developing posterior system and a later-developing anterior system. The details of the posterior system may help to elucidate an understanding of the neurodevelopmental processes involved in the emergence of gaze-following/RJA joint attention (Mundy, Card, and Fox, 2000). The posterior system is thought to begin to come online between birth and 4 months of age, and it regulates reflexive or nonvolitional shifts of attention. Research suggests that aspects of this network develop in the first four months of life and that they are localized to the superior parietal lobe, the pulvinar and reticular nuclei of the thalamus, and the midbrain superior colliculus (Rothbart and Posner, 2001). Specifically, the parietal lobe allows the disengagement of attention from a current focus, the superior colliculus allows the shift of attention to a new location, and the pulvinar and reticular nuclei allow increased processing of information gathered from the new focus. The behavioral concomitants of this pattern of cortical and subcortical control are strikingly similar to the behavioral mechanisms involved in gaze following and responding to joint attention (Mundy, Card, and Fox, 2000). Other research has begun to provide a more detailed picture of the neurodevelopment of the posterior attention system as it may relate to RJA (for an alternative view, see chapter 29 by Richards, this volume). First, between 1 and 2 months of age, infants often engage in periods of extended gaze, also called obligatory

mundy and van hecke: neural systems, gaze following, and joint attention

827

looking or sticky fixation (Hood, 1995), controlled by a pathway from the retina through the basal ganglia, which then inhibits the superior colliculus from ending this period of fixation (M. Johnson, 1990; Rothbart, Posner, and Rosicky, 1994). Second, infants are able to track moving stimuli at this age, but their tracking involves following a target while in their central visual field, reorienting when it moves out of the central field, and following it again. Research has shown that this type of “choppy” tracking is indicative of collicular control of eye movements (M. Johnson, 1990). Last, it has been suggested that the superior colliculus derives most of its input from stimuli in the temporal, versus nasal, visual field. Indeed, it has been shown that newborns orient more readily to stimuli in the temporal visual field (Simion et al., 1998). Evidence of more cortical control of eye movements in infants has been observed by about 3 months of age, when a pathway from the frontal eye fields (BA 8/9) that releases the superior colliculus from inhibition begins to provide additional input to structures controlling visual attention (M. Johnson, 1990). The function of this pathway may underlie 4-month-old infants’ ability to suppress automatic visual saccades in order to respond to a second, more attractive stimulus (M. Johnson, 1995) and 6-month-olds’ ability to respond to a peripheral target when central, competing stimuli are present (Atkinson et al., 1992). These observations are of interest for two reasons. First, it is reasonable to speculate that the functions of this pathway also contribute to the capacity for gaze following that begins to manifest itself in the 3- to 6-month period (D’Entremont, Hains, and Muir, 1997; Hood, Willen, and Driver, 1998; Morales, Mundy, and Rojas, 1998). Second, this pathway is associated with dorsal medial cortical areas (BA 8/9) that have also been associated with social cognitive development. The ability to make saccades to targets may also recruit the capacity to disengage attention, and this is a primary function of the parietal lobe, which develops rapidly in the third and fourth months of life (M. Johnson, Posner, and Rothbart, 1991). Furthermore, PET studies have shown that the metabolism of the parietal lobe is at adult levels by 4 months of age (Chugani, Phelps, and Mazziotta, 1987), perhaps also coinciding with the development of infants’ ability to anticipate visual shifts of attention based upon previous information (Clohessy, Posner, and Rothbart, 2001) and smoothly track moving stimuli (M. Johnson, 1990). Together, these data suggest that multiple structures subserving reflexive shifts of attention are coming online in the first six months of life, at a time when gaze following begins to emerge. According to this model the posterior attention system regulates reflexive attention in response to biologically significant stimuli (Rothbart, Posner, and Rosicky, 1994). Over the

828

emotion⁄cognition interactions

past decades research has suggested that human faces are one of the most salient features of infants’ early visual experience and that there may be an innate tendency to orient toward face or facelike stimuli shortly after birth (Bard et al., 1992). Currently research suggests that faces begin to be perceived as a distinct class of objects in the first six months of life and that attention to faces during an early sensitive period of development leads to perceptual and cortical specialization (Nelson, 2001).

The anterior attention system and initiating joint attention In addition to and in conjunction with this posterior system, attention is also regulated by an anterior system involving the dorsal medial frontal cortex (BA 8/9) and the anterior cingulate (BA 24). This anterior network may become functional after the posterior parietal system and is thought to make numerous contributions to the planning, self-initiation, and self-monitoring of goal-directed behaviors, including visual orienting (Rothbart, Posner, and Rosicky, 1994). One important component process here is the role the anterior system plays in the capacity to share attention across dual tasks, or foci of attention (Stuss et al., 1995), especially with respect to the capacity to maintain and flexibly switch between goal representations in working memory (e.g., Birrell and Brown, 2000; Rushworth et al., 2002). This anterior capacity may play a role in infants’ ability to maintain representations of self, a social partner, and an interesting object spectacle while flexibly switching attention between these foci in initiating joint-attention behaviors (Mundy, Card, and Fox, 2000). The attention-switching facility of the anterior system plays a critical role in the contribution to the supervisory attention system (SAS) (Norman and Shallice, 1986) that functions to guide behavior, especially attention deployment, depending on the motivational context of the task (e.g., Amador, Schlag-Rey, and Schlag, 2000; Bush, Luu, and Posner, 2000). In this regard, this system ultimately comes to participate in monitoring and representing the self and directing attention to internal and external events (Faw, 2003). With respect to self-representation, Craik and colleagues (1999) and S. Johnson and colleagues (2002) have reported studies that reveal that self-referenced memory processes preferentially activate the DMFC component of this anterior system. With respect to self-monitoring, research has led to the observation that, when people make an erroneous saccadic response in an attention-deployment task, there is a negative deflection in the stimulus-and-response-locked event related potential (ERP) called the error-related negativity, or ERN (Luu, Flaisch, and Tucker, 2000; Bush, Luu, and Posner, 2000). Source location suggests the ERN emanates from an area of the DMFC proximal to the anterior cingulate (AC) cortex (Luu, Flaisch, and Tucker, 2000). Observations

of the ERN suggest that there are not only specific cell groups within the DMFC/AC that are active in initiating a behavioral act, such as orienting to a stimulus, but also distinct cell groups involved in the processing of the positive or negative outcome of the response behavior (i.e., accuracy and reward or reinforcement information) (e.g., Amador, Schlag-Rey, and Schlag, 2000; Holroyd and Coles, 2002). Finally, with respect to directing attention to external and internal events, Frith and Frith (1999, 2001) have argued that the DMFC/AC integrates proprioceptive information from the self (e.g., emotions or intentions) with exteroceptive perceptions, processed by the STS, about the goal-directed behaviors and emotions of others. This integrative activity may be facilitated by the abundance of connections between the DMFC/AC and the STS (Morecraft, Guela, and Mesulam, 1993). Indeed, cell groups in and around BA 8/9 may be especially well connected to the STS (Ban, Shiwa, and Kawamura, 1991). We have described this putative facility for the integration of proprioceptive self-information with exteroceptive social perceptions as a social executive function (SEF) of the dorsal medial frontal cortex and anterior cingulate system (Mundy, 2003). Hypothetically, this SEF utilizes the DMFC/AC facility for the maintenance of representation of multiple goals in working memory to compare and integrate the actions of self and others. This integration gives rise to the capacity to infer the intentions of others by matching them with representations of selfinitiated actions or intentions (Stich and Nichols, 1992). Once this integration begins to occur in the DMFC/AC, a fully functional, adaptive, human social-cognitive system emerges with experience (Frith and Frith, 1999, 2001).

A developmental integration There is now a substantial and growing corpus of information on the biobehavioral processes involved in the development of gaze-following, RJA, and other joint-attention skills. Nevertheless, the current status of this literature may be best characterized as raising as many questions as it has answered. For example, some behavioral research suggests that IJA emerges before gaze following and RJA (Carpenter, Nagell, and Tomasello, 1998). Alternatively, the hypothesis that gaze following may be associated with an early-arising posterior/parietal attention system and observations of rudimentary aspects of gaze following in the first 3–6 months of life (e.g., Hood, Willen, and Driver, 1998; Morales, Mundy, and Rojas, 1998; Morales et al., 2000b) suggests that this aspect of joint attention may emerge before IJA (e.g., Mundy, Card, and Fox, 2000). It seems likely that such basic developmental issues will need to be resolved before we can hope to understand the full complexity of the ontogeny of joint attention. Another significant set of questions pertains to the nature and nurture of joint attention development. Is it

sufficient, for example, to ascribe the development of joint attention to an innately determined unfolding of a modular system of neural components that are specific to socialinformation processing and social behavior (Adolphs, 2001; Baron-Cohen, 1995; Emery, 2000; Farah et al., 2000)? Or, do more general transactional and learning processes also play a vital role in joint-attention development (Corkum and Moore, 1998; Karmaloff-Smith, 1992)? Are these truly divergent views, or do they provide complimentary perspectives that offer useful information on different aspects of development in this domain? There is currently insufficient information to resolve these questions and to build a truly veridical model of the development of gaze-following and joint-attention development in infancy. Nevertheless, the ultimate description of such a model is an important goal for both basic and applied developmental research. To contribute to this goal, we outline some of the assumptions and questions that guide our own thinking and research in this area. To start, following Karmaloff-Smith (1992) we assume that initial perceptual biases organize attention and behavior to insure that neonates actively engage in face and socialinformation processing (Mundy and Burnette, 2005; Mundy and Neal, 2001). Then, in an experience-dependent fashion (Greenough, Black, and Wallace, 1987), neural systems begin to organize in response to the relative preponderance of social information input in the first months of life. The nature of the mechanisms that give rise to these initial biases and how these initial biases interact with neural development is one of the essential issues that remain to be resolved in the study of early development (Karmaloff-Smith, 1992: Nelson, 2001). Nevertheless, at present, we assume that initial perceptual biases need not necessarily be specifically social in nature. Neonates preferentially orient to certain stimulus contours (e.g., curved lines) and stimulus contrasts that abound in faces (Fantz et al., 1975). Perhaps even more importantly, people, especially their faces, tend to be a rich and consistent source of synchronous sound and movement information, and this of type intersensory redundancy is a powerful elicitor of attention and information processing in young infants (Bahrick and Lickliter, 2002). Moreover, there is some evidence that intersensory redundancy may facilitate gaze-following-related task performance (e.g., Flom and Pick, 2003). Of course, it may be that there are also early predilections for social stimuli. The human voice itself may be a prepotent stimulus for infant orienting (Alegria and Noirot, 1978), and there may be intrinsic affective motivation factors that tend to predispose neonates to socialinformation processing (Trevarthen and Aitken, 2001). Related to these considerations, it is also important to recognize that parents’ activity is presented to infants in a fashion that may promote and reward social-information processing as part of their nurturing behaviors. Certainly, in

mundy and van hecke: neural systems, gaze following, and joint attention

829

relatively short order, social learning during nurturing interactions is likely to make social-information processing of parents a preferred activity for many infants. All this seems quite reasonable, but we have nowhere near enough information as yet to know how, or if, any of these mechanisms play a fundamental role in organizing early socialinformation processing. Be that as it may, our model holds that, in the next step of development, social information, motivation, and learning begin to interact with parietal attention-regulation and spatial-orienting processes to begin to bootstrap the gazefollowing system by approximately 3 months of age. In particular, the ability to disengage attention, shift attention, and determine the relative spatial orientation of stimuli are all parietal functions that likely contribute to the initial phases of infants’ ability to follow gaze. Interestingly, recent research suggests that infants at risk for autism may have particular difficulty with stimulus visual disengagement in the first year of life, which may contribute to impairments in RJA and other aspects of joint-attention development (Landry and Bryson, 2004). Nevertheless, while parietal attentionregulation mechanisms involve processes that may be necessary for gaze following to emerge, these mechanisms may not be sufficient. Another important mechanism that may be necessary for the early emergence of gaze following is imitation. Neonates and infants have the propensity to imitate facial movements (see Meltzoff and Moore, 1997). Meltzoff and Decety (2003) also have suggested that imitation involves the types of integrated coding between perception of self and perception of others that Frith and Frith (1999, 2001) have suggested is critical to the development of social cognition. Moreover, Meltzoff and Decety (2003) have presented an important research review that indicates that the neural activity associated with imitation clusters in the STS and parietal lobes, as well as the dorsal cortical supplementary motor areas (BA 8/9). In particular, they note that the abundance of mirror neurons in STS and parietal lobes may potentiate the role of these cortical regions in the mediation of imitation. Mirror neurons are a specific class of motor neurons that are involved both when an individual performs a particular action and when an individual observes the same action performed by another person (Gallese and Goldman, 1998). Thus the proposed cortical location of the neural mediators of imitation overlaps with the systems that are thought to mediate the development of gaze following and RJA. Indeed, from a behavioral–task analytic point of view, this conclusion makes a great deal of sense. Gaze following and RJA basically involve copying the eye movements and/or head turn of a social partner. Alternatively, initiating-joint-attention behaviors do not appear to involve a discernable social copying component. Indeed, RJA development has been observed to display a significant path of

830

emotion⁄cognition interactions

association with imitation development in a longitudinal study of typically developing infants. However, in that study there were no observations of significant associations between the development of IJA and imitation (Pomares et al., 2003). Thus it may be useful to better understand the degree to which gaze following/RJA and imitation reflect common and unique biobehavioral processes in early development. Indeed, the importance of this issue has long been recognized in research on autism because this syndrome involves deficits in imitation as well as joint-attention skill development (e.g., Rogers and Pennington, 1991). In the meantime, though, we assume that gaze following and RJA, but not necessarily IJA, are special forms of imitation. Temporal/parietal processes involved in imitation, as well as attention regulation, may play a role in the initial phase of development of this behavior. This initial phase of development may be followed by two “learning” phases (see figure 50.2). In one phase, infants learn to improve the accuracy and consistency of their gaze-following/RJA. For example, the contingent reward of seeing something interesting or receiving social praise after gaze following/RJA may increase the accuracy and efficiency of this behavior (Corkum and Moore, 1998). The course of development here, though, may be better characterized as a reduction in error trials (head or eye turns in a direction opposite to the social partner), rather than an increase in correct gazefollowing/RJA trials between 6 and 10 months of age. Although 6-month-olds display evidence of gaze-following ability (Morales, Mundy, and Rojas, 1998), they certainly display incremental improvements in the accuracy of this behavior from 6 to 8, 10, 12, and 15 months (Morales et al., 2000a). In unpublished observations of this longitudinal progression, it appeared, though, that the initial rates of accurate head turns remained constant over these ages, but the number of inaccurate head turns significantly decreased. This progression may be viewed as a shift in balance from imitative control systems to more control by systems that integrate accurate spatial interpretation of the direction of gaze (see Butterworth and Jarrett, 1991). Initial imitative control may not ensure accurate gaze-following responses, but it does ensure the expression of a pattern of behavior that can be culled or shaped through experience to yield a more useful pattern. This “learning to” process may be mediated by parietal spatial-orienting systems, perhaps in accord with dorsal error-detection systems. Those interested in this possibility may find it useful to refer to Holroyd and Coles (2002) to understand how the anterior cingulate, dorsal medial cortex, and parietal cortex may be integrated in the learning of this critical motor-action sequence. Of course, we should also be aware of the possibility that, in addition to being mediated by neural systems, this learning process serves to organize a subset of social brain processes through experience-expectant mechanisms (Greenough,

Multi-process Model of Joint Attention Spatial Processing

Imitation

Simulation

Representation Attention Regulation Motivation Social Reward

Theory of Mind

Joint Attention 3-12 months

Self-Other Integration Experience

“Learning To” Period 0 to 18 months

“Learning From” Social Cognitive Period From 10 months

Figure 50.2 In the first year, the development of joint attention involves the “learning to” integration of executive, motivation, and imitation processes to support the routine, rapid, and efficient (error-free) execution of patterns of behavior that enable infants to coordinate overt aspects of visual attention with partners in social interaction. In the later part of the first year and the second year, infants can better monitor their own experience and integrate it with information about the social partners during joint-attention events. This ability provides a critical source of information to the infants about the convergence and divergence of self and others’ experience and behavior during sharing of information in social

interactions. Theoretically, this provides the stage for the “learning from” phase of joint-attention development that allows the infant’s incipient understanding of sharing covert aspects of attention in the first year to become elaborated and integral to an understanding of the social coordination of covert aspects of attention (e.g., intentions), as when social partners coordinate attention to psychological phenomena, such as ideas, intentions, or emotions. Thus early development of joint attention and social cognition may be characterized by a transition from learning to share overt attention, which provides the foundation for the later but parallel development—the ability to share covert intentions.

Black, and Wallace, 1987). Only well-designed neurodevelopmental studies of the emergence of gaze-following behavior will be able to address this possibility. Related to the experience-expectant hypothesis that we have outlined is the possibility that infants not only learn to control gaze following/RJA, but also learn from engagement in this pattern of motor behavior (see figure 50.2). That is, infants’ own behavior becomes a critical source of information for subsequent executive and conceptual development (e.g., Piaget, 1952). In this regard, consider the heuristic set of observations from comparative research reviewed by Calder and colleagues (2002). Different sets of cells in the STS of macaque monkeys appear to contribute to the processing of gaze direction versus the processing of the direction and orientation of limb movements (Perrett et al., 1992). However, a subset of limb-movement cells appears to be modulated by activity of the gaze-following system (Jellema et al., 2000). Jellema and colleagues (2000) interpret these data to suggest that the combined analysis of direction of visual attention and body movements of others provides an important source of information that gives rise to the capacity to detect intentionality in others. Translated to human development, this interpretation suggests that gaze following does not occur in isolation, but rather as an

integrated element of processing of additional dimensions of social behavioral information about others. This suggestion is consistent with the observation that, in addition to its social-information-processing specialization, the STS may also include polysensory areas involved in attending to and processing information synchronously from multiple modalities (Hikosaka, 1993). The integrated processing of others’ direction of gaze, limb/postural direction, and vocal behavior ultimately may be an important, if not critical, source of information that enables the infant to learn about self, others, and social intentions. For example, with respect to gaze following, Jellema and colleagues (2000) suggested that one major lesson learned from gaze following is “Where the eyes go, behavior follows.” It is important to understand, though, that this type of epistemological development, social or nonsocial, may only emerge after extensive repetitive experience with motor actions and behaviors, as well as with maturation of cognitive and executive systems (Piaget, 1952; see KarmaloffSmith, 1992, for related arguments). That is, first the infant must master the behavior skill. Then, once this behavior pattern becomes routinized, the behavior pattern itself becomes fodder for higher-order processing and epistemological development. This hypothesis is consistent with

mundy and van hecke: neural systems, gaze following, and joint attention

831

recent observations that suggest that the social cognitive component associated with gaze following may only be evident several months after the onset of this behavior (Brooks and Meltzoff, 2002; Woodward, 2003). Perhaps the best evidence of “learning from” processes in joint attention has been provided by Meltzoff and Decety (2005), who reported that 12-month-olds were inhibited in RJA performance when the social partner was blindfolded but 10month-olds followed head and “blinded” direction of gaze as well as if the social partner were not blindfolded. These data replicate Brooks and Meltzoff’s (2002) observation that RJA is based on greater social cognitive awareness in 12month-olds than 10-month-olds. However, Meltzoff and Decety (2005) also reported that after 10-month-olds were given experience with blindfolds on themselves they were able to project this new self-related information onto social partners and decreased their responding when social partners were blindfolded in an RJA task. Thus they demonstrated some elements of learning about others in “learning from” their own experience and simulated the experience of others in a joint-attention task. Putting all of these assumptions together yields the following preliminary developmental model: Gaze following/RJA emerges between 3 and 6 months of age as a special form of imitation that is regulated to a significant degree by an early-developing parietal attention system with contributions from social perception systems of the STS. Learning, with respect to this skill, occurs between 3 and 10 months so that parietal spatial and attention-regulation processes increasingly regulate imitative responses, in order to yield a reduction in incorrect responses and more consistently correct spatially directed responses. With practice and cognitive maturation (Case, 1987), a lower ratio of available processing capacity needs to be allocated by the child to support accurate gaze following/RJA. As increased processing capacity becomes available, new potentials for integrative cognition arise. This aspect of our model leads to the testable prediction that speed of information processing in gaze-following tasks increases with age. Specifically, the latency between observing a gaze or head turn stimulus and a correct response should decrease significantly in the second six months of life. With the increase in processing capacity (Case, 1987), infants begin to become increasingly capable of integrating their own proprioceptive sense of self as perceiver and actor with exteroceptive information about a related pattern of behaviors of an external agent in the immediate context of displaying gaze-following/RJA behavior. This integrated processing of information from self and other would involve processes currently ascribed to a more anterior dorsal medial cortical system (Frith and Frith, 1999, 2001; Mundy, 2003). Although the developmental period

832

emotion⁄cognition interactions

of this integrative processing phase is not clear, our current conjecture is that this occurs in the 6- to 15-month period of development. That is, the “learning from” phase overlaps considerably with the “learning to” phase of development. In the “learning from” phase, the integrated self and other processing provides information uniquely well suited to the simulation of the intentions of other through the experience of coordinated self-other triadic social interactions. “Simulation theory” suggests that individuals use their awareness (i.e., representations) of their own mental processes to simulate and analyze the intentions of others (Stich and Nichols, 1992). That is, with development people learn to use self-knowledge, derived from self-monitoring, to extrapolate and make inferences about the covert psychological processes that contribute to the behaviors of other people. Several lines of research suggest that the dorsalmedial frontal cortical facility for self-monitoring (see Mundy, 2003, for a review), together with information input from the temporal/parietal systems for processing the behavior of others, may be integrated to yield the information that facilitates social cognitive simulation (Frith and Frith, 1999, 2001). Theoretically, beginning in infancy, this process occurs in vivo in a variety of social interactive contexts. Moreover, the context of triadic social-attentioncoordination interactions that involve the perception of self and other vis-à-vis a common object or event is thought to provide an especially powerful context for simulating the common psychological experience of self and other (Mundy, Sigman, and Kasari, 1993). One unique aspect of the foregoing account of the nature and role of gaze following in early development is the suggestion that processing capacity and processing speed may affect development within this domain. This assumption is derived from the repeated observation that the increasing capacity to process two or more pieces of information rapidly contributes to general aspects of social, intellectual, and language development in infancy (e.g., Case, 1991; Benasich et al., 2002; Rose and Feldman, 1997). An illustration of the importance of processing speed in development may be gleaned from considering the history of personal computing. In the early 1980s, when PCs began to be widely available, they had relatively slow processing speeds, so that the computer could display activity derived from one piece of software at a time. With increases in processing speed, though, computers were increasingly able to switch rapidly back and forth between software applications, so that, on the monitor, it appeared as though several programs (sets of information) were running at one time. These increases in speed enabled the development of the types of multitasking operating systems (e.g., Windows) that now allow us to rapidly hold vast numbers of digital representa-

tions in computer memory so that we can compare and contrast vast amounts of information on our computer displays. It is not too much of an extrapolation from this illustration to suggest that an analogous increase in the human processing speed in the first years of life plays a role in the ability to compare and contrast self and other information in a fashion that contributes to new social-cognitive, conceptual structures. We assume that similar “learning to” and “learning from” phases of development occur for IJA as well (see figure 50.2). However, here imitative processes may not play as primary a role in the early “learning to” phase. Instead, infants may initially be entrained to engage in basic IJA behaviors, such as alternating gaze, when the vocalizations and positive affect of their caregivers stimulate social orienting in passive or supported joint-engagement interactions. It may also be the case that “learning from” gaze following has already begun to establish a rudimentary social-cognitive foundation by 6 to 8 months that facilitates the development of this behavior. As we have noted, though, unlike gaze following and its relatively reflexive control systems, true initiating joint attention involves the spontaneous generation of a pattern of behavior. Spontaneous visual orienting behaviors, like IJA, may be more demanding than reflexive orienting behaviors because they involve the executive integration of motivation processes, memory (e.g. representations of others), and attention-switching processes perhaps regulated by a more anterior cortical system (Rothbart and Posner, 2001). Thus this domain of behavior may develop later than gaze following/RJA. Nevertheless, in the “learning from” phase of IJA there may be a return for the greater processing demands of this type of behavior, in that infants and children may learn something different from behaviors they initiate versus those they copy. In the internal simulation environment concurrent with IJA behaviors, infants may learn that their behavior changes the behavior of others and that they can direct the attention of others to objects. They may also learn that when they express interest or affect regarding some object or event, others take note, and they may share or react to their expression of behavior (Bates, 1976). Thus self-generated IJA behavior may provide information about the self in interaction with others that is different than the information provided by patterns of responsive gaze-following/RJA behavior. This self-generated action may also be regulated to a greater degree by anterior systems such as the dorsal medial frontal cortex. However, like gaze following/RJA, the social simulations that occur in the context of IJA may be mediated, in part, by mirror neurons. In addition to the STS and parietal cortex, the supplementary motor cortex, which overlaps with the DMFC areas implicated in IJA and social cognition, is rich in mirror neurons (Rizzolatti and Arbib, 1998). Moreover, as previously noted, DMFC is richly interconnected with the STS

and parietal cortices (e.g., Morecraft, Geula, and Mesulam, 1993). These observations suggest that, although somewhat different neural systems and social learning may be involved in gaze following/RJA and IJA, there is considerable potential for biobehavioral integration across the development of these joint-attention skills. Understanding the details of integrated neurodevelopment of these important domains of early social behavior remains a substantial goal for developmental science. In addition, information provided by the development of both types of joint-attention skills makes a significant contribution to our understanding of the biobehavioral development of social cognition.

REFERENCES Acra, F., P. Mundy, A. Claussen, K. Scott, and K. Bono, 2003. Infant joint attention and social outcomes in 6- to 7-year-old atrisk children. Paper presented at the Society for Research in Child Development, Tampa, FL, April. Adolphs, R., 2001. The neurobiology of social cognition. Curr. Opin. Neurobiol. 11:231–239. Alegria, J., and E. Noirot, 1978. Neonate orientation behavior towards human voice. Int. J. Behav. Dev. 1:292–312. Amador, N., M. Schlag-Rey, and J. Schlag, 2000. Reward predicting and reward detecting neuronal activity in the primate supplementary eye field. J. Neurophyisol. 84:2166–2170. Astafiev, S., G. Shulman, C. Stanley, A. Snyder, D. Essen, and M. Corbetta, 2003. Functional organization of human intraparietal and frontal cortex for attending, looking and pointing. J. Neurosci. 23:4689–4699. Atkinson, J., B. Hood, J. Wattam-Bell, and O. Braddick, 1992. Changes in infants’ ability to switch attention in the first three months of life. Perception 21:643–653. Bahrick, L., and R. Lickliter, 2002. Intersensory redundancy guides early perceptual and cognitive development. Adv. Child Dev. Behav. 30:153–187. Baldwin, D. A., 1995. Understanding the link between joint attention and language. In C. Moore and P. J. Dunham, eds., Joint Attention: Its Origins and Role in Development, 131–158. Hillsdale, NJ: Lawrence Erlbaum. Ban, T., T. Shiwa, and K. Kawamura, 1991. Cortico-cortical projections from the prefrontal cortex to the superior temporal sulcal area (STS) in the monkey studied by means of HRP method. Arch. Ital. Biol. 129:259–272. Bard, K., K. Platzman, B. Lester, and S. Suomi, 1992. Orientation to social and nonsocial stimuli in neonatal chimpanzees and humans. Infant Behav. Dev. 15:43–56. Baron-Cohen, S., 1995. Mindblindness. Cambridge, MA: MIT Press. Baron-Cohen, S., D. Baldwin, and M. Crowson, 1997. Do children with autism use the speaker’s direction of gaze strategy to crack the code of language? Child Dev. 68:48–57. Baron-Cohen, S., H. Ring, S. Wheelwright, E. Bullmore, M. Brammer, A. Simmons, and S. Williams, 1999. Social intelligence in the normal and autistic brain: An fMRI study. Eur. J. Neurosci. 11:1891–1898. Bates, E., 1976. Language and Context: The Acquisition of Performatives. New York: Academic Press. Benasich, A., J. Thomas, N. Choudhury, and P. Leppaene, 2002. The importance of rapid auditory processing abilities to early

mundy and van hecke: neural systems, gaze following, and joint attention

833

language development: Evidence form converging methodologies. Dev. Psychobiol. 40:278–292. Birrell, J., and V. Brown, 2000. Medial-frontal cortex mediates perceptual attention set shifting in the rat. J. Neurosci. 20:4320– 4324. Block, J., P. Mundy, Y. Pomares, A. Vaughan, C. Delgado, and Y. Gomez, 2003. Different developmental profiles of joint attention skills from 9 to 18 months. Paper presented at the Society for Research in Child Development, Tampa, FL, April. Bono, M., T. Daley, and M. Sigman, 2004. Joint attention moderates the relation between intervention and language development in young children with autism. J. Autism Dev. Disord. 34:495–505. Bretherton, I., 1991. Intentional communication and the development of an understanding of mind. In D. Frye and C. Moore, eds., Children’s Theories of Mind: Mental States and Social Understanding, 49–75. Hillsdale, N J: Lawrence Erlbaum. Brooks, R., and A. Meltzoff, 2002. The importance of eyes: How infants interpret adult looking behavior. Dev. Psychol. 38: 958–966. Brothers, L., 1990. The social brain: A project for integrating primate behavior and neurophysiology in a new domain. Concepts Neurosci. 1:27–51. Bruner, J. S., 1975. From communication to language: A psychological perspective. Cognition 3:255–287. Bush, G., P. Luu, and M. Posner, 2000. Cognitive and emotional influences in the anterior cingulate cortex. Trends Cogn. Sci. 4: 214–222. Butterworth, G., and N. Jarrett, 1991. What minds have in common is space: Spatial mechanisms in serving joint visual attention in infancy. Br. J. Dev. Psychol. 9:55–72. Calder, A., A. Lawrence, J. Keane, S. Scott, A. Owen, I. Christoffels, and A. Young, 2002. Reading the mind from eye gaze. Neuropsychologia 40:1129–1138. Cambell, R., C. Heywood, A. Cowey, M. Regard, and T. Landis, 1990. Sensitivity to eye gaze in prosopagnosic patients and monkeys with superior temporal sulcus ablation. Neuropsychologia 28:1123–1142. Caplan, R., H. Chugani, C. Messa, D. Guthrie, M. Sigman, J. Traversay, and P. Mundy, 1993. Hemispherectomy for early onset intractable seizures: Presurgical cerebral glucose metabolism and postsurgical nonverbal communication patterns. Dev. Med. Child Neurol. 35:582–592. Capps, L., M. Sigman, and P. Mundy, 1994. Attachment security in children with autism. Dev. Psychopathol. 6:249–261. Carpenter, M., K. Nagell, and M. Tomasello, 1998. Social cognition, joint attention, and communicative competence from 9 to 15 months of age. Monogr. Soc. Res. Child Dev. 63(No. 4, Serial No. 255). Carpenter, M., B. Pennington, and S. Rogers, 2002. Interrelations among social-cognitive skills in young children with autism. J. Autism Dev. Disord. 32:91–106. Case, R., 1987. The structure and process of intellectual development. Int. J. Psychol. 22:571–607. Case, R., 1991. Stages in the development of young children’s first sense of self. Dev. Rev. 11:210–230. Charman, T., S. Baron-Cohen, J. Swettenham, G. Baird, A. Cox, and A. Drew, 2000. Testing joint attention, imitation, and play infancy precursors to language and theory of mind. Cogn. Dev. 15:481–498. Charwarska, K., A. Klin, and F. Volkmar, 2003. Automatic attention cuing through eye movement in 2-year-old children with autism. Child Dev. 74:1108–1123.

834

emotion⁄cognition interactions

Chugani, H., M. Phelps, and J. Mazziotta, 1987. Positron emission tomography study of human brain functional development. Ann. Neurol. 22:487–497. Clohessy, A., M. Posner, and M. Rothbart, 2001. Development of the functional visual field. Acta Psychol. (Amet.) 106:51–68. Corkum, V., and C. Moore, 1998. The origins of joint visual attention in infants. Dev. Psychol. 34:28–38. Craik, F., T. Moroz, M. Moscovich, D. Stuss, G. Winocur, E. Tulving, and S. Kapur, 1999. In search of the self: A positron emission tomography study. Psychol. Sci. 10:26–34. Dawson, G., J. Munson, A. Estes, J. Osterling, J. McPartland, K. Toth, L. Carver, and R. Abbott, 2002. Neurocognitive function and joint attention ability in young children with autism spectrum disorder versus developmental delay. Child Dev. 73: 345–358. Dawson, G., S. Webb, G. Schellenberg, S. Dager, S. Friedman, E. Ayland, and T. Richards, 2002. Defining the broader phenotype of autism: Genetic, brain, and behavioral perspectives. Dev. Psychopathol. 14:581–612. D’Entremont, B., S. Hains, and D. Muir, 1997. A demonstration of gaze following in 3- to 6-month-olds. Infant Behav. Dev. 20: 569–572. Eacott, M., C. Heywood, C. Gross, and A. Cowey, 1993. Visual discrimination impairments following lesions of the superior temporal sulcus are not specific for facial stimuli. Neuropsychologia 31:609–619. Emery, N., 2000. The eyes have it: The neuroethology, function, and evolution of social gaze. Neurosci. Biobehav. Rev. 24:581– 604. Fantz, R., J. Fagan, and S. Miranda, 1975. Early visual selectivity as a function of pattern variables, previous exposure, age from birth and conception, and expected cognitive defect. In L. Cohen and P. Salapatek, eds., Infant Perception, Vol. 1, 96–117. New York: Academic Press. Farah, M., C. Rabinowitz, G. Quinn, and L. Grant, 2000. Early commitment of neural substrates for face recognition. Cogn. Neuropsychol. 17:117–123. Farroni, T., S. Massaccesi, and S. Francesca, 2002. Can the direction of gaze of another person shift the attention of a neonate? G. Ital. Psicol. 29:857–864. Faw, B., 2003. Prefrontal executive committee for perception, working memory, attention, long-term memory, motor control and thinking: A tutorial review. Consciousness Cogn. 12:83–139. Filipek, P., P. Accardo, G. Baranek, E. Cook, G. Dawson, B. Gordon, J. Gravel, C. Johnson, R. Kallen, S. Levy, N. Minshew, S. Ozonoff, B. Prizant, I. Rapin, S. Rogers, W. Stone, S. Teplin, R. Tuchman, and F. Volkmar, 1999. The screening and diagnosis of autism spectrum disorders. J. Autism Dev. Disord. 29:439–484. Flom, R., and A. Pick, 2003. Verbal encouragement and joint attention in 18-month-old infants. Infant Behav. Dev. 26:121– 134. Fombonne, E., 2003. The prevalence of autism. JAMA 289:87– 79. Frith, C., and U. Frith, 1999. Interacting minds: A biological basis. Science 286:1692–1695. Frith, U., and C. Frith, 2001. The biological basis of social interaction. Curr. Dir. Psychol. Sci. 10:151–155. Gallese, V., and A. Goldman, 1998. Mirror neurons and the simulation theory of mind-reading. Trends Cogn. Sci. 2:493–501. George, N., J. Driver, and R. Dolan, 2001. Seen gaze direction modulates fusiform activity and its coupling with other brain areas during face processing. NeuroImage 13:1102–1112.

Greenough, W., J. Black, and C. Wallace, 1987. Experience and brain development. Child Dev. 58:539–559. Griffith, E., B. Pennington, E. Wehner, and S. Rogers, 1999. Executive functions in young children with autism. Child Dev. 70:817–832. Hampson, R., T. Pons, T. Sanford, and C. Deadwyler, 2004. Categorization in monkey hippocampus: A possible mechanism for encoding information into memory. Proc. Natl. Acad. Sci. USA 101:3184–3189. Henderson, L., P. Yoder, M. Yale, and A. McDuffie, 2002. Getting the point: Electrophysiological correlates of protodeclarative pointing. Int. J. Dev. Neurosci. 20:449–458. Heyward, C., and A. Cowey, 1992. The role of the “face cell” area in the discrimination and recognition of faces by monkeys. Philos. Trans. R. Soc. London 335:31–38. Hikosaka, K., 1993. The polysensory region in the anterior bank of the caudal superior temporal sulcus of the macaque monkey. Biomed. Res. Tokyo 14:41–45. Hoffman, E., and J. Haxby, 2000. Distinct representation of eye gaze and identity in the distributed human neural system for face perception. Nature Neurosci. 3:80–84. Holroyd, C., and M. Coles, 2002. The neural basis of human error processing: Reinforcement learning, dopamine and the error related negativity. Psychol. Rev. 109:679–709. Hood, B., 1995. Shifts of visual attention in the human infant: A neuroscientific approach. In C. Rovee-Collier and L. Lipsitt, eds., Advances in Infancy Research. Norwood, NJ: Ablex. Hood, B., J. Willen, and J. Driver, 1998. Adult’s eyes trigger shifts of visual attention in human infants. Psychol. Sci. 9:131–134. Hooker, C., 2002. The neurocognitive basis of gaze perception: A model of social signal processing. Dissertation Abstracts International: Science and Engineering 63:2058. Jellema, T., C. Baker, B. Wicker, and D. Perrett, 2000. Neural representation for the perception of intentionality of actions. Brain Cogn. 44:280–302. Johnson, M., 1990. Cortical maturation and the development of visual attention in early infancy. J. Cogn. Neurosci. 2:81–95. Johnson, M., 1995. The inhibition of automatic saccades in early infancy. Dev. Psychobiol. 28:281–291. Johnson, M., and J. Morton, 1991. Biology and Cognitive Development: The Case of Face Recognition. Oxford, UK: Blackwell. Johnson, M., M. Posner, and M. Rothbart, 1991. Components of visual orienting in early infancy: Contingency learning, anticipatory looking, and disengaging. J. Cogn. Neurosci. 3:335– 344. Johnson, S., L. Baxter, L. Wilder, J. Pipe, J. Heiserman, and G. Prigatano, 2002. Neural correlates of self-reflection. Brain 125: 1808–1814. Kampe, K., C. Frith, R. Dolan, and U. Frith, 2001. Reward value of attractiveness and gaze. Nature 413:589. Karmaloff-Smith, A., 1992. Beyond Modularity: A Developmental Perspective on Cognitive Science. Cambridge, MA: MIT Press. Kasari, C., M. Sigman, P. Mundy, and N. Yirmiya, 1990. Affective sharing in the context of joint attention interactions of normal, autistic, and mentally retarded children. J. Autism Dev. Disord. 20:87–100. Kawashima, R., M. Sugiura, T. Kato, A. Nakamura, K. Hatano, K. Ito, H. Fuguda, S. Kojima, and K. Nakamura, 1999. The human amygdala plays an important role in gaze monitoring: A PET study. Brain 122:779–783. Kingstone, A., C.-K. Friesen, and M. Gazzaniga, 2000. Reflexive joint attention depends on lateralized cortical functions. Psychol. Sci. 11:159–166.

Landry, R., and S. Bryson, 2004. Impaired disengagement of attention in young children with autism. J. Child Psychol. Psychiatry 45:115–122. Leekam, S., and C. Moore, 2001. The development of joint attention in children with autism. In J. Barack, T. Charman, N. Yirmiya, and P. Zelazo, eds., The Development of Autism: Perspectives from Theory and Research, 105–130. Mahwah, NJ: Lawrence Erlbaum. Lord, C., H. Floody, D. Anderson, and A. Pickles, 2003. Social engagement in very young children with autism: Differences across contexts. Paper presented at the Society for Research in Child Development, Tampa, FL, April. Luu, P., T. Flaisch, and D. Tucker, 2000. Medial-frontal cortex in action monitoring. J. Neurosci. 20:464–469. Martin, J., 1996. Neuroanatomy: Text and Atlas, 2nd ed. New York: McGraw-Hill. Masten, A., and D. Coatsworth, 1998. The development of competence in favorable and unfavorable environments: Lessons from research on successful children. Am. Psychol. 53:205–220. McEvoy, R., S. Rogers, and R. Pennington, 1993. Executive function and social communication deficits in young autistic children. J. Child Psychol. Psychiatry 34:563–578. Meltzoff, A., and J. Decety, 2003. What imitation tells us about social cognition: A rapprochement between developmental psychology and cognitive neuroscience. Philos. Trans. R. Soc. London 358:491–500. Meltzoff, A., and J. Decety, 2005. The “like me” hypothesis: Neurosciences and developmental perspectives. Paper presented at the Biennial Meeting of the Society for Research in Child Development, April 7, 2005, Atlanta. Meltzoff, A., and M. Moore, 1997. Explaining facial imitation: A theoretical model. Early Dev. Parenting 6:179–192. Moore, C., 1996. Theories of mind in infancy. Br. J. Dev. Psychol. 14:19–40. Morales, M., P. Mundy, M. Crowson, R. Neal, and C. Delgado, 2005. Individual differences in infant attention skills, joint attention and emotion regulation. Int. J. Behav. Dev. 29:259–263. Morales, M., P. Mundy, C. Delgado, M. Yale, D. Messinger, R. Neal, and H. Schwartz, 2000a. Responding to joint attention across the 6- through 24-month age period and early language acquisition. J. Appl. Dev. Psychol. 21:283–298. Morales, M., P. Mundy, C. Delgado, M. Yale, R. Neal, and H. Schwartz, 2000b. Gaze following, temperament, and language development in 6 month olds: A replication and extension. Infant Behav. Dev. 23:231–236. Morales, M., P. Mundy, and J. Rojas, 1998. Following the direction of gaze and language development in 6-month-olds. Infant Behav. Dev. 21:373–377. Morecraft, R., C. Geula, and M. Mesulam, 1993. Architecture of connectivity within a cingula-fronto-parietal neurocognitive network for directed attention. Arch. Neurol. 50:279–284. Mundy, P., 1995. Joint attention and social-emotional approach behavior in children with autism. Dev. Psychopathol. 7:63–82. Mundy, P., 2003. The neural basis of social impairments in autism: The role of the dorsal medial-frontal cortex and anterior cingulate system. J. Child Psychol. Psychiatry 44:793–809. Mundy, P., and C. Burnette, 2005. Joint attention and neurodevelopment. In F. Volkmar, A. Klin, and R. Paul, eds., Handbook of Autism and Pervasive Developmental Disorders, vol. 3., 650–681. Hoboken, NJ: John Wiley. Mundy, P., J. Card, and N. Fox, 2000. EEG correlates of the development of infant joint attention skills. Dev. Psychobiol. 36: 325–338.

mundy and van hecke: neural systems, gaze following, and joint attention

835

Mundy, P., C. Delgado, J. Block, M. Venezia, A. Hogan, and J. Seibert, 2003. A Manual for the Abridged Early Social Communication Scales (ESCS). Available through the University of Miami Psychology Department, Coral Gables, Florida. Available from author at [email protected] Mundy, P., and A. Gomes, 1998. Individual differences in joint attention skills in the second year. Infant Behav. Dev. 21:469– 482. Mundy, P., C. Kasari, and M. Sigman, 1992. Joint attention, affective sharing, and intersubjectivity. Infant Behav. Dev. 15:377– 381. Mundy, P., C. Kasari, M. Sigman, and E. Ruskin, 1995. Nonverbal communication and early language in Down syndrome and in normally developing children. J. Speech Hear. Res. 38:157– 167. Mundy, P., and R. Neal, 2001. Neural plasticity, joint attention and a transactional social-orienting model of autism. Int. Rev. Ment. Retard. 23:139–168. Mundy, P., and M. Sigman, 1989. Specifying the nature of the social impairment in autism. In G. Dawson, ed., Autism: New Perspectives on Diagnosis, Nature, and Treatment, 3–21. New York: Guilford. Mundy, P., and M. Sigman, 2006. Joint attention, social competence and developmental psychopathology. In D. Cicchetti and D. Cohen, eds., Developmental Psychopathology, 2nd ed., vol. 1: Theory and Methods, 293–332. Hoboken, NJ: Wiley. Mundy, P., M. Sigman, and C. Kasari, 1993. The theory of mind and joint attention deficits in autism. In S. Baron-Cohen, H. Tager-Flusberg, and D. Cohen, eds., Understanding Other Minds: Perspectives from Autism, 181–203. Oxford, UK: Oxford University Press. Mundy, P., M. Sigman, and C. Kasari, 1994. Joint attention, developmental level, and symptom presentation in young children with autism. Dev. Psychopathol. 6:389–401. Mundy, P., M. Sigman, J. Ungerer, and T. Sherman, 1986. Defining the social deficits of autism: The contribution of nonverbal communication measures. J. Child Psychol. Psychiatry 27: 657–669. Neal, R., P. Mundy, A. Claussen, S. Mallik, K. Scott, and F. Acra, 2006. The relations between infant joint attention skill and cognitive and language outcome in at-risk children. Submitted. Nelson, C., 2001. The development and neural basis of face recognition. Infant Child Dev. 10:3–18. Norman, D., and T. Shallice, 1986. Attention to action: Willed and automatic control of behavior. In R. Davidson, G. Schwartz, and D. Shapiro, eds., Consciousness and Self-Regulation, 1–18. New York: Plenum. Perrett, D., J. Heitenen, M. Oram, and P. Benson, 1992. Organization and functions of cells responsive to faces in the temporal cortex. Philos. Trans. R. Soc. London 335:23– 30. Piaget, J., 1952. The Origins of Intelligence in Children. New York: Norton. Pomares, Y., P. Mundy, A. Vaughan, J. Block, and C. Delgado, 2003. On the relations between infant joint attention, imitation, and language. Paper presented at the Society for Research in Child Development, Tampa, FL, April. Posner, M., and S. Petersen, 1990. The attention system of the human brain. Annu. Rev. Neurosci. 13:25–42. Puce, A., T. Allison, S. Bentin, J. Gore, and G. McCarthy, 1998. Temporal cortex activation in humans viewing eye and mouth movements. J. Neurosci. 18:2188–2199.

836

emotion⁄cognition interactions

Rizzolatti, G., and M. Arbib, 1998. Language within our grasp. Trends Neurosci. 21:188–194. Rogers, S., S. Hepburn, T. Stackhouse, and E. Wehner, 2003. Imitation performance in toddlers with autism and those with other developmental disorders. J. Child Psychol. Psychiatry 44: 763–781. Rogers, S., and B. Pennington, 1991. A theoretical approach to the deficits in infantile autism. Dev. Psychopathol. 6:635– 652. Rose, S., and J. Feldman, 1997. Memory and speed: Their role in the relation of infant information processing to later IQ. Child Dev. 68:630–641. Rothbart, M., 1981. Measurement of temperament in infancy. Child Dev. 52:569–578. Rothbart, M., and M. Posner, 2001. Mechanism and variation in the development of attention networks. In C. Nelson and M. Luciana, eds., Handbook of Developmental Cognitive Neuroscience, 353–363. Cambridge, MA: MIT Press. Rothbart, M., M. Posner, and J. Rosicky, 1994. Orienting in normal and pathological development. Dev. Psychopathol. 6:635– 652. Rushworth, M., K. Hadland, T. Paus, and P. Siplia, 2002. Role of the human medial frontal cortex in task switching: A combined fMRI and TMS study. J. Neurophysiol. 87:2577– 2592. Russell, T., K. Rubia, E. Bullmore, W. Soni, J. Suckling, M. Brammer, A. Simmons, and T. Sharma, 2000. Exploring the social brain in schizophrenia: Left prefrontal underactivation during mental state attribution. Am. J. Psychiatry 157:2040– 2042. Scaife, M., and J. Bruner, 1975. The capacity for joint visual attention in the infant. Nature 253:265–266. Seibert, J. M., A. E. Hogan, and P. C. Mundy, 1982. Assessing interactional competencies: The Early Social Communication Scales. Infant Ment. Health J. 3:244–245. Sheinkopf, S., P. Mundy, A. Claussen, and J. Willoughby, 2004. Infant joint attention skills and preschool behavior outcomes in at risk children. Dev. Psychopathol. 16:273–291. Sigman, M., and E. Ruskin, 1999. Continuity and change in the social competence of children with autism, Down syndrome, and developmental delay. Monogr. Soc. Res. Child Dev. 64(Serial No. 256):1–108. Simion, F., E. Valenza, C. Umiltà, and B. Dalla, 1998. Preferential orienting to faces in newborns: A temporal-nasal asymmetry. J. Exp. Psychol. Hum. Percept. Perform. 24:1399–1405. Stich, S., and S. Nichols, 1992. Folk psychology: Simulation versus tacit theory. Mind Lang. 7:29–65. Striano, T., and D. Stahl, 2005. Sensitivity to triadic attention in early infancy. Dev. Sci. 8:333–343. Stuss, D., T. Shallice, M. Alexander, and T. Picton, 1995. A multidimensional approach to anterior attention functions. In J. Grafman, K. Holyoak, et al., eds., Structure and Function of the Human Prefrontal Cortex. Ann. NY Acad. Sci. 769:191–211. New York: New York Academy of Sciences. Swettenham, J., S. Baron-Cohen, T. Charman, A. Cox, G. Baird, A. Drew, L. Rees, and S. Wheelwright, 1998. The frequency and distribution of spontaneous attention shifts between social and nonsocial stimuli in autistic, typically developing, and nonautistic developmentally delayed infants. J. Child Psychol. Psychiatry 39:747–753. Tomasello, M., 1995. Joint attention as social cognition. In C. Moore and P. Dunham, eds., Joint Attention: Its Origins and Role in Development, 103–130. Hillsdale, NJ: Lawrence Erlbaum.

Trevarthen, C., and K. Aitken, 2001. Infant intersubjectivity: Research, theory and clinical applications. J. Psychol. Psychiatry 42:3–48. Ulvund, S., and L. Smith, 1996. The predictive validity of nonverbal communicative skills in infants with perinatal hazards. Infant Behav. Dev. 19:441–449. Vaughan, A., P. Mundy, J. Block, C. Burnette, C. Delgado, Y. Gomez, J. Meyer, R. Neal, and Y. Pomares, 2003. Child, caregiver, and temperament contributions to infant joint attention. Infancy 4:603–616. Vaughan Van Hecke, A., P. Mundy, C. F. Acra, J. Block, C. Delgado, M. Venezia, J. Meyer, R. Neal, and Y. Pomares, 2007. Infant joint attention, temperament, and social competence in preschool children. Child Dev. 78:53–69.

Wicker, B., F. Michel, M. Henaff, and J. Decety, 1998. Brain regions involved in the perception of gaze. NeuroImage 8:221– 227. Williams, J., G. D. Waiter, O. Perra, D. I. Perrett, and A. Whitten, 2005. An f MRI study of joint attention experience. NeuroImage 25:133–140. Woodward, A., 2003. Infants’ developing understanding of the link between looker and object. Dev. Sci. 6:297–311. Zelazo, P., L. Qu, and U. Muller, 2005. Hot and cool aspects of executive function: Relations in early language development. In W. Schneider, R. Schumann-Hegsteler, and B. Sodian, eds., Young Children’s Cognitive Development: Interrelations Among Executive Functioning, Working Memory, Verbal Ability and Theory of Mind, 71– 93. Mahwah, NJ: Lawrence Erlbaum.

mundy and van hecke: neural systems, gaze following, and joint attention

837

51

The Biology of Temperament: An Integrative Approach NATHAN A. FOX, HEATHER A. HENDERSON, KORALY PÉREZ-EDGAR, AND LAUREN K. WHITE

Temperament refers to a style or pattern of behavior that characterizes young infants’ ongoing interaction with the world. Individual differences in this behavioral style are thought to reflect biologically based dispositions (Goldsmith et al., 1987), and a large part of the current research is rooted in models of individual differences that emphasize the role of the nervous system in the expression of behavioral styles (Kagan, 1994; Rothbart and Derryberry, 1981). The precise nature of these biological differences has only recently been investigated. This work reflects a developing synthesis of research in the neurosciences and developmental psychology leading toward an understanding of the biology of behavior. The present chapter will provide the reader with a review of research into the underlying biology of infant reactivity and regulation, two factors that comprise the constituent parts of Rothbart’s model of infant temperament (Rothbart and Derryberry, 1981). We will present data from our research program on behavioral inhibition, as a “model system” for examining the biology of infant reactivity and regulation, as well as factors that may influence the stability of individual differences in these two components of temperament. In addition, we will highlight recent work in molecular genetics and gene X environment interactions as an attempt to further understand the developmental processes involved in stability and change in child temperament.

Psychobiological models of temperament The 20th-century neurophysiologist Pavlov noted differences in speed of conditioning and suggested that these effects were a result of individual differences in strength of the nervous system (Pavlov, 1927). Eastern European personality psychologists subsequently adopted this idea and constructed typologies of personality based upon differences in strength of the nervous system (e.g., Strelau, 1983). Gray (1982, 1987) later articulated a motivation-based model of individual differences in temperament in which he proposed three fundamental motivation systems, each defined by a set of behavioral input-output relations and associated with a particular subsystem in the brain. These three systems are

the behavioral inhibition system, the fight-flight system, and the behavioral approach system. There are a number of comprehensive reviews of the comparative behavioral evidence and the underlying neural bases for each system (Gray, 1991). Much of the behavioral work examining Gray’s model and its relation to temperament/personality has centered on the use of Eysenck’s model (Eysenck and Eysenck, 1985). This model makes specific predictions about the susceptibility to conditioning and arousal of introverts and extraverts, as well as the efficacy of certain types of reinforcement as a function of personality. Mary Rothbart expanded on the ideas of Pavlov and the Eastern European personality psychologists, as well as Gray’s notions of activation and inhibition, to articulate a psychobiological model of infant temperament. Rothbart’s theory of temperament has as its foundation the notion that infants differ early in life in the manner in which they respond to sensory stimulation (Rothbart and Derryberry, 1981). Infants also vary widely in their ability to achieve homeostasis or return to baseline subsequent to a reactive response. Interest in individual differences in infant temperament has paralleled research in the development and individual differences in emotional expression and control (N. Fox, 1994a; Henderson and Fox, 2007). Goldsmith and Campos (1986), for example, posit that temperament may be viewed as individual differences among infants in the expression of specific emotions in reaction to specific contexts. For example, some infants may be predisposed to respond with the emotion of fear to novelty or uncertainty, a hallmark of the temperamental style of behavioral inhibition (N. A. Fox et al., 2001; Kagan, Reznick, and Snidman, 1987).

An integrative approach to the biology of temperament Integrative research approaches to the study of temperament incorporate Rothbart’s behavioral model of temperament with the findings from neuroscience research, such as that of Davis (1992, 1998) and LeDoux (LeDoux, Farb, and Ruggiero, 1990; LeDoux et al., 1988). In our view, early

839

manifestations of temperament reflect differences in infant response to different sensory or social stimuli. Overlaying these behavioral and physiological responses is the infant’s affective display. Reactivity is marked by motoric (excluding the motor responses involved in the display of facial expression), physiological, and affective components of response. Infants may vary in their threshold or intensity of reactions to auditory, visual, tactile, or olfactory stimuli. Moreover, the affective bias, or predisposition to respond with a particular emotion or set of emotions, may be modality specific. That is, infants may be more likely to display negative or positive affect in response to stimulation specific to one modality over another. These initial affective dispositions are reflected in certain patterns of central nervous system activity that are evident early in the first year of life and that may bias the infant in its response to the social world. These individual differences in sensory sensitivity may have important implications for an infant’s interactions with his/her environment, including, perhaps most importantly, caregiver-infant interactions. Over the course of the first years of life, reactive responses increase in complexity, as do the competencies that are involved in the regulation of these responses. Behavioral competencies nominated to play a role in regulation include use of language (both private speech and its internalization), motor inhibition, attentional shifting and focusing, selfmonitoring, and the ability to flexibly switch set (Dodge, 1991; Kopp, 1982; Thompson, 1990). These latter skills have also been investigated under the rubric of executivefunction skills, and there is some discussion in the literature on the significance of the coincidental development of these skills and behaviors involved in regulation of emotion (Kopp, 1989; Welsh and Pennington, 1988). In our view, individual differences in the maturation of the prefrontal cortex play an important role in understanding the development of regulatory behaviors and, in particular, behaviors associated with the regulation of emotional reactivity. Areas of the prefrontal cortex have been implicated as subserving executive functions that in turn may play a role in regulating reactive tendencies. The role of cognitive processes as modulators of temperamental reactivity also corresponds to dual-process models describing the neural basis of social behavior, in which cortical regions such as the medial prefrontal cortex and rostral anterior cingulate function to inhibit or regulate lowerlevel limbic reactions and associated emotional responses (e.g., Lieberman et al., 2002; Satpute and Lieberman, 2006). In traditional dual-process models, top-down processes are emphasized in which the motor and autonomic components of responses to sensory stimuli are modulated by the control of higher-order cortical processes in a unidirectional manner. Deficits in cortical control thus contribute to the disinhibition of these more primitive limbic responses.

840

emotion⁄cognition interactions

There are, of course, multiple approaches to understanding the relations between reactive and regulatory processes. One view is that intense emotion may flood or overwhelm the cognitive system, interfering with its efficiency (e.g., Lazarus and Folkman, 1984; Mandler, 1982). Individuals may not necessarily differ in their cognitive skills under affect-neutral conditions, whereas their reactive responses under stress may produce interference in the performance of certain cognitive competencies (Pérez-Edgar et al., 2006). The development of such interference effects may be particularly strong among young behaviorally inhibited children. Developmentally, the experience of intense negative affect might occur in behaviorally inhibited children prior to the appearance of cognitive skills thought important in the regulation of reactivity. Data from our laboratory suggests, for example, that infants exhibiting high frequencies of negative affect and motor reactivity to novel auditory and visual stimuli are more likely to display behavioral inhibition as toddlers (e.g., Calkins, Fox, and Marshall, 1996). Thus, from infancy, these children have had the experience of intense negative affect in response to novelty. As general cognitive skills emerge, there may be attempts to utilize them in the service of regulation of distress or negative affect. If the negative affect is intense, these cognitive skills may not be efficiently evoked to produce successful regulation of negative affect. One would thus expect normative performance in behaviorally inhibited children on cognitive tasks thought to subsume regulation (e.g., attention or executive-function skills) in the absence of emotional challenge. In contrast a decrement in performance on these tasks might be expected when they are overlaid with a negative affect component (Pérez-Edgar and Fox, 2005a). A neural model of this account might include the influence of subcortical activation on those areas of cortex thought to involve these higher-order cognitive skills. This bottom-up view might account for the disruption of ongoing cognitive activity thought to occur during periods of intense negative affect. There are a variety of data that indicate that limbic areas modulate cortical activity (Derryberry and Tucker, 1991). Garcia and colleagues (1999) have reported that in a fear-conditioning paradigm, amygdala activity modulated the activity of the prefrontal cortex. The authors implanted electrodes in prefrontal cortical areas and found a reduction in neuronal activity as a function of amygdala activation. Prefrontal neuronal activity was manipulated by performing bilateral lesions in the amygdala complex of the animals. The authors speculate that abnormal amygdalainduced modulation of prefrontal neuronal activity may be involved in the pathophysiology of certain forms of anxiety disorder. In all probability, neither the top-down model in which cortical processes modulate limbic areas nor the bottom-up

model in which activation of limbic regions modulates cortex is sufficient to explain relations between emotional reactivity and regulation. The behavioral work of Rothbart and others including Kochanska (e.g., Kochanska, Murray, and Coy, 1997) and Eisenberg (e.g., Eisenberg et al., 1994) clearly indicates that certain cognitive processes involving attention and inhibition are critical in regulating emotional distress. However, in individuals with temperamental biases to express negative affect in response to novelty or stress, cognitive processes may be insufficient to down-regulate intense emotional arousal.

Behavioral inhibition: A “model system” for studying the biology of temperament Behavioral inhibition is found in approximately 15 percent of the population and is defined as the tendency to display signs of fear and wariness in response to unfamiliar stimuli (Kagan, Reznick, and Snidman, 1988b; Schmidt et al., 1997). Behaviorally inhibited children are unlikely to initiate interaction and often do not respond positively when social initiations are made toward them (Coplan et al., 1994). Behavioral inhibition is by definition a construct characterized by a constellation of traits at both the behavioral and biological levels. Specifically, inhibited children show a high heart rate and low beat-to-beat variability in heart rate (Marshall and Stevenson-Hinde, 1998), pupillary dilation during cognitive tasks (Kagan, Reznick, and Snidman, 1988b), elevated salivary cortisol levels (N. Fox and Stifter, 1989; Schmidt et al., 1997), an increased startle response (Kagan, Reznick, and Snidman, 1988a, 1988b; Schmidt and Fox, 1999), and right frontal electroencephalogram (EEG) asymmetry (M. Bell and Fox, 1994; Calkins, Fox, and Marshall, 1996; N. A. Fox et al., 2001). Underlying this complex biological and behavioral pattern is the assumption of a hyperaroused limbic system, centered on the amygdala. In proposing the model, Kagan drew on a line of research linking the amygdala to the acquisition of conditioned fear (M. Davis, Walker, and Lee, 1997), the induction of vigorous limb movements (Amaral et al., 1992), and the modulation of distress cries (Newman, 1985). In examining the phenomenology of behavioral inhibition, the studies we have noted all chose behavioral or physiological outcomes that reflect the presumed activation of nuclei within the amygdala. In addition, this work is directly linked to research on the neural circuitry underlying the behaviors of conditioned and unconditioned fear. Brown, Kalish, and Farber (1951) demonstrated that the acoustic startle reflex (an unconditioned response) could be augmented or potentiated by presenting the acoustic startle stimulus in the presence of a cue that had previously been paired with a shock. This state potentiates the acoustic startle reflex, which is thought of as an index of fear. In a series of

elegant experiments Michael Davis and his colleagues identified the neural circuitry involved in the acoustic reflex in the rat and mapped the circuitry involved in the augmentation of that reflex (Davis, 1986, 1989; Davis et al., 1993; Davis, Hitchcock, and Rosen, 1987). These studies showed that the acoustic startle pathway consists of only three synapses and that a single nucleus in the amygdala, the central nucleus, is involved in the potentiation of the reflex (LeDoux et al., 1988, 1990). Davis and colleagues recognized the possible limitations of applying a neural model of conditioned fear to certain forms of fearful or anxious behavior in humans. A fear state is assumed to have a specific elicitor, whereas anxious behaviors may reflect a more generalized state with no specific eliciting stimulus. As a consequence, in their more recent work they have attempted to examine the neural circuits associated with non-fear-potentiated startle (see Davis, 1998). In fear-potentiated startle the central nucleus of the amygdala played a critical role in the augmentation of the reflex. However, Davis and colleagues showed a clear distinction between the central nucleus of the amygdala and the bed nucleus of the stria terminalis (a site adjacent to but anatomically distinct from the amygdala) in relation to fearpotentiated startle versus non-fear-enhanced startle (e.g., Walker and Davis, 1997). While lesions of the central nucleus blocked expression of fear-potentiated startle, they had no effect on light-enhanced startle. Conversely, lesions of the bed nucleus of the stria terminalis significantly attenuated light-enhanced startle without any effect on fear-potentiated startle. Additionally, Davis and colleagues found that the bed nucleus of the stria terminalis (along with other limbic structures) but not the central nucleus was involved in startle enhanced by corticotrophin-releasing factor (CRH), a neuropeptide involved in stress reactivity (Lee and Davis, 1997). Davis and colleagues believe these findings to be critical to their model differentiating conditioned fear and its location in the central nucleus from anxiety and its location in the stria terminalis. A CRH-enhanced state has a longer time course, and this longer action may be more akin to a model of anxiety than to conditioned fear. Importantly, for the study of behavioral inhibition, both neural systems (central nucleus of the amygdala and bed nucleus) have similar outputs to autonomic and motor targets. Kagan’s use of Davis’s fear model is further extended by a growing literature linking behavioral inhibition to anxiety, a class of disorders long associated with an overly sensitive or hyperaroused fear system (Davis, 1992; Pérez-Edgar and Fox, 2005b). For example, a recent report (Schwartz et al., 2003b) found that 15 percent of young adults previously identified as behaviorally inhibited as toddlers were diagnosed with generalized social phobia. Schwartz, Snidman, and Kagan (1999) earlier found that adolescents found to be

fox et al.: biology of temperament

841

inhibited at age 2 were more likely than their uninhibited peers to show symptoms of social anxiety as assessed by a semistructured diagnostic interview (i.e., DISC; Schaffer et al., 1996). Indeed, 61 percent of the adolescents had current symptoms, and a full 80 percent had shown symptoms of anxiety at one point in their lifetime. Biederman and colleagues (2001) found that the rate of social anxiety disorder was significantly higher in inhibited children than in children without behavioral inhibition, particularly if the parent had a current psychiatric diagnosis. In addition, parallel studies have found that children of parents with anxiety disorders are more likely to show extreme behavioral inhibition (Biederman et al., 1991). A summary of these findings and others can be found in table form in a review by Hirshfeld-Becker and colleagues (2003) of the studies linking behavioral inhibition to vulnerability to psychopathology, as well as in Pérez-Edgar and Fox (2005b). Supporting these behavioral observations are documented similarities in the biological patterns associated with both anxiety and behavioral inhibition. Anxious states are often accompanied by autonomic arousal seen in tachycardia, pupil dilation, enhanced reflexes, tremor, and an increased startle response (Cuthbert et al., 2003). These same responses are evident in behaviorally inhibited children when they are confronted with novel or unfamiliar situations (Kagan, Reznick, and Snidman, 1988a, 1988b; Schmidt and Fox, 1999). In the childhood anxiety literature, researchers are now beginning to move beyond secondary measures of the hypothesized amygdala model to direct observations of amygdala reactivity. For example, Monk and colleagues (2003) found that children fearful of an uncomfortable air puff to the larynx showed more right-amygdala activation when faced with the threat of the upcoming air puff. In addition, children with anxiety disorders display hyperresponsive amygdala activity compared to healthy children of the same age when viewing fearful versus neutral faces, particularly in the right hemisphere (Thomas et al., 2001). More recent work (McClure et al., 2007) found that adolescents with general anxiety disorder show increased amygdala reactivity when asked to rate their own subjective fear state during a face-processing task. As in other studies, the effect was strongest when the individual was faced with negative or threatening stimuli. Although research in behavioral inhibition has been driven for the last two decades by the amygdala model, it has relied on behavioral (e.g., motoric reactivity in infancy) and biological (e.g., autonomic reactivity) markers theoretically linked to amygdala functioning (e.g., N. Fox, Henderson et al., 2005; Pérez-Edgar and Fox, 2005b). The first direct observations of amygdala activity in vivo associated with behavioral inhibition have only emerged in the last

842

emotion⁄cognition interactions

few years (Pérez-Edgar et al., 2007; Schwartz et al., 2003a). Schwartz and colleagues (2003a) demonstrated that young adults characterized as behaviorally inhibited in the second year of life exhibited amygdala hyperreactivity to novel neutral facial expressions relative to familiarized neutral expression faces, regardless of their current levels of anxiety. These data were the first to find direct links between early behavioral inhibition and later amygdala activity, laying the foundation for future studies. In a follow-up study (Pérez-Edgar et al., 2007), behaviorally inhibited adolescents completed a face-processing task previously used by McClure and colleagues (2007) to illustrate anxiety-linked differences in amygdala reactivity. Behaviorally inhibited adolescents showed increased amygdala activation while rating subjective fear state, although this effect was evident across all stimuli, without regard to face valence. These data suggest that behaviorally inhibited adolescents are responding less to the valence of the face and more to the novelty of the task demands (e.g., rating how afraid one is of a happy face). In addition, recent work has taken advantage of functional magnetic resonance imaging (fMRI) technology to examine differences in neural functioning underlying behavioral inhibition that move beyond the amygdala and its role in the fear circuit. Until recently, no research on behavioral inhibition has examined attention to cues that engage approach behaviors (e.g., rewards). Such work is important, since behavioral inhibition could be associated with biased responses to an array of motivationally salient stimuli, including both rewards and punishments (S. Reynolds and Berridge, 2002; Roitman, Wheeler, and Carelli, 2005). Indeed, a behavioral study using a monetary incentive delay (MID) task developed by Knutson and colleagues (2001) in shy college students found faster reaction times to potential rewards in shy students, reflecting perhaps an enhanced reward sensitivity (Hardin et al., 2006). Reward-related processing is thought to be mediated in part by a distributed neural circuit encompassing the striatum and associated regions (Di Chiara, 2002). This circuitry facilitates attention to salient stimuli and regulates approach behavior. Similarly, Guyer and colleagues (2006) demonstrated that behaviorally inhibited adolescents present an enhanced neural sensitivity to incentives relative to noninhibited adolescents. This enhanced sensitivity was manifest as greater striatal activation in the anticipation of monetary gain or loss, with striatal activation increasing in the behaviorally inhibited group as the magnitude of incentive grew larger. Again, the enhanced striatal activation found in the study suggests that stimulus salience may be an important factor in relation to behavioral inhibition. The findings may also help interpret the developmental literature linking early behavioral inhibition with later psychopathology,

implicating altered developmental trajectories centered on the reward circuitry (Ernst, Pine, and Hardin, 2006). Future work is poised to expand on these early findings, diversifying the tasks and modalities used, as well as the neural substrates of interest. For example, our understanding of the developmental sequelae of early temperament would benefit from a focus on regulatory mechanisms that may interact with early reactive processes (e.g., fear or reward) to shape behavior. To this end, a focus on the primary regulators of the amygdala, such as the orbitofrontal cortex (OFC) or ventral prefrontal cortex (vPFC), hold a great deal of promise (Bishop, Duncan, and Lawrence, 2004; Monk et al., 2004, 2006). Studies linking reactive and regulatory neural mechanisms add a needed empirical dimension to the theoretical constructs driving much of the current (behavioral) temperament research (Rothbart and Ahadi, 1994).

Frontal EEG asymmetry as a marker of negative reactivity One important biological marker of temperamental biases to approach or withdraw from unfamiliarity can be found in the pattern of EEG recorded over frontal scalp locations. The motivation for examining EEG asymmetries in relation to approach/withdrawal tendencies derives from a variety of data that encompasses work with clinical adult populations, normative adults, and children (see N. Fox, 1994b; Silberman and Weingartner, 1986). There has long been interest in the finding that unilateral lesion or stroke, particularly in anterior portions of cortex, seems to differentially affect mood state. Left anterior lesions are often associated with depressive symptoms and negative affect (Morris et al., 1996), while right anterior lesions are associated with mania and inappropriate positive affect (Sackheim et al., 1982). These clinical findings have prompted theorists to speculate as to the functional implications of prefrontal cortex lateralization for motivation and emotion. One prominent model, postulated by R. Davidson (1992) and N. Fox (1991) holds that anterior regions of the two hemispheres are lateralized for the behavioral/motivational systems involved in either approach or withdrawal. The left anterior region is specialized for the integration and control of those motor and cognitive behaviors associated with approach, novelty seeking, and reward, including most positive affects and fine motor and exploratory behavior. The right anterior region is specialized for the integration and control of those motor and cognitive behaviors associated with withdrawal, flight, or aversive responses. Davidson and colleagues found that subjects exhibiting right frontal EEG asymmetry were more likely to rate video clips as negative compared to subjects exhibiting left frontal EEG asymmetry (Wheeler, Davidson, and Tomarken, 1993).

Sobotka, Davidson, and Senulis (1992) found increased left frontal EEG asymmetry during a task designed to enhance approach-and-reward motivation. In addition, Sutton and Davidson (1997) report an association between frontal EEG asymmetry and self-reported behavioral approach versus inhibition. Subjects high on behavioral approach were more likely to display left frontal EEG asymmetry. Data from developmental studies also support this relation between frontal EEG asymmetry and affective/motivational biases. R. Davidson and Fox (1989) found that 10-month-old infants who exhibited left frontal EEG asymmetry were less likely to cry and show distress to immediate, brief maternal separation than were infants displaying right frontal EEG asymmetry. Fox and colleagues replicated this finding and reported that the pattern of frontal EEG asymmetry appeared to be a stable characteristic in infants over the course of the second half of the first year of life (N. Fox, Bell, and Jones, 1992). In a subsequent study, infants who were selected based on high levels of reactivity to novel sensory stimuli at 4 months of age were more likely to exhibit right frontal EEG asymmetry at 9 months of age than were infants who were either positively reactive or unreactive to the same stimuli (Calkins, Fox, and Marshall, 1996). The combination of behavioral reactivity and negative affect bias as reflected in the pattern of frontal EEG asymmetry is, as well, a strong predictor of temperamental outcome in infants across the first four years of life (Henderson, Fox, and Rubin, 2001). Work with clinical populations has further elaborated the relation of frontal EEG asymmetry to affective/motivational biases. For example, Henriques and Davidson (1991) reported associations between frontal EEG asymmetry and depression, noting that patients exhibiting depressive symptoms were more likely to exhibit right frontal EEG asymmetry. Sutton and colleagues (2005) found that among a sample of children with higher functioning autism, greater left frontal EEG asymmetry was associated with greater social approach and fewer social impairments (Sutton et al., 2005).

Genetic markers of negative reactivity Temperamental biases toward social withdrawal and negative reactivity may have a genetic foundation. Building on the accumulating body of evidence supporting a genetic contribution to temperamental variability, molecular genetic designs seek to determine the specific genes associated with phenotypic variability in temperament. Specifically, through the identification of quantitative trait loci (QTLs) associated with temperamental variability, molecular genetic studies provide information on the mechanisms, including molecular and gene-environment interactions, underlying temperamental variability. One candidate gene in particular that has been associated with behavioral inhibition and psychiatric

fox et al.: biology of temperament

843

conditions such as depression and anxiety is the 5-HTTLPR serotonin transporter gene. Given that observed phenotypic variability in temperament and personality likely results from the combined effects of many genes, it is not surprising that any single variation in a genetic polymorphism can only account for a small percentage of the variance in behavior. Thus several studies suggest that genetic status per se is not related to phenotypic variability but rather that the interaction of genetic status with environmental stressors is the best predictor of behavioral outcomes (e.g., BakermansKranenburg and van IJzendoorn, 2006; N. A. Fox, Nichols, et al., 2005). Variations in the 5-HTT serotonin transporter region have been associated with variations in personality traits including neuroticism, anxiety, and harm avoidance (e.g., Lesch et al., 1996; Munafo et al., 2003). For example, Lesch and colleagues (1996) reported that participants who were homo- or heterozygous for the short 14-repeat allele (s/s or s/l) self-reported higher levels of neuroticism, anxiety, and harm avoidance relative to participants who were heterozygous for the long 16-repeat allele (l/l). The short allele is associated with reduced 5-HTT transcription, lower transporter levels, and reduced serotonin uptake, with functional effects on neural circuits regulated by serotonin (Hariri et al., 2002). Hariri and colleagues (2002) found that the presence of the short allele was associated with heightened amygdala activation in response to fear faces in adults, suggesting that the variations in this polymorphism may be particularly relevant to aspects of temperament related to negative emotionality and reactivity to novelty. Consistent with the findings of Lesch and colleagues (1996), Auerbach and colleagues (1999) reported that infants who were homozygous for the short allele (s/s) were rated as higher on negative emotionality and distress to limitations. It is important to note, however, that others have failed to replicate such an association (e.g., Flory et al., 1999), including several studies with children (e.g., Jorm et al., 2000; Schmidt et al., 2002). For example, Schmidt and colleagues (2002) reported that preschoolers with the short allele (s/s and s/l) did not differ from those with the long allele (l/l) on maternal reports of social adjustment or observational measures of social reticence during interactions with unfamiliar peers. Further, two studies reporting an association between allelic variations and child temperament found the opposite association, such that adolescents who were homozygous for the long allele (l/l) were rated as more anxious than adolescents with the short allele ( Jorm et al., 2000), while the long allele was associated with questionnaire-based reports of grade-school children’s shyness (Arbelle et al., 2003). Several recent studies have examined the combined influence of environmental stressors and genetic status on temperament, behavior problems, and psychopathology. In

844

emotion⁄cognition interactions

relation to 5-HTT promoter polymorphisms, Caspi and colleagues (2003) found that stressful life events were more predictive of depressive disorder for adults with the short allele relative to adults homozygous for the long allele. Kaufman and colleagues (2004) reported that maltreated children with the short allele and no positive social supports had higher depression scores than children with all other combinations of social support and genotype status. In relation to child temperament, N. A. Fox, Nichols, and colleagues (2005) examined the associations between 5-HTT promoter polymorphism status in combination with maternal reports of social support as a predictor of behavioral inhibition in middle childhood. While 5-HTT status did not, on its own, relate to behavioral inhibition, the interaction between genotype and social support did. Specifically, 5HTT status was more strongly associated with behavioral inhibition at increasingly low levels of social support such that children with the short allele (s/s or s/l) who had mothers reporting low levels of social support were significantly more behaviorally inhibited than were children with the long alleles (l/l) whose mothers reported low levels of social support. Importantly, this finding held true regardless of whether laboratory observations or maternal reports of behavioral inhibition were examined. Together these findings suggest that the effects of genetic status on behavioral functioning may be moderated by the level of environmental stress (i.e., caregiver sensitivity, social support) that a child experiences. As such, studies demonstrating such gene-environment interactions provide novel information not only on developmental outcomes but also regarding the ways in which genetic inheritance contributes to developmental processes over time (Bakermans-Kranenburg and van IJzendoorn, 2006; Moffitt, Caspi, and Rutter, 2005; Rutter, 2006; Suomi, 2004).

Regulation of temperamental reactivity In the temperament literature, self-regulation is used to describe processes that function to modulate the timing and intensity of an individual’s reactivity. As described in the previous sections, from birth infants have rudimentary behavioral and physiological capacities to modulate their reactivity. These capacities are primarily reflexive and are elicited in response to changes in the external or internal environment, and thus are often studied under the rubric of temperamental reactivity. With development, increasing cognitive capacities (e.g., working memory, response inhibition, attentional control) and the maturation of the frontal cortex (e.g., including synaptogenesis and pruning, and the refinement in corticocortical and corticolimbic connections) allow for the implementation of more intentional and integrated strategies for managing behavior and emotion. The prefrontal cortex receives input from all sensory modalities,

as well as many interconnections from other brain structures, and as a result is thought to be responsible for behaviors that regulate reactivity from the external environment (Banfield et al., 2004). In the most general sense, self-regulation refers to the ability to generate and voluntarily direct goal-oriented, adaptive responses in the absence of external monitors or supports (Kopp, 1982; Welsh and Pennington, 1988), promoting adaptation through the maintenance of physiological and psychological well-being (Block and Block, 1980; Rothbart and Bates, 1998). It is considered a key aspect of normal development, and deficits in self-regulation are the defining features of numerous developmental disabilities including attention-deficit/hyperactivity disorder (Barkley, 1997), autism (Ozonoff, Pennington, and Rogers, 1991), and conduct disorder (White et al., 1994). In the model of temperament proposed by Rothbart and Derryberry (1981), self-regulative processes are orthogonal to, or relatively independent of, reaction tendencies. As such, an infant’s or child’s pattern of behavior can be thought of as reflecting the intersection or combined influence of relative levels of reactivity and self-regulation. Development of Prefrontal Processes Involved in Regulation Across domains of development and approaches to the study of self-regulation, it is evident that a certain level of cognitive development is required for the transfer of control of regulation from external to internal sources. Thus the strategies used by children to regulate their emotions and behaviors become more complex throughout development. Infants will often use gaze aversion, fussiness, self-soothing, and gross motor activities as regulative strategies (Kopp, 2002; McCabe, Cunnington, and Brooks-Gunn, 2004). However, as their cognitive capabilities mature, the processes used in self-regulation become more advanced. The cognitive capabilities that allow for more efficient self-regulatory strategies belong to a larger group of cognitive functions referred to as executive functions, highlighting their higher-order control, or regulation, over more basic reaction tendencies. Kopp (1982) described fundamental cognitive abilities that are needed for self-initiated behavioral control. Among these cognitive abilities are self-monitoring and other attention processes. The use of attentional strategies to regulate one’s reactivity develops throughout childhood (Rueda, Posner, and Rothbart, 2004). For example, 4-month-old infants who quickly disengaged their attention from a central fixation point upon presentation of another stimulus in the periphery were rated by their mothers as less easily distressed and more easily soothed compared to infants who disengaged their attention less easily ( Johnson, Posner, and Rothbart, 1991). During adult-infant interactions, adults engage and disengage infants’ attention in order to manage

the infants’ level of arousal. States of engaged attention between infants and their caregivers tend to be associated with play, joy, and general positive affect (Gottman, Katz, and Hooven, 1997). Adults also tend to be sensitive to infants’ needs to disengage their attention in order to dampen or reduce levels of arousal (N. A. Fox, Henderson, et al., 2005). When parents respond contingently to their infants’ needs to disengage and reengage interactions, infants learn about the efficacy of attentional control as a means of self-regulation (Gottman, Katz, and Hooven, 1997). Over the first months of life, infants switch from a stimulus-driven, externally reactive attention system to a system with more voluntary attentional control (Rothbart, Posner, and Boylan, 1990; as cited in Rueda, Posner, and Rothbart, 2005), which advances mechanisms of self-regulation. As the infant develops, a more voluntary, executive control system is acquired (Rueda, Posner, and Rothbart, 2004). Children are able to resolve conflict more easily, flexibly shift and adapt their response, and inhibit certain dominant responses, thoughts, and emotions to act more appropriately (M. Davidson et al., 2006). To assess the functioning of executive control, conflict tasks with less demanding or no language requirements have been developed. GerardiCaulton (2000) developed a Strooplike task to assess the development of executive attention in younger children using conflict between the location and identity of an object with no verbal material. By examining the differences in response time to spatially congruent and spatially incongruent trials, 24- to 36-month-olds showed dramatic improvement in their ability to resolve spatial conflict (Gerardi-Caulton, 2000). These results suggest that the brain mechanisms underlying executive attention undergo dramatic changes from 2–3 years of age, a developmental period that is consistent with others’ reports of rapid changes in self-regulation (e.g., Rothbart et al., 2003; Vaughn, Kopp, and Krakow, 1984). The study also found that children with better conflict resolution had less anger and frustration as reported by their parents than children with lower conflict-resolution scores. These results suggest that the development of executive attention allows children to employ strategies to regulate their distress reactivity. An inability to disengage attention from negative stimuli is associated with greater arousal and maintenance of negative affect and distress. For example, work on attention biases to threat (described later) illustrates one mechanism by which attention to negative stimuli may maintain behaviorally inhibited symptoms. The development of executive attention, as examined through conflict resolution, has been studied from 4 years of age to adulthood using the child and adult versions of the Attention Network Task (ANT; Fan et al., 2002). In the ANT, participants respond to the direction of the center stimulus (fish in the child version and arrows in the adult version), which is either congruent or incongruent to the

fox et al.: biology of temperament

845

direction to the flanker stimuli (Rueda, Posner, and Rothbart, 2004). In a series of studies, Rueda, Posner, and Rothbart (2005) found significant improvement in executive attention between 4 and 7 years of age, such that by age 7 children perform at adult levels on the conflict-resolution task of the ANT. Action-oriented plans afford young children reflective control over their behavior and their environment (Zelazo, Reznick, and Pinon, 1995). The ability to use plans to control and inhibit actions involves two processes. This first is the understanding and representation of a rule or contingency specifying the relation between a stimulus condition and the expected response. The second process involves using the representation of the rule to guide behavior. Translating representations into action depends on the maturation of cortical processes, particularly those relating to the inhibition of inappropriate responses and behaviors ( J. Bell and Livesey, 1983). The Dimensional Card Change Sort (DCCS) task, developed by Frye, Zelazo, and Palfai (1995), assesses children’s ability to keep several simple rules in working memory, use the rules to guide their behavior, and then switch between rules (Zelazo, Reznick, and Pinon, 1995). In the DCCS children are asked to sort cards according to one of two dimensions: color or shape. Halfway through the task children are asked to sort according to the other dimension. Zelazo, Reznick, and Pinon (1995) reported that young children ages 3–5 had little difficulty in sorting according to the first rule set. However, when asked to switch to sorting according to the second dimension, using the second rule set, 3-year-olds had great difficulty while most 4- and 5-yearolds performed quite accurately. Interestingly, the children who had difficulty switching to the second rule set could often tell the experimenter the correct rules (i.e., the appropriate way to sort), illustrating a discrepancy between knowing the rule and using that rule set to guide behavior (Zelazo, Frye, and Rapus, 1996; as cited in Rennie, Bull, and Diamond, 2004). This pattern of behavior is analogous to one reported by Diamond, Cruttenden, and Neiderman (1994), who studied the A-not-B task and found that some infants who made the error, reaching to A instead of B, continued to look at the correct location. It has also been argued that the discrepancy in 3-year-olds is a result of decreased inhibitory attentional control (Kirkham, Cruess, and Diamond, 2003). Using a variant of the DCCS, Rennie, Bull, and Diamond (2004) found that 3-year-old performance suffered only when attentional control demands were high. The authors suggest that 3-year-olds have difficulty disengaging from a previous “mind-set” and redirecting their attention to the new task demands (Kirkham, Cruess, and Diamond, 2003; Rennie, Bull, and Diamond, 2004). To assess how attention and behavioral inhibition in childhood relate to similar cognitive abilities in adolescence and adulthood, Eigsti and colleagues (2006) examined

846

emotion⁄cognition interactions

how performance on a delay-of-gratification task at age 4 predicted performance on a go-no-go paradigm more than 10 years later. Both tasks require the use of attentional and behavioral control when the person is faced with salient stimuli. Also, similar neural circuitry is thought to be recruited in both tasks (Casey, Durston, and Fossella, 2001; as cited in Eigsti et al., 2006). During the delay-ofgratification task, children were given the choice between one or two cookies, and all children chose the larger reward (i.e., two cookies). The children were then told that the experimenter had to leave the room for a little while, and if the child could wait to eat the treat until the experimenter returned, they could have both cookies; if the child could not wait, the experimenter would return to the room, but the child would only receive one cookie. During the delay period children’s behaviors and eye gaze were coded. Fouryear-old children who effectively directed their attention away from the reward during the delay period did better at age 14 inhibiting their responses on no-go trials of the go-no-go task compared to the 4-year-olds who spent more time focusing their attention on the reward. The relations between temperament and inhibitory control have only recently been explored. N. Fox and Henderson (2000) assessed inhibitory control in a sample of 2-year-old children, half of whom were identified as behaviorally inhibited and half of whom were identified as exuberant in temperament. Based upon their performance on a delay task, children were divided into those who were able to inhibit their response versus those who could not. The authors reasoned that children high on inhibitory control would display greater social competence. This prediction was confirmed for the exuberant children but not for the children high in behavioral inhibition. Specifically, exuberant children high on inhibitory control were observed to be more socially competent than exuberant children low on inhibitory control. However, the opposite pattern was found for behaviorally inhibited children. Behaviorally inhibited children high on inhibitory control were actually observed to be more socially withdrawn and less socially competent than behaviorally inhibited children low on inhibitory control. The data suggest that inhibitory control may not always work as a process facilitating adaptive emotion control. Rather such cognitive processes as inhibitory control or attention shifting may play different roles in predicting social competence based upon the temperament of the child. Measuring the Neural Circuitry Involved in Regulation Increases in the ability to inhibit dominant responses and perform subdominant responses and flexibly shift attention are attributed to developments in a network of anterior cortical structures including parts of the prefrontal cortex, the anterior cingulate cortex, and the supplementary motor area (see Pardo et al., 1990). The anterior cingulate

may be particularly relevant to the study of temperament, because it appears to serve as a point of integration for the visceral, attentional, and affective information that forms the basis of self-regulation (Devinsky, Morrell, and Vogt, 1995; Lane et al., 1998; Thayer and Lane, 2000). Casey and colleagues (1997) used fMRI to examine possible differences in brain activation between adults and children on a go-no-go task. In the go-no-go task participants have to withhold a dominant response to press a button when specific stimuli appear, requiring cognitive and behavioral control. No differences between children and adults were found in regard to specific areas recruited during the task; thus children and adults appear to use similar areas of the prefrontal cortex during this cognitive control task. However, significant differences in the volume of brain area activated during the task were seen on the no-go trials of the task. The differences were seen mostly in the children’s dorsal and lateral prefrontal areas. The poorer cognitive control of children and the resulting increased difficulty of the task may necessitate increased cognitive resources, resulting in differences in the volume of activated brain regions. A related line of work has relied on the event-related potential (ERP) to examine the neural mechanisms of selfregulation. Within this work, by far, the most common use of the ERP methodology has been to examine the physiological correlates of attentional processes. This observation is true for both children and adults in healthy and clinical populations. Adaptive functioning is often dependent on the individual’s ability to appropriately select aspects of the environment that are of interest from among the constant and simultaneous presentation of competing stimuli. Ridderinkhof and van der Stelt (2000) argue that age-related improvements in this ability, particularly when under strong voluntary control, is one of the most profound advances in information processing that takes place in childhood. These improvements are of particular importance to temperament researchers, as Rothbart and colleagues have shown that individuals better equipped to regulate initial reactivity, particularly through the use of attentional mechanisms, are less likely to show prolonged periods of negative affect. For example, their data suggest that infants prone to distress are less adept at shifting attention away from a distressing stimulus and have difficulty engaging in self-soothing activity (Rothbart, Posner, and Rosicky, 1994: Ruddy, 1993). Attention can act as a coping resource available to the child that may moderate the physiologic and behavioral correlates of reactivity (Mathews and MacLeod, 1994; Nachmias et al., 1996). The importance of attentional mechanisms in shaping developmental trajectories is bolstered by studies demonstrating attentional bias for threat-relevant stimuli in both clinically anxious adults (MacLeod and Mathews, 1991; Mogg, Mathews, and Eysenck, 1992) and

trait-anxious adults (E. Fox et al., 2001; N. Fox and Calkins, 1993; Mogg, Bradley, and Hallowell, 1994). Parallel findings have been noted with clinically anxious children (Taghavi et al., 1999; Vasey et al., 1995), high trait-anxious children (Schippell et al., 2003), and behaviorally inhibited children (Pérez-Edgar and Fox, 2005a). Attention may serve as an important cognitive mechanism for the observed link between temperament and anxiety-related psychopathology. Indeed, recent work suggests that attentional biases may play a causal role in the emergence of anxiety (MacLeod et al., 2002; Wilson et al., 2006). A related series of studies have focused on temperamentlinked differences in selective attention. In the flanker task (Eriksen and Eriksen, 1974; Ridderinkhof and van der Stelt, 2000), subjects are asked to note the direction or identity of a central stimulus while simultaneously ignoring adjacent stimuli that alternately mirror or conflict with the central stimulus (e.g., a central arrow points left, while flankers point right). Using this paradigm, Henderson (2003) found that behaviorally inhibited children show more awareness of their errors in the task, as reflected in greater amplitudes in the error-related negativity (ERN) early in performance. The ERN is thought to reflect either error-monitoring or response conflict, and its presence in the ERP wave has been linked to error correction (Pailing and Segalowitz, 2004). Another series of studies has relied on the Posner paradigm. In this task (Posner and Cohen, 1984), subjects are asked to note the spatial position of a target stimulus. Preceding the target stimulus, the subjects were presented with a cue stimulus that could appear in either the same (valid trial) or alternate (invalid trial) spatial location. Previous studies with both children and adults (Hugdahl and Nordby, 1994; Nobre, Sebestyen, and Miniussi, 2000; Perchet and García-Larrea, 2000) have consistently shown that individuals will respond more quickly to the targets in the valid trial versus the invalid trials and exhibit increased ERP amplitudes to valid trials. This finding is thought to reflect the effect of additional neural processing afforded by the early (correct) cueing. Recent work in children (Perchet and García-Larrea, 2000; Perchet et al., 2001; Pérez-Edgar et al., 2006; Rich et al., 2005; Rich et al., 2007) suggests that the basic mechanisms of selective attention are in place in both healthy and clinically diagnosed young children and that the task is sensitive to individual differences. In terms of temperament-linked variability, data suggest that behaviorally inhibited children show an increased bias for threat cues, particularly when under stress (Pérez-Edgar and Fox, 2005a). In addition to this work, the ERP has proven well suited to expanding our understanding of the psychological concerns marked by temperament and behavioral inhibition. Kagan (1997) argued that behavioral inhibition is marked

fox et al.: biology of temperament

847

by a unique sensitivity to novelty or salience, which is not limited solely to negative or threatening stimuli. Indeed, recruitment in infancy for long-term studies of temperament rely on extreme levels of reactivity to novel, but neutral, auditory and visual stimuli (N. Fox et al., 2001; Kagan et al., 1994). The unique neural sensitivity of behaviorally inhibited children is supported by patterns of early neurophysiological responses to auditory discrepancies (mismatch negativity, MMN) in the form of reduced MMN amplitudes and longer MMN latencies in socially withdrawn children compared to their more sociable peers (Bar-Haim et al., 2003). Similarly, Woodward (2002) found that young children with higher levels of negative reactivity in infancy demonstrated larger ERP responses in the Nc component to oddball and invalid novel stimuli in a visual paradigm. Such Nc responses are thought to reflect the process of novelty detection (Nelson and Collins, 1991; G. Reynolds and Richards, 2005). In line with Kagan, Woodward suggested that highly reactive children have a lower cortical threshold for detecting and responding to novelty, leading to increased ERP amplitudes to unfamiliar events. Currently, research has begun to build on documented individual differences in attentional bias to threat to develop possible avenues of intervention. In particular, an exciting line of work has documented the strength of attention biases as both a marker of psychopathology (Bar-Haim et al., 2007) and a potential mechanism for intervention (Wilson et al., 2006). Recent work (Bar-Haim, Lamy, and Glickman, 2005) found increased ERP amplitudes in ERP components tied to the detection of threat faces (thought to involve amygdala activation) and decreased ERP amplitude in components tied to probe detection and attention, subsuming areas thought to involve prefrontal activation. Using the same task, Bar-Haim and colleagues (personal communication) have documented initial attention biases in anxious children and then trained children’s biases away from threat. As in the adult literature (Wilson et al., 2006), such attentional training is marked by a decrease in anxious symptomatology. Future work will attempt to combine assessment, intervention, and psychophysiological measures in order to trace the course of attention, temperament, and anxiety over time.

Epilogue The study of infant temperament, including research on the biology of infant reactivity and regulation, has reinforced the notion that specific characteristics of the individual can have a significant effect on basic sensory, perceptual, and cognitive processing. For many years, cognitive scientists interested in attention, memory, and perception examined these processes as reflecting general characteristics of the individual. The new interest in affective neuroscience, however, has

848

emotion⁄cognition interactions

emphasized the importance of understanding the role that emotion plays in modulating these basic cognitive processes. Indeed, emotion and cognition are bidirectional in their influence, with evidence accumulating on the effects of emotion on memory and attention, and the effects of attention on emotion. These bidirectional effects are seen clearly in the study of infant and child temperament. From birth, the infant’s characteristic pattern of emotional reactivity influences his or her perception, attention, and memory for events. But as the neural circuitry underlying these cognitive processes develops it will modify basic reactive biases in the young child. Research in neuroscience is just beginning to understand how these reciprocal circuits (between emotion and cognition) work in adult subjects. We know much less about how these circuits develop and how, across development initial reactive tendencies and then environmental influences, they come to shape the neural circuitry that underlies complex social behaviors. The coming years will no doubt see greater exploration of these questions and a greater understanding of the developmental processes underlying the emergence of social behaviors. acknowledgments

Preparation for this chapter was supported in part by a grant from the National Institutes of Health to Nathan A. Fox (HD 17899) and Koraly Pérez-Edgar (MH 073569).

REFERENCES Amaral, D., J. Price, A. Pitkanen, and S. Carmichael, 1992. Anatomical organization of the primate amygdaloid complex. In The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction, 1–66. New York: Wiley. Arbelle, S., J. Benjamin, M. Golin, I. Kremer, R. H. Belmaker, and R. P. Ebstein, 2003. Relation of shyness in grade school children to the genotype for the long form of the serotonin transporter promoter region polymorphism. Am. J. Psychiatry 160:671–676. Auerbach, J., V. Geller, S. Lezer, E. Shinwell, R. H. Belmaker, J. Levine, and R. P. Ebstein, 1999. Dopamine D4 receptor (D4DR) and serotonin transporter promoter (5-HTTLPR) polymorphisms in the determination of temperament in 2-month-old infants. Mol. Psychiatry 4:369–373. Bakermans-Kranenburg, M. J., and M. H. van IJzendoorn, 2006. Gene environment interaction of the dopamine D4 receptor (DRD4) and observed maternal insensitivity predicting externalizing behavior in preschoolers. Dev. Psychobiol. 48(5):406– 409. Banfield, J. F., C. L. Wyland, C. N. Macrae, T. F. Münte, and T. F. Heatherton, 2004. The cognitive neuroscience of self-regulation. In R. F. Baumeister and K. D. Vohs, eds., Handbook of Self-Regulation: Research, Theory, and Applications, 62–83. New York: Guilford Press. Bar-Haim, Y., D. Lamy, and S. Glickman, 2005. Attentional bias in anxiety: A behavioral and ERP study. Brain Cogn. 59:11–12. Bar-Haim, Y., D. Lamy, L. Pergamin, M. J. BakermansKranenburg, and M. H. van IJzendoorn, 2007. Threat-related

attentional bias in anxious and nonanxious individuals: A metaanalytic study. Psychol. Bull. 133:1–24. Bar-Haim, Y., P. J. Marshall, N. A. Fox, E. A. Schorr, and S. Gordon-Salant, 2003. Mismatch negativity in socially withdrawn children. Biol. Psychiatry 54:17–24. Barkley, R. A., 1997. Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD. Psychol. Bull. 121:65–94. Bell, J. A., and P. J. Livesey, 1983. Cue significance and response regulation in 3- to 6-year-old children’s learning of multiple choice discrimination tasks. Dev. Psychobiol. 18:229– 245. Bell, M. A., and N. A. Fox, 1994. Brain development over the first year of life: Relations between EEG frequency and coherence and cognitive and affective behaviors. In G. Dawson and K. Fischer, eds., Human Behavior and the Developing Brain, 314–345. New York: Guilford Press. Biederman, J., D. R. Hirshfeld-Becker, J. F. Rosenbaum, C. Herot, D. Friedman, N. Snidman, J. Kagan, and S. V. Faraone, 2001. Further evidence of association between behavorial inhibition and social anxiety in children. Am. J. Psychiatry 158:1673– 1679. Biederman, J., J. F. Rosenbaum, and D. R. Hirshfeld, 1991. Psychiatric correlates of behavioral inhibition in young children of parents with and without psychiatric disorders. Annu. Prog. Child Psychiatry Child Dev. 269–284. Bishop, S. J., J. Duncan, and A. D. Lawrence, 2004. State anxiety modulation of the amygdala response to unattended threatrelated stimuli. J. Neurosci. 24:10364–10368. Block, J. H., and J. Block, 1980. The role of ego-control and ego-resiliency in the organization of behavior. In W. A. Collins, ed., Minnesota Symposium on Child Psychology, vol. 13, pp. 39–101. Hillsdale, N J: Erlbaum. Brown, J. S., H. I. Kalish, and I. E. Farber, 1951. Conditioned fear as revealed by magnitude of startle response to an auditory stimulus. J. Exp. Psychol. 41:317–328. Calkins, S. D., N. A. Fox, and T. R. Marshall, 1996. Behavioral and physiological antecedents of inhibited and uninhibited behavior. Child Dev. 67:523–540. Casey, B. J., S. Durston and J. A. Fossella, 2001. Evidence for a mechanistic model of cognitive control. Clin. Neurosci. Res. 1:267–282. Casey, B. J., J. Giedd, Y. Vauss, C. Vaituzis, S. Hamburger, P. Kozuch, R. Trainor, and J. L. Rapoport, 1997. The role of the anterior cingulate in automatic and controlled attentional processes: A developmental neuroanatomical study. Dev. Psychobiol. 30:61–69. Caspi, A., K. Snugden, T. E. Moffitt, A. Taylor, I. W. Craig, H. Harrington, J. McClay, J. Mill, J. Martin, A. Braithwaite, and R. Poulton, 2003. Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science 301:386–389. Coplan, R. J., K. H. Rubin, N. A. Fox, S. D. Calkins, and S. L. Stewart, 1994. Being alone, playing alone, and acting alone: Distinguishing among reticence and passive and active solitude in young children. Child Dev. 65:129–137. Cuthbert, B. N., P. J. Lang, C. Strauss, D. Drobes, C. J. Patrick, and M. M. Bradley, 2003. The psychophysiology of anxiety disorder: Fear memory imagery. Psychophysiology 40:407– 422. Davidson, M. C., D. Amso, L. C. Anderson, and A. Diamond, 2006. Development of cognitive control and executive functions from 4–13 years: Evidence from manipulations of

memory, inhibition, and task switching. Neuropsychologia 44: 2037–2078. Davidson, R. J., 1992. Anterior cerebral asymmetry and the nature of emotion. Brain Cogn. 20:125–151. Davidson, R. J., and N. A. Fox, 1989. Frontal brain asymmetry predicts infants’ response to maternal separation. J. Abnorm. Psychol. 98:127–131. Davis, M., 1986. Pharmacological and anatomical analysis of fear conditioning using the fear-potentiated startle paradigm. Behav. Neurosci. 100:814–824. Davis, M., 1989. Sensitization of the acoustic startle reflex by footshock. Behav. Neurosci. 103:495–503. Davis, M., 1992. The role of the amygdala in fear and anxiety. Annu. Rev. Neurosci. 15:353–375. Davis, M., 1998. Are different parts of the extended amygdala involved in fear versus anxiety? Biol. Psychiatry 44:1239–1247. Davis, M., W. A. Falls, S. Campeau, and M. Kim, 1993. Fearpotentiated startle: A neural and pharmacological analysis. Behav. Brain Res. 58:175–198. Davis, M., J. Hitchcock, and J. Rosen, 1987. Anxiety and the amygdala: Pharmacological and anatomical analysis of the fear potentiated startle paradigm. In G. H. Bower, ed., Psychology of Learning and Motivation, vol. 21, pp. 263–305. New York: Academic Press. Davis, M., D. L. Walker, and Y. Lee, 1997. Amygdala and bed nucleus of the stria terminalis: Differential roles in fear and anxiety measured with the acoustic startle reflex. Philos. Trans. R. Soc. London [Biol.] 352:1675–1687. Derryberry, D., and D. M. Tucker, 1991. The adaptive base of the neural hierarchy: Elementary motivational controls on network function. In R. Dienstbier, ed., Nebraska Symposium on Motivation, vol. 38: Perspectives on Motivation, 289–342. Lincoln: University of Nebraska Press. Devinsky, O., M. J. Morrell, and B. A. Vogt, 1995. Contributions of the anterior cingulate cortex to behaviour. Brain 118: 279–306. Diamond, A., L. Cruttenden, and D. Neiderman, 1994. AB with multiple wells. I. Why are multiple wells sometimes easier than two wells? II. Memory or memory inhibition? Dev. Psychol. 30(2):192–205. Di Chiara, G., 2002. Nucleus accumbens shell and core dopamine: Differential role in behavior and addiction. Behav. Brain Res. 137:75–114. Dodge, K. A., 1991. Emotion and social information processing. In J. Garber and K. A. Dodge, eds., The Development of Emotion Regulation and Dysregulation, 159–181. Cambridge, UK: Cambridge University Press. Eigsti, I. M., V. Zayas, W. Mischel, Y. Shoda, O. Ayduk, M. B. Dadlani, et al., 2006. Predicting cognitive control from preschool to late adolescence and young adulthood. Psychol. Sci. 17:478–484. Eisenberg, N., R. A. Fabes, M. Nyman, J. Bernzweig, and A. Pinuelas, 1994. The relations of emotionality and regulation to children’s anger-related reactions. Child Dev. 65:109– 128. Eriksen, B. A., and C. W. Eriksen, 1974. Effects of noise letters upon the identification of a target letter in a nonsearch task. Percept. Psychophys. 16:143–149. Ernst, M., D. S. Pine, and M. G. Hardin, 2006. Triadic model of the neurobiology of motivated behavior in adolescence. Psychol. Med. 36:299–312. Eysenck, H. J., and M. W. Eysenck, 1985. Personality and Individual Differences: A Natural Science Approach. New York: Plenum.

fox et al.: biology of temperament

849

Flory, J. D., S. B. Manuck, R. E. Ferrell, K. M. Dent, D. G. Peters, and M. F. Muldoon, 1999. Neuroticism is not associated with the serotonin transporter (5-HTTLPR) polymorphism. Mol. Psychiatry 4:93–96. Fan, J., B. D. McCandliss, T. Sommer, M. Raz, and M. I. Posner, 2002. Testing the efficiency and independence of attentional networks. J. Cog. Neurocsi. 14:340–347. Fox, E., R. Russo, R. Bowles, and K. Dutton, 2001. Do threatening stimuli draw or hold visual attention in subclinical anxiety? J. Exp. Psychol. [Gen.] 130:681–700. Fox, N. A., 1991. If it’s not left, it’s right: Electroencephalograph asymmetry and the development of emotion. Am. Psychol. 46: 863–872. Fox, N. A., ed., 1994a. The development of emotion regulation: Biological and behavioral considerations. Monogr. Soc. Res. Child Dev. 59(2–3, Serial No. 240). Fox, N. A., 1994b. Dynamic cerebral processes underlying emotion regulation. In N. A. Fox, ed., The development of emotion regulation: Biological and behavioral considerations, 152–166. Monogr. Soc. Res. Child Dev. 59(2–3, Serial No. 240). Fox, N. A., M. A. Bell, and N. A. Jones, 1992. Individual differences in response to stress and cerebral asymmetry. Dev. Neuropsychol. 8:161–184. Fox, N. A., and S. D. Calkins, 1993. Pathways to aggression and social withdrawal: Interactions among temperament, attachment, and regulation. In K. H. Rubin and J. B. Asendorpf, eds., Social Withdrawal, Inhibition, and Shyness in Children, 81–100. Hillsdale, N J: Erlbaum. Fox, N. A., and H. A. Henderson, 2000. Temperament, emotion, and executive function: Influences on the development of self-regulation. In M. Farah and M. Moscovitch, chairs, Life Is More Than a Delayed Saccade: Exploring the Social and Emotional Functions of Frontal Cortex. Symposium conducted at the Annual Meeting of the Cognitive Neuroscience Society, San Francisco, April. Fox, N. A., H. A. Henderson, P. J. Marshall, K. E. Nichols, and M. M. Ghera, 2005. Behavioral inhibition: Linking biology and behavior within a developmental framework. Annu. Rev. Psychol. 56:235–262. Fox, N. A., H. A. Henderson, K. H. Rubin, S. D. Calkins, and L. A. Schmidt, 2001. Continuity and discontinuity of behavioral inhibition and exuberance: Psychophysiological and behavioral influences across the first four years of life. Child Dev. 72:1– 21. Fox, N. A., K. E. Nichols, H. A. Henderson, K. Rubin, L. Schmidt, D. Hamer, M. Ernst, and D. S. Pine, 2005. Evidence for a gene-environment interaction in predicting behavioral inhibition in middle childhood. Psychol. Sci. 16:921–926. Fox, N. A., and C. A. Stifter, 1989. Biological and behavioral differences in infant reactivity and regulation. In G. A. Kohnstamm, J. E. Bates, and M. K. Rothbart, eds., Temperament in Childhood, 169–183. Chichester, UK: John Wiley. Frye, D., P. D. Zelazo, and T. Palfai, 1995. Theory of mind and rule-based reasoning. Cogn. Dev. 10:483–452. Garcia, R., R. M. Vouimba, M. Baudry, and R. F. Thompson, 1999. The amygdala modulates prefrontal cortex activity relative to conditioned fear. Nature 402:294–296. Gerardi-Caulton, G., 2000. Sensitivity to spatial conflict and the development of self-regulation in children 24–36 months of age. Dev. Sci. 3(4):397–404. Goldsmith, H. H., A. H. Buss, R. Plomin, M. K. Rothbart, A. Thomas, S. Chess, et al., 1987. Roundtable: What is temperament? Four approaches. Child Dev. 58:505–529.

850

emotion⁄cognition interactions

Goldsmith, H. H., and J. J. Campos, 1986. Fundamental issues in the study of early temperament: The Denver Twin Temperament Study. In M. E. Lamb, A. L. Brown, and B. Rogoff, eds., Advances in Developmental Psychology, vol. 4, pp. 231–283. Hillsdale, NJ: Lawrence Erlbaum. Gottman, J. M., L. F. Katz, and C. Hooven, 1997. Meta-emotion: How Families Communicate Emotionally. Mahwah, NJ: Lawrence Erlbaum. Gray, J. A., 1982. The Neuropsychology of Anxiety: An Enquiry into the Functions of the Septo-hippocampal System. New York: Oxford University Press. Gray, J. A., 1987. The Psychology of Fear and Stress, 2nd ed. Cambridge, UK: Cambridge University Press. Gray, J. A., 1991. The neuropsychology of temperament. In J. Strelau and A. Angleneitner, eds., Explorations in Temperament: International Perspectives on Theory and Measurement, 105–128. New York: Plenum. Guyer, A. E., E. E. Nelson, K. Perez-Edgar, M. G. Hardin, R. Roberson-Nay, et al., 2006. Striatal function alteration in adolescents characterized by early childhood behavioral inhibition. J. Neurosci. 26:6399–6405. Hardin, M. G., K. Perez-Edgar, A. E. Guyer, D. S. Pine, N. A. Fox, and M. Ernst, 2006. Reward and punishment sensitivity in shy and exuberant adults: A relationship between social and motivated behavior. Personality and Individual Differences 40:699–711. Hariri, A. R., V. S. Mattay, A. Tessitore, B. Kolachana, F. Fera, D. Goldman, M. F. Egan, and D. R. Weinberger, 2002. Serotonin transporter genetic variation and the response of the human amygdala. Science 297:400–403. Henderson, H. A., 2003. Temperamental contributions to problem solving: Cognitive and affective processes. Dissertation Abstracts International, 64. Henderson, H. A., and N. A. Fox, 2007. Considerations in studying emotion in infants and children. In J. A. Coan and J. J. B. Allen, eds., The Handbook of Emotion Elicitation and Assessment, 827–864. New York: Oxford University Press. Henderson, H. A., N. A. Fox, and K. H. Rubin, 2001. Temperamental contributions to social behavior: The moderating roles of frontal EEG asymmetry and gender. J. Am. Acad. Child Adolesc. Psychiatry 40:68–74. Henriques, J. B., and R. J. Davidson, 1991. Left frontal hypoactivation in depression. J. Abnorm. Psychol. 100:535–545. Hirshfeld-Becker, D. R., J. Biederman, S. Calltharp, E. D. Rosenbaum, S. V. Faraone, and J. F. Rosenbaum, 2003. Behavioral inhibition and disinhibition as hypothesized precursors to psychopathology: Implications for pediatric bipolar disorder. Biol. Psychiatry 53:985–999. Hugdahl, K., and H. Nordby, 1994. Electrophysiological correlates cued attentional shifts in the visual and auditory modalities. Behav. Neural Biol. 62:21–32. Johnson, M. H., M. I. Posner, and M. K. Rothbart, 1991. Components of visual orienting in early infancy: Contingency learning, anticipatory looking, and disengaging. J. Cogn. Neurosci. 3:335–344. Jorm, A. F., M. Prior, A. Sanson, D. Smart, Y. Zhang, and S. Easteal, 2000. Association of a functional polymorphism of the serotonin transporter gene with anxiety-related temperament and behavior problems in children: A longitudinal study from infancy to the mid-teens. Mol. Psychiatry 5:542– 547. Kagan, J., 1994. Galen’s Prophecy. New York: Basic Books. Kagan, J., 1997. Temperament and the reactions to unfamiliarity. Child Dev. 68:139–143.

Kagan, J., D. Arcus, N. Snidman, W. Y. Feng, J. Hendler, and S. Greene, 1994. Reactivity in infants: A cross-national comparison. Dev. Psychol. 30:342–345. Kagan, J., J. S. Reznick, and N. Snidman, 1987. The physiology and psychology of behavioral inhibition in children. Child Dev. 58:1459–1473. Kagan, J., J. S. Reznick, and N. Snidman, 1988a. Biological bases of childhood shyness. Science 240:167–171. Kagan, J., J. S. Reznick, and N. Snidman, 1988b. The physiology and psychology of behavioral inhibition in children. Child Dev. 58:1459–1474. Kaufman, J., B. Z. Yang, H. Douglas-Palumberi, S. Houshyar, D. Lipschitz, J. H. Krystal, and J. Gelernter, 2004. Social supports and serotonin transporter gene moderate depression in maltreated children. Proc. Natl. Acad. Sci. USA 101:17316– 17421. Kirkham, N. Z., L. M. Cruess, and A. Diamond, 2003. Helping children apply their knowledge to their behavior on a dimensionswitching task. Dev. Sci. 6(5):449–467. Knutson, B., C. M. Adams, G. W. Fong, and D. Hommer, 2001. Anticipation of increasing monetary reward selectively recruits nucleus accumbens. J. Neurosci. 21:RC159. Kochanska, G., K. Murray, and K. C. Coy, 1997. Inhibitory control as a contributor to conscience in childhood: From toddler to early school age. Child Dev. 68:263–277. Kopp, C. B., 1982. Antecedents of self-regulation: A developmental perspective. Dev. Psychol. 18:199–214. Kopp, C. B., 1989. Regulation of distress and negative emotions: A developmental view. Dev. Psychol. 25:343–354. Kopp, C. B., 2002. Commentary: The codevelopments of attention and emotion regulation. Infancy 3(2):199–208. Lane, R. D., E. M. Reiman, V. B. B. Axelrod, L. -S. Yun, A. Holmes, and G. E. Schwartz, 1998. Neural correlates of levels of emotional awareness: Evidence of an interaction between emotion and attention in the anterior cingulate. J. Cogn. Neurosci. 10:525–535. Lazarus, R. S., and S. Folkman, 1984. Stress, Appraisal, and Coping. New York: Springer. LeDoux, J. E., C. F. Farb, and D. A. Ruggiero, 1990. Topographic organization of neurons in the acoustic thalamus that project to the amygdala. J. Neurosci. 10:1043–1054. LeDoux, J. E., J. Iwata, P. Cicchetti, and D. J. Reis, 1988. Different projections of the central amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear. J. Neurosci. 8:2517–2529. Lee, Y., and M. Davis, 1997. Role of the hippocampus, bed nucleus of the stria terminalis and amygdala in the excitatory effect of corticotrophin releasing hormone (CRH) on the acoustic startle reflex. J. Neurosci. 17:6434–6446. Lesch, K. P., D. Benge, A. Heils, S. Z. Sabol, B. D. Greenberg, S. Petri, et al., 1996. Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 274:1527–1531. Lieberman, M. D., R. Gaunt, D. T. Gilbert, and Y. Trope, 2002. Reflection and reflexion: A social cognitive neuroscience approach to attributional inference. Adv. Exp. Soc. Psychol. 34: 199–249. MacLeod, C., and A. Mathews, 1991. Biased cognitive operations in anxiety: Accessibility of information or assignment of processing priorities. Behav. Res. Ther. 29:599–610. MacLeod, C., E. Rutherford, L. Campbell, G. Ebsworthy, and L. Holker, 2002. Selective attention and emotional vulnerability: Assessing the causal basis of their association through the

experimental manipulation of attentional bias. J. Abnorm. Psychol. 111:107–123. Mandler, G., 1982. Stress and thought processes. In L. Goldberger and S. Breznitz, eds., Handbook of Stress: Theoretical and Clinical Aspects, 88–104. New York: Free Press. Marshall, P. J., and J. Stevenson-Hinde, 1998. Behavioral inhibition, heart period, and respiratory sinus arrhythmia in young children. Dev. Psychobiol. 33:283–292. Mathews, A., and C. MacLeod, 1994. Cognitive approaches to emotion and emotional disorders. Annu. Rev. Psychol. 45:25– 50. McCabe, L. A., M. Cunnington, and J. Brooks-Gunn, 2004. The development of self-regulation in young children: Individual characteristics and environmental contexts. In R. F. Baumeister and K. D. Vohs, eds., Handbook of Self-Regulation: Research, Theory, and Applications, 340–356. New York: Guilford Press. McClure, E. B., C. S. Monk, E. E. Nelson, J. M. Parrish, A. Adler, et al., 2007. Abnormal attention modulation of fear circuit function in pediatric generalized anxiety disorder. Arch. Gen. Psychiatry 64:109–116. Moffitt, T. E., A. Caspi, and M. Rutter, 2005. Strategy for investigating interactions between measured genes and measured environments. Arch. Gen. Psychiatry 62:473–481. Mogg, K., B. P. Bradley, and N. Hallowell, 1994. Attentional bias to threat: Roles of trait anxiety, stressful events, and awareness. Q. J. Exp. Psychol. [A] 47:841–864. Mogg, K., A. M. Mathews, and M. Eysenck, 1992. Attentional biases to threat in clinical anxiety states. Cogn. Emotion 6:149– 159. Monk, C. S., C. Grillon, J. M. P. Baas, E. B. McClure, E. E. Nelson, et al., 2003. A neuroimaging method for the study of threat in adolescents. Dev. Psychobiol. 43:359–366. Monk, C. S., E. E. Nelson, E. B. McClure, K. Mogg, B. P. Bradley, et al., 2006. Ventrolateral prefrontal cortex activation and attention bias in responsive to angry faces in adolescents with generalized anxiety disorder. Am. J. Psychiatry 163:1091– 1097. Monk, C. S., E. E. Nelson, G. Woldehawariat, L. A. Montgomery, E. Zarahn, et al., 2004. Experience-dependent plasticity for attention to threat: Behavioral and neurophysiological evidence in humans. Biol. Psychiatry 56:607–610. Morris, P. L., R. G. Robinson, B. Raphael, and M. J. Hopwood, 1996. Lesion location and poststroke depression. J. Neuropsychiatry Clin. Neurosci. 8:399–403. Munafo, M. R., T. G. Clark, L. R. Moore, E. Payne, R. Walton, and J. Flint, 2003. Genetic polymorphisms and personality in healthy adults: A systematic review and meta-analysis. Mol. Psychiatry 8:471–484. Nachmias, M., M. Gunnar, S. Mangelsdorf, R. H. Parritz, and K. Buss, 1996. Behavioral Inhibition and stress reactivity: The moderating role of attachment security. Child Dev. 67:508– 522. Nelson, C. A., and P. F. Collins, 1991. Event-related potential and looking time analysis of infants’ responses to familiar and novel events: Implications for visual recognition memory. Dev. Psychol. 27:50–58. Newman, J., 1985. The infant cry of primates. In B. Lester and Z. Boukydis, eds., Infant Crying, 307–323. New York: Plenum. Nobre, A. C., G. N. Sebestyen, and C. Miniussi, 2000. The dynamics of shifting visuospatial attention revealed by eventrelated potentials. Neuropsychologia 38:964–974. Ozonoff, S., B. F. Pennington, and S. J. Rogers, 1991. Executive function deficits in high-functioning autistic individuals:

fox et al.: biology of temperament

851

Relationship to theory of mind. J. Child Psychol. Psychiatry 32:1081–1105. Pailing, P. E., and S. J. Segalowitz, 2004. The error-related negativity as a state and trait measure: Motivation, personality, and ERPs in response to errors. Psychophysiology 41:84–95. Pardo, J. V., P. J. Pardo, K. W. Janer, and M. E. Raichle, 1990. The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc. Natl. Acad. Sci. USA 89:5951–5955. Pavlov, I. P., 1927. Conditioned Reflexes. New York: Dover. Perchet, C., and L. García-Larrea, 2000. Visuospatial attention and motor reaction in children: An electrophysiological study of the “Posner” paradigm. Psychophysiology 37:231–241. Perchet, C., O. Revol, P. Fourneret, F. Mauguière, and L. García-Larrea, 2001. Attention shifts and anticipatory mechanisms in hyperactive children: An ERP study using the Posner paradigm. Biol. Psychiatry 50:44–57. Pérez-Edgar, K., and N. A. Fox, 2005a. A behavioral and electrophysiological study of children’s selective attention under neutral and affective conditions. J. Cogn. Dev. 6:89–118. Pérez-Edgar, K., and N. A. Fox, 2005b. Temperament and anxiety disorders. Child Adolesc. Psychiatr. Clin. N. Am. 14:681– 706. Pérez-Edgar, K., N. A. Fox, J. F. Cohn, and M. Kovacs, 2006. Behavioral and electrophysiological markers of selective attention in children of parents with a history of depression. Biol. Psychiatry 60:1131–1138. Pérez-Edgar, K., R. Roberson-Nay, M. G. Hardin, K. Poeth, A. E. Guyer, E. E. Nelson, E. B. McClure, H. A. Henderson, N. A. Fox, D. S. Pine, and M. Ernst, 2007. Attention alters neural responses to evocative faces in behaviorally inhibited adolescents. NeuroImage 35:1538–1546. Posner, M. I., and Y. Cohen, 1984. Components of visual orienting. In H. Bouma and D. Bowhuis, eds., Attention and Performance X, 531–556. Hillsdale, NJ: Erlbaum. Rennie, D. A. C., R. Bull, and A. Diamond, 2004. Executive functioning in preschoolers: Reducing the inhibitory demands of the Dimensional Change Card Sort task. Dev. Neuropsychol. 26:423–443. Reynolds, G. D., and J. E. Richards, 2005. Familiarization, attention, and recognition memory in infancy: An event-related potential and cortical source localization study. Dev. Psychol. 41: 598–615. Reynolds, S. M., and K. C. Berridge, 2002. Positive and negative motivation in nucleus accumbens shell: Bivalent rostrocaudal gradients for GABA-elicited eating, taste “liking”/“disliking” reactions, place preference/avoidance, and fear. J. Neurosci. 22: 7308–7320. Rich, B. A., M. Schmajuk, K. Pérez-Edgar, N. A. Fox, D. S. Pine, and E. Leibenluft, 2007. Response to contingencies and frustration in pediatric bipolar disorder and severe mood dysregulation. Am. J. Psychiatry 164:309–317. Rich, B. A., M. Schmajuk, K. Pérez-Edgar, D. S. Pine, N. A. Fox, and E. Leibenluft, 2005. The impact of reward, punishment, and frustration on attention in pediatric bipolar disorder. Biol. Psychiatry 58:532–539. Ridderinkhof, K. R., and O. van der Stelt, 2000. Attention and selection in the growing child: Views derived from the developmental psychophysiology. Biol. Psychiatry 54:55–106. Roitman, M. F., R. A. Wheeler, and R. M. Carelli, 2005. Nucleus accumbens neurons are innately tuned for rewarding and aversive taste stimuli, encode their predictors, and are linked to motor output. Neuron 45:587–597.

852

emotion⁄cognition interactions

Rosenbaum, J. F., J. Biederman, D. R. Hirshfeld, E. A. Bolduc, S. V. Faraone, and J. Kagan, 1991. Further evidence of an association between behavioral inhibition and anxiety disorders: Results from a family study of children from a nonclinical sample. J. Psychiatr. Res. 25:49–65. Rothbart, M. K., and S. A. Ahadi, 1994. Temperament and the development of personality. J. Abnorm. Psychol. 103:55–66. Rothbart, M. K., and J. E. Bates, 1998. Temperament. In N. Eisenberg, ed., W. Damon, series ed., Handbook of Child Psychology, vol. 3: Social, Emotional, and Personality Development, 5th ed., pp. 105–176. New York: Wiley. Rothbart, M. K., and D. Derryberry, 1981. Development of individual differences in temperament. In M. Lamb and A. Brown, eds., Advances in Developmental Psychology, vol. 1, pp. 37–86. Hillsdale, NJ: Erlbaum. Rothbart, M. K., L. K. Ellis, M. R. Rueda, and M. I. Posner, 2003. Developing mechanisms of temperamental effortful control. J. Pers. 71:1113–1143. Rothbart, M. K., M. I. Posner, and A. Boylan, 1990. Regulatory mechanisms in infant development. In J. T. Enns, ed., The Development of Attention: Research and Theory, 47–66. Oxford, UK: North-Holland. Rothbart, M. K., M. I. Posner, and J. Rosicky, 1994. Orienting in normal and pathological development. Dev. Psychopathol. 6: 635–662. Ruddy, M. G., 1993. Attention shifting and temperament at 5 months. Infant Behav. Dev. 16:255–259. Rueda, M. R., M. I. Posner, and M. K. Rothbart, 2004. Attentional control and self-regulation. In R. F. Baumeister and K. D. Vohs, eds, Handbook of Self-Regulation: Research, Theory, and Applications, 283–300. New York: Guilford Press. Rueda, M. R., M. I. Posner, and M. K. Rothbart, 2005. The development of executive attention: Contributions to the emergence of self-regulation. Dev. Neuropsychol. 28:573–594. Sackeim, H. A., M. A. Greenberg, A. L. Weiman, R. C. Gur, J. P. Hungerbuhler, and N. Geschwind, 1982. Hemispheric asymmetry in the expression of positive and negative emotions. Arch. Neurol. 39:210–218. Satpute, A. B., and M. D. Lieberman, 2006. Integrating automatic and controlled processes into neurocognitive models of social cognition. Brain Res. 1079:86–97. Schippell, P. L., M. W. Vasey, L. M. Cravens-Brown, and R. Bretveld, 2003. Suppressed attention to rejection, ridicule, and failure cues: A specific correlate of reactive but not proactive aggression in youth. J. Clin. Child Adolesc. Psychol. 32:40–55. Schmidt, L. A., and N. A. Fox, 1999. Conceptual, biological, and behavioral distinctions among different categories of shy children. In L. A. Schmidt and J. Schulkin, eds., Extreme Fear, Shyness, and Social Phobia: Origins, Biological Mechanisms, and Clinical Outcomes, 47–66. New York: Oxford University Press. Schmidt, L. A., N. A. Fox, K. H. Rubin, S. Hu, and D. H. Hamer, 2002. Molecular genetics of shyness and aggression in preschoolers. Personality and Individual Differences 33:227–238. Schmidt, L. A., N. A. Fox, K. H. Rubin, E. Sternberg, P. W. Gold, C. Smith, and J. Schulkin, 1997. Behavioral and neuroendocrine responses in shy children. Dev. Psychobiol. 30:127– 140. Schwartz, C. E., N. Snidman, and J. Kagan, 1999. Adolescent social anxiety as an outcome of inhibited temperament in childhood. J. Am. Acad. Child Adolesc. Psychiatry 38:1008–1015. Schwartz, C. E., C. I. Wright, L. M. Shin, J. Kagan, and S. L. Rauch, 2003a. Inhibited and uninhibited infants “grown up”: Adult amygdalar response to novelty. Science 300:1952–1953.

Schwartz, C. E., C. I. Wright, L. M. Shin, J. Kagan, P. J. Whalen, et al., 2003b. Differential amygdalar response to novel versus newly familiar neutral faces: A functional MRI probe developed for studying inhibited temperament. Biol. Psychiatry 53:854–862. Shaffer, D., P. Fisher, M. Dulcan, M. Davies, J. Piacentini, M. Schwab-Stone, B. Lahey, K. Bourdon, P. Jensen, H. Bird, G. Canino, and D. Regier, 1996. The NIMH Diagnostic Interview Schedule for Children Version 2.3 (DISC-2.3): Description, acceptability, prevalence rates, and performance in the MECA study. J. Am. Acad. Child Adolesc. Psychiatry 35:865–877. Silberman, E. K., and H. Weingartner, 1986. Hemispheric lateralization of functions related to emotion. Brain Cogn. 5:322–353. Sobotka, S. S., R. J. Davidson, and J. A. Senulis, 1992. Anterior brain electrical asymmetries in response to reward and punishment. Electroencephalogr. Clin. Neurophysiol. 83:236–247. Strelau, J., 1983. Pavlov’s nervous system typology and beyond. In A. Gale and J. A. Edwards, eds., Physiological Correlates of Human Behavior: Individual Differences and Psychopathology, vol. 3, pp. 139– 154. New York: Academic Press. Suomi, S. J., 2004. How gene-environment interactions influence emotional development in Rhesus monkeys. In C. G. Coll, E. L. Bearer, and R. M. Lerner, eds., Nature and Nurture: The Complex Interplay of Genetic and Environmental Influences on Human Behavior and Development, 35–51. Mahwah, NJ: Lawrence Erlbaum. Sutton, S. K., C. P. Burnette, P. C. Mundy, J. Meyer, A. Vaughan, C. Sanders, and M. Yale, 2005. Resting cortical brain activity and social behavior in higher functioning children with autism. J. Child Psychol. Psychiatry 46:211– 222. Sutton, S. K., and R. J. Davidson, 1997. Prefrontal brain asymmetry: A biological substrate of the behavioral approach and inhibition systems. Psychol. Sci. 8:204–210. Taghavi, M. R., H. T. Neshat-Doost, A. R. Moradi, W. Yule, and T. Dagleish, 1999. Biases in visual attention in children and adolescents with clinical anxiety and mixed depression. J. Abnorm. Child Psychol. 27:215–223. Thayer, J. F., and R. D. Lane, 2000. A model of neurovisceral integration in emotion regulation and dysregulation. J. Affect. Disord. 61:201–216.

Thomas, K. M., W. C. Drevets, R. E. Dahl, N. D. Ryan, B. Birmaher, et al., 2001. Amygdala response to fearful faces in anxious and depressed children. Arch. Gen. Psychiatry 58:1057– 1063. Thompson, R. A., 1990. Emotion and self-regulation. In R. A. Thompson, ed., Socioemotional Development: Nebraska Symposium on Motivation, vol. 36, pp. 367–467. Lincoln: University of Nebraska Press. Vasey, M. W., E. L. Daleeiden, L. L. Williams, and L. Brown, 1995. Biased attention in childhood anxiety disorders: A preliminary study. J. Abnorm. Child Psychol. 23:267–279. Vaughn, B. E., C. B. Kopp, and J. B. Krakow, 1984. The emergence and consolidation of self-control from eighteen to thirty months of age: Normative trends and individual differences. Child Dev. 55:990–1004. Walker, D. L., and M. Davis, 1997. Double dissociation between the involvement of the bed nucleus of the stria terminalis and the central nucleus of the amygdala in light-enhanced versus fear-potentiated startle. J. Neurosci. 17:9375–9938. Welsh, M. C., and B. F. Pennington, 1988. Assessing frontal lobe functioning in children: Views from developmental psychology. Dev. Neuropsychol. 4:199–230. Wheeler, R. E., R. J. Davidson, and A. J. Tomarken, 1993. Frontal brain asymmetry and emotional reactivity: A biological substrate of affective style. Psychophysiology 30:82–89. White, J. L., T. W. Moffit, A. Caspi, D. J. Bartusch, D. J. Needles, and M. Stouthamer-Loeber, 1994. Measuring impulsivity and examining its relationship to delinquency. J. Abnorm. Psychol. 103:192–205. Wilson, E. J., C. MacLeod, A. Mathews, and E. M. Rutherford, 2006. The causal role of interpretive bias in anxiety reactivity. J. Abnorm. Psychol. 115:103–111. Woodward, S. A., 2002. A longitudinal study of infant temperament and brain processing in childhood. Dissertation Abstracts International, Section B: The Sciences and Engineering 63:2093. Cambridge, MA: Harvard University. Zelazo, P. D., D. Frye, and T. Rapus, 1996. An age-related dissociation between knowing rules and using them. Cogn. Dev. 11:37–63. Zelazo, P. D., J. S. Reznick, and D. E. Pinon, 1995. Response control and the execution of verbal rules. Dev. Psychol. 31:508– 517.

fox et al.: biology of temperament

853

52

The Developing Adolescent Brain: Molecular Mechanisms Underlying Nicotine Vulnerability CHARLES F. LANDRY, TERRI L. SCHOCHET, AND ANN E. KELLEY

Adolescence is a unique period of human development that, in general terms, is defined by a progression from dependence on caregivers to independence and self-reliance. Adolescence begins at puberty, and its onset is recognized and celebrated in many cultures (Brooks et al., 2001). It is also a pivotal period in human development because key decisions that influence life trajectory, such as leaving school, moving away from home, or bearing a child, are often made at this time (Rindfuss, 1991). However, adolescence is also a vulnerable stage of life because adverse life-changing experiences such as nonconsensual sex, interpersonal violence, substance abuse, and emotional disorders are more common in this age group (Weaver, 2003; Richter, 2006). Although the teenage years are among the healthiest of the human life span, the highest morbidity and mortality rates in the United States occur within the adolescent period, and over 75 percent of deaths between the ages of 15 and 19 are preventable (R. Anderson and Smith, 2005). Currently, about 20 percent of the world’s population are adolescents, and most live in developing countries (Richter, 2006). Understanding the basis for the inherent vulnerability of this age group to adverse behavior has therefore received renewed attention.

Vulnerability of adolescence Because adolescence is a period of transition, adolescents manifest specific behaviors that are important in gaining confidence and independence. For example, teenagers tend to engage in more peer-oriented activities and spend less time under adult supervision (Csikszentmihalyi, Larson, and Prescott, 1977; Larson and Richards, 1991). Often because of a desire for more autonomy, conflicts with parents and adults in authority become more common (Clemens and Rust, 1979; Riesch et al., 2000). Further, adolescents display specific behavioral traits, including an increase in sensation seeking and reckless behavior, that make them more likely than younger or older individuals to engage in risky activities (Parsons, Siegel, and Cousins, 1997; Donohew et al., 1999). Risk taking and sensation seeking may explain why adolescents are more vulnerable than any other age group to

developing drug-abuse and drug-related problems, why they tend to be polydrug users, and why over 50 percent of 12th grade adolescents report at least some exposure to cigarette smoking (Weaver, 2003; Johnston et al., 2005). Although enhanced exploration of the environment can be viewed at this age as advantageous for developing independence (Spear, 2000), it also places adolescents in a more vulnerable position then other age groups for reckless behavior and drug-related problems. Most adolescents traverse the teenage years with little incident, and in fact, even if faced with adverse living conditions, violence, and parental substance abuse and criminality, the majority of adolescents grow up to lead healthy, productive adult lives (Luthar and Ziglar, 1991). For some, however, poor family relationships and other stresses contribute to conduct problems, an association with deviant peers, and substance abuse (Hawkins, Catalano, and Miller, 1992; Barrett and Turner, 2006). It is not known what predisposes some adolescents to behavioral problems and others to resist aversive behavior even in circumstances that might predict it. However, we are beginning to appreciate that the unique behavioral traits that become pronounced during the adolescent period parallel dramatic developmental changes that occur within specific regions of the adolescent brain (Huttenlocher, 1979; Rosenberg and Lewis, 1995; Giedd et al., 1999; Gogtay et al., 2004; Barnea-Goraly et al., 2005). Understanding the molecular and biochemical basis for these changes may allow us to more fully understand why adolescents are vulnerable to risky behavior and drug abuse, and what programs and therapeutic interventions might be brought to bear when adolescents encounter problems.

What is different about the adolescent brain? Morphological Changes during Adolescence It is now becoming apparent that the brain continues to mature both anatomically and biochemically throughout adolescence and may not reach full maturity until young adulthood (Toga, Thompson, and Sowell, 2006). Although the notion

855

that the brain continues to mature during the adolescent period is not new (Huttenlocher, 1979), magnetic resonance imaging (MRI) studies have enabled clear visualization of these changes and have sparked renewed interest in their neurochemical underpinnings (Giedd et al., 1999; Gogtay et al., 2004; Barnea-Goraly et al., 2005; Casey, Galvan, and Hare, 2005). The most striking anatomical changes evident during the adolescent period in primates are generally confined to forebrain regions and involve two processes: (1) a reduction, reordering, and stabilization of the synaptic connections through which neurons communicate (Bourgeois, Goldman-Rakic, and Rakic, 1994; S. Anderson et al., 1995; Huttenlocher and Dabholkar, 1997; Woo et al., 1997) and (2) an increase in the thickness and extent of the fatty, myelin-containing sheaths that surround axonal processes (Benes et al., 1994; Giedd et al., 1999; Barnea-Goraly et al., 2005). Myelination, which at the gross anatomical level results in an increase in white matter volume, leads to an enhancement of the rate of electrical conduction and therefore an increase in the speed of communication between neurons. A reduction in synaptic connections (called synaptic pruning) likely contributes to the decrease in gray matter volume evident in imaging studies of adolescent brain (Gogtay et al., 2004; Toga, Thompson, and Sowell, 2006). Both these processes occur at the end of the complex program of neural cell migration and circuit maturation and are generally complete in most brain regions by late childhood. However, forebrain regions like the prefrontal cortex, a brain region involved in executive and cognitive functions related to behaviors such as decision making and reward expectancy (Ridderinkhof et al., 2004; Casey, Galvan, and Hare, 2005; Schoenbaum, Roesch, and Stalnaker, 2006), continue to undergo dynamic changes in gray matter volume that parallel changes in synaptic connectivity that occur during adolescence (Huttenlocher, 1979; Huttenlocher and Dabholkar, 1997; Gogtay et al., 2004; Toga, Thompson, and Sowell, 2006). The continued maturation of brain regions during adolescence that are important for more advanced cognitive functions has significant implications for adolescent behavior and, as will be discussed, drug response and addictive processes. Recent neuroimaging studies indicate that adolescent brain activity during specific cognitive tasks is consistent with latent maturation of forebrain cortical regions (Rubia et al., 2000; Luna et al., 2001; Galvan et al., 2006; Scherf, Sweeney, and Luna, 2006; Yurgelun-Todd and Killgore, 2006). For example, although visuospatial working-memory tasks recruit a similar circuitry in adolescents and adults as measured by functional MRI (fMRI), activity in adolescent prefrontal cortical regions is greater, but also more diffuse, than in adults (Scherf, Sweeney, and Luna, 2006). The more focal prefrontal brain activation observed in adults performing a visuospatial memory task occurs in parallel with activity in

856

emotion⁄cognition interactions

regions modulating preparation and execution of responses, activity not evident in adolescents. These results suggest that there are differences in the manner in which the adolescent brain processes working-memory tasks, assesses performance, and executes cognitive control. Additional fMRI evidence for immature cognitive control in adolescents is also evident in tasks that examine reward responsivity. Whereas overall levels of activation during the response to a reward do not differ between adolescents and adults in the nucleus accumbens (one of the major reward centers of the brain), adolescents display a greater degree of change in this brain region in response to different levels of reward than either adults or children (Galvan et al., 2006). This fact suggests that although the adolescent nucleus accumbens may be functioning in an adult manner, adolescents seem to exhibit greater reward sensitivity. In contrast, both adolescents and children display a similar decreased volume of activity in the prefrontal cortex compared to adults in response to a reward, suggesting that this brain region is less mature in adolescents. Although fMRI studies cannot provide conclusive evidence regarding the mechanisms underlying developmental changes during the adolescent period, these studies suggest that connectivity between regions of the corticolimbic reward system and, in particular, those involved in prefrontal control have not fully matured. The gradient of anatomical maturation that moves rostrally through forebrain regions during adolescence is therefore consistent with changes in cognitive development and indicates that the considerable refinement within mesocortical and corticolimbic brain regions observed during the adolescent period is reflected behaviorally. Given the dramatic changes that occur in the development of the mesocorticolimbic reward system during the adolescent period, it is not surprising that adolescents also appear to respond differently than older age groups to drugs of abuse. For many classes of drugs, dependence in teenagers is associated with a lower threshold of frequency and quantity of drug use (K. Chen, Kandel, and Davies, 1997; Kandel and Chen, 2000; K. Chen and Kandel, 2002). Although factors such as drug class, drug metabolism, and gender are important mediators of drug liability in adolescents (Moolchan, Franken, and Jaszyna-Gasiov, 2006; Ridenour et al., 2006), it is intriguing that areas of the brain implicated in reward processes continue to mature during the adolescent period (figure 52.1). Since all classes of addictive substances act on the mesocorticolimbic reward circuitry (Nestler, 2005), it is likely that developmental changes in this system underlie differences in the way that adolescents respond to drugs of abuse. Neurochemical Changes during Adolescence and the Rodent Model We are only beginning to understand the molecular mechanisms that underlie the anatomical and

Figure 52.1 Key pathways in the brain that are responsible for emotional regulation, cognitive function, and reward sensitivity continue to mature during the adolescent period. Anatomical and neurochemical evidence suggests that the continued maturation of these circuits may underlie the increased propensity of adolescents to engage in risk-taking and sensation-seeking behaviors. Further, these changing systems may be uniquely vulnerable to drugs of abuse, which can induce enduring neuroadaptive alterations within these pathways.

physiological changes that are evident during the adolescent period. Much of our current state of understanding has been obtained from studies on lower mammals, such as rodents, which provide basic biochemical information on mechanisms that cannot be examined at the cellular level in humans. Interestingly, changes in morphology and connectivity that are evident in the human and nonhuman primate adolescent brain based on postmortem and functional imaging studies have also been identified in young rodents (Markus and Petit, 1987; Teicher, Andersen, and Hostetter, 1995; Cunningham, Bhattacharyya, and Benes, 2002; Zuo et al., 2005). Adolescence in rodents, which has been defined as the period from 28 to 42 days after birth (Spear, 2000), with a late adolescent period extending through puberty (Adriani et al., 2002), includes changes in behavior, growth, and sexual maturation that parallel changes in higher mammalian species including humans (Spear, 2000). An increase in synaptic pruning, observed in humans, is also evident in rodents (Markus and Petit, 1987; Zuo et al., 2005). In the rodent brain, dramatic alterations in the synaptic connections between neurons of the cortex occur during adolescence and are reflected by a higher number of predendritic protrusions and a higher turnover of synaptic connections compared to adults (Zuo et al., 2005). Projections from the amygdala to medial prefrontal cortex continue to increase in density during rodent adolescence, suggesting a progressive refinement of connectivity between emotional and cognitive

areas (Cunningham, Bhattacharyya, and Benes, 2002). These results enhance the notion that adolescence is a period of dynamic developmental changes that are conserved phylogenetically and that are pronounced in regions governing emotional and cognitive functions. Many receptor systems and protein modulators important in the complex circuitry connecting different brain regions reach functional maturation prior to the adolescent period (Naeff, Schlumpf, and Lichtensteiger, 1992; Lebrand et al., 1998; Petralia et al., 2005). However, some systems continue to mature during adolescence. For example, the dopamine receptor system continues to mature throughout rodent adolescence in brain regions like the striatum and prefrontal cortex that are associated with reward sensitivity (Gelbard et al., 1989; Teicher, Andersen, and Hostetter, 1995; Andersen et al., 2000). Brain-derived neurotrophic factor, a neuropeptide important in the growth and maintenance of neurons and synaptic processes, continues to increase in prefrontal cortical regions throughout adolescence and does not reach peak levels of expression until young adulthood (Webster et al., 2002, 2006). Little information, however, has been obtained for adolescent-related changes in more basic enzymatic and intracellular signaling pathways, and more broad-based screening methods have begun to be applied to identify molecules important in adolescent behavior and drug response (Polesskaya, Smith, and Fryxell, 2007). Microarray Analysis of Rodent Adolescent Forebrain To screen for genes that undergo altered expression profiles in adolescent compared to adult rodent forebrain, we used DNA microarrays, which enable a comparison of messenger RNA (mRNA) expression levels between brain regions from different ages. Since levels of mRNA predict the amount of encoding protein, we used microarrays to compare genes expressed in adolescent and adult prefrontal cortex and identified a number of genes that differed in distribution and expression between these two ages. One of these genes, an enzyme called 12-lipoxygenase (12-lox), is involved in cell signaling in the brain (Shimizu and Wolfe, 1990; Piomelli, 1994) and was present at higher levels in adolescent brain than in adult. Further biochemical analysis revealed that 12-lox underwent interesting changes in distribution and abundance during adolescent development (figure 52.2A and plate 64). Initially distributed throughout the adolescent forebrain, 12-lox became more confined in expression to discrete cortical layers as the brain matured. This finding is particularly provocative because 12-lox generates the eicosanoid 12-hydroxyeicosatetraenoic acid (12-HETE), a fattyacid signaling molecule (Piomelli, 1994), which has been implicated in learning and memory mechanisms (Dumuis et al., 1988) and has been found to modulate the inhibition by opiates of the release of the inhibitory neurotransmitter

landry, schochet, and kelley: adolescent brain development and nicotine

857

Figure 52.2 Differential expression of 12-lipoxygenase (12-lox) and quinoid dihydropteridine reductase (QDR) in adolescent brain. (A) Color-rendered images from coronal forebrain sections subjected to in situ hybridization revealed higher levels of 12-lox in specific cortical brain regions of the adolescent brain (top panels). Bottom right panels are of sections subjected to emulsion autoradiography to illustrate intense expression of 12-lox in neurons of adolescent compared to adult brain. (B) Top panels are coronal forebrain sections depicting heightened expression of QDR in white matter regions of adult brain. Higher expression levels in

adult forebrain were evident in oligodendrocytes (bottom-right panels). Left-bottom panels in A and B are Northern blots depicting an abundance of 12-lox and QDR mRNA in adolescent and adult brain, respectively. The enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to control for equal loading of RNA in each case. Arrows in A represent neurons in cortical layer IV of adolescent or adult brain that express 12-lox. In B, arrows indicate oligodendrocytes that express QDR in a typical “beadson-a-string” histology pattern evident in both adolescent and adult forebrain. (See plate 64.)

γ-aminobutyric acid (GABA; Vaughan et al., 1997). Changes in the abundance and distribution during adolescent cortical development of the enzyme that generates 12-HETE is interesting because GABAergic interneurons, which influence the modulation of monoaminergic systems that include dopamine, undergo prolonged maturation within the mammalian prefrontal cortex during adolescence (Virgili, Barnabei, and Contestabile, 1990; Vincent, Pabreza, and Benes, 1995; Benes et al., 1996; Benes, Taylor, and Cunningham, 2000; Chugani et al., 2001; Erickson and Lewis, 2002; D. Cruz, Eggan, and Lewis, 2003). Ontological changes in a signaling system important in the modulation of inhibitory reward-related processes are likely to play an important role in the adolescent response to drugs of abuse. In fact, adult mice lacking the 12-lox gene show greater locomotor responses to single acute doses of cocaine (Walters et al., 2003). Our results therefore strongly implicate 12-lox in the age-related differences in drug response and indicate that dramatic differences in signaling pathways associated with reward processes are evident during the adolescent period. A second gene that was identified, the enzyme quinoid dihydropteridine reductase (QDR), was expressed at lower levels in the adolescent cortex. This enzyme was expressed primarily within the cells of white matter tracts (called oligodendrocytes) that generate the axon-ensheathing myelin that facilitates electrical connectivity between neurons (figure 52.2B and plate 64). In the human brain, longitudinal functional imaging studies suggest that both the thickness of the myelin sheath and the diameter of axonal processes, both of which facilitate nerve impulse rate, increase in specific cortical axonal projections such as the corticospinal and thalamocortical tracts (Thompson et al., 2000). Although the extent and time course of these changes, as well as their biochemical underpinnings, remain uncertain (Nagel et al., 2006), a decrease in gray matter volume within forebrain regions parallels an increase in white matter volume. In rodent brain, the process of myelination has been less intensely studied in forebrain regions. Furthermore, genes involved in myelination have been traditionally thought to decrease prior to the adolescent period (Campagnoni and Macklin, 1988). The increase in QDR expression from adolescence to adulthood that we observe is consistent with changes in white matter evident in primates and again underscores the phylogenetic parallels in these processes. Additionally, QDR is a critical enzyme in the generation of the signaling molecule nitrous oxide, which has been implicated in the long-term neural plasticity underlying responsiveness to the incentive value of cocaine reward in adolescent mice (Balda, Anderson, and Itzhak, 2006). Its increase within white matter of the developing adolescent cortex underscores dynamic changes in the maturation of cortical connectivity that occur during

this period and, in parallel with studies in humans, suggests a relationship between changes in white matter volume and addictive processes during adolescence.

Adolescent smoking and the influence of nicotine on the adolescent brain The Problem Understanding the influence of nicotine on the adolescent brain is a compelling problem because cigarette smoking is an addiction that begins during adolescence (Mowery, Brick, and Farrelly, 2000). Further, smoking contributes to numerous health problems and increases an individual’s risk of coronary disease, peripheral vascular disease, stroke, and lung cancer (Fisher, 1958; Gill et al., 1989; Wang et al., 1997). Approximately 44 million people smoke in the United States, and a major national health objective is to reduce the number of smokers in the general population from a current 20 percent to less than 12 percent by 2010 (U.S. Department of Health and Human Services, 2000). This initiative has seen a steady decline over the last decade in the number of those who smoke in the United States and in the daily consumption rates among smokers (Centers for Disease Control, 2005). Of those who smoke, only about 10 percent are under the age of 18. However, about 30 percent of smokers smoked their first cigarette before the age of 14, and 90 percent were dependent by the age of 21 (Mowery, Brick, and Farrelly, 2000). Those who begin to smoke as adolescents experience an increased likelihood of nicotine addiction (Taioli and Wynder, 1991; Breslau and Peterson, 1996) and a greater difficulty in quitting during adulthood (Breslau and Peterson, 1996; J. Chen and Millar, 1998; Khuder, Dayal, and Mutgi, 1999). Although the proportion of adolescents who smoke in the United States has also declined over the last 10 years from a 20-year high in the mid-1990s, the decline in the percentage of 12th graders who smoke has slowed and appears to have begun to level off at just less than 25 percent (Johnston et al., 2005, 2006). Further, the burden of smoking-related health problems in developing countries is likely to increase, since up to one third of the population in less developed regions of the world are between the ages of 10 and 19 (Richter, 2006). Since the persistence of smoking in the adolescent population continues to have a long-term influence on the proportion of smokers, understanding why adolescents experiment with smoking and why many become addicted to nicotine is important if reducing the proportion of smokers in the general population is to be achieved. The Mechanism of Nicotine Action Nicotine, the primary addictive substance in cigarettes, is like other addictive substances in that it has a range of physiological actions. However, its ability to induce dependence is based

landry, schochet, and kelley: adolescent brain development and nicotine

859

on its interaction with the mesolimbic and corticolimbic reward system (Laviolette and van der Kooy, 2004). Nicotine acts in the brain by binding neuronal nicotinic acetylcholine receptors (nAChRs), which are pentameric ligand-gated ion channels that are composed of a combination of α and β subunits, the most important of which in terms of the addictive properties of nicotine are the α4β2 and α7 receptor complexes (McGehee and Role, 1995; Gotti and Clementi, 2004). The addictive properties of nicotine are consistent with its ability to induce the release of dopamine into the nucleus accumbens (and prefrontal cortex) from neurons within the ventral tegmental area (VTA; Nisell, Nomikos, and Svensson, 1994; Pidoplichko et al., 1997). Nicotine enhances the release of dopamine from specific neurons within the VTA by stimulating dopaminergic neurons themselves or by enhancing excitatory (McGehee et al., 1995) or reducing inhibitory (Mansvelder, Keath, and McGehee, 2002) inputs to these cells. The culmination of these effects is an enhanced and sustained release of dopamine onto recipient cell populations, which parallels the action of many other addictive substances. Nicotine has a number of well-established molecular effects on adult neural systems that include an upregulation of nAChRs following chronic nicotine treatment (Wonnacott, 1990; Dani and Heinemann, 1996; Perry et al., 1999). Nicotine has also been found to stimulate locomotor activity in rats and to induce changes in the expression of the immediate early genes c-fos, deltaFosB, and CREB (Nisell et al., 1997; Pich et al., 1997; Kelz et al., 1999; Salminen et al., 1999; Pandey et al., 2001), as well as other genes, such as the GABA B receptor subunit 2 and neurotrophic tyrosine kinase receptor type 2, which have been implicated in smoking behavior in genetic association studies (Sun et al., 2007). Nicotine, like other drugs of abuse, is also known to induce contextual conditioning to drug-paired cues in rodents, which has been an important model for drug craving and relapse (Stewart, de Wit, and Eikelboom, 1984; O’Brien et al., 1998; Bossert et al., 2005; Kalivas, Volkow, and Seamans, 2005). For example, clinical observations and human imaging studies suggest that craving and relapse can be triggered by environmental cues that have come to be associated with the subjective drug state (Wikler, 1948; Childress et al., 1999; Ragozzino, Detrick, and Kesner, 1999) and that visual and tactile cues associated with smoking evoke enhanced activity, as measured by fMRI, in anterior cingulate and orbitofrontal cortex (Brody et al., 2002). We found that an induction of the early-response gene c-fos and the activity-regulated, cytoskeletal-associated protein (arc) within prefrontal cortical regions is associated with an increase in locomotor activity in rats that are placed in an environment where they previously received nicotine (Schroeder, Binzak, and Kelley, 2001; Schiltz, Kelley, and

860

emotion⁄cognition interactions

Landry, 2005). Therefore, although nicotine has broad influences on a number of systems in the brain, geneexpression analysis in nicotine-associated conditioning paradigms suggests that drug-related responses induce the most robust changes within mesolimbic and associated cortical regions. Nicotine and the Adolescent Rodent Brain Rodent models used to examine the influence of nicotine on adolescent behavior suggest that adolescents respond differently to nicotine compared to adults. Sensitivity to the effects of nicotine has primarily been measured by two parameters—reward sensitivity and locomotor activation. Adolescent rats are believed to find nicotine more appetitive than adults. Using a biased conditioned-place-preference experiment (CPP), adolescents show enhanced response to the rewarding effects of nicotine (Vastola et al., 2002). However, a nonbiased design suggests this effect is specific only to young adolescent rats, as neither mid to late adolescents nor adult animals develop CPP (Belluzzi et al., 2004; Leslie et al., 2004). Early adolescent animals also show a strong preference for a nicotine solution over water, a preference that reverses in late adolescence (Adriani et al., 2002). Repeated daily injections of nicotine beginning in early adolescence have been shown to result in increasing locomotor activity with time, a measure of behavioral sensitization, in adult but not adolescent rats (Collins and Izenwasser, 2004; Schochet, Kelley, and Landry, 2004; F. Cruz, Delucia, and Planeta, 2005). Furthermore, unlike adults, adolescents fail to display locomotor conditioning to an environment that has been previously paired with nicotine (Schochet, Kelley, and Landry, 2004). However, a later and longer nicotine exposure results in increased sensitization to the locomotor activating effects of nicotine in adolescence (Faraday, Elliott, and Grunberg, 2001), suggesting that neural mechanisms for this behavior continue to develop throughout the adolescent period. Evidence that adolescents respond to nicotine differently than adults is also illustrated in nicotine self-administration models where limited access to nicotine during late adolescence produces higher levels of nicotine intake during adulthood (Levin et al., 2003). Further, limited access to nicotine in periadolescents did not result in self-administration unless it was administered with acetaldehyde, an intermediary in alchohol metabolism and a component of tobacco smoke (Belluzzi et al., 2005). Finally, if adolescents were allowed to self-administer nicotine over a prolonged access schedule, they rapidly acquired nicotine self-administration. In fact, female adolescents acquired the behavior more rapidly and attained higher levels of stable nicotine self-administration compared to adults (H. Chen, Matta, and Sharp, 2007). Clearly, behavioral evidence strongly suggests that

adolescent rodents respond in specific and unique ways to nicotine and that the neurophysiological progression that leads to dependence may be different from that of adults. There is also growing evidence that unique and prolonged structural and neurochemical changes occur within the adolescent brain following exposure to nicotine. Chronic nicotine treatment during adolescence has been found to generate long-lasting inductions of nAChRs within the midbrain and nucleus accumbens (Trauth et al., 1999; Abreu-Villaca et al., 2003; Adriani et al., 2003), as well as increases in dopamine and decreases in serotonin transporter densities in striatum (Collins et al., 2004). Chronic nicotine also selectively increases the density of calcium-binding proteins within inhibitory interneurons of the anterior cingulate cortex, a brain region closely associated with limbic reward circuitry (Liu et al., 2005). More transient effects of nicotine have been identified within the serotonin receptor system, although dramatic suppression of serotonin presynaptic activity upon cessation of nicotine treatment in adolescent rodents may have strong implications for enhanced withdrawal in adolescent humans (Xu et al., 2001, 2002; Slotkin et al., 2006). A more acute nicotine treatment has also been found to influence receptor systems in adolescent brain. For example, three days of single daily injections of nicotine beginning in early adolescence have been found to decrease serotonin synthesis and tryptophan expression (Jang et al., 2002). Whereas we have found that acute nicotine induces the expression of the early-response genes c-fos and arc in cortical and subcortical regions (Schochet, Kelley, and Landry, 2005), others have observed reduced levels of c-fos in prelimbic cortex following acute nicotine treatment in adolescent animals exposed to gestational nicotine (Park, Loughlin, and Leslie, 2006). Furthermore, adolescent nicotine exposure has been shown to result in persistent increases in numbers of dendrite segments and increased dendrite length in the nucleus accumbens shell (McDonald et al., 2005). Recent data show that although adolescence is typified by a loss of spines and spine precursors in the rodent cortex, a majority of spines that are stabilized during this time will persist thorough adulthood (Zuo et al., 2005), suggesting that the changes elicited by drug exposure during this critical period may have persisting effects. Although we are beginning to uncover the neurochemical changes induced by nicotine during adolescence, and we are finding that a number of receptor, calcium-signaling, and monoamine systems are affected, further work will be required to fully understand the relationship between nicotine and its rapid, as well as long-term, neurochemical influences. Acute Effects of Nicotine on Adolescents What distinguishes adolescent from adult smokers is the early onset and infrequent smoking behavior associated with symptoms

of nicotine dependence (DiFranza et al., 2000; O’Loughlin et al., 2003; Riedel et al., 2003; Gervais et al., 2006). The enhanced addictive liability of nicotine in adolescents has been observed for addictive drugs of many classes (K. Chen, Kandel, and Davies, 1997; K. Chen and Kandel, 2002). However, nicotine is unique in that the first indication of nicotine dependence in adolescents can appear within days to weeks of the onset of occasional use, often before daily use begins. Teenage girls, for example, report withdrawal symptoms consistent with nicotine dependence after smoking an average of two cigarettes over an average of only three weeks (DiFranza et al., 2002). More recent longitudinal studies suggest that in some adolescents, signs of nicotine dependence are evident soon after the first inhalation of nicotine (Gervais et al., 2006; Klein, 2006). Even given the problems associated with self-reports in adolescents (Eissenberg and Balster, 2000; Kandel et al., 2006), these studies suggest that nicotine may be having a very rapid effect on specific neural substrates. The mechanisms that underlie the early onset and infrequent smoking behavior that lead to dependence in adolescents are not known, although their relationship to addiction has recently been discussed (DiFranza and Wellman, 2005). Further, little is known about the immediate response that occurs within the adolescent brain following the earliest smoking experience. Work in the adult rodent model suggests that addiction to drugs like nicotine is associated with a number of biochemical and structural changes that eventually become long-lived (Nestler, 2001; Everitt and Wolf, 2002; Kalivas, 2004; Nestler, 2005). These neuroadaptations involve an initial induction of early-response genes (for example, the transcription factor c-fos) in brain regions associated with memory and natural reward (see next paragraph). These immediate changes are followed by changes in signaling molecules and receptors that are longer in duration. Finally, recurrent use in adults of drugs like nicotine cause long-lasting changes that culminate in a remodeling of dendritic communication sites in reward centers of the brain (Brown and Kollo, 2001). Such changes appear to be very long-lasting and are thought to provide the basis for drug dependence. The adolescent brain, however, is unique because regions like the prefrontal cortex, which are important in drug response, are in the midst of a massive remodeling of synaptic connections that occur simultaneously with enhanced myelination of axonal processes (Bourgeois, Goldman-Rakic, and Rakic, 1994; S. Anderson et al., 1995; Huttenlocher and Dabholkar, 1997; Woo et al., 1997; Giedd et al., 1999). The changes that are occurring en masse in forebrain regions of the adolescent cortex are thought to underlie the experiencedriven plasticity mechanisms required for learning such that neuronal processes, along with synaptic connections that receive specific and pronounced activity, become

landry, schochet, and kelley: adolescent brain development and nicotine

861

permanent. Nicotine, in the context of the adolescent brain, is therefore acting on neural substrates undergoing experience-driven changes in connectivity, many of which overlap the intrinsic reward circuitry. Although we are beginning to understand the neurochemical changes that occur following nicotine administration in adolescents, little biochemical information is available regarding the immediate response of the adolescent brain to nicotine. Given that regions of the adolescent brain are “primed” for experience-driven plastic changes, the acute responses to nicotine are likely to be important in understanding the mechanisms that underlie dependence in adolescents. Acute nicotine affects gene expression in the forebrain of adolescent rodents. The earliest response of cells of the nervous system to activation at the level of gene expression involves the induction of a series of early-response genes (Morgan and Curran, 1989). Many of these, like c-fos, act as transcription factors or transcription coactivators and influence the induction of more latent expressed target genes, which, in turn, engage biological mechanisms. Other early-response genes, like arc, have a more immediate role in affecting changes within the synapse. Arc, for example, has direct effects on neuronal processes governing postsynaptic efficacy (Ramirez-Amaya et al., 2005). Arc is necessary for long-term memory formation (Guzowski et al., 2000) and accumulates in dendrites at sites of recent synaptic activity (Guzowski et al., 2000; Steward and Worley, 2001; Vazdarjanova et al., 2002). Additionally, in adult brain, arc is up-regulated following nicotine, amphetamine, and cocaine administration in a number of brain regions (Fosnaugh et al., 1995; Yamagata et al., 2000; Schiltz, Kelley, and Landry, 2005). Since specific regions of the adolescent brain are actively engaged in synaptic modification and remodeling, examining genes involved in plastic events at the synapse might be instructive regarding early events that occur in the adolescent brain following acute nicotine treatment. We found that basal levels of arc and c-fos were higher in cortical and selective subcortical regions in adolescents compared to adult brain (Schochet, Kelley, and Landry, 2005) in agreement with previous reports on c-fos (Leslie et al., 2004). Further, since arc is considered a marker of synaptic density, our results are in keeping with higher synaptic number and enhanced synapse-related plastic changes that occur in adolescent forebrain regions (Zuo et al., 2005). We also found that acute nicotine caused a dramatic induction of arc in specific cortical and subcortical regions of the adolescent brain and that this induction in some prefrontal cortical regions was much more robust than changes in adult brain (Schochet, Kelley, and Landry, 2005) (figure 52.3A, B and plate 65). In addition to an induction of arc, the immediate gene c-fos was also induced in regions that overlapped arc induction, suggesting enhanced neuronal activity in these

862

emotion⁄cognition interactions

regions. However, unlike c-fos, arc underwent a differential induction in expression in selective regions of the prefrontal cortex. This differential induction is interesting given the role of arc in plasticity mechanisms and is consistent with the action of nicotine on regions important in emotional regulation and decision making. To further examine the putative recruitment of plasticity mechanisms at the synapse in adolescent brain, we asked whether the expression of other genes related to synaptic function are induced by acute nicotine. We found that a synaptic protein called dendrin was also differentially induced in the cortex of adolescent rodents acutely treated with nicotine (figure 52.3C, D and plate 65). Dendrin, a molecule associated with the synapse and expressed almost exclusively with corticolimbic brain regions, is synthesized from a dendritically targeted mRNA (Herb et al., 1997). Although its precise function is unknown, dendrin levels decrease following acute sleep deprivation, consistent with the notion that plasticity molecules in cognitive processes are reduced during sleep deprivation (Neuner-Jehle et al., 1996). A putative role for dendrin in processes related to synaptic plasticity is further suggested by recent evidence showing that perturbation of dendrin mRNA localization or dendrin-cytoskeletal synaptic attachment results in the redistribution of dendrin from synapses to the nucleus (Kremerskothen et al., 2006). More interesting in terms of the action of nicotine on the adolescent brain is our unpublished observation that dendrin expression is not influenced by acute cocaine or exposure of adolescents to a novel environment. Our results, therefore, implicate an enhanced level of synaptic plasticity in specific forebrain cortical regions in the acute response of adolescents to nicotine. It may be that nicotine further stimulates specific plasticity mechanisms in the synapse that are in the midst of experience-driven maturation. The role that these immediate effects confer on the addictive process of nicotine in adolescents remains to be determined.

Summary A mechanism that is beginning to emerge is that the exposure of adolescents to nicotine (and other drugs) may interfere with ongoing neuroplastic events and that regions undergoing the most rigorous changes during this period are especially vulnerable. As we have discussed, many of the molecular components found to change during the adolescent period occur within cortical and subcortical regions linked to monoamine and excitatory forebrain circuits (S. Anderson et al., 1995; Moll et al., 2000; Webster et al., 2002). Ontological changes at the neurochemical level in the adolescent brain overlap corticostriatal and mesolimbic dopamine and glutamate systems that are major players in the response of the brain to naturally rewarding and addictive substances (Koob and Le Moal, 2001; Volkow et al.,

Figure 52.3 Arc and dendrin, which are involved in synaptic plasticity, are differentially induced in adolescent forebrain following acute nicotine. (A) Color-rendered images from coronal forebrain sections hybridized to arc probe in situ revealed a dramatic induction of arc in the prefrontal cortex of acutely nicotine-treated adolescent rodents. (B) A much less dramatic up-regulation of arc

following acute nicotine was evident in adult animals. (C) Dendrin mRNA was also induced in specific forebrain regions following acute nicotine treatment of adolescent rodents, while (D) little change was evident in adult animals. Interaction plots showing the differential induction of (E) arc and (F) dendrin in prefrontal cortex of adolescent and adult animals. (See plate 65.)

2004; Wise, 2004; Kelley et al., 2005; Nestler, 2005). Animal and human studies have identified specific changes during the adolescent period that occur within these systems primarily in regions such as the striatum and prefrontal cortex that are critical for the development of reward sensitivity and risk/reward trade-offs. Further, the rapid onset of dependency to drugs like nicotine during adolescence appears to have a strong neurochemical basis, and nicotine may specifically influence plasticity mechanisms in forebrain regions that continue to develop. Clearly, changes within brain regions important in cognitive and attentional functions as well as emotional regulation and behavioral inhibition may underlie both the propensity of adolescents to engage in risky

and sensation-seeking behavior and the manner in which they experience reward from addictive substances like nicotine.

REFERENCES Abreu-Villaca, Y., F. J. Seidler, D. Qiao, C. A. Tate, M. M. Cousins, I. Thillai, and T. A. Slotkin, 2003. Short-term adolescent nicotine exposure has immediate and persistent effects on cholinergic systems: Critical periods, patterns of exposure, dose thresholds. Neuropsychopharmacology 11:1935–1949. Adriani, W., S. Macri, R. Pacifici, and G. Laviola, 2002. Peculiar vulnerability to nicotine oral self-administration in mice during early adolescence. Neuropsychopharmacology 2:212–224.

landry, schochet, and kelley: adolescent brain development and nicotine

863

Adriani, W., S. Spijker, V. Deroche-Gamonet, G. Laviola, M. Le Moal, A. B. Smit, and P. V. Piazza, 2003. Evidence for enhanced neurobehavioral vulnerability to nicotine during periadolescence in rats. J. Neurosci. 11:4712–4716. Andersen, S. L., A. T. Thompson, M. Rutstein, J. C. Hostetter, and M. H. Teicher, 2000. Dopamine receptor pruning in prefrontal cortex during the periadolescent period in rats. Synapse 2:167–169. Anderson, R. N., and B. L. Smith, 2005. Deaths: Leading causes for 2002. National Vital Statistics Reports, vol. 53, no. 17. Hyattsville, MD: National Center for Health Statistics. Anderson, S. A., J. D. Classey, F. Conde, J. S. Lund, and D. A. Lewis, 1995. Synchronous development of pyramidal neuron dendritic spines and parvalbumin-immunoreactive chandelier neuron axon terminals in layer III of monkey prefrontal cortex. Neuroscience 1:7–22. Balda, M. A., K. L. Anderson, and Y. Itzhak, 2006. Adolescent and adult responsiveness to the incentive value of cocaine reward in mice: Role of neuronal nitric oxide synthase (nNOS) gene. Neuropharmacology 2:341–349. Barnea-Goraly, N., V. Menon, M. Eckert, L. Tamm, R. Bammer, A. Karchemskiy, C. C. Dant, and A. L. Reiss, 2005. White matter development during childhood and adolescence: A crosssectional diffusion tensor imaging study. Cerebral Cortex 12:1848– 1854. Barrett, A. E., and R. J. Turner, 2006. Family structure and substance use problems in adolescence and early adulthood: Examining explanations for the relationship. Addiction 1:109– 120. Belluzzi, J. D., A. G. Lee, H. S. Oliff, and F. M. Leslie, 2004. Age-dependent effects of nicotine on locomotor activity and conditioned place preference in rats. Psychopharmacology (Berl ) 174:389–395. Belluzzi, J. D., R. Wang, and F. M. Leslie, 2005. Acetaldehyde enhances acquisition of nicotine self-administration in adolescent rats. Neuropsychopharmacology 4:705–712. Benes, F. M., M. Turtle, Y. Khan, and P. Farol, 1994. Myelination of a key relay zone in the hippocampal formation occurs in the human brain during childhood, adolescence, and adulthood. Arch. Gen. Psychiatry 6:477–484. Benes, F. M., S. L. Vincent, R. Molloy, and Y. Khan, 1996. Increased interaction of dopamine-immunoreactive varicosities with GABA neurons of rat medial prefrontal cortex occurs during the postweanling period. Synapse 4:237–245. Benes, F. M., J. B. Taylor, and M. C. Cunningham, 2000. Convergence and plasticity of monoaminergic systems in the medial prefrontal cortex during the postnatal period: Implications for the development of psychopathology. Cerebral Cortex 10:1014– 1027. Bossert, J. M., U. E. Ghitza, L. Lu, D. H. Epstein, and Y. Shaham, 2005. Neurobiology of relapse to heroin and cocaine seeking: An update and clinical implications. Eur. J. Pharmacol. 1–3:36–50. Bourgeois, J. P., P. S. Goldman-Rakic, and P. Rakic, 1994. Synaptogenesis in the prefrontal cortex of rhesus monkeys. Cerebral Cortex 1:78–96. Breslau, N., and E. L. Peterson, 1996. Smoking cessation in young adults: Age at initiation of cigarette smoking and other suspected influences. Am. J. Public Health 2:214–220. Brody, A. L., M. A. Mandelkern, E. D. London, A. R. Childress, G. S. Lee, R. G. Bota, M. L. Ho, S. Saxena, L. R. Baxter, Jr., D. Madsen, and M. E. Jarvik, 2002. Brain metabolic changes during cigarette craving. Arch. Gen. Psychiatry 12:1162–1172.

864

emotion⁄cognition interactions

Brooks, L., B. K. Curran, D. Gorman, and A. Scott, (last updated) 2001. Rites of passage. Available at http://www.pages. drexel.edu/~ags25/hpmission.htm. College of Information Science, Drexel University. Accessed August 15, 2006. Brown, R. W., and B. Kolb, 2001. Nicotine sensitization increases dendritic length and spine density in the nucleus accumbens and cingulate cortex. Brain Res. 1–2:94–100. Campagnoni, A. T., and W. B. Macklin, 1988. Cellular and molecular aspects of myelin protein gene expression. Mol. Neurobiol. 1:41–89. Casey, B. J., A. Galvan, and T. A. Hare, 2005. Changes in cerebral functional organization during cognitive development. Curr. Opin. Neurobiol. 2:239–244. Centers for Disease Control, 2005. Cigarette smoking among adults—United States, 2003. MMWR Morb. Mortal. Wkly. Rep. 54:509–513. Chen, H., S. G. Matta, and B. M. Sharp, 2007. Acquisition of nicotine self-administration in adolescent rats given prolonged access to the drug. Neuropsychopharmacology 32:700–709. Chen, J., and W. J. Millar, 1998. Age of smoking initiation: Implications for quitting. Health Rep. 4:39–46 (English); 39–48 (French). Chen, K., and D. Kandel, 2002. Relationship between extent of cocaine use and dependence among adolescents and adults in the United States. Drug Alcohol Depend. 1:65–85. Chen, K., D. B. Kandel, and M. Davies, 1997. Relationships between frequency and quantity of marijuana use and last year proxy dependence among adolescents and adults in the United States. Drug Alcohol Depend. 1–2:53–67. Childress, A. R., P. D. Mozley, W. McElgin, J. Fitzgerald, M. Reivich, and C. P. O’Brien, 1999. Limbic activation during cue-induced cocaine craving. Am. J. Psychiatry 1:11–18. Chugani, D. C., O. Muzik, C. Juhasz, J. J. Janisse, J. Ager, and H. T. Chugani, 2001. Postnatal maturation of human GABAA receptors measured with positron emission tomography. Ann. Neurol. 5:618–626. Clemens, P. W., and J. O. Rust, 1979. Factors in adolescent rebellious feelings. Adolescence 53:159–173. Collins, S. L., and S. Izenwasser, 2004. Chronic nicotine differentially alters cocaine-induced locomotor activity in adolescent vs. adult male and female rats. Neuropharmacology 46:349–362. Collins, S. L., D. Wade, J. Ledon, and S. Izenwasser, 2004. Neurochemical alterations produced by daily nicotine exposure in periadolescent vs. adult male rats. Eur. J. Pharmacol. 1–2: 75–85. Cruz, D. A., S. M. Eggan, and D. A. Lewis, 2003. Postnatal development of pre- and postsynaptic GABA markers at chandelier cell connections with pyramidal neurons in monkey prefrontal cortex. J. Comp. Neurol. 3:385–400. Cruz, F. C., R. Delucia, and C. S. Planeta, 2005. Differential behavioral and neuroendocrine effects of repeated nicotine in adolescent and adult rats. Pharmacol. Biochem. Behav. 3:411–417. Csikszentmihalyi, M., R. Larson, and S. Prescott, 1977. The ecology of adolescent activity and experience. J. Youth Adolesc. 6:281–294. Cunningham, M. G., S. Bhattacharyya, and F. M. Benes, 2002. Amygdalocortical sprouting continues into early adulthood: Implications for the development of normal and abnormal function during adolescence. J. Comp. Neurol. 2:116–130. Dani, J. A., and S. Heinemann, 1996. Molecular and cellular aspects of nicotine abuse. Neuron 5:905–908. DiFranza, J. R., N. A. Rigotti, A. D. McNeill, J. K. Ockene, J. A. Savageau, D. St. Cyr, and M. Coleman, 2000. Initial symp-

toms of nicotine dependence in adolescents. Tob. Control 3:313–319. DiFranza, J. R., J. A. Savageau, N. A. Rigotti, K. Fletcher, J. K. Ockene, A. D. McNeill, M. Coleman, and C. Wood, 2002. Development of symptoms of tobacco dependence in youths: 30-month follow-up data from the DANDY study. Tob. Control 3:228–235. DiFranza, J. R., and R. J. Wellman, 2005. A sensitization-homeostasis model of nicotine craving, withdrawal, and tolerance: Integrating the clinical and basic science literature. Nicotine Tob. Res. 1:9–26. Donohew, R. L., R. H. Hoyle, R. R. Clayton, W. F. Skinner, S. E. Colon, and R. E. Rice, 1999. Sensation seeking and drug use by adolescents and their friends: Models for marijuana and alcohol. J. Stud. Alcohol 5:622–631. Dumuis, A., M. Sebben, L. Haynes, J. P. Pin, and J. Bockaert, 1988. NMDA receptors activate the arachidonic acid cascade system in striatal neurons. Nature 6194:68–70. Eissenberg, T., and R. L. Balster, 2000. Initial tobacco use episodes in children and adolescents: Current knowledge, future directions. Drug Alcohol Depend. 59:S41–60. Erickson, S. L., and D. A. Lewis, 2002. Postnatal development of parvalbumin- and GABA transporter-immunoreactive axon terminals in monkey prefrontal cortex. J. Comp. Neurol. 2: 186–202. Everitt, B. J., and M. E. Wolf, 2002. Psychomotor stimulant addiction: A neural systems perspective. J. Neurosci. 9:3312– 3320. Faraday, M. M., B. M. Elliott, and N. E. Grunberg, 2001. Adult vs. adolescent rats differ in biobehavioral responses to chronic nicotine administration. Pharmacol. Biochem. Behav. 4: 475–489. Fisher, R. A., 1958. Lung cancer and cigarettes. Nature 4628:108. Fosnaugh, J. S., R. V. Bhat, K. Yamagata, P. F. Worley, and J. M. Baraban, 1995. Activation of arc, a putative “effector” immediate early gene, by cocaine in rat brain. J. Neurochem. 5:2377–2380. Galvan, A., T. A. Hare, C. E. Parra, J. Penn, H. Voss, G. Glover, and B. J. Casey, 2006. Earlier development of the accumbens relative to orbitofrontal cortex might underlie risk-taking behavior in adolescents. J. Neurosci. 25:6885– 6892. Gelbard, H. A., M. H. Teicher, G. Faedda, and R. J. Baldessarini, 1989. Postnatal development of dopamine D1 and D2 receptor sites in rat striatum. Brain Res. Dev. Brain Res. 1:123– 130. Gervais, A., J. O’Loughlin, G. Meshefedjian, C. Bancej, and M. Tremblay, 2006. Milestones in the natural course of onset of cigarette use among adolescents. CMAJ 3:255–261. Giedd, J. N., J. Blumenthal, N. O. Jeffries, F. X. Castellanos, H. Liu, A. Zijdenbos, T. Paus, A. C. Evans, and J. L. Rapoport, 1999. Brain development during childhood and adolescence: A longitudinal MRI study. Nature Neurosci. 10:861–863. Gill, J. S., M. J. Shipley, S. A. Tsementzis, R. Hornby, S. K. Gill, E. R. Hitchcock, and D. G. Beevers, 1989. Cigarette smoking: A risk factor for hemorrhagic and nonhemorrhagic stroke. Arch. Intern. Med. 9:2053–2057. Gogtay, N., J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, T. F. Nugent, 3rd, D. H. Herman, L. S. Clasen, A. W. Toga, J. L. Rapoport, and P. M. Thompson, 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 21: 8174–8179.

Gotti, C., and F. Clementi, 2004. Neuronal nicotinic receptors: From structure to pathology. Prog. Neurobiol. 6:363– 396. Guzowski, J. F., G. L. Lyford, G. D. Stevenson, F. P. Houston, J. L. McGaugh, P. F. Worley, and C. A. Barnes, 2000. Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory. J. Neurosci. 11: 3993–4001. Hawkins, J. D., R. F. Catalano, and J. Y. Miller, 1992. Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychol. Bull. 1:64–105. Herb, A., W. Wisden, M. V. Catania, D. Marechal, A. Dresse, and P. H. Seeburg, 1997. Prominent dendritic localization in forebrain neurons of a novel mRNA and its product, dendrin. Mol. Cell. Neurosci. 5:367–374. Huttenlocher, P. R., 1979. Synaptic density in human frontal cortex—Developmental changes and effects of aging. Brain Res. 2:195–205. Huttenlocher, P. R., and A. S. Dabholkar, 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 2:167–178. Jang, M. H., M. C. Shin, T. H. Lee, Y. P. Kim, S. B. Jung, D. H. Shin, H. Kim, S. S. Kim, E. H. Kim, and C. J. Kim, 2002. Alcohol and nicotine administration inhibits serotonin synthesis and tryptophan hydroxylase expression in dorsal and median raphe of young rats. Neurosci. Lett. 2:141–144. Johnston, L. D., P. M. O’Malley, J. G. Bachman, and J. E. Schulenberg, (last updated) 2005. Decline in teen smoking appears to be nearing its end. Available at www.monitoringthe future.org/pressreleases/05cigpr_complete.pdf. University of Michigan News and Information Services, Ann Arbor, MI. Accessed August 15, 2006. Johnston, L. D., P. M. O’Malley, J. G. Bachman, and J. E. Schulenberg, 2006. Monitoring the future: National results on adolescent drug use: Overview of key findings, 2005. NIH Publication No. 06–5882. Bethesda, MD: National Institute on Drug Abuse. Kalivas, P. W., 2004. Recent understanding in the mechanisms of addiction. Curr. Psychiatry Rep. 5:347–351. Kalivas, P. W., N. Volkow, and J. Seamans, 2005. Unmanageable motivation in addiction: A pathology in prefrontalaccumbens glutamate transmission. Neuron 5:647–650. Kandel, D. B., and K. Chen, 2000. Extent of smoking and nicotine dependence in the United States, 1991–1993. Nicotine Tob. Res. 3:263–274. Kandel, D. B., C. Schaffran, P. C. Griesler, M. C. Hu, M. Davies, and N. Benowitz, 2006. Salivary cotinine concentration versus self-reported cigarette smoking: Three patterns of inconsistency in adolescence. Nicotine Tob. Res. 4:525–537. Kelley, A. E., B. A. Baldo, W. E. Pratt, and M. J. Will, 2005. Corticostriatal-hypothalamic circuitry and food motivation: Integration of energy, action and reward. Physiol. Behav. 5: 773–795. Kelz, M. B., J. Chen, W. A. Carlezon, Jr., K. Whisler, L. Gilden, A. M. Beckmann, C. Steffen, Y. J. Zhang, L. Marotti, and D. W. Self, et al., 1999. Expression of the transcription factor deltaFosB in the brain controls sensitivity to cocaine. Nature 6750:272–276. Khuder, S. A., H. H. Dayal, and A. B. Mutgi, 1999. Age at smoking onset and its effect on smoking cessation. Addict. Behav. 5:673–677.

landry, schochet, and kelley: adolescent brain development and nicotine

865

Klein, J. D., 2006. Adolescents and smoking: The first puff may be the worst. CMAJ 3:262. Koob, G. F., and M. Le Moal, 2001. Drug addiction, dysregulation of reward, and allostasis. Neuropsychopharmacology 2:97– 129. Kremerskothen, J., S. Kindler, I. Finger, S. Veltel, and A. Barnekow, 2006. Postsynaptic recruitment of dendrin depends on both dendritic mRNA transport and synaptic anchoring. J. Neurochem. 6:1659–1666. Larson, R., and M. H. Richards, 1991. Daily companionship in late childhood and early adolescence: Changing developmental contexts. Child Dev. 2:284–300. Laviolette, S. R., and D. van der Kooy, 2004. The neurobiology of nicotine addiction: Bridging the gap from molecules to behaviour. Nature Rev. Neurosci. 1:55–65. Lebrand, C., O. Cases, R. Wehrle, R. D. Blakely, R. H. Edwards, and P. Gaspar, 1998. Transient developmental expression of monoamine transporters in the rodent forebrain. J. Comp. Neurol. 4:506–524. Leslie, F. M., S. E. Loughlin, R. Wang, L. Perez, S. Lotfipour, and J. D. Belluzzia, 2004. Adolescent development of forebrain stimulant responsiveness: Insights from animal studies. Ann. NY Acad. Sci. 1021:148–159. Levin, E. D., A. H. Rezvani, D. Montoya, J. E. Rose, and H. S. Swartzwelder, 2003. Adolescent-onset nicotine selfadministration modeled in female rats. Psychopharmacology (Berl.) 2:141–149. Liu, J. J., C. A. Mohila, Y. Gong, N. Govindarajan, and S. P. Onn, 2005. Chronic nicotine exposure during adolescence differentially influences calcium-binding proteins in rat anterior cingulate cortex. Eur. J. Neurosci. 10:2462–2474. Luna, B., K. R. Thulborn, D. P. Munoz, E. P. Merriam, K. E. Garver, N. J. Minshew, M. S. Keshavan, C. R. Genovese, W. F. Eddy, and J. A. Sweeney, 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 5:786–793. Luthar, S. S., and E. Zigler, 1991. Vulnerability and competence: A review of research on resilience in childhood. Am. J. Orthopsychiatry 1:6–22. Mansvelder, H. D., J. R. Keath, and D. S. McGehee, 2002. Synaptic mechanisms underlie nicotine-induced excitability of brain reward areas. Neuron 6:905–919. Markus, E. J., and T. L. Petit, 1987. Neocortical synaptogenesis, aging, and behavior: Lifespan development in the motor-sensory system of the rat. Exp. Neurol. 2:262–278. McDonald, C. G., V. K. Dailey, H. C. Bergstrom, T. L. Wheeler, A. K. Eppolito, L. N. Smith, and R. F. Smith, 2005. Periadolescent nicotine administration produces enduring changes in dendritic morphology of medium spiny neurons from nucleus accumbens. Neurosci. Lett. 385:163–167. McGehee, D. S., M. J. Heath, S. Gelber, P. Devay, and L. W. Role, 1995. Nicotine enhancement of fast excitatory synaptic transmission in CNS by presynaptic receptors. Science 5231: 1692–1696. McGehee, D. S., and L. W. Role, 1995. Physiological diversity of nicotinic acetylcholine receptors expressed by vertebrate neurons. Annu. Rev. Physiol. 57:521–546. Moll, G. H., C. Mehnert, M. Wicker, N. Bock, A. Rothenberger, E. Ruther, and G. Huether, 2000. Ageassociated changes in the densities of presynaptic monoamine transporters in different regions of the rat brain from early juvenile life to late adulthood. Brain Res. Dev. Brain Res. 2:251– 257.

866

emotion⁄cognition interactions

Moolchan, E. T., F. H. Franken, and M. Jaszyna-Gasior, 2006. Adolescent nicotine metabolism: Ethnoracial differences among dependent smokers. Ethnicity Dis. 1:239–243. Morgan, J. I., and T. Curran, 1989. Stimulus-transcription coupling in neurons: Role of cellular immediate-early genes. Trends Neurosci. 11:459–462. Mowery, P. D., P. D. Brick, and M. C. Farrelly, 2000. Pathways to Established Smoking: Results from the 1999 National Youth Tobacco Survey. Legacy First Look Report 3. Washington, DC: American Legacy Foundation. Naeff, B., M. Schlumpf, and W. Lichtensteiger, 1992. Pre- and postnatal development of high-affinity [3H]nicotine binding sites in rat brain regions: An autoradiographic study. Brain Res. Dev. Brain Res. 2:163–174. Nagel, B. J., K. L. Medina, J. Yoshii, A. D. Schweinsburg, I. Moadab, and S. F. Tapert, 2006. Age-related changes in prefrontal white matter volume across adolescence. NeuroReport 13: 1427–1431. Nestler, E. J., 2001. Molecular basis of long-term plasticity underlying addiction. Nature Rev. Neurosci. 2:119–128. Nestler, E. J., 2005. Is there a common molecular pathway for addiction? Nature Neurosci. 11:1445–1449. Neuner-Jehle, M., J. P. Denizot, A. A. Borbely, and J. Mallet, 1996. Characterization and sleep deprivation-induced expression modulation of dendrin, a novel dendritic protein in rat brain neurons. J. Neurosci. Res. 2:138–151. Nisell, M., G. G. Nomikos, K. Chergui, P. Grillner, and T. H. Svensson, 1997. Chronic nicotine enhances basal and nicotineinduced Fos immunoreactivity preferentially in the medial prefrontal cortex of the rat. Neuropsychopharmacology 3:151–161. Nisell, M., G. G. Nomikos, and T. H. Svensson, 1994. Systemic nicotine-induced dopamine release in the rat nucleus accumbens is regulated by nicotinic receptors in the ventral tegmental area. Synapse 1:36–44. O’Brien, C. P., A. R. Childress, R. Ehrman, and S. J. Robbins, 1998. Conditioning factors in drug abuse: Can they explain compulsion? J. Psychopharmacol. 12:15–22. O’Loughlin, J., J. DiFranza, R. F. Tyndale, G. Meshefedjian, E. McMillan-Davey, P. B. Clarke, J. Hanley, and G. Paradis, 2003. Nicotine-dependence symptoms are associated with smoking frequency in adolescents. Am. J. Prev. Med. 3:219–225. Pandey, S. C., A. Roy, T. Xu, and N. Mittal, 2001. Effects of protracted nicotine exposure and withdrawal on the expression and phosphorylation of the CREB gene transcription factor in rat brain. J. Neurochem. 3:943–952. Park, M. K., S. E. Loughlin, and F. M. Leslie, 2006. Gestational nicotine-induced changes in adolescent neuronal activity. Brain Res. 1:119–126. Parsons, J. T., A. W. Siegel, and J. H. Cousins, 1997. Late adolescent risk-taking: Effects of perceived benefits and perceived risks on behavioral intentions and behavioral change. J. Adolesc. 4:381–392. Perry, D. C., M. I. Davila-Garcia, C. A. Stockmeier, and K. J. Kellar, 1999. Increased nicotinic receptors in brains from smokers: Membrane binding and autoradiography studies. J. Pharmacol. Exp. Ther. 3:1545–1552. Petralia, R. S., N. Sans, Y. X. Wang, and R. J. Wenthold, 2005. Ontogeny of postsynaptic density proteins at glutamatergic synapses. Mol. Cell. Neurosci. 3:436–452. Pich, E. M., S. R. Pagliusi, M. Tessari, D. Talabot-Ayer, R. Hooft van Huijsduijnen, and C. Chiamulera, 1997. Common neural substrates for the addictive properties of nicotine and cocaine. Science 5296:83–86.

Pidoplichko, V. I., M. DeBiasi, J. T. Williams, and J. A. Dani, 1997. Nicotine activates and desensitizes midbrain dopamine neurons. Nature 6658:401–404. Piomelli, D., 1994. Eicosanoids in synaptic transmission. Crit. Rev. Neurobiol. 1–2:65–83. Polesskaya, O. O., R. F. Smith, and K. J. Fryxell, 2007. Chronic nicotine doses down-regulate PDE4 isoforms that are targets of antidepressants in adolescent female rats. Biol. Psychiatry 61: 56–64. Ragozzino, M. E., S. Detrick, and R. P. Kesner, 1999. Involvement of the prelimbic-infralimbic areas of the rodent prefrontal cortex in behavioral flexibility for place and response learning. J. Neurosci. 11:4585–4594. Ramirez-Amaya, V., A. Vazdarjanova, D. Mikhael, S. Rosi, P. F. Worley, and C. A. Barnes, 2005. Spatial exploration– induced Arc mRNA and protein expression: Evidence for selective, network-specific reactivation. J. Neurosci. 7:1761– 1768. Richter, L. M., 2006. Studying adolescence. Science 5782:1902– 1905. Ridderinkhof, K. R., M. Ullsperger, E. A. Crone, and S. Nieuwenhuis, 2004. The role of the medial frontal cortex in cognitive control. Science 5695:443–447. Ridenour, T. A., S. T. Lanza, E. C. Donny, and D. B. Clark, 2006. Different lengths of times for progressions in adolescent substance involvement. Addict. Behav. 6:962–983. Riedel, B. W., L. A. Robinson, R. C. Klesges, and B. McLainAllen, 2003. Ethnic differences in smoking withdrawal effects among adolescents. Addict. Behav. 1:129–140. Riesch, S. K., L. Bush, C. J. Nelson, B. J. Ohm, P. A. Portz, B. Abell, M. R. Wightman, and P. Jenkins, 2000. Topics of conflict between parents and young adolescents. J. Soc. Pediatr. Nurs. 1:27–40. Rindfuss, R. R., 1991. The young adult years: Diversity, structural change, and fertility. Demography 4:493–512. Rosenberg, D. R., and D. A. Lewis, 1995. Postnatal maturation of the dopaminergic innervation of monkey prefrontal and motor cortices: A tyrosine hydroxylase immunohistochemical analysis. J. Comp. Neurol. 3:383–400. Rubia, K., S. Overmeyer, E. Taylor, M. Brammer, S. C. Williams, A. Simmons, C. Andrew, and E. T. Bullmore, 2000. Functional frontalisation with age: Mapping neurodevelopmental trajectories with fMRI. Neurosci. Biobehav. Rev. 1:13– 19. Salminen, O., T. Seppa, H. Gaddnas, and L. Ahtee, 1999. The effects of acute nicotine on the metabolism of dopamine and the expression of Fos protein in striatal and limbic brain areas of rats during chronic nicotine infusion and its withdrawal. J. Neurosci. 18:8145–8151. Scherf, K. S., J. A. Sweeney, and B. Luna, 2006. Brain basis of developmental change in visuospatial working memory. J. Cogn. Neurosci. 7:1045–1058. Schiltz, C. A., A. E. Kelley, and C. F. Landry, 2005. Contextual cues associated with nicotine administration increase arc mRNA expression in corticolimbic areas of the rat brain. Eur. J. Neurosci. 6:1703–1711. Schochet, T. L., A. E. Kelley, and C. F. Landry, 2004. Differential behavioral effects of nicotine exposure in adolescent and adult rats. Psychopharmacology (Berl.) 3:265–273. Schochet, T. L., A. E. Kelley, and C. F. Landry, 2005. Differential expression of arc mRNA and other plasticity-related genes induced by nicotine in adolescent rat forebrain. Neuroscience 135:285–297.

Schoenbaum, G., M. R. Roesch, and T. A. Stalnaker, 2006. Orbitofrontal cortex, decision-making and drug addiction. Trends Neurosci. 2:116–124. Schroeder, B. E., J. M. Binzak, and A. E. Kelley, 2001. A common profile of prefrontal cortical activation following exposure to nicotine- or chocolate-associated contextual cues. Neuroscience 3:535–545. Shimizu, T., and L. S. Wolfe, 1990. Arachidonic acid cascade and signal transduction. J. Neurochem. 1:1–15. Slotkin, T. A., C. A. Tate, M. M. Cousins, and F. J. Seidler, 2006. Prenatal nicotine exposure alters the responses to subsequent nicotine administration and withdrawal in adolescence: Serotonin receptors and cell signaling. Neuropsychopharmacology 11:2462–2475. Spear, L. P., 2000. The adolescent brain and age-related behavioral manifestations. Neurosci. Biobehav. Rev. 4:417–463. Steward, O., and P. F. Worley, 2001. Selective targeting of newly synthesized Arc mRNA to active synapses requires NMDA receptor activation. Neuron 1:227–240. Stewart, J., H. de Wit, and R. Eikelboom, 1984. Role of unconditioned and conditioned drug effects in the self-administration of opiates and stimulants. Psychol. Rev. 91:251–268. Sun, D., W. Huang, Y. Y. Hwang, Y. Zhang, Q. Zhang, and M. D. Li, 2007. Regulation by nicotine of Gpr51 and Ntrk2 expression in various rat brain regions. Neuropsychopharmacology 32:110–116. Taioli, E., and E. L. Wynder, 1991. Effect of the age at which smoking begins on frequency of smoking in adulthood. N. Engl. J. Med. 13:968–969. Teicher, M. H., S. L. Andersen, and J. C. Hostetter, Jr., 1995. Evidence for dopamine receptor pruning between adolescence and adulthood in striatum but not nucleus accumbens. Brain Res. Dev. Brain Res. 2:167–172. Thompson, P. M., J. N. Giedd, R. P. Woods, D. MacDonald, A. C. Evans, and A. W. Toga, 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 6774:190–193. Toga, A. W., P. M. Thompson, and E. R. Sowell, 2006. Mapping brain maturation. Trends Neurosci. 3:148–159. Trauth, J. A., F. J. Seidler, E. C. McCook, and T. A. Slotkin, 1999. Adolescent nicotine exposure causes persistent upregulation of nicotinic cholinergic receptors in rat brain regions. Brain Res. 1–2:9–19. U.S. Department of Health and Human Services, 2000. Healthy People 2010: Understanding and Improving Health, 2nd ed. Washington, DC: U.S. Department of Health and Human Services. Vastola, B., L. Douglas, E. Varlinskays, and L. Spear, 2002. Nicotine-induced conditioned place preference in adolescent and adult rats. Physiol. Behav. 77:107. Vaughan, C. W., S. L. Ingram, M. A. Connor, and M. J. Christie, 1997. How opioids inhibit GABA-mediated neurotransmission. Nature 6660:611–614. Vazdarjanova, A., B. L. McNaughton, C. A. Barnes, P. F. Worley, and J. F. Guzowski, 2002. Experience-dependent coincident expression of the effector immediate-early genes arc and Homer 1a in hippocampal and neocortical neuronal networks. J. Neurosci. 23:10067–10071. Vincent, S. L., L. Pabreza, and F. M. Benes, 1995. Postnatal maturation of GABA-immunoreactive neurons of rat medial prefrontal cortex. J. Comp. Neurol. 1:81–92. Virgili, M., O. Barnabei, and A. Contestabile, 1990. Regional maturation of neurotransmitter-related and glial markers during

landry, schochet, and kelley: adolescent brain development and nicotine

867

postnatal development in the rat. Int. J. Dev. Neurosci. 2:159– 166. Volkow, N. D., J. S. Fowler, G. J. Wang, and J. M. Swanson, 2004. Dopamine in drug abuse and addiction: Results from imaging studies and treatment implications. Mol. Psychiatry 6:557–569. Walters, C. L., B. C. Wang, M. Godfrey, D. Sun, C. D. Funk, and J. A. Blendy, 2003. Augmented responses to morphine and cocaine in mice with a 12-lipoxygenase gene disruption. Psychopharmacology (Berl.) 2:124–131. Wang, L., M. Kittaka, N. Sun, S. S. Schreiber, and B. V. Zlokovic, 1997. Chronic nicotine treatment enhances focal ischemic brain injury and depletes free pool of brain microvascular tissue plasminogen activator in rats. J. Cereb. Blood Flow Metab. 2:136–146. Weaver, J., 2003. Adolescents are neurologically more vulnerable to addictions. Science Daily News, Yale University Press, June 19. Webster, M. J., M. M. Herman, J. E. Kleinman, and C. Weickert, 2006. BDNF and trkB mRNA expression in the hippocampus and temporal cortex during the human lifespan. Gene Expr. Patterns 6:941–951. Webster, M. J., C. S. Weickert, M. M. Herman, and J. E. Kleinman, 2002. BDNF mRNA expression during postnatal development, maturation and aging of the human prefrontal cortex. Brain Res. Dev. Brain Res. 2:139–150. Wikler, A., 1948. Recent progress in research on the neurophysiologic basis of morphine addiction. Am. J. Psychiatry 105:329–338.

868

emotion⁄cognition interactions

Wise, R. A., 2004. Dopamine, learning and motivation. Nature Rev. Neurosci. 6:483–494. Wonnacott, S., 1990. The paradox of nicotinic acetylcholine receptor upregulation by nicotine. Trends Pharmacol. Sci. 6:216– 219. Woo, T. U., M. L. Pucak, C. H. Kye, C. V. Matus, and D. A. Lewis, 1997. Peripubertal refinement of the intrinsic and associational circuitry in monkey prefrontal cortex. Neuroscience 4:1149–1158. Xu, Z., F. J. Seidler, S. F. Ali, W. Slikker, Jr., and T. A. Slotkin, 2001. Fetal and adolescent nicotine administration: Effects on CNS serotonergic systems. Brain Res. 1–2:166– 178. Xu, Z., F. J. Seidler, M. M. Cousins, W. Slikker, Jr., and T. A. Slotkin, 2002. Adolescent nicotine administration alters serotonin receptors and cell signaling mediated through adenylyl cyclase. Brain Res. 2:280–292. Yamagata, K., K. Suzuki, H. Sugiura, N. Kawashima, and S. Okuyama, 2000. Activation of an effector immediate-early gene arc by methamphetamine. Ann. NY Acad. Sci. 914:22– 32. Yurgelun-Todd, D. A., and W. D. Killgore, 2006. Fear-related activity in the prefrontal cortex increases with age during adolescence: A preliminary fMRI study. Neurosci. Lett. 3:194– 199. Zuo, Y., A. Lin, P. Chang, and W. B. Gan, 2005. Development of long-term dendritic spine stability in diverse regions of cerebral cortex. Neuron 2:181–189.

53

Environmental Influences on Brain-Behavioral Development: Evidence from Child Abuse and Neglect JESSICA E. SHACKMAN, ALISON B. WISMER FRIES, AND SETH D. POLLAK

The past decade has witnessed a burgeoning of research activity with rodents and nonhuman primates aimed at examining how complex sets of neural networks are shaped and refined by environmental experience (Kaufman et al., 2000; Sánchez, 2006). Recent insights regarding the plasticity of neural systems, coupled with an interest in the effects of early experience on behavioral and brain development, have highlighted how adopting a developmental neuroscience perspective can not only advance basic science about cognitive and affective development, but also inform the study of risk, recovery, and resilience in children (Black et al., 1998; Cicchetti and Curtis, 2006; Pollak, 2005). Of course, it is rarely practical or ethical to experimentally manipulate major aspects of children’s life experiences. For this reason, researchers have increasingly focused on children who have been subjected to species-atypical experiences such as various forms of child maltreatment. Such humanitarian tragedies provide unprecedented opportunities, both to generally advance understanding about human brain-behavioral development and also to inform and motivate interventions for vulnerable children. These types of studies with human children are motivated by research with nonhuman animals, which have shown that adverse parental care shapes the development of the neural systems that underlie patterns of emotional and cognitive processing (Heim, Plotsky, and Nemeroff, 2004; Meaney, 2001). In this chapter we discuss some of the difficulties that have been observed in children who have experienced early maltreatment and consider the neurodevelopmental mechanisms that may link children’s experience with their subsequent behavior.

Developmental outcomes associated with early maltreatment Children exposed to stressful and adverse early environments are at an elevated risk for developing a host of emotional and behavioral problems (Cohen et al., 2006;

Cummings, Davies, and Campbell, 2000). Moreover, children harmed during the first few years of life (i.e., before age 5) show especially pervasive developmental deficits (Keiley et al., 2001; Rutter, 1998), suggesting the existence of sensitive periods for acquiring cognitive and affective competencies (Barinaga, 2000). However, current understanding of the effects of maltreatment on development is limited in two important ways. First, behavioral scientists struggle with the issue of multifinality or heterogeneous outcomes. In this regard, although maltreatment places children at risk in general, it is quite difficult to predict the specific kinds of problems that an individual child will develop. In addition, research suggests that it is actually the experience of maltreatment, rather than a genetic predisposition, that leads to the psychological and behavioral difficulties observed in maltreated children (Jaffee, Caspi, Moffitt, Polo-Tomas, et al., 2004; Jaffee, Caspi, Moffitt, and Taylor, 2004). Second, few studies have explicitly focused on the neural mechanisms that causally link childhood experiences to developmental outcomes. One way in which science in this field is advancing is through increased consideration of the importance of individual differences in children’s experiences, such as the specific type, developmental timing, chronicity, severity, and context of maltreatment. These experiential factors are likely to have differential effects on the developing brain. Yet in reality, it is often impossible for researchers to obtain precise objective information of this sort, especially when children’s environments are chaotic and impoverished. Regarding the heterogeneous nature of outcomes, both anxiety and aggressive-behavior problems occur at fairly high rates in physically abused children (Cohen, Brown, and Smailes, 2001; Dodge et al., 1995; Phillips et al., 2005; Pine and Cohen, 2002). However, physical abuse per se, as well as experiences such as family conflict, parental rejection and hostility, lack of parental warmth and care, and the use of harsh discipline techniques have been identified as contributors to behavioral regulation problems in children (Ge

869

et al., 1996; Zahn-Waxler, Klimes-Dougan, and Slattery, 2000). Of particular concern and interest are the ways in which early experiences may alter the ways in which children will respond to stress later in their lives. For example, when assessed prospectively, physically abused children were found to be at increased risk for developing a lifetime diagnosis of posttraumatic stress disorder (PTSD; Widom, 1999), and adults who had experienced physical abuse as children were found to be more likely to develop PTSD when exposed to severe stressors later in life (Bremner et al., 1993). Abused children are also more vigilant to angerrelated stimuli and show deviant and maladaptive patterns of social information processing (e.g., Weiss et al., 1992). However, although abused children are at increased risk for emotional and behavioral problems, the existence of more general deficits in attention, learning, and memory in these children has not been consistently supported by empirical data (Bartholow et al., 2005; Howe, Cicchetti, and Toth, 2006; Rowe and Eckenrode, 1999). Childhood neglect is also associated with regulatory difficulties, including elevations in internalizing and externalizing symptoms, hyperactivity/distractibility, and poor social adaptation (Chisholm et al., 1995; Crittenden and Ainsworth, 1989; Egeland, Sroufe, and Erickson, 1983; Zeanah, 2000). Like other groups of maltreated children, neglected children show more aggressive symptoms than nonmaltreated children. However, a greater incidence of internalizing symptoms sets neglect apart from physical abuse (Erickson et al., 1996; Manly et al., 2001). In terms of social behavior, neglected children’s interpersonal interactions tend to be characterized by withdrawal, passivity, avoidance, and a general lack of social skills (Camras and Rappaport, 1993; Crittenden, 1992). In fact, problems regulating social interactions are the most commonly reported concern of adoptive parents whose children experienced neglect within orphanage settings (Fisher et al., 1997). In light of the many socioemotional problems experienced by maltreated children, the importance of understanding the neural mechanisms underlying such emotion- and cognitive-processing deficits cannot be understated. Our approach to examining the effects of maltreatment on emotion processing is predicated on an assumption of plasticity in the cognitive and affective systems and the associated neural circuitry underlying such processing (Bjorklund, 1997; Pollak, 2005). A developing brain has a limited storage and processing capacity (Bjorklund, 1997), and as a result, children’s cognitive and affective development is likely to be influenced most strongly by aspects of the environment that are most salient for promoting successful adaptation, regardless of the type or quality of the input (Pollak et al., 2000). Although the ability to track associations between stimuli and outcomes in the environment is adaptive in many situations and facilitates rapid learning, such plasticity may

870

emotion⁄cognition interactions

confer risk for maladaptation if the most prominent signals in a child’s environment are ambiguous, inconsistent, excessive, threatening, or absent. Recent work has focused on elucidating a variety of cognitive, affective, and neural mechanisms that may be causally related to the emergence of psychological problems in maltreated children. Thus, in this chapter, we will examine how the development of these mechanisms can be influenced by early maltreatment.

How does plasticity in developing neurobehavioral systems confer risk? In order to clarify the developmental impact of early adversity on biobehavioral outcomes, it is necessary to identify functional neural mechanisms that can account for observed differences in maltreated children’s cognition, affect, and behavior. Here, we focus first on perceptual and attentional processes and then consider how alterations in these processes may influence the regulation of cognition, emotion, and behavior. Our discussion is not designed to be inclusive of all possible neural mechanisms affected by early experience. Rather, we will discuss select neurobiological mechanisms that are particularly relevant for each process in the context of early childhood maltreatment. The term “selective attention” refers to the idea that not all stimuli in one’s environment can be processed equally because of the brain’s limited capacity. The processes that underlie selective attention appear particularly sensitive to environmental input. At birth, human perceptual systems are already sensitive to information from the environment, and this statement appears to be especially true for social stimuli (Johnson et al., 1992). Regular contingencies within the environment, and between one’s behavior and environmental stimuli, help to develop these attentional skills (Capitanio and Mason, 2000). During normal social development, children learn to tune their attentional systems to selectively attend to the most relevant or salient internal and external events in their environments, while at the same time inhibiting or minimizing attention to irrelevant cues. Studies investigating the prioritized processing of emotional information have focused on the ability of threat-related stimuli to capture attention, because of the importance of threat detection for an individual’s survival. As an example, when presented among an array of neutral faces, threat faces (fearful or angry) are typically detected faster and more accurately than are friendly faces (Fox et al., 2000; Lundqvist and Öhman, 2005; Schubö et al., 2006), suggesting that attention becomes tuned to the processing of these signals in the context of normal development. Neural Mechanisms Involved in Cognitive Processing of Emotion Signals Emotional information is a powerful moderator of perception and selective attention, and several

mechanisms have been proposed to explain this relationship (Compton, 2003). Regions of the occipital and posterior temporal visual cortices play a crucial role in the perceptual processing of socially and emotionally relevant visual stimuli (Adolphs, 2002). For example, recordings from the superior temporal sulcus in monkeys have demonstrated neurons that respond selectively to faces (Gross, Rocha-Miranda, and Bender, 1972). Initial sensory processing of a stimulus can also be influenced by neural activity in regions responsible for encoding a stimulus, which likely involves input from the amygdala to sensory cortical regions (Lang et al., 1998). Several studies have shown that the amygdala is an important structure in the encoding of emotional significance (for review, see Öhman, 2005). Exaggerated amygdala activity in response to viewing negative emotional faces has also been observed in individuals with social anxiety (Phan et al., 2006). Furthermore, the amygdala has been shown to exhibit reciprocal connections to cortical sensory areas (Amaral, Behniea, and Kelly, 2003), and functional connectivity between the amygdala and extrastriate cortex has been demonstrated in the context of viewing emotional images (Morris et al., 1998; Vuilleumier et al., 2004). Thus it is possible that the amygdala exerts a bottom-up or sensory-driven influence over cortical areas, which is responsible for attentional amplification in the presence of salient emotional stimuli. The amygdala may actually be more sensitive to learning object-emotion associations than to the perception of emotional information alone (Hooker et al., 2006). In this study, greater right amygdala activation was found when participants learned new object-facial emotion associations compared to facial emotion perception alone. According to this view, one likely function of the amygdala is to analyze emotionally expressive information in order to learn associations about potentially threatening or rewarding environmental events. Emotional signals may therefore be used to enhance vigilance and prepare learning networks for the acquisition of new information. Nonhuman animal studies suggest that the medial dorsal thalamic nucleus (MD), amygdala, and orbitofrontal cortex may function together to aid in assigning emotional significance to stimuli, a key aspect of learning about emotions (Gaffan and Murray, 1990). In primates, the amygdala projects directly to the MD, which in turn projects to the ventromedial prefrontal cortex (PFC). Significant increases of myelination over the first two years of development in humans appear to strengthen both cortical and subcortical pathways and lead to improved efficiency of the connections between the amygdala and cerebral cortex (Kinney et al., 1988). This amygdala-MD-ventromedial PFC circuit is often discussed as the main circuitry involved in emotional learning and memory-related processes (Aggleton and Mishkin, 1984; Sarter and Markowitsch, 1983). Convergent evidence

about the importance of this neural circuit includes observations that monkeys with lesions of the MD demonstrate impaired object-reward association memory (Gaffan and Parker, 2000) and that rats with lesions of this area have impaired contextual fear conditioning (Li et al., 2004). Recordings of neuronal activity within the MD in rats also indicate that the MD is responsive to auditory and visual stimuli predictive of reward (Oyoshi et al., 1996). In addition, neurons in this region change their response patterns during extinction and relearning trials, suggesting that MD is critical for learning the emotional significance of stimuli in the environment. The ability of emotional information to modulate executive attention, which may function to regulate processing priorities, is likely mediated by frontal cortical regions. The orbital (Bishop, Duncan, and Lawrence, 2004), ventromedial (Baxter et al., 2000), and anterior cingulate (Armony and Dolan, 2002; Bush, Luu, and Posner, 2000; Whalen et al., 1998) areas of the PFC appear particularly suited to exert top-down influences upon the amygdala, in service of emotion processing. In addition, attention to emotionally salient cues may be enhanced through the extensive projections from the amygdala to the occipital cortex (Amaral et al., 1992). A region of the brain collectively referred to as the anterior cingulate cortex (ACC) has also been implicated in the processing of emotional information (Fichtenholtz et al., 2004). In support of this notion, enhanced rostral anterior cingulate cortex (rACC) activation was observed when participants were presented with distracting threat-related facial expressions (Bishop, Duncan, Brett, and Lawrence, 2004). In addition, individuals with high levels of anxiety showed both reduced rACC activation overall and reduced recruitment of dorsolateral and ventrolateral PFC during the establishment of expectancy for threat-related distractors. Activation of the ACC has also been observed when subjects are attending to their own internal emotional responses (Lane et al., 1997), further supporting and broadening the role of the ACC in attention to various types of emotion signals. Thus the development of ACC-mediated networks likely contributes to children’s ability to attend to emotionally salient information, though the specifics of this network and its link to emotional behavior are not yet fully understood. Compton and colleagues (2003) found that the left dorsolateral PFC and ventromedial PFC are involved in maintaining attentional set for both neutral and emotional stimuli. The findings suggest that these areas of the brain represent a common system of selective attention, activated when one must attend to some features of the environment and ignore others. Further, this study demonstrated differential activation in posterior regions of the brain when emotional stimuli were to be ignored as compared to when neutral stimuli were irrelevant, suggesting that areas of the right

shackman, wismer fries, and pollak: environment and brain-behavioral development

871

occipitotemporal region are recruited during tasks that require the filtering out of emotional information. The frontal lobes, ACC, and areas of the parietal cortex undergo prolonged periods of postnatal development that extend into late childhood or puberty (Stuss, 1992). Therefore, it is likely that neural attentional mechanisms involved in emotion-information processing continue to be fine-tuned through early childhood (Klingberg, Forssberg, and Westerberg, 2002). In fact, children’s attentional control improves across development, leading to improved efficiency on tasks that require irrelevant information to be ignored while relevant information is being processed (Enns, 1990; Rueda et al., 2004). As expected, better behavioral performance on attention tasks is related to changes in activation in these brain regions across development (Casey et al., 2004, 1997). However, questions remain about whether and how the type and amount of emotional experience children receive throughout development serve to organize the neural systems that subserve the cognitive processing of emotional cues. Effects of early adversity. A fairly large animal literature confirms that adverse early experiences, such as abuse and neglect, exhibit profound influences on the neural systems reviewed in the previous section. Studies of the effects of amygdala lesions in nonhuman primates indicate that lesions early in life increase fear of conspecifics (Bauman et al., 2004; Prather et al., 2001), whereas lesions in adulthood produce the opposite effect (Emery et al., 2001). Because amygdala connectivity matures fairly early in development, it has been suggested that such developmental differences may reflect variations in pathways connecting the amygdala to other areas of cortex (Amaral, 2002). In addition, molecular genetic studies of the mouse serotonin1A receptor indicate that receptor deletion early in life produces longterm increases in anxiety (Gross et al., 2002). Taken together, this type of evidence points to the importance of understanding childhood experiences as long-term organizers of brain systems involved in processing threat-related information. One mechanism that has been proposed to account for the development of an attentional bias toward threat-related information is that individuals who have experienced more aversive events in their lives have had more opportunity to link certain cues (such as negative facial expressions or harsh words) with particular events by means of associative learning (Fulcher et al., 2001). This observation is likely to be especially true for children growing up in physically abusive environments, as anger and hostility are generally more salient in these situations. Several investigations examining children who have experienced early adversity have revealed distorted patterns of attentional deployment in response to threat-related stimuli. These aberrations include a lower threshold for anger detection in ambiguous facial expres-

872

emotion⁄cognition interactions

sions (Pollak and Sinha, 2002), enhanced attention allocation toward processing angry faces (Pollak et al., 2001), difficulty disengaging attention from angry faces (Pollak and Tolley-Schell, 2003), and greater distractibility in the presence of irrelevant auditory anger signals (Shackman, Shackman, and Pollak, 2007). Another study found that children diagnosed with PTSD secondary to severe physical maltreatment exhibited an attentional bias away from threatening faces that increased in relation to maltreatment and PTSD severity (Pine et al., 2005). In search of the biological processes mediating the impact of early adversity on development, studies in children who have experienced traumatic events have documented abnormalities in hypothalamic-pituitary-adrenal (HPA) axis regulation, neurotransmitter function, and brain structure (Bremner and Vermetten, 2001; De Bellis, 2001; Kaufman and Charney, 2001). Furthermore, decreased neuronal metabolism in the ACC has been found in maltreated children and adolescents diagnosed with PTSD (De Bellis et al., 2000). However, fewer studies have thus far been aimed at uncovering functional changes in portions of the neural circuitry involved in emotion and cognition that could be affected by early trauma. Adults with PTSD resulting from various forms of adversity (abuse, rape, combat exposure) have been shown to exhibit regions of abnormal brain function, including greater amygdala activation in response to negative emotional faces (Rauch et al., 2000) and imagery (Shin et al., 1997), as well as reduced rACC (Shin et al., 2001) and medial PFC activation (Bremner et al., 1999). Further work has shown that adult women with childhood abuse-related PTSD evince decreased activity in the ACC, a brain region associated with inhibition of responses as well as attention and emotion regulation, during an emotional Stroop task (Bremner et al., 2004). Indeed, research suggests that neural mechanisms involved in attention to threat are plastic and can be influenced by experience. As evidence in favor of this hypothesis, one study utilized an emotional variant of the Posner cuing task in healthy individuals and found changes over time in the occipitotemporal cortex, an area of the brain thought to be involved in the detailed processing necessary for the identification of stimuli as safe or threatening (Monk et al., 2004). Taken together, the research reviewed in this section suggests that experiences that impact the processing of threat-related information have the potential to increase an individual’s risk for developing aberrant patterns of emotional information processing. We have proposed that children developing in maltreating environments come to develop learned associations between particularly salient stimuli and outcomes, which over time come to be generalized to situations outside of the primary learning context, conferring risk for maladaptation (Pollak, 2003). However, an alternative although not or-

thogonal perspective is that the experience of chronic stress early in life may also alter the developmental trajectory of brain regions involved in the cognitive processing of threatening information, as well as emotion and cognitive regulatory systems (De Bellis et al., 1999, 2002). However, at this point, more empirical evidence is required to link structural changes in the brain to altered functional outcomes in maltreated children. Several nonhuman animal studies have demonstrated the profound effect of social neglect on attentional processes, including problems in the selection of relevant information and an inability to ignore redundant or irrelevant information (see Gilmer and McKinney, 2003). Research with rhesus monkeys has shown a persistent inability to habituate to novelty immediately following early social deprivation (Harlow, Dodsworth, and Harlow, 1965), as well as several years postdeprivation (Fittinghoff et al., 1974; Harlow, Schiltz, and Harlow, 1969; Sackett, 1967), possibly also because of deficits in selective attention. Maternally deprived monkeys tend to allocate their attention to any stimulus in their perceptual field regardless of whether or not the stimulus is relevant (Beauchamp et al., 1991). Another study utilizing isolate-reared monkeys found that social deprivation early in life, even if followed by 3–4 years of social opportunities, leads to deficits in attending to relevant socioemotional cues critical for guiding appropriate behavioral responses (R. Miller, Caul, and Mirsky, 1967). Based upon experimental work with nonhuman animals, it is likely that early experiences of neglect may also impact how human children attend to social information. Indeed, research conducted with neglected children suggests that inadequate, impoverished, or inconsistent social and emotional input may lead to differences in attention to social and emotional cues. For example, Pollak and colleagues (2000) have reported that neglected children had difficulty differentiating between facial expressions of basic emotions. Additionally, clinical descriptions of 3- to 6month-old children in residential nurseries, where social and emotional neglect was prominent, indicate that visual discrimination abilities are impaired for social stimuli (Provence and Lipton, 1963). These findings suggest that the infants attended to features or aspects of the social environment differently than typical infants. Interestingly, their ability to discriminate between the nursing bottle (a stimulus with which they had a great many consistent experiences) and other objects in their environment was nearly normal. However, recent findings suggest that infants and toddlers who had experienced institutional care did not differ on a gross measure of facial expression discrimination (Nelson, Parker, and Guthrie, 2006), suggesting that some cognitive processing abilities may be preserved in the face of early social deprivation.

Studies utilizing event-related potentials have suggested that postinstitutionalized children exhibit disruptions in the neural circuitry implicated in the recognition of emotional facial expressions (Parker and Nelson, 2005b), as well as deficits in discriminating familiar versus unfamiliar faces (Parker and Nelson, 2005a). Investigations of this sort are especially valuable because they provide specific information on important social cognitive processes, the neural bases of which are relatively well understood. Thus, consistent with the data from physically abused children, it appears that infants and children in institutional settings learn to attend to the most salient stimuli in their environment and those stimuli that have negative or positive reinforcing properties. In typically developing children, these stimuli are likely to be cues related to interactions with their primary caregivers (e.g., tactile, auditory, and visual input from caregivers), leading to increased attention and sensitivity to relevant intra- and interpersonal cues. In an institutional setting, where interpersonal interactions are lacking, attentional mechanisms may not be fine-tuned to pick up on relevant social signals. In a study with preschool children who had experienced profound social neglect secondary to orphanage rearing, Pollak and colleagues (Camras et al., 2006; Wismer Fries and Pollak, 2004) found that early childhood neglect led to difficulties in linking facial expressions of emotion with contextual information generally recognized to elicit a particular emotion and recognizing basic facial expressions of emotion. Both these tasks presumably necessitate the utilization of stored information regarding such things as past situations that have been associated with particular emotional expressions or previous instances where certain facial configurations have been given a verbal label. A lack of such associative learning experiences (as in neglected infants who experience fewer pairings of emotional expressions and reliable or predicted outcomes) may lead to an insufficiency of stored associations and subsequent difficulties in interpreting and understanding emotion cues, which would likely have implications for generating adaptive behavioral responses. Summary. Based on the evidence that we have reviewed, the development of neural systems biased toward processing of threat-related information appears to be a critical first step in the pathway leading to the development of affective and behavioral disorders such as anxiety and aggression in physically abused children. The inability to attend to and understand relevant socioemotional cues, secondary to a lack of consistent and reliable socioemotional input, is likely important for the development of social behavioral problems in neglected children. However, if changes in cognitive processing and associated neural mechanisms are ultimately to be linked to specific maladaptive affective and behavioral outcomes, mechanisms relating information-processing

shackman, wismer fries, and pollak: environment and brain-behavioral development

873

disruptions to differences in emotional experience and regulation must be uncovered. In the following section, we review what is known about the relationship between cognitive processing of emotion and the elicitation of emotion. Emotion Elicitation Cognitive processing of emotional information plays a significant role in determining the extent to which environmental events result in the elicitation of emotion. Furthermore, affective states also have the capacity to selectively enhance or undermine cognitive processing (e.g., Shackman, Sarinopoulos, et al., 2006). For example, threat-related stimuli elicit mild to moderate amounts of affect in experimental participants (e.g., Bradley et al., 2001; Smith, et al., 2006). Similarly, startle potentiation, an index of negative affect, is augmented in infants viewing pictures of angry faces, relative to happy or neutral faces (Balaban, 1995), and in anxious children while viewing unpleasant pictures (Waters, Lipp, and Spence, 2005). Among adults, anxious individuals exhibit greater startle responses to threat cues than nonanxious individuals during states of heightened negative affect (M. Miller and Patrick, 2000), suggesting that negative mood states further enhance an individual’s emotional reactivity to threatening information. One investigation using event-related potentials found that a larger positive slow wave occurred in response to pictures that elicited greater autonomic responses and reports of affective arousal (Cuthbert et al., 2000), suggesting that attentional engagement with a stimulus may lead to greater emotional arousal. On the flip side, significant effects of emotional states on cognitive processing and control have also been recognized. For example, limited-capacity models of cognition predict that emotional states use up resources that are then no longer available for controlled cognitive processing (Ellis and Ashbrook, 1989). Individuals high in state anxiety show a reduced ability to regulate attention in the presence of threat-related information, indicated by reduced activation in both rACC and lateral PFC (Bishop, Duncan, Brett, and Lawrence, 2004). Along these lines, some have postulated that under stressful conditions, regulatory control over ongoing cognitions and behavior may be more difficult to initiate (Metcalfe and Mischel, 1999). Moreover, incidental negative affect has the capacity to impact ongoing cognitive processing of nonemotional information (Gray, 2001; Shackman, Sarinopoulos, et al., 2006). Emotion has also been posited to influence encoding of information in memory through its ability to modulate attention and perception (Easterbrook, 1959). This process is thought to occur primarily through the amygdala’s ability to modulate hippocampal activity, such that emotional arousal is thought to enhance hippocampal-dependent episodic memory consolidation (e.g., McGaugh, 2004). Further evidence comes from studies showing that amygdala activation during encoding

874

emotion⁄cognition interactions

of emotional information predicts the degree of memory retention (Packard and Teather, 1998) and that administration of drugs that block beta-adrenergic receptors in the amygdala also eliminate any effect of emotional arousal on memory consolidation (Cahill et al., 1994). Together, these findings suggest that cognitive processes, such as attention and memory, are indeed sensitive to emotional content. Additionally, it appears that interactions between emotion and cognition are bidirectional and interact in both supportive and disruptive ways. Effects of early adversity. Very few studies have explicitly examined emotional reactivity, or the relationship between cognitive processing of emotion and affective arousal, in children who have been subjected to early adversity. In one study, skin conductance responses (SCRs) were recorded while physically abused children saw and heard expressions of anger produced by their mothers. Not only did maltreated children evince larger SCRs than nonmaltreated children in response to anger, but larger SCRs to anger were predicted by the degree of attention allocated toward processing anger expressions, as measured by the P3b component of the event-related potential (Shackman, Shackman, and Pollak, 2006). Although these data are suggestive of a relationship between attentional processing and emotion elicitation, more work is clearly needed in order to determine whether a causal relationship exists between these two processes, as well as the degree to which such a relationship contributes to increased risk for psychological problems. Therefore, most thinking in this area requires translation to humans from nonhuman animal studies. Numerous nonhuman animal and human studies support the idea that early stressful experiences, including social isolation and deprivation, lead to both heightened emotional reactivity and dysregulation of emotion. Repeated stressful emotional experiences alter the balance of excitatory synaptic inputs in the limbic system—such experience-induced modulation of emotional reactivity is likely to affect an individual’s social and cognitive capacities (Ziabreva et al., 2003). Exposure to early stressors, including maternal separation, is associated with behavioral and physiological hyperarousal and increased reactivity to stress. For instance, neonatal rats exposed to 3 hours of daily maternal separation showed hyperresponsiveness and increased anxiety-like behaviors into adulthood (Caldji et al., 2001). Behavioral observations of primates also find that those monkeys exposed to adverse early rearing conditions are more behaviorally responsive to stressors than mother-reared monkeys (Suomi, 1997). In addition, rhesus monkeys reared with inanimate surrogates have a much stronger negative reaction when separated than do mother-reared ones. Presumably the surrogate-reared monkeys, lacking adequate external “scaffolding” for emotion regulation early in life,

were unable to internalize self-regulatory strategies (Kraemer et al., 1989). Consistent with the notion that deprivation can lead to increased stress reactivity, primates that were isolated or nursery reared often function adequately in a stable, familiar environment, but deviate from controls in response to a challenging, uncontrollable situation (Dettling, Feldon, and Pryce, 2002). In an investigation of the neurobiological correlates of these emotional differences, rhesus monkeys were tested using MRI four years following early maternal separation. The researchers found an increase in right PFC volume, specifically in the ventromedial PFC. Given the importance of the right hemisphere for the experience and expression of negative emotion, these researchers suggest that alterations in ventromedial PFC function may be related to alterations of emotional behavior and HPA regulation found in maternally deprived animals (Lyons et al., 2002). Summary. In this section we have raised the possibility that threat-relevant information has the capacity to elicit negative emotion and that it may do so through enhanced attentional processing. Further, we have suggested that traumatic early experiences, such as maltreatment, that have an impact on the development of information-processing patterns may also increase risk for experiencing negative affect. An alternate hypothesis is that problems result not from enhanced emotional experience per se, but rather from the failure to regulate attention, emotion, or behavior. Thus the degree to which early adversity impacts upon the development of brain systems responsible for regulatory functions should predict increased risk for poor socioemotional and behavioral outcomes. In the following section we examine this hypothesis by presenting evidence for emotion-regulatory difficulties in at-risk populations, and we attempt to link these to maladaptive outcomes. Emotion Regulatory Processes The construct of emotion regulation has been difficult to separate from the construct of emotion itself, because of definitional difficulties as well as difficulties in separating the temporally overlapping processes (Campos et al., 1994; Goldsmith and Davidson, 2004; Pollak, 2005). Although emotion regulation has been conceptualized as including the regulation of affect and its associated physiological states as well as the regulation of expressive behaviors (Eisenberg and Spinrad, 2004), the current discussion will refer specifically to changes associated with activated emotion in response to a stressful event. For a young infant, emotion regulation occurs primarily through extrinsic influences, such as interactions with caregivers, and with age, children become increasingly able to independently manage emotions and their expression (Kopp and Neufeld, 2003). In part, this developmental change is likely due to maturation of neural systems responsible for implementing cognitive regulatory functions, and as a result, cognitive and

emotional processes are intricately related in a developing child (Bell and Wolfe, 2004). Findings from cognitive and affective neuroscience suggest that the neural mechanisms underlying emotion regulation may be identical to those underlying certain aspects of cognitive processing, such as attention. As an example, it has been proposed that one component of the executive attention system, the ACC, is involved in regulating both cognitive and emotional processes, as it exhibits connectivity with prefrontal and parietal cortical regions as well as limbic regions, such as the orbitofrontal cortex, amygdala, and hippocampus (Bush, Luu, and Posner, 2000). In support of this proposal, an investigation of the use of cognitive strategies to regulate emotion has revealed that increased activity in the dorsal ACC and medial and lateral regions of PFC during voluntary affect regulation were associated with attenuated activity in the amygdala (Phan et al., 2005). These changes were also correlated with self-reported decreases in negative affect. Furthermore, studies with nonhuman primates have also implicated the ACC in the contextual regulation of freezing behavior (Kalin et al., 2005). Further evidence for the role of attentional control in emotion regulation comes from studies of temperament in young children. Intrinsic temperamental factors, such as regulation of attention, inhibitory control, and other executive functions, are thought to play a critical role in a young child’s developing ability to regulate emotions and behaviors. Rothbart and colleagues (2006) have proposed that, in general, high levels of attentional control are associated with low levels of negative affect. Consistent with this notion, relationships have emerged between neuropsychological indices of executive attentional control and young children’s capacity for affect and stress regulation (Davis, Bruce, and Gunnar, 2002). In this regard, the modulation of attention appears to take on an increasingly important regulatory role in middle childhood. The main reason for this emphasis is that the executive attention system, which functions to coordinate sensory input, regulate reactive attention systems, and organize emotional and behavioral responses, is critical in promoting adaptive flexibility in response to a changing environment (Derryberry and Reed, 1994). Thus children who are better able to flexibly shift attention in service of modulating emotional arousal are better able to cope in the face of negative emotional experiences. Effects of early adversity. Maternal separation and deprivation studies with rats and monkeys indicate that such early rearing experiences can have long-term impact on emotional behaviors and the brain areas that regulate the expression of negative emotions and the physiology of stress (for review, see Sánchez, Ladd, and Plotsky, 2001). Several investigators have demonstrated the existence of deficits in the ability to regulate emotion among physically maltreated children (e.g.,

shackman, wismer fries, and pollak: environment and brain-behavioral development

875

Shipman and Zeman, 2001). One investigation of the specific contributions of attention and emotion regulation to the development of aggression in maltreated children found that physically abused children exhibited an impaired capacity for attention modulation, which mediated the effects of maltreatment on emotion dysregulation (Shields and Cicchetti, 1998). Preschool-age physically maltreated children also demonstrated an impaired ability to regulate autonomic arousal after hearing a simulated angry exchange (Pollak et al., 2005). Similarly, adults raised in family environments characterized by harsh parenting showed reduced amygdala activation when viewing fearful or angry facial expressions and a positive correlation between amygdala and right ventrolateral PFC activation when labeling emotions, a pattern that was reversed in healthy individuals (Taylor et al., 2006). As the ventrolateral PFC has been shown to be critical for regulating responses to threat (Hariri, Bookheimer, and Mazziotta, 2000), the authors suggested that their results indicate dysregulation of emotional responses at the neural level (Taylor et al., 2006). Secure attachments, based on a pattern of sensitive and responsive caregiving, are related to children’s ability to flexibly regulate negative affect on their own. Chronic emotional unresponsiveness leaves the infant with fewer chances to be effective in drawing upon others as a source of support for emotion regulation (Dawson, Hessl, and Frey, 1994). One study of postinstitutionalized children found that the passivity and lack of emotional expression these children displayed while in the institution was replaced by difficulties in modulating and controlling emotional arousal once placed in adoptive homes (Provence and Lipton, 1963). In addition, neglect experienced early in life is associated with difficulty regulating or containing emotional arousal, especially during social interactions, as well as uncontrollable tantrums and anger, which make it difficult for the children to sustain reciprocal interactions with others (Ames, 1997; Marcovitch et al., 1995; O’Connor et al., 2003). Taken together, these data suggest that a lack of appropriate socioemotional input early in life contributes to difficulty regulating arousal or using others as a source of external regulation. Clinical descriptions indicate that postinstitutionalized children have problems with attention regulation and inhibitory control (Provence and Lipton, 1963). In addition, problems with overactivity were noted especially in social group situations (Rutter, Kreppner, and O’Connor, 2001), suggesting a link between these children’s problems with regulationof-attention and emotion-regulation skills. In terms of emotion regulation, extreme forms of emotional overexuberance, nervous excitement, excessive playfulness typical of much younger children, and difficulty regulating or containing excitement and arousal were all reported for neglected children during a separation-reunion procedure (O’Connor et al., 2003). The authors noted that these

876

emotion⁄cognition interactions

emotion-regulation difficulties made it challenging for the children to maintain a sequence of reciprocal interactions. In addition, observational and teacher-report measures have found neglected children to exhibit more overall negative affect than nonmaltreated children (Egeland, Sroufe, and Erickson, 1983). Summary. The evidence presented in this section suggests a critical role for attention and emotion dysregulation in the development of socioemotional problems, which appear to be particularly at risk for maladaptive development in the context of early childhood adversity. Critically, integration of clinical descriptive work on maltreated children with information on relevant neural circuitry gleaned from nonhuman animal studies is required in order to uncover pathways for future research on cognitive and emotion-regulatory mechanisms in children subjected to early adversity. Behavioral Response and Regulation Thus far we have illustrated how experiences of abuse and neglect may impact the development of cognition-emotion interactions by way of alteration of neural systems underlying attention, emotional responding, and regulation, thereby affecting social learning. Although social learning is an important factor, it cannot alone explain why certain children are at greater or less risk for socioemotional and behavioral problems. Moreover, behavior reflects the culmination of multiple processing and regulatory mechanisms interacting over time. Thus our understanding of why children develop particular adaptive or maladaptive patterns of behavior (and how to change them) critically depends on our understanding of the contributory cognitive, affective, and neural mechanisms. For example, although by definition all physically maltreated children are exposed to affective learning environments characterized by excessive anger and/or violence, much heterogeneity exists in behavioral outcomes. Ideally, differences in these outcomes should be explained by differences in how children attend to and regulate emotional responses to information in their environment, the neural underpinnings of which are shaped by early experiences. Other factors are also likely to interact with experiences and contribute to differences in how children process information in their environment. Features of a child’s temperament that have been shown to be especially predictive of aggressive behavior problems include difficulty with inhibition and effortful control (Olson et al., 2005; Schwartz, Snidman, and Kagan, 1996). Studies have also revealed an important contribution of genetic vulnerability to understanding risk for conduct and other externalizing disorders. Specifically, physically maltreated children who were found to possess an allele of the monoamine oxidase A (MAOA) gene conferring high levels of MAOA expression were less

likely to develop antisocial behavior problems (Caspi et al., 2002). Finally, the neural circuitry underlying dysregulated behavior has been shown to include medial and orbital regions of the PFC. In particular, the orbitofrontal cortex is thought to play a role in mediating behavior based on social context (Blair, 2004), and damage to this area has been associated with impulsive aggression and violence (Blair and Cipolotti, 2000). Thus is it possible that potential differences in genetics, temperament, and neural systems underlying poor behavioral regulation, which are independent of early adversity, may distinguish children who are likely to manifest attentional and emotional dysregulation as disruptive and/ or disinhibited behavior in response to environmental stress.

Conclusion and implications for future research Despite a relatively thorough characterization of the psychological consequences of child maltreatment, research in this area still lacks a comprehensive model to explain and integrate neural processes that contribute to maladaptive development. Until now, most research has relied upon descriptions of children’s observable behavior without consideration of underlying neurobehavioral mechanisms. Thus a goal of this chapter has been to encourage the integrated study of emotional and cognitive processes as they relate to childhood development and the experience of early adversity. In this chapter we have highlighted the experience-dependent nature of cognition-emotion interactions and suggested several possible mechanisms with plausible neural substrates whereby early childhood adversity may alter the development of cognitive and emotional processes. The cumulative impact of poor attention- and emotion-regulatory abilities, in combination with individual differences in child temperament, genetic vulnerabilities, and specific features of each child’s emotional learning environment, is likely to contribute to poor socioemotional outcomes, especially in children who have experienced early adversity. acknowledgments

The writing of this chapter was supported, in part, by NIH Research Grants R01-MH61285 and R01-MH68858 to S. Pollak funded by the National Institute of Mental Health and the Children’s Bureau of the Administration on Children, Youth and Families as part of the Child Neglect Research Consortium; a National Research Service Award (F31-MH073313) to J. Shackman; and a Distinguished Graduate Fellowship from the Waisman Center, University of Wisconsin, to A. Wismer Fries. REFERENCES

Adolphs, R., 2002. Recognizing emotion from facial expressions: Psychological and neurological mechanisms. Behav. Cogn. Neurosci. Rev. 1:21–61.

Aggleton, J. P., and M. Mishkin, 1984. Projections of the amygdala to the thalamus in the cynomolgus monkey. J. Comp. Neurol. 222:56–68. Amaral, D. G., 2002. The primate amygdala and the neurobiology of social behavior: Implications for understanding social anxiety. Biol. Psychiatry 51:11–17. Amaral, D. G., H. Behniea, and J. L. Kelly, 2003. Topographic organization of projections from the amygdala to the visual cortex in the macaque monkey. Neuroscience 118:1099–1120. Amaral, D. G., D. L. Price, A. Pitkanen, and S. T. Carmichael, eds., 1992. The Amygdala. New York: Wiley-Liss. Ames, E., 1997. The Development of Romanian Orphanage Children Adopted to Canada (Final Report to the National Welfare Grants Program: Human Resources Development Canada). Burnaby, BC: Simon Fraser University. Armony, J. L., and R. J. Dolan, 2002. Modulation of spatial attention by fear-conditioned stimuli: An event-related fMRI study. Neuropsychologia 40:807–826. Balaban, M. T., 1995. Affective influences on startle in five-monthold infants: Reactions to facial expressions of emotion. Child Dev. 66:28–36. Barinaga, M., 2000. Neuroscience: A critical issue for the brain. Science 288:2116–2119. Bartholow, B. D., M. A. Pearson, C. L. Dickter, K. J. Sher, M. Fabiani, and G. Gratton, 2005. Strategic control and medial frontal negativity: Beyond errors and response conflict. Psychophysiology 42:33–42. Bauman, M. D., P. Lavenex, W. A. Mason, J. P. Capitanio, and D. G. Amaral, 2004. The development of social behavior following neonatal amygdala lesions in rhesus monkeys. J. Cogn. Neurosci. 16:1388–1411. Baxter, M. G., A. Parker, C. C. C. Lindner, A. D. Izquierdo, and E. A. Murray, 2000. Control of response selection by reinforcer value requires interaction of amygdala and orbital prefrontal cortex. J. Neurosci. 20:4311–4319. Beauchamp, A. J., J. P. Gluck, H. E. Fouty, and M. H. Lewis, 1991. Associative processes in differentially reared rhesus monkeys (Macaca mulatta): Blocking. Dev. Psychobiol. 24:175– 189. Bell, M. A., and C. D. Wolfe, 2004. Emotion and cognition: An intricately bound developmental process. Child Dev. 75:366– 370. Bishop, S., J. Duncan, M. Brett, and A. D. Lawrence, 2004. Prefrontal cortical function and anxiety: Controlling attention to threat-related stimuli. Nature Neurosci. 7:184–188. Bishop, S. J., J. Duncan, and A. D. Lawrence, 2004. State anxiety modulation of the amygdala response to unattended threatrelated stimuli. J. Neurosci. 24:10364–10368. Bjorklund, D. F., 1997. The role of immaturity in human development. Psychol. Bull. 122:153. Black, J. E., T. A. Jones, C. A. Nelson, and W. T. Greenough, 1998. Neuronal plasticity and the developing brain. In N. E. Alessi, J. T. Coyle, S. I. Harrison, and S. Eth, eds., Handbook of Child and Adolescent Psychiatric Science and Treatment, 31–53. New York: Wiley. Blair, R. J. R., 2004. The roles of orbital frontal cortex in the modulation of antisocial behavior. Brain Cogn. 55:198–208. Blair, R. J. R., and L. Cipolotti, 2000. Impaired social response reversal: A case of “acquired sociopathy.” Brain 123:1122– 1141. Bradley, M. M., M. Codispoti, B. N. Cuthbert, and P. J. Lang, 2001. Emotion and motivation. I. Defensive and appetitive reactions in picture processing. Emotion 1:276–298.

shackman, wismer fries, and pollak: environment and brain-behavioral development

877

Bremner, J. D., M. Narayan, L. H. Staib, S. M. Southwick, T. McGlashan, and D. S. Charney, 1999. Neural correlates of memories of childhood sexual abuse in women with and without posttraumatic stress disorder. Am. J. Psychiatry 156:1787–1795. Bremner, J. D., S. M. Southwick, D. R. Johnson, and R. Yehuda, 1993. Childhood physical abuse and combat-related posttraumatic stress disorder in Vietnam veterans. Am. J. Psychiatry 150: 235–239. Bremner, J. D., and E. Vermetten, 2001. Stress and development: Behavioral and biological consequences. Dev. Psychopathol. 13: 473–489. Bremner, J. D., E. Vermetten, M. Vythilingam, N. Afzal, C. Schmahl, B. Elzinga, et al., 2004. Neural correlates of the classic color and emotional Stroop in women with abuse-related posttraumatic stress disorder. Biol. Psychiatry 55:612–620. Bush, G., P. Luu, and M. I. Posner, 2000. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn. Sci. 4:215– 222. Cahill, L., B. Prins, M. Weber, and J. L. McGaugh, 1994. Betaadrenergic activation and memory for emotional events. Nature 371:702–704. Caldji, C., D. Liu, S. Sharma, M. Bodnar, D. Francis, P. M. Plotsky, and M. J. Meaney, 2001. Development of individual differences in behavioral and endocrine responses to stress: Role of the postnatal environment. In B. S. McEwen, ed., Handbook of Physiology: Coping with the Environment, 271–292. New York: Oxford University Press. Campos, J. J., D. L. Mumme, R. Kermoian, and R. G. Campos, 1994. A functionalist perspective on the nature of emotion. Monogr. Soc. Res. Child Dev. 59:284–303. Camras, L. A., S. B. Perlman, A. Wismer Fries, and S. D. Pollak, 2006. Post-institutionalized Chinese and eastern European children: Heterogeneity in the development of emotion understanding. Int. J. Behav. Dev. 30:193–199. Camras, L. A., and S. Rappaport, 1993. Conflict behaviors of maltreated and nonmaltreated children. Child Abuse Negl. 17: 455–464. Capitanio, J. P., and W. A. Mason, 2000. Cognitive style: Problem solving by rhesus macaques (Macaca mulatta) reared with living or inanimate substitute mothers. J. Comp. Psychol. 114:115–125. Casey, B. J., M. C. Davidson, Y. Hara, K. M. Thomas, A. Martinez, A. Galvan, et al., 2004. Early development of subcortical regions involved in non-cued attention switching. Dev. Sci. 7:534–542. Casey, B. J., R. Trainor, J. Giedd, and Y. Vauss, 1997. The role of the anterior cingulate in automatic and controlled processes: A developmental neuroanatomical study. Dev. Psychobiol. 30: 61–69. Caspi, A., J. McClay, T. Moffitt, J. Mill, J. Martin, I. W. Craig, et al., 2002. Role of genotype in the cycle of violence in maltreated children. Science 297:851–854. Chisholm, K., M. C. Carter, E. W. Ames, and S. J. Morison, 1995. Attachment security and indiscriminately friendly behavior in children adopted from Romanian orphanages. Dev. Psychopathol. 7:283–294. Cicchetti, D., and W. J. Curtis, 2006. The developing brain and neural plasticity: Implications for normality, psychopathology, and resilience. In D. Cicchetti and D. J. Cohen, eds., Developmental Psychopathology, vol. 2: Developmental Neuroscience, 2nd ed., pp. 1–64. Hoboken, NJ: John Wiley. Cohen, P., J. Brown, and E. Smailes, 2001. Child abuse and neglect and the development of mental disorders in the general population. Dev. Psychopathol. 13:981.

878

emotion⁄cognition interactions

Cohen, R. A., R. H. Paul, L. Stroud, J. Gunstad, B. L. Hitsman, J. McCaffery, et al., 2006. Early life stress and adult emotional experience: An international perspective. Int. J. Psychiatry Med. 36:35–52. Compton, R. J., 2003. The interface between emotion and attention: A review of evidence from psychology and neuroscience. Behav. Cogn. Neurosci. Rev. 2:115–129. Compton, R. J., M. T. Banich, A. Mohanty, M. P. Milham, J. Herrington, G. A. Miller, et al., 2003. Paying attention to emotion: An fMRI investigation of cognitive and emotional Stroop tasks. Cogn. Affective Behav. Neurosci. 3:81–96. Crittenden, P. M., 1992. Children’s strategies for coping with adverse home environments: An interpretation using attachment theory. Child Abuse Negl. 16:329–343. Crittenden, P. M., and M. D. S. Ainsworth, 1989. Child maltreatment and attachment theory. In D. Cicchetti and V. Carlson, eds., Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect, 432–463. New York: Cambridge University Press. Cummings, E. M., P. T. Davies, and S. B. Campbell, 2000. Developmental Psychopathology and Family Process: Theory, Research, and Clinical Implications. New York: Guilford Press. Cuthbert, B. N., H. T. Schupp, M. M. Bradley, N. Birbaumer, and P. J. Lang, 2000. Brain potentials in affective picture processing: Covariation with autonomic arousal and affective report. Biol. Psychol. 52:95–111. Davis, E. P., J. Bruce, and M. R. Gunnar, 2002. The anterior attention network: Associations with temperament and neuroendocrine activity in 6-year-old children. Dev. Psychobiol. 40:43– 56. Dawson, G., D. Hessl, and K. Frey, 1994. Social influences of early developing biological and behavioral systems related to risk for affective disorder. Dev. Psychopathol. 6:759–779. De Bellis, M. D., 2001. Developmental traumatology: The psychobiological development of maltreated children and its implications for research, treatment, and policy. Dev. Psychopathol. 13:539–564. De Bellis, M. D., M. S. Keshavan, D. B. Clark, B. J. Casey, J. N. Giedd, A. M. Boring, et al., 1999. Developmental traumatology: II. Brain development. Biol. Psychiatry 45:1271–1284. De Bellis, M. D., M. S. Keshavan, H. Shifflett, S. Iyengar, S. R. Beers, J. Hall, et al., 2002. Brain structures in pediatric maltreatment-related posttraumatic stress disorder: A sociodemographically matched study. Biol. Psychiatry 52:1066–1078. De Bellis, M. D., M. S. Keshavan, S. Spencer, and J. Hall, 2000. N-acetylaspartate concentration in the anterior cingulate of maltreated children and adolescents with PTSD. Am. J. Psychiatry 157:1175–1177. Derryberry, D., and M. A. Reed, 1994. Temperament and the self-organization of personality. Dev. Psychopathol. 6:653–676. Dettling, A. C., J. Feldon, and C. R. Pryce, 2002. Early deprivation and behavioral and physiological responses to social separation/novelty in the marmoset. Pharmacol. Biochem. Behav. 73: 259–269. Dodge, K. A., G. S. Pettit, J. E. Bates, and E. Valente, 1995. Social information-processing patterns partially mediate the effect of early physical abuse on later conduct problems. J. Abnorm. Psychol. 104:632–643. Easterbrook, J. A., 1959. The effects of emotion on cue utilization and the organization of behavior. Psychol. Rev. 66:183–201. Egeland, B., A. Sroufe, and M. Erickson, 1983. The developmental consequence of different patterns of maltreatment. Child Abuse Negl. 7:459–469.

Eisenberg, N., and T. L. Spinrad, 2004. Emotion-related regulation: Sharpening the definition. Child Dev. 75:334–339. Ellis, H. C., and P. W. Ashbrook, 1989. The “state” of mood and memory research: A selective review. J. Soc. Behav. Pers. 4:1–21. Emery, N. J., J. P. Capitanio, W. A. Mason, C. J. Machado, S. P. Mendoza, and D. G. Amaral, 2001. The effects of bilateral lesions of the amygdala on dyadic social interactions in rhesus monkeys (Macaca mulatta). Behav. Neurosci. 115:515–544. Enns, J. T., 1990. Relations between components of visual attention. In J. T. Enns, ed., The Development of Attention: Research and Theory, 139–158. Oxford, UK: North-Holland. Erickson, M. F., B. Egeland, J. Briere, L. Berliner, J. A. Bulkley, C. Jenny, et al., 1996. Child neglect. In The APSAC Handbook on Child Maltreatment, 4–20. Thousand Oaks, CA: Sage. Fichtenholtz, H. M., H. L. Dean, D. G. Dillon, H. Yamasaki, G. McCarthy, and K. S. LaBar, 2004. Emotion-attention network interactions during a visual oddball task. Cogn. Brain Res. 20:67–80. Fisher, L., E. W. Ames, K. Chisholm, and L. Savoie, 1997. Problems reported by parents of Romanian orphans adopted to British Columbia. Int. J. Behav. Dev. 20:67–82. Fittinghoff, N. A., D. G. Lindburg, J. Gomber, and G. Mitchell, 1974. Consistency and variability in the behavior of mature, isolation-reared, male rhesus macaques. Primates 15:111–139. Fox, E., V. Lester, R. Russo, R. J. Bowles, A. Pichler, and K. Dutton, 2000. Facial expressions of emotion: Are angry faces detected more efficiently? Cogn. Emotion 14:61–92. Fulcher, E. P., A. Mathews, B. Mackintosh, and S. Law, 2001. Evaluative learning and the allocation of attention to emotional stimuli. Cogn. Ther. Res. 25:261–280. Gaffan, D., and E. A. Murray, 1990. Amygdalar interaction with the mediodorsal nucleus of the thalamus and the ventromedial prefrontal cortex in stimulus-reward associative learning in the monkey. J. Neurosci. 10:3479–3493. Gaffan, D., and A. Parker, 2000. Mediodorsal thalamic function in scene memory in rhesus monkeys. Brain 123:816–827. Ge, X., K. M. Best, R. D. Conger, and R. L. Simons, 1996. Parenting behaviors and the occurrence and co-occurrence of adolescent depressive symptoms and conduct problems. Dev. Psychol. 32:717–731. Gilmer, W. S., and W. T. McKinney, 2003. Early experience and depressive disorders: Human and nonhuman primate studies. J. Affect. Disord. 75:97–113. Goldsmith, H. H., and R. J. Davidson, 2004. Disambiguating the components of emotion regulation. Child Dev. 75:361–365. Gray, J. R., 2001. Emotional modulation of cognitive control: Approach-withdrawal states double-dissociate spatial from verbal two-back task performance. J. Exp. Psychol. [Gen.] 130: 436–452. Gross, C. G., C. E. Rocha-Miranda, and D. B. Bender, 1972. Visual properties of neurons in inferotemporal cortex of the macaque. J. Neurophysiol. 35:96–111. Gross, C., X. Zhuang, K. Stark, S. Ramboz, R. Oosting, L. Kirby, et al., 2002. Serotonin-sub(1A) receptor acts during development to establish normal anxiety-like behaviour in the adult. Nature 416:396–400. Hariri, A. R., S. Y. Bookheimer, and J. C. Mazziotta, 2000. Modulating emotional responses: Effects of a neocortical network on the limbic system. NeuroReport 11:43–48. Harlow, H. F., R. O. Dodsworth, and M. K. Harlow, 1965. Total social isolation in monkeys. Proc. Natl. Acad. Sci. USA 54: 90–97.

Harlow, H. F., M. A. Schiltz, and M. K. Harlow, 1969. Effects of social isolation on the learning performance of rhesus monkeys. In Proceedings of the Second International Congress of Primatology, vol. 1, pp. 178–185. New York: Kargen. Heim, C., P. M. Plotsky, and C. B. Nemeroff, 2004. Importance of studying the contributions of early adverse experience to neurobiological findings in depression. Neuropsychopharmacology 29:641–648. Hooker, C. I., L. T. Germine, R. T. Knight, and M. D’Esposito, 2006. Amygdala response to facial expressions reflects emotional learning. J. Neurosci. 26:8915–8922. Howe, M. L., D. Cicchetti, and S. L. Toth, 2006. Children’s basic memory processes, stress, and maltreatment. Dev. Psychopathol. 18:759–769. Jaffee, S. R., A. Caspi, T. E. Moffitt, M. Polo-Tomas, T. S. Price, and A. Taylor, 2004. The limits of child effects: Evidence for genetically mediated child effects on corporal punishment but not on physical maltreatment. Dev. Psychol. 40:1047. Jaffee, S. R., A. Caspi, T. E. Moffitt, and A. Taylor, 2004. Physical maltreatment victim to antisocial child: Evidence of an environmentally mediated process. J. Abnorm. Psychol. 113:44– 55. Johnson, M. H., S. Dziurawiec, J. Bartrip, and J. Morton, 1992. The effects of movement of internal features on infants’ preferences for facelike stimuli. Infant Behav. Dev. 15:129–136. Kalin, N. H., S. E. Shelton, A. S. Fox, T. R. Oakes, and R. J. Davidson, 2005. Brain regions associated with the expression and contextual regulation of anxiety in primates. Biol. Psychiatry 58:796–804. Kaufman, J., and D. Charney, 2001. Effects of early stress on brain structure and function: Implications for understanding the relationship between child maltreatment and depression. Dev. Psychopathol. 13:451–471. Kaufman, J., P. M. Plotsky, C. B. Nemeroff, and D. S. Charney, 2000. Effects of early adverse experiences on brain structure and function: Clinical implications. Biol. Psychiatry 48:778. Keiley, M. K., T. R. Howe, K. A. Dodge, J. E. Bates, and G. S. Pettit, 2001. The timing of child physical maltreatment: A cross-domain growth analysis of impact on adolescent externalizing and interalizing problems. Dev. Psychopathol. 13:891–912. Kinney, H.C., B.A. Brody, A.S. Kloman, and F. Gilles, 1988. Sequence of central nervous system myelination in human infancy. J. Neuropathol. Exp. Neurol. 47:217–234. Klingberg, T., H. Forssberg, and H. Westerberg, 2002. Increased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14:1–10. Kopp, C. B., and S. J. Neufeld, 2003. Emotional development during infancy. In R. J. Davidson, K. R. Scherer, and H. H. Goldsmith, eds., Handbook of Affective Sciences, 347–374. Oxford, UK: Oxford University Press. Kraemer, G. W., M. H. Ebert, D. E. Schmidt, and W. T. McKinney, 1989. A longitudinal study of the effect of different social rearing conditions on cerebrospinal fluid norepinephrine and biogenic amine metabolites in rhesus monkeys. Neuropsychopharmacology 2:175–189. Lane, R. D., G. R. Fink, P. M. L. Chau, and R. J. Dolan, 1997. Neural activation during selective attention to subjective emotional responses. NeuroReport 8:3969–3972. Lang, P. J., M. M. Bradley, J. R. Fitzsimmons, B. N. Cuthbert, J. D. Scott, B. Moulder, et al., 1998. Emotional arousal and activation of the visual cortex: An fMRI analysis. Psychophysiology 35:199–210.

shackman, wismer fries, and pollak: environment and brain-behavioral development

879

Li, X. B., T. Inoue, S. Nakagawa, and T. Koyama, 2004. Effect of mediodorsal thalamic nucleus lesion on contextual fear conditioning in rats. Brain Res. 1008:261–272. Lundqvist, D., and A. Öhman, 2005. Emotion regulates attention: The relation between facial configurations, facial emotion, and visual attention. Visual Cogn. 12:51–84. Lyons, D. M., H. Afarian, A. F. Schatzberg, A. Sawyer-Glover, and M. E. Moseley, 2002. Experience-dependent asymmetric variation in primate prefrontal morphology. Behav. Brain Res. 136:51–59. Manly, J. T., J. E. Kim, F. A. Rogosch, and D. Cicchetti, 2001. Dimensions of child maltreatment and children’s adjustment: Contributions of developmental timing and subtype. Dev. Psychopathol. 13:759. Marcovitch, S., L. Cesaroni, W. Roberts, and C. Swanson, 1995. Romanian adoption: Parents’ dreams, nightmares, and realities. Child Welfare J. 74:993–1017. McGaugh, J. L., 2004. The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu. Rev. Neurosci. 27:1–28. Meaney, M. J., 2001. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu. Rev. Neurosci. 24:1161–1192. Metcalfe, J., and W. Mischel, 1999. A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychol. Rev. 106:3–19. Miller, M. W., and C. J. Patrick, 2000. Trait differences in affective and attentional responding to threat revealed by emotional Stroop interference and startle reflex modulation. Behav. Ther. 31:757–776. Miller, R. E., W. F. Caul, and I. A. Mirsky, 1967. Communication of affects between feral and socially isolated monkeys. J. Pers. Soc. Psychol. 7:231–239. Monk, C. S., E. E. Nelson, G. Woldehawariat, L. A. Montgomery, E. Zarahn, E. B. McClure, et al., 2004. Experiencedependent plasticity for attention to threat: Behavioral and neurophysiological evidence in humans. Biol. Psychiatry 56: 607–610. Morris, J. S., K. J. Friston, C. Buchel, C. D. Frith, A. W. Young, A. J. Calder, et al., 1998. A neuromodulatory role for the human amygdala in processing emotional facial expressions. Brain 121:47–57. Nelson, C. A., S. W. Parker, and D. Guthrie, 2006. The discrimination of facial expressions by typically developing infants and toddlers and those experiencing early institutional care. Infant Behav. Dev. 29:210–219. O’Connor, T. G., R. S. Marvin, M. Rutter, J. T. Olrick, P. A. Britner, et al., 2003. Child-parent attachment following early institutional deprivation. Dev. Psychopathol. 15:19–38. Öhman, A., 2005. The role of the amygdala in human fear: Automatic detection of threat. Psychoneuroendocrinology 30:953– 958. Olson, S. L., A. J. Sameroff, D. C. R. Kerr, N. L. Lopez, and H. M. Wellman, 2005. Developmental foundations of externalizing problems in young children: The role of effortful control. Dev. Psychopathol. 17:25–45. Oyoshi, T., H. Nishijo, T. Asakura, and Y. Takamura, 1996. Emotional and behavioral correlates of mediodorsal thalamic neurons during associative learning in rats. J. Neurosci. 16: 5812–5829. Packard, M. G., and L. A. Teather, 1998. Amygdala modulation of multiple memory systems: Hippocampus and caudateputamen. Neurobiol. Learn. Mem. 69:163–203.

880

emotion⁄cognition interactions

Parker, S. W., and C. A. Nelson, 2005a. An event-related potential study of the impact of institutional rearing on face recognition. Dev. Psychopathol. 17:621–639. Parker, S. W., and C. A. Nelson, 2005b. The impact of early institutional rearing on the ability to discriminate facial expressions of emotion: An event-related potential study. Child Dev. 76:54–72. Phan, K. L., D. A. Fitzgerald, P. J. Nathan, G. J. Moore, T. W. Uhde, and M. E. Tancer, 2005. Neural substrates for voluntary suppression of negative affect: A functional magnetic resonance imaging study. Biol. Psychiatry 57:210–219. Phan, K. L., D. A. Fitzgerald, P. J. Nathan, and M. E. Tancer, 2006. Association between amygdala hyperactivity to harsh faces and severity of social anxiety in generalized social phobia. Biol. Psychiatry 59:424–429. Phillips, N. K., C. L. Hammen, P. A. Brennan, J. M. Najman, and W. Bor, 2005. Early adversity and the prospective prediction of depressive and anxiety disorders in adolescents. J. Abnorm. Child Psychol. 33:13–24. Pine, D. S., and J. A. Cohen, 2002. Trauma in children and adolescents: Risk and treatment of psychiatric sequelae. Biol. Psychiatry 51:519–531. Pine, D. S., K. Mogg, B. P. Bradley, L. Montgomery, C. S. Monk, E. McClure, et al., 2005. Attention bias to threat in maltreated children: Implications for vulnerability to stressrelated psychopathology. Am. J. Psychiatry 162:291–296. Pollak, S. D., 2003. Experience-dependent affective learning and risk for psychopathology in children. Ann. NY Acad. Sci. 1008: 102–111. Pollak, S. D., 2005. Early adversity and mechanisms of plasticity: Integrating affective neuroscience with developmental approaches to psychopathology. Dev. Psychopathol. 17:735–752. Pollak, S. D., D. Cicchetti, K. Hornung, and A. Reed, 2000. Recognizing emotion in faces: Developmental effects of child abuse and neglect. Dev. Psychol. 36:679–688. Pollak, S. D., R. Klorman, J. E. Thatcher, and D. Cicchetti, 2001. P3b reflects maltreated children’s reactions to facial displays of emotion. Psychophysiology 38:267–274. Pollak, S. D., and P. Sinha, 2002. Effects of early experience on children’s recognition of facial displays of emotion. Dev. Psychol. 38:784–791. Pollak, S. D., and S. A. Tolley-Schell, 2003. Selective attention to facial emotion in physically abused children. J. Abnorm. Psychol. 112:323–338. Pollak, S. D., S. Vardi, A. M. P. Bechner, and J. J. Curtin, 2005. Physically abused children’s regulation of attention in response to hostility. Child Dev. 76:968–977. Prather, M. D., P. Lavenex, M. L. Mauldin-Jourdain, W. A. Mason, J. P. Capitanio, S. P. Mendoza, et al., 2001. Increased social fear and decreased fear of objects in monkeys with neonatal amygdala lesions. Neuroscience 106:653–658. Provence, S., and R. C. Lipton, 1963. Infants in Institutions: A Comparison of Their Development with Family-Reared Infants during the First Year of Life. New York: International Universities Press. Rauch, S. L., P. J. Whalen, L. M. Shin, S. C. McInerney, M. L. Macklin, N. B. Lasko, et al., 2000. Exaggerated amygdala response to masked facial stimuli in posttraumatic stress disorder: A functional MRI study. Biol. Psychiatry 47:769–776. Rothbart, M. K., M. I. Posner, J. Kieras, K. McCartney, and D. Phillips, 2006. Temperament, attention, and the development of self-regulation. In Handbook of Early Childhood Development, 338–357. Malden, MA: Blackwell Publishing.

Rowe, E., and J. Eckenrode, 1999. The timing of academic difficulties among maltreated and nonmaltreated children. Child Abuse Negl. 23:813–832. Rueda, M. R., J. Fan, B. D. McCandliss, J. D. Halparin, D. B. Gruber, L. P. Lercari, et al., 2004. Development of attentional networks in childhood. Neuropsychologia 42:1029–1040. Rutter, M., 1998. Developmental catch-up, and deficit, following adoption after severe global early privation. J. Child Psychol. Psychiatry 39:465–476. Rutter, M. L., J. M. Kreppner, and T. G. O’Connor, 2001. Specificity and heterogeneity in children’s responses to profound institutional privation. Br. J. Psychiatry 179:97–103. Sackett, G. P., 1967. Some persistent effects of different rearing conditions on preadult social behavior of monkeys. J. Comp. Physiol. Psychol. 64:363–365. Sánchez, M. M., 2006. The impact of early adverse care on HPA axis development: Nonhuman primate models. Horm. Behav. 50:623–631. Sánchez, M. M., C. O. Ladd, and P. M. Plotsky, 2001. Early adverse experience as a developmental risk factor for later psychopathology: Evidence from rodent and primate models. Dev. Psychopathol. 13:419–449. Sarter, M., and H. J. Markowitsch, 1983. Reduced resistance to progressive extinction in senescent rats: A neuroanatomical and behavioral study. Neurobiol. Aging 4:203–215. Schubö, A., G. H. E. Gendolla, C. Meinecke, and A. E. Abele, 2006. Detecting emotional faces and features in a visual search paradigm: Are faces special? Emotion 6:246–256. Schwartz, C. E., N. Snidman, and J. Kagan, 1996. Early childhood temperament as a determinant of externalizing behavior in adolescence. Dev. Psychopathol. 8:527–537. Shackman, A. J., I. Sarinopoulos, J. S. Maxwell, D. A. Pizzagalli, A. Lavric, and R. J. Davidson, 2006. Anxiety selectively disrupts visuospatial working memory. Emotion 6:40–61. Shackman, J. E., A. J. Shackman, and S. D. Pollak, 2006. Child abuse and risk for psychopathology. I. Attentional control and the processing of emotional information. Manuscript submitted. Shackman, J. E., A. J. Shackman, and S. D. Pollak, 2007. Physical abuse amplifies attention to threat and increases anxiety in children. Emotion 7:838–852. Shields, A., and D. Cicchetti, 1998. Reactive aggression among maltreated children: The contributions of attention and emotion dysregulation. J. Clin. Child Psychol. 27:381–395. Shin, L. M., S. M. Kosslyn, R. J. McNally, and N. M. Alpert, 1997. Visual imagery and perception in posttraumatic stress disorder: A positron emission tomographic investigation. Arch. Gen. Psychiatry 54:233–241. Shin, L. M., P. J. Whalen, R. K. Pitman, G. Bush, M. L. Macklin, N. B. Lasko, et al., 2001. An fMRI study of anterior cingulate

function in posttraumatic stress disorder. Biol. Psychiatry 50: 932–942. Shipman, K. L., and J. Zeman, 2001. Socialization of children’s emotion regulation in mother-child dyads: A developmental psychopathology perspective. Dev. Psychopathol. 13:317–336. Smith, J. C., A. Löw, M. M. Bradley, and P. J. Lang, 2006. Rapid picture presentation and affective engagement. Emotion 6:208– 214. Stuss, D. T., 1992. Biological and psychological development of executive functions. Brain Cogn. 20:8–23. Suomi, S. J., 1997. Nonverbal communication in nonhuman primates: Implications for the emergence of culture. In U. C. Segerstrale and P. Molnar, eds., Nonverbal Communication: Where Nature Meets Culture, 131–150. Hillsdale, NJ: Lawrence Erlbaum. Taylor, S. E., N. I. Eisenberger, D. Saxbe, B. J. Lehman, and M. D. Lieberman, 2006. Neural responses to emotional stimuli are associated with childhood family stress. Biol. Psychiatry 60: 296–301. Vuilleumier, P., M. P. Richardson, J. L. Armony, J. Driver, and R. J. Dolan, 2004. Distant influences of amygdala lesion on visual cortical activation during emotional face processing. Nature Neurosci. 7:1271–1278. Waters, A. M., O. V. Lipp, and S. H. Spence, 2005. The effects of affective picture stimuli on blink modulation in adults and children. Biol. Psychol. 68:257–281. Weiss, B., K. A. Dodge, J. E. Bates, and G. S. Pettit, 1992. Some consequences of early harsh discipline: Child aggression and a maladaptive social information processing style. Child Dev. 63: 1321–1335. Whalen, P. J., G. Bush, R. J. McNally, S. Wilhelm, S. C. McInerney, M. A. Jenike, et al., 1998. The emotional counting Stroop paradigm: A functional magnetic resonance imaging probe of the anterior cingulate affective division. Biol. Psychiatry 44:1219–1228. Widom, C. S., 1999. Posttraumatic stress disorder in abused and neglected children grown up. Am. J. Psychiatry 156:1223–1229. Wismer Fries, A. B., and S. D. Pollak, 2004. Emotion understanding in postinstitutionalized Eastern European children. Dev. Psychopathol. 16:355–369. Zahn-Waxler, C., B. Klimes-Dougan, and M. J. Slattery, 2000. Internalizing problems of childhood and adolescence: Prospects, pitfalls, and progress in understanding the development of anxiety and depression. Dev. Psychopathol. 12:443–466. Zeanah, C. H., 2000. Disturbances of attachment in young children adopted from institutions. J. Dev. Behav. Pediatr. 21: 230–236. Ziabreva, I., R. Schnabel, G. Poeggel, and K. Braun, 2003. Mother’s voice “buffers” separation-induced receptor changes in the prefrontal cortex of Octodon degus. Neuroscience 119:433– 441.

shackman, wismer fries, and pollak: environment and brain-behavioral development

881

54

Neurocognitive Development of Performance Monitoring and Decision Making EVELINE A. CRONE AND MAURITS W. VAN DER MOLEN

Cognitive developmental theories have provided important insights into developmental changes in self-regulation, defined here as the ability to monitor and regulate behavioral outcomes. This ability allows rapid changes in behavior to accommodate to environmental change, an ability that underlies age-related improvements in a broad range of intellectual and social behaviors. In this chapter we describe studies that were inspired by our desire to better understand the neural systems contributing to the development of two aspects of self-regulation, decision making and performance monitoring. This approach has the potential to move cognitive developmental theories toward incorporating the effects that neural system interactions have on reasoning, monitoring, and decision making. We argue that the prefrontal cortex is a key brain region contributing to developmental changes in decision making and performance monitoring. The frontal lobes comprise a substantial area of the human brain, and there is evidence that this region has reached its maximum size in humans compared to other organisms, therefore allowing for a greater complexity in intellectual abilities (Grafman, 1994). Moreover, this region of the brain is thought to be the latest to mature, reaching adult levels sometime during adolescence (Casey et al., 2005; Giedd, 2004; Paus, 2005; Sowell et al., 2004). Recently, many neuroimaging studies have investigated subprocesses that are relevant for self-regulation functions, such as reward processing, guessing, planning, inductive reasoning, and manipulating complex information in working memory (Klingberg, Forssberg, and Westerberg, 2002; Knutson et al., 2005; Ullsperger and von Cramon, 2006). These studies emphasized the importance of the prefrontal cortex in higher cognitive processing, but also pointed out that this region may be fractionated according to separate subprocesses and the neural connectivity of separable prefrontal regions to other brain regions. The orbitofrontal cortex, which involves the ventromedial prefrontal cortex and the lateral ventral brain region, is presumed to be involved in best-guess estimations and in emotional experiences associated with gains and losses (Breiter et al., 2001;

Elliott et al., 2003; Knutson et al., 2005; Rogers et al., 1999). Studies of humans with orbitofrontal brain injury and neuroimaging studies indicate that this region is highly relevant for processing many types of rewards and punishments. The dorsolateral prefrontal cortex, in contrast, appears to be relevant for manipulating information online, considering options, and updating performance outcomes (Holyoak and Kroger, 1995; E. Miller and Cohen, 2001). Both the orbitofrontal and dorsolateral regions of the prefrontal cortex have connections with the anterior cingulate cortex, a region involved in conflict monitoring and processing behavioral outcomes (Botvinick et al., 1999; Carter, Botvinick, and Cohen, 1999). The anterior cingulate cortex is thought to be involved in both cognitive and emotional aspects of control (Bush, Luu, and Posner, 2000), and also plays an important role in the regulation of bodily arousal states (Critchley, 2005). This fractionation should not be taken to suggest that separate regions carry out functions in isolation from other brain regions, but rather that different areas of the prefrontal cortex appear to be engaged by dedicated neural systems involved in separable cognitive functions. By definition, these neural systems function in such a way that the prefrontal cortex interacts with many other brain areas (Krawczyk, 2002). Understanding the differential functioning of these networks is important, since subprocesses of performance monitoring may rely on differentiated frontal regions that appear separable both regionally and functionally; these regions include the orbitofrontal cortex and the dorsolateral prefrontal cortex, and their functional connections with the anterior cingulate cortex. The anatomical segregation between these prefrontal cortex regions will serve as a reference point for the remainder of this chapter and will be particularly relevant to the description of developmental patterns in decision making and performance monitoring.

The orbitofrontal network and the development of decision making Decision making, or the ability to anticipate future outcomes of our choices, is required for a variety of behavior and often

883

involves consideration of multiple alternatives and reasoning about distant future consequences. In contrast to theories suggesting that emotions interfere with decision making, it is now understood that emotions may have an adaptive evolutionary human function and can help us to make advantageous decisions in real life (Damasio, 1994). Damasio’s (1994) somatic marker hypothesis was developed to explain specifically the behaviors of patients with damage to the orbitofrontal region. Such patients have relatively intact intellectual functions, but suffer from poor dailylife decision making. Emotions play a critical role within this theory and are defined as “somatic states,” referring to “the musculoskeletal, visceral, and internal milieu components of the soma” (Damasio, 1996, 1414). The somatic marker hypothesis suggests that the orbitofrontal cortex plays a critical role in forming associations between somatic responses associated with previously learned outcomes of situations and the reinstatement of these somatic states when a similar decision has to be made. These associations potentially reactivate emotions by acting on the appropriate cortical or subcortical structures and become highly relevant in situations where future outcomes cannot be easily predicted on the basis of logical cost-benefit comparisons. Given a certain situation, the orbitofrontal cortex establishes a simple linkage between the disposition for a certain aspect of the situation and the disposition for the type of emotion that in the past experience has been associated with the situation. The somatic markers, which are manifested in autonomic states such as heart-rate and skin-conductance differences, normally help constrain the decision-making space by making that space manageable for logic-based cost-benefit analyses (see Dunn, Dalgleish, and Lawrence, 2006, for a critical evaluation of the somatic marker theory). Empirical support for the operation of autonomic signals in decision making comes from studies using the Iowa Gambling Task, which mimics real-life decisions in the way it allocates rewards and punishments in the context of uncertain outcomes (Bechara et al., 1994). This task (typically computerized) requires individuals to sample from four decks of cards, two of which can result in immediate high gain whereas two others result in immediate low gain. The uncertainty in outcomes lies in the way delayed punishment

is presented. The two decks that result in high gain are accompanied by large delayed punishment, but one of the decks is accompanied by frequent, relatively small punishment (50%) whereas the other is accompanied by infrequent, relatively large punishment (10%). Similarly, of the decks that result in low immediate gain, one is accompanied by frequent but small delayed punishment (50%) and the other by infrequent but large delayed punishment (10%). The first two decks (A and B) are disadvantageous in the long run, because they result in a net loss, whereas the last two decks (C and D) are advantageous in the long run, because they result in net gain (table 54.1). Intact individuals typically learn to adopt an advantageous response strategy during the course of the task, and develop anticipatory skin-conductance responses preceding disadvantageous choices (Bechara et al., 1997, 1996; Tomb et al., 2002). Patients with damage to the orbitofrontal regions, however, keep selecting the disadvantageous decks, and somatic markers in the form of skin-conductance responses preceding disadvantageous options are absent. Thus the anticipatory skin-conductance arousal preceding disadvantageous choices on the Iowa Gambling Task may represent risk-related behavior that is associated with the magnitude of future loss. This anticipatory response is presumed to rely on orbitofrontal convergence zones, which are the connections between the regulating orbitofrontal cortex and the limbic regions including the amygdala (Bechara et al., 1996; Damasio, 1996). To examine the extent to which the development of autonomic warning signals is associated with learning across trials, we asked 100 healthy adults to perform a computerized version of the Iowa Gambling Task while heart rate and skin conductance were recorded. Based on their performance, we categorized these individuals into “good performers” (mostly advantageous choices), “intermediate performers” (slow learning of advantageous choices), and “bad performers” (mostly disadvantageous choices). All groups demonstrated increases in skin conductance and slowing of heart rate following loss relative to gain, and this autonomic response was directly related to the magnitude of loss. Importantly, these responses did not differentiate between good, intermediate, and bad performers. In contrast, only good performers developed skin-conductance

Table 54.1 Reward and punishment schedules in the Iowa Card Gambling Task Deck A Deck B Deck C

Deck D

Reward

+100

+100

+50

+50

Occasional punishment

−200, −250, −300 (50% of the trials) 1000 − 1250 = −250 Disadvantageous

−1250 (10% of the trials) 1000 − 1250 = −250 Disadvantageous

−25, −50, −75 (50% of the trials) 500 − 250 = +250 Advantageous

−250 (10% of the trials) 500 − 250 = +250 Advantageous

Net result per 10 trials

884

emotion⁄cognition interactions

amplifications and heart-rate slowing preceding disadvantageous decks. These results were taken to suggest that primary autonomic responses, which are directly observed following gain and loss, are present in all individuals. In contrast, secondary autonomic responses, which are defined as the autonomic responses that are acquired across the task, are related to learning across trials and may help to restrain the decision-making space (Crone, Somsen, et al., 2004). In a subsequent study, we examined whether children would share with orbitofrontal patients and worst-performing adults the inability to anticipate future outcomes of their behavior. This idea was inspired by the converging evidence of a protracted development of prefrontal cortex throughout childhood and adolescence suggesting an important parallel between brain maturation and cognitive development (Casey, Giedd, and Thomas, 2000; Luciana and Nelson, 1998; Sowell et al., 2004). Prior behavioral studies have shown that children become increasingly able to understand proportional reasoning and hypothetical thought evaluation, but mature levels are not reached until adolescence (Baird and Fugelsang, 2004; Boyer, 2006; Falk and Wilkening, 1998). In this study, participants aged 6–25 performed a developmentally appropriate analogue of the Iowa Gambling Task—the Hungry Donkey Task. The basic format of the card gambling task was retained, but card gambling was changed into a prosocial game inviting the player to assist a hungry donkey to win as many apples as possible. The change of card gambling into a prosocial game served the purpose of making the Iowa task more meaningful for children and of motivating their involvement—“You cannot let a hungry donkey down” (Falk and Wilkening, 1998). Figure 54.1 presents a display of the Hungry Donkey Task and performance findings that emerged from this task. A clear pattern can be seen of a developmental increase in the

ability to learn to adopt advantageous choices. Six-to-nineyear-old children and 10- to 12-year-old children had a strong bias toward disadvantageous choices, even though they switched decks immediately after receiving punishment, just like adults. Adolescents aged 13–15 years made more advantageous choices than younger children, but still made more disadvantageous choices than adults, suggesting that advantageous decision making does not reach adult levels until late adolescence (Crone and van der Molen, 2004). This finding was reinforced by an analysis of reaction times. The information provided by the reaction-time analysis indicated that for adults, advantageous choices were made faster than disadvantageous choices, whereas for children disadvantageous choices were made faster than advantageous choices. This pattern was interpreted in terms of the decisional-field theory proposed by Busemeyer and Townsend (1993). These authors suggest that decision making is biased by preference states derived from previous choices. This bias causes the more favorable choice to be selected more quickly. Finally, exit interviews showed that there was also a developmental change in reported preference. Whereas children younger than 12 mostly reported the disadvantageous choices as their preference, adolescents showed a shift toward a more advantageous choice preference. In contrast, adults reported this preference the most consistently. A further precaution was taken to examine the specificity of this developmental pattern. Two versions of the task were constructed in accord with the strategy adopted by Bechara and coworkers for differentiating between sensitivity to future consequences versus sensitivity to reward or insensitivity to punishment (Bechara, Tranel, and Damasio, 2000). In one version, reward was given on each trial and punishment was presented occasionally and unpredictably. In the other version, the reversed schedule was presented,

Figure 54.1 Stimulus display of the Hungry Donkey Task and performance on the standard and reversed gambling task of four age groups. With age, volunteers make more advantageous choices on both the standard and the reversed version of the task. (Figure adapted from E. A. Crone and M. W. Van der Molen, 2004,

Developmental changes in real life decision making: Performance on a gambling task previously shown to depend on the ventromedial prefrontal cortex. Dev. Neuropsychol. 25(3):251–279. Reprinted with permission.)

crone and van der molen: performance monitoring and decision making

885

with punishment occurring at each trial and reward given occasionally and unpredictably. Bechara, Tranel, and Damasio (2000) found that their orbitofrontal patients were more influenced by the immediate prospects, positive or negative, than by the future consequences. This finding provides strong support for an interpretation of the patients’ deficits in terms of inability to anticipate future consequences. Indeed, a similar pattern was seen for children’s performance. Using variants of the Iowa Gambling Task, Hooper and colleagues (2004) and Overman and colleagues (2004) showed that developmental changes in gambling performance still continue during late adolescence. A comparison of performance of 11- to 13-year-old children and 14- to 17-yearold children in the study by Hooper and colleagues revealed that both groups learned to make advantageous choices over the course of the task, but this learning curve was faster for older adolescents (although not yet at adult level, in comparison to other published studies). Importantly, some studies suggest gender differences in how adolescents and adults perform the task. In general, boys outperform girls, in such a way that they make more advantageous choices, whereas girls are influenced by the frequency with which the punishment is presented (Overman, 2004; Reavis and Overman, 2001). These findings indicate the importance of including gender as a predictor in developmental research. In addition to the Iowa Gambling Task, Hooper and colleagues (2004) asked all participants to complete a go-nogo task indexing response inhibition and a digit-span task indexing working memory. Developmental differences were observed for all tasks, but hierarchical regressions did not support a specific relationship between the development of inhibition, working memory, and gambling performance. Thus each task may tap into specific cognitive processes with separate underlying neural structures. Together, these studies suggest that young children share with orbitofrontal patients a “myopia for the future” and that the ability to make advantageous future choices continues to develop in adolescence. Although at first glance the choice pattern of children is quite similar to the performance of orbitofrontal patients, there are also important differences evident from an analysis of their actual choices. For both advantageous and disadvantageous choices, one option is always associated with frequent, smaller-magnitude punishment and one option is always associated with infrequent, higher-magnitude punishment. These modifications were added to make the task more complex and unpredictable (Bechara, Damasio, and Damasio, 2000). Although orbitofrontal cortex patients have no preference for one frequency over the other, we observed that all age groups in our studies had a bias toward choices that were associated with infrequent, high-magnitude punishment, relative to frequent, low-magnitude punishment

886

emotion⁄cognition interactions

(Dunn, Dalgleish, and Lawrence, 2006). In a separate experiment, we used a between-subjects design in which we asked children aged 7–12 and adults to perform two-choice versions of the task. In one version, the advantageous and disadvantageous decks were associated with frequent punishment (50% of the trials), and in the other version the advantageous and disadvantageous decks were associated with infrequent punishment (10% of the trials). Children had a preference for disadvantageous choices but only on the version where punishment was presented infrequently. In contrast, when punishment was presented frequently, even the youngest children learned to make advantageous choices over the course of the task (Crone, Bunge, et al., 2005). Indeed, in a study using this task with infants, Kerr and Zelazo (2004) found that 4-year-olds already learned to make advantageous choices when punishment was presented on 50 percent of the trials (Garon and Moore, 2004). Recently, we used mixture analysis to reexamine the data from the Crone and Van der Molen (2004) study, and we showed that children’s performance can be categorized according to the ability to use increasingly complex rules. First, children’s focus is mainly on the frequency of punishment, followed by a focus on magnitude of punishment, and eventually adults integrate these two rules to select the most advantageous choice (Huizenga, Crone, and Jansen, 2007). These results are consistent with Piaget and Inhelder’s (1974) conclusion that a full understanding of probability (including the notion that the actual situation may differ from long-term outcome) is not present until the developmental stage of formal operations. Thus the myopia for the future that is seen in patients with orbitofrontal cortex damage is mirrored in the developing child, suggesting that these developmental changes may indicate slow development of the brain circuitry that involves orbitofrontal cortex. The analysis of children’s performance on variants of the gambling task has also led to insights into the atypical development of decision-making competence. For example, recent theories of attention-deficit/hyperactivity disorder (ADHD) have discussed a specific motivational style characterized by delay aversion (Sonuga-Barke, 2002) or dopamine system differences (Li et al., 2006) as important processes in the development of ADHD. Extending this conceptualization, researchers have examined whether decision making as indexed with the Iowa Gambling Task is impaired in children with ADHD. In general, adolescents with ADHD made less optimal selections than controls by selecting more cards from the disadvantageous decks (specifically deck B) (Toplak, Jain, and Tannock, 2005). However, the differences were subtle, and more research is necessary to understand the relation between cognitive and emotional cognition deficits in ADHD. In a study including adolescents and university students, we examined the role of behavioral and cognitive disinhibi-

tion in decision-making performance (Crone, Vendel, and van der Molen, 2003). Behavioral inhibition referred to inhibition as an important aspect of executive control, and cognitive inhibition referred to inhibition as a trait dimension. Participants were selected on the basis of extreme scores on the disinhibition subscale of Zuckerman’s Sensation Seeking Scale, and the scores were either in the high or low range (Zuckerman, 1979). Participants with high cognitive-disinhibition scores performed less advantageously on the standard but not the reversed gambling task, suggesting a reward-oriented response style. This pattern of results was independent of developmental changes in decision-making performance for both the standard and the reversed task, which were observed even in late adolescence, consistent with Hooper and colleagues (2004). Behavioral disinhibition (as assessed with the Matching Familiar Figures Test), however, was not related to gambling performance (Kagan et al., 1964). The implication of these and the current results is that individual differences in behavioral inhibition are seemingly unrelated to the ability to anticipate future outcomes. A potential caveat of this study was that participants were selected on the basis of extreme cognitiveinhibition scores and not on the basis of behavioralinhibition scores; therefore, it is possible that a relation between behavioral inhibition and decision-making performance would be observed in more extreme samples (Franken and Muris, 2005). Also, the concept of inhibition is complex, and different inhibition indices generally have low correlations (Huizinga, Dolan, and van der Molen, 2006). Therefore, future research, involving different concepts of inhibitory ability, is needed to assess the robustness of these results. Recently, the notion of protracted orbitofrontal cortex development vis-à-vis decision making was confirmed in an elegant f MRI study including children, adolescents, and adults. Galvan and colleagues (2006) had participants within three age groups (7–11 years, 13–17 years, and 23–29 years) perform a delayed-response two-choice task, in which a cue indexed whether the response would be followed by a small, medium, or large reward. Prior f MRI studies in adults (Galvan et al., 2005) and animal studies (Winstanley et al., 2004) have shown that OFC is important for updating incentive values during delayed-response evaluations. Galvan and colleagues (2006) demonstrated that increases in reward magnitude resulted in increased activity in the nucleus accumbens and orbitofrontal cortex in all age groups. However, adolescents showed a larger increase in nucleus accumbens activation relative to children and adults (Ernst et al., 2005; May et al., 2004), whereas both children and adolescents showed more activity in orbitofrontal cortex relative to adults. These results were interpreted in terms of protracted maturational changes in top-down control systems relative to subcortical regions implicated in appetitive behaviors. Therefore, adolescents may be more driven by appeti-

tive systems than control systems, leading to suboptimal choices in decision-making tasks.

The dorsolateral PFC-ACC network and the development of performance monitoring Whereas the orbitofrontal sector of the prefrontal cortex is probably most directly involved in the representation of positive and negative states under conditions of uncertainty, the dorsolateral PFC is thought to be involved in the representation of the goal states toward which these positive and negative states are directed (Davidson and Irwin, 1999). The overarching function of the dorsolateral prefrontal cortex may be best described as “executive attention,” referring to the capability whereby memory representations are maintained in a highly active state in the presence of interference. These representations may reflect action plans, goal states, or task-relevant stimuli in the environment (cf. Kane and Engle, 2002), or remaining context information in an active state (cf. E. Miller and Cohen, 2001). Several classical neuropsychological models have indicated that the ability to monitor ongoing performance in a dynamically changing environment is an important aspect of goal-directed behavior (G. Miller, Galanter, and Pribram, 1960), especially when there is a need to adjust behavior after the detection of errors or changes in the environment (Norman and Shallice, 1986). More recent neuropsychological and neuroimaging studies have demonstrated that the dorsolateral prefrontal cortex is important for performance monitoring and updating and that the dorsolateral PFC works closely with the anterior cingulate cortex, which detects conflict after an erroneous response (Gehring and Knight, 2000; Kerns et al., 2004). Error Detection The detection of errors has been studied extensively in the adult literature using electrophysiological techniques. These studies have incorporated a focus on the error-related negativity (ERN), a scalp potential that is observed when participants make errors in choice-reactiontime tasks (Falkenstein et al., 1991; Scheffers et al., 1996). Source localization studies have localized this potential to the anterior cingulate cortex (Van Veen and Carter, 2002), and therefore this region is thought to be important for error monitoring (Holroyd and Coles, 2002). There have been few studies examining error processing in children using simple error-detection tasks in the context of event-related potential (ERP) recordings, but the studies that have been performed show that the ERN is minimal around 7–8 years of age and increases in magnitude during adolescence (Davies, Segalowitz, and Gavin, 2004; Ladouceur, Dahl, and Carter, 2004). This finding is illustrated in figure 54.2 showing that the amplitude of the ERN increases with age. Despite the slowly developing ERN,

crone and van der molen: performance monitoring and decision making

887

Figure 54.2 Increasing negative event-related potentials following performance errors with age. (Figure adapted from P. L. Davies, S. J. Segalowitz, and W. J. Gavin, 2004, Development of response-

monitoring ERPs in 7- to 25-year-olds. Dev. Neuropsychol. 25(3):355– 376. Reprinted with permission.)

young children already show evidence of error detection at a behavioral level, because they slow down following an erroneous response, similar to adults (Davies, Segalowitz, and Gavin, 2004; Jones, Rothbart, and Posner, 2003). Also, young children show evidence of a positive brain potential following the ERN, the so-called error positivity. This potential has been linked to the awareness of an error (Nieuwenhuis et al., 2001). Taken together, the monitoring system of errors/conflict, purportedly controlled by the ACC, indicates a slow developmental trajectory, as indexed by the smaller ERN response in children relative to

adults. In contrast, the error awareness system is developed at an earlier age, as indexed by an adultlike error-positivity potential and slowing of reaction time following an error. The interplay of the error-detection and error-awareness systems during the course of development remains to be investigated in future research.

888

emotion⁄cognition interactions

Feedback Processing A related aspect of performance monitoring is feedback processing. The research on feedback processing has benefited considerably from results obtained from the use of the classic Wisconsin Card Sorting Task

(WCST). The WCST is probably the most widely used neuropsychological task to examine feedback processing, or the external detection of an error. It has traditionally been used to characterize patients with damage to the dorsolateral prefrontal cortex (Barcelo and Knight, 2002), although the WCST impairments are also seen following damage to other brain regions (Anderson et al., 1991). In its original form, the WCST uses stimulus cards and response cards that display varying forms (crosses, circles, squares, and triangles), colors (red, green, blue, yellow), and numbers (one, two, three, four). Four cards with different combinations of the described characteristics are placed before the subject. The subject is handed a deck of response cards and instructed to place each consecutive card from that deck in front of one of the four stimulus cards, wherever the subject thinks it should go. The subject is informed only whether each response is right or wrong, and is not told the correct sorting principle. Once the subject has made a specified number of consecutive sorts according to the specific sorting principle, the criterion is changed without warning. The test proceeds through a specified number of shifts of the three sorting principles (Heaton et al., 1993). The most important requirements of this task are therefore (1) the ability to use feedback information to find the correct sorting rule, that is, efficient use of environmental cues, (2) the ability to maintain the correct sorting rule in working memory and monitor ongoing performance, and (3) the ability to inhibit responses to the previously correct sorting rule. Cardiac concomitants of feedback processing. In a series of experiments, we examined the cardiac changes that accompany feedback monitoring using several variants of the WCST. These experiments were inspired by an initial finding of Somsen and colleagues (2000), who had adolescents complete a computerized WCST. Somsen and colleagues (2000) observed that the shift toward cardiac acceleration that usually occurs at response onset (Somsen et al., 1985) was significantly delayed when participants received negative feedback relative to positive feedback. This cardiac slowing was larger when the feedback was unexpected (after a rule change) and for good relative to bad performers. Somsen and colleagues (2000) concluded that a monitoring/evaluation process must have triggered the cardiac response. The authors suggested that this heart-rate slowing following negative feedback bears a functional similarity to the ERN. Thus the heart rate deceleration occurring when individuals receive negative feedback might reflect a performance-monitoring mechanism that is responsible for immediate error correction and strategic adjustments that reduce the likelihood of errors in the future (Bernstein, Scheffers, and Coles, 1995). Possibly, the anterior cingulate cortex is also important for external feedback processing

(Holroyd et al., 2004), although this hypothesis is currently not universally accepted (Nieuwenhuis et al., 2005). To examine the extent to which this cardiac pattern is sensitive to the informative value of the feedback, we asked healthy adults to perform a reinforcement-learning task based on a study by Holroyd and Coles (2002). The experimental task required participants to press one of two buttons in response to a series of six stimuli. The participants were told to infer the stimulus-response mappings by trial and error, using information provided by positive or negative feedback, presented at the end of each trial. A critical aspect of the task was that the six stimuli differed in the degree to which the response was predictive of the value of the feedback. For two of these stimuli, participants could learn to control the value of feedback by acquiring the stimulus-response mapping. In this condition the feedback was 100 percent valid (the “100% condition”). For two other stimuli, feedback was unrelated to the selected response, providing 50 percent positive and 50 percent negative feedback (the “50% condition”). The last two stimuli provided always-positive feedback or alwaysnegative feedback, independent of the response (the “always condition”). The results that emerged from the probability-learning task are plotted in figure 54.3. Similarly to previous findings by Somsen and colleagues, heart rate slowed following an erroneous response in the 100 percent condition. More importantly, this slowing was absent in the “always” condition, showing that it is specifically related to monitoring of performance-related negative feedback and not to negative feedback in general. Indeed, in the 50 percent condition, in which feedback mappings changed from trial to trial, heartrate slowing was observed following both positive and negative feedback (Crone, van der Veen, et al., 2003). Thus, on the basis of a reinforcement-learning paradigm, we showed that heart-rate slowing is sensitive to violation of performance-based expectations. The monitoring hypothesis of cardiac slowing is consistent with substantial literature reporting phasic heart-rate changes in preparation for or following informative events (Jennings, 1992; Jennings, van der Molen, and Brock, 1997; Pribram and McGuinness, 1975) and phasic heart-rate slowing when a control system needs to detect the priority of actions (Kleiter and Schwarzenbacher, 1989). The cardiac concomitants of feedback processing and error processing extend the findings that emerged from ERP studies. The demonstration that feedback processing is manifested in the cardiac response is important in at least two respects. First, this finding makes a contribution to the cognitive psychophysiology literature on performance monitoring. Second, and equally important, the feedback-related cardiac response provides a window on the central regulation of the autonomic nervous system— a research domain that is growing rapidly because of the

crone and van der molen: performance monitoring and decision making

889

Figure 54.3 Performance and heart rate changes on the reinforcement learning task based on Holroyd and Coles (2002) of three age groups (Crone, Jennings, and van der Molen, 2004). Heart rate slows following negative feedback relative to positive feedback. For adults, this slowing is only seen for performancerelated negative feedback, whereas 8-year-old children also show

heart rate slowing to uninformative negative feedback. (Figure adapted from E. A. Crone, J. R. Jennings, and M. W. Van der Molen, 2004, Developmental change in feedback processing as reflected by phasic heart rate changes. Dev. Psychol. 40(6):1228– 1238. Reprinted with permission.)

availability of increasingly more powerful brain-imaging techniques (Critchley, 2005). These studies provide the context for examining developmental changes in performance monitoring.

the youngest children would be able to learn the 100 percent mapping. As expected, 8-year-old children and 12-year-old children learned this mapping slower than adults, but all participants improved over the course of a block. Consistent with the slower learning, 8-year-old children and 12-yearold children also showed less heart-rate slowing following an erroneous response relative to adults (see figure 54.3). Like adults, all children showed heart-rate slowing following feedback in the 50 percent mapping condition, demonstrating that all participants tried to interpret the feedback that followed performance choices. However, children also showed heart-rate slowing following negative feedback in the “always negative” condition, demonstrating that they failed to differentiate between relevant and irrelevant negative feedback (Crone, Jennings, and van der Molen, 2004). Subsequently, we reexamined the finding by Somsen and colleagues (2000) in a developmental study. Analysis of heart-rate changes following feedback in a WCST-analogue task revealed that children and adolescents showed similar heart-rate slowing following the presentation of an external signal (negative feedback) indicating that the previous sorting rule was no longer correct. When making a perseverative

Developmental change in feedback processing. Although there are numerous behavioral studies that have examined set shifting following positive and negative feedback across development (Luciana and Nelson, 1998; Zelazo, 2004), including the WCST (Somsen, in press; Welsh, Pennington, and Groisser, 1991), there are few studies that have examined to what extent set shifting is specifically related to performance monitoring. To examine how children monitor performancerelated feedback, we asked 8-year-old children, 12-year-old children, and young adults to perform the reinforcementlearning task developed by Holroyd and Coles (2002) while heart rate was sampled during the task. The task was slightly modified and presented in a between-subjects design, in which participants either performed the 100 percent mapping rule together with the always mapping rule or the 100 percent mapping rule together with the 50 percent mapping rule. This change was necessary to ensure that even

890

emotion⁄cognition interactions

160 18–20 years 14–15 years 9–11 years

140 120 100 80 60 40 20 0

First warning

Efficient negFB

Error negFB

First posFB

Correct posFB

Negative FB > Positive FB 18–20 yrs (n = 20)

14–15 yrs (n = 20)

9–11 yrs (n = 15)

RCI contrast value

2 1.5 1

18–20 years 14–15 years 9–11 years

0.5 0 –0.5 –1 First warning Efficient negFB Error negFB First posFB Correct posFB 2

RCI contrast value

error, adolescents’ heart rate significantly decreased, indicating that the error was evaluated for the purpose of adjusting future behavior. In contrast, young children did not show such heart-rate slowing, suggesting that the system that monitors the need for performance adjustments does not register a perseverative error as such (Crone, Somsen, et al., 2006). A tentative interpretation is that the failure to monitor these types of errors may account for the large number of perseverative errors seen in young children. Recently we started to examine the neural basis of performance monitoring using fMRI. Children aged 9–11 years, adolescents aged 14–15 years, and young adults were asked to perform a rule-switching task in which they had to switch between three possible location rules based on positive and negative feedback. The location rules were trained in advance, and participants were informed that rules could change unexpectedly while they performed the task. The location rules were chosen such that there were no ambiguous trials; therefore, sorts could only be correct according to one of the three rules. Based on a scoring method developed by Barcelo and Knight (2002), we distinguished between three types of negative feedback. The first-warning error was the first negative feedback that indicated that prior performance was no longer correct (similar to the rule changes in the WCST). The efficient negative feedback was a negative feedback that could be used to find the correct sorting rule. Following a first-warning error, participants had a 50 percent chance to find the new rule on the subsequent trial. In case they first chose the incorrect rule followed by the correct rule, then this negative feedback was labeled “efficient,” because it guided the participants toward finding the correct rule. Finally, the third type of negative feedback was related to performance errors. The spatial WCST task was previously used in a behavioral study (Crone, Ridderinkhof, et al., 2004), and in this study we demonstrated that adult levels of performance were not reached until adolescence. In the fMRI study, we showed that negative performance feedback was associated with increased activation in dorsolateral prefrontal cortex (DLPFC) and anterior cingulate cortex relative to positive feedback. In figure 54.4 (and plate 66) it can be seen that both regions were active to the same extent following performance errors. However, DLPFC was more active following efficient negative feedback, which is the feedback that indicates that the rule that was selected following first-warning feedback was incorrect, whereas the reversed pattern was seen for anterior cingulate cortex (Zanolie et al., 2007). These findings were interpreted in terms of a role for the anterior cingulate cortex in detection of conflict (errors and first warnings) and a role for the DLPFC in hypothesis testing (errors and efficient feedback). In addition, following negative feedback, children aged 9–11 recruited both DLPFC and anterior cingulate cortex differently than adults.

1.5 1 0.5 0 –0.5 –1 First warning Efficient negFB Error negFB First posFB Correct posFB

Figure 54.4 Neural activity associated with the processing of positive and negative performance feedback for children, adolescents, and adults. The pattern of activation in anterior cingulate cortex shows an adult pattern in adolescence, whereas the pattern of activation in DLPFC does not reach adult levels until late adolescence. (See plate 66.)

Adolescents aged 14–15 years showed neural activity in medial PFC similar to that of adults following the different types of negative feedback, but failed to recruit DLPFC to the same extent as adults following efficient negative feedback and performance errors (Zanolie et al., 2006). These results were interpreted as demonstrating that the

crone and van der molen: performance monitoring and decision making

891

monitoring of performance feedback continues to develop until late adolescence and that the late developmental progression is mostly associated with functional development of DLPFC, a region important for implementing cognitive control (E. Miller and Cohen, 2001). These findings are also consistent with our prior fMRI work demonstrating slow DLPFC development in other domains of cognitive control, for example, in the ability to manipulate information in working memory (Crone, Wendelken, et al., 2006).

Conclusions and future directions This chapter was concerned with the development of decision making and performance monitoring mediated by functionally and regionally differentiated prefrontal regions: the orbitofrontal, dorsolateral prefrontal, and anterior cingulate cortex. Current models of frontal lobe maturation suggest that functions that rely on the prefrontal cortex develop slowly (Diamond, 2002; Zelazo, 2004). This claim is supported by longitudinal studies showing that the frontal lobes are among the latest to mature in terms of gray matter volume (Gogtay et al., 2004; Shaw et al., 2006; Sowell et al., 2004), by the striking similarity between performance of children and deficits observed in adult patients with prefrontal damage (Dempster, 1993), and by recent fMRI studies showing a link between cognitive development and functional brain activity in prefrontal cortex (Casey et al., 2005). Generally, it has been hypothesized that prefrontal maturation is associated with developmental changes in the ability to self-regulate behavior in cognitive and affective domains (Posner and Rothbart, 2000). In this chapter we demonstrated that physiological measures such as skin conductance, heart rate, and event-related potentials may provide a solid basis for relating developmental changes in performance on tasks known to rely on different brain regions to actual brain maturation. Recent advances in studies relating heart rate to brain activation have shown that the anterior cingulate cortex represents arousal states that are measured through the autonomic nervous system (Critchley, 2005), and source-localization studies have demonstrated that the ERN potential following errors originates from the anterior cingulate (Van Veen and Carter, 2002). The developmental fMRI studies to date are largely complementary to the conclusions we draw from behavioral data (Casey et al., 2005). Directly measuring brain activation in children and adults is important for examining the relative contribution of separable brain regions to deficiencies in performance (Crone, Wendelken, et al., 2006). In future studies, it will be important to also examine within-person changes. A longitudinal study by Durston and colleagues (2006) showed that developmental changes in response inhibition within individuals were associated with additional recruitment of ventrolateral prefrontal

892

emotion⁄cognition interactions

cortex over time, a region known to be important for response inhibition in adults (Aron, Robbins, and Poldrack, 2004). These longitudinal findings are consistent with the argument that developmental changes in specific cognitive control functions are associated with increased recruitment of domain-specific regions in the prefrontal cortex. We suggest that developmental models of prefrontal maturation can be articulated further on the basis of the anatomical differentiation between subregions of the prefrontal cortex. Previous studies examining the development of prefrontal functioning typically used complex tasks (such as variants of the WCST) and used different behavioral indices (such as number of completed categories or perseverative errors) to describe developmental changes. One of the major virtues of decomposing functions of the prefrontal cortex is that this method may allow assessment of each function in terms of its anatomical basis and psychophysiological manifestation. The studies described in this chapter illustrate this approach. It is important to emphasize that the prefrontal cortex regions involved in performance monitoring and decision making should be interpreted as part of a network including posterior and subcortical regions (Fuster, 2001). Important connections for affective decision making include a network of regions including the orbitofrontal cortex, amygdala, and nucleus accumbens (Schoenbaum, Roesch, and Stalnaker, 2006), whereas error and feedback processing relies on a network including dorsolateral prefrontal cortex, anterior cingulate cortex, and basal ganglia (Ullsperger and von Cramon, 2006). The task for the future is to gain converging evidence from behavioral, neuroimaging, and connectivity analysis that may provide a working model for functional and regional prefrontal segregation and that may account for the developmental changes as well as individual differences in various aspects of self-regulation. For example, through the use of diffusion tensor imaging (DTI), it is possible to track the neural organization of white matter tracks across development (Barnea-Goraly et al., 2005; Mukherjee and McKinstry, 2006) and how this organization relates to the development of peripheral arousal systems (Roberts, 2006). One aspect of decision making that has been neglected in this chapter is the interplay between social and cognitive aspects of behavior in childhood and adolescence. It is only recently that researchers have started to integrate these previously separated lines of research. For example, Gardner and Steinberg (2005) demonstrated that risk taking in adolescence, relative to childhood and adulthood, is greatly exaggerated under the influence of social manipulations (peers present or absent). These changes are possibly associated with the concurrent development of the ability to make advantageous decisions and the ability to mentalize, or to take another person’s perspective. The mentalizing system is active when there is a need to think about other people’s

feelings, beliefs, and actions, and depends on medial prefrontal cortex, a region also referred to as the “social” brain (Adolphs, 2003). The slow emergence of the ability to make favorable long-term decisions in adolescence, together with the increasing influence of the environment, can lead to deviations of behavior in adolescence; therefore, adolescence can be a period of increased vulnerability to behavioral disruptions as a result of immature brain development (Raine, 2002). In future studies we should focus on the interplay between developmental, cognitive, and social neuroscience studies to examine how the prefrontal cortex contributes to behavioral changes that can be observed across development. REFERENCES Adolphs, R., 2003. Cognitive neuroscience of human social behaviour. Nature Rev. Neurosci. 4(3):165–178. Anderson, S. W., H. Damasio, R. D. Jones, and D. Tranel, 1991. Wisconsin Card Sorting Test performance as a measure of frontal lobe damage. J. Clin. Exp. Neuropsychol. 13:909–922. Aron, A. R., T. W. Robbins, and R. A. Poldrack, 2004. Inhibition and the right inferior frontal cortex. Trends Cogn. Sci. 8(4): 170–177. Baird, A. A., and J. A. Fugelsang, 2004. The emergence of consequential thought: Evidence from neuroscience. Philos. Trans. R. Soc. London [Biol.] 359(1451):1797–1804. Barcelo, F., and R. T. Knight, 2002. Both random and perseverative errors underlie WCST deficits in prefrontal patients. Neuropsychologia 40(3):349–356. Barnea-Goraly, N., V. Menon, M. Eckert, L. Tamm, R. Bammer, A. Karchemskiy, et al., 2005. White matter development during childhood and adolescence: A cross-sectional diffusion tensor imaging study. Cerebral Cortex 15(12):1848–1854. Bechara, A., A. R. Damasio, H. Damasio, and S. W. Anderson, 1994. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 50(1–3):7–15. Bechara, A., H. Damasio, and A. R. Damasio, 2000. Emotion, decision making and the orbitofrontal cortex. Cerebral Cortex 10(3):295–307. Bechara, A., H. Damasio, D. Tranel, and A. R. Damasio, 1997. Deciding advantageously before knowing the advantageous strategy. Science 275(5304):1293–1295. Bechara, A., D. Tranel, and H. Damasio, 2000. Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions. Brain 123(Part 11):2189–2202. Bechara, A., D. Tranel, H. Damasio, and A. R. Damasio, 1996. Failure to respond autonomically to anticipated future outcomes following damage to prefrontal cortex. Cerebral Cortex 6(2):215– 225. Bernstein, P. S., M. K. Scheffers, and M. G. Coles, 1995. “Where did I go wrong?” A psychophysiological analysis of error detection. J. Exp. Psychol. [Hum. Percept.] 21(6):1312–1322. Botvinick, M., L. E. Nystrom, K. Fissell, C. S. Carter, and J. D. Cohen, 1999. Conflict monitoring versus selection-foraction in anterior cingulate cortex. Nature 402(6758):179–181. Boyer, T. W., 2006. The development of risk-taking: A multiperspective review. Dev. Rev. 3:291–345. Breiter, H. C., I. Aharon, D. Kahneman, A. Dale, and P. Shizgal, 2001. Functional imaging of neural responses to expec-

tancy and experience of monetary gains and losses. Neuron 30(2):619–639. Busemeyer, J. R., and J. T. Townsend, 1993. Decision field theory: A dynamic-cognitive approach to decision making in an uncertain environment. Psychol. Rev. 100(3):432–459. Bush, G., P. Luu, and M. I. Posner, 2000. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn. Sci. 4(6): 215–222. Carter, C. S., M. M. Botvinick, and J. D. Cohen, 1999. The contribution of the anterior cingulate cortex to executive processes in cognition. Rev. Neurosci. 10(1):49–57. Casey, B. J., J. N. Giedd, and K. M. Thomas, 2000. Structural and functional brain development and its relation to cognitive development. Biol. Psychol. 54(1–3):241–257. Casey, B. J., N. Tottenham, C. Liston, and S. Durston, 2005. Imaging the developing brain: What have we learned about cognitive development? Trends Cogn. Sci. 9(3):104–110. Critchley, H. D., 2005. Neural mechanisms of autonomic, affective, and cognitive integration. J. Comp. Neurol. 493(1):154–166. Crone, E. A., S. A. Bunge, H. Latenstein, and M. W. Van der Molen, 2005. Characterization of children’s decision making: Sensitivity to punishment frequency, not task complexity. Child Neuropsychol. 11(3):245–263. Crone, E. A., J. R. Jennings, and M. W. Van der Molen, 2004. Developmental change in feedback processing as reflected by phasic heart rate changes. Dev. Psychol. 40(6):1228–1238. Crone, E. A., K. R. Ridderinkhof, M. Worm, R. J. Somsen, and M. W. Van der Molen, 2004. Switching between spatial stimulus-response mappings: A developmental study of cognitive flexibility. Dev. Sci. 7(4):443–455. Crone, E. A., R. J. Somsen, B. Van Beek, and M. W. Van der Molen, 2004. Heart rate and skin conductance analysis of antecedents and consequences of decision making. Psychophysiology 41(4):531–540. Crone, E. A., R. J. Somsen, K. Zanolie, and M. W. Van der Molen, 2006. A heart rate analysis of developmental change in feedback processing and rule shifting from childhood to early adulthood. J. Exp. Child Psychol. 95:99–116. Crone, E. A., and M. W. Van der Molen, 2004. Developmental changes in real life decision making: Performance on a gambling task previously shown to depend on the ventromedial prefrontal cortex. Dev. Neuropsychol. 25(3):251–279. Crone, E. A., F. M. van der Veen, M. W. Van der Molen, R. J. Somsen, B. Van Beek, and J. R. Jennings, 2003. Cardiac concomitants of feedback processing. Biol. Psychol. 64(1–2):143– 156. Crone, E. A., I. Vendel, and M. W. Van der Molen, 2003. Decision-making in disinhibited adolescents and adults: Insensitivity to future consequences or driven by immediate reward? Personality and Individual Differences 35:1625–1641. Crone, E. A., C. Wendelken, S. Donohue, L. van Leijenhorst, and S. A. Bunge, 2006. Neurocognitive development of the ability to manipulate information in working memory. Proc. Natl. Acad. Sci. USA 103(24):9315–9320. Damasio, A. R., 1994. Descartes’ Error. New York: Grosset/Putnam. Damasio, A. R., 1996. The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philos. Trans. R. Soc. London [Biol.] 351(1346):1413–1420. Davidson, R. J., and W. Irwin, 1999. The functional neuroanatomy of emotion and affective style. Trends Cogn. Sci. 3(1):11–21. Davies, P. L., S. J. Segalowitz, and W. J. Gavin, 2004. Development of response-monitoring ERPs in 7- to 25-year-olds. Dev. Neuropsychol. 25(3):355–376.

crone and van der molen: performance monitoring and decision making

893

Dempster, F. N., 1993. Resistance to Interference: Developmental Changes in a Basic Processing Mechanism. New York: Springer-Verlag. Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: Cognitive functions, anatomy and biochemistry. In K. Sa, ed., Principles of Frontal Lobe Function, 466–503. London: Oxford University Press. Dunn, B. D., T. Dalgleish, and A. D. Lawrence, 2006. The somatic marker hypothesis: A critical evaluation. Neurosci. Biobehav. Rev. 30(2):239–271. Durston, S., M. C. Davidson, N. Tottenham, A. Galvan, J. Spicer, J. A. Fossella, et al., 2006. A shift from diffuse to focal cortical activity with development. Dev. Sci. 9(1):1–8. Elliott, R., J. L. Newman, O. A. Longe, and J. F. Deakin, 2003. Differential response patterns in the striatum and orbitofrontal cortex to financial reward in humans: A parametric functional magnetic resonance imaging study. J. Neurosci. 23(1):303–307. Ernst, M., E. E. Nelson, S. Jazbec, E. B. McClure, C. S. Monk, E. Leibenluft, et al., 2005. Amygdala and nucleus accumbens in responses to receipt and omission of gains in adults and adolescents. NeuroImage 25(4):1279–1291. Falk, R., and F. Wilkening, 1998. Children’s construction of fair chances: Adjusting probabilities. Dev. Psychol. 34(6):1340–1357. Falkenstein, M., J. Hohnsbein, J. Hoormann, and L. Blanke, 1991. Effects of crossmodal divided attention on late ERP components. II. Error processing in choice reaction tasks. Electroencephalogr. Clin. Neurophysiol. 78(6):447–455. Franken, I. H. A., and P. Muris, 2005. Individual differences in decision-making. Personality and Individual Differences 39(5):991– 998. Fuster, J. M., 2001. The prefrontal cortex—An update: Time is of the essence. Neuron 30(2):319–333. Galvan, A., T. A. Hare, M. Davidson, J. Spicer, G. Glover, and B. J. Casey, 2005. The role of ventral frontostriatal circuitry in reward-based learning in humans. J. Neurosci. 25(38): 8650–8656. Galvan, A., T. A. Hare, C. E. Parra, J. Penn, H. Voss, G. Glover, et al., 2006. Earlier development of the accumbens relative to orbitofrontal cortex might underlie risk-taking behavior in adolescents. J. Neurosci. 26(25):6885–6892. Gardner, M., and L. Steinberg, 2005. Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Dev. Psychol. 41(4): 625–635. Garon, N., and C. Moore, 2004. Complex decision-making in early childhood. Brain Cogn. 55(1):158–170. Gehring, W. J., and R. T. Knight, 2000. Prefrontal-cingulate interactions in action monitoring. Nature Neurosci. 3(5):516–520. Giedd, J. N., 2004. Structural magnetic resonance imaging of the adolescent brain. Ann. NY Acad. Sci. 1021:77–85. Gogtay, N., J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. USA 101(21):8174–8179. Grafman, J., 1994. Alternative Frameworks for the Conceptualization of Prefrontal Lobe Functions. Amsterdam: Elsevier. Heaton, R. K., G. K. Chelune, J. L. Talley, G. G. Kay, and G. Curtiss, 1993. Wisconsin Card Sorting Test Manual: Revised and Expanded. Odessa: Psychological Assessment Resources. Holroyd, C. B., and M. G. Coles, 2002. The neural basis of human error processing: Reinforcement learning, dopamine, and the error-related negativity. Psychol. Rev. 109(4):679–709. Holroyd, C. B., S. Nieuwenhuis, N. Yeung, L. Nystrom, R. B. Mars, M. G. Coles, et al., 2004. Dorsal anterior cingulate

894

emotion⁄cognition interactions

cortex shows fMRI response to internal and external error signals. Nature Neurosci. 7:497–498. Holyoak, K. J., and J. K. Kroger, 1995. Forms of reasoning: Insight into prefrontal functions? Ann. NY Acad. Sci. 769:253– 263. Hooper, C. J., M. Luciana, H. M. Conklin, and R. S. Yarger, 2004. Adolescents’ performance on the Iowa Gambling Task: Implications for the development of decision making and ventromedial prefrontal cortex. Dev. Psychol. 40(6):1148– 1158. Huizenga, H., E. A. Crone, and B. Jansen, 2007. Decisionmaking in healthy children and adults explained by the use of increasingly complex proportional reasoning rules. Dev. Sci. 10(6):814–825. Huizinga, M., C. V. Dolan, and M. W. Van der Molen, 2006. Age-related change in executive function: Developmental trends and a latent variable analysis. Neuropsychologia 44(11):2017– 2036. Jennings, J. R., 1992. Is it important that the mind is in a body? Inhibition and the heart. Psychophysiology 29(4):369–383. Jennings, J. R., M. W. van der Molen, and K. Brock, 1997. Mnemonic search, but not arithmetic transformation, is associated with psychophysiological inhibition. J. Exp. Psychol. [Hum. Percept.] 23(1):154–167. Jones, L. B., M. K. Rothbart, and M. I. Posner, 2003. Development of executive attention in preschool children. Dev. Sci. 6(5):498–504. Kagan, J., B. L. Rosman, L. Day, J. Albert, and W. Philips, 1964. Information processing in the child: Significance of analytic and reflective attitudes. Psychol. Monogr. 78(1):578. Kane, M. J., and R. W. Engle, 2002. The role of prefrontal cortex in working-memory capacity, executive attention, and general fluid intelligence: An individual-differences perspective. Psychon. Bull. Rev. 9(4):637–671. Kerns, J. G., J. D. Cohen, A. W. MacDonald, 3rd, R. Y. Cho, V. A. Stenger, and C. S. Carter, 2004. Anterior cingulate conflict monitoring and adjustments in control. Science 303(5660): 1023–1026. Kerr, A., and P. D. Zelazo, 2004. Development of “hot” executive function: The children’s gambling task. Brain Cogn. 55(1): 148–157. Kleiter, G. D., and K. Schwarzenbacher, 1989. Beyond the answer: Post-error processes. Cognition 32(3):255–277. Klingberg, T., H. Forssberg, and H. Westerberg, 2002. Increased brain activity in frontal and parietal cortex underlies the development of visuospatial working memory capacity during childhood. J. Cogn. Neurosci. 14(1):1–10. Knutson, B., J. Taylor, M. Kaufman, R. Peterson, and G. Glover, 2005. Distributed neural representation of expected value. J. Neurosci. 25(19):4806–4812. Krawczyk, D. C., 2002. Contributions of the prefrontal cortex to the neural basis of human decision making. Neurosci. Biobehav. Rev. 26(6):631–664. Ladouceur, C. D., R. E. Dahl, and C. S. Carter, 2004. ERP correlates of action monitoring in adolescence. Ann. NY Acad. Sci. 1021:329–336. Li, D., P. C. Sham, M. J. Owen, and L. He, 2006. Meta-analysis shows significant association between dopamine system genes and attention deficit hyperactivity disorder (ADHD). Hum. Mol. Genet. 15(14):2276–2284. Luciana, M., and C. A. Nelson, 1998. The functional emergence of prefrontally-guided working memory systems in four- to eightyear-old children. Neuropsychologia 36(3):273–293.

May, J. C., M. R. Delgado, R. E. Dahl, V. A. Stenger, N. D. Ryan, J. A. Fiez, et al., 2004. Event-related functional magnetic resonance imaging of reward-related brain circuitry in children and adolescents. Biol. Psychiatry 55(4):359–366. Miller, E. K., and J. D. Cohen, 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24:167–202. Miller, G. A., E. Galanter, and K. H. Pribram, 1960. Plans and the Structure of Behavior. New York: Holt, Rinehart & Winston. Mukherjee, P., and R. C. McKinstry, 2006. Diffusion tensor imaging and tractography of human brain development. Neuroimaging Clin. N. Am. 16(1):19–43, vii. Nieuwenhuis, S., K. R. Ridderinkhof, J. Blom, G. P. Band, and A. Kok, 2001. Error-related brain potentials are differentially related to awareness of response errors: Evidence from an antisaccade task. Psychophysiology 38(5):752–760. Nieuwenhuis, S., H. A. Slagter, N. J. von Geusau, D. J. Heslenfeld, and C. B. Holroyd, 2005. Knowing good from bad: Differential activation of human cortical areas by positive and negative outcomes. Eur. J. Neurosci. 21:3161–3168. Norman, D. A., and T. Shallice, 1986. Attention to Action: Willed and Automatic Control of Behavior. New York: Plenum. Overman, W. H., 2004. Sex differences in early childhood, adolescence, and adulthood on cognitive tasks that rely on orbital prefrontal cortex. Brain Cogn. 55(1):134–147. Overman, W. H., K. Frassrand, S. Ansel, S. Trawalter, B. Bies, and A. Redmond, 2004. Performance on the Iowa card task by adolescents and adults. Neuropsychologia 42(13):1838– 1851. Paus, T., 2005. Mapping brain maturation and cognitive development during adolescence. Trends Cogn. Sci. 9(2):60–68. Piaget, J., and B. Inhelder, 1974. The Child’s Construction of Quantities. London: Routledge. Posner, M. I., and M. K. Rothbart, 2000. Developing mechanisms of self-regulation. Dev. Psychopathol. 12(3):427–441. Pribram, K. H., and D. McGuinness, 1975. Arousal, activation, and effort in the control of attention. Psychol. Rev. 82(2):116– 149. Raine, A., 2002. Biosocial studies of antisocial and violent behavior in children and adults: A review. J. Abnorm. Child Psychol. 30(4): 311–326. Reavis, R., and W. H. Overman, 2001. Adult sex differences on a decision-making task previously shown to depend on the orbital prefrontal cortex. Behav. Neurosci. 115(1):196–206. Roberts, A. C., 2006. Primate orbitofrontal cortex and adaptive behaviour. Trends Cogn. Sci. 10(2):83–90. Rogers, R. D., A. M. Owen, H. C. Middleton, E. J. Williams, J. D. Pickard, B. J. Sahakian, et al., 1999. Choosing between small, likely rewards and large, unlikely rewards activates inferior and orbital prefrontal cortex. J. Neurosci. 19(20):9029–9038. Scheffers, M. K., M. G. Coles, P. Bernstein, W. J. Gehring, and E. Donchin, 1996. Event-related brain potentials and errorrelated processing: An analysis of incorrect responses to go and no-go stimuli. Psychophysiology 33(1):42–53.

Schoenbaum, G., M. R. Roesch, and T. A. Stalnaker, 2006. Orbitofrontal cortex, decision-making and drug addiction. Trends Neurosci. 29(2):116–124. Shaw, P., D. Greenstein, J. Lerch, L. Clasen, R. Lenroot, N. Gogtay, et al., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440(7084):676– 679. Somsen, R. J., 2007. The development of attention regulation in the Wisconsin Card Sorting Test. Dev. Sci. 10(5):664–680. Somsen, R. J., M. W. van der Molen, J. R. Jennings, and J. F. Orlebeke, 1985. Response initiation, not completion, seems to alter cardiac cycle length. Psychophysiology 22(3):319–325. Somsen, R. J., M. W. van der Molen, J. R. Jennings, and B. van Beek, 2000. Wisconsin Card Sorting in adolescents: Analysis of performance, response times and heart rate. Acta Psychol. (Amst.) 104(2):227–257. Sonuga-Barke, E. J., 2002. Psychological heterogeneity in AD/ HD—A dual pathway model of behaviour and cognition. Behav. Brain Res. 130(1–2):29–36. Sowell, E. R., P. M. Thompson, C. M. Leonard, S. E. Welcome, E. Kan, and A. W. Toga, 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24(38):8223–8231. Tomb, I., M. Hauser, P. Deldin, and A. Caramazza, 2002. Do somatic markers mediate decisions on the gambling task? Nature Neurosci. 5(11):1103–1104; author reply, 1104. Toplak, M. E., U. Jain, and R. Tannock, 2005. Executive and motivational processes in adolescents with attention-deficithyperactivity disorder (ADHD). Behav. Brain Funct. 1(1):8. Ullsperger, M., and D. Y. von Cramon, 2006. The role of intact frontostriatal circuits in error processing. J. Cogn. Neurosci. 18(4): 651–664. Van Veen, V., and C. S. Carter, 2002. The timing of actionmonitoring processes in the anterior cingulate cortex. J. Cogn. Neurosci. 14(4):593–602. Welsh, M. C., B. F. Pennington, and D. B. Groisser, 1991. A normative-developmental study of executive function in children. Dev. Neuropsychol. 7:131–149. Winstanley, C. A., D. E. Theobald, R. N. Cardinal, and T. W. Robbins, 2004. Contrasting roles of basolateral amygdala and orbitofrontal cortex in impulsive choice. J. Neurosci. 24(20): 4718–4722. Zanolie, K., L. Van Leijenhorst, S. A. R. B. Rombouts, and E. A. Crone, 2007. Separable neural mechanisms contribute to feedback processing in a rule-learning task. Neuropsychologia 46(1):117–126. Zanolie, K., C. S. Van Meel, L. Van Leijenhorst, S. Rombouts, P. M. Westenberg, and E. A. Crone, 2006. Development of feedback processing. Presented at the Cognitive Neuroscience Meeting, San Francisco, April. Zelazo, P. D., 2004. The development of conscious control in childhood. Trends Cogn. Sci. 8(1):12–17. Zuckerman, M., 1979. Sensation Seeking: Beyond the Optimal Level of Arousal. Hillsdale, NJ: Erlbaum.

crone and van der molen: performance monitoring and decision making

895

Contributors Hajnalka Ábrahám, M.D. Central Electron Microscopic Laboratory, University of Pécs, Pécs, Hungary David G. Amaral, Ph.D. The MIND Institute, University of California, Davis, Sacramento, California Rosa M. Angulo-Barroso, Ph.D. Motor Development Laboratory, Division of Kinesiology, University of Michigan, Ann Arbor, Michigan Wendy Ark, B.S. Department of Cognitive Science, University of California, San Diego, La Jolla, California Richard N. Aslin, Ph.D. Department of Brain and Cognitive Sciences, University of Rochester, Rochester, New York Jocelyne Bachevalier, Ph.D. Department of Psychology, Emory University, Atlanta, Georgia Neil P. Bardhan, B.S. Department of Brain and Cognitive Sciences, University of Rochester, Rocheseter, New York Melissa D. Bauman, Ph.D. The MIND Institute, University of California, Davis, Sacramento, California Andrew D. Blackwell, Ph.D. Department of Psychiatry, University of Cambridge School of Clinical Medicine, Cambridge, United Kingdom Michael H. Bloch, M.D. Yale Child Study Center, New Haven, Connecticut Stephanie M. Carlson, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Christopher H. Chatham, B.A. Department of Psychology, University of Colorado, Boulder, Colorado Meghan A. Clayards, Ph.D. Department of Brain and Cognitive Sciences, University of Rochester, Rochester, New York Wendy Comeau, Ph.D. Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, Canada Eveline A. Crone, Ph.D. Department of Psychology, University of Leiden, Leiden, The Netherlands Gergely Csibra, Ph.D. Department of Psychology, Birbeck College, University of London, London, United Kingdom Ronald E. Dahl, M.D. Professor of Psychiatry, Pediatrics, and Psychology, University of Pittsburgh, Pittsburgh, Pennsylvania James L. Dannemiller, Ph.D. Department of Psychology, Rice University, Houston, Texas Michelle de Haan, Ph.D. Institute of Child Health, University College London, London, United Kingdom Kathryn DeRoche, M.A. University of Northern Colorado, Greeley, Colorado Mischa de Rover, Ph.D. Department of Experimental Psychology, University of Cambridge, Cambridge, United Kingdom

Guinevere F. Eden, Ph.D. Center for the Study of Learning, Georgetown University Medical Center, Washington, D.C. Ansgar Endress, Ph.D. International School for Advanced Studies, Cognitive Neuroscience, Trieste, Italy Linda Ewing-Cobbs, Ph.D. Children’s Learning Institute, University of Texas Health Science Center, Houston, Texas D. Lynn Flowers, Ph.D. Center for the Study of Learning, Georgetown University Medical Center, Washington, D.C. Erika E. Forbes, Ph.D. Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Nathan A. Fox, Ph.D. Child Development Lab, University of Maryland, College Park, Maryland Angela D. Friederici, Ph.D. Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany Susanna L. Fryer, M.S. Center for Behavioral Teratology, San Diego State University, San Diego, California Anita J. Fuglestad, B.A. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Michael K. Georgieff, M.D. Department of Pediatrics, University of Minnesota, Minneapolis, Minnesota Judit Gervain, Ph.D. International School for Advanced Studies, Cognitive Neuroscience, Trieste, Italy Robbin Gibb, Ph.D. Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, Canada David Gilliam, Ph.D. School of Psychological Sciences, University of Northern Colorado, Greeley, Colorado Ian M. Goodyer, M.D. Department of Psychiatry, Cambridge University, Cambridge, United Kingdom Elizabeth Gould, Ph.D. Department of Psychology, Princeton University, Princeton, New Jersey Tobias Grossmann, Ph.D. School of Psychology, Birkbeck College, University of London, London, United Kingdom Megan R. Gunnar, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Ahmad R. Hariri, Ph.D. Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Khader M. Hasan, Ph.D. Department of Diagnostic and Interventional Imaging, University of Texas Health Science Center, Houston, Texas Heather A. Henderson, Ph.D. Department of Psychology, University of Miami, Miami, Florida Claus C. Hilgetag, Ph.D. School of Engineering and Science, International University Bremen; Department of Health Sciences, Boston University, Boston, Massachusetts Myron A. Hofer, M.D. Department of Psychiatry, Columbia University, New York, New York

897

Bogdan F. Iliescu, M.D. Department of Psychology, Rice University, Houston, Texas Oskar G. Jenni, M.D. Child Development Center, Department of Pediatrics, University Children’s Hospital Zurich, Zurich, Switzerland Mark H. Johnson, Ph.D. Department of Psychology, Birbeck College, University of London, London, United Kingdom Miloš Judaš, M.D. Department of Neuroscience, Croatian Institute for Brain Research, University of Zagreb, Zagreb, Croatia Canan Karatekin, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Annette Karmiloff-Smith, Ph.D. School of Psychology, Birbeck College, University of London, London, United Kingdom Ann E. Kelley, Ph.D. Department of Psychiatry, University of Wisconsin, Madison, Madison, Wisconsin Amanda Kesek, M.A. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Maria Kharitonova, B.S. Department of Psychology, University of Colorado, Boulder, Colorado Torkel Klingberg, M.D., Ph.D. Karolinska Institutet Pediatric Neurology, Stockholm Brain Institute, Stockholm, Sweden Bryan Kolb, Ph.D. Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, Canada Ivica Kostovic´, M.D., Ph.D. Department of Neuroscience, Croatian Institute for Brain Research, School of Medicine, University of Zagreb, Zagreb, Croatia Yevgenia Kozorovitskiy, Ph.D. Department of Psychology, Princeton University, Princeton, New Jersey Elena Kushnerenko, Ph.D. School of Psychology, Birkbeck College, University of London, London, United Kingdom Zoë Kyte, Ph.D. Department of Psychiatry, Cambridge University, Cambridge, United Kingdom Charles F. Landry, Ph.D. Department of Psychiatry, University of Wisconsin, Madison, Madison, Wisconsin Eric M. Langlois, Ph.D. Yale Child Study Center, New Haven, Connecticut James F. Leckman, M.D. Yale Child Study Center, New Haven, Connecticut Paul Letourneau, Ph.D. Department of Neuroscience, University of Minnesota, Minneapolis, Minnesota Pat Levitt, Ph.D. Vanderbilt Kennedy Center for Research on Human Development, Vanderbilt University School of Medicine, Nashville, Tennessee Terri L. Lewis, Ph.D. Department of Psychology, Neuroscience and Behaviour, McMaster University, Toronto, Ontario, Canada Kelvin O. Lim, M.D. Department of Psychology, University of Minnesota, Minneapolis, Minnesota Monica Luciana, Ph.D. Center for Neurobehavioral Development, University of Minnesota, Minneapolis, Minnesota Denis Mareschal, Ph.D. Department of Psychology, Birbeck College, University of London, London, United Kingdom Rachel Marsh, Ph.D. Department of Psychiatry, Columbia University, New York, New York Sarah N. Mattson, Ph.D. Center for Behavioral Teratology, San Diego State University, San Diego, California

898

contributors

Daphne Maurer, Ph.D. Department of Psychology, Neuroscience and Behaviour, McMaster University, Hamilton, Ontario, Canada Linda C. Mayes, M.D. Yale Child Study Center, New Haven, Connecticut Christie L. McGee, M.S. Center for Behavioral Teratology, San Diego State University, San Diego, California Jacques Mehler, Ph.D. International School for Advanced Studies, Cognitive Neuroscience, Trieste, Italy Teresa V. Mitchell, Ph.D. Eunice Kennedy Shriver Center, University of Massachusetts Medical School, Waltham, Massachusetts Catherine J. Mondloch, Ph.D. Department of Psychology, Brock University, St. Catherines, Ontario, Canada Christopher S. Monk, Ph.D. Department of Psychology, University of Michigan, Ann Arbor, Michigan Sharon Morein-Zamir, Ph.D. Department of Experimental Psychology, University of Cambridge, Cambridge, United Kingdom Bryon A. Mueller, Ph.D. Department of Psychology, University of Minnesota, Minneapolis, Minnesota Yuko Munakata, Ph.D. Department of Psychology, University of Colorado, Boulder, Colorado Peter Mundy, Ph.D. Department of Psychology, University of California, Davis, Davis, California Ponnada A. Narayana, Ph.D. Department of Diagnostic and Interventional Imaging, University of Texas Medical School at Houston, Houston, Texas Charles A. Nelson III, Ph.D. Harvard Medical School, Children’s Hospital Boston, Boston, Massachusetts Marina Nespor, Ph.D. Department of Psychology, University of Milano, Bicocca, Italy Theresa A. Nick, Ph.D. Center for Neurobehavioral Development and Department of Neuroscience, University of Minnesota, Minneapolis, Minnesota Elizabeth D. O’Hare, A.B. Laboratory of Neuro Imaging, Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, California Elizabeth A. Olson, Ph.D. Center for Neurobehavioral Development, University of Minnesota, Minneapolis, Minnesota Sally Ozonoff, Ph.D. MIND Institute, University of California, Davis, Davis, California Brianna Paul, Ph.D. Clinical Psychology, San Diego State University, San Diego, California Koraly Pérez-Edgar, Ph.D. Department of Psychology, George Mason University, Fairfax, Virgina Jonas Persson, Ph.D. Department of Psychology, University of Michigan, Ann Arbor, Michigan Zdravko Petanjek, M.D. Department of Neuroscience, Croatian Institute for Brain Research, University of Zagreb, Zagreb, Croatia Bradley S. Peterson, M.D. Department of Psychiatry, Columbia University, New York, New York Daniel S. Pine, M.D. National Institute of Mental Health (NIMH) Intramural Research Program, Bethesda, Maryland Seth D. Pollak, Ph.D. Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin Mary R. Prasad, Ph.D. Children’s Learning Institute, University of Texas Health Science Center, Houston, Texas

Karina Quevedo, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Raghavendra Rao, M.D. Department of Pediatrics, University of Minnesota, Minneapolis, Minnesota Patricia A. Reuter-Lorenz, Ph.D. Department of Psychology, University of Michigan, Ann Arbor, Michigan John E. Richards, Ph.D. Department of Psychology, University of South Carolina, Columbia, South Carolina Jenny Richmond, Ph.D. School of Psychology, University of New South Wales, Sydney, Australia Edward P. Riley, Ph.D. Center for Behavioral Teratology, San Diego State University, San Diego, California Barbara J. Sahakian, Ph.D. Department of Experimental Psychology, University of Cambridge, Cambridge, United Kingdom Balasrinivasa Rao Sajja, Ph.D. Department of Radiology, University of Nebraska Medical Center, Omaha, Nebraska Terri L. Schochet, Ph.D. Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin Robert T. Schultz, Ph.D. Center for Autism, Children’s Hospital; Department of Pediatrics, University of Pennsylvania, Philadelphia, Pennsylvania László Seress, M.D., Ph.D. Central Electron Microscopic Laboratory, University of Pécs, Pécs, Hungary Jessica E. Shackman, M.S. Department of Psychiatry, University of Wisconsin, Madison, Madison, Wisconsin Mohinish Shukla, Ph.D. International School for Advanced Studies, Cognitive Neuroscience, Trieste, Italy Mikle South, Ph.D. Department of Psychology, Brigham Young University, Provo, Utah Elizabeth R. Sowell, Ph.D. Laboratory of Neuro Imaging, Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, California Gregg D. Stanwood, Ph.D. Vanderbilt Kennedy Center for Research on Human Development, Vanderbilt University School of Medicine, Nashville, Tennessee Jennifer Merva Stedron, Ph.D. Department of Psychology, University of Colorado, Boulder, Colorado

Joan Stiles, Ph.D. Department of Cognitive Science, University of California, San Diego, La Jolla, California Regina M. Sullivan, Ph.D. Department of Zoology, University of Oklahoma, Norman, Oklahoma Amanda R. Tarullo, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Kathleen M. Thomas, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Chad W. Tiernan, M.S. Motor Development Laboratory, Division of Kinesiology, University of Michigan, Ann Arbor, Michigan Angela Tseng, B.A. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota Maurits W. van der Molen, Ph.D. Department of Psychology, University of Amsterdam, Amsterdam, The Netherlands Amy Van Hecke, Ph.D. Department of Psychiatry, University of Illinois, Chicago, Illinois Essi Viding, Ph.D. Department of Psychology, University College London, London, United Kingdom Marilyn Welsh, Ph.D. School of Psychological Sciences, University of Northern Colorado, Greeley, Colorado Lauren K. White, M.S. University of Maryland, College Park, Maryland Tonya White, M.D. Division of Child and Adolescent Psychiatry, Youth Psychosis Clinic, University of Minnesota, Minneapolis, Minnesota Douglas E. Williamson, Ph.D. Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Alison B. Wismer Fries, M.S. Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin Jeffrey R. Wozniak, Ph.D. Department of Psychology, University of Minnesota, Minneapolis, Minnesota Yanki Yazgan, M.D. Department of Child Psychiatry, Marmara University, Istanbul, Turkey Philip David Zelazo, Ph.D. Institute of Child Development, University of Minnesota, Minneapolis, Minnesota

contributors

899

INDEX

A ABB grammar vs. ABC grammar, in young infants, 328 ABPs (actin-binding proteins), in regulation of actin filament organization, 7–8 Absolute thresholds auditory, 98–99 visual, 128 Abuse and attachment, 790 child, 869–870. See also Maltreatment, childhood cocaine. See also Cocaine, prenatal exposure to in adults, 654 Acetylcholine, and recovery from early brain injury, 394 Acoustic environment, in auditory cortex organization, 102–103 Acoustic parameters, nonphonemic, 108 Acoustic startle reflex, 841 Acquired prosopagnosia, childhood onset of, face processing impairment and, 516 ACTH (adrenocorticotropic hormone), secretion of, 63, 65 Actin filament(s) adhesive contacts of, 10 microtubule interactions with, 9 in neurons, 6–7 Actin filament organization in cells, 7 regulation of actin-binding proteins in, 7–8 and dynamics of cytoplasmic signaling pathways, 8–9 Actin-binding proteins (ABPs), in regulation of actin filament organization, 7–8 Action approach, to motor development, 147–148 Activation functions mathematical expression of, 369 in neural network models, 369–370 Active fixation tasks, in eye tracking, 264 Active memory representations, vs. latent memory traces, 555 Activity, level of, in motor development, 149–151 Acuity, visual. See Visual acuity ADC (apparent diffusion coefficient), 301 in newborns, 304 ADHD. See Attention-deficit/hyperactivity disorder (ADHD) Adhesion receptors, of actin filaments, 10 Adolescent(s) acute effects of nicotine on, 861–862 behavioral traits of, 855

brain development in, 855–863 acute effects of nicotine and, 861–862 fMRI of, 856 mechanism of nicotine action and, 859–860 microarray forebrain analysis in, 857, 859 morphological changes during, 855–856 MRS imaging of, 341, 343 neurological changes during, 856–857 smoking and influence of nicotine associated with, 859–862 brain maturation in, 772–773 eye tracking in atypical development and, 276–288. See also specific disorder normative development saccades and, 269–276 frontal circuitry organization in, 231 generalized anxiety disorder in, attention avoidance and, 764 LHPA regulation in, 73–74 prefrontal cortex in behavioral inhibition and, 585–586 development of, 575–587 structural, 577–578 dorsolateral, 575–576 development of, 582–584, 586 functions of, 578–581 in different prefrontal areas, 582–586 planning ability and, 582–583 process vs. content of, 581–582 recognition memory and, 581 selected tasks and, 583–584 self-ordered search tasks and, 582 self-organized behavior and, 583 span tasks and, 581–582 spatial delayed-response tasks and, 582 structure and connectivity of, 575–577 ventrolateral, 575–576 ventromedial, 576 development of, 584–585, 586 prenatally cocaine-exposed. See also Cocaine, prenatal exposure to outcomes for, 663–666 smoking among, 859–862 acute effects of nicotine and, 861–862 mechanism of nicotine action and, 859–860 vulnerability of, 855 Adrenal steroids, regulating adult neurogenesis, 53 Adrenalectomy, 53 Adrenocorticotropic hormone (ACTH), secretion of, 63, 65 Adult(s) cocaine-abusing, 654. See also Cocaine, prenatal exposure to

face processing in, 509–511 core cortical system of, 509–511 fusiform face area of, 510–511 superior temporal sulcus/gyrus of, 511 extended cortical-subcortical system of, 511 subcortical system of, 509 localization of auditory system in, recalibration of, 101–102 neurogenesis in, 51–59. See also Neurogenesis, adult recognition-memory abilities in, emergence of, 501–502 relational-memory abilities in, emergence of, 502–503 Adversity, resilience resulting from, 780 Affective disorders, genetic influence on, 353 Affective neuroscience, and pediatric anxiety, 763–766 Affective policy domains, relevance of sleep to, 813 Afferent pathways, in prefrontal cortex corticocortical connections in, 224 extrathalamic, 214–215, 222 long associative connections in, 224– 225 sequential development of, 221–225 thalamocortical, 215, 222–224 Affiliation, in autism, 709 Age at cochlear implantation, performance affected by, 447 at early brain injury, functional outcome associated with, 388–389 in language acquisition, 374–375 at traumatic brain injury, impact of, 406–407 Aggression in abused and neglected children, 869, 870 sleep deprivation and, 813–814 Aging and cognition dementia and, 607–616. See also Alzheimer’s disease; Dementia reserve, underactivation, and compensation in, 600–601 typical development in, 591–601 differential effects of, on retrieval processing, 600 executive functions and, 595–597 in childhood, 556–562 and memory long-term, 597–600 working, 592–595 structural brain changes associated with, 591–592

901

Alcohol use disorders, development of, affective pathway for, 361 Alzheimer’s disease diagnosis of CANTAB-PAL scores in, 610 criteria for, 607 early-stage, cognition in, 608–611 memory in, neuropsychological tests of, 609–611 mild cognitive impairment in, 608–609 vs. dementia with Lewy bodies, 613–614 vs. frontotemporal dementia, 612 vs. vascular dementia, 612–613 Alzheimer’s Disease Assessment Scale– Cognitive Scale (ADAS-cog), 610 Amacrine–ganglion cell interaction, in retina, 130 American Sign Language, 444–446 visuospatial nature of, 444–445 vs. spoken language, 444 Ammon’s horn, in hippocampal development, 187, 188, 189 cytoarchitectonics of, 191, 207 Amnesia, developmental, pattern of impairment in, 548–549 Amygdala and attachment, 792–793 corticosterone action on, 794–795 cytoarchitectonic maturation of, 222 functional studies of, 165–167 innervation of, 5-HT receptors in, 356 lesion studies of, 167 neuroanatomy of, 165 prefrontal cortex connection to, 215 reactivity of, 5-HTTLPR S allele affecting, 356–358 role of in autism, 709–711 in object-emotion associations, 871 in social behavior, 165–168 Amygdala-frontal connections, in autism, 710 Amygdala-fusiform connections, in autism, 710 Anesthesia, bird’s own song memory during, 459–460 Animal models of autism, 706 LHPA system in, 70–71 nonprimate, memory development in, 499–504. See also Memory plasticity of speech in, 453–461. See also Birdsong of social development, 173–175 Anterior cingulate cortex emotion processing in, 871 functional studies of, 169 lesion studies of, 169 neuroanatomy of, 168–169 performance monitoring in, 558–560, 887. See also Performance monitoring role of, in social behavior, 168–169 Anterior forebrain pathway in neural song system, 458 in song plasticity, 460

902

index

Antisaccade response times, age-related changes in, 269, 272 Antisaccade tasks in eye tracking, 264 warning stimulus on, age-related changes in CNV to, 269 Antisocial behavior, development of, 360–361 Anxiety. See also Fear entries chronic. See Posttraumatic stress disorder (PTSD) definition of, 755 in maltreated children, 869 normal vs. pathological, 755–756 “stranger,” 789 Anxiety disorder(s) clinical integrative framework of, 758–759 model of, 758–763 DSM-IV definitions of, 755 generalized, 755 genetic influence on, 353, 844 major advances in, 756–758 pediatric, 756 affective neuroscience and, 763–766 clinical classification of, 763–764 clinical questions concerning, 758 information-processing and fear-circuit functions in, 764–766 separation, 755 social, 755 Anxious behaviors, 841 Aortic stenosis, supravalvular, in Williams syndrome, 691, 692 Apparent diffusion coefficient (ADC), 301 in newborns, 304 Appraisal theory, of depression, 775 Approach-avoidance decisions, of executive function in childhood, 557 arc gene, basal levels of, in adolescent vs. adult brain, 862, 863 Arginine vasopressin, secretion of, LHPA stress response and, 63 Arousal attention system, 479–480 in prenatally cocaine-exposed children, 664 Asperger syndrome. See also Autism spectrum disorders (ASDs) DMS-IV definition of, 701 Association genetic studies of affective disorders, 353 involved in emotion regulation, 352–353 limitations of, 353–354 of personality, 353 traditional, 352 Atkinson’s model, of visual development, 138 Attachment abuse and, 790 amygdala involved in, 792–793 attenuated aversion learning and, neural basis of, 792–795 behavioral system of, 789 emotion in, emergence of, 795–797 and implications for human development, 798–799

and infant development, 797–798 and maternal behavior in next generation, 797–798 and maternal separation responses, 795–797 mother-infant interaction patterns of, 797–798 neurobiology of, 787–799 in newborns, 789 olfactory preference learning and, neural basis of, 790–792 perinatal transition in, 788–789 postnatal learning and, 789–790 prenatal origins of, 788 specific, initial formation of, 787–795 vocal communication in, origins of, 795–797 Attention in adolescence generalized anxiety disorder and, 764 prefrontal development and, 581–582 anterior system of, joint attention initiation and, 828–829 as coping resource, 847 effect of, on cognitive processes, 490 familiar stimulus and, 489–490 impaired depression and, 779–780 fetal alcohol syndrome and, 647 joint, development of, 819–833. See also Joint attention negative central (Nc) component in, 490–491 novel stimulus and, 489–490 posterior system of, 481–482 gaze following/joint attention and, 827–828 in preterm children, 402 pupillary dilation studies of, in children and adolescents, 276 selective, 870 in executive function in childhood, 561–562 shifting, brain activation patterns associated with, 526 spatial, development of, 525–527 specific systems of, 480–482 sustained, 482 phases of, 482–483 role of, in behavior, 487 visual-spatial after treatment of congenital cataracts, 425 impaired in fetal alcohol syndrome, 647 in schizophrenia, 280 in young infants arousal system of, 479–480 brain systems involved in, 479–482 briefly presented stimuli and, recognition of, 489–494 covert (orienting), 482 definition of, 479 developmental psychological perspective on, 479–495 direct measure of, 484

Attention (continued) event-related potentials and, scalprecorded, 483–484 eye movements and, 484–489. See also Eye movements heart rate in, 482–483, 489 indirect measure of, 484 psychophysiological measures of, 482–484 Attention Network Task (ANT), 845–846 Attention-deficit/hyperactivity disorder (ADHD) clinical depression and, 774 in cocaine-exposed children, 89 decision making in, 886 dyslexia with, 745 eye tracking studies in, 284, 285–288 review of, 289–291 in fetal alcohol syndrome, 647 MRS imaging of, 345–346 onset of, after traumatic brain injury, 405 overactivity in, 156 risk of, preterm birth and, 402 treatment effects on, using fMRI studies, 319–320 Attenuated aversion learning, neural basis of, 792–795 Attribution theory, of depression, 775 Audition, interactions between vision and, after cochlear implantation, 446–447 Auditory feedback, in song system, study of, 456–457 Auditory perception, in autism, 709 Auditory stimuli, ERP response to in cognitive development, 249–252 components related to lexical and syntactic processing in, 252 mismatch responses in, 251–252 Auditory system fundamental capacity of absolute thresholds in, 98–99 in infants, 97–100 intensity, frequency, and duration discrimination in, 99–100 onset of hearing and, 97–98 lexicon in, reference and learning of, 107–109 localization of in infants, 101 in nonhumans, 100–101 recalibration of, in adults, 101–102 neural specializations in, 100–104 phonetic reorganization of, 104–107 distributional modifications in, 105 distributional sensitivity in, 105–106 second language and, 106–107 universal inventories in, 104 plasticity of brain correlates of, 109–111 in humans, 103–104 in nonhumans, 102–103 reorganization of mechanisms of, 97–111 phonetic, 104–107 visual system and, 100

Auditory/visual speech fusion, sensitive period for, in young recipients of cochlear implants, 444–448 Autism affiliation in, 709 amygdala in, 709 role of, 710 amygdala-frontal connections in, 709–711 amygdala-fusiform connections in, 709 animal models of, 706 auditory perception in, 709 brain development in, 703–706 complex information processing in, 705–706 executive function in, 704–705 head/brain growth and, 703–704 neurochemistry in, 704 prenatal course of, 703 theory of mind in, 705 weak central coherence in, 705 brain imaging studies of, 344–345, 703– 704, 708–709 cognitive models in, 704–706 developmental course of, 701 diagnosis of, 701 DSM-IV criteria for, 165 etiology of, 702–706 facial expression processing in, 708 facial identity processing in, 706–708 gaze in abnormal, 281 eye tracking studies of, 282–284, 708 genetic basis of, 702–703 gyrification index in, 46 joint-attention impairments in, 822–823 mirror neurons in, 708–709 neurocognitive development in, 701–712 prevalence of, 702 repetitive behavior in, 711 and social brain, 706–710 social deficits in, 165 social motivation in, 709–710 social perception in, 706–710 Autism spectrum disorders (ASDs), 701 executive function in, 704 facial expression processing deficits in, 708 facial identity processing deficits in, 706–707 fMRI abnormalities observed in, 707 fusiform face area in, 707–708, 710 mode of transmission of, 702 serotonin levels in, 704 weak central coherence tasks in, 705 Autopsy studies, in fetal alcohol syndrome, 643–644 Axial diffusivity, in DTI data processing, 302 Axon(s) corticospinal, navigation of, 16 corticothalamic navigation of, 15–16, 17 prefrontal, 215 differentiation of, 13 distinguishing microtubule features in, 6 formation of, developing neurons in, 5–20 initial growth of, neuronal polarization and, 13–14

injury to, 403–404. See also Brain injury myelination of, in central nervous system, 544 targets of patterning distribution in, 17–18 stereotypical routes to, 14–15 Axonal-tension concept, of brain development, 42 Axo-somatic inhibitory cell(s), parvalbumincontaining, 203

B Backpropagation learning algorithm, 370 Basal forebrain–prefrontal cholinergic system, of prefrontal cortex, 215 Basal ganglia in fetal alcohol syndrome, neuroimaging studies of, 645 involved in song learning and production, 454–455 Basilar membrane, in auditory development, 102 BclI polymorphism, and individual LHPA function, 76 BDNF. See Brain-derived neurotropic factor (BDNF) Behavior(s) adaptive and maladaptive patterns of, in maltreated children, 876–877 antisocial, development of, 360–361 anxious, 841. See also Anxiety entries habitual. See also Obsessive-compulsive disorder (OCD); Tourette’s syndrome disturbances in self-regulatory control of, 717–731 in prenatally cocaine-exposed children, 665 repetitive, in autism, 711 self-organized, 583 social. See also Social behavior neurobiology of, 161–172 of songbirds, 453. See also Birdsong understanding, from neural network models, 367–368 Behavior genetics, 351–354 association studies in affective disorders and, 353 involving emotion regulation, 352–353 limitations of, 353–354 personality and, 353 traditional, 352 study of genes in, 351–352 Behavioral attachment system, 789. See also Attachment Behavioral development, fMRI studies of, 318–319 Behavioral evidence, of dorsal and ventral dissociation, in infants, 471–474 Behavioral inhibition, 74–75, 585–586 in temperament, 841–843 in unipolar depression, 778 Behavioral paradigms, in fMRI research, 315–316 Behavioral profile, in dyslexia, 740–741

index

903

Behavioral sequelae, after early brain injury, 387–389 measurement of, 389 Behavioral studies, of neonates’ perception, attention, and learning abilities, 326–328 Behavioral therapy, for early brain injury, 391–393 Behavioral traits, in adolescence, 855 Biases in learning mechanisms, 373 in unipolar depression, 778–779 Bilingualism auditory development in, 106–107 brain responses in, 110–111 facilitating executive function in childhood, 565 Biocular deprivation congenital cataracts and, 417, 420–421. See also Cataract(s), congenital restrictions after, 418 Bird’s own song (BOS) memory during anesthesia and sleep, 459–460 response to, 456 vs. tutor song memory, 457 Birdsong brain areas involved in, 454–456 features of bird behavior in, 453 future research into, 461 as model of speech, 453 neural song system in, 453–458 anterior forebrain pathway of, 458 motor loop of, 457 sensorimotor integration loop in, 456–457 sensorimotor system in, 458–461 behavior and role of sleep in, 458 development of synapses in, 459 and implications for human speech, 460 neuromodulators in, 458–459 sensory development in, 459–460 vocal development in, genetic aspects of, 459 vocal learning/production of brain-behavior relationship in, 453–454 nuclei involved in, 454–455 Birth LHPA system at, 71 preterm. See Preterm birth Birth weight, very low, neuropsychological outcome associated with, 402 Blind person auditory cortical plasticity in, 103 visual cortical plasticity in, 427 Blood Phe levels, in phenylketonuria, 678, 682 Blood-oxygenation-level-dependent (BOLD) imaging, 313 Brain. See also specific part activation patterns of, with shifting attention, 526 adult, neurogenesis and, 51–59. See also Neurogenesis, adult areas of, involved in birdsong, 454–456 arousal system of, 479–480

904

index

changes in age-related structural, 591–592 motor development and, 151–155 postnatal, gyrification and, 43–44 functional abnormalities of, in dementia, 614–615 increased growth of, autism and, 703–704 influence of corticotropin-releasing hormone on, 68–69 influence of cortisol on, 67–68 language processing in, 109–111. See also Language acquisition lexical processing in, 119–122 phonotactic knowledge and, 121 stress patterns of words and, 119–121 word familiarity and, 121 locations of, involved in attention, 479–482 maturation of, during childhood and adolescence, 772–773 measurement of activity of, in early development, 117–118 metabolism of in fetal alcohol syndrome, 645 role of sleep in, 809 normal, changes in, 385–387 postnatal changes in, gyrification and, 43–44 prenatal and postnatal, dendritogenesis in, 19 prosodic processing in, 118–119 semantic processing in sentences in, 122–123 words in, 121–122 social. See Social brain structural abnormalities of in dementia, 614–615 in Williams syndrome, 691 structural changes in, age-related, 591–592 syntactic processing in, 123–124 Brain development, 23–36 in adolescence, 855–863. See also Adolescent(s), brain development in after early brain injury, 389–391 in autism, 703–706 complex information processing in, 705–706 executive function in, 704–705 head/brain growth and, 703–704 neurochemistry in, 704 prenatal course of, 703 theory of mind in, 705 weak central coherence in, 705 basics in, 83 choline in, 627, 634–635 differential effects of sleep on, 810 dynamic and regressive changes in, agerelated, 23–24 effect of selected nutrients on, 625–635 experience-dependent changes in, 386–387 fMRI studies of, 317–318 gray matter changes in, cognitive correlates of, 33–36

gray matter density in, 26–30 gray matter loss in, 25, 26 gray matter thickness in, 30–33 growth factors in, role of, 624–625 gyrification in disorders of, clinical conditions associated with, 44–46 heritability of, 43 mechanical folding hypothesis of, 42–43 phylogeny in, 41 postnatal morphology changes in, 43–44 theories of, 41–44 impact of prenatal cocaine exposure on, 88–91, 653–666. See also Cocaine, prenatal exposure to in infancy, declarative memory performance and, 543. See also Declarative memory, in infancy influence of corticotropin-releasing hormone on, 70 influence of cortisol on, 69–70 introduction to, 23 iodine in, 626, 633–634 iron in, 626, 630–632 measurement of activity in, 117–118 monoamines in balance of receptor signaling and, 91–92 effects of, 83–92 modulatory influence of, 88–91 neuropharmacology of, 85–88 MRS imaging of, 337–347 in children and adolescents, 341, 343 developmental disorder profiles in, 343–346 in fetuses, 339–340 limitations of, 346–347 in neonates, 340–341, 342, 343 neuronal changes during, 385–386 normal, 385–387 postmortem studies of, synaptic modification and myelination in, 23–25 primary neurulation in, 771–772 prior to gyrification, 39–40 protein-energy status and, 625, 627–630 role of sleep in, 813 selenium in, 627, 634 stress influencing, 69–77. See also Limbichypothalamic-pituitaryadrenocortical (LHPA) system whole-brain mapping methods of, voxelbased morphometry in, 25–26 zinc in, 626, 632–633 Brain injury early, 385–396 behavioral sequelae of, 387–389 measurement of, 389 behavioral therapy for, 391–393 brain development after, 389–391 cortical connectivity changes in, 390 developmental outcomes after, 399–409 endogenous changes in, manipulation of, 391–395

Brain injury (continued) experience-dependent changes in, 386–387 frontal cortical, modification of effects of, 392 gonadal hormones for, 394–395 Hebb hypothesis of, 387–388, 400 impact of age at, functional outcome associated with, 388–389 Kennard hypothesis of, 387–388 moderators influencing outcome of, 407–408 neurogenesis changes in, 390–391 neuromodulators for, 394 neurotrophic factors in, 393 plasticity and, 385–396, 399–409 preterm birth resulting from, 400 neurodevelomental changes and, 400–403 psychoactive drug therapy for, 395 synaptic space and, 389 timing of, 400 type of, 400 focal, spatial analytical functioning disorders associated with, 528 prenatal, face processing impairment and, 515–516 traumatic, 403–407 impact of age at, 406–407 long-term neuropsychological outcome associated with, 405–406 neurodevelopmental process disruption in, 403–404 neuroimaging studies of, 404 Brain models, of repetitive behavior, 711 Brain stem, involved in song learning and production, 454–455 Brain-behavior relationship in visual development, 127–141. See also Visual system cortically motivated, models of, 135–139 subtleties and assumptions in, 139–141 in vocal learning/production of birdsong, 453–454 Brain-derived neurotropic factor (BDNF), 625 cocaine-induced suppression of, 661 gene transcription of, 75 suppression of, 70 Branch formation, along neurites, 13 BrdU-labeled cells, in dentate gyrus, 52, 53, 56 Breast milk, in promotion of CNS development, 629–630 5-Bromo-2′-deoxyuridine (BrdU) labeling, in neurogenesis, 52

C Cajal-Retzius cells calretinin-secreting, 197 hippocampal, 196–200 prefrontal, 221 reelin-secreting, 196–197 Calbindin marker, of granule cells, 200–203

Calbindin-containing interneuron(s), 204 Callous-unemotional (CU) traits, in antisocial behavior, 360 Calretinin-containing interneuron(s), 204 Cambridge Neuropsychological Test Automated Battery (CANTAB) for Alzheimer’s disease, 609–610 for depression, 779 Cambridge Neuropsychological Test Automated Battery (CANTAB) Paired Associates Learning (PAL) scores, 609–610 utility of, in Alzheimer’s disease diagnosis, 610 cAMP response element-binding (CREB) protein, cocaine-induced suppression of, 661 Carbohydrate(s), in cognitive development, 623 Cardiac concomitants, of feedback processing, 889–890 Cardiovascular function, effects of cocaine on, 654 Cataract(s), congenital bilateral, treatment of, sleeper effects in, 428–429 high-level vision and, 420–426 low-level vision and, 417–420 sensitive period for damage with, variability of, 429–431 sensitive period for recovery from, prediction of, 431 treatment of contrast sensitivity after, 417 pace perception after, 422–426 peripheral vision after, 417–418 sensitivity to global form after, 421–422 sensitivity to global motion after, 420–421 sensitivity to motion perception after, 419 visual acuity after, 416–417 visual spatial attention after, 426 unilateral, treatment of, 418 vision deprivation from, 416 Catecholamine, synthesis of, 86 C-domain, of growth cone, 10 Cell(s). See also specific cell type actin organization in, 7 microtubule formation in, 5 Cell count, cortical, age-related changes in, 44 Cell death, 83 cocaine-induced, 654–655 hippocampal, evidence of, 195 programmed, in brain development, 772 Cell migration, hippocampal, 195–196 in preterm infants, 203, 207–208 Cell proliferation, hippocampal, in preterm infants, 203, 207–208 Cellular events, neurogenetic, in prefrontal cortex development, 216, 218–219 Central nervous system (CNS). See also Brain entries development of breast milk promoting, 629–630

impact of prenatal cocaine exposure on, 88–91, 653–666. See also Cocaine, prenatal exposure to nutrients in, 626, 627, 630–635 dysfunction of in autism, 704 in fetal alcohol syndrome, 643 myelination of axons in, 544 Central pattern generators, in development of walking, 150 Central visual field (CVF), in deaf and hearing subjects, 440 Cerebellum in fetal alcohol syndrome, neuroimaging studies of, 644 in Williams syndrome, 691 Cerebral cortex. See also specific cortical specialization axonal-tension–based morphogenesis of, 42 Cajal-Retzius cells in, 197 cell counts in, age-related changes and, 44 development of anomalies in, 83 fundamentals of, 83–85 growth cone navigation along pathways during, 15–18 early injury to. See also Brain injury, early connectivity changes after, 390 histogenesis of, 83–85 intricate architecture of, 41 mechanical folding hypothesis in, implications of, 42–43 morphology of, cocaine-related effects on, 654–655 myelin proliferation into, age-related, 24 neuronal migration into, 40 plasticity of, in auditory development brain correlates of, 109–111 in humans, 103–104 in nonhumans, 102–103 and projections to superior colliculus, in saccade triggering, 265, 266 subcortical structures of, connectivity between, 40–41 Cerebral hemisphere(s) left. See Left hemisphere right. See Right hemisphere c-fos gene, basal levels of, in adolescent vs. adult brain, 862, 863 Change, mechanisms of, in neural network models, 372–375 age-of-acquisition effects and, 374–375 learning about words and semantic categories and, 373–374 Children. See also Adolescent(s); Infant(s); Newborn(s) acquired prosopagnosia onset in, face processing impairment and, 516 antisocial behavior in, CU traits and, 360–361 anxiety disorders in, 756 affective neuroscience and, 763–766 clinical classification of, 763–764 clinical questions concerning, 758 information-processing and fear-circuit functions in, 764–766

index

905

Children (continued) brain development in, MRS imaging of, 341, 343 brain maturation in, 24 cataracts in, 416. See also Cataract(s), congenital circuitry elements in, overproduction of, 230–231 EEG of, 248 executive function in age-related changes in, 556–562 approach-avoidance decisions in, 557 bilingualism facilitating, 565 correlates of, 562–564 cognitive, 563 demographic, 564 socioemotional, 563–564 decision making in, 558 delay-of-gratification paradigms in, 557–558 development of, 553–567 theories of, 554–556 influences on, 564–566 labeling facilitating, 564–565 measurement issues in, 566–567 performance monitoring in, 558–560 reward learning in, 558 rule use at various complexity levels in, 560–561 selective attention in, 561–562 symbolism facilitating, 565–566 task set selection in, 562 training facilitating, 566 working memory in, 561 eye tracking in atypical development and, 276–288 normative development saccades and, 269–276 maltreatment of, 869–877 behavioral response and regulation in, 876–877 cognitive processing of emotion signals in, neural mechanisms involved in, 870–874 consequences of, on LHPA system, 72–73 developmental outcomes associated with, 869–870 emotion elicitation in, 874–875 emotion regulatory processes in, 875–876 prefrontal cortex in, development of, 553–554 prenatally cocaine-exposed. See also Cocaine, prenatal exposure to attention-deficit/hyperactivity disorder in, 89 neurobehavioral and developmental findings in, 664–666 neurochemical findings in, 664 outcomes of, 663–666 prosodic, lexical, semantic, and syntactic processing in, 117– 125

906

index

regulation of LHPA function in, 72–73 sleep patterns and neurocognitive functioning in, 810–811 Cho/Cr ratio in attention-deficit/hyperactivity disorder, 345–346 in autism, 344 in developmental delays, 344 Choline, in brain development, 627, 634–635 Cholinergic afferents, prefrontal, 222 Cholinergic system in arousal aspect of attention, 480 basal forebrain–prefrontal, 215 Cingulate cortex, anterior functional studies of, 169 lesion studies of, 169 neuroanatomy of, 168–169 role of, in social behavior, 168–169 Circadian process, 809 Circular reaction, in infant motor development, 148 Clinical policy domains, relevance of sleep to, 813 Closure positive shift (CPS), in processing intonational phrase boundaries, 119 CNS. See Central nervous system (CNS) CNV (contingent negative variation) in adolescence, 586 age-related changes in, to warning stimulus on pro- and antisaccade tasks, 269 Cocaine pharmacological sites of action of, 88–89 prenatal exposure to affecting neurotransmitter function, 655–660 altered CNS development in, 88–91, 653–666 cortical morphology in, 654–655 dopaminergic system and, 656–657 and expression of immediate early genes, 660–661 GABAergic system and, 657–658 human model of, 663–666 incidence of, 653 neuroadaptive changes in response to, 89–90 neurobehavioral and developmental findings in infants and children after, 664–666 neurobehavioral teratologic effects of, candidate mechanisms for, 653–654 neurochemical findings in infants and children after, 664 noradrenergic system and, 660 preclinical models of, 654–660 behavioral correlates in, 661–663 serotonin system and, 658–659 Cochlear implantation advent of, 446 age at, performance affected by, 447 speech perception after, 446 cross-modal plasticity in, 446–448

success of, 103–104 young recipients of, auditory/visual speech fusion in, sensitive period for, 444–448 Cognition aging and dementia and, 607–616 reserve, underactivation, and compensation in, 600–601 typical development in, 591–601 diffusion-tensor-imaging studies of, 305 in early-stage Alzheimer’s disease, 608–611 impaired, in dyslexia, 742–743 negative, mood-valent theories of, depression and, 775–776 social. See also Social brain definition of, 162 elements of, 162 neurobiology of, 161–172 unipolar depression and, 774–775 Cognition-emotion interactions, abuse and neglect impacting, 876 Cognitive correlates, of executive function in childhood, 563 Cognitive development, 247–258 electroencephalogram in, 247–248 event-related oscillations in, 256–257 event-related potentials in, 248–249 auditory stimuli and, 249–252 components related to lexical and syntactic processing in, 252 mismatch responses in, 251–252 multimodal stimuli and, 255–256 saccade-related, 256 visual stimuli and, 252–254 future directions in, 257–258 macronutrients in, 623 effects of, 629–630 micronutrients in, 623–624 neural network models of, 367–379. See also Neural network models benefits of, 367–368 challenges to, 375–376 contributions of, 371–375 critical elements of, 368–371 nutrients in categories of, 623–624 deprivation and subsequent repletion of, role of timing in, 635–636 effect of selected, 625–635 nutrition in, 623–637 future directions of, 636 role of, 624–625 in prenatally cocaine-exposed children, 665 Cognitive function, relationship between myelination and, 240 Cognitive models, in autism, 704–706 Cognitive neuroscience developmental behavior genetics in, 351–354 fMRI studies of, 316–320 behavioral development and, 318–319 behavioral paradigms and, 315–316

Cognitive neuroscience (continued) brain development and, 317–318 history of, 316–317 intervention or treatment effects using, 319–320 imaging genetics in conceptual basis of, 354–355 findings of, 355–358 importance of, 358–360 ongoing investigations in, 360–362 transitional aspects of, 361–362 of sleep, 807. See also Sleep Cognitive process(ing) effect of attention on, 490 of emotion signals effects of early adversity in, 872–873 neural mechanisms involved in, 870–874 of emotional states, 874 effects of early adversity in, 874–875 role of, as modulators of temperament, 840 Cognitive reserve, age-related, 600 Cognitive resilience, 780 Cognitive resource regulation, pupillary dilation studies of, in children and adolescents, 276 Cognitive skills, slower gains of, in preterm children, 403 Cognitive-linguistic profile, in Williams syndrome, 693 Coherence thresholds in global motion sensitivity, 420 weak central, in autism, 705 Communication, vocal, origins of, 795–797 Comorbid disorders, associated with dyslexia, 745–746 Compensation, cognitive, age-related, 600–601 Compensation-related utilization of neural circuits hypothesis (CRUNCH), evidence supporting, 594 Complementary nutritional measures, cognitive development and, 636 Complex information processing, in autism, 705–706 Complexity levels of, in executive function in childhood, 560–561 neural network models dealing with, 367–368 Compulsions, 717, 729. See also Obsessive compulsive disorder (OCD) CSTC circuitry in, 726–729 Computerized decision-making task, development of, 776–777 Conditioned head-turning procedure, in auditory development, 99 Conditioned tone stimulus, in auditory cortex organization, 103 Conduct disorder, clinical depression and, 774 Cones, in retina, 129–130 neonatal morphology of, 129 Congenital prosopagnosia, face processing impairment and, 516

Connectivity development of, between cortical and subcortical structures, 40–41 intricate pattern of, in prefrontal cortex, 576–577 Consonants, in linguistic theory, 329–330 Contingent negative variation (CNV) in adolescence, 586 age-related changes in, to warning stimulus on pro- and antisaccade tasks, 269 Contrast sensitivity, after treatment of congenital cataracts, 417 Control, in neural network models, 367 Coping resource, attention as, 847 Core cortical system, in face processing in adults, 509–511 fusiform area of, 510–511 superior temporal regions of, 511 eye gaze and, 514–515 in infants, 512–515 fusiform area of, 512–514 superior temporal regions of, 514–515 sensitivity to emotion and, 514 Corpus callosum corticocortical fibers forming, guidance decisions of, 16–17 effect of traumatic brain injury on, 404 in fetal alcohol syndrome, neuroimaging studies of, 644–645 peak growth rate of, 553 in Williams syndrome, 691 Cortex cerebral. See Cerebral cortex; specific cortical specialization visual. See Visual cortex Cortical mapping methods of gray matter density, 26–30, 403 of gray matter thickness, 30–33 Cortical parcellation hypothesis, of visual information processing, 140–141 Cortical plate glial cell migration to, 40 neuron cell migration to, 40–41 subcortical structures of, development of connectivity between, 40–41 Cortical-subcortical system, extended, in adult face processing, 511 Corticocortical connections, prefrontal, 215–216 development of, 224 long associative, 224–225 Corticohypothalamic connections, prefrontal, 215 Corticospinal axons, navigation of, 16 Corticosterone action of, on amygdala, 794–795 in learning avoidance, 793 natural fluctuations of, 793–794 Corticostriatal fibers, prefrontal, 215 Corticostriatothalamocortical (CSTC) circuitry components of, 718, 726 in perceptually cued learning, 729–730 in tics and compulsions, 726–729

Corticothalamic axons navigation of, 15–16, 17 prefrontal, 215 Corticotropin-releasing hormone (CRH) influence of on brain, 68–69 on brain development, 70 involved in stress reactivity, 841 secretion of, LHPA stress response and, 63 Cortisol elevated in children attending day-care, 73 in hippocampus, 69 in puberty, 73–74 influence of on brain, 67–68 on brain development, 69–70 salivary, collection of, 65 synthetic form of, 69 Covert orienting attention, 482 CPS (closure positive shift), in processing intonational phrase boundaries, 119 Cretinism, iodine deficiency causing, 633–634 CRH. See Corticotropin-releasing hormone (CRH) Critical elements, of neural network models, 368–371 challenges to, 376–377 learning algorithms as, 370–371 net input and activation functions as, 369–370 units and weights as, 368–369 CSTC circuitry. See Corticostriatothalamocortical (CSTC) circuitry CVF (central visual field), in deaf and hearing subjects, 440 CYLYN2 gene, in Williams syndrome, 692 Cytoplasmic signaling pathways, dynamics of, regulation of microtubule and actin organization and, 8–9

D Day-care, elevated cortisol in children attending, 73 DCCS (Dimensional Change Card Sort), 558, 559, 561, 566, 846 Deafness, 439–449 American Sign Language and, 444–446. See also American Sign Language cochlear implantation for, 446–448. See also Cochlear implantation cortical plasticity in, 103–104 effects of on processing of visual motion, 439–443 on processing of visual space, 443– 444 Decision making in attention-deficit/hyperactivity disorder, 886 autonomic signals in, using Iowa Card Gambling Task, 884–885

index

907

Decision making (continued) behavioral and cognitive disinhibition in, 886–887 development of, 883–887 in executive function in childhood, 558 approach-avoidance, 557 fMRI studies in, 887 future directions of, 892–893 Hungry Donkey Task in, 885 Iowa Card Gambling Task in, variants of, 886 in unipolar depression, 776–778 Declarative memory adult recognition abilities in, emergence of, 501–502 adult relational abilities in, emergence of, 502–503 beyond infancy, 547–548 development of, 501–503 cognitive neuroscience approach to, 541–550 hippocampus in, 543 early damage to, 548–549 in infancy, 542–543 brain development and, 543 deferred imitation and, 543 development of, 543–547 encoding and, 543–544 retention and, 544–545 retrieval and, 545–547 visual paired-comparison (VPC) task and, 542–543 neural circuits mediating, maturation of, 503 Deferred-imitation task, infant performance on, declarative memory and, 543 Delayed nonmatch to sample (DNMS) task, 541 in assessing recognition-memory abilities, 502 in joint-attention study, of autistic children, 823–824 Delay-of-gratification paradigms, in executive function in childhood, 557–558 Delay-of-gratification task, in assessing attention and behavioral inhibition, 846 Deliberate phase, of walking, 150 Dementia, 607–616. See also Alzheimer’s disease brain abnormalities in, functional and structural, 614–615 differential cognitive profiles of, 612–614 measuring change over time in, 611 types of, 607 vs. normal aging, 607 Dementia with Lewy bodies, 607 vs. Alzheimer’s disease, 613–614 Demographic correlates, of executive function in childhood, 564 Dendrite(s) development of, 18–19 differentiation of, 13 distinguishing microtubule features in, 6 formation of, developing neurons in, 5–20

908

index

initial growth of, neuronal polarization and, 13–14 in normal brain, 386 structural abnormalities of, prenatal cocaine exposure associated with, 655 Dendritogenesis, in prenatal and postnatal brain, 19 Dentate gyrus Cajal-Retzius cells in, 197 cell proliferation suppression in, 54 development of, 52 granule cells in, 187, 200 calbindin immunoreactivity of, 200– 202 in hippocampal formation cell formation and, 191, 192 location of, 189 Depression associated with 5HTT gene, in maltreated children, 76 clinical characteristics of, 773–774 cognitive resilience and, 780 cognitive theories of, 775–776 genetic influence on, 353 onset of, critical stressor in, 359 symptoms of, 773 unipolar behavioral inhibition and biases in, 778–779 decision making in, 776–778 developmental neuropsychology of, 771–781 diagnosis of, 773–774 executive dysfunctions of, 776–780 impaired attention in, 779–780 malfunction of brain regions and, 774–775 Detroit principle, of developmental plasticity, 429 Developmental cognitive neuroscience behavior genetics in, 351–354 fMRI studies of, 316–320 behavioral development and, 318–319 behavioral paradigms and, 315–316 brain development and, 317–318 history of, 316–317 intervention or treatment effects using, 319–320 imaging genetics in conceptual basis of, 354–355 findings of, 355–358 importance of, 358–360 ongoing investigations in, 360–362 transitional aspects of, 361–362 Developmental delay(s), MRS imaging of, 343–344 Developmental neuropsychology, of unipolar depressions, 771–781. See also Depression, unipolar Dexamethasone, fetal exposure to, 69 Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV), criteria in for Alzheimer’s disease, 607 for anxiety disorders, 755

for Asperger syndrome, 701 for autism, 165 Dialects, 108–109 Diet, low phenylalanine, for phenylketonuria, 683 Dietary therapy, for phenylketonuria, 678 Diffusion coefficient, apparent, 301 in newborns, 304 Diffusion spectral imaging (DSI), 307 Diffusion tensor imaging (DTI), 301–307 advanced methods of, 305–307 anatomical and physiological correlates of, 303–304 of brain in autistic children, 704 in preterm-born children, 401 in socioemotional deprivation studies, 408 of corpus callosum, memory performance and, 592, 593 data acquisition in, 301–302 methodological challenges of, 304 data processing in, 302–303 developmental correlates of, 304–305 limitations of, 306 principles of, 301 of traumatic brain injury, 404, 405 of white matter, 578 Dimensional Change Card Sort (DCCS), 558, 559, 561, 566, 846 Dissociation, dorsal and ventral, behavioral evidence of, 471–474 Distributional modifications, in phonetic reorganization, 105 Distributional sensitivity, in phonetic reorganization, 105–106 of second language, 106–107 Division-of-labor hypothesis, in linguistic theory, 330 Dopamine abnormalities of, disorders associated with, 86 developmental neuropharmacology of, 85–87 influence of, in emotion regulation, 352 synthesis of, 86 tyrosine in, 683 in vitro studies of, 87 Dopamine receptor(s), 727 subfamilies of, 86 transcripts for, 86–87 Dopamine receptor coupling, reduced, prenatal cocaine exposure and, 91 Dopamine receptor signaling balance of, 91–92 loss of, prenatal cocaine exposure and, 90 Dopamine transmission, abnormalities of, in Tourette’s syndrome, 727 Dopaminergic system in arousal aspect of attention, 480 effect of prenatal cocaine exposure on, 656–657 of prefrontal cortex, 214–215 Dorsal visual pathway. See Visual pathways, dorsal “Double bouquet” neurons, 221

Double-deficit hypothesis, of dyslexia, 742 Down syndrome, vs. Williams syndrome, 695 DSI (diffusion spectral imaging), 307 DSM-IV. See Diagnostic and Statistical Manual of Mental Disorders-IV edition (DSM-IV) entries DTI. See Diffusion tensor imaging (DTI) Duration discrimination, in auditory development, 99–100 Dynamic systems approaches (DSAs), to motor development, 148–149 critical implications of, 149 Dynein, in regulation of microtubule organization, 5 Dyslexia, 739–749 attention-deficit/hyperactivity disorder and, 745 behavioral profile in, 740–741 comorbid disorders with, 745–746 core cognitive impairment in, 742–743 definition of, 739–740 functional anatomy and, 743–744 future goals for, 749 genetic studies of, 744–745 reading intervention for, 746–747 neurobiological basis of, 747–749 reading-related deficits in, 740 recognition of, 739 risk assessment and prediction of, 746 specific language impairment with, 745–746 theories of, 741–743

E Earlier left anterior negativity (ELAN), in sentence processing, 124 Echo time (TE), in proton MRS, 338 Eigenvalues, in DTI data processing, 302 Eigenvectors, in DTI data processing, 302 ELAN (earlier left anterior negativity), in sentence processing, 124 Electrode(s), placement of, for electroencephalography, 247 Electroencephalography (EEG) asymmetry of, as maker of negative reactivity in temperament, 843 in cognitive neuroscience, 247–248 of joint-attention development, 826 ELN gene, deletion of, in Williams syndrome, 692 Emotion in attachment, emergence of, 795–797 elicitation of, 874 effects of early adversity in, 874–875 expression of, deficit identification of, orbitofrontal cortex lesions in, 168 regulation of genes involved in, 352–353 sleep and, 813–814 sensitivity to, in face processing, 514 Emotion processing effects of maltreatment on, 870 regulatory, 875 effects of early adversity in, 875–876

Emotion signals, cognitive processing of effects of early adversity in, 872–873 neural mechanisms involved in, 870–874 Encoding of information age-related differences in, 544 developmental changes in, 544 episodic, long-term memory and, 598–599 in infancy, declarative memory and, 543–544 Endophenotype(s), 759 in autism, 702 Entorhinal cortex, in declarative memory, 503 Environment acoustic, in auditory cortex organization, 102–103 influence of on brain-behavioral development, 869–877 on early brain injury recovery, 393, 407 on traumatic brain injury, 407–408 stressful and adverse, children exposed to, 869–877 Environmental factors in fear circuit function variations, 762–763 and pediatric anxiety, 763–764 Epigenetic processes, in neurodevelopment, 636 Epigenetics, 797 Equifinality, in LHPA function, 74 ER22/23EK polymorphism, and individual LHPA function, 76 EROs. See Event-related oscillations (EROs) ERPs. See Event-related potentials (ERPs) Error-driven learning algorithms, 370 Error-related negativity (ERN) in behavioral inhibition, 847 in performance monitoring, 558, 887–888 Estrogen, regulating adult neurogenesis, 53–54 Event-related oscillations (EROs), 256–257 and object processing, 471 Event-related potentials (ERPs) abnormal, in Williams syndrome, 693–694 in behavioral inhibition, 847–848 in brain activity, measurement of, 117–118 in cognitive development, 248–249 auditory stimuli and, 249–252 lexical and syntactic processing and, components related to, 252 mismatch responses in, 251–252 multimodal stimuli and, 255–256 saccade-related, 256 visual stimuli and, 252–254 in language development, 119–122, 124 scalp-recorded, as psychophysiological measure of attention, 483–484 in self-regulation, 847 studies of in deafness and visual motion processing, 439–440 in dorsal and ventral visual pathways, 469–471

Excitatory neurons, hippocampal, 200–203 Executive function and aging, 595–597 in autism, 704–705 development of, in childhood, 553–567. See also Children, executive function in impaired in fetal alcohol syndrome, 647–648 in phenylketonuria, 681 in unipolar depression, 776–780 Experience, and brain development, normal, 386–387 Experience-dependent processes, in early brain injury, 407–408 Experience-dependent synaptogenesis, 386–387 Experience-expectant processes, in early brain injury, 407–408 Explicitness, in neural network models, 368 Extinction, fear conditioning and, 760 Extrathalamic afferent system, of prefrontal cortex, 214–215 basal forebrain–prefrontal cholinergic fibers in, 215 development of, 222 dopaminergic fibers in, 214–215 Eye. See also Vision; Visual entries; specific part optics of, 129–130 Eye contact, diminished, temporal cortex lesions and, 170 Eye gaze. See Gaze entries Eye monitors, 263, 265 Eye movements and attention, 484–489 ballistic, 263, 264, 265. See also Saccades control of, brain area involved in, 485 main sequence in, 487–489 reflexive saccadic effect of attention in, 487 in tracking visual stimuli, 485 during scene and face perception, 264, 266–267 in children and adolescents, 273 smooth pursuit, 266, 267 in children and adolescents, 269, 273 effect of attention on, 486–487 in tracking visual stimuli, 485 types of, 485 effect of attention on, 486–487 voluntary saccadic, in tracking visual stimuli, 485 Eye tracking studies, 263–294 in children and adolescents atypical development and, 276–288. See also specific disorder normative development saccades and, 269–276 limitations of, 292 potentials of, conclusions regarding, 292–294 review of, in normative and atypical development, 284, 289–292 tasks and measures in, 263–269

index

909

F Face(s), detection of, 511–512 after treatment of congenital cataracts, 422, 423 Face perception, 422–426 deficits in, autism and, 707–708 face detection in, 422, 423 facial identity in, 425–426 holistic processing in, 422, 424–425 neural substrates of, 442 visual processing in, effects of deafness on, 442 Face perception tasks, in eye tracking, 264, 266–267 in children and adolescents, 273 Face processing adult, 509–511 core cortical system in, 509–511 fusiform face area of, 510–511 superior temporal sulcus/gyrus of, 511 extended cortical-subcortical system in, 511 subcortical system in, 509 in autism, 706–708 development of, 511–515 core cortical system in fusiform area of, 512–514 superior temporal regions of, 514–515 impairments in, 515–516 neurocognitive mechanisms for, 509–517 subcortical system in, 511–512 holistic composite face effect in, 422 deaf vs. hearing subjects and, 443 visual deprivation preventing, 424–425 whole-part advantage in, 422, 424 impairments in childhood onset of acquired prosopagnosia and, 516 congenital prosopagnosia and, 516 development of, 515–516 fusiform gyrus lesions and, 171 perinatal brain injury and, 515–516 Williams syndrome and, 693–694 right and left hemispheres in, 442 Face-voice pairs, emotionally congruent and incongruent, infants’ ERP responses to, 255 Facial anomalies, in fetal alcohol syndrome, 643 Facial expressions imitation of, in neonates and infants, 830 processing of. See also Face processing deficits in, autism spectrum disorders and, 708 Facial identity processing of. See also Face processing in autism, 706–708 recognition of, 425–426 Familiar stimulus, and attention, 489–490 Familiarity, retrieval process of, differential effects of aging on, 600

910

index

FAS. See Fetal alcohol syndrome (FAS) FASD (fetal alcohol spectrum disorders), 643 Fast-spiking GABAergic interneurons, 730 Fat(s), in cognitive development, 623 “Fate map” concept, in neurulation, 39 Fatty acids, long-chain polyunsaturated, in breast milk, 629, 630 Fear. See also Anxiety entries definition of, 755 elicitor of, 841 Fear circuit function information-processing and, 761–762 in pediatric anxiety, 764–766 modulators of, 762–763 Fear conditioning conceptualization of, 761 and extinction, 760 neural circuit engaged during, 759–760 research on, 760 Fear-potentiated startle, 841 Fear-related impairment, amygdala lesions causing, 167 Feedback processing cardiac concomitants of, 889–890 developmental change in, 890–892 in performance monitoring, 888–892 Fetal alcohol spectrum disorders (FASD), 643 Fetal alcohol syndrome (FAS), 643–648 attention deficits in, 647 autopsy studies of, 643–644 diagnosis of, criteria for, 643 executive function deficits in, 647–648 fMRI studies of, 645–646 future research in, 648 intellectual function deficits in, 646 learning and memory impairments in, 646–647 methodological considerations in, 643 motor dysfunction in, 648 neuroimaging studies of, 644–646 basal ganglia in, 645 brain metabolism in, 645 cerebellum in, 644 corpus callosum in, 644–645 neuropsychological studies of, 646–648 speech and language deficits in, 648 visual-spatial dysfunction in, 647 Fetal origins hypothesis, of nutritional environment, 636 Fetus auditory development in, 97–98 brain development in cocaine-compromised, 654. See also Cocaine, prenatal exposure to MRS imaging of, 339–340 exposure of, to dexamethasone, 69 late-gestation, iron deficiency in, 632 malnutrition in insulin-like growth factor-1 and, 625 protein-energy, effects of, 625, 627–628 origins of attachment in, 788 prefrontal intrinsic circuitry in, 227, 229 prenatal experiences in, influencing postnatal brain structure, 387 FFA. See Fusiform face area (FFA)

Fibroblast growth factor (FGF), 625 Fibroblast growth factor-2 (FGF-2), administration of, regeneration of injured cortical tissue after, 390–391 Flasker task, in assessment of temperament, 847 Fluoxetine, prenatal exposure to, brain development and, 395 fMRI. See Functional magnetic resonance imaging (fMRI) Focal brain insults, 403. See also Brain injury Folic acid deficiency, 624 Forebrain adolescent rodent microarray analysis of, 857, 859 nicotine affecting gene expression in, 862 anterior pathway of. See Anterior forebrain pathway developmental milestones in, 84 FOXP2 gene mutation, 459 Fractional anisotropy (FA), in DTI data processing, 302, 578 Frequency discrimination, in auditory development, 99–100 Frontal cortex dorsal medial, joint attention and socialcognitive performance involving, 827 early injury to. See also Brain injury, early modification of effects of, 392 microstructural asymmetry and hemispheric specialization of, 231–232 studies of, in obsessive-compulsive disorder, 722 Frontotemporal dementia, 607 vs. Alzheimer’s disease, 612 Functional magnetic resonance imaging (fMRI), 311–321 in auditory cortex mapping, 104 basic principles of, 312–316 of brain, 118 in adolescents, 856 in autistic children, 704, 708–709 speech stimuli and, 111 in decision making, 887 in developmental cognitive research, 316–320 behavioral development and, 318–319 behavioral paradigms and, 315–316 brain development and, 317–318 history of, 316–317 intervention or treatment effects using, 319–320 of fetal alcohol syndrome, 645–646 future directions of, 320–321 of gaze following activities, 824, 825 of neural activation associated with reading, 327 of obsessive-compulsive disorder, 723 of pediatric anxiety disorders, 764–766 in performance monitoring, 891 physics of, 312 physiologic basis of, 312–314

Functional magnetic resonance imaging (fMRI) (continued) in reading processing, 743 safety considerations for, 314–315 of schizophrenia, 825 signal artifact in, sources of, 314 Fusiform face area (FFA) in adult face processing, 510–511 in autism spectrum disorders, 707–708, 710 in infant face processing, 512–514 Fusiform gyrus functional studies of, 170–171 lesions of, 171 neuroanatomy of, 170 role of, in social behavior, 170–171

G Gain, in smooth-pursuit eye movement, definition of, 266 Gamma-aminobutyric acid (GABA) neurons, in prefrontal cortex, 214 Gamma-aminobutyric acid (GABA) neurotransmitter, inhibitory, 135, 726 Gamma-aminobutyric acid (GABA) receptor genes, repetitive behavior in autism and, 711 Gamma-aminobutyric acid (GABA) system effect of prenatal cocaine exposure on, 657–658 role of, in autism, 704 Ganglia, basal. See Basal ganglia Ganglion cells, of retina, 130, 139 Gap and overlap tasks, in eye tracking, 264 Gaze alternating, 822 in autism abnormal, 281 eye tracking studies of, 282–284, 708 direction of, in face processing, 514–515 Gaze following activity in early development importance of, 821–822 individual differences in, 821 nature of, 819–821 fMRI of, 824, 825 and ventral social brain, 823–825 Gaze shifting, 819–820 Gene(s). See also specific gene candidate, in autism, 702–703 identification of, in adolescent rodent brain, 857–859 involved in emotion regulation, association studies of, 352–353 regulation of, recognized system of, 797–798 study of, 351–352 Gene expression immediate early, prenatal cocaine exposure and, 660–661 nicotine affecting, in adolescent rodent brain, 862 Generalized anxiety disorder, 755. See also Anxiety disorder(s) attention avoidance in adolescents with, 764

Genetic aspects, of vocal development, 459 Genetic markers, of negative reactivity in temperament, 843–844 Genetic polymorphism impact of, on behavior, 354 in individual LHPA function, 75–76 Genetic studies association, 352–354. See also Association genetic studies of behavior, 351–354. See also Behavior genetics of dyslexia, 744–745 of phenylketonuria, 684–685 Genotype, in Williams syndrome, 691–692 Geodesic sensor net, for electroencephalography, 247 Germinal matrices, in hippocampal development, 187–195 proliferating cells of, 189 GGF (glial growth factor), 625 GH (growth hormone), mother-infant interactions and, 795 GI (gyrification index), 42 utilization of, in neuropsychiatric disorders, 45–46 Glial cells, migration of, to cortical plate, 40 Glial growth factor (GGF), 625 Global form, sensitivity to, after treatment of congenital cataracts, 421–422 Global motion, sensitivity to after treatment of congenital cataracts, 420–421 in high-level vision, 420–421 Glucocorticoid(s), regulating adult neurogenesis, 53 Glucocorticoid receptor(s), in cortisol binding in brain, 67–68 Glycogen synthase kinase-3β, cocaineinduced suppression of, 661 Goldman-Rakic’s model, of prefrontal function, 579 Gonadal hormone(s), and recovery from early brain injury, 394–395 Go-no-go task, 585–586 in assessment of brain activation, 847 impairments of, obsessive-compulsive disorder and, 722 GR gene, and individual LHPA function, 76 Granule cell(s), of dentate gyrus, 187, 200 calbindin immunoreactivity in, 200–203 Granule cell neurogenesis, in dentate gyrus, 52, 53 Grasping, precision in, 150 Gratification, delay of, in executive function in childhood, 557–558 Grating acuity, deficits in, congenital cataracts and, 416–417 Gray matter cortical rim of, 39–40 development of back-to-front pattern of, 577–578 changes in, cognitive correlates of, 33–36 effect of traumatic brain injury on, 404

volume decrease of age-related, 25, 26, 44, 553 rate and pace of, 772 Gray matter density, cortical mapping methods of, 26–30 Gray matter thickness, cortical mapping methods of, 30–33 Growth cone elongation, in neurite maturation, 10–12 Growth cone navigation in neurite maturation, 12 pathways for, 14–15 during cerebral cortex development, 15–18 thalamocortical, 17 Growth cone organization, in neurite maturation, 10 Growth factor(s), 624–625 Growth hormone (GH), mother-infant interactions and, 795 Growth hormone releasing factor, motherinfant interactions and, 795 Growth retardation, intrauterine, 627–628, 632 GTF2IRD1 gene, in Williams syndrome, 692 Gyrification abnormalities of, clinical conditions associated with, 44–46 brain development prior to, 39–40 heritability of, 43 ontogeny of, theories of, 41–44 phylogeny in, 41 postnatal changes in brain morphology and, 43–44 Gyrification index (GI), 42 utilization of, in neuropsychiatric disorders, 45–46 Gyrus alterations in, 45 dentate. See Dentate gyrus fusiform, role of, in social behavior, 170–171 middle temporal, activation in, deaf vs. hearing subjects and, 440 superior temporal in face processing, 511, 514–515 in social perception, 823

H Habit-learning systems, 723–726 in obsessive compulsive disorder, 725–726 PCL assessment of, 724 in Tourette’s syndrome, 725 Habitual behaviors. See also Obsessivecompulsive disorder (OCD); Tourette’s syndrome disturbances of self-regulatory control in, 717–731 Hansen and Fulton model, of visual development, 128–129 HARDI (high-angular-resolution diffusion imaging), 307 Head, increased growth of, autism and, 703–704

index

911

Hearing. See also Auditory system loss of. See Deafness onset of, 97–98 Heart rate, as psychophysiological measure of infant attention, 482–483, 489 Hebb hypothesis, of early brain injury, 387– 388, 400 Hebbian algorithm, 370 Heel-toe progression phase, of walking, 150 Hemispheric asymmetry reduction in older adults (HAROLD) model, 598–599 Hemispheric encoding/retrieval asymmetry (HERA) model, 598 Heritability, of gyrification, 43 High variability phonetic training (HVPT) regime, 107 High-angular-resolution diffusion imaging (HARDI), 307 Hilar mossy cells, spines of, formation of, 202 Hippocampus activation of, during transitive interference task, 546 in declarative memory performance, 543 early damage to, 548–549 dentate gyrus in cell formation in, 191, 192 location of, 189 effect of traumatic brain injury on, 404 elevated cortisol in, 69 in expression of novelty preference, 542–543 formation of Cajal-Retzius cells in, 196–200 calretinin-secreting, 197 in newborns, 200 reelin-secreting, 196–197 cell migration in, 195–196, 203, 207–208 cell proliferation in, 203, 207–208 evidence of cell death in, 195 germinal matrices and cell proliferation in, 187–195 morphological changes of cytoarchitectonics in, 187, 190, 191 postmortem studies in, 188 pyramidal cells in, 187, 200, 202–203 neurogenesis in in adults, 51–59 discovery of, 51–52 methodological advances leading to, 52 experience regulating, 54–57 functional significance of, 57–58 hormones regulating, 52–54 learning and, 55–57 blockade of, 58 parallel changes in, 57–58 stress affecting, 54–55 neurons of, excitatory and inhibitory, 200–203 in Williams syndrome, 691

912

index

Holistic face processing. See also Face processing composite face effect in, 422 deaf vs. hearing subject and, 443 visual deprivation preventing, 424–425 whole-part advantage in, 422, 424 Hormone(s). See also specific hormone regulating adult neurogenesis, 52–54 5-HT. See Serotonin (5-HT) 5-HT gene, in threat response behavior, 762 5-HTT gene, 75 depression associated with, in maltreated children, 76 in personality genetics, 353, 844 polymorphism of, 355–358 temperament associated with, 844 in threat response behavior, 762 5-HTTLPR S allele in depression and anxiety, 844 effects of on amygdala reactivity, 356–358 on brain circuitry, 358 Hungry Donkey Task, in decision making, 885 Hyperactivity disorders, in fetal alcohol syndrome, 647 Hyperphenylalaninemia, 678 Hyperserotoninemia, in autism spectrum disorders, 704 Hypersociability, in Williams syndrome, 165 Hypothalamus, prefrontal cortex connection to, 215

I IEGs (immediate early genes), expression of, prenatal cocaine exposure and, 660–661 IGF-1 (insulin-like growth factor-1), 625 Imaging genetics conceptual concept of, 354–355 findings in, 355–358 importance of, in understanding development and psychopathology, 358–360 ongoing developmental studies in, 360–362 transitional aspects of, 361–362 Immediate early genes (IEGs), expression of, prenatal cocaine exposure and, 660–661 Implants, cochlear. See also Cochlear implantation advent of, 446 success of, 103–104 Impulsivity, increased, in unipolar depression, 777 In utero cocaine exposure. See Cocaine, prenatal exposure to Independent stepping phase, of walking, 150 Indexical factors, in speech processing, 108–109 Indoleamine 5-HT, 87

Infant(s). See also Children; Newborn(s) attachment of development of, 797–798 neurobiology of, 787–799 attention in arousal system of, 479–480 brain systems involved in, 479–482 briefly presented stimuli and, recognition of, 489–494 covert (orienting), 482 definition of, 479 developmental psychological perspective on, 479–495 direct measure of, 484 event-related potentials and, scalprecorded, 483–484 eye movements and, 484–489. See also Eye movements heart rate in, 482–483, 489 indirect measure of, 484 psychophysiological measures of, 482–484 auditory development in absolute thresholds and, 98–99 bilingual exposure and, 106–107 distributional sensitivity of, 105–106 fundamental capacity of, 97–100 intensity, frequency, and duration discrimination and, 99–100 reference and learning in, 107–109 sound thresholds in, 99 in utero experiences and, 97–98 auditory localization in, 101 cognitive development studies in, 247–258 electroencephalogram in, future directions in, 247–248 event-related oscillations in, 256–257 event-related potentials in, 248–249 auditory stimuli and, 249–252 multimodal stimuli and, 255–256 saccade-related, 256 visual stimuli and, 252–254 future directions in, 257–258 declarative memory in, 542–543 brain development and, 543 deferred imitation and, 543 development of, 543–547 encoding and, 543–544 retention and, 544–545 retrieval and, 545–547 visual paired-comparison (VPC) task and, 542–543 face processing in core cortical system in fusiform area of, 512–514 superior temporal regions of, 514–515 development of, 511–515 impairments in, 515–516 neurocognitive mechanisms for, 509–517 subcortical system in, 511–512 imitation of facial expressions in, 830 language acquisition in, NIRS studies of, 326–328

Infant(s) (continued) movement in, 147–148. See also Motor system, development of overall level of motor activity in, 151 prenatally cocaine-exposed. See also Cocaine, prenatal exposure to neurobehavioral and developmental findings in, 664–666 neurochemical findings in, 664 outcomes of, 663–666 physical abnormalities in, 89 preterm. See Preterm birth prosodic, lexical, semantic, and syntactic processing in, 117–125 reaching abilities of, 149–150 walking phases in, 150–151 Infant start state, vs. phenotypic end state in Williams syndrome, 695 Information processing complex, in autism, 705–706 and fear circuit function, 761–762 in pediatric anxiety, 764–766 Inhibition behavioral, 74–75, 585–586 in temperament, 841–843 in unipolar depression, 778 and pediatric anxiety, 763 Inhibitory cells, axo-somatic, parvalbumincontaining, 203 Inhibitory control, relations between temperament and, 846 Inhibitory neurons, hippocampal, 200– 203 Inhibitory processes, age-related impairment in, 595 Insulin-like growth factor-1 (IGF-1), 625 Integrated walking phase, 150 Intelligence, impaired, in fetal alcohol syndrome, 646 Intelligence quotient (IQ) score after traumatic brain injury, 406 in children with phenylketonuria, 678 everyday-memory function impairment and, 548 fetal alcohol syndrome and, 646 in preterm children, 402 relationship between executive function and, 647 relationship between gray matter structure and, 34–35 Williams syndrome and, 693 Intelligence quotient (IQ) tests, reading achievement and, 740 Intensity discrimination, in auditory development, 99–100 Interindividual variability, of sleep patterns, 811 Interneuron(s) calbindin-containing, 204 calretinin-containing, 204 parvalbumin-containing, 203, 205, 206 Intonational phrase boundaries, sentential prosody marking, 118–119 Intra-Dimensional, Extra-Dimensional (IDED) Set Shifting Task, in depression, 779

Intraindividual stability, of sleep patterns, 811 Intrauterine growth retardation (IUGR), 627–628, 632 Intrinsic circuitry, prefrontal, 214 endogenous and sensory-driven, 229, 230 neonatal, 229–230 overproduction of elements of, in infancy and childhood, 230–231 preterm infant (endogenous), 229 prolonged plasticity and reorganization of, in adolescence and postadolescence, 231 Iodine in brain development, 626, 633–634 deficiency of, 633–634 Iowa Gambling Task (IGT), 884–885 in decision making, 884–885 variants of, 886 reward and punishment schedules in, 884 Iron in brain development, 626, 630–632 deficiency of, 631 in late-gestation fetuses, 632 Isolation, ultrasonic vocalization response to, 796 IUGR (intrauterine growth retardation), 627–628, 632

J Joint attention. See also Attention adoption of term, 819 anterior attention system and, 828–829 development of, 819–833 gaze following, 829–833 individual differences in, 821 and dorsal social brain, 825–827 impairment of, in autism, 822–823 importance of, 821–822 initiation of, 820 related to language outcome, 821 multi-process model of, 831 nature of, 819–821 posterior attention system and, 827–828 responding to, 820 and ventral social brain, 823–825

K Katanin, in regulation of microtubule organization, 5–6 Kennard hypothesis, of early brain injury, 387–388 Kinesin, in regulation of microtubule organization, 5 Knowledge neural network models of, 371–372 phonotactic, and lexical form, 121

L Labeling, facilitating executive function in childhood, 564–565 Lactose, in breast milk, 629–630 LAD (language acquisition device), 332

Lamination, in prefrontal cortex development, 219–221 Language(s). See also Speech entries deficits of, in fetal alcohol syndrome, 648 multiple. See also Bilingualism simultaneous acquisition of, 326 outcome of, initiating-joint-attention skills related to, 821 spoken, vs. American Sign Language, 444. See also American Sign Language stress patterns of words in, 119–121 visual plasticity in, 444–446 Language acquisition distributional information and linguistic categories in, interaction of, 329–330 evolutionary accounts of, 326 in infants, NIRS studies of, 326–328 mechanisms of, 325–333 nonlinear change in, 374–375 perceptual primitives in, 330–332 productivity in, 326 statistics and prosodic structures in, interaction of, 328–329 theories of, 325–326 Williams syndrome and, 694–695 Language acquisition device (LAD), 332 Language impairment, specific, 741–742 diagnosis of, 741 dyslexia with, 745–746 Larmor relationship, in diffusion tensor imaging, 301 Latent memory traces, vs. active memory representations, 555 Lateral geniculate nucleus (LGN), in visual cortex, 131, 133, 139 Learning and adult neurogenesis, 55–57 blockade of, 58 parallel changes in, 57–58 attenuated aversion, and attachment, 792–795 brain correlates of, 109–111 disabilities in, dyslexia and, 739, 740. See also Dyslexia fetal and attachment, 788 transition of, 788–789 habit, 723–726 in obsessive compulsive disorder, 725–726 PCL assessment of, 724 in Tourette’s syndrome, 725 impaired, in fetal alcohol syndrome, 646–647 lexicon, 107–109 mechanisms of, 373–374 olfactory preference and attachment, 790–792 end sensitive period in, 792 role of neurotransmitters in, 791–792 perceptually cued, CSTC circuits in, 729–730 postnatal, and attachment, 789–790

index

913

Learning (continued) reward, in executive function in childhood, 558 role of sleep in, 813 Learning algorithms, in neural network models, 370–371 Learning environment, retrieval cues in, infants’ memory and, 545–546 Left hemisphere in face processing, 442 in forward and backward speech, 118 injury to, spatial functioning after, 530–531 visuospatial processing in, 528–529 Lewy bodies, dementia with, 607 vs. Alzheimer’s disease, 613–614 Lexical form development of, in Williams syndrome, multiple factors contributing to, 696 identification of, 119–121 phonological familiarity of words in, 121 phonotactic knowledge in, 121 stress patterns of words in, 119–121 Lexicon, auditory, reference and learning of, 107–109 LGN (lateral geniculate nucleus), in visual cortex, 131, 133, 139 LHPA system. See Limbic-hypothalamicpituitary-adrenocortical (LHPA) system Limbic-hypothalamic-pituitaryadrenocortical (LHPA) system, 63–77 anatomy and physiology of, 63, 65 in animal models, 70–71 function of, individual differences in, 74–76 hyperreactivity and hyporeactivity of, 74 measurement of, in developmental research, 65–67 regulation of, social experience and, 71–74 stress responses of, 63, 65 influence of cortisol and corticotropinreleasing hormone on, 67–69 regions associated with, 64 LIMK-1 gene, deletion of, in Williams syndrome, 692 Linguistic theory, in language, 325 12-Lipoxygenase (12-lox), in rodent adolescent brain, 857–859 Lissencephaly, 45 Local motion, sensitivity to, after treatment of congenital cataract, 419 Localization auditory in infants, 101 recalibration of, 101–102 spatial, in dorsal visual pathway, 523–525 Locomotor development, nervous system maturation in, 151 Long-chain polyunsaturated fatty acids, in breast milk, 629, 630

914

index

Long-term memory aging and, 597–600 episodic deficits in, 598–599 episodic retrieval in, 599–600 Low phenylalanine diet, for phenylketonuria, 683

M MAA (minimum audible angle), in auditory localization in infants, 101 Macrocephaly, benign, definition of, 703 Macronutrients, 623. See also specific macronutrient specific, effects of, 629–630 Magnetic resonance imaging (MRI) of brain after traumatic injury, 404 in autistic children, 703 in children born preterm, 401 functional. See Functional magnetic resonance imaging (fMRI) of myelination, 237 vs. MRS, 337 Magnetic resonance spectroscopy (MRS) of brain, 337–347 in children and adolescents, 341, 343 developmental disorder profiles in, 343–346 in fetuses, 339–340 limitations of, 346–347 in neonates, 340–341, 342, 343 multivoxel mode of, 338 proton, 338 single-voxel mode of, 338 spectral quantitative, 339 using MR sensitive nuclei, 337 vs. MRI, 337 Malnutrition, protein-energy and brain development, 625, 626, 627–630 postnatal, effects of, 628–629 prenatal, effects of, 625, 627–628 risk of, weaning and, 628 Maltreatment, childhood behavioral response and regulation in, 876–877 cognitive processing of emotion signals in, neural mechanisms involved in, 870–874 consequences of, on LHPA system, 72–73 developmental outcomes associated with, 869–870 emotion elicitation in, 874–875 emotion regulatory processes in, 875–876 MAPs (microtubule-associated proteins), in regulation of microtubule organization, 5–6 Material bias, in learning mechanisms, 373 Maternal attachment, in next generation, 797–798 Maternal behavior, postpartum, role of oxytocin in, 798 Maternal separation responses, 795–797 Maternal voice, newborn preference for, 98

Maturation hypothesis, differential, in visual development, 139–141 MCI. See Mild cognitive impairment (MCI) Measurement issues, in executive function in childhood, 566–567 Mechanical folding hypothesis, in brain development, implication of, 42–43 Medial temporal lobe age-related changes in, 592, 599 damage to, 547 in declarative memory, 546–547 Memory in Alzheimer’s patients, neuropsychological tests of, 609–611 bird’s own song during anesthesia and sleep, 459–460 response to, 456 vs. tutor song memory, 457 declarative. See Declarative memory development of interneurons role in, 203 nonprimate models in, 499–504 vs. human memory, 503–504 impaired, in fetal alcohol syndrome, 646–647 long-term aging and, 597–600 episodic deficits in, 598–599 episodic retrieval in, 599–600 procedural, 725 adult abilities in, emergence of, 500 development of, 499–501 neural circuits mediating, maturation of, 500–501 recognition adult abilities in, emergence of, 501– 502, 581 attention and, 490–494 relational, adult abilities in, emergence of, 502–503 sensory, physical location of, reassessment of, 460 working. See Working memory Memory consolidation, role of sleep in, 809 Memory-guided saccade task, in eye tracking, 264 Mental arithmetic problems, pupillary dilation studies of, in children and adolescents, 273, 276 Mental retardation, phenylketonuria and, 678 Mental rotation developmental studies of, 527–525 in spatial operations, 527 Metabolism, brain in fetal alcohol syndrome, 645 restorative, role of sleep in, 809 Microarray analysis, of adolescent rodent brain, 857, 859 Micronutrients, 623–624. See also specific micronutrient

Microtubule(s) actin interactions with, 9 formation of, 5 support and transport provided by, 5 Microtubule organization, regulation of and dynamics of cytoplasmic signaling pathways, 8–9 microtubule-associated proteins in, 5–6 Microtubule-associated proteins (MAPs), in regulation of microtubule organization, 5–6 Middle temporal gyrus, activation in, deaf vs. hearing subjects and, 440 Mild cognitive impairment (MCI) in Alzheimer’s disease, 608–609 markers of, 609 in dementia, 614–615 Mineralocorticoid receptors, in cortisol binding in brain, 67–68 Minerals, 623 Minimum audible angle (MAA), in auditory localization in infants, 101 Minnesota study, of long-term neuropsychological outcome, in preterm-born children, 402 Mirror neurons in autism, 708–709 in social behavior, 171 Mismatch negativity (MMN), 251 in behavioral inhibition, 848 Mismatch response (MMR) in ERPs to auditory stimuli, 251–252 during processing of word stress patterns, 120 Monetary incentive delay (MID) task, in behavioral inhibition studies, 842 Monoamine(s), in brain development balance of receptor signaling and, 91–92 effects of, 83–92 modulatory influence of, 88–91 neuropharmacology of, 85–88 Monoamine oxidase A (MAOA) gene, in maltreated children, 876–877 Monoaminergic afferents, prefrontal, 222 Monocular deprivation congenital cataracts and, 416–417, 420– 421. See also Cataract(s), congenital restrictions after, 418 Mood, negative, sleep deprivation and, 813 Mood-valent theories, of negative cognition, depression and, 775–776 Mother-infant interaction patterns, of attachment, 797–798 Motion perception, sensitivity to, after treatment of congenital cataracts global, 420–421 local, 419 Motor activity linked to behavioral and mental health conditions, 156 overall level of, in infants, 151 Motor behavior, neural substrates of, 151–155 Motor dysfunction, in fetal alcohol syndrome, 648

Motor loop, of brain, in neural song system, 457 Motor response inhibition, 585–586 Motor system components, pathways, and functions of, 153 development of, 147–157 level of activity in, 149–151 motor-cognitive relationships in, 155–156 motor-emotional relationships in, 156 neural substrates in, 151–155 reaching in, 149–150 relevance of movement in, 147–148 theoretical approaches to, 148–149 therapeutic interventions in, 156–157 walking in, 150–151 Motor-cognitive relationships, in motor development, 155–156 Motor-emotional relationships, in motor development, 156 Movement in motor development, 149–151 relevance of, in developing human, 147–148 MRI. See Magnetic resonance imaging (MRI) MRS. See Magnetic resonance spectroscopy (MRS) Multifinality, in LHPA function, 74 Multimodal stimuli, ERP response to, in cognitive development, 255–256 Musty odor, associated with phenylketonuria, 677 Myelination age and, 24 of axons aberrant, disrupted dopamine synthesis and, 683 in central nervous system, 544 in brain development, 24–25 Flechsig’s map of, 237, 238 functional consequences of, 239–241 magnetic resonance imaging of, 237 in prefrontal cortex, during childhood, 553 relationship between cognitive function and, 240 Myosin(s), 8

N N170 adult, 513 infant, 513, 515 N170/N290, in ERP responses to visual stimuli, 253 N290, infant, 513 N400 index, of lexical-semantic processes, 121–123 NAA metabolite, concentration of in children and adolescents, 343 in fetuses, 339–340 in neonates, 340–341

NAA/Cho ratio in autism, 344 in children and adolescents, 341, 343 in developmental delays, 344 NAA/Cr ratio in attention-deficit/hyperactivity disorder, 345–346 in autism, 344 in children and adolescents, 341 in developmental delays, 344 n-back task, in assessment of working memory, 594–595 Near-infrared spectroscopy (NIRS) study of brain activity, 118 of language dispositions in infants, 326–328 Negative central (Nc) component in attention, 490–491 of ERP research, 254 Negative reactivity, in temperament, EEG asymmetry as maker of, 843 Negative slow wave activity, in event-related potential, 254 Neglect, childhood, 870. See also Maltreatment, childhood in residential nurseries and orphanages, 873 Neocortex Cajal-Retzius cells in, 197, 199 circuitry elements in, postnatal development of, 225–227 histogenesis of, 83–85 Net input mathematical expression of, 369 in neural network models, 369–370 Neural bases of attenuated aversion learning, 792–795 of developmental processes, techniques for studying, 175–178 of olfactory preference learning, 790–792 Neural circuits compensation-related utilization of, hypothesis involving, 594 engaged by threats, 759–761 involved in temperament regulation, 846–848 maturation of mediating declarative memory, 503 mediating procedural memory, 500–501 Neural groove, 39 Neural network models benefits of, 367–368 challenges to, 375–376 of change mechanisms, 372–375 age-of-acquisition effects in, 374–375 learning about words and semantic categories in, 373–374 of cognitive development, 367–379 contributions of, 371–375 challenges to, 377–378 control in, 367 critical elements of, 368–371 challenges to, 376–377 explicitness of, 368 learning algorithms in, 370–371

index

915

Neural network models (continued) net input and activation factors in, 369–370 of origins of knowledge, 371–372 object continuity in, 371–372 object permanence in, 372 understanding behavior from, 367–368 units and weights in, 368–369 Neural plate, 39 Neural song system, 453–458. See also Birdsong anterior forebrain pathway in, 458 motor loop in, 457 sensorimotor integration loop in, 456–457 Neural specializations, in auditory system, 100–104 Neural substrates of face perception, 442 of motor behavior, 151–155 Neural system, reorganization of, during adolescence, 772–773 Neural tube, formation of, 39 neuronal migration in, 39–40 Neurite(s), initiation and growth of branching in, 13 growth cone elongation in, 10–12 growth cone organization in, 10 growth cone turning in, 12 mechanism for, 10–13 Neurobehavioral profiles, cocaine-related human model of, 664–666 preclinical model of, 661–663 teratologic effects of, 653–654 Neurobehavioral systems, developing, early adversity associated with, 870–877 Neurochemical maturation, hippocampal, in preterm infants, 203, 207–208 Neurochemistry studies of adolescent brain, 856–857 of arousal aspects of attention, 480 of autism, 704 of prenatally cocaine-exposed children and infants, 664 Neurocognitive development in autism, 701–712 in face processing, 509–517. See also Face processing in prenatally cocaine-exposed children, 664 Neurocognitive functioning, in children, sleep patterns and, 810–811 Neurodevelopment changes in, associated with preterm birth, 400–403 of social cognition, 161–178. See also Social cognition Neurodevelopmental processes, disruption of, after traumatic brain injury, 403–404 Neurogenesis adult, 51–59 discovery of, 51–52 methodological advances leading to, 52 experience regulating, 54–57 functional significance of, 57–58

916

index

in hippocampus, 51–59 hormones regulating, 52–54 learning and, 55–57 blockade of, 58 parallel changes in, 57–58 stress affecting, 54–55 changes in, early brain injury and, 390–391 cortical, 83–85 in perirhinal cortex, 503 Neurogenetic cellular events, in prefrontal cortex development, 216, 218 Neuroimaging studies, of fetal alcohol syndrome, 644–646 basal ganglia in, 645 brain metabolism in, 645 cerebellum in, 644 corpus callosum in, 644–645 Neuromodulator(s) and recovery from early brain injury, 394 in song system, 458–459 Neuromotor system, changes of, during first year, 150 Neuron(s) actin filaments in, 6–7 changes in, during development, 385–386 dendritic arborization of, 18–19 development of, axon and dendrite formation in, 5–20 differentiation of, in brain development, 83 GABA, in prefrontal cortex, 214 hippocampal, excitatory and inhibitory, 200–203 migration of, 13 into cortex, 40 in neural tube formation, 39–40 reduced, prenatal cocaine exposure associated with, 655 mirror in autism, 708–709 in social behavior, 171 organization of, in prefrontal cortex, 213 polarization of, initial growth of axons and dendrites and, 13–14 preplate, 84, 221 responding to social stimuli, 165 spindle-shaped (Von Economo), 169 subplate, 84, 221 supraplate, 84 tonically active, 730 Neuronal cells, formation of, in hippocampal development, 187–195 Neuronal morphogenesis dynamic properties of, 5–10 actin filaments and, 6–8 cytoplasmic signaling pathways and, 8–9 microtubules and, 5–6 in vivo regulation of, 13–15 axonal guidance in, 14–15 neuronal migration in, 13 neuronal polarization and axon/ dendrite growth in, 13–14 Neuronal phenotypes, prenatal development of, in prefrontal cortex, 221

Neuronal polarity, 85 Neuropathology, of phenylketonuria, 678–680 Neuropsychiatric disorders, 717–731. See also Obsessive compulsive disorder (OCD); Tourette’s syndrome associated with altered gyrification, 45–46 Neuropsychological outcome, long-term, associated with preterm birth, 402–403 Neuropsychological studies, of fetal alcohol syndrome, 646–648 Neuropsychology, developmental, of unipolar depressions, 771–781. See also Depression, unipolar Neuroscience affective, and pediatric anxiety, 763–766 cognitive. See Cognitive neuroscience; Sleep Neurotransmitter(s). See also specific neurotransmitter in brain development, 771–772 function of, effects of prenatal cocaine exposure on, 655–660 role of in emotion regulation, 352–353 in olfactory preference learning, 791–792 Neurulation, primary, in brain development, 771–772 Newborn(s). See also Children; Infant(s) attachment system in, 789 auditory development in, 98 brain development in, MRS imaging of, 340–341, 342, 343 brain maturation in, 23–24 frontal circuitry in, 229–230 hippocampal formation in, Cajal-Retzius cells in, 200 imitation of facial expressions in, 830 LHPA system in, 71 movement in, 147–148. See also Motor system, development of perception, attention, and learning abilities of, behavioral studies of, 326–328 phenylketonuria screening in, 677 preterm. See Preterm birth vision in, limitations of, 415. See also Visual entries Nicotine acute effects of, 861–862 and adolescent rodent brain, 860–861 influence of, on adolescent brain, 859–862 mechanism of action of, 859–860 N-methyl-D-aspartate receptors (NMDARs), role of, in song learning, 459 Noise-vocoded speech stimuli, 109 Non–Cajal-Retzius cells, reelin-positive, in hippocampal formation, 202 Non–fear-potentiated startle, 841 Nonnutritive sucking, in neonates, 326–327 Nonprimate models, of memory development, 499–504

Non–rapid eye movement (NREM) sleep, 807. See also Sleep effects of, on brain development, 810 Nonverbal learning, impaired, in fetal alcohol syndrome, 646 Noradrenaline, and recovery from early brain injury, 394 Noradrenergic system in arousal aspect of attention, 480 effect of prenatal cocaine exposure on, 660 Norepinephrine developmental neuropharmacology of, 88 role of, in olfactory preference learning, 791–792 Novel stimulus, and attention, 489–490 Novelty preference expression of, hippocampus in, 542–543 in memory testing, 542 NREM (non–rapid eye movement) sleep, 807. See also Sleep effects of, on brain development, 810 Nucleotides, in breast milk, 630 Nucleus basalis of Meynert, 222 Nutrients, in cognitive development categories of, 623–624 deprivation and subsequent repletion of, role of timing in, 635–636 selected, effect of, 625–635 Nutrition, in cognitive development, 623–637 future directions of, 636 role of, 624–625

O Object continuity, neural network model of, 371–372 Object permanence, neural network model of, 372 Object processing, dorsal and ventral visual pathways in computational model of, 474–475 event-related oscillations and, 471 Observer-based procedure, in auditory development, 99 Obsessive compulsive disorder (OCD) age of onset of, 718–719 components of, 717 habit-learning systems in, 725–726 self-regulatory systems in, 722–723 tics in, 719 with Tourette’s syndrome, 717–718 Ocular dominance (OD) columns, development of, 133 Oculomotor suppression task, impairments of, obsessive-compulsive disorder and, 722 Odor, musty, associated with phenylketonuria, 677 Olfactory bulb, neonatal, norepinephrine input in, 791–792 Olfactory preference learning end sensitive period in, 792 neural basis of, 790–792 role of neurotransmitters in, 791–792

Oligosaccharides, in breast milk, 629–630 Optics, of eye, 129–130. See also Visual entries Orbitofrontal cortex in approach-avoidance decisions, 557 in decision making, 883–887 development of, 554 neuroanatomy of, 168 role of in behavior mediation, 877 in social behavior, 168 Oregon Adolescent Depression Project (OADP), 775 Orphanages, preschool children reared in, profound social neglect of, 873 Ovarian steroids, regulating adult neurogenesis, 53–54 Oxytocin, role of, in postpartum maternal behavior, 798

P P1, in ERP responses to visual stimuli, 252–253 P1-N1-P2 complex, in ERP response, to auditory stimuli, 249–250 P400 in ERP responses to visual stimuli, 253 infant, 513 Pachygyria, 45 PAH gene mutations, 678, 683, 684 Pallium, involved in song learning and production, 454–455 Parvalbumin interneurons, 203, 205, 206 Parvalbumin-containing axo-somatic inhibitory cell(s), 203 PDDNOS (pervasive developmental disorder not otherwise specified), 701 P-domain, of growth cone, 10 PEM. See Protein-energy malnutrition (PEM) Performance monitoring development of, 887–892 error detection in, 887–888 of executive function in childhood, 558–560 feedback processing in, 888–892 cardiac concomitants of, 889–890 developmental change in, 890–892 future directions of, 892–893 measurement of, error-related negativity in, 558 Perinatal transition, in attachment, 788–789 Peripheral vision, after treatment of congenital cataracts, 417–418 Peripheral visual field (PVF), in deaf and hearing subjects, 440 Perirhinal cortex, in declarative memory, 503 Personality changes in, orbitofrontal cortex lesions causing, 168 genetic influence on, 353, 844 Personality profile, in Williams syndrome, 693 Pervasive developmental disorder(s), eye tracking studies in, 280–284

Pervasive developmental disorder not otherwise specified (PDDNOS), 701 Phe tolerance test, 677 Phenotype, in Williams syndrome, 693 Phenylalanine, diet low in, for phenylketonuria, 683 Phenylalanine hydroxylase deficiency, tetrahydrobiopterin-responsive form of, 684–685 Phenylketonuria (PKU), 677–686 atypical, 678 blood Phe levels in, 678 classical, 678 early-treated blood Phe levels in, 682, 683 dopamine/tyrosine theory in, 683 executive function impairment in, 681 low phenylalanine diet in, 683 myelin/axonal theory in, 683 neurocognitive outcomes in, metaanalytic approach to, 680–683 genetic studies of, 684–685 musty odor associated with, 677 neonatal screening for, 677 neuropathology of, 678–680 PAH mutations in, 678, 683, 684 prefrontal dysfunction in, 679 tetrahydrobiopterin for, 684 white matter abnormalities in, 679–680 Phenylpyruvic acid, identification of, 677 Phobia, specific, 756 Phonetic reorganization, 104–107 distributional modifications in, 105 distributional sensitivity in, 105–106 in second language, 106–107 universal inventories in, 104 Phonological processing, developmental changes in gray matter and, 33–34 Phonological representation, in dyslexia, 742–743 Phonological skills, poorly developed, in dyslexia, 740 Phonotactic knowledge, and lexical form, 121 Photon absorption, in visual acuity, 129–130 Photoreceptors, development of, 128–130 Photosensitive pigment, developmental increase in, 128 Phrase boundaries intonational, 118–119 syntactic, 118 Physics, of MRI, 312 PKU. See Phenylketonuria (PKU) Plasticity of auditory system brain correlates of, 109–111 in humans, 103–104 in nonhumans, 102–103 cross-modal in development, 439–449 in speech perception, after cochlear implantation, 446–448 early brain injury and, 385–396, 399–409 of speech, in animal model, 453–461. See also Birdsong

index

917

Plasticity (continued) synaptic, mechanisms of, in adolescent brain, 862, 863 of visual system, 415–431 in blind person, 427 congenital cataracts and, 426–427 Detroit principle of, 429 developmental changes in, 427–428 high-level vision and, 420–426 in language, 444–446 low-level vision and, 416–420 Policy domains, relevance of sleep to, 813 Polymicrogyria, 45 Polyunsaturated fatty acids, long-chain, in breast milk, 629, 630 Positive slow wave activity, in event-related potential, 254 Positron emission tomography (PET) study of neural activation associated with reading, 743 of obsessive-compulsive disorder, 722 Posterior superior temporal sulcus, sine-wave speech in, 110 Postmortem studies, of brain, synaptic modification and myelination in, 23–25 Postnatal growth deficiency, in fetal alcohol syndrome, 643 Postnatal learning, and attachment, 789–790 Posttraumatic stress disorder (PTSD), 756, 762–763 in abused children, 870 Posture, in motor development, 149–151 Predictive saccade task, in eye tracking, 264 Prefrontal cortex afferent pathways in corticocortical connections in, 215–216, 224 extrathalamic, 214–215, 222 long associative connections in, 224–225 sequential development of, 221–225 thalamocortical, 215, 222–224 architecture and connectivity of, 213–216 basal forebrain–prefrontal cholinergic system of, 215 behavioral inhibition and, 585–586 corticocortical connections in, 215–216, 224 long associative, 224–225 corticohypothalamic connections in, 215, 217 cytoarchitectonics of, 213 laminar development of, 216, 218 delineation of, 213 development of in adolescence, 577–578 afferent pathways in, 221–225 in children, 553–554 early, 216–225 laminar events in, 219–221 neurogenetic cellular events in, 216, 218 structural, 213–232 dopaminergic system of, 214–215

918

index

dorsal, in obsessive-compulsive disorder, 722 dorsolateral, 575–576 development of, 582–584 vs. ventromedial development, 586 emotion processing in, 871 in feedback processing, 892–893 functions of, 579–580 in performance monitoring, 887. See also Performance monitoring reprocessing of rules in, 560–561 selective attention in, 561 functions of by different prefrontal areas, 582–586 impaired, in phenylketonuria, 679 immature circuitry elements in, development of, 225–227 dendritic spine development in, 225, 227 development of intrinsic circuitry in, 227 interneuron development in, 225, 227, 228 neuronal differentiation in, 225, 226 injury to, spontaneous tissue regeneration after, 390–391 intrinsic circuitry in, 214 endogenous and sensory-driven, 229, 230 neonatal, 229–230 overproduction of elements of, in infancy and childhood, 230–231 preterm infant (endogenous), 229 prolonged plasticity and reorganization of, in adolescence and postadolescence, 231 neuronal organization of, 213 neuronal phenotypes in, prenatal development of, 221 planning ability of, 582–583 process vs. content of, 581–582 recognition memory and, 581 rostrolateral, task set selection in, 562 rule of representation in, 556, 557 selected tasks and, 583–584 self-ordered search tasks and, 582 self-organized behavior and, 583 span tasks and, 581–582 spatial delayed-response tasks and, 582 structure and connectivity of, 575–577 subregions of, 553 thalamocortical pathways in, 215 development of, 222–224 ventrolateral, 575–576 emotion processing in, 871 reprocessing of rules in, 560–561 ventromedial, 576 development of, 584–585 vs. dorsolateral development, 586 functions of, 579–580 Prefrontal processes, in temperament regulation, 845–846 Preplate neurons, 84, 221 Preschool children, orphanage-reared, profound social neglect in, 873

Preterm birth cell proliferation, migration, and neurochemical maturation with, 203, 207–208 long-term neuropsychological outcome associated with, 402–403 neurodevelopmental changes associated with, 400–403 neuroimaging studies in, 401–402 prefrontal endogenous and sensory-driven circuitry in, 229 Proactive interference (PI), in working memory, 595 Probabilistic classification learning (PCL) tasks, in assessment of habit learning, 724 Procedural memory adult abilities in, emergence of, 500 development of, 499–501 neural circuits mediating, maturation of, 500–501 Process C, in sleep regulation, 809 Process S, in sleep regulation, 808–809 Productivity, in language acquisition, 326 Progenitor cells, development of, in normal brain, 39–40, 83, 385 Programmed cell death, in brain development, 772 Proliferation process, in brain development, 83 Proliferative zones, neurons arising from, 84–85 Prosaccade response times, age-related changes in, 269, 272 Prosaccade tasks neurological bases of, 265, 266 warning stimulus on, age-related changes in CNV to, 269 Prosopagnosia acquired and congenital, face processing impairment and, 516 lesions resulting in, 707–708 Protein(s). See also specific protein actin-binding, in regulation of actin filament organization, 7–8 in cognitive development, 623 microtubule-associated, in regulation of microtubule organization, 5–6 Protein-energy malnutrition (PEM) and brain development, 625, 626, 627–630 postnatal, effects of, 628–629 prenatal, effects of, 625, 627–628 risk of, weaning and, 628 Protein-energy status, in brain development, 625, 627–630 Psychoactive drug therapy, for early brain injury, 395 Psychobiological models, of temperament, 839 Psychopathology developmental research areas in, 756–758 theoretical perspective of, 757 developmental course of, 358–360

PTSD. See Posttraumatic stress disorder (PTSD) Pupillary dilation phasic changes in, 267 task-specific, 268–269 Pupillary dilation tasks, in eye tracking, 264, 267–269 in children and adolescents, 273, 276 Purkinje cells, 500 Pursuit studies, in eye tracking, 264, 266 in children and adolescents, 269, 273 Pursuit-rotor task, in assessment of motorskill habit learning, 725 PVF (peripheral visual field), in deaf and hearing subjects, 440 Pyramidal cells hippocampal, 187, 200, 202–203 development of, 187 in prefrontal cortex, 215

Q Quantitative trait loci (QTLs), associated with temperamental variability, 843 Quinoid dihydropteridine reductase (QDR), in rodent adolescent brain, 858, 859

R Radial diffusivity, in DTI data processing, 302 Rapid eye movement (REM) sleep, 807–808. See also Sleep effects of, on brain development, 810 Rapid odor preference learning, termination of, 792 Reaching, development of, 149–150 Reading, neural mechanisms for, development of, 743 Reading acquisition, skills necessary for, 741, 746 Reading intervention, 746–747 neurobiological basis for, 747–749 Reading skills, genes influencing, 744–745 Reading-related deficits, in dyslexia, 740, 741 Reciprocal connections, in prefrontal cortex, 576 Recognition memory adult abilities in, emergence of, 501–502, 581 attention and, 490–494 Recollection, retrieval process of, differential effects of aging on, 600 Reelin, secretion of, in hippocampal formation, 196 Reelin gene mutation, 196 Referential mapping errors, in gaze following/joint attention, 820 Reflex, acoustic startle, 841 Reflexive saccadic eye movement. See also Eye movements effect of attention in, 487 in tracking visual stimuli, 485

Reflexive stepping phase, of walking, 150 Refocusing pulse, in diffusion tensor imaging, 302 Regions of interest (ROI), in DTI data processing, 303 Relational memory, adult abilities in, emergence of, 502–503 RELN gene, in autism, 702 REM (rapid eye movement) sleep, 807–808. See also Sleep effects of, on brain development, 810 Repetitious action, in development, 148 Repetitive behavior, in autism, 711 Residential nurseries, social and emotional neglect in, 873 Resilience, cognitive, 780 Retention of information age-related changes in, 545 in infancy, declarative memory and, 544–545 Retina, development of, 128–130 cones in, 129–130 rods in, 128–129 Retrieval cues age-related changes in, 546 in infancy, declarative memory and, 545–547 Retrieval processing, differential effects of aging on, 600 Reward, hyposensitivity to, in unipolar depression, 777 Reward learning, in executive function in childhood, 558 Rey Osterrieth Complex Figure (ROCF), in evaluation of spatial planning, 530–531 Rho family, of GTPase proteins, 9 Rhodopsin, developmental increase in, 128 Right hemisphere activation of in face processing, 442 in perception of American Sign Language, 444–445 injury to, spatial functioning after, 530–531 visuospatial processing in, 528–529 RMSE (root-mean-square error), in eye movement performance, 266 ROCF (Rey Osterrieth Complex Figure), in evaluation of spatial planning, 530–531 Rod(s) in retina, 128–129 rhodopsin in, 128 Rod outer segment lengths, in infants vs. adults, 128–129 Rodent brain, adolescent microarray analysis of, 857, 859 nicotine affecting gene expression in, 862 nicotine and, 860–861 Root-mean-square error (RMSE), in eye movement performance, 266 Running, as positive stressor, in adult neurogenesis, 54–55

S Saccade(s) externally guided, 263 in eye tracking, 263, 264, 265 internally guided, 265 memory-guided, impaired, in schizophrenia, 280 neural correlates of, in children, 269 in normative development, 270–271 physiologic characteristics of, 487 Saccade sequence, two-dimensional, in response to visual targets, 470–471 Saccade-related event-related potentials, 256 Saccadic response times, 263, 265 Safety considerations, in fMRI studies, 314–315 Salivary cortisol, collection of, 65 Scalp-recorded event-related potentials, as psychophysiological measure of attention, 483–484 Scene perception tasks, in eye tracking, 264, 266–267 in children and adolescents, 273 Schizophrenia eye tracking studies in, 277, 278– 279, 280 fMRI of, 825 gyrification index in, 45–46 MRS imaging of, 346 onset of, critical stressor in, 359 School performance, in prenatally cocaineexposed children, 665 Selective attention, 870 in executive function in childhood, 561–562 Selenium, in brain development, 627, 634 Self-ordered search task, 582, 583 Self-organizing facilities, in gaze following/ joint attention, 820 Self-organizing learning algorithms, 370 Self-regulation, 719–723 in obsessive-compulsive disorder, 722–723 of temperament, 847 in Tourette’s syndrome, 720–722 Semantic categories, acquisition of, 373–374 Semantic processes at sentence level, 122–123 at word level, 121–122 Sensorimotor integration loop, of brain, in neural song system, 456–457 Sensorimotor phase, of motor development, 148 Sensorimotor system, in birdsong, 458–461. See also Birdsong behavior and role of sleep in, 458 development of synapses in, 459 and implications for human speech, 460 neuromodulators in, 458–459 sensory development in, 459–460 vocal development in, genetic aspects of, 459 Sensory memory, physical location of, reassessment of, 460 Sentences, semantic processing of, 122–123

index

919

Sentential prosody definition of, 118 processing of, 118–119 Separation anxiety disorder, 755. See also Anxiety disorder(s) Serial reaction time (SRT) task, in assessment of obsessive-compulsive disorder, 725 Serotonin (5-HT). See also 5-HT gene; 5HTT gene; 5-HTTLPR S allele developmental neuropharmacology of, 87–88 effect of prenatal cocaine exposure on, 658–659 influence of, on emotion regulation, 352–353 levels of, in autism spectrum disorders, 704 Serotoninergic system, in arousal aspect of attention, 480 Severe mood dysregulation syndrome, 774 Sex hormone(s), and recovery from early brain injury, 394–395 Shape bias, in learning mechanisms, 373 Sign Language, American, 444–446 visuospatial nature of, 444–445 vs. spoken language, 444 Signal artifact, sources of, in fMRI studies, 314 Simon spatial incompatibility task performance of, Tourette’s syndrome and, 720, 721 vs. Stroop task, 719–720 Sine-wave speech stimuli, 109, 110 Single nucleotide polymorphism (SNP), in BDNF gene, 75 Single-photon emission computed tomography (SPECT) study, of obsessive-compulsive disorder, 722 Sleep bird’s own song memory during, 459–460 cognitive neuroscience of, 807 differential effects of, on brain development, 810 and emotion regulation, 813–814 functions of, 809 homeostatic component of, 808–809 NREM, 807 organization and regulation of, during human development, 807–809 patterns of and developmental outcome, 811–812 and pediatric neurocognitive functioning, 810–811 prolonged deprivation of, effects of, 54, 55, 813–814 REM, 807–808 role of in learning and memory, 813 in song learning, 458 slow-wave, 807, 808 synaptic homeostasis hypothesis of, 812–813 Sleep-wake states, control of, 795 SLI. See Specific language impairment (SLI)

920

index

Slow waves, in event-related potential, negative and positive, 254 Slow-wave activity (SWA), in sleep, 809 homeostatic regulation of, 812 Slow-wave sleep (SWS), 807, 808. See also Sleep Smoking, in adolescence, 859–862 acute effects of nicotine and, 861–862 mechanism of nicotine action and, 859–860 Smooth pursuit eye movement. See also Eye movements effect of attention on, 486–487 gain in, definition of, 266 impaired, in schizophrenia, 277 in normative development, 274–275 substrates and connections in, 267 in tracking visual stimuli, 485 Social adversity developmental implications of, 74 resilience resulting from, 780 Social anxiety disorder, 755. See also Anxiety disorder(s) Social behavior definition of, 162 evolutionary perspective on, 164 input vs. output of, 825 of neglected children, 870 neurobiology of, 161–172 selective changes in, developmental disorders and, 165 trajectory characterizing, 164–165 Social brain autism and, 706–710 components of, 163 development of, 172–178 animal model in, 173–175 neural bases in, 175–178 dorsal, joint attention/social cognition and, 825–827 lesion research of, 177–178 mirror neuron system in, 171 neural components of, 162 neuroanatomical studies of, 175–176 neurophysiology and imaging of, 176 plausibility of, 163–165 putative structures of, 165–172 role of amygdala in, 165–168 role of anterior cingulate cortex in, 168–169 role of fusiform gyrus in, 170–171 role of orbitofrontal cortex in, 168 role of temporal cortex in, 169–170 ventral, gaze following/joint attention and, 823–825 Social cognition and dorsal social brain, 825–827 elements of, 162 gaze following/joint attention reflecting, 820–821 impaired, in Williams syndrome, 693 neurobiology of, 161–172 unipolar depression and, 774–775 Social context, affecting speech, 460 Social deficits, in autism, 165

Social development, animal model of, 173–175 Social dominance impact of, on adult neurogenesis, 55, 56 orbitofrontal cortex damage affecting, 168 Social experience, in regulation of LHPA system, 71–74 Social knowledge, vs. other knowledge domains, 164 Social motivation, in autism, 709–710 Social partner, gaze of, 819 Social perception, in autism, 706–710 Social policy domains, relevance of sleep to, 813 Social processing components of, 162 model of, 162, 163 neural regions implicated in, 165, 166 Social skills, slower gains of, in preterm children, 403 Social stimuli amygdala response to, 166–167 neurons responding to, 165 Socially relevant information, orbitofrontal cortex processing of, 168 Socioeconomic status, impact of, on early brain injury recovery, 407 Socioemotional correlates, of executive function in childhood, 563–564 Socioemotional deprivation, of Romanian orphans, neuroimaging study of, 408 Somatosensory information, in mother-infant interactions, 790 Somatostatin, mother-infant interactions and, 795 Songbirds, behavior of, 453. See also Birdsong Sound thresholds, in infants, assessment of, 99 Spatial analytical processing complexity and sophistication of, changes in, 529 disorders of, focal brain injury associated with, 528, 530 neural system associated with, development of, 530 Spatial attention, development of, 525–527 Spatial delayed-response task, 582 Spatial localization, in dorsal visual pathway, 523–525 Spatial location processing, in dorsal visual pathway during infancy, 524–525 in older children, 525 Spatial location task, 523 Spatial planning, ROCF in evaluation of, 530–531 Spatial planning ability, 582–583 Spatial processes, associated with dorsal visual pathways, 523–528 Spatial span test, 581–582 Spatial working memory, neuroimaging studies of, 525 Specific language impairment (SLI), 741–742

Specific language impairment (SLI) (continued) diagnosis of, 741 dyslexia with, 745–746 Speech. See also Language(s) entries birdsong as model of, 453. See also Birdsong forward and backward, left-hemispheric asymmetry in, 118 human, implications of birdsong in, 460 impaired, in fetal alcohol syndrome, 648 infant-directed, 98 maternal input in, 98 noise-vocoded, 109 plasticity of, in animal model, 453–461. See also Birdsong sine-wave, 109, 110 Speech fusion, auditory/visual, sensitive period for, in young recipients of cochlear implants, 444–448 Speech perception cross-modal plasticity in, after cochlear implantation, 446–448 talker-specific effects in, 109 Speech processing, indexical factors in, 108–109 Speed-of-proccessing deficits, in phenylketonuria, 680 Spelling ability, poor, in dyslexic children, 741 S-shaped curve, of activation function, 369–370 Startle, fear-potentiated vs. non–fearpotentiated, 841 Startle reflex, acoustic, 841 Statelike effects, of sleep patterns, 810–811 Static phase, of walking, 150 Stem cells, development of, in normal brain, 385 Steroids, regulating adult neurogenesis adrenal, 53 ovarian, 53–54 STG. See Superior temporal gyrus (STG) Stimulation theory, of gaze following/joint attention, 832 Stop-signal task, 585 “Stranger” anxiety, 789 Stress, effect of, on adult neurogenesis, 54–55 Stress hyporesponsive period in attachment-learning period, 794 in humans and animals, 71–72 postnatal, 69 Stress patterns, of words, 119–121 Stress responses, of LHPA system, 63, 65 influence of cortisol and corticotropinreleasing hormone on, 67–69 regions associated with, 64 Striatal activation, enhanced, in behavioral inhibition studies, 842–843 Stroop task impairments of, obsessive-compulsive disorder and, 722 limitations of, 719–720 performance of, Tourette’s syndrome and, 720, 721

in study of self-regulation, 719 STS. See Superior temporal sulcus (STS) Subcortical plate, development of connectivity between cortical structures and, 40–41 Subcortical system, in face processing in adults, 509 in infants, 511–512 Subplate neurons, 84, 221 Subthalamic nucleus (STN), 726, 730 Sulcus alterations in, 45 superior temporal. See Superior temporal sulcus (STS) Superior temporal gyrus (STG) in face processing, 511, 514–515 in social perception, 823 Superior temporal sulcus (STS) activation in, deaf vs. hearing subjects and, 440 in face processing, 511, 514–515 autism spectrum disorders and, 708 in social perception, 823, 824 Supplementary nutritional measures, cognitive development and, 636 Supraplate neurons, 84 Supravalvular aortic stenosis, in Williams syndrome, 691, 692 SWA (slow-wave activity), in sleep, 809 homeostatic regulation of, 812 SWS (slow-wave sleep), 807, 808. See also Sleep Symbolism, facilitating executive function in childhood, 565–566 Synapse(s) in brain development, modification of, 23–24 formation of. See Synaptogenesis from Purkinje cells, 500 in song learning, development of, 459 Synaptic homeostasis hypothesis, purpura fulminans sleep, 812–813 Synaptic plasticity, mechanisms of, in adolescent brain, 862, 863 Synaptic pruning, in brain development, 23–24 Synaptic space, early brain injury and, 389 Synaptic targets, axonal patterning distribution in, 17–18 stereotypical routes to, 14–15 Synaptogenesis in brain development, 23–24, 83, 385–386 experience-dependent, 386–387 Syntactic phrase boundary, 118 Syntactic processes, in brain, 123–124

T Tactile stimulation, for early brain injury, 391–392 prenatal, 387 TANs (tonically active neurons), 730 Task set selection, in executive function in childhood, 562 Temperament behavioral inhibition in, 841–843

biology of, 839–848 integrative approach to, 839–841 model system studying, 841–843 in individual LHPA function, 74–75 modulators of, cognitive processes as, 840 negative reactivity in EEG asymmetry as maker of, 843 genetic markers of, 843–844 and pediatric anxiety, 763 psychobiological models of, 839 reactivity in, regulation of, 844–848 regulation of neural circuitry measurement involved in, 846–848 prefrontal processes involved in, 845–846 relations between inhibitory control and, 846 self-regulation of, 847 Temporal cortex functional studies of, 169–170 lesions of, 170 neuroanatomy of, 169 role of, in social behavior, 169–170 Temporal lobe, medial age-related changes in, 592, 599 damage to, 547 in declarative memory, 546–547 Temporal sulcus, posterior superior, sinewave speech in, 110 Test of Everyday Attention for Children (TEA-Ch), 779 Tetrahydrobiopterin (BH4), 684 Thalamocortical afferent pathways, prefrontal, 215 development of, 222–224 Thalamus involved in arousal aspects of attention, 480 involved in song learning and production, 454–455 Theory of mind (ToM) task in autism, 705 performance of, 822, 827 Threats neural circuits engaged by, 759–761 response to 5 HT gene and 5HTT gene in, 762 role of environment in, 757 Tics age of onset of, 718 chronic, 718 CSTC circuitry in, 726–729 in obsessive-compulsive disorder, 719 in Tourette’s syndrome, 717 Tissue regeneration, spontaneous, after prefrontal injury, 390–391 Tonically active neurons (TANs), 730 Tourette’s syndrome habit-learning systems in, 725 natural history of, 718–719 neural basis for, 718 with obsessive compulsive disorder, 717–718 self-regulatory systems in, 720–722 tics in, 717

index

921

Tower of London test, 582–583 TPs (transition probabilities), in language acquisition, 328–329 Trace elements, 623. See also specific element Training, facilitating executive function in childhood, 566 Traitlike effects, of sleep patterns, 811–812 Transition phase, of walking, 150 Transition probabilities (TPs), in language acquisition, 328–329 Transitive interference (TI) task, hippocampal activation during, 546 Trauma, brain. See Brain injury Tutor song memory, 457 Twin studies gyral patterns evaluation in, 43 traditional behavioral genetics in, 352 Twin-level theoretical framework, for dyslexia, 742 Tyrosine hydroxylase, in dopamine synthesis, 86

U UCS-CS+ relationship, threats and, 759 Ultrasonic vocalization response, to isolation, 796 Underactivation, cognitive, age-related, 600 Unipolar depression, developmental neuropsychology of, 771–781. See also Depression, unipolar Units, in neural network models, 368–369 Universal inventories, in phonetic reorganization, 104 Urbach-Wiethe disease, 167

V Valproic acid, maternal ingestion of, risk of autism with, 703 Vascular dementia, 607 vs. Alzheimer’s disease, 612–613 Vasopressin, secretion of, LHPA stress response and, 63 Velocardiofacial syndrome, schizophrenia and, 46 Ventral visual pathway. See Visual pathways, ventral Ventricular germinal layers, in hippocampal formation, cell proliferation in, 191, 193 Verbal learning, impaired, in fetal alcohol syndrome, 646 Very low birth weight, neuropsychological outcome associated with, 402 Vision high-level, 420–426 face detection in, 422, 423 face perception in, 422–426 holistic face processing in, 422, 424–425 sensitivity to global form in, 421–422 sensitivity to global motion in, 420–421 visual spatial attention in, 425

922

index

interactions between audition and, after cochlear implantation, 446– 447 low-level, 416–420 acuity in, 416–417 contrast sensitivity in, 417 peripheral, 417–418 sensitivity to local motion in, 419 summary of, 419–420 Visual acuity after treatment of congenital cataracts, 416–417 development of, 129–130 in infants and children, 415–416 in newborn vs. adult, 129, 415 Visual attention difficulty, in preterm infant, 402 Visual cascade, 128 Visual cortex of blind person, plasticity of, 427 cross-modal plasticity of, postlingual deafness and, 446 development of, 130–136 anatomical data on, 139 events in, time line illustrating, 134 ocular dominance (OD) columns in, 133 GABA neurotransmitter in, 135 lateral geniculate nucleus in, 131, 133 pathways in, 130–133 postnatal reorganization of, factors responsible for, 133 striate absence of, 136, 138 in newborn vs. 6-month-old infant, 136–137 visual processes in, 133 Visual fields central and peripheral, in deaf and hearing subjects, 440 receptive, responses of, to attention, 480–481 restricted, after treatment of congenital cataracts, 417–418 shift attention in, infant’s ability to, 526–527 Visual motion processing, effects of deafness on, 439–443 Visual paired-comparison (VPC) task, 541 in assessing recognition-memory abilities, 501–502 in infancy declarative memory and, 542–543 performance of, 546 Visual pathways dorsal anatomy of, 521–522 development of, 467–476 dissociation of, with ventral pathway, behavioral evidence of, 471–474 function of, 468 high-density ERP studies of, 469–471 mental rotation in, developmental studies of, 527–528

processing in computational model of, 474–475 event-related oscillations and, 471 spatial systems in, 468 spatial attention in, 525–527 spatial localization in, 523–525 during infancy, 524–525 in older children, 525 spatial processes associated with, 523–528 as “where” or “action” pathway, 467 information transmission in, 128–130 retinocortical, 131–133 subcortical and cortical, 130–131, 132 ventral anatomy of, 521–522 development of, 467–476 dissociation of, with dorsal pathway, behavioral evidence of, 471–474 function of, 468 high-density ERP studies of, 469–471 processing in, 468–469 computational model of, 474–475 event-related oscillations and, 471 spatial processes associated with, 528–531 as “what” or “perception” pathway, 467 Visual perception, in newborns, limitations of, 415 Visual processing dorsal and ventral routes of, 467–469 identifying presence of visual streams in, 469–475 Visual search task, effects of color and motion in, deaf vs. hearing subjects and, 440–441 Visual space processing, effects of deafness on, 443–444 Visual stimuli briefly presented, recognition in, 489–494 ERP response to, in cognitive development, 252–254 Visual system abnormal, in dyslexia, 742 auditory system and, 100 development of absolute threshold in, 128 acuity in, 129–130. See also Visual acuity Atkinson’s model of, 138 Banks and colleagues model of, 129 bottle-neck developmental theories in, 141 brain-behavior relationships in, 127–141 cortical, 130–136. See also Visual cortex cortical parcellation hypothesis in, 140–141 cortically motivated models in, 136–139 differential maturation hypothesis in, 139–141 Hansen and Fulton model of, 128–129 involved in visual tracking, 485 at neural level, 127 optical properties in, 129

Visual system (continued) retinal, 128–130 cones in, 129–130 rods in, 128–129 plasticity of, 415–431 in blind person, 427 congenital cataracts and, 426–427 developmental changes in, 427–428 high-level vision and, 420–426 in language, 444–446 low-level vision and, 416–420 Visual tracking, reflexive, 136 Visual-spatial attention after treatment of congenital cataracts, 425 impaired in fetal alcohol syndrome, 647 in schizophrenia, 280 Visuospatial nature, of American Sign Language, 444–445 Visuospatial processing, development of, 521–533 associated with dorsal stream, 523–528. See also Visual pathways, dorsal trajectories of, 531–533 associated with ventral stream, 528–531. See also Visual pathways, ventral trajectories of, 531–533 Vitamin(s), 623–624 deficiency of, 624 Vitamin A deficiency, 624 Vocal behavior, development of genetic aspects of, 459 in song learning, 458 Vocal communication, origins of, 795–797 Vocal learning and production, in song birds. See also Birdsong brain-behavior relationship in, 453–454 nuclei involved in, 454–455 Voice, mother’s, newborn preference for, 98 Voice onset time (VOT), 105

Voice quality, of talker, 108 Volumetric studies, in brain development, 25 Voluntary saccadic eye movement. See also Eye movements in tracking visual stimuli, 485 Von Economo neurons, 169 VOT (voice onset time), 105 Vowels, in linguistic theory, 329–330 Voxel-based morphometry methods, of brain development study, 25–26

W Walking, development of, 150–151 Weak central coherence (WCC), in autism, 705 Weaning, and risk of protein-energy malnutrition, 628 Weights, in neural network models, 368–369 Weschler Intelligence Scale for Children-III (WISC-III), 581 White matter abnormalities of, in phenylketonuria, 679– 680, 683 development of, 237–238 effect of traumatic brain injury on, 404 myelination of Flechsig’s map of, 237, 238 functional consequences of, 239–241 magnetic resonance imaging of, 237 volume increase in age-related, 25, 577 rate and pace of, 772 Whole-brain mapping methods, voxel-based morphometry in, 25–26 Williams syndrome, 691–697 brain size in, 691 building developmental trajectories in, 695 clinical features of, 691 genetic basis of, 692 genotype in, 691–692

gyrification index in, 46, 691 hypersociability in, 165 lexical development in, multiple factors contributing to, 696 phenotype in, 693 phenotypic outcome in face processing and, 693–694 language and, 694–695 vs. infant start state, 695 prevalence of, 692 Wisconsin Card Sort task in depression, 779 feedback processing using, 888–889 Wisconsin Card Sorting Test (WCST), 553, 556 Woodstock Johnson III Broad Reading score, 746 Words learning about, 373–374 phonemic structure of, dyslexic children and, 740–741 phonological familiarity of, 121 semantic processing of, 121–122 stress patterns of, 119–121 Working memory in adolescence, prefrontal development and, 581–582 aging and, 592–595 assessment of, n-back task in, 594–595 development of, during childhood, 240 effect of traumatic brain injury on, 406 in executive function in childhood, 561 importance of, 555 neuroimaging studies of, 593–594 role of experience in, 241–242 spatial, neuroimaging studies of, 525

Z Zinc in brain development, 626, 632–633 deficiency of, 633

index

923

Plate 1 Actin filaments and microtubules are polarized polymers. Actin filaments are polarized polymers for which the addition of ATP-actin is more likely at the barbed end than the pointed end. After hydrolysis of ATP-actin to ADP-actin, subunits dissociate at the pointed end. Microtubules are also polarized structures with GTP-tubulin dimers adding to the plus or growing end and GDP-tubulin dimers dissociating from the minus end. Microtubules also exhibit posttranslational modifications (detyrosination shown here) that correlate with the age and stability of the polymer. (From Dent and Gertler, 2003.) (See figure 1.1.)

Plate 2 The distribution of microtubules and actin filaments in developing neurons and in axonal growth cones. Microtubules (green) are densely packed with the neuronal cell bodies (S) and are bundled in the axons and branches. Actin filaments are arrayed in filament networks and bundles in the peripheral domains (P) of the growth cones and along the shafts of the axons, where small areas of actin filament dynamics may give rise to collateral branches (B). In a growth cone, the microtubules from the central bundle of the central domain (C) splay apart, and individual microtubules extend into the P domain and into filopodia (arrows). (See figure 1.2.)

Plate 3 The trajectory of growing thalamocortical and corticothalamic fibers involves multiple steps and both attractive and repulsive guidance cues. The expression of guidance molecules is related to each of these steps: Slit is a repellent that steers thalamic axons emerging from the diencephalon and in the ventral telencephalon. Ephrin-A5 is involved in sorting thalamocortical axons in the ventral telencephalon. Netrin-1 is an attractive factor for

both populations of fibers in the internal capsule. Semaphorins 3A and 3C steer cortical fibers to penetrate the intermediate zone and then turn. EphA4 in the thalamus and ephrin-A5 in the cortex are involved in the establishment of topographic connections. Th, thalamus; Hyp, hypothalamus; IC, internal capsule; GE, ganglionic eminence; Ncx, neocortex. (From Uziel et al., 2006.) (See figure 1.8.)

Plate 4 Schematic diagrams of coronal sections through the developing forebrain showing the trajectory of corticospinal (red), corticothalamic (blue), and thalamocortical (purple) axons in relation to regions that express slit-1 (blue) and slit-2 (yellow) at selected

levels. Regions depicted in green express both slit-1 and slit-2. CGE, caudal ganglionic eminence; H, hippocampus; ic, internal capsule; NCx, neocortex. (From Bagri et al., 2002.) (See figure 1.9.)

Plate 5 Mechanisms and molecules controlling retinotopic mapping in chicks and rodents. The names and/or distributions of molecules known, or potentially able, to control the dominant mechanisms at each stage are listed. The gradients represent the consensus distribution for a combination of related molecules

(i.e., ephrin-A’s), which are not listed individually owing to distinctions in the individual members expressed and the precise distributions between species. Molecules other than those listed are likely to participate. (From McLaughlin and O’Leary, 2005.) (See figure 1.10.)

Plate 6 VBM reveals the full spectrum of gray matter density loss across childhood, adolescence, and adulthood. The top panel shows the child minus adolescent statistical map for negative age effects, the bottom panel shows the same maps for adolescent minus adult. Areas in color represent clusters of gray matter density reduction observed between these age groups. These maps are three-dimensional renderings of statistical maps shown inside the transparent cortical surface rendering of one representative subject’s brain. Color coding is applied to each cluster based on its location within the representative brains. Clusters are shown in the frontal lobes (purple), parietal lobes (red), occipital lobes (yellow), temporal lobes (blue), and subcortical regions (green) (Sowell, Thompson, et al., 1999a, 1999b). (Reproduced with permission from Sowell, Thompson, and Toga, 2004.) (See figure 2.1.)

Plate 7 Cortical pattern matching methods and gray matter density. Top left: Three representative brain image data sets with the original MRI, tissue-segmented images, and surface renderings with sulcal contours shown in pink. Top right: Surface rendering of one representative subject with cutout showing tissue-segmented coronal slice and axial slice superimposed within the surface. Sulcal lines are shown where they would lie on the surface in the cutout region. Note the sample spheres over the right hemisphere inferior frontal sulcus (lower sphere) and on the middle region of the precentral sulcus (upper sphere) that illustrate varying degrees of gray matter density. In the blown-up panel, note that the upper sphere has a higher gray matter density than the lower sphere as it contains only blue pixels (gray matter) within the brain. The lower sphere also contains green pixels (white matter) that would lower the gray matter proportion within it. In actual analyses, gray matter proportion is measured within 15-mm spheres centered across every point over the cortical surface. Bottom: Sulcal anatomical delineations are defined according to color. These are the contours drawn on each individual’s surface rendering according to a reliable, written protocol (see also figure 2.5). (Reproduced with permission from Sowell, Thompson, and Toga, 2004.) (See figure 2.2.)

Plate 8 Reductions in gray matter density are occurring in the same locations as brain growth. Composite statistical maps (top) showing the correspondence in age effects for changes in brain growth (defined here as distance from center, or DFC) and changes in gray matter in the child-to-adolescent contrast (A). Shown in green is the Pearson’s R map of all positive correlation coefficients for DFC, and in blue is the probability map of all regions of significant gray matter loss (surface point significance threshold P = 0.05). In red are regions of overlap in the gray and DFC statistical maps. A similar composite map for the adolescent-to-adult age effects is also shown (B). Note the highly spatially consistent relationship between brain growth and reduction in gray matter density. The shapes of the regions of greatest age-related change for the two maps (gray matter and DFC) are nearly identical in many frontal regions in the adolescent-to-adult contrast. Very few regions of gray matter density reduction fall outside regions of increases in DFC. Shown in images in the lower part of this figure (left, right, and top

views) are the difference between Pearson’s correlation coefficients for the age effects for gray matter density and the age effects for DFC between childhood and adolescence (C) and between adolescence and adulthood (D). The color bar represents corresponding Z scores ranging from −5 to +5 for the difference between correlation coefficients for DFC and gray matter. Highlighted in red are regions of significant negative correlation between DFC and gray matter density (P = 0.05), showing that the relationship between regions of greatest gray matter density reduction are statistically the same as the regions with the greatest brain growth, particularly in the adolescent-to-adulthood years. Highlighted in white are the regions where the difference between correlation coefficients for the gray matter and DFC maps is positive, indicating that the change with age is in the same direction for both variables (i.e., increased DFC change goes with increased gray matter density change). (Reproduced with permission from Sowell, Thompson, et al., 2001.) (See figure 2.3.)

Plate 9 Plots of the relationship between age and gray matter density reveal different trajectories of gray matter changes for different brain regions. Shown is a surface rendering of a human brain (left hemisphere; left is anterior, right is posterior) with scatter plots for gray matter density at various points over the brain surface. The graphs are laid over the brain approximately where the measurements were taken. The axes for every graph are identical, with gray matter density plotted on the x-axis and age (in years) plotted on the y-axis (Sowell et al., 2003). (See figure 2.4.)

Plate 10 Cortical pattern matching methods and gray matter thickness. The skull-stripped, 3 D, gray-scale image volume is shown in the upper left for one representative subject. Surface renderings (upper right) are automatically rendered for each subject using the signal value that best differentiates cortical surface sulcal CSF from cortical gray matter. Thirty-five sulcal landmarks on the lateral and medial surfaces are identified and manually traced. After sulcal patterns are demarcated, surface renderings are flattened to a 2 D planar format. In the bottom left the flattened sulcally delineated surface renderings are shown for four individual subjects. Note the crosshairs in each map: while at slightly different locations in the image, they represent the same sulcal anatomy in each subject (i.e., homologous surface points). A complex deformation, or warping transform, is then applied that aligns the sulcal anatomy of each subject with an average sulcal pattern derived for

the group. Given that the deformation maps associate cortical locations with the same relation to the primary sulcal pattern across subjects (i.e., the crosshairs in all for subjects illustrated here), a local measurement of cortical thickness can be made in each subject and averaged across equivalent cortical locations in all subjects. This is illustrated in the bottom right panel. Cortical thickness, defined as the 3 D distance (in mm) between the inner gray matter/ white matter border and the closest point on the outer CSF/gray matter boundary, is calculated using the Eikonal fire equation (illustrated in more detail in figure 2.6). Using these methods, the average thickness value within a 15-mm sphere can be calculated and averaged across subjects to estimate cortical thickness within groups of individuals. On the bottom right is a group average map of cortical thickness. (See figure 2.5.)

Plate 11 Cortical thickness maps: (A) original T1-weighted image for one representative subject, (B) tissue segmented image, (C) gray matter thickness image where thickness is progressively coded in millimeters from inner to outer layers of cortex using the 3 D Eikonal fire equation. Note the images were resampled to a voxel size of 0.33 mm cubed, so the thickness measures are at a submillimeter level of precision, according to the color bar on the right (mm). Figures A through C are sliced at the same level in all three image volumes from the same subject. Shown in (D) is an in

vivo average cortical thickness map created from these 45 subjects at the first scan. The brain surface is color coded according to the color bar where thickness is shown in millimeters. Our average thickness map can be compared to an adapted version of the 1929 cortical thickness map of von Economo (von Economo, 1929) (E). Color coding has been applied over his original stippling pattern, respecting the boundaries of his original work, to highlight the similarities between the two maps. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See figure 2.6.)

Plate 12 These maps show the statistical significance of annualized change in cortical thickness measurements. Color coding represents t values at each cortical surface point according to the color bar at the near right (ranging from t = −3.00 to t = 3.0). Significant values are overlaid in shades of red (significant thickness decreases,

TD) and white (significant thickness increases, TI), according to the color bar at the far right. Arrows point to three regions of significant increases in gray matter thickness. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See figure 2.7.)

Plate 13 Brain-behavior maps showing the relationship between vocabulary scores and cortical thickness. These maps show the p value for negative correlations between change in cortical thickness (time 2 value − time 1 value) and change in vocabulary scores (time 2 score − time 1 score). Regions in color represent negative p values,

that is, regions where greater thinning was associated with greater vocabulary improvement. Regions in white were not significant. No positive correlations reached significance. (Reproduced with permission from Sowell, Thompson, Leonard, et al., 2004.) (See figure 2.8.)

Plate 14 Pearson’s R correlations between changes in gray matter thickness and phonological processing (panel A) and between thickness and motor processing (panel B). Regions in white represent positive correlations with a threshold of p < 0.05. Regions in red represent negative correlations with a threshold of p < 0.05. Increases in gray matter thickness in left inferior frontal regions significantly correlate with improvement in pho-

nological processing, but not with motor processing, while decreases in gray matter thickness in primary motor cortex significantly correlate with improvement in motor processing, but not with phonological processing. (Reproduced with permission of Oxford University Press from Lu et al., 2007.) (See figure 2.9.)

Plate 15 Schematic representation of developmental milestones in the forebrain, depicting proliferation and migration, aggregation, circuit formation, and synaptogenesis. The approximate times of each event in the human brain are labeled on the left, and corresponding days in the rodent brain are on the right. Neurons and glia are produced from proliferative zones, and a variety of physical and chemical cues contribute to their migration to specific locations in the brain. Transient structures, such as the subplate in the cerebral cortex, provide cues that assist in the generation of topographic

axon projections. Some elements of cell differentiation and synaptogenesis occur over protracted periods of time and peak after birth. Abbreviations: E, embryonic day; P, postnatal day; W, age in gestational weeks; M, age in months after birth; PZ, proliferative zone; MZ, marginal zone; Bf, basal forebrain; Th, thalamus; SP, subplate; CP, cortical plate; CC, corpus callosum; TC, thalamocortical projections; CTX, cerebral cortex. (Figure is modeled on Rakic, 1995.) (See figure 6.1.)

Plate 16 Schematic representation of cellular effects of in utero cocaine exposure in the rabbit. Rabbits exposed to low doses of cocaine intravenously exhibit alterations in the structure and function of neurons in the anterior cingulate cortex. The apical dendrites of pyramidal neurons (red) exhibit an undulating trajectory. The number of interneurons in which immunoreactivity for GABA is detectable (black) is increased. There also is an increase in parvalbumin (green) immunostaining in the dendrites of a subset of

these neurons. D1 dopamine receptor coupling to Gs protein (blue) is reduced, whereas coupling of the D2 receptor to Gi (orange) is unaffected. These changes likely influence the balance of excitatory and inhibitory influences in the cingulate cortex and produce the aberrant behavioral phenotypes exhibited by these rabbits, including psychostimulant responsiveness and attentional deficits (see text for more detail). (Adapted from Levitt et al., 1997.) (See figure 6.2.)

Plate 17 D1- and D2-like DA receptors produce distinct, and sometimes opposite, neuromodulatory effects on cell signaling, elec-trophysiological responses, and behavior. They are often coexpressed within specific functional brain circuits, and sometimes even by the same neurons. A reduction in D1 receptor signaling could shift the balance of activity to D2 recep-tor subtypes. Normal physiological responsiveness may not necessarily depend on the

levels of receptor (above a certain minimum), but rather a balance between opposing activities. Thus, in considering strategies for restoring normal DA responsiveness, blockade of D2 receptors may have beneficial effects on the net outflow of DA-dependent neural circuits. (Adapted from Stanwood and Levitt, 2004.) (See figure 6.3.)

Plate 18 Model of social processing. (A) Simplified diagram of the component processes essential for social behavior. An organism must first have an overarching motivation to interact with conspecifics. Given this proviso, the brain must be able to perceive a social stimulus through sensory processing. Once the social stimulus has been perceived, there must be an evaluative process in order to determine its intent. The intent of the social stimulus will have different consequences depending on the context in which it takes place. Context includes situational variables as well as the characteristics of the particular conspecific. Whether a social behavior is executed will depend not only on the development of a motor plan but also on whether modulatory influences are consistent with the implementation of the behavior. (B) Amygdalocentric network mediating danger detection. A variety of sensory stimuli indicative of dangers are perceived by the ventral visual processing system. This information is directed to the lateral nucleus of the amygdala, where an evaluation of the potential danger of the stimulus is carried out. Internal connections within the amygdala convey information from the lateral to the basal nucleus, which, in turn, receives a prominent input from the orbitofrontal cortex. This input may provide context information that is utilized to determine whether an escape behavior is executed. It is not surprising that the amygdala, as a multipurpose danger detector, is attuned to facial stimuli such as fear expressions.(See figure 11.1.)

Plate 19 Cell formation in the hippocampal formation of a 16-week-old fetus. Photomicrographs of MIB-1-labeled cells in cresyl-violet–counterstained coronal sections of the hippocampal formation of a 16-week-old fetus. (A) The ventricular zone (vz) along the CA1 area of Ammon’s horn and the region below the dentate gyrus (DG) and CA3 area is fully packed with labeled cells. There are far fewer labeled cells in the hilus (h) below the granule

cell layer (g). The area marked with a single asterisk is shown with higher magnification in B, whereas the region marked with double asterisks is shown in C. Arrows in A and C point to the section of the ventricular wall, where the proliferative germinal layer terminates. Calibration bars, 250 μm for A and 75 μm for B and C. (See figure 12.2.)

Plate 20 Cell formation in the dentate gyrus of 16- and 22-weekold fetuses. Photomicrographs of MIB-1-labeled and cresyl-violet– counterstained coronal sections of the dentate gyrus of (A) 16-week-old and (B) 22-week-old fetuses. Equally large numbers of

MIB-1-labeled cells appear in the hilus (h), below the granule cell layer (g), while only a few labeled cells are in the molecular layer (m) and in Ammon’s horn (CA3). Calibration bars, 75 μm. (See figure 12.3.)

Plate 21 Cell formation in the ventricular germinal layers. Photomicrographs of cresyl-violet–counterstained sections (A, B, C) from 22-week-old fetus and (D) from newborn child. In the hippocampal ventricular zone (v) there are only a few MIB-1-labeled cells, whereas both the ventrical zone (v) and the subventricular zone (SVZ) of the temporal cortex is fully packed with MIB-1positive cells. The area marked with an asterisk is shown with

higher magnification from an adjacent section in B, whereas a similar area marked with double asterisks is shown in C. In the newborn child MIB-1-positive proliferating cells are not visible in the hippocampal ventricular germinal zone (v) and are sparse in the zone of the temporal neocortex (D). Calibration bars, 100 μm for A and 50 μm for B, C, and D. (See figure 12.4.)

Plate 22 Postnatal cell proliferation and migration. Photomicrographs of coronal sections showing the border between the granule cell layer (g) and the hilus (h) of the dentate gyrus. The subgranular zone contains large numbers of migrating cells (a few pointed out by arrows) in a newborn (A), whereas such cells are absent in a one-year-old child (B). In a one-year-old child the neurons in the

granule cell layer (g) are easily distinguishable from astroglial (arrows) and oligodendroglial (open arrow) cells (B). In the newborn child a few MIB-1-positive cells are in the hilus (arrows) (C). Labeled cells (arrow) are rare in the hilus of the one-year-old child (D). Calibration bars, 20 μm. (See figure 12.5.)

Plate 23 Cajal-Retzius cells in the developing hippocampal formation. (A) Calretinin-immunoreactive Cajal-Retzius cell, displaying characteristic filopodia (open arrows) at the hippocampal fissure of a newborn child. (B) Reelin-immunostained Cajal-Retzius cells (arrows) at the hippocampal fissure of a newborn child. (C) A large, bipolar calretinin-immunoreactive Cajal-Retzius cell with long

dendrites running along the hippocampal fissure in a 2-year-old child. (D) Photomontage of a Golgi-impregnated large, bipolar Cajal-Retzius type cell at the hippocampal fissure of a newborn child. Filopodium-like processes (open arrows) are on both the soma and the dendrites. Calibration bars, 25 μm. (See figure 12.6.)

Plate 24 Cajal-Retzius cells in layer I of the temporal neocortex. (A) Photomicrographs of calretinin-positive Cajal-Retzius cells (arrows) in the temporal neocortex of a newborn child. Arrowheads point to calretinin-immunoreactive axonal plexus deeper in layer I, which may correspond to the axonal plexus of the Cajal-Retzius cells as shown on (F). (B, C) Calretinin-positive Cajal-Retzius cell (arrow) of different morphology in 5-month-old infant. (D) Reelinimmunoreactive Cajal-Retzius cells (arrows) and reelin-positive interneuron (curved arrow) in a 3-month-old infant. (E, F) Golgiimpregnated Cajal-Retzius-type cells (arrows) in the temporal

cortex of a newborn child corresponding to the cells first described by Retzius. (E) One of the cells has a long dendrite that runs parallel with the pial surface displaying the characteristic side branches (open arrows) that can be observed on many drawings of Retzius. (F) The other cell shows a fuzzy cell body (arrow) with several dendritic branches that run in different directions. The thin side branches leave the main dendrites perpendicularly to them (open arrows). Arrowheads show axonal plexus of Cajal-Retzius cells in the deep layer I. Calibration bars, 25 μm for A–C and E, 20 μm for D, 50 μm for F. (See figure 12.7.)

Plate 25 Calbindin immunoreactivity in granule cells of the dentate gyrus and in the pyramidal cells of Ammon’s horn. (A) Pyramidal cells that are located closer to the ventricular germinal layer express calbindin at the 16th GW in the pyramidal cell layer of CA1 area. (B) At the 21st GW, virtually all pyramidal cells of the CA1 area are calbindin immunoreactive. In contrast, only the oldest granule cells are calbindin positive at the border of the granule cell (g) and molecular layers (m). (C) At term, in the dorsal blade (arrow) of the granule cell layer (g) most cells display strong immunoreactivity for calbindin, while in the ventral blade (open curved arrows) only cells in the outer part of the granule cell layer

(g) close to the molecular layer (m) are immunoreactive. In the hilus (h) of the dentate gyrus a few scattered calbindin-positive cells are probably interneurons, whereas in the pyramidal layer of the CA3 area all cells are immunonegative. (D) Calbindin-immunoreactive cells in granule cell layer (g) of the dentate gyrus in a 5-month-old child. Many of the cells display immunoreactive dendrites that run toward the molecular layer (m). In contrast, large numbers of cells with elongated cell nuclei (one pointed out by an arrow) are immunonegative at the hilar border (h). These are probably the newly generated, still-migrating granule cells. Calibration bars, 50 μm for A, 100 μm in B and C, 20 μm for D. (See figure 12.8.)

Plate 26 Development of the calbindin- or calretinin-containing interneurons. (A) Calbindin-immunoreactive local circuit neurons (arrows) with immature morpholgy in the hilus (h), the stratum moleculare (m) of the dentate gyrus, and the stratum radiatum (r) of the CA3 area of the hippocampal formation of a 22-week-old fetus. (B) Large numbers of calretinin-positive interneurons at the border of the strata radiatum (r) and lacunosum-moleculare (l-m)

of the CA1 region of a 22-week-old fetus. The location is similar to that found in the adult, although the cellular morphology of these cells is still immature. (C) Small calretinin-immunoreactive cells in the ventricular germinative zone (vz) along the CA1 area of the hippocampus. These calretinin-positive cells may migrate later to the strata oriens (o) and pyramidale (p) of Ammon’s horn. Calibration bars, 50 μm. (See figure 12.9.)

Plate 27 Parvalbumin immunoreactive interneurons in the CA1 region. (A) Large parvalbumin-positive cell with immature morphology and sparse dendritic branches in a 1-month-old infant. Arrowheads point to parvalbumin-immunoreactive axons with terminal-like axonal swellings. (B) A multipolar, large parvalbuminimmunoreactive interneuron with lightly stained branching dendrites in a 3-month-old infant. Arrowheads point to rare terminal-like boutons. (C) Large, multipolar parvalbumin-positive cell in the pyramidal layer (p) of Ammon’s horn with dendrites running

both toward stratum oriens and through stratum radiatum (r) in an 10-year-old child. The morphology of this cell is comparable to parvalbumin-positive cells found in adults. The parvalbuminpositive axonal network is confined to the pyramidal cell layer (p). (D) High-magnification photomicrograph of the parvalbumin-positive axonal network in the pyramidal layer (p). Arrowheads point to axonal swellings (terminal-like boutons) that appear to surround individual neurons (n). Calibration bars, 20 μm for A and D, 25 μm for B, and 50 μm for C. (See figure 12.10.)

Plate 28 Parvalbumin-immunoreactive interneurons in the dentate gyrus. (A) Parvalbumin-immunoreactive axonal branches displaying large swellings (arrowheads) in the granule cell layer (g) and in the hilus (h) of the dentate gyrus in a 1-month-old infant. (B) Soma and main dendrites of a parvalbumin-immunoreactive cell in the granule cell layer of the dentate gyrus in a 3-month-old infant. Axonal branching (arrowheads) is similarly sparse as in 1month-old child. (C) Large-magnification photomicrograph of parvalbumin-immunoreactive axon terminals in the granule cell layer of an 10-year-old child. The axons (arrowheads) display large

numbers of boutons that appear to surround somata of granule cells (n). (D) A large parvalbumin-immunoreactive hilar (h), neuron with long dendrites that cross the granule cell layer (g). Inside the granule cell layer (g) the dense parvalbumin-immunoreactive axonal network appears to be denser at the hilar border (h), suggesting uneven perisomatic innervation of granule cells in the width of the granule cell layer. Density of axonal branches is low in the hilus (h), corresponding with a lower cellular density in the hilus than in the granule cell layer. Calibration bars, 20 μm for A, 25 μm for B, 10 μm for C, 50 μm for D. (See figure 12.11.)

Plate 29 Schematic illustration of the local circuitry of monkey prefrontal cortex formed by the axons of layer III pyramidal neurons (blue) and parvalbumin (red)- and calretinin (green)containing cells. Collaterals of pyramidal axons make excitatory synapses (filled blue circles) on dendrites of interneurons, and axons of specific classes of interneurons (A, B, C) make synapses on specific parts of pyramidal somata and dendrites (open red and green circles). (See figure 13.1.)

Plate 30 Histological laminar development correlated with postmortem MR images: (A) cresyl violet staining; (B, C, E) acetylcholinesterase histochemistry; (D) T1-weighted MRI scan; (A, B) 18–19 postconceptional weeks; (C, D) 23 postconceptional weeks;

Plate 31 Early origin of cholinergic afferents (arrows) from the nucleus basalis area (dark area that surrounds the asterisk). (See figure 13.6.)

(E) 29 postconceptional weeks. Note the prominent subplate zone (5) in the prefrontal cortex. (Reproduced with permission from Kostovic´ et al., 2002.) (See figure 13.4.)

Plate 32 Wiring diagram of prefrontal connections. The vast majority of connections shown are monosynaptic; however, some may be polysynaptic. Terminology for cortical areas according to Mesulam (1998): green, modulatory pathways; red, glutamatergic excitatory pathways; black, inhibitory neurons and pathways; blue, cholinergic projection; interconnection between different heteromodal areas is not completely shown. (See figure 13.2.)

Plate 33 Laminar development of cytoarchitectonic layers in the frontal cortex, from the stage before the appearance of the cortical plate to the newborn stage. All layers are transient, and their developmental changes reflect neurogenetic events (proliferation, migration, differentiation, ingrowth of afferent pathways). (A) Seven postconceptional weeks; (B) 10 postconceptional weeks; (C) 13 postconceptional weeks; (D) 23 postconceptional weeks; (E) 34 postconceptional weeks; (F) newborn. Abbreviations for this and subsequent figures: C, caudate nucleus; CP, cortical plate; G, ganglionic eminence; IZ, intermediate zone; MZ, marginal zone; P, putamen; SP, subplate zone; SPf, the subplate zone in formation; SVZ, subventricular zone; SVZf, subventricular fibrillar zone; VZ, ventricular zone; WM, white matter; I–VI, cortical layers I–VI. Arabic numerals in Figure 13.5E correspond to central (1), intermediate (2), and corona radiata (3), segments of the white matter. (See figure 13.3.)

Plate 34 Laminar development of the frontal lobe revealed by in vitro (A) and in vivo (C, E) imaging, correlated with Nissl-stained histological sections (B, D, F). Early consolidation of cortical plate at 10 postconceptional weeks (A, B). Typical fetal lamination with

prominent subplate zone at 21 postconceptional weeks (C, D). Gradual disappearance of the subplate zone with appearance of gyri, corona radiata, and “mature” pattern at 34–35 postconceptional weeks. (E, F). (See figure 13.5.)

Plate 35 Sequential development of afferents from the dorsomedial nucleus of the thalamus (acetylcholinesterase-stained histological sections, with superimposed line drawings). Early outgrowth (A, 10.5 postconceptional weeks), spread through the subplate zone

(B, 20 postconceptional weeks), waiting in the upper part of the subplate zone (C, 23 postconceptional weeks), and penetration into the cortical plate (D, 28 postconceptional weeks) during ingrowth and address selection are visible. (See figure 13.7.)

Plate 36 Transient circuitry of the human fetal cortex coexists with permanent circuitry elements. CP, cortical plate; PYR, pyramidal neurons; SP, subplate zone. (See figure 13.11.)

Plate 37 Imaging of white matter structure with diffusion tensor imaging (Klingberg et al., 2000). The direction of axons is color coded: yellow, up/down; red, anterior/posterior; blue, left/right. The encircled region (VOI) is where subjects with dyslexia have a disturbance of white matter. STG, superior temporal gyrus. IC, internal capsule. Inset, a lateral view of the white matter region affected in dyslexia. (See figure 14.2.)

Plate 38 A frontoparietal network involved in development of working memory (Olesen et al., 2003). Red: frontal and parietal regions where brain activity (BOLD signal) correlates with development of working-memory capacity; white: region where myelination (FA) correlates with development of working-memory capacity. Lines indicate correlations between brain activity and myelination. (See figure 14.3.)

Plate 39 Gamma-band EROs time locked to a tunnel-lifting event in 6-month-old infants. In the “unexpected-disppearance” condition the infants had just seen a train entering the tunnel. The

difference map represents the scalp distribution of the oscillatory activity. (Adapted from Kaufman, Csibra, and Johnson, 2003.) (See figure 15.4.)

Plate 40 The major substrates of smooth-pursuit eye movements and their connections. Broken lines indicate connections that are still hypothetical or have not been elucidated in sufficient detail. The scheme considers observations, not discussed in the main text, that suggest that signals for horizontal and vertical smooth pursuit are dealt with by different parts of the vestibular complex: namely, horizontal smooth pursuit by medial vestibular nuclei; and vertical smooth pursuit by the ygroup—a small cell group that caps the inferior cerebellar peduncle and that, similar to vestibular complex neurons, receives primary vestibular afferents. Abbreviations: FEF, frontal eye field; LGN, lateral geniculate nucleus; MST, middle superior temporal; MT, middle temporal; NRTP, nucleus reticularis tegmenti pontis; PN, pontine nuclei; SEF, supplementary eye field; V1, primary visual cortex; VN, vestibular nuclei. (Figure and figure caption reprinted from Thier and Ilg, 2005, with permission from Elsevier.) (See figure 16.3.)

Plate 41 Fixations made by an observer while making a peanut butter and jelly sandwich. Images were taken from a camera mounted on the head, and a composite image mosaic was formed by integrating over different head positions using a method described in Rothkopf and Pelz (2004). (The reconstructed panorama shows artifacts due to the incomplete imaging model that

does not take the translational motion of the subject into account.) Fixations are shown as yellow circles, with diameter proportional to fixation duration. Red lines indicate the saccades. Note that almost all fixations fall on task-relevant objects. (Figure and figure caption reprinted from Hayhoe and Ballard, 2005, with permission from Elsevier.) (See figure 16.4.)

Plate 42 Sagittal view of the cingulum bundle as rendered by stream-based tractography (left panel) and un-interpolated DTI fractional anisotropy image of the same data in the coronal plane (right panel) illustrating the low spatial resolution of DTI. (See figure 17.1.)

A

B

Plate 43 Functional magnetic resonance images of the brain. (A) T2*-weighted blood-oxygenation-level-dependent (BOLD) magnetic resonance signal overlaid on T1-weighted structural image at the same axial slice location. (B) Regions demonstrating a statisti-

cally significant difference in activation for an experimental condition relative to a baseline condition. Color changes indicate the magnitude of the F-statistic. Only pixels showing an F-value significant at p = 0.005 or better are displayed. (See figure 18.2.)

Plate 44 The artificial speech streams used in Shukla, Nespor, and Mehler (2007). The upper panel shows the design of a monotonous speech stream with nonce words inserted at the indicated locations. The lower panel shows an intonated stream, obtained by

overlaying prosody (i.e., intonational phrase [IP] contours) on the previously monotonous syllable string. Now the nonce words fall within a contour or straddling the boundary of two adjacent contours. (See figure 19.1.)

Plate 45 Conventional MRI (top row) and diffusion tensor tractography (bottom row) in three 9-year-old children. Images show white matter pathways in a healthy child (left) and in two children who sustained early TBI (middle and right). The middle image is from a child sustaining bilateral subdural hematomas, subarachnoid hemorrhage, and frontal lobe contusions at 2 months of age; follow-up MRI disclosed focal atrophy in left parietal cortex, deep white periventricular atrophy with associated thinning of the corpus callosum, and compensatory enlargement of the lateral ventricles. Tractography shows reduced callosal fibers (left-right fibers indi-

cated in red), diminution of the corticospinal tract (inferior-superior fibers in blue), and diminution of association fibers (anteriorposterior fibers in green) in prefrontal regions and in arcuate and superior longitudinal fasciculi. Images on the right side are from a child sustaining a parietal depressed skull fracture with underlying subdural hematoma at 3 months of age; follow-up MRI revealed left parietal encephalomalacia and moderate thinning of the posterior body of the corpus callosum. Tractography shows focal reduction in posterior callosal and association fibers consistent with focal parietal lobe injury. (See figure 24.1.)

Plate 46 Functional MRI activation in Broca’s and Wernicke’s areas in the LH and their homologues in the RH, as elicited by written English in native English speakers (top) and ASL in native signers (bottom). Activation is depicted in bar graphs, as percent signal change as well as spatial extent (mm3). (Reprinted with per-

mission from Bavelier, D., D. Corina, P. Jezzard, V. Clark, A. Karni, A. Lalwani, J. P. Rauschecker, A. Braun, R. Turner, and H. J. Neville, 1998. Hemispheric specialization for English and ASL: Left invariance–right variability. NeuroReport 9(7):1537– 1542.) (See figure 26.2.)

Plate 47 Gamma-band activity presented as a time-frequency analysis of the average EEG at four electrodes over the right temporal cortex during and after the object’s disappearance behind a visible or invisible occluder. Gray indicates the times of statistical difference. There was significantly greater gamma EEG activity in

the occlusion event than in the disintegration event before and just after the disappearance of the object. (Reprinted from J. Kaufman, G. Csibra, and M. J. Johnson, 2005. Oscillatory activity in the brain reflects object maintenance. Proc. Natl. Acad. Sci. USA 102(42):15271–15274, with permission.) (See figure 28.2.)

Familiarisation trials

Test trials

Baseline trial

Locations (ST) trial

Identity (SF) trial

Binding trial

Plate 48 Examples of test and familiarization trials in Mareschal and Johnson (2003). The top four panels illustrate a single familiarization event, whereas the bottom four panels illustrate the four possible test trials. (Reprinted from D. Mareschal, and M. H. Johnson, 2003. The “what” and “where” of infant object representations in infancy. Cognition 88: 259–276. Copyright 2003, with permission from Elsevier.) (See figure 28.3.)

Cholinergic

Noradrenergic

Dopaminergic

Serotoninergic

Plate 49 The neurochemical systems involved in attention and arousal. Abbreviations: III, oculomotor nucleus; T, thalamus; HC, hippocampal formation; RF, reticular formation; PSG, parasympathetic ganglion cell; X, dorsal motor nucleus of the vagus; H, hypo-

thalamus; LC, locus ceruleus; C, caudate nucleus; P, putamen; S, septal nuclei; V, ventral striatum. (Reprinted with permission of the publisher from J. Nolte and J. B. Angevine, The Human Brain, pp. 134–137, St. Louis: Mosby. Copyright 1995.) (See figure 29.2.)

InFrequent Familiar

InFrequent Familiar

μV +10

InFrequent Novel

InFrequent Novel

Plate 50 The Nc component during attention and inattention. The ERP recording from 100 ms prior to stimulus onset through 1 s following stimulus is shown for the FZ and CZ electrodes for attentive (top figures) and inattentive (bottom figures) periods, combined over the three testing ages. The topographical scalp potential maps show the distribution of this component for the three memory stimulus types in attention and inattention. The topographical maps represent an 80-ms average of the ERP for the Nc component at the maximum point of the ERP response. The data are plotted with a cubic spline interpolation algorithm and represent absolute amplitude of the ERP. (From Richards, 2003a, figure 2.) (See figure 29.8.)

-10 μV

Frequent Familiar

Nc During Inattention

Frequent Familiar

Nc During Attention

-20μV

ICA Projection

Plate 51 The ICA component cluster for the prefrontal component. The topographical map of the average ICA loadings is similar to the topographical map of the grand average ERP of the Nc component. The ECD locations are displayed on several MRI slices, and each location represents an ICA from one individual. (From Reynolds and Richards, 2005, figure 4.) (See figure 29.9.)

μV 20μ

Grand Average ERP

Prefrontal ICA Cluster: Medial Fronta l Gyrus (25) , Inferior Frontal Gyrus (47), Anterior CingulateCortex (8) (Talaira chcoordinates: 9.4, 42.9, 16.4)

Plate 52 The face perception system. The three rectangles with beveled edges show the core system for face perception. Solid lines indicate cortical pathways, and dashed lines show the subcortical route. Areas in yellow represent regions involved in processing identity and associated semantic information. Areas in red represent regions involved in emotion analysis, and areas in blue indicate

those involved in spatial attention as it interacts with the face processing system. (Reprinted form R. Palermo and G. Rhodes, 2007. Are you always on my mind? A review of how face perception and attention interact. Neuropschologia, 45:75–92, with permission from Elsevier.) (See figure 31.1.)

Plate 53 Visual processing pathways in monkeys. Solid lines indicate connections arising from both central and peripheral visual field representations; dotted lines indicate connections restricted to peripheral field representations. Red boxes indicate ventral stream areas related primarily to object vision; green boxes indicate dorsal stream areas related primarily to spatial vision; and white boxes indicate areas not clearly allied with either stream. The shaded region on the lateral view of the brain represents the extent of the cortex included in the diagram. Abbreviations are as follows: DP, dorsal prelunate area; FST, fundus of superior temporal area; HIPP, hippocampus; LIP, lateral intraparietal area; MSTc, medial superior temporal area, central visual field representation; MSTp, medial superior temporal area, peripheral visual field representation; MT, middle temporal area; MTp, middle temporal area, peripheral visual field representation; PO, parieto-occipital area;

PP, posterior parietal sulcal zone; STP, superior temporal polysensory area; V1, primary visual cortex; V2, visual area 2; V3, visual area 3; V3A, visual area 3, part A; V4, visual area 4; and VIP, ventral intraparietal area. Inferior parietal area 7a; prefrontal areas 8, 11 to 13, 45, and 46; perirhinal areas 35 and 36; and entorhinal area 28 are from Brodmann (1909). Inferior temporal areas TEO and TE, parahippocampal area TF, temporal pole area TG, and inferior parietal area PG are from von Bonin and Bailey (1947). Rostral superior temporal sulcal (STS) areas are from Seltzer and Pandya (1978), and VTF is the visually responsive portion of area TF (Boussaoud, Desimone, and Ungerleider, 1991). (Figure and caption reprinted from L. G. Ungerleider, 1995. Functional brain imaging studies of cortical mechanisms for memory. Science 270:770.) (See figure 32.1.)

DL-PFC

VL-PFC

ACC

RL-PFC

Plate 54 The human brain, showing various regions of prefrontal cortex on the lateral surface (left) and the medial surface (right). OFC, orbitofrontal cortex; ACC, anterior cingulate cortex; DL-

OFC

PFC, dorsolateral prefrontal cortex; VL-PFC, ventrolateral prefrontal cortex; RL-PFC, rostrolateral prefrontal cortex. (See figure 34.2.)

Plate 55 Right lateral and top views of the human brain, showing age-related declines in gray matter volume. (Reprinted with permission from N. Gogtay, J. N. Giedd, L. Lusk, K. M. Hayashi, D. Greenstein, A. C. Vaituzis, et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proceedings of the National Academy of Sciences of the USA 101:8174–8179.) (See figure 34.3.)

Plate 56 A hierarchical model of rule representation in PFC. A lateral view of the human brain is depicted at the top of the figure, with regions of PFC identified by the Brodmann areas (BA) that comprise them: orbitofrontal cortex (BA 11), ventrolateral PFC (BA 44, 45, 47), dorsolateral PFC (BA 9, 46), and rostrolateral PFC (BA 10). The PFC regions are shown in various shades of gray, indicating which types of rules they represent. Rule structures are depicted below, with darker shades of gray indicating increasing levels of rule complexity. The formulation and maintenance in working memory of more complex rules depends on the reprocessing of

information through a series of levels of consciousness, which in turn depends on the recruitment of additional regions of PFC into an increasingly complex hierarchy of PFC activation. Abbreviations: S, stimulus; check mark, reward; cross, nonreward; R, response; C, context, or task set. Brackets indicate a bivalent rule that is currently being ignored. (Reprinted with permission from S. Bunge and P. D. Zelazo, 2006. A brain-based account of the development of rule use in childhood. Current Directions in Psychological Science 15:118–121.) (See figure 34.4.)

Plate 57 Sample target and test cards in the standard version of the Dimensional Change Card Sort (DCCS). (Reprinted with permission from P. D. Zelazo, 2006. The Dimensional Change Card

Sort [DCCS]: A method of assessing executive function in children. Nature Protocols 1:297–301.) (See figure 34.5.)

Plate 58 Functional MRI abnormalities observed in autism spectrum disorders (ASD). A, These coronal MRI images show the cerebral hemispheres above, the cerebellum below, and a circle over the fusiform gyrus of the temporal lobe. The examples illustrate the frequent finding of hypoactivation of the fusiform gyrus to faces in an adolescent male with ASD (right) compared with an age- and IQ-matched healthy control male (left). The red/yellow signal shows brain areas that are significantly more active during perception of faces; signals in blue show areas more active during perception of nonface objects. Note the lack of face activation in the boy with ASD but average levels of nonface object activation. B, Schematic diagrams of the brain from lateral and medial orien-

tations illustrating the broader array of brain areas found to be hypoactive in ASD during a variety of cognitive and perceptual tasks that are explicitly social in nature. Some evidence suggests that these areas are linked to form a “social brain” network. IFG, Inferior frontal gyrus (hypoactive during facial expression imitation); pSTS, posterior superior temporal sulcus (hypoactive during perception of facial expression and eye gaze tasks); SFG, superior frontal gyrus (hypoactive during theory of mind tasks, i.e. when taking another person’s perspective); A, amygdala (hypoactive during a variety of social tasks); FFA, fusiform face area (hypoactive during perception of personal identity). (From DiCicco-Bloom et al., 2006.) (See figure 43.1.)

Plate 59 Simon versus Stroop activations. Axial slices are labeled with Talairach Z-coordinates at far left. “Contrast”: comparison of Simon versus Stroop activity with p < 0.05, cluster filter 9 pixels. The Simon tends to activate superior parietal cortices slightly more than does the Stroop, and the Stroop tends to activate posterior temporal areas more than does the Simon, consistent with their greater demands on spatial and receptive language functions, respectively. (See figure 44.1.)

Plate 60 Main effects of Stroop task performance in (A) Tourette’s syndrome and (B) normal control groups. Diagnosis-byperformance interactions (C) were detected in frontostriatal and posterior cingulate regions in which interactions of diagnosis with age were observed. The TS subjects likely needed to engage frontostriatal regions more to maintain task performance. (See figure 44.2.)

GABA Glutamate

S

Striosomal medium spiny neuron

M

Medium spiny neuron in the matrix

Acetylcholine Dopamine

FS

Tan

Midline thalamic nuclei

Inputs to striosomes

Fast spiking aspiny neuron

Tonically active neuron

Cortical inputs to the matrix

Hippocampus Prefrontal cortex

Striosome M Anterior cingulate cortex

FS

S

M

Tan Orbital frontal cortex

S

Motor and premotor cortex

FS M Sensorimotor cortex

Amygdala

Matrix

Substantia nigra, pars compacta

Plate 61 A schematic diagram of the major inputs into the medium spiny GABAergic projection neurons of the striatum. Tonically active neurons (TANs) synapse on fast-spiking neurons (FSNs), which play a key role in modulating the medium spiny neurons (M). (See figure 44.6.)

Direction of Attention Avoid Anger

Attend to Anger

vs. * Plate 62 This figure shows a group contrast of an event-related analysis in 18 pediatric GAD patients and 15 age-matched healthy adolescents, drawing data from Monk and colleagues (2006). The specific contrast shown maps brain areas where GAD patients show enhanced activation for a contrast of events where asterisk targets

* appear proximal to neutral faces, relative to when they appear proximal to angry faces. As shown in the upper left, a group difference emerges in the prefrontal cortex, at coordinates of x, y, z = 51, 30, −2, in Talairach space, with t = 3.3 (p = .001). (See figure 46.3.)

B

A

C

Plate 63 Illustrations of RJA following gaze and point (A), IJA pointing (B), and IJA alternating gaze (C) from the Early Social Communication Scales (ESCS; Mundy, Delgado, et al., 2003). (See figure 50.1.)

Plate 64 Differential expression of 12-lipoxygenase (12-lox) and quinoid dihydropteridine reductase (QDR) in adolescent brain. (A) Color-rendered images from coronal forebrain sections subjected to in situ hybridization revealed higher levels of 12-lox in specific cortical brain regions of the adolescent brain (top panels). Bottom right panels are of sections subjected to emulsion autoradiography to illustrate intense expression of 12-lox in neurons of adolescent compared to adult brain. (B) Top panels are coronal forebrain sections depicting heightened expression of QDR in white matter regions of adult brain. Higher expression levels in adult forebrain

were evident in oligodendrocytes (bottom-right panels). Leftbottom panels in A and B are Northern blots depicting an abundance of 12-lox and QDR mRNA in adolescent and adult brain, respectively. The enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to control for equal loading of RNA in each case. Arrows in A represent neurons in cortical layer IV of adolescent or adult brain that express 12-lox. In B, arrows indicate oligodendrocytes that express QDR in a typical “beads-on-astring” histology pattern evident in both adolescent and adult forebrain. (See figure 52.2.)

Plate 65 Arc and dendrin, which are involved in synaptic plasticity, are differentially induced in adolescent forebrain following acute nicotine. (A) Color-rendered images from coronal forebrain sections hybridized to arc probe in situ revealed a dramatic induction of arc in the prefrontal cortex of acutely nicotine-treated ado-

lescent rodents. (B) A much less dramatic up-regulation of arc following acute nicotine was evident in adult animals. (C) Dendrin mRNA was also induced in specific forebrain regions following acute nicotine treatment of adolescent rodents, while (D) little change was evident in adult animals. (See figure 52.3.)

160 18–20 years 14–15 years 9–11 years

140 120 100 80 60 40 20 0

First warning

Efficient negFB

Error negFB

First posFB

Correct posFB

Negative FB > Positive FB 18–20 yrs (n = 20)

14–15 yrs (n = 20)

9–11 yrs (n = 15)

RCI contrast value

2 1.5 1

18–20 years 14–15 years 9–11 years

0.5 0 –0.5 –1 First warning Efficient negFB Error negFB First posFB Correct posFB

RCI contrast value

2 1.5 1 0.5 0 –0.5 –1 First warning Efficient negFB Error negFB First posFB Correct posFB

Plate 66 Neural activity associated with the processing of positive and negative performance feedback for children, adolescents, and adults. The pattern of activation in anterior cingulate cortex

shows an adult pattern in adolescence, whereas the pattern of activation in DLPFC does not reach adult levels until late adolescence. (See figure 54.4.)