The Biological Basis for Mind Body Interactions

PROGRESS IN BRAIN RESEARCH

VOLUME 122 THE BIOLOGICAL BASIS FOR MIND BODY INTERACTIONS

Other volumes in PROGRESS IN BRAIN RESEARCH Volume96: Neurobiology of Ischemic Brain Damage, by K. Kogure, K.-A. Hossmann and B.K. Siesjo (Eds.) - 1993, ISBN 0-444-89603-1. Volume 97: Natural and Artificial Control of Hearing and Balance, by J.H.J. Allum, D.J. AllumMecklenburg, F.P. Harris and R. Probst (Eds.) - 1993, ISBN 0-444-81252-0. Volume98: Cholinergic Function and Dysfunction, by A.C. Cuello (Ed.) - 1993, ISBN 0-444-89717-8. Volume 99: Chemical Signalling in the Basal Ganglia, by G.W. Arbuthnott and P.C. Emson (Eds.) 1993, ISBN 0-444-8 1562-7. Volume 100: Neuroscience: From the Molecular to the Cognitive, by F.E. Bloom (Ed.) - 1994, ISBN 0-444-81678-X. Volume 101: Biological Function of Gangliosides, by L. Svennerholm et al. (Eds.) - 1994, ISBN 0-444-81658-5. Volume 102: The Self-organizing Brain: From Growth Cones to Functional Networks, by J. van Pelt, M.A. Corner, H.B.M. Uylings and F.H. Lopes da Silva (Eds.) - 1994, ISBN 0-444-818 19-7. Volume 103: Neural Regeneration, by F.J. Seil (Ed.) - 1994, ISBN 0-444-81727-1. Volume 104: Neuropeptides in the Spinal Cord, by F. Nyberg, H.S. Sharma and Z. Wiesenfeld-Hallin (Eds.) - 1995, ISBN 0-444-81719-0. Volume 105: Gene Expression in the Central Nervous System, by A.C.H. Yu et al. (Eds.) - 1995, ISBN 0-444-8 1852-9. Volume 106: Current Neurochemical and Pharmacological Aspects of Biogenic Amines, by P.M. Yu, K.F. Tipton and A.A. Boulton (Eds.) - 1995, ISBN 0-444-81938-X. Volume 107: The Emotional Motor System, by G. Holstege, R. Bandler and C.B. Saper (Eds.) - 1996, ISBN 0-444-81962-2. Volume 108: Neural Development and Plasticity, by R.R. Mize and R.S. Erzurumlu (Eds.) - 1996, ISBN 0-444-82433-2. Volume 109: Cholinergic Mechanisms: From Molecular Biology to Clinical Significance, by J. Klein and K. Loffelholz (Eds.) - 1996, ISBN 0-444-82166-X. Volume 110: Towards the Neurobiology of Chronic Pain, by G. Carli and M. Zimmermann (Eds.) 1996, ISBN 0-444-82149-X. Volume I 1 1: Hypothalamic Integration of Circadian Rhythms, by R.M. Buijs, A. Kalsbeek, H.J. Romijn, C.M.A. Pennartz and M. Mirmiran (Eds.) - 1996, ISBN 0-444-82443-X. Volume 1 12: Extrageniculostriate Mechanisms Underlying Visually-Guided Orientation Behavior, by M. Norita, T. Bando and B.E. Stein (Eds.) - 1996, ISBN 0-444-82347-6. Volume 113: The Polymodal Receptor: A Gateway to Pathological Pain, by T. Kumazawa, L. Kruger and K. Mizumura (Eds.) - 1996, ISBN 0-444-82473-1. Volume 114: The Cerebellum: From Structure to Control, by C.I. de Zeeuw, P. Strata and J. Voogd (Eds.) - 1997, ISBN 0-444-82313-1. Volume 115: Brain Function in Hot Environment, by H.S. Sharma and J. Westman (Eds.) - 1998,ISBN 0-444-82377-8. Volume 116: The Glutamate Synapse as a TherapeuticalTarget: Molecular Organization and Pathology of the Glutamate Synapse, by O.P. Ottersen, I.A. Langmoen and L. Gjerstad (Eds.) 1998, ISBN 0-444-82754-4. Volume 1 17: Neuronal Degeneration and Regeneration: From Basic Mechanisms to Prospects for Therapy, by F.W. van Leeuwen, A. Salehi, R.J. Giger, A.J.G.D. Holtmaat and J. Verhaagen (Eds.) - 1998, ISBN 0-444-82817-6. Volume 118: Nitric Oxide in Brain Development, Plasticity, and Disease, by R.R. Mize, T.M. Dawson, V.L. Dawson and M.J. Friedlander (Eds.) - 1998, ISBN 0-444-82885-0. Volume 119: Advances in Brain Vasopressin, by I.J.A. Urban, J.P.H. Burbach and D. De Wied (Eds.) - 1999, ISBN 0-444-50080-4. Volume 120: Nucleotides and their Receptors in the Nervous System, by P. Illes and H. Zimmermann (Eds.) - 1999, ISBN 0-444-50082-0. Volume 121: Disorders of Brain, Behavior and Cognition: The Neurocomputational Perspective, by J.A. Reggia, E. Ruppin and D. Glanzman (Eds.) - 1999, ISBN 0-444-50175-4.

PROGRESS IN BRAIN RESEARCH VOLUME 122

THE BIOLOGICAL BASIS FOR MIND BODY INTERACTIONS EDITED BY E.A. MAYER U C W C U R E Neuroenteric Disease Program, Department of Medicine and Physiology, UCLA School of Medicine, LQS Angeles, CA 90024, USA

C.B. SAPER Department of Neurology, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215, USA

ELSEVIER AMSTERDAM - LAUSANNE - NEW YORK - OXFORD - SHANNON - SINGAPORE - TOKYO 2000

ELSEVIER SClENCE B.V. Sara Burgerhartstraat 25 P.O. Box 21 I, 1000 AE Amsterdam, The Netherlands 0 2000 Elsevier Science B.V. All rights reserved.

This work is protected under copyright by Elsevier Science, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may he made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier Science Rights & Permissions Department, PO Box 800, Oxford OX5 IDX, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: [email protected]. You may also contact Rights & Permissions directly through Elsevier's home page (http://www.elsevier.nl). selecting first 'Customer Support'. then 'General Information'. then 'Permissions Query Form'. In the USA, users may clear permissions and make payments through the Copyright Clearance Center. Inc.. 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (978) 7508400, fax: (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90Tottenham Court Road. London WIPOLP. UK: phone: (+44) 171 631 5555; fax: ( + 4 4 ) 171 631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, hut permission of Elsevier Science is required for external resale or distribution of such material. Permission of the Puhlisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or pan of a chapter. Except as outlined above, no part of this work may he reproduced, stored in a retrieval system or transmitted in any form or hy any means, electronic, mechanical, photocopying, recording or otherwise. without prior written permission of the Publisher. Address permissions requests to: Elsevier Science Rights & Permissions Department. at the mail, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury andor damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should he made. First edition 2000 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congrcss has heen applied for.

ISBN: 0-444-50049-9 (volume) 0-444-80104-9 (series)

8

The paper used in this publication meets the requiremenls of ANSVNISO 239.48-1992 (Permanence of Paper). Printed in The Netherlands.

List of Contributors

K.J.S. Anand, Division of Critical Care Medicine, 800 Marshall Street, Arkansas Children’s Hospital, Little Rock, AR 72202, USA V.P. Bakshi, Department of Psychiatry, University of Wisconsin at Madison, Wisconsin Psychiatric Institute and Clinics, 6001 Research Park Blvd., Madison, WI 53719, USA R. Bandler, Department of Anatomy and Histology, F13, University of Sydney, Sydney, NSW 2006, Australia J.J. Barrell, Departments of Oral and Maxillofacial Surgery and Neuroscience, University of Florida, 2210 N W 24th Ave., Gainesville, Fl32605, USA L. Chang, UCLNCURE Neuroenteric Disease Program, West LA VA Medical Center, Bldg. 115, Room 223, 11301 Wilshire Blvd., Los Angeles, CA 90073, USA N. Clerc, Laboratoire de Neurobiologie, CNRS, 3 1 Chemin J. Aiguier, 13402 Marseilles Cedex 20, France 0. Devinsky, New York University - Mount Sinai Comprehensive Epilepsy Center, 560 First Avenue, Rivergate 4th Floor, New York, NY 10016, USA G. Devroede, Centre Universitaire de Sant6 de L‘Estrie, Pavillon Heurimont, DtSpartement de Chirurgie, 3001 12e Avenue Nord, Fleurimont, Que. J1H 5N4, Canada D.L. Diehl, UCLA School of Medicine, 16300 Sand Canyon Avenue, #loo, Irvine, CA 92618, USA D. Eisenberg, Center for Alternative Medicine Research, Beth Israel/Deaconess Hospital, 330 Brookline Avenue, Boston, MA 02215, USA A. Ericsson, Laboratory of Neuronal Structure and Function, The Salk Institute for Biological Studies and the Foundation for Medical Research, 10010 N. Torrey Pines Rd., La Jolla, CA 92037, USA D.L. Felten, Departments of Pathology and Neurology, Lorna Linda University School of Medicine, Alumni Hall for Basic Sciences 321, 11021 Campus St., Lorna Linda, CA 92350, USA H.L. Fields, Box 01 14 - Neurology, UC San Francisco, 513 Parnassus Avenue, S-784, San Francisco, CA 94143-0114, USA R.D. Foreman, University of Oklahoma Health Sciences Center, Department of Physiology, P.O. Box 26901, BMSB 653, Oklahoma City, OK 73190, USA J.B. Furness, Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia J.M. Fuster, UCLA Neuropsychiatric Institute, 760 Westwood Plaza, Los Angeles, CA 90024, USA

vi

T.L. Gruenewald, Department of Psychology, Franz Hall, UCLA, Los Angeles, CA 90095, USA H.-J. Habler, Physiologisches Institut, Universitat Kiel, Olshausenstrasse 40,24098 Kiel, Germany R.L. Huot, Stress and Neurobiology Laboratory, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta GA 30322, USA W. Janig, Physiologisches Institut, Universitat Eel, Olshausenstrasse 40, 24098 Gel, Germany R. Jessum, Emperors’ College of Traditional Oriental Medicine, Santa Monica, CA 90403, USA D.H. Johnson, California Institute of Integral Studies, 1453 Mission Street, San Francisco, CA 94103, USA N.H. Kalin, University of Wisconsin at Madison, Department of Psychiatry, 6001 Research Park Blvd., Madison, WI 53719, USA K.A. Keay, Department of Anatomy and Histology, F13, University of Sydney, Sydney NSW 2006, Australia M.E. Kemeny, Department of Psychology, UCLA, 5625 Franz Hall, Los Angeles, CA 90095-1563, USA S.G. Khasar, Departments of Anatomy, Medicine and Oral Surgery and Division of Neuroscience Biomedical Sciences Program and NIH Pain Center (UCSF), University of California, San Francisco, CA 94143-0440, USA C.O. Ladd, Stress Neurobiology Laboratory, Department of Psychiatry and Behavioral Sciences, 1639 Pierce Drive, WMRB Room 4105, Atlanta, GA 30322, USA J.D. Levine, Departments of Anatomy, Medicine and Oral Surgery and Division of Neuroscience Biomedical Sciences Program and NIH Pain Center (UCSF), University of California, San Francisco, CA 94143-0440, USA H.-Y. Li, Laboratory of Neuronal Structure and Function, Salk Institute for Biological Studies and Foundation for Medical Research, 10010 N. Torrey Pines Rd., La Jolla, CA 92037, USA T.J. Maher, Massachusetts College Pharmacy, 179 Longwood Avenue, Boston, MA 021 15, USA D.J. Mayer, Department of Anesthesiology, Medical College of Virginia, Box 516, Richmond, VA 23298-0695, USA E.A. Mayer, UCLNCURE Neuroenteric Disease Program, Department of Medicine and Physiology, UCLA School of Medicine, Los Angeles, CA 90024, USA B . S . McEwen, Rockefeller University, Laboratory of Neuroendocrinology, 1230 New York Avenue, New York, NY 10021, USA M.J. Meaney, Developmental Neuroendocrinology Laboratory, Departments of Psychiatry, and Neurology and Neurosurgery, Douglas Hospital Research Centre, McGill University, Montreal, Que. H4H 1R3, Canada F.J.-P. Miao, Departments of Anatomy, Medicine and Oral Surgery and Division of Neuroscience Biomedical Sciences Program and NIH Pain Center (UCSF), University of California, San Francisco, CA 94143-0440, USA M.A. Mittleman, Institute for Prevention of Cardiovascular Disease, Beth Israel Deaconess Medical Center, Harvard University Medical School, 1 Autumn Street, Boston, MA 02215, USA

vii

J. Munakata, UCLNCURE Neuroenteric Disease Program, West LA VA Medical Center, Bldg. 115, Room 223, 11301 Wilshire Blvd., Los Angeles, CA 90073, USA D.L. Musselman, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 1639 Pierce Drive, Suite 4000, Atlanta, GA 30322, USA B.D. Naliboff, UCLNCURE Neuroenteric Disease Program, West LA VA Medical Center, Bldg. 115, Room 223, 11301 Wilshire Blvd., Los Angeles, CA 90073, USA C.B. Nemeroff, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 1639 Pierce Drive, WMB, Suite 4000, Atlanta, GA 30322, USA P.M. Plotsky, Stress Neurobiology Laboratory, Department of Psychiatry and Behavioral Sciences, 1639 Pierce Drive, WMRB, Room 4105, Atlanta, GA 30322, USA D.D. Price, Departments of Oral and Maxillofacial Surgery and Neuroscience, University of Florida, 2210 NW 24th Ave., Gainesville, FL 32605, USA J.L. Price, Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, MO 63 110, USA L. Rapgay, UCLA Mind Body Medicine Group, Division of Head and Neck Surgery, No. 550,200 Medical Plaza, Los Angeles, CA 90024, USA N.W. Read, Centre for Human Nutrition, Northern General Hospital, Sheffield S5 7AU, UK V.L. Rinpoche, Gaden Shartse Monastic College, Mundgod, Karnataka State, India C.B. Saper, Department of Neurology, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215, USA P.E. Sawchenko, Laboratory of Neuronal Structure and Function, The Salk Institute for Biological Studies and the Foundation for Medical Research, 10010 N. Torrey Pines Rd., La Jolla, CA 92037, USA S.E. Shelton, Department of Psychiatry, Wisconsin Psychiatric Institute and Clinics, 6001 Research Park Blvd., Madison, WI 53719, USA R.W. Sikes, Department of Physical Therapy, Northeastern University, 6 Robinson Hall, 360 Huntington Avenue, Boston, MA 021 15, USA G.P. Smith, E.W. Borne, Behavioral Research Laboratory, New York Presbyterian Hospital, 21 Bloomingdale Rd., White Plains, NY 10605, USA D.S. Sobel, Kaiser Permanente Northern California, Regional Health Education, 1950 Franklin, 13th Floor, Oakland, CA 94612, USA R. Sovik, 841 Delaware Avenue, Buffalo, NY 14209, USA E.M. Sternberg, National Institute of Mental Health NIH, Building 10, Rm. 2D-46, 10 Center Drive, MSC 1284, Bethesda, MD 20892-1284, USA K.V. Thrivikraman, Stress Neurobiology Laboratory, Department of Psychiatry and Behavioral Sciences, 1639, Pierce Drive, WMRB, Room 4105, Atlanta, GA 30322, USA D. Tranel, Department of Neurology, University of Iowa Hospitals and Clinics, 200 Hawkins Drive, Iowa City, IA 52242, USA R.L. Verrier, Institute for Prevention of Cardiovascular Disease, Beth Israel Deaconess Medical Center, Harvard University Medical School, 1 Autumn Street, Boston, MA 02215, USA B.A. Vogt, Department of Physiology and Pharmacology,Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, NC 27 157-1083, USA

This Page Intentionally Left Blank

ix

Preface

This book is based on a symposium entitled ‘The Science and Practice of Mind/Body Interactions’ which was held in Sedona, Arizona on March 15-18, 1998. The symposium brought together thought leaders from the areas of neuroscience, practitioners of different forms of ‘mind body medicine’, and representatives of ancient healing traditions. The chapters in this book are not merely ‘proceedings of the conference’, but represent state of the art reviews of different aspects of braidmindhody interactions in health and disease. The invited speakers were not asked to bring their papers to the symposium, but to prepare them after the conference thereby enabling them to incorporate many of the contents of the discussions occurring during the conference. Although the organizers realized that it would lead to some delay in the publication time, thorough peer review of all the chapters was enforced to maximize the quality of the publication.

This Page Intentionally Left Blank

xi

Acknowledgements

The symposium on the ‘Science and Practice of Mind/Body Interactions’ was organized by Emeran A. Mayer (UCLNCURE Neuroenteric Disease Program, Department of Medicine and Physiology, UCLA School of Medicine, Los Angeles, CA 90024, USA) and Clifford Saper (Department of Neurology, Harvard University, Beth Israel Deaconess Hospital, Boston, USA). Contributions .of the following institutions made it possible to organize the symposium, which was held March 15-18, 1998 in Sedona, Arizona: Procter & Gamble Balm Foundation. The organizers are especially grateful to Mrs. Janet Stein for her generous donation, to Mrs. Joyce Fried for organizing the conference and to Teresa Olivas for preparing the manuscripts.

This Page Intentionally Left Blank

...

Xlll

Contents ..................................

v

........................................ Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

List of contributors Preface

xi

I. Introduction 1. Minding the mind E.A. Mayer and C.B. Saper (Los Angeles, CA and Boston, MA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

11. Relationship between mind, brain and emotions 2. Topography and relationships of mind and brain B.A. Vogt and 0. Devinsky (Winston-Salem, NC and New York, NY,USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

HI. The neurobiology of the stress response 3. Protective and damaging effects of stress mediators: central role of the brain B.S. McEwen (New York, NY, USA) . . . . . . . . . . . . . . . .

25

4. Interactions between the immune and neuroendocrine systems E.M. Sternberg (Bethesda, MD, USA)

...............

35

5 . Depression really does hurt your heart: stress, depression and cardiovascular disease D.L. Musselman and C.B. Nemeroff (Atlanta, GA, USA) . . . .

.

43

6. Circuits and mechanisms governing hypothalamic responses to stress: a tale of two paradigms P.E. Sawchenko, H.-Y. Li and A. Ericsson (La Jolla, CA, USA) .

.

61

IK Early life experiences and the developing brain 7. Long-term behavioral and neuroendocrine adaptations to adverse early experience C.O. Ladd, R.L. Huot, K.V. Thrivikraman, C.B. Nemeroff, M.J. Meaney and P.M. Plotsky (Atlanta, GA, USA and Montreal, Que., Canada). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

xiv

8. Neurobiological correlates of defensive behaviors

..

105

...............

117

V.P. Bakshi, S.E. Shelton and N.H. Kalin (Madison, WI, USA) 9. Effects of perinatal pain and stress K.J.S. Anand (Little Rock, AR,USA).

10. Early life abuses in the past history of patients with gastrointestinal tract and pelvic floor dysfunctions G. Devroede (Fleurimont, Que., Canada) . . . . . . . . . . . . . . 131

V. Influences of the internal environment on the brain 11. Responses of afferent neurons to the contents of the digestive tract, and their relation to endocrine and immune responses J.B. Furness and N. Clerc (Parkville, Australia and Marseille, France). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

159

12. The controls of eating: brain meanings of food stimuli G.P. Smith (White Plains, NY, USA) . . .

173

13. Effects of nutrients on brain function T.J. Maher (Boston, MA, USA)

.............

...................

187

14. The evolving neurobiology of gut feelings

E.A. Mayer, B. Naliboff and J. Munakata (Los Angeles, CA, USA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

195

VI. Influences of the body on the brain 15. Integration of viscerosomatic sensory input at the spinal level R.D. Foreman (Oklahoma City, OK, USA) . . . .

.........

16. The medial pain system, cingulate cortex, and parallel processing of nociceptive information B.A. Vogt and R.W. Sikes (Winston-Salem, NC and Boston, MA, USA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17. Pain as a visceral sensation C.B. Saper (Boston, MA, USA)

209

223

. . . . . . . . . . . . . . . . . . . 237

18. Pain modulation: expectation, opioid analgesia and virtual pain H.L. Fields (San Francisco, CA, USA) . . . . . . .

........

19. Mechanisms of analgesia produced by hypnosis and placebo suggestions D.D. Price and J.J. Barrel1 (Gainesville, FL, USA) . . . . . . .

..

245 255

20. The role of vagal visceral afferents in the control of nociception W. Jhig, S.G. Khasar, J.D. Levine and EJ.-P. Mia0 (Kiel, Germany and San Francisco, CA, USA) . . . . . . . . . . . . . . 273

xv

VII. The influence of brain and mind on the body 21. Affect, cognition, the immune system and health M.E. Kemeny and T.L. Gruenewald (Los Angeles, CA, USA) 22. Memory networks in the prefrontal cortex J.M. Fuster (Los Angeles, CA, USA)

...

291

................

309

23. Non-conscious brain processing indexed by psychophysiological measures D. Tranel (Iowa City, IA, USA) . . . . . . . . . . . . . . . . .

. . 317

24. Brain mediation of active and passive emotional coping R. Bandler, J.L. Price and K.A. Keay (Sydney, St. Leonards, NSW, Australia and St. Louis, MO, USA) . . . . . . . . . . . . . 333 25. Specificity in the organization of the autonomic nervous system: a basis for precise neural regulation of homeostatic and protective body functions W. Janig and H.-J. Habler (Kiel, Germany) . . . . . . . . . . . . . 351 26. The impact of emotions on the heart R.L. Verrier and M.A. Mittleman (Boston, MA, USA) . . . . . . . 369 27. Neural influence on immune responses: underlying suppositions and basic principles of neural-immune signaling D.L. Felten (Loma Linda, CA, USA) . . . . . . . . . . . . . . . . 381

VIII. Practical use of mind-body interactions in medicine 28. The cost-effectiveness of mind-body medicine interventions D.S. Sobel (Oakland, CA, USA). . . . . . . . . .

. . . . . . . . . 393

29. Towards an integrative model of irritable bowel syndrome B.D. Naliboff, L. Chang, J. Munakata and E.A. Mayer (Los Angeles, CA, USA) . . . . . . . . . . . . . . . . . . . . . . .

..

30. Bridging the gap between mind and body: do cultural and psychoanalytic concepts of visceral disease have an explanation in contemporary neuroscience? N.W. Read (Sheffield, UK) . . . . . . . . . . . . . . . . . . . . .

413

425

3 1. Complementary and alternative medicine (CAM): epidemiology and implications for research D.L. Diehl and D. Eisenberg (Sylmar, CA and Boston, MA, USA) 445 32. Biological mechanisms of acupuncture D.J. Mayer (Richmond, VA, USA)

.................

457

xvi

33. Intricate tactile sensitivity: a key variable in western integrative bodywork D.H. Johnson (San Francisco, CA, USA) . . . . . . . . . . . . . . 479 34. The science of breathing - the yogic view R. Sovik (Buffalo, NY, USA) . . . . . . . . . . . . . . . . . . . .

491

35. Exploring the nature and functions of the mind: A Tibetan Buddhist meditative perspective L. Rapgay, V.L. Rinpoche and R. Jessum (Los Angeles and Santa Monica, CA, USA and Karnataka State, India) . . . . . . . . . . . 507 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

517

SECTION I

Introduction

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Bruin Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 1

Minding the mind Emeran Mayer’.” and Clifford B. Sape? ’UCWCURE Neuroenteric Disease Program, Department of Medicine and Physiology, UCLA School of Medicine, Los Angeles, CA 90024, USA ’Department of Neurology, Beth Israel Hospital, 330 Brookline Avenue, Boston, MA 02215, USA

Minding the mind For many centuries and in virtually every society, long before the appearance of evidence-based medicine, people with medical problems have turned to healers. It is striking to realize that only a handful of the drugs that were prescribed at the turn of the last century by Western physicians are still in clinical use (opiates, digitalis, aspirin, and quinine). Presumably, most of the rest of what physicians were doing at that time, and what they did for millennia before that, relied upon the healing effects of endogenous physiological systems, activated by state of mind. The universal concepts of traditional healing practices, ranging from the prehistoric shamans to the practitioners of Western Hippocratic, Chinese, Aryuvedic and Native American Medicine include the following beliefs: (a) the belief in a universal life force (pneuma, chi, prana, animal magnetism, vis medicatrix naturae); (b) the belief in the unity between mind, body and universe; (c) the conceptualization of health as a state of harmony between mind and body, and between the organism and nature; and (d) the conceptualization of disease as a loss of such harmony. The role of the healer was seen as a catalyst who uses subtle interventions to stimulate the body’s own healing abilities, with the goal to reestablish harmony and the undisturbed flow of the universal life force. *Corresponding author. Tel.: + 1 (3 10)-312-9276; Fax: + 1 (3 10)-794-2864; e-mail: [email protected]

In the last century, in the movement to compare treatment to placebo, and to relinquish practices that were not superior to placebo, we have moved medical technology forward by enormous strides. At the same time, many patients and some physicians have come to expect that treatments that meet this standard should be available for all conditions, and that evidence-based medicine should be able to provide cures or at least effective treatment for all ills. Alas, this has not proved to be the case. As a result, many patients in disillusionment have sought care from ‘alternative’ healers and treatments, and many have received benefits that they value. The range of ‘alternative’ treatments is enormous, from such generally accepted modalities as acupuncture or relaxation therapies to herbal treatments, ‘high colonics’, magnets and homeopathy to spiritual practices such as prayer and meditation. There is a tendency for many physicians to look down upon these therapies as ‘unscientific’ or at best ‘unproven’, and therefore not acceptable for patient management. Other physicians have recognized that patients often respond to kindness, human contact with perceived healers and the attractiveness of a ‘holistic’ view of health and disease as providing benefits, despite the absence of an effective evidence-based therapy. Interestingly, most of the popular ‘alternative’ healing practices (from homeopathy to acupuncture, body work, hypnosis and Native American healing) share ‘holistic’ views. These views are reminiscent of the historic concepts of health as a

4

state of internal and external harmony, self-healing abilities of the body and the health promoting effects of re-establishing the flow of ‘energy’ within the bodylmind continuum. The disparity between patients’ preferences and demands and traditional belief systems of health care providers has produced a national debate on the validity of ‘alternative’ medicine, and induced a major investment by the NIH in finding which parts of it may include therapies that can pass the test as evidence-based, meaning superior to placebo. Not only has the NIH recently established an Institute of Comprehensive and Alternative Medicine, but major resources have been made available to establish national NIH-sponsored Mind Body Research Centers. The current wave of interest in ‘holistic’ (i.e. mind-brain-body)medicine practices has been generated by an unlikely alliance of forces, including breakthroughs in neuroscience, economic considerations of managed care organizations, a popular demand for more natural healing approaches and a general political concept of self reliance (‘self-efficacy’) applied to health and disease. In this volume, which represents the proceedings of a ‘Mind-Brain-Body Medicine’ symposium held in Sedona, Arizona in March of 1998, we address key issues relevant to the ongoing debate over the validity of traditional, holistic, alternative or complementary approaches to health, disease and healing. Even though the placebo effect is considered Western Medicine’s eternal and most formidable competitor, it is agreed by all that placebo is effective at relieving pain and suffering. In fact, some of our best and most established therapies fall far short of the 3040% placebo response rates seen in controlled clinical trials in many chronic disorders. In the context of MindBrain-Body Medicine, the placebo response (even though poorly understood) is the manifestation of the remarkable self-healing abilities of the organism against a wide spectrum of diseases, ranging from coronary heart disease and cancer to rheumatoid arthritis and chronic pain. If we are honest with ourselves, practices do not necessarily have to be better than placebo to find a useful place in medical practice, and honest practitioners have used this approach to the benefit of their patients for many

years. Conversely, practitioners who eschew the healing arts in favor of only offering their patients scientifically based therapies may be justifiably accused of doing their patients a serious disservice. In this volume we are asking the question of how the mind and body interact with each other and with the environment, from the first days of life to adulthood, and in this process actively maintain health and prevent disease. We are examining mechanisms by which early disruptions of the mindhodylenvironment homeostasis can predispose the individual to chronic disease. Finally, we are asking the question how a few selected ‘alternative’ practices, either aimed at the body or at the mind, are able to re-establish balance and thereby a state of health. This volume is divided into two parts. In the first part (sections I1 to VII), we examine the physiological basis for the bi-directional interactions between mindibrain and the body. The scientists asked to review this material are outstanding and internationally recognized leaders in physiology and anatomy, and we received a gratifying response. An overwhelming number of them, when contacted about this symposium and volume, were prepared to bring an open mind, and a thoughtful approach to these issues. The results of their labor is one of the most complete reviews of the scientific basis of mind-body medicine extant. The second component of this volume and the meeting (section VIII), involves the work of various practitioners of ‘mind body medicine’. Again, the response and interest of this group, all chosen for their international prominence and serious scholarship in their fields and representing a wide cross-spectrum of practices, was remarkable. Even more remarkable was the interactions between these two disparate groups of individuals, who under other circumstances would not be likely to meet, let alone spend several days listening to one another’s viewpoints. The emerging principles of mind-body interactions, which came from this meeting, form the core of this book. The different programmatic sections of the book address different aspects of the mind/brain/body interactions in health and disease. The first two sections (I11 and IV) deal with neurobiological mechanisms under-

5

lying the organism’s response to stress and how this response is altered by events occurring early in life. The next two sections (V and VI) address different aspects of how the body (from its internal and external interfaces with the environment) communicates to the brain, and section VII reviews different mechanisms by which the mindhrain communicates back to the body. Finally, in section VIII, authors from widely differing disciplines address aspects of the practical application of mind body approaches to health and disease. In the first section of the book, Vogt and Devinsky address one of the key issues of this volume: where in the brain is the mind located? Following a review of theoretical considerations, reports about patients with brain lesions and recent evidence from brain imaging studies the authors propose that the fundamental components of mind are located in the cingulofrontal and parietotemporal confluence regions, that the mind is lateralized and that the distributed mental activity is associated with a unified mind through various binding mechanisms. Section III addresses the neurobiological mechanisms underlying the response of an organism to different types of stressors. This integrated response is mediated by the emotional motor system, with its neuroendocrine, autonomic and antinociceptive arms. B. McEwen reviews the biological evidence demonstrating how the attempt of the organism to adapt to chronic stress can result in a maladaptive response with harmful effects on health. He calls the accommodation to chronic stress the allostatic load, referring to the wear and tear that results from chronic overactivity or underactivity of systems that try to achieve stability through change. The effects of hyper- and hyporesponsiveness of the neuroendocrine arm of the stress response, the HPA axis, on the immune system is discussed by E. Sternberg. She provides evidence to suggest that while hypo-responsiveness of the HPA axis appears to predispose to overactivity of the immune system, overactivity of the HPA axis, predisposes to enhanced susceptibility to infectious disease. D. Musselman and C.B. Nemeroff discuss the evidence and the underlying mechanisms which link affective disorders, in particular major depression with the development

of atherosclerotic heart disease and with the increased morbidity and mortality after an index myocardial infarction. They identify hyperactivity of the HPA axis, increased sympathoadrenal responses and diminished heart rate variability (corresponding to decreased cardiovagal tone) as major components in providing the link between the mind and cardiovascular disease. The question how different types of stressors result in specific hypothalamic responses to stress, even though both types of stressors result in activation of HPA axis and sympathoadrenal systems is addressed by P. Sawchenko. By comparing the cellular activation patterns in the CNS in response to acute cytokine (systemic or interoceptive stressor) and footshock challenges (neurogenic or exteroceptive stressor), he demonstrates that while interoceptive stress effects on hypothalamic (PVH) effector neurons may be conceived as simple reflex responses, exteroceptive stress effects mediate adaptive visceromotor responses. Section IV deals with the effect of environmental factors occurring early in life of an organism which are able to permanently alter the neurocircuitry of the developing central nervous system, in particular the different components of the emotional motor system. Such permanent biasing of the mind body relationship and the way an organism responds to environmental stressors occurring later in life is likely to have important consequences for the mental and physical health of the affected adult. As illustrated by G. Devroede for the relationship between physical and sexual abuse and organic and functional gastrointestinal disorders, and by K. Anand for neonatal stress and pain, early traumatic life events serve as major risk factors for chronic disease later in life. P. Plotsky and co-authors describe a rodent model of perinatal stress in the form of moderate maternal separation. The changes in neurocircuitry resulting from this type of perinatal stressor are associated with the expression of enhanced anxiety-like behavior, anhedonia, alcohol preference and HPA axis hyper-responsiveness to psychological stressors in the adult animals. The evidence related to the developmental biology of pain in man is reviewed by K. Anand. His review focuses on the increased pain sensitivity in neonates, the development of hyperalgesia following

6

an acute noxious stimulus and on the neurobehavioral and developmental consequences of neonatal pain experiences. N. Kalin and co-authors summarize work in non-human primates related to the ontogeny and neural substrates of defensive behaviors, an organism's behavioral response to fear. The authors present a hypothesis which conceptualizes pathological anxiety as the inappropriate expression of fear-related behaviors, related to specific neural substrates. Section V deals with influences of the internal environment on brainfmind functions. Food, diet and special dietary supplements, including herbs and animal products play an important role in most traditions of mindhody medicine. Due to its large and leaky surface, the gut is protected against external influences by distinct subsystems of the immune, the nervous and the endocrine systems. The unique role of the gastrointestinal tract and its sensory innervation as an interface with the environment is discussed by J. Furness and N. Clerc. A significant amount of preprocessing of this sensory information is achieved at the level of the gut, before it reaches the central nervous system. G. Smith discusses the process of eating, starting from the rhythmicity of eating related to the activity of a central pattern generator and the regulation of this central mechanism by intestinal feedback. Even though many of the reflex responses involved in eating occur without conscious awareness, there are important effects of food intake on the conscious mind, including pleasure and tranquilization. The evidence that dietary components, primarily in form of precursors of different amino acids can cross the blood brain barrier and have distinct effects on a variety of behavioral and affective responses is reviewed by T. Maher. E. Mayer and co-workers discuss evolving concepts about the role of visceral afferent information in modulating mental functions, including both affective and cognitive aspects. Starting with the lay concept of 'gut feelings' they discuss recent results from functional brain imaging studies identifying distinct brain regions which are activated in association with perception of and autonomic responses to visceral stimuli. Section VI deals with topics related to sensory feedback from the body to the brain and the ways

by which the brain is able to modulate the subjective experience of pain. R. Foreman discusses the integrative role of the upper cervical spinal cord in processing sensory information from distal visceral organs, providing not only a way for intraspinal communication between different viscera but also for the convergence of somatic and visceral input. Specifically, he discusses the possible role of convergence of craniovascular afferent input and somatic input in the cervical spine in the etiology of referred headaches. B. Vogt and R.W. Sikes discuss the role of the medial pain system as a mental process which is primarily aimed at predicting, rather than localizing the outcome of life-threatening events. Drawing on extensive evidence from studies in animals and functional brain imaging studies in humans they describe the role of subregions within the anterior cingulate cortex, projections to other cortical and subcortical regions in the experience of, response to and prediction of nociceptive events. C. Saper proposes that pain is not aimed at exploring the properties of objects of the outer world but rather pain guards against tissue damage during that exploration. According to this view, pain is not a primary sensory modality but rather a visceral sensation concerned with the internal body, beginning one cell layer under the cornified epidermis. H. Fields reviews the role of endogenous pain modulation systems in altering the experience of pain. Inhibitory and facilitatory systems act on the subjective experience of pain which is produced by a specific spatiotemporal pattern of neural activity called a representation. These systems are able to produce analgesia under conditions of threat virtual pain under situations of suggestions and expectations. The different mechanisms of analgesia produced by hypnosis and placebo suggestions is reviewed by D. Price and J.J. Barrell. In the case of hypnotic analgesia, these mechanisms include inhibition of afferent nociceptive signals originating at the level of the spinal cord, as well as cerebral mechanisms selective for modulation of pain affect with the anterior cingulate cortex. In the case of placebo analgesia, these mechanisms include a change in affective state, response bias and activation of descending control systems. The authors suggest that these differences could be useful in the design of specific therapeutic

7

approaches to chronic pain. W. Janig and coworkers discuss the role of subdiaphragmatic vagal visceral afferents in the modulation of mechanical hyperalgesic behavior. They provide evidence that the effector pathway of this antinociceptive response involves an endocrine signal from the adrenal medulla. Their results show how the brain is able to regulate sensitivity of nociceptors in the periphery by an endocrine signal and how the sensitivity of nociceptors can be influenced by changes in body parts which are remote from the location of the sensitized nociceptor. Section VII deals with different aspects of the way the mind/brain communicates with the body. These aspects include the role of belief systems in mediating effects on health and disease, the role of brain regions and networks in temporal organization of action, in unconscious emotional processing, in different emotional copings styles, in function-specific autonomic regulation, including autonomic regulation of the heart and of the immune system. M. Kemeny discusses the evidence supporting the links between the mind and the immune system, focusing on newer evidence that cognitive and emotional states have immunological correlates. She provides evidence suggesting that specific cognitive factors (such as negative expectancies, self-blame, psychological inhibition) may be related to immunological processes relevant to health. The role of the prefrontal cortex, which is at the top of the perception-action cycle, in the mediation of contingencies of action across time, an essential aspect of temporal organization is discussed by J. Fuster. He reviews the evidence showing that the role of cross-temporal mediation is based on the interplay of two shortterm cognitive functions of the prefrontal cortex: one retrospective, of short-term active perceptual memory, and the other prospective, of attentive set. D. Tranel reviews evidence supporting the range of neural and cognitive operations that support the part of mental life that takes place beneath the level of conscious awareness. He provides experimental evidence showing that non-conscious processing can proceed independently from conscious operations, that the two modes appear to have different neural substrates, and that non-conscious processing can trigger emotions which provide key

influences in reasoning and decision-making. Functional neuroanatomical evidence supporting the concept that the lateral and ventrolateral columns of the PAG coordinate opposite modes of emotional coping in the form of engagement or disengagement with the external environment is reviewed by R. Bandler and coworkers. They also present evidence for function-specific networks involved in different patterns of emotional coping, made up of projections between discrete subregions of prefrontal cortex, PAG, amygdala and hypothalamus. W. Janig and H.-J. Habler review neurophysiological evidence which demonstrate the function-specific organization of sympathetic outflow from the central neuroaxis to the body. This organization is the basis for precise neural regulations of all homeostatic body functions, and is in sharp contrast to the traditional concept of the functioning of the sympathetic nervous system as an all-or-none system, stereotypically activated in the context of stress. R. Verrier and M.A. Mittleman discuss distinct physiological processes which play a role in the mediation of potentially life threatening effects of stressors and emotions in patients with ischemic heart disease. These effects are mediated by output patterns of the autonomic nervous system characterized by high sympathetic and low vagal tone. They discuss a variety of new diagnostic methodologies aimed at identifying patients at increased risk for stress-induced cardiac vulnerability. Aspects of neural-immune signaling, in particular the role of norepinephrine in modulation of immune function is discussed by D. Felten. He suggests the possibility that both behavioral and pharmacological manipulation of sympathetic nerve to immune cell signaling may be useful for treating certain immune-mediated diseases. Section VIII deals with applications of mind body medicine concepts in the clinical setting where D. Sobel gives an overview of mind body medicine interventions from a health economics point of view. He makes a strong argument that simple, safe and relatively inexpensive behavioral medicine interventions can dramatically improve health outcomes and reduce the need for more expensive treatments of both chronic organic disease and of so-called functional disorders. Costeffective mind body interventions, particularly

8

when offered to groups of patients, are likely to become important strategies for health care delivery systems to deal with escalating costs, rising public expectations, limited access and an aging population with multiple chronic illnesses. B. Naliboff and co-authors focus on a common functional disorder, irritable bowel syndrome, to present a biobehavioral model which integrates neurophysiological, perceptual and behavioral processes. This model may be generalizable to other common functional syndromes, such as fibromyalgia, dyspepsia and chronic pelvic pain, and forms a rationale basis for effective behavioral treatment strategies. An alternative model of functional disease, which is based on a psychoanalytical understanding of how disease may result from the impact of psychosocial stresses on a fragile and vulnerable personality structure is presented by N. Read. Even though the psychoanalytical model and the preceding biobehavioral model arise from fundamentally different traditions of psychology, in the end, they both converge with neurobiological concepts presented in this volume. The issue of complementary and alternative medicine (CAM) is addressed by D. Diehl and D. Eisenberg. They give a brief overview of the different approaches that are usually referred to as ‘alternative’,and give information about the widespread utilization of such treatment options by the US population. The authors also present the type of support that investigations into the effectiveness of CAM treatment options have recently received from national funding agencies. D. Mayer focuses on one of the

most established ‘alternative’ treatment strategies, acupuncture, and discusses possible neurobiological mechanisms, in particular the role of endogenous opioids underlying the effectiveness of such treatments. He also addresses the problems of how to evaluate the clinical efficacy of acupuncture as a treatment modality for various medical conditions. D. Johnson discusses the concepts and general therapeutic strategies shared by different schools of Western Integrative Bodyworks. The author suggests that this ‘Intricate Tactile Sensitivity’ which includes the therapist’s sensitivity for certain patterns of the patient’s body, and a unique contact between therapist and patient may become amenable for scientific evaluation. The last two chapters address concepts related to mind body interactions developed by the great Eastern traditions of Aryuvedic and Tibetan philosophy. The central role of breathing in the practice of yogic techniques is reviewed by R. Sovik. Within yogic traditions, voluntary control of breathing is used to foster self-awareness and to reduce unnecessary autonomic arousal, thereby exerting important effects on physiologic and psychologic functioning. L. Rapgay and Lati Rinpoche review some key aspects of the Tibetan view of the mind, and discuss a series of meditative practices are aimed at strengthening the abilities of the mind. The Tibetan Buddhist view of the mind, its divisions and the elaborate practices to achieve ideal states is a complex presentation of psychological, behavioral and spiritual concepts that requires thorough understanding before it is possible to assess their value.

SECTION I1

Relationship between mind, brain and emotions

This Page Intentionally Left Blank

E.A. Mayer and C.B.Saper (Eds.) Progress in Brain Research, Vol 122 0 ZOO0 Elsevier Science BV. All rights reserved.

CHAPTER 2

Topography and relationships of mind and brain Brent A. Vogt’**and Orrin Devinsky2

’ Cingulum NeuroSciences Institute, 101 N. Chestnut St, Winston-Salem, NC 27101, USA and Wake Forest Universiv School of Medicine, Medical Center Blvd, Winston-Salem, NC 27157, USA ’New York University-Mount Sinai Comprehensive Epilepsy Center, 560 First Avenue-Rivergate 4th Floor, New York, W 10016, USA

Introduction Although the primary neuroscience literature does not directly address the mind, mentalistic perspectives are often used to frame broad issues of brain function. These concepts include attention, perception, stress, emotion, and pain. In order for neuroscience to participate in the Mind/Body debate, a physical definition of mind and its relationship to brain function is needed. The concept of mind considered by Descartes, Sherrington and others was an introspective view of self and emphasized the unitary and continuous nature of mental activity. Early efforts to distinguish the mind from the body, however, led to a dualism in which the physicalheurological origin and constraints on mind were not apparent. This view still dominates Western views of the mind even though the dependence of mental activity on internal and external events makes such a distinction counterproductive. As the Mind/Brain problem is clarified, the mechanics of Mind/Brain/Body interactions are open to discourse, investigation, and hypothesis testing. Some critical questions for neuroscience include the following: To what extent can the properties of mind be specified? Is there a neural basis for the theory of mind? To what extent is *Corresponding author. Tel.: 336-716-8588; Fax: 336-7 16-8501 ; e-mail: [email protected]

mental activity localized or distributed in the brain? How do various processing modes and links with executive systems relate to consciousness? La Peyronie in 18th century France sought relations between mind and brain scientificallywith postmortem tissue (Kaitaro, 1996). Although one may argue over his conclusions, he was among the first to presume that mind did not have a uniform distribution in the brain. The next two centuries of neurological observations and the recently acquired body of functional brain imaging under highly controlled neuropsychological conditions provide specific information about relationships between mind and brain that are pivotal to understanding Mind/Body interactions. The mind is often defined as ‘awareness of self’, yet this concept itself can be misconstrued. If the mind is solely awareness of self, the mind might be expected to thrive outside the context of the body including neuronal activity in the brain. However, isolated, premature babies fail to gain weight and thrive (Schanberg and Field, 1987) and sensory stabilization experiments in isolation chambers lead to reduced intellectual performance, altered perceptions, hallucinations, and emotional changes (Bexton et al., 1954; Doane et al., 1959; Davis et al., 1960; Solomon et al., 1961). The notion that ‘self’ can exist independent of internal and external events is an extreme conclusion based on the philosophical MindEIody duality and is not a part of neuroscience discourse.

12

At the other extreme, one might posit that all sensory and motor activity is mental activity and that the mind is equivalent to activity in most of the brain (e.g., Kandel, 1991). Spinal cord injury followed by appropriate rehabilitation, however, does not impair mental activity. Furthermore, Crick and Koch (1995, 1998) have argued that mammals are not aware of events in primary visual cortex and that explicit interpretation of visual scenes occurs in visual and polymodal association cortices. Logothetis and Schall (1989) provide elegant support for this notion with single unit recordings in monkey at different points along the visual cortical processing hierarchy. Although many neurons were driven by the retinal features of a stimulus, those in cortex of the superior temporal sulcus appeared to be driven by the perception of motion. Finally, synthesizing the visual image may depend on simultaneous and synchronous activity in primary visual and visual association areas because patients with blindsight following primary visual cortical lesions have non-conscious access to visual information but are unaware of what they can see (Devinsky, 1997). Since mental activity may not extend to the spinal motor neuron and primary sensory cortices, the mind localization problem is a question of finding the essential networks that mediate mental activity in the forebrain. These networks comprise the focal point of Mind/Body research and therapy. The importance of neuroanatomical organization in assessing awareness and mental functions has been explicitly stated by Crick and Koch (1995) and supports the anatomical orientation of this review. Since it is unlikely that the entire brain is engaged equally in mental activity, the mind has a high probability of being located in parts of the cerebral cortex that underlie perception and mental activity. Neuroscientific observations have not established the relationships between the mind and consciousness and there is no strategy for identifying the mind much less to localize it. One theoretical approach is to search for the neuronal correlates of consciousness as suggested by Crick and Koch (1995), however, such information is not available and it is unlikely that current methodologies for multiunit recording will succeed at this task. Frith (1 992) suggests that understanding

consciousness in terms of information processing requires a cognitive mechanism that is constantly associated with consciousness, whether or not the processing mechanism itself is available for observation. This may require coordinated activity in a number of distributed processing modules and/or persistent processing in one or two profoundly important regions. Several models relating specialized processing modules, phenomenal consciousness, executive and other functional and cognitive subsystems have been proposed (Block, 1995). One of the problems of neuroscience is to identify the morphological substrates in these and similar models and to assess interactions among their processing subsystems. They are, however, beyond the scope of this assessment. The present analysis considers the localization of mental activity in cortical regions of confluence among cytoarchitectural areas. It suggests that the binding problem may be solved by understanding the interconnections of these regions. Finally, the location of mind per se and mechanisms of consciousness are not addressed here. Instead, the high probability of association between mental activity and the mind is used here to suggest where the mind is located in the brain.

Mental activity and the cortical confluence regions In order to avoid the pitfalls inherent in a strict Mind/Body dualism, the definition of mind can be extended to the following: Mind is mental activity associated with internal and external awareness, intentions, and memories of self.

Each of these components are critical to mental activity and this perspective does not make mind dependent on a sensory modality or motor output. Indeed, the intention to move is an essential activity of mind, not a particular movement itself. Similarly, long-term memories establish the motivational context of intentions or willed actions. Although little is known about the very long-term storage of self-oriented information in the brain, lesion studies indicate that some regions are more important to internal and external awareness of self

13

than are others, while functional imaging lends insight into localizing the intentions, imagery, and logical activities of mind. As stated earlier, it is unlikely that the entire cerebral cortex participates equally in mental function. Additionally, it has been known for some time that serial processing of cortical information from single sensory modalities to ‘higher order’ integration sites does not occur in a fashion that can account for unified perceptions and a coherent awareness of self. Efforts to solve the binding problem emphasize the failure of serial processing models. We propose that there are two sites in the cortex that fulfill the definition of mind provided above. These are the cingulofrontal and parietotemporal confluence regions. The cingulofrontal and parietotemporal confluence regions bind the internal and external features of self and employ the various motor and cognitive systems to implement intentions.

These sites are probably not involved in binding features into entities as discussed by Damasio (1989) nor do they address the broader issues of binding throughout the cerebral cortex (Sejnowski, 1986). The neural codes from early convergences and memories of self are used by the cingulofrontal and parietotemporal convergence regions to implement the intentions of self.

Localizing mental activity with lesions Anterior cingulate cortex is essential for mental activity. Bilateral lesions of the anterior cingulate gyri produce akinesia, mutism, indifference to noxious stimuli and surroundings, incontinence, and lack of spontaneity (Nielsen and Jacobs, 1951; Barris and Schuman, 1953). Akinetic mute patients appear conscious, their eyes are open, deep tendon reflexes and muscle tone may be normal, but these patients show minimal or no spontaneous motor or verbal behavior. They lack intention and willed actions. Other evidence supports the hypothesis that anterior cingulate cortex is essential for mental activity and initiation and regulation of behavior. Cannon and his colleagues emphasized the critical

role of anterior cingulate cortex in regulating aggressive behaviors that are initiated in the midbrain (Bard and Mountcastle, 1948). Medial cortical lesions remove sensory guidance and control of reflexes and suggest a cortical localization of mind. Glucose metabolism is reduced in perigenual cingulate cortex in depressed individuals and may indicate that this region is involved in the disease and, hence, the initiation of behavior (Drevets et al., 1997; Mayberg, 1997). Finally, seizure activity in midcingulate cortex (area 24’) is associated with obsessive-compulsive disorder and this behavior resolves following surgical removal of this region (Levin and Duchoway, 1991). The crucial role of anterior cingulate cortex to willed actions and mental activity is provided by observations of individuals with unilateral lesions in this region following occlusion of the proximal part of the anterior cerebral artery. Such lesions resulted initially in mutism and paucity of movement or akinesia, although these patients were alert and oriented to time and place (Chan and Ross, 1997). Two to four weeks after the vascular event there was a continued reduction in spontaneous speech which was monotonic as well as reduced contralateral movements that did not appear to have been consciously driven by the patient such as those of the hand. In fact, one patient with a large right anterior cingulate/supplementary motor cortex infarct reported that his left hand was alien because it operated ‘against his will’. One interpretation of these findings is that his willed actions were generated in his left anterior cingulate cortex. Review of a broad neurological and neurobiological literature (Devinsky et al., 1995) suggests that anterior cingulate cortex is pivotal to internal and external awareness of self and intentions. Many of the neurological impairments suffered by Phineas P. Gage may have been due to damage in the anterior cingulate and adjacent orbitofrontal and dorsomedial prefrontal cortices. Although his sensorimotor faculties were intact and his linguistic and cognitive functions fairly well preserved, he could not select among appropriate responses, had an altered personality, and was socially inappropriate. He was described as fitful, irreverent, indulging at times in the grossest of profanity, impatient, vacillating, a child in his intellectual

14

capacity and manifestations. Recent analysis of the skull and tamping iron that produced the cortical damage in Gage’s brain has been made by Damasio et al. (1994). This reconstruction of the placement of the tamping iron in the brain suggests that the primary lesion site in Gage’s brain was in perigenual cingulate cortex and adjacent orbitofrontal and mediodorsal prefrontal areas. Since sociopathic behaviors and altered personality have also been reported in patients with medial and orbitofrontal cortex damage (Damasio et al., 1990), this region appears to be necessary for mental activity. Contralateral neglect, impaired volition and movement, and reductions in spontaneous speech are not unique consequences of lesions in anterior cingulate cortex. Although an account of lesion size and placement must be made, lesions in posterior parietal and prefrontal cortices can produce these effects. Neglect following frontal and parietal lesions may be specialized such that ophthalmokinetic neglect is more frequent following parietal lobe lesions, while melokinetic neglect more often follows frontal lesions (Bisiach et al., 1995), and both regions have also been implicated in intention (Stuss and Benson, 1984; Andersen, 1995). Prefrontal lesions can produce motor neglect and interfere with specific executive functions (Rizzolatti et al., 1983) and, to the extent that they disrupt an internal sense of self and produce apathy and denial of illness (anosognosia), are likely critical for mental functions. Although large lesions of prefrontal cortex severely impair mental activity, it is possible that the mind is much less impaired by loss of small parts of lateral prefrontal cortex. The extent to which subregions in frontal and parietal cortices contribute to mental function and how this relates to activity in anterior cingulate cortex can be addressed to some extent with functional imaging techniques as discussed below. Failure to produce deficits associated with intentions and self awareness following lesions in the parietotemporal region may be due to lack of equivalent and bilateral destruction in most cases, whereas bilateral destruction of homologous regions in anterior cingulate cortex is more likely due to the pattern of blood flow impairment and hemorrhage produced by anterior cerebral and communicating artery lesions.

Before moving to functional imaging assessments, let us contrast the neurological deficits following damage in regions that are essential to mental activity with those that are not essential to the internal and external awareness of self. One dramatic instance of an ineffective ablation is that in retrosplenial cortex and possibly the fornix after removal of an arteriovascular malformation (Valenstein et al., 1987). Following surgical recovery, this individual could not return to work because of persistent anterograde memory impairment. Although he was oriented to time and place and had relatively intact remote memory, recall of events over the past four years was difficult. Remarkably, his general intellectual functions, language, and praxis were intact. There are no reports of impaired executive or other ‘frontal lobe’ functions including personality, behavioral initiation or suppression problems. It seems, therefore, that cortical and subcortical regions involved in short-term memory formation may limit the range of new behaviors but they are not essential for mental activity and personality per se. Some structures involved in short-term memory that may be viewed as nonessential to mental activity include the anterior thalamic nuclei, mammillary bodies, retrosplenial cortex, parahippocampal cortex, and the hippocampus. Similarly, lesions in primary and association sensory and motor cortices impair restricted aspects of perceptual or motor behavior, but not of mind or personality.

Localizing mental activity with functional imaging Although brain lesion outcomes implicate many parts of the cerebral cortex in mental activity, localization is difficult in highly interconnected and distributed processing systems. Damage to a single component of a distributed network may reflect composite deficits associated with direct damage as well as disruption of one or more deafferented areas. In many instances cortical lesions may not be equal in extent and bilateral providing for recovery of function. Although bilateral lesions of anterior cingulate cortex are available, for example, it is heavily interconnected with prefrontal area 46 and these two regions are frequently but not always

15

coactivated in functional imaging studies. In contrast to brain lesion studies, functional imaging provides localization information for repetitive tasks and methods for assessing subfunctions of components in a distributed network. Information gleaned with this strategy, however, is influenced by signal smoothing, subtraction protocols, system habituation, signal averaging, and statistical criteria. Each of these analytical issues raises difficulties in defining the limits of an active region. This strategy also requires that activated sites be uniform in each case in a standardized stereotaxic space; a requirement that is often not met in the highly variable human brain. In spite of these shortcomings, many testing paradigms support the contention that anterior cingulate and adjacent medial prefrontal cortices as well as parietotemporal cortex are pivotal to mental activity. Activation of anterior cingulate and dorsolateral prefrontal cortices during a visually guided, divided attention task led Corbetta et al. (1991) to suggest that these two regions are involved in response selection in cognitively challenging situations. The proposition of these authors that anterior cingulate cortex was engaged in response selection was a pivotal step toward understanding the contribution of this region to brain function. Support for this hypothesis was provided later in a verbal response selection task (Raichle et al., 1994) as well as during the standard Stroop (Pardo et al., 1990; Derbyshire et al., 1998) and counting Stroop (Bush et al., 1998) tasks. In addition to activation of midcingulate cortex in the various Stroop paradigms, other cortical areas are activated including prefrontal areas 46, 44/45, 9 and 10, the inferior parietal and retrosplenial areas, and the anterior insula. The persistent involvement of midcingulate cortex and its extension to dorsal perigenual area 32 in functional imaging studies emphasizes the necessary contribution of this region to response selection. Another perspective on response selection is embodied in the studies of willed actions (Frith et al., 1991). When responses in a response selection task were open-ended and involved a deliberate choice, blood flow increased in anterior cingulate areas 24 and 32 as well as prefrontal area 46 and posterior temporal cortex. Response selection/

willed action tasks require assessment of the motivational significance of a particular outcome and working memory during the selection process. Such a dual function likely requires involvement of anterior cingulate cortex in response selection for motivationally relevant cognitive and behavioral outcomes. The massive and reciprocal connections between these anterior cingulate and prefrontal cortices are well known (Vogt et al., 1979; Baleydiere and Mauguiere, 1980; Goldman-Rakic et al., 1984; Vogt and Pandya, 1987; Barbas and Pandya, 1989). Such connections suggest that these areas often operate in parallel, since the response selection function requires an intact working memory and contributions from prefrontal cortex (Goldman-Rakic, 1987; Fuster, 1995). During motor imagery, Decety et al. (1994) observed activation of anterior cingulate areas 24 and 32, lateral prefrontal areas 6, 9, 8, and 46 and inferior parietal area 40. These authors noted that supplementary motor cortex was not active during motor imagery suggesting a functional dissociation of medial and lateral premotor areas. In this context, it is interesting that Bancaud and Talairach (1992) stimulated midcingulate cortex and produced the desire to leave the room; cognitive activity that precedes the actual formation of a movement strategy and belies the motivational functions of midcingulate cortex. Thus, an important part of establishing motivationally relevant responses is associated with imagery, motivational significance of motor outcomes, and goal orientation (Vogt et al., 1997). Another strategy for localizing mental activity is a consideration of those structures involved in the logic of mental activity. Goel et al. (1997) engaged subjects in deductive and inductive reasoning problems involving three sentences that did or did not, respectively, support a particular conclusion during scanning sessions. They found that subtraction of blood flow during deductive reasoning from that during inductive reasoning showed elevated blood flow in medial prefrontal cortex including areas 8, 9, 24, and 32. In addition to localizing the site crucial for inductive reasoning, they did not coactivate prefrontal area 46 or lateral parietotemporal areas. This is one of the few instances in which anterior cingulate and prefrontal cortical

16

activity has been dissociated and suggests the essential role of the former region in mental activity. The close association between mental processing and inductive reasoning may be due to the persistent use of inconclusive information to interpret the significance of sensory information and predict outcomes. One facet of mental activity is the ability to interpret the mental state of others. Mentalizing story comprehension tasks activate medial cortex between areas 8 and 9 (Mazoyer et al., 1993). ‘Theory of mind’ stories that require attribution of mental states to characters also activate medial area 8 and midcingulate cortex (Fletcher et al., 1995). The latter activation was demonstrated by subtracting blood flow produced during ‘physical’ stories that did not require mental state attribution or stories with unconnected sentences from blood flow evoked by the ‘theory of mind’ stories. Although posterior cingulate cortex was additionally activated in the latter study, this does not mean that areas 23 and 3 1 are involved in mental activity but, rather, that the stories themselves require a personal orientation in space that is organized in posterior cingulate cortex. Hence, area 23 has neuronal discharges coded for large visual fields (i.e., picture content stimuli and grating rather than simple shapes and other features; Olson et al., 1993). Hirono et al. (1998) have shown a relationship between glucose hypometabolism in posterior cingulate cortex in Alzheimer’s disease and spatial disorientation. Therefore, coactivation of posterior cingulate cortex in mental imagery and mentalizing tasks likely represents a visuospatial orientation rather than a function of mind per se and, if the posterior cingulate cortex were missing, it is likely that subjects could still perform ‘theory of mind’ tasks.

The cingulofrontal and parietotemporal confluence regions: primary processors of the mind Lesion and functional imaging studies together show that the primary sites of mental activity are the confluence regions between cingulate and medial prefrontal cortices and that between lateral parietal and temporal cortices. These two con-

fluence regions are outlined in Fig. 1 as are the relevant cytoarchitectural areas. Although many other areas may contribute to activity in these regions including lateral prefrontal and posterior cingulate cortices in the normally active brain, lesion and imaging studies do not support a pivotal contribution to mental activity of these areas. The confluence regions, however, may not be sufficient for mental activity or normal functioning of the mind because complete uncoupling of them from internal and external stimuli as well as skeletomotor and visceral outputs would render these regions and the mind dysfunctional. Functional imaging studies emphasize unique or coactivation of a number of areas in the rostra1 and medial cerebral cortex. These include areas 8,9,32, 24, and 24’. This region may participate in the activity of many networks to the extent that even ‘purely sensory’ processing may require significance coding and an assessment of motivational relevance. It is highly interconnected with lateral prefrontal area 46 that provides information for working memory and the temporal binding of behavioral sequences and plans (Fuster, 1995). Thus, the cingulofrontal confluence region provides the data upon which response selection among motivationally relevant cognitive and behavioral outputs is made. Activity in other cortical areas may not be required for reasoning, willed actions, and the motivational relevance of information processing. Certainly the amygdala has significance codes for simple sensory stimuli that require stereotypical reflexes. However, when these responses need modification or complex interpretations and selection among responses, cortical inputs are employed. Since the cingulofrontal and parietotemporal confluence regions appear to be critical for decision making in relation to the internayexternal and motivational parameters of self, they are essential links in networks engaged by mental activity for processing self-significant information. Therefore, fundamental components of mind are located in the cingulofrontal and parietotemporal confluence regions.

Lateralization of mental activity The right hemisphere may dominate awareness and image of self and the relation of self, visuospatially

17

Fig. 1. The medial and lateral surfaces of the human brain are shown with Brodmann’s areas. The two arrow pairs on the medial surface indicate cortex that was warped dorsally to account for opening part of the cingulate sulcus; the fundus of which is marked with a dotted line. The dashed line at the splenium of the corpus callosum represents the fundus of the callosal sulcus and shows that the retrosplenial areas 29 and 30 are in the depths of this sulcus. The two shaded regions are an approximate outline of the regions that we propose dominate mental activity and, as such, represent the primary processors of the mind as discussed in the text. The cingulate motor areas in the depths of the cingulate sulcus and visceromotor control cortex in area 25 are not viewed as part of the cingulofrontal confluence area that is necessary for mental function.

and psychically, to the environment (Devinsky, 1997). Acute lesions of the right hemisphere more severely disrupt the sense of self than left-sided lesions. Right hemisphere lesions can result in failure to recognize profound deficits as noted above, such as cortical blindness or left-sided hemiplegia (anosognosia),respond appropriately to recognized deficits (anosodiaphoria), or attend to the left half of extrapersonal and personal space. They can also cause the delusional belief that duplicate persons are impersonating well-known persons or delusional reduplications (Ruff and Volpe, 1981; Malloy et al., 1992; Signer, 1992). The right hemisphere may dominate our sense of self and lesions in the right parietal lobe can impair our body image. Left-sided neglect, anosognosia, and anosodiaphoria can be explained by destruction of a module of cortex controlling body image and physical relations of self to the environment. Although the left hemisphere has been most often implicated in language, the right hemisphere

has a role in prosody or interpreting the meaning of words by the way in which they are stated (Bryden and Ley, 1983). Goldberg and Podell (1995) discuss some of the observations that lead to the suggestion that the right hemisphere is critical for exploratory processing of novel cognitive situations for which there are no pre-existing codes or strategies, whereas the left hemisphere is critical for processing based on pre-existing representations and routine cognitive strategies. As they point out, the novelty-routinization hypothesis emphasizes the importance of instructional biases that accompany a task. An ambiguous task is likely to depend on the right hemisphere. The novelty-routinization hypothesis can be assessed with functional imaging studies that engage some part of the cingulofrontal confluence region. Cognitively challenging tasks can activate the right cingulofrontal confluence region (Corbetta et al., 1991; Bench et al., 1993; Raichle et al., 1994; Bush et al., 1998). Although the left region is

18

required for inductive reasoning (Goel et al., 1997) and performance on the sad Stroop task (George et al., 1994), interpreting ‘theory of mind’ stories appears to involve both hemispheres to some extent (Fletcher et al., 1995). Reasoning, processing of cognitively challenging information, assessment of emotionally charged faces, and ‘mentalizing’ functions may require both hemispheres even though studies of cortical damage lead to intriguing hypotheses about lateralization of functions. The specific contributions of each hemisphere to mental activity are not fully understood.

Distributed mental activity and the unified mind To the extent that there are two cortical sites in both hemispheres engaged in mental activity, perception of a single mind requires an explanation. Bogen (1986) considered issues relating to the duality of mind in the intact brain from the perspective of lateralization of function. The duality of mind associated with lateralization to either of the confluence regions alone is partially solved by contralateral callosal connections for each region. Either hemisphere can contribute to recovery of function following unilateral lesions and different contributions may be made by each hemisphere but joined by callosal connections. Thus, lateralization of functions do not interfere with a unified perception of self in either confluence region. Furthermore, if the cingulofrontal confluence region were the only primary site necessary for mental activity, activity in this region would represent the perceptual whole. However, the presence of two confluence regions requires solving what is often termed the ‘binding problem’. How do we experience existence through a unified sense of self that seamlessly joins external and internal stimuli, memory, plans, emotions, and reflective thought in spite of distributed parallel processing in the cerebral cortex? One attempt to solve the binding problem evolved naturally from the study of corticocortical connections. Serial connections from primary sensory to sensory association and then to multimodal areas is a sequence that would solve the problem by having a terminal site in the cortex that receives

unique and highly processed inputs and is also interconnected with motor systems to control skeletomotor and autonomic activity. Numerous studies of monkey cortex failed to identify such an area. Furthermore, functional imaging studies support the concept of two or more cortical areas operating in parallel rather than converging on a single ‘higher order’ association area. In terms of the two confluence regions, crucial corticocortical connections could subserve mental unity only if it can be shown that the cingulofrontal and parietotemporal confluence regions are interconnected. connections of the human cerebral cortex are not known, however, and the connections of key parts of the confluence regions in human cortex do not appear to have counterparts in the monkey. Much of midcingulate cortex and areas 39 and 40 are difficult to homologize among primate species and this will continue to hamper efforts to understand mechanisms that subserve the unity of mind. Another mechanism for binding activity in the two confluence regions is with a common input from a subcortical structure such as the intralaminar and midline thalamic nuclei. The system for processing noxious stimuli provides a means of assessing the potential role of these thalamic nuclei in consciousness and binding. As discussed in more detail in Chapter 16 of this volume, the midline and intralaminar thalamic nuclei have profound connections with anterior cingulate cortex. Although we do not know if cortex in the parietotemporal confluence region in human receives similar projections, we do know that noxious stimuli activate area 40 as well as other somatosensory areas (Svensson et al., 1997). These nuclei contain nociceptive neurons and blocking activity in this part of thalamus blocks nociceptive activity in anterior cingulate cortex (Sikes and Vogt, 1992). Since the cingulofrontal and parietotemporal confluence regions are primary components of the mind, noxious stimuli can directly drive neurons in both regions and engage affective, body orientation, and response selection patterns. In the instance of pain processing, the noxious stimulus directly drives the relevant response circuits via the intralaminar and midline thalamic nuclei. Thus, sensory or cognitive driving in combination with common circuitry in a

19

parallel processing system may provide for unified perceptual events. The common input mechanism provides for temporal linking of activity in divergent cortical regions and is similar to the proposition that massive and parallel networks are functionally synchronized to produce 40 Hz discharges by projections from the midline and intralaminar thalamus (Llinas and Ribary, 1993). Although this assures a common temporal pattern of discharge, it is not clear how this synchronization itself via a common input leads to binding of mental activity into a unified mind. Although it is unlikely that these thalamic nuclei operate like a ‘searchlight’, this mechanism might help to explain binding at the level of confluence regions, if the regions were themselves interconnected. The binding problem for uniform mental activity may be solved in a number of ways. First, there may be just a few primary confluence regions that provide for the essence of mental activity such as willed actions, intentions, imagery, intuitive logic, and integration of the internal and external features of self. Second, sensory driving via the midline and intralaminar thalamic nuclei could provide for synchronizing activity in the confluence regions themselves andor in other relevant networks. Third, homotopic callosal and ipsilateral connections between the cingulofrontal and parietotemporal regions would result in the final stage of binding parallel activities in distributed structures into a single mental event.

Therapeutic implications of localizing the mind to two confluence regions From a neuroscientific perspective, an understanding of functional activity in the cingulofrontal and parietotemporal confluence regions is essential for studies that seek to identify biological mechanisms whereby mental activity regulates the body and, hence, Minmody interactions. This emphasis on a few critical regions does not preclude involvement in many other telencephalic structures that are engaged in particular sensorimotor events and in the activities of parallel and distributed networks such as those in lateral prefrontal cortex and sequential and distributed processing in

sensory systems.These localizations simply emphasize that there are regions essential to most mental activity, that are likely the location of the mind, and that cannot be overlooked in terms of dominant network functions during mental activity. The profound importance of the anterior cingulate cortex itself has been observed by others, including Damasio (1994), who observes that it is a ‘fountainhead’ that constitutes the source of energy for both external and internal actions. The cingulofrontal confluence region provides a target for drug, rehabilitation, and meditation strategies for modifying MindBody interactions. l k o examples of such possibilities can be provided in terms of psychological modifications of Mind BraidBody relationships. The first example is provided in studies of cerebral blood flow during hypnosis. In a study by Rainville et al. (1997), hypnosis was used to modify the unpleasantness of noxious stimuli. High levels of unpleasantness were associated with elevated blood flow in midcingulate cortex, while blood flow in the anterior insula was not selectively altered by unpleasantness of the stimuli. Although affect is not localized in midcingulate cortex (Vogt et al., 1997)’ hypnotic regulation of midcingulate cortex may be related to its involvement in mental imagery and response selection in relation to motivationally relevant stimuli. A second example of the potential of activations of the cingulofrontal confluence region to provide therapeutic relief is that suggested for motor imagery. Decety (1995) suggested that motor imagery itself may provide a strategy for stimulating recovery of motor function following damage to the central nervous system. From the perspective of Minmody interactions, such responses and their close relationship to mental activity suggest that the cingulofrontal confluence region should be a morphological target of future efforts to improve Minmody interactions and to counteract pathological processes that influence mental functions. Therapies based on mentalistic strategies can be guided by testable hypotheses and neuroscience is on the verge of assessing the mechanisms of action of complex approaches to neurological and psychiatric diseases like those embodied by traditionalholistic medicine. A crucial link to

20

future efforts will be hypothesis-driven experiments using functional imaging modalities. The neurobiology of Mind/Brain/Body interactions will inevitably focus on those regions in the brain that are primarily associated with mental activity including the cingulofrontal and parietotemporal confluence regions. References Andersen, RA. (1995) Encoding of intention and spatial location in the posterior parietal cortex. Cereb. Cort., 5: 457-469. Baleydier, C. and Mauguikre, F. (1980) The duality of the cingulate gyms in monkey. Neuroanatomical study and functional hypotheses. Brain, 103: 525-554. Bancaud, J. and Talairach, J. (1992) Clinical semiology of frontal lobe seizures. Adv. Neuml., 57: 3-58. Barbas, H. and Pandya, D.N. (1989) Architecture and intrinsic connections of the prefrontal cortex in the rhesus monkey. J. Comp. Neurol., 286: 353-375. Bard, P. and Mountcastle, V.B. (1948) Some forebrain mechanisms involved in expression of rage with special reference to suppression of angry behavior. Res. Nerv. Ment. Dis., 27: 362-404. Barris, R.W. and Schuman, H.R. (1953) Bilateral anterior cingulate gyms lesions. Neurology, 3: 44-52. Bexton, W.H., Heron, W. and Scott, T.H. (1954) Effects of decreased variation in the sensory environment. Can. J. Psychol., 8: 70-76. Bench, C.J., Frith, C.D., Grasby, M., Friston, K.J., Paulesu, E., Frackowiack, R.S.J. and Dolan, R.J. (1993) Investigations of the functional anatomy of attention using the Stroop test. Neuropsychologia, 3 1: 907-922. Bisiach, E., Tegner, R., Ladavas, E., Rusconi, M.L., Mijovic, D. and Hjaltason, H. ( 1995) Dissociation of ophthalmokinetic and melokinetic attention in unilateral neglect. Cereb. Cort., 5: 439-447. Block, N. (1995) On a confusion about the function of consciousness. Behav. Brain Sci., 18: 227-287. Bogen, J.E. (1986) Mental duality in the intact brain. Bull. Clin. Neumsci., 5 1: 3-29. Bryden, M.P. and Ley, R.G. (1983) Right-hemispheric involvement in the perception and expression of emotion in normal humans. In: K.M. Heilman and P. Satz (Eds.), Neuropsychology of Human Emotion, Guilford Press, New York, pp. 6-45.

Bush, G., Whalen, P.J., Rosen, B.R., Jenike, M.A., McLnerney, S.C. and Rauch, S.L. (1998) The counting Stroop: an interference task specialized for functional neuroimagingvalidation study with functional MRI. Hum. Brain Mapp., 6: 270-282. Chan, J.-L. and Ross, E.D. (1997) Alien hand syndrome: influence of neglect on the clinical presentation of frontal and callosal variants. Cortex, 33: 287-299.

Corbetta, M., Miezin, EM., Dobmeyer, S., Shulman, G.L. and Peterson, S.E. (1991) Selective and divided attention during visual discriminations of shape, colorm and speed: functional anatomy by positron emission tomography. J. Neurosci., 11: 2383-2402. Crick, F. and Koch, C. (1995) Are we aware of neural activity in primary visual cortex? Nature, 375: 121-123. Crick, F. and Koch, C. (1998) Consciousness and neuroscience. Cereb. Cort., 8: 97-107. Damasio, A.R. (1989) The brain binds entities and events by multiregional activation from convergence zones. Neural Comp., 1: 123-132. Damasio, A.R. (1994) Descartes’ Error. G.P. Putnam’s Sons, New York. Damasio, A.R., Tranel, D. and Damasio, H. (1990) Individuals with sociopathic behavior caused by frontal damage fail to respond autonomically to social stimuli. Behav. Brain Res., 41: 81-94. Damasio, H., Grabowski, T., Frank, R., Galaburda, A.M. and Damasio, A.R. (1994) The return of Phineas Gage: clues about the brain from the skull of a famous patient. Science. 264: 1102-1 105. Davis, J.M., McCourt, W.F. and Solomon, P. (1960) Effect of visual stimulation on hallucinations and other mental experience during sensory deprivation. Am. J. Psychiatry, 116: 889-908. Decety, J., Perani, D., Jeannerod, M., Bettinardi, V., Tadary, B., Woods, R., Mazziotta, J.C. and Fazio, F. (1994) Mapping motor representation with positron emission tomography. Nature, 371: 600-602. Decety, J. (1995) Can motor imagery be used as a form of therapy? J. NIH Res., 7: 4 7 4 8 . Derbyshire, S.W.G., Vogt, B.A. and Jones, A.K.P. (1998) Pain and Stroop interference tasks activate separate processing moduIes in anterior cingulate cortex. Enp. Brain Res., 118: 52-60. Devinsky, 0. (1997) Neurological aspects of the conscious and unconscious mind. Ann. NYAcad. Sci., 835: 321-329. Devinsky, O., Morrell, M.J. and Vogt, B.A. (1995) Contributions of anterior cingulate cortex to behavior. Brain, 118: 279-306. Doane, B.K., Mahatoo, W., Heron, E. and Scott, T.H. (1959) Changes in perceptual function after isolation. Can. J. Psychol., 13: 210-215. Drevets, W.C., Price, J.L., Simpson, J.R., Todd, R.D., Reich, T., Vannier, M. and Raichle, M.E. (1997) Subgenual prefrontal cortex abnormalities in mood disorders. Nature, 386: 824-827. Fletcher, P., Happ6, F., Frith, U., Baker, S.C., Dolan, R.J., Frakowiack, R.S.J. and Frith, C.D. (1995) Other minds in the brain: a functional imaging study of ‘theory of mind’ in story comprehension. Cognition, 57: 109-128. Frith, C.D. (1992) Consciousness, information processing and the brain. J. Psychopharmacol., 6: 436-449. Frith, C.D., Friston, K., Liddle, P.F. and Frackowiack, R.S.J. (1991) Willed actions and the prefrontal cortex in man: a study with PET. Pmc. R. SOC. Lond. B., 244: 241-246.

21 Fuster, J.M. (1995) Memory and planning: ' h o temporal perspectives of frontal lobe function. In: H.H. Jasper, S.P.S. Riggio and Goldman-Rakic (Eds.), Epilepsy and the Functional Anatomy of the Frontal Lobe, Advances in Neurology, Vol66, Raven Press, New York, pp. 9-20. George, M.S., Ketter, T.A., Parekh, P.I., Rosinsky, N., Ring, H., Casey, B.J., Trimble, M.R., Honvitz, B., Herscovitch, P. and Post, R.M. (1994) Regional brain activity when selecting a response despite interference: An H,'*O PET study of the Stroop and an emotional Stroop. Hum. Brain Mapp., 1: 194-209. Goel, V., Gold, B., Kapur, S. and Houle, S. (1997) The seats of reason? An imaging study of deductive and inductive reasoning. Cog. Neurosci. Neuropsychol., 8: 1305-1 309. Goldberg, E. and Podell, K. (1995) Lateralization in the frontal lobes. In: H.H. Jasper, S. Riggio and P.S. Goldman-Rakic (Eds.),Epilepsy and the Functional Anatomy of the Frontal Lobe, Advances in Neurology, Vol 66, Raven Press, New York, pp. 85-96. Goldman-Rakic, P.S. (1987) Circuitry of the prefrontal cortex and the regulation of behavior by representational memory. In: V.B. Mountcastle, F. Plum and S.R. Geiger (Eds.), Handbook of Physiology, Part 1 The Nervous System, Chapter 9, American Physiological Society, Bethesda, MD, pp. 373417. Goldman-Rakic, P.S., Selemon, L.D. and Schwartz, M.L. (1984) Dual pathways connecting the dorsolateral prefrontal cortex with the hippocampal formation and parahippocampal cortex in the rhesus monkey. Neuroscience, 12: 719-743. Hirono, N., Mori, E., Ishii, K., Ikejiri, Y.,Imamura, T., Shimomura, T., Hashimoto, M., Yamashita, H. and Sasaki, M. (1998) Hypofunction in the posterior cingulate gyms correlates with disorientation for time and place in Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry, 64: 552-554. Kaitaro, T. (1996) La Peyronie and the experimental search for the seat of the soul: neuropsychological methodology in the eighteenth century. Cortex, 32: 557-564. Kandel, E.R. (1991) Brain and behavior. In: E.R. Kandel, J.H. Schwartz and T.M. Jessell (Eds.), Principles of Neural Science, Elsvier, New York. Levin, B., Duchowny, M. (1991) Childhodd obsessive-compulsive disorder and cingulate epilepsy. Biol. Psychiatry, 30: 1049-1055. Llinas, R. and Ribary, U. (1993) Coherent 40-Hz oscillation characterizes dream state in humans. Proc. Natl. Acud. Sci. USA, 90: 2078-2081. Logothetis, N.K. and Schall, J.D. (1989) Neuronal correlates of subjective visual perception. Science, 245: 761-763. Malloy, P., Cimino, C. and Westlake, R. (1992) Differential diagnosis of primary and secondary Capgras delusions. Neuropsychiatry Neurophysiol. Behav. Neurol., 5: 83-96. Mayberg, H.S. (1997) Limbic-cortical dysregulation: a proposed model of depression. J. Neuropsych. Clin. Neurosci., 9: 47 1 4 81. Mazoyer, B.M., Tzourio, N., Frak, V., Syrota, A., Murayama, N., Levrier, 0.. Salamon, G., Dehaene, S., Cohen, L. and

Mehler, J. (1993) The cortical representation of speech. J. Cog. Neurosci., 5: 467469. Nielson, J.M. and Jacobs, L.L. (1951) Bilateral lesions of the anterior cingulate gyri. Bull. Los Angeles Neurol. SOC.,16: 23 1-238. Olson, C.R., Musil, S.Y. and Goldberg, M.E. (1993) Posterior cingulate cortex and visuospatial cognition: Properties of single neurons in the behaving monkey. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp. 366-380. Pardo, J.V., Pardo, P.J., Janer, K.W. and Raichle, M.E. (1990) The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proc. Natl. Acad. Sci. USA, 87: 256-259. Raichle, M.E., Fiez, J.A., Videen, T.O., MacLeod, A.-M.K., Pardo, J.V., Fox, P.T. and Peterson, S.E. (1994) Practicerelated changes in human brain functional anatomy during non-motor learning. Cereb. Cort., 4: 8-26. Rainville, P., Duncan, G.H., Price, D.D., Carrier, B. and Bushnell, M.C. (1997) Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science, 277: 968-97 1. Rizzolatti, G., Matelli, M. and Pavesi, G. (1983) Deficits in attention and movement following the removal of postarcuate (area 6) and prearcuate (area 8) cortex in macaque monkeys. Brain, 106: 655-673. Ruff, R.L. and Volpe, B.T. (198 1) Environmental reduplication associated with right frontal and parietal lobe injury. J. Neurol. Neurosurg, Psychiatry, 44: 382-386. Schanberg, S.M. and Field, T. (1987) Sensory deprivation, stress, and supplemental stimulation in rat pup and preterm human neonates. Child Dev., 58: 1431-1447. Sejnowski, T.J. (1986) Open questions about computation in cerebral cortex. In: J.L. McClelland and D.E. Rummelhart (Eds.), Parallel Distributed Processing, MIT Press, Cambridge, pp. 372-389. Signer, S.F. (1992) Psychosis in neurologic diseases: Capgras symptom and delusions of reduplication in neurologic disorders. Neuropsychiatry Neuropsychol. Behav. Neurol., 5: 138- 143. Sikes, R.W. and Vogt, B.A. (1992) Nociceptive neurons in area 24 of rabbit cingulate cortex. J. Neurophysiol., 68: 1720-1732. Solomon, P. (1961) Sensory Deprivation, Harvard University Press, Cambridge. Stuss, D.T. and Benson, D.F. (1984) Neuropsychological studies of the frontal lobes. Psychol. Bull., 95: 3-28. Svensson, P., Minoshima, S.,Beydoun, A., Morrow, T.J. and Casey, K.L. (1997) Cerebral processing of acute skin and muscle pain in humans. J. Neurophysiol., 78: 450-460. Valenstein, E., Bowers, D., Verfaellie, M., Heilman, K.M., Day, A. and Watson, R.T. (1987) Retrosplenial amnesia. Brain, 110: 1631-1646. Vogt, B.A. and Pandya, D.N. (1987) Cingulate cortex of rhesus monkey. 11. Cortical afferents. J. Comp. Neurol., 262: 27 1-289.

22 Vogt, B.A., Rosene, D.L. and Pandya, D.N. (1979) Thalamic and cortical afferents differentiate anterior from posterior cingulate cortex in the monkey. Science, 204: 205-207. Vogt, B.A., Vogt, L.J., Nimchinsky, E.A. and Hof, P.R. (1997) Primate cingulate cortex chemoarchitecture and its disruption

in Alzheimer’s disease. In: F.E. Bloom, A. Bjorklund and T. Hokfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 13: The Primate Nervous System, Part I, Elsevier Science B.V., Netherlands.

SECTION III

The neurobiology of the stress response

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 3

Protective and damaging effects of stress mediators: central role of the brain Bruce S . McEwen Harold and Margaret Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller Universiw, 1230 York Avenue, New York, NY 10021, USA

Introduction When the body is challenged physically or psychologically, the physiological responses of the autonomic nervous system and neuroendocrine system promote adaptation and survival in the short run. However, the same physiological response systems can cause damage and exacerbate disease processes over longer periods of time. The relationship between these paradoxical aspects of stress hormone action has been with us since the late Hans Selye introduced the stress concept in 1936 (Selye, 1936). Understanding the basis for the paradoxical role of stress hormones is closely tied to the further elaboration of two aspects of mind-body interactions (see Fig. 1); first, are the ways in which genetics, developmental processes and experience influence how the brain processes events and responds to them as potential stressors, and second, the nature of the behavioral and physiological responses to the stressful challenge, particularly the ways in which these responses may protect the body or contribute to pathophysiological changes. In the case of behavior, the concern is whether the response gets the individual out of trouble or if it further exacerbates the situation and contributes to Corresponding author. Tel.: 212-327-8624; Fax: 212-327-8634; e-mail: [email protected]

the eventual wear and tear on the body. In the case of physiological responses, the key point is how efficiently these responses are turned on and then turned off again when no longer needed. This paper summarizes a new formulation of this relationship and illustrates it with examples from the cardiovascular/metabolic and immune systems and the brain. The plasticity and vulnerability of the brain, particularly the hippocampus, is discussed as an example of the key interpretive and regulatory role that the brain plays in the response to stress.

Allostasis and allostatic load DeJnitions

When the body is challenged physically or psychologically, the physiological responses such as activation of the autonomic nervous and neuroendocrine systems promote adaptation and survival in the short run. Hormones act via receptors and set into motion changes at the cellular level (see Fig. 2). This has been referred to as ‘allostasis’ or literally ‘re-establishing stability through change‘ (Sterling and Eyer, 1988). Allostasis is therefore the means by which the body re-establishes homeostasis in the face of a challenge. Homeostasis is a term coined in the era of classical thermodynamics, whereas allostasis reflects modem open system thermodynamics in

26

Fig. 1. The stress response and development of allostatic load. Perception of stress is influenced by one’s experiences, genetics, and behavior. When the brain perceives an experience as stressful, physiologic and behavioral responses are initiated leading to allostasis and adaptation. Over time, allostatic load can accumulate, and the overexposure to neural, endocrine, and immune stress mediators can have adverse effects on various organ systems, leading to disease. Reprinted from McEwen, 1998, by permission.

which the ‘steady state’ replaces ‘equilibrium’. Physiological systems operate within a dynamic range of steady states, being higher at some times of day or as they are driven by ongoing events, and yet they maintain internal balance or homeostasis, in terms of blood pH, electrolyte concentrations, oxygen tension, etc. During allostasis, physiological systems sometimes work for periods of time at higher or lower levels than average: for example, physical exertion or emotional turmoil may lead to elevated heart rate and blood pressure, as well as elevated glucocorticoid levels; the elevated glucocorticoids, in turn, suppress the production of inflammatory cytokines as well as producing a variety of effects on metabolism and brain function (McEwen, 1998). The increased activity of these systems can accelerate pathological processes if the elevation occurs chronically over long times, because the products of the persistently elevated activity produce various effects on tissue of the body that are not so likely to happen if the allostatic systems shut off when they are no longer needed (McEwen, 1998). The wear and tear caused by overactivity of allostatic systems has been termed ‘allostatic load’ (McEwen and Stellar, 1993; McEwen, 1998), and allostatic load is a better description for the

sometimes subtle long-term influences that can compromise health than is the term ‘chronic stress’, which covers only some of the possibilities that are described below. Types of allostatic load As long as allostatic responses shut off when they are no longer needed, then the body is able to adapt and survive the immediate challenge and, at the same time, not to suffer many long-term consequences. However, when stresses occur very frequently, then wear and tear results. Wear and tear from overexposure to stress hormones also results when the allostatic responses are inefficiently managed and remain active when not needed. Allostatic load is seen as impaired immunity, atherosclerosis, abdominal obesity, bone demineralization, muscular weakening and atrophy of nerve cells in the brain. At least four situations are recognized that contribute to allostatic load (see Fig. 3): (1) The first is simply too frequent or too much stress, and this corresponds to the general notion of ‘chronic stress’. For example, rats or tree shrews exposed to repeated restraint or

27

Fig. 2. Allostasis in the autonomic nervous system and the hypothalamic-pituitary-adrenal axis. Top. Allostatic systems respond to stress by initiating the adaptive response, sustaining it until the stress ceases, and then shutting it off (recovery). Bottom. Allostatic responses are initiated by an increase in circulating catecholamines from the autonomic nervous system and glucocorticoids from the adrenal cortex. This sets into motion adaptive processes that alter the structure and function of a variety of cells and tissues. These processes are initiated via intracellular receptors for steroid hormones, plasma membrane receptors, and second messenger systems for catecholamines. Cross-talk between catecholamines and glucocorticoid receptor signaling systems can occur. Reprinted from McEwen, 1998, by permission.

psychosocial stress display atrophy of nerve cells in the hippocampus (Magarinos and McEwen, 1995; Magarinos et al., 1996), and rats with more active HPA axes from birth show earlier loss of hippocampal-dependent cognitive function as they age (Dellu et al., 1994). In the same vein, human beings who have experienced considerable life-long economic hardship show declines in physical and mental functioning and increased depression as they age (Lynch et al., 1997). (2) The second is the failure to adapt to repetitions of the same stressor, leaving the organism exposed to higher levels of stress hormones. For example, there is the failure to habituate during the repetition of a public speaking challenge, which is characteristic of a minority of subjects given the Treier Social Stimulation Test; these individuals continue to produce a cortisol elevation after a number of repetitions of the same situation (Kirschbaum et al., 1995). (3) The third is the failure to shut-off stress hormone secretion or autonomic activity. An example of this type of allostatic load is the persistence of blood pressure elevations in the aftermath of job-related stress that was seen in individuals with a familial history of hypertension (Azmitia, 1981). Blood pressure elevation is a known contributor to atherosclerosis (Manuck et al., 1995). Another example is the persistence of cortisol elevations, seen in depression and leading to accelerated calcium loss from bone (Michelson et al., 1996) and in chronic, moderate sleep deprivation and leading to elevated glucose and insulin levels (Van Cauter et al., 1992; Van Cauter, personal communication). (4) The fourth is the failure to mount an adequate response, leading to enhanced activity of other allostatic systems that are normally opposed by other systems. This illustrates the concept of counter-regulation by stress hormones of other components of the stress response. For example, glucocorticoids oppose the activation of some of the major brain neurotransmitter systems that are activated by stress, such as the noradrenergic system (Akana et al., 1988;

28

Fig. 3. Four types of allostatic load. The top panel illustrates the normal allostatic response, in which a response is initiated by a stressor, sustained for an appropriate interval, and then turned off. The remaining four panels illustrate four conditions that lead to allostatic load chronic stress (repeated ‘hits’ by multiple novel stressors); lack of adaptation to repetitions of the same stressor; prolonged response (delayed shut down); and inadequate response that leads to compensatory hyperactivity of other mediators (e.g., inadequate secretion of glucocorticoid, resulting in increased levels of cytokines that are normally counter-regulated by glucocorticoids). Figure drawn by Dr. Firdaus Dhabhar, Rockefeller University. Reprinted from McEwen, 1998, by permission.

McEwen et al., 1992). Both glucocorticoids and catecholamines oppose activation of some of the components of the immune system: e.g., the production of inflammatory cytokines (Madden and Felten, 1995; McEwen et al., 1997). In rats of the Lewis strain, a deficiency of glucocorticoids leads to enhanced autoimmune and inflammatory disorders (Sternberg et al., 1992). The Lewis rat may be a model for a number of human disorders involving inflammatory or autoimmune disorders, in which there is a reduced levels of adrenal steroids (McEwen, 1998).

Contrasting aspects of allostasis and allostatic load

Thus the acute and chronic actions of stress mediators can often have quite different and sometimes opposite effects, as exemplified by the actions of glucocorticoids. For metabolism, glucocorticoids promote hunger and facilitate the mobilization of energy; as a result, however, chronic glucocorticoid hyperactivity leads to breakdown of muscle proteins into energy and leads to insulin resistance and elevated blood glucose levels, leading to fat deposition and ultimately in

29

Q p e I1 diabetes. For the brain, glucocorticoids promote memory storage by acting in the hippocampus, amygdala and other brain regions, and they counter-regulate neurochemical activities in the brain, keeping the noradrenergic system in check while facilitating some activities and inhibiting others of the serotonergic system; yet chronic elevation of glucocorticoids promotes atrophy of hippocampal neurons and contributes to neuronal loss. In the immune system, glucocorticoids participate in activation of immune responses by promoting cell trafficking to sites where they are needed for immune defense; this can be seen as enhancement of delayed-type hypersensitivity. Yet glucocorticoids also contribute to immunosuppression, and chronic stress has the opposite effect on delayed-type hypersensitivity as well as increasing susceptibility to infectious disease.

Allostasis, allostatic load and the brain

Key role of the brain in perceiving and responding to stress The brain is the key to understanding allostasis and allostatic load because it controls both the physiological and behavioral coping responses and because the brain is itself a target of allostasis and allostatic load through the actions of stress hormones on receptors in the brain (see Fig. 1). Two structures in the brain, the amygdala and the hippocampus, are instrumental in interpreting what is stressful and in helping to choose a response. The amygdala is the focal point of ‘emotional memories’ and becomes hyperactive in post-traumatic stress disorder and depressive illness (LeDoux, 1996). Both types of stress hormones, the catecholamines and glucocorticoids, play an important role in the consolidation of passive avoidance behavior in rats and mice (McGaugh et al., 1996; Roozendaal and McGaugh, 1997). The hippocampus is responsible for memories of events and contexts and is vulnerable to insults and trauma, showing loss of nerve cells after stroke and head injury. The hippocampus is also vulnerable to stress, having receptors that enable it to respond to stress hormones in the blood. As will be described below, the hippocampus undergoes a variety of structural and functional changes in response to

stress. When the hippocampus and other brain regions are damaged, individuals lose the ability to remember and perform optimally, thus contributing further to impaired ability to cope. In the brain, there is an additional constraint related to the fact that nerve cells for the most part are born early in life and are not replaced during the rest of the lifespan. Moreover, it has been assumed that neural circuits and nerve cell structures are largely static in adult life and that only during development is the brain plastic. However, this idea is changing, at least in part, because of studies showing in the adult brain the occurrence, in certain brain regions, of cyclic synaptogenesis,atrophy and elongation of dendrites after stress and hibernation, and regulated neurogenesis. Moreover, the response of the adult brain to its environment, including some of the plastic structural changes seen in the adult brain, is shaped by events that occur early in brain development. Hormones of the gonads, adrenals and thyroid gland play an important role in these types of plasticity. Developmental events In contrast to the early ‘nature-nurture’ debates, when genes and environment were regarded as isolated and independent entities, modern cell and molecular biology is teaching us that genes are regulated by environmental signals and that this regulation goes on for the entire lifespan. A common pattern in development is for genes to be made available, or unavailable, in different tissues for regulation by hormonal or other inter- and intracellular messengers at a later time. Individual differences in brain function and behavior are shaped, in part, by the effects of early experience and by hormones. Sexual differentiation of the brain is an example of an hormonallydirected event in which the presence or absence of testosterone during fetal and neonatal life causes brain development to diverge in two directions: male or female. As a result, male and female brains differ in subtle ways, both in terms of structure and connectivity as well as different responses to hormonal signals (Witelson, 1989; McEwen, 1991a; Kimura, 1992; Witelson et al., 1995). That is, besides differences in numbers and distributions of synaptic connections and groups of neurons,

30

male and female brains differ in that they use the same hormonal signals, androgens, estrogens and progestins, to achieve somewhat different purposes in various brain regions. For example, estrogens induce progestin receptors in female brain cells to a greater extent than in the same cells in males, and estrogens induce new synapses to form in female hypothalamus and hippocampus and do not have the same effect in these regions of the male brain (McEwen, 1991a). Androgens regulate synaptic and dendritic plasticity of specific regions of the male nervous system that are not present in the female, such as the spinal nucleus that innervates muscles of the penis (Forger and Breedlove, 1991). Androgens also regulate sensitivity of the brain to serotonin in the ‘serotonin behavioral syndrome’ (Fischette et al., 1984). Likewise, effects of prenatal stressful experiences, or the opposite effects of postnatal handling, appear to involve adrenal and thyroid hormone actions, respectively, that program the brain to have either higher or lower reactivity to novel experiences later in life (Catalani et al., 1993; Meaney et al., 1994). According to this model, once the reactivity of the adrenocortical system is established by events early in life, it is the subsequent actions of the hypothalamo-pituitary-adrenal (HPA) axis in adult life that play a major role in determining the rate of brain and body aging. Increased I-PA activity is associated with increased brain aging, whereas the opposite is true of animals with reduced I-PA reactivity to novel situations (see below). The hippocampal formation of the brain turns out to be one of the most vulnerable and plastic of brain regions and one in which many of these processes can be studied, and recent studies have shown that rats that show high levels of stress hormone reactivity show earlier decline of cognitive function as they age (Dellu et al., 1994), whereas rats with a lower stress hormone reactivity have a slower rate of cognitive aging and a reduced loss of hippocampal neurons and function (Catalani et al., 1993; Meaney et al., 1994). Plasticity of the adult brain The hippocampus is an important brain structure for working and spatial memory in animals and

humans, and in episodic memory as well as memory for ‘context’ in which positive and negative experiences have taken place (Eichenbaum and Otto, 1992; LeDoux, 1995). The hippocampus is also a vulnerable brain structure as far as sensitivity to epilepsy, ischemia, head trauma, stress and aging (Sapolsky, 1992). The hippocampus is also a target brain area for the actions of hormones of the steroidlthyroid hormone family (McEwen et al., 1995a,b), which traditionally have been thought to work by regulating gene expression (Miner and Yamamoto, 1991). ‘Genomic’ actions of steroid hormones involve intracellular receptors, whereas ‘non-genomic’ effects of steroids involve putative cell surface receptors (McEwen, 199lb). Although this distinction is valid, it does not go far enough in addressing the variety of mechanisms that steroid hormones use to produce their effects on cells. This is because cell surface receptors may signal changes in gene expression, while genomic actions sometimes affect neuronal excitability, often doing so quite rapidly (Orchinik and McEwen, 1995). Moreover, steroid hormones and neurotransmitters may operate together to produce effects, and sometimes these effects involve collaborations between groups of neurons. For example, a number of steroid actions in the hippocampus involves the co-participation of excitatory amino acids (McEwen et al., 1995a,b). These interactions are evident for the regulation of synaptogenesis by estradiol in the CA1 pyramidal neurons of hippocampus (Woolley and McEwen, 1994b; McEwen et al., 1995b) and for the induction of dendritic atrophy of CA3 neurons by repeated stress as well as by glucocorticoid injections (Magarinos and McEwen, 1995; McEwen et al., 1995a). Estrogen-regulated synaptogenesis is a cyclic event in the 4-5 d estrous cycle of the female rat, and it involves the formation of new synapses on dendritic spines by a process in which estradiol is involved in the induction of synapses and the proestrus surge of progesterone is responsible for the down-regulation of the synapses (Woolley and McEwen, 1993). In contrast to these rapid changes, stress-induced atrophy of dendrites is slower, taking almost three weeks to occur in rats and tree shrews. However, atrophy of dendrites of CA3 ~

31

pyramidal neurons has been reported to occur in a matter of hours as a result of hibernation in squirrels and hamsters, and the reversal of this atrophy has been found in 1-2 h after waking the animals (Popov et al., 1992; Popov and Bocharova, 1992; Magarinos and Pevet, unpublished). Besides synaptic and dendritic plasticity, neurogenesis occurs in the adult and developing dentate gyrus of rats, mice, tree shrews and primates (Cameron et al., 1995; Gould et al., 1997; Kempermann et al., 1997; Gould et al., 1998). This process is ‘contained’ by adrenal steroids as well as by excitatory amino acids, and these events may play a role in long-term, e.g., seasonal changes in the size and functional capacity of the hippocampus, at least in small mammals (Gould and McEwen, 1993; Cameron et al., 1995). Moreover, neurogenesis is inhibited by acute stress (Gould et al., 1997, 1998) and stress may therefore contribute to reductions in the size of the dentate gyrus and the stress-induced impairment of hippocampal function. At the same time, survival of dentate gyrus neurons is increased by placing mice in an enriched environment, with the result that dentate gyrus volume increases and with it there are increases in the functional capacity of the hippocampus in spatial learning and memory (Kempermann et al., 1997).

Implications of the plasticity and vulnerability of the brain for mind-body interactions Importance of the brain for coping with the environment

The brain, and in particular the limbic system including the hippocampus and amygdala, plays a pivotal role in processing of psychosocial experiences and in determining behavioral responses. Important functions of the hippocampus and amygdala involve episodic memory and memory of emotionally-laden events, respectively, and of the context in which they have occurred (Eichenbaum and Otto, 1992; Phillips and LeDoux, 1992). A sensitized amygdala can contribute to over-reactions to normally non-threatening events, such as loud noises in the case of post-traumatic stress disorder (Morgan et al., 1995). An impaired

hippocampus compromises the ability of the individual to remember events in daily life and make distinctions among complex cues in social and other situations (Gray, 1982). The medial prefrontal cortex (mPFC) might also be a target of stress and stress hormones, in light of the atrophy of this structure reported in depressive illness (Drevets et al., 1997) and the role of structure as a target of glucocorticoids in the regulation of HPA activity (Diorio et al., 1993). Thus, the health and functional capacity of the brain is an essential factor in human behavior and health, because if the brain fails to perform its memory functions adequately or is programmed to over-react to certain stimuli, as in post-traumatic stress disorder, then aberrant or inappropriate behavior is likely to result. One outcome of such failure is captured in the popular adage: ‘Stress makes you stupid’. Importance of sex differences

Sex differences in brain structures and mechanisms occur in other brain regions besides hypothalamus, such as hippocampus, and they appear to be involved in aspects of cognitive function and other processes that go beyond the reproductive process itself, such as the higher incidence of depression in women and of substance abuse in males (Regier et al., 1988). There are also sex differences in the seventy of brain damage resulting from transient ischemia (Hall et al., 1991) and sex differences in the response of the brain to lesions (Morse et al., 1992) and to severe, chronic stress (Uno et al., 1989; Mizoguchi et al., 1992). Genetic traits as risk factors for allostatic load

So far we have neglected discussing genotype as a factor involved in determining individual differences. There is no question that the genetic traits are important in this regard and genes that underlie diabetes, atherosclerosis, obesity, susceptibility to depressive illness, schizophrenia and Alzheimer’s disease increase the risk that experiences and other aspects of the life-long environment will ultimately produce disease (Plomin, 1990). In other words, genetic traits contribute to allostatic load by making it more potent in certain individuals.

32

However, it is not sufficient simply to find and describe the genes responsible for diseases, but rather it is necessary to understand how these genes are involved in the pathophysiology of disease and, in particular, how they are regulated.

Psychosocial determinants of health Psychosocial interactions and social hierarchies appear to be powerful determinants of allostatic load, not only in animal societies but also in human societies, where there are reported gradients of health and disease across the entire spectrum of socioeconomic status (Adler et al., 1993). Moreover, the collapse of societies, such as occurred in Russia after the fall of communism, has led to dramatic and frightening increases in mortality (Bobak and Marmot, 1996). Social interactions regulate neuroendocrine function, and social support is related to lower levels of adrenocortical activity (Seeman and McEwen, 1996). The repeated psychosocial stress of being subordinate has been reported, in tree shrews, to cause dendritic atrophy in hippocampus and, in vervet monkeys, to cause actual loss of hippocampal neurons (Uno et al., 1989; Fuchs et al., 1995). Psychosocial stress produces an allostatic load in cynomologous monkeys that increases atherosclerosis in dominant males in unstable social hierarchies (Manuck et al., 1995) and in subordinate females (Shively and Clarkson, 1994). These are some examples of what further investigation may show is a wide range of psychosocial and environmental influences that create allostatic load and alter the progression toward disease.

Conclusion The life-long interplay between genes and the environment is instrumental in shaping the structure and function of the body, and this now includes the brain as a plastic, ever-changing and vulnerable organ of the body. Key brain areas like the hippocampus and amygdala are vital to the processing of information that affect how each individual adapts to, and responds to, potentially stressful life events, and the response of the brain through its control of endocrine and autonomic function in turn

determines the degree of allostatic load that an individual will experience. The allostatic load in turn works with the intrinsic genetic susceptibility and capacities and propensities of the brain shaped by development to determine the progression toward declining health. Allostatic load affects cardiovascular function, muscle strength, bone mineral density, the deposition of fat, immune function and structure and function of neurons in the brain. New information is providing insights into the ways in which changes in neuron structure in the hippocampus participates in the vulnerability and sensitivity of this brain region to stress.

References Adler, N., Boyce, W.T., Chesney, M., Folkman, S. and Syme, S.L. (1993) Socioeconomic inequalities in health: no easy solution. JAMA. 269: 3 140-3 145. k a n a , S.F., Jacobson, L., Cascio, C.S., Shinsako, I. and Dallman, M.F. (1988) Constant corticosterone replacement normalizes basal adrenocorticotropin (ACTH) but permits sustained ACTH hypersecretion after stress in adrenalectomized rats. Endocrinol., 122: 1337-1342. Azmitia, E.C. (I98 1) Bilateral serotonergic projections to the dorsal hippocampus of the rat: simultaneous localization of 'H-5HT and HRP after retrograde transport. J. Comp. Neuml., 203: 737-743. Bobak, M. and Marmot, M. (1996) East-West mortality divide and its potential explanations: proposed research agenda. BMJ, 312: 421425. Cameron, H.A., McEwen, B.S. and Gould, E. (1995) Regulation of adult neurogenesis by excitatory input and NMDA receptor activation in the dentate gyrus. J. Neurosci., 15: 46874692. Catalani, A., Marinelli, M., Scaccianoce, S., Nicolai, R., Muscolo, L.A., Porcu, A,, Koranyi, L., Piazza, P.V. and Angelucci, L. (1993) Progeny of mothers drinking corticosterone during lactation has lower stress-induced corticosterone secretion and better cognitive performance. Brain Res., 624: 209-2 15, Dellu, F., Mayo, W., Vallee, M., Le Moal, M. and Simon, H. (1994) Reactivity to novelty during youth as a predictive factor of cognitive impairment in the elderly: a longitudinal study in rats. Brain Res., 653: 51-56. Diorio, D., Viau, V. and Meaney, M.J. (1993) The role of the medial prefrontal cortex (cingulate gyrus) in the regulation of hypothalamo-pituitary-adrenalresponses to stress. J. Neurosci., 13: 3839-3847. Drevets, W.C., Price, J.L., Simpson, Jr., J.R., Todd, R.D., Reich, T., Vannier, M. and Raichle, M.E. (1997) Subgenual prefrontal cortex abnormalities in mood disorders. Nature, 386: 824-827. Eichenbaum, H. and Otto, T. (1992) The hippocampus - what does it do? Behav. Neural. Biol., 57: 2-36.

33 Fischette, C., Biegon, A. and McEwen, B.S. (1984) Sex steroid modulation of serotonin behavioral syndrome. Life Sci., 35: 1197-1206. Forger, N.G. and Breedlove, S.M. (1991) Steroid influences on a mammalian neuromuscular system. Seminars Neurosci., 3: 459-468. Fuchs, E., Uno, H. and Flugge, G. (1995) Chronic psychosocial stress induces morphological alterations in hippocampal pyramidal neurons of the tree shrew. Brain Res., 673: 275-282. Gould, E., McEwen, B.S., Tanapat, P., Galea, L.A. and Fuchs, E. (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., Tanapat, P., McEwen, B.S., Flugge, G. and Fuchs, E.( 1998) Proliferation of granule cell precursors in the dentate gyms of adult monkeys is diminished by stress. Proc. Natl. Acad. Sci. USA, 95: 3168-3171. Gould, E. and McEwen, B.S. (1993) Neuronal birth and death. Curs Opin. Neurobiol., 3: 676-682. Gray, J. (1982) Precis of the neuropsychology of anxiety: an enquiry into the functions of the septo-hippocampal system. Behav. Brain Sci., 5: 469-534. Hall, E.D., Pazara, K.E. and Linseman, K.L. (1991) Sex differences in postischemic neuronal necrosis in gerbils. J. Cerebs Blood Flow Metab., 11: 292-298. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More hippocampal neurons in adult mice living in an enriched environment. Nature, 586: 493495. Kimura, D. (1992) Sex differences in the brain. Sci. Ames, 267: 119-125. Kirschbaum, C., Prussner, J.C., Stone, A.A., Federenko, I., Gaab, J., Lintz, D., Schommer, N. and Hellhammer, D.H. (1 995) Persistent high cortisol responses to repeated psychological stress in a subpopulation of healthy men. Psychosom. Med., 57: 468474. LeDoux, J.E. (1995) In search of an emotional system in the brain: leaping from fear to emotion and consciousness. In: M. Gazzaniga (Ed.), The Cognitive Neurosciences, MIT Press, Cambridge, pp. 1049-1061. LeDoux, J.E. (1996) The Emorional Brain, Simon and Schuster, New York. Lynch, J.W., Kaplan, G.A. and Shema, S.J. (1997) Cumulative impact of sustained economic hardship on physical, cognitive, psychological, and social functioning. NEJM, 337: 1889-1895. Madden, K.S. and Felten, D.L. (1995) Experimental basis for neural-immune interactions. Physiol. Rev., 75: 77-106. Magarinos, A.M., McEwen, B.S., Flugge, G. and Fuchs, E. (1996) Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J. Neurosci., 16: 3534-3540. Magarinos, A.M. and McEwen, B.S. (1995) Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: involvement of glucocorticoid secretion and excitatory amino acid receptors. Neuroscience, 69: 89-98.

Manuck, S.B., Kaplan, J.R., Adams, M.R. et al. (1995) Studies of psychosocial influences on coronary artery atherosclerosis in cynomolgus monkeys. Healfh Psychol., 7: 113-124. McEwen, B.S. (1991a) Our changing ideas about steroid effects on an ever-changing brain. Seminars Neurosci., 3: 497-507. McEwen, B.S. (1991b) Steroids affect neural activity by acting on the membrane and the genome. Trends Pharmacol. Sci., 12: 141-147. McEwen, B.S., Angulo, J., Cameron, H., Chao, H.M., Daniels, D., Gannon, M.N., Gould, E., Mendelson, S. Sakai, R., Spencer, R. et al. (1992) Paradoxical effects of adrenal steroids on the brain: protection versus degeneration. Biol. Psych., 31: 177-179. McEwen, B.S., Albeck, D., Cameron, H., Chao, H.M., Gould, E., Hastings, N., Kuroda, Y., Luine, V., Magarinos, A.M., McKittrick, C.R. et al. (1995a) Stress and the brain: a paradoxical role for adrenal steroids. In: G.D. Litwack (Ed.), Wtamins and Hormones, Academic Press, Inc., pp. 371402. McEwen, B.S., Gould, E., Orchinik, M., Weiland, N.G. and Woolley, C.S. (1995b) Oestrogens and the structural and functional plasticity of neurons: implications for memory, ageing and neurodegenerative processes. In: J. Goode (Ed.), The Non-Reproductive Actions of Sex Steroids, Ciba Foundation Symposium # I 9 1, CIBA Foundation, London, pp. 52-73. McEwen, B.S., Biron, C.A., Brunson, K.W., Bulloch, K., Chambers, W.H., Dhabhar, F.S., Goldfarb, R.H., Kitson, R.P., Miller, A.H., Spencer, R.L. et al. (1997) Neural-endocrineimmune interactions: the role of adrenocorticoids as modulators of immune function in health and disease. Brain Res. Rev., 23: 79-133. McEwen, B.S. (1998) Protective and damaging effects of stress mediators. N. Engl. J. Med., 338: 171-179. McEwen, B.S. and Stellar, E. (1993) Stress and the individual: mechanisms leading to disease. Arch. Int. Med., 153: 2093-2101. McGaugh, J.L., Cahill, L. and Roozendaal, B. (1996) Involvement of the amygdala in memory storage: interaction with other brain systems. Proc. Natl. Acad. Sci. USA, 93: 13508-13514. Meaney, M.J., Tannenbaum, B., Francis, D. et al. (1994) Early environmental programming hypothalamic-pituitary-adrenal responses to stress. Seminars Neurosci., 6: 247-259. Michelson, D., Stratakis, C., Hill, L., Reynolds, J., Galliven, E., Chrousos, G. and Gold, P. (1996) Bone mineral density in women with depression. N. Engl. J. Med., 335: 1176-1 181. Miner, J.N. and Yamamoto, K.R. (1991) Regulatory crosstalk at composite response elements. Trends Biochem. Sci., 16: 423-426. Mizoguchi, K., Kunishita, T., Chui D.H. and Tabira, T. (1992) Stress induces neuronal death in the hippocampus of castrated rats. Neurosci. Lett., 138: 157-160. Morgan, C.A., Grillon, C., Southwick, S.M., Davis, M. and Charney, D.S. (1995) Fear-potentiated startle in posttraumatic stress disorder. Biol. Psychiatry, 38: 378-385.

34 Morse, J.K., Dekosky, S.T. and Scheff, S.W. (1992) Neurotrophic effects of steroids on lesion-induced growth in the hippocampus. Exp. Neurol., 118: 47-52. Orchinik, M. and McEwen, B.S. (1995) Rapid actions in the brain: a critique of genomic and non-genomic mechanisms. In: M. Wehling (Ed.), Genomic and Non-Genomic Effects of Aldosterone, CRC Press, Boca Raton, pp. 77-108. Phillips, R.G. and LeDoux, J.E. (1992) Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behav Neurosci., 106: 274-285. Plomin, R. (1990) The role of inheritance in behavior. Science, 48: 183-188. Popov, V.I., Bocharova, L.S. and Bragin, A.G. (1992) Repeated changes of dendritic morphology in the hippocampus of ground squirrels in the course of hibernation. Neuroscience, 48: 45-5 I. Popov, V.I. and Bocharova, L.S. (1992) Hibemation-induced structural changes in synaptic contacts between mossy fibres and hippocampal pyramidal neurons. Neuroscience, 48: 53-62. Regier, D.A., Boyd, J.H., Burke, J.D., Rae, D.S., Myers, J.K., Kramer, M., Robins, L.N., George, L.K., Kamo, M. and Locke, B.Z. (1988) One-month prevalence of mental disorders in the U.S. Arch. Gen. Psychiatry, 45: 977-986. Roozendaal, B. and McGaugh, J.L. (1997) Glucocorticoid receptor agonist and antagonist administration onto the basolateral but not central amygdala modulates memory storage. Neurobiol. Learn. Mem., 67: 176-179. Sapolsky, R. (1992) Stress, the Aging Brain and the Mechanisms ofNeumn Death. Cambridge MIT Press, 1423. Seeman, T.E. and McEwen, B.S. (1996) The impact of social environment characteristics on neuroendocrine regulation. Psychosom. Med., 58: 459471.

Selye, H. (1936) A syndrome produced by diverse nocuous agents. Nature, 138: 32 Shively, C.A. and Clarkson, T.B. (1994) Social status incongruity and coronary artery atherosclerosis in female monkeys. Arteriosclerosis Thrombosis, 14: 721-726. Sterling, P. and Eyer, J. (1988) Allostasis: a new paradigm to explain arousal pathology. In: S. Fisher and J. Reason (Eds.), Handbook of Life Stress, Cognition and Health, John Wiley & Sons, New York, pp. 629-649. Sternberg, E.M., Chrousos, G.P., Wilder, R.L. and Gold, P.W. (1992) The stress response and regulation of inflammatory disease. Ann. Inr. Med., 1 17: 854-866. Uno, H., Tarara, R., Else, J., Suleman, M.A. and Sapolsky, R.M. (1989) Hippocampal damage associated with prolonged and fatal stress in primates. J. Neurosci., 9: 1705-17 11. Van Cauter, E., Shapiro, E.T., Tillil, H. and Polonsky, K.S. ( 1992) Circadian modulation of glucose and insulin responses to meals: Relationship to cortisol rhythm. Am. J. Physiol., 262: E467-E475. Witelson, S. (1989) Hand and sex differences in the isthmus and genu of the human corpus callosum. Brain, 112: 799-835. Witelson, S.F., Glezer, 1.1. and Kigar, D.L. (1995) Women have greater density of neurons in posterior temporal cortex. J. Neurosci., 15: 3418-3428. Woolley, C. and McEwen, B.S. (1993) Roles of estradiol and progesterone in regulation of hippocampal dendritic spine density during the estrous cycle in the rat. J. Comp. Neurol., 336: 293-306. Woolley,C. and McEwen, B.S. (1994) Estradiol regulates hippocampal dendritic spine density via an N-methyl-oaspartate receptor dependent mechanism. J. Neurosci., 14: 7680-7687.

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Bmin Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 4

Interactions between the immune and neuroendocrine systems Esther M. Sternberg Section on Neumendocrine Immunology and Behaviol; NIMHLVIH, Bldg. 10, Rm. 20-46, 3 Center DI:MSC 1284, Bethesda, ID 20892-1284, USA

Introduction The concept that the mind can affect the body, that emotions can affect disease, is not new. From before the time of Hippocrates until today, the popular culture has championed the belief that stress and the emotions can affect bodily disease. Until recently, however, this concept was dismissed by the scientific and medical community as unfounded (Sternberg, 1997a). However the availability of technological tools in neurobiology and immunology, and in the related clinical disciplines of clinical psychiatry, psychology and rheumatology, has permitted definition of the molecular, biochemical and neuroanatomical pathways by which the immune system and central nervous systems communicate. Through defining these communication networks, we can better understand the diseases which result when these pathways are interrupted. Such understanding is leading not only to the scientific community’s acceptance of the concept of wholeness of mind and body, but to development of novel approaches to treat illnesses which result when such communication networks are disturbed. There are many levels and mechanisms by which the immune and central nervous system commuCorresponding author. Tel.: 301-402-2773; Fax: 301-402- 1561; e-mail: ems@ codon.nih.gov

nicate, and interruptions or perturbations of each can alter the severity, course and susceptibility to, and resistance to inflammatory, autoimmune and infectious diseases (Sternberg, 1997b). The immune system signals the central nervous system at multiple levels and through many mechanisms, depending on the route of administration, dose and specific cytokine involved. Cytokines can signal the brain via systemic vascular routes (Ericsson et al., 1995), by crossing the blood brain barrier at leaky points (the circumventricularorgans), or by specific active transport uptake mechanisms (Banks and Kastin, 1997). One of the most important mechanisms by which cytokines signal the brain is through activation of second messenger synthesizing enzymes and resultant production of second messengers such as the nitric oxide synthase (NOS)/nitric oxide and the cyclooxygenase/prostaglandin system. Rapid signaling of the central nervous system from the peritoneum can also occur, by cytokine binding to receptors on paraganglia, and subsequent early activation of the nucleus of the tractus solitarius (NTS) in the brainstem via stimulation of the vagus nerve (Bluthe et al., 1994; Watkins et al., 1995a,b). Such nerve-associated lymphoid tissue, or (NALT), may be akin to other peripheral associated lymphoid tissues, such as the bronchial-associated lymphoid tissue (BALT) or gut-associated lymphoid tissue (GALT). This immune system-nervous system communication signalling results in a characteristic

36

set of CNS responses, including fever, and a set of behaviors known as sickness behavior (Elmquist et al., 1997). In turn the nervous system regulates the immune system at multiple levels. The neuroendocrine hypothalamic-pituitary-adrenal (HPA) axis modulates immune responses systemically through the potent anti-inflammatory effects of the glucocorticoids (Sternberg et al., 1992); the sympathetic and parasympathetic nervous systems regulate the immune system regionally through innervation of immune organs such as the spleen, thymus and lymph nodes (Felten and Felten, 1991); and peripheral nerves regulate immune responses locally through innervation at sites of inflammation and release of neuropeptides (Payan, 1989). In general, underactivity of the neuroendocrine stress response predisposes to enhanced susceptibility to inflammatory disease due to lack of suppression by glucocorticoids (Wick et al., 1987; Mason et al., 1990; Sternberg et al., 1992), while overactivity and excess glucocorticoid responses, as occur during stress, tend to suppress immune responses and predispose to infection (Bonneau et al., 1993). Perturbations of the sympathetic nervous system have differential effects on inflammation, depending on the precise site at which the innervation is interrupted (Madden et al., 1994a,b). Local neuropeptides released during inflammation have an overall pro-inflammatory effect (Payan and Goetzl, 1987).

Stress definitions The concept that stress could modify immune responses, and thus inflammatory or infectious diseases, was dismissed until recently by the medical community as unfounded in part because of the vague definition of stress. The popular culture, and early studies, often did not distinguish between the stressful stimulus and the organism’s response to that stimulus. However, recent advances have permitted more precise definition of the triggering events and the many specific neuronal and neuroendocrine response pathways which come into play during stressful situations. These studies indicate that when exposed to a threat, whether physical or psychological, a series of

neuroendocrine and neuronal responses occur which produce a characteristic set of physiological responses. Different stressors may induce different patterns of response and different neural and neuroendocrine pathways are activated. Such specificity in turn can result in varying effects on immune responses and disease outcome. The sub-cortical hypothalamic-pituitary-adrenal (HPA) axis hormonal cascade is the final common pathway of the neuroendocrine stress response. It can be activated by many different forms of stimuli, including psychological, physical and chemical. After exposure to stressful stimuli, the hypothalamus secretes corticotrophin releasing hormone (CRH), which in turn stimulates the pituitary gland to secrete adrenocorticotrophic hormone (ACTH), and the adrenal glands to produce glucocorticoids (Chrousos and Gold, 1992). Glucocorticoid feedback suppresses the hypothalamic-pituitary-adrenal HPA axis cascade at every level. More recent studies indicate that peripheral infectious or inflammatory signals can also stimulate the hypothalamus and pituitary to secrete CRH and ACTH, via the action of cytokines (Berkenbosch et al., 1987; Bernton et al., 1987; Sapolsky et al., 1987). In addition to the HPA axis, stressful stimuli also activate brainstem adrenergic pathways and the sympathetic nervous system. Activation of both these neuronal systems induces a characteristic set of behaviors and physiological responses, including increased heart rate, sweating, focused attention, and decreased vegetative functions, such as feeding and reproductive behavior, together known as the fight-or-flight response.

Neuroendocrine effects on the immune system Most of the effects of the central stress response on the immune system occur through the powerful anti-inflammatory effects of the glucocorticoids. When the anti-inflammatory effects of glucocorticoids was discovered (Hench et al., 1950), it was believed that these effects were pharmacological rather than physiological. More recently, the study of lower, more physiological concentrations of glucocorticoids, has revealed that glucocorticoids play a physiological role of keeping the immune system in check. Furthermore, gluco-

31

corticoids are not wholly anti-inflammatory and immunosuppressive. At lower concentrations, they cause a shift in immune responses, with relatively greater suppression of the pro-inflammatory cytokines tumor necrosis factor alpha, (TNFa), interleukin-1 (IL-1) and less suppression, or actual stimulation of the anti-inflammatory cytokines, interleukin-10 (IL- 10) and the endogenous IL- 1 receptor antagonist (IL-Ira) (DeRijk et al., 1997). The ultimate effect of HPA axis suppression of immune responses through adrenal glucocorticoids, is to suppress irnmunehnflammatory responses as soon as they begin, and prevent the immune response from proceeding unchecked. The physiological role of the HPA axis in regulating the immune system has been most conclusively shown in animal models of inflammatory disease. We showed that inflammatory disease susceptibility in Lewis rats was associated with HPA axis hypo-responsiveness, and inflammatory disease resistance in largely histocompatible Fischer rats was associated with HPA axis hyperresponsiveness (Sternberg et al., 1989). We also showed that the differential ACTH and corticosterone responsiveness in these strains is associated with their relative hypo- and hyper-responsiveness of hypothalamic CRH secretion and synthesis in response to inflammatory and other stress stimuli (Sternberg et al., 1989). These findings provided evidence for the physiologic role of interactions between the immune and central nervous system in susceptibility and resistance to inflammatory disease. Because a simple association between HPA axis hypo-responsiveness and susceptibility to inflammatory disease did not prove a cause and effect relationship between the endocrine and inflammatory disease traits, we treated Fischer rats with the glucocorticoid receptor antagonist RU 486, and Lewis rats with low dose dexamethasone. Interruption of glucocorticoid effects with RU 486 rendered relatively inflammatory resistant Fischer rats susceptible to inflammatory disease, while reconstitution with low dose dexamethasone significantly diminished Lewis inflammatory disease (Sternberg et al., 1989). These studies suggested that the HPA axis played a physiological role in regulating the immune system through the anti-

inflammatory and immunosuppressive effects of the glucocorticoids.

Animal models of autoimmune disease related to hypofunction of the HPA axis Several animal models of enhanced susceptibility to, or severity of, different autoimmunehnflammatory diseases have been associated with blunted HPA axis activity. Thus, similar pre-morbid HPA axis hyporesponsiveness to that seen in Lewis rats, has been shown in animals susceptible to autoimmune inflammatory diseases, including obese strain chickens which develop thyroiditis, lupus prone MRL and NZB mice, and insulin-dependent diabetes mellitus prone NOD mice and BB rats (Wick et al., 1993). Furthermore, not only RU 486, but also surgical intervention, by adrenalectomy or hypophysectomy, renders inflammatory resistant strains susceptible to inflammatory disease (MacPhee et al., 1989; Edwards et al., 1991). Together, these animal studies provide the best evidence that disruption of the immune systemcentral nervous system communication is not simply an epi-phenomenon, but has physiologic relevance for susceptibility to inflammation. Thus, genetic, surgical, pharmacological or toxic disruptions of the HPA axis in animals are associated with enhanced susceptibility to or severity of inflammatory disease. Such interruptions can occur at the level of the hypothalamus, pituitary or adrenal glands, either on a genetic basis, through surgical manipulations, or through pharmacological treatment with drugs which interrupt or reconstitute the axis at various anatomical, biochemical or molecular levels. The converse, reconstitution of the HPA axis in inflammatory susceptible animals can reverse inflammatory susceptibility. Treatment with low dose dexamethasone prevents expression of streptococcal cell wall induced arthritis (Sternberg et al., 1989). Transplantation of fetal hypothalamic tissue intra-cerebroventricularly from inflammatory resistant Fischer rats into inflammatory susceptible Lewis rats, decreases subcutaneous carrageenan inflammation in Lewis rats over 85% (Misiewicz et al., 1997). This effect occurs at least in part through reconstitution of the HPA axis, as transplanted rats

38

show a plasma corticosterone and hypothalamic CRH response equal to that of Fischer rats. The graft tissue in these animals only variably expresses CRH, suggesting that the main effects of such transplants on the HPA axis occur through stimulation of the host hypothalamic CRH expression.

Sympathetic and peripheral nerve-immune system interactions and inflammatory and infectious disease Interruptions of sympathetic innervation of immune organs also modulate inflammatory and infectious disease outcome and susceptibility. Denervation of lymph node noradrenergic (NA) fibers is associated with exacerbation of inflammation, while systemic sympathectomy or denervation of joints is associated with decreased severity of inflammation (Madden et al., 1994a,b). Treatment of neonatal rats with 6-OH-dopamine, which interrupts both central and peripheral NA systems, is associated with exacerbation of experimental allergic encephalomyelitis (EAE). Pharmacologic studies show decreased inflammation in experimental arthritis with beta-blockade, and decreased severity of EAE in rats treated with p adrenergic agonists. Dual activation of the sympathetic nervous system and HPA neuroendocrine response during stress plays a combined role in modulating host defenses to infection (Hermann et al., 1994). The precise effect of these neural regulatory systems on infectious disease outcome depends on the mechanism of pathogenesis of the infection - whether disease results from the deleterious effects of the organism, or from the host’s inflammatory responses to the organism. In the latter case, suppression of the HPA axis will predispose to more severe sequelae of infection from increased host inflammatory responses. In the former case, greater activation of the HPA axis during stress predisposes to greater sequelae of infection by suppressing the host’s ability to suppress or clear the organism. Thus, enhanced mortality from septic shock is seen in hypophysectomized animals exposed to salmonella (Edwards et al., 1991). Interruptions or stimulation of neural-immune communications at the level of peripheral nerves can also affect inflammatory disease outcome. The

peripheral nervous system affects inflammation through neuropeptides, such as substance P (SP) or vasoactive intestinal polypeptide (VIP), which in general, are pro-inflammatory. These molecules may be released from nerve endings or synapses at sites of inflammation, or they can be synthesized and released by immune cells depending on the type of inflammation involved. Both local capsaicin denervation of small substance P fibers at the level of the lymph nodes, and systemic capsaicin treatment are associated with diminished peripheral inflammation. Helme and Andrews (1985) showed a dissociation of neural effects on the cellular and exudative components of inflammation, in rats pre-treated either with 6-OHDA (chemical sympathectomy), capsaicin (ablation of SP afferent neurons) or surgical denervation. Under these conditions, 6-OHDA treated rats showed a decrease in exudate but no decrease in the cellular component of inflammation, while capsaicin treated rats showed a decrease in both the volume and cellular components of inflammation.

Human autoimmune disease and the neuroendocrine stress response A variety of human autoimmunehnflammatory diatheses have recently been associated with blunted HPA axis responses. These include rheumatoid arthritis (Cash et al., 1992), atopic dermatitis (Buske-Kirschbaum et al., 1997), allergic asthma and the fatigue states, chronic fatigue syndrome (Demitrack et al., 1991) and fibromyalgia (Crofford et al., 1994). Children with juvenile rheumatoid arthritis show both a relatively blunted HPA axis response and perturbations in sympathoneuronal responses to stress. Asymptomatic children with a history of atopic skin disease or allergic asthma, off medication, show blunted salivary cortisol responses to a public speaking and mental arithmetic stress, compared to their age and sex-matched controls. Adults with rheumatoid arthritis show relatively blunted HPA axis responses to the stress of surgery. Other clinical syndromes associated with blunted HPA axis responses include psychiatric syndromes without accompanying immune activation: atypical depression (Gold et al., 1995). These different clinical syndromes show different patterns of blunted HPA

39

axis response, some appearing to be more central and others more peripheral in origin. One major difficulty in carrying out and interpreting neuroendocrine response studies in humans with inflammatory disease, is that inflammation itself activates and re-sets the central stress response. Chronic inflammation acts as a stressor to the HPA axis and alters acute HPA axis responses to other stressors (Harbuz et al., 1993). Furthermore, inflammation and cytokine stimulation of the HPA axis induces a long-lived shift in the stress response from one primarily driven by CRH, to one primarily driven by arginine vasopressin (AVP) (Schmidt et al., 1995). At present, clinical studies show an association but do not prove a cause and effect relationship between blunted HPA axis responses and enhanced inflammatory susceptibility in humans. However, animal studies in which interventions prove cause and effect provide a strong indication that this association between HPA axis responsiveness and inflammatory disease does play an important role in the pathogenesis of many of the clinical features of these syndromes in humans. Overactivity of the neuroendocrine stress response and disease While underactivity of the HPA axis responses predispose to overactivity of the immune system and enhanced susceptibility to inflammation, overactivity of the HPA axis as occurs during stress, predisposes to enhanced susceptibility to infectious disease. Thus, excess secretion of glucocorticoids during chronic stress activation of the HPA axis tends to shift cytokine response patterns away from a pro-inflammatory pattern, and towards an antiinflammatory pattern, by relatively greater suppression of pro-inflammatory cytokines and less suppression, or even stimulation of anti-inflammatory cytokines. The resultant effect on the immune response is to make immune cells less effective in responding to infectious agents, thus predisposing the host to greater susceptibility to, or severity of, infectious disease. The effects of stress on susceptibility to, or severity of, infection has been most clearly shown in animals. In these situations, the effects of stress on the course of infection depend on the pathogene-

sis of the illness, that is, if the pathology is caused by excess of the infectious agent, activation of the stress response and increased glucocorticoids enhance pathology. If the pathology is related to excessive inflammatory responses of the host to the infectious agent, the disease course may be ameliorated by the increased HPA axis and glucocorticoid response. In mice exposed to viral infection with influenza, restraint stress worsens the pulmonary disease (Hermann et al., 1993). However, the glucocorticoid receptor antagonist RU 486 alone does not reverse the effects of restraint stress on infection. Addition of beta-adrenergic antagonists to block the sympathetic nervous system are required to completely abrogate the deleterious effects of restraint stress on viral infection. Similarly, increased glucocorticoid responses during restraint stress increases susceptibility to infection with mycobacterium tuberculosis in susceptible mice (Brown et al., 1993). In humans, during excessive physical stress, as in army rangers, immune responses are suppressed. Immune responses are similarly suppressed in caregivers exposed to chronic psychological stress, and in divorcing couples (Kiecolt-Glaser et al., 1987, 1991). Interestingly, the more severe immune and neuroendocrine effects of the stress of divorce are seen in the female spouse who repeatedly confronts a withdrawing partner, with the wife showing greater sympathetic and HPA axis activation as well as decreased immune and cytokine responses in vitro compared to the male spouse. Medical students exposed to exam stress also show blunted immune responses, decreased antibody formation to hepatitis B vaccination (Glaser et al., 1992) and enhanced susceptibility and severity of viral infection such as the common cold (Cohen et al., 1991). Although direct evidence is still required, accumulating evidence supports the possibility that psychobiological variables mediated through neuroendocrine factors may influence the course of HIV-1 infection (Cole and Kemeny, 1997).

Implications of neuroendocrine-immune interactions for relaxation and wellness strategies Most of the studies to date examining the impact of neuroendocrine-immune interactions in animals

40

and humans have focused on the effects of these interactions on disease. Furthermore, the majority of such studies focus on the effects of excess stimulation or pathological disruptions of the neuroendocrine stress response on immune disease outcomes. To date, little attention has been paid to the effects of stress reduction on neuroendocrine hormone responses, and concomitant effects of stress reduction on immune responsiveness and health. Part of the reason for this is that such studies are difficult to perform in a carefully controlled fashion. The neuroendocrine stress response is tightly regulated at baseline, and in the resting organism, stimulation can increase stress hormone levels, but further relaxation does not blunt these hormone responses below baseline. The immune response is also tightly regulated at baseline, with a strong tendency in the absence of inflammatory stimuli to remain in the inactive state. Thus simply reducing the tone of neuroendocrine regulatory systems through stress reduction would be unlikely to reveal easily detectable differences in immune responsiveness. However, the presence of these inherent measurement difficulties should not be interpreted to mean that stress reduction could not play a role in enhancement of health. Furthermore, such technical challenges and our current minimal understanding of the biological and psychological mechanisms by which different forms of stress reduction operate, should not cast doubt on the ameliorative effects of the many different modes of stress management that have evolved throughout the centuries in different cultures. Thus, meditation, psychotherapy, prayer and exercise all share components which could change an individual's perception of stressful stimuli, thus changing the degree to which stress-responsive neuroendocrine systems are brought into play in response to such stimuli. If psychological and neurohormonal responses can be altered by relaxation techniques, such approaches could also ultimately modulate immune responses after stress exposure, through modulating the many neurohormonal and neurotransmitter pathways through which the nervous and immune systems communicate. In attempting to predict the beneficial effects of such stress reduction paradigms on health, it is

important to remember that each individual brings to a stressful situation not only a set of learned responses which could potentially be modified by such approaches, but also an inherent genetically determined degree of stress- and immune-responsiveness. Thus, as in inbred rat strains, there are humans with greater or lesser neuroendocrine responsiveness to various stressful stimuli (Petrides et al., 1994; Buske-Kirschbaum et al., 1997). It may be that in a stress-hyperresponsiveindividual, such relaxation paradigms might have different degrees of effectiveness than in stress-hyporesponsiveindividuals. Such data is not available at present, but must be considered in designing studies to address these questions. Further complicating the effects of stress reduction paradigms on inflammatory or infectious disease, is the important role played in such diseases of potency and dose of the antigen or infectious agent to which the host is exposed. It is unlikely that stress reduction will have a major effect on inflammatory responses to a specific antigen in an individual with a particularly strong immune reactivity to that antigen. Similarly, if the individual is exposed to a potent infectious agent, or to a large dose of antigen, the ability of stress modification to affect disease outcome would be expected to be relatively low. Removal of the antigen or eradication of the infectious agent with antibiotics must remain first-line approaches to therapy in such situations. However, in the future, better definition of the many specific molecular, neuroanatomical and neuroendocrine mechanisms by which the nervous system regulates the immune system, and by which the immune system signals the nervous system, will permit better understanding of the modes by which relaxation paradigms can improve health and prevent disease. And in the course of such scientific understanding, skepticism and doubt regarding the value and effectiveness of such relaxation techniques in the medical therapeutic armamentarium will be dissipated.

References Banks, W.A. and Kastin, A.J. (1997) Relative contributions of peripheral and central sources to levels of IL-1 alpha in the

41 cerebral cortex of mice: assessment with species-specific enzyme immunoassays. J. Neuroimmunol., 79: 22-8. Berkenbosch, F., van Oers, J., del Rey, A., Tilders, F. and Besedovsky, H. (1 987) Corticotropin-releasing factor-producing neurons in the rat activated by interleukin-l. Science, 238: 524-536. Bernton, E.W., Beach, J.E., Holaday, J.W., Smallridge, R.C. and Fein, H.G. (1987) Release of multiple hormones by a direct action of interleukin-l on pituitary cells. Science, 238: 5 19-52 1. Bluthe, R.M., Walter, V., Parnet, P., Laye, S., Lestage, J., Venier, D., Poole, S., Stenning, B.E., Kelley, K.W. and Dantzer, R. (1994) Lipopolysaccharide induces sickness behaviour in rats by a vagal mediated mechanism. C.R.Acad. Sci. Iii, 317: 499-503. Bonneau, R.H., Sheridan, J.F., Feng, N. and Glaser, R. (1993) Stress-induced modulation of the primary cellular immune response to herpes simplex virus infection is mediated by both adrenal-dependent and independent mechanisms. J. Neuroimmunol., 42: 167-176. Brown, D.H., Sheridan, J., Pearl, D. and Zwilling, B.S. (1993) Regulation of mycobacterial growth by the hypothalamuspituitary-adrenal axis: differential responses of Mycobacterium bovis BCG-resistant and -susceptible mice. Infect. Immun., 61: 4793-800. Buske-Kirschbaum, A., Jobst, S., Psych, D., Wustmans, A., Kirschbaum, C., Rauh, W. and Hellhammer, D. (1997) Attenuated free cortisol response to psychosocial stress in children with atopic dermatitis. Psychosom. Med., 59: 419-26. Cash, J.M., Crofford, L.J., Gallucci, W.T., Sternberg, E.M., Gold, P.W., Chrousos, G.P. and Wilder, R.L. (1992) Pituitaryadrenal axis responsiveness to ovine corticotropin releasing hormone in patients with rheumatoid arthritis treated with low dose prednisone. J. Rheumatol., 19: 1692-1696. Chrousos, G.P. and Gold, P.W. (1992) The concepts of stress and stress system disorders. J.A.M.A., 267(9): 1244-1252. Cohen, S., Qrrell, D.A. and Smith, A.P. (1991) Psychological stress and susceptibility to the common cold. N. Engl. J. Med., 325: 606-612. Cole, S.W. and Kemeny, M.E. (1997) Psychobiology of HIV infection. Crit. Rev. Neumbiol., 1I : 289-32 1. Crofford, L.J., Pillemer, S.R., Kalogeras, K.T., Cash, J.M., Michelson, D., Kling, M.A., Sternberg, E.M., Gold, P.W., Chrousos, G.P. and Wilder, R.L. (1994) Hypothalamicpituitary-adrenal axis perturbations in patients with fibromyalgia. Arthritis Rheum., 37: 1583-1592. Demitrack, M.A., Dale, J.K., Straus, S.E., Laue, L., Listwak, S.J., Kruesi, M.J., Chrousos, G.P. and Gold, P.W. (1991) Evidence for impaired activation of the hypothalamicpituitary-adrenal axis in patients with chronic fatigue syndrome. J. Clin. Endocrinol. Metab., 73: 1224-1234. DeRijk, R., Michelson, D., Karp, B., Petrides, J., Galliven, E., Deuster, P., Paciotti, G., Gold, P.W. and Sternberg, E.M. (1997) Exercise and circadian rhythm-induced variations in plasma cortisol differentially regulate interleukin-1 beta (IL1 beta), IL-6, and tumor necrosis factor-alpha (TNF alpha)

production in humans: high sensitivity of TNF alpha and resistance of IL-6. J. Clin. Endocrinol. Metab., 82: 2182-2 19I . Edwards, C.K.I., Yunger, L.M., Lorence, R.M., Dantzer, R. and Kelley, K.W. (1991) The pituitary gland is required for protection against lethal effects of Salmonella typhimurium. Pmc. Natl. Acad. Sci. USA, 88: 2274-2277. Elmquist, J.K., Scammell, T.E. and Saper, C.B. (1997) Mechanisms of CNS response to systemic immune challenge: the febrile response. Trends Neurosci., 20: 565-570. Ericsson, A., Liu, C., Hart, R.P. and Sawchenko, P.E. (1995) Type 1 interleukin-l receptor in the rat brain: distribution, regulation, and relationship to sites of L - I -induced cellular activation. J. Comp. Neurol., 361: 681-698. Felten, S.Y. and Felten, D.L. (1991) Innervation of lymphoid tissue. In: R. Ader, D.L. Felten and N. Cohen (Eds.), Psychoneumimmunology, Academic Press, Inc., San Diego, pp. 27-61. Glaser, R., Kiecolt-Glaser, J.K., Bonneau, R.H., Malarkey, W., Kennedy, S. and Hughes, J. (1992) Stress-induced modulation of the immune response to recombinant hepatitis B vaccine. Psychosom. Med., 54: 22-29. Gold, P.W., Licinio, J., Wong, M.L. and Chrousos, G.P. (1995) Corticotropin releasing hormone in the pathophysiology of melancholic and atypical depression and in the mechanism of action of antidepressant drugs. Ann. NY Acad. Sci., 771: 716-729. Harbuz, M.S., Rees, R.G. and Lightman, S.L. (1993) HPA axis responses to acute stress and adrenalectomy during adjuvantinduced arthritis in the rat. Am. J. Physiol., 264: 179-185. Helme, R.D. and Andrews, P.V. (1985) The effect of nerve lesions on the inflammatory response to injury. J. Neumscience Res., 13: 453459. Hench, P. S., E. C. Kendall, et al. (1950) Effects of cortisone acetate and pituitary ACTH on rheumatoid arthritis, rheumatic fever and certain other conditions. Arch. Intern. Med., 85: 545-666. Hermann, G., Beck, EM., Tovar, C.A., Malarkey, W.B., Allen, C. and Sheridan, J.F. (1994) Stress-induced changes attributable to the sympathetic nervous system during experimental influenza viral infection in DBN2 inbred mouse strain [published erratum appears in J. Neumirnmunol., 1994 55: 2211. J. Neuroimmunol., 53(2): 173-180. Hermann, G., Tovar, C.A., Beck, EM., Allen, C. and Sheridan, J.F. (1993) Restraint stress differentially affects the pathogenesis of an experimental influenza viral infection in three inbred strains of mice. J. Neuroimmunol., 47: 83-94. Kiecolt-Glaser, J.K., Dura, J.R., Speicher, C.E., Trask, O.J. and Glaser, R. (1991) Spousal caregivers of dementia victims: longitudinal changes in immunity and health. Psychosom. Med., 53: 345-362. Kiecolt-Glaser, J.K., Fisher, L.D., Ogrocki, P., Stout, J.C., Speicher, C.E. and Glaser, R. (1987) Marital quality, marital disruption, and immune function. Psychosom. Med., 49: 13-34.

42 MacPhee, I.A.M., Antoni, F.A. and Mason, D.W. (1989) Spontaneous recovery of rats from experimental allergic encephalomyelitis is dependent on regulation of the immune system by endogenous adrenal corticosteroids. J. Exp. Med., 169: 431445. Madden, K.S., Felten, S.Y., Felten, D.L., Hardy, C.A. and Livnat, S. (1994) Sympathetic nervous system modulation of the immune system. 11. Induction of lymphocyte proliferation and migration in vivo by chemical sympathectomy. J. Neuroimmunol., 49: 67-75. Madden, K.S., Moynihan, J.A., Brenner, G.J., Felten, S.Y., Felten, D.L. and Livnat, S. (1994) Sympathetic nervous system modulation of the immune system. ID.Alterations in T and B cell proliferation and differentiation in vitro following chemical sympathectomy. J. Neuroimmunol., 49: 77-87. Mason, D., MacPhee, I. and Antoni, F, (1990) The role of the neuroendocrine system in determining genetic susceptibility to experimental allergic encephalomyelitis in the rat. Immunology, 70: 1-5. Misiewicz, B., Poltorak, M., Raybourne, R.B., Gomez, M., Listwak, S. and Sternberg, E.M. (1997) Intracerebroventricular transplantation of embryonic neuronal tissue from inflammatory resistant into inflammatory susceptible rats suppresses specific components of inflammation. Exp. Neurol., 146: 305-314. Payan, D.G. (1989) Neuropeptides and inflammation: the role of substance P. Annu. Rev. Med., 40: 341-352. Payan, D.G. and Goetzl, E.J. (1987) Dual roles of substance P: Modulator of immune and neuroendocrine functions. Ann. NYAcad. Sci., 512: 465-475. Petrides, J.S., Mueller, G.P., Kalogeras, K.T., Chrousos, G.P., Gold, P.W. and Deuster, P.A. (1994) Exercise-induced activation of the hypothalamic-pituitary-adrenalaxis: marked differences in the sensitivity to glucocorticoid suppression. J. Clin. Endocrinol. Metab., 79: 377-383. Sapolsky, R., Rivier, C., Yamamoto, G., Plotsky, P. and Vale, W. (1987) Interleukin-1 stimulates the secretion of hypothalamic corticotropin releasing factor. Science, 238: 522-524.

Schmidt, E.D., Janszen, A.W., Wouterlood, F.G., Tilders, F.J. (1995) Interleukin-1-induced long-lasting changes in hypothalamic corticotropin-releasing hormone (CRH)-neurons and hyperresponsiveness of the hypothalamus-pituitaryadrenal axis. J. Neurosci., 15: 7417-7426. Sternberg, E.M. (1997a) Emotions and disease: from balance of humors to balance of molecules. Nat. Med., 3: 264267. Sternberg, E.M. (1997b) Neural-immune interactions in health and disease. J. Clin. Invest., 100: 2641-2647. Sternberg, E.M., Chrousos, G.P., Wilder, R.L. and Gold, P.W. (1992) The stress response and the regulation of inflammatory disease. Ann. Intern. Med., 117: 854-866. Sternberg, E.M., Hill, J.M., Chrousos, G.P., Kamilaris, T., Listwak, S.J., Gold, P.W. and Wilder, R.L. (1989) Inflammatory mediator-induced hypothalamic-pituitary-adrenal axis activation is defective in streptococcal cell wall arthritissusceptible Lewis rats. Proc. Natl. Acad. Sci. USA, 86: 23742378. Sternberg, E.M., Young, W.S., Bernardini, R., Calogero, A.E., Chrousos, G.P., Gold, P.W. and Wilder, R.L. (1989) A central nervous system defect in biosynthesis of corticotropinreleasing hormone is associated with susceptibility to streptococcal cell wall-induced arthritis in Lewis rats. Proc. Natl. Acad. Sci. USA, 86: 4771-4775. Watkins, L.R., Goehler, L.E., Relton, J.K., Tartaglia, N., Silbert, L., Martin, D. and Maier, S.F. (1995) Blockade of interleukin-1 induced hyperthermia by subdiaphragmatic vagotomy: evidence for vagal mediation of immune-brain communication. Neurosci. Lett., 183: 27-3 1. Watkins, L.R., Maier, S.F. and Goehler, L.E. (1995) Cytokineto-brain communication: a review and analysis of alternative mechanisms. Life Sci., 57: 101 1-1026. Wick, G., Hu, Y., Schwarz, S. and Kroemer, G. (1993) Immunoendocrine communication via the hypothalamopituitary-adrenal axis in autoimmune diseases. Endocrine Rev., 14: 539-563. Wick, G., Kromer, G., Neu, N., Fassler, R., Ziemiecki, A., Muller, R.G., Ginzel, M., Beladi, I., Kuhr, T.and Hala, K. (1987) The multi-factorial pathogenesis of autoimmune disease. Immunol. Lett.. 16: 249-257.

E.A. Mayer and C.B.Saper (Fds.) Pmgress in Brain Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 5

Depression really does hurt your heart: stress, depression, and cardiovascular disease Dominique L. Musselman and Charles B. Nemeroff * Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 1639 Pierce Drive, Suite 4000, Atlanta, GA 30322, USA

Introduction The interplay of stress, personality traits, and psychiatric symptoms syndromes with the cardiovascular system, have long intrigued investigators in their search for psychosocial and psychiatric contributions for development and progression of atherosclerotic heart disease. Surveillance of the established risk factors for atherosclerotic heart disease (IHD), e.g. smoking, hypertension, hypercholesterolemia, and age, leave a substantial portion of the differences in rates of IHD unexplained. Indeed, the Type A personality pattern has been extensively studied as a risk factor for coronary artery disease, (CAD) (Friedman and Rosenman,1959; Booth-Kewley and Friedman, 1987; Hayward, 1995). However, increasing evidence is accumulating suggesting that an affective disorder, major depression, is a major contributing factor, not only to elevated morbidity and mortality after an index myocardial infarction (MI), but as an independent risk factor in the development of atherosclerotic heart disease.

Cornorbid depression in patients with cardiovascular disease Major depression and depressive symptoms are considerably more common in patients with a *Corresponding author. Tel.: (404) 727-8382; Fax: (404) 727-3233

variety of medical illnesses than in individuals in the general population. Indeed receiving a diagnosis of a serious, life-threatening illness often results in a cascade of emotional responses, including feelings of sadness, shock, disbelief, anxiety, or a myriad of other emotional responses. However, when these symptoms are of such magnitude and duration that they fulfill DSM-IV (American Psychiatric Association, 1994) criteria for major depression, accurate diagnosis and treatment is essential. For a patient to fulfill diagnostic criteria for major depression, the patient must first have symptoms of dysphoria and/or anhedonia (defined as a loss of interest or pleasure), pervasively for at least a 2-week period. Patients must also have at least four of the following symptoms (or at least three of the following symptoms if they have both dysphoria and anhedonia): sleep disturbance, change in appetite, fatigue, psychomotor retardation and/or agitation, low-self-esteem and/or guilt, poor concentration and/or indecisiveness, and thoughts of suicide and/or suicidal ideation (American Psychiatric Association, 1994). Major depression is the most common disorder observed in the primary care setting (Katon and Sullivan, 1990) with prevalence rates of major depression in primary care outpatients that range from 2-16%, and 9-20% for all affective disorders (Hoeper et al., 1979; Leeper et al., 1985; Schulberg et al., 1985; Blacker and Clare, 1987; Barrett et al., 1988; Von Korff et al., 1987; Cohen-Cole and Kaufman,

44

1993). The Epidemiologic Catchment Area (ECA) study, an extensive survey completed after publication of DSM-I11 (American Psychiatric Association, 1980) interviewed 20 291 adults (ages 18 and older) between 1980 and 1984 using the Diagnostic Interview Schedule (DIS) (Robins et al., 1981). Subjects were interviewed in a variety of settings: community households, long-term treatment mental hospitals, nursing homes, and correctional institutions. The ECA study [based on DSM-I11 (American Psychiatric Association,1980) criteria] reported the following 1-month prevalence rates (a guide in determining the number in the population with a disorder at a given point in time) per 100 persons: major depression (1.8), dysthymia (3.3), bipolar I (0.4), and bipolar I1 disorder (0.2) (Regier et al., 1988). Another comprehensive epidemiologic project, the National Comorbidity Study, has been more recently completed (Kessler et al., 1994). Data was derived by interviewing 8098 non-institutionalized persons aged 15 to 54 years in the 48 coterminous states. Seventeen percent of the respondents reported a history of major depression, 1.6% a manic episode, and 6% dysthymia (lifetime prevalence rates based on DSM-111-R criteria). Studies of medical inpatients that have employed structured interviews have reported even higher prevalence rates: 8% for major depression and 1536% for all mood disorders (Magni et al., 1986; Feldman et al., 1987). Early studies reported the prevalence of depression to be from 18% to 60% in patients with CAD

(Wynn 1967; Hackett et al., 1968; Cay et al., 1972; Stem et al., 1977; Kurosawa et al., 1983). Later studies reported more consistent prevalence rates of depression in patients with cardiovascular disease (CVD), ranging from 16% to 23% (mean, 19%; median, 18%), despite the potential methodological weaknesses of some of the studies summarized in Table 1. These include utilization of unmodified psychiatric diagnostic instruments to determine the prevalence of depression, excluding patients because of age or severity of CVD, measuring the presence of depression at different time intervals after admission, and excluding patients with cognitive deficits. Moreover methodological differences between the various studies also exist including dissimilar patient populations, different diagnostic instruments, hospitalization status of some subjects, and unspecified type of heart disease. Unfortunately, depression in patients with CAD is rarely diagnosed by primary care physicians and cardiologists (Wynn, 1967; Mayou et al., 1979; Kurosawa et al., 1983; Carney et al., 1987; Schleifer et al., 1989; Frasure-Smith et al., 1993). Although severity of physical illness is generally considered one of the most important variables associated with depression in patients with other medical illnesses, most studies of patients with CVD do not document a higher prevalence rate of depression in patients with measures of more advanced CVD or with a greater level of disability (Carney et al., 1987; Schleifer et al., 1989; FrasureSmith et al., 1993, 1995). In studies examining psychological factors in patients with CVD (Ruber-

TABLE 1 Prevalence of major depression in patients with cardiovascular disease (Musselman et al., 1998) Study

# m p e of pts.

Diagnostic Method

Prevalence ~~~~

Carney et al. 1988 Schleifer et al. 1989 Frasure-Smith et al. 1993 Gonzalez et al., 1996

52 CAD pts. undergoing elective cardiac catheterization 283 pts. hospitalized with MI 222 pts. hospitalized with MI 99 hospitalized pts. with CAD

DIS (Robins et al., 1981)

18%

SADS (Endicott and Spitzer, 1978)

18%

DIS

16%

DIS

23%

Abbreviations:CAD = coronary artery disease; MI =myocardial infarction; DIS = Diagnostic Interview Schedule,Version III; SADS =Schedule for Affective Disorders and Schizophrenia. Used with permission (Archives of General Psychiatry).

45

man et al., 1984; Frasure-Smith and Prince, 1985; Ragland and Brand, 1988; Wiklund et al., 1988; Ahern et al., 1990; Frasure-Smith, 1991; Case et al., 1992; Williams et al., 1992; Frasure-Smith et al., 1992, 1993), those at greatest risk are unmarried (Wiklund et al., 1988; Case et al., 1992; Williams et al., 1992), women (Schleifer et al., 1989; Frasure-Smith et al., 1993), with low income (Williams et al., 1992), and low educational level (Ruberman et al., 1984), a Type A or Type B personality pattern (Ragland and Brand, 1988; Schleifer et al., 1989; Ahern et al., 1990), with a history of a previous cardiac event (Allison et al., 1995), who continued smoking (Allison et al., 1995)’ and have ongoing life stress (Ruberman et al., 1984; Schleifer et al., 1989). However several of these studies were comprised solely of male subjects (Frasure-Smith et al., 1985, 1992; FrasureSmith and Prince, 1985; Wiklund et al., 1988; Ragland and Brand, 1988; Frasure-Smith, 1991).

Depression as a risk factor for ischemic heart disease Patients with major depression have been observed to have elevated mortality rates from CVD ever since the pioneering report of Malzberg (1937). The finding that a psychiatric illness such as major depression renders one more susceptible for developing ischemic heart disease (IHD) remains controversial, and has been often intuitively ‘explained’ by the hypothesis that individuals with psychiatric disorders generally have an increase in established risk factors for the development of CAD, e.g. poor diets, minimal exercise, increased rates of smoking, etc. Table 2 summarizes the most rigorously designed studies; they are prospective in design, controlled for demographic factors (such as age, sex, socioeconomic status), utilizing structured clinical interviews or diagnostic instruments, and included analysis of other known risk factors for CVD. Nearly all of these recent studies document increased cardiovascular morbidity and mortality in patients with depressive symptoms or major depression thereby implicating depression as an independent risk factor in the pathophysiologic progression of cardiovascular disease, rather than

merely a secondary emotional response to cardiovascular illness. Advances in biological psychiatry have included the discoveries of a number of neurochemical, neuroendocrine, and neuroanatomic alterations in unipolar depression. Certain of these markers may also reflect important pathophysiologic alterations associated with the syndrome of major depression which may contribute to the increased vulnerability of depressed patients to CVD: hyperactivity of the sympathoadrenal (SA) and hypothalamic-pituitaryadrenal (HPA) axis, diminished heart rate variability, ventricular instability and myocardial ischemia in reaction to mental stress, and alterations in platelet activation and aggregation.

Biology: potential pathophysiologic mechanisms Hypothalamic-pituitary-adrenocortical and sympathomedullary hyperactivity

Two primary components central to the ‘fight or flight’ stress response observed by Cannon in 1911 (Vingerhoets, 1985) and the ‘general adaptation syndrome’ described by Selye (1956) are HPA axis and SA system. A myriad of studies have documented evidence of HPA axis hyperactivity in drug-free patients with major depression, i.e. elevated CSF CRF concentrations (Nemeroff et al., 1984; Arato et al., 1986; Banki et al., 1987, 1992; France et al., 1988; Risch et al., 1992), blunting of the ACTH response to CRF administration, nonsuppression of cortisol secretion following dexamethasone administration, hypercortisolemia, pituitary and adrenal gland enlargement, as well as direct evidence of increased numbers of hypothalamic CRF neurons in post-mortem tissue of depressed patients (Raadsheer et al., 1994, 1995). Administered corticosteroids have long been known to induce hypercholesterolemia, hypertriglyceridemia, and hypertension. Other atherosclerosis-inducing actions of steroids include injury of vascular endothelial cells (Bjorkerud 1973) and intima (Kemper et al. 1957; Nahas et al. 1958; Valigorsky 1969), and inhibition of normal healing (Ross and Harker, 1976). Indeed elevated morning plasma cortisol concentrations have been

46

significantly correlated with moderate to severe coronary atherosclerosis in young and middle-aged men (Troxler et al. 1977). Many patients with major depression also exhibit dysregulation of the SA system. The adrenal medulla and sympathetic nervous system (SNS) together comprise the SA system. Although the CNS regulation of the SA system is only partially characterized, hypothalamic CRF-containing neu-

rons provide stimulatory input to several autonomic centers involved in regulating sympathetic activity (Merchenthaleret al., 1982; Cummings et al., 1983; Swanson and Sawchenko, 1983). Nerve impulses from CNS regulatory centers control catecholamine release from the SA system. Physiological and pathological conditions causing SA activation include physical activity, coronary ischemia and heart failure, and mental stress.

TABLE 2 Antecedent depression and subsequent risk of cardiovascular disease (Musselman et al., 1998) Source

# m p e of Pts.

1990 male pts. Western Electric employees Brozek et al., 1966 258 men Goldberg et al., 1979 82 pairs (male and female) of case/ control subjects randomly selected from two communities Murphy et al., 1987 1003 male and female subjects from the community Ostfeld et al., 1964

Anda et al., 1993

2832 men and women age 45-77 Aromaa et al., 1994 5,355 men and women age 40-64 Ford et al., 1994 1198 men Vogt et al. 1994 1187menand 1386 women (aged 18 and older) in a HMO Simonsick et al. 1995 1063 men and 2398 women (aged 65 years and older with hypertension) Everson et al., 1996 2428 men (age 42 to 60) Barefoot and Schroll. 1996 Pratt et al. 1996 WassertheilSmoller et al., 1996

Diagnostic Method

Relative Risk (RR) of Major Depression or Depressive Sxs for Cardiac Disease or Cardiac Disease-Related Death

MMPI (Hathaway and McKinley, 1951) 16 PF (Catell et al., 1957)

None

MMPI (Dahlstrom and Welsh, 1960) CES-D Scale (Radloff, 1977) (as well as 4 other depression scales)

None None

DPAX algorithm (Murphy et aL.1985)

None For cardiovascular diseaserelated death: Men: 2.5 Women: 1.5 For IHD-related death: 1.5-2.1

Depression subscale of GWS (Dupuy, 1977) PSE (Wing et al., 1974)

For MI: men 2.62 women 1.90

Depression questionnaire For MI: 1.9 Depression scale (McFarland et al. 1985) Depressive symptoms nor related to incidence of CVD CES-D Scale (Radloff, 1977)

Elevated rates of CVD-related death in women with high

scores of depressive symptoms Hopelessness selfreport questionnaire

409 men MMPI (Greene, 1991) 321 women (all born in 1914) 1551 men DIS (Robins et al., 1981) and women 4736 men and women CES-D (Berkman et al., 1986) over 60 with hypertension

For CVD-related death: RH: 2.52 (moderate hopelessness score) RH: 3.90 (high hopelessness scores) For MI: 1.7 For MI: 2.07 (hx of dysphoria) 4.54 (hx of MDE) Baseline CES-D score 5 16 did not predict future MI; RR of future MI per 5-unit increase in CES-D score: women 1.2

Abbreviations: RH = relative hazard; IHD = ischemic heart disease: MI = myocardial infarction: hx = history; MDE = episode of major depression. Used with permission (Archives of General Psychiatry).

41

Hypersecretion of NE in unipolar depression has been documented by elevated plasma NE and NE metabolite concentrations (Wyatt et al., 1971; Louis et al., 1975; Roy et al., 1988; Veith et al., 1994), and elevated urinary concentrations of NE and its metabolites (vide infra). Not only do depressed patients exhibit higher basal plasma concentrations of NE, but those with melancholia exhibit even greater elevations in plasma NE when subjected to orthostatic challenge than normal control subjects and depressed patients without melancholia (Roy et al., 1987). Following treatment with TCAs, both urinary excretion of NE and its metabolites and plasma NE concentrations fall (Sulser et al., 1978; Scubee-Moreau et al., 1979; Charney et al., 1981; Linnoila et al., 1982, 1986; Golden et al., 1988), though Veith and colleagues (1994) reported that long-term treatment with desipramine increased plasma NE concentrations. Even normotensive depressed patients have been found to exhibit greater heart rates at rest, after orthostasis and exercise, in comparison to normal controls. These depressed patients also exhibited increased plasma concentrations of NE and serotonin (5HT)at rest (Lechin et al., 1995). SA hyperactivity contributes to the development of CVD through effects of catecholamines upon the heart, blood vessels, and platelets. SA activation modifies the function of circulating platelets through direct effects upon the platelet, catecholamine-induced changes of hernodynamic factors (increased shear-stress), circulating lipids, and inhibition of vascular eicosanoid synthesis (Anfossi and Trovati, 1996). Arachidonic acid metabolites, e.g. prostaglandins and leukotrienes, contribute to diverse circulatory and hemostatic functions including inhibition of platelet aggregation, and vascular contractility and permeability (Gerritsen, 1996). Thus the SA hyperactivity observed in many patients with major depression may contribute to the development of CVD via effects of catecholamines upon cardiac function and platelets.

Diminished heart rate variability Alterations in autonomic nervous system activity, as demonstrated by reduced heart rate variability (HRV), represent another potential mechanism

contributing to the diminished survival of depressed patients with CVD. It is believed that the beat-to-beat fluctuations in hemodynamic parameters reflect the dynamic response of the cardiovascular control systems to a myriad of naturally occurring physiological perturbations, e.g. fluctuations in heart rate associated with respiration. Therefore fluctuations in heart rate may provide a sensitive measure of the functioning of the rapidly reacting sympathetic and parasympathetic systems. Cardiovascular homeostasis is maintained by the two divisions of the autonomic nervous system, the parasympathetic and sympathetic, via afferent pressoreceptors and chemoreceptors, and efferents which alter heart rate, atrioventricular conduction, and contractility, and impinge upon the peripheral vasculature, altering arterial and venous vasomotor tone (Akselrod et al., 1981). HRV is the standard deviation of successive Rto-R intervals in sinus rhythm and reflects the interplay and balance between sympathetic and parasympathetic input on the cardiac pacemaker. Peripheral control of HRV is mainly via the parasympathetic cholinergic vagus nerve (Low, 1993). Central generation and control of heart rate is regulated by the hypothalamus, the limbic system, and the brainstem. Numerous CNS neurotransmitters are involved in modulating HRV, including acetylcholine, NE, 5HT,and dopamine (Spyer, 1988; Shields, 1993). A high degree of HRV is observed in normal hearts with good cardiac function, whereas HRV can be significantly decreased in patients with severe CAD or heart failure (Dalack and Roose, 1990). Moreover the relative risk of sudden death after acute MI is significantly higher in patients with decreased HRV (Wolf et al., 1978; Billman et al., 1982; Kleiger et al., 1987; Bigger et al., 1988; LaRovere et al., 1988; Cripps et al., 1991). HRV is one of many prognostic factors post-infarction [e.g. age, left ventricular ejection fraction (LVEF), frequency of arrythmias]. Although positive predictive accuracy is not high when HRV is considered in combination with other prognostic factors, (Viskin and Belhassen, 1992; ArayaGomez et al., 1994), clinically useful levels of negative predictive accuracy can be achieved

48

(Viskin and Belhassen, 1992; Araya-Gomez et al., 1994; Campbell, 1996). Autonomic tone, of many arrhythmogenic factors, is the most difficult to measure (Campbell, 1996) - thus interest in HRV continues. The low frequency (LF) power of the heart period power spectrum reflects modulation of sympathetic and vagal tone by baroreflex activity (Koizumi et al., 1985), while high frequency (HF) power reflects modulation of vagal tone, primarily by respiratory frequency and depth, i.e. the respiratory sinus arrhythmia (Katona and Jih, 1975; Fouad et al., 1984). The physiological mechanisms contributing to ultra low (ULF) and very low (VLF) frequency power of the heart period spectrum (which account for more than 90% of the total power in a 24-hour period) remain obscure. In a study of 715 patients post-MI, certain frequency bands (total, ULF and VLF) of the heart period power spectrum were strongly associated with mortality during four years of follow-up, even after adjustment for other major risk factors. Indeed VLF power was most strongly associated with death secondary to arrhythmia (Bigger et al., 1992). Reduced HF HRV has been observed in depressed patients in comparison to non-depressed groups (Dalack and Roose, 1990; Miyawaki and Salzman, 199l), though discrepant reports exist (Yeragani et al., 1992; Rechlin et al., 1994). In patients with angiographically confirmed CAD, diminished HRV during 24-hour Holter monitoring was significantly more common in depressed patients than in matched non-depressed patients (Carney et al., 1995). Diminished HF HRV is thought to reflect decreased parasympathetic tone, possibly predisposing to ventricular arrhythmia, and perhaps to the excessive cardiovascular mortality found in CVD patients with comorbid MDD (Roose et al., 1989). One study (without a placebo control group) revealed normalization of reduced HRV of depressed patients after effective treatment (Balogh et al., 1993). The prognostic importance of antidepressant-induced improvement in diminished HRV in depressed patients remains an intriguing area of research. Subsequent investigation will seek to determine the processes underlying ULF and VLF frequency bands of the heart power spectrum; whether or not these bands are altered in depressed patients (with or without CVD), remains obscure.

Myocardial ischemia and ventricular instability in reaction to mental stress It has long been thought that the combination of a vulnerable myocardium after MI, acute ischemia, and negative emotional arousal can trigger fatal ventricular arrhythmias (Verrier, 1990). Indeed the interplay of these factors in CAD patients is being increasingly scrutinized. Jiang and colleagues (1996) followed 126 patients with CAD over a 5-year period; mental stress-induced myocardial ischemia at baseline in CAD patients was associated with significantly higher rates of subsequent fatal and non-fatal cardiac events, independent of age, baseline LVEF, and previous MI. This study proposed that the relationship between psychological stress and adverse cardiac events is mediated by myocardial ischemia. Although myocardial ischemia is likely to be the most significant factor in predisposition to ventricular instability, other factors contribute. CNS control mechanisms can significantly decrease the threshold for ventricular fibrillation (Lown et al., 1980). Ventricular fibrillation is believed to be the mechanism underlying sudden cardiac death, the most common cause of fatality among patients with CAD (Lown and Verrier, 1976). Indeed psychological stress in humans with CAD increases ventricular ectopic activity and increases the risk of ventricular fibrillation (Tavazzi et al., 1986; Follick et al., 1988). The vagus nerve, however, exerts antiarrhythmic activity by direct action on the ventricular myocardium and interference with sympathetic activity (Zaza and Schwartz, 1985). Increased parasympathetic activity has a protective effect on myocardium electrically destabilized by increased adrenergic tone (Lown and Verrier, 1976). Frasure-Smith and colleagues (1995) proposed that depression worsens prognosis post-MI via another mechanism, premature ventricular contractions (PVCs). Indeed the risk of sudden cardiac death associated with significant depressive symptoms was greatest among patients with > 10 PVCs per hour (60% of these patients died within 18 months), suggesting arrhythmia as the link between depression and sudden cardiac death (FrasureSmith et al., 1995). Depressed CAD patients are not more likely to have arrhythmias than CAD

49

patients without depression, but the risk associated with depression is largely confined to patients with PVCs. Patients who were not depressed experienced little increase in risk associated with PVCs, even in the presence of low LVEF (Frasure-Smith et al., 1995). Thus the prognostic impact of PVCs may be more related to depression than to PVCs per se. In the Cardiac Arrhythmia Suppression Trial (CAST) (Echt et al., 1991), suppression of PVC frequency in post-MI patients did not reduce, and actually increased mortality even though previous studies revealed that PVCs are associated with increased post-MI mortality. Treatment of depression may be one necessary component to improve survival in patients with PVCs.

Alterations in platelet receptors and/or reactivity The adverse effects of depression on cardiovascular disease may be plausibly mediated via platelet mechanisms. Markovitz and Matthews (1991) first proposed that enhanced platelet responses to psychological stress might trigger adverse coronary artery ischemic events. This association between platelet activation and vascular disease is indirectly supported by studies linking cerebrovascular disease and depression (Simonsick et al., 1995; Morris et al., 1993). Platelets play a central role in hemostasis, thrombosis, development of atherosclerosis, and acute coronary syndromes (Lefkovits et al., 1995) through their interactions both with subendothelial components of damaged vessel walls, and plasma coagulation factors, primarily thrombin. Human platelets contain adrenergic, serotonergic and dopaminergic receptors. Through stimulation of the a,-adrenoceptors on platelet membranes, increases in circulating catecholamines potentiate the effects of other agonists and, at higher concentrations, initiate platelet responses, including secretion, aggregation, and activation of the arachidonate pathway. Following injury to vessel endothelium, platelets and circulating leukocytes attach to the newly exposed subendothelial layer. Platelet adhesion to collagen (and other components of the subendothelial matrix) exposed within a denuded area of the vascular endothelium and thrombin

stimulates platelet activation. Activation converts platelet membrane GPIIbflIIa complexes into functional receptors for fibrinogen. Activation is also accompanied by extrusion or secretion of platelet storage granule contents into the extracellular environment. Platelets activated at the interface with a vessel wall injury accelerate the local formation of thrombin and release a variety of products from their storage granules, including chemotactic and mitogenic factors inducing leukocyte migration from the blood stream and vascular cell proliferation. These secreted platelet products, e.g. platelet factor 4 (PF4), P-thromboglobulin (pTG), and 5HT, stimulate and recruit other platelets, cause irreversible platelet-platelet aggregation, leading to formation of a fused platelet thrombus. Platelets also contribute to vascular damage through stimulation of lipoprotein uptake by macrophages and mediating vasoconstriction through the production and/or release of substances such as thromboxane A2 (TXAJ, platelet-activating factor (PAF) and 5HT (Anfossi and Trovati, 1996). Clinical trials have confirmed the importance of the platelet in vascular damage; antiaggregating medications are useful in secondary prevention (Hess et al., 1985; Antiplatelet Trialists’ Collaboration, 1988; Verstraete, 1991) and delay progression of atherosclerotic lesions (Ridker et al., 1991). We wondered whether heightened susceptibility to platelet activation might be a mechanism by which depression in physically healthy, young individuals acts as a significant risk factor for heart and cerebrovascular disease and/or mortality after myocardial infarction. Utilizing fluorescence-activated flow cytometric (FAFC) analysis we discovered that, in comparison to normal controls, depressed patients exhibited enhanced baseline platelet activation and responsiveness (Musselman et al., 1996). Of note is that these depressed patients did not have any of the commonly accepted risk factors for CVD, including nicotine or alcohol abuse, hypercholesterolemia,obesity, hypertension, or have a family history of premature heart disease. In another recent study, patients suffering from comorbid CVD and major depression exhibited increased platelet activation as measured by markedly elevated plasma concentrations of platelet

50

secretion products, PF4 and P-TG, when compared to healthy control subjects and non-depressed patients with CVD (Laghrissi-Thode et al., 1997). Although the mechanism(s) responsible remain unknown, we believe that heightened susceptibility to platelet activation and secretion underlies, at least in part, the increased vulnerability of depressed patients to CVD and/or mortality after myocardial infarction. 5HT secreted by platelets induces both platelet aggregation and coronary vasoconstriction, both mediated by 5HT2receptors. Vasoconstriction especially occurs when normal endothelial cell counterregulatory mechanisms of vascular relaxation are defective, as often occurs in patients with CAD (DeClerck, 1991; Weyrich et al., 1992; Laghrissi-Thode et al., 1997). Indeed, essential hypertension, elevated plasma cholesterol, older age, and smoking, well-known predisposing factors for development of CVD, all contribute to 5HTmediated platelet activation. 5HT-mediated platelet activation can contribute to the development of atherosclerosis, thrombosis, and vasoconstriction. Even though 5HT itself is a weak platelet agonist, it markedly amplifies platelet reactions to a variety of other agonists such as ADP, thromboxane A2, catecholamines, or thrombin. By an action on 5HT2 receptors, 5HT enhances the extent of platelet aggregation and the release of intragranular products and arachidonic acid metabolites in response to otherwise ineffective agonist concentrations (DeClerck, 1991). Such serotonergic platelet amplification occurs at low concentrations attained when 5HT is released from seeping platelets subjected to shear stresses (Osim and Wyllie, 1982) and from platelet activation by contact with an arterial wall lesion (Ashton et al., 1986, 1987). Platelets of depressed patients exhibit significantly increased elevations of [Ca”], after 5HT-stimulation in comparison to controls (Kusumi et al., 1991; Mikuni et al., 1991; Eckert et al., 1993). Even functionally trivial increases in intraplatelet calcium ‘prime’ the platelet secretion and aggregation response to stimulation by even a ‘weak’ agonist (such as 5HT) (Ware et al., 1987) or in response to increased rate of blood flow. Thus platelets with elevated [Ca”],, as observed in

depressed patients, would likely exhibit increased activation in comparison to normal comparison subjects under basal conditions or in response to shear-induced aggregation (e.g. following an orthostatic challenge). Future investigations will seek to confirm and interconnect the pathophysiological mechanisms of sympathoadrenal hyperactivity, exaggerated platelet reactivity, and alterations in the platelet 5HT system in depressed patients to their propensity for the development of CVD.

Treatment of depression in patients with cardiovasculardisease Recognition and treatment of major depression is crucial, especially for patients after myocardial infarction (MI), because depressive disorders adversely affect compliance with medical therapy (Blumenthal et al., 1982) and psychosocial rehabilitation (Stem et al., 1977; Mayou et al., 1978), increase medical comorbidity (Stem et al., 1977), predict future cardiac events (Carney et al., 1987; 1988; Aromaa et al., 1994), and hasten mortality (Ahern et al., 1990; Frasure-Smith et al., 1993, 1995). Treatment of the 15 to 23% of post-MI patients who fulfill criteria for major depression, as well as those with significant dysphoria but who have subsyndromal depression, may have a significant impact on both medical morbidity and mortality. Although not all have produced positive results, some intervention trials in patients with CVD have reported effective reduction of Q p e A behavior and emotional distress and reduction in cardiac morbidity and mortality. Friedman and colleagues studied 1013 post MI patients who were observed for 4.5 years to determine whether their type A behavior could be altered and what the effect of such a behavioral change would be on subsequent cardiac morbidity and mortality rates. Two hundred and seventy people were randomized to a control group who received group counseling (advice and information concerning diet, exercise, drugs, possible surgical regimens, and heart pathophysiology) while 592 people were assigned to an experimental group who received both group counseling and Q p e A behavioral counseling (instruction in progressive muscle relaxation, behavior alteration techniques,

51

changes in certain belief systems, restructuring of various environmental situations, etc . . .). The remaining 151 patients served as a ‘no counseling’ comparison group. v p e A behavior was markedly reduced in 35% of the participants given cardiac and type A behavior counseling compared with 10% of participants receiving cardiac counseling only. The cumulative rate of cardiac recurrence (non-fatal infarctions and cardiac deaths) was 13% in the experimental group that received type A counseling and group counseling, significantly less than that observed in the group counseling group (21%) or in the no treatment group (28%). After the first year, a significant difference in the number of cardiac deaths between the experimental and control participants was observed during each of the remaining 3.5 years of the study (Friedman et al., 1986).

The impact of high levels of psychological stress (Frasure-Smith, 1991) was examined over five years among 461 men post-MI who took part in a trial of psychological stress monitoring and intervention. Psychological distress was assessed using the 20-item General Health Questionnaire (GHQ) one to two days before hospital discharge. Once discharged, patients in the treatment group responded to the GHQ by telephone on a monthly basis, and, when they reported high levels of stress symptoms, they received visits from nurses to help them deal with their life problems. Nursing interventions involved an individually tailored combination of teaching, support and consultation or referral strategies. Visits continued until the patient’s stress score decreased to a normal level and until problems were resolved. The average highly stressed patient required only five to six hours over a 6-month period. Control patients received routine medical care after discharge. For patients receiving routine care after discharge, highly distressed patients exhibited a near threefold increase in risk of cardiac mortality over 5 years ( p = 0.0003). Those highly stressed patients who took part in the 1-year program of stress monitoring and intervention, however, did not experience any significant long-term increase in risk. The program’s impact was significant in terms of reduction of both cardiac mortality (p=0.006) and MI recurrences ( p = 0.004) among highly distressed

patients. There was little evidence of impact among patients with low levels of stress in the hospital. Of note is that the average treated patient averaged only one to two more contacts with physicians during the program year than did the average control patient. Therefore improved medical care was not the major factor responsible for the program’s impact nor was there significant reduction in risk factors for CAD by patients even in the treatment group. A recent study randomly assigned 2328 post-MI patients to a control group or a rehabilitation program consisting of seven weekly two-hour sessions including relaxation training, cardiac education, group discussion, and individual counseling. These investigators observed no significant improvement in anxiety and depressive symptoms at 6 months or cardiovascular morbidity or mortality at 12 months post-MI in the rehabilitation program patients assigned compared to the control group (Jones and West, 1996). This ‘lack of effect’ was likely due to the relatively non-specific nature of the treatment program as well as the fact that patients in the ‘control’ group likely received some cardiac education and behavioral counseling during outpatient clinic visits. More recent studies report no benefit to patients (Taylor et al., 1997) or an even worse outcome for women in comparison to usual care (Frasure-Smith, 1997). In summary, the psychological interventions with post-MI patients (Frasure-Smith and Prince, 1984; Friedman et al., 1986; Frasure-Smith, 1991; Frasure-Smith et al., 1992) specifically targeting individuals suffering emotional consequences (designated as ‘distress’) of a perceived threat (myocardial infarction) (Frasure-Smith, 1991) or altering a particular personality type (Friedman et al., 1986) have reported a reduction in risk of recurrent cardiovascular events (Friedman et al., 1986) and an increase in long-term (5-year) survival rates (Frame-Smith and Prince, 1985; Frasure-Smith 1991; Frasure-Smith et al., 1992). Such behavioral interventions have been hypothesized to act in part by reducing sympathetic arousal and silent ischemia (Frasure-Smith, 1991). The efficacy of psychotherapeutic treatment of post-MI patients with comorbid major depression is unknown.

52

Somatic treatment of major depression in patients with CVD reduces dysphoria. Whether psychopharmacologic treatment of major depression per se has a positive impact on post-MI survival remains a tantalizing, and unanswered, question. Avery and Winokur (1976) observed 519 hospitalized depressed patients over a three-year follow-up period with particular attention to causes of mortality, i.e. suicide, MI, cerebrovascular accident, cancer, or other causes (pneumonia, cirrhosis, car accident). Depressed patients adequately treated with either tricyclic antidepressants or ECT exhibited a lower mortality from heart disease than inadequately treated depressed patients. Indeed a study conducted in Austria, Denmark, Canada, and Germany reported that lithium treatment for more than two years reduced mortality from CAD in 827 patients with major affective disorder, including those with major depression, mania and schizoaffective disorder (Ahrens et al., 1995). Because of fewer potential side effects upon the cardiovascular system and the lack of lethality in overdose, somatic treatment with selective 5HT reuptake inhibitors (SSRIs) or other ‘atypical’ antidepressants (such as buproprion or nefazodone) may offer significant advantages in depressed patients with CVD. The cardiac toxicity of tricyclic and related cyclic antidepressants limit their clinical use in patients with cardiovascular disease. The reader is directed toward excellent reviews on the safety and efficacy of tricyclic treatment of patients with cardiac disease (Muskin and Glassman, 1983; Roose and Dalack, 1992). Monoamine oxidase inhibitors (MAOIs) and trazodone are generally free of effects on cardiac conduction, but, like the tricyclic and related antidepressants, may cause postural hypotension (Arana and Hyman, 1991). With the introduction of selective 5HT reuptake inhibitors (SSRIs), more medically ill patients may be treated without fear of complicating cardiovascular side effects. Because of fewer potential side effects and the lack of lethality in overdose, these agents offer significant advantages in depressed patients with cardiovascular disease. The only known cardiac effect of SSRIs is severe sinus node slowing, to date reported in only a few cases (Ellison et al., 1990; Enemark, 1993). After short-

term treatment with paroxetine or fluvoxamine, depressed patients exhibited no changes in HRV (Rechlin, 1994). Because these agents are newer than TCAs, little systematic research on SSRI efficacy in the elderly or in patients with CAD has been performed, including large-scale, randomized, treatment trials of post-MI patients with comorbid major depression (Roose et al., 1991, 1994). A recent randomized, double-blind multicenter study compared the efficacy of nortriptyline and paroxetine in depressed patients with ischemic heart disease (Roose et al., 1998). Both antidepressants were effective in the treatment of depression but not surprisingly, there were more dropouts due to side effects and more adverse cardiac effects with the TCA. There have been some reports of alterations of hemostasis (Humphries et al., 1990; Evans et al., 1991; Yaryura-Tobias et al., 1991) and platelet aggregation (Alderman et al., 1992) following treatment with fluoxetine. Because of inhibition of cytochrome P450 isoenymes, SSRIs must be used with caution particularly in those patients receiving medications metabolized by the P450 2D6 isoenzyme (e.g. lipophilic P-blockers, type 1C antiarrhythmics: encainide, flecainide, mexiletene, propafenone) and the P450 3A4 isoenzyme (e.g. calcium channel blockers and warfarin) (Callahan et al., 1996). An ongoing multicenter study sponsored by Pfizer, SADHART, a randomized, double-blind trial of sertraline vs. placebo in the treatment of post-MI patients with comorbid major depression seeks to determine the efficacy of this antidepressant in the depressed post-MI patient. Although these agents may be as effective as TCAs in depressed patients with CVD, their safety is not yet well established in this patient population. Psychotherapeutic and/or psychopharmacologic treatment of the 15 to 23% of post-MI patients who fulfill criteria for major depression, as well as those with significant dysphoria but who have subsyndromal depression, may have a significant impact on both medical morbidity and mortality. Although the belief that the outcomes of CVD patients may be improved if their comorbid depressive symptoms are treated is tantalizing, due to advances in medical management of CVD patients,

53

therapeutic trials determining improvement in survival must be quite large (Frasure-Smith et al., 1997),e.g. the 22-month CAST trial was comprised of 1489 subjects (Echt et al., 1991). Such past experience cautions against the raising of hopes of demonstrating improving cardiac outcome via antidepressant treatment of depression in patients with CVD. However, awaiting the completion of a largescale mortality trial similar to the CAST may not be appropriate given the interpersonal, social, and medical burden of depression and early indications of SSRI efficacy in depressed CVD patients.

Comment Over the last 30 years, numerous prospective studies have determined that depression is a major risk factor in both the development of CVD, and in death after an index myocardial infarction. Meanwhile psychiatric research on the neurochemical, neuroendocrine, and neuroanatomic alterations in unipolar depression (and the response of certain of these markers to somatic treatments) has led to the discovery of so-called biological markers, measures that presumably reflect the underlying pathophysiologic processes of this ‘mental’ disorder. A number of such pathophysiologic perturbations may play a mechanistic role in the elevated morbidity and mortality of patients with CVD and comorbid depression including sympathoadrenal system hyperactivity, alterations in serotonergic platelet system, and exaggerated platelet reactivity. Illumination of the interplay between CNS, platelet, and cardiovascular processes, particularly in those patients with CVD and major depression, will undoubtedly lead to the development of new treatment modalities that will not only improve these patients’ quality of life but potentially decrease their morbidity and improve long-term survival rates.

Acknowledgements The authors greatly acknowledge the assistance of Laurence A. Harker, MD, Anne Griswell, Bettina T. Knight, BSN, Ulla Marzec, Paul Robinson, MD, and Gail Uwaifo. They were supported by NIH

Grants MH-01399, MH49523, RR-00039, and an Established Investigator Award from the National Alliance for Research on Schizophrenia and Depression (NARSAD) to Charles B. Nemeroff, MD, PhD.

References Ahern, D.K., Gorkin, L., Anderson, J.L., Tiemey, C., Hallstrom, A., Ewart, C., Capone, R.J., Schron, E., Kornfeld, D., Herd, J.A., Richardson, D.W. and Follick, M.J. (1990) Biobehavioral variables and mortality or cardiac arrest in the Cardiac Arrhythmia Pilot Study (CAPS). Am. J. Cardiol., 66: 59-62. Ahrens, B., Muller-Oerlinghausen, B., Schou, M., Wolf, T., Ada, M., Grof, E., Grof, P., Lenz, G., Simhandl, C., Thau, K., Vestergaard, P., Wolf, R. and Moller, H.J. (1995) Excess cardiovascular and suicide mortality of affective disorders may be reduced by lithium prophylaxis. J. Affect. Disord., 33: 67-75. Akselrod, S., Gordon, D., Ubel, F.A., Shannon, D.C., Barger, A.C., Cohen, R.J. (1981) Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat cardiovascular control. Science, 213: 220-222. Alderman, C.P., Moritz, C.K. and Ben-Tovim, D.I. (1992) Abnormal platelet aggregation associated with fluoxetine therapy. Ann. Pharmucothex, 26: 1517-1519. Allison, T.G., Williams, D.E., Miller, T.D., Patten, C.A., Bailey, K.R., Squires, R.W. and Gau, G.T. (1995) Medical and economic costs of psychologic distress in patients with coronary artery disease. Mayo Clin. Pmc., 70: 734-742. American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders, 3rd edn, American Psychiatric Association, Washington, DC. American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, 4th edn, American Psychiatric Association, Washington, DC, pp. 3 17-327. Anda, R., Williamson, D., Jones, D., Macera, C., Eaker, E., Glassman, A. and Marks, J. (1993) Depressed effect, hopelessness, and the risk of ischemic heart disease in a cohort of U.S. adults. Epidemiology, 4: 285-294. Anfossi, G. and Trovati, M. (1996) Role of catecholamines in platelet function: pathophysiological and clinical significance. Eux J. Clin. Invest., 26: 353-370. Antiplatelet Trialists’ Collaboration (1988) Secondary prevention of vascular disease by prolonged antiplatelet treatment. Bx Med. J. (Clin. Res. Ed.), 296: 320-31. Arana, G.W. and Hyman, S.E. (1991) Handbook of Psychiatric Drug Therapy, 2nd edn, Little, Brown and Company, Boston, Massachusetts, p. 61. Arato, M., Banki, C.M., Nemeroff, C.B. and Bissette, G . (1986) Hypothalamic-pituitaq-adrenalaxis and suicide. Ann. N.!l Acad. Sci., 487: 263-270. Araya-Gomez, V., Gonzalez-Hermosillo, J., Casanova-Garces, J., Colin, L., Kershenovich, S . and Iturralde, P. (1994) Identification of patients at risk of malignant arrythmia in the

54

1st year after myocardial infarction (in Spanish). Arch. Inst. Cardiol. Ma.,64: 145-159. Aromaa, A,, Raitasalo, R., Reunanen, A., Impivaara, 0.. Heliovaara, M., Kneckt, P., Lehtinen, V., Joukamaa, M. and Maatela, J. (1994) Depression and cardiovascular diseases. Acta Psychiatry Scand., 377: 77-82. Ashton, J.H., Benedict, C.R., Fitzgerald, C., Raheja, S., Taylor, A.L., Campbell, W.B., Buja, L.M. and Willerson, J.T. (1986) Serotonin as a mediator of cyclic flow variations in stenosed canine coronary arteries. Circulation, 73: 572-578. Ashton, J.H., Ogletree, M.L.. Michel, I.M., Golino, P., McNatt, J.M., Taylor, A.L., Raheja, S., Schmitz, J., Buja, L.M., Campell, W.B. and Willerson, J.T. (1987) Cooperative mediation by serotonin S2 and thromboxane A2Iprostaglandin H2 receptor activation of cyclic flow variation in dogs with severe coronary artery stenosis. Circulation, 76: 952-959. Avery, D. and Winokur, G. (1976) Mortality in depressed patients treated with electroconvulsive therapy and antidepressants. Arch. Gen. Psychiatry, 33: 1029-1037. Balogh, S., Fitzpatrick, D.F., Hendricks, S.E. and Paige, S.R. (1993) Increases in heart rate variability with successful treatment in patients with major depressive disorder. Psychophannacol. Bull., 29: 201-206. Banki, C.M., Bissette, G., Arato, M., O’Connor, L. and Nemeroff, C.B. (1987) Cerebrospinal fluid corticotropinreleasing factor-like immunoreactivity in depression and schizophrenia. Am. J. Psychiatry, 144: 873-877. Banki, C.M., Karmasci, L., Bissette, G. and Nemeroff, C.B. (1992) CSF corticotropin-releasing and somatostatin in major depression: Rresponse to antidepressant treatment and relapse. Eu,: Neuropsychopharmacol., 2: 107-1 13. Barefoot, J.C. and Schroll, M. (1996) Symptoms of depression, acute myocardial infarction, and total mortality in a community sample. Circulation, 93: 1976-1980. Barrett, J.E., Barrett, J.A., Oxman, T.E. and Gerber, P.D. (1988) The prevalence of psychiatric disorders in a primary care practice. Arch. Gen. Psychiatry, 45: 1100-1 106. Bigger, T.J. Jr., Fleiss, J.L., Steinman, R.C., Rolnitzky, L.M., Kleiger, R.E. and Rottman, J.N. (1992) Frequency domain measures of heart period variability and mortality after myocardial infarction. Circulation, 85: 164-171, Bigger, J.T., Kleiger, R.E., Fleiss. J.L., Rolnitzky, L.M., Steinman, R.C. and Miller, J.P. (1988) Components of HR variability measured during healing of acute myocardial infarction. Am. J. Cardiol., 61: 208-215. Billman, G.E., Schwartz, P.J. and Stone, H.L. (1982) Baroreceptor reflex control of heart rate: a predictor of sudden cardiac death. Circulation, 66: 814-880. Bjorkerud, S. (1973) Effect of adrenocortical hormones on the integrity of rat aortic endothelium. In: G. Schettler and A. Weizel (Eds), Proceedings of the 3rd International Symposium on Atherosclerosis III., Springer-Verlag, Berlin, p. 245. Blacker, C.V.R. and Clare, A.W. (1987) Depressive disorder in primary care. B,: J. Psychiatry. 150: 737-751. Blumenthal, J.A., Williams, R.S., Wallace, A.G., Williams, R.B. and, Needles, T.L. (1982) Physiological and psychological

variables predict compliance to prescribed exercise therapy in patients recovering from myocardial infarction. Psychosom. Med., 44: 519-527. Booth-Kewley, S. and Friedman, H.S. (1987) Psychological predictors of heart disease: a quantitative review. Psych. Bull., 101: 343-362. Brozek, J., Keyes, A. and Blackburn, H. (1966) Personality differences between potential coronary and noncoronary subjects. Ann. N.E Acad. Sci., 134: 1057-1064. Callahan, A.M., Marangell, L.B. and Ketter, T.A. (1996) Evaluating the clinical significance of drug interactions: a systematic approach. Harvard Rev. Psychiatry, 4: 153-158. Campbell, R.W.F. (1996) Can analysis of heart rate variability predict arrhythmias and antimhythmic effects? In: A.M. Oto (Ed.), Practice and Progress in Cardiac Pacing and Electrophysiology, Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 63-69. Carney, R.M., Rich, M.W., Freedland, K.E. and Saini, J. (1988) Major depressive disorder predicts cardiac events in patients with coronary artery disease. Psychosom. Med., 50: 627-633. Carney, R.M., Rich, M.W., teVelde, A., Saini, J., Clark, K. and Jaffe, A.S. (1987) Major depressive disorder in coronary artery disease. Am. J. Cardiol., 60: 1273-1275. Carney, R.M., Saunders, R.D., Freedland, K.E., Stein, P., Rich, M.W. and Jaffe, A.S. (1995) Association of depression with reduced heart rate variability in coronary artery disease. Am. J. Cardiol., 76: 562-564. Case, R.B., Moss, A,J., Case, N., McDermott, M. and Eberly, S. (1992) Living alone after myocardial infarction: impact on prognosis. JAMA, 261: 515-519. Cattell, R.B., Saunders, D.R. and Stice, G. (1957) Handbook for the Sixteen Personality Factor Questionnaire, Institute for Personality and Ability Testing, Champaign, Illinois. Cay, E.L., Vetter, N., Philip, A.E. and Dugard, P. (1972) Psychological status during recovery from an acute heart attack. J. Psychosom. Rex, 16: 425435. Charney, D.S., Menkes, D.B. and Henninger, G.R. (1981) Receptor sensitivity and the mechanism of action of antidepressant treatment. Arch. Gen. Psychiatry, 38: 1160-1180. Cohen-Cole, S.A. and Kaufman, K.G. (1993) Major depression in physical illness: diagnosis, prevalence, and antidepressant treatment (a ten year review: 1982-1992). Depression, 1: 181-204. Cripps, T., Malik, M., Farrell, T. and Camm, A.J. (1991) Prognostic value of reduced heart rate variability after myocardial infarction: clinical evaluation of a new analysis method. BE Heart J., 65: 14-19. Cummings, S., Elde, R., Ellis, J. and Lindall, A. (1983) Corticotropin-releasing factor immunoreactivity is widely distributed with the central nervous system of the rat: an immunohistochemical study. J. Neurosci., 8: 1355-1368. Dahlstrom, W.G. and Welsh, G.S. (1960) A MMPI Handbook: A Guide to Use in Clinical Practice and Research, Univ. of Minnesota Press, Minneapolis, Minnesota.

55

Dalack, G.W. and Roose, S.P. (1990) Perspectives on the relationship between cardiovascular disease and affective disorder. J. Clin. Psychiatry, 51 (Suppl. 7): 4-9, discussion, 10-11. DeClerck, F. (1991) Effects of serotonin on platelets and blood vessels. J. Cardiovasc. Pharmacol., 17 (Suppl. 5): S 1 - S . Dupuy, H.J. (1977) A concurrent validational study of the NCHS General Well-Being Schedule, Vital and Health Statistics, Series 2, No. 73. DHEW Pub. No. (HRS) 78-1347. US Government Printing Office, Washington, DC. Echt, D.S., Liebson, P.R., Mitchell, L.B., Peters, R.W., ObiasManno, D., Barker, A.H., Arensberg, D., Baker, A., Friedman, L., Greene, H.L., Huther, M.L., Richardson, D.W., and the CAST Investigators (1991) Mortality and morbidity in patients receiving ecainide, flecenaide or placebo: The Cardiac Arrythmia Suppression Trial. N. Engl. J. Med., 324: 781-788. Eckert. A., Gann, H., Riemann, D., Aldenhoff, J. and Muller, W.E. (1993) Elevated intracellular calcium levels after 5-HT2 receptor stimulation in platelets of depressed patients. Biol. Psychiatry, 34: 565-568. Ellison, J.M., Milofsky, J.E. and Ely, E. (1990) Fluoxetineinduced bradycardia and syncope in two patients. J. Clin. Psychiatry, 51: 385-386. Endicott, J. and Spitzer, R.L. (1978) A diagnostic interview: The Schedule for Affective Disorders and Schizophrenia. Arch. Gen. Psychiatry, 35: 837-844. Enemark, B. (1993) The importance of ECG monitoring in antidepressant treatment. Nord. J. Psychiatry, 47 (Suppl. 30): 57-65. Evans, T.G., Buys, S.S. and Rodgers, G.M. (1991) (letter to the editor). N. Engl. J. Med., 324: 1671. Everson, S.A., Goldberg, D.E., Kaplan, G.A., Cohen, R.D., Pukkala, E., Tuomilehto, J. and Salonen, J.T. (1996) Hopelessness and risk of mortality and incidence of myocardial infarction and cancer. Psychosom. Med., 58: 113-121. Feldman, E., Mayou, R., Hawton, K., Ardern, M. and Smith, E.B. (1987) Psychiatric disorder in medical inpatients. QJM, 63: 405-412. Follick, M.J., Gorkin, L., Capone, R.J., Smith, T.W., Ahern, D.K., Stablein, D., Niaura, R. and Visco, J. (1988) Psychological distress as a predictor of ventricular arrhythmias in a post-myocardial infarction population. Am. Heart J., 116: 32-36. Ford, D.E., Mead, L.A., Chang, P.P., Levine, D.M. and Klag, M.J. (1994) Depression predicts cardiovascular disease in men: the Precursors Study (Abstract). Circulation, 90 (Suppl. I): 614. Fouad, EM., Tarazzi, R.C., Gemario, C.M., Fighaly, S. and Alicandri, C. (1984) Assessment of parasympathetic control of heart rate by a non-invasive method. Am. J. Physiol., 246: H838-842. France, R.D., Urban, B., Krishnan, K.R.R., Bissett, G., Csaba, M.B., Nemeroff C. and Speilman, F.J. (1988) CSF Corticotropin-releasing factor-like immunoreactivity in chronic pain patients with and without major depression. Biol. Psychiatry, 23: 86-88.

Frasure-Smith, N. (1991) In-hospital symptoms of psychological stress as predictors of long-term outcome after acute myocardial infarction in men. Am. J. Cardiol., 67: 121-127. Frasure-Smith, N., Lesperance, F. and Juneau, M. (1992) Differential long-term impact of in-hospital symptoms of psychological stress after non-Q-wave and Q-wave myocardial infarction. Am. J. Cardiol., 69: 1128-1 134. Frasure-Smith, N., Lesperance, F., Prince, R.H., Verrier P., Garber, R.A., Juneau, M., Wolfson, C. and Bourassa, M.G. (1997) Randomised trial of home-based psychosocial nursing intervention for patients recovering from myocardial infarction. Lancet, 350: 473479. Frasure-Smith, N., Lesperance, F. and Talajic, M. (1993) Depression following myocardial infarction: impact on 6-month survival. JAMA, 270: 1819-1 86 1. Frasure-Smith, N., Lesperance, F., Talajic, M. (1995) Depression and 18-month prognosis after myocardial infarction. Circulation, 91: 999-1005. Frasure-Smith, N. and Prince, R. (1985) The ischemic heart disease life stress monitoring program: impact on mortality. Psychosom. Med., 47: 4 3 1 4 5 . Friedman, M. and Rosenman, R. H. (1959) Association of specific overt behavior pattern with blood and cardiovascular findings. JAMA, 169: 1286-1296. Friedman, M., Thoresen, C.E., Gill, J.J., Ulmer, D., Powell, L.H., Price, V.A., Brown, B., Thompson, L., Rabin, D.D., Breall, W.S., Bourg, E., Levy, R. and Dixon, T. (1986) Alteration of type A behavior and its effect on cardiac recurrences in post myocardial infarction patients: summary results of the recurrent coronary prevention project. Am. HeartJ., 112: 653-665. Gerritsen, M.E. (1996) Physiological and pathophysiological roles of eicosanoids in the microcirculation. Cardiovasc. Res., 32: 720-132. Goldberg, E.L., Comstock, G.W. and Hornstra, R.K. (1979) Depressed mood and subsequent physical illness. Am. J. Psychiatry, 136: 530-534. Golden, R.N., Markey, S.P., Risby, E.D., Rudorfer, M.V., Cowdry, R.W. and Potter, W.Z. (1988) Antidepressants reduce whole-body norepinephrine turnover while enhancing 6-hydroxymelatonin output. Arch. Gen. Psychiatry, 45: 150-154. Gonzalez, M.B., Snyderman, T.B., Colket, J.T., Arias, R.M., Jiang, J.W., O’Connor, C.M. and Krishnan, K.R. (1996) Depression in patients with coronary artery disease. Depression, 4: 57-62. Greene, R.L. (1991) The MMP1-2MMPI: An Interpretive Manual, Allyn and Bacon, Boston, Massachussetts. Hackett, T.P., Cassem, N.H. and Wishnie, H.A. (1968) The coronary-care unit. An appraisal of its psychological hazards. N.Engl. J. Med., 279: 1365-1370. Hathaway, S.R. and McKinley, J.C. (1951) Minnesota Multiphasic Personalily Inventory Manual (revised edition), The Psychological Corporation, New York. Hayward, C. (1995) Psychiatric illness and cardiovascular disease risk. Epidemiologic Rev., 17: 129-138.

56 Hess, H., Mietaschk, A. and Deichsel, G. (1985) Drug-induced inhibition of platelet function delays progression of peripheral occlusive arterial disease. A prospective double-blind arteriographically controlled trial. Lancet, l(8426): 415419. Hoeper, E.W., Nycz, G.R. and Cleary, P.D. (1979) Estimated prevalence of RDC mental disorders in primary care. In,. J. Ment. Health, 8: 6-15. Humphries, J.E., Wheby, M.S. and VandenBerg, S.R. (1990) Fluoxetine and the bleeding time. Arch. Pathol. Lab. Med., 114: 727-728. Jiang, W., Babyak, M., Krantz, D.S., Waugh, R.A., Coleman, R.E., Hanson, M.M., Frid, D.J., McNulty, S., Moms, J.J., O’Connor, C.M. and Blumenthal, J.A. (1996) Mental stressinduced myocardial ischemia and cardiac events. JAMA, 21 : 1651-1656. Jones, D.A. and West, R.R. (1996) Psychological rehabilitation after myocardial infarction: multicentre randomised controlled trial. BMJ, 313: 1517-1521. Katon, W. and Sullivan M.D. (1990) Depression and chronic medical illness. J. Clin. Psychiatry, 5 1 (Suppl. 6): 3-1 1. Katona, P.G. and Jih, F. (1975) Respiratory sinus arrhythmia: Non-invasive measure of the parasympathetic cardiac control. J. Appl. Physiol., 39: 801-805. Kemper, J.W., Baggenstoss, A.H. and Slocumb, C.H. (1957) The relationship of therapy with cortisone to the incidence of vascular lesions in rheumatoid arthritis. Ann. Intern. Med., 46: 831-851, Kessler, R.C., McGonagle, K.A., Zhao, S., Nelson, C. B., Hughes, M., Eschleman, S., Wittchen, H.-A. and Kendler, K.S. (1994) Lifetime and 12-month prevalence of DSM-111-R psychiatric disorders in the United States. Arch. Gen. Psychiatry, 51: 8-19. Kleiger, R.E., Miller, P.J., Bigger, T.J., Moss, A.J. and the Multicenter Post-Infarction Research Group ( 1987) Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. Am. J. Cardiol., 59: 256-262. Koizumi, K., Terui, N. and Kollai, M. (1 985) Effect of cardiac vagal and sympathetic nerve activity on heart rate in rhythmic fluctuations. J. Auton. Nen! Syst., 12: 251-259. Kurosawa, H., Shimizu, Y., Nishimatsu, Y., Hirose, S. and Takano, T. (1983) The relationship between mental disorders and physical severities in patients with acute myocardial infarction. Jpn. Circ. J., 47: 723-125. Kusurni, I., Koyama, T. and Yarnashita, I. (1991) Serotoninstimulated Ca2+ response is increased in the blood platelets of depressed patients. Bid. Psychiatry, 30: 3 10-3 12. Laghrissi-Thode, F., Wagner, W.R., Pollock, B.G., Johnson, P.C. and Finkel, M.S. (1997) Elevated platelet factor 4 and pthromboglobulin plasma levels in depressed patients with ischemic heart disease. Biol. Psychiatry, 42: 290-295. LaRovere, M.T., Specchia, G., Mortana, A. and Schwartz, P.J. (1988) Baroreflex sensitivity, clinical correlates, and cardiovascular mortality among patients with a first myocardial infarction: a prospective study. Circulation, 78: 8 16-824.

Lechin, F., van der Dijs, B., Orozco ,B., Lechin, M.E., Baez, S., Lechin, A.E., Rada, I., Acosta, E., Arocha, L., Jimenez, V., Leon, G. and Garcia, Z. (1995) Plasma neurotransmitters, blood pressure, and heart rate during supine-resting, orthostasis, and moderate exercise conditions in major depressed patients. Biol. Psychiatry, 38: 166-173. Leeper, J., Badger, L. and Milo, T. (1985) Mental disorders among physical disability determination patients. Am. J. Public Health, 75: 78-79. Lefkovits, J., Plow, E.F. and Topol, E.J. (1995) Platelet glycoprotein IIb/IIIa receptors in cardiovascular medicine. N. Engl. J. Med., 332: 1553-1559. Linnoila, M., Guthrie, S., Lane, E.A., Karoum, F.,Rudorfer, M. and Potter, W.Z. (1986) Clinical studies on NE metabolism: How to interpret the numbers. Psychiatry Res., 17: 229-239. Linnoila, M., Karoum, F.,Calil, H.M., Kopin, I.J. and Potter, W.Z. (1982) Alteration of NE metabolism with desipramine and zimelidine in depressed patients. Arch. Gen. Psychiarry, 39: 1025-1028. Louis, W.J., Doyle, A.E. and Anavekar, S.N. (1975) Plasma noradrenaline concentration and blood pressure in essential hypertension, phaeochromocytoma and depression. Clin. SOC.,48: 239s-2428. Low, P.A. (1993) Autonomic nervous system function. J. Clin. Neurophysiol., 10: 14-27. Lown, B., DeSilva, R.A., Reich, P. and Murawski, B.J. (1980) Psychophysiological factors in sudden cardiac death. Am. J. Psychiatry, 137: 1325-1335. Lown, B. and Venier, R.L. (1976) Neural activity and ventricular fibrillation. N. Engl. J. Med., 294: 1 165-1 170. Magni, G., Schifano, F. and DeLeo, D. (1986) Assessment of depression in an elderly medical population. J. Affect. Disord., 11: 121-124. Malzberg, B. (1937) Mortality among patients with involution melancholia. Am. J. Psychiatry, 93: 1231-1238. Markovitz, J.H. and Matthews, K.A. (1991) Platelets and coronary heart disease: potential psychophysiological mechanism. Psychosom. Med., 53: 643-668. Mayou, R., Foster, A. and Williamson, B. (1979) Medical care after myocardial infarction. J. Psychosom. Res., 23: 23-26. Merchenthaler, I., Vigh, S., Petruscz, P. and Schally, A.V. (1982) Immunocytochemical localization of corticotropinreleasing factor (CRF) in the rat brain. Am. J. Anatomy, 165: 385-396. Mikuni, M., Kusumi, I., Kagaya, A., Kuroda, Y., Mori, H. and Takahashi, K. (1991) Increased 5-HT2 receptor function as measured by serotonin-stimulated phosphoinositide hydrolosis in platelets of depressed patients. Prog. Neuropsychopharmacol. Biol. Psychiatry, 15: 49-6 1. Miyawaki, E. and Salzman, C. (1991) Autonomic nervous system tests in psychiatry: implications and potential uses of heart rate variability. Integrated Psychiatry, 7: 21-28. Moms, P.L.P., Robinson, R.G.,Andrzejewski, P., Samuels, J. and Price, T.R. (1993) Association of depression with 10-year poststroke mortality. Am. J. Psychiatry, 150: 124- 129.

57 Murphy, J.M., Monson, R.R., Olivier, D.C., Sobol, A.M. and Leighton, A.H. (1987) Affective disorders and mortality. Arch. Gen. Psychiatry, 44:473480. Murphy, J.M., Neff, R.K., Sobol, A.M., Rice, J.X. and Olivier, D.C. (1985) Computer diagnosis of depression and anxiety: The Stirling County Study. Psychol. Med., 15: 99-112. Muskin, P.R. and Glassman, A.H. (1983) The use of tricyclic antidepressants in a medical setting. In: J.B. Finkel (Ed.), Consultation-Liason Psychiatry: Current Trends and Future Perspectives, Gmne and Stratton, New York, pp. 137-158. McFarland, B.H., Freeborn, D.K., Mullooly, J.P. and Pope, C.R. (1985) Utilization patterns among long-term enrollees in a prepaid group practice health maintenance organization. Med. Care, 23: 1221-1233. Musselman, D.L., Evans, D.L. and Nemeroff, C.B. (1998) The relationship of depression to cardiovascular disease: epidemiology, biology, and treatment. Arch. Gen. Psych., 55: 580-592. Musselman, D.L., Knight, B.T., Baron, A. and Nemeroff, C.B. (1998) Further studies on altered platelet reactivity in major depression. Biological Psychiatry, 43: 99s. Musselman, D.L., Tomer, A., Manatunga, A.K., Knight, B.T., Porter, M.R., Kasey, S., Marzec, U., Harker, L.A. and Nemeroff, C.B. (1996) Exaggerated platelet reactivity in major depression. Am. J. Psychiarry, 153: 1313-1317. Nahas, G.G., Brunson, J.G., King, W.M. and Cavert, H.M. (1958) Functional and morphologic changes in heart lung preparations following administration of adrenal hormones. Am. J. Clin. Pathol., 34: 717-729. Nemeroff, C.B., Widerlov, E., Bissette, G., Walleus, H., Karlsson, I., Eklund, K., Kilts, C.D., Loosen, P.T. and Vale, W. (1984) Elevated concentrations of CSF corticotropinreleasing factor-like immunoreactivity in depressed patients. Science, 226: 1342-1344. Osim, E.E. and Wyllie, J.H. (1982) Evidence for loss of 5-hydroxytryptamine from circulating platelets. J. Physiol. ( b n d . ) , 326: 25P-26P. Ostfeld, A.M., Lebovits, B.Z., Shekelle, R.B. and Paul, 0. (1964) A prospective study of the relationship between personality and coronary heart disease. J. Chronic Dis., 17: 265-276. Pratt, L.A., Ford, D.E., Crum, R.M., Armenian, H.K., Gallo, J.J. and Eaton, W.W. (1996) Depression, psychotropic medication, and risk of myocardial infarction: prospective data from the Baltimore ECA follow-up. Circulation, 94: 3 123-3129. Raadsheer, F.C., Hoogendijk, W.J.G., Stam, F.C., Tilders, F.J.H. and Swaab, D.F. (1994) Increased numbers of corticotropinreleasing hormone expressing neurons in the hypothalamic paraventricular nucleus of depressed patients. Neuroendocrinology, 60: 436-444. Raadsheer, F.C., van Heerikhuize, J.J., Lucassen, P.J., Hoogendijk, W.J.G., Tilders, F.J.H. and Swaab, D.F. (1995) Corticotropin-releasing hormone mRNA levels in the paraventricular nucleus of patients with Alzheimer's disease and depression. Am. J. Psychiatry, 152: 1372-1376.

Radloff, L.S. (1977) The CES-D scale: a self-report depression scale for research in the general population. J. Appl. Psychol. Meas., 1: 385401. Ragland, D.R. and Brand, R.J. (1988) Type A behavior and mortality from coronary heart disease. N.Engl. J. Med., 3 18: 65-69. Risch, S.C., Lewine, R.J., Kalin, N.H., Jewart, R.D., Risby, E.D., Caudle, J.M., Stipetic, M., Turner, J., Eccard, M.B. and Pollard, W.E. (1992) Limbic-hypothalamic-pituitary-adrenal axis activity and ventricular-to-brain ratio studies in affective illness and schizophrenia. Neuropsychophamcology, 6~95-100. Rechlin, T., Weis, M. and Claus, D. (1994) Heart rate variability in depressed patients and differential effects of paroxetine and amitriptyline on cardiovascular autonomic functions. Pharmacopsychiatry, 27: 124-128. Regier, D.A., Boyd, J.H., Burke, J.D. Jr., Rae, D.S., Myers, J.K., Kramer, M., Robins, L.N., George, L.K.,Karno, M. and Locke, B.Z. (1988) One-month prevalence of mental disorders in the United States: Based on five Epidemiologic Catchment Area sites. Arch. Gen. Psychiatry, 45: 977-986. Ridker, P.M., Manson, J.E., Burning, J.E., Goldhaber, S.Z. and Hennekens, C.H. (1991) The effect of chronic platelet inhibition with low-dose aspirin on atherosclerotic progression and acute thrombosis: clinical evidence from the Physicians' Health Study. Am. Heart J., 122: 1588-1592. Risch, S.C., Lewine, R.J., Kalin, N.H., Jewart, R.D., Risby, E.D., Caudle, J.M., Stipetic, M., Turner, J., Eccard, M.B. and Pollard, W.E. ( 1992) Limbic-hypothalamic-pituitary-adrenmal axis activity and ventricular-to-brain ratio studies in affective illness and schizophrenia. Neuropsychopharmacology, 6: 95-100. Robins, L.N., Helzer, J.E., Croughan, J., Williams, J.B.W. and Spitzer, R.L. (1981) National Institute of Mental Health Diagnostic Interview Schedule: its history, characteristics, and validity. Arch. Gen. Psychiatry, 38: 381-389. Roose, S.P. and Dalack, G.W. (1992) Treating the depressed patient with cardiovascular problems. J. Clin. Psychiatry, 53: 25-3 1, Roose. S.P., Dalack, G.W., Glassman, A.H., Woodring, S., Walsh, T. and Giardina, G.V. (1991) Cardiovascular effects of buproprion in depressed patients with heart disease. Am. J. Psychiatry, 148: 512-516. Roose. S.P., Glassman, A.H., Attia, E. and Woodring, S. (1994) Comparative efficacy of selective serotonin reuptake inhibitors and tricyclics in the treatment of melancholia. Am. J. Psychiatry, 151: 1735-1 739. Roose, S.P., Glassman, A.H. and Dalack, G.W. (1989) Depression, heart disease, and tricyclic antidepressants. J Clin. Psychiatry, 50 (Suppl. 7): 12-16. Roose, S.P., Laghrissi-Thode, F., Kennedy, J., Nelson, C., Bigger, J.T. Jr., Pollock, B., Gafney, A., Narayan, M., Finkel, M.S., McCafferty, J. and Gergel, I. (1998) Comparison of paroxetine and norhiptyline in depressed patients with ischemic heart disease. JAMA, 279: 287-291.

58

Ross, R. and Harker, L. (1976) Hyperlipidemia and atherosclerosis. Science, 193: 1094-1 100. Roy, A., Guthrie, S., Pickar, D. and Linnoila, M. (1987) Plasma NE responses to cold challenge in depressed patients and normal controls. Psychiatry Res., 21: 161-168. Roy, A,, Pickar, D., DeJong, J., Karoum, F. and Linnoila, M. (1988) Norepinephrine and its metabolites in cerebrospinal fluid, plasma and urine: relationship to hypothalamicpituitary-adrenal axis function in depression. Arch. Gen. Psychiatry, 45: 849-857. Ruberman, W., Weinblatt, E., Goldberg, J.G. and Chaudary, B.S. (1984) Psychosocial influences on mortality after myocardial infarction. N. Eng. J. Med., 31 1: 552-559. Sayers, B.M. (1973) Analysis of heart rate variability. Ergonomics, 16: 17-32. Schleifer, S.J., Macari-Hinson. M.M., Coyle, D.A., Slater, W.R., Kahn, M., Gorlin, R. and Zucker, H.D. (1989) The nature and course of depression following myocardial infarction. Arch. Int. Med., 149: 1785-1789. Schulberg H.C., Saul, M. and McClelland M. (1985) Assessing depression in primary medical and psychiatric practices. Arch. Gen. Psychiatry, 42: 1164-1 170. Scubee-Moreau, J.J. and Dresse, A.E. (1979) Effect of various antidepressant drugs on the spontaneous firing rate of locus coeruleus and raphe dorsalis neurons of the rat. Eu,: J. Phamacol., 57: 219-25. Selye, H. (1956) The Stress of Life, McGraw-Hill, New York, NY. Shields, R.W. (1993) Functional anatomy of the autonomic nervous system. J. Clin. Neurophysiol., 10: 2-13. Simonsick, E.M., Wallace, R.B., Blazer, D.G. and Berkman, L.F. (1995) Depressive symptomatology and hypertensionassociated morbidity and mortality in older adults. Psychosom. Med., 57: 427-435. Spyer, K.M. (1988) Central nervous system control of the cardiovascular system. In: R. Bannister (Ed.), Autonomic Failure: A Textbook of Clinical Disorders of the Autonomic Nervous System, Oxford University Press, Oxford, England, pp. 5 6 7 9 . Stem, J.J., Pascale, L. and Ackerman, A. (1977) Life adjustment post myocardial infarction: determine predictive variables. Arch. Intern. Med., 137: 1680-1685. Sulser, F., Vetulani, J. and Mobley, P.L. (1978) Mode of action on antidepressant drugs. Biochem. Pharmacol., 27: 257-261. Swanson, L.W. and Sawchenko, P.E. (1983) Organization of ovine corticotropin-releasing factor immunoreactive cells and fibers in the rat brain: an immunohistochemical study. Neuroendocrinology, 36: 165-186. Tavazzi, L., Zotti, A.M. and Rondanelli, R. (1986) The role of psychologic stress in the genesis of lethal arrhythmias in patients with coronary artery disease. E m Heart J., 7 (Suppl. A): 99-106. Taylor, C.B., Miller, N.H., Smith, P.M. and DeBusk, R.F. (1997) The effect of a home-based, case-managed, multifactorial risk-reduction program on reducing psychological

distress in patients with cardiovascular disease. J. Cardiopulm. Rehabil., 17: 157-162. Troxler, R.G., Sprague, E.A., Albanese, R.A., Fuchs, R. and Thompson, A.J. (1977) The association of elevated plasma cortisol and early atherosclerosis as demonstrated by coronary angiography. Atherosclerosis, 26: 15 1-162. Valigorsky, J.M. (1969) Metaplastic transformation of aortic smooth muscle cells in cortisone-induced dissecting aneurysms in hamsters. Fed. Pmc., 28: 802. Veith, R.C., Lewis, L., Linares, O.A., Barnes, R.F., Raskin, M.A., Villacres, E.C., Murburg, M.M., Ashleigh, E.A., Castillo, S., Peskind, E.R., Pascualy, M. and Halter, J.B. (1994) Sympathetic nervous system activity in major depression: basal and desipramine-induced alterations in plasma NE kinetics. Arch. Gen. Psychiatry, 51: 411-242. Verrier, R.L. (1990) Behavioral stress, myocardial ischemia, and arrhthmias. In D.P. Zipes and J. Jalife (Eds), Cardiac Electmphysiology from Cell to Bedside, W.B. Saunders, Toronto, Canada, pp. 343-352. Verstraete, M. (1991) Risk factors, interventions and therapeutic agents in the prevention of artherosclerosis-related ischaemic diseases. Drugs,42 (Suppl. 5): 22-38. Vingerhoets, A. (1 985) Psychosocial Stress: An Experimental Approach. Groningen, the Netherlands: Swets & Zeitlinger. Viskin, S. and Belhassen, B. (1992) Non-invasive and invasive strategies for the prevention of sudden death after myocardial infarction. Value, limitations and implications for therapy. Drugs, 44: 336-355. Vogt, T., Pope, C., Mullooly, J. and Hollis, J. (1994) Mental heath status as a predictor of morbidity and mortality: a 15-year follow-up of members of a health maintenance organization. Am. J. Public Health, 84: 227-23 1. Von Korff, M., Shapiro, S., Burke, J.D., Teitlebaum, M., Skinner, E.A., German, P.,Turner, R.W., Klein, L. and Burns, B. (1987) Anxiety and depression in a primary care clinic: Comparison of Diagnostic Interview Schedule, General Health Questionnaire, and practitioner assessments. Arch. Gen. Psychiatry, 44: 152-156. Ware, J.A., Smith, M. and Salzman, E.W. (1987) Synergism of platelet aggregating agents. Role of elevation of cytoplasmic calcium. J. Clin. Invest., 80: 267-27 I . Wassertheil-Smoller, S., Applegate, W.B., Berge, K., Chang, C.J., Davis, B.R., G r i m , R., Kostis, J., Pressel, S., Schron, E. (1996) Change in depression as a precursor of cardiovascular events. Arch. Intern. Med., 156: 553-561. Weyrich, A.S., Solis, G.A., Li, K.S.,Tulenko, T.N. and Santarnore, W.P. (1992) Platelet amplification of vasospasm. Am. J. Physiol., 263(2 Pt 2): H349-358. Wiklund, I., Oden, A., Sanne, H., Ulvenstam, G., Wilhelmsson, C. and Wilhelmsen, L. (1988) Prognostic importance of somatic and psychosocial variables after a first myocardial infarction. Am. J. Epidemiol., 128: 786795. Williams, R.B., Barefoot, J.C., Calift, R.M., Haney, T.L., Saunders, W.B., Pryor, D.B., Hlarnky, M.A., Siegler, I.C.and Mark, D.B. (1992) Prognostic importance of social and economic resources among medically tested patients with

59 angiographically documented coronary artery disease. JAMA, 267: 520-524. Wing, J.K., Cooper, J.E. and Sartorius, N. (1974) The measurement and class$cation of psychiatric symptoms, Cambridge University Press, London. Wolf, M., Varigos, G., Hunt, D. and Sloman, J.G. (1978) Sinus arrhythmia in acute myocardial infarction. Med. J. Aust., 2: 52-53. Wyatt, R.J., Portnoy, B., Kupfer, D.J., Snyder, F. and Engelman, K. (1971) Resting plasma catecholamine concentrations in patients with depression and anxiety. Arch. Gen. Psychiatry, 2 4 65-70.

Wynn, A. (1967) Unwarranted emotion distress in men with ischaemic heart disease. Med. J. Aust., 2: 847-85 1. Yaryura-Tobias, J.A., Kirschen, H., Ninan, P. and Mosberg, H.F. (1991) Fluoxetine and bleeding in obsessive-compulsive disorder (letter). Am. J. Psychiatry, 148: 949. Yeragani, V.K., Pohl, R., Rarnesh, C., Glitz, D., Weinberg, P. and Merlos, B. (1992) Effect of imipramine treatment on heart rate variability measures. Neumpsychobiology, 26: 27-32. Zaza, A. and Schwartz, P.J. (1985) Role of the autonomic nervous system in the genesis of early ischemic arrhythmias. J. Cardiovasc. Pharmacol., 7 (Suppl. 5): S8-Sl2.

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Brain Research, Vol 122 0 ZOO0 Elsevier Science BV. All rights reserved.

CHAPTER 6

Circuits and mechanisms governing hypothalamic responses to stress: a tale of two paradigms P.E. Sawchenko", H.-Y. Li and A. Ericsson Laboratory of Neuronal Structure and Function, The Salk Institute for Biological Studies, and The Foundationfor Medical Research, La Jolla. CA 92037. USA

Introduction Significant threats to homeostasis, be they real or perceived, and whether posed by events in the outside world or from within, evoke adaptive responses that serve to defend the stability of the internal environment. This commonly includes activation of (1) the neuroendocrine cascade of events whereby the pituitary, under hypothalamic control, releases ACTH into the general circulation to stimulate production and release of glucocorticoids from the adrenal cortex; and (2) the more purely neuronal mechanisms that mediate the release of catecholamines into plasma and sympathetically innervated tissues in the periphery. Collectively, the end products of these hypothalamo-pituitary-adrenal (HPA) and sympathoadrenal systems act and interact to mobilize and redistribute bodily resources to enable men and animals to cope with emergency situations, that is, to facilitate 'fight-or-flight' responses. So prevalent is the activation of one or both of these systems in response to diverse challenges that their appearance is commonly taken as defining stressful circumstances. Despite the fact that the HPA and sympatho-adrenal systems sit squarely at the core of bodily defense systems, it is important to recognize that the entire remainder of the animal's *Corresponding author. Tel.: (619) 453-4100 X1562; Fax: (619) 453-8104; E-mail: [email protected]

physiologic and behavioral repertoires are available to sculpt the overall response of the animal to meet the demands of the particular challenge at hand. Though more commonly considered as a manifestation of basic bodily housekeeping than a province of mind, a consideration of the organization of stress-related neural systems is important to the topic of this volume for several reasons. First, the capacity to mobilize neuroendocrine and autonomic defense mechanisms provides an important avenue of emotional expression. Accordingly, brain systems governing the stress response are extensively interconnected with the limbic region of the telencephalon, which clearly provides an important mind-body interface in its capacity to evaluate the emotional significance of environmental events, and to orchestrate appropriate response patterns to cope with them (Le Doux, 1987). Second, the mediators of the neural and, particularly, the endocrine arms of the stress response have widespread and profound influences on bodily functions that extend to brain systems that are intimately involved in higher cognitive functions, such as learning and memory (see Chapter 3 by McEwen in this volume). Dysregulation of the HPA axis and sympatho-adrenal systems has been linked not only to systemic disorders, such as hypertension and autoimmune disease (Ramsey, 1982; Wick et al., 1993), but to affective ones as well, including major depression and anorexia nervosa (Owens and Nemeroff, 1993). Thus, the influence of stress

62

systems on the loftiest of mental functions extends well beyond the mere provision of an internal environment that is conducive to their emergence. Adopting the view that an understanding of the neuronal underpinnings of the stress response is a necessary prerequisite to any broader appreciation of one major manifestation of mind-body interactions, the present chapter will summarize recent progress in delineating the functional organization of brain pathways that provide for adaptive responses in animal models of stress.

Stress-related hypothalamic effector populations Any consideration of the circuits and mechanisms that provide for integrated whole-animal responses to stress must necessarily focus on the hypothalamus, and one particular cell group within it, the paraventricular nucleus (PVH). Within the PVH are contained neurosecretory neurons that express the neuropeptide, corticotropin-releasing factor (CRF),

and comprise the final common pathway through which HPA axis responses to stress are initiated, cells that project to the brainstem and spinal cord and are in a position to directly modulate central autonomic function, including sympatho-adrenal output, and additional cell types that govern multiple ancillary, or stressor-specific, neuroendocrine, autonomic and behavioral mechanisms (Fig. 1; Swanson and Sawchenko, 1983, Sawchenko et al., 1996). This single cell group is unique in its content of ample representations of each of the major physiological response avenues that the brain has at its disposal to mobilize bodily resources in response to environmental insults. The cells that give rise to major classes of visceromotor projections are separate from one another (Swanson and Kuypers, 1980; Swanson et al., 1980), suggesting they are not necessarily called into play in a stereotyped, all-or-none, manner, but rather that the potential exists for differential recruitment. The physiological and functional anatomical literatures provide ample evidence that this potential is realized, and that the mix of regulatory responses that emanate from the PVH varies lawfully, and on

Fig. 1. Organization of the paraventricular nucleus. Photomicrographs to show the cellular architecture (upper left), some major visceromotor cell types (upper right), and the functional organization (bottom) of the PVH. At the upper right are shown, in single fluorescence photomicrograph, cells immunoreactive fdr CRF (red) and vasopressin (green); cells retrogradely labeled following an injection of a tracer in the spinal cord (blue) are representative of the autonomic-related outputs of the nucleus.

63

a moment-to-moment basis, as a function of physiological state. An obvious starting point in considering the substrates that provide for situation-specific recruitment of cell types within the PVH rests in the organization of its neural inputs. This topic has been reviewed in some detail (e.g. Swanson and Sawchenko, 1983; Sawchenko et al., 1996), and only a few generalizations will be repeated here. First, in keeping with the range of specific stimuli that are capable of eliciting stress responses, the PVH receives a rich and diverse afferent supply, which includes potential routes by which all major modalities of sensory information may come to influence PVH mechanisms. Second, the PVH is not known to receive a substantial input from the cerebral cortex or the dorsal thalamus. Its major direct inputs arise, instead, from a number of distinct cell groups in the brainstem, limbic forebrain and from most other hypothalamic cell groups. Third, inputs from individual afferent sources vary markedly with respect to the manner and extent to which they distribute to individual effector neuron populations represented within the nucleus. And finally, as visceral effector neurons, all PVH cell types are powerfully innervated by pathways conveying visceral sensory information. Two sets of projections provide a substantial portion of the visceral sensory control of PVH outputs. The nucleus of the solitary tract is the principal central recipient of first-order sensory information carried by the vagus and glossopharyngeal nerves, which supply the bulk of the thoracic and abdominal viscera. The NTS projects to the PVH both directly and via relays in the ventrolateral medulla; both the direct and relayed projections arise principally from catecholaminesynthesizing neurons (Cunningham and Sawchenko, 1988; Cunningham et al., 1990). These neural pathways are complemented by a series of descending projections from a triad of structures that line the rostral margin of the third ventricle, or lamina terminalis. Components of this complex lay outside the blood-brain barrier, and possess morphological specializationsthat permit them to serve as important transducers of signals borne by circulating macromolecules and the ionic composition of the blood (Gross, 1987). These structures

are extensively interconnected with one another, and each projects strongly to the PVH (Miselis et al., 1979; Sawchenko and Swanson, 1983).

Categorization of stress models It is common among workers in the field to partition stress models into two basic categories, referred to in the older literature as systemic (also referred to as homeostatic or physiological) and neurogenic (emotional or psychological) paradigms (e.g. Fortier, 1951; Allen et al., 1973). While there exists no consensus as to the validity or utility of such distinctions, systemic stressors are generally conceived as targeted perturbations in physiologic parameters that are transduced by a manageably small number of peripheral or central receptors, and whose essential features are not consciously appreciated. Commonly employed systemic stress models include cardiovascular, osmotic and immune challenges. Neurogenic stresses, by contrast, involve manipulations whose effective stimuli are less easily specified, often rely upon somatosensory or nociceptive pathways for their initial transduction, and involve a distinct cognitive and/or affective component. Restraint, immobilization and electrical footshock are widely exploited neurogenic stress paradigms. In the remainder of the chapter, we will compare and contrast attempts to unravel the circuitry mediating adaptive responses in these putative classes of stressors, focusing on an immune challenge model (intravenous interleukin1 injection) and intermittent electrical footshock as representative of systemic and neurogenic paradigms, respectively. Electrical footshock is a time honored stress model, well known to provoke increases in secretory and biosynthetic activity at each level of the HPA axis (e.g., Kant et al., 1984; Imaki et al., 1993), and in the sympathoadrenal system (e.g., McCarty and Kopin, 1978). The upregulation of CRF expression in the PVH seen following repeated footshock can be blocked by concurrent local administration of drugs that interfere with synaptic transmission (Sawchenko et al., 1993), suggesting that synaptic input to the CRF neuron is involved in the elicitation andor maintenance of increased central drive to the HPA axis that accompanies this model.

64

IL-1 is a cytokine that mimics most of the central acute phase responses to immune insult, including a capacity to powerfully stimulate HPA output. This effect has come to provide a popular venue for studying what are now recognized as bidirectional interactions between the central nervous and immune systems. Thus, on one hand, glucocorticoids, the end product of the HPA cascade, are broad-spectrum inhibitors of immune and inflammatory functions (Sternberg and Wilder, 1993), and have been exploited clinically for this purpose for decades. On the other, certain immune system principles, prominently including IL- 1, are capable of potently stimulating HPA output (Besedovsky et al., 1975). This effect has been interpreted as indicative of a usurping of a neuroendocrine mechanism by the immune system to negatively regulate its own activity, that is, to provide a brake on excess cytokine production and immune cell proliferation following episodes of infection or inflammation (Besedovsky et al., 1986). Dysfunction of this regulatory mechanism has been implicated as playing a role in certain forms of autoimmune disease (Wick et al., 1993). The questions of the identity of the pathways and mechanisms by which such immune influences come to be exerted on central systems is a daunting one, which has commanded a great deal of attention among workers in the nascent field of neuroimmunology (Dantzer, 1994; Watkins et al., 1995; Ericsson et al., 1996; Elmquist et al., 1997b). IL-1 is a 17.5 kDa protein that would normally not be expected to freely traverse the blood-brain barrier. A number of alternative mechanisms have been proposed as to how the cytokine may access the central nervous system in general, and HPA control systems, in particular. These include: (1) facilitated transport across the barrier (Banks et al., 1989), which would permit cytokine interactions with any central elements bearing appropriate receptors; (2) entry at circumventricular organs, such as the vascular organ of the lamina terminalis or the area postrema, which lie outside the barrier and are known to project either directly or indirectly to the PVH; (3) peripheral transduction by nerves such as the vagus, whose sensory components exhibit IL-1 sensitivity (Ek et al., 1998), and terminate in the NTS, affording them direct access

to the endocrine hypothalamus by way of ascending catecholamine pathways alluded to above.

Neural systems activated in systemic and neurogenic stress Recognition of the ability of certain immediateearly genes (IEGs), particularly the c-fos proto-oncogene, to provide generic indices of cellular activation in the central nervous system (Morgan and Curran, 1991), has greatly facilitated the identification and characterization of hypothalamic effector neurons and extended neural systems that are targeted in a range of acute stress paradigms (e.g., Chan et al., 1993; Hoffman et al., 1993). This, in turn, has enabled hypotheses to be formulated concerning how components of these systems may come to comprise situation-specific functional circuits. A comparison of cellular activation patterns seen in response to acute cytokine and footshock challenges reveals them to be largely distinct, and in keeping with those reported in other systemic and neurogenic paradigms, respectively, though they share some key features in common that may be relevant to understanding how adaptive visceromotor responses are mediated in each. Intravenous administration of IL-1 at a dose moderately above the threshold for eliciting HPA secretory responses provokes transient expression of c-fos mRNA and its protein product, Fos, in the PVH that peak at 1 and 2-3 hr, respectively, after the challenge, and abate rapidly thereafter (Ericsson et al., 1994). A single exposure to a 30 min intermittent footshock session elicits similar responses with a somewhat accelerated time course (Pezzone et al., 1992; Imaki et al., 1993; Li and Sawchenko, 1998). In fact, the patterns of cellular activation elicited by the two stressors are indistinguishable, and include a pervasive activation of parvocellular CRF neurons, with secondary involvement of autonomic-related projection neurons and oxytocin-expressing magnocellular neurosecretory cells. Thus, both models activate cell groups involved in the initiation of the endocrine and neural arms of the stress response. Although the responses of oxytocin-containing magnocellular neurosecretory neurons is consistent with the fact that many stressors, particularly

65

neurogenic models, elicit increased oxytocin secretion, the adaptive significance of this is unclear (Li and Sawchenko, 1998). Beyond the endocrine hypothalamus, patterns of IL- 1 and footshock-induced cellular activation are generally quite distinct. Cytokine injection provokes Fos induction in a handful of cell groups in the limbic forebrain (bed nucleus of the stria terminalis and central nucleus of the amygdala) and brainstem (lateral parabrachial nucleus, NTS and ventrolateral medulla) that are acknowledged as nodal points in central autonomic regulation (Fig. 2; Saper, 1995). The linchpin in this pattern is the NTS, as all of the others are known receive direct inputs from it. All are also known to project extensively to one another, and to the PVH. All were found to respond to IL-1 with sensitivities that rivaled or even exceeded those of target cells in the PVH, while circumventricular organs responded only at doses an order of magnitude above the threshold required for hypothalamic activation, and other cell groups implicated as potential afferent mediators of IL-1 effects on the HPA axis, such as the hippocampal formation and midbrain raphe nuclei (Cunningham et al., 1992), were unresponsive over the range of doses examined. By contrast, extrahypothalamic cell groups activated in response to acute footshock prominently involved limbic (prefrontal), and to a lesser extent somatosensory, regions of the cerebral cortex, aspects of subcortical limbic structures (septum, amygdala, bed nucleus) whose dominant connections are with neocortical regions and one another, and cell groups the thalamus (midline and intralaminar nuclei) and brainstem (periaqueductal gray, mesopontine cholinergic cell groups, locus coeruleus) involved in the processing of somatic sensory information, particularly that carried by the spinothalamic tract (Fig. 3; Bullit et al., 1990; Willis et al., 1995; Li and Sawchenko, 1998). Despite the fact that this basic pattern of footshock-inducedFos expression differed in fundamental ways from that provoked by cytokine injection, as in the IL-1 situation, catecholamine-containing cell groups of the NTS and ventrolateral medulla were also prominently activated in response to acute footshock (Pezzone et al., 1993; Li and Sawchenko, 1998).

To identify candidate mediators of IL-1 and footshock effects on the PVH, retrograde transport methods were used in experimental settings to ask which of the cells that exhibit sensitivity in each of the paradigms also project directly to the PVH. In rats challenged with a moderate dose of IL-1, catecholamine containing neurons of the NTS and ventrolateral medulla comprised the only loci at which such doubly labeled cells were reliably found (Ericsson et al., 1994). Although these were found to be more widely distributed in footshock situation, with representations in brainstem cell groups involved in the processing or modulation of somatosensory information, and in the limbic forebrain and hypothalamus, medullary catecholamine neurons once again comprised the dominant seat of these, accounting for more than 60%of the total population of footshock-sensitive neurons identified as contributing afferent projections to the PVH (Li and Sawchenko, 1998). Despite some fundamental differences in the manner in which the two stress paradigms impact the brain, aminergic neurons present themselves as prime candidates for mediating the recruitment of stress-related PVH effector populations in each. These IL-1 and footshock-induced patterns of cellular activation are quite generally representative of those reported in other systemic and neurogenic stress paradigms. Hemorrhage, for example, activates a very similar complement of central autonomic cell groups as is seen in the IL-1 situation, albeit with distinctive emphases, and also without significant involvement of the cerebral cortex (Li and Dampney, 1994). Similarly, comparison of Fos induction patterns seen in footshock, physical restraint or immobilization reveal only very subtle differences (Senba et al., 1993; Chen and Herbert, 1995; Cullinan et al., 1995; Viau and Sawchenko, 1995). Notwithstanding the distinctive nature of way in which the brain responds to what we have termed systemic and neurogenic stress, the fact remains that the footshock and L-1 models, among a number of others, both activate a similar complement of stress-related effector populations in the PVH, and medullary catecholamine neurons, which provide a major source of input to it. Although these catecholamine cell groups are best known for their key roles in the processing of

66

CP

67

visceral sensory information, some of them are known to receive somatosensoryhociceptive input, as well (e.g. Smith and Day, 1994). Collectively, these considerations raise the possibility that by virtue of their projections to the hypothalamus, aminergic cell groups may comprise a substrate for a generalized (or nonspecific) alarm response to any acute challenge, thereby functioning in a manner coarsely analogous to their phenotypic counterparts in the sympatho-adrenal system.

Catecholaminergicinvolvement in stress effects on PVH mechanisms The identification in both the footshock and IL-1 injection paradigms of medullary aminergic neurons as the dominant seats of stress-activated neurons that project to the PVH implicate this afferent source as a mediator of hypothalamic effects in each instance. Ablation experiments were carried out to test this hypothesis. Unilateral transections of ascending aminergic projections near their origins in the medulla that were effective in markedly depleting the PVH of its catecholaminergic innervation, were found to essentially eliminate IL- 1 stimulated increases in Fos induction and CRF mRNA expression on the lesioned side of the brain, while leaving these responses fully intact on the non-lesioned side (Ericsson et al., 1994; see Fig. 4). Compatible findings have been reported using a catecholaminespecific neurotoxin as a lesioning instrument, and plasma corticosterone secretion as an endpoint (Chuluyan et al., 1992). By contrast, these same manipulations fail to modify increases in cellular activation, or in indices of the biosynthetic and secretory activity of the HPA axis, seen in response to footshock or restraint stress (Chuluyan et al., 1992; Li et al., 1996). These findings suggest that ascending aminergic pathways are specifically involved in mediating IL-1 effects on PVH effector

mechanisms, including CRF-expressing neurons that comprise the central limb of the HPA axis. In the foregoing example, a Fos based approach was used to generate a hypothesis concerning the afferent mediation of cytokine effects on hypothalamic stress-related effector mechanisms, which was tested, supported and remains viable (see below). Similar strategies have been used to probe the circuitries that may underlie hypothalamic visceromotor responses in other systemic stress paradigms (Fig. 5). Data have been gathered to suggest that the widespread activation of the magno- and parvocellular neurosecretory systems, and of pre-autonomic neurons in the PVH, seen in response to hemorrhage also depend heavily, though not exclusively, on the integrity of ascending aminergic pathways (Chan and Sawchenko, 1996, and unpublished observations), while the differential effects of chronic hyperosrnolality (saltloading) on the parvocellular (inhibition) and magnocellular (activation) neurosecretory systems are achieved via descending inputs from the ventral part of the lamina terminalis (Kovrics and Sawchenko, 1993). Taken together, these results support the view that systemic stress effects on hypothalamic effector mechanisms may be conceived as visceral reflex responses that are mediated in a challenge-specific manner by wellknown central autonomic pathways and associated circumventricular structures.

Mechanisms of IL-1 effects on the central limb of the HPA axis Lesion experiments described above support an involvement of ascending catecholaminergic projections in mediating IL-1 effects on PVH effector mechanisms, that is, that we have extended our understanding of the underlying circuitry one synapse removed from the target. But the initial

4 Fig. 2. IL-1-induced cellular activation in extrahypothalamic regions. Schematic representations of Fos induction patterns seen in response to intravenous IL-1 injection, imposed on drawings of rat brain regions modified from Swanson (1992). The pattern of activation is overwhelmingly subcortical, and is most prominent in the oval subnucleus of the bed nucleus of the stria terminalis (BSTov), the central nucleus of the amygdala (CeA), the lateral parabrachial nucleus (PBI) and catecholaminergic regions of the nucleus of the solitary tract (NTS) and ventrolateral medulla (VLM). Each of these regions is known to be involved in the processing of visceral sensory information conveyed initially to the NTS, and to engage in extensive interconnections with one another and with the PVH.

68

I

I I

69

Fig. 3. Footshock-inducedcellular activation in extrahypothalamic regions. Schematic representations of Fos induction patterns seen in response to a single exposure to a 30 min intermittent footshock session, organized after Figure 2. Shock-induced activation patterns contrast sharply with those resulting from IL- 1 injection in including prominent responses in prefrontal cortical regions (PL, LA), and other aspects of the limbic forebrain including the lateral septa1 nuclei (LSi, LSv), bed nucleus (BST), and amygdala (BLA, MEA), whose dominant connections are with the cerebral cortex and other limbic structures. In addition, acute footshock activates cell groups associated with the processing and/or modulation of somatosensory and nociceptive information, including the periaqueductal gray, the laterodorsal (LDT) and pedunculopontine tegmental nuclei (PPN) and the locus coeruleus (LC). Despite fundamental differences with the pattern of activation seen in response to IL-1, footshock shares with the cytokine injection model a capacity to prominently activate catecholaminergic neurons in the nucleus of the solitary tract (NTS) and the ventrolateral medulla (VLM)

A.

IL-1

Contra lpsi

Contra lpsi

PVH

NTS

B. Footshock

Contra lpsi

PVH

Contra lpsl

NTS

Fig. 4. Medullary knife cut effects on IL-1 and footshock-induced Fos expression in hypothalamus and medulla. Diagrammatic summary and mean f SEM number of Fos-ir neurons in the NTS and PVH on the side of the brain ipsilateral to a medullary knife cut, expressed as a percentage of immunoreactive cells on the contralateral side, following acute IL-1 (A; top) and footshock (B; bottom) challenges. Lesion effects are summarized in a schematic drawing of a horizontal section through the brain at the left of each graph. In the IL-1 model, transections significantly reduced Fos induction in the PVH, while leaving the activation of medullary aminergic neurons intact. Opposite effects were noted in the footshock paradigm. **, p c0.01 vs. contralateral side. Modified after Li et al., 1998.

70

projections are clearly in a position to influence medullary aminergic neurons (Cunningham et al., 1993). But neither abdominal vagotomy nor lesions of the area postrema were found to affect IL-1 induced activational responses at the level of either the hypothalamus or medulla (Ericsson et al., 1997). While there is ample evidence to support an involvement of vagal mechanisms in IL-1 and endotoxin effects on HPA activity, this comes principally from paradigms in which the immune challenge is administered via the intraperitoneal

question as to how IL-1 gains initial access to the brain parenchyma to affect HPA control circuitry remains. In view of the dominant role aminergic neurons play in the processing of visceral sensory information, the possibility of peripheral transduction by the vagus nerves remains viable (Ek et al., 1998). Similarly, the circumventricular component of the NTS complex, the area postrema, must be entertained as a potential route of access on the basis of its capacity to express IL-1 receptors (Ericsson et al., 1995) and the fact that its neural

I'\

\*c1

SOMATOSENSORY 7

/Jd +

ACTH

Effects

Model

Paw0 Magno

cut 1

cut 2

Cytokine (IL-1)

t

t

intact

eliminated

Salt-loading

I

t

eliminated

intact

t t

t t

intact

attenuated

intact

intact

Hemorrhage Footshock

Fig. 5. Differential dependence of stress-induced activation of PVH mechanisms on projections from the lamina terminals and medullary aminergic neurons. The approximate positions of knife cuts and the projections they sever are shown schematically in a drawing of a sagittal view of the rat brain (top), and their effects summarized in tabular form, below. Hypothalamic effects of the three systemic stress models (intravenous IL-1 injection, chronic salt loading and hemorrhage) require the integrity of one or another of these inputs to be manifest, while those of the single neurogenic stress model (footshock) do not.

71

route (Wan et al., 1994; Gaykema et al., 1995; Kapcala et al., 1996). Studies of the central distribution of the type 1 IL- 1receptor, the subtype acknowledged as mediating the cytokine’s biological effects (e.g. Sims et al., 1994), initially offered little in the way of insight into the problem. The small handful of neuronal cell groups that were found to express the receptor have no obvious or particularly direct involvement in hypothalamic regulation. Instead the dominant sites of receptor expression include non-neuronal elements that comprise the great barriers between the brain and its fluid environments, including the meninges, the ependyma, the choroid plexus and cells lining substantial portions of the cerebral vasculature (Wong and Licinio, 1994; Yabuuchi et al., 1994; Ericsson et al., 1995; see Fig. 6). Of these, only cells associated with the vasculature were found to display activational

responses to a moderate-dose intravenous IL- 1 challenge (Ericsson et al., 1995), raising the possibility that the initial transduction step may occur at the blood-brain interface. Among the local signaling molecules that may be produced by perivascular cells, certain prostaglandins have been most strongly implicated as playing a role in IL-1 effects. Blockade of prostanoid biosynthesis has been shown to reduce or eliminate immune challenge effects on HPA output (Watanabe et al., 1990). This effect has been interpreted as suggesting a possible role for circulating prostaglandins in mediating the effects of immune challenges on the brain. But in response to endotoxin challenges, brain perivascular cells have been shown to be capable of manifesting induced expression of prostaglandin E2 (PGE2; Van Dam et al., 1993), and a key enzyme in prostanoid biosynthesis, cyclooxygenase-2 (COX-2; Elmquist

Fig. 6. Expression of the type 1 IL-1 receptor in rat brain. Darkfield photomicrograph to show the distribution of cells showing positive hybridization signals (white) for IL-1R1 mRNA, in a coronal section at level of the PVH (left).The drawing at the right shows key features of the distribution schematically. This receptor is expressed principally by non-neuronal elements in the meninges, choroid plexus, ependyma, and lining cerebral blood vessels. Neuronal expression at this level (seen in a small group of cells in the amygdala), and throughout the brain, is limited.

12

et al., 1997; Matsumura et al., 1998), raising the possibility that prostaglandin-dependent step may reside within the brain. Figure 7 presents a proposed model for IL-1-induced activation of neurons comprising the central limb of the HPA axis (Ericsson et al., 1997), which involves: (1) IL1 binding of its type 1 receptor on cells lining the medullary vasculature with consequent local release of prostanoids, most likely PGE2, into the extracellular space; (2) PGE2 interacting with cognate receptors expressed on or near medullary aminergic neurons, resulting in activation of this population; and (3) consequent synaptic excitation of hypothalamic CRF neurons, by way of direct axonal projections. All the requisite components of this proposed signaling cascade are in place in the caudal medulla to support this mechanism, though there remain

significant unresolved questions. The vascularassociated cell type@) that express the IL- 1 receptor have not been identified, and there are ostensibly conflicting reports that endothelial cells (e.g. Matsumura et al., 1998) and perivascular microglia (e.g. Elmquist et al., 1997a) comprise the dominant seat of induced COX-2 expression following endotoxin challenges. This leaves open to question whether a single cell type may mediate both the initial transduction and prostanoid release, or whether more complex interactions among cells associated with the vasculature, perhaps involving additional mediators (e.g. Wong et al., 1996) must be considered. Even more challenging is the question of how, in the face of widespread activation and COX-2 induction throughout the cerebral microvasculature, specific targeting of medullary aminergic neurons may be achieved. A

Fig. 7. A candidate mechanism for IL-I-mediated stimulation of HPA control systems. The image at the left shows combined localization of perivascular cells displaying IL- 1RI mRNA and IL-1-stimulated Fos-ir in the ventrolateral medulla. Evidence is summarized in the text to suggest that circulating IL-1 binds its cognate receptor on perivascular receptors in the region, inducing them to synthesize PGE2, which, in turn diffuses through the extracellular space to (directly or indirectly) stimulate nearby aminergic neurons, and consequently CRF-expressing targets of their axonal projections in the endocrine hypothalamus. Listed at the right are markers relevant to this signaling cascade that have been localized in the requisite regions under basal or stimulated conditions. Markers that have been colocalized in individual cell types are bracketed.

13

recent examination of the central distribution of the dominant PGE2 receptor subtype expressed in brain (Ericsson et al., 1995, 1996), the EP3 subtype, may provide some initial insight. This receptor has been found to be expressed by IL1-sensitive neurons in the NTS and ventrolateral medulla, but not by target cells in the PVH, or to any substantial extent in other cell groups known to provide major inputs to it. Recent studies have shown that IL-1 induced activation of medullary catecholamine neurons and their hypothalamic targets are mitigated to comparable extents by graded levels of systemic prostaglandin synthesis blockade (Ericsson et al., 1997), thus fulfilling a prediction that follows logically from the hypothesized mechanism. More importantly, microinjections of PGE2 into the major seat of IL-1 sensitive medullary aminergic inputs to the PVH provokes surprisingly discrete activation of catecholaminergic neurons in the region, and essentially recapitulates the effects of systemic IL-1 injection in the PVH, and, indeed, throughout the forebrain (Ericsson et al., 1997). Thus, these findings offer initial support for the mechanism outlined above as a component of the minimum essential circuitry required for the activation of the central limb of the HPA axis by increased circulating IL- 1. It is important to emphasize that the available evidence is quite clear in indicating that central pathways and mechanisms that are responsive to challenges posed by treatment with individual cytokines, or to more complex immune stimuli, can vary markedly as a function of the nature of the stimulating agent(s), dose and route of administration. For example, it is clear that more strenuous and complex immune challebes can elicit activation of additional sources of afferent input to the PVH (e.g. Elmquist and Saper, 1996), and may well be involved in governing PVH output under such conditions. The challenge for the field, in general, is to establish a hierarchy of the particular routes of access and associated circuitry that underlie adaptive responses to specific, and ever more naturalistic, immune challenges. In a more general sense, attempts to clarify the nature of immune system influences on brain have identified a number of novel mechanisms by which immune

mediators may come to influence both closely regulated physiologic functions and, under extreme conditions, central nervous system function in a more global sense. The barriers between the brain and its fluid environment seem much less imposing and monolithic than they did a few short years ago.

Possible pathways mediating footshock stress effects Despite the fact that the footshock and IL-1 models share a capacity to activate a common complement of effector neurons in the PVH, and a major source of neural input to them, the extent to which the recruitment of hypothalamic effectors seen in response to these stressors depends on the integrity of that afferent appears to be starkly differential (Fig. 4).While disruption of medullary aminergic projections failed to mitigate footshock effects on the PVH, this is not to say that these lesions were without effect in this paradigm. Indeed, prior unilateral transections markedly attenuated footshock induced activational responses in the medullary aminergic cell groups, themselves, on the ipsilateral side of the brain, and this again contrasted with the situation in the IL-1 challenge model, where knife cuts exerted no significant activational responses in these neurons on the lesioned side (Li et al., 1996). These findings are consistent with the view that the activation of aminergic neurons seen in the footshock model is a secondary consequence of stress, mediated by a descending pathway that is disrupted by the lesion. While neither the origin of such an influence nor the significance of the recruitment of aminergic population in the footshock situation is clear, aminergic pathways are not required for shockinduced recruitment of PVH mechanisms. What cell groups, then, may govern hypothalamic responses in the footshock situation? Apart from medullary aminergic neurons, relatively small complements of footshock-sensitive neurons that were identified as projecting to the PVH were found in discrete aspects of the brainstem, limbic forebrain and hypothalamus. What links these disparate cell groups together is that each of them may be viewed as lying along the path of the direct

74

and extended projections of spinothalamic and/or spinohypothalamic tracts, which are known to convey nociceptive information to the forebrain (Bullitt, 1990; Cliffer et al., 1991; see Fig. 8). Thus, footshock-sensitive, PVH-projecting neurons were identified in the periaqueductal gray, the pedunculopontine and laterodorsal tegmental nuclei and the locus coeruleus, each of which is known to be involved in the processing and/or modulation of somatosensory and nociceptive information (Willis et al., 1995). And while the dorsal thalamic termini of the spinothalamic pathway are not known to issue direct projections to the PVH, aspects of the limbic forebrain in receipt of inputs from the midline and/or intralaminar nuclei, including the prefrontal cortex and the lateral septa1 nucleus (e.g. Groenewegen and Berendse, 1994), were consistently identified as sources of footshock-sensitive projections to the PVH. Neither cell group is known to issue prominent inputs to the PVH, itself, but do project to and/or through immediately adjoining regions where local inhibitory interneurons have been identified as projecting

into the PVH, proper (Roland and Sawchenko, 1993; Boudaba et al., 1996). Both have been implicated as playing a modulatory, principally inhibitory, role in neuroendocrine regulation (Herman and Cullinan, 1997). In addition, two additional sites in which footshock-sensitive,retrogradely labeled neurons were found, the dorsomedial nucleus of the hypothalamus and the posterolateral part of the bed nucleus, both provide substantial projections to the PVH, whose terminal distributions closely mimic that of footshockinduced Fos expression (Sawchenko and Swanson, 1983). Both receive direct inputs from the medial prefrontal cortex and/or the lateral septum, and both may be envisioned as lying one step distal along a putative path of footshock-related sensory information flow through the medial thalamus and limbic forebrain. Determining the extent to which these candidate afferent mediators may underlie footshock effects on PVH mechanisms will address the key question of whether the visceromotor and affective responses to a neurogenic stress are organized in

ACTH Fig. 8. Possible pathways mediating acute footshock-induced activation of PVH mechanisms. A schematic sagittal view of the rat brain on which are represented the distribution of neurons displaying acute footshock-induced Fos expression in cells that were (filled circles) and were not (open circles) identified as projecting to the region of the PVH. Imposed upon this is an indication of the trajectory of the spinothalamic and spinohypothalamic tracts, and possible downstream pathways in a position to convey this information on to the limbic forebrain and hypothalamus, including the PVH. It remains to be determined whether footshock-related sensory information accesses the PVH in a relatively direct manner, or requires processing through the limbic system to become affixed with an appropriate emotional valence before being in a position to elicit activation of stress-related effector populations.

75

series or in parallel. Does footshock-related sensory information access PVH effectors in a relatively direct manner, or must it first be processed through the limbic forebrain for evaluation of its emotional significance? Swnmary and conclusions

The results of recent studies support a partitioning of stress models into at least two basic classes. While these have been referred to as ‘systemic’and ‘neurogenic’, we would suggest that the terms intemceptive and exteroceptive, respectively, are more apt descriptors. This is based on the similarities in the overall patterns of activational responses seen as a consequence of exposure to a range of perturbations in the internal versus external environments. While stressors of each class may share in common such fundamental features as a capacity to enlist certain PVH effector populations and medullary catecholamine-containing neurons, both the capacity to involve specific output neuron classes and the dependence of hypothalamic effects on the integrity of aminergic afferents in at least some interoceptive and exteroceptive models, are clearly differential.The available evidence suggests that interoceptive stress effects on PVH effector populations may be conceived essentially as simple reflex responses, mediated at a subcortical level by cell groups and associated circumventricularorgans that comprise the core of a system involved in the processing of visceral sensory information. Based on the general pattern of acute footshock-induced Fos expression and commonalities of cellular activation profiles seen in this and other acute exteroceptive paradigms, it seems a reasonable assumption that pathways that convey somatosensoryhociceptive information to the PVH are apt to mediate adaptive visceromotor responses in these models. Multiple candidates for such roles have been identified at various levels of what may be viewed as the ascent of the spinothalamic pathway through the brainstem and thalamus, and on through the limbic forebrain and hypothalamus. Dissecting the relative contributions of these in determining PVH output will speak to important conceptual issues concerning the extent to which the affective and visceromotor responses to exter-

oceptive stressors are organized, and the level(s) at which these different avenues of emotional expression may be integrated.

Acknowledgements Work from our laboratory summarized here was supported by grants NS-21182 and HL-35137 from the National Institutes of Health, and was conducted in part by the Foundation for Medical Research. PES is an Investigator of the Foundation for Medical Research. We gratefully acknowledge the excellent assistance of Belle Wamsley (editorial) and Kris Trulock (photography and graphics) in the preparation of the manuscript.

References Allen, J.P., Allen, C.F., Greer, M.A. and Jacobs, J.J. (1973) Stress-induced secretion of ACTH. In: S. Brodish and E.S. Redgate (Eds), Brain-PituitapAdrenal Interrelationships, Karger, Basel, pp. 99-127. Banks, W.A., Kastin, A.J. and Durham, D.A. (1989) Bidirectional transport of interleukin-1 alpha across the blood-brain barrier. Brain Res. Bull., 23: 433437. Besedovsky, H., DelRey, A., Sorkin, E. and Dinarello, C.A. (1986) Immunoregulatory feedback between interleukin-1 and glucocorticoid hormones. Science, 233: 652-654. Besedovsky, H., Sorkin, E., Keller, M. and Muller, J. (1975) Changes in blood hormone levels during the immune response. Proc. SOC.Exp. Biol. Med., 150: 466-470. Boudaba, C., Szabo, K., and Tasker, J.G (1996) Physiological mapping of local inhibitory inputs to the hypothalamic paraventricular nucleus. J. Neurosci., 16: 7151-7160. Bullitt, E. (1990) Expression of c-fos-like protein as a marker for neuronal activity following noxious stimulation in the rat. J. Comp. Neurol., 296: 517-530. Chan, R.K.W. and Sawchenko, P.E. (1996) The roles of medullary catecholaminergic and lamina terminalis projections in mediating hemorrhage-induced increases in CRF and vasopressin expression in the paraventricular nucleus. Proceedings of the 10th International Congress of Endocrinology,Vol. 2, p. 52 1. Chen, X. and Herbert, J. (1995) Regional changes in c-fos expression in the basal forebrain and brainstem during adaptation to repeated stress: Correlations with cardiovascular, hypothermic and endocrine responses. Neuroscience, 64: 675-685. Chan, R.K.W., Brown, E.R., Ericsson, A., Kovics, K.J. and Sawchenko, P.E. (1993) A comparison of two immediateearly genes, c-fos and NGFI-B, as markers for functional activation in stress-related neuroendocrine circuitry. J. Neurosci., 13: 5125-5138. Chuluyan, H.E.,Saphier, D., Rohn, W.M. and Dunn, A.J. (1992) Noradrenergic innervation of the hypothalamus

16 participates in adrenocortical responses to interleukin- 1. Neumendocrinology, 56: 106-1 11. Cliffer, K.D., Burstein, R. and Giesler, G.J., Jr. (1991) Distributions of spinothalamic, spinohypothalamic, and spinotelencephalic fibers revealed by anterograde transport of PHA-L in rats. J. Neurosci., 11: 852-868. Cullinan, W.E., Herman, J.P., Battaglia, D.F., Akil, H. and Watson, S.J. (1995) Pattern and time course of immediate early gene expression in rat brain following acute stress. Neuroscience, 64: 471-505. Cunningham, E.T., Jr., Bohn, M.C. and Sawchenko, P.E. (1990) Organization of adrenergic inputs to the paraventricular and supraoptic nuclei of the hypothalamus in the rat. J. Comp. Neurol., 292: 65 1-667. Cunningham, E.T., Jr. and Sawchenko, P.E. (1988) Anatomical specificity of noradrenergic inputs to the paraventricular and supraoptic nuclei of the rat hypothalamus. J. Comp. Neurol., 214: 60-16. Cunningham, E.T., Jr., Wada, E., Carter, D.B., Tracey, D.E., Battey, J.F. and De Souza, E.B. (1992) In situ histochemical localization of type 1 interleukin-I receptor RNA in the central nervous system, pituitary, and the adrenal gland of the mouse. J. Neurosci.. 12: 1101-1 114. Dantzer, R. (1994) How do cytokines say hello to the brain? Neural versus humoral mediation. EUKCytokine Network, 5: 211-273. Ek, M., Kurosawa, M., Lundeberg, T. and Ericsson, A. (1998) Activation of vagal afferents after intravenous injection of interleukin-1 beta: role of endogenous prostaglandins. J .Neurosci., 18: 9471-9. Elmquist, J.K., Breder, C.D., Sherin, J.E., Scammell, T.E., Hickey, W.F., DeWitt, D. and Saper, C.B. (1997a) Intravenous lipopolysaccharide induces cyclooxygenase 2-like immunoreactivity in rat brain perivascular microglia and meningeal macrophages. J. Comp. Neurol., 381: 119-129. Elmquist, J.K. and Saper, C.B. (1996) Activation of neurons projecting to the paraventricular hypothalamic nucleus by intravenous lipopolysaccharide. J. Comp. Neurol., 314: 3 15-33 1. Elmquist, J.K., Scammell, T.E. and Saper, C.B. (1997b) Mechanisms of CNS response to systemic immune challenge: the febrile response. Trends Neurosci., 20: 565-570. Ericsson, A., Arias, C. and Sawchenko, P.E. (1997) Evidence for an intramedullary prostaglandin-dependent mechanism in the activation of stress-related neuroendocrine circuitry by intravenous interleukin-1. J. Neurosci., 11: 7166-7179. Ericsson, A., Ek, M., Wahlstrom, I., Kovacs, K., Liu, C.-L., Hart, R. and Sawchenko, P.E. (1996) Pathways and mechanisms for interleukin-1 mediated regulation of the hypothalamo-pituitary-adrenal axis. In: R. McCarty, G . Aguilera, E. Sabban and R. Kvetnansky (Eds), Stress: Molecular Genetic and Neurobiological Advances, Gordon and Breach, New York, pp. 101-120. Ericsson, A., Kovics, K.J. and Sawchenko, P.E. (1994) A functional anatomical analysis of central pathways subserving the effects of interleukin-1 on stress-related neuroendocrine neurons. J. Neurosci., 14: 897-9 13.

Ericsson, A., Liu, C., Hart, R.P. and Sawchenko, P.E. (1995) Type 1 interleukin-1 receptor in the rat brain: distribution, regulation, and relationship to sites of IL-1-induced cellular activation. J. Comp. Neurol., 361: 681698. Fortier, C. (1951) Dual control of adrenocorticotropin release. Endocrinology, 49: 782-788. Gaykema, R.P.A., Dijkstra, I. and Tilders, F.J.H. (1995) Subdiaphragmatic vagotomy suppresses endotoxin-induced activation of hypothalamic corticotropin-releasing hormone neurons and ACTH secretion. Endocrinology, 136: 41174720. Groenewegen, H.J. and Berendse, H.W. (1994) The specificity of the ’non-specific’ midline and intralaminar thalamic nuclei. Trends Neurosci., 17: 52-57. Gross, P.M. (1987) Circumventriculur Organs and Body Fluids. Boca Raton, CRC Press. Herman, J.P. and Cullinan, W.E. (1997) Neurocircuitry of stress: central control of the hypothalamo-pituitary-adrenocortical axis. Trends Neurosci., 20: 78-84. Hoffman, G.E., Smith, M.S. andverbalis, J.G. (1993) C-fos and related immediate early gene products as markers of activity in neuroendocrine systems. Front. Neuroendocrinol., 14: 173-2 13. Imaki, T., Shibasaki, T., Hotta, M. and Demura, H. (1993) Intracerebroventricular administration of corticotropinreleasing factor induces c-fos mRNA expression in brain regions related to stress responses: comparison with pattern of c-fos mRNA induction after stress. Brain Res., 616: 114-125. Kant, G.J., Mougey, E.H., Pennington, L.L. and Meyerhoff, J.L. (1984) Graded footshock stress elevates pituitary cyclic amp and plasma P-endorphin, P-LPH, corticosterone and prolactin. Life Sci., 33: 2657-2663. Kapcala, L.P., He, J.R., Gao, Y., Pieper, J.O. and DeTolla, L.J. (1996) Subdiaphragmatic vagotomy inhibits intra-abdominal interleukin-1 (3 stimulation of adrenocorticotropin secretion. Brain Res., 728: 241-254. Koviics, K.J. and Sawchenko, P.E. (1993) Mediation of osmoregulatory influences on neuroendocrine CRF expression by the ventral lamina terminalis. Proc. Natl. Acad. Sci. USA, 90: 7681-7685. LeDoux, J.E. (1987) Emotion. In F. Plum (Ed.), Handbook of Physiology - The Nervous System American Physiological Society, Washington, DC, pp. 419-459. Li, H.-Y., Ericsson, A. and Sawchenko, P.E. (1996) Distinct mechanisms underlie activation of hypothalamic neurosecretory neurons and their medullary catecholaminergic afferents in categorically different stress paradigms. Proc. Narl. Acad. Sci. USA, 93: 2359-2364. Li, H.-Y. and Sawchenko, P.E. (1998) Hypothalamic effector neurons and extended circuitries activated in ‘neurogenic’ stress: a comparison of footshock effects exerted acutely, chronically, and in animals with controlled glucocorticoid levels. J. Comp. Neurol., 393: 244-266. Li, Y.-W. and Dampney, R.A.L. (1994) Expression of fos-like protein in brain following sustained hypertension and

77 hypotension in conscious rabbits. Neuroscience, 61: 6 13-634. Matsumura, K., Cao, C., Ozaki, M., Morii, H., Nakadate, K. and Watanabe, Y. (1998) Brain endothelial cells express cyclooxygenase-2 during lipopolysaccharide-induced fever: light and electron microscopic immunocytochemical studies. J. Neurosci., 18: 6279-6289. McCarty, R. and Kopin, I.J. (1978) Sympatho-adrenal medullary activity and behavior during exposure to footshock stress: a comparison of seven rat strains. Physiol. Behav., 21: 567-572. Miselis, R.R., Shapiro, R.E. and Hand, P.J. (1979) Subfornical organ efferents to neural systems for control of body water. Science, 205: 1022-1025. Morgan, J.L. and Curran, T. (1991) Stimulus-transcription coupling in the nervous system: involvement of the inducible proto-oncogenes fos and jun. Ann. Rev. Neurosci., 14: 42 1-45 1. Owens, M.J. and Nemeroff, C.B. (1993) The role of corticotropin-releasing factor in the pathophysiology of affective and anxiety disorders: laboratory and clinical studies. In D.J. Chadwick, J. Marsh and K. Ackrill (Eds), Corticotropinreleasing Factos Ciba Foundation Symposium 172, Wiley, Chichester, pp. 296-3 16. Pezzone, M.A., Lee, W.-S., Hoffman, G.E., Pezzone, K.M. and Rabin, B.S. (1993) Activation of brainstem catecholaminergic neurons by conditioned and unconditioned aversive stimuli as revealed by c-fos immunoreactivity. Brain Rex, 608: 310-318. Pezzone, M.A., Lee, W.-S., Hoffman, G.E. and Rabin, B.S. (1992) Induction of c-fos immunoreactivity in the rat forebrain by conditioned and unconditioned aversive stimuli. Brain Rex, 597: 41-50. Ramsey, J.M. (1982) Basic Pathophysiology: Modern Stress and the Disease Process, Addison Wesley, Palo Alto, CA. Roland, B.L., and Sawchenko, P.E.(1993) Local origins of some GABAergic projections to the paraventricular and supraoptic nuclei of the hypothalamus in the rat. J. Comp. Neurol., 332: 123-143. Saper, C.B. (1995) Central autonomic system. In: G. Paxinos (Ed.), The Rat Nervous System, 2nd edn., Academic Press, San Diego, pp. 107-128. Sawchenko, P.E., Arias, C.A. and Mortrud, M. (1993) Evidence for synaptic mediation of chronic stress effects on corticotropin-releasing factor and vasopressin mRNAs in hypophysiotropic neurons. J. Neuroendocrinol., 5: 341-348. Sawchenko, P.E., Brown, E.R., Chan, R.K.W., Ericsson,A., Li, H.-Y., Roland, B.L. and Kovacs, K.J. (1996) The paraventricular nucleus of the hypothalamus and the functional neuroanatomy of visceromotor responses to stress. In G. Holstege, R. Bandler and C.B. Saper (Eds), Progress in Brain Research, The Emotional Motor System, Vol. 107, Amsterdam,pp. 201-222. Sawchenko, P.E. and Swanson, L.W. (1983) The organization of forebrain afferents to the paraventricular and supraoptic nuclei of the rat. J. Comp. Neurol., 218: 121-144.

Senba, E., Matsunaga, K., Tohyama, M. and Noguchi, K. (1993) Stress-induced c-fos expression in the rat brain: Activation mechanism of sympathetic pathway. Brain Res. Bull., 31: 329-344. Sims, J.E., Gin, J.G. and Dower, S.K. (1994) The two interleukin-1 receptors play different roles in IL-1 actions. Clin. Immunol. Immunopathol., 72: 9-14. Smith, D.W. and Day, T.A. (1994) C-fos expression in hypothalamic neurosecretory and brainstem catecholamine cells following noxious somatic stimuli. Neuroscience, 58: 765-775. Stemberg, E.M. and Wilder, R.L. (1993) Corticosteroids. In D.J. McCarthy and D.J. Koopman (Eds), Arthritis and Allied Conditions: A Textbook of Rheumatology, 12th edn., Lea & Febiger, Philadelphia, PA, pp. 665-682. Swanson, L.W. (1992) Brain Maps: Structure of the Rat Brain, Amsterdam, Elsevier. Swanson, L.W. and Kuypers, H.G.J.M. (1980) The paraventricular nucleus of the hypothalamus: cytoarchitectonic subdivisions and organization of projections to the pituitary, dorsal vagal complex, and spinal cord as demonstrated by retrograde fluorescence double-labeling methods. J. Comp. Neurol., 194: 555-570. Swanson, L.W. and Sawchenko, P.E. (1983) Hypothalamic integration: organization of the paraventricular and supraoptic nuclei. Ann. Rev. Neurosci., 6: 269-324. Swanson, L.W., Sawchenko, P.E., Wiegand, S.J. and Price, J.L. (1980) Separate neurons in the paraventricular nucleus project to the median eminence and to the medulla or spinal cord. Brain Res., 198: 190-195. Van Dam, A.-M., Brouns, M., Man-A-Hing, W., and Berkenbosch, F. (1993) Immunocytochemical detection of prostaglandin E, in microvasculature and in neurons of rat brain after administration of bacterial endotoxin. Brain Rex, 613: 331-336. Viau, V. and Sawchenko, P.E. (1995) The pattern of cellular activation seen in response to acture restraint suggests commonalties among “neurogenic” stress models. SOC. Neurosci. Abstr., 21: 1626. Wan, W., Wetmore, L., Sorensen, C.M., Greenberg, A.H. and Nance, D.M. (1994) Neural and biochemical mediators of endotoxin and stress-induced c-fos expression in the rat brain. Brain Res. Bull., 34: 7-14. Watanabe, T., Morimoto, A., Sakata, Y. and Murakami, N. (1990) ACTH response induced by interleukin-1 is mediated by CRF secretion stimulated by hypothalamic PGE. Enperientia, 46: 481434. Watkins, L.R., Maier, S.F. and Goehler, L.E. (1995) Cytokineto-brain communication: a review and analysis of alternative mechanisms. Life Sci., 57: 1011-1026. Wick, G., Hu, Y., Schwartz, S. and Kroemer, G. (1993) Immunoendocrine communication via the hypothalamopituitary-adrenal axis in autoimmune diseases. Endoc: Rev., 14: 539-563. Willis, W.D., Westlund, K.N. and Carlton, S.M. (1995) Pain. In: G. Paxinos (Ed.), The Rat Nervous System, 2nd edn., Academic Press, New York,pp. 725-750.

78 Wong, M.-L. and Licinio, J. (1994) Localization of interleukin 1 type I receptor mRNA in rat brain. Neuroimmunomodularion, 1: 110-115. Wong, M.L., Bongiorno, P.B., al-Shekhlee, A.. Esposito, A., Khatri, P.and Licino, J. (1996) IL-1beta, IL-1 receptor type

I and iNOS gene expression in rat brain vasculature and perivascular areas. NeuroReport, 7: 2445-2448. Yabuuchi, K., Minami, M., Katsumata, S. and Satoh, M. (1994) Localization of type I interleukin-1 receptor mRNA in the rat brain. Molec. Brain Rex, 27: 27-36.

SECTION IV

Early life experiences and the developing brain

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Brain Research, Vol 122 Q 2000 Elsevier Science BV. AU rights reserved.

CHAPTER 7

Long-term behavioral and neuroendocrine adaptations to adverse early experience Charlotte 0. Ladd’, Rebecca L. Huot’, K.V. Thrivikraman’, Charles B. Nemeroe, Michael J. Meaney3,and Paul M. Plotsky’,* Stress Neurobiology Laboratory. Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta, GA 30322, USA Laboratory of Neumpsychopharmacology, Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta, GA 30322, USA Developmental Neuroendocrinology Laboratory, Deparmtents of Psychiatry, and Neurology and Neuorsurgery, Douglas Hospital Research Centre, McGill University, Montreal, Que. H4H I R3, Quebec, Canada

’

Introduction Adaptation of an organism to the environment occurs through numerous processes beginning in the prenatal period and continuing through the neonatal and early adolescent period. Environmental signals, through processes such as activity dependent plasticity (Ben-Ari, 1995; Katz and Shatz, 1996; Kolb and Whishaw, 1998; Wheal et al., 1998), interact with the concurrently unfolding genetic blueprint for the central nervous system giving rise to a stable, individual phenotype governing perception of, and responsiveness to, salient features of the environment (Sroufe, 1997; Rutter et al., 1997). This process of adaptation, illustrated in Fig. 1, may be viewed as ‘fine-tuning’ or ‘environmental programming’ of neural circuitry. Thus, our studies have initially focused on a comprehensive description of the phenotype resulting from exposure to neonatal handling or maternal separation at the level of behavior, neuroendocrine responsiveness, and central nervous system circuitry. *Corresponding author. Tel.: 404-727-8258; Fax: 404-727-3233; e-mail: pplotsky @emory.edu

Seemingly beneficial adaptations in the shortterm, may, under challenging environmental conditions, actually be maladaptive over the life span of the individual. Increasing basic, clinical, and epidemiological evidence supports the thesis that exposure to an adverse early environment may underlie vulnerability to, and later expression of, physio- (Leserman et al., 1996) andor psychopathology (Ambelas, 1987; Brown et al., 1987, 1993; Kendler et al., 1993; DeBellis et al., 1994; Cui and Vaillant, 1996; Heim et al., 1997;Young et al., 1997). Among the most important early life influence is the interaction between the primary caregiver and the offspring. Along with a genetic influence on personality (Robinson et al., 1992; Cloninger et al., 1996; Jang et al., 1998; McGue and Bouchard, 1998), the development of attachment to the primary caregiver in primates holds a central position in the development of personality and psychopathology (Freud, 1966; Bowlby, 1977a, b). Harlow demonstrated this phenomenon dramatically during his two decades of research studying maternally deprived rhesus monkeys. Rhesus monkeys raised in partial isolation (i.e., raised with peers in the absence of the mother) for the first six months of life exhibited learning disabilities,

82

exaggerated oral behaviors, stereotypic movements, heightened fear and aggression, and an inability to cope with daily stressors (Chamove et al., 1973; Arling et al., 1976; Levine et al., 1993). Maternally separated primates also display psychomotor alterations (Spencer-Booth and Hinde, 1971). Neuroanatomical studies of monkeys separated using the methodology of Harlow revealed profound neuroanatomical changes (Martin et al., 1991). Furthermore, the offspring of monkey mothers in which it was necessary for them to devote considerable attention to foraging for food demonstrated increased anxiety and elevations in cerebrospinal fluid levels of corticotropin releasing factor (CRF) several years following this period (Coplan et al., 1996). Attempts to reverse the effects of early maternal deprivation were only partially successful; many monkeys were able to function normally under basal conditions but were

unable to cope with psychosocial stressors (Suomi et al., 1976). Kalin (Kalin et al., 1998; Kalin and Shelton, 1998; Kalin, Chapter 8, this volume) has studied factors associated with individual differences in infant rhesus monkeys, focusing on freezing behavior, vocalization, and cortisol. In both mothers and infants, basal cortisol levels were positively correlated with freezing duration. In contrast, the number of offspring a mother had was negatively correlated with her infant’s cortisol level. These findings suggest a mechanism by which maternal experience may affect infants’ cortisol levels. In humans, a childhood history of abuse, neglect, or trauma increases later susceptibility to affective and gastrointestinal disorders when compared to those adults who were reared in a nurturing environment (Kendler et al., 1993; Leserman et al., 1996; Heim et al., 1997). The expression of

D EVELO PM ENTA L< GENES 111111, ADAPTATION

ENVIRONMENT

n

SOCIAL BUFFERING

PHENOTYPE TRAUMA

A \

4

\

I(PUBERTY

VU LNERABILlTY

LONG-TERM (MAL)-ADAPTATION Fig. 1. Interaction between early experience and the genetic endowment leads to an individual phenotype. During the perinatal period such interactions may temporarily or permanently alter levels of gene expression thus initiating a cascade resulting in stable adaptations of numerous neurocircuits affecting behavior, neuroendocrine, and autonomic nervous system responsivity throughout life. During the perinatal period, social buffering may mute the effects of adverse early experience. Other factors including puberty and other adverse experiences throughout the life span, such as medical illness or major trauma, may exacerbate the underlying vulnerabilities resulting in expression of physio- or psycho-pathology. (Modified and used with permission from Plotsky et al., 1998).

83

these effects is partially regulated by the genetic endowment of the individual which appears to exert a substantial influence on personality as well as the level of ‘social buffering’ (Sroufe, 1997; Rutter et al., 1997). It is currently postulated that each individual shapes their environment as well as responds to it, thus providing a mechanism to amplify any modest genetic effects on behavior or personality (Scarr and McCartney, 1983). In lower vertebrates it is also clear that the primary care-giver plays an important role in the development of regulatory systems in the offspring (Hofer, 1973, 1984, 1994; Alberts, 1994). Ader, Denenberg, Levine, and others (Levine, 1957, 1975; Denenberg, 1964;Ader and Grota, 1969) first reported that early handling (i.e. infantile stimulation) reduced novelty-induced fearfulness and adrenal responsiveness to stressors. In contrast, maternal separation exerted the opposite effect on adrenal responsivity and, when assessed, fearfulness (Levine et al., 1991; Pihoker et al., 1993; Plotsky and Meaney, 1993; Suchecki et al., 1995; Vazquez et al., 1996; Ladd et al., 1996; Kehoe et al., 1998). Prenatal stress of the mother also leads to alterations in the behavioral and neuroendocrine responsiveness of the offspring (Weinstock, 1997). Furthermore, this response is not limited to rats, but appears to represent a general response to withdrawal of maternal care (Henessey, 1997). The lack of particular aspects of maternal care (i.e. contact, feeding, licking) appear to be the primary mediators of numerous developmental effects including neuroendocrine responsiveness, behavioral responsiveness, and growth (Evoniuk et al., 1979; Schanberg et al., 1984; Jans and Woodside, 1990; Suchecki et al., 1993; Liu et al., 1997). These classic studies demonstrated that the development of rudimentary adaptive responses to stressors could be influenced by postnatal environmental events.

The hypothalamic-pituitary-adrenal axis and the stress response How can something as seemingly simple as the quality of the primary caregiver-infant bond influence an individual’s ability to adapt to challenges later in life? Accumulating research suggests that the answer may lie in two primary processes:

activity-induced plasticity in response to environmental input (vide supra) and alterations in the regulation of the so-called ‘stress response’. Furthermore, these processes, themselves, may interact to strengthen the net effect on the individual. The mammalian response to any actual or perceived threat (i.e. stressor) has four branches: endocrine, autonomic, immunologic, and behavioral. The hypothalamic-pituitary-adrenal (HPA) axis represents a major counter-regulatory system activated upon exposure to many stressors. A schematic of the HPA axis and important neurocircuits modulating its activity is shown in Fig. 2. In response to a perceived stressor, external or internal cues of threat, pain, or failure of expectations are communicated to the hypothalamus via visceral and somatosensory input from several noradrenergic nuclei in the brainstem including the ventral tegmental area (VTA), nucleus of the solitary tract (NTS), and the locus coeruleus (LC); limbic input via the bed nucleus of the stria terminalis (BNST) including projections from the ventral hippocampus (vHP), amygdala (AMYG), and prefrontal cortex (PFC); internal regulatory input via the intrahypothalamic nuclei, and circumventricular organs such as the subfornical organ (Sawchenko et al., 1993; Cullinan et al., 1995). Interoceptive stimuli, such as internal hemorrhage, do not need to be experienced at the conscious level to elicit a stress response. Following activation of one or more of these neural pathways, corticotropin-releasing factor (CRF) stored in parvocellular neurons of the hypothalamic paraventricular nucleus (pPVN) is released from median eminence nerve terminals into the hypophysial portal circulation (Merchenthaler et al., 1982; Swanson et al., 1983; Plotsky, 1991). By virtue of its role as a hypothalamic-hypophysiotropic hormone, CRF stimulates the synthesis and release of adrenocorticotrophin hormone (ACTH) from the anterior pituitary gland. ACTH, in turn, stimulates the production and release of glucocorticoids, the end products of the HPA axis. The glucocorticoids, cortisol in primates and corticosterone in rodents, mobilize energy substrates during stress, and synergize with catecholamines released from the adrenal medulla and sympathetic nerve terminals to elevate circulating glucose levels as

84

Fig. 2. Scheme illustrating the hypothalamic-pituitary-adrenal (HPA) axis. Viscerosensory and emotional stimuli are encoded within the central nervous system and funnel to the hypothalamic pPVN, adjacent to the third ventricle (3V), where they activate secretion of corticotropin releasing factor (CRF) and arginine vasopressin (AVP) from nerve terminals in the external zone of the median eminence (ME) ending in the penvascular spaces of the primary portal capillary plexus. These peptides enter the hypophysial portal vascular system via fenestrations in the walls of these capillaries and are carried by the long portal vessels to the anterior pituitary gland where they diffuse from capillaries of the secondary portal plexus to act at corticotropes. CRF stimulates transcription of the propiomelanocortin (POMC) gene, providing the precursor peptide for adrenocorticotropin (ACTH). Both CRF and AVP facilitate the secretion of stored ACTH into the systemic circulation. ACTH binds to its membrane receptor on adrenocortical cells resulting in the de novo biosynthesis and release of glucocorticoids (corticosterone in the rat, cortisol in primates) which then acts throughout the organism via cyoplasmic receptors. The circulating glucocorticoids also complete a negative feedback loop to damp ongoing and subsequent activity of the HPA axis through actions at cytoplasmic glucocorticoid receptors (GR) distributed in pituitary corticotropes, the hypothalamic PVN, the hippocampus (HP), and numerous other regions within the central nervous system. Input from regions such as the brainstem and corticolimbic regions modulates HPA axis activity (see Fig. 4 for schematic of this portion of the neurocircuitry).

85

well as to increase the heart rate and blood pressure. Prolonged elevation of circulating glucocorticoids has negative physiological and psychological consequences (Munck et al., 1984; Magarinos et al., 1997; McEwen, 1998). The levels of circulating glucocorticoids are tightly controlled by the process of glucocorticoidmediated negative feedback (Dallman et al., 1987; Jacobson and Sapolsky, 1991). This process occurs via ligand-induced activation of cytoplasmic mineralocorticoid receptors (MR) localized in the hippocampus and septum and/or activation of cytoplasmic glucocorticoid receptors (GR) which are widely distributed throughout the CNS as well as in the adenohypophysis (DeKloet, 1991; Rosenfeld et al., 1993). Both M R and GR are transcriptional regulators which, upon activation, translocate to the nucleus where they bind in homoor heter-dimeric form to DNA recognition sites to alter levels of gene expression. Research on glucocorticoid binding and actions also provides evidence for the possibility of specific glucocorticoid receptors localized to the plasma membrane of cells or of glucocorticoid binding sites on other neurotransmitter receptors (Orchinik et al., 1991; Sze and Towle, 1993; Orchinik and McEwen, 1994; Joels, 1997). These observations may provide an explanation for the rapid actions of glucocorticoids on behavior (Orchinik, 1998). Interestingly, there is considerable individual variability in the regulation of the HPA axis, particularly with respect to rearing conditions (Shoenfeld et al., 1980; Sapolsky et al., 1986; Levine et al., 1991). This is an adaptive process which is essential for survival in adulthood, especially during periods of extreme stressors, but may be maladaptive in infancy while the central nervous system is still developing (Bohn, 1980, 1984; Meaney et al., 1993). In fact, under normal circumstances, at least in the rat, the developing central nervous system is partially protected from catabolic and other effects of glucocorticoids due to reduced sensitivity of the axis to activation during a substantial portion (PND4-14) of the neonatal period - an epoch referred to as the stress hyporesponsive period or SHRP (Sapolsky and Meaney, 1986; Walker et al., 1986; Van Oers et al., 1998). This period may, in part, be essential to

permit normal brain development (Gould et al., 199la, b; McEwen, 1995). The hyporesponsiveness appears to be maintained by internal factors (i.e. neurodevelopmental status) as well as social buffering in the form of maternal cues including feeding and licking (Suchecki et al., 1993). However, it is worth noting that this reduced responsiveness is not absolute and may be circumvented in the presence of strong psychological or physical stressors. Maternal behavior in the rat consists of several rituals associated with feeding that are vital for proper infant development. During each nursing bout, the dam may assume an arched-back posture to encouraging suckling or a blanket posture which is less conducive to suckling. Each nursing bout concludes with anogenital licking/grooming to stimulate urination and defecation. Studies have shown that, in the rat, feeding (suckling) stimulates the infant heart rate (Hofer, 1973; Hofer, 1984), while stroking (licking) stimulates the secretion of growth hormone and expression of ornithine decarboxylase (Evoniuk et al., 1979). Suchecki and colleagues (1993) reported that these maternal behaviors suppress the HPA axis at different levels, thus preventing a pup from mounting an adult-like HPA response to a moderate stressor. In the neonatal rat, feeding desensitizes the adrenal gland to circulating ACTH, while stroking inhibits the release of ACTH from the anterior pituitary. These observations explain why maternal separation is one of the few stressors that can induce a substantial HPA response in rat pups during the SHRP. Thus, when the pups are removed from the home cage, the pituitary secretes ACTH, stimulating corticosterone release from, and hypertrophy of, the adrenal cortex, which then loses its refractoriness to ACTH.

A rodent model of adverse early experience Given the aforementioned data supporting a preeminent role for maternal-infant interactions in the development of coping strategies in sub-human primates and humans, coupled with observations that maternal separation in rodents is capable of modifying both HPA axis and behavioral responses in rodents, we hypothesized that early maternal separation in rodents might serve as a suitable model of adverse early experience. In collaboration

86

with Michael Meaney’s group (Meaney et al., 1993), we designed a neonatal maternal separation protocol which mimicked patterns of separation observed in the wild. Rodent dams often leave the nest to forage for periods of 15-30 min periods (Leon et al., 1978; Jans and Woodside, 1990), although low ranking females may have to leave the nest for longer periods of time due to the greater distance of their burrows from sources of food and water. For the studies described below, we used outbred, timed-pregnant Long Evans hooded rats (Charles River, Boston, MA). On postnatal day (PND) 2, all pups were removed from their home cages, randomized, and culled to 8-10 male pups per dam. Each litter was then exposed to one of three rearing conditions from PND2 to 14: (a) animal facility rearing (AFR) which comprised home cage bedding material changes and brief handling twice weekly beginning on PND5 with no other handling or separation; (b) handled (HMS15) animals which were removed from the home cage daily for 15 min periods; and (c) maternal separation (HMS180) in which pups were removed from the home cage for 180 min daily. These manipulations occurred between 0800-1300 hours. In some instances a non-handled group was run: these animals were not handled for any purpose nor were their cages disturbed from PND3- 14. Prior to manipulation of the HMS15 and HMS 180 pups, when the dam was off her pups, she was removed to an adjacent cage where she was housed for the duration of the daily separation period. During the separation period, rat litters of the HMS180 and HMS15 groups were housed in individual cages placed into a temperature-controlled incubator. At the end of each separation period, both HMS15 and HMS180 pups were returned to their home cage, rolled in the soiled bedding, and then reunited with their dam. No manipulation of the rats occurred other than that described above. Litters were weaned on PND2223 and housed 2-3 per cage until adulthood ( > PND60).

Neonatal handling and separation effects on maternal behavior Similar to previous reports (Leon et al., 1978; Jans and Woodside, 1990); brief disturbance of the dam

and litter resulted in increased maternal behavior. Upon reunion of the dam with her litter after brief separation, the rate of maternal licking and grooming was increased in the HMS15 animals as compared to the AFR group; in contrast, maternal separation (HMS 180) was associated with reduced and deranged maternal behavior (Huot et al., 1997). Specifically, dams in the HMS15 paradigm retrieved their pups with a shorter latency upon reunion and engaged in more frequent bouts of nursing, in effect increasing the total time spent licking and grooming the pups, but not the total time feeding. This increase in licking and grooming, coupled with the observed increase in arched back nursing in these dams, appeared to be largely responsible for the long-term resistance to stress in their offspring as cross-fostering of HMS15 pups to HMS 180 dams and vice versa blunted the effects of the particular pup rearing experience (Plotsky, unpublished studies). In support of this hypothesis, our group has demonstrated an inverse correlation between the amount of time spent licking, grooming, and arched-back nursing and subsequent adult stress responsiveness (Liu et al., 1997). In short, it appears that the quality of maternal behavior encountered upon reunion played a large role in determining individual differences in stress responsiveness. In a continuing series of studies, we also examined these adult animals for rearing-associated differences in anxiety- or fear-like behavior, anhedonic behavior, alcohol preference, neuroendocrine stress responsiveness, and various neurotransmitter circuits which facilitate or inhibit the stress response and are associated with mood disorders in the CNS.

Neuroendocrine adaptations in response to maternal separation The original observations of Ader, Denenberg, Levine, and others (Levine, 1957, 1975; Denenberg, 1964; Ader and Grota, 1969) were replicated and extended by collaborative studies of Meaney and Plotsky (Viau et al., 1993; Plotsky and Meaney, 1993; Bhatnagar et al., 1995) which focused on the central mechanisms underlying differences in HPA responsiveness to stressors. We first observed that

87

handled (HMS15) rats exhibited lower CRF and AVP concentrations in the hypophysial-portal circulation which was accompanied by decreased median eminence CRF and AVP peptide content as well as decreased hypothalamic CRF mRNA (Plotsky and Meaney, 1993). Even under restingstate conditions, hypothalamic CRF and AVP synthesis was lower in the handled animals, a difference that occurred in the presence of basal glucocorticoid levels (Viau et al., 1993). These differences in CRF/AVP synthesis and ACTH responses to stressors were found to be associated with altered glucocorticoid negative feedback sensitivity. Administration of either corticosterone or dexamethasone suppressed plasma ACTH responses to stressors more effectively in handled than in non-handled animals (Meaney et al., 1989). Furthermore, HMS 180 rats displayed impaired glucocorticoid-mediated negative feedback in response to a dexamethasone suppression test as compared to either AFX or HMS 15 rats (Plotsky, in preparation). Adrenalectomy, which removed the glucocorticoid negative feedback signal, completely eliminated the differences in plasma ACTH responses to stress while replacement with low levels of corticosterone significantly decreased adrenalectomy-induced ACTH hypersecretion during stress only in the handled animals (Viau et al., 1993). Interestingly, in comparison to non-handled rats, HMS 15 animals displayed increased glucocorticoid receptor binding capacity (reviewed in Meaney et al., 1996) in hippocampus and frontal cortex, increased glucocorticoid receptor protein levels, and increased steady-state levels of glucocorticoid receptor mRNA throughout the hippocampus as compared to non-handled rats. In contrast to AFR or HMS15 rats, maternally separated (HMS180) rats displayed a down regulation of the hippocampal GR system and upregulation of the M R system (Ladd et al., 1998). These changes were highly localized as no differences in GR or MR expression were apparent in septum, amygdala, hypothalamus, or pituitary. It seems logical that this decreased GR receptor density in the HMS 180 rats would attenuate the sensitivity of the hippocampus and frontal cortex to circulating glucocorticoids, decreasing the efficacy of negative-feedback inhibi-

tion over HPA activity, thus serving to increase CRF and AVP synthesis in the hypothalamic pPVN as well as stress-induced release from the median eminence nerve terminals of this system. The enhanced stress-induced glucocorticoid response may then increase expression of CRF mRNA in the central nucleus of the amygdala (Makino et al., 1994; Hsu et al., 1998). In fact, basal levels of CRF mRNA in the pPVN, central nucleus of the amygdala, and bed nucleus of the stria terminalis were found to be elevated in adult HMS 180 rats as compared to AFR or HMS15 animals and these elevations were mirrored by increased CRF peptide content in terminal fields of these neurons (Plotsky et al., in press). Although both basal trough levels of circulating ACTH and corticosterone were normal in adults of all rearing conditions, the response to a psychological stressor such as airpuff startle (Engelmann et al., 1996), but not a physical stressor such as controlled hemorrhage was greatly accentuated in the HMS180 rats as shown in Fig. 3 (Plotsky et al., submitted). Surveying stressors it appeared that psychological stressors (e.g. novel environment, airpuff startle, restraint), but not physical stressors (e.g. hemorrhage, nitroprusside-induced hypotension, cold) elicited hyperactivity of the HPA axis suggesting alterations in the corticolimbic pathways involved in the processing of this information and its transfer to the hypothalamic pPVN. The marked increase in the ACTH response to psychological stressors in the HMS180 animals was followed by an enhanced and prolonged release of corticosterone from the adrenal cortex, which is sensitive to both absolute and integrated concentrations of ACTH. No differences were found in metabolism of corticosterone or concentrations of corticosterone binding globulin between the handled and maternally separated groups (Rosenfeld et al., 1992). These data suggest that early environmental factors exerted a significant impact on the developing HPA axis resulting in persistent alterations in the adult organism’s responsiveness to psychological stressors. Furthermore, these observations suggest that adverse early events may alter the function of pathways proximal to the hypothalamic CRF system (vide infra). Mechanisms underlying these stable changes in CRF systems

88

and in localized GR expression resulting from neonatal maternal separation remain to be determined. In a series of studies examining the regulation of glucocorticoid negative feedback in HMS 15 vs. non-handled animals, Meaney's group accrued in vitro and in vivo evidence supporting a neurochemical cascade underlying changes in hippocampal GR expression involving thyroxin (T3), serotonin (5-HT), and numerous transcription factors (Francis et al., 1996). These studies demonstrate at least one possible scenario that may lead to differential regulation of the HPA axis secondary to early rearing environment. Whether these same systems are involved in mediating the opposite effects of maternal separation or other adverse early events remains untested.

Behavioral adaptations in response to maternal separation Anhedonia

Most organisms seek pleasurable stimuli especially in the forms of sex or highly palatable food. Anhedonia, which is characteristic of major depressive disorder, is the inability to perceive pleasure and the lack of effort to seek pleasure. In rats, pleasure seeking is often measured as the quantity of sucrose or saccharine solution that they will ingest within some defined period of time (Willner et al., 1987; Muscat and Willner, 1992). Chronic mild stress has been shown to reduce the amount of sweetened solution ingested by rats, an effect reversed with antidepressant administration (Willner et al., 1987).

B

A 400

800

300

600

az

200

1n

100

F.

22

400

2

200

v ) U

n

0

0 AFR HMSl5 *HIUS180

I

V

Q)

:=

300

.-0O 0C

2oo

-I--

;

100

0

400

.-

0

-10

0

lo

20

30

Time (min)

40

I

al

/ -.-.

AFR H MSl5 * H M S ~80 0

0

400

ZE

I

V

0-

O n

:2

HEM

--. -.-.

5'0

60

;-5=E

300

gs

200

;

100

.=

V

OY c

0

0 -10

0

10

20

30

40

50

60

Time (min)

Fig. 3. Rearing-associated differences in adult responsiveness to psychological (A) and physical stressors (B). In the maternally separated adults (HMS180),the ACTH and corticosterone responses to the psychological stressor of airpuff startle (APS) was enhanced relative to the responses in animal facility reared (AFR) colony controls or handling controls (HMSIS). In contrast, no Data difference in responsiveness was evident among the groups in response to the physical stressor of 15% hemorrhage (HEM). shown as mean f sem. Further details may be found in the text.

89

Assessment of anhedonia was recently performed in adult animals previously exposed to different rearing conditions (vide supra). In the initial set of experiments, animals were presented with a home cage two-bottle choice with one bottle containing tap water and the other containing a 1% sucrose solution in water. Intake from each bottle was recorded over a 24-hour period on three separate occasions. The maternally separated rats drank approximately 35% less sucrose solution as compared to the AFR and HMS15 groups although neither total fluid intake or body weight differed among the groups. Another test was run in which overnight food-deprived rats were presented with a saccharin (0.02%) solution for a period of 60 min. Similar results were obtained indicating anhedonia in the HMS 180 rats. Anxiety- or fear-like behavior On the basis of previous studies (Denenberg, 1964; Levine, 1975; Arling et al., 1976), we postulated that adult rats previously exposed to varying degrees of neonatal handling - maternal separation would exhibit individual differences in noveltyinduced fearfulness, with HMS 180 and HMS 15 rats representing the high and low ends of the spectrum, respectively. Behavior was assessed using several behavioral tests of anxiety, including defensive withdrawal, elevated plus maze, open field exploration, acoustic startle reflex and novelty-induced suppression of appetitive behavior. In the defensive withdrawal test (Takahashi et al., 1989), each rat was coaxed into an enclosed dark cylinder which was closed at one end. The cylinder was then placed in a brightly lit, novel open field. The latency to exit the dark cylinder into the open field, the number of transits between these compartments, and the total time spent in each compartment were monitored as indices of locomotor activity and of anxiety-like behavior (Blanchard et al., 1974). In a 15 min defensive withdrawal test, adult HMS180 rats exhibited a greater latency to enter the open field compared to either the HMS 15 or AFR groups, a greater amount of time in the dark cylinder, and increased freezing in the open field (Huot et al., 1997; Plotsky et al., submitted). These observations suggested that maternally separated

rats expressed greater anxiety or fearfulness in a novel environment compared to handled (HMS 15) and normally reared (AFR) rats. Interestingly, nonhandled rats exhibited anxiety-like behavior similar to the HMS 180 animals. Similar results were found in an open-field test of exploration. Animals were placed in a novel circular open field (1.6 m in diameter) and observed for 10 min by an observer blinded to the rearing condition of the animal. HMS15 animals spent 2-4 times as long exploring the inner arena ( > 10 cm from the wall) than did the HMS180 rats, suggesting that the latter group experienced heightened fearfulness in the novel arena (Caldji et al., in press). The elevated plus maze represents a classical test of anxiety or fearfulness (Pellow et al., 1985). Rats placed at the central point of the plus maze were scored on the amount of time spent in the open vs. the closed arms of the apparatus as well as the number of transits among compartments. During a 5-min test, the HMS 180 rats spent less time in the open arms as compared to either HMS15 or AFR rats. Interestingly, the HMS 180 rats often simply remained at the choice point; displaying profound freezing behavior or they jumped off the 50 cm high apparatus and headed for a dark comer of the testing room. This behavior was consistent with the interpretation of enhanced anxiety or fearfulness in the HMS 180 animals. We also assessed the acoustic startle reflex (ASR) in differentially reared animals as adults. In this test (Davis, 1986; Blaszczyk and Tajchert, 1997), rats were placed in a plastic cage within a sound-attenuatedchamber and, after acclimation to a 60 dB white noise for 5 min, rats were subjected to five repetitions (30 msec each) of an 80-120 dB tone with an inter-stimulus interval averaging 15 s. A piezoelectric strain platform beneath the cage floor detected the startle in response to each tone. We observed a significant startle effect in all animals; the HMS180 rats exhibited much greater startle responses than did either HMS15 or AFR animals to tones in the 110-120 dB range (Plotsky, unpublished studies; Caldji et al., in press). No individual differences were observed between any of the groups in the 80-1 10 dB range. These results demonstrated that HMS 180 rats had a greater reflex

90

to strong sensory stimuli than did HMS15 or AFR rats. Finally, we determined the performance of these animals in a test of novelty-induced suppression of appetitive behavior (Britton and Britton-Thatcher, 1981). In this test, animals were food-deprived for 24 hours prior to testing, then presented with food for a period of 6 min either in the center a novel open arena or in their home cage. Several parameters of fearful behavior were measured, including the time spent eating, the latency to approach the food, and the latency to feed. Compared to the HMS15 animals, HMS180 rats exhibited a greater latency to approach and eat the food and spent less time eating (Caldji et al., in press). Collectively, these observations (see Table 1) provide compelling support for the thesis that neonatal maternal separation profoundly influences anxiety- or fearfulness-like behavior which lasts to adulthood. Thus, these data imply that experiences encountered during some critical period of development, particularly those inherent in the caregiver-child interaction, significantly affect individual perception (e.g. threatening, neutral, stimulating) of the environment and responses to it throughout life. We postulate that this process arises through a combination of stimulus-induced plasticity during development, activation of extrahypothalamic CRF neurocircuits, and glucocorticoid effects mediated through a cascade of transcriptional actions throughout critical sensory and limbic neurocircuits. TABLE 1

Summary of rearing-associated differences in anxiety-like behavior _____

~~

~

~~

Test

Variable

Defensive withdrawal Open field locomotion Elevated plus maze Acoustic startle Conflict test

Latency to exit tube Time in center Time in open arms Startle response Latency to eat

HMSl5 HMSl80 vsAFR vs AFR

+

+ +

V V

AAA VVV VVV AAA AAA

Details of behavioral tests can be found in the text. AFR = animal facility reared colony controls; HMS15 = neonatal handlingmaternal separation 15 minute handling controls; HMSllO = neonatal handling-maternal separation 180 minute; =no change; V = decrease; A =increase.

+

Central nervous system adaptations to maternal separation Corticotropin-releasingfactor (CRF)

CRF is most well known as the obligatory hypothalamic ACTH secretagogue (Plotsky, 1991); however, it is also a putative neurotransmitter within the CNS (Dunn and Berridge, 1990; Owens and Nemeroff, 1991). CRF-containing neurons and CRF receptors are widely distributed in cortical, limbic, and brainstem regions (De Souza et al., 1985; Sawchenko et al., 1993). Central CRF neurocircuits, specifically the hypothalamic- and amygdala-brainstem projections, have been implicated in the behavioral and autonomic expression of stress and fear (Butler et al., 1990; Cador et al., 1992; Fisher, 1993; Heinrichs et al., 1995; Lee and Davis, 1997). Increasing anatomical and functional evidence suggests that the amygdala may play a pivotal role in central CRF neurotransmission. Intra-amygdala CRF neurotransmissionis tightly regulated by CRF binding protein, CRF receptors, and GR, all expressed in either the central nucleus or peripheral nuclei (Makino et al., 1994; Gray and Bingaman, 1996). Activation of the amygdala is modulated via peptidergic and aminergic projections from numerous cortical, limbic and midbrain structures, which, in turn,receive reciprocal CRF projections from the central nucleus of the amygdala (Gray and Bingaman, 1996). For example, the hypothalamus receives both monosynaptic and polysynaptic projections from the central nucleus of the amygdala; the polysynaptic projections travel via the BNST and brainstem nuclei. These neurocircuits, some of which contain CRF, stimulate the HPA axis (Gray, 1990; Gabr et al., 1995; Marcilhac and Siaud, 1997; Feldman and Weidenfeld, 1998). Similarly, the amygdala also appears to mediate fearful behavior via CRF projections to brainstem nuclei, particularly the LC. In various behavioral tests, activation of amygdala CRF neurocircuits appears to mediate the acoustic startle reflex (Lee and Davis, 1997), stress-induced freezing (Swiergiel et al., 1993), and stress-induced emotionality in an elevated plus maze (Heinrichs et al., 1992). It is interesting to note that, unlike in the pPVN, GR activation in the amygdala enhances CRF mRNA

91

expression (Makino et al., 1994). Thus, repeated elevations in plasma corticosterone could potentially increase CRF expression within the amygdala. The locus coeruleus (LC) is the largest of seven brainstem nuclei from which all noradrenergic projections in the brain radiate. Ultrastructural studies have demonstrated that CRF terminals synapse onto catecholamine dendrites in the rostral LC (Van Bockstaele et al., 1996, 1998), supporting previous evidence that acute stressors sequentially activate CRF and noradrenergic neurotransmission in this region (Valentino et al., 1993, 1998). Recent studies have shown that previous exposure to stress can modulate CRF neurotransmission in the LC, as can chronic antidepressant treatment (Curtis et al., 1994, 1995). In addition, defensive withdrawal can be attenuated by local infusion of a CRF antagonist into the LC (Smagin et al., 1996). Given the apparent involvement of CRF neurocircuits in the central nucleus of the amygdala and the LC in the expression of fear, coupled with the observed increase in novelty-induced fear in HMS180 versus HMS 15 animals, we hypothesized that neonatal maternal separation would increase expression of CRF and CRF receptors in the central nucleus of the amygdala and LC (Plotsky et al., submitted). Using in situ hybridization, we observed a two-fold increase in basal concentrations of CRF mRNA in the pPVN of HMS180 animals compared to the HMS 15 and AFR rearing groups. In addition, basal pPVN CRF mRNA concentrations in the non-handled group were significantly greater than in the HMS15 and AFR animals. The HMS180 rats also exhibited an increase in CRF mRNA concentration in the central nucleus of the amygdala compared to all other groups, but there were no other differences between groups in this region. There were no significant differences in CRF mRNA expression in the LC although CRF peptide content was increased in this region in the HMS180 animals. These data support previous observations in our laboratory that maternal separation increased CRF mRNA expression in the pPVN and central nucleus of the amygdala (Plotsky and Meaney, 1993, 1996). In this same study, we also examined CRF receptor binding in the cortex, the lateral and

central nuclei of the amygdala, the pPVN, and the LC. We did not observe any rearing effects for CRF receptor binding in the amygdaloid nuclei or in the cortex. However, in both the pPVN and the LC, CRF receptor binding was increased more than 4-fold in both the HMS180 and non-handled animals compared to the AFR and HMS15 groups. These data suggest that differential rearing conditions can alter the sensitivity of discrete nuclei to CRF. Thus, the pPVN and the LC in both nonhandled and HMS180 animals would be expected to be more sensitive to extracellular CRF because these regions have a 400% greater density of CRF receptors than do the central nucleus of the amygdala and LC in HMS15 and AFR animals; additionally, these areas may receive increased CRF signaling as well. Anatomical and functional evidence suggests that the expression of fearfulness is at least partly mediated by CRF projections from the central nucleus of the amygdala to the LC (Gray et al., 1990; Koegler-Muly et al., 1993). In experiments investigating the long-term effects of early rearing on central CRF neurocircuits, we have found that HMS180 and, to a lesser extent, non-handled animals, exhibited evidence of increased in CRF neurotransmission in the pPVN, central nucleus of the amygdala, and LC. Thus, it is likely that the CRF neurocircuits connecting these three regions are sensitive to the effects of early experience. CRF stimulation at the level of the LC and the pituitary could theoretically subserve the characteristic enhanced fearfulness and HPA activation, respectively, in the HMS180 animals. These results divulge a possible mechanism underlying the individual differences in adult stress responsiveness due to early rearing environment. Noradrenaline (NA)

A major set of afferents to the pPVN arises from the brainstem nuclei that are predominantly noradrenergic, including the ventral tegmental area or A1 (70%), the nucleus of the solitary tract or A2 (20%), and the LC or A6 (10%). These nuclei are activated following visceral and somatosensory stressors, such as restraint or hemorrhage. An increase in NA release into the pPVN following

92

such a stressor induces the release and synthesis of CRF, thus stimulating the HPA axis (Plotsky et al., 1989). Given the observed adaptations in CRF regulation in the LC with respect to differential rearing environments coupled with the exaggerated behavioral and humoral stress responses in maternally-separated animals, we hypothesized that HMS 180 animals would exhibit heightened NA release from brainstem afferents innervating the pPVN. In a series of experiments (Liu et al., in press), we investigated the long-term effects of early experience on the brainstem-pPVN circuit by measuring NA release in pPVN during and following a 60 min restraint stressor as well as assessing a 1- and 1x2-adrenergic receptor concentrations in the pPVN, LC, and nucleus of the solitary tract. The al-receptor is a postsynaptic facilitatory receptor at the level of the pPVN, while the a2receptor may function as either a post-synaptic receptor or a presynaptic autorecptor which serves to inhibit the activity of the presynaptic neuron. Plasma ACTH concentrations were also measured to correlate hypothalamic NA release with HPA axis activation following initiation of restraint. In these studies, we observed a marked increase in pPVN NA levels in response to acute restraint in all rearing groups. However, the response in the HMS 180 rats was significantly greater than that observed in the other rearing groups. Furthermore, pre-restraint levels of NA were greater in the HMS180 rats relative to the other groups. This elevation in NA corresponded to a similar increase in the integrated ACTH response to restraint in HMS180 versus HMS15, AFR, and non-handled animals suggesting that NA afferents to the hypothalamus w&re at least partially responsible for activating the HPA axis during restraint stress. Additionally, HMS 180 and non-handled rats exhibited a significant decrease in 1x2-receptor concentration in both the LC and NTS compared to HMS15 or AFR animals. There were no rearingassociated changes in 1x2-receptor density in the pPVN or in al-receptor density in any region examined. These results highlight another mechanism potentially accounting for the rearing-associated differences observed in stressor responsiveness in

our model. The observed decrease in a2-adrenergic receptor density in the LC and NTS suggests that noradrenergic feedback in these critical circuits is impaired leading to greater and more prolonged NA release at the terminal fields of these neurons resulting in increased drive to CRF neurons in the pPVN and, thus, enhanced ACTH and corticosterone secretion in response to a stressor. Gamma-aminobutyric acid (GABA) and benzodiazepine (BZ) receptors

Recently, we have turned our focus to the possible involvement of GABAergic neurocircuits in mediation of the maternal separation phenotype. GABAergic inhibitory circuits are widely distributed throughout the brain with particular concentrations in the central and lateral amygdala, the cortex, and brainstem nuclei (De Blas, 1996; Wilson, 1996). Signaling occurs via postsynaptic GABAA- and GABA,-receptors; the multimeric GABAA-receptors also contain a benzodiazepine binding site which, upon binding endogenous BZ, enhances chloride conductance of the GABAA receptor complex (Luddens and Korpi, 1995; De Blas, 1996; Clement, 1996). Changes in local GABAergic systems may at least partially mediate the behavioral and neuroendocrine consequences of maternal separation as it has been shown that: (1) GAl3A systems modulate fearful behavior in novel situations (Bodnoff et al., 1989; File, 1995;Wilson, 1996); (2) GABA systems modulate the stress responses to psychosocial but not physical stressors (Malizia et al., 1995; Bhatnagar et al., 1995); (3) benzodiazepines are thought to mediate their effects at least in part through inhibition of CRF neurocircuits in the central nucleus of the amygdala (Owens et al., 1991); and (4)environmental stressors can modulate the BZ binding site (Deutsch et al., 1994). Two mechanisms by which maternal separation may modulate GABAergic tone include changes in the expression or subunit composition of the GABAA/BZ-receptor. This idea is supported by our experimental observations that handled rats show greater habituation of plasma ACTH responses to repeated stressors than do non-handled rats (Bhatnagar et al.,

93

1995) and that the handled rats exhibit increased BZ receptor binding capacity as compared to nonhandled animals (Bodnoff et al., 1987). Thus, in addition to glucocorticoid inhibitory tone, there appears to be a second source of inhibitory regulation to the HPA axis in response to psychological stressors which may differ among rearing groups. One central question concerns the critical target(s) of this inhibitory influence. A primary candidate is the ascending NAergic projection system to the pPVN. Noradrenergic input to the pPVN is at least partially responsible for providing drive to CRFIAVP release during many types of stressors (Plotsky et al., 1989). Moreover, stressinduced increases in hypothalamic NA activity are attenuated by glucocorticoids (Pacak et al., 1993, 1995) and by BZ receptor agonists (Wilson, 1996) suggesting that these multiple inhibitory systems may share a common target. Initially, we assessed GABAA and BZ receptor binding in adult rats using [3H]muscimol or [3H]flunitrazepamfor autoradiography. The results (Caldji et al., in press) revealed clear effects of early environmental manipulations. Differences in GABA, and BZ receptor binding were apparent in the amygdala as well as in NAergic cell body regions. HMS 180 and non-handled animals showed significant reductions in GABAA receptor density in the LC and nucleus of the solitary tract, as well as significantly decreased BZ receptor density in the lateral and central nuclei of the amygdala, the LC, and NTS. We observed no differences in GABA, or BZ receptor binding density in the hypothalamus, frontal cortex or hippocampus, thus demonstrating that the effect of maternal separation on GABAergic receptor systems was anatomically quite distinct from that on GR systems (vide supra). In addition, expression of the 72 subunit of the GABAA receptor complex, which confers BZ binding activity, was reduced in HMS180 vs. HMS15 or AFR rats. These findings provide a putative mechanism for decreased GABAergic inhibition of activity in ascending NAergic neurons and dovetail with our recent findings on the amygdaloid CRF system, which is a potent regulator of ascending NAergic systems. Nemeroff’s group (Owens et al., 1991) reported evidence for BZ inhibition of CRF expression in the amygdala,

notably in the CRFergic projection from the central nucleus of the amygdala to the LC and in recent experiments, we (Plotsky et al., submitted) observed increased CRF peptide content in the LC region and increased CRF mFWA expression in the amygdala in non-handled and maternally separation animals. Together with the GABAA and BZ receptor binding differences, these findings form the basis for our focus on the amygdaloid nuclei in mediating the effects of maternal separation on HPA and behavioral responsiveness to stressors.

Adverse early experience as a vulnerability factor in depression-likesyndrome The stress diathesis model postulates the interaction between a genetic vulnerability or predisposition and adverse life events in the genesis of major depressive disorder. Considerable research supports the contribution of adverse early experience and/or exposure to a major trauma as precipitating factors in the onset of major depression (Dunner et al., 1979; Anisman and Zacharko, 1982; Ambelas, 1987; Brown et al., 1987; Nemeroff, 1991; Heim et al., 1997). While many theories of the primary defect leading to the onset of depression have been offered (Duman et al., 1997), much research in the current decade has been focussed on two theories: dysfunction of the central glucocorticoid receptor system (Holsboer et al., 1994, 1995) and dysregulation of central CRF systems (Nemeroff, 1996; Heit et al., 1997). These theories, of course, are not mutually exclusive. The underlying etiology and pathophysiological adaptations in the central nervous system occurring during depression have been difficult to elucidate due to the lack of appropriate laboratory animal models (Kessler et al., 1994). Willner (1995) offered multiple criteria for the validation of animal models of depression including face and construct validity. Unfortunately, several of the proposed criteria require a priori knowledge of the etiology of the disease and, thus, cannot be fulfilled by any model. The chronic mild stress (CMS) model, consisting of daily exposure of adult rats to a variety of stressors over a prolonged period of weeks, has shown success in replicating much of the symptomology of depression and these effects

94

can be reversed by antidepressant treatment (Papp et al., 1996; Willner, 1997). The model has good predictive validity, face validity, and construct validity; however, the duration of effects is variable and the model lacks a genetic component. Pucilowski and colleagues (1993) applied CMS to the hypercholinergic Flinders Sensitive Line (FSL) of rats, a putative genetic animal model of depression, and found that stress-induced anhedonia was increased in the FSL vs. the control Flinders Resistant Line (FRL)rats. On the basis of our studies, we believe that the neonatal maternally separated rat provides a suitable model of at least a vulnerability to the development of a depression-like syndrome. These animals exhibit dysregulation of the HPA axis including CRF hypersecretion and dexamethasonemediated negative feedback resistance, enhanced anxiety-like behavior, and anhedonia. Furthermore, many of the neurocircuits postulated to mediate the pathophysiology observed in major depressive disorder exhibit stable changes in function in the adult HMS 180 animal. Finally, chronic treatment of these adult animals with antidepressants at least partially reverses all of the dysfunctions thus far observed. Many of the symptoms observed in major depressive disorder and in animal models can be elicited by central administration of exogenous CRF, a neuropeptide which coordinates the mammalian endocrine, autonomic, behavioral, and immunologic responses to stress (Heinrichs et al., 1995). Numerous preclinical and clinical studies have shown that both maternally separated rats and depressed patients exhibit an apparent increase in CRF neurotransmission, as evidenced by heightened HPA axis activity and increased CRF concentrations in the cerebrospinal fluid (CSF) (Heit et al., 1997). As a consequence of these observations, increased limbic and hypothalamic CRF activity has been linked to the psychopathology of affective disorders. Clinical studies have repeatedly shown that drug-free depressed patients exhibit elevated concentrations of serum cortisol, failure of cortisol suppression after the administration of the synthetic glucocorticoid dexamethasone (Evans et al., 1983a, b), increased concentrations of cerebrospinal fluid CRF (Nemeroff et al., 1984;

Banki et al., 1987), decreased CRF receptorbinding in the frontal cortex (Nemeroff et al., 1988), a blunted ACTH response to exogenous CRF (Gold et al., 1986; Amsterdam et al., 1987), and hypertrophied pituitary and adrenal glands (Kathol et al., 1989; Nemeroff et al., 1992). These apparent increases in C W neurotransmission and HPA axis activity are currently thought to represent a state rather than a trait marker of depression, since hypercortisolemia and elevated CSF CRF concentrations normalize after electroconvulsive therapy or following clinical recovery (Nemeroff et al., 1991; Amsterdam et al., 1998). However accumulating evidence suggests that there may be subtle trait markers in the function of these systems among populations with genetic or environmental loading for development of major depressive disorder (Holsboer et al., 1995; Lauer et al., 1998; Model1 et al., 1998). In addition to dysregulation of hypothalamic and extra-hypothalamic CRF neurocircuitry, HMS 180 rats and depressed patients also appear to share dysregulation of noradrenergic and serotonergic systems (Owens and Nemeroff, 1994; Mongeau et al., 1997). Indeed, the pharmacological mechanism of action of most antidepressants is to increase NA and/or 5-HT neurotransmission. Antidepressant drugs are divided into several classes based on their pharmacological mechanisms of action. These classes include tri- and tetra-cyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), monoamine oxidase inhibitors (MAOIs), and atypical antidepressants. However, the neurochemical cascade(s) initiated by antidepressants resulting in clinical efficacy remains to be determined. Antidepressants of these various classes have similar clinical efficacy (approximately 65%) and generally require 4-8 weeks of treatment to produce their full therapeutic activity. Much research investigating chronic effects of antidepressant drugs has been conducted on normal, non-stressed animals. This approach, while convenient, will not likely provide much insight into the ultimate mechanisms of clinical recovery following antidepressant therapy. Antidepressants do not elevate the mood of non-depressed individuals (Sindrup et al., 1990). Therefore, it is unlikely that they will cause the same neurochemical

95

cascade of events in normal rats as they might in those which have been exposed to early adverse experience. In support of this thesis, chronic antidepressant treatment has no consistent effects on basal CRF expression in normal rats but can prevent a stress-induced increase in CRF expression (Brady et al., 1992; Heilig M and Ekman, 1995; Stout et al., 1997). Furthermore, heightened pituitary-adrenal responses and CSF CRF concentrations are normalized by chronic antidepressant treatment in both depressed patients and maternally separated rats but are unaltered in control populations. Because antidepressants alter HPA axis activity and alter central components of the HPA axis, Barden and colleagues (Barden et al., 1995) have postulated that at least part of their mechanisms of action is via these changes. We believe that the maternal separation model is suitable for investigating the pathophysiology of major depression and the mechanism(s) of action of antidepressant drugs. In support of this hypothesis, we have obtained preliminary evidence that various classes of antidepressant drugs attenuate or reverse the maternal separation phenotype. For example, we have found that increased regional CRF expression in maternally separated animals is attenuated by chronic treatment with the antidepressant paroxetine (Plotsky et al., unpublished communication). In addition, chronic treatment with paroxetine or the atypical antidepressant mirtazapine normalizes behavioral and endocrine stress responses in maternally separated rats (Plotsky et al., 1996; Ladd et al., 1997). These observations validate the maternal separation paradigm as a model of depression-like syndrome and, therefore, a means by which we can investigate the pathophysiology of this disease and the mechanism(s) of action of antidepressant drugs. Approximately 50% of patients who discontinue pharmacological antidepressant therapy during the first few months relapse into a depressive episode (Hirschfeld, 1996). This observation suggests that antidepressant therapy is required not only to reach clinical recovery but also to maintain it. Cessation of therapy removes the stabilizing effects of the drug, increasing the frequency and severity of relapse. It is our hypothesis that the neurochemical cascade of events underlying this relapse parallels

that initiating the primary affective episode. Thus, we will attempt to elucidate the pathophysiology of depression by investigating the neurochemical cascade(s) associated with antidepressant withdrawal. Preliminary data from our laboratory has revealed that normalization of the maternal separation phenotype following paroxetine administration is reversed upon drug withdrawal in adult HMS 180 rats, suggesting that the maternal separation paradigm is suitable for investigating the pathophysiology of affective states.

Synthesis and conclusions We are just beginning to unravel the cascade of neurochemical and microstructural adaptations which arise from our perinatal experiences. Although the psychological impact of adverse early experience has long been appreciated, the physiological and humoral imprint of early trauma, illness, or neglect is far deeper than previously imagined. In a series of studies we have demonstrated the long-term consequences of early handling-maternal separation on adult behavior, neuroendocrine responses to stressors, and several neurocircuits within the CNS. These are changes largely initiated as a result of specific aspects of maternal-infant interactions. A hypothetical model of our current understanding of the neural mechanisms that mediate the effects of early rearing on the expression of HPA responses to stress is presented in Fig. 4. This model illustrates sites which we believe play a role in tonic regulation of CRF/AVP synthesis as well as sites we believe to be involved in dynamic regulation of CRF/AVP release. We propose that the glucocorticoid feedback signal at the level of the hippocampus subserves, in part, tonic inhibition of CRF/AVP synthesis under resting-state conditions. Thus, increased hippocampal GR expression as a function of maternal separation would contribute to increases in resting-state levels of CRF and AVP in pPVN neurons while the increased MR expression in the hippocampus would restrain basal release of this CRF into the hypophysial-portal circulation. The increased CRF signal from the central nucleus of the amygdala and bed nucleus of the stria terminalis would function to increase

96

CRF

NE

Fig. 4. Central adaptations in response to maternal separation. Specific hypothalamic and extrahypothalamic neurocircuits exhibit upregulation in corticotropin releasing factor (CRF) mRNA, CRF peptide content, and up-regulation of CRF-RI receptors. This may be due in part to down regulation of glucocorticoid receptors (GR) in the hippocampus and medial prefrontal cortex (mpFrontal Cortex). Inhibitory GABAergic tone on brainstem and limbic sites may also mediate these effects via down-regulation of GABA, receptors and of the benzodiazepine (BZ) binding site. Brainstem noradrenergic (IW) systems arising from the locus coeruleus (LC) and the nucleus of the solitary tract (NTS) may also be more responsive due to the down-regulation of alpha2-adrenergic autoreceptors (a2-AR). Together, these changes would be expected to give rise to HPA axis and behavioral hyper-responsiveness to psychological stressors. Further details may be found in the text. AVP, arginine vasopressin; BNST, bed nucleus of the stria terminalis; DMH, dorsomedial hypothalamus; GABA, gamma-aminobutyric acid; pPVN, parvicellular paraventricular nucleus of the hypothalamus.

stress-induceddrive to ascending NAergic systems; as such, this influence would be greater in nonhandled and HMS180 animals as compared to HMS15 and AFR rats, thus contributing to enhanced NA release at pPVN (and other) terminal sites during periods of stress. We propose inhibitory sites that serve to regulate CRF/AVP release during psychological, but not physical stressors via their actions either at the level of the CRF neurons in the central nucleus of the amygdala or directly on NAergic cell bodies in the LC and NTS. These sites include GR sensitive neurons in the frontal cortex whose effect is mediated either by projections to the NAergic cell body regions, to the central nucleus of the amygdala, or both. In addition, we propose that BZ receptor activation at

the level of central and/or lateral nucleus of the amygdala, or directly at the level of the LC/NTS could serve to inhibit the magnitude of the ascending NAergic signal and, thus, CRF/AVP release. It also seems likely that the CRF pathway from the central nucleus of the amygdala and the bed nucleus of the stria terminalis to the LC, pathways which are postulated to activate the ascending NAergic system, could also serve as targets for BZ inhibition. These are not mutually exclusive possibilities; it is likely that the very substantial differences associated with early rearing condition occur as a function of multiple, converging systems. A synthesis of this model postulates that maternal separation (or adverse early experience) results in a localized insufficiency in the

97

function of two primary inhibitory systems, the glucocorticoid and GABAergic systems. We postulate that these changes lead to changes observed in excitatory neurocircuits (e.g. C W , NA, 5-HT). Together, these adaptations mediate the complex phenotype observed in the maternally separated rats. Both the major systems involved and order of this neurochemical cascade, however, remain to be confirmed. Burgeoning evidence indicates that there are critical developmental windows during which central nervous system neurocircuitry may be quite susceptible to environmental influences. We are compelled to acknowledge the importance of caregiver-child bonds during these periods which, through multiple mechanisms, sculpt the nervous system. The processes of activity-dependent plasticity and ‘hormonal programming’ act to ‘fine-tune’ the basic genetically determined neurocircuitry by altering the strength of connections, changing the morphology of neurons, and altering expression levels of critical signaling molecules. Together these changes result in robust differences in how an individual decodes and responds to its environment. While the road to full understanding remains long, perhaps these ideas provide a bridge between current molecular medicine and systems neuroscience and the earlier ideas of Freud and Bowlby. In integrating these concepts, we may be able to prevent many of the mental and physical diseases for which modern medicine has no cure.

Acknowledgements We would like to thank Seymour Levine for his insights and intellectual contributions to our research. We also acknowledge the National Institutes of Health for supporting our research via grants MH50113 (PhW and MJM) and MH42088 (CBN and PMP).

References Ader, R. and Grota, L.J. (1969) Effects of early experience on adrenocortical reactivity. Physiol. Behav., 4 303-305. Alberts, J.R. (1994) Learning as adaptation of the infant. Acta Paedit,: Suppl., 397: 77-85. Ambelas, A. (1987) Life events and mania: a special relationship? Brit. J. Psychiat., 150 235-240.

Amsterdam, J.D., Maislin, G., Winokur, A., Kling, M. and Gold, P. (1987) Pituitary and adrenocortical responses to the ovine corticotropin-releasing hormone in depressed patients and healthy volunteers. Arch. Gen. Psychiat., 44: 775-781. Amsterdam, J.D., Maislin, G.,Winokur, A., Berwish, N., Kling, M. and Gold, P. (1998) The oCRF stimulation test before and after clinical recovery from depression. J. Afecr. Disord., 14: 2 13-222. Anisman, H. and Zacharko, R.M. (1982) Depression: the predisposing influence of stress. Behav. Brain. Sci., 5 : 89-137. Arling, G.L., Harlow, H.F., Sackett, G.P., Suomi, S.J. and Ruppenthal, G.C. (1976) A 10-year perspective of motherless-mother monkey behavior. J. Abn. Psychol., 85: 341-349. Banki, C.M., Bissette, G., Arato, M., O’Conner, L. and Nemeroff, C.B. (1987) Cerebrospinal fluid corticotropinreleasing factor-like immunoreactivity in depression and schizophrenia. Am. J. Psychiat., 144: 873-877. Ben-Ari, Y. (1995) Activity-dependent forms of plasticity. J. Neurobiol., 26: 295-298. Barden N., Reul J.M. and Holsboer F. (1995) Do antidepressants stabilize mood through actions on the hypothalamic-pituitary-adrenocorticalsystem? Trends Neurosci., 18: 6-11. Bhatnagar, S., Shank, N. and Meaney, M.J. (1995) Hypothalamic-pituitary-adrenal function in handled and non-handled rats in response to chronic stress. J. Neuroendocrinol., 7: 107-1 19. Blanchard, R.J., Kelley, M.J. and Blanchard, D.C. (1974) Defensive reactions and exploratory behavior in rats. J. Comp. Physiol. Psychol., 87: 1129-1 133. Blaszczyk, J.W. and Tajchert, K. (1997) Effect of acoustic stimulus characteristics on the startle response in hooded rats. Acta Neurobiol. Exp. (Warsz), 57: 315-321. Bodnoff, S.R., Suranyi-Cadotte, B., Quirion, R. and Meaney, M.J. (1987) Postnatal handling reduces novelty-induced fear and increases [3H]flunitrazepam binding in rat brain. Eur: J. Pharmacol., 144: 105-107. Bodnoff, S.R., Suranyi-Cadotte, B.E., Quirion, R. and Meaney, M.J. (1989) Role of the central benzodiazepine receptor system in behavioral habituation to novelty. Behav. Neurosci., 103: 209-212. Bohn, M.C. (1980) Granule cell genesis in the hippocampus of rats treated neonatally with hydrocortisone. Neuroscience, 5 : 2003-2012. Bohn, M.C. (1984) Glucocorticoid induced teratologies of the nervous system. In: Yanai, J. (Ed.), Neurobehavioral Teratology. Elsvier, New York: pp. 365-387. Bowlby, J. (1977a) The making and breaking of affectional bonds. I. Aetiology and psychopathology in the light of attachment theory. An expanded version of the Fiftieth Maudsley Lecture, delivered before the Royal College of Psychiatrists, 19 November 1976. B,: J. Psychiatry, 130: 20 1-2 10.

98 Bowlby, J. (1977b) The making and breaking of affectional bonds. 11. Some principles of psychotherapy. The Fiftieth Maudsley Lecture. B,: J. Psychiatry, 130: 421-431. Brady, L.S., Gold, P.W., Herkenham, M., Lynn, A.B. and Whitfield, H.J., Jr. (1992) The antidepressants Buoxetine, idazoxan, and phenelzine alter corticotropin-releasing hormone and tyrosine hydroxylase mRNA levels in the rat brain: therapeutic implications. Brain Res., 572: 117-125. Britton, D.R. and Thatcher-Britton, R. (1981) A sensitive openfield measure of anxiolytic drug action. Pharm. Biochem. Behav., 15: 577-580. Brown, G.W., Bifulco, A. and Harris, T.O. (1987) Life events, vulnerability and onset of depression. Brit. J. Psychiat., 150: 30-42. Brown, G.W. and Harris, T.O. (1993) Aetiology of anxiety and depressive disorders in an inner-city population. 1. Early adversity. Psychol. Med., 23: 143-154. Butler, P.D., Weiss, J.M., Stout, J.C. and Nemeroff, C.B. (1990) Corticotropin-releasing factor produces fear-enhancing and behavioral activating effects following infusion into the locus coeruleus. J. Neurosci., 10: 176-183. Cador, M., Ahmed, S.H., Koob, G.F., LeMoal, M. and Stinus, L. (1992) Corticotropin-releasing factor produces a fear place aversion independent of its neuroendocrine role. Brain Res., 597: 304-309. Caldji, C., Francis, D., Sharma, S., Plotsky, P.M. and Meaney, M.J. (1999) The effects of early rearing environment on the development of GABA, and central benzodiazepine receptor levels and novelty-induced fearfulness in the rat. J. Neumsci., in press. Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P.M. and Meaney, M.J. (1998) Maternal care during infancy regulates the development of neural systems mediating the expression of fearfulness in the rat. Pmc. Nut. Acad. Sci. USA, 95: 5335-5340. Chamove, A.S., Rosenblum, L.A. and Harlow, H.F. (1973) Monkeys (Macaca mulatta) raised only with peers. A pilot study. Animal. Behav., 21: 316-325. Clement, Y. (1996) Structural and pharmacological aspects of the GABAA receptor: involvement in behavioral pathogenesis. J. Physiol. Paris, 90: 1-13. Cloninger, C.R., Adolfsson, R. and Svrakic, N.M. (1996) Mapping genes for human personality. Nature Genetics, 12: 3-4. Coplan, J.D., Andrews, M.W., Rosenblum, L.A., Owens, M.J., Friedman, S., Gorman, J.M. and Nemeroff, C.B. (1996) Persistent elevations of cerebrospinal fluid concentrations of corticotropin-releasing factor in adult non-human primates exposed to early-life stressors: Implications for the pathophysiology of mood and anxiety disorders. Proc. Nut. Acad. Sci. USA, 93: 1619-1623. Cui, X. and Vaillant, G.E. (1996) Antecedents and consequences of negative life events in adulthood: a longitudinal study.Am. J. Psychiat., 152: 21-26. Cullinan, W.E., Herman, J.P., Helmreich, D.L. and Watson, S.J. (1995) A neuroanatomy of stress. In: M.J. Friedman, D.S. Charney and A.Y. Deutch (Eds), NeurobiologicaE and

Clinical Consequences of Stress: Fmm Normal Adaptation to PTSD, Lippincott-Raven, Philadelphia, pp. 3-25. Curtis, A.L., Pavcovich, L.A., Grigordiadis, D.E. and Valentino, R.J. (1995) Previous stress alters corticotropin-releasing factor neurotransmission in the locus coeruleus. Neuroscience, 65: 541-550. Curtis, A.L. and Valentino, R.J. (1 994) Corticotropin-releasing factor neurotransmisson in locus coeruleus: A possible site of antidepressant action. Brain Res. Bull., 35: 58 1-587. Dallman, M.F., Akana, S., Cascio, C.S., Darlington, D.N., Jacobson, L. and Levin, N. (1987) Regulation of ACTH secretion: variations on a theme of B. Rec. Pmg. Horn. Res., 43: 113-173. Davis, M. (1986) Pharmacological and anatomical analysis of fear conditioning using the fear-potentiated startle paradigm. Behav. Neumsci., 100: 8 14-824. DeBellis, M.D., Chrousos, G.P., Dom, L.D., Burke, L., Helmers, K., Kling, M.A., Trickett, P.K. and Putnam, F.W. (1994) Hypothalamic pituitary adrenal dysregulation in sexually abused girls. J. Clin. Endocrind Metab., 78: 249-255. De Blas, A.L. (1996) Brain GABAA receptors studied with subunit-specific antibodies. Mol. Neurobiol., 12: 55-71. De Kloet, E.R. (1991) Brain corticosteroid receptor balance and homeostatic control. Frontiers in Neumendocrinology, 12: 95-164. Denenberg, V.H. (1964) Critical periods, stimuli input, and emotional reactivity: a theory of infantile stimulation. Psychol. Rev., 7 1: 335-35 1. De Souza, E.B., Insel, T.R., Perrin, M.H., Rivier, J., Vale, W.W. and Kuhar, M.J. (1985) Corticotropin-releasing factor receptors are widely distributed within the rat central nervous system: an autoradiographic study. J. Neumsci., 5 : 3 189-3203. Deutsch, S.I., Rosse, R.B. and Mastropaolo, J. (1994) Environmental stress-induced functional modification of the central benzodiazepine binding site. Clin. Neuropharmacol., 17: 205-228. Duman, R.S., Heninger, G.R. and Nestler, E.J. (1997) A molecular and cellular theory of depression. Arch. Gen. Psychiatry, 54: 597-606. Dunn, A.J. and Berridge, C.W. (1990) Physiological and behavioral responses to corticotropin-releasing factor administration: is CRF a mediator of anxiety and of stress responses? Brain Res. Rev., 15: 71-100. Dunner, D.L., Patrick, V. and Fieve, R.R. (1979) Life events at the onset of bipolar affective illness. Am. J. Psychiar., 136: 508-5 1 1. Engelmann, M., Thrivikraman, K.V., Su, Y., Nemeroff, C.B.. Montkowski, A., Landgraf, R., Holsboer, F. and Plotsky, P.M. (1996) Endocrine and behavioral effects of airpuff-startle in rats. Psychoneuroendocrinology,2 1: 39 1-400. Evans, D.L., Burnett, G.B. and Nemeroff, C.B. (1983a) The dexamethasone suppression test in the clinical setting. Am. J. Psych., 140: 586-589.

99 Evans, D.L., Burnett, G.B. and Nemeroff, C.B. (1983b) Use of the dexamethasone suppression test using DSM ZZZ criteria on an inpatient psychiatric unit. Biol. Psych., 18: 505-51 1. Evoniuk, G.E., Kuhn, C.M. and Schanberg, S.M. (1979) The effect of tactile stimulation on serum growth hormone and tissue ornithine decarboxylase activity during maternal deprivation in rat pups. Commun. Psychophannacol., 3: 363-370. Feldman, S. and Weidenfeld, J. (1998) The excitatory effects of the amygdala on hypothalamo-pituitary-adrenocortical responses are mediated by hypothalamic norepinephrine, serotonin, and CRF4I. Brain Res. Bull., 45: 389-93. File, S.E. (1995) Animal models of different anxiety states. In: G. Biggio and E. Costa (Eds), GABA, Receptors anddnxiety: From Neurobiology to Treatment, Raven Press, New York, pp. 93-1 13. Francis, D., Diorio, J., LaPlante, P., Weaver, S., Seckl, J.R. and Meaney, M.J. (1996) The role of early environmental events in regulating neuroendocrine development. Ann. N. Y: Acad. Sci., 794: 136-52. Freud, S. (1966) The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 1, Hogarth, London, p. 28 1. Gabr, R.W., Birkle, D.L. and Azzaro, A.J. (1995) Stimulation of the amygdala by glutamate facilitates corticotropin-releasing factor release from the median eminence and activation of the hypothalamic-pituitary-adrenal axis in stressed rats. Neuroendocrinology, 62: 333-339. Gold, P.W., Loriaux, D.L., Rby, A., Kling, M.A., Calabrese, J.R., Kellner, C.H., Nieman, L.K., Post, R.M., Pickar, D., Gallucci, W., Avgerinos. P., Paul, S., Oldfield, E.H., Cutler, G.B., Jr. and Chrousos, G. (1986) Responses to corticotropinreleasing hormone in the hypercortisolism of depression and Cushing’s disease. Pathophysiologic and diagnostic implications. N. Eng. J. Med., 314: 1329-1355. Gould, E., Woolley, C.S. and McEwen, B.S. (1991a) Adrenal steroids regulate postnatal development of the rat dentate gyrus: I. Effects of glucocorticoids on cell death. J. Comp. Neurol., 313: 479435. Gould, E., Woolley, C.S., Cameron, H.A., Daniels, D.C. and McEwen, B.S. (1991b) Adrenal steroids regulate postnatal development of the rat dentate gyrus: 11. Effects of glucocorticoids and mineralocorticoids on cell birth. J. Comp. Neurol., 313: 486-493. Gray, T.S. and Bingaman, E.W. (1996) The amygdala: corticotropin-releasing factor, steroids, and stress. Critical Rev. Neurobiol., 10: 155-168. Gray, T.S. (1990) The organization and possible function of amygdaloid corticotropin-releasing factor pathways. In: E.B. DeSouza and C.B. Nemeroff (Eds), Corticotropin-releasing Factor: Basic and Clinical Studies of a Neuropeptide, CRC Press, Boca Raton, pp. 53-68. Heilig, M. and Ekman, R. (1995) Chronic parenteral antidepressant treatment in rats: unaltered levels and processing of neuropeptide Y (NPY) and corticotropin-releasing hormone (CRH). Neurochem. Znt., 26: 351-355.

Heim, C., Owens, M.J., Plotsky, P.M. and Nemeroff, C.B. (1997) The role of early adverse life events in the etiology of depression and posttraumatic stress disorder: focus on corticotropin-releasing factor. Ann. N. Y: Acad. Sci., 821: 194-207. Heinrichs, S.C., Pich, E.M., Miczek, K., Britton, K.T. and Koob, G.F. (1992) Corticotropin-releasing factor antagonist reduces emotionality in socially defeated rats via direct neurotropic action. Brain Res., 581: 190-197. Heinrichs, S.C., Menzaghi, F,, Pich, E.M., Britton, K.T. and Koob, G.F. (1995) The role of CRF in behavioral aspects of stress, Ann. N.E Acad. Sci., 771: 92-104. Heit S., Owens, M.J., Plotsky, P.M. and Nemeroff, C.B. (1997) Corticotropin-releasing factor, stress, and depression. The Neuroscientist, 3: 186-194. Hennessy, M.B. (1997) Hypothalamic-pituitary-adrenal responses to brief social separation. Neurosci. Biobehav. Rev., 21: 11-29. Hirschfeld, R.M.A. (1996) Long-term drug treatment of unipolar depression. Zntl. Clin. Psychophannacol., 11: 211-217. Hofer, M.A. (1973) The role of nutrition in the physiological and behavioral effects of early maternal separation on infant rats. Psychosom. Med., 35: 350-359. Hofer, M.A. (1984) Early stages in the organization of cardiovascular control. Proc. SOC. Exp. Biol. Med., 175: 147-157. Hofer, M.A. (1994) Early social relationships as regulators of infant physiology and behavior. Acta Paediatl: Suppl., 397: 9-18. Holsboer, F., Grasser, A., Fries, E., Wiedemann, K. (1994) Steroid effects on central neurons and implications for psychiatric and neurological disorders. Ann. NI: Acad. Sci., 746: 345-359. Holsboer, F., Lauer, C.J., Schreiber, W., Krieg, J.C. (1995) Altered hypothalamic-pituitary-adrenocortical regulation in healthy subjects at high familial risk for affective disorders. Neuroendocrinology, 62: 340-347. Hsu, D.T., Chen, EL., Takahashi, L.K., Kalin, N.H. (1998) Rapid stress-induced elevations in corticotropin-releasing hormone mRNA in rat central amygdala nucleus and hypothalamic paraventricular nucleus: an in situ hybridization analysis. Brain Res., 788: 305-310. Huot, R., Smith, M., Plotsky, P.M. (1997) Alterations of maternal-infant interaction as a result of maternal separation in Long Evans rats and its behavioral and neuroendocrine consequences. International SOCPsychoneuroendocrinology, XXVIIIth Congress, July 26-30, San Francisco, CA. Huot, R.L., Thrivikraman, K.V., and Plotsky, P.M. (1997) Developmentof adult ethanol preference as a consequence of maternal separation in neonatal Long Evans rats. Abstract for the 27th Annual Society for Neuroscience: 724.24. Jacobson, L. and Sapolsky, R.M. (1991) The role of the hippocampus in feedback regulation of the hypothalamicpituitary-adrenal axis. Endocl: Rev., 12: 118-134. Jang, K.L.,McCrae, R.R., Angleitner, A., Riemann, R. and Livesley, W.J. (1998) Heritability of facet-level traits in a

100 cross-cultural twin sample: support for a hierarchical model of personality. J. Pers. SOC.Psychol., 74: 15561565. Jans, J. and Woodside. B.C. (1990) Nest temperature: effects on maternal behavior, pup development, and interaction with handling. Develop. Psychobiol., 23: 5 19-534. Joels, M. (1997) Steroid hormones and excitability in the mammalian brain. Front. Neuroendocrinol., 18: 2-48. Kalin, N.H., Larson, C., Shelton, S.E. and Davidson, R.J. (1998) Asymmetric frontal brain activity, cortisol, and behavior associated with fearful temperament in rhesus monkeys. Behav. Neurosci., 112: 286292. Kalin, N.H. and Shelton, S.E. (1998) Asymmetric frontal brain ontogeny and stability of separation and threat-induced defensive behaviors in rhesus monkeys during the first year of life. Am. J. Primatol., 44: 125-135. Kathol, R.G., Jaeckle, R.S., Lopez, J.F. and Meller, W.H. (1989) Consistent reduction of ACTH responses to stimulation with CRH, vasopressin and hypoglycemia in patients with major depression. BE J. Psychiat., 155: 468-478. Katz, L.C., Shatz, C.J. (1996) Synaptic activity and the construction of cortical circuits. Science, 274: 1133-1 138. Kehoe, P., Shoemaker, W.J., Triano, L., Callahan, M. and Rappolt, G . (1998) Adult rats stressed as neonates show exaggerated behavioral responses to both pharmacological and environmental challenges. Behav. Neurosci., 112: 116-125. Kendler, K.S., Kessler, R.C., Neale, M.C., Heath, A.C. and Eaves, L.J. (1993) The prediction of major depression in women: toward an integrated etiological model. Am. J. Psychiat., 150: 1139-1 148. Kessler, R.C., McGonagle, K.A., Zhao, S., Nelson, C.B., Hughes, M., Eshleman, S., Wittchen, H.U. and Kendler, K.S. (1994) Lifetime and 12-month prevalence of DSM-111-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch. Gen. Psychiat., 5 1: 8-19. Koegler-Muly, S.M., Owens, M.J., Ervin, G.N., Kilts, C.D. and Nemeroff, C.B. (1993) Potential corticotropin-releasing factor pathways in the rat brain as determined by bilateral electrolytic lesions of the central amygdaloid nucleus and the paraventricular nucleus of the hypothalamus. J. Neuroendocrinol., 5 : 95-98 Kolb, B. and Whishaw, LQ. (1998) Brain plasticity and behavior. Ann. Rev. Psychol., 49: 43-64. Ladd, C.O., Owens, M.J. and Nemeroff, C.B. (1996) Persistent changes in corticotropin-releasing factor neuronal systems induced by maternal deprivation. Endocrinology, 137: 1212-1218. Ladd, C.O., Huot, R.L., Thrivikraman, K.V. and Plotsky, P.M. (1998) Persistent alterations in the negative feedback regulation of the hypothalamic-pituitary-adrenal (HPA) axis in maternally-separated adult Long Evans hooded rats. Abstract for the 28th Annual Society for Neuroscience. Ladd, C.O., Lyss, P.J., Owens, M.J. and Plotsky, P.M. (1997) Effects of mirtazapine on behavioral, hormonal, and neurochemical parameters in maternally separated rats. Abstract for the 27th annual Society for Neuroscience: 391.1.

Lauer, C.J., Schreiber, W., Modell, S . , Holsboer, F. and Krieg, J.C. (1998) The Munich vulnerability study on affective disorders: overview of the cross-sectional observations at index investigation. J. Psychiatr: Res., 32: 393-401. Lee, Y. and Davis, M. (1997) Role of the hippocampus, the bed nucleus of the stria terminalis, and the amygdala in the excitatory effect of corticotropin-releasing hormone on the acoustic startle reflex. J. Neurosci., 17: 6434-6446. Leon M., Croskerry P.G. and Smith G.K. (1978) Thermal control of mother-infant contact in rats. Physiol. Beh., 21: 793-81 1. Leserman, J., Drossman, D.A., Li, Z., Toomey, T.C., Nachman, G . and Glogau, L. (1996) Sexual and physical abuse history in gastroenterology practice: how types of abuse impact health status. Psychosom. Med., 58: 4-15. Levine, S. (1957) Infantile experience and resistance to physiological stress. Science, 126: 405-406. Levine, S. (1975) Psychosocial factors in growth and development. In: L. Levi (Ed.), Society, Stress, and Disease, Oxford University Press, London, pp. 43-50. Levine, S., Wiener, S.G. and Coe, C.L. (1993) Temporal and social factors influencing behavioral and hormonal responses to separation in mother and infant squirrel monkeys. Psychoneuroendocrinology,18: 297-306. Levine, S., Huchton, D.M., Wiener, S.G. and Rosenfeld, P. (1991) Time course of the effect of maternal deprivation on the hypothalamic-pituitary-adrenal axis in the infant rat. Dev. Psychobiol., 24: 547-558. Liu, D., Tannenbaum, B., Caldji. C., Francis, D., Freedman, A., Sharma, S., Pearson, D., Plotsky, P.M. and Meaney, M.J. (1997) Maternal care, hippocampal glucocorticoid receptor gene expression and hypothalamic-pituitary-adrenal responses to stress. Science, 277: 1659-1662. Liu, D., Caldji C., Sharma S., Plotsky, P.M. and Meaney, M.J. (1999) Influence of neonatal rearing condition on stressinduced hypothalamic-pituitary-adrenal responses and norepinephrine release in the hypothalamic paraventricular nucleus. J. Neuroendocrinology, in press. Luddens, H. and Korpi, E.R. (1995) Biological function of GABAJbenzodiazepine receptor heterogeneity. J. Psychiatry Res., 29: 77-94. Magarinos, A.M., Verdugo, J.M. and McEwen, B.S. (1997) Chronic stress alters synaptic terminal structure in hippocampus. Proc. Nail. Acad. Sci. USA, 94(25): 14002-14008. Makino, S., Gold, P.W. and Shulkin, J. (1994) Corticosterone effects on corticotropin-releasing hormone mRNA in the central nucleus of the amygdala and the parvocellular region of the paraventricular nucleus of the hypothalamus. Brain Res., 640: 105-112. Makino, S., Gold, P.W. and Shulkin, J. (1994) Effects of corticosterone on CRH mRNA and content in the bed nucleus of the stria terminalis; comparison with the effects in the central nucleus of the amygdala and the paraventricular nucleus of the hypothalamus. Brain Res., 657: 141-149. Malizia, A.L., Coupland, N.J. and Nutt, D.J. (1995) Benzodiazepine receptor function in anxiety disorders. In: E. Biggo, F. Sanna and E. Costa (Eds), GABA, Receptors and Anxiety:

101

Fmm Neumbiology to Treatment, Raven Press, New York, pp. 115-133. Marcilhac A. and Siaud P. (1997) Identification of projections from the central nucleus of the amygdala to the paraventricular nucleus of the hypothalamus which are immunoreactive for corticotrophin-releasinghormone in the rat. Exp. Physiol., 82: 273-81. Martin, L.J., Spicer, D.M., Lewis, M.H., Gluck, J.P. and Cork, L.C. (1991) Social deprivation of infant rhesus monkeys alters the chemoarchitecture of the brain in subcortical regions. J. Neurosci., 11: 33443358. McEwen, B.S. (1995) Stressful experience, brain, and emotions: Developmental, genetic, and hormonal influences. In: M.S. Gazzaniga (Ed.), The Cognitive Neurosciences, MIT Press, Cambridge, pp. I 1 17-1 135. McEwen, B.S. (1998) Stress, adaptation, and disease. Allostasis and allostatic load. Ann. NE Acad. Sci., 840: 33-44. McGue, M. and Bouchard, T.J. Jr. (1998) Genetic and environmental influences on human behavioral differences. Ann. Rev. Neumsci., 21: 1-24 Meaney, M.J., Aitken, D.H., Sharma, S., Viau, V. and Sarrieau, A. (1989) Posmatal handling increases hippocampal type 11, glucocorticoid receptors and enhances adrenocortical negative-feedback efficacy in the rat. Neuroendocrinology, 5: 597404. Meaney, M.J., Bhatnagar, S., Larocque, S., McCormick, C., Shanks, N., Sharma, S., Smythe, J., Viau, V. and Plotsky, P.M. (1993) Individual differences in the hypothalamic-pituitaryadrenal stress response and the hypothalamic CRF system. Ann. NY:Acad. Sci. USA, 697: 70-85. Meaney, M.J., Diorio, J., Francis, D., Widdowson, J., LaPlante, P.. Caldji, C., Sharma, S., Seckl, J. and Plotsky, P.M. (1996) Early environmental regulation of forebrain glucocorticoid receptor gene expression: implications for adrenocortical responses to stress. Dev. Neurosci., 18: 49-72. Meaney, M. J., O’Donnell, D., Viau, V., Bhatnagar, S., Sarrieau, A., Smythe, J., Shanks, N. and Walker, C.D. (1993) Corticosteroid receptors in the rat brain and pituitary during development and hypothalamic-pituitary-adrenal function. In: 1,s. Zagon and P.J. McLaughlin (Eds), Receptors in the Developing Nervous System, Chapman & Hall, London, pp. 163-201. Merchenthaler, I., Vigh, S., Petrusz, P. and Schally, A.V. (1982) Immunocytochemical localization of corticotropin-releasing factor (CRF) in the rat brain. Am. J. Anat., 165: 385-396. Modell, S., Lauer, C.J., Schreiber, W., Huber, J., Krieg, J.C. and Holsboer, F, (1998) Hormonal response pattern in the combined DEX-CRH test is stable over time in subjects at high familial risk for affective disorders. Neuropsychopharmacology, 18: 253-262. Mongeau, R., Blier, P. and de Montigny, C. (1997) The serotonergic and noradrenergic systems of the hippocampus: their interactions and the effects of antidepressant treatments. Brain Res. Rev., 23: 145-195. Munck, A., Guyre, P.M. and Holbrook, N.J. (1984) Physiological functions of glucocorticoids in stress and their relations to pharmacological actions. Endocr: Rev., 5: 2 5 4 .

Muscat, R. and Willner, P. (1992) Suppression of sucrose drinking by chronic mild unpredictable stress: a methodological analysis. Neurosci. Biobehav. Rev., 16: 507-5 17. Nemeroff, C.B. (1996) The corticotropin-releasing factor (CRF) hypothesis of depression: New findings and new directions. Molec. Psychiatry, 1: 336-342. Nemeroff, C.B., Widerlov, E., Bissette, G., Walleus, H., Karlsson, I., Eklund, K., Kilts, C.D., Loosen, P.T. and Vale, W. (1984) Elevated concentrations of CSF corticotropinreleasing factor-like immunoreactivity in depressed patients. Science, 226: 1342-1344. Nemeroff, C.B., Owens, M.J., Bissette, G., Andom, A. and Stanley, M. (1988) Reduced corticotropin-releasing factor binding sites in the frontal cortex of suicide victims. Arch. Gen. Psychiat., 45: 577-579. Nemeroff, C.B. (1991) Chronic stress and the psychopathology of affective disorders. In: H.G.M. Westenberg (Ed.), Stress, Biological Rhythms and Psychiatric Disorders, Third International Symposium Proceedings, The Netherlands, Medidact, Houten, pp. 13-16. Nemeroff, C.B., Bissette, G., Akil, H. and Fink, M. (1991) Neuropeptide concentrations in the cerebrospinal fluid of depressed patients treated with electroconvulsive therapy: corticotropin-releasing factor, @-endorphinand somatostatin. Br. J. Psychiat., 158: 59-63. Nemeroff, C.B., Krishnan, K.R.R., Reed, D., Leder, R., Beam, C. and Dunnick, N.R. (1992) Adrenal gland enlargement in major depression: a computed tomography study. Arch. Gen. Psychiat., 49: 196-202. Orchinik, M. (1998) Glucocorticoids, stress and behavior: Shifting the timeframe. Hormones Behav., 34: 320-327. Orchinik, M. and McEwen, B.S. (1994) Rapid steroid actions in the brain: A critique of genomic and non-genomic mechanisms. In: M. Wehling (Ed.), Genomic and Non-Genomic Effects of Aldosterone, CRC Press, Boca Raton, Florida, pp. 77-108. Orchinik, M., Murray, T.F. and Moore, EL. (1991) A corticosteroid receptor in neuronal membranes. Science, 252: 1848-185 1. Owens, M.J., Vargas, M.A., Knight, D.L. and Nemeroff, C.B. (1991) The effects of alprazolam on corticotropin-releasing factor neurons in the rat brain: acute time course, chronic treatment and abrupt withdrawal. J. Pharm. Exp. Thex, 258: 349-356. Owens, M.J. and Nemeroff, C.B. (1991) The physiology and pharmacology of corticotropin-releasing factor. Pharmacol. Rev., 43: 425-473. Owens, M.J. and Nemeroff, C.B. (1994) Role of serotonin in the pathophysiology of depression: focus on the serotonin transporter. Clin. Chem., 40: 288-295. Pacak, K., Kvetnansky, R., Palkovits, M., Fukuhara, K., Yadid, G., Kopin, I. and Goldstein, D.S. (1993) Adrenalectomy augments in vivo release of norepinephrine in the paraventricular nucleus during immobilization stress. Endocrinology, 133: 1404-1410. Pacak, K., Palkovits, M., Kvetnansky, R., Matem, P., Hart, C., Kopin, I. and Goldstein, D.S. (1995) Catecholaminergic

102 inhibition by hypercortisolemia in paraventricular nucleus of conscious rats. Endocrinology, 136: 4814-4819. Papp, M., Moryl, E. and Willner, P. (1996) Pharmacological validation of the chronic mild stress model of depression. Eur: J. Pharmacol., 296: 129-136. Pellow, S., Chopin, F?, File, S.E. and Briley, M. (1985) Validation of opedclosed arm entries in an elevated plus maze as a measure of anxiety in the rat. J. Neurosci. Methods, 14: 149-167. Pihoker, C., Owens, M.J., Kuhn, C.M., Schanberg, S.M. and Nemeroff, C.B. (1993) Maternal separation in neonatal rats elicits activation of the hypothalamic-pituitary-adrenocortical axis: a putative role for corticotropin-releasing factor. Psychoneuroendocrinology,18: 485-493. Plotsky, P.M., Cunningham E.T. and Widmaier E.P. (1989) Catecholaminergic modulation of corticotropin-releasing factor and dexamethasone secretion. Endocrin. Rev., 10: 437-458. Plotsky, P. (1991) Pathways to the secretion of adrenocorticotrophin: a view from the portal. J. Neuroendocrinol., 3: 1-9. Plotsky, P.M., Ladd, C.O., Huot, B.L., Thrivikraman K.V., Owens, M.J., Meaney M.J. and Nemeroff, C.B. (1999) A rat adverse early experience model of depression-like syndrome: reversal by paroxetine. Science, submitted. Plotsky, P.M. and Meaney, M.J. (1993) Early, postnatal experience alters hypothalamic corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stressinduced release in adult rats. Mol. Brain Res., 18: 195-200. Plotsky, P.M. and Meaney, M.J. (1996) Neonatal rearing conditions alter HPA axis function, central CRF mRNA, CSF CRF levels, and behavior: reversal by SSRI treatment. Abstract for the 26th ISPNE Congress. Plotsky, P.M., Ladd, C.O., Huot, R.L., Thrivikraman, K.V. and Nemeroff, C.B. (1997) Long-term consequences of maternal separation in Long Evans rats and their reversal by antidepressants. Abstract for the 28th ISPNE Congress: S 174. Plotsky, P.M., Owens, M.J. and Nemeroff, C.B. (1998) Psychoneuroendocrinolgyof depression: Hypothalamic-pituitary-adrenal axis. In: C.B. Nemeroff (Ed.), Psychiatry Clinics of North America, Vol. 21, W.B. Saunders, Philadelphia, pp. 293-307. Plotsky P.M., Thrivikraman, K.V., Caldji, C., Sharman, S. and Meaney, M.J. (1999) The effects of neonatal rearing environment on CRF mRNA and CRF receptor levels in adult rat brain. Proc. Natl. Acad. Sci. USA, submitted. Pucilowski, O., Overstreet, D.H., Rezvani, A.H. and Janowsky, D.S. (1993) Chronic mild stress-induced anhedonia: greater effect in a genetic rat model of depression. Physiol. Behav., 54: 1215-1220. Robinson, J.L., Kagan, J., Reznick, J.S. and Corley, R. (1992) The heritability of inhibited and uninhibited behaviors. Developmental Psychology, 28: 1030-1037. Rosenfeld, P., Suchecki, D. and Levine, S. (1992) Multifactorial regulation of the hypothalamic-pituitary-adrenal axis during development. Neurosci. Biobehav. Rev., 16: 553-568.

Rosenfeld, P., Van Eekelen, J.A.M., Levine, S. and De Kloet, E.R. (1993) Ontogeny of corticosteroid receptors in the brain. Cell. Mol. Neurobiol., 13: 295-317. Rutter, M., Dunn, J., Plomin, R., Simonoff, E., Pickles, A., Maughan, B., Ormel, J., Meyer, J. and Eaves, L. (1997) Integrating nature and nurture: Implications of personenvironment correlations and interactions for developmental psychopathology. Dev. Psychopathology, 9: 335-364. Sapolsky, R.M. and Meaney, M.J. (1986) Maturation of the adreno-cortical stress response: neuroendocrine control mechanisms and the stress hyporesponsive period. Brain Res. Rev., 11: 65-76. Sawchenko, P.E., Imaki, T., Potter, E., Kovics, K., Imaki, J., Vale, W. (1993) The functional neuroanatomy of corticotropin-releasing factor. In: D.J. Chadwick, S. Marsh and K. Ackrill (Eds), Corticotmpin-Releasing Factor; CIBA Found. Symp. 172, Wiley and Sons, Chichester, pp. 5-29. Scam, S. and McCartney, K. (1983) How people make their own environment: A theory of genotype-environmental effects. Child Dev., 54: 424-435. Schanberg, S.M., Evoniuk, G. and Kuhn, C.M. (1984) Tactile and nutritional aspects of maternal care: specific regulators of neuroendocrine function and cellular development. Proc. SOC. Exp. Biol. Med., 175: 135-146. Schoenfeld, N.M., Leathem, J.H. and Rabii, J. (1980) Maturation of adrenal stress responsiveness in the rat. Neuroendocrinology, 3 1: 101-105. Sindrup, S.H., Gram, L.F., Brosen, K., Eshoj, 0. and Mogensen, E.F. (1990) The selective serotonin reuptake inhibitor paroxetine is effective in the treatment of diabetic neuropathy symptoms. Pain, 42: 135-144. Smagin, G.N., Harris, R.B.S. and Ryan, D.H. (1996) Corticotropin-releasing factor receptor antagonist infused into the locus coeruleus attenuates immobilization stress-induced defensive withdrawal in rats. Neumsci. Lett., 220: 167-170. Spencer-Booth, Y. and Hinde, R.A. (1971) Effects of brief separations from mothers during infancy on behavior of rhesus monkeys 6-24 months later. J. Child Psycho1 Psychiat., 12: 157-172. Sroufe, L.A. (1997) Psychopathology as an outcome of development. Dev. Psychopathology, 9: 25 1-268. Stout, S.C., Owens, M.J., Heim, C. and Nemeroff, C.B. (1997) Attenuation by venlafaxine of stress effects on CRF systems and the HPA axis of rats. Abstract for the 27th annual Society for Neuroscience: 424.5. Suchecki, D., Nelson, D.Y., Van Oers, H. and Levine, S. (1995) Activation and inhibition of the hypothalamic-pituitaryadrenal axis of the neonatal rat: effects of maternal deprivation. Psychoneuroendocrinology, 20: 169-1 82. Suchecki, D., Rosenfeld, D. and Levine, S. (1993) Maternal regulation of the hypothalamic-pituitary-adrenal axis in the infant rat: the roles of feeding and stroking. Dev. Brain. Res., 75: 185-192. Suomi, S.J., Delizio, R. and Harlow, H.F. (1976) Social rehabilitation of separation-induced depressive disorders in monkeys. Am. J. Psychiatry, 133: 1279-1285.

103 Swanson, L.W., Sawchenko, P.W., Rivier J. andvale W. (1983) Organization of ovine corticotropin-releasing factor (CRF)immunoreactive cells and fibers in the rat brain: an immunocytochemical study. Neuroendocrinol., 36: 165-1 86. Sweirgiel, A.H., Takahashi, L.K. and Kalin, N.H. (1993) Attenuation of stress-induced behavior by antagonism of corticotropin-releasing factor receptors in the central amygdala in the rat. Brain. Res., 623: 229-234. Sze, P.Y. and Towle, A.C. (1993) Developmental profile of glucocorticoid binding to synaptic plasma membrane from rat brain. Int. J. Dev. Neurosci., 11: 339-346. Takahashi, L.K., Kalin, N.H., Vandenburgt, J.A. and Sherman, J.E. (1 989) Corticotropin-releasing factor modulates defensive-withdrawal and exploratory behavior in rats. Behav. Neurosci., 103: 648-654. Valentino, R.J., Foote, S.L. and Page, M.E. (1993) The locus coeruleus as a site for integrating corticotropin-releasing factor and noradrenergic mediation of stress responses. Ann. N.Z Acad. Sci., 697: 173-188. Valentino, R.J., Curtis, A.L., Page, M.E., Pavcovich, L.A. and Florin-Lecher, S.M. (1998) Activation of the locus ceruleus brain noradrenergic system during stress: circuitry, consequences, and regulation. Adv. Phamacol., 42: 781-784. Van Bockstaele, E.J., Colago, E.E.O. and Valentino, R.J. (1996) Corticotropin-releasing factor-containing axon terminals synapse onto catecholamine dendrites and may presynaptically modulate other afferents in the rostra1 pole of the nucleus locus coeruleus in the rat brain. J. Comp. Neurol., 364: 523-534. Van Bockstaele, E.J., Colago, E.E. and Valentino, R.J. (1998) Amygdaloid corticotropin-releasing factor targets locus coeruleus dendrites: substrate for the co-ordination of emotional and cognitive limbs of the stress response. J. Neuroendocrinol., 10: 743-757. Van Oers, H.J., de Kloet, E.R., Li, C. and Levine, S. (1998) The ontogeny of glucocorticoid negative feedback: influence of maternal deprivation. Endocrinology, 139: 2838-2846. Vazquez, D.M., Van Oers, H., Levine, S. and Akil, H. (1996) Regulation of glucocorticoid and mineralocorticoid receptor

mRNAs in the hippocampus of the maternally deprived infant rat. Brain Res., 731: 79-90. Viau, V., Sharma, S., Plotsky, P.M. and Meaney, M.J. (1993) Increased plasma ACTH responses to stress in non-handled compared with handled rats require basal levels of corticosterone and are associated with increased levels of ACTH secretagogues in the median eminence. J. Neuroscience, 13: 1097-1105. Walker, C.D., Perrin, M., Vale W. and Rivier, C. (1986) Ontogeny of the stress response in the rat: role of the pituitary and the hypothalamus. Endocrinology, 80: 1177-1 179. Weinstock, M. (1997) Does prenatal stress impair coping and regulation of the hypothalamic-pituitary-adrenal axis? Neurosci. Biobehav. Rev., 21: 1-10. Wheal, H.V., Chen, Y., Mitchell, J., Schachner, M., Maerz, W., Wieland, H., Rossum, D.V. and Kirsch, J. (1998) Molecular mechanisms that underlie structural and functional changes at the postsynaptic membrane during synaptic plasticity. Prog. Neurobiol., 55: 611-640. Willner, P. (1995) Animal models of depression: validity and applications. In: G. Gessa, W. Fratta, L. Pani and G. Serra (Eds), Depression and Mania: From Neurobiology to Treatment, Raven Press, New York, pp. 19-41. Willner, P. (1997) Validity, reliability and utility of the chronic mild stress model of depression: a 10-year review and evaluation. Psychophannacology (Berl), 134: 3 19-329. Willner, P., Towell, A., Sampson, D., Sophokleous, S. and Muscat, R. (1987) Reduction of sucrose preference by chronic mild stress and its restoration by a tricyclic antidepressant. Psychophamacology, 93: 358-364. Wilson, M.A. (1996) GABA physiology: modulation by benzodiazepines and hormones. Crit. Rev. Neurobiol., 10: 1-37. Young, E.A., Abelson, J.L., Curtis, G.C. and Nesse, R.M. (1997) Childhood adversity and vulnerability to mood and anxiety disorders. Depression Anxiety, 5: 6 6 7 2 .

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.)

Progress in B M i n Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 8

Neurobiological correlates of defensive behaviors Vaishali P. Bakshi, Steven E. Shelton and Ned H. Kalin* Department of Psychiatq University of Wisconsin at Madison, Wisconsin Psychiatric Institute and Clinics, 6001 Research Park Blvd., Madison, WI 53719, USA

Defensive behaviors In an attempt to understand the basic neural mechanisms underlying psychiatric conditions involving fear and anxiety, our group has focused on identifying the ontogeny and neural substrates of defensive behaviors. Defensive behaviors are exhibited by a wide array of species including rats, non-human primates, and humans in response to perceived threats from the environment, and are essential components of an organism’s behavioral repertoire that ensure its protection and survival. Although the specific behavioral responses that comprise ‘defensive behaviors’ are dependent on the environmental context and vary from species to species, a common element that unites this crossspecies phenomenon is that defensive behaviors represent an organism’sbehavioral response to fear. For example, vertebrates have evolved defensive behavioral responses that facilitate survival from threatening stimuli such as predators. Because defensive behaviors are expressed in response to an immediate threat, they characteristically supercede and interrupt the expression of other normal homeostatic behaviors such as feeding and reproduction that the organism may be engaging in at the time of the perceived threat (Schaller, 1972; Ficken and Witkin, 1977; Magurran and Girling, 1986; Ficken, 1990; Morse, 1993). One defensive response pattern expressed by many species is to *Correspondingauthor. Tel.: (608) 263-6079; Fax: (608) 263-9340

inhibit all body movements and assume an immobile or freezing posture. This phenomenon of behavioral inhibition is effective in preventing detection and attack by predators (Palmer, 1909; Curio, 1976), may have special relevance for understanding psychopathology, and will be discussed in detail in this review. Rodent species express a variety of defensive responses including vocalizations, fleeing, fighting, and freezing. Rodents develop the ability to engage in threat-induced behavioral inhibition or freezing at two weeks of age. Prior to this age, rat pups emit ultrasonic vocalizations in response to threatening stimuli (Noirot, 1972; Takahashi, 1992). These ultrasounds may serve to elicit and direct maternal attention to the pup (Noirot, 1972; Hofer, 1996) and may be associated with the fact that this species possesses immature sensory and motor systems at the time of birth and thus requires significant maternal care during the immediate postnatal period. Similarly, in non-human primates, defensive behaviors are comprised of a constellation of responses that include vocalizations, freezing, fleeing, or defensive hostility and aggression. The particular set of responses that is emitted depends upon, among other variables, the nature of the perceived threat (Kalin and Shelton, 1989; Kalin, 1993). We have been studying defensive, or fearrelated behaviors, and their physiological concomitants in rhesus monkeys to examine the factors mediating the development of individual differences in fearful temperament. Studies of

106

defensive behaviors in rhesus monkeys may provide valuable information that could aid in the understanding of fear and anxiety-related psychopathology in humans, since extreme fearful or defensive responses occur in dispositionally fearful humans who have an increased risk to develop psychopathology. For example, extremely inhibited children are at greater risk to develop anxiety and depressive disorders (Biederman et al., 1990; Hirshfeld et al., 1993), and these children are also more likely to have parents that suffer from anxiety disorders (Rosenbaum et al., 1988). Extremely inhibited children may have elevated levels of the stress-related hormone cortisol (Kagan et al., 1988), greater sympathetic nervous system activity (Kagan et al., 1987), and asymmetric electrical activity of right frontal brain regions (Davidson, 1992). These physiological characteristics may not only increase the vulnerability to develop psychopathology, but may also have important consequences for physical health. Our current working hypothesis is that psychiatric illnesses such as anxiety disorders and depression might involve the aberrant expression of defensive behaviors. In other words, pathological anxiety could be conceptualized as the inappropriate expression of defensive or fear-related behaviors, consisting of either an exaggerated or overly fearful response to an appropriate context, or a fearful response to an inappropriate or neutral context. Thus, while appropriate levels of defensive behaviors in response to environmental threats is an adaptive response that ensures one’s survival, the overly intense or context-inappropriate display of fear-related defensive behaviors may represent a liability that might ultimately contribute to certain forms of fear-related psychopathology. An understanding of the specific neural substrates underlying the expression and regulation of defensive behaviors may therefore ultimately shed insight into the processes that become dysregulated in these forms of psychiatric illness.

Use of non-human primates in the study of psychopathology Rhesus monkeys are ideally suited for studies examining mechanisms underlying human tem-

perament because they share key biological and social characteristics with humans (Harlow and Harlow, 1965; Hinde and White, 1974; Kalin and Carnes, 1984; Kalin et al., 1991a). Additionally, an extensive literature supports the importance of this species in modeling human psychopathology (Kaufman and Rosenblum, 1967; McKinney and Bunney, 1969; Harlow and Suomi, 1970; Suomi, 1986). Studies in rhesus monkeys can provide a bridge between findings elucidated from more basic work in rodents with those derived from human clinical research. In addition, mechanistic studies can be performed in monkeys that enable a further understanding of important scientific leads derived from clinical research. Also, compared to humans, rhesus monkeys have a relatively short life span that allows for longitudinal studies to be performed within a relatively short time frame. Our approach has been to develop a laboratory paradigm to characterize monkeys’ fearful behavioral responses to enable us to identify animals with fearful dispositions. Developing the ability to characterize different types of fearful responses has allowed us to understand the cues and environmental contexts that elicit different defensive behaviors. In addition, we have characterized an animal’s ability to adaptively regulate its defensive responses as environmental conditions change. Understanding individual differences in contextspecific defensive responses, their long-term stability, and the abilities of animals to adaptively regulate these responses has been a major area of focus.

Studies of defensive behaviors in non-human primates: human intruder paradigm To understand the conditions that elicit different defensive responses, as well as their ontogeny and neurochemical regulation, we developed the Human Intruder Paradigm. In this paradigm, young monkeys encounter a situation where they must cope with the adversity induced by maternal separation, in the presence and absence of a potential predatorial threat (Kalin and Shelton, 1989; Kalin et al., 1991a; Kalin, 1993). Thus, the infant is briefly separated from its mother and placed in a test cage where it remains for 30-40

107

min while its behavior is recorded on videotape. The test session consists of three consecutive 9 min conditions: alone (‘A’, animal left alone in cage); no eye contact (“EC’, animal presented with the facial profile of a human standing 2.5 m away); stare (‘ST’, animal presented with a human who faces it and engages it in direct eye contact). Qpically, animals respond the A condition by increasing their levels of locomotion and by emitting frequent coo vocalizations, which have been likened to the human cry and function to signal the infant’s location and facilitate maternal retrieval (Harlow and Harlow, 1965; Newman, 1985). The NEC condition causes a reduction in cooing and an increase in behavioral inhibition, which functions to help the monkey remain inconspicuous in the face of a predator and is often manifested as hiding behind the food bin and freezing. The ST condition elicits aggressive (openmouth threats, lunges, cage shaking, barking vocalizations) and submissive (lip smacking, feargrimacing) behaviors that represent adaptive responses to the perceived threat of the staring experimenter. Coo calls also are frequently emitted during ST. We have observed that the different test conditions (A, NEC, ST) reliably elicit responses in young or adult laboratory-reared monkeys or in feral animals (Kalin and Shelton, 1989; Kalin et al., 1991a; and unpublished data). Moreover, these condition-specific defensive responses are not dependent on the gender of the intruder, and can also be elicited by showing the animal a videotape of the intruder (unpublished data). Using the Human Intruder Paradigm, some of the neurochemical substrates mediating various components of defensive behaviors have been identified. Based on the findings that the p-opiate agonist morphine decreases cooing in the (A) condition, but has no effect on either NEC-induced freezing or ST-induced barking, and also that the opiate antagonist naloxone selectively increases Ainduced cooing, we have hypothesized that the opiate system mediates cooing associated with attachment bond disruption (Kalin et al., 1988; Kalin and Shelton, 1989; Kalin et al., 1995). These findings are in agreement with studies from numerous other laboratories that have suggested a primary role for brain opiatergic systems in the

regulation of attachment behavior between mother and infant (Panksepp et al., 1978; Fabre-Nys et al., 1982; Kehoe and Blass, 1986; Keverne et al., 1989; Martel et al., 1993). In addition to the opiate system, y-amino-butyric acidhenzodiazepine (GABA/BZ) receptors as well as the corticotropin releasing hormone (CRH) system are thought to modulate defensive behaviors in non-human primates. Studies in rodents (Vale et al., 1981; Kalin and Takahashi, 1991; Koob et al., 1993) and primates (Kalin et al., 1983; Kalin et al., 1989) suggest that CRH functions to integrate the coordinated activation of endocrine, autonomic, and behavioral responses essential to adaptive responding to stressful situations. Interestingly, these systems appear to modulate a different aspect of defensive behaviors than opiates. It has been found that benzodiazepine agonists such as diazepam decrease NEC-induced freezing and ST-induced barking without affecting A-induced cooing; conversely, intracerebroventricular administration of CRH increases freezing (Kalin and Shelton, 1989; Kalin et al., 1989; Kalin et al., 1991b). Thus, it is hypothesized that GABA/BZ and CRH systems might selectively regulate the expression of threatrelated behaviors rather than those elicited by disruption of maternal bond attachment.

Ontogeny of defensive behaviors Using the Human Intruder Paradigm, we investigated the developmental timecourse of defensive behaviors in rhesus monkeys. This work along with findings from other laboratories, summarized below, suggests the critical importance of the early postnatal period in the development of defensive responses. Similar to humans, infant rhesus monkeys require considerable and prolonged parental nurturance. During the first few months of life, if threatened by conspecifics or a predator, the infant completely depends on its mother for protection. Around 2 4 months of age, infant rhesus monkeys become increasingly self-reliant, venturing farther away from their mothers as they are drawn into playful and exploratory activities with their peers (Hinde and White, 1974; Berman, 1980). Around this time, the young monkeys begin to develop the ability to defend themselves and if threatened they

108

respond similarly to older animals by engaging in adult-like adaptive defensive responses (Kalin et al., 1991a). We characterized the age at which infant rhesus monkeys first express defensive responses as well as the age that they begin to adaptively regulate these responses. From observations of motherinfant interactions (Hinde and White, 1974; Berman 1980), we predicted that infant monkeys would be capable of regulating their defensive responses prior to four months of age. We suggested that this would be similar to the period at the end of the first year in human infants when they begin to discriminate strangers and novel situations by responding with increased fearfulness. Thus, we tested different age groups of infants that had no prior experience with the paradigm. During the first two weeks of life, infants were emotionally expressive but their defensive responses were indiscriminately displayed regardless of the context. For example, the infants frequently emitted coos and barks and engaged in freezing and hostile behaviors. However, they did not modulate the expression of these responses in relation to the presence or absence of the human intruder. However, by 9-12 weeks of age the infants regulated their defensive responses to the A, NEC, and ST conditions in an adult-like manner (Kalin et al., 1991). These findings are consistent with those from other studies in rhesus monkeys demonstrating that by two months of age rhesus infants perceive visual cues related to faces (Mendelson et al., 1982). Around this age, they respond fearfully to the presentation of pictures of conspecifics displaying threat faces and fear grimaces (Sackett, 1966) and also develop the ability to perform tasks based on visual working memory. Visual working memory is likely important in processing information conveying changes in environmental cues and may be fundamental in regulating defensive responses. Given the role of the dorsolateral prefrontal cortex in visual working memory and its connections to the hippocampus (which may be involved in mediating emotional appraisal and defensive behaviors), it will be interesting in future studies to examine the role of CRF receptors within these regions in the ontogeny of defensive responses,

since we have found that CRH receptors are first expressed in the dentate gyrus at around the same time that infant monkeys begin to be able to regulate their defensive behaviors (Goldman-Rakic et al., 1984; Goldman-Rakic, 1987; Grigoriadis et al., 1995). Indeed, it is likely that there is a critical period of brain development when environmental and hormonal influences could contribute to the development of individual differences in fearfulness, defensive responses, and perhaps CRH systems. In another study, we examined the development of defensive responses throughout the first year of life. In a different group of animals, we tested the maturational patterns of A-induced cooing and NEC-induced freezing at 4, 8, and 12 months of age. As expected, 4-month old infants regulated their defensive responses in an adult-like manner. By eight months, the infants cooed significantly less during the A condition a response which further decreased by 12 months. In contrast, levels of NEC-induced freezing remained unchanged from 4 to 12 months (Kalin and Shelton, 1998). The decrease in cooing that occurred during the A condition is consistent with the observation that infant rhesus monkeys are less dependent on their mothers during the second half of the first year of life. Importantly, the 12 month old animals did not lose their ability to coo, but rather the contexts that elicit the response changed. For example, older animals increased their cooing during the ST condition, which likely functions to recruit support from conspecifics during directly threatening situations. Converging lines of evidence from a number of species point to the importance of the early postnatal period, and in particular the bond between mother and infant, in the development of normal defensive behaviors and presumably underlying emotional responses. Indeed, a large body of work in rats indicates that a number of deleterious and long-lasting effects are produced as a result of separating rat pups from their mothers prior to weaning. The notion that perturbations in the early postnatal environment might have enduring neuroendocrine, neurochemical, and behavioral effects was originally put forth several decades ago by Levine and colleagues (Levine, 1957). It has since

I09

been demonstrated that a likely source of these alterations is a disruption of the interaction between mothers and rat pups (Sapolsky, 1997; Kuhn and Schanberg, 1998). Indeed, it has been shown recently that brief periods of maternal separation cause increased CRH expression, as well as alterations in CRH, benzodiazepine, noradrenergic, and glucocorticoid receptors in adult rats (Plotsky and Meaney, 1993; Liu et al., 1997; Caldji et al., 1998). Moreover, maternally separated rats are found in adulthood to exhibit increased defensive or fear-related behaviors and blunted responsiveness to reinforcers (Matthews et al., 1996; Caldji et al., 1998). These findings in rodents complement the classic reports by Harlow and colleagues, who studied the effects of maternal separation in primates (Harlow, 1959; Harlow and Zimmerman, 1959; Harlow and Harlow, 1965) and found that in addition to lifesupporting nourishment, physical contact and comfort are necessary for primates’ normal social and emotional development. During the first months of life, the attachment between mother and infant is intense and as a consequence the infant remains in close proximity to its mother (Harlow, 1959; Berman, 1980). Indeed, monkeys who have been separated from their mothers for prolonged periods during this time exhibit profound symptoms of enhanced defensive or fear-related behavioral responses as well as other abnormal behaviors into adulthood, a phenomenon that has led to the suggestion that the behavioral and neuroendocrine sequelae of maternal separation might provide a model for some of the dysfunction that is observed in anxiety disorders and depression (Harlow et al., 1964; McKinney and Bunney, 1969; Suomi et al., 1972; Reite, 1977; Rosenblum and Paully, 1987; Matthews et al., 1996). These findings are syntonic with the observations of Bowlby that children who were placed in nurseries that lacked adequate social stimulation developed a syndrome of ‘protest, despair, and detachment’ that may be analogous to an increase in defensive responses (Bowlby, 1973). Furthermore, recent reports suggest that children reared without the appropriate nurturance can display neuroendocrinological abnormalities and may develop long-term behavioral and emotional difficulties including an

increased risk for psychiatric illness (Frank et al., 1996; Carlson and Earls, 1997).Thus, disruption of normal attachment behavior at critical developmental phases can in a number of species lead to marked and persistent disturbances in behaviors and brain systems that are thought to participate in the regulation of fear-related responses; this disruption may ultimately contribute to an individual’s propensity to develop exaggerated or inappropriate defensive responses.

Individual differences in defensive behaviors Several lines of evidence support the notion that an individual’s level of defensive responding is a relatively stable trait characteristic which in part may be derived from the nature of early postnatal maternal interactions (Liu et al., 1997). Extreme individual differences detected early in life may be predictive of future psychopathology. For example, extremely inhibited children are at greater risk to develop anxiety and depressive disorders and are more likely to have parents that suffer from anxiety disorders (Biederman et al., 1990; Hirshfeld et al., 1993; Rosenbaum et al., 1993; Pollock et al., 1995). Moreover, behavioral inhibition in childhood (based on retrospective self-reports) is highly associated with anxiety in adulthood (Mick and Telch, 1998). Some of the physiological correlates that have been observed in extremely inhibited children are elevated levels of the stress-related hormone cortisol (Kagan et al., 1988), greater sympathetic nervous system activity (Kagan et al., 1987), and asymmetric electrical activity of frontal brain regions (Davidson, 1992). In our primate work, we have investigated individual differences in defensive behaviors in an attempt to further elucidate the neuroendocrine and neurobiological concomitants of extreme behavioral inhibition or fearfulness. Marked individual differences among rhesus monkeys have been noted with regard to the intensity of context-specific defensive responses. For example, some monkeys tend to coo frequently during the A condition, whereas other same-aged animals engage in little or no cooing. We have also observed large individual differences in the duration of NEC-induced freezing and ST-induced

110

hostility. Some animals freeze the entire length of the test period while at the other extreme, some never freeze and act relatively undisturbed by the human intruder. These individual differences in fear-related responses seen in the laboratory are similar to those we have observed in rhesus monkeys who inhabit Cay0 Santiago, a 45-acre island with approximately 800 free-ranging monkeys (unpublished data). Importantly, we have found that monkeys’ individual differences in the magnitude of defensive responses are relatively stable over time. Initially, we demonstrated that the duration of NEC-induced freezing behavior remained stable in 12 animals tested twice with an interval of four months (r = 0.94). Using a larger sample size, we found that NEC-induced freezing and ST-induced hostility are also relatively stable (unpublished data). Interestingly, within an animal we did not find significant correlations between the magnitude of the different types of defensive responses. Thus, monkeys that exhibited extreme levels of NECinduced freezing did not necessarily display extreme levels of ST-induced hostility. This lack of correlation between different types of defensive responses suggests that cooing, freezing, and defensive hostility represent different and somewhat unrelated characteristics of animals’ defensive styles. Along with the evidence for differential neurochemical regulation of these responses, these data support the contention that the different context-specific defensive responses may have different underlying neural substrates. Finally, to identify some of the mechanisms underlying the development of these individual differences in defensive responding, we examined the relationships between the stress-related hormone cortisol or asymmetric frontal EEG activity and individual differences in fearful behavior. Thus, in 28 mother-infant pairs, we found that in both mothers and infants, freezing duration was significantly correlated with baseline (non-stressed) cortisol levels, (mothers, rS = 0.53; p c 0.01; infants, rS = 0.62; p c 0,001) (Kalin et al., 1998b). Maternal cortisol levels ranged from approximately 13-50 p,g/dl; infant cortisol levels ranged from approximately 8-40 p,g/dl in these studies. These data are consistent with findings from human

studies demonstrating that extremely inhibited children have elevated levels of salivary cortisol (Kagan et al., 1987; Kagan et al., 1988), and is also consistent with findings in rodents that corticosterone (the rodent analogue of cortisol) is required for rat pups to develop the ability to freeze when threatened (Takahashi and Rubin, 1993). Because of the potential importance of cortisol in mediating the development of defensive responses, we examined factors that were expected to affect infant cortisol concentrations. Our study revealed that maternal cortisol levels were moderately correlated with those of their infants (r=0.34; p c 0.05) (Kalin et al., 1998b). Interestingly, it was also found that maternal parity was negatively correlated with infant cortisol levels (rS = - 0.55; pc0.01) such that the current infants of mothers that previously had more offspring were likely to have lower cortisol levels. This finding indicates that a mother’s past infant rearing andor pregnancy experience may contribute to individual differences in infant baseline cortisol levels, and provides further support for the notion that the mother-infant interactions may be a critical factor in determining the future fearful disposition of the offspring (Kalin et al., 19981.3).Although the precise mechanism for this interaction remains to be determined, one might imagine that mothers with little rearing experience would interact differently with their infants than mothers with more experience. Indeed, it has been shown in rodents that mothers that more actively lick and groom their pups produce adult offspring that have less reactive HPA systems, higher densities of hippocampal glucocorticoid receptor mRNA, and are less fearful (Liu et al., 1997). In adult humans, asymmetric right frontal brain activity has been associated with negative emotional responses (Tomarken et al., 1992). This pattern of brain activity also occurs in children with extreme behavioral inhibition (Davidson, 1992). Our studies of asymmetric frontal activity in rhesus monkeys have demonstrated similarities in this measure between monkeys and humans (Davidson et al., 1993). We performed studies to examine the hypothesis that monkeys with extreme right frontal electrical activity would have higher cortisol levels and would be more fearful when compared with

111

monkeys with extreme left frontal activity. It was found that individual differences in asymmetric frontal activity in the 4-8 Hz range are a stable characteristic of an animal (Kalin et al., 1998a; Davidson et al., 1993). We also found a significant positive correlation between relative right asymmetric frontal activity and basal cortisol levels in 50 one-year-old animals ( r = 0.41; p ~0.003).As predicted, the more right frontal an animal was, the higher was its cortisol level. An extreme groups analysis revealed that extreme right compared to extreme left frontal animals had greater cortisol concentrations as well as increased defensive responses, such as freezing and hostility. The association between extreme right frontal activity and increased cortisol appeared to be long-lasting because the right frontal animals continued to demonstrate elevated cortisol levels at three years of age. These findings are the first to link individual differences in asymmetric frontal activity with circulating levels of cortisol. This is important because both factors have been independently associated with fearful temperamental styles. The finding suggests that fearful temperament should be conceptualized as a constellation of hormonal, electrophysiological and behavioral characteristics.

Implications for psychopathology In humans, anxiety and fear are adaptive emotions that when experienced in appropriate contexts and at appropriate levels, help to recruit defensive behavioral responses that ensure one’s survival. However, when these responses are excessive, or are experienced as out of control and interfere with normal functioning, they cause significant distress and are considered pathological. The similarities in the expression of individual differences in fearful and defensive responding between monkeys and humans makes the aforementioned monkey studies relevant to understanding human anxiety disorders. As already mentioned, we have identified a wide range of individual differences in the magnitude of rhesus monkeys’ context-appropriate defensive responses that tend to be stable over time, and are thus thought of as trait markers. In monkeys, defensive responses that are overly intense but

appropriate for the context could result in performance deficits and related psychological distress and physiological dysregulation. For example, animals that freeze for very long durations are likely to be timid and socially avoidant. We have suggested that these animals are analogous to extremely behaviorally inhibited children who are exceedingly shy (Kalin and Shelton, 1989). This trait may not pose a problem for an infant monkey that is protected by its mother, but is likely to be disadvantageouswhen the monkey is older and challenged with more complex social situations. For example, on Cay0 Santiago, adolescent male monkeys leave their natal group attempting to join a new group of unfamiliar animals. Compared to their less inhibited counterparts, extremely inhibited animals might be expected to experience greater distress and have more difficulty in successfully emigrating to a new group. Demonstrations in the laboratory show that more inhibited animals already have higher levels of basal cortisol. Therefore, it is likely that when facing the stresses of leaving familiar environments and forming new relationships, ort tisol levels in inhibited animals would be markedly elevated. Later in life, chronic elevations in cortisol can have serious health effects (Sapolsky et al., 1985; McEwen and Stellar, 1993) and can also have deleterious effects on neurons. There is now a convincing body of rodent and primate data documenting the toxic effects of high levels of glucocorticoids on hippocampal neurons (Sapolsky et al., 1985; Uno et al., 1989; Sapolsky et al., 1990). In addition to overly intense responses, a small proportion of monkeys engage in defensive responses that are inappropriate for the environmental conditions they are confronting. Approximately 5% of the animals that freeze during NEC continue to freeze at considerable levels during ST. Interestingly, the animal’s level of freezing during NEC is not necessarily predictive of whether, or how much, the animal continues to freeze during ST. For example, some monkeys that continue to freeze during ST have high levels of freezing during NEC, the appropriate context for freezing. Other monkeys that freeze during ST freeze only moderately during NEC. Thus, some animals with high levels of freezing during NEC

112

are able to rapidly and effectively switch their adaptive responses as the environmental conditions change from the NEC to the ST condition whereas other animals are not able to do so. This contextinappropriate expression of freezing behavior (e.g. freezing during ST) could be due to a number of factors. In the Human Intruder paradigm, the NEC condition precedes ST, therefore it is possible that some animals have difficulty shutting off the freezing response. This would result in a failure to adaptively switch their behavior in accordance with the changing context. Alternatively, some animals may either have a limited repertoire of defensive responses or their propensity to freeze may be so strong that it is consistently expressed in threatening situations regardless of the environmental cues. It is also possible that these animals have an impaired ability to differentiate among cues that signal different types of environmental threats. Further studies must be performed to distinguish among the above possibilities, none of which are mutually exclusive. The inappropriate freezing could result in emotional and behavioral dysfunction that differs from that observed in animals with overly intense but context-appropriate responses. Under certain circumstances inappropriate freezing can have serious consequences. For example, when discovered by a predator, flight or defensive aggressions are protective responses which increase the likelihood of survival. Freezing is not helpful under this condition and likely increases the risk of injury and/or death from predatorial attack. In clinical studies, laboratory-induced anxiety responses are commonly elicited by contexts that reliably produce an anxious response in the majority of subjects tested. Inevitably, these are contexts in which the anxious behavior has an adaptive function, and experiments that utilize this approach are likely to be studying the expression of contextappropriate responses. It is unusual to perform tests examining the degree to which anxiety and fearrelated responses are context-independent or are inappropriately modulated. Our data suggests that this dimension (e.g. the expression of contextinappropriate defensive behaviors) should be added to clinical research strategies aimed at examining anxiety-related responses. It is possible that the overly intense but context-appropriate responses

are linked to different types of psychopathology and have different underlying neurobiological mechanisms than responses that are context-inappropriate. For example, in considering anxiety disorders, overly intense but context-appropriate responses might include social and simple phobias, as well as the cued and context-dependent symptoms associated with post-traumatic stress disorder. In contrast, such illnesses as generalized anxiety and panic disorder may be thought of as responses occurring out of context. In conclusion, we have comprehensively characterized the expression and regulation of defensive responses in rhesus monkeys. These behaviors are expressed early in life and monkeys are capable of regulating them by 3-4 months of age. Individual differences in these responses are relatively stable within animals and their magnitude is associated with basal cortisol levels and asymmetric frontal brain activity. In addition, the defensive responses are modulated by opiate, benzodiazepine and CRH systems. The Human Intruder paradigm, which challenges monkeys to regulate their defensive responses to changing environmental contexts, has revealed two types of extreme individual differences in defensive responding. Overly intense fear-related responses occur in the expected context but due to their magnitude could be considered maladaptive. Other animals express defensive responses in contexts that are clearly inappropriate. These observations have led us to conceptualize human anxiety disorders as the maladaptive expression of normal and adaptive emotional responses and to further classify them based on whether the anxiety/fearful response is context-appropriate or context-inappropriate. Furthermore, we predict that different neurobiological mechanisms may underlie these different types of maladaptive defensive responses. Future studies on the neural mechanisms of defensive behaviors will undoubtedly aid in further extending our understanding of the neural substrates of fear- and anxiety-related psychopathology.

Acknowledgments We gratefully acknowledge the contributions of Helen Van Valkenberg.This work was supported by

113

National Institute of Mental Health grants MH46792 and P50-MH52354, the Health Emotions Research Institute and Meriter Hospital.

References Berman, C. M. (1980) Mother-infant relationships among freeranging rhesus monkeys on Cay0 Santiago: a comparison with captive pairs. Anim. Behav., 28: 860-873. Biederman, J., Rosenbaum, J. F., Hirshfeld, D. R., Faraone, S. V., Bolduc, E. A., Gersten, M., Meminger, S. R., Kagan, J., Snidman, N., and Reznick, J. S. (1990) Psychiatric correlates of behavioral inhibition in young children of parents with and without psychiatric disorders. Arch. Gen. Psychiatry, 47: 21-26. Bowlby, J. (1973) Attachment and Loss, Vol. 11: Separation, Basic Books, New York. Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P. M., and Meaney, M. J. (1998) Maternal care during infancy regulates the development of neural systems mediating the expression of fearfulness in the rat. Proc. Nut. Acad. Sci. USA, 95: 5335-5340. Carlson, M. and Earls, F. (1997) Psychological and neuroendocrinological sequelae of early social deprivation in institutionalized children in Romania. Ann. NY Acad. Sci., 807: 419428. Curio, E. (1976) The Ethology of Predation. Springer-Verlag, New York. Davidson, R. J. (1992) Anterior cerebral asymmetry and the nature of emotion. Brain Cogn., 20: 125-151. Davidson, R. J., Kalin, N. H., and Shelton, S. E. (1993) Lateralized response to diazepam predicts temperamental style in rhesus monkeys. Behav. Neurosci., 107: 1106-1 110. Fabre-Nys, C., Meller, R. E., and Keverne, E. B. (1982) Opiate antagonists stimulate affdiative behaviour in monkeys. Pharmacol. Biochem. Behav., 16: 653-659. Ficken, M. S. (1990) Acoustic characteristics of alarm calls associated with predation risk in chickadees. Anim. Behav., 39: 400-401. Ficken, M. S. and Witkin, S. R. (1977) Responses of blackcapped chickadees to predators. Auk, 94: 156-157. Frank, D. A., Klass, P. E., Earls, F., and Eisenberg, L. (1996) Infants and young children in orphanages-one view from pediatrics and child psychiatry. Pediatrics, 97: 569-578. Goldman-Rakic, P. S. (1987) Development of cortical circuitry and cognitive function. Child Dev., 58: 601-622. Goldman-Rakic, P. S., Selemon, L. D., and Schwartz, M. L. (1984) Dual pathways connecting the dorsolateral prefrontal cortex with the hippocampal formation and parahippocampal cortex in the rhesus monkey. Neuroscience, 12: 719-743. Grigoriadis, D. E., Dent, G. W., Turner, J. G., Uno, H., Shelton, S. E., De Souza, E. B., and Kalin, N. H. (1995) Corticotrophin-releasing factor (CRF) receptors in infant rhesus monkey brain and pituitary gland: biochemical characterization and autoradiographic localization. D o . Neurosci., 17: 357-367.

Harlow, H. F. (1959) Love in infant monkeys. Sci. Am., 200: 68-74. Harlow, H. F., and Harlow, M. K. (1965) The affectional systems. In: A. Schrier, H. Harlow, and F. Stollnitz (Eds), Behavior of Non-human Primates, Academic Press, New York. Harlow, H. F., Rowland, G. L., and Griffin, G. A. (1964) The effect of total social deprivation on the development of monkey behavior. Psychiat. Res. Rep., 19: 116-135. Harlow, H. F., and Suomi, S. J. (1970) Induced psychopathology in monkeys. Eng. Sci., 33: 8-14. Harlow, H. F., and Zimmerman, R. R. (1959) Affectional responses in the infant monkey. Science, 130: 421-432. Hinde, R. A., and White, L. E. (1974) Dynamics of a relationship: Rhesus mother-infant ventro-ventral contact. J. Compara. Physiol. Psychol., 86: 8-23. Hirshfeld, D. R., Rosenbaum, J. F., Biederman, J., Bolduc, E. A., Faraone, S. V., Snidman, N., Reznick, J. S., and Kagan J. (1993) A 3-year follow-up of children with and without behavioral inhibition. J. Am. Acad. Child Adolescent Psychiatry, 32: 814-821. Hofer, M. A. (1996) Multiple regulators of ultrasonic vocalizations in the infant rat. Psychoneumendocrinology, 21 : 203-2 17. Kagan, J., Reznick, J. S., and Snidman, N. (1987) The physiology and psychology of behavioral inhibition in children. Child Dev., 58: 1459-1473. Kagan, J., Reznick, J. S., and Snidman, N. (1988) Biological bases of childhood shyness. Science, 240: 167-171. Kalin, N. H. (1993) The neurobiology of fear. Sci. Am., 268: 94-101. Kalin, N. H., and Carnes, M. (1984) Biological correlates of attachment bond disruption in humans and non-human primates. Prog. Neuro-Psychophamacol. Biol. Psychiatry, 8: 459469. Kalin, N. H., Larson, C., Shelton, S. E., and Davidson, R. J. (1998a) Asymmetric frontal brain activity, cortisol, and behavior associated with fearful temperaments in rhesus monkeys. Behav. Neurosci., 112: 286-292. Kalin,N. H., and Shelton, S. E. (1989) Defensive behaviors in infant rhesus monkeys: environmental cues and neurochemical regulation. Science, 243: 1718-1721. Kalin, N. H., and Shelton, S. E. (1998) The Ontogeny of separation and threat-induced defensive behaviors in rhesus monkeys during the first year of life. Am. J. Primatol., 44: 125-1 35. Kalin, N. H., Shelton, S. E., and Barksdale, C. M. (1988) Opiate modulation of separation-induced distress in nonhuman primates. Brain Res., 440: 285-292. Kalin, N. H., Shelton, S. E., and Barksdale, C. M. (1989) Behavioral and physiologic effects of CRH administered to infant primates undergoing maternal separation. Neuropsychopharmacology, 2: 97-104. Kalin, N. H., Shelton, S. E., Kraemer, G. W., and McKinney, W. T. (1983) Corticotropin-releasing factor administered intraventricularly to rhesus monkeys. Peptides, 4: 217-220.

114 Kalin, N. H., Shelton, S. E., and Lynn, D. E. (1995) Opiate systems in mother and infant primates coordinate intimate contact during reunion. Psychoneuroendocrinology, 20: 735-742. Kalin, N. H., Shelton, S. E., Rickman, M., and Davidson, R. J. (1998b) Individual differences in freezing and cortisol in infant and mother rhesus monkeys. Behav. Neurosci., 112: 25 1-254. Kalin, N. H., Shelton, S. E., and Takahashi, L. K. (1991a) Defensive behaviors in infant rhesus monkeys: ontogeny and context-dependent selective expression. Child Dev., 62: 1175-1 183. Kalin, N. H., Shelton, S. E., and Turner, J. G. (1991b) Effects of alprazolam on fear-related behavioral, hormonal and catecholamine responses in infant rhesus monkeys. Life Sci., 49: 2031-2044. Kalin, N. H. and Takahashi, L. K. (1991) Animal studies implicating a role of corticotrophin-releasing hormone in mediating behavior associated with psychopathology. In: S. C. Risc (Ed.), Central Nervous System Peptide Mechanisms in Stress and Depression, American Psychiatric Press, Washington, DC, pp. 53-72. Kaufman, I. C., and Rosenblum, L. A. (1967) The reaction to separation of infant monkeys: anaclitic depression and conservation-withdrawal. Psychosom. Med., 29: 648-675. Kehoe, P., and Blass, E. M. (1986) Opioid-mediation of separation distress in 10-day-old rats: Reversal of stress with maternal stimuli. Dev. Psychobiol., 19: 385-398. Keveme, E. B., Martensz, N. D., and Tuite, B. (1989) Betaendorphin concentrations in cerebrospinal fluid of monkeys are influenced by grooming relationships. Psychoneuroendocrinology, 1 4 155-161. Koob, G. F., Heinrichs, S. C., Pich, E. M., Menzaghi, F., Baldwin, H., Miczek, K., and Britton, K. T. (1993) The role of corticotropin-releasing factor in behavioural responses to stress. Ciba Found. Symp., 172: 277-289. Kuhn. C. M., and Schanberg, S. M. (1998) Responses to maternal separation-mechanisms and mediators. Int. J. Dev. Neurosci., 16: 261-270. Levine, S. (1957). Infantile experience and resistance to physiological stress. Science, 126: 405406. Liu, D., Dorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., S h m a , S., Pearson, D., Plotsky, P., and Meaney, M. (1997) Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-adrenal response to stress. Science, 277: 1659-1662. Magurran, A. F., and Girling, S. L. (1986) Predator model recognition and response habituation in shoaling minnows. Anim. Behav., 34: 510-518. Martel, F. L., Nevison, C. M., Rayment, F. D., Simpson, M. J. A., and Keverne, E. B. (1993) Opioid receptor blockade reduces maternal affect and social grooming in rhesus monkeys. Psychoneuroendocrinology,18: 307-32 1. Matthews, K., Wilkinson, L. S., and Robbins, T. W. (1996) Repeated maternal separation in preweanling rats attenuates behavioral responses to primary and conditioned incentives in adulthood. Physiol. Behav., 59: 99-107.

McEwen, B.S. and Stellar, E. (1993) Stress and the individual. Mechanisms leading to disease. Arch. Intern. Med., 153: 2093-2101. McKinney, W. T., and Bunney, W. E. (1969) Animal model of depression. I. Review of evidence: implications for research. Arch. Gen. Psychiatry, 21: 240-248. Mendelson, M. J., Haith, M. M., and Goldman-Rakic, P. S. (1982) Face scanning and responsiveness to social cues in infant rhesus monkeys. Dev. Psychology, 18: 222-228. Mick, M. A. and Telch, M. J. (1998) Social anxiety and history of behavioral inhibition in young adults. J. A m . Dis., 12: 1-20. Morse, D. H. (1993) Interactions between tit flocks and spmowhawks. Accipiter nisus Ibis, 115: 591-593. Newman, J. D. (1985) The infant cry of primates: an evolutionary perspective. In B. M. Lester, and C. F. Zachariah Boukydis (Eds), Infant Crying, Plenum, New York. Noirot, E. (1972) Ultrasounds and maternal behavior in small rodents. Dev. Psychobiol., 5 : 371-387. Palmer, W. (1909) Instinctive stillness in birds. Auk, 26: 23-36. Panksepp, J., Herman, B. H., Vilberg, T., Bishop, P., and DeEskinazi, F. G. (1978) Endogenous opioids and social behavior. Neurosci. Biobehav. Rev., 4: 473487. Plotsky, P. M., and Meaney, M. J. (1993) Early, postnatal experience alters hypothalamic corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stressinduced release in adult rats. Mol. Bruin Res., 18: 195-200. Pollock, R. A., Rosenbaum, J. F., Marrs, A,, Miller, B. S., and Biederman, J. (1995) Anxiety disorders of childhood, Implications for adult psychopathology. Psychiatric Clinics N. Am., 18: 745-766. Reite, M. (1977) Maternal separation in monkey infants: a model of depression. In: I. Hanin and E. Usdin (Eds), Animal Models in Psychiatry and Neurology, Pergamon Press, Oxford, UK, pp. 127-140. Rosenbaum, J. F., Biederman, J., Bolduc-Murphy, E. A., Faraone, S. V.,Chaloff, J., Hirshfeld, D. R., and Kagan, J. R. (1993) Behavioral inhibition in childhood: a risk factor for anxiety disorders. Harvard Rev. Psychiatry, 1: 2-16. Rosenbaum, J. F., Biederman, J., Gersten, M., Hirshfeld, D. R., Meminger, S. R., Herman, J. B., Kagan, J., Reznick, J. S.. and Snidman, N. (1988) Behavioral inhibition in children of parents with panic disorder and agoraphobia. Arch. Gen. Psychiatry, 45: 463470. Rosenblum, L. A. and Paully G. S. (1987) Primate models of separation-induced depression. psychiatric Clin. N. Am., 10: 437447. Sackett, G. P. (1966) Monkeys reared in isolation with pictures as visual input: evidence for an innate releasing mechanism. Science, 154: 1468-1473. Sapolsky, R. M. (1997) The importance of a well-groomed child. Science, 277: 1620-1621. Sapolsky, R. M., Krey, L. C., and McEwen, B. S. (1985) Prolonged glucocorticoid exposure reduced hippocampal

115 neuron number: Implications for aging. J. Neurosci., 5: 1222-1227. Sapolsky, R. M., Uno, H., Rebert, C. S., and Finch, C. E. (1990) Hippocampal damage associated with prolonged glucocorticoid exposure in primates. J. Neurosci., 9: 2897-2902. Schaller, G. B. (1972) The Serengeti Lion, University of Chicago Press, Chicago. Suomi, S. J. (1986) Anxiety-like disorders in young nonhuman primates. In: R. Gittleman (Ed.), Anxiety Disorders of Childhood, Guilford Press, New York. Suomi, S. J., Harlow, H. F,, and McKinney, W. T. (1972) Monkey psychiatrists. Am. J. Psychiatry, 128: 927-932. Takahashi, L. K.(1992) Developmental expression of defensive responses during exposure to conspecific adults in preweanling rats ( R a m s norvegicus). J. Comp. Psychol., 106: 69-77.

Takahashi, L. K., and Rubin, W. W. (1993) Corticosteroid induction of threat-induced behavioral inhibition in preweanling rats. Eehav. Neurosci., 107: 860-866. Tomarken, A. J., Davidson, R. J., Wheeler, R. E., and Kinney, L. (1992) Individual differences in anterior brain asymmetry and fundamental dimensions of emotion. J. Personality Social Psychol., 62: 676-687. Uno, H., Tarara, R., Else, J. G., Suleman, M. A., and Sapolsky, R. M. (1989) Hippocampal damage associated with prolonged and fatal stress in primates. J. Neurosci., 5 : 1705-1711. Vale, W., Speiss, J., Rivier, C., and Rivier, J. (1981) Characterization of a 41-residue ovine hypothalamic peptide that stimulates secretion of corticotropin and beta-endorphin. Science, 213: 1384-1397.

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 9

Effects of perinatal pain and stress K.J.S. Anand Pain Neurobiology Laboratory, University of Arkansas for Medical Sciences and Arkansas Children S Hospital, Little Rock, AR 72202, USA

Introduction Surgical operations without adequate anesthesia were performed routinely in newborn infants less than a decade ago. Widespread medical beliefs supported this practice, which questioned the sensory capability of pain perception in neonates and may have exaggerated the risks of complications or side effects from the use of anesthetic drugs in newborns (Anand and Hickey, 1987). Some of these practices changed after randomized clinical trials of potent vs. inadequate anesthesia showed significant reductions in the incidence of postoperative complications following major surgery in preterm and term neonates (Anand and Aynsley-Green, 1985; Anand et al., 1987; Anand et al., 1988). These data stimulated systematic changes in the clinical practice of neonatal anesthesia (Rogers, 1992) and a greater scientific interest in the development of pathways and mechanisms associated with pain perception in early life. Clinical and experimental research have contributed to a major reorientation in our understanding of the developmental biology of pain (Fitzgerald et al., 1988; Anand and Can, 1989; Craig et al., 1993; Johnston et al., 1993; Fitzgerald, 1994), and have stimulated a re-evaluation of the definition of pain (Anand and Craig, 1996).

*Corresponding author. Tel.: 501-320-1008; F ~ x 50 : 1-320-3188; e-mail: [email protected]

Complementary investigations in preterm neonates and neonatal rat pups have identified a predominant role for early pain and stress in altering neonatal clinical outcomes, brain development and subsequent behavior. Much of these data were obtained from infant rats, because of similar pain pathways and mechanisms in the human and rodent species and similar neurological maturity of the rat pup at birth and the viable preterm neonate at 23-24 weeks gestation. Based on these data, we propose the following four hypotheses: (a) the perinatal period is associated with an increased sensitivity to pain, mediated by immature peripheral and central mechanisms in the developing pain system; (b) a hyperalgesic state occurs following acute painful stimuli, and the duration of this physiological state is prolonged in preterm neonates and newborn rat pups; (c) the exposure to multiple invasive procedures required for the resuscitation and clinical management of preterm neonates after birth stimulates acute physiologic and behavioral responses and increases their vulnerability to gross neurologic damage (intraventricular hemorrhage and/or periventricular leukomalacia); (d) the plasticity of the developing pain system provides a critical window for producing longterm changes in subsequent behavior, responses to stress, and susceptibility to psychosomatic complaints and psychiatric disorders in later life.

118

Increased pain sensitivity in neonates The traditional view, supported by the early experiments of McGraw and others, held that newborn infants were relatively insensitive to pain (McGraw, 1941). More recently, thresholds for the withdrawal reflex were investigated using calibrated Von Frey hairs to stimulate the dorsal cutaneous flexor withdrawal reflex in preterm and term newborn infants. These thresholds correspond with the pain threshold and are inhibited by opioid or other analgesics (Woolf and Wall, 1986). In marked contrast to earlier findings, Fitzgerald and colleagues found that flexor reflex thresholds were directly related to increasing gestational age and postnatal age, with significantly lower thresholds in preterm neonates than in term neonates (Fitzgerald et al., 1988). Lower thresholds and exaggerated behavioral responses to noxious stimuli were also noted in neonatal rat pups as compared with older pups or adult rats. Newborn rat pups at less than one week

of age commonly responded with limb flexion, shaking and licking of the stimulated paw, decreased sleep, squirming and kicking; these responses were transient after saline injection and prolonged after formalin injection (McLaughlin et al., 1990; Guy and Abbott, 1992). Other specific responses included kicking and jerking in 3 to 15-day old pups. The overall intensity and duration of these behavioral responses decreased during development, such that only the youngest rats exhibited extreme responses, or whole body convulsions, or noxious responses to injected saline (McLaughlin et al., 1990; Guy and Abbott, 1992;Yi and Ban, 1995). When formalin concentrations were increased in the formalin pain test (Teng and Abbott, 1998), pain behaviors in 50% of the rats tested were noted at formalin concentrations of 0.52% for 3-day old rats, 1.21% for 15-day old pups, 4.56% for 25-day old rats, and 5.60% for adult rats. These data suggest that the thresholds for inflammatory pain increased 2.5-fold from 3 days to 15 days of age,

TABLE 1 Mechanisms supporting increased pain sensitivity in neonates Components of Pain System

Developmental processes in the newborn rat pup

Cutaneous receptors: polymodal nociceptors high-threshold mechanoceptors Cutaneous nerve terminals

mature soma membrane properties (membrane thresholds, firing frequencies, neuro- transmitter expression) in the late fetal rat and at birth immature soma membrane properties (low thresholdsflow frequencies) superficial plexus of nerve terminals, becomes subepidermal with development of the stratum corneum, density regulated by NGF expression

Skin or tissue injury

marked hyperinnervation of neonatally injured skin, decreased threshold

C-fiber terminals in the dorsal horn A-fiber terminals in the dorsal horn

delayed entry into the substantia gelatinosa, synaptogenesis at 3 weeks early synaptogenesis in laminae 3-5,collaterals to substantia gelatinosa

Nociceptive neurons in the substantia gelatinosa (laminae 1, 2)

large receptive fields, marked overlapping of adjacent neurons low thresholds, early expression of excitatory neurotransmitters/receptors delayed development of descending inhibition (5-HT, NE, dopamine) single stimuli evoke long-lasting excitation which lowers the threshold to subsequent stimuli; repeated stimuli cause immense background activity

Dorsal horn substance P receptors NMDA receptors in the dorsal horn

six-fold greater density as compared to the adult, diffuse localization peak density at 6-10 days age, increased affinity for NMDA & [Ca"] release fast inhibitory transmitters in the adult spinal cord actually depolarize immature spinal neurons and cause increases in intracellular [Ca"]

Inhibitory transmitters: GABA, glycine Somatosensory cortex pyramidal cells

large receptive fields which may be due to a lack of surround inhibition

119

9-fold to 25 days, and 11-fold from the neonate to the adult rat (Teng and Abbott, 1998). Stimulation of the dorsal flexor reflex with calibrated Von Frey hairs also demonstrated lower thresholds to mechanical stimulation in newborn rat pups as compared with older pups or adults (Fitzgerald et al., 1988; Andrews and Fitzgerald, 1994). Similar findings were noted for the nociceptive thresholds to thermal stimulation using latencies to paw withdrawal from a hot plate (Hu et al., 1997) or tail withdrawal from a heated water bath (Falcon et al., 1996), with thermal pain thresholds noted at 37.5"C, 46.5"C, and 47.2"C for 3-day-old, 15-dayold, and adult rats respectively (Falcon et al., 1996). Mechanisms underlying the increased pain sensitivity during early development are summarized in Table 1. Proposed mechanisms for the low pain thresholds noted in newborn rats and preterm neonates include immature inhibitory interneurons within the dorsal horn and the lack of descending inhibitory controls from supraspinal centers (Bicknell and Beal, 1984; Fitzgerald and Anand, 1993). Synaptogenesis of incoming C-fibres with dorsal horn cells coincided with decreases in the size of their receptive fields, development of inhibitory interneurons, withdrawal of the AG-fiber terminals from the substantia gelatinosa, and changes in the expression of excitatory and inhibitory neurotransmitters (Marti et al., 1987; Pignatelli et al., 1989). These developmental changes mostly occur during the two weeks after birth in newborn rat pups and in the last trimester of human fetal development. This period also coincides with the development of spinothalamic and thalamocortical connections, which follow the general pattern of mature excitatory mechanisms before maturation of the inhibitory mechanisms (Dani et al., 1991). Support for the concept of increased pain sensitivity in newborns may also be inferred from other clinical studies. An increased magnitude of hormonal, metabolic or cardiovascular responses to surgical operations was noted in preterm and term neonates (Anand et al., 1985; Anand, 1990) and higher plasma concentrations of analgesics and anesthetics were required to produce analgesia or anesthesia in neonates as compared with older age groups (Yaster, 1987; Greeley and de Bruijn, 1988; Chay et al., 1992).

Prolonged hyperalgesia following acute painful stimuli Afferent impulses in the late embryonic and early postnatal period produced a long-lasting excitation of dorsal horn cells, and repetitive stimuli caused considerable background activity in areas of the spinal cord above and below the site of stimulation (Fitzgerald, 1985). The exaggerated nociceptive reflexes, large receptive fields of dorsal horn cells, and prolonged excitation following stimulation in newborn rats parallel the low pain thresholds and sensory hypersensitivity noted in preterm newborn infants. Repeated stimulation and local tissue injury were associated with prolonged periods of sensory hyperalgesia in neonatal rat pups or preterm neonates (Fitzgerald, 1987; Fitzgerald et al., 1988), which was prevented by applying topical analgesia to areas of local tissue injury (Fitzgerald et al., 1989). Exposure to an acute painful stimulus, such as heel lancing, causes 'windup' or an increased excitability of nociceptive neurons in the substantia gelatinosa of the dorsal horn, with further increases in the sensitivity of neonates to subsequent painful stimulation (Fitzgerald et al., 1989; Basbaum, 1996). This phenomenon is mediated by NMDA and tachykinin receptor activity, and spreads in nociceptive neurons of the dorsal horn above and below the spinal level innervating the area of tissue damage. Windup was associated with increases in the receptive area of the stimulated neurons and the duration of windup was prolonged in neonatal as compared with adult animals (Fitzgerald et al., 1988; Fitzgerald et al., 1989; Fitzgerald et al., 1991). The importance of windup in human newborns can be estimated from recent clinical data. Preterm neonates receiving intensive care are exposed to multiple, randomly occurring invasive procedures, interspersed with several other, nonnoxious stimuli (handling, physical examination, nursing procedures, etc.). Barker and Rutter (1995) recorded 488 invasive procedures during the hospital stay of one preterm infant (23 weeks gestation, birthweight 560 g), and a longitudinal study noted an average of 766 invasive procedures experienced by preterm neonates (Porter, 1998). Invasive proce-

120

dures stimulate heightened activity in the immature nociceptive pathways of preterm neonates which is further prolonged by physical handling, nursing care, or mechanical ventilation, thus producing a chronic noxious stimulation and prolonged physiologic stress (Anand, 1993; McIntosh, 1997). In addition to invasive procedures, sensitization of the CNS may occur following repeated physical handling of preterm neonates as well. For example, baseline physiological and behavioral measures in preterm neonates exposed to physical handling were no different from unhandled neonates, yet the handled neonates developed greater tachycardia in response to a heel lance as compared with unhandled preterm neonates (Porter et al., 1998). There were no obvious clinical signs to diagnose the central sensitization caused by physical handling. Thus, prolonged activity may occur in the immature nociceptive pathways of preterm neonates receiving neonatal intensive care, thereby exposing them to chronic noxious stimulation for several days and weeks (Anand and McGrath, 1993; McIntosh, 1997). Research attention focused on the assessment and management of chronic pain and stress may have a much greater biological and clinical significance for the care and outcomes of prematurity.

Exposure to multiple invasive procedures There is an increasing awareness amongst physicians and nurses regarding pain perception in neonates and that that repetitive pain and stress during intensive care may worsen the clinical outcomes of preterm neonates (Barker and Rutter, 1996; Anand et al., 1998). However, clinical practices incorporating neonatal analgesia are rare and nascent. In answering recent questionnaires, almost all physicians and nurses felt that infants experience the same degree or greater degrees of pain than adults (Tohill and McMorrow, 1990; McLaughlin et al., 1993; Porter et al., 1997). These results are in marked contrast to their professional opinions 10 years ago, when clinicians thought that neonates were incapable of experiencing pain (Franck, 1987; Purcell-Jones et al., 1988; Bauchner et al., 1992; McRae et al., 1997). Differences in the beliefs about current and optimal pain management

amongst neonatal physicians and nurses were determined by their professional training, years of clinical experience and a history of personal pain experiences (Porter et al., 1997). There were large discrepancies between how often caregivers believed that analgesic interventions should be performed and how often the interventions were actually performed. Neonatal nurses did not provide comfort measures despite believing that they should (Porter et al., 1997). In order to document these differences between opinion and practice, the actual use of analgesia and sedation at the bedside of all neonates admitted to the Neonatal Intensive Care Unit (NICU) were studied. A total of 1068 neonates admitted to 109 NICUs in the USA and 239 neonates from 14 NICUs in Canada were studied prospectively during a one-week period (Anand et al., 1996; Johnston et al., 1997). The results revealed that procedure-related pain was the most common type of pain experienced, but was rarely treated ( < 2%). In contrast, postoperative pain in NICUs appeared to be consistently treated in >75% neonates, primarily with opioid analgesics. Analgesia or sedation to treat the ongoing paidstress associated with intensive care was prescribed to 26.5% of the neonates. Despite the recent advances in our understanding of neonatal pain assessment, the clinical impact of repetitive pain, and the neurobiology of pain, these studies indicate that most human neonates continue to receive inadequate pain management.

Increased vulnerability to early neurologic injury Painful procedures in neonates are generally associated with diaphragmatic splinting, forced expiratory movements (crying), and sympathetic activation leading to tachycardia and hypertension. Changes in the intrathoracic pressure of ventilated neonates cause substantial oscillations in the intracranial pressure, cerebral oxygen delivery (Pokela, 1994) and intracranial blood volume (Perlman and Thach, 1988; Zernikow et al., 1994). The magnitude and rapidity of these physiological changes is sufficient to cause or extend venous hemorrhage into the germinal matrix or brain parenchyma

121

(Ghazi-Birry et al., 1997) and produce the ischemidreperfusion injury associated with periventricular leukomalacia (Evans and Kluckow, 1998). More than 95% of all intraventricular hemorrhage occurs within five days after birth, and hemorrhage develops within the first 24 hours after birth in more than half of these preterm neonates (Vohr and Ment, 1996). After an initial intraventricular hemorrhage, the recurrence of intracranial bleeding may occur in 1045% of preterm neonates, and forms an important predictor for poor neurologic outcomes and early mortality (Abdel-Rahman and Rosenberg, 1994; Wells and Ment, 1995). The long-term functional impact of early neurologic injury may be ameliorated during later development because of the plasticity of developing neurons. The vagal bradycardia and hypertension caused by repeated invasive procedures after birth were correlated with marked changes in cerebral blood flow and oxygen delivery (Ramaekers et al., 1993; Zernikow et al., 1994; Haxhija et al., 1995), as well as the sonographic findings predictive of cerebral palsy and other developmental sequelae (Low et al., 1992; Pinto-Martin et al., 1995). The cumulative effect of physiologic disruptions caused by venepuncture (Fiselier et al., 1983), tracheal suctioning (Greisen et al., 1985), feeding tube insertion, heel sticks, nursing procedures (Lagercrantz et al., 1986; Pokela, 1994), and mechanical ventilation (Greenough et al., 1987, 1990) serves to accentuate the vulnerability of preterm neonates to neurologic injury and other complications (Anand et al., 1987; Anand, 1998). Thus, morphine analgesia reduced the vulnerability of preterm neonates to early neurologic injury in a recent placebo-controlled trial (Anand et al., 1999a).

Long-term effects of pain and stress The perinatal period is recognized as a critical period of increased plasticity in the early development of most mammalian species, including the human. Perinatal synaptic activity is crucial because inactive synapses are solubilized and inactive neurons undergo apoptosis in early development (Rakic, 1985; Rabinowicz et al., 1996). Thus, repetitive exposure to early pain or stress

would cause more profound effects on brain development than similar experiences in later life (Bower, 1990; Leon, 1992; Philbin et al., 1994). Experimental studies on the long-term effects of neonatal stressors have focused on maternal separation rather than pain, although limited experimental data using pain paradigms are now available. In contrast, the recent interest in the long-term effects of neonatal pain has spawned an abundance of clinical data reviewed below. These studies can be arbitrarily divided into those investigating the longterm behavioral effects of neonatal pain and those investigating the long-term effects of neonatal drug therapy on subsequent development of the pain system.

Long-term effects of early stress in neonatal rat PUPS

Temporary separation of the mother from rat pups, causes an interruption of maternal care, tactile stimulation (licking, grooming, urogenital toilet), source of nutrition and warmth, and contact with littermates. Neonatal rat pups exposed to maternal separation for 15 min each day (from P2 to P14) developed an adult phenotype (Meaney et al., 1988, 1991; Viau et al., 1993) characterized by increased exploratory activity and decreased pituitary-adrenal responses to stress, resulting from an increased inhibition (negative-feedback) of the HPA axis (Ader and Grota, 1969; Meaney et al., 1994). Inhibition of the HPA axis was associated with increased expression of glucocorticoid receptors in the hippocampus, reduced basal and stress-related expression of CFW mRNA, lower plasma CRF and ACTH responses, and lower corticosterone responses (O’Donnell et al., 1994; Liu et al., 1997). Regular handling in the neonatal period also produced naloxone-reversible increases in the pain threshold, increased sensitivity to opioids, and reversed the locomotor hyperactivity induced by isolation (Gentsch et al., 1988; Pieretti et al., 1991; See also chapter by Ladd, et al. Chapter 7 this volume). In contrast, newborn rat pups separated for 180 minutes each day from P2 to P14 and then grown to adulthood showed significantly increased ACTH and corticosterone responses to stress as compared I

122

with handled and nonhandled animals (Plotsky and Meaney, 1993). Adult rats exposed to prolonged maternal separation as neonates had marked increases in CRF mRNA in the hypothalamic PVN, increased CRF content in the median stalk, and early escape from dexamethasone suppression tests; suggesting hypothalamic mechanisms for the increased ACTH and corticosterone responses to stress (Plotsky and Meaney, 1993). Glucocorticoid receptor density was significantly reduced in the hippocampus and median frontal cortex of adult animals exposed to neonatal stress, which promoted the increased HPA axis responses resulting from decreased feedback inhibition of the HPA axis (Liu et al., 1997). Hyper-responsiveness of glucocorticoid stress responses was associated with an accelerated loss of hippocampal neurons, early cognitive deficits, increased anxiety, depression, neophobia and fearfulness, hyperactivity, elevated pain thresholds, and a reduced life span of these rats (Gentsch et al., 1988; McEwen, 1994). The transition from low to high serum corticosterone concentrations within the range of physiologic circadian changes or from high basal to stressed values, increased the susceptibility of CA3 hippocampal neurons to excitotoxic damage (Stein-Behrens et al., 1994). Adult hypersecretion of CRF resulting from maternal deprivation in the neonatal period was noted in rodents (Caldji et al., 1998) and in non-human primates as well (Coplan et al., 1998). Clinical correlates of these experiments may be obtained from victims of sexual or physical abuse who are faced with stressful life events (Mullen et al., 1993; Goodyer, 1994). (See Chapter 10, this volume).

Long-Term effects of pain in neonatal rat pups Preliminary data suggest that invasive procedures associated with neonatal skin injury and repetitive pain may cause permanent changes in the peripheral and/or central pain systems. Skin injury occurring on the day of birth was associated with subsequent hyperinnervation from exuberant nerve sprouting in the area of injured skin. The degree of nerve sprouting in neonatal rats occurred earlier, was greater in magnitude, and lasted longer than the innervation following skin injury at 7 days, 14

days of age, or in adult rats (Reynolds and Fitzgerald, 1995). Thresholds for the flexor withdrawal reflex were reduced despite complete healing in the area of injured skin. These changes lasted a few weeks following skin injury in adulthood, and were permanent if the skin injury occurred in the neonatal period. Thus permanent changes in the pain threshold follow local hyperinnervation of skin injuries caused in the neonatal period, whereas these changes are relatively shortlived if the skin injury occurs during infancy (7 days, 14 days) or during adult life (Reynolds and Fitzgerald, 1995). Further experiments suggest that repetitive pain may lead to an altered central processing of pain and other behavioral effects. Neonatal rat pups stimulated with painful (needle prick) stimuli four times a day from birth to seven days age were noted to have significantly lower pain thresholds at 16 and 22 days of age, as compared to rat pups randomized to receiving tactile stimuli with the same schedule. During adulthood, pain stimulated rats had an increased preference for alcohol, increased anxiety and defensive withdrawal behavior, prolonged memory for chemosensory cues, and decreased expression of Fos in the somatosensory cortex following exposure to a hot plate (Anand et al., 1999b). Adult rats exposed to neonatal pain exhibited freezing behavior, autotomy, and learned helplessness when placed on the hot plate, whereas adult rats exposed to neonatal tactile stimulation showed marked exploratory and escape behavior. Thus, repetitive neonatal pain may cause decreased pain thresholds during development, leading to stress vulnerability and anxiety-mediated adult behavior, correlated with early-onset cognitive defects. Changes in the central processing of pain and stress may lead to these behavioral differences, although differences in pain threshold may also be related to cutaneous hyperinnervation following skin wounds in the neonatal period (Blass et al., 1995; Reynolds and Fitzgerald, 1995; Anand et al., 1999b).

Long-term effects of pain and stress in fullterm human neonates Few clinical studies have investigated the long-term effects of pain in full-term healthy infants, with

123

limited data suggesting the prolonged effects of neonatal pain on subsequent behavior. Altered behavior following circumcision was associated with poor orientation, decreased coordination of motor processes, impaired ability to regulate their behavioral states, and altered feeding or sleep patterns (Emde et al., 1971; Marshall et al., 1980; Dixon et al., 1984). These behavioral changes altered the adaptation of newborn infants to their postnatal environment, and persisted for several days following circumcision. Healthy term neonates who were born after a stressful delivery also showed increased behavioral responses to a subsequent neonatal blood draw and to vaccination pain at four and six months of age (Ramsay and Lewis, 1995). Gunnar et al. (1992), found that obstetric complications altered the neonatal responses to physical handling, but not to heel lancing. Adrenocortical responses were sensitized in term neonates receiving a second heel lance after 24 hours (Gunnar et al., 1991). Greater reactivity to a neonatal heel lance (more crying, increased heart rates, lower vagal tone, and higher cortisol) was subsequently associated with lower scores on the ‘distress-to-limitations’temperament test at six months age. This finding was consistent with the expectation that the capacity to react strongly to an aversive stimulus reflects greater neurobehavioral organization in the newborn. Recovery measures of cardiac activity were correlated with baseline measures indicating the strong self-righting properties of the healthy newborn (Gunnar et al., 1995). Full-term male neonates who were subjected to unanesthetized circumcision subsequently showed a stronger behavioral pain response to routine vaccination at four or six months of age compared with the responses of uncircumcised boys (Taddio et al., 1995, 1997). Some of these differences were attenuated in neonates who received topical anesthesia (EMLA) during the neonatal circumcision (Taddio et al., 1997). Because a multitude of other factors (environmental, social, maternal, genetic, co-morbidity) can exert their effects during early infancy, it is questionable whether neonatal pain can explain these long-term changes. Epidemiological studies relating stressful events at birth with adult self-destructive behavior demonstrated the

impact of such neurobiological changes on subsequent development (Jacobson et al., 1987, 1998). Biological mechanisms for ‘imprinting’ were implicated to explain these findings, although our current understanding of the long-term effects of early paidstress in term neonates is limited. Preliminary epidemiological data also suggests that major psychoses, such as schizophrenia, may be related to the neurobiological changes caused by fetal or neonatal stress or adverse conditions in the perinatal period (Cannon and Murray, 1998). These clinical data suggest the importance of further research on the long-term neurobiological and behavioral effects of neonatal pain and stress.

Long-term effects of pain or stress in premature infants Follow-up studies of ex-premature infants in their early school years have reported an increased incidence of non-specific pain symptoms, decreased affective responses to subsequent pain, neurological and developmental deficits, social difficulties, cognitive and learning defects. Based on parent report, 18-month old ex-preterm neonates who were exposed to multiple noxious stimuli while in the NICU, were noted to be underresponsive to subsequent pain (Grunau et al., 1994a). Follow-up of older age groups (4-5 years) showed clinically abnormal somatization and other behaviors in 25% of ex-preterm neonates (Grunau et al., 1994b; Sommerfelt et al., 1996). In later childhood (8-10 years age) these children showed accentuated emotional responses to pictures depicting painful events as compared to ex-full-term peers (Grunau et al., 1998). The length of NICU stay was correlated with increased somatization (Grunau et al., 1994b) and the emotional responses to depicted pain (Grunau et al., 1998), indicating that these findings were related to some cumulative experiences associated with intensive care. Of these experiences, painful stimuli were perhaps the most frequent, physiologically disruptive, and perceptually prominent as compared to all other sensory inputs. An etiological role for repetitive pain in producing these long-term changes is supported by changes in the pattern of neonatal responses associated with exposure to four weeks of NICU

124

care. The number of invasive procedures experienced by each neonate (Johnston and Stevens, 1996) most significantly predicted changes in the pattern of physiological and behavioral responses to a heel lance in preterm neonates. Ex-premature infants were also found to have more educational, behavioral and emotional difficulties during school age (Hack et al., 1994; Whitfield et al., 1997) and adolescence as compared with their peers (Levy-Shiff et al., 1994; Botting et al., 1997). From a follow-up of 137 expreterm neonates at the age of 12 years, 23% were noted to have attention deficit-hyperactivity disorder, 34% had depressive symptoms, 8% had generalized anxiety, with increased behavioral difficulties based on parent and teacher ratings (Botting et al., 1997). Similar behaviors were noted in adult rats exposed to repetitive pain or stress during their neonatal period (Liu et al., 1997; Anand et al., 1999b), supporting the possibility that these behaviors may be linked to the multiple pain exposures during routine NICU care. However, other factors related to repeated stressful experiences in the NICU cannot be ruled out.

Long-term effects of analgesia or comfort measures Selective pharmacological interventions may reduce the adverse consequences of pain in human neonates (Fitzgerald et al., 1989; Taddio et al., 1997; Anand et al., 1999a). In a blinded, randomized multicenter trial, low-dose continuous infusions of morphine significantly decreased the composite incidence of neonatal death and severe neurologic damage (defined as grade III or IV intraventricular hemorrhage, or periventricular leukomalacia) (Anand et al., 1999a). These differences may result from decreased stress responses (Pokela, 1993; Quinn et al., 1993; Als et al., 1994; Barker and Rutter, 1996), increased blood pressure stability (Goldstein and Brazy, 1991;Quinn et al., 1992), greater ventilator synchrony (Goldstein and Brazy, 1991; Barker et al., 1995; Dyke et al., 1995), and improved oxygenation (Pokela, 1994; Dyke et al., 1995) in the neonates given morphine. Intravenous opioid infusions were associated with decreased behavioral and hormonal responses of ventilated

preterm neonates in previous clinical trials (Barker and Rutter, 1996; Orsini et al., 1996). These findings were correlated with a decreased neonatal mortality in one study (Barker and Rutter, 1996) and prolonged mechanical ventilation in the other (Orsini et al., 1996). Other stress relieving interventions were also correlated with a decreased incidence of neurologic sequelae in preterm neonates (Anand, 1998). Preterm neonates exposed to maternal opioid intake in utero and postnatally treated for opioid withdrawal were noted to have a lower incidence of intraventricular hemorrhage than matched controls (Cepeda et al., 1987). The alterations in arterial blood pressure caused by diaphragmatic splinting were related to early intraventricular hemorrhage (Perlman and Thach, 1988), these diaphragmatic movements and the incidence of hemorrhage were reduced by neuromuscular paralysis in preterm neonates (Perlman et al., 1985). Nursing interventions designed to minimize stress in preterm neonates and support their neurobehavioral development was also associated with a decreased incidence of intraventricular hemorrhage and improved developmental outcomes (Becker et al., 1991; Als et al., 1994). Prevention of severe intraventricular hemorrhage by early indomethacin therapy was related to its hemodynamic effects (Ment et al., 1994, 1996), although its potent analgesic effects (Sims et al., 1994) must also be contributory. The most frequent invasive procedure performed in preterm neonates is the heel lance for blood sampling. Application of a topical anesthetic cream (EMLA, a eutectic mixture of 2.5% lidocaine and 2.5% prilocaine) to the heel tissue damaged by repetitive heel lancing reversed the cutaneous hypersensitivity and lower pain thresholds as compared to a placebo cream (Fitzgerald et al., 1989). Pre-treatment with EMLA, prior to neonatal circumcision, was also found to lessen the pain response to subsequent routine vaccinations at 4-6 months (Taddio et al., 1997), though it was not as effective as dorsal penile block or ring block (Lander et al., 1997). These data suggest that the clinical benefits of optimal assessment and management of neonatal pain may persist well beyond the neonatal period. Confirmation of recent find-

125

ings from randomized analgesia trials may portend major improvements in the neurologic outcomes of preterm neonates.

Summary Neonatal intensive care exposes preterm neonates to a series of repeated, randomly occurring invasive procedures and handling, resulting in acute pain, chronic pain, and prolonged stress during a critical window associated with epochal brain development. Characteristics of the immature pain system in preterm neonates (such as a low pain threshold, prolonged periods of windup, overlapping receptive fields, immature descending inhibition) predisposes them to greater clinical and behavioral sequelae from inadequately treated pain than older age groups. Evidence for developmental plasticity in the neonatal brain suggests that repetitive painful experiences during this period or prolonged exposure to analgesic drugs may alter neuronal and synaptic organization permanently. Traditionally, clinicians have chosen the perspective that routine use of analgesic or sedative drugs in preterm neonates may create more problems than minimal therapy. However, the immediate and long-term consequences of inadequately treated pain have forced them to reconsider the risk-benefit ratios for such therapy. Whereas the short-term consequences of prolonged analgesic therapy in human neonates are well-known (tolerance, withdrawal, ventilator dependency), long-term consequences are relatively unknown. Advances in the study of repetitive pain associated with routine NICU care have challenged the perspective that prolonged pain and stress were inevitable consequences of premature birth.

Acknowledgements Supported by grants from the University of Arkansas for Medical Sciences and the National Institute for Child Health and Human Development (HDOll23-02).

References Abdel-Rahman, A.M. and Rosenberg, A.A. (1994) Prevention of intraventricular hemorrhage in the premature infant. Clin. Perinatol., 21: 505-521.

Ader, R. and Grota, L.J. (1969) Effects of early experience on adrenocortical reactivity. Physiol. Behav., 4: 303-305. McAnulty, G. and Als, H., Lawhon, G., Gibes, R., Duffy, F.H., Blickman, J.G. (1994) Individualized developmental care for the very low-birth weight preterm infant: medical and neurofunctional effects. JAMA, 272: 853-858. Anand, K.J.S. (1990) Neonatal stress responses to anesthesia and surgery. Clin. Perinatol., 17: 207-214. Anand, K.J.S. (1993) Relationships between stress responses and clinical outcome in newborns, infants, and children. Crit. Care Med., 21: S3586359. Anand, K.J.S. (1998) Clinical importance of pain and stress in preterm newborn infants. Biol. Neonate, 73: 1-9. Anand, K.J.S. and Aynsley-Green, A. (1985) Metabolic and endocrine effects of surgical ligation of patent ductus arteriosus in the human preterm neonate: are there implications for further improvement of postoperative outcome? Mod. Pmbl. Paediat,:, 23: 143-157. Anand, K.J.S., Brown, M.J., Causon, R.C., Christofides, N.D., Bloom, S.R. and Aynsley-Green, A. (1985) Can the human neonate mount an endocrine and metabolic response to surgery? J. Pediat,: Surg., 20: 41-48. Anand, K.J.S. and Can; D.B. (1989) The neuroanatomy, neurophysiology and neurochemistry of pain, stress, and analgesia in newborns, infants, and children. Ped. Clin. N. Am., 36: 795-822. Anand, K.J.S., Coskun, V., Thrivikraman, K.V., Nemeroff, C.B. and Plotsky, P.M. (1999b) Long-term behavioral effects of repetitive pain in neonatal rat pups. Physiol. Behav., 66: 627-637. Anand, K.J.S. and Craig, K.D. (1996) Editorial: New perspectives on the definition of pain. Pain, 67: 3-6. Anand, K.J.S. and Hickey, P.R. (1987) Pain and its effects in the human neonate and fetus. N. Eng. J. Med., 317: 1321-1329. Anand, K.J.S. and McGrath, P.J. (1993) Future directions. In: K.J.S. Anand and P.J. McGrath (Eds.), Pain in Neonates, Elsevier Science Publishers, Amsterdam, pp. 321-333. Anand, K.J.S., McIntosh, N., Lagercrantz, H., Young, T.E., Vasa, R. and Barton, B.A. (1999a) Analgesia and sedation in ventilated preterm neonates: results from the pilot NOPAIN trial. Arch. Pediatl: Adolesc. Med., 153: 33 1-338. Anand, K.J.S., Selanikio, J.D. and the SOPAIN Study Group (1996) Routine analgesic practices in 109 neonatal intensive care units (NICUs). Ped. Res., 39: 192A. Anand, K.J.S., Sippell, W.G. and Aynsley-Green, A. (1987) Randomized trial of fentanyl anaesthesia in preterm babies undergoing surgery: effects on the stress response. Lancer, I: 243-248. Anand, K.J.S., Sippell, W.G., Schofield, N.M. and AynsleyGreen, A. (1988) Does halothane anaesthesia decrease the metabolic and endocrine stress responses of newborn infants undergoing operation? Brit. Med. J., 296: 668-672. Andrews, K. and Fitzgerald, M. (1994) The cutaneous withdrawal reflex in human neonates: sensitisation, receptive fields and the effects of contralateral stimulation. Pain, 56: 95-101.

126 Barker, D.P. and Rutter, N. (1995) Exposure to invasive procedures in neonatal intensive care unit admissions. Arch. Dis. Childhood, 72: F47-F48. Barker, D.P. and Rutter, N. (1996) Stress, severity of illness, and outcome in ventilated preterm infants. Arch. Dis. Childhood, 75: 187-190. Barker, D.P., Simpson, J., Pawula, M., Barrett, D.A., Shaw, P.N. and Rutter, N. (1995) Randomised, double blind trial of two loading dose regimens of diamorphine in ventilated newborn infants. Arch. Dis. Child Fetal Neonatal Edn., 73(1): F22F26. Basbaum, A.I. (1996) Memories of Pain. Science Med. Novl Dec: 22-31. Bauchner, H., May, A. and Coates, E. (1992) Use of analgesic agents for invasive medical procedures in pediatric and neonatal intensive care units. J. Ped., 121: 647-649. Becker, P.T., Grunwald, P.C., Moorman, J. and Stuhr, S. (1991) Outcomes of developmentally supportive nursing care for very low birth weight infants. Nurs. Res., 40: 150-155. Bicknell, H.R. and Beal, J.A. (1984) Axonal and dendritic development of substantia gelatinosa neurons in the lumbosacral spinal cord of the rat. J. Comp. Neurol., 226: 508-522. Blass, E.M., Shide, D.J., Zaw-Mon, C. and Sorrentino, J. (1995) Mother as shield differential effects of contact and nursing on pain responsivity in infant rats -evidence for nonopioid mediation. Behav. Neurosci., 109: 342-353. Botting, N., Powls, A,, Cooke, R.W.I. and Marlow, N. (1997) Attention deficit hyperactivity disorders and other psychiatric outcomes in very low birth weight children at 12 years. J. Child Psychol. Psychiat., 38: 931-941. Bower, A.J. (1990) Plasticity in the adult and neonatal central nervous system. Brit. J. Neurosurg., 4: 253-264. Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P.M. and Meaney, M.J. (1998) Maternal care during infancy regulates the development of neural systems mediating the expression of fearfulness in the rat. Pmc. Natl. Acad. Sci. USA, 95: 5335-5340. Cannon, M. and Murray, R.M. (1998) Neonatal origins of schizophrenia. Arch. Dis. Child, 78: 1-3. Cepeda, E.E., Lee, M.I. and Mehdizadeh, B. (1987) Decreased incidence of intraventricular hemorrhage in infants of opiate dependent mothers. Acta Paediatr: Scand., 76: 16-1 8. Chay, P.C.W., Duffy, B.J. and Walker, J.S. (1992) Pharmacokinetic-pharmacodynamic relationships of morphine in neonates. Clin. Pharmacol. The,:, 5 1: 334-342. Coplan, J.D., Trost, R.C., Owens, M.J., Cooper, T.B., Gorman, J.M., Nemeroff, C.B. and Rosenblum, L.A. (1998) Cerebrospinal fluid concentrations of somatostatin and biogenic amines in grown primates reared by mothers exposed to manipulated foraging conditions. Arch. Gen. Psychiatry, 55: 473-477. Craig, K.D., Whitfield, M.F., Grunau, R.V., Linton, J. and Hadjistavropoulos, H.D. (1993) Pain in the preterm neonate: behavioural and physiological indices. Pain, 52: 287-299. Dani, J.W., Armstrong, D.M. and Benowitz, L.I. (1991) Mapping the development of the rat brain by GAP-43 immunocytochemistry. Neuroscience, 40: 277-287.

Dixon, S., Snyder, J., Holve, R. and Bromberger, P. (1984) Behavioral effects of circumcision with and without anesthesia. J. Dev. Behav. Pediat,:, 5: 246-250. Dubnau, J. and Tully, T. (1998) Gene discovery in Drosophila: new insights for learning and memory. Ann. Rev. Neurosci, 21: 4071144. Dyke, M.P., Kohan, R. and Evans, S. (1995) Morphine increases synchronous ventilation in preterm infants. J. Paediat,: Child Health, 3 l(3): 176-179. Emde, R.N., Harmon, R.J., Metcalf, D., Koenig, K.L. and Wagonfield, S. (1971) Stress and neonatal sleep. Psychosomatic Med., 33: 491-497. Evans, N.J. and Kluckow, M. (1998) The hemodynamic basis of intraventricular haemorrhage (IVH). Early Human Dev. (Abstract) Falcon, M., Guendellman, D., Stolberg, A., Frenk, H. and Urca, G. (1996) Development of thermal nociception in rats. Pain, 67: 203-208. Fiselier, T., Monnens, L., Moerman, E., Munster, P., Jansen, M. and Peer, P. (1983) Influence of the stress of venepuncture on basal levels of plasma renin activity in infants and children. Int. J. Pediatr: Nephrol., 4: 181-185. Fitzgerald, M. (1985) The postnatal development of cutaneous afferent fibre input and receptive field organization in the rat dorsal horn. J. Physiol. (Lond.),364: 1-18. Fitzgerald, M. (1987) Spontaneous and evoked activity of foetal primary afferents ‘in vivo’. Nature, 326: 603-605. Fitzgerald, M. (1994) Neurobiology of fetal and neonatal pain. In: P.D. Wall and R. Melzack (Eds), The Textbook of Pain, Churchill Livingstone, Edinburgh, pp. 153-163. Fitzgerald, M. and Anand, K.J.S. (1993) The developmental neuroanantomy and neurophysiology of pain. In: N. Schechter, C. Berde and M. Yaster (Eds), Pain Management in Infants, Children and Adolescents, Williams and Williams, Baltimore, pp. 11-32. Fitzgerald, M., Millard, C. and MacIntosh, N. (1988) Hyperalgesia in premature infants. Lancet, 8580( 1): 292. Fitzgerald, M., Millard, C. and McIntosh, N. (1989) Cutaneous hypersensitivity following peripheral tissue damage in newborn infants and its reversal with topical anaesthesia. Pain, 39: 31-36. Fitzgerald, M., Reynolds, M.L. and Benowitz, L.I. (1991) G A P 4 3 expression in the developing rat lumbar spinal cord. Neuroscience, 41: 187-199. Fitzgerald, M., Shaw, A. and MacIntosh, N. (1988) The postnatal development of the cutaneous flexor reflex: a comparative study in premature infants and newborn rat pups. Dev. Med. and Child Neurol., 30: 520-526. Franck, L.S. (1987) A national survey of the assessment and treatment of pain and agitation in the neonatal intensive care unit. J. Obstet. Gynecol. Neonatal Nurs., Nov./Dec.: 387-393. Gentsch, C., Lichtsteiner, M., Frischknecht, H.R., Feer, H. and Siegfried, B. (1988) Isolation-induced locomotor hyperactivity and hypoaldesia in rats are prevented by handling and reversed by resocialization. Physiol. Behav., 43: 13-16.

127 Ghazi-Birry, H.S., Brown, W.R., Moody, D.M., Challa, V.R., Block, S.M. and Reboussin, D.M. (1997) Human germinal matrix: venous origin of hemorrhage and vascular characteristics. A m J. Neumrudiol., 18: 219-229. Goldstein, R.F. and Brazy, J.E. (1991) Narcotic sedation stabilizes arterial blood pressure fluctuations in sick premature infants. J. Perinutol., 11: 365-37 1. Goldstone, R.L. (1998) Perceptual learning. Ann. Rev. Psychol., 49: 585-612. Goodyer, I.M. (1994) Development psychopathology: The impact of recent life events in anxious and depressed schoolage children. J. Roy. SOC.Med., 87: 327-329. Greeley, W.J. and de Bruijn, N.P. (1988) Changes in sufentanil pharmacokineticswithin the neonatal period. Anesth. Analg., 67: 86-90. Greenough, A., Lagercrantz, H., Pool, J. and Dahlin, I. (1987) Plasma catecholamine levels in preterm infants. Acta. Puediutr. Scund., 76: 54-59. Greenough, A., Nicolaides, K.H. and Lagercrantz, H. (1990) Human fetal sympathoadrenal responsiveness. Early Human Dev., 23: 9-13. Greisen, G.S., Frederiksen, P.S., Hertel, M. and Christensen, N.J. (1985) Catecholamine response to chest physiotherapy and endotracheal suctioning in preterm infants. Actu Puediatr. Scand., 74: 525-529. Grunau, R.V.E., Whitfield, M.F. and Petrie, J.H. (1994a) Pain sensitivity and temperament in extremely-low-birthweight premature toddlers and preterm and full-tern controls. Pain, 58: 341-346. Grunau, R.V.E., Whitfield, M.F., Petrie, J.H. and Fryer, E.L. (1994b) Early pain experience, child and family factors as precursors of somatization: a prospective study of extremely premature and fullterm children. Pain, 56: 353-359. Grunau, R.V.E., Whitfield, M.F. and Petrie, J.H. (1998) Children’s judgements about pain at age 8-10 years: do extremely low birthweight ( c 1000 gm) children differ from full birthweight peers? J. Child. Psychol. Psychiats., 39: 5 87-594. Gunnar, M., Hertsgaard, L., Larson, M. and Rigatuso, J. (1991) Cortisol and behavioral responses to repeated stressors in the human newborn. Dev. Psychobiol., 24: 487-506. Gunnar, M.R., Larson, M.C., Hertsgaard, L., Harris, M.L. and Bmderson, L. (1992) The stressfulness of separation among nine-month-old infants: Effects of social context variables and infant temperament. Child Dev., 63: 290-303. Gunnar, M.R., Porter, EL., Wolf, C.M., Rigatuso, J. and Larson, M.C. (1995) Neonatal stress reactivity: predictions to later emotional temperament. Pediatrics, 66: 1-1 3. Guy, E.R. and Abbott, F.V. (1992) The behavioral response to formalin in pre-weaning rats. Pain, 51: 81-90. Hack, M., Taylor, H.G., Klein, N., Eiben, R., Schatschneider,C. and Mercuri-Minich, N. (1994) School-age outcomes in children with birth weights under 750g. N.Eng. J. Med., 331: 753-803. Haxhija, E.Q., Rosegger, H. and Prechtl, H.F. (1995) Vagal response to feeding tube insertion in preterm infants: has the key been found? Early Hum.Dev., 4 1: 15-25.

Hu, D., Hu, R. and Berde, C.B. (1997) Neurologic evaluation of infant and adult rats before and after sciatic nerve blockade. Anesthesiology, 86: 957-965. Izquierdo, I. and Medina, J.H. (1997) Memory formation: the sequence of biochemical events in the hippocampus and its connection to activity in other brain structures. Neumbiol. b u r n . Mem., 68: 285-316. Jacobson, B., Eklund, G., Hamberger, L., Linnarsson, D., Sevdall, G . and Vlaverius, M. (1987) Perinatal origin of adult self-destructive behavior. Acta Psychiatry S c a d . , 76: 364-371. Jacobson, B. and Bygdeman, M. (1998) Obstetric care and proneness of offspring to suicides as adults: case-control study. BMJ, 317: 1346-1349. Johnston, C.C., Collinge, J.M., Henderson, S.J. and Anand, K.J.S. (1997) A cross-sectional survey of pain and pharmacological analgesia in Canadian neonatal intensive care units. Clin. J. Pain, 13: 308-312. Johnston, C.C. and Stevens, B.J. (1996) Experience in a neonatal intensive care unit affects pain response. Pediatrics, 98(5): 925-930. Johnston, C.C., Stevens, B., Craig, K.D. and Grunau, R.V.E. (1993) Developmental changes in pain expression in premature, full-term, two- and four-month-old infants. Pain, 52: 201-208. Lagercrantz, H., Nilsson, E., Redham, I. and Hjemdahl, P. (1986) Plasma catecholamines following nursing procedures in a neonatal ward. Early Human Dev., 14: 61-65. Lander, J., Brady-Fryer, B., Metcalfe, J.B., Nazarali, S. and Muttitt, S. (1997) Comparison of ring block, dorsal penile nerve block, and topical anesthesia for neonatal circumcision: a randomized controlled trial. JAMA, 278: 2157-2162. Leon, M. (1992) The neurobiology of filial learning. Ann. Rev. Psychol., 43: 377-398. Levy-Shiff, R., Einat, G., Har-Even, D., Mogilner, M., Mogilner, S., Lerman, M. and Krikler, R. (1994) Emotional and behavioral adjustment in children born prematurely. J. Clin. Child Psychol., 23: 323-333. Liu, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., Sharma, S., Pearson, D., Plotsky, P.M. and Meaney, M.J. (1997) Maternal care, hippocampal glucocorticoid receptors, and hypothalamic-pituit-adrenal responses to stress. Science, 277: 1659-1662. Low, J.A., Froese, A.B., Smith, J.T., Galbraith, R.S., Sauerbrei, E.E. and Karchmar, E.J. (1992). Hypotension and hypoxemia in the preterm newborn during the four days following delivery identify infants at risk of echosonographically demonstrable cerebral lesions. Clin. Invest. Med., 15: 60-65. Marshall, R.E., Stratton,W.C., Moore, J.A. and Boxerman, S.B. (1980) Circumcision: Effects Upon Newborn Behavior. Infant Behav. Dev., 3: 1-14. Marti, E., Gibson, S.J., Polak, J.M., Facer, P., Sringall, D.R., van Aswegen, G., Aitchison, M. and Koltzenburg, M. (1987) Ontogeny of peptide- and amine-containing neurones in motor, sensory and autonomic regions of rat and human

128

spinal cord, dorsal root ganglia and rat skin. J. Comp. Neurol., 266: 332-359. McEwen, B.S. (1994) The Plasticity of the hippocampus is the reason for its vulnerability. Neurosciences, 6: 239-246. McGraw, M.B. (1941) Neural maturation as exemplified in the changing reactions of the infant to pin prick. Child Dev., 12: 3142. McIntosh, N. (1997) Pain in the newborn, a possible new starting point. Euro.J. Pediatr:, 56: 173-178. McLaughlin, C.F., Lichtman, A.H., Fanselow, M.S. and Cramer, C.P. (1990) Tonic nociception in neonatal rats. Phamacol. Biochem. Behav., 36: 859-862. McLaughlin, C.R., Hull, J.G., Edwards, W.H., Cramer, C.P. and Dewey, W.L. (1993) Neonatal pain: a comprehensive survey of attitudes and practices. J. Pain Symptom Manag., 8 : 7-16. McRae, M.E., Rourke, D.A., Imperial-Perez, EA., Eisenring, C.M. and Ueda, J.N. (1997) Development of a researchbased standard for assessment, intervention, and evaluation of pain after neonatal and pediatric cardiac surgery. Pediatr: Nurs., 23: 263-27 1. Meaney, M., Mitchell, J.B., Aitken, D.H., Bhatnagar, S., Bodnoff, S.R., Iny, L.J. and Sarrieau, A. (1991) The effects of neonatal handling on the development of the adrenocortical response to stress: implications for neuropathology and cognitive deficits in later life. Psychoneuroendocrinology, 16: 85-103. Meaney, M.J., Sharma, S., Sarrieau, S. and Plotsky, P.M. (1988) Postnatal development and environmental regulation of hippocampal glucocorticoid and mineralocorticoid receptors in the rat. Dev. Brain Res., 43: 158-162. Meaney, M.J., Tannenbaum, B., Bhatnagar, S., Shanks, N., Viau, V. and O’Donnel, D. (1994) Early environmental programming of hypothalmic-pituitary-adrenalresponses to stress. Neurosciences, 6: 247-259. Ment, L., Oh, W., Ehrenkranz, R.A., Philip, A.G.S., Vohr, B., Allan, W., Duncan, C.C., Scott, D.T., Taylor, K.J.W., Katz, K.H., Schneider, K.C. and Makuch, R.W. (1994) Low-dose indomethacin and prevention of intraventricular hemorrhage: a multicenter randomized trial. Pediatrics, 93: 543-550. Ment, L.R., Vohr, B., Oh, W., Scott, D.T., Allan, W.C., Westerveld, M., Duncan, C.C., Ehrenkranz, R.A., Katz, K.H., Schneider, K.C., et al. (1996) Neurodevelopmental outcome at 36 months’ corrected age of preterm infants in the multicenter indomethacin intraventricular hemorrhage prevention trial. Pediatrics, 98: 714-718. Mullen, P.E., Martin, J.I., Anderson, J.C. and Herbison, G.P. (1993) Childhood sexual abuse and mental health in adult life. Br: J Psych., 163: 721-732. O’Donnell, D., Larocque, S., Seckl, J.R. and Meaney, M.J. (1994) Postnatd handling alters glucocorticoid, but not minerallocorticoid messenger RNA expression in the hippocampus of adult rats. Brain Res. Mol. Brain Res., 26: 242-248. Orsini, A.J., Leef, K.H., Costarino, A., Dettorre, M.D. and Stefano, J.L. (1996) Routine use of fentanyl infusions for

pain and stress reduction in infants with respiratory distress syndrome. J. Pediarr., 129: 140-145. Perlman, J. and Thach, B. (1988) Respiratory origin of fluctuations in arterial blood pressure in premature infants with respiratory distress syndrome. Pediatrics, 8 1: 399-403. Perlman, J.M., Goodman, S., Kreusser, K.L. and Volpe, J.J. (1985) Reduction in intraventricular hemorrhage by elimination of fluctuating cerebral blood flow velocity in preterm infants with respiratory distress syndrome. New Engl. J. Med., 312: 1353-1357. Philbin, M.K., Ballweg, D.D. and Gray, L. (1994) The effect of an intensive care unit sound environment on the development of habituation in health avian neonates. Dev Psychobiol., 27: 11-21. Pieretti, S., d’Amore, A. and Loizzo, A. (1991) Long-term changes induced by developmental handling on pain threshold effects of morphine and naloxone. Behav. Neurosci., 105: 215-218. Pignatelli, D., Silva, D.A. and Coimbra, A. (1989) Postnatal maturation of primary afferent terminations in the substantia gelatinosa of the rat spinal cord: An electron microscope study, brain. Brain Rex, 491: 33-44. Pinto-Martin, J.A., Riolo, S., Cnaan, A., Holzman, C., Susser, M.W. and Paneth, N. (1995) Cranial ultrasound prediction of disabling and nondisabling cerebral palsy at age two in a low birth weight population. Pediatrics, 95: 249-254. Plotsky, P.M. and Meaney, M.J. (1993) Early postnatal experience alters hypothalamic corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stressinduced release in adult rats. Mol. Brain Res., 18: 195-200. Pokela, M.L. (1993) Effect of opioid-induced analgesia on beta-endorphin, cortisol and glucose responses in neonates with cardiorespiratory problems. Biol. Neonate, 64(6): 360-367. Pokela, M.L. (1994) Pain relief can reduce hypoxemia in distressed neonates during routine treatment procedures. Pediatrics, 93: 379-383. Porter, EL., Wolf, C.M. and Miller, J.P. (1999) Procedural pain in newborn infants: the influence of intensity and development. Pediatrics, 104: e13. Porter, EL., Wolf, C.M., Gold, J., Lotsoff, D. and Miller, J.P. (1997) Pain and pain management in newborn infants: a survey of physicians and nurses. Pediatrics, 100: 626-632. Porter, EL., Wolf, C.M. and Miller, J.P. (1998) The effect of handling and immobilization on the response to acute pain in newborn infants. Pediatrics, 102: 1383-1389. Purcell-Jones, G., Sumner, E. and Dormon, F. (1988) Paediatric anaesthetists’ perceptions of neonatal and infant pain. Pain, 33: 181-187. Quinn. M.W., Otoo, F., Rushforth, J.A., Dean, H.G., Puntis, J.W., Wild, J. and Levene, M.I. (1992) Effect of morphine and pancuronium on the stress response in ventilated pretenn infants. Early Human Dev., 30: 241-248. Quinn, M.W., Wild, J., Dean, H.G., Hartley, R., Rushforth, J.A., Puntis, J.W. and Levene, M.I. (1993) Randomised doubleblind controlled trial of effect of morphine on catecholamine

129 concentrations in ventilated preterm babies. Lancet, 342: 324-327. Rabinowicz, T., de Courten-Myers, G.M., Petetot, J.M., Xi, G. and de 10s Reyes, E. (1996) Human cortex development: estimates of neuronal numbers indicate major loss late during gestation. J. Neuropathol. Exp. Neurol., 55: 320-328. Rakic, P. (1985) Limits of neurogenesis in primates. Science, 227: 1054-1055. Ramaekers, YT., Casaer, P. and Daniels, H. (1993) Cerebral hyperperfusion following episodes of bradycardia in the preterm infant. Early Hum.Dev., 34: 199-208. Ramsay, D.S. and Lewis, M. (1995) The effects of birth condition on infants’ cortisol response to stress. Pediatrics, 95: 546-549. Reynolds, M.L. and Fitzgerald, M. (1995) Long-term sensory hyperinnervation following neonatal skin wounds. J. Comp. Neuml., 358: 487-498. Rogers, M.C. (1 992) Do the right thing - pain relief in infants and children. N. Eng. J. Med., 326: 55-56. Sims, C., Johnson, C.M., Bergesio, R., Delfos, S.J. and Avraamides, E.A. (1994) Rectal indomethacin for analgesia after appendectomy in children. Anaesth. Int. Care, 22: 272-275. Somrnerfelt, K., Troland, K., Ellertsen, B. and Markestad, T. (1996) Behavioral problems in low-birthweight preschoolers. Dev. Med. Child Neuml., 38: 927-940. Squire, L.R. (1986) Mechanisms of memory. Science, 231: 729-73 1. Stein-Behrens, B., Mattson, M.P., Chang, I., Yeh, M. and Sapolsky, R.M. (1994) Stress exacerbates neuron loss and cytoskeletal pathology in the hippocampus. J. Neumsci., 14: 5373-5380. Taddio, A., Goldbach, M., Ipp, M., Stevens, B. and Koren, G. (1995) Effect of neonatal circumcision on pain responses during vaccination in boys. Lancet, 345: 291-292.

Taddio, A., Katz, J., Ilersich, A.L. and Koren, G. (1997) Effect of neonatal circumcision on pain response during subsequent routine vaccination. Lancet, 349: 599-603. Teng, C.J. and Abbott, F.V. (1998) The formalin test: a doseresponse analysis at three development stages. Pain, 76: 337-347. Tohill, J. and McMorrow, 0. (1990) Pain relief in neonatal intensive care. Lancet, 336: 569 Viau, V., Sharma, S., Plotsky, P. and Meaney, M. (1993) Increased plasma ACTH responses to stress in non-handled compared with handled rats require basal levels of corticosterone and are associated with increased levels of ACTH secretagogues in the median eminence. J. Neuroscience, 13: 1097-1 105. Vohr, B. and Ment, L.R. (1996) Intraventricular hemorrhage in the preterm infant. Early Human Dev., 44: 1-16. Wells, J.T. and Ment, L.R. (1995) Prevention of intraventricular hemorrhage in preterm infants. Early Human Dev., 42: 209-233. Whitfield, M.F., Grunau, R.V.E. and Holsti, L. (1997) Extreme prematurity ( < 800 g) at school age: multiple areas of hidden disability. Arch. Dis. Child, 77: F85-F90. Woolf, C.J. and Wall, P.D. (1986) The relative effectiveness of C primary afferent fibres of different origins in evoking a prolonged facilitation of the flexor reflex in the rat. J. Neurosci., 6: 1433-1443. Yaster, M. (1987) The dose response of fentanyl in neonatal anesthesia. Anesthesiology, 66: 433435. Yi, D.K. and Ban; G.A. (1995) The induction of fos-like immunoreactivityby noxious thermal, mechanical and chemical stimuli in the lumbar spinal cord of infant rats. Pain, 60: 257-265. Zernikow, B., Michel, E., Kohlmann, G., Steck, J., Schmitt, R.M. and Jorch, G. (1994) Cerebral autoregulation of preterm neonates - a non-linear control system. Arch. Dis. Childhood, 70: F166-Fl73.

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research, Vol 122 Q 2000 Elsevier Science BV. All rights reserved.

CHAPTER 10

Early life abuses in the past history of patients with gastrointestinal tract and pelvic floor dysfunctions Ghislain Devroede Centre universitaire de santd de l'Estrie, Campus Fleurimont, 3001 - 12e Avenue Nord, Fleurimont, Que. J1H 5N4, Canada

Introduction The experience of a life-threatening event may precede onset of bowel disturbance, as commonly seen clinically, particularly when it occurs in the early part of life, when the mind is progressively being shaped up into individual idiosyncrasies. Among negative events, extreme family tensions are often recorded among patients with functional abdominal pain (Creed et al., 1988). Following anecdotal reports about the association of sexual abuse and gastrointestinal disorders (Devroede et al., 1989; Devroede, 1990), Drossman and his colleagues demonstrated the high prevalence of such a history among women, seen in a university context for functional or organic gastrointestinal disorders (Drossman et al., 1990). But the impact of an abuse history is highly unlikely to be linked specifically to a diagnosis of irritable bowel syndrome, as exemplified by Drossman's study, which also indicated a negative impact of an abuse history on the number of medical consultations and surgical interventions. Moreover, severe sexual dysfunction has been shown to exist in women with irritable bowel syndrome (Guthrie et al., 1987) and this is highly likely to have a negative impact on the

*Corresponding author. Tel.: (819) 564-5231;Fax: (819) 820-6411 ; e-mail: 1tessierecourrier.usherb.ca

life of these patients, with, presumably, a psychosocia1 dysfunctional cost and subsequent somatization. When a victim of sexual abuse speaks out and the perpetrator recognizes the act, one cannot question the data and call into action the fantasy theory popularized by Freud, but even then, physical symptoms may persist long after the recognition of the crime (Devroede, 1995; Lalonde, 1995). We've known about the association between abuse during the early part of life and health problems in the gastrointestinal sphere for less than ten years, and the impact on gynecological health problems has followed through, while investigation about urological consequences has barely begun. We thus dearly lack long-term follow-up studies to be able to understand and describe all involved parameters at work. Moreover, a simple association between abuse and bowel dysfunction does not imply a causative element, and other elements, much more deeply imbedded in the past and in the unconscious mind of this type of patients may be at work. The purpose of this chapter is three-fold: (a) To report a number of long-term follow-up studies of patients who have been sexually abused. (b) To do a brief review of presently available data on the links between sexual abuse and gastrointestinal complaints; what is known about pelvic floor dysfunction will also be discussed,

132

because the pelvic floor is a functional neuromuscular unit. (c) To attempt to provide an integrated vision of objective findings and subjective experience. All science is made of measurements but subjects are not objects, and, thus, illness behavior is different from disease activity. A process of desomatization cannot begin if an individual, often dissociated as a consequence of the trauma, is approached in a way which is dissociated between the mind and the body, i.e. a purely scientific approach, which has to reduce a suffering human being to a sick organ. Long-term follow-up studies About surgery for constipation: MARTHE This 35-year old white woman was referred for total colectomy in March, 1981, because of severe chronic idiopathic constipation, with a bowel movement every two months, and severe abdominal pain that had provoked nine prolonged stays in hospital over the previous ten years with no specific diagnosis. Hysterectomy and left ovarectomy had been performed to no avail; both uterus and ovary were normal at pathological examination. The patient was not returning the enemas she was given, and none of the laxatives she had been given in gigantic amounts were working; in despair she had been placed on vasopressin by nasal spray (DDAVP 5-10 p,g intra-nasally once or twice daily), which was mildly effective through its side effect of cramping abdominal pain, presumably via an increase in propagated electrical activity in the colon. The gastroenterologist, who had referred the patient for colectomy, thought she had colonic inertia on the basis of psychogenic origin. He was advised not to proceed for two reasons. First, the patient threatened to commit suicide if surgery was not effective in releasing her symptoms, a not too uncommon experience for colorectal surgeons dealing with this kind of problem. This blackmail approach is often used by victims of sexual abuse and is an acting out of anxiety of the order of

“operate or I shall kill myself.” The second reason for the advice against surgery was that the patient said her father had raped her when she was sixteen; it was not until five years later that she said the rape was a sodomy; it was not until fifteen years later that she was able to recount in detail the experience of the rape. On October 3, 1986, the father of the patient died. By then, she had been seen in consultation 125 times over the course of these five years. She called in, in emergency, not at the hospital, but at my residence, telling me about the death and about the fact her family did not understand why she did not cry. She was offered either to tell the family about the rape (in fact, she did not know her four sisters had also been abused), or to understand the fact that they did not understand, a clearly unacceptable alternative, therefore conducive to a situation of Ericksonian hypnosis, of dissociative nature. She therefore proposed a third solution, namely to write a letter in which she would write to him about pent-up feelings, a proposal which was encouraged as being a powerful transitional object. She placed a six-page long letter to her dead father in his suit prior to burial, and was cured overnight. By October 28, i.e. three weeks later, she was having a daily bowel movement. Colorectal transit

MARTHE COLORECTALTRANSIT TlME

150-

TRANSIT

-

TIME (HOURS) 100-

50-

n”

RlCM COLON LEFT COLON ECrOSIGMOID

TOTAL

Fig. 1 . Changes in colorectal transit times of Marthe, before (1981) and after (1986) she put into the coffin of her dead father a six-page long letter about her feelings toward him for having sodomized her at age 16. The transit times studies were repeatedly abnormal from 1981 to the death (see details in reference 2). There was further improvement later on (1997), particularly in the rectosigmoid area (see text for explanations about the concomitant life history).

133

Marthe SEQUENTIAL CHANGES IN PERSONALITY PROFILE

.

1981 4

--

1982 -0 19a ............ 1987 ---d 1988 ---.--o

2ol 0

L f K

HsDHyPdMfPdPtSChSi

A R E S L B C a D y D o R C P R S t0 tn

SCALES

Fig. 2. MMF'I (Minnesota Multiple Personality Inventory) profiles of Marthe over successive years. In 1981, when a colectomy was proposed by a gastroenterologist, the pattern was almost that of a psychotic depression. Note that the scale of depression (D) is higher than that of hypochondria (Hs) and hysteria (Hy) in 1981 and 1982 but that it is slowly becoming of a lesser value. The cutoff seems to be after the letter to the dead father.

Marthe PERSONALITY PROFILE 100

90 MMPl

TSCORES 8o (K corrected) 70

60 50

40

30 20 10

0

L

F

K

HS

D

H Y P D M F P A

PT

X M A S 1

Fig. 3. Personality profile of Marthe at onset of follow up and in 1997. The amount of psychopathology has markedly decreased (values above 70 are abnormal) during follow up. Of the three validation scales (L, F, K),the F score (faking bad) is markedly reduced, indicating a much greater self-esteem of the patient. The configuration of the first three scales (HS,D, HY) is changed: initially, the depression (D)score is higher than the hypochondria (HS) and the hysteria (HY) scores, while, in 1997, a typical somatization pattern has appeared, with depression score being of lower value (psychosomatic "V" or hysterical valley). The patient's enhanced trust is suggested by the lowering of the schizophrenia (SC) score, and her decreased anxiety level by the lowering of the psychasthenia (PT) score. Decrease in rage is suggested by the decrease in psychopathy (PD) level.

134

Marthe PERSONALITY PROFILE

MMPI TSCOW

y

A

R

ES

LB

C A D Y D O R F .

PR

ST

CN

Fig. 4. Research scales of the MMPI persondity test of Marthe. Ego strength (ES)of the patient has been boosted between 1981 and 1997, as well as lessening of anxiety (A) and repression (R).Initially, the patient was much more dependent (DY) than dominant (DO), and this pattern has disappeared.

times, which were markedly prolonged, none of twenty ingested radiopaque markers being excreted within a week, normalized. Segmental transit time was initially normal in the right colon, the colonic delay being limited to the hindgut portion of the large bowel. These suddenly fell within normal limits (Fig. 1). Propagated electrical activity in the left colon, which was virtually absent initially, reappeared. The details of the repeated physiological measurements have been published elsewhere (Devroede et al., 1989) and they confirm the sudden turnabout both of the symptomatology and the accompanying objective measurements of the bowel dysfunction underlying this symptomatology. Shortly thereafter, the patient's husband, so far faithful despite a miserable sexual compatibility, began to have numerous extramarital affairs, which eventually led to a divorce. Simultaneously, the patient began to have 10-30 liquid stools per day. Organic evaluations, quite extensive, revealed normal appearance of the entire gastrointestinal tract,

no infection and no fat malabsorption. A bile salts turnover study demonstrated bile salts malabsorption (14% retention of isotope one week after ingestion of a capsule of SeHCAT - 75 Selenium Taurocholic acid - 370 Rbq or 10 microcurie). A few months after her divorce in February, 1991, the patient began to have a relationship with a new man, which was never very sexual, in contrast to the obligatory ritual of daily, 2-minute intercourse with her ex-husband. She broke off this new relationship herself in April, 1994 and the diarrhea suddenly disappeared. She began to produce 2 to 3 normal stools every day. The evolution of her bowel transit times is shown in Fig. 1. Between 1986, when the constipation suddenly disappeared, and 1994, when the diarrhea did, she had been seen 108 times. She dismissed herself. Over these years, there was constant improvement in the pathological configuration of her MMPI profile (Minnesota Multiple Personality Inventory), but without the same sudden changes that happened for the large bowel function (Fig. 2).

135

l b o years later, she came back, ostensibly not for abdominal symptoms nor bowel problems, but because her relationship with her three sons was very conflictual. On September 3, 1996, she asked for an emergency appointment and turned up in the late afternoon. She had seen her former lover again and had been unable to refuse to have sex with him. She broke off again with him afterwards and suddenly discovered, much to her, and my, surprise, that cutaneous sensation, until then totally absent to the point where she could touch burning objects without hurting herself, had reappeared all over her body except for her breasts and pelvis. She had never taken a bath in her life, and showers lasted only fifteen seconds, for fear of having to touch herself. She entered psychotherapy with a woman who helped her to express her rage, and for the first time, she told in detail about the rape, at times with regressive episodes of emotional turmoil. Soon thereafter, she began to masturbate herself. By November, 1997, she experienced orgasm for the

first time by self-manipulation of the clitoris and defecated stools of approximately the length of the entire hindgut. She thought these were not ‘normal’ at first. She began to have intravaginal sensations. All together, she had been followed up for 262 visits over a span of seventeen years. Her personality profile has practically normalized (Figs 3 and

4).

* * * About diarrhea and massive fecal incontinence: DENISE The patient consulted me in October, 1986, at the age of 45, for chronic abdominal pain and diarrhea. She had been having 6-7 liquid stools day and night for the last eleven years. She obviously did not have malabsorption since she was also evidently morbidly obese. She was extremely angry at the gastroenterologists she had seen, claiming they

TABLE 1 DENISE Functional gastrointestinal motor evaluation

Colorectal Transit Times (hours)

Right Colon Left Colon Rectosigmoid Colorectal

Rectometrogram (volumes in ml water)

Constant sensation volume (CSV) Maximum tolerable volume (MTV) Sensitivity index*

Anorectal manometry (Pressure. in cm water)

PRESSURE PROFILE Rectum Upper anal canal Lower anal canal Rectoanal inhibitory reflex

AMSMUS Straining effort Upper anal canal Lower anal canal VOLUNTARY CONTRACTION Upper anal canal Lower anal canal

1987

1988

1990

1992

12 35 18 65

8 10 26

14

44

25 13 52

20 25 22 67

220 240 0.08

290 320 0.1

310 340 0.08

36 105 44 80 Present and normal 14 80

+ 29 + 30 + 10

+ I5

+ 12 + 30

0 0 + 38

40 70

70 135

30 19

136

Denise A BRIEF GENOSOCIOGRAM Charles

I

i

"..

Jean-Paul

.... ....._.._..

-......

_.._..

Christiane

'.

...'..

-..-.. -..-'.... mdtal relatiomhip ..%..

,id''

....

Nelson

.,"

8..

_..... ...... ......_._........_.._..-. .---"

..'." ...-"..'

_..."

serunl mlatlonship

~11pmthaod to

Fig. 5. The notion of a 'genosociogram' implies that there is a 'behavioral' heredity which is not chromosomic, organic, in the genes, but that the family history has an impact on the offsprings. It has been recently well exemplified by Anne Ancelin Schulzenberger, in her book 'The ancestors syndrome' (Rutledge, London, New Yurk, 1998). In this figure, important links for the well-fare of Denise, which would normally not appear in a classic genealogy tree are those of Charles with Christiane, his daughter, with whom he had a sexual relationship. Denise does not know if she was conceived by Charles, her grandfather, or Jean-Paul, her legal father. Denise, when she consulted, was living with Nelson, officially her uncle, since he is the brother of her mother Christiane, but possibly her brother, since he is the son of Charles, who is possibly her father too. Thus, this evidence of a very dysfunctional family is shown vividly in a genosociogram.

were not caring for her and were incompetent to treat her bad case of diarrhea. An extensive organic evaluation failed to demonstrate any disease along the gastrointestinaltract. In particular, she exhibited no steatorrhea on several occasions. The etiology of the watery diarrhea turned out, as in Marthe, to be idiopathic bile salts malabsorption (5% retention one week after ingestion). Functional evaluation during follow up revealed the overall colorectal transit time to remain normal

throughout the ten-year follow-up period until recovery, with a trend for the right colon transit to slightly prolong and the left colon transit to shorten. Rectal hypoesthesia persisted until the last examination, with a very small gap between rectal sensation and maximum tolerable volume in the rectum. The rectoanal inhibitory reflex was present and normal. She had a very weak capacity to voluntarily contract the anal canal and anismus was persistently found. Bladder evaluation was a mirror image of the rectal accommodation studies in that

137

she had normal compliance but no sensation until a capacity of 950 ml, with 200 ml residue. The flow at debimetry was normal at 33mUsec. Details are given in Table 1. During the first two years of follow-up, Denise evolved from chronic diarrhea to having bouts of constipation, where she would defecate only once a week. In July, 1988, she brought a letter in which she had written the reasons why she did not want to be followed up anymore, while adding that she had changed her mind and asked for another appointment when it was possible to do this. It took another two years, in March, 1989, for her to say that from then on she would follow a weight reduction program with the Weight-Watchersgroup rather than continue a medical evaluation. Eight months later, in November, she showed up again, apparently because of the failure of the program. She asked if bariatric surgery was a possibility and was referred to another surgeon to see if he would consider her for a Scopinaro

TABLE 2 DENSE Abuse history Type of abuse

Age Perpetrators

Y

Grandfather (maternal) Mother

7

14 Father 15

Husband

30 Stepfather Mother

Satanic rituals Enemas Sexual touches Possibility of repressed memories (dissociative episodes) Exhibitionism Sexual touches Rapes by sodomy Forced intercourse with husband‘s friends Forced intercourse by mother

procedure, which would entail removing 90% of her stomach and bypassing half of the small bowel. By then, she had been gaining weight constantly

Denise PERSONALITY PROFILE

MMPi

T SCORES

(Kcorrected)

L

F

K

HS

D

HY

PD

MF

PA

PT

SC

MA

SI

Fig. 6. Although to a lesser extent, because there was less psychopathology initially than in Marthe, as shown in Fig. 3, the trend towards normalization over the years, in Denise, is the same. There is less self-depreciation (lowering of the F scale), and a decrease in anger (PD) and mistrust (SC). In contrast, however, to what is shown in Fig. 2, the psychosomatic ‘V’ configuration is present from onset, and the mania (MA) scale was higher initially, protecting Denise from depression.

138

Denise PERSONALITY PROFILE

MMPI 70

T SCORES

60 50

40 30 20 10

0 A

R

ES

LB

CA

DY

DO

R€

PR

ST

CN

Fig. 7. Research scales of the MMPI. Changes in Denise over seven years. As compared to Fig. 4, where similar data on Marthe are shown, there is also an improvement in ego strength (ES), and an enhancement of her dominating attitudes (DO) as compared to her dependency (DY) attitudes.

and her Quetelet index was 44 (1 12.7 kg/1.6 m’). Therefore, she was considered eligible for surgery, placed on the waiting list, and referred to a number of consultants, including a psychiatrist, as part of the routine preoperative process in preparation for the morbid obesity surgery. The psychiatrist considered her anxiety ‘normal’ in terms of the surgery to come and merely proposed to have her visit the operating suite once, to get accustomed to the environment. This process, however, triggered an enormous amount of rage and she exclaimed, “Si ce docteur pense qu’il va me toucher!” (“If this doctor believes he is going to touch me!”). In Quebec, the word ‘toucher’ carries a very clear double meaning, i.e. touching-neutral or touching-sexual (Bkrubk et al., 1994). Indeed, by then, the heavy abuse history of the patient had begun to be explored. When she began consulting, she was living maritally with her uncle, but since her grandfather was having sex across generations, they thought they were brother and sister (Fig. 5 ) .

The abuse history was unusually heavy (Table 2). Denise started experiencing emotional catharsis about her life history. She voiced a lot of rage about what she had been submitted to, and physically lost, in her own terms, “one hundred pounds of anger.” The diarrhea subsided and bile salts malabsorption disappeared: a turnover bile salts study was normal in November, 1993 with 46% retention of the isotope one week after ingestion, in sharp contract to the mere 5 % retention initially. Yet, in April, 1992, she began to complain of fecal incontinence, and biofeedback for anismus was undertaken, over 20 sessions between August 2, 1992 and July 28, 1993. In parallel to the medical and the functional evaluation after revelation of the very heavy abuse history, Denise began to evolve. On her own and supported by her uncle-brother-lover, who had asked for help from Alcoholics Anonymous, she entered what she said was ‘transactional analysis’ on February 15, 1990. She had run away from her

139

family home at age 15, when her father began to be insistent about having sex with her. Her husband not only violently sodomized her, but also forced her to have sex with his male friends. They had three children, two boys and one girl, and her husband had sex with all three. At age 30, she attempted suicide and divorced immediately when she left the hospital. She went back to her mother, by then remarried, who actively pushed her to have sex with her second husband, and then had violent quarrels with her afterwards. Therefore, the patient left them and had a ten-year peaceful homosexual relationship with an Indian woman. Then she entered into a relationship with her uncle-brotherlover, and this incestuous relationship, paradoxically, was the best and most healing relationship she had had in her life. Very shortly after she had entered ‘transactional analysis’, her older son committed suicide on August 23, 1990, after having spent many years in psychiatric institutions. She barely recovered from this trauma when she learned that her second son’s body was found in Texas, three years after he had disappeared, apparently having committed suicide. This triggered several years of emotional turmoil. In July 1993, she founded a self-help group, in parallel to the consultation service in colorectal disease. A few months later, her father died. She and her uncle-brother-lover moved apart two days later. By then, she had begun raising questions about her gender identity. She said her mother had wished to call her ‘Denis’, and she began to have numerous dreams about being a man, seeing a pregnant man, and searching for a little girl. She then threw away a doll she had kept preciously for many years and had baptized ‘Denis’. Shortly thereafter, cutaneous sensations reappeared. Also, she realized the existence of a link between food and sexual contacts with her uncle-brother-lover. Out of sexual desire mixed with fear, she would eat a lot before going to his apartment to make love with him. But sex was, by then, extremely frustrating for both of them, and she would eat again afterwards out of dissatisfaction. She eventually stopped having any sexual contacts with him, but they developed a tender, physical, but asexual, proximity. Thereafter, she presented herself as a volunteer in a palliative care unit. She became one

of the key persons of the unit, while fully aware that she had chosen this type of activity in part because it would constantly confront her with death, bereavement and letting go. During the early part of this period, her former female lover reappeared. Denise refused to have sex with her, and the former female lover, in return, had sex with her uncle-brother-former lover. This led to both the Indian lover and her brother to attempt suicide, and she cared for them intensely. In 1996, she accompanied a woman to her death, then accompanied her widower, a man of 50 (five years younger than Denise) and she and the man eventually fell in love. She claimed this was the first loving relationship of her life. They both experienced intense pleasure and cried a lot while making love but she voiced that these were not tears of sadness. In January, 1997, she came in the clinic just to say goodbye. She and her lover had moved in together and intended to get married. She was totally asymptomatic. Over the ten years follow-up, she had been seen 213 times. Figures 6 and 7 show the normalization of her personality profile.

* * * About surgery for fecal incontinence: CLAUDETTE This woman was referred at age 42, to have a colostomy performed. She complained of massive chronic diarrhea and severe daily fecal incontinence for which she was constantly wearing diapers. ’helve years before, she had an anal sphincter injury at delivery. A sphincteroplasty had been performed six months before consultation and had failed to change the symptomatology. The medical evaluation of Claudette was unremarkable. She had a normal gastrointestinal morphology, no malabsorption and a normal bile salts turnover study. There was a lot of tension in the endoscopy room where she was to have a rigid proctoscopy with evaluation of the anorectal reflexes. The examina-

140

tion was cancelled and the patient was referred immediately to psychiatry. No psychiatric diagnosis was made. Functional evaluation revealed that the transit time was extremely rapid (Table 3). The anal canal was hypotonic at rest and the rectoanal inhibitory reflex present and of normal configuration. She was unable to strain and had great difficulty to squeeze the anal canal. The rectal volume was normal, but there was clear indication of hypoesthesia. At electromyography, there was no evidence of pudendal neuropathy and mapping around the anus showed normal activity without deficit anywhere. Here too, she had great difficulty to squeeze voluntarily the anal sphincter and upon straining, she had minimal decline in electrical activity. Bronchoprovocation with metacholine was positive, despite the absence of a history of asthma, and compatible with a diagnosis of irritable bowel syndrome. Defecography showed some degree of perineal descent, an open anal canal, and again, no

change in anorectal angulation upon straining or squeezing. The psychiatrist had concluded that the patient was normal and that the consultation had been totally unnecessary. Over the following weeks, however, it was learned that at age seven, Claudette had been raped in the vagina and in the rectum by her 24-year old uncle, who was the babysitter. For thirty-four years subsequently, she had nightmares every night about the rape. She had never confided in her parents, her teachers, her friends, or anyone else, including her numerous physicians. She had made several suicidal attempts and no one had asked questions about the abuse history. During follow-up, Claudette evolved from having constant diarrhea to alternate between diarrhea and incontinence, and painful constipation without incontinence. She had 16 sessions of anorectal biofeedback in order to regain control of her pelvic floor musculature. She was seen 119 times over four years. She moved from her hometown 500

TABLE 3 CLAUDETTE Functional gastrointestinal motor evaluation

Colorectal transit time (hours)

Right colon Left colon Rectosigmoid Colorectal

Rectometrogram (volumes in ml water)

Constant sensation volume (CSV) Maximum tolerable volume (MTV) Sensitivity index*

Anorectal manometry (pressure in ml water)

*- MTV-CSV MTV

PRESSURE PROFILE RECTUM Upper anal canal Lower anal canal Rectoanal inhibitory reflex ANISMUS Straining effort Upper anal canal Lower anal canal VOLUNTARY CONTRACTION Upper anal canal Lower anal canal

1988

1989

1990

0 0 1.2 1.2

13.2 28.8 10.8 52.8

10.8 51.5 31.2 93.6

100

120 0.17

30 40 Present and normal 0

0 0

10 8

141

Claudette PERSONALITY PROFILE

"1 0 '

s8

1 I

1 l

1 l

L

F

K

I

1

1

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I Hst D Hy Pdt Mf Pa Ptt Sct Mat Si 5K 4K 1K 1K 2K I

SCALES Fig, 8. Changes in personality profile of Claudette over two years of follow up. Note again the normalization of the MMPI profile. Again, the F scale has decreased, indicating the patient became less self-depreciativewhen her gastrointestinal disappeared. Here, the depression scale, although lowered to acceptable levels, remains higher than the hypochondria and hysteria scales at the end of follow up. As in Denise, note a lowering of the masculinity-femininityscale (MF) below 40, indicating Claudette has become more feminine. Although arbitrary, the MF scale of constipated women has been found in earlier studies to be completely independent of all other variables of the MMPI, and much lower (p c 0.001) than the scores of a control group of women with arthritis.

miles away to live near the hospital. Each visit consisted of reevaluation or her medical condition and discussion of the motility data; she could then discuss any matter of her choice about her life. Her friend came visiting once in a while and, since she was unable to refuse having sex, she was raped every time and went into a dissociative state. Eventually, she asked that he be informed about the childhood abuse; the man said he was not angry at her for the abuse, and quickly thereafter, she broke off the relationship. When she entered a new relationship, she discovered for the first time that she could have intercourse without dissociating. She began making drawings about the rape, in very vivid signs: the more her drawings were enriched in sordid details, the more her nightmares became void of them, until

she eventually stopped having any and began to have peaceful sleep. In 1992, she returned to her home town. By then her personality profile had normalized (Fig. 8), and she was well. An update in 1998 reveals that she is asymptomatic, without fecal incontinence, diarrhea, constipation or abdominal pain. She is working full time. She is back with her husband and only once in a while dissociates during intercourse. The follow-up of Claudette was much shorter than it was for Marthe and Denise. She was seen only 119 times over 4 years.

* * * About pelvic floor as a dysfunctional unit:

142

GINETTE The case and life history of Ginette exemplify that a history of sexual abuse exposes to a high risk of medicalisation if details of the case history are not placed in perspective and in parallel with those of the life history. It is also sending a message against over-generalizations which have often characterized this area of investigation, and about the difficulties to entangle simple correlations (for instance irritable bowel syndrome and sexual abuses) and more specific links. From 1986, when Ginette was 29, to 1993, she consulted numerous times in urology, in the outpatient clinic or in the emergency room, for pollakyuria, severe dysuria, interrupted micturition, and bladder retention. Debimetry and urodynamic studies led to a diagnosis of vesicourethral dyssynergia. Urethral dilation, done repeatedly, did not lead to remission until the nurse in charge of the teaching of selfcatheterization elicited a history of sexual abuse, first by her brother Roney, then by her father RenC. The brother went as far as an attempt to penetration; the father kissed her on the mouth and fondled her breasts. The nurse inquired about the entire pelvic floor symptomatology (Devroede, 1999). Ginette was also constipated, with a bowel movement every two weeks, but nobody had inquired about it. She had dyspareunia and never had an orgasm except when swallowing her mate’s sperm. Ginette was very articulate. At first visit, she said, “My father gave me quarters so that I would shut up,” to which she was told “What an obedient little girl!” Asked what I meant, I told her she had somatized her father’s words by shutting up bladder, vagina, and anus. Hearing this triggered in her a storm of emotions, sorrow first, rage later, to the point of slapping. She emerged from this crisis over the following weeks without urinary, genital or gastrointestinal symptoms. The sudden remission in the three spheres led to the concept of a ‘cloacal’ way of thinking (Devroede, 1997). She entered into a process of psychoanalysis, delving into the family background, and remained asymptomatic on a somatic point of view, desomatization remaining permanent. There were some longer-term stigmas, however. In particular, and this would have been difficult to evaluate objectively in the context of a

medical setting, she discovered four years later that until then, just as Marthe, she had experienced complete anesthesia of her clitoris, perineum and nipples, pleasure and cutaneous sensations returning much more slowly along the recovery process. This process also led her to question her social environment. Her mate had told the referring urologist early on when she came to the emergency room in acute bladder retention “If you want her to be anesthetized, doctor, I can hit her on the head.” She divorced him. She also broke off completely all links with the abusive father and brother and the condoning mother for almost two years. Today, she does not take appointments anymore, after having all along determined the frequency of them. Every six months or so, she comes in for advice and discussion on issues that she has already worked upon by herself. The ways to accompany Ginette along her process were multiple. Over one year, she had eleven sessions of biofeedback until anismus (Leroi et al., 1995a; Leroi et al., 1995b) was not recognized anymore at anorectal manometry. She entered a self-help group of patients with mixed pathologies, organic and functional, aware of the psychosocial elements of pathology in the biopsycho-social model of disease. She analyzed many of her dreams and did a lot of free association along the way of a psychoanalytical process. Early on, several of the dreams were nightmares turning around the theme of the abuses; later on, she became aware of the symbolic confusion between orifices and appendices (Sami-Ali, 1984). For instance, for a while, she had dreams about ‘red stools’, unrelated whatsoever with any hint of rectal bleeding, and spontaneously verbalized about “it is like if the front was in the rear,” meaning she thought that unconsciously there was, for her, a confusion between stools and menses (Devroede, 1999). On occasion, she asked to undergo hypnotic age regression and went back as far as age two. Finally, she was encouraged to use drawings and work on art production as tools of non-verbal communication (Feldman et al., 1993; Devroede, 1995a; Devroede, 1995b) with very deep unconscious meanings (McDougall, 1995). Today, she is completely asymptomatic and she says she is happy.

143

Lessons to be learned from patients The stories of Marthe, Denise, Claudette and Ginette have been selected out of a group of over 200 patients sexually abused during their childhood and seen in a consultation service of colorectal diseases. They had consulted for a somatic complaint without initially mentioning the abuse as primary complaint. The reason for selecting these stories is first to demonstrate the kaleidoscope and the plasticity of the array of symptoms covered under the umbrella of ‘somatization’. We have previously demonstrated that constipation (the problem of Marthe and Ginette at onset) and diarrhea (problem of Denise and Claudette at onset) were significantly associated with a history of sexual abuse (Leroi et al., 1995a). But if Marthe had consulted after the death of her father, and Denise during a bout of constipation, they would have been assigned to a different group when statistical analysis was performed. The alternation of constipation and diarrhea is a well-known feature of irritable bowel syndrome, but nothing is known either about the reasons for alternation or about the persistence and duration of one of the symptoms, except that more than fifty percent of patients are symptomatic five years after the onset of follow-up. Moreover, there is a major overlap between systems, as shown by the case history of Ginette. Obviously, different symptoms have different underlying pathophysiological mechanisms. These can be followed as scientific assurance of quality control by the available tests used to perform functional exploration. Often patients, particularly those with functional symptoms, overemphasize the reality of their complaint; conversely, once taken care of, they may minimize them, claiming they are better and that their bowels function more actively, for instance, in cases of constipation, while in fact, objective measurements of colorectal transit times with radiopaque markers have not changed. Another lesson to be learned from these longterm follow-ups is the importance of symptoms that are ignored by the majority of physicians and dismissed as ‘normal’ by some patients. For instance, Marthe and Ginette were what some people describe as ‘frozen’: they had cutaneous

hypo or anesthesia, a clear reminder of the witches of the Middle Ages who were burned when they did no feel the pain of needles. In these times, incest was common and quasi ‘normal’ (Michelet, 1966). Yet the ‘witches’ of modern times described in this chapter had never complained about the cutaneous anesthesia. They were amazed at the reappearance of skin sensations. Of course, this would not have occurred outside of a context of adequate communication and good doctor-patient relationship built over several years. I have pointed out elsewhere that transference and its omnipresent simultaneous counter-transference attitude does not occur solely on the couch of the analyst, but is seldom analyzed outside of this context (Devroede, 1995b). Issues related to the abuses often originate way before the Oedipal period and often turn around very archaic issues of gender identity and the notion of Otherness as coined by Joyce MacDougall (McDougall, 1995). A third reason for the need of prolonged followup is that short-term follow-up studies are often misleading, even if positive results are statistically significant. For instance, short-term follow-up for anorectal myectomy demonstrated its usefulness in patients with chronic idiopathic constipation but longer tern follow-up studies have failed to confirm this. Vice versa, using a rigid algorithm with at least one year of follow-up has succeeded in drastically reducing the number of indications for surgery in patients with chronic idiopathic constipation. In patients with irritable bowel syndrome, all available medications merely have a placebo effect; only psychotherapy has been found in one study to be useful. But the duration of follow-up in this excellent study was only three months long. Constipation was not altered by psychotherapy. Had she been part of this study, Marthe’s constipation would not have disappeared and she would have been rated as a failure. Even if I am a strong advocate of the psychoanalytical method as the inner counterpart of the scientific method for all external, measurable components in a human being, I must add that the extensive psychoanalytical literature on the treatment of patients with somatic disorders and the processes of somatization and desomatization is appallingly poor in terms of measuring bodily functions and diseases.

144

Thus, neither approach is mutually exclusive. Experiential phenomena are probably more conducive to desomatization than any purely scientific approach could achieve. Conversely, only bodily measurements through, in the present situation, transit studies with radiopaque markers, electromyography of the colon and the pelvic floor, anorectal manometry and now barostat studies can bring the proof that a process of desomatization and cure has occurred.

Sexual abuses and gastrointestinal disorders Epidemiology

Anecdotal reports associating bowel dysfunction and a history of sexual abuse during childhood (Devroede et al., 1989; Devroede, 1990) were followed by the pioneer paper of Drossman (Drossman et al., 1990). Of 206 patients seen consecutively in a university-based gastroenterology practice, 44% reported a history of sexual of physical abuse during childhood. Almost one-third of the abused patients had never discussed their experience with anyone, 60% had not discussed it with their family, and only 17% had informed their gastroenterologist. Patients with functional disorders were twice as likely as those with organic disease to report a history of forced intercourse, and more than ten times as likely to report a history of frequent physical abuse. They also complained more often about having chronic or recurrent abdominal pain and to having been submitted to more surgical procedures. By themselves, abuses were associated to pelvic pain, multiple somatic symptoms, and excess lifetime surgeries. Half of the abused patients reported sexual abuse both during childhood and adulthood, suggesting a pattern of revictimization. No explanation was offered for the link between abuse and surgery. The hypothesis of the author of this pioneer study that an abuse history was less prevalent in non-referral-based practices was not supported by our study comparing a university practice and a private practice of gastroenterology (Leroi et al., 1995a). Forty percent of patients suffering from lower functional digestive disorder gave a history of having been victims of sexual abuse in contrast

to only ten percent of patients with organic diseases, but the prevalence was the same in both settings. It was, however, four times greater in patients with lower than in those with upper functional motor disorders of the gastrointestinal tract. Presenting complaints of abused patients were mainly constipation or diarrhea, and anismus was recognized as a key objective finding related to abuse. When anismus is present, and it is easy to detect simply by asking the patient to strain during the rectal examination, the likelihood of a history of sexual abuse is ten times greater than when no contraction is felt (OR of 9.5; CI= 1.12-80). Thus this simple finding permits occasionally, in patients with severe abdominal pain, to make the hypothesis of a past history of sexual abuse, when no clear cut organic cause for the pain is recognized through the usual clinical process of diagnosis, and this without having to ask the question in a setting which is not always adequate for this, such as the emergency room. The prevalence of childhood abuse is clearly related to the severity of the symptomatology of irritable bowel syndrome as shown in a study of close to 2000 members of a large health maintenance organization (Longstreth and Wolde-Tsadik, 1993). The authors compared three groups: one without irritable bowel syndrome, one with less severe symptomatology, and one with more severe symptomatology. For instance, forced attempt of having sex was found in the history of 5% of the asymptomatic group, in 10% of those with mild symptoms, and 22%in those with severe problems. Some form of abuse was found four times as often in women as compared to men, and when people had been physically abused, they were much more likely (35%) to also have been sexually abused than not (9%). This association is deeply rooted in history. In Brittany, France, the association is vividly depicted in ‘Les enclos paroissiaux’, small cemeteries surrounding medieval churches. Soldiers whipping Christ are in full erection. Much more work is needed to understand why violence is so often associated with sexuality and will perhaps shed some light on the etiology of incest and sexual abuses. A review of all research articles and observational data was published in 1995 (Drossman et al.,

145

1995). Essentially it came to conclusions quite similar to the proband publication, namely that an abuse history is associated with gastrointestinal illness and psychological disturbance, appears more often in women and patients with functional, rather than organic, gastrointestinal disorders; is not usually known by the physician, and is associated with poorer adjustment to illness and adverse health outcome. Women are more often victims than men. The generalizability of these findings to other clinical settings seems to be acceptable (Scarinci et al., 1994). One important question is that of validation. The issue of false memory is now a fashionable subject, in a movement of pendulum exactly identical to that which made Sigmund Freud shift from the recognition of sexual abuses as a major factor during childhood with an impact on adult behavior to his elaboration of the theory of fantasies. Confirmation by the perpetrator of the abuse is of course ideal (Devroede, 1995; Mak et al., 1995), followed by adequate police records and confirmation with family and acquaintances. This is why the frequency of abuse may be grossly underestimated by the general population and the medical community alike, and why estimates widely differ from 6%-62%. It has even been written that up to 98% of abuse probably remains unreported (Russell, 1986; Arnold et al., 1990). A Canadian general population survey of a random sample of 9953 residents aged 15 years and older has shown that male children were more often (3 1%) physically abused than females (21%), but that conversely, little girls were three times as likely to be sexually abused (12.8% vs. 4.3%) (MacMillan et al., 1997). It is also possible to look at the problem the other way around and see how many abused subjects suffer medical consequences from the abuses. They indeed frequently develop not only gastrointestinal but also genitourinary symptoms (Felice et al., 1978; Rimsza et al., 1988), particularly if the abuses persisted over some period of time. Thus seventy-one percent of children abused for more than 24 months develop symptoms (Rimsza et al., 1988). One key finding is that in the short term after a rape, psychological symptoms are prominent, although in the long term physical symptoms may emerge and not be associated spontaneously to the

sexual abuses, such as chronic abdominal pain (Felitti, 1991). Another key finding is that abused subjects not only suffer from gastrointestinal and genitourinary symptoms, but respiratory and neurologic symptoms and mental problems (Lechner et al., 1993). Thus a history of abuse is deleterious to health and leads to symptoms and surgery. A lot of research remains to be done in terms of the epidemiology of sexual abuses and simultaneous or subsequent somatization in terms of gastrointestinal disorders. Even if we know that most abuses have been perpetrated at a young age, with victimization later on (Drossman et al., 1990), the exact impact of the age at which the victim has been abused and of the difference in respective ages of the perpetrator is not known. The nature of their relationship is also probably very important: whether it is intrafamilial, of which proximity, and which side (paternal or maternal), as well as extrafamilial and which proximity, are most likely key factors. It is not known if a homosexual and a heterosexual abuse trigger the same stigmata although weak data indicate they have a similar impact (Leroi et al., 1995a). The only female homosexual abuse we have seen occurred in a woman referred for ‘bowel obstruction’.In fact she had no pain, contrary to her claim, and enteroclysis was normal. She had had elsewhere a subtotal gastrectomy and sigmoid resection on the basis of similar, phony complaints, with the findings of normal specimens. When she was six, her mother forced her to cunnilingus; later on she gave the little girl for abuse by her brother first, her mate next. A husband who beat her followed, then many lovers, then a much older, verbally abusive husband from whom she dare not divorce. She was diagnosed as having factitious disorder (Munchhausen’s syndrome), was referred to psychiatry, and divorced. The repetitive nature of abuses, the multiplicity of partners, and the possibility for the victim to communicate with parents immediately after the crime are factors that need to be investigated. A final point to investigate is why a history of sexual abuse is found two to three times more often in patients with functional bowel syndrome than in those with organic digestive diseases, a fact uniformly confirmed (Drossman et al., 1990; Talley et

146

al., 1994; Leroi et al., 1995a; Delvaux et al., 1997). Body clues

Sexually abused women undergo almost four times as much surgery as healthy, non-abused controls, hysterectomy in particular (Leroi et al., 1995b). Regardless of the debate presently raging around the question of ‘false memories’ vs. real abuse (Devroede, 1995; Mak et al., 1995), it must be kept in mind that when a woman says she has been abused, she is at risk of more surgery. Moreover, another study has shown that a woman who has been raped by her father will have on average eight surgical procedures during her lifetime and that 70% of these surgeries will be performed on an absolutely unnecessary basis. She will also see eighteen non-psychiatric consultants and be hospitalized 30 times. Overall, it will take sixteen years between the time of first medical contact the identification of the abuse (Arnold et al., 1990). Gynecological and Gastroenterological procedures are particularly likely to be practiced unnecessarily, Thus normal hysterectomy, laparoscopy, appendectomy, and exploratory laparotomy are situations where the issue of abuse obligatorily must be raised preoperatively when a specific diagnosis has not yet been obtained. Whatbdoes this mean? Is the link between the surgeon and the patient in this situation of a symbolic nature? Is there phallic symbolism to a surgical knife? Certainly this is the view of some analysts, and the nature of many jokes in surgical circles and operating rooms could lend some support to this view. For sure, more questioning of what leads to unnecessary surgery in victims of sexual abuse must be done. The compulsion to repeat is a psychological process which leads people to repeat time and again the same situations, like the patient with acute Crohn’s colitis and multiple perineal fistulae who had an alcoholic father who physically abused her, then married a man with similar behavior, divorced him, then had a relationship with another alcoholic man who was not physically abusive, separated from him, then had a lover who was a former alcoholic which whom she had her first orgasm. He was twenty

years older than she was and had a son with Crohn’s disease, He abandoned her, then she returned to her ex-husband, who was in therapy and did not drink anymore and did not beat her anymore. But she longed for the sexual pleasure she had with the man who had abandoned her, so she left again and had numerous brief affairs before she realized she craved tenderness and wanted to be loved, and eventually went into long-term remission. This anecdote illustrates both the pathological repetition of the notion of what a man is, imprinted upon her by her father, and the curative evolution away from this pattern. It is possible that multiple ‘abusive’ surgeries in victims of sexual abuses stem from the victim looking for a magical surgical cure and from surgeons who are unable to resist the demands of the patients but are unable to resolve their patient’s dramatic complaints. Whatever the mechanisms, the issue of unnecessary surgery must be dealt with. Moreover, the fact that the act of abuse is linked to the act of surgery raises many questions about non-verbal communication as it pertains to sexual abuse, somatization, and medical care. The Freudian theory of fantasies, although it may be true unconsciously and may even be true in real situations of false memories, is very hard to follow when anorectal motility is found to be specifically perturbed in victims of sexual abuse as compared to non-abused women with anismus and healthy controls (Leroi et al., 1995b). Not only anismus is found in quasi all abused women (Leroi et al., 1995b) and is found to be present more often than in non-abused women, but also other abnormalities are found in an abused population: rectal hypoesthesia, a decreased amplitude of the rectoanal inhibitory reflex (the normal upper anal canal relaxation triggered by sudden brief rectal distension and related to rectal wall viscoelastic properties, and internal smooth muscle and sphincter relaxation), and a hypertonic anal canal. Moreover, when asked to perform simple voluntary movements, they have a hard time to squeeze the anus voluntarily, and they strain hardly at all when asked to do so (Leroi et al., 1995b). These are simple clinical maneuvers, much more objective than questioning about a possible history of abuse, that may eventually lead to a laboratory diagnosis

147

of abuse if properly tested in a prospective double blind study.

Sexual abuses and other medical conditions I have pointed out elsewhere (Devroede, 1999) that the pelvic floor is a functional unit, and that the tendency of modern scientific medicine to fragment it into different clinical problems variously referred to urologists, gynecologists, gastroenterologists or gastrointestinakolorectal surgeons is bound to lead to many diagnostic mistakes, without an integrated view of the patient’s diseases and dysfunctions. Moreover, forgetting about the psychosocial issues will prevent sexologists, psychologists and sociologists to be in touch with these patients, or vice versa, if patients consult them first, they will not have adequate care of their physical ailments. An integrated approach is exemplified by the good example of Ginette’s case and life history. Relatively few papers have emerged from the urological literature about a possible link between abuse and urinary symptoms (Fenster and Patterson, 1995; Ellsworth et al., 1995; McCarthy et al., 1996). Some authors report that a typical history which should immediately bring the possibility of abuse begins as such: “An 18-year old female was first seen for urinary retention, She had been using self-catheterization since the age of 15 years due to a history of urinary retention, intermittent flank pain and incontinence. She had abortions at 13 and 14 years of age, gave birth to a daughter at age 15, and had a spontaneous abortion when she was 20. A subsequent emergency admission at age 20 elicited a history of child abuse by her father beginning at age four and continuing for 11 years. The patient had a history of bed wetting for 11 years, that began shortly after the onset of sexual abuse” (Fenster and Patterson, 1995). In a series of 18 patients who had been sexually abused and had dysfunctional voiding (Ellsworth et al., 1995), an inquiry was also made about gastrointestinal symptoms. Urinary tract infections, and incontinence were prominent symptoms. Bladder capacity was often decreased but at times increased, and residual urine was commonly present. The urinary stream was intermittent with vesicourethral dyssynergia in the two

patients where it was looked for, a dysfunction completely akin to anismus. A past history of child abuse is found in one quarter of women with urinary tract infections. One fifth reported being victims of violent crimes and almost one sixth reported spouse abuse in the past year. These data were obtained from 1599 participants in a randomized, sociodemographically representative sample of United States women (Plichta and Abraham, 1996). There are more papers in the gynecology literature about abuse history. In the study just mentioned about urinary tract infection, there were also many gynecologic problems (Plichta and Abraham, 1996), chronic pelvic pain being a major one of them (Walker et al., 1988). The incidence is much greater than in control subjects with headache or pain-free controls (Walling et al., 1994a; Walling et al., 1994b). Pain (odds ration of 1-16) and gynecological surgery (1-17) are the two gynecological markers pointing towards a history of abuse in a gynecological population (Kirkengen et al., 1993). As was found in a gastroenterological (Drossman et al., 1990), although to a somewhat lesser extent, the majority of women who complained of chronic pelvic pain and had been sexually abused, had also been physically abused. A multivariate analysis demonstrated that childhood physical abuse was more strongly related to depression, anxiety and somatization that childhood sexual abuse (Walling et al., 1994b). Since chronic pelvic pain and irritable bowel syndrome are both associated with a past history of sexual abuse, and are both common disorders, it is logical to look for overlaps. Chronic pelvic pain occurs in 35% of women with an irritable bowel syndrome. Compared to women with irritable bowel syndrome alone, there is a greater likelihood of a history of abuse in women with emotionally, and, predictably, they also undergo an excess hysterectomy (Walker et al., 1996). From the previous paragraphs, a notion strongly emerges that the pelvic floor is a functional unit (Devroede, 1999) and that it would be a major mistake to focus solely on one or another system. This is why this chapter, although dealing mainly with gastrointestinal tract diseases and dysfunctions, also covers other pelvic floor dysfunctions.

148

But the pelvic floor is not apart and separate from the rest of a human being, despite the fact that our collective rituals, as well established by anthropologists (Frexinos et al., 1998) have tended to pretend that the anus and its vicinity belonged to someone else. Thus it is appropriate to look at the general well-being of victims of sexual abuse. Several studies have shown that an abuse history is associated independently with some other medical syndromes and with increased symptom reporting, such as pain (headache, back pain, myofascial pain, pelvic pain), eating disorders (Brown et al., 1997), alcohol and drug abuse (Drossman et al., 1995). French adolescent victims of rape experience similar difficulties as well as sleep disturbances, tobacco consumption, depressive symptoms, behavioral problems and suicidal attempts (Choquet et al., 1997). The increased nocturnal activity and impaired sleep maintenance has been confirmed (Glod et al., 1997). A large number of children, of the order of 50%, followed for several years after abuse remain sad or depressed, with low self-esteem and behavioral dysfunction (Tebbutt et al., 1997; Garnefski and Diekstra, 1997), sexually abused boys having considerably more emotional and behavioral problems (Garnefski and Diekstra, 1997). Similar findings were found in Australia in a five-year follow up case-control study (Swanston et al., 1997), and this makes it easier to understand better long term outcomes such as those exemplified earlier in this chapter. Childbirth complications, such as reported in Claudette, have also been found more frequently in abused women than in control subjects (Farley and Keaney, 1997). Well-designed epidemiologic studies are needed to determine whether the putative preferential association of abuse history with gastrointestinal illness is valid, or whether these observations are related to the greater attention recently given to the links between abuse and gastrointestinal disorders. Overall, it appears clearly however, that an abuse history is associated with poor health status. Abuse severity is also related to health status, a further agreement in favor of medical consequences to abuse. In cases of sexual abuse, 24% of the variance of health status is related to injury during abuse and to the presence of multiple perpetrators. In those of physical abuse,

39% of the variance is associated to rape or the multiplicity of episodes (Leserman et al., 1997). A meta-analysis of seven general population studies over 7550 women and 245 1 men has concluded that a history of sexual assault is associated with a poor subjective health, particularly in cases of multiple assaults or assaults by a spouse, and this regardless of gender, ethnicity, sampling method and the presence or not of depression (Golding et al., 1997). The problem is generating an extremely high cost (Irazuzta et al., 1997). Childhood sexual abuse is a poor preparation to happy sexualized relationships. Not surprisingly, and because of the compulsion to repeat, it was found to lead to early onset consensual sexual activity, teenage pregnancy, a multiplicity of sexual partnership, unprotected intercourse, sexually transmitted diseases, and revictimization, i.e. sexual assault after age 16 (Stock et al., 1997; Fergusson et al., 1997). This kind of life condition fits logically with a poorer health status as described above. The number of sexual partners of an abused victim, as compared to non abused adolescents, is twice as great when the abuse is ongoing, and four times as great when there is a past abuse history. The number of partners is even greater if the victim experienced both physical and sexual abuse (Luster and Small, 1997). Abused Afro-American women feel their partner does not care and are afraid he will leave. He is often physically abusive, particularly when asked to wear a condom. Many have had sexually transmitted disease, abortions, and worry about AIDS. Again, suggestive of a preferential link between abuse and the gastrointestinal tract, they are four times as likely to have a history of anal sex, as compared to those who have not been abused (Wingood and Di Clemente, 1997). Only one population-based study has been reported (Talley et al., 1994) and it has shown that abused subjects tended to be separated or divorced, a trend also observed in France (Delvaux et al., 1997). A final important element further obscures the global evaluation of the effects of early life abuse on adult behavior. Dissociation, common in sexually abused subjects (GBlinas, 1983), such as in Ginette, who dichotomized her cutaneous sensations, as well as Claudette and Marthe, permits the

149

patient to adapt to the experience. The process may range psychologically from isolated, brief flashback and nightmares to amnesia, and physically from relative cutaneous insensitivity to a total lack of sensations in case of trauma such as bums, as was the case in Marthe. It leads to a dissociative split between the psychic element and the somatic element in psychosomatic experiences, which makes a patient lose the awareness of what (s)he is living and what symptoms are concomitant (Farley and Keaney, 1997). Cognitive restructuring of stress then becomes an important element along the path of reassociation.

The process of desomatization A process is not a treatment. It is the personal way of an individual subject to grow and mature, and to recover from post trauma and unmet needs. Thus, if the source of the association between sexual abuse and subsequent bodily dysfunction is very archaic, and may antedate by many years the actual trauma, it is clear that no standard approach can be adopted to such patients, and that it has to be tailored made, i.e. patient-based rather than doctor-based, as too often is the case. Recognizing the abuse and letting the patient express one’s feelings about the trauma is but the first step. It was, however, fraught by a long repressed attitude of the medical community when faced with all sexual issues, including trauma, which made it for a long time abstain not only to ask questions about it, but as well on religious and politics. This was in an attempt to respect ‘values’ of the patient, without wondering about the value of these ‘values’, neither if they were hiding a hidden agenda. In general, physicians shy away from violence and sexuality. Reasons for this are complex. The majority of physicians are not trained to screen for, or manage patients who have been harmed by violence and sexual invasion; consequently, many do not want to open ’Pandora’s boxes’ by asking violence-related and sexualityrelated questions. They may also fear offending a patient and upsetting their office routine by taking too much time with the medical interview. Physicians’ own attitude towards violence, towards sexuality, and towards their impotence and frustra-

tion, when a patient is unable to run away from an abusive situation, can also be a major handicap (Plichta and Abraham, 1996). This calls for a need for physicians to educate themselves and learn how to deal with the process of transference and countertransference. But the fact remains that, given the large prevalence of violence and sexual abuses, they need to ask questions and be prepared both to deal with the reactions to their questions and to provide information about local community services (Plichta and Abraham, 1996). It has been demonstrated that self-exposure and cognitive restructuring is a better approach to victims of sexual aggression than simple progressive relaxation training (Echeburua et al., 1997). This may be partly explained by the fact that somatization and childhood abuse may involve a paradoxical pattern of hiding feelings and reality, while seeking acknowledgement of suffering. ‘Saying’ permits to break the vicious circle that occurs when the current physician denies the physical pain during adulthood, just the same as the abuser denied the emotional and physical pain during childhood (Morse et al., 1997). There are many ways to ask about the abuse history, but they should all be gentle, and respect the capacities of the victim to cope with the expression of the abuse (Drossman et al., 1995). Referral to a mental health professional should be consistent with the patient’s needs and expectations, and the clinician should not be surprised to discover that, at least at first, the patient prefers to see the non-mental health professional to whom (s)he confided, and even prefer the invasiveness of biofeedback reeducation of the pelvic floor (Leroi et al., 1996). Joint follow up between the mental health and non-mental health professionals is to be done if the patient so wishes, and if one of them requires an exclusive commitment immediately against the wish of the patient, (s)he should be replaced by another professional. Talking helps, but it would be grossly oversimplified to state that this is the only thing that matters. People express themselves via their intellect and words, via their emotions, via their bodily functions and via their sexuality, and this in various proportions according to their life histories. Exchange with male therapies, female therapies

150

and couples of therapists are also enormously different. The purpose of this chapter is not to expand on each of these variables, as it would take too much space to cover the spectrum of the different types of available psychotherapies. I have described elsewhere the algorithm of what goes on between a doctor and a patient (Fig. 9) and detailed some of the ways to foster their exchanges (Devroede, 1995b). There are important facts to remember, when accompanying victims of sexual abuses, and particularly the fact that they were invaded into their physical boundaries. Thus, when offered a choice, a patient who is constipated, has anismus, but was abused, will surprisingly refuse to see a psychologist or a psychiatrist, or even go to a

group of victims of abuses, but will accept to be sent for biofeedback reeducation of the pelvic floor, which implies anal penetration and all its symbolic meaning for the unconscious mind. Not surprisingly, also, they will react intensely with a lot of emotions to the biofeedback. Even then they will prefer to continue seeing the laboratory technician who performs the biofeedback than a psychologist, providing of course (s)he not only takes care of the patient but cares for the patient. This, of course, indicates a need for exchanges not only at technical level, but humanistic level between the involved professionals (Leroi et al., 1996). Another point of importance is the use of artwork, both as a way of symbolization of the

CONSULTING A PHYSICIAN Case History

v + Constipation -7

I

v

Attention

J -

v

I

Competence

-Wdght st001 -freqwncy

I

e Studies

I

-output

measures

-Evacuation

+

Problem

I

?

Solvlng

I

I

New

a

t

Fig. 9. An Algorithm of what occurs when a constipated patient, as often is the case for victims of sexual abuse, meets a physician. The scientific method reduces a suffering human being to a sick organ, to be measured. The areas in black indicate everything in the doctor-patient relationship which is not scientific, because it should be on an equal basis of two independent, autonomous subjects. At first, the patient is totally dependent and in need for help. Communication not only is more humane, but probably has a major impact on treatment outcome.

15 1

abuses, and as a nonverbal way of communicating any kind of trauma including sexual abuses (Feldman et al., 1993; Devroede, 1995a; Devroede, 1995b). We have found this extremely useful, as well as others (Burgess et al., 1982; Goodwin, 1982; Kelley, 1985; Manning, 1987; Briggs and Lehmann, 1989; Riordan and Verdel, 1991; Hagood, 1992; Trent, 1992; Sadowski and Loesch, 1993; Carpenter et al., 1997; Zinni, 1997). Although there is disagreement about the diagnostic value of artwork (Cohen and Phelps, 1985; Yates et al., 1985; Hibbard and Hartman, 1990), its therapeutic value has been demonstrated (Feldman et al., 1993), and this not only for victims of abuse, but as a tool to work through conflicts symbolically. There are relatively few data about adult artwork, but some psychoanalysts emphasize rightly that there are unconscious exchanges between sexuality, creativity and psychosomatic relationships (McDougall, 1995). From an excellent study in preschool children, it is very clear that very young children usually are not able to communicate their feelings clearly, particularly in a dysfunctional family (Carpenter et al., 1997). Thus, artwork, of any kind, at any level, in any context, social or therapeutic, is a canvas upon which the artist, or the patient, projects a glimpse of hisher inner world, their traits and their behavioral characteristics (Wohl and Kauffman, 1977; Carpenter et al., 1997). Artworks speak out things that words are unable to say. Of course, this means that in adult subjects, their artistic production speaks out about their inner child, and a lot of associated unconscious elements. Thus, it is only one of the key elements of the entire spectrum of non verbal ways of communicating with the patient (Devroede, 1995b). Another note of caution has to do with the diagnostic value and diagnostic meaning of drawings. Although it has been well established that family drawings of abused or neglected children significantly differ, in a number of items, from those of a control group, this, as indicated throughout this chapter, does not mean in a way that abuse is the bottom source of the problem, but is merely a key element. In depth interviews have been conducted with patients who have somatization disorder and a history of childhood abuse, in order to understand their illness behavior (Morse et al., 1997). This

study has shown that these patients have a paradoxical pattern of hiding feelings of reality, but at the same time seek acknowledgment of suffering. It has also been shown that increased insight decreases health care use, and therefore that exploration of patients experiences is useful. Themes that can be explored, with tact and sensitivity, trying never to break the patient’s defense mechanisms, are the abuse experiences, the emotional and behavioral reactions to the abuse, the relationship of abuse to the clinical symptomatology, the relationship of abuse to health care use, the past attempts to tell about abuses, and the relationship to previous physicians and physicians’ behavior (Morse et al., 1997). In a psychoanalytically-oriented approach, one might also explore the compulsion of repetition patterns, and help the patient discover links between present events and present symptomatology, without telling himher this link, for fear of depriving himher of the emotional and desomatization impact of the discovery. A systematic recall of the past as today reflects it is also very useful. One must always remember also that the medical community, in general, has a tendency to confuse imaginary diseases, i.e., fictitious disorders, with somatization disorders, which is really in the body. Even the official definition of ‘somatizationdisorder’ is a bit confused in this regard, which calls for “a tendency to experience and communicate somatic distress and symptoms unaccounted for by a pathologic findings, to attribute them to physical illness, and to seek medical help for them” (Lipowski, 1988). The key word of course in this definition is what means ‘pathologic’. For instance, anger makes the colon contract, more so in patients with an irritable bowel syndrome, who have abdominal pain, and are often abuse survivors, but anger is not a ‘pathology’, the contraction is in the large bowel and not in the head, and the patient, often abused, complains of abdominal pain only. Georg Groddeck, the pioneer in studies of psychosomatic relationship used to speak about the physician as being a catalyst. It could even better be said, since a human being is not just a chain of chemical reactions but a conglomerate of biological events that an ideal goal for physicians, particularly when they deal with victims of major childhood

152

trauma, should be to become an enzyme. To be discarded after use. Rather than the reverse, as it is when the doctor dismisses the patient, and, of course, remains in charge and control.

Is early life abuse a marker of lack of love and recognition? One of the most impressive clinical experiences, when accompanying subjects who try to recover from early life sexual and physical abuse, is that, regardless of the amount of emotions and the degree of verbalization they are capable to reach about the episodes of abuse, the symptoms which are part of the somatization background do not disappear until the subject begins to explore the family context during the period of time which antedates the abuse. One obtains the impression that the abuse, however traumatic it be, is like a black veil, or a screen memory, thrown over and hiding worse problems present sometimes quasi from conception time. Quite often, abused subjects have major gender identity problems. Joyce McDougall correctly said that the most difficult issues to bereave are those that separate psychic bisexuality from somatic monosexuality (McDougall, 1995). Psychoanalysts often refer to latent homosexual components present in every human being, many people believe we are all somewhat bisexual, admirers of the Orient claim we are all yin-yang. What one can hear for sure if one cares to listen and let patients speak out, is the number of women who wished to have been boys in order to avoid the trauma of abuse, and those who also say they were never accepted as women, because of the wishes of their mother, or father, or both. This, of course, is transmitted from generation to generation. Children are not the clones of their parents, and parents who are poorly identified to their own sex cannot generate children who are not somewhat confused about their gender identity because they are a poor mirror of a child trying to grow up as who he or she really is, apart and different from both parents (Dumas, 1985; Ancelin-Schutzenberger, 1998; Canault, 1998). Moreover, maternal distress, in case of disclosure of sexual abuse, correlates much more to her own assessment of child functioning than the child’s own assessment

of the situation, meaning she projects a lot, views herself in the child, and functions more at the transferential level than in an attitude of love (Leonard et al., 1997). Childhood sexual abuse is not only much more prevalent in a tribal context (Robin et al., 1997) than in a community setting (MacMillan et al., 1997; Fleming, 1997), but the perpetrator is also more often originating from an intrafamilial surrounding. This suggests that the more closely knit the family and community setting, the greater is the likelihood of abuse. In fact, it provides support to those who claim that the major source of violence in all societies is intrafamilial. Gender identity disturbances are definitely not limited to victims of childhood abuse, and in this regard should provide hope for those who have been victims that they can not only survive but recover and be well, if they face other issues than merely those of abuses. Indeed, artists quite often portray some degrees of gender identity disturbances, as vividly portrayed in an exhibit put together in Paris, in 1995-1996, in the Centre Georges-Pompidou (Femininmasculin- in one word- : le sexe de l’art) (1995). At a yet deeper level on the path of recovery, one must question the concept of the self. Who is really a subject? There must be a level of consciousness that cannot be damaged by severe traumas such as early life abuses. Here comes a convergence of millenary philosophical concepts such as those of Buddhism, and modem scientific findings. There are still two clearly different concepts of the self. For the Occident, the self is what refers to the original individual, entering into relationship with significant others, starting with the parents. But the Gerland sutra says something else as a Buddhist concept. There is no innate self-nature, and each individual is made of an infinite number of relations (s)he has with others. There are other major differences between East and West. In the West, marked by Christian values, the miracle is God, i.e. Love, made flesh; yet the Christian churches, along with countries, have had major difficulties to integrate this beautiful concept with a profound respect and love of the body. In the East, on the contrary, the nirvana is reached, when through karmic reincarnations, coming back into flesh is not necessary any more; yet the East has

153

always used the body with profound respect, and tantric values even claim that sexuality is the surest path to real spirituality, i.e, not religiosity, a purely mental activity of intelligence, reflection and memory. This is not the place to conclude about the respect values of two different concepts of Love but when it comes to the fact that sexuality is involved in abuses on children, some comparisons are in order. In the Occident, Love is incarnated and separated; it is not of myself but of somebody who is Other than I. In the Orient, at a different level of consciousness, the demarcation between subjects or objects? - becomes blurred. Everything becomes one being which cannot be named, the nothingness, close to the Christian notion of the communion of saints, but different because it includes everything in the Universe. There, nothingness is not the emptiness of a vacuum anymore, and can be better expressed as no-thingness. Even in the Occident, philosophers debate on how to solve these existential issues. Luc Ferry concludes there is such a thing as a transcendental attitude which includes truth, beauty, goodness and Love, the highest in the order of hierarchy. On the contrary, Andre ComteSponville says that the existential anxiety is only abolished when the subject accepts to be part of the Universe (Ferry and Comte-Sponville, 1998).What the victims of early childhood abuse raise as question, bound to disturb and induce countertransferential attitudes in clinicians of all obediences, is something as this: “I am I, and You are You, and if We meet, it is beautiful”, to paraphrase Frederic Perls in a different context. Along this path, we must, as my wife, Eve Beauskjour, once told me: “Learn to go beyond identifying to somebody, and identify that somebody”, i.e. an Other person. We are not the patients we meet, even if we project a lot on them. We must therefore, as Francis Peabody wrote over fifty years ago learn that the secret for taking care of a patient, is caring for the patient.

References Ancetin-Schutzenberger,A. (1998) Aie, mes dieux!, 10th edn. La MkridiennelDesc16e. de Brouwer, Paris.

Arnold, R.P., Rogers, D. and Cook, D.A.G. (1990) Medical problems of adults who were sexually abused in childhood. BMJ, 300: 705-708. Btrubt, J., Feldman, P., Devroede, G. and Lacombe, G. (1994) Le toucher, agent thtrapeutique. Sexologies, 3( 13): 43-48 (1lbnpartie); 3(14): 28-36 (,Iernepartie). Briggs, F. and Lehmann, K. (1989) Significance of children’s drawings in cases of sexual abuse. Early Ch. Dev. Care, 47: 131-147. Brown, L., Russell, J., Thornton, C. and Dunn, S. ( 1997) Experiences of physical and sexual abuse in Australian general practice attenders and an eating disordered population. Aust. NZ. J. Psychiatry, 31: 398-404. Burgess, A.W., McCausland, M.P. and Wolvert, W.A. (1982) Children’s drawings as indicators of sexual trauma. Perspect. Psychiatric Care, 19: 50-58. Canault, N. (1998) Comment paye-t-on les fautes de ses ancktres?, Desclte de Brouwer, Paris. Carpenter, M., Kennedy, M., Armstrong, A.L. and Moore, E. (1997) Indicators of abuse or neglect in preschool children’s drawings. J. Psychosoc. Nurs., 35: 10-17. Choquet, M., Darves-Bornoz, J.M., Ledoux, S., Manfredi, R. and Hassler, C. (1997) Self-reported health and behavioral problems among adolescent victims of rape in France: Results of a cross-sectional survey. Child Abuse Neg., 21: 823-832. Cohen, F.W. and Phelps, R.E. (1985) Incest markers in children’s artwork. Arts Psychother., 12: 265-283. Creed, F., Craig, T. and Farmer, R. (1988) Functional abdominal pain, psychiatric illness, and live events. Gut, 29: 235-242. Delvaux, M., Denis, P. and Allemand, H. (1997) Sexual abuse is more frequently reported by IBS patients than by patients with organic digestive diseases or controls. Results of a multicenter inquiry. Eul: J. Gastroent. Hepat., 9: 345-352. Devroede, G. (1990) Constipation and sexuality. Medical Aspects of Human Sexuality, February: 40-46. Devroede, G. (1995a) Cancer du cblon, personnalitk et immunitt ptritumorale: Rtflexions observations, hypothtses. In: L. Bessette, (Ed.) Le Deuil Comme Processus de Gutrison. Bereavement As A Healing Process, MNH Beauport, Qukbec, pp. 137-156. Devroede, G. (1995) Love, sex and incest. Humane Med., 11: 6-7. Devroede, G. (1995b) Psychophysiological considerations in subjects with chronic idiopathic constipation. In: S.D. Wexner and D.C.C. Bartolo, (Eds) Constipation: Etiology, Evaluation and Management, Butterworth Heinernann, Oxford, pp. 103-134. Devroede, G. (1997) La penste claocale. Sexologies, 6(24): 19-23 (1lhepartie); 6(25): 16-23 partie). Devroede, G. (1999) Front and rear: The pelvic floor in an integrated functional structure. Med. Hypoth., 52(2): 147-153. Devroede, G., Bouchoucha, M. and Girard, D. (1989) Constipation, anxiety and personality: What comes first? In: L. Bueno,

154

S . Collins and J.L. Junior, (Eds) Stress and Digestive Motility, John Libbey eurotext, London, Paris, pp. 55-60. Drossman, D.A., Leserman, J., Nachman, G., Li, Z.M., Gluck, H., Toomey, T.C. and Mitchell, C.M. (1990) Sexual and physical abuse in women with functional or organic gastrointestinal disorders. Ann. Int. Med., 113: 828-833. Drossman, D.A., Talley, N.J., Leserman, J., Olden, K.W. and Barreiro, M.A. (1995) Sexual and physical abuse and gastrointestinal illness. Review and recommendations. Ann. Int. Med., 123: 782-794. Dumas, D. (1985) L‘ange et le fanthme., Les tditions de minuit, Paris. Echeburua, E., De Corral, P., Zubizarreta, I. and Sarasua, B. (1997) Psychological treatment of chronic posttraumatic stress disorders in victims of sexual aggression. Behav. Mod., 21 : 433-456. Ellsworth, P.I., Merguerian, P.A. and Copening, P.E. (1995) Sexual abuse: Another causative factor in dysfunctional voiding. J. Urol., 153: 773-776. Farley, M. and Keaney, J.C. (1997) Physical symptoms, somatization and dissociation in women survivors of childhood sexual assualt. Women Health, 25: 33-45. Feldman, P.C., Villanueva, S., Lanne, V. and Devroede, G . (1993) Use of play with clay to treat children with intractable encopresis. J. Pediat., 122: 483-487. Felice, M., Grant, J., Reynolds, B., Gold, S., Wyatt, M. and Heald, F.P. (1978) Follow-up observations of adolescent rape victims. ‘Rape may be one of the more serious afflictions of adolescence with respect to long-term psychological effects’. Clin. Pediat., 17: 311-315. Felitti, V.J. (1991) Long-term medical consequences of incest, rape and molestation. South Med. J., 84: 328-331. Ftmininmasculin (1995). Le sexe de l’art. GallimardlElectra Centre Georges-Pompidou, Paris. Fenster, H. and Patterson, B. (1995) Urinary retention in sexually abused women. Can. J. Uml., 2: 185-188. Fergusson, D.M., Honvood, L.J. and Lynskey, M.T. (1997) Childhood sexual abuse, adolescent sexual behavior, ans sexual revictimization. Child Abuse Neg., 21 : 789-803. Ferry, L. and Comte-Sponville, A. (1998) Y a-t-il un sens B l’existence? In La Sagesse des Modernes, Robert Laffont, Paris. Fleming, J.M. (1997) Prevalence of childhood sexual abuse in a community sample of Australian women. Med. J. Ausfralia, 166: 65-68. Frexinos, J., Le Breton, D., Natrel, J.L., Denis, P., Edouard, A., Filali, A., Ndjitoyap-Ndam, E.C., Devroede, G., Zahar, A., Khneisser, H., Kawamura, T. and Desoutier, P. (1998) Voyage Sans Transit. Gkographie Mondiale de la Constipation., Editions Medigone Beaufour, Paris. Gamefski, N. and Diekstra, R.F. (1997) Child sexual abuse and emotional behavioral problems in adolescence: Gender differences. J. Am. Acad. Child Adol. Psychiatry, 36: 323-329. Gtlinas, D.J. (1983) The persisting negative effects of incest. Psychiatry, 46: 3 12-332.

Glod, C.A., Teicher, M.H., Hartman, C.R. and Harakal, T. (1997) Increased nocturnal activity and impaired sleep maintenance in abused children. J. Am. Acad. Child Adol. Psychiatry, 36: 1236-1243. Golding, J.M., Cooper, M.L. and George, L.K. (1997) Sexual assault history and health perceptions: Seven general population studies. Health Psychol., 16: 417-425. Goodwin, J. (1982) Use of drawings in evaluating children who may be incest victims. Child. Youth Sew. Rev., 4: 269-278. Guthrie, E., Creed, F.H. and Whonvell, P.J. (1987) Severe sexual dysfunction in women with the imtable bowel syndrome: Comparison with inflammatory bowel disease and duodenal ulceration. BMJ, 295: 577-578. Hagood, M.M. (1992) Diagnosis or dilemma. Drawings of sexually abused children. BE J. Proj. Psychol., 37: 22-23. Hibbard, R.A. and Hartman, G.I. (1990) Emotional indicators in human figure drawings of sexually victimized and nonabused children. J. Clin. Psychol., 46: 21 1-219. Irazuzta, J.E., McJunkin, J.E., Donadian, K., Arnold, F. and Zhang, J. (1997) Outcome and cost of child abuse. Child Abuse Negl., 21: 751-757. Kelley, S. (1985) Drawings: Critical communications for sexually abused children. Ped. Nurs., 11: 42 1-426. Kirkengen, A.L., Schei, B. and Steine, S. (1993) Indicators of childhood sexual abuse in gynaecological patients in a general practice. Scand. J. Prim. Health Care, 11: 276-280. Lalonde, M.O. (1995) Breaking the silence: Recovery from incest. Humane Med., 11: 29-33. Lechner, M.E., Vogel, M.E., Garcia-Shelton, L.M., Leichter, J.L. and Steibel, K.R. (1993) Self-reported medical problems of adult female survivors of childhood sexual abuse. J. Fam. Pract., 36: 633-638. Leonard, B.J., Hellerstedt, N.L. and Josten, L. (1997) Association of maternal psychological functioning to pathology in child sexual abuse victims. Iss. Ment. Health Nurs., 18: 587-601. Leroi, A.M., Bernier, C., Watier, A., Hkmond, A., Goupil, G., Black, R., Denis, P. and Devroede, G . (1995a) Prevalence of sexual abuse among patients with functional disorders of the lower gastrointestinal tract. Int. J. Colorect. Dis., 10: 200-206. Leroi, A.M., Berkelmans, I., Denis, P., HBmond, A. and Devroede, G . (1995b) Anismus, as a marker of sexual abuse: Consequences of sexual abuse on anorectal motility. Dig. Dis. Sci., 40: 1411-1416. Leroi, A.M., Duval, V., Roussignol, C., Berkelmans, I., Peninque, P. and Denis, P. (1996) Biofeedback for anismus in 15 sexually abused women. Int. J. Colorect. Dis., 11: 187-190. Leserman, J., Li, Z., Drossman, D., Toomey, T.C., Nachnian, G . and Glogau, L. (1997) Impact of sexual and physical abuse dimensions on health status: Development of an abuse seventy measure. Psychosomat. Med., 59: 152-160. Lipowski, Z.J. (1988) Somatization: The concept and its clinical application. Am. J. Psychiatry, 145: 1358-1368. Longstreth, G.F. and Wolde-Tsadik, G . (1993) Irritable boweltype symptoms in HMO examinees: Prevalence,

155 demographics and clinical correlates. Dig. Dis. Sci., 38: 1581-1589. Luster, T. and Small, S.A. (1997) Sexual abuse history and number of sex partners among female adolescents. Farn. Plan. Perspect., 29: 204-211. MacMillan, H.L., Fleming, J.E., Troeme, E., Boyle, M.H., Wong, M., Racine, Y.A., Beardslee, W.R. and Offord, D.R. (1997) Prevalence of child physical and sexual abuse in the community. Results from the Ontario Health Supplement. JAMA, 278: 131-135. Mak, A.J.W., Qroler, P.M., Jeffery, A.J., Wright, C.J. and Devroede, G. (1995) Sexual abuse and ‘false memories’. Humane Med., 11: 125-128. Manning, T.M. (1987) Aggressions depicted in abused children’s drawings. Arts PsychotheK, 1 4 15-25. McCarthy, T., Roberts, L.W. and Hendrickson, K. (1996) Urologic sequellae of childhood genitourinary trauma and abuse in men: Principles of recognition with fifteen case illustrations. Urology, 47: 617-621. McDougall, J. (1995) The Many Faces of Eros., WW Norton and Company, New York. Michelet, J. (1966) La Sorcibre., G.F. Flammarion, Paris. Morse, D.S., Suchman, A.L. and Frankel, R.M. (1997) The meaning of symptoms in 10 women with somatization disorders and a history of childhood abuse. Arch. Farn. Med., 6: 468476. Plichta, S.B. and Abraham, C. (1996) Violence and gynaecologic health in women < 50 years old. Am. J. Obstet. Cynaecol., 174: 903-907. Rimsza, M.E., Berg, R.A. and Locke, C. (1988) Sexual abuse: Somatic and emotional reactions. Ch. Abuse Negl., 12: 201-208. Riordan, R.J. and Verdel, A.C. (1991) Evidence of sexual abuse in children’s art products. School Counsellor, 39: 116-121. Robin, R.W., Chester, B., Rasmussen, J.K., Jaranson, J.M. and Goldman, D. (1997) Prevalence, characteristics and impact of childhood sexual abuse in a Southwestern American Indian tribe. Ch. Abuse Negl., 21: 769-787. Russell, D.E.H. (1986) The Secret Trauma: Incest in the Lives of Girls and Women., Basic Books, New York. Sadowski, P.M. and Loesch, L.C. (1993) Using children’s drawings to detect potential child sexual abuse. Elem School Guid. Counsel., 28: 115-123. Sami-Ali,M. (1984) Corps rkel, corps imaginaire., Dunod, Paris. Scarinci, LC., McDonald-Hale, M., Bradley, L.A. and Richter, J.E. (1994) Altered pain perception and psychosocial features

among women with gastrointestinal disorders and history of abuse: A preliminary model. Am. J. Med., 97: 108-1 18. Stock, J.L., Bell, M.A., Boyer, D.K. and Connell, F.A. (1997) Adolescent pregnancy and sexual risk-taking among sexually abused girls. Farn. Plann. Perspect., 29: 200-203. Swanston, H.Y., Tebbutt, J.S., O’Toole, B.I. and Oates, R.K. (1997) Sexually abused children 5 years after presentation: A case-control study. Pediatrics, 100: 600-608. Talley, N.J., Fett, S.L., Zinsmeister, A.R. and Melton, L.J.3. (1994) Gastrointestinal tract symptoms and self-reported abuse: A population-based study. Gastroenterology, 107: 1040-1049. Tebbutt, J., Swanston, H., Oates, R.K. and O’Toole, B.I. (1997) Five years after child sexual abuse: Persisting dysfunction and problems of prediction. Adol. Psychiatry, 36: 330-339. Trent, B. (1992) Art therapy can shine a light into the dark history of a child’s sexual abuse. Can. Med. Assoc. J., 146: 1412-1422. Walker, E., Katon, W., Harrop-Griffiths, J., Holin, L., Russo, J. and Hickok, L.R. (1988) Relationship of chronic pelvic pain to psychiatric diagnoses and childhood sexual abuse. Am. J. Psychiatry, 145: 75-79. Walker, E.A., Gelfand, A.N., Gelfand, M.D., Green, C. and Katon, W.J. (1996) Chronic pelvic pain and gynaecological symptoms in women with irritable bowel syndrome. J. Psychosom. Obstet. Gynaecol., 17: 3 9 4 6 . Walling, M.K., O’Hara, M.W., Reiter, R.C., Milburn, A.K., Lilly, G. and Vincent, S.D. (1994b) Abuse history and chronic pain in women. 11. A multivariate analysis of abuse and psychological morbidity. Obstet. Gynecol., 84: 200-206. Walling, M.K., Reiter, R.C., O’Hara, M., Milburn, A.K., Lilly, G. and Vincent, S.D. (1994a) Abuse history and chronic pain in women. I. Prevalence of sexual abuse and physical abuse. Obstet. Gynecol., 84: 193-199. Wingood, G.M. and Di Clemente, R.V. (1997) Child sexual abuse, HIV sexual risk, and gender relations of AfricanAmerican women. Am. J. Prevent. Med., 13: 380-384. Wohl, A. and Kauffman, B. (1977) Silent Screams and Hidden Cries., BrunnerMazel, New York. Yates, A., Beutler, L. and Crago, M. (1985) Drawings by child victims of incest. Ch. Abuse. Negl., 9: 183-189. Zinni, V.R. (1997) Differential aspects of sandplay with 10- and 11-year-old children. Ch. Abuse. Negl., 21: 657-668.

This Page Intentionally Left Blank

SECTION V

Influences of the internal environment on the brain

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Fds.) Progress in Brain Research, Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAFTER 11

Responses of afferent neurons to the contents of the digestive tract, and their relation to endocrine and immune responses John B. Furness'" and Nadine Clerc2

' Anatomy and Cell Biologx University of Melbourne, Parkville, Victoria 3052, Australia Laboratoire de Neurobiologie, CNRS, 31 Ch J. Aiguiec 13402 Marseille Cedex 20, France

Introduction The lining of the gastrointestinal tract is our largest external surface. This surface performs a difficult task: it needs to be in immediate contact with the contents of the intestine so that nutrients are efficiently absorbed, and it needs to protect against the intrusion of harmful entities, such as toxins and bacteria that may enter the digestive system with food. Thus the state of the gut needs to be monitored, and the gut itself needs to react to its contents. Signalling is through the immune system, endocrine hormones and the nervous system (Fig. 1). It is therefore no surprise that the digestive tract has three control systems that are more extensive than those of any other organ: the gut immune system, in which 70% of the body's immune cells are found (Heel et al., 1997); the gastroenteropancreatic endocrine system, which uses more than 30 identified hormones (Brand and Schmidt, 1995); and the enteric nervous system, which contains of the order of lo8 neurons (Furness and Bornstein, 1995). Thus, the gastrointestinal tract has an integrated response to changes in its lumenal contents. When this response is maladjusted, or is overwhelmed by *Corresponding author. e-mail: [email protected]

injurious substances, the consequences can be severe, as in cholera intoxication, or debilitating, as in the irritable bowel syndrome. Thus it is essential to obtain a full understanding of the manner in which the lumenal content of the gut is sensed, and how the body reacts to the information. This review deals principally with the neurons of the gut, but also includes descriptions of the other major gut signalling systems, the gut immune and gut endocrine systems.

The gut immune system The vast surface area presented by the lining of the gastrointestinal tract must defend the body against many potentially injurious substances in the food that accompanies food or drink, or is produced by degradation from food. At the same time, it must welcome and absorb nutrients into the body. To defend the otherwise highly permeable epithelial membrane, the small and large intestines have developed a number of specialisations, collectively called gut associated lymphoid tissue (GALT). Within the gut wall, the GALT includes antigen presenting cells (M cells) in the epithelial lining, collections of lymphocytes and immune-associated cells, namely, macrophages, eosinophils, mast cells and neutrophils (Fig. 2) (Blumberg and Stenson, 1995). The GALT consists of organized lymphoid aggregates, represented by Peyer's patches in the

160 Mucosal Epithelium

7

To Systemic Circulation

To Central

Fig. 1. Three types of message originate from the gastrointestinal rnucosa. Endocrine messages are in the form of hormones released from cells in the epithelium. The hormones enter the circulation, and can act at remote sites, but they also act locally, on nerve endings, on the epithelium and possibly on cells of the immune system. Immune messages are conveyed by circulating lymphocytes, that are activated by antigens presented to them from the lumen, or if the mucosal epithelium is breached, from their local tissue environment. Neural messages are conveyed by neurons whose sensitive endings are in the lamina propria, beneath the mucosal epithelium. These include neurons with cell bodies in the gut wall (IPANs) and extrinsic primary afferent neurons that carry information to the central nervous system (see also Fig. 3).

Fig. 2. The gut immune system includes M cells, that sample the lumen and present antigen molecules to macrophages and dendritic cells. After processing by these cells, the antigen is presented to lymphocytes, which enter the circulation, and after maturing return via the circulation to the mucosa. Other cells, such as mast cells, are activated when the mucosa becomes inflamed. E enterocyte; T,T-lymphocyte; B, B-lymphocyte; L, circulating lymphocyte.

small intestine, the mesenteric lymph nodes and solitary lymphoid nodules, plus numerous immune cells that are found in the mucosa at every point, even away from the Peyer’s patches; these constitute the non-organized lymphoid elements that include the intraepithelial lymphocytes, and lymphocytes in the lamina propria. The mucosa associated lymphoid tissue (MALT) of the gastrointestinal tract is similar to MALT of other organs, such as the respiratory tract and breast. MALT is characterized by the predominance of local IgA production and by localization signals through which activated lymphocytes derived from one mucosal surface can recirculate and localize selectively to the same and other mucosal surfaces. The connection between different mucosal surfaces permits immunity initiated at one anatomical site to protect other mucosal sites. The immune system of the intestine is involved whenever there is inflammation of the mucosa, whatever its initiating cause, e.g. inflammatory bowel disease, infection, allergy or injury. The mucosa becomes infiltrated with increased numbers of granulocytes, lymphocytes, macrophages, eosinophils and mast cells. These release a host of soluble mediators of inflammation, including cyto-

161

kines, prostaglandins, leucotrienes and histamine. Several of these substances act on enteric neurons and on the extrinsic afferent endings in the gut. In the intestine, antigens are sampled from the lumen by M cells, which are modified enterocytes (intestinal epithelial cells). The antigen that is transported across the epithelium by the M cell is taken up and processed by macrophages and dendritic cells and then presented to local T lymphocytes, which in turn stimulate local B lymphocytes. The B lymphocytes proliferate in the lamina propria of the mucosa, and produce antibody, primarily IgA. A proportion of the lymphocytes migrates to mesenteric lymph nodes and continues maturation and proliferation. These then enter the circulation and localise to intestinal and non-intestinal MALT, utilising specific receptors on post-capillary venules to guide their localisation (Heel et al., 1997). There is evidence for innervation of Peyer’s patches by enteric neurons (Krammer and Kuhnel, 1993), for the innervation of lymphoid follicles by extrinsic primary afferent fibers (Clerc and Mazzia, 1994) and for the presence of receptors for transmitters of extrinsic primary afferent and enteric neurons on lymphocytes (Heel et al., 1997).

The gut endocrine system The endocrine cells of the gastrointestinal tract are dispersed amongst the epithelial cells of its lumenal surface, and react to changes in the gut contents by releasing hormones that are, in general, targeted to

other parts of the digestive system. For example, cholecystokinin (CCK) is released from the duodenum in response to a meal, the major chemicals signalling this release being the products of the breakdown of fats and proteins. The major actions of CCK are on the pancreas to release digestive enzymes and on the gallbladder to trigger the emptying of bile salts into the duodenum. The targets for gut hormones include neurons (Table 1). CCK stimulates the endings of vagal afferent neurons, resulting in reflex inhibition of gastric emptying and in satiation (Smith et al., 1985; Blackshaw and Grundy, 1991). CCK also stimulates the endings of vagal neurons in the gallbladder; this appears to be the major mechanism through which gallbladder emptying is induced when CCK is released (Mawe, 1991). Vagal afferent neurons are also stimulated by 5-HT, in this case causing nausea. Motilin, which is released from duodenal endocrine cells, activates a motor program in the enteric nervous system, which results in migrating myoelectric complexes that sweep the contents of the small intestine in an anal direction. Conversely, gut endocrine cells are under neural control. For instance, release of gastrin is caused by activity in vagal nerve pathways, and vagal stimulation also causes 5-HT release.

The gut nervous system Monitoring and control of the digestive system through the nervous system is hierarchical. The gut contains an extensive collection of neurons, the

TABLE 1 Gastroenteropancreatic hormones that act on neurons Hormone

Major source

Major actions

Neural interactions

Gastrin

Gastric antrum

1gastric acid secretion

Release stimulated by gastric secretomotor neurons

Cholecystokinin

Duodenum

pancreatic secretion, gallbladder contraction

Excites motor neurons of gallbladder, stimulates vagal nerve endings

Motilin

Duodenum

Triggers migrating complexes

Stimulates neurons of MMC motor program in stomach and duodenum

5-HT (serotonin)

Small intestine

Initiates enteric reflexes, initiates nauseous responses

Stimulates vagal afferent and IPAN endings

162

enteric nervous system, within its walls (Furness and Bornstein, 1995).This intrinsic nervous system is capable of generating appropriate reflex responses to the contents of its lumen; for example, intrinsic reflexes generate mixing movements of the muscle, cause local changes in blood flow, and modulate secretion of water and electrolytes. The enteric nervous system also participates in reflexes between organs, for example, between the duodenum and stomach, to regulate gastric emptying. However, the enteric nervous system is itself under central nervous system control. Central direction of function is particularly prominent for the esophagus and stomach, which are controlled from the central nervous system through the vagus nerves, and for the large intestine, which receives central commands via the pelvic nerves. Throughout the gut, centrally controlled sympathetic neurons can override local reflexes; sympathetic neurons slow transit by inhibitory actions on the intrinsic neurons that control the non-sphincter muscle and by contracting the sphincters, restrict fluid loss into the gut lumen by acting on secretomotor neurons, and constrict the arteries that supply the gut (Furness, 1991). To determine the appropriate activities of motor neurons to the gut, it is necessary for the state of the gastrointestinal tract, and gut-dependent metabolic signals, to be monitored and communicated. Monitoring is by both neurons and endocrine cells. The neurons that detect the states of tissues are known as primary afferent neurons, primary because they are the first neurons in reflex pathways, and afferent because they run towards reflex control centres (neurons that run from control centres to effectors, such as muscle, are efferent neurons). Three broad classes of primary afferent neurons are associated with the gut: intrinsic primary afferent neurons (IPANs) with cell bodies and connections in the gut wall, extrinsic primary afferent neurons that have cell bodies in vagal and dorsal root (spinal) ganglia, and intestinofugal neurons that have cell bodies in the gut, but send processes to neurons outside the gut wall (Fig. 3; Furness et al., 1998). Intrinsic primary afferent neurons

Experiments conducted in the first decade of this century showed that motility reflexes could be

elicited in the intestine after the axons of extrinsic neurons had been severed and allowed to degenerate (Langley and Magnus, 1905). The obvious implication of this discovery was that the complete circuits of reflexes, including primary afferent neurons, were in the gut wall. However, it was not until 90 years later that IPANs were unequivocally identified (Furness et al., 1998). To this date, these neurons have been only identified in the small intestine: in fact, they have only been studied in detail in one species, the guinea pig. PANS have yet to be shown to exist in the stomach. In colon, reflexes elicited by distension may depend on axon reflexes in the endings of extrinsic primary afferent neurons, whereas mucosal reflexes in the colon appear to involve IPANs (Grider and Jin, 1994; see below). IPANs react to three types of stimuli, chemical changes in the intestinal lumen, distension of the intestine and mechanical distortion of the mucosa. Chemosensitive IPANs Intracellular records taken from nerve cell bodies in the guinea-pig small intestine have identified a class of intrinsic neuron that responds to chemicals (e.g. inorganic acid and short chain fatty acids at neutral pH) applied to the lumenal surface of the mucosa of the small intestine (Kunze et al., 1995; Bertrand et al., 1997). These neurons have a distinctive shape, known as Dogiel type Il morphology. Interestingly, when Dogiel (1899) originally described these neurons, he predicted that they would be primary afferent neurons. The neurons are multi-axonal with one or more axons that leads to, and branches into, the lamina propria of the mucosa just beneath the absorptive epithelium, and axons that lead into the myenteric ganglia and supply terminals around several nerve cells, that have been shown to include other IPANs, interneurons and motor neurons. In addition to their unique shapes and projections, these neurons have distinct electrophysiological properties. They have broad action potentials that are carried by both sodium and calcium currents and the action potentials are followed by early and late afterhyperpolarizing potentials. Moreover, these neurons, unlike all other neurons in the guinea-pig small intestine, do not receive fast excitatory

163

synaptic potentials. However, they do receive synaptic inputs through which slow excitation is mediated. This is unusual for primary afferent neurons, which generally do not receive any synapses on their cell bodies (Willis and Coggeshall, 1991). Thus IPANs seem to be unique in that their excitabilities can be modified by synapses at

the soma, and that action potentials traverse the soma, and are therefore subjected to the possibility of soma gating or enhancement (Furness et al., 1998). A majority of the inputs to IPANs arise from other IPANs and it has been suggested that these neurons form self-reinforcing networks (Furness et al., 1998).

Y Vagal Primary AIlerent Neurons

-I-

Spinal Primary Afferent Neuron

LM

I

>

I

Mucosa Mechanosensitive Neuron

Chernosensitive & Stretch Sensitive Neurons

~

MP

lntestinofugal Neuron

CM

SM

,Muc

rig. 3 . ine arrerent neurons 01 me intesune: inrnnsic pnmary arrerent neurons (IYANS), vagiu ana spinal pnmary arrerent neurons, and intestinofugal neurons. IPANs are multipolar and their terminals are confined within the wall of the intestine. Vagal and spinal primary afferent neurons are pseudounipolar and have collaterals that run to enteric ganglia. Vagal primary afferent neurons have cell bodies in the nodose ganglia, and their outputs are via terminals in the nucleus tractus solitarius, within the brain stem. The cell bodies of spinal primary afferent neurons are in dorsal root ganglia, their central processes end in the dorsal horns of the spinal cord and their peripheral axons pass via sympathetic ganglia to the intestine. They also provide collaterals in prevertebral sympathetic ganglia. Intestinohgal neurons are part of the afferent limb of intestino-intestinal reflexes that pass though sympathetic ganglia; their cell bodies are in the myenteric plexus. LM, longitudinal muscle; MP, myenteric plexus; CM, circular muscle; SM submucosa; Muc, mucosa.

164

The intrinsic primary afferent neurons that detect the presence of a bolus (mechanical stimulation of the villi) or changes in the chemical content of the gut lumen may do so indirectly, via the release of hormones from entero-endocrine cells. The reason to suppose an indirect action is that the mucosal epithelium separates the nerve endings from the lumenal environment (Fig. 1). A possible intermediate in enteric reflexes is 5-HT, which is a potent stimulant of the endings of IPANs (Kirchgessner et al., 1992; Bertrand et al., 1997). 5-HT is released when the mucosa is mechanically stimulated to elicit motility reflexes and the reflex responses are antagonised by drugs that block receptors for 5-HT (Foxx-Orenstein et al., 1996; Grider et al., 1996). Mechanical stimulation also causes c-fos induction in neurons that are deduced to be IPANs with cell bodies in submucosal ganglia, and this induction is blocked by 5-HT receptor antagonists (Kirchgessner et al., 1992). Other hormones that are contained in gut endocrine cells, such as CCK and motilin, are released by nutrients and act on neurons, but have not been tested for their possible roles as intermediates in enteric reflexes. Stretch sensitive IPANs Quite recently, neurons that respond directly to tension in the muscles have been identified (Kunze et al., 1998). These neurons also have Dogie1 type I1 morphology. Action potentials were recorded from the neurons when the intestine was stretched in vitro. However, the discharge of action potentials was abolished if the muscle contraction was prevented by muscle relaxants, either isoprenaline (an agonist of P-receptors for catecholamines) or nicardipine, a blocker of L-type Ca2+ channels. This indicates that active tension in the muscle contributes to the excitation of the tension-sensitive IPANs. The involvement of the muscle cells is interesting, because it has long been known that intestinal-muscle cells are directly sensitive to stretch and respond to it by depolarisation and contraction (Bulbring, 1955). This reaction of the smooth muscle may be integral to the response of IPANs during sustained stretch. The neurons themselves appear to possess mechanosensitive ion

channels and, when the neurons are distorted, they discharge action potentials (Wood, 1973). Consistent with this, the myenteric ganglia are distorted by muscle movement (Gabella and Trigg, 1984). Mucosal mechanoreceptors Functional evidence for IPANs with cell bodies in submucous ganglia comes from experiments in which activity-dependent induction of c-fos and activity-dependent uptake of dyes have been localized. C-fos immunoreactivity was detected in submucous nerve cells after the mucosa had been distorted by puffs of nitrogen gas that were ejected from a pipette (Kirchgessneret al., 1992). The c-fos expression was abolished by TTX, but not by blocking fast excitatory transmission between neurons with hexamethonium, suggesting that c-fos was produced in the cell bodies of IPANs that had processes in the mucosa. Styryl dyes, which are taken up by the endings of active neurons and transported back to the cell bodies, have also been used to identify PANS that are mucosal mechanoreceptors (Kirchgessner et al., 1996). Distortion of the villi by puffs of nitrogen gas caused styryl dye labelling in numerous submucosal and in few myenteric ganglia in the presence of hexamethonium. Fibers in myenteric ganglia, presumed to be axons from submucosal IPANs, were labelled. These results suggest that cell bodies of mucosal mechanoreceptors are in submucous ganglia and project to the myenteric plexus (Fig. 3). In the normal situation, in vivo, distension stimuli activate both mucosal mechanoreceptors and distension-sensitive neurons. This can explain why both myenteric and submucous neurons were revealed by uptake of activity dependent dyes after distension (Kirchgessner et al., 1996). Whether cell bodies of stretch-sensitive neurons are present in the submucosal ganglia, as well as in myenteric ganglia, has not yet been determined.

Rolesoflpms Intrinsic reflexes that affect motility, water and electrolyte secretion and blood flow all occur in the small intestine. Each of these is evoked by similar

165

stimuli, although it is not known whether the same, different or overlapping populations of IPANs contribute to motility, secretomotor, and vasomotor reflexes. Motor reflexes that have been studied as stereotyped responses of the circular muscle, excitation oral and relaxation anal, are evoked by distension of the muscle (which can be carried out without distorting the mucosa), chemicals applied to the mucosal surface, and by mucosal distortion (Hukuhara et al., 1958; Smith and Furness, 1988; Yuan et al., 1991). Secretomotor reflexes are initiated physiologically by chemical or mechanical interaction of lumenal contents with the mucosa, or pathologically by toxins, such as cholera toxin or enterotoxins, in the lumen (Frieling et al., 1992; Cooke and Reddix, 1994). Enteric reflexes also cause bicarbonate secretion in response to duodenal acidification (Flemstrom, 1994). The enteric secretomotor circuits consist of IPANs with their endings in the mucosa and nerve circuits that pass through the myenteric and submucosal plexuses and feed back to secretomotor neurons with cell bodies in the submucosal ganglia (Cassuto et al., 1983; Diener and Rummel, 1990; Frieling et al., 1992; Cooke and Reddix, 1994). The secretomotor neurons stimulate the epithelial cells to pump chloride ions, which are accompanied by water, into the lumen. Local vasodilator reflexes in the small intestine are caused by mechanical or chemical irritation of the mucosa, and substantial evidence indicates that the vasodilator neurons are intrinsic to the intestine and transmission from them is predominantly noncholinergic (Vanner et al., 1993; Vanner and Surprenant, 1996). It is presumed that the first neurons in these reflexes are IPANs, but this has not been directly shown. In fact, of the reflexes in the intestine, the intrinsic vasomotor reflexes are the least studied. Histochemical studies suggest that the same motor neurons have axons that branch to supply both the secretory epithelium and arterioles, thus some secretomotor and vasodilator reflexes may share the same final neurons (Furness et al., 1987). This makes physiological sense, as the secreted water and electrolyte comes indirectly from the vasculature.

Thus the roles of IPANs are to signal changes in the state of the intestine that are consequences of the presence and nature of its contents. The information is conveyed to other neurons of the enteric nervous system that integrate the information and cause appropriate changes in mixing and propulsive activity, in water and electrolyte transport and in local blood flow. Extrinsic primary aferent neurons

There are two groups of extrinsic primary afferent neurons, vagal primary afferent neurons with cell bodies in the nodose ganglia and axons that reach the gut via the vagus nerves, and spinal primary afferent neurons, with cell bodies in dorsal root ganglia (Fig. 3). The axons of spinal primary afferent neurons in the thoracic and lumber regions pass through sympathetic ganglia to reach the gut via splanchnic and mesenteric nerves. The axons of most spinal primary afferent neurons with cell bodies in sacral ganglia follow the pelvic nerves to reach the colon and rectum. Extrinsic primary afferent neurons have been recorded from directly, by electrodes placed in nerves or on their cell bodies in sensory ganglia, or their existence has been deduced indirectly, by monitoring behavioural or physiological consequences of their activation. Some of the information conducted by extrinsic primary afferent neurons is directly perceived, for example, sensation of gastric or intestinal fullness, various pain sensations (cramp, colicky pain, sharp pain), warmth and the sensation of gastric emptiness. Perception of the state of the gastrointestinal tract is subject to psychosensory modulation (Coffin et al., 1994; Accarino et al., 1997). Perception is enhanced when attention is paid to the gut and is diminished with inattention. Sensation is also diminished when painless somatic stimuli are simultaneously applied (Coffin et al., 1994). Other information is indirectly perceived, for example, we feel sated after a meal, but the feeling of satiety is not directed to a particular organ (although it may be felt at the same time as an organ-directed sensation). Finally, much of the activity of gastrointestinal primary afferent neurons is not perceived. This is very well demonstrated by tests on fistula

166

patients who report no sensation when the healthy stomach is probed, or in patients in whom the intestinal lining is cut to take a biopsy under endoscopic guidance (Sengupta and Gebhart, 1994). A classification of extrinsic primary afferent neurons by the stimuli to which they respond, by the sensations that they give rise to, and by the functions that they regulate is given in Table 2. It is possible to define low and high threshold mechanosensitive afferent neurons, the low threshold neurons responding to tissue states that occur regularly in the normal gastrointestinal tract, and high threshold neurons responding to stimuli that are at the extreme of normal, or are pathological. For example, moderate distension of the stomach stimulates low threshold mechanoreceptors that cause reflex relaxation of the proximal stomach and

rhythmic mixing contractions in the distal stomach. This occurs during normal digestion, and is unnoticed by the subject. Greater distension activates higher threshold mechanoreceptors and causes feelings of fullness. Further distension can lead to discomfort and pain. Painful sensations are more likely to be felt, and occur at lower thresholds, if the gut is inflamed. This is because the nerve endings in inflamed tissues are sensitised. In general, low threshold afferent endings are vagal in origin and high threshold nerve endings are of spinal origin. Thus, pain is conducted to the central nervous system via spinal afferent neurons. In humans, cutting the spinal afferent nerves abolishes pain due to strong distension of the stomach, intestine or gallbladder (Ray and Neill, 1947; Bingham et al., 1950). However, some spinal

TABLE 2 Gastrointestinal afferent neurons - vagal and spinal. Simplified classification of neurons, sensations and regulated functions REGION

AFFERENT NEURON V = vagal; S =spinal

SENSATIONS

REGULATED FUNCTION

Esophagus

V Mechanoreceptor (L) V Thermoreceptor S Mechanoreceptor

None Heat Discomfort, pain

Propulsion

V Mechanoreceptor (volume) V Mucosal receptor (chemical, acid?) V Hormone receptor S Mechanoreceptor

Fullness, satiety

Gastric volume

None

Acid secretion, vomiting Feeding Behaviour, vomiting

Stomach

Satiety Discomfort, pain, nausea

Vomiting

Gastric emptying and acid secretion

Duodenum

V Mechanoreceptor (L) V Hormone receptor S Mechanoreceptor

None Satiety Discomfort, pain

Jejunum, ileum

V Mechanoreceptor (L) S Mechanoreceptor*

None Discomfort, pain

Entero-enteric inhibitory reflexes

Liver

V Osmoreceptor V Hormone receptor V Glucose receptor

Thirst Satiety

Osmoregulation? Feeding, glucose metabolism

Gall bladder

S Mechanoreceptor*

Discomfort, pain, biliary colic

Entero-enteric inhibitory reflexes

Pancreas

V Insulin receptor S Chemoceptor?

Pain

Colon, rectum

S Mechanoreceptor* S Chemoreceptor

Fullness, urge to defecate Discomfort, pain

Defecation Entero-enteric inhibitory reflexes

Anal canal

S Mechanoreceptor, Thermoreceptor, Nociceptor

Touch Discomfort, pain

Defecation Guarding

* These and probably other receptors are sensitised and become reactive in inflammation L = low threshold receptor

I67

primary afferent neurons can also be activated by low threshold stimuli (Clerc and Mei, 1983; Sengupta et al., 1990; Pan and Longhurst, 1996). Vagal primary afferent fibers respond to various stimuli. Some are activated by mechanical probing of the esophageal, gastric or duodenal mucosa (Leek, 1977; Cottrell and Iggo, 1984). Some are also activated by distension of the gut wall or by muscle contraction (Sengupta and Gebhart, 1994). Responses have been recorded to intralumenal chemicals, including acid (Clarke and Davison, 1978), absorbable carbohydrates (Mei, 1978), lipids (Melone, 1986) and fatty acids (Jeanningros, 1982).Vagal afferent fibers also respond to changes in osmolarity (Mei and Gamier, 1986) and to temperature (El Ouazzani and Mei, 1982). It is not clear whether or not vagal afferents are selectively and directly activated by nutrients. Activation of vagal primary afferent neurons by carbohydrates might be secondary to motility reflexes (Cottrell, 1984) or secondary responses to endocrine hormones that are released by nutrients. For example, extrinsic primary afferent neurons that innervate the mucosa are activated by 5-HT (Blackshaw and Grundy, 1993) and CCK (Blackshaw and Grundy, 1990). CCK and 5-HT are both released from epithelial cells, but their roles in initiation or modulation of extrinsic primary afferent messages remain to be determined. Because the majority of the information that is communicated to the central nervous system does not come to conscious sensation, it is impossible to know how specific and fine-detailed is the information. It may in fact be a summed information that informs the central nervous system of the general state of the digestive organs. This is suggested by the substantial sizes of receptive fields of the afferent neurons (e.g., Powley et al., 1994). It is also suggested by the responses of the organs, e.g., acid or enzyme secretion, gallbladder contraction or gastric relaxation, none of which appear to be finely graded. Moreover, hormones can be expected to diffuse locally and, via the blood stream, also affect many afferent nerve endings. Finally, mucosal receptors appear to be polymodal, and there is no evidence to suggest that their responses to different stimuli might be decoded in the central nervous system.

Roles of extrinsic prirnaiy afferent neurons

The roles of these neurons are to signal to the central nervous system information that is necessary to regulate organs and behaviours that are beyond the immediate, local territories of the afferent endings, local control being primarily the province of IPANs. The information that reaches the central nervous system is obtained from the detection of several qualities of the gastrointestinal tract, including the state of distension, chemicals in the lumen, and the presence and degrees of tissue injury and inflammation. This can be decoded and interpreted consciously as satiety, pain or hunger. The afferent information can also be used to direct functions automatically, for example, esophageal propulsion, gastric relaxation in response to a meal, gastric acid secretion, all through the vagus, defecation through the pelvic nerves, and control of water and electrolyte transport and blood flow in relation to the relative needs of all organs, via sympathetic motor pathways. Intestinojigal neurons

The intestinofugal neurons represent a very unusual class of neurons. Their cell bodies are in the gut wall, and their processes run towards the central nervous system and form synapses in prevertebral sympathetic ganglia (Fig. 3). They were deduced to be present by Kuntz and his collaborators in the 1930s. These investigators found that distension of one region of the gastrointestinal tract caused inhibition of motility in other regions, and that these entero-enteric inhibitory reflexes persisted after connections with the central nervous system were severed, so long as the integrity of connections with prevertebral sympathetic ganglia were maintained (Kuntz and Saccomanno, 1944). Methods to study these reflexes in vitro were developed in the early 1970s (Crowcroft et al., 1971) and since that time the organisation of the pathways has been studied in considerable detail (Szurszewski and Miller, 1994). The cell bodies of intestinofugal neurons are in the myenteric plexus. They are most numerous in the large intestine, in the small intestine they increase in number distally, and they are rare in the stomach. The axons of intestinofugal

168

neurons make excitatory, cholinergic, synapses with the cell bodies of sympathetic neurons which project back to the gut (Fig. 3). Roles of intestinofugal neurons

substance cause transmitter release by acting on sites on nerve endings, for example calcium channels, without generating action potentials. Cells in damaged or inflamed tissue produce neuroactive substances, including cytokines [interleukins, tumor necrosis factor a (TNFa) and

The roles of intestinofugal neurons have been analysed almost exclusively in relation to motility control, although they also innervate sympathetic neurons whose function is to inhibit secretion of water and electrolytes in the intestine (Furness, 1991). The neurons that affect motility are in the afferent limbs of entero-enteric inhibitory reflexes. These reflexes appear to act primarily on sites in the gastrointestinal tract that are more proximal than the sites from which they are initiated (Szurszewski and Miller, 1994). Thus the reflexes are one of the mechanisms by which more distal parts of the intestine regulate the more proximal regions from which they receive products of digestion; other mechanisms of entero-enteric regulation include actions of hormones that are released in response to nutrients, and reflexes via the vagus (Malagelada and Azpiroz, 1989).

The interactions between afferent neurons, immune and endocrine systems The lamina propria is a milieu in which the secreted products of inflammatory cells, endocrine hormones and afferent nerve endings interact with receptors on nerve endings and cells of each of the other two systems (Fig. 4).Amongst the events that occur are the activation or sensitisation of afferent nerve endings by inflammatory mediators, the actions of neurotransmitters released by axon reflexes on other axons, immune cells, arteriole diameter, vascular permeability and on the epithelium and the actions of endocrine cell products on nerve endings (and possibly on immune cells). Axon reflex is the term used to describe the passage of an action potential from one afferent nerve ending to a place of axon bifurcation and then to another nerve ending, to cause transmitter release. It is also feasible that hormones, cytokines or other

Fig. 4. Illustration of interactions between endocrine hormones, neurotransmitters and the products of immune and inflammatory cells in the lamina propria of the mucosa. Irritation of the mucosa can release prostaglandins (PG) and endocrine hormones. In the inflamed or immunologically challenged lamina propria, macrophages (Mph) and mast cells release cytokines (Cyt K), as well as PG and histamine, which can act on nerve endings. The nerve endings also take up substances such as nerve growth factor (NGF) and cytokines that are transported back to the nerve cell body and change gene expression. Capillaries carry hormones and immune signals from other regions, and can also release active compounds, such as bradykinin (BK). The mixture of endocrine, epithelial, neural and inflammatory products act on afferent neurons, the gut epithelium and immunocytes to produce integrated reactions to local stimuli.

169

leukemia inhibitory factor (LIF)] from macrophages and histamine from mast cells. Moreover, inflammation promotes the formation of bradykinin. Each of these substances acts on neurons. It is long established that intestinal inflammation causes intestinal hyperalgesia (Kinsella, 1940), which may be contributed to by a number of substances released as part of the inflammatory reaction. For example, visceral afferents are stimulated by bradykinin (Lew and Longhurst, 1986), and in human, intraperitoneal bradykinin causes pain of abdominal origin (Lim et al., 1967). 5-HT, presumably released from gut endocrine cells, is released in experimental intestinal anaphylaxis and stimulates the endings of vagal afferent fibers via 5-HT3 receptors (Castex et al., 1995). Pharmacological experiments suggest that tachykinins, presumably released by axon reflexes, contribute to sensitization of gut afferent nerve endings (Julia and BuCno, 1997). Moreover, inflammatory mediators stimulate increased synthesis of nerve growth factor (NGF) by fibroblasts and its production from mast cells (Lewin, 1995). NGF is a neurotrophic factor for unmyelinated afferent neurons that is produced in target tissues and is retrogradely transported to their cell bodies, where it stimulates the production of neurotransmitters by the neurons and may be involved in promoting axonal proliferation within tissues (Lewin and Mendell, 1993). Retrograde transport of LIF also affects neurotransmitter production in spinal primary afferent neurons (Thompson et al., 1998). Tachykinins, usually measured as substance P, and CGRP are contained in a majority of spinal afferent neurons (Green and Dockray, 1988; Julia and BuCno, 1997) and are also in intrinsic primary afferent neurons, at least in the guinea-pig (Costa et al., 1996). Tachykinins and CGRP cause vasodilatation and plasma cell extravasation, thus contributing to increased blood flow and the access of immunocytes to the tissue. These substances may have significant effects in restricting the deleterious consequences of tissue damage. In the stomach, CGRP released from afferent nerve endings after injury contributes to mucosal protection and reduces the degree of ulceration (Holzer and Holzer-Petsche, 1997). Transmitter release from afferent endings in the colon also reduces the

severity of damage consequent on inflammation, at least in the acute phase (Reinshagen et al., 1996).

Conclusions It is well established that neurons, endocrine hormones and the immune system all react to changes occurring in the gut. Although many of the effects mediated through these systems have been elegantly dissected, the subtleties of their interactions are still being unravelled. In pathological conditions, such as inflammation, it has been easy to reveal the participation of many factors released from the inflamed tissue, including substances from neurons, immune cells and endocrine cells, but it remains difficult to analyse how the interactions between these factors produce the overall inflammatory response. It is probable that changed neuronal sensitivities are involved in disorders such as irritable bowel syndrome, but how these changes are brought about, and how they may be modified to alleviate the pathology requires further investigation.

Acknowledgements This work was supported by a grant from the National Health & Medical Research Council of Australia (grant no. 963213). We thank Soibhan Lavin for excellent assistance in production of the figures and manuscript.

References Accarino, A.M., Azpiroz, F. and Malagelada, J.-R. (1997) Attention and distraction: effects on gut perception. Gastroenterology, 113: 415422. Bertrand, P.P., Kunze, W.A.A., Bornstein, J.C., Fumess, J.B. and Smith, M.L. (1997) Analysis of the responses of myenteric neurons in the small intestine to chemical stimulation of the mucosa. Am. J. Physiol., 273: G422(3435. Bingham, J.R., Ingelfinger, F.J. and Smithwick, R.H. (1950) The effect of sympathectomy on abdominal pain in man. Gastroenterology, 15: 18-3 1. Blackshaw, L.A. and Grundy, D. (1990) Effects of cholecystokinin (CCK-8) on two classes of gastroduodenal vagal afferent fibre. J. Auton. New. Syst., 3 1: 191-202.

170 Blackshaw, L.A. and Grundy, D. (1991) Locally and reflexly mediated effects of cholecystokinin-octapeptideon the ferret stomach. J. Auton. New. Syst., 36: 129-138. Blackshaw, L.A. and Grundy, D. (1993) Effects of 5-hydroxytryptamine on discharge of vagal mucosal afferent fibres from the upper gastrointestinal tract of the ferret. J. Auton. New. Syst., 45: 41-50. Blumberg, R.S. and Stenson, W.F. (1995) The immune system. In: T. Yamada, D.H. Alpers, C. Owyang, D.W. Powell and F.E. Silverstein (Eds), Textbook of Gastroenterology, 2nd Edn., J.B.Lippincott, Philadelphia, pp. 111-140. Brand, S.J. and Schmidt, W.E. (1995) Gastrointestinal hormones. In: T. Yamada, D.H. Alpers, C. Owyang, D.W. Powell, and F.E. Silverstein (Eds), Textbook of Gastroenterology, 2nd Edn., J.B.Lippincott, Philadelphia, pp. 25-7 1. Bulbring, E. (1955) Correlation between membrane potential, spike discharge and tension in smooth muscle. J. Physiol.(Lond.), 128: 200-22 1. Cassuto, J., Siewert, A., Jodal, M. and Lundgren, 0. (1983) The involvement of intramural nerves in cholera toxin induced intestinal secretion. Acta Physiol. Scand., 117: 195-202. Castex, N., Fioramonti, J., Fargeas, M.J. and BuCno, L. (1995) C-fos expression in specific rat brain nuclei after intestinal anaphylaxis: Involvement of 5-HT3 receptors and vagal afferent fibers. Brain Res., 688: 149-160. Clarke, G.D. and Davison, J.S. (1978) Mucosal receptors in the gastric antrum and small intestine of the rat with afferent fibres in the cervical vagus. J. Physiol. (Lond.), 284: 55-67. Clerc, N. and Mazzia, C. (1994) Morphological relationships of choleragenoid horseradish peroxidase-labeled spinal primary afferents with myenteric ganglia and mucosal associated lymphoid tissue in the cat esophagogastricjunction. J. Comp. Neurol., 347: 171-186. Clerc, N. and Mei, N. (1983) Thoracic esophageal mechanoreceptors connected with fibers following sympathetic pathways. Brain Res. Bull., 10: 1-7. Coffin, B., Azpiroz, F. and Malagelada, J.-R. (1994) Somatic stimulation reduces perception of gut distention in humans. Gastroenterology, 107: 1636-1642. Cooke, H.J. and Reddix, R.A. (1994) Neural regulation of intestinal electrolyte transport. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 3rd Edn., Raven Press, New York, pp. 2083-2132. Costa, M., Brookes, S.J.H., Stele, P.A., Gibbins, I., Burcher, E. and Kandiah, C.J. (1996) Neurochemical classification of myenteric neurons in the guinea-pig ileum. Neuroscience, 75: 949-967. Cottrell, D.F. (1984) Mechanoreceptors of the rabbit duodenum. Q.J. Exp. Physiol., 69: 677-684. Cottrell, D.F. and Iggo, A. (1984) Mucosal enteroreceptors with vagal afferent fibres in the proximal duodenum of the sheep. J. Physiol. (Lond.),354: 496-522. Crowcroft, P.J., Holman, M.E. and Szurszewski, J.H. (1971) Excitatory input from the distal colon to the inferior mesenteric ganglion in the guinea-pig. J. Physiol. (Lond.), 219: 443-461.

Diener, M. and Rummel, W. (1990) Distension-induced secretion in the rat colon: Mediation by prostaglandins and submucosal neurons. Eur. J. Pharmacol., 178: 47-57. Dogiel, AS. (1899) Uber den Bau der Ganglien in den Geflechten des Dames und der Gallenblase des Menschen und der Siiugetiere. Arch. Anat. Physiol. (Leipzig), Anat. Abt. Jg., 1899: 130-158. El Ouazzani, T. and Mei, N. (1982) Electrophysiologic properties and role of vagal thermoreceptors of lower esophagus and stomach of cat. Gastroenterology, 83: 995-1001. Flemstrom, G. (1994) Gastric and duodenal mucosal secretion of bicarbonate. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 3rd Edn., Raven Press, New York, pp. 1285-1309. Foxx-Orenstein, A.E., Kuemmerle, J.F. and Grider, J.R. (1996) Distinct 5-HT receptors mediate the peristaltic reflex induced by mucosal stimuli in human and guinea pig intestine. Gastroenterology, 111: 1281-1290. Frieling, T., Wood, J.D. and Cooke, H.J. (1992) Submucosal reflexes: distention-evoked ion transport in the guinea-pig distal colon. Am. J. Physiol., 263: (391496. Furness, J.B. (1991) Roles of sympathetic noradrenergic neurons supplying the gastrointestinal tract. Brain Gut Interact., 7: 91-102. Furness, J.B. and Bornstein, J.C. (1995) The enteric nervous system and its extrinsic connections. In: T. Yamada, D.H. Alpers, C. Owyang, D.W. Powell and F.E Silverstein (Eds), Textbook of Gastroenterology, 2nd Edn., J.B. Lippincott, Philadelphia, pp. 2-25. Furness, J.B., Costa, M., Rokaeus, A,, McDonald, T.J. and Brooks, B. (1987) Galanin-immunoreactive neurons in the guinea-pig small intestine: their projections and relationships to other enteric neurons. Cell Tissue Res., 250: 607-6 15. Furness, J.B., Kunze, W.A.A., Bertrand, P.P., Clerc, N. and Bornstein, J.C. (1998) Intrinsic primary afferent neurons of the intestine. Pmg. Neurobiol., 54: 1-18. Gabella, G. and Trigg, P. (1984) Size of neurons and glial cells in the enteric ganglia of mice, guinea-pigs, rabbits and sheep. J. Neurocytol., 13: 49-71. Green, T. and Dockray, G.J. (1988) Characterization of the peptidergic afferent innervation of the stomach in the rat, mouse and guinea-pig. Neuroscience, 25: 181-193. Grider, J.R. and Jin, J.-G. (1994) Distinct populations of sensory neurons mediate the peristaltic reflex elicited by muscle stretch and mucosal stimulation. J. Neurosci., 14: 2854-2860. Grider, J.R., Kuemmerle, J.F. and Jin, J.-G. (1996) 5-HT released by mucosal stimuli initiates peristalsis by activating 5-HT4/5-HT,, receptors on sensory CGRP neurons. Am. J. Physiol., 270: G7784782. Heel, K.A., McCauley, R.D., Papadimitriou, J.M. and Hall, J.C. (1997) Peyer's patches. J. Gastmenterol. Heptatol., 12: 122-136. Holzer, P. and Holzer-Petsche, U. (1997) Tachykinins in the gut. Part 2. Roles in neural excitation, secretion and inflammation. Pharmacol. Ther., 73: 219-263.

171 Hukuhara, T., Yamagami, M. and Nakayama, S. (1958) On the intestinal intrinsic reflexes. Jap. J. Physiol., 8: 9-20. Jeanningros, R. (1982) Vagal unitary responses to intestinal amino acid infusions in the anaesthetised cat: a putative signal for protein induced satiety. Physiol. Behav., 28: 9-21. Julia, V. and BuBno, L. (1997) Tachykininergic mediation of viscerosensitive responses to acute inflammation in rats: evidence for a CGRP-induced release of tachykinins. Am. J. Physiol., 272, G1414146. Kinsella, V.J. (1940) Sensibility in the abdomen. B,: J. Surg., 27: 449-463. Kirchgessner, A.L., Liu, M.-T. and Gershon, M.D. (1996) In situ identification and visualization of neurons that mediate enteric and enteropancreatic reflexes. J. Comp. Neurol., 371: 270-286. Kirchgessner, A.L., Tamir, H. and Gershon, M.D. (1992) Identification and stimulation by serotonin of intrinsic sensory neurons of the submucosal plexus of the guinea pig gut: activity-induced expression of Fos immunoreactivity. J. Neurosci., 12: 235-248. Krammer, H.-J. and Kuhnel, W. (1993) Topography of the enteric nervous system in Peyer’s patches of the porcine small intestine. Cell Tissue Res., 272: 267-272. Kuntz, A. and Saccomanno, G. (1944) Reflex inhibition of intestinal motility mediated through decentralized prevertebra1 ganglia. J. Neurophysiol., 7: 163-170. Kunze, W.A.A., Bornstein, J.C. and Furness, J.B. (1995) Identification of sensory nerve cells in a peripheral organ (the intestine) of a mammal. Neuroscience, 66: 1 4 . Kunze, W.A.A., Furness, J.B., Bertrand, P.P. and Bornstein, J.C. ( 1998) Intracellular recording from myenteric neurons of the guinea-pig ileum that respond to stretch. J. Physiol. (Lond.), 506: 827-842. Langley, J.N. and Magnus, R. (1905) Some observations of the movements of the intestine before and after degenerative section of the mesenteric nerves. J. Physiol. (Lond.), 33: 34-5 1. Leek, B.F. (1977) Abdominal and pelvic visceral receptors. B,: Med. Bull., 33: 163-168. Lew, W.Y.W. and Longhurst, J.C. (1986) Substance P, 5-hydroxytryptamine and bradykinin stimulate abdominal visceral afferents. Am. J. Physiol., 250: 465473. Lewin, G.R. (1995) Neurotrophic factors and pain. Neuroscience, 7: 227-232. Lewin, G.R. and Mendell, L.M. (1993) Nerve growth factor and nociception. Trends Neurosci., 16: 353-359. Lim, R.K.S., Miller, D.G., Guzman, F., Rodgers, D.W., Rogers, R.W., Wang, S.K., Chao, P.Y. and Shih, T.Y. (1967) Pain and analgesia evaluated by the intraperitoneal bradykinin-evoked pain method in man. Clin. Phamzacol. The,:, 8: 521-542. Malagelada, J.-R. and Azpiroz, F. (1989) Determinants of gastric emptying and transit in the small intestine. In: J.D.Wood (Ed.), Handbook of Physiology, Sect 6: the Gastrointestinal System, American Physiological Society, Bethesda MD, pp. 909-938.

Mawe, G.M. (1991) The role of cholecystokinin in ganglionic transmission in the guinea-pig gall-bladder. J. Physiol. (Lond.), 439: 89-102. Mei, N. (1978) Vagal glucoreceptors in the small intestine of the cat. J. Physiol. (Lond.), 282: 485-506. Mei, N. and Gamier, L. (1986) Osmosensitive vagal receptors in the small intestine of the cat. J. Auton. New. Syst., 16: 159-170. Melone, J. (1986) Vagal receptors sensitive to lipids in the small intestine of the cat. J. Auton. New. Syst., 11: 231-241. Pan, H.-L. and Longhurst, J.C. (1996) Ischaemia-sensitive sympathetic afferents innervating the gastrointestinal tract function as nociceptors in cats. J. Physiol. (Lond.), 492: 841-850. Powley, T.L., Holst, M.-C., Boyd, D.B. and Kelly, J.B. (1994) Three-dimensional reconstructions of autonomic projections to the gastrointestinal tract. Microsci. Res. Tech., 29: 297-309. Ray, B.S. and Neill, C.L. (1947) Abdominal visceral sensation in man. Ann. Surg., 126: 709-724. Reinshagen, M., Patel, A., Sottili, M., French, S., Sternini, C. and Eysselein, V.E. (1996) Action of sensory neurons in an experimental rat colitis model of injury and repair. Am. J. Physiol., 270: (379486. Sengupta, J.N. and Gebhart, G.F. (1994) Gastrointestinal afferent fibers and sensation. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 3rd Edn., Raven Press, New York, pp. 483-5 19. Sengupta, J.N., Saha, J.K. and Goyal, R.K. (1990) Stimulusresponse function studies of oesophageal mechanosensitive nociceptor in sympathetic afferents of opossum. J. Neurophysiol., 64: 796-812. Smith, T.K. and Furness, J.B. (1988) Reflex changes in circular muscle activity elicited by stroking the mucosa: An electrophysiological analysis in the isolated guinea-pig ileum. J. Auton. New. Syst., 25: 205-218. Smith, G.P., Jerome, C. and Norgren, R. (1985) Afferent axons in abdominal vagus mediate satiety effect of cholecystokinin in rats. Am. J. Physiol., 249: R638-R643. Szurszewski, J.H. and Miller, S.M. (1994) Physiology of prevertebral ganglia. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 3rd Edn., Raven Press, New York, pp. 795-877. Thompson, S.W.N., Priestley, J.V. and Southall, A. (1998) GPl30 cytokines, leukemia inhibitory factor and interleukin6, induce neuropeptide expression in intact adult rat sensory neurons in vivo: time-course, specificity and comparison with sciatic nerve axotomy. Neuroscience, 84: 1247-1255. Vanner, S., Jiang, M.-M. and Surprenant, A. (1993) Mucosal stimulation evokes vasodilation in submucosal arterioles by neuronal and nonneuronal mechanisms. Am. J. Physiol., 264: (32024212. Vanner, S. and Surprenant, A. (1996) Neural reflexes controlling intestinal microcirculation. Am. J. Physiol., 27 1: G223-G230.

172 Willis, W.D. and Coggeshall, R.E. (1991) Sensory Mechanisms ofthe Spinal Cord, Plenum Press, New York. Wood, J.D.(1973) Electrical discharge of single enteric neurons of guinea-pig small intestine. Am. J. Physiol., 225: 1107-1 113.

Yuan, S.Y., Fumess, J.B., Bomstein, J.C. and Smith, T.K. (1991) Mucosal distortion by compression elicits polarized reflexes and enhances responses of the circular muscle to distension in the small intestine. J. Auton. Nerv. Sysr.. 35: 2 19-226.

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Brain Research,Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 12

The controls of eating: brain meanings of food stimuli Gerard P. Smith Department of Psychiatry, Joan and Sanford I. Weill Medical College of Cornell University and the Edward W Bourne Laboratory, New York-Presbyterian Hospital, Westchester Division, 21 Bloomingdale Road, White Plains, NY 10605, USA

Introduction Eating is a simple behavior with complex functions. Eating not only serves metabolism, nutrition, and energy balance, but it is also used for individual psychological needs and to satisfy social imperatives of culture, ethnicity, and religion. In this chapter, I use results from experiments in the rat to trace the recent progress in understanding eating. This progress includes new information about the central pattern generator for the rhythmic movements of eating, the recognition that the meal is the functional unit of analysis, and a new classification of the controls of eating based on preabsorptive receptors in the gastrointestinal mucosa from the tip of the tongue to the end of the small intestine, visceral afferent feedbacks, and reciprocal connections between the forebrain and hindbrain. The chapter ends with a discussion of three prominent conscious experiences of eating - pleasure, fullness, and tranquilization. The complete ignorance about how these experiences relate to the activity of the central neural networks that control eating reminds us of the deep problem of the relationships among brain, mind, and behavior that lurks under the surface of eating behavior.

Central pattern generator of eating Eating is a behavior that consists of a sequence of rhythmic oral movements. The movements depend *Corresponding author. Tel: (914) 997-5935; Fax: (914) 682-3793; e-mail: [email protected]

on the kind of food - mastication of solids, licking and lapping of liquids. We know most about licking of liquid nutrients because computer techniques permit the recording of every lick emitted during a bout of eating and facilitate the analysis of the rate and pattern of licking, i.e. its microstructure. Microstructural analysis has demonstrated two characteristics of licking that reveal fundamental aspects of its neural organization (Davis and Smith, 1992; Davis, 1998). The first is that rats lick different liquid foods at a relatively constant rate of 5-8 sec; this range is more narrow in an individual rat. Such a stereotyped motor response is characteristic of the output of a neural network functioning as a pattern generator (Travers et al., 1997). Anatomical and physiological studies suggest that the cpg is in the caudal brainstem distributed within the lateral, medial, and intermediate reticular formation. Second and third order neurons of the cpg are widely distributed throughout the brainstem and extend cephalad as far as the substantia nigra (Fay and Norgren, 1997). The second characteristic of licking is that licks occur in clusters separated by intervals of nonlicking > 500 msec. The interval between clusters is caused by inhibition of the central pattern generator (cpg) and the clusters of licking are caused by excitation of the cpg that exceeds contemporary inhibition or removes it. Whenever meal size changes in response to changes of diet or other experimental and genetic manipulations, the pattern of licking changes, but the momentary lick rate does not (Davis, 1998).

174

This demonstrates that all the controls of eating are exerted upstream of the motor neurons of the network that produce the constant momentary rate of licking. Thus, the most fundamental neurological problem posed by the controls of eating is to identify and understand the excitatory and inhibitory input mechanisms that turn the central pattern generator (cpg) on and off. In Sherringtonian terms, the cpg is the final common path for licking. Its location in the caudal brainstem means that this region is the spinal cord for ingestion in that it has the same neurological relationship to eating as the spinal cord has to posture and movement. If the caudal brainstem is the spinal cord for ingestion, it should receive afferent input from its relevant sensory surface - the mucosa of the gut from the mouth to the end of the small intestine and it should also organize motor outputs to the same regions. The caudal brainstem’s connections with the gut fulfill these requirements. The distributions of the afferent and efferent innervation in the nucleus tractus solitarius (NTS) and the dorsal motor nucleas of the vagus are grouped according to the gross functional compartments of the mouth, stomach, and small intestine identified by Pavlov in his pioneering analysis of the control of gastrointestinal glandular secretions (Pavlov, 1910). (Afferent stimuli from and efferent output to the large intestine may affect eating under circumstances of emotional disturbances or disease, but they are probably not normally involved in the control of meal size.)

A meal - the functional unit of eating A bout of eating separated by an interval of noneating longer than 5 min is a meal. The successful analysis of the controls of meal size must account for two important characteristics of meals. One is their binary behavioral form - a meal is occurring or not. The other is the large range of meal sizes. Even though the form is fixed - meals begin and end abruptly - the duration of a meal and the amount of food consumed during the meal varies over at least an order of magnitude in different situations. (Although meal size and duration covary

under many circumstances, this is not always the case because different patterns of ingestion can produce different intakes within the same duration of eating.) Note that the dynamic form of meals is innate, not learned. This is true whether preweanling mammals are suckled or artificially fed. Thus, the dynamic form of eating we call meals is a residue of natural selection represented in the genome and sensitive to mutation. Indeed, most kinds of genetic obesity in rodents are characterized by abnormally large meals. Food intake is completely determined by the size and number of meals. Thus, the meal is the functional behavioral unit of eating (Smith, 1996). Because a meal has a dynamic behavioral form, the neurological controls of a meal must be complex enough to account for the form and its dynamic range. If meal size were fixed, the neural control system would be simpler. What is known about the neural organization of the controls of meals other than that it contains a cpg in the caudal brainstem? Until about twenty years ago, most of the neural investigation of eating attempted to account for the initiation of eating by the arousal of motivational processes driven by hypothesized metabolic deficits. When the hypothesized metabolic deficits that could account for the initiation of eating under normal conditions were not found and it became clear that the initiation of eating and meal size were under separate controls (Campfield and Smith, 1990), a number of researchers began to investigate the controls of meal size independent of the problem of what initiated eating.

Peripheral feedbacks The behavioral analysis of meal size revealed an important characteristic of the neural organization of eating - it is under strong, peripheral feedback control produced by food stimuli as they pass from the mouth to the stomach and into the small intestine during a meal (Smith, 1996). The mucosal surface from the tip of the tongue to the distal small intestine is a sensory sheet peppered with receptors responsive to the mechanical and chemical stimuli of ingested and digested food. The importance of peripheral sensory feedback was demonstrated by sham feeding, a procedure

175

that prevents ingested food from accumulating in the stomach or small intestine by allowing it to drain out of an esophageal or gastric cannula or to be withdrawn through a gastric catheter (Smith, 1998a). When sham feeding occurs, eating is prolonged and meal size always increases. The magnitude of the increase depends on the type of diet, the duration of food deprivation, and other experimental conditions. The fact that preventing ingested food from accumulating in the stomach or small intestine results in an increase in meal size demonstrates that postingestive stimuli provide negative-feedback control of eating. The adequate stimuli for this negative feedback are volume in the stomach (Smith, 1998a), and volume and chemicals in the small intestine (Greenberg, 1998). The negativefeedback effects of these preabsorptive stimuli are carried to the medulla by afferent fibers in the abdominal vagal nerves. Visceral afferent fibers of neurons in the dorsal root ganglia that reach the medulla by ascending spinal pathways are probably also involved. Sham feeding also revealed that orosensory stimuli provide positive feedback that maintains eating. For example, the rate of intake during sham feeding is a function of nutrient concentration in the liquid food (Weingarten and Watson, 1982). The orosensory receptors are sensitive to gustatory, somatosensory, thermal, and ‘common chemical’ stimuli. Activation of these preabsorptive receptors is transduced into afferent neural activity in cranial nerves 5 , 7, 9, and 10; these afferent fibers also project directly into the caudal brainstem. An additional sensory input can come from olfactory receptors stimulated by volatile molecules sprayed up the retronasal passage during eating. Some of these olfactory neurons influence the caudal brainstem by multisynaptic pathways. The relative potencies of the peripheral positiveand negative-feedbacks determine meal size. Eating continues as long as positive feedback exceeds negative feedback. When the relative potencies are computed to be equal by an unidentified comparator mechanism(s) in the brain, eating stops and the meal ends (Smith, 1996). If these peripheral feedbacks were the only control of meal size, then the size of a meal of a

specific nutrient, e.g. 10% sucrose, would be the same under a variety of conditions. But this is not the case. The size of a meal of 10% sucrose varies over a wide range depending on the experimental situation. This means that there must be controls of meal size that are not dependent on the peripheral orosensory and postingestive feedbacks.

Direct and indirect controls of meal size The adequate stimuli for these other controls include light, temperature, time, space, hormones (e.g. estrogen and leptin), circulating nutrients and metabolites, the presence of conspecifics, and the relative density of foods and predators in the ecological niche (Smith, 1996). Furthermore, all of these stimuli can acquire control over eating through associative mechanisms active during prior experience (see Table 1). Despite their diversity, these numerous controls are distinguished from the peripheral feedback controls by the fact that they do not activate the preabsorptive gut receptors directly. Because peripheral sensory feedback controls of eating are active during every meal, these other controls must act indirectly by modulating the potency of the peripheral positive or negative feedbacks on eating. I have used the criterion of direct access to gut receptors to classify these two kinds of controls as Direct Controls and Indirect Controls (Smith, 1996).

Neurological criterion of direct and indirect controls I have recently added a neurological criterion to differentiate these controls (Smith, 1998’13). This TABLE 1 Indirect controls of meal size Rhythmic (diurnal, ovarian) Metabolic Adipose Experiential Cognitive Ecological Social Cultural Pathological (infection, cancer) Note: The names of the controls refer to non-exclusivecategories of stimuli. This is especially true of associative learning as a result of prior experience; learning can occur in all of these categories.

176

criterion is that the caudal brainstem is necessary and sufficient for the Direct Controls, but the forebrain and its connections with the caudal brainstem are required for the Indirect Controls. This criterion was derived from the neurological fact that whenever food stimuli activate mechanical and chemical receptors during a meal, primary afferent fibers carry the transduced information in the rate and pattern of their activity directly to the caudal brainstem where it influences the network of premotor and motor neurons that constitute the cpg of licking. Afferent activity produced by acceptable food in the mouth stimulates the network of the cpg, while afferent activity from the stomach and small intestine inhibits it.

Chronic decerebrate rat Thus, the caudal brainstem receives all of the afferent input produced by the Direct Controls of eating. Does it also have the comparator function necessary for the integration of the stimulating and inhibiting afferent activity into the dynamic form of a meal? This is analogous to the question of whether an animal with a section of the spinal cord at the level of C, can integrate afferent input into the sequence of stepping that constitutes walking. The answer is that the high spinal cat walked when its feet touched a moving treadmill (Miller and van der Meche, 1976). To determine if the caudal brainstem can integrate afferent inputs to produce a meal, Grill, Norgren, and their colleagues investigated the controls of eating in rats after decerebration at the level of the superior colliculus (Grill and Norgren, 1978a). This lesion disconnects the midbrain, pons, and medulla from the forebrain. When liquid foods, such as milk and carbohydrate solutions, are infused into the mouth of the chronic decerebrate rat, the rat emits rhythmic ingestive movements and swallows the food until it lets the infused liquid drain out of its mouth. This is the behavioral sign that the meal has ended. Meal size varies under these conditions as a function of concentration and kind of nutrient, and time since the last meal (Grill and Kaplan, 1990). It is decreased by an intragastric preload and it increases during sham feeding (Grill and Kaplan, 1992). This integration

of afferent input from the gut into meals of various sizes indicates that the disconnected brainstem has a comparator function that integrates the relative potencies of the positive and negative feedbacks into a motor command that turns the cpg on or off. Thus, like walking in the high spinal cat, integrated eating occurs in the decerebrate rat when food contacts orosensory receptors. Although the capacity of the decerebrate rat to eat meals demonstrates that the disconnected brainstem has sufficient neural complexity to integrate the Direct Controls of meal size, the decerebrate rat does not respond to any Indirect Control that has been tested. These include experience, e.g. learning and remembering taste aversions (Grill and Norgren, 1978b), and metabolism (Seeley et al., 1994). Because the decerebrate rat will only eat when food is placed in its mouth, it also does not respond to social or ecological Indirect Controls. Therefore the forebrain and its reciprocal connections with the caudal brainstem are necessary for these Indirect Controls and presumably all of the others not yet tested.

Forebrain - brainstem connections and indirect controls The necessity of the forebrain and its connections with the caudal brainstem for the Indirect Controls could be based on a variety of functional relationships. For example, the forebrain receives afferents from first and second order neurons in the caudal brainstem that, in turn, receive afferent projections from preabsorptive gut receptors activated by food stimuli. The effect of Indirect Controls in the forebrain could involve the processing of this afferent input from the caudal brainstem to change the potencies of the positive and negative feedbacks from the gut. This change in potency would then be communicated by descending connections to the comparator function of the caudal brainstem. Another possibility is that the forebrain is the site of action of a hormonal mechanism of a rhythmic (estrogen) or lipostatic (leptin) control and its projections to the hindbrain carry the neural effects of the hormone to the sites of peripheral afferent feedbacks, the comparator mechanism, or the premotor components of the network for ingestion.

1I1

Although at the present time we do not know the details of these or other possible functional relationships, I propose that the reciprocal connections between the forebrain and the hindbrain are the anatomical basis for the modulation of the potency of peripheral feedbacks by all Indirect Controls of meal size.

Forebrain - hindbrain connections in normal rat There is extensive experimental evidence for the importance of the reciprocal connections between forebrain and hindbrain in the controls of eating in the neurologically intact rat. I shall review five examples.

Sham feeding of sucrose and orosensory positive feedback The rate of licking and volume ingested during the sham feeding of sucrose is a monotonic function of sucrose concentration (Weingarten and Watson, 1982; Smith, 1995). The strong correlation between sucrose concentration and reward means that sham feeding is a form of oral self-stimulation. Because the orosensory effects (gustatory, thermal, and tactile) of a sucrose solution stimulate and maintain eating, sucrose is an adequate stimulus for a Direct Control of eating, specifically an orosensory positive feedback (Table 2). The anatomy of the gustatory system in the rat is relevant to the functions of positive feedback and

reward (Norgren, 1995). Following transduction by sweet receptors, the peripheral gustatory information is carried to the brain over afferent fibers of cranial nerves 7, 9, and 10. Tactile and thermal information is carried by some of these afferent fibers as well as those of the trigemimal nerve. The gustatory afferents project to the rostra1 nucleus tractus solitarius (RNTS), second order neurons of the RNTS project to the medial and waist region of the parabrachial nucleus (PBN) in the rat. Third order PBN neurons project up to the diencephalon where they bifurcate into two major divisions. One is thalamocortical and the other is to the ventral forebrain (Norgren, 1995). The thalamocortical pathway is formed by bilateral projections from the PBN to the medial part of the ventral posteromedial parvocellular thalamic nucleus (VPpc). Fibers from the VPpc pass rostrolaterally through the zona incerta and ascend in the internal capsule through the striatum to the external capsule to terminate in the gustatory cortex that is on both sides of the middle cerebral artery just dorsal to the rhinal fissure. This is a classical thalamocortical sensory system that functions for the discrimination and intensity of the sucrose stimulus. It is noteworthy that neurons responsive to taste, temperature, and tactile stimulation in the mouth are significantly segregated at thalamic and cortical levels (Norgren, 1995). The projections from the PBN to the ventral forebrain are numerous and widespread. Fibers

TABLE 2 Molecular mechanisms of direct controls of meal size ~

~~

DIRECT CONTROLS

PERIPHERAL

CENTRAL

Orosensory

Gustatory and olfactory transducers

Gastric

CCK* at CCK, vagal mechanoreceptors, other mechanoreceptors,and bombesin-lie peptides CCK* at CCK, receptors on vagal mechano- and chemoreceptors; Glucagon*, Amylin, Enterostatin, Apolipoprotein IV, and Insulin released by contact with mucosal receptors or by the release of incretins by nutrient or digestive stimuli

Dopamine* Opioids* Amino acids from gastric vagal afferent terminals in NTS Serotonin* Amino acids from duodenal and hepatic vagal afferent terminals in NTS

Small Intestinal

Serotonin*

Note: *indicates that a molecule has been demonstrated to be a physiological mechanism. The physiological status of the other molecules is uncertain.

178

terminate in the central nucleus of the amygdala, the bed nucleus of the stria terminalis, the lateral hypothalamus, the lateral preoptic area, substantia innominata, and part of the diagonal band of Broca. It is not clear whether all of these terminal fields are involved in processing gustatory information. The working hypothesis is that the processing in the ventral forebrain is necessary for the motivating, reinforcing, rewarding, and pleasurable effects of sucrose in the intact rat (see Conscious Experiences of Eating for discussion of pleasure). Dopaminergic synapses are critical for the forebrain processing of the afferent information from the hindbrain produced by sham feeding sucrose. Peripheral administration of D, and D, antagonists produce a dose-dependent inhibition of sham feeding. This inhibition is due to the antagonism of the reinforcing and incentive functions of the central dopamine released by the orosensory stimulation by sucrose and not to sensory, motor, or aversive functions of the drugs (Smith, 1995). The accumbens is one site of this action because microinjections of the antagonists into the accumbens produce the same effect as peripheral injections. These results with antagonists suggest that the sham feeding of sucrose increases dopaminergic activity in the forebrain. This has been confirmed (Smith, 1995). In addition, the real feeding of highcarbohydrate solid diets increases DA release in microdialysates from the accumbens (Hernandez and Hoebel, 1988), anteromedial hypothalamus (Orosco and Nicolddis, 1992), and amygdala (Lenard et al., 1992). Furthermore, the quantity released is correlated with the palatability of the diet (Martel and Fantino, 1996). This necessity of forebrain DA for the normal orosensory positive-feedback effect of sucrose on eating is unconditioned because it can be demonstrated in the preweanling rat as early as postnatal day 7 (Tyrka and Smith, 1991;v r k a et al., 1992) as well as in the adult rat without prior ingestive experience with sucrose solutions. The fact that DA antagonists do not inhibit sucrose intake in preweanling or adult rats when the sucrose solution is infused intraorally in the same manner as is required for intake of sucrose in chronic decerebrate rats (Smith, 1995) supports the

argument I am making. The decerebrate rat has sufficient neural complexity to discriminate differences in sucrose concentration of orally infused solutions and use them to maintain differential rates of licking, but without forebrain mechanisms, including dopaminergic systems, this information cannot be processed into the motivating effects necessary to reinitiate contact with sucrose in the environment during a meal (Smith, 1995). The effect of forebrain lesions on the orosensory positive-feedback effect of sucrose during sham feeding is further evidence for forebrain-hindbrain interactions. Bilateral ibotenic acid lesions of the basolateral nuclei of the amygdala reduced intake of weak solutions of sucrose, but not of more concentrated solutions (Siege1 et al., 1988). Bilateral lesions of the ventromedial hypothalamus and of the stria terminalis, however, increased intake of concentrated solutions more than dilute solutions (Black and Weingarten, 1988). The role of DA in the rewarding effect of food is not exclusive. There is considerable pharmacological evidence for central opioids antagonized by naloxone being necessary for normal sham feeding, but the central site and specificity of the effect of naloxone has not been analyzed as completely as the effect of DA antagonists (Smith, 1995).

Fat and postingestive negative feedback Duodenal infusion of fat emulsions inhibits sham feeding (Greenberg and Smith, 1996). The inhibition is proportional to the load (concentration x volume) of fat infused. Intraluminal digestion of fat is necessary for the effect, at least in the rat, because infusion of fatty acids, but not triglycerides, decreases intake during sham feeding. The inhibitory potency of fatty acids is a function of their chain length and degree of saturation. Abdominal vagal afferent fibers are necessary for this negative-feedback effect of fatty acids. Fatty acids can stimulate vagal afferent terminals directly (Melone, 1986) and indirectly through the release of cholecystokinin (CCK) from mucosal cells in the upper small intestine. The release of CCK stimulates vagal afferent fibers through CCK, receptors (Schwartz et al., 1994). Both direct and indirect stimulation of vagal afferent fibers occurs because blockade of the CCK-dependent mechanism

179

reduces, but does not abolish, the inhibitory effect of fat infusions (Greenberg et al., 1989; Yox et al., 1992). Vagal afferent fibers stimulated by CCK project to the caudal part of the NTS (CNTS) and recent evidence suggests that an excitatory amino acid, perhaps glutamine (Burns et al., 1998), mediates their synaptic action. The subsequent projection of the vagal information follows a similar pattern to that described for the orosensory information produced by sucrose stimulation. Neurons of the CNTS project to lateral and external medial PBN (Saper, 1995). These PBN neurons project to the lateral part of the VPpc; most of this projection is contralateral. From here fibers go to the granular insular cortex caudal to the gustatory cortex. These regions of the PBN also project to many ventral forebrain sites, including those structures that receive gustatory information. Forebrain sites of vagal afferent stimulation concerned with ingestion have been identified with the c-fos technique. Vagal stimulation produced by peripherally-administered CCK results in increased c-fos (specifically c-fos-like immunoreactivity) in the supraoptic nucleus (SON), paraventricular nucleus (PVN), and in the central nucleus of the amygdala (Day et al., 1994). Activation of the parvocellular region of the PVN is particularly interesting because this is an important site for integration of forebrain information for the control of eating (Weingarten et al., 1985) and lesion of the PVN blocks the inhibitory effect of peripheral CCK (Crawley and Kiss, 1985). This demonstrates the necessity of forebrain sites for determining the potency of CCK, a mechanism of a postingestive, inhibitory Direct Control, the load of fat in the small intestine. In contrast to the lack of effect of DA antagonists, peripheral administration of CCK inhibits intake of intraorally-infused sucrose in intact and chronic decerebrate rats (Grill and Smith, 1988). Because only one dose was tested in this experiment, it is not known if the inhibitory potency of CCK on intraoral infusions in intact and chronic decerebrate rats is equivalent.

Estrogen and indirect control by ovarian rhythm The increase of estrogen during the night of proestrus in the rat produces decreased food intake

during the day of estrus about 24 hours later. This decreased food intake is due to decreased meal size without a change in the number of meals (Geary and Eckel, 1997). Thus, estrogen acts as a rhythmic, inhibitory Indirect Control of meal size. Although the anteromedial preoptic area is the critical forebrain site of action for the other phasic behavioral effects of estrogen that occur during estrus, such as increased activity and increased sexual behaviors, the locus of the inhibitory action of estrogen on meal size is less certain. Recent evidence suggests that the PVN (Butera et al., 1996) and the anteromedial preoptic area (Dagnault and Richard, 1997) are involved, but the problem requires further investigation. The recent analysis of estrogen’s effect on meal size is a clear example of the relationship between Direct and Indirect Controls. The first important observations were that exogenous estrogen did not decrease sham feeding of sucrose in ovariectomized rats, but it decreased real feeding of sucrose under the same experimental conditions (Geary et al., 1995). This demonstrated that estrogen enhances postingestive, negative-feedback Direct Controls during real feeding without changing the orosensory positive feedback of sucrose during sham feeding. Because CCK is a mechanism of postingestive negative feedback from the small intestine (see above), its interaction withvstrogen was investigated. The satiating potency of exogenous CCK-8 was reduced in ovariectomized rats (Geary et al., 1994). Systemic replacement of estrogen increased the potency of CCK-8 (Geary et al., 1994) and local administration of estrogen to the PVN reproduced this effect (Butera et al., 1996). These results with exogenous estrogen and exogenous CCK-8 suggested that the inhibitory effect of endogenous estrogen on meal size was mediated, at least in part, by modulating the potency of endogenous CCK. This hypothesis was tested by giving devazepide, a CCK, antagonist, on different days of the estrus cycle. Devazepide blocked the decrease of food intake on the day of estrus, but had no effect on intake during diestrus (Eckel and Geary, 1999). This result is consistent with the hypothesis and illustrates the modulation of the potency of a mechanism of a postingestive Direct Control by a

180

rhythmic Indirect Control; and at least in the case of the enhanced potency of CCK-8 by local administration of estrogen in the PVN (Butera et al., 1996), the modulation must depend on connections between the forebrain and hindbrain.

Leptin, a lipostatic indirect control Leptin, a hormone secreted by adipocytes, produces inhibition of food intake and increased energy expenditure by acting on the hypothalamus (Campfield et al., 1996). Because leptin secretion is not stimulated by perabsorptive food stimuli during a meal and its duration of inhibitory action on food intake extends over several meals, it is a mechanism for the Indirect Control of meal size produced by the peripheral fat mass originally proposed by Kennedy (Kennedy, 1953). Decreased food intake after central or peripheral administration of leptin measured over hours or days, however, does not tell us whether leptin decreased the number of meals or decreased the size of meals. Three recent experiments answered this question by showing that central or peripheral administration of leptin decreased meal size without changing meal number in male and female rats (Eckel et al., 1998; Flynn et al., 1998; Kahler et al., 1998). The discovery that leptin inhibits meal size demonstrates that leptin is a hormonal mechanism of the Indirect Control of meal size produced by fat mass. It also opens up this effect of leptin to the behavioral and physiological techniques that are available to determine whether the decrease in meal size is produced by a decrease of orosensory positive feedback or an increase of postingestive negative feedback. Because the mechanisms of positive and negative feedbacks are different (Table 2), identification of the feedback potency that is changed by leptin will facilitate the analysis of the mechanisms of Direct Controls that leptin modulates. Note that the results of Eckel et al. (Eckel et al., 1998) appear to exclude an interaction between leptin and estrogen in the control of meal size in female rats. Leptin also fulfills the neurological criterion of an Indirect Control. Its initial site of action is in the ventrobasal hypothalamus (Campfield et al., 1996)

and there is evidence for transmission of its effect to the hindbrain through the action of endogenous agonists of the melanocortin receptors, especially of the MC3 and MC4 subtypes (Seeley et al., 1997; Grill et al., 1998). Recent anatomical studies have provided apparent connections between the site of leptin action on arcuate neurons containing neuropeptide Y and melanocortin neurons in the lateral hypothalamus that project to the hindbrain and many other regions of the brain (Broberger et al., 1998; Elias et al., 1998).

Conditioned taste aversion and learned indirect control Conditioned taste aversion is a type of learning in which the hedonic response to a gustatory stimulus is changed as a result of association with a toxic visceral stimulus (Grill, 1985). It is an unusual kind of conditioning because it can occur when the interval between the gustatory stimulus and the toxic stimulus is hours, it can occur in one trial, and it is resistant to extinction. A common procedure for producing a CTA is to pair the ingestion of sucrose with toxic doses of lithium chloride. Before conditioning, sucrose is avidly consumed and the rat emits ingestive behaviors. After conditioning, the rat rejects sucrose and emits aversive behaviors characteristic of oral stimulation by quinine, such as oral gapes, forelimb flailing, and chin rubbing. Thus, the formation of the CTA reverses the hedonic response to sucrose from preference to aversion; the rat responds to gustatory stimulation by sweet sucrose as if it were bitter quinine. There is evidence that the hedonic change is due to changes in the central processing of sucrose rather than to its peripheral transduction. Although altered central processing can be detected electrophysiologically in the RNTS (Giza et al., 1996) and PBN lesions prevent the acquisition of a CTA (Grigson et al., 1998), the neural networks in the caudal brainstem are not sufficient to mediate a CTA because the chronic decerebrate rat can neither acquire a CTA nor express a CTA acquired prior to decerebration (Grill and Norgren, 1978b). This demonstrates that the learned control of sucrose intake produced by the formation of a CTA is dependent on interactions between the forebrain and hindbrain.

181

concentrate on three conscious effects that ingested food has on an individual person in most, if not all circumstances. My criterion of a conscious effect is that it is accessible to self-report that can be validated by objective techniques. The three conscious experiences are pleasure, fullness, and tranquilization. They occur in that sequence during a meal; pleasure dominates the early period of eating, fullness and tranquilization

The c-Fos technique has revealed some of the details of the forebrain and hindbrain interactions. One hour after oral stimulation with sucrose or saccharin, rats that have acquired a CTA express cFos in the intermediate NTS (INTS) (Swank and Bernstein, 1994; Houpt et al., 1994). In fact, the expression of c-Fos in the INTS is the best neural correlate of a CTA currently available (Houpt et al., 1996). Its relevance to forebrain-hindbrain interactions is that its expression is abolished by lesions of the amygdala (Schafe and Bernstein, 1996) or insular cortex (Schafe and Bernstein, 1998).

TABLE 3

Summary

Candidate central mechanisms of direct and indirect controls of meal size

These five examples demonstrate that the modulation of the Direct Controls of eating by the Indirect Controls depend upon the reciprocal neural connections between the forebrain and caudal brainstem. That such connections are important for the integration of the diverse internal and external stimuli for the control of eating is not a new idea. What is new is that the forebrain is necessary for all Indirect Controls, and that Indirect Controls of eating exert their effects by modulating the potency of the positive and negative feedbacks of the Direct Controls that project into the caudal brainstem. Having demonstrated that the reciprocal connections between the forebrain and hindbrain are the basis for the functional relationship between Indirect and Direct Controls, the experimental task is to identify the specific connections and the synaptic mechanisms necessary for the neural integration underlying the Direct and Indirect Controls of eating and meal size in a variety of conditions. The current candidate transmitters and neuromodulators are listed in Tables 2 and 3.

Conscious experiences of eating The conscious experiences of eating in humans are numerous and complex. They depend on the food ingested, the person's experience with the food, the person's health, the social and physical circumstances in which the food is eaten, and the cultural meaning of the food, especially its relevance to social status, ethnic history, and religious significance. For the purposes of this chapter, I shall

Increase Intake

Decrease Intake

Norepinephrine Dopamine" NeuropeptideY Galanin Growth hormone-releasing hormone Orexin A and B (hypocretins 1 and 2) Dynorphin

Serotoninh Dopamine Norepinephrine Estrogen' Leptin"

Beta-endorphin Agouti-related protein

Insulin' Corticotropin-releasing hormone Bombesin-like peptides Cholecystokinin Glucagon-likepeptide-1 Alpha-Melanocyte stimulating hormone Melanocortin Agonists

Note: Although these amines and peptides have been shown to increase or decrease meal size under some conditions, their relative physiological importance and their mediation of Direct and Indirect Controls of meal size is uncertain except in the following cases: ' Dopamine mediates orosensory positive feedback of sweet taste and other foods. Central serotonin mediates the postingestive negative-feedback effect of small intestinal CCK (Poeschla et al., 1993). It is probably involved in numerous Direct and Indirect inhibitory Controls. It is also possible that peripheral serotonin mediates inhibition of intake (Simansky, 1996). ' Estrogen mediates the ovarian rhythmic Indirect Control by increasing the satiating potency of postingestive negative-feedback mechanisms, especially CCK (Geary, et al., 1994). d.e Centrally administered leptin and insulin mediate the inhibitory Indirect Control of meal size exerted by adipose mass (Woods, 1995; Campfield et al., 1996). At least part of insulin's effect is produced by increasing the satiating potency of CCK (Reidy et al., 1995).Leptin's action depends on the melanocortin system (Seeley et al., 1997); its synergism with CCK is uncertain. Note that dopamine and norepinephrine increase or decrease intake depending on the site of injection, phase of the diurnal rhythm, etc. This underlines the fact that the meaning of any central molecule is determined by the function it has in the neural network it is part of.

182

come later. This sequence suggests that food in the mouth elicits the pleasure of eating and that postingestive food stimuli act preabsorptively in the stomach and small intestine to elicit fullness and tranquilization.

Pleasure The ability of orosensory food stimuli to elicit pleasure is well documented in the extensive literature concerned with the psychophysics of gustatory and olfactory stimuli (Frijters, 1987). For example, when a person samples a series of sucrose solutions of different concentrations and is not permitted to swallow the solutions (‘sip and spit’ technique), she or he discriminates these solutions along two dimensions. One is intensity of sweetness and the other is intensity of pleasure. Pleasure is inferred from positive answers to questions, such as “How much would you like to drink this solution?’ or “How good does it taste?’ Note that these discriminations between intensity of sweetness and intensity of pleasure can be made by preschool children as well as by older people, require little or no training, and are reliable across different measures. Although the reports of intensity and pleasure tend to be monotonic and highly correlated for the dilute range of sucrose concentrations, at higher concentrations the ratings of pleasure often plateau or even decrease while the ratings of intensity continue to increase monotonically. Central dopaminergic and opioid mechanisms are involved in the rewarding effects of orosensory stimuli in rodents (see above), but it is not clear if these systems mediate the conscious pleasure of orosensory stimuli in humans. The problem deserves experimental attention and it can be approached through psychopharmacological and neuropsychological techniques in the context of brain imaging.

Fullness The intensity of fullness increases during a meal and ratings of fullness are inversely related to hunger in normal people, but not in some patients with eating disorders (Kissileff et al., 1996).

Distention of the stomach is sufficient to elicit reports of fullness. This is consistent with the evidence that the volume of ingested food, not its chemical characteristics, is the major stimulus for gastric negative feedback on eating (Smith, 1998a). The peripheral mechanism for the effects of gastric distention on eating and conscious experience is probably stimulation of gastric mechanoreceptors that activate vagal afferent fibers. Gastric vagal afferent fibers are not the only mechanism, however, because postprandial fullness has been reported after abdominal vagotomy. Although gastric distention can elicit fullness, it is likely that reports of fullness during a meal are elicited by the synergistic action of oral, gastric, and small intestinal mechanisms (Cecil et al., 1998). Some of this synergism is peripheral and is the result of slowing of gastric emptying by neural and endocrine mechanisms of the small intestine, stimulated by the digestive products of ingested food. But some of the synergism is probably central because fullness can be increased by infusions of fat into the small intestine that do not change gastric emptying (Welch et al., 1988). Cholecystokinin is one of the molecular mechanisms that mediates the negative-feedback effect of food in the small intestine on eating (see above). When it decreases intake, CCK does not change self-reports of fullness or a variety of other experiences produced by the ingestion of food (Geary et al., 1992) (Fig. 1). Thus, the vagal afferent activation produced by exogenous CCK is processed by the brain as if it represented vagal afferent activation by ingested food in the small intestine. It is remarkable that an infusion of CCK mimics the vagal activation by food so closely that the conscious experience of these events do not differ.

Wanquilization To watch a baby become drowsy and then sleep at the end of a feeding is to be convinced of the tranquilizing effect of ingested food. Tranquilization at the end of eating can also be observed in numerous animal species. For example, when rats eat a familiar diet alone during the light phase, the occurrence of resting or slow-wave sleep (Man-

183

URGE TO EAT HUNGRY STOM. FULL SATIATED RESTLESS SLEEPY ALERT RELAXED DIZZY COLD/WARM STOM. ACHE STOM. SICK

-p p

STOM. TENSE l

1

.

I

2

.

I

3

.

I

4

.

1

5

Fig. 1. Intravenous infusion of cholecystokinin (2 ngkglmin) in eight lean men reduced test meal size significantly without changing psychophysical ratings made from 15cm visual analogue scales compared to the same ratings made after saline infusions. Black bars are ratings after cholecystokinin; stippled bars are ratings after saline (from Geary et al., 1992 with permission of the publisher).

sbach and Lorenz, 1983) at the end of eating is so constant that it can be used as the terminal item of a behavioral sequence characteristic of satiety that marks the end of the meal (Antin et al., 1975). The tranquilization in adults has recently been demonstrated to require postingestive effects of eating because sham feeding did not produce it (Harnish et al., 1998).

Summary and perspective Eating consists of rhythmic oral movements that are organized by a cpg in the caudal brainstem. This means that the duration of eating and the size of a meal are determined by the sensory control of the cpg. A major source of the sensory control comes from peripheral feedbacks produced by ingested and digested food stimuli acting on preabsorptive chemical and mechanical receptors distributed in the mucosa of the mouth, stomach, and small intestine. Positive feedback from the mouth stimulates the cpg; negative feedback from the stomach and small intestine inhibits the cpg. Taken together they constitute the Direct Controls of meal size. A meal lasts as long as the positive

feedback is judged to exceed the negative feedback by an unidentified central comparator mechanism. The relative potency of the positive and negative feedbacks, however, is not only determined by the direct stimulation of preabsorptive receptors, but also by a large number of metabolic, neuroendocrine, and environmental stimuli. These diverse stimuli require the forebrain and its reciprocal connections with the hindbrain to produce their effects on meal size by modulating the potency of the positive and negative feedbacks that project into the caudal brainstem. They constitute the Indirect Controls of meal size. This account of the controls of eating considers the investigation of eating as a problem in Behavioral Neuroscience. This is a significant shift from the traditional view that eating is driven by metabolic deficits and is only concerned with energy balance. This new view is more comprehensive because it not only encompasses the traditional view, but also addresses psychological, social, clinical, and behavioral issues that the traditional view did not. The new view is also more neurologically coherent because it identifies the reciprocal connections between the forebrain and the caudal brainstem as the neural basis of the functional relationships between the Direct and Indirect Controls. Finally and most important, this new view provides a neurological framework in which to search for the meaning of the numerous molecules that affect food intake under experimental conditions. That search is necessary in order to move beyond preliminary wiring diagrams and potential cellular mechanisms of candidate molecules to a comprehensive physiology of the role of specific molecules in the functional networks that control meal size, integrate eating into the behavioral stream, and produce the conscious experiences of eating.

Acknowledgements I thank Ms. Laurel Torres for processing this manuscript and Dr. Norcross Geary, Dr. James Gibbs, and Dr. Ralph Norgren for constructive criticism of it. I am supported by research grants MH00149 and MH15455 from the National Institute of Health.

184

References Antin, J., Gibbs, J., Holt, J., Young, R.C. and Smith, G.P. (1975) Cholecystokinin elicits the complete behavioral sequence of satiety in rats. J. Comp. Physiol. Psychol., 89: 784-790. Black, R.M. and Weingarten, H.P. (1988) A comparison of taste reactivity changes induced by ventromedial hypothalamic lesions and stria terminalis transections. Physiol. Behav., 44: 699-708. Broberger, C., De Lecea, L., Sutcliffe, J.G. and Hokfelt, T. (1998) Hypocretin/orexin- and melanin-concentrating hormone-expressing cells form distinct populations in the rodent lateral hypothalamus: relationship to the neuropeptide Y and agouti gene-related protein systems. J. Comp. Neurol., 402: 460-474. Bums, G.A., Fleischmann, L.G. and Ritter, R.C. (1998) MK801 interferes with nutrient-related signals for satiation. Appetite, 3 0 1-12. Butera, P.C., Xiong, M., Davis, R.J. and Platania, S.P. (1996) Central implants of dilute estradiol enhance the satiety effect of CCK-8. Behav. Neumsci., 110: 823-830. Campfield, L.A., Smith, F.J. and Bum, P. (1996) The OB protein (leptin) pathway - a link between adipose tissue mass and central neural networks. Horn. Metab. Res., 28: 619-632. Campfield, L.A. and Smith, FJ. (1990) Systemic factors in the control of food intake. In: E.M. Stricker (Ed.), Handbook of Behavioral Neurobiology, Vol. 10, Neurobiology of Food and Fluid Intake, Plenum Press, New York, pp. 183-206. Cecil, J.E., Francis, J. and Read, N.W. (1998) Relative contributions of intestinal, gastric, oro-sensory influences and information to changes in appetite induced by the same liquid meal. Appetite, 3 1: 377-390. Crawley, J.N. and Kiss, J.Z.(1985) Paraventricular nucleus lesions abolish the inhibition of feeding induced by systemic cholecystokinin. Peptides, 6: 927-935. Dagnault, A. and Richard, D. (1997) Involvement of the medial preoptic area in the anorectic action of estrogens. Am. J. Physiol., 272: R3114317. Davis, J.D. (1998) A model for the control of ingestion - 20 years later. In: S.J. Fluharty and A.R. Morrison (Ed.), Progress in Psychobiology and Physiological Psychology, Vol. 17, pp. 127-173. Davis, J.D. and Smith, G.P. (1992) Analysis of the microstructure of the rhythmic tongue movements of rats ingesting maltose and sucrose solutions. Behav. Neurosci., 106: 217-228. Day, H.E., McKnight, A.T., Poat, J.A. and Hughes, J. (1994) Evidence that cholecystokinin induces immediate early gene expression in the brainstem, hypothalamus and amygdala of the rat by a CCKA receptor mechanism. Neurophamcology, 33: 719-727. Eckel, L.A. and Geary, N. (1999) Endogenous cholecystokii n ’ s satiating action increases during estrus in female rats. Peptides (in press). Eckel, L.A., Langhans, W., Kahler, A., Campfield, L.A., Smith, F.J. and Geary, N. (1998) Chronic administration of OB

protein decreases food intake by selectively reducing meal size in female rats. Am. J. Physiol., 275: R186-R193. Elias, C.F., Saper, C.B., Maratos-Flier, E., Tritos, N.A., Lee, C., Kelly, J., Tatro, J.B., Hoffman, G.E., Ollmann, M.M., Barsh, G.S., Sakurai, T., Yanagisawa, M. and Elmquist, J.K. (1998) Chemically defined projections linking the mediobasal hypothalamus and the lateral hypothalamic area. J. Comp. Neurol., 402: 442-459. Fay, R.A. and Norgren, R. (1997) Identification of rat brainstem multisynaptic connections to the oral motor nuclei using pseudorabies virus. 111. Lingual muscle motor systems. Brain Res. Rev., 25: 291-311. Flynn, M.C., Scott, T.R., Pritchard, T.C. and Plata-Salaman, C.R. (1998) Mode of action of OB protein (leptin) on feeding. Am. J. Physiol., 275: R174-179. Frijters, J.E.R. (1987) Sensory sweetness perception, its pleasantness, and attitudes to sweet foods. In: J. Dobbing (Ed.), Sweetness, Springer-Verlag, New York, pp. 67-78. Geary, N., Kissileff, H.R., Pi-Sunyer, EX. and Hinton, V. (1992) Individual, but not simultaneous, glucagon and cholecystokinin infusions inhibit feeding in men. Am. J. Physiol., 262: R975-R980. Geary, N., Trace, D., McEwen, B. and Smith, G.P. (1994) Cyclic estradiol replacement increases the satiety effect of CCK-8 in ovariectomized rats. Physiol. Behav., 56: 281-289. Geary, N., Trace, D. and Smith, G.P. (1995) Estradiol interacts with gastric or postgastric food stimuli to decrease sucrose ingestion in ovariectomized rats. Physiol. Behav., 57: 155-158. Geary, N. and Eckel, L.A. (1997) Decreased meal size and increased activity during estrus in rats. Int. Behav. Neumsci. Soc., 6: 23. (Abstract) Giza, B.K., McCaughey, S.A., Zhang, L. and Scott, T.R. (1996) Taste responses in the nucleus of the solitary tract in saccharin-preferring and saccharin-averse rats. Chem. Senses, 21: 147-157. Greenberg, D., Torres, N.I., Smith, G.P. and Gibbs, J. (1989) The satiating effects of fats is attenuated by the cholecystokinin antagonist lorglumide. Ann. NY Acud. Sci., 575: 5 17-520. Greenberg, D. (1998) Intestinal Satiety. In: G.P. Smith (Ed.), Satiation: From Gut to Brain, Oxford University Press, New York, pp. 40-70. Greenberg D. and Smith, G.P. (1996) The controls of fat intake. Psychosom. Med., 58: 559-569. Grigson, P.S., Reilly, S., Shimura, T. and Norgren, R. (1998) Ibotenic acid lesions of the parabrachial nucleus and conditioned taste aversion: further evidence for an associative deficit in rats. Behav. Neurosci., 112: 160-171. Grill, H.J. (1985) Introduction: Physiological mechanisms in conditioned taste aversions. Ann. NY Acad. Sci., 443: 67-88. Grill, H.J. and Kaplan, J.M. (1990) Caudal brainstem participates in the distributed neural control. In: E.M. Stricker (Ed.), Handbook of Behavioral Neurobiology, Vol. 10, Neurobiol-

185 ogy of Food and Fluid Intake, Plenum Press, New York, pp.

125-1 49. Grill, H.J. and Kaplan, J.M. (1992) Sham feeding in intact and chronic decerebrate rats. Am. J. Physiol., 262: R1070R1074. Grill, H.J. and Norgren, R. (1978a) Neurological tests and behavioral deficits in chronic thalamic and chronic decerebrate rats. Brain Res., 143: 299-312. Grill, H.J. and Norgren, R. (1978b) Chronic decerebrate rats demonstrate satiation, but not bait shyness. Science, 201: 267-269. Grill, H.J. and Smith, G.P. (1988) Cholecystokinin decreases sucrose intake in chronic decerebrate rats. Am. J. Physiol., 254: R8534856. Grill, H.J., Ginsberg,A.B., Seeley, R.J. and Kaplan, J.M. (1998) Brainstem application of melanocortin receptor ligands produces long-lasting effects on feeding and body weight. J. Neurosci., 18: 10128-10135. Harnish, M.J., Greenleaf, S.R. and Orr, W.C. (1998) A comparison of feeding to cephalic stimulation on postprandial sleepiness. Physiol. Behav., 64: 93-96. Hernandez, L. and Hoebel, B.G. (1988) Feeding and hypothalamic stimulation increase dopamine turnover in the accumbens. Physiol. Behav., 44: 599-606. Houpt, T.A., Philopena, J.M., Wessel, T.C., Joh, T.H. and Smith, G.P. (1994) Increased c-fos expression in nucleus of the solitary tract correlated with conditioned taste aversion to sucrose in rats. Neumsci. Lett., 172: 1-5. Houpt, T.A., Philopena, J.M., Joh, T.H. and Smith, G.P. (1996) c-fos induction in the rat nucleus of the solitary tract correlates with the retention and forgetting of a conditioned taste aversion. Learn. Mem., 3: 25-30. Kahler, A,, Geary, N., Eckel, L.A., Campfield, L.A., Smith, F.J. and Langhans, W. (1998) Chronic administration of OB protein decreases food intake by selectively reducing meal size in male rats. Am. J. Physiol., 275: R180-Rl85. Kennedy, G.C. (1953) The role of depot fat in the hypothalamic control of food intake in the rat. Proc. Roy. SOC.Lond. Sex B., 140: 578-592. Kissileff, H.R., Wentzlaff, T.H., Guss, J.L., Walsh, B.T., Devlin, J.J. and Thronton, J.C. (1996) A direct measure of satiety disturbance in patients with bulimia nervosa. Physiol. Behuv., 60: 1077-1085. Lenard, L., Hajnal, A., Czurko, A., Vida, I., Abraham, I. and Karadi, Z. (1992) Hunger and satiation releated changes of CA metabolites in the amygdala: an in vivo microdialysisHPLC study. Appetite, 19: 197. Mansbach, R.S. and Lorenz, D.N. (1983) Cholecystokinin (CCK-8) elicits prandial sleep in rats. Physiol. Behav., 30: 179-1 83. Martel, P. and Fantino, M. (1996) Influence of the amount of food ingested on mesolimbic dopaminergic system activity: a microdialysis study. Pharmucol. Biochem. Behav., 55: 297-302. Melone, J. (1986) Vagal receptors sensitive to lipids in the small intestine of the cat. J. Auton. Nerv. Syst., 17: 331-341.

Miller, S. and van der Meche, F.G. (1976) Coordinated stepping of all four limbs in the high spinal cat. Brain Res., 109: 395-398. Norgren, R. (1995) Gustatory system. In: G. Paxinos (Ed.), The Rat Nervous System, 2nd Edn, Academic Press, Inc, New York, pp. 75 1-77 1. Orosco, M. and Nicoldidis, S. (1992) Spontaneous food-related monoaminergic changes in the rostromedial hypothalamus revealed by microdialysis. Physiol. Behav., 52: 1015-1019. Pavlov, LP. (1910) The Work of the Digestive Glands, Charles Griffin & Company, Ltd., UK. Poeschla, B.D., Gibbs, J., Simansky, K.J., Greenberg, D. and Smith, G.P. (1993) Cholecystokinin-induced satiety depends upon activation of 5-HTlC receptors. Am. J. Physiol., 264: R62-R64. Riedy, C.A., Chavez, M., Figlewicz, D.P. and Woods, S.C. (1995) Central insulin enhances sensitivity to cholecystokinin. Physiol. Behav., 58: 755-760. Saper, C.B. (1995) Central Autonomic System. In: G. Paxinos (Ed.), The Rat Nervous System, Academic Press, New York, pp. 107-136. Schafe, G.E. and Bernstein, LL. (1996) Forebrain contribution to the induction of a brainstem correlate of conditioned taste aversion: I. The amygdala. Brain Res., 741: 109-1 16. Schafe, G.E. and Bemstein, I.L. (1998) Forebrain contribution to the induction of a brainstem correlate of condtioned taste aversion. 11. Insular (gustatory) cortex. Brain Res., 800: 4047. Schwartz, G.J., McHugh, P.R. and Moran, T.H. (1994) Pharmacological dissociation of responses to CCK and gastric loads in rat mechanosensitive vagal afferents. Am. J. Physiol., 267: R3034308. Seeley, R.J., Grill, H.J. and Kaplan, J.M. (1994) Neurological dissociation of gastrointestinal and metabolic contributions to meal size control. Behav. Neurosci., 108(2): 347-352. Seeley, R.J., Yagaloff, K.A., Fisher, S.L., Bum, P., Thiele, T.E., van Dijk. G, Baskin, D.G. and Schwartz, M.W. (1997) Melanocortin receptors in leptin effects [letter]. Nature, 390: 349. Siege], A., Joyner, K. and Smith, G.P. (1988) Effect of bilateral ibotenic acid lesions in the basolateral amygdala on the sham feeding response to sucrose in the rat. Physiol. Behav., 42: 231-235. Simansky, K.J. (1996) Serotonergic control of the organization of feeding and satiety. Behav. Brain Res., 73: 37-42. Smith, G.P. (1995) Dopamine and food reward. In: S. Fluharty, A.R. Morrison, J. Sprague, and E. Stellar (Eds), Progress in Psychobiology and Physiological Psychology, Vol. 16, Academic Press, New York, pp. 83-144. Smith, G.P. (1996) The direct and indirect controls of meal size. Neurosci. Biobehav. Rev., 20: 41-46. Smith, G.P. (1998a) Pregastric and Gastric Satiety. In: G.P. Smith (Ed.), Satiation: From Gut to Brain, Oxford University Press, New York, pp. 10-39. Smith, G.P. (1998b) Control of food intake. In: M.E. Shils, J.O. Olson, M. Shike, and A.C. Ross (Eds), Modem Nutrition in

186 Health and Disease, 9th Edn, Williams & Wilkins, Baltimore, pp. 631-644. Swank, M.W. and Bernstein, I.L. (1994) c-fos induction in response to a conditioned stimulus after single trial taste aversion learning. Brain Res., 636: 202-208. Travers, J.B., Dinardo, L.A. and Kariminamazi, H. (1997) Motor and premotor mechanisms of licking. Neurosci. Biobehav. Rev., 21: 631-647. Tyrka, A., Gayle, C. and Smith, G.P. (1992) Raclopride decreases sucrose intake of rat pups in independent ingestion tests. Pharmacol. Biochem. Behav., 43: 863-869. Tyrka, A. and Smith, G.P. (1991) Potency of SCH 23390 for decreasing sucrose intake in rat pups depends on mode of ingestion. Phamacol. Biochem. Behav., 39: 955-961, Weingarten, H.P., Chang, P. and McDonald, T.J. (1985) Comparison of the metabolic and behavioral disturbances

following paraventricular- and ventromedial-hypothalamic lesions. Brain Res. Bull., 14: 551-559. Weingarten, H.P. and Watson, S.D. (1982) Sham feeding as a procedure for assessing the influence of diet palatability on food intake. Physiol. Behav., 28: 401-407. Welch, I.M., Sepple, C.P. and Read, N.W. (1988) Comparisons of the effects on satiety and eating behavior of infusion of lipid into the different regions of the small intestine. Gut, 29: 306-311. Woods, S.C. (1995) Insulin and the brain: a mutual dependency. In: S. Fluharty, A.R. Momson, J. Sprague and E. Stellar (Eds). Progress in Psychobiology and Physiological Psychology, Vol. 16, Academic Press, New York, pp. 53-81. Yox, D.P., Brenner, L. and Ritter, R.C. (1992) CCK-receptor antagonists attenuate suppression of sham feeding by intestinal nutrients. Am. J. Physiol., 262: R554R561.

E.A. Mayer and C.B.Saper (Fils.) Progress in Brain Research, Vol 122 6 Zoo0 Elsevier Science BV. All rights reserved.

CHAPTER 13

Effects of nutrients on brain function Timothy J. Maher* Sawyer Professor of Pharmaceutical Sciences, Dean of Research and Sponsored Programs, Division of Pharmaceutical Sciences, Massachusetts College of Pharmacy and Health Sciences, 179 Longwood Avenue, Boston, MA 02115, USA

Introduction Nerve cells in the brain communicate with one another and with effector sites in the periphery using signals generated by chemical compounds termed neurotransmitters, neurohormones and neuromodulators. The synthesis, storage and release of these compounds must normally be subject to strict regulatory control mechanisms so as to maintain homeostasis in the face of constant challenges from the external environment. It is not surprising therefore to find that alterations in the levels of circulating neurotransmitter molecules in the periphery typically have little direct effect on the neurons contained within the central nervous system. However, all the precursor compounds that are eventually metabolized into these neurotransmitters derive from the diet. Thus, the possibility that diet can affect the rate of synthesis of neurotransmitters important in the control of many bodily functions requires investigation. In fact, studies have demonstrated the ability of various dietary manipulations including both macronutrient (e.g. carbohydrate, protein or fat) and micronutrient (e.g. vitamins, minerals and individual components of the diet) to influence the chemistry of the brain. Additionally, in other studies with animals and humans involving such dietary manipulations, certain behaviors believed to be associated with a particular neurotransmitter have been monitored *Corresponding author. Tel.: 6 17-732-2940; Fax: 617-732-2963; e-mail: [email protected]

and found to be changed. Therefore the potential of diet to influence brain chemistry and associated behaviors appears to be a fruitful avenue of investigation. The measurement of chemical changes in the brains of experimental animals following dietary interventions has demonstrated, in some instances, powerful influences on the levels of neurotransmitter compounds. However the ability to demonstrate functional changes associated with such chemical changes has been more of a challenge, especially in humans. Although these changes tend to be subtle in nature, they may be of significance during times of altered homeostasis, e.g. stress, disease, etc. And, while most studies have investigated the acute effects of such dietary manipulations, few studies have looked at the effects of long-term exposure on body-mind interactions. The present review will discuss the ability of specific amino acids found in the diet to influence the synthesis and release of their corresponding neurotransmitter compounds. A number of important functions of the brain are controlled by complex neuronal systems that utilize these chemicals and are messenger molecules. For instance, an individual’s response to psychological and physical stress, changes in blood pressure, anxiety-provoking situations, painful stimuli, etc. are influenced by the catecholamine neurotransmitter norepinephrine. As norepinephrine is synthesized directly from its dietary amino acid precursor tyrosine, the availability of this amino acid in the diet may influence the individual’s response to the above

188

environmental stressors. Other situations that may have an endogenous neurochemical component, e.g. mood, depression, hypertension, may be influenced by treatments that alter the synthesis and/or release of norepinephrine. Other neurotransmitter systems to be discussed include serotonin derived from tryptophan, which is involved with sleep, mood, temperature regulation, pain sensations and appetite; nitric oxide derived from arginine, which is involved with learning and memory; and glycine derived from threonine which is involved with the central control of motor function.

Nutrient entry into the brain: the blood-brain barrier In order for any of the dietary components to influence chemical neurotransmission in the central nervous system they must traverse the blood-brain barrier (BBB). The BBB functions to effectively separate the peripheral chemical compartment from the central chemical compartment. Characterized by adjacent capillary endothelial cells with tight junctions (zona occludens), a paucity of micropinocytotic transport vesicles, and astrocytic foot processes, the BBB selectively determines which compounds pass between the periphery and the brain. Qpically lipid-soluble (e.g. gaseous general anesthetics) and small molecular weight compounds (e.g. water, ions) are able to easily traverse the BBB. While most neurotransmitters, due to their water-soluble, polar characteristics, fail to pass the BBB, their precursor compounds, which are contained in the diet, are usually capable of passing the BBB via a facilitated diffusion process. A number of carrier-protein molecules have been identified that participate in the selective, competitive and saturable transport of amino acids, nucleic acids, amines and sugars across the BBB. The amino acids that are neurotransmitters themselves (e.g. glycine, glutamate, aspartate) typically have great difficulty passing the BBB. This is extremely important since consumption of protein that contains these amino acid neurotransmitters leads to elevations in their plasma concentrations, which if also were to lead to elevations in their levels in brain would most likely result in seizures,

neurotoxicity and possibly death following a standard meal. However, some of the amino acids that are precursors for their respective neurotransmitters in the brain (e.g. L-tryptophan : serotonin; L-tyrosine :dopamine and norepinephrine; L-arginine : nitric oxide; L-threonine :glycine) do cross the BBB with greater ease. In this way the diet provides the precursor amino acids required by the brain for the synthesis of important neurotransmitters. The synthesis of these neurotransmitters is thus potentially influenced by the availability of these precursor molecules provided by the diet. In order for an increased availability of a precursor substance to enhance neurotransmission in the brain, specific criteria must be met. First, if given orally as a bolus or in the diet, or if given parentally, as is usually the case in initial studies in experimental animals, the precursor molecule must reach the general circulation. The molecule must then be transported via the circulation to the neurons of interest. As many of the neurons of interest are located in the central nervous system, the BBB must be traversed in such cases. The precursor molecule must then be taken up into the neurons of interest and enzymatically converted into the neurotransmitter substance. This requires that the rate-limiting enzyme(s) be unsaturated with substrate at normal precursor concentrations. Once synthesized the neurotransmitter must be located in a releasable pool. Additionally, no negative feedback systems can be allowed to operate along the way that would blunt the ability of the enhanced precursor to accelerate synthesis and release.

Tryptophan, carbohydrates and serotonin Circulating levels of the precursor amino acids tryptophan and tyrosine are found to fluctuate depending on the composition of the diet (Glaeser et al., 1983; Maher et al., 1984). Ingestion of a high protein meal leads to increases in the plasma levels of the large neutral amino acids (LNAA) valine, leucine, isoleucine, phenylalanine, trytophan and tyrosine. Since the transport into the brain of these amino acids is competitive in nature and via facilitated diffusion at the BBB, the consumption of a high protein meal, which contains some trypto-

189

phan, but which also contains much more of the branched-chain amino acids (BCAA) valine, leucine and isoleucine, actually leads to a decrease in the flux of tryptophan into the brain. Studies indicate that to increase the flux of trytophan into the brain a carbohydrate-rich meal would need to be consumed (Glaeser et al., 1983). Carbohydrates elicit the release of insulin, which in addition to enhancing the uptake of glucose into peripheral sites, also enhances the uptake of the BCAA into muscle. The removal of a portion of the BCAA from the circulation results in a decrease in the competition tryptophan is exposed to for passage across the BBB. Thus, it is actually the ratio of the concentration of a particular amino acid in plasma to that of its competitors, the so-called ‘plasma ratio’, that determines the flux of that amino acid into brain. Tryptophan is an essential amino acid, i.e. it cannot be synthesized by the body and must be consumed in the diet, that is the precursor for the neurotransmitter serotonin. The synthesis of serotonin occurs in two steps: initial conversion to 5-hydroxy-tryptophan followed by conversion to 5-hydroxy-tryptamine,or serotonin. The rate-limiting enzyme, tryptophan hydroxylase is normally unsaturated at typical brain concentrations of tryptophan, and thus the administration of tryptophan to animals leads to increases in the synthesis of serotonin. Additionally, diurnal variations in the levels of tryptophan in the circulation due to consumption of foods produces predictable changes in brain serotonin (Fernstrom and Wurtman, 1971). Numerous experiments have evaluated the ability of endogenously administered tryptophan to alter serotonergic neurotransmission. Tryptophan has been shown to reduce pain in animals and humans, reduce food intake (especially carbohydrates), improve depression, and decrease sleep latency (Van Praag and Korf, 1974; Hartmann, 1977; King, 1980). Additionally, animals given a diet deficient in tryptophan (e.g. certain types of corn), that results in reductions in brain serotonin, can increase pain sensitivity, a response that is reversed by tryptophan administration (Lytle et al., 1975). Following the demonstration in experimental animals that the consumption of a meal high in

carbohydrate enhances the synthesis of serotonin in the brain, studies were performed in human volunteers in an attempt to further explore the functional significance of these findings. Initial studies by Spring (1986) demonstrated that when men consumed a high-carbohydrate lunch, a significant increase in feelings of fatigue were noted. The high-carbohydrate lunch did not have to be composed of simple sugars to trigger fatigue; a high-starch, protein-poor meal was similarly effective. Others have demonstrated carbohydratecraving by obese patients and suggested that this ultimately involves a serotonergic mechanism of food self-selection (Lieberman et al., 1986). A number of studies have demonstrated the utility of various serotoninergic agents (e.g. fluoxetine, dexfenfluramine) to ameliorate the mood and cognitive distrubances associated with premenstrual syndrome (PMS).Because of the relationship between macronutrient consumption in the diet and serotonin synthesis, the ability of a specially designed carbohydrate-rich beverage to ameliorate some of the dysphoric symptoms of PMS in women was investigated in a placebo-controlled, crossover design study (Sayegh et al., 1995). Three isocaloric beverages were designed that would provide for different degrees of insulin release following consumption. One of the beverages containing dextrose and maltodextrin did adequately stimulate insulin release to increase significantly the ratio of tryptophan to LNAA by 29%, as compared with a placebo drink, and thus presumably would have enhanced serotonin synthesis in the brain. Women consuming this beverage had significantly lower scores for tension, anger, depression and confusion as compared with placebo. Additionally, their performance on a test of cognition (Auditory Consonant Trigrams Recognition test) was significantly improved. llvo other drinks containing either protein or a carbohydrate with a much smaller glycemic index were also tested but both failed to alter the tryptophan to LNAA ratio. Neither of these drinks affected mood or cognition in the test subjects. While it is impossible to predict what proportion of those women suffering from PMS might realize symptomatic relief from such a dietary intervention, this approach appears to have associated with it essen-

190

tially no significant risk to the patient and thus should be tried in larger samples of women.

Qrosine and the catecholamines Unlike tryptophan, which is an essential amino acid, the catecholamine precursor amino acid tyrosine, besides being obtained from the diet, can also be derived slowly from the hepatic hydroxylation of phenylalanine. The catecholamine norepinephrine is extremely important in our response to stress, whether the stress is psychological or physiological in nature, or a combination of both. Alterations in catecholaminergic neuronal function has been shown to influence a number of the responses to and consequences of stress. While there is usually ample tyrosine to meet the biochemical and metabolic needs of an individual, during situations when catecholamine-containing neurons are made to fire rapidly, the availability of tyrosine may become limiting and compromise the rate of synthesis of dopamine, norepinephrine and/ or epinephrine. Under resting conditions the rate-limiting enzyme in the biosynthesis of the catecholamines, tyrosine hydroxylase, is normally saturated with its substrate tyrosine, and it is the availability of its cofactor tetrahydrobiopterin that determines the rate of catecholamine synthesis (Lovenberg and Victor, 1974). However, when such neurons are made to fire rapidly, tyrosine hydroxylase is believed to become phosphorylated, resulting in a conformational change of the enzyme, and thereby making the availability of tyrosine the limiting factor. While normally feedback inhibition of catecholamine synthesis keeps catecholaminergic neurotransmission carefully regulated, during periods of increased neuronal firing this feed-back inhibition appears to be inoperative. To explore this relationship between firing frequency of a group of neurons and their dependence on ample precursor supplies to support synthesis, a number of experiments have been performed which take advantage of the body's ability to selectively stimulate groups of neurons in response to pertubations of homeostasis. When a portion of the neurons in the brain are damaged leading to a decrease in the release of neurotransmitter, as is seen in certain disease states

or experimentally when neurotoxins are employed, the remaining neurons typically respond initially by increasing their firing frequency to maintain neurotransmitter homeostasis. This feed-back mechanism allows the organism to maintain neurotransmission at pre-perturbation levels. To demonstrate this point, the neurotoxin 6-hydroxydopamine, which selectively destroys dopaminergic neurons, was administered unilaterally in the striatum of rats to produce at least a 75% reduction in the number of nigrostriatal neurons (Melamed et al., 1980). The surviving neurons increased their firing rate as demonstrated by the increase in the major dopamine metabolite, homovanillic acid (HVA) on the lesioned side when compared with levels of HVA on the intact contralateral side. Administration of tyrosine via injection led to a further increase of HVA on the lesioned side, while failing to influence HVA levels on the intact, non1esioned side. Similar dopamine synthesis-enhancingresults could also be obtained by blocking postsynaptic dopamine receptors with haloperidol (Scally et al., 1977), or by depleting catecholamine stores with reserpine (Sved et al., 1979a). Some of the more important homeostatic systems in the body that control blood pressure utilize selectively neurons that contain catecholamines. For example, when blood pressure is elevated central neurons in the brain stem increase their firing frequency to produce sympathoinhibition in an attempt to restore blood pressure back towards normal. The neurons that the released norepinephrine interact with are the same group of neurons that a-methylDOPA's (Aldomet) metabolite (a-methylnorepinephrine) acts on as a therapeutically useful antihypertensive agent. On the other hand, when blood pressure is decreased sympathoadrenal neurons are activated which result in increased peripheral resistance, heart rate, and thus blood pressure. During periods of hypotension, the depressor brain stem neurons are quiescent so as not to interfere with the adrenal responses, and similarly during periods of hypertension the pressor adrenal neurons are not activated. Thus, these reciprocal systems utilize catecholamines selectively to both increase and decrease blood pressure. As expected, administration of tyrosine to sponta-

191

neously hypertensive rats ( S H R ) decreases blood pressure (Sved et al., 1979b), while tyrosine increases blood pressure in rats made hypotensive by blood removal (Conlay et al., 1981). As predicted, the ability of tyrosine to decrease blood pressure in hypertensive rats was associated with an increase in the major norepinephrine metabolite, 3-methoxy-4-hydroxy-phenylethylglycol(MHPG) sulfate in the brain stem (Sved et al., 1979b). Additionally, since the hypotensive action of tyrosine requires its entry into the brain, coadministration of the LNAA valine, which competes for passage across the BBB and entry into the brain, attenuated the response. However, in animals made hypotensive by hemorrhage while tyrosine dose-dependently increased blood pressure, the response was not influenced by other LNAA administration, since this involves a peripheral mechanism not influenced by the BBB. Feeding rats a diet supplemented with tyrosine (five times the normal amount found in protein) has also been demonstrated to prevent the dramatic drop in blood pressure following intravenous administration of the nitrovasodilator agent hydralazine, and to significantly attenuate the rapid fall in blood pressure during a controlled hemorrhage, thereby prolonging survival (Moya-Huff et al., 1989). Some sympathomimetic amines derive all or a portion of their activity via the release of stored catecholamines, with little or no direct interaction with adrenoceptors. One of the characteristics of such indirect-acting sympathomimetic drugs is the rapid development of tachyphylaxis, or a significant diminution in the observed response. When isolated perfused rat hearts are exposed to repeated doses of such a drug, e.g. tyramine, this tachyphylaxis is rapidly produced. However, when small amounts of tyrosine are included in the perfusion solution, a situation not typically employed with in vitro perfusion solutions for isolated tissues, the development of tachyphylaxis is completely prevented (Pinto and Maher, 1986). Similarly, the anorectic effects of several mixed-acting sympathomhetic amines, which produce a portion of their response via the release of stored catecholamines, e.g. d-amphetamine, 2-ephedrine, dl-norephedrine, are enhanced by tyrosine administration in hyperphagic rats (Hull and Maher,

1990). Qrosine itself had no anorectic effects in this model. The ability of opioids such as morphine and codeine to produce analgesia is also partly dependent upon the release of norepinephrine in the central pain-pathways. Administration of tyrosine significantly potentiated the centrally mediated analgesic effects of these opioids (Hull et al., 1994). Additionally their duration of action was significantly prolonged. The response was not mimicked by the unnatural enantiomer, D-tyrosine, and the response was attenuated by co-administration of the LNAA valine. The degree of potentiation correlated with the magnitude of the increment in brain tyrosine, and co-administration of the tyrosine hydroxylase inhibitor alpha-methylp-tyrosine prevented the potentiation. Opioids also have well known peripherally-mediated effects such as decreased gastric motility and altered body temperature. Neither of these responses were influenced by increased tyrosine availability since these responses are not catecholamine-dependent. These studies have now been extended to include dietary supplementation of tyrosine (five and 10 times the normal amount for five days) in which it was found that tyrosine was still effective at enhancing the potency and prolonging the duration of action of morphine in the tail-flick analgesia assay (Hollenbach and Maher, 1998). While it was possible that the inclusion of tyrosine in higher than normal amounts in the diet might have led to an activation of tyrosinase, an enzyme that would divert the metabolism of tyrosine away from catecholamine synthesis, these studies suggest that this did not occur at least to an extent that tyrosine availability was still appropriately enhanced when needed for supporting the actions of these opioids. Stress is known to activate catecholarninecontaining neurons in the locus coeruleus. Experiments performed in rats exposed to restraint and electric shock stress demonstrated the protective effects of tyrosine added to the diet (Lehnert et al., 1984). As stress is known to rapidly deplete stores of norepinephrine in this brain area, the supplementation of tyrosine in the diet was associated with a maintenance of norepinephrine levels and an increase in its major metabolite MHPG. Further studies involving stress have been

192

performed in human volunteers subjected to high altitude (4700 meters) and cold temperature (15OC) stress with the aid of an environmental chamber (Banderet and Lieberman, 1989). Tyrosine supplementation (100 mgkg, orally over a 40 minute period) significantly diminished the normal reactions to this hypobaric, hypoxic cold stress (e.g. headache, distress, muscular discomfort, fatigue, hostility, tension, feelings of cold, confusion) and prevented the cognitive deficits (e.g. poor performance on pattern recognition tests, vigilance) normally observed in subjects exposed to such severe stress. Qrosine may find utility in modifying function in a number of catecholaminedependent systems.

Arginine :nitric oxide Arginine is a basic amino acid that serves as the precursor of the free radical nitric oxide (NO) (Bredt and Snyder, 1992). Besides its long-time recognized role in the immune system, NO also plays a significant role in the peripheral regulation of blood pressure and in the central nervous system mediation of memory and learning. In the periphery NO appears to be identical to the previously described endothelium-derived relaxing factor (EDRF),which functions to relax many vascular beds via an activation of guanylyl cyclase (Palmer et al., 1987). Studies have demonstrated the utility of arginine to prevent or delay the development of hypertension in S H R and also to decrease blood pressure in animals with existing hypertension (Calvier et al., 1990). These effects of arginine are generally prevented if one of the NO synthase inhibitors (e.g. L-NAME, L-NOArg) is employed, suggesting that the effects of arginine occur via NO production. Arginine has a promising therapeutic role in modifying cardiovascular function. Within the central nervous system NO appears to function as a retrograde neuromodulator which is associated with long-term potentiation (LTP) and learning/memory formation (Bohme et al., 1991). The process of LTP involves an increase in the efficiency of neuronal neurotransmission associated with repeated traffic through a neuronal circuit. The administration of arginine (50-400 mgkg, i.p.) has been demonstrated to improve learning and mem-

ory in rats tested in a Moms water maze (Sato and Maher, 1995). The beneficial effects of arginine were dose-dependent and attenuated by the coadministration of NO synthase inhibitors. Interestingly, very high doses of arginine led to a decrement in performance in this behavioral task. This could have been the result of the formation of two other arginine metabolites, agmatine and the polyamine spermine, both of which are known to interfere with mechanisms associated with memory formation, or due to an overall toxic response to an exaggerated amino acid imbalance (Maher, 1994).

Threonine :glycine The inhibitory amino acid neurotransmitter glycine plays an important role in the control of motor function in the brain and spinal cord. Early attempts to increase central nervous system glycine concentrations by administering large doses of glycine peripherally failed since the transport of this amino acid across the BBB is very poor. Using the precursor approach, Maher and Wurtman (1980) demonstrated the ability of threonine administration to enhance central glycine concentrations in rats. Threonine was then tried in patients with spasticity and found to have beneficial effects (Barbeau et al. 1982; Lee et al., 1990; Growden et al., 1991). Currently, threonine has 'orphan drug' status with the Food and Drug Administration.

Summary and conclusions While many of the above examples support a role of these dietary components in modifying the synthesis, storage, release and actions of various neurotransmitter molecules in the central nervous system, most of the responses to eating everyday foods are expected to produce subtle changes in physiological and/or behavioral parameters. However, the observed subtle changes may have significant consequences when present in individuals with altered homeostasis as might be present in various disease states or certain environmental situations (e.g. depression, PMS, stress). Studies in the future should investigate the effects of various diets, e.g. vegetarian, macrobiotic, traditional Eastem, etc. on physiological and psychological

193

functioning. Care should be taken to differentiate between the responses of subgroups of subjects, e.g. male vs. female, old vs. young, and lean vs. obese, as some differences in the rate of neurotransmitter synthesis and receptor dynamics have been reported in some studies. Chronic consumption of these diets may lead to long-term alterations in the neurotransmitter systems’ dynamics, or as is often the situation with long-term pharmacological treatments, may result in adaptive changes to minimize the acute effects of such treatments. To date, no such studies have been performed that have systematically addressed many of these issues. Future studies will require careful design so as to enhance the chances of detecting such alterations in function. However, the most significant alterations in function occur when a dietary component is administered in a purified form, separate from the normal diet. In this case the compound should be treated more like a pharmacological agent than a nutrient since adverse (i.e. antinutritive) effects may result. The most difficult studies however will use everyday foods with the aim of detecting changes based on the underlying biochemical changes.

References Banderet, L.E. and Lieberman, H.R. (1989) Treatment with tyrosine, a neurotransmitter precursor, reduces environmental stress in humans. Brain Res. Bull., 22: 759-762. Barbeau, A., Roy, M. and Chouza, C. (1982) Pilot study of threonine supplementation in human spasticity. Can. J. Neurol. Sci., 9: 141-145. Bohme, G.A., Bon, C., Stutzmann, J.M., Doble, A. and Blanchard, A. (1991) Possible involvement of nitric oxide in long-term potentiation. Eur. J. Phanacol., 199: 379-381. Bredt, D.S. and Snyder, S.H. (1992) Nitric oxide, a novel neuronal messenger. Neuron, 8: 3-1 1. Calvier, A, Collier, J. and Vallance, P. (1990) L-Arginineinduced hypotension. Lancet, 336: 1016-1017. Conlay, L.A., Maher, T.J. and Wurtman, R.J. (1981) Tyrosine increases blood pressure in hypotensive rats. Science, 212: 559-560. Fernstrom, J.D. and Wurtman, R.J. (1971) Brain serotonin content: physiological regulation by plasma neutral amino acids. Science, 178: 414-416. Glaeser, B.S., Maher, T.J. and Wurtman, R.J. (1983) Changes in brain levels of acidic, basic, and neutral amino acids after consumption of single meals containing various proportions of protein. J. Neurochem., 41: 1016-1021.

Growdon, J.H., Nader, T.M., Schoenfield, J. and Wurtman, R.J. (1991) L-threonine in the treatment of spasticity. Clin. Neumphannacol., 14: 403412. Hartmann, E. (1977) L-tryptophan: a rational hypnotic with clinical potential. Am. J. Psychiatry, 134: 366-370. Hollenbach, E. and Maher, T.J. (1998) Effects of dietary Ltyrosine on opioid-induced analgesia utilizing the tail-flick test. Naunyn-Schmiedeberg’s Arch. Pharmacol., 358 (Suppl. 1): R49. Hull, K.M. and Maher, T.J. (1990) L-tyrosine potentiates the anorexia induced by mixed-acting sympathomimetic drugs in hyperphagic rats. J. Pharmacol. Exp. Ther., 255: 403409. Hull, K.M., Tolland, D.E. and Maher, T.J. (1994) L-tyrosine potentiation of opioid-induced analgesia in mice utilizing the hot plate test. J. Phanacol. Exp. Ther., 269: 1190-1 195. King, R.B. (1980) Pain and tryptophan. J. Neurosurg., 53: 4452. Lee, K.C., Patterson, V., Roberts, G. and Trimble, E. (1990) The antispastic effect of L-threonine. In: G. Lubec and G.A. Rosenthal (Eds), Amino Acids: Chemistry, Biology and Medicine, ESCOM Science Publishers, Leiden, pp. 658-663. Lehnert, H., Reinstein, D.K., Strowbridge, B.W. and Wurtman, R.J. (1984) Neurochemical and behavioral consequences of acute, uncontrollable stress: effects of dietary tyrosine. Brain Res., 303: 215-223. Lieberman, H. Wurtman, J. and Chew, B. (1986) Changes in mood after carbohydrate consumption among obese individuals. Am. J. Clin. Nutr., 44: 772-778. Lovenberg, W. and Victor, S.J. (1974) Regulation of tryptophan and tyrosine hydroxylase. Life Sci., 14: 2337-2353. Lytle, L.D., Messing, R.B., Fisher, L. and Phebus, L. (1975) Effects of long-term corn consumption on brain serotonin and the response to electric shock. Science, 190: 692-694. Maher, T.J. (1994) Safety concerns regarding supplemental amino acids: results of a study. In: B.M. Manion (Ed.), Food Components to Enhance Perjormance, National Academy Press, Washington, D.C., pp. 455-460. Maher, T.J., Glaeser, B.S. and Wurtman, R.J. (1984) Diurnal variations in plasma concentrations of basic and neutral amino acids and in red cell concentrations of aspartate and glutamate: effects of dietary protein. Am. J. Clin. Nut,:, 39: 722-729. Maher, T.J. and Wurtman, R.J. (1980) L-Threonine administration increases glycine concentrations in the rat central nervous system. Life Sci., 26: 1283-1286. Melamed, E., Hefti, F. and Wurtman, R.J. (1980) Tyrosine administration increases striatal dopamine release in rats with partial nigrostriatal lesions. Proc. Natl. Acad. Sci. USA., 77: 43054309. Moya-Huff, F.A., Pinto, J.M.B., Kiritsy, P.J. and Maher, T.J. (1989) Dietary supplementation of tyrosine prevents the rapid fall in blood pressure during hemorrhage. J. Neural Transm., 78: 159-165. Palmer, R.M.J., Ferrige, A.G. and Moncada, S. (1987) Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Narure, 327: 524-526.

194 Pinto, J.M.B. and Maher, T.J. (1986) Tyrosine availability prevents tyramine-induced tachyphylaxis in the isolated rat heart. Neumchem. Res., 9: 533-537. Sato, K.and Maher, T.J. (1995) Effects of L-arginine on spatial memory in rats: role of nitric oxide. SOC.Neumsci. Abstc, 21: 948. Sayegh, R., Wurtman, J., Spiers, P., McDermott, J. and Wurtman, R. (1995) The effect of a carbohydrate-rich beverage on mood, appetite, and cognitive function in women with premenstrual syndrome. Obstet. Gynecol., 86: 520-528. Scally, M.C., Ulus, I.H. and Wurunan, R.J. (1977) Brain tyrosine level controls striatal dopamine synthesis in haloperidol-treated rats. J. Neural Transm.,41: 1-6.

Spring, B. (1986) Effects of foods and nutrients on the behavior of normal individuals. In: R.J. Wurtman and J.J. Wurtman (Eds), Nutrition and the Brain, Vol. 7, Raven Press, New York, pp. 1-47. Sved, A.F., Femstrom, J.D. and Wurtman, R.J. (1979a) Tyrosine administration decreases prolactin levels in chronically reserpinized rats. Life Sci., 25: 1293-1300. Sved, A.F., Fernstrom, J.D. and Wurtman, R.J. (1979b) Tyrosine administration reduces blood pressure and enhances brain norepinephrine release in spontaneously hypertensive rats. Pmc. Natl. Acad. Sci. USA, 76: 3511-3514. Van Praag, H.M. and Korf, J. (1974) Serotonin metabolism in depression: clinical application of the probenecid test. Int. Phannacopsychiatry,9: 35-51.

E.A. Mayer and C.B. S a p (Eds.) Pmgress in Brain Research, Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAPTER 14

The evolving neurobiology of gut feelings Emeran A. Mayer*, Bruce Naliboff and Julie Munakata U C W C U R E Neumenteric Disease Program, Department of Medicine and Physiology, UCLA School of Medicine, Los Angeles, CA 90024, USA

Introduction The English language is full of expressions that suggest a long history of implicit understanding amongst its users of the important role of the viscera, in particular the digestive and cardiovascular systems, in the emotional and cognitive functions of the brain. Hippocratic, pre-scientific Western Medicine emphasized the important role of visceral secretions (‘bile’) in the generation of emotion and other cultures have used reference to the viscera as a way to describe imbalances within the energy system of the body, manifesting in part as emotional disturbances (for example, Chinese Medicine refers to excesses or deficiencies of liver, gallbladder, kidneys, etc.) (O’Connor and Bensky, 1983). In contrast, it has only recently been that science has come to consider a possible biological basis for such mind body interactions (Damasio, 1994). The common expressions ‘lump in the throat’, ‘butterflies in the stomach’, ‘stomach tied in knots’ or hating somebody’s ‘guts’ all point towards an implicit understanding in lay language about the intricate relationship between anger, fear or anxiety and specific, generally aversive, visceral experiences related to the gut. The fact that everybody knows intuitively what is meant by these expressions indicates that most people have experienced such sensations themselves and have formed a *Corresponding author. Tel.: (310) 312-9276; Fax: (3 10) 794-2864;e-mail: [email protected]

memory thereof. In many, but not all cases the memories involving gastrointestinal sensations refer to aversive situations. In contrast, expressions referring to visceral sensations attributed to the heart (‘broken heart’, ‘heartache’, ‘heartfelt wishes’, from ‘all my heart’) recall memories of affection, interpersonal relationships, love, loss, etc. Some of the most commonly used expressions, such as ‘my gut feelings tell me’ or you should ‘listen to your heart’ have additional implications, which refer to a type of pre-rational insight, or ‘emotional intelligence’ related to visceral sensations (Goleman, 1995). Even though, the person who refers to hisher gut feelings has an inner certainty that the decision is correct, no good reasons or hard facts can be given for this certainty. ‘Gut feelings’ refer to a type of intelligence that appears to be based on previous experiences, the encoding of which somehow was related to the experience of a visceral sensation. Nauta (1971) and more recently Damasio (1994) have taken this concept one step further. They have postulated that part or most of our rational decision making may be based on the encoded emotional content related to a positive or negative outcome of previous decisions, with all its visceral and somatic associations.

Transcendence of the Cartesian separation between mind and body For the past three centuries, in the spell of the Cartesian view of the human organism, the aim of

196

biological and medical research has been the understanding of the physiology and pathology of the body proper. With very few exceptions, such as psychoanalysis, and behavioral medicine approaches, emotions were largely left out of the focus of scientific investigations. Descartes’ primary reason for this selective focus on the body, leaving the mind as an exclusive domain for religion and philosophy, is assumed to be related to his survival strategy, vis-a-vis a powerful Church. The Church was intolerant towards any attempt to compete on its self-proclaimed territory of the human mind. While Descartes may have performed this ‘selective attention’ to the body for purely selfish reasons, the more recent rationales for this biased scientific view of the human organism are more related to arguments, that something as elusive and subjective as human emotions can never be investigated in scientific, reductionistic terms. Even after the mind became the focus of a specific discipline, psychology, it did not gain entry into biology and medicine until recently and it continues to be viewed in both disciplines as fundamentally separate from somatic, ‘organic’ processes (Mayer et al., 1997). Some of the primary reasons for the failure of various psychological disciplines to view the mind and emotions in terms conducive to neurobiological investigation have recently been summarized by LeDoux (1996) as follows: (a) The proper level of analysis of a psychological function is the level at which that function is represented in the brain. The various classes of emotions are mediated by separate neural systems that have evolved for separate reasons. (b) The brain systems that generate emotional behaviors are highly conserved through many levels of evolutionary history. (c) Emotional responses are, for the most part, generated unconsciously. However, when these neural systems function in an animal that also has the capacity for conscious awareness, and when attentional or arousal systems are simultaneously activated, then conscious emotionalfeelings occur. In this view, feelings are commonly associated, but not required for the expression of emotions. (d) The conscious feelings that we know and value our emotions by, are red herrings, in the study of emotions. Feelings of fear, for example, occur as

part of the overall reaction to danger and are probably less central to the overall emotional response to danger than the behavioral or the visceral physiological responses that also occur, such as diarrhea, sweating or palpitations. (e) Conscious emotional feelings are in a certain sense not different from other states of consciousness. These states of consciousness occur when attentional systems are activated simultaneously with the unconscious processing systems of specific emotional systems. These attentional systems appear to be more effective in the context of emotional feelings than in the context of rational thoughts. (f) Once emotions occur, the memory of images and associated bodily responses can become powerful motivators of future behaviors (‘somatic markers’). Several conclusions can be drawn from these statements which are pertinent for the following discussion: (i) If emotional responses and emotional feelings are effects caused by the activity of common underlying systems, the objectively measurable emotional response (autonomic, neuroendocrine, brain response) can be used to investigate the underlying neurobiological mechanisms of both emotions and feelings. (ii) If the majority of emotional processes occur without conscious awareness, characterizations of emotions based solely on assessment of cognitive symptoms (including instruments to assess anxiety or depression) will not yield significant insights into underlying mechanisms. This conclusion has been summarized by LeDoux in a beautiful way by the statement: “From the point of a lover, the only thing important about love is the feeling. But from the point of view of trying to understand what a feeling is, why it occurs, where it comes from, and why some people give or receive it more easily than others, love, the feeling, may not have much to do with it at all” (LeDoux, 1996).

Recent interest in neurobiology of emotions and the role of interaction between body and brain in both cognitive and emotional functions The intricate interdependence of feelings, emotions, cognitive function and visceral sensations has recently been discussed by Damasio (1994) and LeDoux (1996). Both authors have summarized

197

extensive animal experimental and human data and formulated a theory of brain function, which is inextricably connected to afferent feedback from the body, including the viscera. Starting with the original theory of emotions of William James (1890), a concept has been proposed according to which emotions are generated by CNS networks, in particular those involving the prefrontallanterior cingulate cortices, amygdala, hippocampus, and the hypothalamus. These brain regions in turn produce changes in the periphery via rapid autonomic and slower neuroendocrine modulation of target organ function. Altered target organ function, such as tachycardia, increased blood pressure, altered respiration, increased sweat production, increased activity of bladder and distal colon, is encoded by visceral afferents which feed back to some of the very same brain regions which were involved in triggering the central autonomic response in the first place. Similarly, neuroendocrine responses associated with emotions (epinephrine, cortisol) feed back to the brain on a much slower time scale (see also Chapter 7 by Plotsky in this volume). In addition to possible modulatory effects on the experienced emotional feeling, epinephrine through an action mediated by vagal afferents plays an important role in enhancing the memory of a particular experience associated with high epinephrine release (Cahill et al., 1994; McGaugh et al., 1995; Cahill and McGaugh, 1998). In contrast to W. Cannon’s concept of a stereotypic pattern of autonomic and neuroendocrine responses associated with emotional arousal, considerable evidence suggests that different types of emotions and stressors are associated with target-specific patterns of autonomic (and possibly neuroendocrine) modulation (Jihig and McLachlan, 1992) (see also Chapter 25 by Janig et al. in this volume). Thus the afferent feedback reaching the brain should also reflect this emotion-specific pattern of target organ modulation. There is an extensive literature describing the differential autonomic nervous system effects associated with different emotional states on the gut. Starting with Beaumont’s classic report on the emotion-specific changes in the mucosa of the exposed stomach of his patient Alexis St. Martin (Beaumont, 1833), a series of reports in the 1940s

by Wolf and Wolff (1943), Engel et al. (1956) and others demonstrated the sensitivity of autonomically regulated gut functions such as motility, secretion and mucosal blood flow to even minor events when the latter were invested with emotional significancefor the subject (Almy, 1989). Friedman and Snape reported blanching and hyperemia of children’s colostomies following brief painful stimulation of skin adjacent to the stoma, or simply during the expectation of repeated stimulation (Friedman and Snape, 1946). Almy and coworkers performed a series of studies monitoring sigmoid motor activity in human subjects during stress interviews (Almy et al., 1949; Almy et al., 1950). Increased phasic contractions were noted during emotional arousal and coping behavior manifested by aggressive gestures and facial and verbal expressions of hostility and self-assertion. In contrast, whenever the patient indicated verbally or non-verbally feelings of guilt, depression, helplessness and personal inadequacy, a prompt, transient cessation of phasic sigmoid motor activity was observed. The hypermotilekoping pattern and the hypomotilehelpless pattern were of variable duration, and sometimes alternated during the same experimental session (Almy et al., 1950). Even though the objectivity of these studies may be criticized from a methodological standpoint today, they are consistent with more recent studies by Welgan et al. (1988), who induced a state of anger by using more standardized laboratory stressors. In subjects suffering from a chronic visceral pain syndrome (irritable bowel syndrome, IBS) and to a lesser degree in healthy subjects, anger was associated with an increase in phasic motor activity in the sigmoid colon, and the increase was correlated with the intensity of the subjective rating of the associated emotional experience (Welgan et al., 1988). In another study, the fear induced during a sigmoidoscopic examination by the investigator’s report that the subject may have a malignancy resulted in increase in blood flow and motility of the sigmoid colon (Almy, 1951). In summary, these reports suggest that the experience of emotional feelings such as fear, anger and hostility are associated with an increased ratio of parasympathetic/sympathetic output to the lower gastrointestinal tract, whereas the feeling of depression

198

and helplessness may be associated with a reduction in this ratio. The observation during all of the cited studies that the emotion-specific alterations in gastrointestinal function were not perceived by the subjects is consistent with the concept that the great majority of visceral feedback to the brain occurs on a subliminal level. On the basis of reduced or absent skin conductance responses in frontal patients, Nauta (1971) hypothesized that the loss of prefrontal cortex would produce a state of ‘interoceptive agnosia’, due to the absence of emotion-related autonomic responses to the various alternatives the individual may be considering (Neafsey, 1990). Based on the specific cognitive defects of a series of well characterized patients with lesions of brain regions concerned with higher processing of visceral information (perigenual ACC, ventromedial PFC), Damasio recently developed the concept of the ‘somatic marker’ hypothesis to explain how rational decision-making is strongly influenced by memories of past positive or negative outcomes of previous decisions. The term is based on the observation that the recall of an outcome of a given response option (good or bad) is associated with the experience (consciously or subconsciously) of a positive or negative gut feeling. Since the feeling is about the body (including the viscera), Damasio refers to it as a somatic state. Since the gut feeling marks an image, he refers to it as a marker (Damasio, 1994). For example, when a person is faced with a decision between two options, the decision-making process will be strongly influenced by recall of bodily sensations experienced during previous experiences which may again be experienced depending on which one of the options is chosen: “When a negative somatic marker is juxtaposed to a particular future outcome the combination functions as an alarm bell. When a positive somatic marker is juxtaposed instead, it becomes a beacon of incentive” (Damasio, 1994). When a person is faced with a decision between a large number of options, the somatic marker-based decision making process may result in a reduction of the response options to those with an anticipated positive outcome. Damasio suggests that the experience of ‘gut feelings’ in these situations can result from either the re-enactment of the original

situation in autonomic terms, i.e. autonomic responses associated with the decision-making process produce peripheral changes which are encoded by afferent pathways feeding back to the brain. Alternatively, the experience of the original autonomic response is stored in memory and recalled in form of an ‘as if loop’ (Damasio, 1994). In the latter case, the ‘gut feelings’ associated with a decision making process no longer have to be created by autonomic responses in the periphery, but will be available for recall directly as visceral memory. Such a short circuiting of the peripheral ‘body loop’ would greatly enhance the speed with which such information may become available during the decision-making process. The close interplay between emotions, autonomic responses and activation of specific brain regions involved in the processing of ‘gut feelings’ is further illustrated by recent research efforts into the neurobiology of sleep states (Hobson et al., 1998). Recent evidence from brain imaging studies in humans have provided evidence that cortical brain regions involved in the processing of visceral information in the waking state are also activated during those portions of the sleep cycle which are characterized by high autonomic activity, i.e. the rapid eye movement (REM) sleep. For example, autonomic output to the heart (Somers et al., 1993) and to the colon (Fukudo and Suzuki, 1987) are significantly increased during REM sleep compared to non-REM sleep phases. Recent imaging studies indicate a preferential activation during REM sleep of limbic and paralimbic regions of the forebrain (including bilateral amygdala, anterior cingulate cortex (BA 24), parietal association cortex (BA 40) and entorhinal cortex) (Maquet et al., 1996; Braun et al., 1997; Nofzinger et al., 1997) and simultaneously, significant deactivation of the dorsolateral prefrontal cortex (Maquet et al., 1996; Braun et al., 1997). Based on these findings, it has been suggested that REM sleep and waking are similar brain-mind states. However, compared with waking, REM sleep is characterized by a unique pattern of modulation of ascending arousal systems. It is accompanied by aminergic demodulation (due to inhibition of neurons within locus coeruleus (noradrenergic) and raphe nuclei (serotonergic) with

199

ascending projections) and by cholinergic hypermodulation (due to disinhibition of pontine cholinergic neurons with ascending projections). The pattern of modulation of these ascending projections may explain the characteristic dream feature of ‘emotional intensification’ (Hobson et al., 1998). The brain imaging results are consistent with the postulated role for REM sleep in the processing of emotion in memory systems (Braun et al., 1997). Thus, while we see and judge the world in the waking state using dual systems of rational thinking as well as gut feelings, gut feelings dominate our experience of the reality of dreams during REM sleep, undisturbed by purely cognitive influences. One may speculate about the relationship of these recent neurobiological insights and the underlying assumptions of the various psychological schools emphasizing the importance of dream interpretations. The focus on a state of awareness that is dominated by ‘gut feelings’ may allow the individual to change cognitions and attitudes in a way enhancing his or her ‘emotional intelligence’.

Unique functions of visceral afferents and visceral pain conducive to the generation of gut feelings and visceral memories In contrast to signals processed by our special senses, the majority of interoceptive, visceral signals are subliminal, i.e. they are generally not consciously perceived. Normal visceral sensations occur in the form of vague states of feelings associated with satiety, hunger, fullness, urgency and relief of urgency only when associated arousal brings these visceral afferent signals into conscious awareness: The sense of urgency preceding the emptying of urinary bladder or rectum, or the sense of satiety preceding the cessation of food intake. The sensation can fall below the perception threshold again, if the attention is shifted to another focus. Thus the intensity of perception is not directly related to the degree of afferent stimulation, but rather to the relative influences of arousal and afferent input. Pain arising from the viscera in the animal living in the wild may be rare, primarily occurring in the context of acute visceral inflammation or the acute

irritation of the gastrointestinal tract due to intake of poisons, infective organisms or spoiled food. In contrast to somatic pain, which is localized and generally results in purposeful motor responses to escape the painful stimulus, the behavioral response is mediated by the autonomic nervous system, aimed at expulsion of the ingested substance. This occurs in the form of stereotypic visceromotor responses, such as vomiting and diarrhea. The associated pain is poorly localized with a strong emotional component. The primary function of the associated emotional feeling is not known, but is likely to play an important role in the enhancement of memory of this event to avoid future exposure to similar situations of harmful oral intake. Such visceral memory formation of an aversive event can be considered the forerunner of ‘gut feelings’ which reminds the affected organism of a negative outcome of a given response option. Depending on the intensity of the emotional response associated with the visceral event, an animal is likely to avoid the same choice of food intake or even the site of the meal, when faced with the same response option again in the future. In contrast to the enhanced memory formation of the emotionally charged response to acute visceral pain, there appears to be less evolutionary advantage of experiencing chronic visceral pain. Chronic somatic pain engages the organism in behaviors characterized by protection of the affected body part from further injury and to promote healing. In contrast, the appropriate response options of an animal to chronic visceral pain are poorly understood but may be less specific. They may include altered vital functions, and general withdrawal of attention from the environment.

The unique functions of visceral afferents are reflected in their neuroanatomy The differences between perception of somatic and visceral signals summarized above are reflected in differences in the functional neuroanatomy between visceral and somatic afferent pathways (Ness and Gebhart, 1990; Mayer and Gebhart, 1994). Visceral information is encoded by fewer specialized sense organs, and the encoded information is transmitted by fewer and slower conducting

200

afferent fibers. Spinal visceral afferents project divergently onto a large number of spinal segments and there is considerable overlap of projections from different organs. Spinal visceral afferent fibers converge with somatic fibers from deep structures onto the same dorsal horn neurons and there appear to be no viscero-specific ascending spinal pathways. At the supraspinal level, visceral afferent information carried by spinal and vagal afferents is primarily processed in brain regions belonging to the medial pain system (thalamus, insula, perigenual anterior cingulate and prefrontal cortex) concerned with autonomic, affective, and antinociceptive responses, and not in brain regions concerned with sensory discrimination and motor responses (sensorimotorcortex, mid anterior cingulate cortex) (for details, see below). These brain regions show the highest density of p opioid receptors within the brain (Vogt et al., 1996) and are therefore well suited to attenuate the conscious experience of pain. In summary, the special features of visceral afferent pathways match closely the unique functions summarized earlier: They are not designed for precise and fast stimulus assessments but rather for slow sampling of information from general body regions. They are not ‘hardwired’ to fast motor systems capable of selective behavioral responses, but to stereotypic autonomic responses. There is a close connection between brain regions involved in the generation of emotions and the formation of emotional memory.

Human brain imaging studies of visceral afferent input to the brain In a series of ongoing studies, our group has begun to characterize the regions of the brain that play a role in the regulation of autonomic responses to visceral stimuli, and those that are engaged during the conscious perception of visceral sensations generated by experimental balloon distension of the colorectum. In general, these studies are aimed to enhance our understanding of how ‘gut feelings’ are generated in health and disease by studying brain responses associated with visceral sensations. Even though the methodology is crude compared to the subtleties of physiologically occurring, bi-

directional interactions between the brain and the viscera, it nevertheless allows for the characterization of important aspects of the concepts outlined above. In particular, the studies were designed to address the following questions: (a) Which brain areas are activated by distension of the rectosigmoid and are correlated with the conscious perception of such stimuli? (b) Which brain regions are activated in correlation with changes in autonomic responses? (c) Which brain regions are activated following the delivery of a repetitive phasic stimulus of noxious intensity?

Brain regions concerned with conscious perception of visceral stimuli as compared to somatic stimuli Derbyshire et al. recently reviewed published studies which have used positron emission tomography to determine changes in regional brain activity associated with the experience of somatic pain (Derbyshire et al., 1997). The region most commonly reported was the mid-anterior cingulate cortex (ACC), followed by insula and thalamus. Other regions, which were less consistently reported, include prefrontal cortex, lentiform nucleus and somatosensory cortex. In order to determine if brain regions activated during experimental visceral stimulation (ranging from non-painful to painful) differ from those activated during somatic stimuli, Munakata et al. performed activation studies using H,I5O positron emission tomography (PET) during visceral stimulation (Munakata et al., 1999). Covariate analysis of regional brain activity with subjective ratings of stimulus intensity (Gracely et al., 1978) was performed in 69 subjects who underwent a visceral stimulation paradigm using graded rectal balloon distension (phasic distensions to 20-60 mmHg) (Munakata et al., 1999). The protocol also included ‘simulated’, but undelivered distensions, where the subjects were warned by a taped message to expect a more intense stimulus than the preceding one, yet no stimulus was delivered. In the type of analysis used (covariate analysis) brain regions were identified which correlate with the conscious processing of the visceral stimulus (and or the recall of such perception).

20 1

The 69 subjects (33 males, 36 females) were a heterogeneous group comprised of 17 healthy control subjects, 43 patients with irritable bowel syndrome (IBS) and 9 patients with quiescent ulcerative colitis (UC), a chronic inflammatory bowel disease. This analysis was based on the a priori hypothesis that brain networks involved in the generation of subjective sensations should not differ qualitatively between these subject populations, despite likely difference in the magnitude or time course of activation and deactivation. Subjective sensory ratings showed significant positive correlations with bilateral activity in the lateral (but not medial) prefrontal cortex (BA 9 and lo), the bilateral ACC (BA 32), the premotor cortex (BA 6/44), the supramarginal gyrus of the parietal cortex (BA 40), the inferior gyrus of the frontal cortex (BA 46) and the cerebellum. Significant negative correlations with bilateral activity were seen in the hippocampus, the posterior cingulate cortex (BA 31), the orbital gyri (BA 19), and the middle temporal gyrus (BA 39). A subsequent analysis showed that these results were not confounded by absence or presence of disease conditions (IBS or UC). Thus, similar to concepts proposed for central processing of somatic pain, the conscious processing of physiological and painful visceral stimuli involves a network(s) of brain regions. It has to be emphasized that in the reported study, stimulus intensities were not limited to the painful range and that even the highest intensity of tolerated visceral stimuli is probably lower than the comparable intensity of somatic pain stimuli used in experimental studies (capsaicin injection, laser heat, etc). In contrast to somatic pain studies, no significant activations were seen in somatosensory areas, in the thalamus, insula and lentiform nucleus. Several of the identified brain regions overlap with brain regions referred to as the ‘medial pain system’ (see also Chapter 16 by Vogt in this volume).

using H2”0 positron emission tomography while simultaneously recording autonomic responses in form of heart rate (Munakata et al., 1999). Visceral stimulation was performed by controlled tonic rectal balloon distensions of 60 s durations. Following the delivery of a 20 and 45 mmHg stimulus (separated by a 15 min rest period), the subjects were ‘warned’ to expect a series of pressure pulses of much greater intensity and discomfort (expectation period; Exp.). Three conditions occurred during this expectation period, each separated by a 15 min rest period: during the first expectation condition (Exp. l), delivery of a pressure pulse was simulated, but without actual balloon inflation. This was followed by an actual pressure pulse of 60 mmHg, which again was followed by a second expectation condition (Exp. 2). No significant heart rate changes were observed during Exp. 1 and 2, while mean heart rate increased by 5% during the 60 mmHg stimulus. Covariate analyses for regional brain activity and heart rate response during the three stimuli of the expectation period (Exp. 1, 60 mmHg, Exp. 2) showed a significant negative correlation of heart rate response with activity in the medial (but not lateral) PFC bilaterally (BA 9/10) and the perigenual ACC (BA 24,32). In addition, a positive correlation was seen with activity in the midline region in the dorsal pons, close to the periaqueductal grey (PAG). These findings suggest that expectation of an aversive visceral stimulus, even in the absence of detectable autonomic output to the heart, is associated with activity of brain regions within the visceromotor cortex (see Chapter 24 by Bandler in this volume). Activity in these brain regions in the current study may have been related to autonomic output to viscera other than the heart which was not measured in this study (i.e. gastrointestinal tract), or they may reflect activity changes in premotor regions of the visceral motor cortex in anticipation of an unpleasant visceral experience.

Brain regions concerned with processing of autonomic function

Brain regions activated by noxious repetitive visceral stimuli

In order to identify brain regions that participate in the central regulation of autonomic function, Munakata et al. studied 12 healthy control subjects

In a recent study by Mayer et al. (Mayer et al., 1998), a modified rectostimulation paradigm was used to study the effect of a visceral stimulus of

202

noxious stimulus intensity on regional brain activity. In this study brain activity in response to a 45 mmHg distension as well as during the expectation of an intense rectal pressure stimulus was assessed before (Exp,,) and after (Exp,,,) a train of noxious repetitive sigmoid distensions in 10 healthy control subjects. It had previously been shown that this type of repetitive noxious distension induces rectal hyperalgesia in patients suffering from IBS, but not in healthy control subjects (Munakata et al., 1997) or in patients with quiescent ulcerative colitis, a chronic inflammatory condition of the colon (Chang et al., 1996). We had interpreted the results from these earlier studies as consistent with activation of anti-nociceptive systems in healthy control subjects and ulcerative colitis patients (who do not report significant pain), but not in IBS patients. In contrast to the two studies discussed previously, in which a covariate analysis with either perceptual ratings or autonomic responses were used, a contrast analysis was used in the current study. Thus, observed changes in regional brain activity could be related to either activity changes in networks concerned with perception, autonomic control or antinociception. During the 45 mmHg distensions and during both expectation conditions (i.e. without any actual delivery of a visceral stimulus), activation of orbitofrontal (BA1l), perigenual ACC and medial PFC (BA 24/32) and anterior insula was observed. Following the train of the noxious repetitive sigmoid distension, additional activation in the thalamus and the PAG region was observed supporting a broader activation of antinociceptive systems in response to noxious input. Only one of the subjects reported pain during this paradigm.

Central processing of visceral and emotional pain within the ‘medial pain system’ When viewed together, the results from these three studies allow us to make several conclusions: (a) Within the PFC, activity changes in lateral regions are more associated with perceptual responses, while activity changes in the medial region are more associated with autonomic responses. Within ACC, changes within mid ACC are associated with perceptual responses, while changes within peri-

genual ACC are associated with autonomic responses. Even though comparisons between brain regions described in non-human primates with those found in the human brain have to be made with caution, the observed regional differences in activation may be related to the lateral and medial networks described by Price and coworkers (An et al., 1998; Ongiir et al., 1998) [(see also R. Bandler et al., Chapter 24) in this issue]. (b) The anticipation of an imminent aversive visceral sensation from the lower GI tract (with little or no associated conscious perception or heart rate response) showed the same activation in these regions as did the actual visceral stimulus, suggesting that recall of aversive visceral memory is processed in the same brain regions in which the actual experience is processed (Fuster, 1997). Since some subjects also reported sensations during these simulated distensions, the observed activation may be part of the ‘as if loops’ which generate a similar type of gut feeling without involvement of the peripheral ‘body loop’ (Damasio, 1994). (c) There are similarities as well as important differences in regional brain activation by anticipated and by experienced painful and non-painful visceral stimuli with those reported from somatic pain studies. Brain regions involved in the processing of visceral stimuli and in generating the appropriate autonomic, affective responses to these stimuli (i.e. medial PFC and perigenual ACC) overlap with brain regions identified as part of the medial pain system (Vogt and Gabriel, 1993). As reviewed elsewhere in this volume (see Vogt and Sikes, Chapter 16), activity changes within this brain region have also been observed in a variety of conditions with strong emotional context, such as sadness (George et al., 1993, 1995), while deep brain stimulation of this region in both animals and humans has been found to produce verbal reports or behaviors characteristic of emotional responses (Meyer et al., 1973; Bancaud and Talairach, 1992). The unique role of perigenual ACC/medial PFC in the generation of emotional feelings, antinociceptive responses and autonomic responses is further highlighted by older studies implicating this brain structure in the ‘family-related triad’ maternal behavior, playful behavior and separation cry (MacLean, 1993). Stamm reported that pregnant

203

rats with cingulate lesions failed to construct a preparatory nest and were deficient in performing expected postpartum forms of maternal behavior, and only 12% of the pups survived (Stamm, 1955). Similar findings were observed in hamsters (Murphy et al., 1981). One of the earliest forms of emotional pain is related to the separation of newborns from their mothers [see also contributions to this volume by Plotsky (Chapter 7) and Read (Chapter 30)]. The behavioral response of newborns to this maternal separation has been called the ‘separation cry’. Pharmacological interventions primarily aimed at the treatment of externally caused pain, such as morphine (Panksepp et al., 1978) and antagonists of the substance P receptor (Kramer et al., 1998) are able to attenuate this audiovocal expression of early emotional pain. MacLean and Newman showed that ablation of a continuous strip of perigenual and infragenual cingulate cortex in Saimiri monkeys results in a failure of these animals to produce an isolation cry (MacLean and Newman, 1988), while electrical stimulation of the same brain region elicits vocalization in the monkey (Smith, 1945).

Possible correlation of emotional disturbances with altered central processing of visceral afferent and efferent information Different types of cognitive and emotional disturbances have been described in the context of altered processing of visceral afferent information. While in some of these pathologies (autism, inflammatory bowel disease), a causal effect relationship has not been demonstrated, the most conclusive evidence comes from patients with traumatic or surgical brain lesions. In addition to the description of the classic patient Phineas Gage, Damasio summarized reports from four other patients with lesions of ventromedial PFUperigenual ACC which shared distinctive cognitive defects related to judgement and future planning (Bechara et al., 1994; Damasio et al., 1994). Several studies point towards a role of altered function of the PFC/ACC region in patients with depression (Frith and Dolan, 1998). In depressed patients, perigenual ACC fails to activate normally when patients perform a complex planning task (Elliott et al., 1997a). Functional activity

in a similar brain region has been shown to be predictive of drug responsiveness, with hypometabolism predicting non-responsiveness and hypermetabolism predicting responsiveness (Mayberg et al., 1997). Drevets et al. have demonstrated a localized functional and morphometric abnormality in perigenual ACC/medial PFC of patients with bipolar disorder and unipolar patients with a family history of depression (Drevets et al., 1997). In view of the evidence implicating perinatal events (see Chapter 7 in this volume by Plotsky), in particular early maternal separation in the etiology of affective disorders, one may speculate that early structural and functional alterations in perigenual ACC, induced either by neural inputs or by chronically enhanced cortisol feedback may play a crucial role in the development of certain types of affective disorders. Subgenual ACCPFC (a region involved in central autonomic and visceral afferent processing) is activated when normal subjects monitor feedback in a guessing task (Elliott et al., 1997b). In contrast, depressed patients, when performing an identical task, fail to activate this region when processing feedback (Elliott et al., 1998). These findings have been interpreted to suggest that the perigenuakubgenual subregion of ACC may play a role in evaluating feedback, in terms of rewards and punishment, of current and future behavior, and that this function is compromised in patients with certain affective disorders (Frith and Dolan, 1998). Patients suffering from a common functional abdominal pain syndrome (IBS) commonly show evidence for affective ‘comorbidity’. Up to 60% of patients with TSS show evidence for psychiatric disorders, with depression and anxiety being the most common diagnoses (Creed, 1998). Conversely, patients seen in mental health clinics show a high prevalence for IBS. There is a 60% prevalence in patients seeking medical care for dysthymia and 46% for panic disorder (Kaplan et al., 1996; Masand et al., 1997). It has been suggested that the prevalence of psychiatric comorbidity varies with the patient population studied: while so-called non-consulters (symptomatic individuals who do not seek healthcare for their symptoms) appear to have the same prevalence as healthy control subjects, up to 60% of IBS

204

patients seen in tertiary referral centers show affective comorbidity (Drossman et al., 1995). In contrast, in a large population-based survey by Lydiard et al., an association of IBS symptoms with psychiatric illness was observed, regardless of treatment seeking status of the individual respondents (Lydiard et al., 1994). Considerable evidence suggests that it is primarily the severity of symptoms, rather than the associated affective morbidity that determines if a patient seeks medical advice (Talley et al., 1997). One may speculate that psychosocial and affective disturbances in mildly symptomatic IBS patients are not consciously perceived and are not detectable with traditional psychometric instruments (but may be seen in structured interviews administered by mental health professionals) for the same reason that the majority of emotional responses occur without associated emotional feelings. Only when the symptoms become severe enough to trigger sufficient arousal associated with the subliminal emotional response, will feelings of anxiety and depression become clinically detectable.

Summary and conclusion The bi-directional communication between limbic regions and the viscera play a central role in the generation and expression of emotional responses and associated emotional feelings. The response of different viscera to distinct, emotion-specific patterns of autonomic output is fed back to the brain, in particular to the cingulofrontal convergence region. Even though this process unfolds largely without conscious awareness, it plays an important role in emotional function and may influence rational decision making in the healthy individual. Alterations in this bi-directional process such as peripheral pathologies within the gut or alterations at the brain level may explain the close association between certain affective disorders and functional visceral syndromes.

Acknowledgements This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK 48351 and the Department of Veterans Affairs Medical Research Program.

References Almy, T.P. (195 I ) Experimental studies on the imtable bowel syndrome. Am. J. Med., 10: 60-67. Almy, T.P. (1989) Historical perspectives of functional bowel disease. In: W.J. Snape Jr. (Ed.), Pathogenesis of Functional Bowel Disease, 1, Plenum Medical Book Company, New York and London, pp. 1-1 1. Almy, T.P., Abbot, F.K. and Hinkle, L.E. (1950) Alterations in colonic function in man under stress. IV: Hypomotility of the sigmoid colon and its relationship to the mechanism of functional diarrhea. Gastroenterology, 15: 95-105. Almy, T.P., Hinkle, L.E. and Berle, B. (1949) Alterations in colonic function in man under stress. 111 Experimental production of sigmoid spasm in patients with spastic constipation. Gastroenterology, 12: 437-449. An, X., Bandler, R., Ongiir, D. and Price, J.L. (1998) Prefrontal cortical projections to longitudinal columns in the midbrain periaqueductal gray in macaque monkeys. J. Comp. Neurol. (in press) Bancaud, J. and Talairach, J. (1992) Clinical semiology of frontal lobe seizures. Adv. Neurol., 57: 3-58. Beaumont, W. (1833) Experiments and Observations on the Gastric Juice, and the Physiology of Digestion, F.P.Allen, Plattsburg. Bechara, A., Damasio, A.R., Damasio, H. and Anderson, S.W. (1994) Insensitivity to future consequences following damage to human prefrontal cortex. Cognition, 50: 7-15. Braun, A.R., Balkin, T.J., Wesensten, N.J., Carson, R.E., Varga, M., Baldwin, P., Selbie, S., Belenky, G. and Herscovitch, P. (1997) Regional cerebral blood flow throughout the s l e e p wake cycle. Brain, 120: 1173-1 197. Cahill, L. and McGaugh, J.L. (1998) Mechanisms of emotional arousal and lasting declarative memory. TINS, 21: 294-299. Cahill, L., Prins, B., Weber, M. and McGaugh, J.L. (1994) pAdrenergic activation and memory for emotional events. Nature, 371: 702-704. Chang, L., Munakata, J., Saba, L., Naliboff, B., Procaccino, F. and Mayer, E.A. (1996) Evidence for the activation of endogenous pain modulation system in ulcerative colitis (UC). Gastroenterology, 110: A645 (Abstract) Creed, F.H. (1998) Psychopathology of functional disorders of the gut. In: S.F. Phillips and D.L. Wingate (Eds), Funcrional Disorders of the Gut: A Handbook for Clinicians, WB Saunders Co. Ltd., London. Damasio, A.R. (1994) Descartes' Error, G.P. htnam's Sons, New York, pp. 1-312. Damasio, H., Grabowski, T., Frank, R., Galaburda, A.M. and Damasio, A.R. (1994) The return of Phineas Gage: The skull of a famous patient yields clues about the brain. Science, 264: 1102-1 105. Derbyshire, S.W.G., Jones, A.K.P., Gyulai, F., Clark, S., Townsend, D. and Firestone, L.L. (1997) Pain processing during three levels of noxious stimulation produces differential patterns of central activity. Pain, 73: 431445. Drevets, W.C., Price, J.L., Simpson, J.R., Jr., Todd, R.D., Reich, T., Vannier, M. and Raichle, M.E. (1997) Subgenual

205 prefrontal cortex abnormalities in mood disorders. Nature, 386 824827. Drossman, D.A., Creed, F.H. and Fava, G.A. (1995) Psychosocial aspects of the functional gastrointestinal disorders. Gastmenteml. Intl., 8: 47-90. Elliott, R., Baker, S.C., Rogers, R.D., OLeary, D.A., Paykel, E.S.,Frith, C.D., Dolan, R.J. and Sahakian, B.J. (1997a) Refrontal dysfunction in depressed patients performing a complex planning task a study using positron emission tomography. Psychol. Med., 27: 93 1-942. Elliott, R., Frith, C.D. and D o h , R.J. (1997b) Differential neural response to positive and negative feedback in planning and guessing tasks. Neuropsychologia, 247: 1395-1404. Elliott, R., Sahakian, B.J., Michael, A., Paykel, E.S. and Dolan, R.J. (1998) Abnormal neural response to feedback on planning and guessing tasks in patients with unipolar depression. Psychol. Med., 28: 559-571. Engel, G., Reichsman, F. and Segal, H.L. (1956) A study of an infant with gastric fistula. Psychol. Med., 18: 374-398. Friedman, M.H.F. and Snape, W.J. (1946) Color changes in the mucosa of the colon in children as affected by food and psychic stimuli. Fed. Proc., 5: 30. Frith, C. and D o h , R.J. (1998) Images of psychopathology. Cur,: Opin. Neumbiol., 8: 259-262. Fukudo, S. and Suzuki, J. (1987) Colonic motility, autonomic function, and gastrointestinal hormones under psychological stress on irritable bowel syndrome. Tohoku. J. Exp. Med., 151: 373-385. Fuster, J.M. (1997) Network memory. TINS, 20: 451-459. George, M.S ., Ketter, T.A., Gill, D.S., Haxby, J.V., Ungerleider, L.G. and Herscovitch, P. (1993) Brain regions involved in recognizing facial emotion or identity: an oxygen-15 PET study. J. Neuropqchiatry Clin. Neurosci., 5 : 384-394. George, M.S., Ketter, T.A., Parekh, P.I., Horwitz, B., Herscovitch, P. and Post, R.M. (1995) Brain activity during transient sadness and happiness i n healthy women. Am. J. Psychiatry, 152: 341-351. Goleman, D. (1995) Emotional Intelligence: Why it Can Matter More Than IQ, Bantam Books, New York. Gracely, R.H., McGrath, P. and Dubner, R. (1978) Ratio scales of sensory and affective verbal pain descriptors. Pain, 5: 5-18. Hobson, J.A., Pace-Schott, E.F., Stickgold, R. and Kahn, D. (1998) To dream or not to dream? Relevant data from new neuroimaging and electrophysiological studies. Curr. Opin. Neumbiol., 8: 239-244. Jaenig, W. and McLachlan, E.M. (1992) Characteristics of function-specific pathways in the sympathetic nervous system. TINS, 15: 475-481. James, W. (1890) Principles of Psychology, Holt, New York. Kaplan, D.S., Masand, P.S. and Gupta, S. (1996) The relationship of irritable bowel syndrome (IBS) and panic disorder. Ann. C h i . Psychiatry, 8: 81-88. Kramer, M.S., Cutler, N., Feighner, J., Shrivastava, R., Carman, J., Sramek, J.J., Reines, S.A., Liu, G., Snavely, D. and WyattKnowles, E. (1998) Distinct mechanism for antidepressant

activity by blockade of central substance P receptors. Science, 281: 1640-1645. LeDoux, J.E. (1996) The Emotional Brain: the Mysterious Underpinnings of Emotional Life, Simon & Schuster, New York. Lydiard, R.B., Greenwald, S., Weissman, M.M., Johnson, J., Drossman, D.A. and Ballenger, J.C. (1994) Panic disorder and gastrointestinal symptoms: Findings from the NIMH Epidemiologic Catchment Area Project. Am. J. Psychiatry, 151: 64-70. MacLean, P.D. (1993) Perspectives on cingulate cortex in the limbic system. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus: a Comprehensive Handbook, Birkhauser, Boston, pp. 1-15. MacLean, P.D. and Newman, J.D. (1988) Role of midline frontolimbic cortex in production of the isolation call of squirrel monkeys. Brain Res., 45: 111-123. Maquet, P., Peters, J.M., Aerts, J., Delfiore, G., Degueldre, C., Luxen, A. and Franck, G. (1996) Functional neuroanatomy of human rapid-eye-movement sleep and dreaming. Nature, 383: 163. Masand, P.S., Kaplan, D.S., Gupta, S. and Bhandary, A.N. (1997) Irritable bowel syndrome and dysthymia. is there a relationship? Psychosomatics, 38: 63-69. Mayberg, H.S., Brannan, S.K., Mahurin, R.K., Jarabek, P.A., Brickman, J.S., Tekell, J.L., Silva, J.A., McGinnis, S., Glass, T.G., Martin, C.C. and Fox, P.T. (1997) Cingulate function in depression: a potential predictor of treatment response. NeuroReport, 8: 1057-1061. Mayer, E.A. and Gebhart, G.F. (1994) Basic and clinical aspects of visceral hyperalgesia. Gastroenterology, 107: 271-293. Mayer, E.A., Liu, M., Munakata, J., Mandelkern, M., Berman, S., Silverman, D.H.S., Schmulson, M., Williams, J., Kovalik, E. and Naliboff, B. (1998) Regional brain activity before and after visceral sensitization. SOC. Neurosci. Abst,:, 24: 529. (Abstract) Mayer, E.A., Munakata, J. and Chang, L. (1997) Does mindbody medicine have a role in gastroenterology? Curr. Opin. Gastroenterol., 13: 1 4 . McGaugh, J.L., Cahill, L., Parent, M.B., Mesches, M.H., Coleman-Mesches, K. and Salinas, J.A. (1995) Involvement of the amygdala in the regulation of memory storage. In: J.L. McGaugh, F. Bermudez-Rattoni and R.A. Prado-Alcala (Eds), Plasticity in the Central Nervous System: Learning and Memory, Erlbaum, Hasdale. Meyer, G., McElhaney, M., Martin, W. and McGraw, C.P. (1973) Stereotactic cingulotomy with results of acute stimulation and serial psychological testing. In: L.V. Laitinen and K.E. Livingston (Eds), Surgical Approaches in Psychiatry, MTP, Lancaster, pp. 39-58. Munakata, J., Naliboff, B., Chang, L., Mandelkem, M., Derbyshire, S.W.G. and Mayer, E.A. (1999) Gender differences in brain regions associated with subjective intensity ratings of visceral stimuli. Gastroenterology, 116 (Abstract) Munakata, J., Naliboff, B., Harraf, F.,Kodner, A., Lembo, T., Chang, L., Silverman, D.H. and Mayer, E.A. (1997)

206 Repetitive sigmoid stimulation induces rectal hyperalgesia in patients with irritable bowel syndrome. Gastroenterology, 112: 55-63. Murphy, M.R., MacLean, P.D. and Hamilton, S.C. (1981) Species-typical behavior of hamsters deprived from birth of the neocortex. Science, 213: 459461. Nauta, W.J.H. (1971) The problem of the frontal lobe: a reinterpretation. J. Psychiatric Res., 8: 167-187. Neafsey, E.J. (1990) Prefrontal cortical control of the autonomic nervous system: Anatomical and physiological observations. Progr: Brain Res., 85: 147-166. Ness, T.J. and Gebhart, G.F. (1990) Visceral pain: a review of experimental studies. Pain, 41: 167-386. Nofzinger, E.A., Mintun, M.A., Wiseman, M.B., Kupfer, D.J. and Moore, R.Y. (1997) Forebrain activation in REM sleep: an FDG PET study. Brain Res., 770: 192-201. OConnor, J. and Bensky, D. Eds. (1983) Acupuncture. A Comprehensive Test. (Shanghai College of Traditional Medicine), Eastland Press, Chicago, pp. 1 4 3 . Ongiir, D., An, X. and Price, J.L. (1998) Prefrontal cortical projections to the hypothalamus in macaque monkeys. J. Comp. Neurol. (in press) Panksepp, J., Herman, B., Conner, R., Bishop, P. and Scott, J.P. (1978) The biology of social attachments: Opiates alleviate separation distress. Biol. Psychiatry, 13: 607-6 18.

Smith, W.K. (1945) The functional significance of the rostra1 cingular cortex as revealed by its responses to electrical excitation. J. Neurophysiol., 8: 24 1-255. Somers, V.K., Dyken, M.E., Mark, A.L. and Abboud, EM. (1993) Sympathetic-nerve activity during sleep in normal subjects. N. Engl. J. Med., 328: 303-307. Stamm, J.S. (1955) The function of the median cerebral cortex in maternal behavior of rats. J. Compar: Physiol. Psychol., 47: 21-27. Talley, N.J., Boyce, P.M. and Jones, M. (1997) Predictors of health care seeking for irritable bowel syndrome: A population based study. Gut, 41: 394-398. Vogt, B.A., Derbyshire, S. and Jones, A.K.P. (1996) Pain processing in four regions of human cingulate cortex localized with co-registered PET and MR imaging. Eur. J. Neurosci., 8: 1461-1473. Vogt, B.A. and Gabriel, M. Eds. (1993) Neurobiology of Cingulate Cortex and Limbic Thalamus: a Comprehensive Handbook, Birkhaeuser, Boston. Welgan, P., Meshkinpour, H. and Beeler, M. (1988) The effect of anger on colon motor and myoelectric activity in irritable bowel syndrome. Gastroenterology, 94: 1150-1 156. Wolf, S. and Wolff, H.G. (1943) Human Gastric Function, Oxford University Press, London.

SECTION VI

Influences of the body on the brain

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research, Vol 122 0 2000 Elscvier Science BV. All rights reserved.

CHAPTER 15

Integration of viscerosomatic sensory input at the spinal level Robert D. Foreman* George Lynn Cross Research Professor and Chair; Samuel Roberts Noble Foundation Presidential Professor; Universily of Oklahoma Health Sciences Center; College of Medicine, 940 Stanton L. Young Blvd., BSMB, Room 653, Oklahoma City, OK 73190, USA

Introduction Throughout the history of medicine, headache pain has afflicted human beings. People have sought relief from the physician, medicine man, shaman and others who use complementary treatments. Significant advances have been made to treat many of the various types of headaches, but much more work needs to be done to understand how the associated pain is perceived and all the related autonomic, psychological, emotional, and behavioral response are generated in the central nervous system (Goadsby, 1997). These investigations should also address the treatments that are used with alternative medicines. Of special interest for this chapter is the pain of cervical headache that originates from the cervical spine but is referred to the head (Bogduk, 1997). This term, cervical headache, combines critical elements of the headache and the cervical origin, but this description may be misleading because it implies that headache is in the neck. The description is useful, however, in that it encompasses all types of headaches of cervical origin (Bogduk, 1997). Our interest in headache developed because this pain seemed to have some characteristics similar to angina pectoris that was referred to the neck and jaw. *Corresponding author. Tel.: (405) 27 1-2226; Fax: (405) 271-3 181; e-mail: [email protected]

Headaches of cervical origin and the accompanying referred pain most likely occur because of the neuroanatomical organization in the upper cervical segments of the spinal cord. Gray matter of the Cl-C3 spinal segments converge with par caudalis of the spinal nucleus of the trigeminal nerve to form a column of gray matter in these cervical segments called the trigeminal cervical nucleus (Fig. 1) (Kerr, 1961; Pfaller and Arvidsson, 1988). Nociceptive afferent fibers transmit their impulses from the trigeminal nerve and from the first three cervical spinal nerves to this converged column. The trigeminal and spinal afferent fibers form multiple collateral terminals that overlap and converge on common second order neurons of the upper cervical segments to provide the basis for referred pain (Bogduk, 1997). Investigations for mechanisms of cervical headache have focused on the anatomy and physiology of somatic structures that innervate the upper cervical segments (Bogduk, 1980, 1982; Bogduk and Marsland, 1986, 1988; Pollmann, et al., 1997). Recent studies from our laboratory have shown that afferent input from visceral organs converge with somatic input onto second order neurons and propriospinal neurons of the upper cervical segments. These neurons might be the same ones that were implicated in cervical headache. The purpose of this chapter is to show the complex organization of visceral and somatic input in these segments of the spinal cord. The chapter will be divided into

2 10

Trigeminal Ganglion

C2 Dorsal Root Ganglion

I

I

1

I I

I 1

I

b

C

b'

C'

d

Fig. I . Transverse sections from the caudal medulla (a, a') and the C1 (b, b'), C2 (c, c'), and C3 (d, d') spinal segments. The sections show the labeling pattern following WGA-HRP injections into the C2 dorsal root ganglion (left panel) and the trigeminal ganglion (right panel). Modified from Pfaller and Arvidsson (1988) with permission from Alan R. Liss, Inc.

21 1

three sections to address these findings. Section I will address the activation of Cl-C3 spinothalamic tract cells with vagal and thoracic sympathetic input and the convergence of this visceral information with the somatic input. Section 11 will discuss the propriospinal relay in the upper cervical segments that provides intraspinal communication between visceral organs. Finally, section I11 will address pain referral that could result from the convergence of craniovascular afferent input and somatic inputs in these segments. This information will deepen our understanding about the complexity of upper cervical segments and expand possibilities for additional causes of cervical headaches. An important feature is the convergence of input from distant visceral organs that might converge to produce headaches of cervical origin.

I. Convergence of cardiac afferent inputs in upper cervical segments Viscerosomatic convergence is an important underpinning of osteopathic physicians for understanding communication between visceral organs and somatic structures. These physicians believe that altered or impaired function in the somatic system frequently occurs as presymptomatic sign of visceral disease (Beal, 1985). These symptoms may be expressed as somatovisceral or viscerosomatic reflexes. The viscerosomatic reflex results from increased afferent activity arising from the diseased visceral organ. These afferent fibers activate reflexes in the gray matter of the spinal cord where the information is then transmitted to the peripheral somatic structures. Osteopathic physicians use palpatory cues of the body surface, especially along the spinal column, to identify muscle hypertonicity and irritability or subcutaneous edema associated with preclinical visceral disease. During the chronic phase, muscles especially along the vertebral processes become hard and tense, and they may be hypersensitive to palpation (Burns, 1928). Patients with coronary artery disease produce a viscerosomatic reflex that results in hypertonicity of the Tl-T5 and occasionally Cl-C2 vertebral muscles (Fig. 2) (Beal, 1985). The tonicity changes in the upper thoracic somatic structures correlate well with the innervation of the Tl-T5 spinal

segments by cardiac sympathetic afferent fibers and the convergence of somatic spinal nerves (Vance and Bowker, 1983; Kuo, et al., 1984; Beal, 1985; Hopkins and Armour, 1989). This correlation does not hold for the patients having changes in upper cervical muscle tonicity (Fig. 2). This raised the question about what produced these changes in these upper cervical segments. Was there a correlation between the upper cervical somatic hypertonicity and referral of angina pectoris to the neck and jaw regions (Harrison and Reeves, 1968; Sampson and Cheitlin, 1971; Procacci and Zappi, 1985)? Commonly angina pectoris is referred to the chest and left arm. Surgical sympathectomies that disrupted sympathetic afferent pathways abolished or relieved angina pectoris in 6040% of the patients (Lindgren and Olivercrona, 1947; White and Bland, 1948; White, 1957). After surgical interruption of the sympathetic afferent fibers, ischemic episodes occasionally unmasked pain in the neck and jaw, or patients still experienced pain in these regions. This pain location led to the suggestion that vagal afferent fibers may produce the referred pain associated with myocardial ischemia (Lindgren and Olivercrona, 1947; White and Bland, 1948; Meller and Gebhart, 1992). This information gave us a basis to do a systematic study of vagal and sympathetic afferent input onto spinothalamic tract cells of upper cervical segments. The spinothalamic tract (STT) system serves to transmit noxious visceral information associated with painful perceptions and referred pain from the heart to the thalamus, and then to cortical sites that are important for interpreting the noxious stimuli. A noxious stimulus in the heart or overlying somatic structures generates impulses that are transmitted to the spinal cord in small myelinated and unmyelinated afferent fibers. These fibers penetrate the gray matter and synapse, either directly or indirectly, on cells of origin of the STT (Blair, et al., 1981, 1982; Ammons, et al., 1985a, b; Foreman, 1989; Foreman, 1997). The STT ascends in the anterolateral quadrant and terminates directly in the lateral thalamus and can send collaterals along the way as it ascends into the medial thalamus. The information can ascend to the somatosensory area in the postcentral g y m of the

212

Fig. 2. Graphic representation of segments of somatic dysfunction associated with heart disease. Palpatory examination was used to identify the segments where viscerosomatic reflexes resulting from stimulationof cardiac afferent fibers affected vertebral muscle. The gray histogram represents the usual segments involved with angina pectoris and the black histogram represents the segments less commonly involved with angina pectoris. The abscissa is the cervical (C) and thoracic (T) segments of the spinal cord and the ordinate is the number of patients examined from different studies reported in the osteopathic literature. From Beal(l985) with permission from the American Osteopathic Association.

cortex for localization (Price and Dubner, 1977; Melzak and Wall, 1982). The information also ascends to the association cortex that includes the cingulate g y m , prefrontal cortex and insular (Rosen, et al., 1994, 1996; Vogt, et al., 1994; Hautvast, et al., 1997; Foreman, 1998). The complex organization of the association cortex provides the basis for motivational affective behavior and autonomic adjustments (Melzack and Casey, 1968; Albe-Fessard and Besson, 1973; Casey and Jones, 1978). Authors of other chapters provide important insights about the processing of information in this pathway. Our study of cervical (Cl-C3) STT cells supports the idea that these neurons receive input from activation of thoracic vagal afferent fibers from the heart and cardiopulmonary sympathetic afferent fibers (Chandler, et al., 1995, 1996). Electrical stimulation of the left (ipsilateral to cell) thoracic vagus increases activity of 50% of the STT cells, decreases activity of 5% and does not affect the activity of 45%. Input from the contralateral vagus and from the vagus below the heart is much less

effective than the input from the ipsilateral side. Responses of Cl-C3 STT neurons to electrical stimulation of the cardiopulmonary afferent fibers also increases activity in 55% of the cells, decreased activity in 7% and does not affect activity in the rest. Based on changes in cell activity, vagal stimulation serves as a more potent stimulus than the input from cardiopulmonary afferent fibers (Fig. 3). Noxious somatic input converges onto most of the STT cells that receives one or both of the visceral inputs. We found that 97% of the wide dynamic range neurons and 88% of the high threshold neurons respond to one or both visceral inputs. Somatic receptive fields for these C1-C3 STT neurons are found most commonly in the neck, jaw, ear and upper arm (Fig. 4). Variable locations of the fields may occur because primary afferent fibers from different regions have overlapping terminations. Somatic fields from the ipsilateral neck andor shoulder regions enter the spinal cord at the upper cervical dorsal root ganglia (Hekmatpanah, 1961). The input from the more caudal somatic fields may come from somatic

213

I

0

5

1

0

I

1

s

I

I

2

0

2

3

3

,

0

3

5

Stimulus Intensity (V) Fig. 3. Stimulus-response relationships of thoracic vagus ( 0 ) and cardiopulmonarysympathetic afferent fibers (m)in C1 and C2 STT neurons. Curves show effects of varying stimulus intensity (2-33V, 1 Hz and 0.1 ms, single pulse or 2 pulses 3 ms apart). Data (mean*S.E) of the thoracic vagus ( n = 9 ) and cardiopulmonary sympathetic afferents ( n = 14) are number of action potentials produced in the first peak of penstimulus histograms (50 sweeps).

afferent fibers of C5-C7 segments that innervate cells of the upper cervical segments (Shriver, et al., 1968;Webster and Kempley, 1987). Cardiopulmonary afferent input to the STT cells in the Cl-C3 segments may depend on propriospinal pathways. These cells are a long distance from Tl-T6 dorsal root ganglia where cardiac afferent fibers enter the spinal cord and terminate in the gray matter of these segments (Kuo, et al., 1984; Hopkins and Armour, 1989). Most likely cardiac afferent input enters the spinal cord and synapses on cells with axons ascending in the ventrolateral quadrant and innervate the Cl-C3 neurons (Molenaar and Kuypers, 1975, 1978; Zhang, et al., 1997). In summary, C1-C3 STT neurons provide a neural substrate for referred pain that results from activation of cardiac nociceptors and is perceived in the neck and jaw region. The exact mechanism has not been defined to explain why some patients feel neck and jaw pain with angina and others do not.

11. Relay of intraspinal pathways in upper cervical segments Modulation of visceral and somatic sensory information in neurons of the spinal cord results from

activation of descending pathways originating from supraspinal nuclei such as the periaqueductal gray, parabrachial-subcoerleus nuclei, and nucleus raphe magnus (Beall, et al., 1976; Mayer and Price, 1976; Basbaum and Fields, 1978, 1984; McCreery, et al., 1979; Gerhart, et al., 1981; Willis 1982, 1984, 1988;Ammons, et al., 1984b; Brennan, et al., 1987; Girardot, et al., 1987; Jones, 1992). These pathways are implicated to explain inhibition of lumbosacral STT cells by activation of inputs with noxious stimulation of the contralateral hindlimb, either forelimb, or cardiopulmonary sympathetic or splanchnic afferents (Gerhart, et al., 1981; Dickenson and LeBars, 1983; Foreman, et al., 1988; Hobbs, et al., 1992). Much less attention, however, has been paid to the possibility that modulation of sensory processing can occur within the spinal cord and does not require supraspinal mechanisms. Propriospinal pathways are suggested because transection or anesthetic blockades of the C2-C4 spinal cord do not consistently eliminate the inhibitory effects of noxious stimulation of somatic structures and visceral nerves (Gerhart, et al., 1981; Cadden, et al., 1983). This variable effect most likely occurs because these segments have many descending propriospinal neurons (Burton and Loewy, 1976;Molenaar and Kuypers, 1978;Yezierski, et al., 1980; Svensson, et al., 1985; Flink and Svensson, 1986; Verburgh, et al., 1989, 1990). Thus, disruption of the pathways between the C2 and C4 segments may interfere with processing of propriospinal information in these segments. This section provides evidence to show that propriospinal pathways originating from the upper cervical segments can participate in the modulation of sensory processing between visceral organs. A common assumption is that sensory input from a visceral organ or somatic structure such as the hindlimb excites cells in the segments where the afferent fibers penetrate the spinal cord and has little influence on sensory processing in segments distant from these inputs. This is based on patient studies showing that pain resulting from urinary bladder disease is felt in the suprapubic and perineal regions and is not referred to the chest (Head, 1893; White, 1943). We have shown that this assumption is not correct because stimulation of visceral and somatic structures that innervate the

214

lumbosacral segments suppresses activity of upper thoracic STT neurons. Urinary bladder distension and pinching the hindlimb at noxious intensities

inhibit STT cells in the upper thoracic region (Brennan, et al., 1989). Bladder distension inhibits activity in 80% of the cells and does not affect the

B

C

Fig. 4. Representativeexcitatory somatic receptive fields and locations drawn in black on monkey figurines. Fields were mapped for 10 of 66 C1 and C2 STT neurons. (A) small somatic receptive fields (total=21); (B) medium-sized somatic receptive fields (total= 14); (C) large somatic receptive fields (total of 21). From Chandler, et al., (1996) with permission from the American Physiological Society.

215

remaining cells in the T2 to T5 segments. These same thoracic cells are excited by electrical stimulation of the cardiopulmonary afferent fibers and pinching the skin and muscle of the arm and chest. Differential modulation of thoracic and lumbosacral spinal neuron activity exists because not only does urinary bladder distension inhibit thoracic S l T cell activity but cardiopulmonary afferent stimulation inhibits activity of cells in the lumbosacral segments (Hobbs, et al., 1992; Zhang, et al., 1996). We were interested in the possible neural mechanisms and pathways that could produce excitation in one region of the spinal cord and inhibit neurons that receive visceral input from a distant organ. We wanted to know if inhibition occurred directly between the segments that receive the visceral input and the segments that were inhibited, required a supraspinal pathway or depended on the upper cervical segments. To learn how this inhibition occurred, sequential transections were made in the upper cervical spinal cord while recording extracellular activity from the same lumbosacral STT or dorsal horn cell before and after the transections (Zhang, et al., 1996). In the intact preparation, electrical stimulation of the cardiopulmonary sympathetic afferent fibers inhibits spontaneous activity of 95% of lumbosacral dorsal horn cells and STT cells. The transection was made first at the rostral C1 segments and then at the C4-C6 segments (Fig. 5). Approximately 60min elapsed after the transections and before neuronal responses were recorded. After the rostral C1 transection, inhibition still appeared, although it was reduced, but the vagal effects were eliminated (Fig. 5). These results suggested that supraspinal pathways were not required to produce inhibition of lumbosacral neurons by cardiopulmonary stimulation and noxious forelimb pinch. We still needed to determine if the inhibitory effects depended on a pathway descending from the upper thoracic segments to the lumbosacral spinal cord or required the integrity of the upper cervical segments. Sequential transections at the C4-C6 segments abolish inhibitory effects of cardiopulmonary stimulation on lumbar cells (Fig. 5). Since afferent input from cardiopulmonary afferent fibers to the upper thoracic spinal cord was still intact after the

"1

Intad

C, transection

C&

bsnsedion

OJ

Con Stim

Con Stim

Con Stim

Fig. 5. Responses of lumbosacral cell activity to cardiopulmonary sympathetic afferent stimulation in the intact condition and after sequential upper cervical spinal cord transections in the same animal. Inhibitory responses are seen in the intact preparations ( n = 10) (p0.6). Control (Con) activity was determined before each stimulation (Stim) from rate meter records. Stimulation parameters were 33V, 0.1 ms, 20 Hz for approximately 15 s.

transections, most likely stimulation of cardiopulmonary afferent fibers activated ascending pathways that relayed information to cells in the Cl-C3 segments before inhibition of lumbosacral cell activity could occur. The ascending propriospinal pathway is most likely similar to the one described in the previous section for activation of upper cervical STT cells. Furthermore, this pathway does not seem to require supraspinal pathways because cardiopulmonary stimulation still excites cells after rostral C1 transection. Since inhibition of lumbosacral cell activity by cardiopulmonary afferents required the integrity of the upper cervical spinal cord, we needed to show that descending propriospinal pathways were present to transmit the information to cells of the lumbosacral segments. Neuroanatomical studies show that neurons in the C1 segment project down the spinal cord (Miller, et al., 1998). After Flurogold or WGA-HRP was injected into gray matter of the lumbar spinal cord cells with retrograde tracer were found in the intermediate and ventral gray matter, and the lateral cervical nucleus. In addition, a particular concentration of neurons was found around the central canal. These results showed that some of the Cl-C2 neurons project to the lumbar sacral spinal cord. To determine if cells with

216

descending projections received cardiopulmonary sympathetic afferent input, extracellular recordings were made from Cl-C2 propriospinal neurons (Zhang, et al., 1995). These neurons were activated antidromically by stimulating axons with concentric bipolar electrodes in the ventrolateral white matter of the lower thoracic and lumbar spinal cord. A constant latency, train of stimuli and collision of orthodromic with an antidromic stimulus were used to show that these cells were cells of origin and not synaptically activated. Stimulation of the cardiopulmonary afferents showed a stimulus response relationship of these Cl-C2 neurons. These neurons could also be strongly activated by stimulation of the cervical vagal afferent fibers. Thus, recordings from these propriospinal neurons show that they receive visceral input from the cardiopulmonary region and transmit this information to the lumbar sacral spinal cord. In general, these cells were found in concentrations around the central canal and primarily in the lamina V segments. These recording sites correlated quite well with the anatomical location of C1-C2 cells that project to the lumbar spinal cord (Miller, et al., 1998). The neuroanatomical and electrophysiological studies provide evidence for the existence of a descending propriospinal pathway, but these studies do not show if activation of Cl-C2 propriospinal neurons can inhibit the activity of the lumbosacral cells. To demonstrate this possibility,

30

very small pledgets soaked with glutamate (1M) were placed on the surface of the upper cervical spinal cord to excite cell bodies, but not the axons passing through these segments (Goodchild, et al., 1982; Schramm and Livingstone, 1987; Sandkuhler, et al., 1993). The pledgets were applied to the C 1 -C2 segments while recording spontaneous activity of lumbosacral spinal neurons and their evoked responses to colorectal distension. After control responses were taken, a small glutamate pledget was placed on the spinal cord. Both the spontaneous activity and the evoked activity were markedly reduced after a very short latency (Fig. 6). After the pledget was removed and saline was used to wash off the glutamate, cell activity returned to control. Glutamate placed on the Cl-C2 segments inhibited evoked activity during colorectal distension in about 60% of the lumbosacral cells. Transection of the spinal cord at the rostra1 end of the C1 segment did not abolish spontaneous activity and colorectal distension evoked responses of lumbosacral cells when glutamate was placed on the Cl-C2 segments. Thus, results of these studies show that excitation of cells in the Cl-C2 segments of the spinal cord can produce inhibitory responses of lumbosacral neurons. In summary, afferent information arising from the cardiopulmonary region enters the spinal cord in the T1-T5 segments. STT cells are excited by this stimulus and ascend to the thalamus and other

.

I

408

GI; on C I - 2

off

Fig. 6. Spontaneous activity and evoked responses to colorectal distension recorded from a lumbar spinal neuron before and during placement of a glutamate (Glu) (1 M) pledget on the dorsal surface of the Cl-C2 spinal segments. The horizontal line is duration of colorectal distension with a latex balloon at 80 mmHg. The upward arrow represents the time when Glu was placed on the surface and the downward arrow is the removal of the pledget and flushing the surface of the spinal cord with saline.

217

areas of the supraspinal system. In addition, the cardiopulmonary afferents also excite a pathway that ascends to the cervical segments, primarily C1C3 segments. This ascending pathway appears to ascend bilaterally in the ventrolateral quadrants the spinal cord. After the information is processed, propriospinal neurons descend down the spinal cord and either indirectly or directly inhibit STT cells that receive input from the pelvic viscera, namely the colon, rectum and urinary bladder.

III. Convergence of craniovascularaerent and somatic input in upper cervical segments Nociceptive sensitive intracranial structures, such as the venous sinuses and other cranial blood vessels in the dura mater contribute to pain associated with headaches (Ray and Wolff, 1940; Goadsby, et al., 1991; Goadsby and Zagami, 1991). The pathophysiology of migraine headaches, for example, is commonly believed to depend on the trigeminal innervation of the head (Wolff, 1963; Lance, 1993). The central distribution of trigeminal neurons involved with vascular nociceptors is important for understanding headache and its treatment (Goadsby and Hoskin, 1997). The use of c-fos immunocytochemistry to identify activated cells shows that electrical stimulation of the superior sagittal sinus activates neurons, not only in the most caudal part of the trigeminal nucleus caudalis, but also in the dorsal horn of the Cl-C3 segments of the spinal cord (Kaube, et al., 1993a; Strassman, et al., 1994; Goadsby and Hoskin, 1997). These data show that nociceptive afferent fibers from the superior sagittal sinus innervate upper cervical segments where trigeminal afferent system and the cervical spinal system converge anatomically (Fig. 7). The intracranial vessels produce a diffuse activation of caudal trigeminal nucleus and Cl-C2 neurons associated with visceral pain (Goadsby, 1997). This concept provides an anatomic substrate to explain referral of pain to the back of the head in migraine. Craniovascular afferent innervation of the upper cervical segments also raises the possibility that convergence of afferent inputs from visceral organs such as the heart and gastrointestinal tract might activate upper cervical cells that refer pain to the neck and head. Very little direct information is

available to support this point. An interesting study conducted in children with migraine shows that these children often complain of gastrointestinal pain (Mavromichalis, et al., 1995). Of 31 children with migraines, 29 had an inflammatory lesion that explained their gastrointestinal symptoms. The authors posed the possibility that minute amounts of vasoactive chemical substances may enter the bloodstream and contribute to the migraines. While this is one possibility, another may be that nociceptive sympathetic afferent fibers from the gut activate C1-C3 neurons that receive input from craniovascular structures and from somatic structures. In a recent study, Chandler et al. (1999) show that electrical stimulation of the superior sagittal sinus excited spinothalamic tract cells that also received afferent input from the heart and from somatic afferents of the face and neck. Some very

T

N

C

W

c1

Fig. 7. Representative sections from the cervical (Cl, C2) and from the trigeminal nucleus caudalis (TNC) of the caudal medulla showing Fos expression during electrical stimulation of the superior sagittal sinus. The superior sagittal sinus was stimulated for 1 hr (0.3 Hz,250 ms, 15OV). Cells with fos expression are represented as black dots. Modified from Hoskin and Goadsby (1998) with permission from Academic Press.

218

early evidence suggests that craniovascular afferent fibers and visceral afferent fibers from visceral organs converges on the same neurons. This early evidence supports the idea that some types of headaches might result from diseases of the visceral organs.

Summary The major point of this chapter is that there is evidence to support the idea that cervical headache

might not only result from injured somatic structures in the neck but also occur because of interactions with visceral organs. The complex arrangement of convergent inputs from somatic and visceral afferent fibers and of the propriospinal pathways in the upper cervical segments may create an environment to precipitate such headaches (Fig. 8). It is possible that the soreness experienced in the muscles innervating the neck may not be due to direct injury but may occur as

Fig. 8. Schematic diagram summarizing the somatic, sympathetic, and vagal afferent convergence onto the spinothalamic tract system (-) and the propriospinal system (-.-.-.-) of the upper cervical segments. The ophthalmic branch of the trigeminal nerve enters the spinal tract of V and descends to the cervical segments. Cervical dorsal roots and the trigeminal nerve innervate the somatic receptive fields from the neck and jaw, respectively. The vagal afferent fibers (-.-.-.-) synapse primarily in the nucleus of the tractus solitarius (NTS) and then either directly or indirectly synapse (NTS-Cl-C2) on the spinothalamic tract system and the propriospinal . .) from the heart enter the spinal cord at the upper thoracic segments and then some neurons. The sympathetic afferent fibers (. of them ascend in the ventrolateral quadrant to the upper cervical segments. The propriospinal pathway descends to synapse either directly or directly on spinal neurons of the lumbosacral segments. These neurons were excited by colorectal distension and inhibited by the descending propriospinal pathway. +

+

9 .

219

muscle hyperalgesia that is often associated with visceral pain (Giamberardino, et al., 1993). Much more research is required to understand these complex interactions before patients who suffer pain of cervical headache can be treated satisfactorily.

Acknowledgements I thank Ms Carrie Hulka for typing and organizing the manuscript and Mr. Patrick Whelan for preparing the figures. I also thank Mr. Sam Nourani and Mr. Jerry Jou for reading and discussing the manuscript. This work was supported by the National Institutes of Health grants HL22732, HLS2986, and NS35471 and by the Presbyterian Health Foundation.

References Albe-Fessard,D. and Besson, J. M. (1973) Convergent thalamic and cortical projections. The non-specific system. In: A. Iggo (Ed.), Handbook of Sensory Physiology, Springer-Verlag, Berlin, pp. 489-560. Ammons, W.S., Blair, R.W. and Foreman, R.D. (1984b) Raphe magnus inhibition of primate Tl-T4 spinothalamic cells with cardiopulmonaryvisceral input. Pain, 20: 247-260. Ammons, W.S., Girardot. M.-N. and Foreman, R.D. (1985a) T2-T5 spinothalamic neurons projecting to medial thalamus with viscerosomatic input. J. Neurophysiol., 54: 73-89. Ammons, W.S., Girardot, M.-N. and Foreman, R.D. (1985b) Effects of intracardiac bradykinin on T2-T5 medial spinothalamic cells. Am. J. Physiol., 249: R147-R152. Basbaum,A.I. and Fields, H.L. (1978) Endogenous pain control mechanisms: review and hypothesis. Ann. Neurol., 4: 451462. Basbaum,A.I. and Fields, H.L. (1984) Endogenous pain control systems: brainstem spinal pathways and endorphin circuitry. Ann. Rev. Neurosci., 7: 309-338. Bed, M.C. (1985) Viscerosomaticreflexes: A review. JAOA, 85: 786-801. Beall, J.E., Martin, R.F., Applebaum, A.E. and Willis, W.D. (1976) Inhibition of primate spinothalamic tract neurons by stimulation in the region of the nucleus raphe magnus. Brain Res., 114: 328-333. Blair, R.W., Weber, R.N. and Foreman, R.D. (1981) Characteristics of primate spinothalamic tract neurons receiving viscerosomatic convergent inputs in T3-T5 segments. J. Neurophysiol., 46: 797-8 11. Blair, R.W., Weber, R.N. and Foreman, R.D. (1982) Responses of thoracic spinothalamic neurons to intracardiac injection of bradykinin in the monkey. Circ. Res., 5 1: 83-94. Bogduk, N. (1980) The anatomy of occipital neuralgia. Clin. Exp. Neurol., 17: 167-184.

Bogduk, N. (1982) The clinical anatomy of the cervical dorsal rami. Spine, 7: 319-330. Bogduk, N. and Marsland, A. (1986) On the concept of third occipital headache. J. Neurol. Neurosurg., Psychiatty, 49: 775-780. Bogduk, N. and Marsland, A. (1988) The cervical zygapophyseal joints as a source of neck pain. Spine, 13: 610-617. Bogduk, N. (1997) Headache and the neck. In: P.J. Goadsby and S.D. Silberstein (Eds), Headache, Butterworth-Heinemann, New York, pp. 369-38 1. Brennan, T.J., Oh, U.-T., Girardot, M.-N., Ammons, W.S. and Foreman, R.D. (1987) Inhibition of cardiopulmonary input to thoracic spinothalamic tract cells by stimulation of the subcoeruleus-parabrachial region in the primate. J. Auton. New. Syst., 18: 61-72. Brennan, T.J., Oh, U.-T., Hobbs, S.F., Garrison, D.W. and Foreman, R.D. (1989) Urinary bladder and hindlimb afferent input inhibits activity of primate T2-T5 spinothalamic tract neurons. J. Neurophysiol., 61(3): 573-588. Burns, L. (1928) Qualities distinguishing muscles affected by primary vertebral lesions from those affected by viscerosomatic nerve reflexes. JAOA, 27: 542-545. Burton, H. and Loewy, A.D. (1976) Descending projections from the marginal cell layer and other regions of the monkey spinal cord. Brain Res., 116: 485491. Cadden, S.W., Villanueva, L, Chitour, D. and Lebars, D. (1983) Depression of activities of dorsal horn convergent neurons by propriospinal mechanisms triggered by noxious inputs; comparison with diffuse noxious inhibitory controls (DNIC). Brain Res., 275: 1-1 1. Casey, K.L.and Jones, E.G. (1978) Supraspinal mechanisms: an overview of ascending pathways: brainstem and thalamus. Neurosci. Res. Prog. Bull., 16: 103-1 18. Chandler, M.J., Qin, C., Yuan, Y. and Foreman, R.D. (1999) Convergence of trigeminal input with visceral and phrenic inputs on primate Cl-C2 spinothalamic tract cells. Brain Res., 829(1-2): 204-208. Chandler, M.J., Zhang, J. and Foreman, R.D. (1995) Pericardial injections of inflammatory chemicals excite upper cervical (Cl-C3) spinothalamic tract (STT) cells in monkeys. SOC. Neurosci., 21(1): 260.4 (Abstr.). Chandler, M.J., Zhang, J. and Foreman, R.D. (1996) Vagal, sympathetic and somatic sensory inputs to upper cervical (Cl-C3) spinothalamic tract neurons in monkeys. J. Neurophys., 76(4): 2555-2567. Dickenson, A.H. and Le Bars, D. (1983) Diffuse noxious inhibitory controls (DNIC) involve trigeminothalamic and spinothalamic neurons in the rat. Exp. Brain Res., 49: 174-180. Flink, R. and Svensson, B.A. (1986) Fluorescent doublelabeling study of ascending and descending neurons in the feline lateral cervical nucleus. Exp. Brain Res., 62: 479-485. Foreman, R.D., Hobbs, S.F., Oh, U.-T. and Chandler, M.J. (1988) Differential modulation of thoracic and lumbar spinothalamic tract cell activity during stimulation of cardiopulmonary sympathetic afferent fibers in the primate. A new

220 concept for visceral pain. In: R. Dubner, G.F. Gebhart and M.R. Bond (Eds), Proceedings of the 5th World Congress on Pain, Elsevier, New York, pp. 227-23 1. Foreman, R.D. (1989) Organization of the spinothalamic tract as a relay for cardiopulmonary sympathetic afferent fiber activity. In: D. Ottoman (Ed.), Progress in Sensory Physiology, Springer-Verlag, Heidelberg, pp. 1-5 1. Foreman, R.D. (1997) Organization of visceral input. In: T.L. Yaksh, C. Lynch 111, W.M. Zapol, M. Maze, J.F. Biebuyck and L.J. Saidman (Eds), Anesthesia Biologic Foundations, Lippincott-Raven, Philadelphia. Foreman, R.D. (1999) Mechanisms of cardiac pain. In: J.F. Hoffman (Ed.), Annual Review of Physiology, Vol. 61, 143-167. Gerhart, K.D., Wilcox, T.K., Chung, J.M. and Willis, W.D. (1981) Inhibition of nociceptive and non-nociceptive responses of primate spinothalamic cells by stimulation in medial brainstem. J. Neurophysiol., 45: 121-136. Gerhart, K.D. Yezierski, R.P., Giesler, G.J., Jr. and Willis, W.D. (198 I) Inhibitory receptive fields of primate spinothalamic tract cells. J. Neurophysiol., 46: 1309-1325. Giamberardino, M.A., Valente, R. and Vecchiet, L. (1993) Muscular hyperalgesia of renalhreteral origin. In: D. AlbeFessard, L. Lindblom (Eds) and M.S. Giamberardinao (Assoc. Ed.), New Trends in Referred Pain and Hyperalgesia, Elsevier Sci. Pub. B.V., pp. 149-160. Girardot, M.-N., Brennan, T.J., Ammons, W.S. and Foreman, R.D. (1987) Effects of stimulating the subcoeruleus-parabrachial region on the non-noxious and noxious responses of T2-T4 spinothalamic tract neurons in the primate. Brain Res., 409: 19-30. Goadsby, P.J. (1997) Pathophysiology of migraine: A disease of the brain. In: P.J. Goadsby and S.D. Silberstein (Eds), Headache, Butterworth-Heinemann, New York, pp. 5-24. Goadsby, P.J. and Hoskin, K.L. (1997) The distribution of trigeminovascular afferents in the non-human primate brain Macaca nemestrina: a c-fos immunocytochemical study. J. Anat., 190: 367-375. Goadsby, P.J. and Zagami, A.S. (1991) Stimulation of the superior sagittal sinus increases metabolic activity and blood flow in certain regions of the brainstem and upper cervical spinal cord in the cat. Brain, 114: 1001-101 1. Goadsby, P.J., Zagami, AS. and Lambert, G.A. (1991) Neural processing of craniovascular pain: a synthesis of the central structures involved in migraine. Headache, 3 I: 365-371. Goodchild, A.K., Dampney, R.A. and Bandler, R. (1982) A method for evoking physiological responses by stimulation of cell bodies, but not axons of passage, within localized regions of the central nervous system. J. Neurosci. Methods, 6: 351-363. Harrison, T.R. and Reeves, T.J. (1968) Patterns and causes of chest pain. In: Principles and Problems of Ischemic Heart Diseuse, Year Book Medical Publishers, Chicago, pp. 197-204. Hautvast, R.W., TerHorst, G.J.. DeJong, B.M., DeJongste, M.F., Blanksma, P.K., Paans, A.M. and Korf, J. (1997) Relative changes in regional cerebral blood flow during spinal cord

stimulation in patients with refractory angina pectoris. Eu,: J. Neurosci., 9(6): 1178-1183. Head, H. (1893) On disturbances of sensation with especial reference to the pain of visceral disease. Brain, 16: 1-132. Hekmatpanah, J. (1961) Organization of tactile dermatomes, C1 in cat. J. Neurophysiol,, 264: 582-586. Hobbs, S.F., Oh, U.T., Chandler, M.J., Fu, Q.-G., Bolser, D.C. and Foreman, R.D. (1992) Evidence that C1 and C2 propriospinal neurons mediate the inhibitory effects of viscerosomatic spinal afferent input on primate spinothalamic tract neurons. J. Neurophysiol., 67(4): 852-860. Hopkins, D.A. and Armour, J.A. (1989) Ganglionic distribution of afferent neurons innervating the canine heart and cardiopulmonary nerves. J. Auto New. Syst., 26: 213-222. Hoskin, K.I. and Goadsby, P.J. (1988) Comparison of more and less lipophilic serotonin (SHTIB/lD) agonists in a model of trigeminovascular nociception in cat. Exptl. Neurol., 150(1): 45-5 1. Jones, S.L. (1992) Descending control of nociception. In: A.R. Light (Ed.), The Initial Processing of Pain and Its Descending Control: Spinal and Trigeminal Systems, Karger, New York, pp. 203-295. Kaube, H., Keay, K., Hoskin, K.L., Bandler, R. and Goadsby, P.J. (1993a) Expression of c-fos-like immunoreactivity in the trigeminal nucleus caudalis and high cervical cord following stimulation of the sagittal sinus in the cat. Bruin Res., 629: 95-102. Ken; F.W.L. (1961) Structural relation of the trigeminal spinal tract to upper cervical roots and the solitary nucleus in the cat. Exp. Neurol., 4: 134-148. Kuo, D.C., Oravitz, J.J. and de Groat, W.C. (1984) Tracing of afferent and efferent pathways in the left inferior cardiac nerve in the cat using retrograde and transport of horseradish peroxidase. Brain Res., 321: 111-1 18. Lance, J.W. (1993) Mechanisms and Management of Headache, 5th edn., Butterworth Scientific, London. Lindgren, I. and Olivercrona, H. (1947) The surgical treatment of angina pectoris. J. Neurosurg., 4: 19-39. Mavromichalis, I., Zaramboukas, T. and Giala, M.M. (1995) Migraine of gastrointestinal origin. Eur: J. Pediatl:, 154: 406-410. Mayer, D.J. and Price, D.D. (1976) Central nervous system mechanisms of analgesia. Pain, 2: 379-404. McCreery, D.B., Bloedel, J.R. and Hames, E.G. (1979) Effects of stimulating in raphe nuclei and in reticular formation on response of spinothalamic neurons to mechanical stimuli, J. Neurophysiol., 42(1): 166-182. Meller, S.T., Gebhart, G.F. (1992) A critical review of the afferent pathways and the potential chemical mediators involved in cardiac pain. Neuroscience, 48(3): 501-524. Melzack, R. and Casey, K.L. (1968) Sensory, motivational and central control determinants of pain. In: D.R. Kenshalo (Ed.), The Skin Senses, Charles C. Thomas, Springfield, pp. 423-443. Melzack, R. and Wall, P.D. (1982) The Challenge of Pain, Basic Books, New York.

22 1 Miller, K.E., Douglas, V.D., Richards, A.B., Chandler, M.J. and Foreman, R.D. (1998) Propriospinal neurons in the Cl-C2 spinal segments project to the L5-SI segments of the rat spinal cord. Brain Res. Bul., 47: 43-47. Molenaar, I. and Kuypers, H.G.J.M. (1975) Identification of cells of origin of long fiber connections in the cat = s spinal cord by means of retrograde axonal horseradish technique. Neurosci. Lett., 1: 193-197. Molenaar, I. and Kuypers, H.G.J.M. (1978) Cells of origin of propriospinal fibers and of fibers ascending to supraspinal levels. A HRP study in cat and rhesus monkey. Brain Res., 152: 429450. Pollmann, W., Keidel, M. and Pfaffenrath, V. (1997) Headache and the cervical spine: a critical review. Cephalagia, 17(8): 80 1-8 16. Pfdler, K. and Arvidsson, J. (1988) Central distribution of trigeminal and upper cervical primary afferents in the rat studied by anterograde transport of horseradish peroxidase conjugated to wheat germ agglutinin. J. Comp. Neurol., 268: 91-108. Procacci, P. and Zoppi, M. (1985) Heart Pain. In: P.D. Wall and R. Melzack (Eds), Pain, Churchill Livingstone, Edinburgh, pp. 309-3 18. Price, D.D. and Dubner, R. (1977) Neurons that subserve the sensory-discriminative aspects of pain. Pain, 3: 307-338. Ray, B. S. and Wolff, H.G. (1940) Experimental studies on headache. Pain sensitive structures of the head and their significance in headache. Arch. Surg., 41: 813-856. Rosen, S.D., Paulesu, E., Frith, C.D., Frackowiak, R.S.J., Davies G.J., Jones, T. and Camici, P.G. (1994) Central nervous pathways mediating angina pectoris. Lancet, 344: 147-150. Rosen, S.D., Paulseu, E., Nihoyannopoulos, P., Tousoulis, D., Frackowiak, R.S., Frith, C.D., Jones, T. and Camici, P.G. ( I 996) Silent ischemia as a central problem: regional brain activation compared in silent and painful myocardial ischemia. Ann. Int. Med., 124: 939-949. Sampson, J.J. and Cheitlin, M.D. (1971) Pathophysiology and differential diagnosis of cardiac pain. Prog. Cardiovasc. Dis., 23: 507-531. Sandkuhler, J., Stelzer, B. and Fu, Q.G. (1993) Characteristics of propriospinal modulation of nociceptive lumbar spinal dorsal horn neurons in the cat. 54(4): 957-967. Schramm, L.P. and Livingstone, R.H. (1987) Inhibition of renal nerve sympathetic activity by spinal stimulation in rat. 252(3 R 2): R514R525. Shriver, J.E., Stein, B.M. and Carpenter, M.B. (1968) Central projections of spinal dorsal roots in the monkey: I. Cervical and upper thoracic dorsal roots. Am. J. Anat., 123: 27-74 Strassman, A.M., Mineta, Y. and Vos, B.P. (1994) Distribution of fos-like immunoreactivity in the medullary and upper cervical dorsal horn produced by stimulation of dural blood vessels in the rat. J. Neurosci., 14: 3725-3735. Svensson, B.A., Westman, J. and Rostad, J. (1985) Light and electron microscopic study of neurons in the feline lateral cervical nucleus with a descending projection. Brain Res., 361: 114-124.

Vance, W.H. and Bowker, R.C. (1983) Spinal origins of cardiac afferents from the region of the left anterior descending artery. Brain Res., 258: 96-100. Verburgh, C.A., Kuypers, H.G.J.M., Voogd, J. and Stevens, H.P.J.D. (1989) Spinocerebellar neurons and propriospinal neurons in the cervical spinal cord: a fluorescent double labeling study in the rat and the cat. Exp. Brain Res., 75: 73-82. Verburgh, C.A., Voogd, J., Kuypers, H.G.J.M. and Stevens, H.P.J.D. (1990) Propriospinal neurons with ascending collaterals to the dorsal medulla, the thalamus, and the rectum: a retrograde fluorescent double-labeling study of the cervical cord of the rat. Exp. Brain Res., 80: 577-590. Vogt, B.A., Sikes, R.W. andVogt, L.J. (1994) Anterior cingulate cortex and the medial pain system. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp. 3 13-344. Webster, K.E. and Kemplay, S.K. (1987) Distribution of primary afferent fibres from the forelimb of the rat to the upper cervical spinal cord in relation to the location of spinothalamic neuron populations. Neurosci. Lett., 76: 18-24. White, J.C. (1943) Sensory innervation of the viscera. Res. Publ. Assoc. Res. New. Ment. Dist., 23: 373-390. White, J.C. (1957) Cardiac pain. Anatomic pathways and physiological mechanisms. Circulation, 16: 644-655. White, J.C. and Bland, E.F. (1948) The surgical relief of severe angina pectoris. Methods employed and end results in 83 patients. Medicine, 27: 1-42. Willis, W.D. (1982) Control of nociceptive transmission in the spinal cord. In: D. Ottosson (Ed.), Progress in Sensory Physiology 3, Springer-Verlag, Berlin, pp. 77-106. Willis, W.D. (1984) The raphe-spinal system. In: C.D. Barnes (Ed.), Brainstem Control of Spinal Cord Function, Academic Press, New York, pp. 141-214. Willis, W.D. (1988) Anatomy and physiology of descending control of nociceptive responses of dorsal horn neurons: comprehensive review. In: H.L. Fields and J.M. Besson (Eds), Pain Modulation, Progress in Brain Research, Vol. 77, Elsevier, Amsterdam, pp. 1-29. Wolff, H.G. (1963) Headache and Other Head Pain, Oxford University Press, New York. Yezierski, R.P., Culberson, J.L. and Brown, P.B. (1980) Cells of origin of propriospinal connections to cat lumbosacral gray as determined with horseradish peroxidase. Exp. Neurol., 69: 493-5 12. Zhang, J., Chandler, M.J. and Foreman, R.D. (1995) Cardiopulmonary sympathetic and vagal afferents excite C1-C2 propriospinal cells with long descending axons. SOC. Neurosci., 21(1): 383, #158.12. Zhang, J., Chandler, M.J. and Foreman, R.D. (1996) Thoracic visceral inputs use upper cervical segments to inhibit lumbar spinal neurons in rats. Brain Res., 709: 337-342. Zhang, J., Chandler, M.J., Miller, K.E. and Foreman, R.D. (1997) Cardiopulmonary sympathetic afferent input does not require dorsal column pathways to excite C I -C3 spinal cells in rats. Brain Rex. 771: 25-30.

This Page Intentionally Left Blank

E.A. Mayer and C.B.Saper (Eds.) Pmgress in Brain Research,Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 16

The medial pain system, cingulate cortex, and parallel processing of nociceptive information Brent A. Vogt’i* and Robert W. Sikes2 Cingulum NeumSciences Institute, 101 N. Chestnut St, Winston-Salem, NC 27101 and Department of Physiology and Pharmacology, Wake Forest University School of Medicine, Medical Center Blvd, Winston-Salem, NC 27157, USA Department of Physical Therapy, Northeastern University 6 Robinson Hall, 360 Huntington Ave, Boston, MA 02115, USA

The medial pain system Nociceptive information in the cerebral cortex is thought to be processed according to discriminative properties, including localization and intensity, and affective associations. Stimulus localization is assessed mainly in somatosensory and posterior parietal cortices, while affective responses are processed in limbic regions. It has been known for four decades that ablations of anterior cingulate cortex (ACC) and its underlying white matter, the cingulum bundle, reduce or abolish affective responses to noxious stimuli, while sensory localization remains intact (Foltz and White, 1962; Ballantine et al., 1967). Alterations in pain thresholds have been reported following damage to the insula (Berthier et al., 1988; Greenspan and Winfield, 1992) and early functional imaging studies demonstrated activation of the anterior cingulate and insular cortices, thalamus, and periaqueductal gray, in addition to somatosensory areas, with noxious stimuli applied to the skin (Jones et al., 1991; Talbot et al., 1991). Electrophysiological recording documented the nociceptive properties of neurons in the midline and intralaminar thalamic nuclei (Casey, 1966; Dong et al., 1978) and the central nucleus of the amygdala (Bernard et al., *Corresponding author. Tel.: 336-716-8588; Fax: 336-716-8501; e-mail: [email protected]

1992) which also has a role in conditioned analgesia (Watkins et al., 1993). Thus, it was proposed that “cortical and subcortical structures engaged in affect and motivation form the medial pain system. Structures of the medial system have little or no somatotopic organization and are comprised of the medial and intralaminar thalamic nuclei, the amygdala, limbic cortical areas, and projections from these latter areas to nociception regulating centers like the periaqueductal gray” (Vogt et al., 1993). Processing of noxious information in the medial pain system is a problem of mind and the ability to predict or mentalize the outcome of life-threatening events

Of what value is nociceptive information if it cannot be used to predict and avoid painful and potentially life-threatening outcomes? The primary contribution of the medial pain system is to predict and avoid noxious stimuli rather than localization and evaluation of the stimulus on the body. The features of noxious stimuli assist the organism in establishing the consequences of such stimuli. As discussed in Chapter 2 of this volume, a confluence of medial frontal and anterior cingulate areas are engaged in mental activity including inductive reasoning. Inductive reasoning is likely employed in mental processes to evaluate the probability that an event will be painful.

224

Although ACC is often treated as a single functional entity, it is a structurally and functionally heterogeneous region. The part of ACC that is repeatedly activated in pain imaging studies is the midrostrocaudal or midcingulate region. This region is not solely activated by noxious stimuli and there are other divisions of ACC that may contribute to processing of noxious information. In addition to activation with noxious stimuli, midcingulate cortex has been activated by many cognitively challenging tasks which are best referred to as response selection (Corbetta et al., 1991; Bush et al., 1998). Electrical stimulation studies further suggest a role in early premotor planning (Bancaud and Talairach, 1992), and positron emission tomography (PET) studies show that midcingulate cortex is involved in motor imagery (Decety et al., 1994). Therefore, midcingulate cortex is not a pain center and may have a lesser role in affect and a greater involvement in motivation or goal orientation (Vogt et al., 1997). It is also possible that during functional imaging studies of nociception the midcingulate cortex is engaged in anticipation and inhibition of movement during the task. An overview of the structural and functional organization of the entire human cingulate gyrus is provided here in order to evaluate the contributions of different parts to pain processing. The primary view derived during this undertaking and from a previous PET study (Vogt et al., 1995b) is that “perigenual area 24 is involved in affect and midcingulate area 24’ is involved in selection among motivationally relevant response options or goal orientation. This selection process includes motor imagery and an assessment of outcomes and avoiding tissue damage equates to self preservation.”

Overview of human cingulate cortex Cingulate cortex forms a cingulum around the genual, dorsal, and splenial parts of the corpus callosum. The human cingulate sulci can form single or double parallel patterns which make averaging across cases difficult in functional imaging studies. The multiple sulcal patterns are related in turn to different depths with the single cingulate sulcus having the greatest depth of more than 1.5

cm. In order to expedite conversations about cingulate cortex, we routinely divide the cingulate gyrus into four regions that have unique cytoarchitectures, connections, and functions. As shown in Fig. 1, the four regions and associated areas are as follows: perigenual areas 25, 24, and 32; midcingulate areas 24’ and 32’; posterior areas 23 and 31; retrosplenial areas 29 and 30. The rationale for distinguishing between anterior area 24 and posterior area 24’ has been discussed elsewhere (Vogt, 1993). As functional imaging studies differentiate between rostral and caudal parts of ACC, the confusing nomenclature is now arising which identifies an anterior anterior cingulate and a posterior anterior cingulate cortex. Since early distinctions between anterior and posterior cingulate cortices are not adequate for either structural or functional studies, the designation of a midcingulate region provides a simple regional designation to avoid such concepts as a posterior anterior cingulate cortex. The following comments on each of these regions is oriented toward issues relating to pain processing even in instances where such activity does not have a prominent role such as in posterior cingulate cortex. Perigenual cortex The perigenual areas are associated with affective experience and autonomic regulation. Area 25 is a visceromotor control cortex and it has projections to the nucleus of the solitary tract, the dorsal motor nucleus of the vagus and the thoracic intermediolateral cell column as reviewed by Neafsey et al. (1993). Visceromotor changes are the most consistent responses evoked by electrical stimulation of areas 25 and 24 and include increases and decreases in respiratory and cardiac rate and blood pressure, mydriasis, piloerection, and facial flushing (Escobedo et al., 1973; Talairach et al., 1973). Visceral responses include nausea, vomiting, epigastric sensation, salivation, or bowel and bladder evacuation (Pool and Ransohoff, 1949; Lewin and Whitty, 1960; Meyer et al., 1973). This region has elevated blood flow when healthy women recall sad experiences (George et al., 1995) and area 24 has elevated blood flow during recognition of faces that express emotional content (George et al., 1993).

225

Fig. 1. Distribution of regions and cytoarchitectural areas on the medial surface of the human brain in a flat map format where the major sulci have been opened (Vogt et al., 1995a). The surface borders of each sulcus are shown, although the fundi are not marked in order to simplify the diagram. In A, the approximate topography of four regions are demarcated with different grades of shading within cingulate cortex that is outlined with the dashed line. In B, the detailed borders of each cytoarchitectural area are shown with doted lines. The dash-dot line is the fundus of the callosal sulcus (CaS). Notice that the cingulate motor areas are contained in the cingulate sulci and include areas Dc‘, 2 4 , and 23c. The cingulate sulcus is segmented (CSl, CS2) and the splenial sulcus (SpS; also subparietal sulcus), and the paracentral sulcus (PCS) are shown.

Electrical stimulation in this region in monkey is associated with vocalizations that reflect internal states such as fear or happiness (Vogt and Barbas, 1988). Finally, areas 25 and 24 have prominent projections into the periaqueductal gray and these may be associated with vocalizations, affective defense and other behaviors associated with flight and immobility (Siege1 and Brutus, 1990; Bander et al., 1991; Holstege, 1992). Electrical stimulation to the dorsal perigenual area 24 produces fear, pleasure, and agitation (Meyer et al., 1973).The most frequent response in

this series (11 of 75 cases) was intense or ‘overwhelming’ fear including one individual who reported the feeling that death was imminent. Pivotal to understanding the differences between anterior area 24 and posterior area 24‘ are the electrical stimulation studies of Bancaud and Talairach (1992). Stimulation of area 24 produced this response, “I was afraid and my heart started to beat”, whereas stimulation of midcingulate area 24’ evoked the report that, “I felt something, as though I were going to leave”. The former experience is one of almost pure fear, while the latter is one of an

226

early planning event related to motivation and goal orientation. In this context, the claim that ACC is involved in affect during pain processing appears to require activation of areas 25 andor 24. Midcingulate cortex Many different cognitive and behavioral paradigms produce elevated blood flow in ACC. However, since the entire perigenual region is often not activated, confusion has arisen as to how midcingulate cortex relates to the ACC activation sites. Midcingulate cortex is defined by topographical, connectional and cytoarchitectural criteria and represents one of the main expansions of human cingulate cortex when compared to that in the monkey brain (Vogt, 1993). Midcingulate cortex is the caudal division of ACC and the prime designation (area 24’; 32’) provides a means of identifying the essential link of this region with ACC. While the perigenual region has an output from area 25 to autonomic centers, midcingulate cortex has outputs to the skeletomotor system. The depths of the cingulate sulcus contains the cingulate motor areas that have gigantopyramidal neurons in layer V (Braak, 1976),project to the spinal cord (Biber et al., 1978; Dum and Strick, 1993), control somatotopically organized movements (Luppino et al., 1991), project to motor cortex (Morecraft and Van Hoesen, 1992; Van Hoesen et el., 1993; Nimchinsky et al., 1996) and contain neurons with premotor discharge properties (Shima et al., 1991). Electrical stimulation of midcingulate cortex in human elicits gestures such as touching, kneading, rubbing or pressing the fingers or hands together, and lip puckering or sucking (Escobedo et al., 1973; Meyer et al., 1973; Talairach et al., 1973). These movements are often adapted to the environment, they can be modified with sensory stimuli, and, at times, resisted. Motivation or goal orientation appears to summarize the functional significance of midcingulate cortex. It does not require overt movement because this region is engaged in cognitively challenging tasks as reported by Corbetta et al. (1991). Although a button press is a consistent feature of all parts of their task and movement is involved in most other imaging paradigms, it is assumed that

activation associated with consistent movements is subtracted from the final measurements of blood flow. The emotional Stroop which employs words with emotional content (George et al., 1994) and recognition of faces with emotional content (George et al., 1993) elevate blood flow in midcingulate cortex. The notion of goal orientation and motivational relevance as early premotor planning is confirmed by Decety et al. (1994) who show that motor imagery also activates midcingulate cortex. The role of the perigenual and midcingulate regions in affect and motivation, respectively, has profound consequences for information processing in the medial pain system. An explicit test of the functional dissociation of these two regions has been made recently in the same subjects. While the counting Stroop task activated the midcingulate cortex (Bush et al., 1998), performance of an emotional counting Stroop task activated the perigenual region (Whalen et al., 1998). The former study also provides an excellent summary of previous PET work that validates the hypothesis that these two regions are functionally distinct. Posterior and retrosplenial regions The retrosplenial areas do not appear on the surface of the posterior cingulate g y m but are in the depths of the callosal sulcus and continue in the dorsal bank of the calcarine sulcus. Most importantly, the retrosplenial and posterior cingulate cortices have no documented role in affect or motivation. Electrical stimulation studies in conscious human subjects routinely fail to produce reports of emotional experiences, fear or other affective responses, and no autonomic responses that might be associated with such events. Indeed, Olson et al. (1993; 1996) report on the large visual field activation of posterior cingulate neurons and their discharges in relation to the orbital position of the eye. Hirono et al. (1998) report that glucose hypometabolism is related to disorientation to place and time in Alzheimer’s disease. Thus, posterior cingulate and retrosplenial cortices are involved in visuospatial orientation. In terms of nociception, Hsieh et al. (1995) reported elevated blood flow in posterior cingulate

227

areas 23 and 31 in patients with chronic neuropathic pain. In contrast, studies of acute nociceptive responses showed reduced blood flow in these same areas (Coghill et al., 1994; Vogt et al., 1995b). Since posterior cingulate cortex has no known role in affect or motivation, elevated activity here in patients with neuropathic pain indicates a pathological condition. Although posterior cingulate cortex has few connections with midcingulate cortex, it does have connections with perigenual cingulate areas (Vogt and Pandya, 1987). Therefore, chronic pain may cause a pathological enhancement of neuronal activity associated with some intracingulate connections.

Nociceptive properties of rat and rabbit anterior cingulate neurons Nociceptive responses in ACC are not somatotopically organized in rat (Yamamura et al., 1996; Hsu and Shyu, 1997) or rabbit (Sikes and Vogt, 1992). The ACC neurons have large nociceptive fields that usually include the entire animal and are often polymodal nociceptors in that they respond to both noxious heat and noxious pressure. The only innocuous stimulus that activates these neurons is tap stimuli. The large receptive fields and activation by innocuous tap are similar to nociceptive neurons in the midline and intralaminar thalamic nuclei (Casey, 1966; Dong et al., 1978) and this suggests that ACC nociceptive activity arises from these nuclei. In support of this hypothesis, lidocaine injections into this region block nociceptive activity in ACC and complete disconnection of ACC from somatosensory cortex does not abolish nociceptive responses (Sikes and Vogt, 1992). Figure 2 shows the location of nociceptive neurons in ACC, the effects of lidocaine in the midline thalamus focussed primarily on the parafascicular nucleus, and failure of a knife cut lesion to block nociceptive responses generated with transcutaneous electrical stimulation (TCES) in area 24. The nociceptive properties of ACC neurons in rat may be somewhat different than those in rabbit, although sampling biases associated with each preparation could also play a role in the observed differences. In an elegant study of the morphological and receptive field properties of ACC

(Yamamura et al., 1996), most nociceptive neurons were in layer V rather than layer 111as in the rabbit. Also, these neurons were located more anteriorally to include area 32. In another study of nociceptive neurons, they were displaced more laterally in medial agranular motor cortex (Hsu and Shyu, 1997). Although nociceptive neurons were also in medial agranular cortex (area 8) in the rabbit as shown in Fig. 2, most were in area 24b. In view of the more dorsal localization of nociceptive units in the rat medial cortex, it is interesting that Pastoriza et al. (1996) were able to impair performance on the hot plate test in rats with lesions mainly in medial agranular cortex without effects on formalin and tail-flick tests. Although these lesions had an involvement of area 24b, they appear to be centered more dorsally and suggest a direct role for medial agranular motor cortex in nociception in this species. Medial agranular cortex has substantial connections with most of cingulate cortex, and the primary and secondary visual and somatosensory cortices (Reep et al., 1984, 1990). In the context of the broad topographical origin of cortical connections, it appears that medial agranular cortex in the rat is engaged in orienting to and avoidance responses of acute noxious stimuli. The rat and rabbit studies of nociception provide insight into the organization of pain processing in ACC in all mammals. First, the origin of nociceptive responses in ACC may be the medial and intralaminar thalamic nuclei. In addition to thalamic lidocaine injections blocking ACC nociceptive responses in the rabbit (Fig. 2), Hsu and Shyu (1997) antidromically activated a large proportion of nociceptive neurons in ACC with electrical stimulation in the intralaminar thalamic nuclei. Second, nociceptive responses in ACC are independent of those in somatosensory cortex. Although no known connections are made between these two cortical regions (Vogt et al., 1986), medial agranular motor cortex in the rat receives significant input from somatosensory cortex and it projects heavily to area 24 (Reep et al., 1990). In order to disconnect secondary connections between somatosensory, medial agranular, and anterior cingulate cortices, scalpel blade lesions that transected potential interactions between these regions were made as shown in Fig. 2. Severing connections

228 A. Noclceptive Neurons

.

AREA24b

AREA 24a

B. Lldocane Block of Midllne Thalamus 15-

BEFORE

10-

CANNULA TRACT

: a

6-

E

0 v)

1.

S W

J

rn

10

z

220 SECONDS AFTER ldJ1ll

,I

I-

5-

mm 1

TIME ( 1 secldivision)

C. Knife-cut Cortical Seperatlon

IPSILATERAL

CONTRALATERAL

TCES

I,. 4

1360 SECONDS AFTER

1 secldlvi~lon

229

between the medial and lateral systems at the cortical level failed to abolish nociceptive activity in ACC. This finding suggests that the medial and lateral pain systems are activated in parallel and that activity in each does not depend on the other. Cognitive processes associated with sensory localization and coding, therefore, may be distinct from those involved in affect and motivation. Finally, most nociceptive neurons in the anesthetized rabbit ACC are in area 24b, not midcingulate area 24b’. Hence, discharges observed in this preparation are likely associated with affect rather than motivation, response selection, or motor imagery.

Functional imaging of nociceptive responses in anterior cingulate cortex Numerous studies of human nociception report activation of midcingulate cortex with noxious stimuli to all parts of the body (Jones et al., 1991; Coghill et al., 1994; Casey et al., 1994; Silverman et al., 1997; Svensson et al., 1997). A study of activation sites in single cases with either noxious heat or Stroop interference suggested that separate modules are activated within midcingulate cortex by each task (Derbyshire et al., 1998). In addition to this latter finding, there are many reasons why this region is not a pain center and this section seeks to expand on a unified perspective of the function of midcingulate cortex to account for its involvement in a range of tasks. A study of cerebral blood flow in individuals during noxious heat stimulation to the back of the hand elevated flow in perigenual and midcingulate cortices (Vogt et al., 1995b). Figure 3 shows three cases that were selected to emphasize the range of variation in individual activation sites. Although each of the three cases has elevated flow in both perigenual and midcingulate cortices, the statistical analysis showing two significant sites dorsal to the

-----

corpus callosum understates the involvement of cortex rostra1 and ventral to the genu of the corpus callosum. In addition, some cases had small activation sites in posterior cingulate cortex. Do these latter activation sites indicate that in some individuals pain processing occurs in areas 23 and 3 1? Probably not for the many reasons cited earlier. Rather, it is likely that some individuals employ different coping or mentalizing strategies when asked to lay flat and still during noxious stimulation to the back of their hand. Among the strategies might be altering their internal orientation in space such that posterior areas are activated in some cases. Finally, there are substantial expanses of cortex adjacent to the cingulate gyms including areas 6, 8, and 9 rostrally and posterior cingulate areas 29, 30, 23, and 31 that have reduced blood flow. This latter observation emphasizes that posterior areas are not directly engaged in pain processing. Pain processing in this particular paradigm, therefore, may involve the following events based on these findings and the theoretical orientation enumerated above. There is an elevation of activity in perigenual cortex associated with affect and an elevation of activity in midcingulate cortex associated with mental imagery and inhibition of motor activity due to the instruction set provided before scanning. There is a massive reduction in blood flow in rostromedial prefrontal areas as the subject attempts to block anguish produced by the noxious stimulation and orient mental activity to implement the instruction set provided by the examiner. Finally, reductions in blood flow in posterior cingulate cortex occur as subjects suppress their orientation to place in the scanner. Although there are few reports of involvement of perigenual cortex in statistical group analyses of noxious stimulation (Vogt et al., 1995b; Derbyshire

Fig. 2. This is a compilation of findings in the rabbit relating the distribution of nociceptive neurons and the source of such activity in the telecephalon (Sikes and Vogt, 1992). A. The primary location of nociceptive neurons is in area 24b. B. Activity in these units can be almost completely blocked with lidocaine injections into the medial and intralaminar thalamic nuclei. This particular injection into the midline blocked activity in the parafascicular nucleus as documented by the blockade of neuronal activity at the star. C. Complete severance of any possible connections between the medial and lateral systems with a knife-cut failed to block neuronal discharges in area 24b. In this instance noxious transcutaneous electrical stimulation (TCES) to both sides of the rabbit body were intact. The neuron from which the discharges were recorded is marked in the transverse section with a star. CC, corpus callosum, SS, splenial sulcus.

230 rCBF Elevatlons

rCBF Reductions

Fig. 3. Examples of individual cases from a study of regional cerebral blood flow (CBF) in individuals during noxious heat stimulation to the back of the hand following subtraction of control values associated with innocuous heat to the same part of the hand (Vogt et al., 1995b). These cases were selected to demonstrate the variability in responses within each of the regions of the cingulate cortex. Although there is substantial variability, the perigenual and midcingulate regions are activated in all cases, while there are always reductions in rCBF in posterior cingulate and dorsomedial frontal areas.

et al., 1997; Silverman et al., 1997), the thermal grill experiments by Craig et al. (1996) provide an interesting perspective on pain processing in ACC. The thermal grill employs a grid of innocuous hot and cold stimuli which together produce the perception of painful burning. Since this illusion may be explained by subcortical mechanisms, the information provided to ACC appears to be noxious. Also, only noxious stimuli elevated blood flow in ACC and activity in this region is not selecting among innocuous and noxious stimuli, but appears to be treating the thermal grill as a noxious stimulus. Although the anterior insula was activated by all innocuous, noxious, and thermal grill stimuli, ACC was only activated by noxious stimuli and the thermal grill. The thermal grill activation site in horizontal sections (z = 38 mm) is located in the dorsal part of the perigenual region. These observations support the conclusion that perigenual cingulate cortex is responsible for the affective component in processing nociceptive information.

Source of nociceptive input to anterior cingulate cortex Since the observation that cingulotomy and cingulumotomy lesions alleviate responses to noxious

stimuli without disrupting sensory localization (Ballantine et al., 1967; Foltz and White, 1968), questions arose about the source of such a signal. Recent neuroanatomical, single cell electrophysiology, and functional imaging studies lead to a number of explicit criteria: (a) The source should have neurons with nociceptive properties similar to those in ACC. (b) The source should have direct connections with ACC and disconnection of the source from ACC should abolish nociceptive activity in ACC. (c) There should be joint activation of the source and ACC in human functional imaging studies and both regions should share similar parameters of activation. Midcingulate responses to noxious stimuli to all parts of the body have been reported, while innocuous and just painful stimuli fail to consistently activate this region (Jones et al., 1991; Talbot et al., 1991; Coghill et al., 1994; Casey et al., 1994; Silverman et al., 1997; Svensson et al., 1997). Since there is some variability in the extent of ACC involvement and a consensus has not developed about the behavioral conditions that consistently activate midcingulate cortex, strict application cannot be made of the above criteria for a source of nociceptive inputs to ACC. Some possibilities, however, can be tentatively eliminated.

23 1

Regions that often coactivate with ACC and are candidates for the source of nociceptive input to ACC include the following: SI, SII, areas 7b and 40, anterior insula, prefrontal areas 10 and 46, thalamus, and the periaqueductal gray. The periaqueductal gray can be discounted because it does not project to ACC but only receives input from ACC (Hardy and Liechnetz, 1981). Although SI, SII, and area 7b do not project directly to ACC (Vogt and Pandya, 1987), human area 40 in inferior parietal cortex could have such a projection but there may not be a cortical corollary for this region for experimental analysis in the monkey. Although Svensson et al. (1997) observed a positive correlation between blood flow activation in area 40 and that in midcingulate cortex, we observed a reduction in blood flow in area 40 during cutaneous noxious stimulation (Vogt et al., 1995b) and others have not observed changes here (Coghill et al., 1994). It appears unlikely, therefore, that area 40 drives the ACC responses to noxious stimuli. In addition, acute rabbit experiments with complete disconnection of lateral somatosensory areas and ACC failed to abolish nociceptive responses in ACC (Fig. 2C) suggesting a general independence of ACC responses from those in parietal cortex. Finally, it is unlikely that prefrontal areas are uniquely engaged in pain processing but, rather, coding for working memory and response inhibition (Goldman-Rakic, 1987; Fuster, 1995). Since area 46 is heavily and reciprocally connected with midcingulate cortex (Baleydier and Mauguidre, 1980; Vogt and Pandya, 1987), it is possible that prefrontal activation during noxious stimulation is due to its ACC connections. Alternatively, areas 46 and 24 may share a common nociceptive input from the thalamus. The anterior insula is a candidate for driving nociceptive responses in ACC because there are connections between these regions (Mesulam and Mufson, 1982; Vogt and Pandya, 1987). These connections, however, tend to be most pronounced to cortex in the depths of the cingulate sulcus rather than on the gyral surface where the greatest pain activation is located and the anterior insula appears to have different stimulus response properties. Elevated blood flow in the anterior insula and SII but not ACC are correlated during noxious cuta-

neous stimulation (Svensson et al., 1997), hypontic alterations in stimulus unpleasantness influence ACC responses but not those in the insula (Rainville et al., 1997), and the thermal grill and noxious cold and heat, and innocuous stimuli activated the anterior insula, while ACC was the only region activated uniquely by noxious stimuli (Craig et al., 1996). To the extent that both the anterior insula and ACC are engaged in pain processing, they appear to operate independently as part of a parallel and distributed network. The midline and intralaminar thalamic nuclei are the only known sites of nociceptive neurons that also project to ACC. The nuclei with nociceptive neurons, inputs from the spinothalamic tract, and telencephalic projections include the following: parafascicular, centrolateral, submedial, paralaminar, paraventricular, paratenial, ventromedial, and reunions. The nociceptive properties of the midline and intralaminar thalamic neurons were assessed in the single unit studies of Casey (1966) and Dong et al. (1978). These studies showed that neurons in this part of the thalamus have large receptive fields and respond mainly to noxious stimuli such as intense mechanical pressure and high temperatures, although innocuous tap stimuli would also drive these units. No somatotopic organization of neurons was found in these nuclei. The midline and intralaminar thalamic nuclei project to areas 24 and 24‘ (Vogt et al., 1979, 1987), the anterior insula (Mufson and Mesulam, 1984), and prefrontal cortex (Barbas et al., 1991). Thus, these thalamic nuclei are the only likely source of nociceptive inputs to ACC.

Midlinehntralaminarthalamic projections are diffuse but not ‘non-specific’ Although the midline and intralaminar thalamic nuclei are the most likely candidates for transmitting nociceptive information to ACC, there is a strong bias in neuroscience against such an idea. How is it possible that the ‘non-specific’ thalamus and its projections to the cerebral cortex provide information that activates ACC? This question poses the key pitfall; although these nuclei have diffuse projections, they are not non-specific. Neither does the diffuse projection necessarily

232

imply non-specific functions such as stress and anxiety as is often presumed. Indeed, the somatic stimuli that activate the midlinehtralaminar thalamus are mainly noxious with tap being the only innocuous stimulus to activate these neurons. Thus, this thalamocortical system can have a specific role in processing noxious stimuli as well as some more general contribution to arousal without invoking the extreme view that these cortical inputs provide only ‘non-specific’driving. There is also no reason to expect that, because a projection is diffuse, it necessarily has the same influence on driving all neurons in all areas to which it projects. Driving of diffuse projections to multiple cortical areas could activate multiple parallel circuits to perform a number of different but related information processing tasks such as working memory in prefrontal cortex, cardiovascular regulation by the central nucleus of the amygdala, affective assessment and motor planning by ACC, and localization and coding by the anterior insula and parietal somatosensory areas. There is no reason to expect that nociceptive responses in these many areas will have the same properties, latencies, durations, and modulations by other sensory or contextual or memory events. Thus, the diffuse projections of the midline and intralaminar thalamic nuclei engage many specific information processing subroutines within the cerebral cortex.

Parallel processing and the problem of mentally uniform pain perceptions In view of the fundamentally different roles that the medial and lateral pain systems have in affect/ motivation and stimulus localization, respectively, there is little reason to expect that these systems share significant interconnections. In fact, such crosstalk between a system with no localization information and another finely tuned to localize and assess noxious stimuli could be counterproductive. The case has been made above for proposing that the medial and lateral systems are activated in parallel and that there are few direct links between them. These observations, however, raise another key problem in cortical pain processing; the ‘binding problem’.

The independence of the lateral and medial systems raises the binding problem for producing a unified mental perception in a parallel and distributed network as discussed in Chapter 2 of this volume. The binding problem is particularly difficult in the present context because there are no known direct connections between the two systems and they do not have a final common higher order association cortex. Although the ACC and anterior insula receive common connections from the medial and intralaminar thalamic nuclei, these joint inputs may simply assure that multiple regions are coactivated in a temporally coherent fashion. It is not clear how having a common input from the thalamus solves the binding problem. Thus, the mechanism by which paidaffect is so precisely and tightly linked to localization of the noxious stimulus to a specific part of the body remains a mystery.

References Baleydier, C. and MauguiBre, F. (1980) The duality of the cingulate gyrus in monkey. Neuroanatomical study and functional hypotheses. Brain, 103: 525-554. Ballantine, H.T., Cassidy, W.L., Flanagan, N.B. and Marino, R. Jr. (1967) Stereotaxic anterior cingulotomy for neuropsychiatric illness and intractable pain. J. Neumsurg., 26: 488495. Bancaud, J. and Talairach, J. (1992) Clinical semiology of frontal lobe seizures. Adv. Neurol., 57: 3-58. Bandler, R., Carrive, P. and Zhang, S.P. (1991) Integration of somatic and autonomic reactions within the midbrain periaqueductal gray: viscerotopic, somatotopic and functional organization. Pmg. Brain Res., 87: 269-305. Barbas, H., Henion, T.H.H. and Dermon, C.R. (1991) Diverse thalamic projections to the prefrontal cortex in the rhesus monkey. J. Comp. Neurol., 313: 65-94. Bernard, J.F., Huang, G.F.and Besson, J.M. (1992) Nucleus centralis of the amygdala and the globus pallidus ventralis: electrophysiological evidence for an involvement in pain processes. J. Neurophysiol., 68: 55 1-569. Berthier, M., Starkstein,S. and Leiguarda, R. (1988) Asymbolia for pain: a sensory-limbic disconnection syndrome. Ann. Neurol., 24: 41-49. Biber, M.P., Kneisley, L.W. and LaVail, J.H. (1978) Cortical neurons projecting to the cervical and lumber enlargements of the spinal cord in young and adult monkeys. Exp. Neurol., 59: 492-508. Braak, H. (1976) A primitive gigantopyramidal field buried in the depth of the cingulate sulcus of the human brain. Bruin Res., 109: 219-233. Bush, G., Whalen, P.J., Rosen, B.R., Jenike, M.A., McLnemey, S.C. and Rauch, S.L. (1998) The counting Stroop: An

233 interference task specialized for functional neuroimaging Validation study with functional MFU. Hum.Brain Mapp., 6: 27@2-82. Casey, K.L. (1966) Unit analysis of nociceptive mechanisms in the thalamus of the awake squirrel monkey. J. Neurophysiol., 29: 127-750. Casey, K.L., Minoshima, S., Berger, K.L., Koeppe, R.A., Morrow, T.J. and Frey, K.A. (1994) Positron emission tomographic analysis of cerebral structures activated specifically by repetitive noxious heat stimuli. J. Neurophysiol., 71: 802-807. Coghill, R.C., Talbot, J.D., Evans, A.C., Meyer, E., Gjedde, A., Bushnell, M.C. and Duncan, G.H. (1994) Distributed processing of pain and vibration by the human brain. J. Neurosci., 14: 40954108. Corbetta, M., Miezin, EM., Dobmeyer, S., Shulman, G.L. and Peterson, S.E. (1991) Selective and divided attention during visual discrimination of shape, color, and speed: functional anatomy by positron emission tomography. J. Neurosci., 11: 2383-2402. Craig, A.D., Reiman, E.M., Evans, A. and Bushnell, M.C. (1996) Functional imaging of an illusion of pain. Nature, 384: 258-260. Decety, J., Perani, D., Jeannerod, M., Bettinardi, V., Tadary, B., Woods, R., Mazziotta, J.C. and Fazio, F. (1994) Mapping motor representations with positron emission tomography. Nature, 371: 600-602. Derbyshire, S.W.G., Jones, A.K.P., Gyulani, F., Clark, S., Townsend, D. and Firestone, L.L. (1997) Pain processing during three levels of noxious stimulation produces differential patterns of central activity. Pain, 73: 4 3 1 4 5 . Derbyshire, S.W.G., Vogt, B.A. and Jones, A.K.P. (1998) Pain and Stroop interference tasks activate separate processing modules in anterior cingulate cortex. Exp. Brain Res., 118: 52-60. Dong, W.K., Ryu, H. and Wagman, LH. (1978) Nociceptive responses of neurons in medial thalamus and their relationship to spinothalamic pathways. J. Neurophysiol., 41: 1592-1613. Durn, R.P. and Strick, P.L. (1993) Cingulate motor areas. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhiiuser, Boston, pp. 415-441.

Escobedo, F., Femandez-Guardiola, A. and Solis, G. (1973) Chronic stimulation of the cingulum in humans with behavior disorders. In: L.V. Laitinen and K.E. Livingston (Eds), Surgical Approaches in Psychiatry, MTP, Lancaster (UK), pp. 65-68. Foltz, E.L. and White, L.E. (1962) Pain ‘relief‘ by frontal cingulumotomy. J. Neurosurg., 19: 89-100. Fuster, J.M. (1995) Memory and planning: Two temporal perspectives of frontal lobe function. In: H.H. Jasper, S. Riggio and R.S. Goldman-Rakic (Eds), Advances in Neurology, Vol. 66: Epilepsy and the Functional Anatomy of the Frontal Lobe, Raven Press, New York, pp. 9-20. George, M.S., Ketter, T.A., Gill, D.S., Haxby, J.V., Ungerleider, L.G. and Herscovitch, P. (1993) Brain regions involved in

recognizing facial emotion or identity: an oxygen-15 PET study. J. Neuropsychiatry Clin. Neurosci., 5: 384-394. George, M.S., Ketter, T.A., Parekh, P.I., Rosinksy, N., Ring, H. and Casey, B.J. (1994) Regional brain activity when selecting a response despite interference: an H,”O PET study of the Stroop and an emotional Stroop. Hum. Brain Mapp., 1: 194-209. George, M.S., Ketter, T.A., Parekh, P.I., Horwitz, B., Hershcovitch, P. and Post, R.M. (1995) Brain activity during transient sadness and happiness in healthy women. Am. J. Psychiatry, 152: 341-351. Goldman-Rakic, P.S. (1987) Circuitry of the prefrontal cortex and the regulation of behavior by representational memory. In: V.B. Mountcastle, F. Plum and S.R. Geiger (Eds), Handbook of Physiology, Part I. The Nervous System, Chapter 9, American Physiological Society, Bethesda, MD, pp. 373417. Greenspan, J.D. and Winfield, J.A. (1992) Reversible pain and tactile deficits associated with a cerebral tumor compressing the posterior insula and parietal operculum. Pain, 50: 29-39. Hardy, S.G.P. and Leichnetz, G.R. (1981) Cortical projections to the periaqueductal gray in the monkey: A retrograde and orthograde horseradish peroxidase study. Neurosci. Lett., 22: 97-1 0 1. Hirono, N., Mori, E., Ishii, K., Ikejiri, Y., Imamura, T., Shimomura, T., Hashimoto, M., Yamashita, H. and Sasaki, M. (1998) Hypofunction in the posterior cingulate gyrus correlates with disorientation for time and place in Alzheimer‘s disease. J. Neurol. Neurosurg. Psychiatry, 64: 552-554. Holstege, G. (1992) The emotional motor system. Eul: J. Morphol., 30: 67-79. Hsieh, J.-C., Belfrage, M., Stone-Elander, S., Hansson, P. and Ingvar, M. (1995) Central representation of chronic ongoing neuropathic pain studied by positron emission tomography. Pain, 63: 225-236. Hsu, M.-M. and Shyu, B.-C. (1997) Electrophysiological study of the connection between medial thalamus and anterior cingulate cortex in the rat. NeuroReport, 8: 2701-2707. Jones, A.K.P., Brown, W.D., Friston, K.J.,Qi. L.Y. and Frackowiack, R.S.J. (1991) Cortical and subcortical localization of response to pain in man using positron emission tomography. Proc. R. SOC.Lond. Biol., 244: 3 9 4 . Lewin, W. and Whitty, C.W.M. (1960) Effects of anterior cingulate stimulation in conscious human subjects. J. Neurophysiol,, 23: 4 4 5 4 7 . Luppino, G., Matelli, M., Camarda, R.M., Gallese, V. and Rizzolatti, G. (1991) Multiple representations of body movements in mesial area 6 and the adjacent cingulate cortex: an intracortical microstimulation study in the macaque monkey. J. Comp. Neurol., 31 1: 463482. Mesulam, M.-M., and Mufson, E.J. (1982) Insula of the Old World monkey. Part 111. Efferent cortical output. J. Comp. Neurol., 212: 38-52. Meyer, G., McElhaney, M., Martin, W. and McGraw, C.P. (1973) Stereotactic cingulotomy with results of acute stim-

234 ulation and serial psychological testing. In: L.V. Laitinen and K.E. Livingston (Eds), Surgical Approaches in Psychiatry, MTP, Lancaster (UK), pp. 39-58. Mufson, E.J. and Mesulam, M.M. (1984) Thalamic connections of the insula in the rhesus monkey and comments on the paralimbic connectivity of the medial pulvinar nucleus. J. Comp. Neurol., 227: 109-120. Morecraft, R.J. and Van Hoesen, G.W. (1992) Cingulate input to the primary and supplementary motor cortices in the rhesus monkey: evidence for somatotopy in areas 24c and 23c. J. Comp. Neurol., 322: 471-489. Neafsey, E.J., Terrebeny, R.R., Hurley, K.M., Ruit, K.G. and Frysztak, R.J. (1993) Anterior cingulate cortex in rodents: Connections, visceral control functions, and implications for emotion. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp.206-223. Nimchinsky, E.A., Hof, P.R., Young, W.G. and Morrison, J.H. (1996) Neurochemical, morphologic, and laminar characterization of cortical projection neurons in the cingulate motor areas of the macaque monkey. J. Comp. Neurol., 374: 136-160. Olson, C.R., Musil, S.Y. and Goldberg, M.E. (1993) Posterior cingulate cortex and visuospatial cognition: Properties of single neurons in the behaving monkey. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp.366-380. Olson, C.R., M u d , S.Y. and Goldberg, M.E.(1996) Single neurons in posterior cingulate cortex of behaving macaque: Eye movement signals. J. Neumphysiol., 76: 3285-3300. Pastoriza, L.N., Morrow, T.J. and Casey, K.L. (1996) Medial frontal cortex lesions selectively attenuate the hot plate response: possible nocifensive apraxia in the rat. Pain, 64: 11-17. Pool, J.L. and Ransohoff, J. (1949) Autonomic effects on stimulating rostra1 portion of cingulate gyri in man. J. Neurophysiol., 12: 385-392. Rainville, P., Duncan, G.H., Price, D.D., Carrier, B. and Bushnell, M.C. (1997) Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science, 277: 968-97 1. Reep, R.L., Corwin, J.V., Hashimoto, A. and Watson, R.T. (1984) Afferent connections of medial precentral cortex in the rat. Neurosci. Lett., 44:247-252. Reep, R.L., Goodwin, G.S. and Corwin, J.V. (1990) Topographic organization in the corticocortical connections of medial agranular cortex in rats. J. Comp. Neurol., 294: 262-280. Shima, K., Aya, K., Mushiake, H., Inase, M., Aizawa, H. and Tanji, J. (1991) n o movement-related foci in the primate cingulate cortex observed in signal-triggered and self-paced forelimb movements. J. Neurophysiol., 65: 188-202. Siegel, A. and Brutus, M. (1990) Neural substrates of aggression and rage in the cat. Prog. Psychobiol. Physiol. Psychd., 14: 135-233.

Sikes, R.W. and Vogt, B.A. (1992) Nociceptive neurons in area 24 of rabbit cingulate cortex. J. Neurophysiol., 68: 1720-1731. Silverman, D.H.S., Munakata, J.A., Ennes, H., Mandelkern, M.A., Hoh, C.K. and Mayer, E.A. (1997) Regional cerebral activity in normal and pathological perception of visceral pain. Gastroenterology, 112: 64-72. Svensson, P., Minoshima, S., Beydoun, A., Morrow, T.J. and Casey, K.L. (1997) Cerebral processing of acute skin and muscle pain in humans. J. Neurophsyiol., 78: 450-460. Talairach, J., Bancaud, J., Geier, S., Bordas-Ferrer, M., Bonis, A. and Szikla, G . (1973) The cingulate gyrus and human behavior. Electroencephalogr. Clin. Neumphysiol., 34: 45-52. Talbot, J.D., Marrett, S., Evans, A.C., Meyer, E., Bushnell, M.C. and Duncan, G.H. (1991) Multiple representations of pain in human cerebral cortex. Science, 251: 1355-1358. Van Hoesen, G.W., Morecraft, R.J. and Vogt, B.A. (1993) Connections of the monkey cingulate cortex. In: B.A. Vogt and M. Gabriel (Eds),Neurobiology of Cingulate Cortex and Limbic Thalamus. Birkhauser, Boston, 249-284. Vogt, B.A. (1993) Structural organization of cingulate cortex: Areas, neurons, and somatodendritic transmitter receptors. In: B.A. Vogt and M. Gabriel (Eds), Neumbiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp. 19-70. Vogt, B.A. and Barbas, H. (1988) Structure and connections of the cingulate vocalization region in the rhesus monkey. In: J.D. Newman (Ed.), The Physiological Control of Mammalian Vocalization, Plenum Press, New York, pp. 203-225. Vogt, B.A., Nimchinsky, E.A., Vogt, L.J. and Hof, P.R. (1 995a) Human cingulate cortex: Surface features, flat maps, and cytoarchitecture. J. Comp. Neurol., 359: 490-506. Vogt, B.A. and Pandya, D.N. (1987) Cingulate cortex of the rhesus monkey: 11. Cortical afferents. J. Comp. Neural., 262: 27 1-289. Vogt, B.A., Pandya, D.N. and Rosene, D.L. (1987) Cingulate cortex of the rhesus monkey: I. Cytoarchitecture and thalamic afferents. J. Comp. Neurol., 262: 256-270. Vogt, B.A., Rosene, D.L. and Pandya, D.N. (1979) Thalamic and cortical afferents differentiate anterior from posterior cingulate cortex in the monkey. Science, 204: 205-207. Vogt, B.A., Sikes, R.W., Swadlow, H.A. and Weyand, T.G. (1986) Rabbit cingulate cortex: Cytoarchitecture, physiological border with visual cortex, and afferent cortical connections of visual, motor, postsubicular, and intracingulate origin. J. Comp. Neurol., 248: 74-94. Vogt, B.A., Sikes, R.W. and Vogt, L.J. (1993) Anterior cingulate cortex and the medial pain system. In: B.A. Vogt and M. Gabriel (Eds), Neurobiology of Cingulate Cortex and Limbic Thalamus, Birkhauser, Boston, pp. 3 13-344. Vogt. B.A., Vogt, L.J., Nimchinsky, E.A. and Hof, P.R. (1997) Primate cingulate cortex chemoarchitecture and its disruption in Alzheimer's disease. In: F.E. Bloom, A. Bjbrklund and T. Hokfelt (Eds), Handbook of Chemical Neuraanatomy, Vol 13. The Primate Nervous System, Part I, Elsevier Science B.V., pp. 455-528.

235 Vogt, B.A., Derbyshire, S.W.J. and Jones, A.K.P. (1995b) Pain processing in four regions of human cingulate cortex localized with coregistered PET and MR imaging. Eul: J. Neumsci., 8: 1461-1473. Whalen, P.J., Bush, G., McNally, R.J., Wilhelm, S., McInemey, S.C., Jenike, M.A. and Rauch, S.L. (1998) The emotional counting Stroop paradigm: an fMRI probe of the anterior cingulate affective division. Biolog. Psychiatry, 44: 1219-1228.

Watkins, L.R., Wiertelak, E.P. and Maier, S.F. (1993) The amygdala is necessary for the expression of conditioned by not unconditioned analgesia. Behav. Neumsci., 107: 402-405. Yamamura, H., Iwata, K., Tsuboi, Y.,Toda, K., Kitajima, K., Shimizu, N., Nomura, H., Hibiya, J., Fujita, S. and Sumino, R. (1996) Morphological and electrophysiological properties of ACCx nociceptive neurons in rats. Bruin Res., 735: 83-92.

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Pmgress in Brain Research, Vol 122 0 2000 Elsevier Science BV. AU rights reserved.

CHAPTER 17

Pain as a visceral sensation Clifford B. Saper* Department of Neurology and Program in Neuroscience, Harvard Medical School, Beth Israel Deaconess Medical Center; Boston, MA 02215. USA

Introduction Pain has traditionally been considered a primary sensory modality. Head (1920) considered pain to be a more primitive sensation than fine discrimination senses, classing it as ‘protopathic’, in comparison to the more aristocratic ‘epicritic’ sensations used to perceive tiny variations and subtle patterns in the external world and one’s relationship to it. Although this conceptualization has been criticized, pain has traditionally been paired with temperature sensation as a ‘spinothalamic’ sensation, as distinct from the fine skin and joint discrimination senses, which run in the dorsal column system. Although pain mediates responses to injury of all bodily tissues, the vast majority of pain research has focused upon cutaneous sensation, the most superficial of the body components and the most experimentally accessible. Pain has therefore been viewed as an externally directed, somatosensory modality. Visceral sensation has consequently been conceptualized along entirely different lines. ‘Parasympathetic’ visceral afferents, which are carried in the facial, glossopharyngeal, and vagus nerves, mediate chemosensory and mechanosensory modalities important for reflex parasympathetic control of internal organs. ‘Sympathetic afferents’, so-called because they enter the spinal cord along with the thoracic spinal nerves *Corresponding author. Tel.: 617-667-2622; Fax: 617-667-2987; e-mail: [email protected]

which carry sympathetic efferents, convey visceral sensory information that is important for spinal and spino-bulbo-spinal sympathetic reflexes. However, the distinction between ‘parasympathetic’ and ‘sympathetic’ visceral afferents loses sight of the fact that visceral afferents capable of eliciting autonomic reflexes enter the spinal cord at all levels, from cervical to sacral. In addition, the distinction between visceral and more superficial afferents ignores the role of superficial afferents in modulating a wide variety of autonomic reflexes (Sato and Schmidt, 1973). There are other indications that the distinction between visceral and more superficial afferents is artificial. Pain can originate from any organ, including the muscles and bones, as well as the deep organs of the thoracic and abdominal cavities. Pain fibers from these organs are carried by the same spinal nerves that convey pain from cutaneous, muscular, and bone sites. Recent advances in understanding the pathways taken by visceral sensory afferents, as well as revelations about the pathways taken by pain afferents, have steadily eroded the classical concepts that kept these modalities in distinctly different classifications. In fact, the data reviewed in this chapter will suggest that pain is itself a visceral sensory modality, originating from virtually all parts of the organism, but reaching perception in a way that is more typical of other visceral sensations than it is of the dorsal column systems with which it is more usually compared. Rather than being an outwardly directed sensory

238

modality, whose purpose is to explore the external world, pain is intrinsically an inwardly directed modality. The role of pain is to identify tissue injury, whether it is due to an external source or an internal one, and to modify reflex responses and behaviors to minimize tissue damage. In this sense, pain can be conceptualized as a visceral sensory modality concerned with the internal body beginning one cell layer under the cornified epidermis.

Visceral sensory pathways in the CNS Prior to the introduction of the axonal tracer methods for identifying neural pathways in the early 1970s, it had proven impossible to trace the fine, generally unmyelinated fibers that transmit visceral sensory information into the CNS. Hence it was not until the late 1970s and early 1980s that neuroanatomical and neurophysiological studies established for the first time the overall pathways by which visceral afferent information, carried in the facial, glossopharyngeal, and vagal nerves is distributed in the brain (Ricardo and Koh, 1978; Contreras et al., 1982). All of these afferents end in the nucleus of the solitary tract (NTS), and by tracing the outputs of the NTS, it was possible to determine the central pathways taken by visceral influences (Ricardo and Koh, 1978) (see Fig. 1). The NTS itself is topographically organized, with oral visceral sensation (taste) located in the rostra1 third of the NTS; gastrointestinal sensation from nerves innervating the esophagus, stomach, and gut terminating successively more caudally in the intermediate third of the NTS; and cardiopulmonary and respiratory afferents terminating in the caudal third of the nucleus (Altschuler et al., 1989). Projections from the NTS take three main routes. First, one set of efferents relays visceral input to the spinal cord, providing inputs to the phrenic motor nucleus and the intermediolateral cell column (Loewy and Burton, 1978; Ricardo and Koh, 1978; Holstege and Kuypers, 1982). Second, a major output from the NTS goes to the medullary reticular formation. These second order sensory fibers provide feedback to a variety of reflex responses and coordinated behaviors related to the cranial nerve nuclei. For example, taste afferents relay into the dorsal parvocellular reticular forma-

tion, where they merge with inputs related to tongue temperature and somatic sensation in providing a more complete picture of intraoral sensation that is necessary for chewing, licking, and swallowing reflexes (Norgren, 1984). Cardiorespiratory afferents relay to the ventrolateral medulla, in which there are pools of neurons that coordinate cardiovascular reflexes, such as the baroreceptor reflexes, and coordinate respiratory responses, such as yawning, sighing, sneezing, and coughing (Ross et al., 1984; Feldman, 1986). The third main output from the NTS is its ascending projection to supply visceral sensory information to the forebrain. The bulk of these fibers synapse in a relay station, the parabrachial nucleus, which obtains its name from the brachium conjunctivum (the superior cerebellar peduncle) which it surrounds (Ricardo and Koh, 1976; Herbert et al., 1990). The parabrachial nucleus relays visceral information from the NTS to a wide range of sites (Saper and Loewy, 1980; Fulwiler and Saper, 1984). It provides descending inputs to the ventrolateral and parvocellular medullary reticular formation and to the spinal cord, including essentially the same targets as the direct input from the NTS (Saper and Loewy, 1980; Holstege and Kuypers, 1982; Herbert and Saper, unpublished data). In addition, the parabrachial nucleus, joined by a much smaller number of afferents from the NTS itself, provides visceral sensory input to the hypothalamus, amygdala, and basal forebrain (Fulwiler and Saper, 1984; Moga et al. 1990; Bernard et al., 1991, 1993, 1994). The sites of innervation in the hypothalamus include the lateral hypothalamic area and paraventricular nucleus, which in turn provide descending projections back to the central autonomic system (Saper et al., 1976; Luiten et al., 1987; Cechetto and Saper, 1988). In addition, the parabrachial nucleus and NTS innervate the anteroventral third ventricular region, which is intimately involved in regulating cardiovascular responses, as well as more complex physiological responses that depend upon them, such as fluid and electrolyte balance, thermoregulation and reproductive cycles (Saper and Levisohn, 1983; Saper et al., 1983; Simerly and Swanson, 1987). In the basal forebrain, the parabrachial nucleus and the NTS innervate the central nucleus

239

of the amygdala, the bed nucleus of the stria terminalis, and the region that connects the two, which has been termed the ‘extended amygdala’ (Moga et al., 1990).

Pain Pathways

The visceral afferent input to the thalamus is mainly from the parabrachial nucleus in rats (Norgren, 1976; Fulwiler and Saper, 1984). In primates, some neurons in the rostra1 NTS, con-

Visceral Pathwavs

Cortex

Ventroposterior

Insular (Visceral) Cortex

ThalamIUS

Pain (S2) Cortex

\

I

i

me11 area I

Pons

Medulla

Vagus Nerve

Spinal

Fig. 1. A summary diagram of the brain and spinal cord comparing aspects of the ascending pathways taken by the pain system (left) and the traditional visceral sensory system (right). Note that the ascending nociceptive inputs converge with visceral inputs in the nucleus of the solitary tract and the parabrachial complex in the brainstem. At the level of the thalamus and the cerebral cortex, visceral and pain afferents are topographically organized in adjacent fields (see text for full explanation). At every level of the nervous system, pain is handled by the brain as if it were a visceral sensation.

240

cerned with taste, also project as far as the thalamus, but other parts of the NTS seem to relay via the parabrachial nucleus (Beckstead et al., 1980). The thalamic visceral sensory area is a narrow strip, just ventral to the classical somatosensory nuclei, the ventroposteromedial (VPM) and ventroposterolateral (VPL) nuclei, concerned with sensation in the face and body, respectively (Cechetto and Saper, 1987). The visceral sensory strip has been explored electrophysiologically in the rat, in which it is organized in a roughly topographic order (Cechetto and Saper, unpublished data). The most medial neurons underlying VPM are mainly concerned with taste; those concerned with gastrointestinal sensation are more laterally placed; and those concerned with cardiorespiratory sensation are furthest lateral, near the medial border of VPL. The visceral sensory nucleus in the rat receives afferents from neurons in the parabrachial nucleus containing calcitonin gene-related peptide (CGRP) (Yasui et al., 1989). Afferents containing the same peptide can be seen in both the parabrachial nucleus and the thalamus in monkeys and in humans, suggesting that this pathway be conserved through evolution (de Lacalle and Saper, unpublished). The parabrachial nucleus in the rat also innervates visceral sensory areas in the cerebral cortex. These include a portion of the insular cortex which receives inputs from the visceral sensory thalamus, as well as the infralimbic cortex, which acts as an autonomic motor cortex (Cechetto and Saper, 1990; Hurley et al., 1991).

New views of the central nociceptive system Although the pain system in the brain is often equated with the spinothalamic pathway, in fact even very early evidence indicated that the spinothalamic component of this system is only a small part of it. Examination of the distribution of the anterolateral column fibers in monkeys, by Mehler and his colleagues (1960), indicated this ascending fiber pathway gives off a variety of branches before it reaches the thalamus. In fact, the sites that receive this spinal input match very closely those that receive input from the vagal visceral sensory system, including the nucleus of the solitary tract

itself, the ventrolateral medulla, and the parabrachial nucleus. These inputs were rediscovered in the 1980s and it was found that the inputs to the NTS and the parabrachial nucleus derive from neurons both in the superficial laminae of the spinal cord, most usually associated with nociceptive afferents, as well as in part from the deeper layers of the dorsal horn (Cechetto et al., 1985; Menetrey and Basbaum, 1987). Recordings of neurons in the spinal dorsal horn confirmed that neurons in these sites that project to the parabrachial nucleus are, for the most part, nociceptive, but that many of them have visceral sensory fields (Foreman et al., personal communication). For example, Foreman, who has demonstrated convergence of visceral (including cardiac) and superficial pain in individual dorsal horn neurons in the monkey and cat spinal cord (Foreman et al., 1984; Hobbs et al., 1992), finds that as many as half of these neurons can be backfired from the parabrachial nucleus. Even more surprisingly, Burstein, Geisler, and their colleagues find that some spinal or trigeminal dorsal horn nociceptive neurons send their axons all the way to the hypothalamus or the amygdala (Burstein et al., 1987; Cliffer et al., 1991; Burstein and Potrebic, 1993). Spinal afferents to the parabrachial complex end in a distinct set of nuclei that overlap only in part with those receiving visceral input from the NTS (Cechetto et al., 1987; Feil and Herbert, 1995). Some parabrachial subnuclei that receive spinal input, such as the internal lateral nucleus, which projects to the intralaminar thalalmus, receive no NTS input at all, and may be primarily concerned with pain (Fulwiler and Saper, 1984). These parabrachial projections overlap substantially with the medial thalamic inputs from the spinothalamic tract itself. On the other hand, the external lateral parabrachial nucleus, which contains neurons that project to the central nucleus and the ‘extended’ amygdala, is activated by painful stimuli, but also receives massive inputs from the NTS (Herbert et al., 1990; Bernard et al., 1994). This cell group may be involved in emotional responses to visceral stimuli, including pain. Another part of the parabrachial nucleus, the superior lateral nucleus, which receives pain as well as visceral afferent

24 1

inputs related to the gastrointestinal tract and feeding, projects to the ventromedial nucleus of the hypothalamus, and is thought to be involved in feeding responses and perhaps the anorexic response to pain (Fulwiler and Saper, 1985; Bester et al., 1997; Hermanson and Blomqvist, 1997). Finally, the careful mapping of the spinothalamic inputs to the lateral thalamus has indicated that pain afferents are more segregated from dorsal columnmedial lemniscal afferents than had once been realized. In rats and in cats, spinothalamic fibers mainly innervate a region along the outer shell of the VPM and VPL, which is adjacent to and extends beyond the CGRP-delimited visceral sensory nucleus (Yokota et al., 1988; Yokota, 1989; Craig,1995). In monkeys and in humans, Craig and colleagues have identified a calbindin-immunoreactive nucleus, which they call the ventromedial posterior nucleus, which is just lateral to the visceral sensory nucleus, and which receives nociceptive inputs (Craig et al., 1994). Thus, the visceral and nociceptive afferents to the thalamus form a shell which virtually encases the classic somatosensory VPM and VPL nuclei. The cortical territories of these thalamic cell groups are also of interest. The visceral sensory nucleus, in both the rat and in primates, projects to a rostral, dysgranular part of the insular cortex (Pritchard et al., 1986; Kosar et al., 1986; Cechetto and Saper, 1987). The slightly more posterior nociceptive cell group projects into a part of the second somatosensory representation that is located just caudal to the visceral sensory representation in the insular region (Robinson and Burton, 1980). Hence, at both the levels of the thalamus and the cortex, pain is represented as a topographic extension of the visceral sensory map.

hence emotional response. The profound nature of pain responses, which can color all of behavior, are similar to the behavioral responses to visceral sensations, such as hunger, air hunger, abdominal distention, or bladder fullness. In fact, it is difficult, in retrospect, to conceive why pain, which monitors the integrity of virtually all tissues, a key visceral modality, should ever have been considered anything other than a visceral sensation. The answer may be that the original conceptualization of pain came from experiments involving the cutaneous surface, and comparing pain to other sensations arising from that surface. These experimenters were, for obvious reasons, concerned with sensory systems that monitor the outside world. However, pain, by its nature, like any visceral sensation, is really a modality that monitors the inner world. Pain is not aimed at exploring the properties of objects in the outer world but rather pain guards against tissue damage during that exploration. For example, it may be discovered by sense of touch that a pin is cool, smooth, and elongated, and even that the point is quite sharp, but it is not intrinsically painful; the latter sensation is evoked only when the point of the pin threatens or causes actual tissue damage and exemplifies the internal state, not a property of the pin. Skin is a critical organ whose integrity is of as great survival value to the organism as any other. By classing cutaneous pain with sensations whose purpose is quite different, i.e., to explore the outer world and the individual’s relationship to it, we have done a major disservice to the understanding of what is, ultimately, the most pervasive and arguably the most important visceral sensory modality in the body.

Behavioral and autonomic responses to pain and visceral sensation

Alden, M., Besson, J.M. and Bernard, J.F. (1994) Organization of the efferent projections from the pontine parabrachial area to the bed nucleus of the stria terminalis and neighboring regions - a PHA-L study in the rat. J. Comp. Neurol., 341: 289-3 14. Altschuler, S.M., Bao, X., Bieger, D., Hopkins, D.A. and Miselis, R.R. (1989) Viscerotopic representation of the upper alimentary tract in the rat: sensory ganglia and nuclei of the solitary and spinal trigeminal tracts. J. Comp. NeuroL, 283: 248-268. Beckstead, R.M., Morse, J.R. and Norgren, R. (1980) The nucleus of the solitary tract in the monkey: projections to the

At the behavioral and autonomic levels, pain is also treated as a visceral sensation. Like other visceral inputs, pain can elicit profound visceral motor responses. These may be produced at a brainstem level, by pain inputs to the visceral motor system (Sato and Schmidt, 1973), or at a forebrain level, by the influences of pain on cortical function, and

References

242 thalamus and brain stem nuclei. J. Comp. Neurol., 190: 259-282. Bernard, J.F., Alden, M. and Besson, J.M. (1993) The organization of the efferent projections from the pontine parabrachial area to the amygdaloid complex - a phaseolusvulgaris leucoagglutinin (PHA-L) study in the rat. J. Comp. Neuml., 329: 201-229. Bernard, J.F., Cmoue, J. and Besson, J.M. (1991) Efferent projections from the external parabrachial area to the forebrain - a phaseolus-vulgaris leucoagglutinin study in the rat. Neumsci. Lett., 122: 257-260. Bernard, J.F., Huang, G.F. and Besson, J.M. (1994) The parabrachial area: Electrophysiological evidence for an involvement in visceral nociceptive processes. J. Neumphysiol., 71: 1646-1660. Bester, H., Matsumoto, N., Besson, J.M. and Bernard, J.F. (1997) Further evidence for the involvement of the spinoparabrachial pathway in nociceptive processes: a c-fos study in the rat. J. Comp. Neuml., 383: 439-458. Burstein, R., Cliffer, K.D. and Giesler, G.J. (1987) Direct somatosensory projections from the spinal cord to the hypothalamus and telencephalon. J. Neumsci., 7: 41594164. Burstein, R. and Potrebic, S. (1993) Retrograde labeling of neurons in the spinal cord that project directly to the amygdala or the orbitol cortex in the rat. J. Comp. Neuml., 335: 469-485. Cechetto, D.F. and Saper, C.B. (1987) Evidence for a viscerotopic sensory representation in the cortex and thalamus in the rat. J. Comp. Neurol., 262: 27-45. Cechetto, D.F. and Saper, C.B. (1988) Neurochemical organization of the hypothalamic projection to the spinal cord in the rat. J. Comp. Neurol., 272: 579-604. Cechetto D.F. and Saper C.B. (1990) Role of the cerbral cortex in autonomic function. In: A.D. Loewy and K.D. Spyer (Eds), Central Regulation of Autonomic Functions, Oxford University Press, New York, pp. 208-223. Cechetto, D.F., Standaert, D.G. and Saper, C.B. (1985) Spinal and trigeminal dorsal horn projections to the parabrachial nucleus in the rat. J. Comp. Neuml., 240: 153-160. Cliffer, K.D., Burstein, R. and Giesler, G.J. (1991) Distributions of spinothalamic, spinohypothalamic, and spinotelencephalic fibers revealed by anterograde transport of PHA-L in rats. J. Neurosci., 11: 852-868. Contreras, R.J., Beckstead, R.M. and Norgren, R. (1982) The central projections of the trigeminal, facial, glossopharyngeal and vagus nerves: An autoradiographic study in the rat. Brain Res. Bull., 6: 303-322. Craig, A.D. (1995) Distribution of brainstem projections from spinal lamina I neurons in the cat and the monkey. J. Comp. Neuml., 361: 225-248. Craig A.D., Bushnell, M.C., Zhang, E.T. and Blomqvist, A. (1994) A thalamic nucleus specific for pain and temperature sensation. Nature, 372: 770-773. Feil, K. and Herbert, H. (1995) Topographic organization of spinal and trigeminal somatosensory pathways to the rat

parabrachial and Kolliker-Fuse nuclei. J. Cornp. Neurol., 353: 506-528. Feldman, J.L. (1986) Neurophysiology of breathing in mammals. In: F.E. Bloom (Ed.), Handbook of Physiology - The Nervous System IV - Intrinsinc Regulatory Systems of the Bruin. American Physiological Society, Bethesda, MD. Foreman, R.D., Blair, R.W. and Weber, R.N. (1984) Viscerosomatic convergence onto T2-T4 spinoreticular, spinoreticularspinothalamic, and spinothalamic tract neurons in the cat. Exp. Neurol., 85: 597-619. Fulwiler, C.E. and Saper, C.B. (1984) Subnuclear organization of the efferent connections of the parabrachial nucleus in the rat. Brain Res. Rev., 7: 229-259. Fulwiler, C.E. and Saper, C.B. (1985) Cholecystokinin-immunoreactive innervation of the ventromedial hypothalamus in the rat: Possible substrate for autonomic regulation of feeding. Neurosci. Lett., 53: 289-296. Head, H. (1920) Herbert, H., Moga, M.M. and Saper, C.B. (1990) Connections of the parabrachial nucleus with the nucleus of the solitary tract and the medullary reticular formation in the rat. J. Comp. Neurol., 293: 540-580. Hermanson, 0. and Blomqvist, A. (1997) Subnuclear localization of FOS-like immunoreactivity in the parabrachial nucleus after orofacial nociceptive stimulation of the awake rat. J. Comp. Neurol., 387: 114-123. Hobbs, S.F., Chandler, M.J., Bolser, D.C. and Foreman, R.D. (1992) Segmental organization of visceral and somatic input onto C3-T6 spinothalamic tract cells of the monkey. J. Neurophysiol., 68: 1575-1 588. Holstege, G. and Kuypers, H.G.J.M. (1982) The anatomy of brain stem pathways to the spinal cord in cat. A labeled amino acid tracing study. In: H.G.J.M. Kuypers and G.F. Martin (Eds), Descending Pathways to the Spinal Cord, Elsevier Biomedical Press, pp. 145-175. Hurley, K.M., Herbert, H., Moga, M.M. and Saper, C.B. (1991) Efferent projections of the infralimbic cortex of the rat. J. Comp. Neuml., 308: 249-216. Kosar, E., Grill, H.J. and Norgren, R. (1986) Gustatory cortex in the rat. I. Physiological properties and cytoarchitecture. Brain Res., 379: 329-341. Loewy, A.D. and Burton, H. (1978) Nuclei of the solitary tract: Efferent projections to the lower brain stem and spinal cord. J. Comp. Neuml., 181: 421-450. Luiten, P.G.M., Ter Horst, G.J. and Steffens, A.B. (1987) The hypothalamus, intrinsic connections and outflow pathways to the endocrine system in relation to the control of feeding and metabolism. Prog. Neurobiol., 28: 1-54. Mehler, W.R., Feferman, M.E. and Nauta, W.J.H. (1960) Ascending axon degeneration following anterolateral cordotomy. An experimental study in the monkey. Brain, 83: 7 18-752. Menetrey, D. and Basbaum, A.I. (1987) Spinal and trigeminal projections to the nucleus of the solitary tract: a possible substrate for somatovisceral and viscerovisceral reflex activation. J. Comp. Neurol., 255: 439-450.

243 Moga, M.M., Herbert, H., Hurley, K.M., Yasui, Y., Gray, T.S. and Saper, C.B. (1990) Organization of cortical, basal forebrain, and hypothalamic afferents to the parabrachial nucleus in the rat. J. Comp. Neurol., 295: 624-661. Norgren, R. (1976) Taste pathways to hypothalamus and amygdala. J. Comp. Neurol., 166: 17-30. Norgren R (1984) Central neural mechanisms of taste. In: D. Smith I (Ed.), Handbook of Physiology, Section I: The Nervous System, Sensory Processes, Vol. 111, American Physiological Society, Bethesda, MD, pp. 1087-1 128. Pritchard, T.C., Hamilton, R.B., Morse, J.R. and Norgren, R. (1986) Projections of thalamic gustatory and lingual areas in the monkey, macaca fascicularis. J. Comp. Neurol., 244: 213-228. Ricardo, J.A. and Koh, E.T. (1978) Anatomical evidence of direct projections from the nucleus of the solitary tract to the hypothalamus, amygdala, and other forebrain structures in the rat. Brain Res., 153: 1-26. Robinson, C.J. and Burton, H. (1980) Somatic submodality distribution within the second somatosensory (SII), 7b, retrohsular, postauditory, and granular insular cortical areas of M. fascicularis. J. Comp. Neurol., 192: 93-108 Ross, C.A., Ruggiero, D.A., Park, D.H., Joh, T.H.,Sved, A.F., Fernandez-Pardol, J., Saavedra, J.N. and Reis, D.J. (1984) Tonic vasomotor control by the rostra1 ventrolateral medulla: effect of electrical or chemical stimulation of C1 adrenaline containing neurons on arterial pressure, heart rate and plasma catecholamines and vasopressin. J. Neurosci., 4: 474-494. Saper, C.B. and Levisohn, D. (1983) Afferent connections of the median preoptic nucleus in the rat: anatomical evidence

for a cardiovascular integrative mechanism in the anteroventral third ventricular (AV3V) region. Bruin Res., 288: 21-31. Saper, C.B. and Loewy, A.D. (1980) Efferent connections of the parabrachial nucleus in the rat. Brain Res., 197: 291-317. Saper, C.B., Loewy, A.D., Swanson, L.W. and Cowan, W.M. (1976) Direct hypothalamo-autonomic connections. Brain Res., 117: 305-312. Saper, C.B., Reis, D.J. and Joh, T. (1983) Medullary catecholamine inputs to the anteroventral third ventricular cardiovascular regulatory region in the rat. Neurosci. Lett., 42: 285-291. Sato, A. and Schmidt, R.F. (1973) Somatosympatheic reflexes: afferent fibers, central pathways, discharge characteristics. Physiol. Rev., 53: 916-947. Simerly, R.B. and Swanson, L.W. (1987) The distribution of neurotransmitter-specificcells and fibers in the anteroventral periventricular nucleus: Implications for the control of gonadotropin secretion in the rat. Brain Res., 400: 11-34. Yasui, Y., Saper, C.B. and Cechetto, D.F. (1989) Calcitonin gene-related peptide immunoreactivityin the visceral sensory cortex, thalamus, and related pathways in the rat. J. Comp. Neuml., 290: 487-501. Yokota, T. (1989) Thalamic mechanism of pain: shell theory of thalamic nociception. Jpn. J. Physiol., 39: 335-348. Yokota, T., Asato, F., Koyama, N., Masuda, T. and Taguchi, H. (1988) Nociceptive body representation in nucleus ventralis posterolateralis of cat thalamus. J. Neurophysiol., 60: 1714-1 727.

This Page Intentionally Left Blank

E.A. Mayer and C.B.Saper (Fds.) Progress in Brain Research,Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 18

Pain modulation: expectation, opioid analgesia and virtual pain Howard L. Fields Department of Neurology, Box 0453, University of California, 513 Parnassus Avenue, S-784, San Francisco, CA 94143, USA

Introduction This volume is a reflection of the nascent, but still only vaguely appreciated idea that the nervous system is the origin of what we call the mind, the body and the world. The brain weaves an intriguing tale for us. It informs us that there is a physical world beyond us. It also tells us that our subjective awareness, (what some people call mind), is a different kind of stuff. What it doesn’t tell us is that both the external world and the self we experience are (generated by) spatiotemporalpatterns of neural activity called representations. It isn’t that there is no physical body or external world, only that our experience of them is indirect. In some ways our subjective experiences bear the same relationship to the physical entities they represent as a television image has to the physical things whose image is projected onto the screen. The image on the screen is generated by a spatiotemporal pattern of charged particles. As with neural representations, this pattern encodes meaning, i.e. it represents the actual thing whose image appears on the screen. Understandingpain: representations of the mind, the physical body and the subjectively perceived body

mind, is also a representation. Since the awareness of both bodily pain and mind are neurally generated, they are basically the same kind of thing and they can interact. Because they are both generated by neural activity, the dualism of subjectively experienced body and mind is illusory. In contrast to the subjectively experienced body, the physical body (the viridical body that others can see and touch) is a completely different sort of thing than either perceived pain or mind. Thus there is no obvious way for the mind and physical body to directly interact. The dualism of the physical body and the mind is therefore not illusory. In contrast to the subjective experiences they generate, neural ensembles (representations) have a physical existence in the body. There is no philosophical problem with the physical body affecting and being affected by the brain. Thus unless the agency of the brain is clearly appreciated, the mind body ‘problem’ has an inherent mystery which has provided fascinating contemplation for countless numbers of people. By putting the brain into the issue, modem neuroscience establishes a satisfying framework for greater understanding of the mind body interaction. Because they have a physical existence and generate subjective experiences the activity patterns of the brain provide- a nexus between the physical body and the subjectively experienced self (mind).

The Physical correlate of the subjective experience of bodily pain is generated by a neural representation. The physical correlate of our sense of self, or

Projections and the virtual body

*Corresponding author. Tel.: 415-476-4201; Fax: 415-476-4201; e-mail: [email protected]

Let me illustrate the primacy and sufficiency Of the brain for pain with a few dramatic examples.

246

Although injury to the body is the usual cause of pain, the body per se is neither necessary nor sufficient for pain. The phenomenon of projection is a clear illustration of this fact. One striking example of projection is that electrical stimulation of certain brain regions (for example the ventral caudal thalamus) can produce pain in a specific body part in the absence of any tissue injury (Davis et al., 1998). What is felt is the activity of a group of neurons. These neurons are located in the brain, but the senqtion they produce is projected to a location determined by the ensemble of active neurons. Another illustration of projection is phantom limb. Phantom limb is the robust sensation that an amputated body part is still there. These examples of projection inform us that what we ‘feel’ when the car door smashes our thumb is not the thumb in the car door but the neural representation of the thumb. If you will, this is a virtual thumb. To repeat, the subjective experience of pain is a projection created by a neural representation. Through some poorly understood process the sensation is projected from the physical location of the neural activity (in the brain) to the site in the body where it is perceived to occur. Although pain is often produced by stimuli to the body, the body is not necessary for pain to be experienced. In contrast, neural activity is both necessary and sufficient to produce the subjective experience of pain. The virtual nature of sensation is a corollary of an absolutely critical property of the brain: that it can only be fully understood in the context of what it does. What the brain does is to create representations: of the body, of the outer world, of past events and of potential futures. The brain has intentionality; it is about things other than itself. It is a metaphorical organ. Just as a book is not about ink and paper, the brain is not ‘about’ action potentials and synapses. Books and brains are about meaning and, for both, the meanings are about things outside themselves. The meaning of pain is that damage is occurring or that there is potential damage to the body in the current situation and there is benefit in terminating (or avoiding in future) the cause of the damage. This semi-philosophical introduction is crucial for understanding how psychological factors can

have such powerful effects on pain perception. There is no question that the body can affect the mind, that is what the visceral and somatic sensory systems and the various CNS chemoreceptors are designed to do. It is also clear from what is presented by others in this volume, that through the autonomic and endocrine systems, the brain can exert powerful effects on the physical body. In this chapter, I will take a somewhat different tack and present a neural model for understanding how psychological factors can either reduce or enhance pain. The powerful potential of psychological factors to modify pain is much easier to accept after one fully appreciates that the pain we experience as occurring in the body is a neural representation with no necessary correlate in the physical body. Pain modulation

Under conditions of strong emotion (e.g. anger, fear, elation) major injuries may be essentially painless. Conversely, in situations when pain is anticipated, subjects often report the occurrence or worsening of pain without an imposed noxious stimulus. The point is that whether or not a noxious stimulus produces pain depends as much upon the state of the subject as it does on stimulus parameters. The discovery and elucidation of CNS circuits that selectively modulate pain transmission has provided us with a potential explanation for how attention, arousal and expectation alter pain perception (Basbaum and Fields, 1984; Fields et al., 1991; Fields and Basbaum, 1994). Psychologicalfactors influence the $ring of dorsal horn pain transmission neurons

The clearest evidence that pain transmission neurons are modulated by attentional factors was provided by electrophysiological studies of pain transmission neurons in the trigeminal nucleus caudalis in the medulla. This nucleus, which is the homologue of the spinal cord dorsal horn, receives direct input from primary afferent nociceptors that innervate cranial structures. In elegant experiments from the U.S. National Institutes of Health (Duncan et al., 1987), monkeys were trained to detect small noxious heat pulses to their upper lip. The monkeys

247

learned to press and hold a lever in response to a light cue signaling the task to be performed. To obtain a reward, the monkey had to hold the lever down until he detected a brief noxious heat pulse. All neurons discharged at higher frequencies when stimuli of increasing intensity were applied to their receptive fields. The critical finding in these experiments was that many of these pain transmission neurons increased (or decreased) their firing rates in response to either the signal light or the act of pushing the lever, before any noxious stimulus was applied! The cells had acquired the property of changing their activity in anticipation of a noxious stimulus. These experiments show that in the primate nervous system there are pathways that mediate the modulatory actions of attention and learning on paintransmission neurons. What are these pathways? Pain modulation and opioid analgesia

The discovery of stimulation produced analgesia was a critical step toward understanding brain mechanisms of pain modulation. Electrical stimulation of specific brain regions produces analgesia in rats and humans with chronic pain (Baskin et al., 1986). The midbrain periaqueductal gray (PAG) was the first brain pain modulating site discovered (see Chapter 24 by Bandler et al. in this volume). Activation of the PAG suppresses pain behaviors in rats. In humans, stimulation of the PAG produces a gradual fading of ongoing pain without any other consistent sensory or motor effects. The PAG is part of a circuit that controls nociceptive neurons in the dorsal horn (Fig. 1). The PAG receives major inputs from the frontal neocortex, the hypothalamus and the amygdala. One major projection from the PAG is caudally to the rostral ventromedial medulla (RVM) which in turn projects massively and selectively to pain transmitting neurons in the dorsal horn of the spinal cord and the trigeminal nucleus caudalis. Electrical stimulation of either the PAG or RVM produces analgesia and inhibits dorsal horn pain transmission neurons. Another major brainstem region involved in pain modulation is the dorsolateral pontine tegmentum (DLPT). The DLPT is directly linked to both the PAG and RVM and projects directly to the spinal cord dorsal horn.

Opioid analgesics work in part by activating these brainstem pain-modulating sites. Thus, when opioid agonists such as morphine are microinjected into the frontal cortex, amygdala, PAG or RVM a powerful analgesic effect is produced (Fields et al., 1991; Helmstetter et al., 1993; Burkey et al., 1996). The amygdala, hypothalamus, PAG, DLPT and RVM each contain high concentrations of endogenous opioid peptides and they are linked to each other by opioid synapses (Ma et al., 1992; Kiefel et al., 1993; Roychowdhury and Fields, 1996). Furthermore, activation of PAG neurons inhibits spinal cord pain transmission neurons through a locally released endogenous opioid (Budai and Fields, 1998). In addition to this opioid linked circuit, there are other important descending inputs to dorsal horn nociceptive neurons. For example, there are direct projections from somatosensory cortex, hypothalamus and the subnucleus reticularis dorsalis (Holstege, 1988; Villanueva et al., 1995). Any of these pathways could contribute to the clinical variability of pain and to the demonstrated modulatory effects of expectation and attention on pain intensity. In this chapter, however, the focus will be on the amygdala-PAG-RVM to dorsal horn circuit since our knowledge of its connectivity, pharmacology and function is most extensive. Physiological activation of opioid-mediated pain modulating circuits

The most studied form of physiological activation of opioid linked pain modulating circuits comes under the general rubric of stress induced analgesia. Although there are many ways to stress rodents, arguably the most productive approach has been to study conditioned fear-evoked defense responses. In the conditioned fear paradigm, rats are subjected to an inescapable noxious stimulus, leading to stress and apparent analgesia (Watkins and Mayer, 1982; Watkins et al., 1983; Fanselow, 1991; Helmstetter and Tershner, 1994). Animals recover rapidly, but when returned to the contex in which they were shocked, the environment is sufficient to produce an analgesic effect. This form of conditioned analgesia can be blocked by the opioid antagonist naloxone or by lesions of the amygdala-

248

PAG-RVM-dorsal horn circuit described above (Kimet al., 1993; Helmstetter and Tershner, 1994). These animal studies demonstrate that analgesia can be learned and indicate that specific circuitry and neurotransmitters can contribute to the analgesic effect. What is the evidence that similar pain modulating circuitry exists in people?

Although technically possible, it is not feasible to carry out in humans the type of invasive anatomical and physiological experiments needed to study the relevant neural circuits at the required level of spatial and temporal resolution. Since it is very

difficult to directly test a specific neural circuit hypothesis about a complex human behavior, we must primarily rely upon indirect evidence. There are several independent lines of evidence supporting the idea that humans have opioid-mediatedpain modulating circuitry similar to that discovered in animals. The brainstem-to-spinal cord circuitry implicated in conditioned analgesia is highly conserved in a variety of mammalian species, including rodents, carnivores, primates and marsupials. Importantly, the distribution of neurotransmitters,including opioid peptides, in this pathway also appears to be similar in a number of species, including humans (Emson et al., 1984;

Fig. 1. Diagram of Pain Transmission and Modulation Pathways (Adapted from Fields, H.L., Pain, McGraw-Hill, 1987). Left: A noxious stimulus applied to the skin (lower left) elicits a train of impulses in peripheral nociceptors which are propagated to the dorsal horn of the spinal cord where they activate the nerve cells of origin of the spinothalamic tract. The spinothalamic tract activates thalamic neurons which, in turn, project to and activate neurons in the cingulate cortex (C), frontal cortex (F) and somatosensory cortex (SS). Right: A variety of stimuli can activate pain modulation circuits. Frontal and cingulate cortex neurons, as well as afferents from the amygdala (A) and hypothalamus (H), converge on midbrain periaqueductal gray (PAG) neurons, which, through a relay in the rostral ventromedial medulla (RVM) control spinothalamic pain transmission neurons.

249

Pittius et al., 1984). The highly conserved anatomy and pharmacology of pain-modulating circuitry across this diversity of species leaves little doubt that similar circuitry is present in humans. Furthermore, opioid drugs that relieve clinically significant pain in humans are effective in inhibiting escape behaviors in all tested mammalian species. In animals, this analgesic effect is exerted in part through actions upon the brainstem-tospinal cord circuitry that mediates conditioned analgesia. Perhaps the strongest support for the idea that homologous circuitry produces analgesia in humans is that, as mentioned above, patients with chronic pain report relief during stimulation of specific sites in the human midbrain (Baskin et al., 1986). The effect is selective for pain and is elicited by stimulating the PAG. Another approach to this issue is the use of pharmacological interventions. Thus, if similar endogenous opioid circuitry modulates pain transmission in humans and rodents, it should be possible to define conditions under which analgesia is produced in humans and can be reversed with the selective opioid antagonist naloxone. The situations that produce endogenous opioid-mediated analgesia in rats would be difficult to study in a controlled manner in human subjects. On the other hand, it has been possible to demonstrate naloxone hyperalgesia in humans. We were the first to clearly demonstrate that, compared to placebo, 10 mg of the opioid antagonist naloxone significantly worsens postoperative pain (Levine et al., 1978b). Because it produces no subjective effect that could consistently betray its administration, it can be given in a true double blind manipulation (Wolkowitz and Tinklenberg, 1985). Naloxone hyperalgesia has been replicated in studies using both clinical and experimental pains (Grevert and Goldstein, 1985; Benedetti and Amanzio, 1997). The significance of these naloxone studies is twofold. First, they demonstrate, under controlled conditions, that an endogenous pain-modulating system in humans can be reproducibly activated. Second, the hyperalgesic response to naloxone suggests that the pain-modulating system had opioid links. In summary, the weight of evidence supports the notion that endogenous opioid linked pain modulating circuits produce analgesia in

humans. One question of critical importance is what factors activate these pain-modulating circuits. Expectation and placebo analgesia

In rats, threat is clearly a robust determinant, and anecdotal evidence suggests that threat can produce powerful analgesic effects in humans (Melzack and Wall, 1982). In humans, expectation is another very potent pain modulator. The well-known phenomenon of placebo analgesia is a familiar example of the power of expectation to control pain (Fields and Price, 1997). Placebo analgesia is a robust phenomenon that can be easily demonstrated in group comparisons (Gelfand et al., 1963; Liberman, 1964; Gracely et al., 1983; Levine and Gordon, 1984; Grevert and Goldstein, 1985). Since previous exposure to a similar appearing effective analgesic treatment significantly enhances the effectiveness of a placebo, it is likely that expectation is a critical factor in generating placebo analgesia (Laska and Sunshine, 1973;Voudouris et al., 1990). Does placebo analgesia in humans involve the opioid-linked PAG-RVM-dorsal horn painmodulating circuitry described above?

In our first study of this issue (Levine et al., 1978a), dental postoperative pain patients were given a placebo and then randomized to receive either a second placebo or naloxone. Since the typical natural history of pain under these conditions is a steady increase for up to five hours (Levine et al., 1979), those patients whose pain increased following the first placebo were labeled placebo non-responders; those whose pain decreased or did not change were called responders. Naloxone increased pain in placebo responders but had no effect on non-responders. In a subsequent study that used identical methods, a natural history group that received hidden infusions of either naloxone or vehicle was included (Levine and Gordon, 1984). In this study, naloxone reversed placebo analgesia but had no effect upon patients that had had no overt treatment. Together, these studies demonstrate that placebo analgesia can have a major

250

naloxone-reversible component. Studies using ischemic ann pain in normal volunteers have confirmed this conclusion (Grevert and Goldstein, 1985; Benedetti and Amanzio, 1997). Although not conclusive, these data provide an evidentiary link between the behavioral phenomenon of placebo analgesia in humans and specific pain-modulating circuitry, and they are consistent with the hypothesis that any therapeutic act interpreted by the patient as potentially effective can trigger pain-modulating circuitry with opioid links. Such circuitry appears to be present in the human central nervous system, and its rodent homolog is well-described. On and off cells in the RVM: bi-directional control of pain

There is evidence that the same circuit that produces opioid analgesia can, under other circumstances, enhance pain. This pain enhancing function raises the intriguing possibility that the brain can generate a pain signal in the absence of a peripheral stimulus. Electrophysiological recordings in the RVM have revealed unexpected and interesting properties of its neurons. One class of RVM neuron, the On cell, shows a burst of activity beginning just prior to withdrawal from a noxious stimulus. The other major class of cell, the Off cell, has the opposite firing pattern (Fig. 2), pausing during withdrawal from noxious heat. On cells receive an enkephalinergic input, are inhibited by opioids through an action at the p opioid receptor and project to the spinal cord dorsal horn (Fields et al., 1991; Fields et al., 1995). Off cells, which also project to the spinal cord dorsal horn, are inhibited by On cells, and thus are activated indirectly (disinhibited) by opioids. Several independent lines of evidence support the hypothesis that On cells facilitate, while Off cells inhibit pain transmission (Fields et al., 1991; Heinricher et al., 1994). Both On and Off cells project directly to those laminae of the dorsal horn that contains pain-transmission neurons (Fields et al., 1995). If On cells do indeed excite dorsal horn pain transmission neurons they provide a substrate for the generation of a pain signal through central nervous system activity. Is there

7

MORPHINE INHIBITS

MORPHINE EXCITES

-F

pain transmlsslon cell

pain receptor

Fig. 2. Pain modulating neurons exert bi-directional control of pain transmission. Upper left: Neurons in the rostroventral medulla (RVM), receive input from the midbrain periaqueductal gray (PAG) and directly control spinal cord dorsal horn neurons. Upper right. There are two classes of RVM pain modulating neuron which exert opposite effects on pain transmission. On-cells show a burst of activity just prior to the heat evoked tail flick (TF) reflex. Off-cells pause just prior to the TF. Lower: Morphine excites Off-cells that inhibit pain transmission. Morphine inhibits On-cells that facilitate spinal pain transmission

any evidence that psychological factors can enhance or generate pain in human subjects? Attention and expectation can generate or worsen pain

Attention and affect are influenced by and can powerfully modulate subjective responses to noxious stimuli. The effect of attention on pain is

25 1

widely appreciated by the lay public and many people use distraction as a common tool for coping with pain. Formal psychophysical studies have confirmed that drawing attention away from a noxious stimulus raises the pain threshold at that point (Miron et al., 1989). Conversely, pain intensity ratings are increased when a subject's attention is focused on a pain assessment task (Bushnell et al., 1985). One important study indicating that attention and expectation or suggestion may induce clinically significant pain (headache) was carried out by Bayer and colleagues (Bayer et al., 1991). They observed that placing dummy electrodes on the head of normal volunteers can induce head pain if the subjects believe they are being electrically stimulated. In their study, head pain was not correlated with increases in heart rate or electrodermal responses, which might be expected if stress/anxiety alone were responsible for the pain. Emotional stress, anxiety and negative aflect are correlated with increased pain

The most common and commonly studied affective variable related to pain perception is depressive symptomatology. Patients with chronic pain, including those with frequent moderate to severe migraine or tension type headaches are very likely to have depressive symptoms (Fishbain et al., 1986; Merikangas et al., 1990; Breslau et al., 1991; Ziegler and Paolo, 1995).A very high percentage of depressed patients report somatic pain (Wells et al., 1989). More relevant to the current discussion is the limited evidence that depression can cause or worsen pain. For example, inducing sadness in normal volunteers reduces pain tolerance (Zelman et al., 1991). Furthermore, in patients undergoing surgery, postoperative pain ratings are positively correlated with preoperative Beck Depression Inventory scores (Taenzer et al., 1986). Although the circuitry underlying the pain enhancing effect of mood and attention is unknown, it is clear that brain regions involved in the generation of emotion (e.g. medial prefrontal, insular and anterior temporal cortex, hypothalamus and amygdala, see Chapter 17 by Saper and Chapter 16 by Brent Vogt and Sikes in this volume) project massively to

brainstem structures (PAG and RVM) involved in pain modulation.

Summary and conclusions To summarize, although there are multiple potential target nuclei for modulating pain transmission and several candidate efferent pathways that exert modulatory control, the most completely described pain modulating circuit includes the amygdala, PAG, DLPT and RVM in the brainstem. Through descending projections, this circuit controls both spinal and trigeminal dorsal horn pain transmission neurons and mediates both opioid and stimulation produced analgesia. Several different neurotransmitters are involved in the modulatory actions of this circuit, which exerts bi-directional control of pain through On cells that facilitate and Off cells that inhibit dorsal horn nociceptive neurons. There is evidence that this circuit contributes to analgesia in humans and may be activated by acute stress or the expectation of relief. Conversely, through the facilitating effect of On cells, this circuit is theoretically capable of generating or enhancing perceived pain intensity. Such an effect could provide a physiological mechanism for the pain enhancing actions of mood, attention and expectation.

Acknowledgement This work was supported by U.S.P.H.S. grants NS21445 and DA01949. References Basbaum, A.I. and Fields, H.L. (1984) Endogenous pain control systems: brainstem spinal pathways and endorphin circuitry. Annu. Rev. Neumsci., 7 : 309-338. Baskin, D.S., Mehler, W.R., Hosobuchi, Y., Richardson, D.E., Adams, J.E. and Flitter, M.A. (1986) Autopsy analysis of the safety, efficacy and cartography of electrical stimulation of the central gray in humans. Brain Res., 371: 231-236. Bayer, T.L., Baer, P.E. and Early, C. (1991) Situational and psychophysiologicalfactors in psychologically induced pain. Pain, 44:45-50. Benedetti, F. and Amanzio, M. (1997) The neurobiology of placebo analgesia: from endogenous opioids to cholecystokinin. Pmg. Neumbiol., 52: 109-25. Breslau, N., Davis, G.C. and Andreski, P. (1991) Migraine, psychiatric disorders, and suicide attempts: an epidemiological study of young adults. Psychiatry Res., 37: 11-23. Budai, D. and Fields, H.L. (1998) Endogenous opioid peptides acting at m-opioid receptors in the dorsal horn contribute to

252 midbrain modulation of spinal nociceptive neurons. J. Neurophysiol., 79: 677-687. Burkey, A.R., Carstens, E., Wenniger, J.J., Tang, J. and Jasmin, L. (1996) An opioidergic cortical antinociception triggering site in the agranular insular cortex of the rat that contributes to morphine antinociception. J. Neumsci., 16: 6612-6623. Bushnell, M.C., Duncan, G.H., Dubner, R., Jones, R.L. and Maixner, W. (1985) Attentional influences on noxious and innocuous cutaneous heat detection in humans and monkeys. J. Neurosci., 5: 1103-11 10. Davis, K.D., Kiss, Z.H.T., Luo, L., Tasker, R.R., Lozan0,A.M. and Dostrovsky, J.O. (1998) Phantom sensations generated by thalamic microstimulation. Nature, 391: 385-387. Duncan, G.H., Bushnell, M.C., Bates, R. and Dubner, R. (1987) Task-related responses of monkey medullary dorsal horn neurons. J. Neurophysiol., 57: 289-3 10. Emson, P.C., Corder, R., Ratter, S.J., Tomlin, S., Lowry, P.J.. Ress, L.H., Arregui, A. and Rosser, M.N. (1984) Regional distribution of pro-opiomelanocortin-derivedpeptides in the human brain. Neuroendocrinology, 38: 45-50. Fanselow, M.S. (1991) The midbrain periaqueductal gray as a coordinator of action in response to fear and anxiety. In: A. Depaulis and R. Bandler (Eds), The Midbrain Periaqueductal Gray Matter, Plenum Press, New York, pp. 151-173. Fields, H.L. and Basbaum, A.I. (1994) Central nervous system mechanisms of pain modulation. In: P.D. Wall and R. Melzack (Eds), Textbook of Pain, 3rd Edn, Churchill Livingston, New York, pp. 243-257. Fields, H.L., Heinricher, M.M. and Mason, P. (1991) Neurotransmitters in nociceptive modulatory circuits. Annu. Rev. Neurosci., 14: 2 19-245. Fields, H.L., Malick, A. and Burstein, R. (1995) Dorsal horn projection targets of ON and OFF cells in the rostral ventromedial medulla. J. Neurophysiol., 74: 1742-1759. Fields, H.L. and Price, D.D. (1997) Toward a neurobiology of placebo analgesia. In: A. Harrinton and J. Dowling (Eds), Placebo, Harvard University Press, Cambridge, MA, pp. 93-1 16. Fishbain, D.A., Goldberg, M., Meagher, B.R., Steele, R. and Rosomoff, H. (1986) Male and female chronic pain patients categorized by DSM-I11 psychiatric diagnostic criteria. Pain, 26: 181-197. Gelfand, S.,Ullman, L.P. and Krasner, L.I. (1963) The placebo response: an experimental approach. J. Nerv. Ment. Dis., 136: 379-387. Gracely, R.H., Dubner, R., Wolskee, P.J. and Deeter, W.R. (1983) Placebo and naloxone can alter post-surgical pain by separate mechanisms. Nature, 306: 264-265. Grevert, P. and Goldstein, A. (1985) Placebo analgesia, naloxone and the role of endogenous opioids. In: L. White, B. Tursky and G.E. Shwartz (Eds), Placebo: Theory, Research and Mechanisms, The Guilford Press, New York, pp. 332-350. Heinricher, M.M., Morgan, M.M., Tortorici, V. and Fields, H.L. (1994) Disinhibition of off-cells and antinociception produced by an opioid action within the rostralventromedial medulla. Neuroscience, 63: 279-288.

Helmstetter, F.J., Bellgowan, P. and Tershner, S.A. (1993) Inhibition of the tail flick reflex following microinjection of morphine into the amygdala. NeuroReporr, 4: 47 1-474. Helmstetter, F.J. and Tershner, S.A. (1994) Lesions of the periaqueductal gray and rostral ventromedial medulla disrupt antinociceptive but not cardiovascular aversive conditional responses. J. Neurosci., 14: 7099-7 108. Holstege, G. (1988). Direct and indirect pathways to lamina I. in the medulla oblongata and spinal cord of the cat. In: H.L. Fields and J.M. Besson (Eds), Pain Modulation, Progress in Brain Research, Elsevier, Amsterdam, pp. 47-94. Kiefel, J.M., Rossi, G.C. and Bodnar, R.J. (1993) Medullary p and 6 opioid receptors modulate mesencephalic morphine analgesia in rats. Brain Res., 624: 151-161. Kim, J.J., Rison, R.A. and Fanselow, M.S. (1993) Effects of amygdala, hippocampus, and periaqueductal gray lesions on short- and long-term contextual fear. Behav. Neurosci., 107: 1093-1098. Laska, E. and Sunshine, A. (1973) Anticipation of analgesia: a placebo effect. Headache, 13: 1-1 1. Levine, J.D. and Gordon, N.C. (1984) Influence of the method of drug administration on analgesic response. Nature, 312: 755-756. Levine, J.D., Gordon, N.C., Bornstein, J.C. and Fields, H.L. (1979) Role of pain in placebo analgesia. Proc. Natl. Acad. Sci. USA, 76: 3528-3531. Levine, J.D., Gordon, N.C. and Fields, H.L. (1978a) The mechanism of placebo analgesia. Lancet, 2: 654-657. Levine, J.D., Gordon, N.C., Jones, R.T. and Fields, H.L. (1978b) The narcotic antagonist naloxone enhances clinical pain. Nature, 272: 826-827. Liberman, R. (1964) An experimental study of the placebo response under three different situations of pain. J. Psychiatry Res., 2: 233-246. Ma, Q.P., Shi,Y.S. and Han, J.S. (1992) Naloxone blocks opioid peptide release in N. accumbens and amydgala elicited by morphine injected into periaqueductal gray. Brain Res. Bull., 28: 351-354. Melzack, R. and Wall, P.D. (1982) Acute pain in an emergency clinic: latency of onset and descriptor patterns related to different injuries. Pain, 14: 33-43. Merikangas, K.R., Angst, J. and Isler, H. (1990) Results of the Zurich cohort study of young adults. Arch. Gen. Psychiatry, 47: 849-853. Miron, D., Duncan, G.H. and Bushnell, M.C. (1989) Effects of attention on the intensity and unpleasantness of thermal pain. Pain, 39: 345-352. Pittius, C.W., Seizinger, B.R., Pasi, A,, Mehraein, P. and Herz, A. (1984) Distribution and characterization of opioid peptides derived from proenkephalin A in human and rat central nervous system. Brain Res., 304: 127-136. Roychowdhury, S.M. and Fields, H.L. (1996) Endogenous opioids acting at a medullary p, opioid receptor contribute to the behavioral antinociception produced by GABA antagonism in the midbrain periaqueductal grey. Neuroscience, 74: 863-872.

253 Taenzer, P., Melzack, R. and Jeans, M.E. (1986) Influence of psychological factors on postoperative pain, mood and analgesic requirements. Pain, 24: 331-342. Villanueva, L., Bernard, J.F. and Le Bars, D. (1995) Distribution of spinal cord projections from the medullary subnucleus reticularis dorsalis and the adjacent cuneate nucleus: a Phaseolus vulgaris-leucoagglutinin study in the rat. J. Comp. Neuml., 352: 11-32. Voudouris, N.J., Peck, C.L. and Coleman, G . (1990) The role of conditioning and verbal expectancy in the placebo response. Pain, 43: 121-128. Watkins, L.R. and Mayer, D.J. (1982) Organization of endogenous opiate and nonopiate pain control systems. Science, 216: 1185-1192. Watkins, L.R., Young, E.G.,Kinscheck, I.B. and Mayer, J.D. (1983) The neural basis of footshock analgesia: the role of

specific ventral medullary nuclei. Brain Res., 276: 305-315. Wells, K.B., Stewart, A., Hays, R.D., Burnam, M.A., Rogers, W., Daniels, M., Berry, S., Greenfield, S. and Ware, J. (1989) The functioning and well-being of depressed patients. Results from the Medical Outcomes Study. JAMA, 262: 914-919. Wolkowitz, O.M. and Tinklenberg, J.R. (1985) Naloxone's effects on cognitive functioning in drug-free and diazepamtreated normal humans. Psychophannacology, 85: 221-223. Zelman, D.C., Howland, E.W., Nichols, S.N. and Cleeland, C.S. (1991) The effects of induced mood on laboratory pain. Pain, 46: 105-111. Ziegler, D.K. and Paolo, A.M. (1995) Headache symptoms and psychological profile of headache-prone individuals. A comparison of clinic patients and controls. Arch. Neurol., 52: 602-606.

This Page Intentionally Left Blank

E.A. Mayer and C.B.Saper (Eds.) Progress in Brain Research, Vol 122 0 ZOO0 Elsevier Science BV.All rights reserved.

CHAPTER 19

Mechanisms of analgesia produced by hypnosis and placebo suggestions Donald D. Price* and James J. Barrel1 Departments of Oral and Maxillofacial Surgery and Neuroscience, University of Florida. 2210 NW 24th Avenue, Gainesville, FL 32605, USA

Introduction Studies of hypnotic and placebo analgesia have labored under a double burden. Both the independent variables of hypnotic and placebo treatments and the multiple components of pain experience, which are the dependent variables, are subjective phenomena. Partly as a consequence, precise measurement of subjective independent and dependent variables of hypnotic and placebo analgesia experiments may be considered to lack the precise control that is present in physiological or pharmacological studies. The purpose of this chapter is to present an alternative view of the possibility of precise analysis and measurement of both the independent and dependent variables associated with placebo and hypnotic analgesia. In the context of providing this alternative view, a general explanation of the neural and psychological mechanisms that may underlie these forms of analgesia will be presented. An understanding of these mechanisms has implications for the treatment and management of pain. Hypnotic analgesia

It is complicated to consider the mechanisms by which hypnosis may induce analgesia, because of

*Corresponding author. Tel.: 352 846 2718; Fax: 352 846 0588; e-mail: [email protected]

the difficulty in separating the factors that evoke pain reduction and in defining the nature of the pain reduction itself. The factors that evoke pain reduction extend from psychosocial, including interactions between therapist and patient, to neurophysiological factors that influence the actual transmission of pain signals within the patient. The following discussion will consider first, the demand characteristics of the hypnotic analgesia situation; second, the role of hypnotic state and the possible interactions between hypnotic state and incorporation of hypnotic suggestions, and third, psychological and neural mechanisms that help explain hypnotic analgesia. This discussion will include a consideration of how hypnotic interventions influence the multiple dimensions of pain and, at least in a general sense, the neural processing of pain. Is hypnotic analgesia simply the result of demand characteristics of the experiment or clinical situation? One major hypothesis about the nature of hypnotic analgesia is that the phenomenon represents compliance with demand characteristics of the experimental or clinical situation. Thus, after hypnotic induction and suggestions, subjects/ patients cognitively re-label their reports of pain as less intense, not because they perceive them to be less intense but simply to act in the role of someone

256

who has less pain (Spanos, 1986). There are two interrelated claims in this explanation. The first is that there is nothing special about the hypnotic situation and that a change in conscious state is not required to evoke responses to hypnotic suggestion. The second related claim is that there is nothing special about the hypnotic response of reducing one’s rating of pain intensity. It does not necessarily involve an actual reduction in pain but rather a willingness to use lowered ratings to describe unaltered pain sensations. According to the ‘role enactment’ theorists, the elaborate ritual of hypnotic induction only serves to strengthen the demand characteristics of the situation and encourage subjects to simply follow instructions and emit the desired behavior. Some studies show no difference in analgesia levels between two groups of subjects given analgesia suggestions with and without an induction of a hypnotic state (Barber and Hahn, 1962; Evans and Paul, 1985; Barber and Wilson, 1977). These studies indicate that a change in state of consciousness is not necessary for, nor does it contribute to, hypnotic analgesia. On the other hand, there are considerably more studies that demonstrate that greater analgesia occurs when subjects enter a hypnotic state and that hypnotic susceptibility is at least somewhat predictive of hypnotic analgesia (Hilgard and Hilgard, 1983). A related claim is that hypnotically induced reports of reduced pain do not necessarily reflect actual reductions in perceived pain. The claim is supported to a limited extent by observations that physiological responses to pain, such as increased heart rate and blood pressure, often still occur during hypnotically induced reports of greatly reduced pain (Hilgard and Hilgard, 1983). Are a hypnotic induction and/or a hypnotic state necessary for hypnotic analgesia? It is possible that a hypnotic state is not required to evoke reductions in pain report. It is also possible that subjects of hypnotic analgesia experiments are simply enacting an elaborate role. If both of these possibilities are true, then subjects who are deliberately instructed to simulate analgesia should be able to tolerate intense pain as well as those who

undergo the unnecessary ritual of hypnotic induction. Greene and Reyher (1972) tested this hypothesis by assigning highly hypnotizable subjects randomly to the hypnotized and simulating groups. They instructed the simulators to remain out of a hypnotic state while deceiving the hypnotist into believing they were hypnotized and to react to the painful stimulus as if they were analgesic. Pain tolerance and pain intensity reports were obtained in response to increasing electric shock intensities. Despite the attempt to behave like hypnotized subjects while not in fact hypnotized, the simulators were clearly less tolerant of the pain during each of the experimental conditions than the hypnotized subjects. For example, the truly hypnotized and the simulators increased their tolerance of pain by 45% and 16% respectively, a difference that was statistically significant. The shocks were more bearable for the truly hypnotized than for the equally hypnotizable but non-hypnotized role enactors. These results cast some doubt on the idea that hypnotized subjects feel pain but report less pain to satisfy the demand characteristics of the hypnotist. How are responses to suggestions facilitated by a hypnotic state? Hypnotic states are produced by hypnotic inductions, particularly in subjects who have high scores on standard tests of hypnotic susceptibility (Hilgard and Hilgard, 1983). In fact, hypnotic states have recently been shown (using positron emission tomographic brain imaging along with EEG) to be associated with a pattern of increases in activity in occipital cortical regions and decreases in posterior parietal cortical regions (Rainville et al., 1999). It is also clear from brain imaging and psychophysical studies that the presence of a hypnotic state is not sufficient to produce reduction in pain or painevoked activity in the cerebral cortex (Rainville et al., 1997, 1999). However, there is evidence that analgesia is greater in response to suggestions for analgesia when a hypnotic state is produced (Hilgard and Hilgard, 1983). Based on decades of both experimental as well as clinical observations, it is evident that hypnotic state can at least facilitate the analgesia produced by direct or indirect suggestions. This brings us to the general question as to

257

how a hypnotic state does this. On the basis of experiential and path analysis studies, Price and Barrel1 (1990) proposed the following common elements of a hypnotic state: A feeling of relaxation or inner calmness (not necessarily physical relaxation). An absorbed and sustained focus of attention on one or few targets. An absence of judging, monitoring, and censoring. A suspension of usual orientation toward time, location, and/or sense of self. Ones own responses are experienced as automatic (i.e. without deliberation and/or effort). Using visual analogue scale ratings of these factors, strong interrelationships were shown to exist among them, suggesting that some of the common elements are likely to be necessary for others (Price, 1996). Element 1. (‘relaxation, becoming at ease’) appeared to provide a supportive general background for element 2 (‘absorbed and sustained focus’) which in turn appeared to result in elements 3 (‘absence of judging, monitoring, censoring’) and 4 (‘suspensionof usual orientation toward time and location’). The latter two elements, in turn, appeared to maintain element 5 (‘automaticity’). Finally, elements 4 (‘suspension’) and 5 (‘automaticity’) directly contributed to perceived hypnotic depth. The experiential and conceptual basis for this model is also generally supported by the work of others who have independently arrived at many of these same common elements (Bowers, 1978; Pekula and Kumar, 1984). Returning then to the question as to how a hypnotic state facilitates incorporation of suggestions, such as that for analgesia, this model implicitly provides a basis for increased responsiveness to suggestion that is unique and distinguishablefrom other types of psychologically mediated increases in responsiveness to suggestion (e.g. placebo). This basis is directly evident in the phenomenology of the interrelationships between the common elements of this model as follows: A hypnotic state begins with an absorbed and sustained focus on something. It can occur during fascination, watching an absorbing movie, or watching ripples in a stream. It captures us. At first,

it can be effortful but with time one proceeds from an active form of concentration to a relaxedpassive form. There is often (though perhaps not necessarily) an inhibition or reduction in the peripheral range of ones experience. At the same time, this relaxation and/or reduction in range of attention supports a lack of monitoring and censoring of that which is allowed into experience. Hence, inconsistencies are now more tolerable. Contradictory statements, which once arrested attention and caused confusion or disturbance, now no longer do so. The uncensored acceptance of what is being said by the hypnotist is not checked against ones own associations. Consequently, one no longer chooses or validates the correctness of incoming statements. This allows thinking and meaning-initself that is disconnected from active reflection. From this way of experiencing, there emerges the sense of automaticity wherein thinking doesn’t precede an action but action precedes thought. Thus, if the hypnotist suggests a bodily action, a sensation, or a lack of sensation (e.g. pain), there is no experience of deliberation of effort on the part of the subject. The subject simply and automatically identifies with the suggested action, sensation, or lack of sensation, as it is suggested. The possibility of not carrying out the action or experiencing the suggested changes in sensation is not considered or is considered very little. In this way, a hypnotic state facilitates the incorporation of and responses to suggestions, including that of analgesia. Suggestions of pain reduction are more believable in the absence of monitoring, censoring, or reflection normally present in ordinary states of consciousness. The role of hypnotic state in facilitating analgesia in response to suggestions is supported by a considerable number of studies (Hilgard and Hilgard, 1983). It is important to emphasize, however, that suggestions for analgesia are alone sufficient to produce analgesia in at least some individuals. Some people that score high on standard tests of hypnotic susceptibility readily develop analgesia in response to suggestions, regardless of whether the suggestions are given during ‘waking’ or hypnotic conditions (Tenenbaum et al., 1990). It may well be that what is unique about the hypnotic factors required for pain reduction is not the state of

25 8

consciousness nor even the suggestions, but the way in which the suggestions implicitly or explicitly refer to the source of experiential change as coming automatically from within. The lack of censoring in combination with a sense of automaticity provides an experiential context in which subjects or patients can more readily incorporate suggestions and develop a tendency to respond to suggestions. The considerations raised so far indicate that multiple factors within the psychosocial context and within the experience of subjects influence the alteration of pain as a result of hypnotic suggestions for analgesia. The influences of individual factors that comprise the hypnotic state on responses to various types of hypnotic suggestion have received relatively little explicit recognition. The relative influence of elements within the hypnotic state on analgesic responses to suggestions could be tested in experiments wherein participants provide judgments of their experienced magnitudes. For example, self-ratings of such elements as ‘automaticity’ and ‘depth of hypnotic state’ could constitute independent variables in studies of the relationship between depth of hypnotic state and responsiveness to analgesic suggestions. What are the types of hypnotic suggestions for analgesia? The suggestions for alteration of the experience of pain in studies of hypnotic analgesia relate closely to the dimensions of pain and to the psychological stages of pain processing. Thus, there are suggestions that specifically target the affective-motivational dimension of pain as distinguished from the sensory-discriminative dimension. These would include suggestions for reinterpreting sensations as pleasant rather than unpleasant or suggestions for reducing or eliminating the implications of threat or harm from the sensations. Then there are suggestions designed to specifically alter the quality andor intensity of sensations so that they become less painful, not at all painful, or absent altogether. These would include suggestions for replacing sensations of pain with those of numbness, warmth, or other sensa-

tions as well as suggestions for the complete absence of sensation. The latter include suggestions for dissociation wherein subjects do not feel parts of their bodies that would otherwise be painful or wherein subjects experience themselves in another location and context altogether. Just as studies are needed to assess the role of hypnotic depth and individual components of hypnosis on pain, so also there need to be studies of differential effects of various types of suggestion on sensory and affective dimensions of pain experience. For example, what are the effects on pain of suggestions designed to reinterpret the meanings of the sensations so that they are less threatening or unpleasant? How does hypnosis affect the sensory and affective dimensions of pain? Just as there are multiple psychological dimensions that contribute to hypnotic analgesia, analgesia itself is likely to be comprised of multiple dimensions. Pain has sensory-discriminative, cognitive-evaluative, and affective-motivational dimensions. The strategy discussed above for characterizing and measuring the factors in hypnotic treatments could be interfaced with one which assesses how these factors influence the three dimensions of pain as well as the different general stages of pain processing (i.e. spinal, cortical, etc.). Since pain is comprised of these three dimensions, then a number of questions can be raised about how a hypnotic intervention influences the various dimensions and stages of pain processing. Does hypnotic analgesia reduce the affective dimension more than the sensory dimension of pain? To what extent does hypnotic analgesia involve descending inhibition of pain transmission at spinal cord levels or intracortical mechanisms that prevent awareness of pain? Does hypnotic analgesia utilize an endogenous opioid system? As surprising as it may seem, these questions are at least partly answerable in experiments that utilize multiple measures of pain experience and, in some cases, physiological indices of pain processing at different levels of the nervous system. The question as to whether differential effects of hypnotic suggestion on sensory and ugective

259

dimensions was addressed in a study of the factors that contribute to the magnitudes of reduction in these dimensions after indirect hypnotic suggestions (Price and Barber, 1987). %o groups of human volunteers made responses on extensively validated visual analogue scales (VAS) of pain sensation intensity (sensory VAS) and pain unpleasantness (affect VAS) to noxious skin temperature stimuli (44.5 to 51.5"C) before and after hypnotic suggestions were given for analgesia. Group 1 was given suggestions for developing a hypnotic state only once just before analgesic testing and did not have significantly reduced VAS responses to experimental pain after hypnosis. Group 2 was continuously given cues for maintaining a hypnotic state during their analgesic testing session and had a 44.5% mean overall reduction in pain sensation intensity and a 87.4% mean overall reduction in pain affect. As shown in Table 1, the reduction in pain affect was much larger and more consistent across group 2 subjects than reduction in pain sensation intensity. Similar to several previous studies (Hilgard and Hilgard, 1983), a small but statistically reliable correlation was found between hypnotic susceptibility test scores and overall magnitude of reduction in VAS-sensory ratings but not VAS affect ratings (Table 1). The hypnotic

susceptibility test used (Stanford clinical scale for adults) assesses the degree of both ideational and motoric responses to hypnotic suggestions. It is not immediately apparent why hypnotically induced reductions in VAS affect ratings to experimental pain were greater and more consistent than reductions in VAS sensory ratings. Affective responses to experimental heat pain were reduced even in participants who had low susceptibility scores and who had very little change in perceived sensation intensities (Table 1). A possible answer to this question may be obtained through consideration of the nature of sensory and affective responses to experimental pain and of the degree of hypnotic involvement required to experience alterations in these pain dimensions. Affective responses associated with pain are more influenced by the perceived context of the experimental situation than are sensory responses (Price, 1988). Thus, factors related to the psychological context of the person, such as those related to the immediate implications of the pain and expectations, can selectively and often powerfully reduce affective responses to experimental pain (Price et al., 1980; Price, 1988). Hypnotic suggestions of this study were directed toward; (a) experiencing the testing situation as more pleasant, (b) experiencing the heat stimuli as

TABLE 1 Hypnotic susceptibility and analgesia Subject

Hypnotic susceptibility

%VAS sensory change

%VAS affective change

1 1 1 1 1 1 2 2 2 3 3 3 3 4 5 5

56.8 9.0 65.5 63.1 3.6 60.0 17.8 17.5 20.7 28.3 24.7 50.0 46.0 82.7 83.9 81.7

74.5 90.0 100.0 100.0 100.0 100.0 91.4 78.6 51.5 78.0 74.6 100.0 100.0 99.1 62.5 98.7

44.5

87.4

~

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

M=U

.4

- .2

260

more pleasant, and (c) experiencing the heat stimuli as less intense. It is clear that these three alterations in experience would require different degrees of hypnotic involvement (Weitzenhoffer, 1957; Shor and Ome, 1965). Experiencing the testing situation and test stimuli as less unpleasant would require less hypnotic involvement than experiencing direct reductions in sensations evoked by noxious heat stimuli. In some instances, selective reduction in affect could occur without a hypnotic state. Therefore, one component of a hypnotic intervention may involve responses to suggestion for reduced unpleasantness that do not require a hypnotic state. This explanation is consistent with the result that reductions in unpleasantness were not at all correlated with hypnotic susceptibility. The reduction in pain-related unpleasantness beyond that accountable by a simple reduction in pain sensation intensity, and in some cases without any reduction in pain sensation (e.g. Table l), is not likely the result of reduction of the pain signal at peripheral or even spinal levels. Rather, it is likely the result of alteration in the meanings that normally attend painful experience. As such, the selective reduction in pain affect by cognitive mechanisms is likely to reflect neural events at higher levels of pain processing, including intracerebral mechanisms. What are the mechanisms of hypnotically induced reductions in pain sensation intensity? Although it is clear that hypnotic suggestions may exert a more powerful reduction of pain affect than pain sensation, it is also quite apparent that both dimensions are reduced, as has been amply demonstrated in several experimental laboratories (Hilgard and Hilgard, 1983; Price, 1988). Moreo-

ver, it is the reduction in pain sensation itself that is statistically associated with hypnotic susceptibility, albeit at modest levels. Therefore, the component of a hypnotic intervention that relies on hypnotic ability and a hypnotic state is the one most influential on pain sensation intensity. Interestingly, the association becomes stronger with increasing levels of pain intensity, as shown in Table 2. To the extent that hypnotic analgesia represents a negative illusion, it makes sense that the reduction of stronger pains requires more hypnotic ability than the reduction of weaker pains (Weitzenhoffer, 1957). The modest correlation between hypnotic susceptibility and sensory analgesia and the lack of significant correlation between hypnotic ability and reductions in pain affective ratings (Table 2) reflect multiple factors involved in hypnotic analgesia. These factors include all of those that result from the hypnotic intervention, including some that are unrelated to hypnotic susceptibility and perhaps even to a hypnotic state. Several multiple factors are potentially related to different proposed meehanisms of hypnotic analgesia. General neural mechanisms. Based on current knowledge, there are two general mechanisms by which pain sensations could be reduced in intensity during hypnosis. Ernest and Josephine Hilgard ( 1983) proposed that during hypnotic analgesia there is reduced awareness of pain that occurs after nociceptive information has reached higher centers. According to this neo-dissociation theory, the body registers the pain and there is covert awareness of pain during hypnotic analgesia, yet it is masked by an amnesia-like barrier between dissociated streams of consciousness. This dissociation in awareness has been demonstrated through 'auto-

TABLE 2 Hypnotic susceptibility, analgesia, and stimulus intensity Stimulus temperature

44.5"C 47.5"C 49.5"C 51.5'C

Sensory analgesia Spearman correlation

Affective analgesia

+ 0.04

- 0.23

+0.21

-0.11 - 0.08 +0.10

+ 0.43* +0.56* (*p<0.05)

coefficient

26 1

matic writing’ and through the phenomenon of the ‘hidden observer’ (Hilgard et al., 1975; Hilgard, 1977; Hilgard and Hilgard, 1983). Neo-dissociation and ‘intracerebral’ mechanisms. Hilgard and colleagues had subjects report covert levels of cold pressor pain through automatic key pressing ratings of pain, whereas their overtly experienced pain produced by the same cold pressor was reported through magnitude estimation (Hilgard et al., 1975). The result was that during non-hypnotic suggestions for analgesia there was about a 40% reduction in both overtly and covertly reported pain intensity. Suggestions for analgesia after inducing a hypnotic state produced an additional reduction in overtly reported pain but not covertly reported pain. This additional reduction is said to reflect amnesia or dissociative mechanisms that are available only when a hypnotic state is induced. A component of the perception of pain may be immediately forgotten or somehow diverted from conscious awareness. This interpretation of hypnotic analgesia as dissociation in consciousness suggests an explanation for the paradox that physiological indices of stress often continue during hypnotic analgesia, even though the subject consciously feels little or no pain. Interestingly, Hilgard and colleagues found that with highly hypnotizable subjects the rise in heart rate caused by cold pressor pain was somewhat less than during waking non-hypnotic control conditions but that some increase in heart rate still occurred (Hilgard and Hilgard, 1983). They explained this partial reduction as reflecting two components of pain reduction. The first was said to result from nonhypnotic factors and was accompanied by reductions in autonomic and reflex responses to pain. The second component of hypnotic analgesia was considered to be associated with dissociative mechanisms and would not be related to decreases in autonomic responses. Descending spinal cord inhibitory mechanisms. A second general mechanism by which hypnotic suggestions could reduce pain is by activation of an endogenous pain inhibitory system that descends to the spinal cord where it prevents the transmission of pain-related information to the brain. There are multiple lines of indirect evidence for and against

such a mechanism. The question of whether hypnotic analgesia involves a brain-to-spinal cord descending control mechanism is indirectly related to another question as to whether endogenous opioids mediate hypnotic analgesia. If they do, it would support the hypothesis that a descending control system is involved since it has been well established that opioid analgesic mechanisms rely heavily on a brain-to-spinal cord descending control system. A number of observations indicate that hypnotic analgesia does not depend on endogenous opioid mechanisms. First, different groups of investigators have found that naloxone, an opioid antagonist, does not reverse analgesia produced by hypnotic suggestions. For example, Barber and Mayer (1977) found that hypnotic suggestions elevated pain thresholds produced by tooth-pulp stimulation and that these elevations in threshold were completely unaffected by naloxone. Goldstein and Hilgard (1975) obtained similar negative results. Other characteristic differences also exist between opioid analgesia and hypnotic analgesia. Once analgesia is repeatedly established in a highly hypnotizable subject, it can be induced again very quickly (sometimes within seconds) in the same subject and can also be quickly terminated. Endogenous opioid mechanisms, by contrast, typically have a delayed onset to maximum effect (e.g. several minutes) and are slow to dissipate. However, the lack of demonstration of an endogenous opioid mechanism involved in hypnotic analgesia does not exclude the possibility of a descending control system, only that of an opioid descending control mechanism. Non-opioid brain-to-spinal cord descending control mechanisms have been identified (Fields and Basbaum, 1994). Although physiological studies address the possibility that hypnotic analgesia involves a brain-to-spinal cord descending inhibitory mechanism, they focus on autonomic (Barber and Hahn, 1962), neurochemical (Goldstein and Hilgard, 1975; Mayer et al., 1976; Barber and Mayer, 1977), or electrocortical changes associated with hypnotic analgesia (Crawford and Gruzelier, 1992). A limitation common to all of these studies is that it is difficult to identify the general neuroanatomical sites at which the relevant modulatory mechanisms take place. Evi-

262

dence that hypnotic analgesia involves a descending system that inhibits nociceptive processing at spinal levels could be simply provided if some measure of spinal nociceptive function and pain perception could be simultaneously provided during hypnotic analgesia. The feasibility of such an approach is indicated by considerable studies of Willer (1984, 1985) who has shown different types of somatosensory stimulation and attentional manipulations simultaneously reduce pain and the electrically evoked flexion reflex. He also has demonstrated that graded doses of morphine reduce the electrically evoked flexion reflex and pain intensity on a near equal percentage-wise basis, thereby providing a standard of assessing descending pain inhibitory mechanisms. All of these observations point to the possibility of simultaneous measurement of pain and the flexion reflex during hypnotic analgesia. Indeed, this idea may have originated in Hagbarth and Finer’s (1963) preliminary demonstration of marked suppression of the flexion reflex in a few subjects during hypnotic analgesia. Though this result is provocative, it was demonstrated in subjects who were aware of the physiological response that was measured. A more recent and extensive analysis of the question of a possible descending inhibitory mechanism of hypnotic analgesia was made by examining changes in the R-111, a nociceptive spinal reflex, during hypnotic reduction of pain sensation and unpleasantness (Kiernan et al., 1995). The R-I11 was measured in 15 healthy volunteers who gave VAS-sensory and VAS-affective ratings of an electrical stimulus during conditions of resting wakefulness without suggestions and during hypnosis with suggestions for hypnotic analgesia. A critically important feature of this study was that subjects were blind to the physiological index being measured, and, when later informed that measurements were being made of the R-111 flexion reflex, failed to intentionally reduce the magnitude of this reflex. Hypnotic sensory analgesia was partially yet reliably related to reduction in the R-111 (R-square=0.51, p c 0.003), suggesting that hypnotic sensory analgesia is at least in part mediated by descending anti-nociceptive mechanisms that exert control at

spinal levels in response to hypnotic suggestion. Hypnotic affective analgesia was not quite significantly related to reduction in R-III (p=0.053). Reduction in R-III was 67% as great as, and accounted for 51% of the variance in reduction of pain sensation. In turn, reduction in pain sensation was 75% as great as, and accounted for 77% of the variance in reduction of unpleasantness. The results suggest three general mechanisms may be involved in hypnotic analgesia: The first, implicated by reductions in R-111, is related to spinal cord antinociceptive mechanisms. The second, implicated by reductions in pain sensation over and beyond reductions in R-111, may be related to brain mechanisms that serve to prevent awareness of pain once nociception has reached higher centers, as suggested by the neo-dissociation theory discussed above (Hilgard and Hilgard, 1983). That the percent reduction in pain sensation intensity was greater than that of the R-I11 is consistent with Hilgards’ and others’ finding that some autonomic responses to pain remain even under deep hypnotic analgesia. The third, implicated by reductions in unpleasantness over and beyond reductions in pain sensation, may be related to selective reduction in the affective dimension, possibly as a consequence of reinterpretation of meanings associated with the painful sensation, as previously suggested by Price and Barber (1987). This analysis strategy provided in this study is useful because it conceptualizes the possible multiple actions of a hypnotic pain control intervention and it provides a strategy for independent evaluation of multiple stages of pain processing. Similar to the work of Hilgard and Hilgard (19831, this study shows that multiple components of pain reduction are produced by a hypnotic intervention. Most critically, the study demonstrates that hypnosis has some measurable effect on a person beyond the worse case scenario of only changing the verbal labels used to describe an unaltered pain experience. How are mechanisms of hypnotically induced analgesia clarified through the use of Positron Emission Tomography? Even more precise analysis of the central nervous system levels at which pain processing is

263

modulated by hypnotic suggestion is provided by two recent studies that have used positron emission tomography (PET). These studies address questions si.mi1a.r to that of Kiernan et al. (1995) but analyzed neural images of cerebral cortical activity rather than the pain-related flexion reflex (R-III). In both PET studies, pain was experimentally induced by immersion of the left hand in a 47°C water bath for approximately one minute. Significant pain-related activation occurred within somatosensory area I, anterior cingulate cortex (area 24), and anterior insular cortex both before and during hypnosis. In the first study, hypnotic suggestions were given to selectively alter the affective dimension of pain without changing the perceived intensity of the painful sensation (Rainville et al., 1997). The experimental conditions included immersion of the left hand in moderately painful water (47°C) without hypnosis, with hypnosis but without suggestions, with hypnosis with suggestions for increased unpleasantness, and finally with hypnosis with suggestions for decreased unpleasantness. These manipulations were successful in providing much larger unpleasantness ratings during the high unpleasantness condition as compared to the low unpleasantness condition, but no differences in pain sensation ratings. Consistent with these changes, no differences occurred across high and low unpleasantness conditions in somatosensory area I, which is one region wherein sensory processing relevant for pain is considered to occur. In striking contrast and consistent with unpleasantness ratings, activity in anterior cingulate cortical area 24 was much greater in the high as compared to the low unpleasantness condition. A separate regression analysis, controlling for factors such as individual differences in global cerebral blood flow and pain sensation intensity ratings, showed that pain unpleasantness ratings were significantly associated only with anterior cingulate activity in a specific region of area 24 (R=0.55,p
that pain thresholds are unchanged after prefrontal lobotomy yet preoccupation with and rumination about clinical pain appear substantially diminished (Hardy et al., 1955). The second hypnosisPET study was designed exactly like the first, except that the suggestions were selectively targeted toward the sensory intensive dimension of pain (Hofbauer et al., 1998). However, the suggestions were effective in modulating both sensory and affective dimensions of pain experience, as measured by subjects' ratings of these dimensions. Unlike the first study, both activity in somatosensory area I and pain sensation ratings were higher in the high as compared to the low sensory intensity condition. There was a similar but non-significant trend in the anterior cingulate cortex. The most important result of the second study is that it provides an independent confirmation of the hypothesis that hypnotic suggestions for sensory analgesia result, at least in part, in a reduction of nociceptive afferent input into S-1 cortex. This result is complementary to that of the flexion reflex study of Kiernan et al. (1995). The combined results of PET and R-111 flexion reflex studies demonstrate multiple neural mechanisms for pain modulation. These include an inhibition of afferent nociceptive signals originating within the spinal cord and arriving at somatosensory cortex as well as a cerebral mechanism that is selective for modulation of pain affect within cortical areas heavily interconnected with limbic system circuits (e.g. anterior cingulate cortex). From a psychological perspective, this latter mechanism is likely to involve changes in the meaning of pain and does not necessarily require large hypnotic ability or involvement, as pointed out above. Placebo analgesia Many of the same issues and strategies for understanding the mechanisms of hypnotic analgesia apply at least in a general way to that of placebo analgesia. We have presented our views about the possible psychological and neural mechanisms of placebo analgesia elsewhere (Price and Fields, 1997; Fields and Price, 1997). Therefore, the following discussion will briefly summarize these

264

views, elaborate on some of the possible mechanisms, and then focus on the possible similarities and differences between placebo analgesia and hypnotic analgesia. The desire/expectation hypothesis Under conditions wherein patients have a strong need to be relieved of pain and/or they have expectations that pain relief will occur as a result of a treatment, and/or the treatment situation reproduces in some way a previously effective treatment, pain reduction may result from psychological factors. This constitutes the placebo analgesic response. A better understanding of how placebo manipulations (e.g. saline injections) can reduce pain requires explicit attention to determine which dimensions of pain are most affected by such manipulations and what are the most important psychological factors that contribute to the placebo effect. We have proposed that two general factors mediate placebo analgesia: (a) a desire or need for relief of pain; (b) an expectation that a given procedure or agent will relieve the pain. What are the contributions of classical conditioning and expectation to placebo analgesia? Prior experience with an effective analgesic or other types of learning are likely to result in expectations about the efficacy of any agent given to relieve pain, including a placebo agent. One hypothesis about the learning involved in placebo analgesia is that it represents an instance of classical conditioning. Consistent with the classical conditioning hypothesis, a landmark human study clearly showed that prior experience with an effective analgesic drug enhances the analgesic effectiveness of a subsequent placebo (Laska and Sunshine, 1973). In this study, placebo given as a second treatment was more effective as an analgesic when it followed a more potent analgesic. Their results support learning, even classical conditioning as a major factor in placebo analgesia. However, this result does not distinguish between the clas-

sical conditioning hypothesis and expectancy as a mediating factor in placebo analgesia. A recent experiment was designed to test alternative hypotheses related to the stimulus substitution model of classical conditioning and expectancy. Montgomery and Kirsch (1996b) tested opposing models of classical conditioning (the stimulus substitution variant) and expectancy by using an experimental paradigm in which a placebo skin cream was administered under the guise of a strong analgesic. Moderately painful shocks to the skin were given before administration of the cream. Manipulation (i.e. conditioning) trials that followed then consisted of giving surreptitiously lowered shock intensities after applying the cream. In the post-conditioning trials, they raised stimulus strength back to its original baseline value and administered these shocks to a different skin location after applying the cream to the new location. The mean decrease in pain report during these post-conditioning trials (as compared to the original baseline) was taken as a measure of the placebo effect. However, they provided two conditions of unconditional stimulus (lowered stimulus strength)-conditioned stimulus (placebo cream) pairings. These two conditions included informing one group of subjects about the lowering of painful stimulus intensity and not informing the other group. The uninformed group was thereby provided with an experience of cream-induced analgesia during conditioning and, as expected, demonstrated placebo analgesia in the subsequent test trials. In contrast, expectations of analgesia were lowered in the informed group and no overall placebo analgesic effect occurred in this group. Furthermore, although conditioning trials significantly enhanced placebo responding, this effect was eliminated by adding expectancy values of the subjects to the regression equation. This experiment shows that conditioning is sufficient for placebo analgesia but that its effect in this experiment is mediated by expectancy. Finally, the magnitude of the placebo effect increased significantly over 10 extinction trials, a result opposite to that predicted by a stimulus substitution model of classical conditioning. Although the results of this study do not refute the role of conditioning, they indicate that expectancy is necessary for conditioning to exert its

265

effects and they provide evidence against the role of the automatic stimulus substitution model of classical conditioning in placebo analgesia. What is the role of desire for relief or motivation in placebo analgesia? Although expectation may be a salient factor that influences the magnitude of the placebo response, it does not appear to operate alone. Placebo effects are commonly observed in circumstances wherein it is likely that subjects not only expect therapeutic effects but also strongly want these effects to occur. Therefore, a similar approach needs to be taken to evaluate the additional mediating role of motivation or desire for pain relief. Although direct evidence for the mediating role of desire for relief has yet to be obtained, the possible contribution of this factor in placebo analgesia is supported by some indirect evidence. First, there is evidence that the magnitude of placebo analgesia is influenced by the degree of threat that is present in the context in which placebo treatments are given. Presumably, the degree of threat would contribute to desire for pain relief. Based on comparisons of placebo analgesic effects across studies of different types of pain, both Beecher (1955; 1959) and Jospe (1978) asserted that the magnitudes of analgesic response to an explicit placebo manipulation are, in general, much greater in studies of clinical pain than in studies of experimental pain. Although both authors base this assertion on considerable numbers of studies, a serious limitation in this comparison is that the natural history of the subjects’ pain is rarely taken into account in clinical studies. However, among studies utilizing experimental pain, for which assessments of the natural history and/or baseline reliability are available, placebo analgesic effects are larger for those forms of experimental pain that are of longer duration and/or are more stressful (Jospe, 1978). These types of experimental pain are more likely to simulate the psychological conditions of most acute clinical pains. Placebo treatment produces large reductions in experimental limb ischemic pain, a form of pain that continuously increases in intensity over several minutes (Grevert et al., 1983; Grevert and Goldstein, 1985). In contrast, placebo treatments have

little or no effects on brief pains produced by 5 sec heat stimuli applied to the skin (Price et al., 1985) or one sec electrical stimuli applied to the tooth pulp (Gracely, 1979). These results suggest a significant role for desire for relief in placebo analgesia. What are the possible mechanisms of pain modulation by desire and expectation? We have provided evidence for the possibility that both pain and placebo analgesic responses are strongly influenced by desire for pain relief and expectancy. However, questions remain as to how these two factors influence pain and pain relief. In this section, the possible general mechanisms by which these factors could mediate pain and placebo analgesia are discussed. One possibility is that desire and expectation are integral components of pain affect and reduction in pain affect by placebo administration. The other is that desire for pain relief and expectancy produce a response bias that is reflected in changes in the way people report pain ratings, pain thresholds, or pain tolerance without an actual change in strength of pain-related signals in ascending pathways for pain. The third is that desire for pain relief and expectancy somehow are associated with brain mechanisms that trigger descending modulation of the pain-related signal. The following discussion focuses on these possibilities and some of their implications. Desire for pain relief and expectancy are direct components of pain and pain reduction. Desire for pain reduction and expectancy could be integral components of pain and placebo analgesia if two conditions were fulfilled:

(a) desire for pain relief and expectancy were components of pain-related affect itself, and (b) placebo administration directly influenced pain affect through changes in these two factors. There is empirical support for the role of these two factors in pain affect (Price and Barrell, 1999, in press). The mechanism by which these two factors effect pain and placebo responses could be made more clear if it could be shown that desire and expectation are components of emotions in general and hence in emotions that operate during pain.

266

One study has shown a selective effect of a placebo on the affective dimension of pain. Gracely (1979) found that a saline placebo injection reduced the affective but not the sensory ratings of experimental pain. This effect was greater toward the low end of the nociceptive stimulus range, a pattern of effect similar to that produced by an anxiety-reducing agent (Gracely et al., 1976) and by cognitive manipulations likely to reduce anxiety (Price et al., 1980). Given the likely associations between desire for relief, uncertainty of relief, and anxiety, the effect of some placebo treatments may be mediated through a reduction in uncertainty a n d or in desire for relief and hence a reduction in anxiety. This possibility would be consistent with the proposed associations between placebo effects and anxiety reduction (Bootzin, 1985;Evans, 1985; White et al., 1985). Desire for and expectation of pain relief produce a response bias. The explanation of a general placebo mechanism that reduces only pain-related affect is at variance with evidence for placebo mechanisms that reduce both pain intensity and pain affect. Furthermore, there exists evidence that placebo-induced reductions in pain may include multiple components and that not all placebo manipulations are selective for pain affect. Thus, some studies have shown placebo effects on sensory intensive aspects of pain (Levine et al., 1978, 1979; Grevert and Goldstein, 1985) and others have shown placebo effects on both sensory and affective dimensions of pain (Montgomery and Kirsch, 1996b). A study by Montgomery and Kirsch (1996a) provides some evidence that psychological mediation of placebo analgesic effects involves much more specific mechanisms than the simple reduction of anxiety or other global effects on emotions. They demonstrated that the application of a placebo in the guise of a topical anesthetic produced reduction in pain only at the body site at which the placebo anesthetic agent was administered. Controlled mechanical pain stimuli were administered simultaneously to treated and untreated fingers for one group of participants and sequentially for another. For both groups, reduction in pain occurred only on the finger that was treated with

the placebo anesthetic, thereby indicating a highly specific mechanism. This result suggests that placebos cannot all be mediated by such global mechanisms as anxiety reduction or endogenous release of opioids. Such specificity in response would be consistent with a highly specific response expectancy on the part of the participant (Kirsch, 1990). Their results indirectly suggest a second general mechanism for placebo analgesia, the development of a response bias. This latter possibility has been more directly tested using signal detection theory. Clark (1969) demonstrated that the administration of a placebo described as a strong analgesic increased the criterion for labeling a stimulus intensity as painful (also considered a measure of response bias) without changing the capacity to distinguish different levels of painful stimulus intensity. Unfortunately, there have been no other studies of placebo analgesia that have a bearing on the possibility that placebo analgesia is mediated by response bias. Nevertheless, such a possibility is consistent with, though not confirmatory of, the hypothesis that placebo analgesic responses are mediated by the development of highly specific response expectancies. Kirsch has provided the alternative explanation that response expectancies, defined as the anticipation of non-volitional responses, are capable of eliciting the expected response in much the same way that intentions elicit voluntary behaviors. However, it is not clear how Kirsch’s proposal differs from the hypothesis that specific expectations and desires lead to the development of a response bias or perceptual bias. Compliance with demand characteristics of the experimental or clinical situation can also be conceptualized as a type of response bias that could mediate apparent placebo analgesic responses. In such cases, patients/participants would reduce their ratings of pain after a placebo treatment in response to both their desire to please the person administering the treatment and their expectation that reduction of pain ratings would do so. Desire for and expectation of pain relief trigger descending control of pain signals. The possibility of a descending control mechanism of placebo analgesia is consistent with some evidence that,

261

under certain conditions, placebo analgesia can be mediated physiologically by endogenous opioid systems. The evidence partly consists of demonstrations that placebo-induced reductions of experimental and clinical pain can be reversed or antagonized by naloxone, an opioid antagonist (Levine et al., 1978, 1979; Grevert et al., 1983). The evidence also partly consists of more recent experiments that demonstrate that blockade of cholecystokinen receptors potentiates the placebo analgesic effect, thereby suggesting an inhibitory role of cholecystokinen in placebo analgesia (Benedetti and Amanzio, 1997). Since cholecystokinen is known to inhibit exogenous and endogenous opioid analgesic effects, then potentiating effects of cholecystokinen antagonists on placebo analgesia further implicate opioid mechanisms in placebo analgesia. At a psychological level, desire for and expectation of pain relief may be necessary for placebo analgesia. However, if placebo analgesia is mediated physiologically by endogenous opioid control systems, then brain states associated with desire and expectation of pain relief may be sufficient to trigger descending control systems that rely on opioid mechanisms (see Chapter 18 by Fields). If so, then the reduction in pain would represent an anti-nociceptive effect (i.e. reduction of the ascending signal for pain) and not a production of a response bias or a general selective change in affective state. Moreover, the reduction would occur both in sensory and affective dimensions of pain because it is likely that descending modulatory effects target both dimensions at once (Price, 1988). These three general mechanisms (direct change in affective state, response bias, and activation of descending control mechanisms) do not exclude the hypothesis that desire and expectation are proximate psychological mediators of pain affect or placebo-induced pain reduction (or at least report of pain reduction). However, from the foregoing discussion, it is apparent that desires and expectations may be targeted toward different responses or aspects of future experience. Thus, the desire and expectation may be focused on the avoidance of negative consequences associated with the pain condition, the avoidance of the pain sensation itself,

or even on the avoidance of disappointing the person who administers the therapeutic agent. The possibility must be considered that there are multiple types of responses to placebo analgesic manipulations because of multiple types of desires and expectations that can be developed. Clearly, different desires and expectations could be induced in study participants by the way placebo suggestions are framed and even the same placebo instructions could lead some people to desire and expect reductions in unpleasantness and others to desire and expect reductions in both pain sensation intensity and unpleasantness. Furthermore, some people may expect global changes whereas others may expect highly specific locations andor types of therapeutic effects. Still others may simply desire and expect to please the person administering the therapeutic agent. This potential diversity in types of placebo responses is indirectly supported by studies of hypnotic analgesia wherein these same general possibilities have been proposed to explain variation in changes in pain reports and behavior after hypnotic analgesic interventions (Barber and Adrian, 1982; Price and Barber, 1987; Kiernan et al., 1995). Despite this potential diversity in types of desires and expectations that may mediate different types of placebo responses, the desire/expectation model may provide the most parsimonious model of factors that mediate placebo responses. It may partly account for the diversity of types of emotional, perceptual, and behavioral responses to placebo manipulations. To the extent that one can conceptualize placebo responses in terms of changes in emotions, changes in tendency to respond, or learned changes in pain responsiveness, the potential application of a desire/expectancy model to all of these possible variants of placebo analgesic response appears promising and testable. What are the differences between hypnotically mediated and placebo mediated analgesia? Several general differences appear to exist between hypnotically mediated and placebo mediated analgesia. For example, at least some types of placebo analgesia may be mediated by endogenous opioids, whereas hypnotic analgesia does not depend on an

268

opioid mechanism. However, the greatest distinctions between the two types of analgesia appear to be psychological. Hypnotic analgesia depends on a self-generated automatic expectancy, whereas placebo analgesia is based on a perception of the therapeutic agent as fulfilling a desire and an expectation of pain relief. This distinction between the two types of psychologically mediated analgesia perhaps can be illustrated by comparing the nature of hypnotic suggestions with those that occur during placebo administration. The ‘suggestion’ provided in the case of placebo analgesia can readily be distinguished from that provided during hypnotic analgesia in that the former refers to an outside authoritative source as the origin of the pain relief. For example, injections or tablets provided by a healthcare professional indicate that the agency of therapeutic relief comes from a medicine and a person experienced in the knowledge of the efficacy of the treatment. Within the placebo literature, there exists some evidence that greater placebo effects occur as a result of more believable and technically convincing agents. Thus, placebo injections are more effective than placebo pills and placebo morphine is more effective than placebo aspirin (Traut and Passarelli, 1957). Implicit in the overall suggestion inherent to a placebo analgesic manipulation is the idea that in the absence of an outside authoritative agent, there would be unrelieved pain. The nature of hypnotic suggestions for analgesia, on the other hand, refers to a more innate and self-directed capacity to alter one’s own experience, often to the effect that one can experience sensations differently and including the possibility that there is no pain to be experienced (Barber and Adrian, 1982). The incorporation of suggestions for hypnotic analgesia lead to an unquestioned expectancy for altered experience, including analgesia. All of these types of suggestion are learned, regardless of whether or not they reflect the intentions of the person who administers them. This experiential distinction between hypnotic and placebo analgesia may at least partly account for experimental results showing a complete lack of relationship between the magnitude of hypnotic analgesia and the magnitude of placebo analgesia tested in high and low hypnotizable subjects

(Hilgard and Hilgard, 1983). These studies determined that for subjects not susceptible to hypnosis, about the same modest degree of analgesia is achieved through hypnosis and placebo administration. For subjects highly susceptible to hypnosis, pain reduction evoked by hypnosis is far greater than the negligible or even negative effects produced by placebo. Therefore, hypnotic analgesia is more than a placebo and very likely involves somewhat different psychological mediating factors. What are the implications of the general mechanisms of hypnotic and placebo analgesia? Taken together, investigations about the psychological and neural mechanisms of reduction of pain by both hypnotic and placebo interventions strongly indicate the existence of multiple factors and mechanisms. Both hypnotic and placebo analgesia depend upon subjects’ incorporation of and response to suggestions as well as the mediating influence of these suggestions on multiple dimensions of pain experience and behavior. In the case of hypnosis, a hypnotic state is at least likely to facilitate the incorporation of and response to suggestions for analgesia. Future studies of hypnotic analgesia should be able to better assess the experiential factors and psychosocial contextual factors that influence changes in self-reported pain intensity that occur as a result of a hypnotic intervention. New approaches and methods are available with which to measure experiential factors such as ‘perceived automaticity’ and ‘degree of absorption’, for example (Bowers, 1978; Pekula and Kumar, 1984; Price and Barrell, 1990). In this way, the factors that serve to optimally influence responses to hypnotic suggestions for analgesia could not only benefit studies of hypnotic analgesia but also pain patients. Both patients and healthcare providers could benefit from knowledge of the factors that facilitate hypnotic analgesia. The concept of ‘hypnotic therapeutic manipulations’ may shift in emphasis from reliance on outside authority to the patients’ active participation in developing psychological conditions for optimal therapeutic effects. In the case of placebo interventions, non-specific factors related to the perception of the therapy serve

269

to determine patients’ desire for and expectation of pain reduction. New approaches and methods for assessing these two factors could serve to enhance their influence in clinical contexts. These approaches and methods could also improve clinical and experimental studies of placebo analgesia because they would provide an addition and/or alternative to that of comparing placebo and natural history groups to assess the contribution of placebo effects. If desire for pain relief and expectancy are the actual mediating variables of placebo analgesia, then the contribution of placebo influences could be assessed within all groups of a study and not just the group that receives placebo treatment. Assessment of placebo effects could be especially facilitated in studies in which subjects can easily distinguish between active and control treatments, a condition that probably applies to the vast majority of analgesic studies. More precise analysis of different components of hypnotic and placebo analgesic effects could lead to important therapeutic improvements because different mechanisms could be utilized. These include selective reduction of pain-related affect (i.e. unpleasantness) among patients with low to moderate hypnotic ability. That reductions in pain affect are larger and more prevalent than reductions in pain sensation and are unrelated to hypnotic ability indicates that a large percentage of people could benefit by a hypnotic intervention. Another mechanism would include reductions in pain sensation as a result of processes that divert pain from conscious awareness once nociceptive information has reached higher centers. To the extent this component is manifested within an individual, the normal somatomotor reflexes and autonomic, neuroendocrine, and neuroimmunological consequences of pain would not be attenuated. Thus, stress related responses associated with pain would still occur often to the physiological detriment of the individual. Finally, both placebo and hypnotic analgesia may include a component mechanism wherein pain signals are inhibited at the spinal level of processing. In contrast to the mechanism just described, negative physiological consequences of pain would be attenuated by this mechanism, since inhibition at spinal levels would interrupt the supraspinal activation of brain structures involved

in autonomic and neuroendocrine responses to pain. The reduction of pain sensation intensity by this mechanism implies that the physiological consequences of pain, such as reduced immune response and prolongation of healing, would also be ameliorated (Liebeskind, 1991). Different individuals may utilize different proportions of these mechanisms as has been suggested previously (Price, 1988; Kiernan et al., 1995). Knowledge about the prevalence of these multiple mechanisms and the factors that influence them has far reaching implications for treatment of medical conditions, including pain.

References Barber, J. and Adrian, C. (1982) Psychological Approaches to the Management of Pain, BrunnerNazel, New York. Barber, J. and Mayer, D. (1977) Evaluation of the efficacy and neural mechanism of a hypnotic analgesia procedure in experimental and clinical dental pain. Pain, 4: 41-48. Barber, T.X. and Hahn, K.W. (1962) Physiological and subjective responses to pain producing stimulation under hynotically-suggested and waking-imagined “analgesia”. J. Abnorm. SOC. Psychol., 65: 411-418. Barber, T.X. and Wilson, S.C. (1977) Hypnosis, suggestions, and altered states of consciousness: experimental evaluation of the new cognitive-behavioral theory and the traditional trance-state theory of ‘hypnosis’. In: W.E. Edmonston, Jr. (Ed.), Annals of the New York Academy of Sciences, Vol. 296, Conceptual and Investigative Approaches to Hypnotic Phenomena, New York Academy of Sciences, New York, pp. 34-47. Beecher, H.K. (1955) The powerful placebo. J. Amer: Med. ASS.,159: 1602-1606. Beecher H.K. (1959) Measurement of Subjective Responses: Quantitative Efects of Drugs, Oxford University Press, New York. Benedetti, F. and Amanzio, M. (1997) The neurobiology of placebo analgesia: from endogenous opioids to cholecystokinen. Pmg. Neurobiol., 52(2): 109-125. Bootzin, R.R. (1985) The role of expectancy in behavior change. In: L. White, B. Turskey and G.E. Schwartz (Eds), Placebo: Theory, Research, and Mechanisms, The Guilford Press, New York. Bowers, K.S. (1978) Hypnotizability, creativity, and the role of effortless experiencing. Int. J. Clin. f i p e c Hypnosis, 26: 184-202. Clark, W.C. (1969) Sensory decision theory analysis of placebo effect on the criterion for pain and thermal sensitivity ( d ‘ ) .J. Abnorm. Psychology, 74: 361-371. Crawford, H.J. and Gruzelier, J.H. (1992) A midstream view of the neuropsychology of hypnosis: recent research and future directions. In: E. F r o m and M.R. Nash (Eds), Contemporary Hypnosis Research, Guilford, New York, pp. 221-266.

270 Evans, F.J. (1985) Expectancy, therapeutic instructions, and the placebo response. In: L. White, B. Turskey and G.E. Scwhartz (Eds), Placebo: Theory, Research, and Mechanisms, The Guilford Press, New York. Evans, M.B. and Paul, G.L. (1985) Effects of hypnotically suggested analgesia on physiological and subjective responses to cold stress. J. Consult. Clin. Psychol., 35: 362-37 1. Fields, H.L., and Basbaum. A.I. (1994) Endogenous pain control mechanisms. In: P.D. Wall and R. Melzack (Eds), Textbook of Pain, 3E, Churchill Livingston, Edinburgh. Fields, H.L. and Price, D.D. (1997) Toward a neurobiology of placebo analgesia. In: Ann Harrington (Ed.), Placebo: Probing the SelfHealing Brain, Boston: Harvard University Press. Goldstein, A., and Hilgard, E.R. (1975) Lack of influence of the morphine antagonist naloxone on hypnotic analgesia. Proc. Nut. Acad. Sci., 72: 2041-2043. Gracely, R.H. (1979) Psychophysical assessment of human pain. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fessard (Eds), Advances in Pain Research and Therapy, Vol. 3, Raven Press, New York. Gracely, R.H., McGrath, P. and Dubner, R. (1976) Validity and sensitivity of ratio scales of sensory and affective verbal pain descriptors: manipulation of affect by diazepam. Pain, 2: 19-29. Greene, R.J. and Reyher, J. (1972) Pain tolerance in hypnotic analgesia and imagination states. J. Abnorm. Psychol., 79: 29-38. Grevert, P., Albert, L.H. and Goldstein, A. (1983) Partial antagonism of placebo analgesia by naloxone. Pain, 16: 129-143. Grevert, P. and Goldstein, A. (1985) Placebo analgesia, naloxone, and the role of endogenous opioids. In: L. White, B. Turskey and G.E. Schwartz (Eds), Placebo: Theory, Research, and Mechanisms, The Guilford Press, New York. Hagbarth, K.E. and Finer, B.L. (1963) The plasticity of human withdrawal reflexes to noxious skin stimuli in lower limbs. Prog. Brain Rex, 1: 65-78. Hilgard, E.R. (1977) Divided Consciousness: Multiple Controls in Human Thought and Action, John Wiley and Sons, New York, pp. 300. Hilgard, E.R. and Hilgard, J.R. (1983) Hypnosis in the Relief of Pain, William Kaufmann, Los Altos, CA, pp. 294. Hilgard, E.R., Morgan, A.H. and MacDonald, H. (1975) Pain and dissociation in the cold pressor test: a study of hypnotic analgesia with hidden reports: through automatic keypressing and automatic talking. J. Abnorm. Psychol., 81: 170-1 74. Hofbauer, R.H., Rainville, P., Duncan, G.H. and Bushnell, M.C. (1998) Cognitive modulation of pain sensation alters activity in human cerebral cortex. SOC. Neumsci. Abstracts 24: (in press). Jospe, M. (1978) The Placebo Effect in Healing, Lexington Books, Lexington, Mass. Kieman, B.D., Dane, J.R., Phillips, L.H. and Price, D.D. (1995) Hypnotic analgesia reduces R-III nociceptive reflex: further

evidence concerning the multifactorial nature of hypnotic analgesia. Pain, 60:39-47 Kirsch, I. (1990) Changing Expectations: a Key to Effective Psychotherapy, BrooksICole, Pacific Grove, CA. Lasagna, L., Laties, V.G. and Dohan, J.L. (1958) Further studies on the 'pharmacology' of a placebo administration. J. Clin. Invest., 37: 533-537. Laska, E. and Sunshine, A. (1973) Anticipation of analgesia a placebo effect. Headache, 13: 1-1 1. Levine, J.D., Gordon, N.C., Bornstein, J.C. and Fields, H.L. (1979) Role of pain in placebo analgesia. PNAS, 76: 3528-3531. Levine, J.D., Gordon, N.C. and Fields, H.L. (1978) The mechanism of placebo analgesia. Lancet, 2: 654-657. Liebeskind, J.C. (1991) Pain can kill. Pain, 44:3 4 . Mayer, D.J., Price, D.D., Barber, J. and Rafii, A. (1976) Acupuncture analgesia: evidence for activation of a pain inhibitory system as a mechanism of action. In: J.J. Bonica and D. Albe-Fessard (Eds), Advances of Pain Research and Therapy, Vol. 1, Raven, New York, pp. 75 1-754. Montgomery, G.H. and Kirsch, I. (1996a) Mechanisms of placebo pain reduction: an empirical investigation. Psychol. Sci., 7: 174-176. Montgomery, G.H. and Kirsch, I. (1996b) Classical conditioning and the placebo effect. Pain, 72: 103-113. Pekula, R.J. and Kumar, V.K. (1984) Predicting hypnotic susceptibility by a self-report phenomenological instrument. Am. J. Clin. Hypnosis, 27: 114-121. Price, D.D. (1988) Psychological and Neural Mechanisms of Pain. Raven, New York, pp. 241, Price, D.D. (1996) The neurological mechanisms of hypnotic analgesia. In: J. Barber (Ed.), Hypnosis and Suggestion in the Treatment of Pain, W.W. Norton and Company, New York, pp. 117-136. Price, D.D. and Barber, J. (1987) An analysis of factors that contribute to the efficacy of hypnotic analgesia. J. Abnorm. Psychol., 96: 46-51. Price, D.D. and Barrell, J.J. (1990) The structure of the hypnotic state: a self-directed experiential study. In: J.J. Barrell (Ed.), The Experiential Method: Exploring the Human Experience, Copely Publishing group, Acton, Massachusetts, pp. 85-97. Price, D.D. and Barrell, J.J. (1999) The role of expectancy in pain and pain relief. In: Irving Kirsch (Ed.), Expectancy, Experience, and Behavior, APA Books, Washington D.C. Price, D.D. and Fields, H.L. (1997) The contribution of desire and expectation to placebo analgesia: implications for new research strategies. In: Ann Hanington (Ed.), Placebo: Probing the Self-Healing Brain, Boston: Harvardc University Press. Price, D.D., Barrell, J.E. and Barrell, J.J. (1985) A quantitativeexperiential analysis of human emotions. Motiv. Emor., 9: 19-38. Price, D.D. and Barrell, J.J. (1990) The structure of the hypnotic state: a self-directed experiential study. In: J.J.

27 1 Barrel1 (Ed.), The Experiential Method: Exploring the Human Experience, Copely Publishing Group, Acton, Massachusetts, pp. 85-97. Price, D.D., Barrell, J.J. and Gracely, R.H. (1980) A psychophysical analysis of experiential factors that selectively influence the affective dimension of pain. Pain, 8: 137-179. Rainville, P., Carrier, B., Hofbauer, R.K., Duncan, G.H. and Bushnell, M.C.(1999) Dissociation of sensory and affective dimensions of pain using hypnotic modulation. Pain, 82(2): 159-171. Rainville, P., Duncan, G.H., Price, D.D., Carrier, B. and Bushnell, M.C. (1997) Pain affect encoded in human anterior cingulate but not somatosensoty cortex. Science, 277: 968-97 I. Shor, R.E. and Ome, M.T. (1965) The Nature of Hypnosis: Selected Basic Readings, Holt, Rhinehart, and Winston, New York. Spanos, N.P. (1986) Hypnotic behavior: a social-psychological interpretation of amnesia, analgesia, and ‘trance logic’. Behav. Brain Sci., 9: 440467.

Tenenbaum, S.J., Kurtz, R.M. and Bienias, J.L. (1990) Hyponotic susceptibility and experimental pain reduction. Am. J. Clin. Hyp., 33(1): 4 0 4 9 . Traut, E.F. and Passarelli, E.W. (1957) Placebos in the treatment of rheumatoid arthritis and other rheumatic conditions. Ann. Rheum. Dis., 16: 18-22. Weitzenhoffer, A.M. (1957) General Techniques of Hypnotism, Grune and Stratton, New York. White, L., Turskey, B. and Schwartz, G.E. (Eds), (1985) Placebo: Theory, Research, and Mechanisms, The Guilford Press, New York. Willer, J.C. Comparative study of perceived pain and nociceptive flexion reflex in man. Pain, 3: 69-80. Willer, J.C. (1984) Nociceptive flexion reflex as a physiological correlate of pain sensation in humans. In: B. Bromm (Ed.), Pain Measurement in Man. Neurophysiological Correlates of Pain, New York: Elsevier, pp. 87-1 10. Willer, J.C. (1985) Studies on pain. Effects of morphine on a spinal nociceptive flexion reflex and related pain sensation in man. Brain Res., 331: 105-114.

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research, Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAPTER 20

The role of vagal visceral afferents in the control of nociception W. Jiinig'?*,S.G. Khasa?, J.D. Levine and EJ.-P. Miao2

' Physiologisches Institut, Christian-Albrechts- Universitat zu Kiel, 24098 Kiel, Germany 'Departments of Anatomy, Medicine and Oral Surgery and Division of Neuroscience, Biomedical Sciences Program and NIH Pain Center (UCSF), University of California, San Francisco, San Francisco, CA 94143-0440, USA

Abdominal vagal afferents and protection of the body Nerve fibers in the abdominal vagus nerves contain 8-85% visceral afferent and 15-20% preganglionic parasympathetic fibers (Precht and Powley, 1990; Berthoud et al., 1991). The vagal afferents innervate the gastrointestinal tract (GIT) and associated organs (e.g. liver, pancreas). These vagal afferents encode in their activity mechanical and chemical events of the GIT (see Furness and Clerc, Chapter 11, this volume). They project viscerotopically to the nucleus of the solitary tract (NTS; Loewy and Spyer, 1990; Ritter et al., 1992). The second-order neurons in the NTS project to various sites in the lower brain stem, upper brain stem, hypothalamus and amygdala, establishing well-organized neural pathways which are the basis for distinct regulations of GIT functions. The preganglionic parasympathetic uxons have their cell bodies in the dorsal motor nucleus of the vagus nerve. These efferent neurons are also topographically organized and establish final efferent pathways for the regulation of various functions of the GIT (see Loewy and Spyer, 1990; Ritter et al., 1992). The small intestine is a very vulnerable portal of entry into the body and therefore it is of great *Corresponding author. Tel.: + 43 118802036; Fax: + 43 118802036; e-mail: [email protected]

teleological advantage to have both a potent local defense system as well as an early warning system to the rest of the body for entry of microorganisms and toxins located there. This part of the GIT does, in fact, contain a powerful immune system (the gut associated lymphoid tissue [GALT] [Shanahan, 19941). The small intestine is innervated by vagal afferents that project through the celiac branches of the abdominal vagus nerve. Specific modulation of activity in these afferents in conjunction with the reaction of the GALT may act as an early warning system to the rest of the body by transmitting important information to the brain about toxic events and agents in the intestine which are dangerous for the organism. Recent experimental investigations support this idea that abdominal vagal afferents may be important for protective functions of the GIT and the body

(1) Electrical stimulation of abdominal vagal afferents exerts inhibition or excitation of the central nociceptive system and depresses nociceptive behavior depending on whether unmyelinated or myelinated afferents are stimulated (Gebhart and Randich, 1992; Randich and Gebhart, 1992; for details see below). (2) When injected intraperitoneally in rats, illnessinducing agents, such as the bacterial cell wall endotoxin lipopolysaccharide (LPS) produce behavioral hyperalgesia. This is mediated by

214

activity in subdiaphragmatic vagal afferents, specifically afferents running in the hepatic branch. It is suggested that LPS activates hepatic macrophages (Kupffer’s cells) which release interleukin-1p (IL-1p) and tumor necrosis factor a (TNFa). This in turn activates vagal afferents from the liver. By the same token, IL-1p and TNF-a injected intraperitoneally themselves generate behavioral hyperalgesia (enhancement of the thermal nociceptive tail-flick reflex) which is abolished by vagotomy (Watkins et al., 1994a, b, 1995a, b, c, d). These results suggest that vagal afferents innervating the liver (another internal defense line in the body) are activated by cytokines released by activated leukocytes and signal these events to the brain resulting in pain behavior (i.e. an enhanced tail-flick reflex). Watkins and coworkers suggest that the IL-1 (and possibly TNF-a)released by the Kupffer’s cells binds specifically to glomus cells in the abdominal paraganglia which are innervated by vagal afferents; the vagal afferents are activated and signal the peripheral events to the brain, leading to hyperalgesia and other illness responses (see below). (3) Recently it has been shown that the first (fast) phase of the fever response which can be generated in the rat and guinea-pig by intravenous injection of the endotoxin lipopolysaccharide (LPS) is mediated by vagal afferents which possibly innervate the liver (Sehic and Blatteis, 1996; Blatteis and Sehic, 1997). This fast component of fever generated by LPS is abolished after abdominal vagotomy. Activation of vagal afferents by LPS leads to activation of neurons in the nucleus of the solitary tract (NTS) and subsequently of noradrenergic neurons in the A1 and A2 areas of the brain stem which project to the hypothalamus. It is proposed that Kupffer’s cells of the liver are activated via the complement fragments C3a and C5a, that these activated Kupffer’s cells synthesize and release cytokines as well as eicosanoids (e.g. prostaglandin E2) and that this leads to the activation of vagal afferents. (4) Pain behavior mediated by vagal afferents, which are activated by intraperitoneal injection

of LPS, is part of a general sickness behavior characterized by various protective illness responses (e.g. immobility, decrease in food intake, formation of taste aversion to novel foods, decrease of digestion, loss of weight (anorexia), fever, increase of sleep, change in endocrine functions, malaise, etc.) and is correlated with marked alterations of brain function. For example, food aversions and anorexia are generated in rats by TNF (injected intraperitoneally) and subcutaneous implantation of Leydig LTW(m) tumor cells. Both sickness behaviors are abolished or attenuated by subdiaphragmatic vagotomy (Bret-Dibat et al., 1995; Bernstein, 1996). By the same token it has been shown that endotoxin injected intraperitoneally generates profound changes in various brain areas in rodents (such as c-fos expression in neurons of the nucleus of the solitary tract and in hypothalamic nuclei as well as induction of IL-1p mRNA in the pituitary gland, hypothalamus and hippocampus). These changes do not occur or are significantly attenuated in subdiaphragmatically vagotomized animals (Ericsson et al., 1994; Wan et al., 1994; Lay6 et al., 1995). ( 5 ) Vagal afferents in the celiac branches of the abdominal vagus nerve monitor chemical and mechanical events that occur in the intestine under physiological and pathophysiological conditions (for review see Grundy and Scratcherd, 1989), i.e. are related to meals, ingestion of toxic substances, inflammation, obstruction, etc. Neurophysiological recordings have shown that these afferents respond to distension or contraction of the small intestine and to intraluminal chemical stimulation. Cholecystokinin released locally, in response to chemical stimuli, may activate these afferents (Richards et al., 1996) and sensitize them for mechanical and other stimuli (Schwartz and Moran, 1994, 1996; Berthoud et al., 1995). Cells of the mucosa may function as secondary sensory cells and modulate activity in vagal afferents via paracrine actions (Grundy, 1992). These afferents may be important for preabsorptive detection of energy-yielding molecules and probably for other properties of nutrient solu-

215

tions which may be toxic and deleterious for the gastrointestinal tract and for the body (Walls et al., 1995a, b). Additionally some vagal afferents which innervate the small intestine and the liver respond to cytokines (e.g. IL-lp); these afferents may encode events which are related to the immune system of the GIT and liver (Niijima, 1992, 1996). Here we will discuss experiments which show that activity in abdominal vagal afferents influence central nociceptive pathways and pathways which are involved in neuroendocrine control of mechanical hyperalgesic behavior. It will be shown that nociception and pain in the somatic domain is potentially under remote control from the viscera via vagal afferents and that central as well as peripheral neuroendocrine mechanisms are involved.

Control of central nociceptive pathways by activity in vagal visceral afferents It is assumed that vagal visceral afferents are normally not associated with the direct generation of visceral pain although this has recently been questioned as far as the heart is concerned (Mehler and Gebhart, 1992). However, there are indications that vagal afferents are important in control of nociceptive impulse transmission in the spinal cord and probably elsewhere. This control is mainly inhibitory (see below). Only unmyelinated afferents which innervate the mucosa of the trachea may have a nociceptive function. Stimulation of these afferents induce vasodilation and plasma extravasation in the mucosa (see McDonald, 1990) as well as discomfort and probably pain. Most impulses in vagal visceral afferents never reach consciousness. They generate otherwise general feelings like hunger, satiety, nausea, gut fullness, feelings of gastric acidity and motility, air hunger, etc. (see Ritter et al., 1992; Jbig, 1996c) The idea that vagal afferents are involved in control of nociception and pain is based on experiments on monkeys and rats. Electrical stimulation of cervical vagal afferents in monkeys suppresses transmission of impulse activity in spino-thalamic relay neurons with nociceptive

function at all levels of the spinal cord. Electrical stimulation of subdiaphragmatic vagal afferents has no effect on thoracic spino-thalamic relay neurons which are involved in nociception of the heart in this species. It is concluded from this observation that particularly cardio-pulmonary afferents are involved in the inhibitory control of these spinothalamic relay neurons (see Foreman, 1989, Foreman, Chapter 15, this volume). The central pathways which mediate this effect are neurons in the locus ceruleus and in the A7 group (dorsolateral pontine tegmentum; noradrenergic and non-noradrenergic) and neurons in the nucleus raphe magnus of the rostro-ventral medulla (serotonergic, non-serotonergic) which project to the dorsal horn (Kwiat and Basbaum, 1992; Yeoman et al, 1992; see Fields and Basbaum, 1999; Fields, Chapter 18, this volume). Similar results were obtained in the rat. Here transmission of nociceptive impulses from skin and colon in the dorsal horn and the tail-flick-reflex elicited by noxious heat stimulation of the tail were enhanced by electrical stimulation of myelinated vagal afferents and depressed by electrical stimulation of unmyelinated vagal afferents. In this species the subdiaphragmatic vagal afferents were particularly powefil in eliciting these modulatory effects. As in the monkey, the inhibitory effects were generated via descending systems from the pontine reticular nuclei and from the rostro-ventral medulla. The facilitatory effect was mediated by suprapontine pathways (Fig. 1; see Gebhart and Randich, 1992; Randich and Gebhart; 1992). These results are compelling; however, several questions remain unanswered in both sets of experiments: First, in which functional context does the control of nociception and pain by vagal afferents occur? Second, are all vagal (cardiovascular, pulmonary, gastrointestinal) afferents involved in this control or subsets of functionally identifiable vagal afferents or vagal afferents, which cannot be identified with one of the organ functions? Answers to these and other questions will show how important these modulatory effects really are.

216

From the clinical point of view it may be speculated as to whether these general vagal effects are related to clinical entities such as silent myocardial infarction, non-cardiac chest pain, nonulcer dyspepsia, irritable bowel syndrome and fibromyalgia. In these diseases the regulation of transmission of nociceptive impulse activity may be altered, leading to decreases or increases in pain. No peripheral morphological, physiological or biochemical markers exist which could explain these changes in nociception and pain.

Cutaneous mechanical hyperalgesic behavior is enhanced after abdominal vagotomy Withdrawal threshold to stimulation of the rat hindpaw with a linearly increasing mechanical stimulus to the dorsum of the paw decreases dosedependently after intradermal injection of bradykinin (Fig. 2). Following a single injection of bradykinin this decrease lasts for more than one hour for mechanical stimulation (Taiwo and Levine, 1988). This type of mechanical hyperalgesic

L=l, forebrain

spinal

COI

Fig. 1. Schematic diagram of the central pathways and brain stem nucA that mediate ...s effects of activity in vagal afferents from visceral organs on spinal nociceptive transmission. Vagal afferents synapse in the nucleus tractus solitarius (NTS). Inhibitory eflects of electrical stimulation of unmyelinated abdominal vagal afferents are mediated by relays in the locus coeruleuslsubcoeruleus (A6) and in the nucleus raphe magnus (NRM). The descending fibers probably project through the dorsolateral and ventrolateral funiculi (DLF,VLF) of the spinal cord. Facilitatory eflecfsof electrical stimulation of myelinated abdominal vagal afferents are mediated by ascending fibers to and through the A6 cell group to the forebrain before eventually descending in the spinal VLF. These pathways and relays are organized bilaterally. PAG, periaqueductal gray. Based on experiments on rats. Modified from Gebhart and Randich (1992).

277

behavior is mediated by the B, bradykinin-receptor (Khasar et al., 1995) and is not present when bradykinin is injected subcutaneously (Khasar et al., 1993). The decrease in paw-withdrawal threshold provided by bradykinin is significantly reduced after surgical sympathectomy and is blocked by indomethacin. This shows that the terminals of the population of cutaneous nociceptive neurons are sensitized by a prostaglandin which is believed to

v h

l’;i

~

-0T

80 -

-

Normal

+Sham Vagx

-

+Vagx 7 d I

100

-

be released from the sympathetic varicosities (Gonzales et al., 1989) or from other cells in association with the sympathetic varicosities. Interestingly, decentralization of the lumbar paravertebral sympathetic ganglia (denervating the postganglionic neurons by cutting the preganglionic sympathetic axons) does not abolish bradykinin-induced mechanical hyperalgesic behavior. This indicates that the sensitizing effect

4 Baseline threshold

9

L

60 -

40

I

I

I

I

I

Fig. 2. Decrease of paw-withdrawal threshold to mechanical stimulation of the dorsum of the rat hindpaw induced by bradykinin (bradykinin-inducedbehavioral mechanical hyperalgesia) in normal control (open circles, n = 26), vagotomized (triangles, n = 16) and sham vagotomized (closed circles, n = 18) rats. Experiments conducted 7 days after vagotomy. Post hoc test showed significant differences between vagotomized and normal (p < 0.05) as well as between vagotomized and sham vagotomized (p< 0.05) rats, in response to bradykinin. Cutaneous mechanoreceptors in the hairy skin were stimulated by a linearly increasing mechanical force using a Basile Algesimeter (Stoelting, Chicago, IL). Threshold is defined as the minimum force (8) at which the paw is withdrawn by a rat. Ordinate scale expresses paw-withdrawal threshold in grams. The abscissa scale is the log dose of BK (in ng) injected in a volume of 2.5 pg saline into the dermis of the skin of the dorsal aspect of the hindpaw. In this and subsequent figures,Vagx = subdiaphragmatic vagotomy. Data from Khasar et al. (1998a).

278

of bradykinin is not dependent on activity in the sympathetic neurons innervating skin and therefore not on release of norepinephrine (Khasar et al., 1998a). Unfortunately, while evidence for sensitization of cutaneous nociceptors to mechanical stimulation is poor or absent (see Treede et al., 1992) some sensitization to mechanical stimulation by bradykinin has been demonstrated for afferents from the knee joint (Neugebauer et al., 1989) and from skeletal muscle (Mense and Meyer, 1988); however, the discrepancies may be due to technical difficulties with this experimental approach. If vagal inhibition acts continuously on the central nociceptive pathway one would expect, on the basis of studies reported by Gebhart, Randich and coworkers (see Gebhart and Randich, 1992; Randich and Gebhart, 1992),that subdiaphragmatic vagotomy might enhance the mechanical hyperalgesic behavior, irrespective of the way the nociceptive afferents had been sensitized (e.g. by bradykinin or another hyperalgesic agent), and lower baseline threshold to mechanical stimulation. Therefore, we have examined the effect of subdiaphragmatic vagotomy on the baseline paw-withdrawal threshold. We have also investigated the effects of sub-diaphragmatic vagotomy on decreases of mechanical paw-withdrawal threshold induced by bradykinin injected intradermally and on decreases of mechanical paw-withdrawal threshold induced by intradermally injected prostaglandin E,. Baseline paw-withdrawal threshold in normal and sham-vagotomized rats was 109f 2.1 g (meanfSEM) and 107+2.8 g, respectively. This mechanical baseline threshold significantly decreased to 89 r 1.7 g, seven days after subdiaphragmatic vagotomy (Fig. 2). Intradermal injection of bradykinin produced a dose-dependent decrease in mechanical nociceptive threshold (i.e., mechanical hyperalgesic behavior) in normal rats and in rats seven days after subdiaphragmatic sham-vagotomy (Fig. 2). Bradykinin-induced hyperalgesia was significantly enhanced seven days after subdiaphragmatic vagotomy (Fig. 2). There are three important characteristics of the effect of vagotomy on mechanical baseline threshold and on bradykinin-induced decrease of paw-withdrawal threshold to mechanical stimulation:

(1) The dramatic enhancement of bradykinininduced mechanical hyperalgesic behavior occurs also when only the celiac vagal branches are interrupted but not when the gastric andor hepatic branches of the abdominal vagus nerves are interrupted. Thus, the vagal afferents involved project through the celiac branches of the abdominal vagus nerves, which innervate the small intestine and proximal part of the large intestine, and not through the hepatic or gastric branches (Khasar et al., 1998a). Surprisingly, the baseline paw-withdrawal threshold to mechanical stimulation does not decrease when only the celiac vagal branches are interrupted (Khasar et al., 1998a). (2) Both vagotomy-induced changes (decrease in baseline paw-withdrawal threshold, bradykinin-induced hyperalgesic behavior) take about two weeks to reach maximum and remain stable over five weeks (Fig. 4; Khasar et al., 1998a, b). (3) Finally and most important, subdiaphragmatic vagotomy does not have a significant effect on cutaneous mechanical hyperalgesic behavior produced by intradermal injection of prostaglandin E, (which is supposed to act directly to sensitize nociceptors) (Khasar et al., 1998a). Thus, the effect of vagotomy, firstly, is not a general effect of all abdominal vagal afferents and, secondly, cannot readily be explained by an immediate removal of inhibition from the central nociceptive system (e.g. in the dorsal horn) as predicted by the experiments of Foreman, Gebhart, Randich and coworkers (see above). We therefore looked for an alternative explanation and found out in our experiments that the sympatho-adrenal system is the important link mediating these changes.

Enhancement of mechanical hyperalgesic behavior after vagotomy is mediated by the sympathoadrenal system The experiments were primarily designed to test whether the enhancement or induction of the mechanical hyperalgesic behavior generated in rats after subdiaphragmatic vagotomy is entirely or partially produced by activation of a neuroendo-

279 Sham SDV

1

-0- Adrenal Medx

Baseline threshold

B

120

I

I

-1

0

I

I

1 2 LOG [BK] ng

I

3

1 +AM denerv + SDV

100 0

r m

E Barnline threshold

60

40

I

I

-1

0

I

I

1 2 LOG [BK] ng

3

I

Fig. 3. (A) Baseline and decrease of paw-withdrawal threshold to mechanical stimulation of the dorsum of the rat hindpaw, induced by bradykinin (bradykinin-induced behavioral mechanical hyperalgesia) in sham vagotomized rats (closed circles, n = 18), in rats whose adrenal medullae were removed (Adrenal Medx, open squares, n = 12) and in rats with removed adrenal medullae and which were also vagotomized (closed squares, n = 12). Experiments conducted 5 weeks after removal of the adrenal medullae and 7 days after additional vagotomy. (B) Data from experiments in which the adrenal medullae were denervated: rats with denervated adrenal medullae (AM denerv, open squares, n = 6); vagotomizedrats with denervated adrenal medullae (AM denerv plus SDV, closed squares; n = 10). Experiments were conducted 7 days after surgery. Paw-withdrawal thresholds of vagotomized rats in which the adrenal medullae were removed or denervated were significantly higher than those of rats which were only vagotomized (see open triangles in Fig. 2; p < 0.05). Paw-withdrawal thresholds of rats in which the adrenal medullae were removed or denervated were significantly higher than those of rats which were additionally vagotomized (peO.05; compare open with closed squares). Data on shamvagotomized rats in A and B are the same as in Fig. 2. Modified from Khasar et al. (1998b).

280 -@-

ShamSDV

-br SDV

-br SDV +Adrenal denenr.

+SDV +Adrenal denew. 120 'I

A

t p

10

s

t

I

I

I

I

I

1

0

7

14

21

28

35

-

B

r

0

T

h

0-

I! 0

f6 -10 -

B

E

-20

-

-30

-

.-c

t

t

I 0

120

-

B 100

-

L P r Y

I

r

I

I

I

1

14

21

28

35

C \\

L A

t

-0

z

1 gE

I

T

80-

60-

k

n 40

- Vagotow

Adrenal danervation

r

I

0

7

I

1

14 21 llme (days)

I

1

28

35

28 1

crine system which acts directly or indirectly on the nociceptor population. The two candidate neuroendocrine systems which may be involved are the sympatho-adrenal (SA) system and the hypothalamo-pituitary adrenal (HPA) system. In preliminary experiments we found out that the HPA system is probably not directly involved; furthermore, hypophysectomy led to changes of the behavior of the animals which made interpretation of changes of nociceptive behavior difficult. However, surgical interventions at the SA system clearly showed that the SA is involved. So we tested the mechanical hyperalgesic behavior before and after subdiaphragmatic vagotomy in various groups of rat with the following interventions at the SA system: (a) Removal of the adrenal medullae five weeks before vagotomy and animals without vagotomy (experiments five weeks later). (b) Denervation of the adrenal medullae (cutting the sympathetic preganglionic axons) during vagotomy or seven days before vagotomy (experiments seven days after the last surgery). (c) Denervation of the adrenal medullae 14 days after vagotomy. Effect of adrenal medullectomy or denervation of the adrenal medullae Adrenal medullectomy or denervation of the adrenal medullae generated both a small increase in baseline paw-withdrawal threshold and paw-withdrawal threshold to intradermal bradykinin compared to the controls (Fig. 3). Under this condition of defunctionalized adrenal medullae, subdiaphragmatic vagotomy is followed by only a small decrease in paw-withdrawal threshold (compare open and closed squares in Fig. 3) but not by the large changes seen in animals with functioning

adrenal medullae. These changes are significant, with the exception of the change of baseline threshold in animals with denervated adrenal medullae. The changes can fully be explained by removal of central inhibition of nociceptive impulse transmission occurring probably in the dorsal horn (see above; Khasar et al., 1998b). Effect of denervation of adrenal medullae 14 days following subdiaphragmatic vagotomy If the decrease of baseline mechanical pawwithdrawal threshold and enhanced decrease of paw-withdrawal threshold to mechanical stimulation, generated by intradermal injection of bradykinin, are related to a signal released from the adrenal medullae, which is dependent on activity in sympathetic preganglionic axons, one would expect both changes to be reversed when the adrenal medullae are denervated. This was in fact the case. Figure 4 shows this reversal for the baseline pawwithdrawal threshold (a), for the decrease in paw-withdrawal threshold to intradermal injection of 1 ng bradykinin alone (b) and for both effects together (c). The reversal of the vagotomy effect after denervation of the adrenal medullae had a slow time course similar to the time course of the decrease in baseline paw-withdrawal threshold following vagotomy (open triangles in Fig. 4) over a five-week period. Repeated testing of shamvagotomized control rats over the same period of time did not reveal a decrease in paw-withdrawal threshold produced by 1 ng bradykinin (closed circles in Fig. 4). The paw-withdrawal thresholds, 14 and 21 days after denervation of the adrenal medullae, were significantly higher than those measured in the animals which were only vagoto-

w

Fig. 4. Baseline paw-withdrawal threshold (A), difference between baseline paw-withdrawalthreshold and paw-withdrawal threshold in response to 1 ng BK injected intradermally (B), and total change of paw-withdrawal threshold in response to intradermal injection of 1 ng bradykinin (C) in rats before and 7 to 35 days after vagotomy (open triangles, n=6), before and 7 to 35 days after shamvagotomy (closed circles, n = 8) and in rats which were first vagotomized and whose adrenal medullae (AM) were denervated 14 days after vagotomy and measurements taken up to 35 days after initial surgery. The latter group of animals consists of two subgroups: rats which were tested after vagotomy and after additional denervation of the adrenal medullae (half-closed squares, n = 6) and rats which were only tested after additional denervation of the adrenal medulla (closed inverted triangles, n=4). Ordinate scale is threshold in grams. Data of the sham-vagotomy and the vagotomy group of rats were significantly different 7 days after vagotomy ( p < 0.01). Data of vagotomized rats with denervated AM and rats that were only vagotomized were significantly different on days 28 and 35 ( p< 0.01). Data between sham-vagotomized rats and vagotomized rats in which the adrenal medullae were denervated were not significantly different on days 28 and 35 (p>0.05). Modified from Khasar et al. (1998b).

282

mized (compare closed triangles with open triangles in Figs. 4). Furthermore, the paw-withdrawal thresholds in response to 1 ng bradykinin at 14 and 21 days after denervation of the adrenal medullae in vagotomized animals were not significantly different from those in shamvagotomized animals which had repeatedly been tested over a period of five weeks after surgery (compare closed triangles with closed circles in Fig. 4).

Interpretation and implications These results suggest that two mechanisms contribute to the decrease of baseline paw-withdrawal threshold to mechanical stimulation and to enhancement of mechanical hyperalgesic behavior following vagotomy: (1) Ongoing central inhibition of nociceptive impulse transmission (occurring probably in the dorsal horn) which is normally maintained by spontaneous activity in vagal afferents is removed after vagotomy (Fig. 5 ) resulting in an enhancement of nociceptive behavior to mechanical stimulation (see Fig. 3). This enhancement is in accordance with the idea that nociception and pain is centrally under inhibitory control from the visceral domain via vagal afferents (see Foreman, 1989, this volume; Gebhart and Randich, 1992; Randich and Gebhart, 1992; see above and Fig. 1). However, it is small when compared to the overall enhancing effect of vagotomy on nociceptive behavior in animals with functioning adrenal medullae. (2) Vagotomy triggers the activation of sympathetic preganglionic neurons innervating the adrenal medullae (Fig. 5 ) , probably by removing central inhibition acting at this sympathetic pathway, thus leading to release of a hormonal signal from the adrenal medullae. Interruption of these sympathetic preganglionic axons (by denervation of the adrenal glands [Fig. 5 ] ) , stops the release of this hormonal signal and therefore prevents or reverses the decrease of baseline mechanical paw-withdrawal threshold and the enhancement of bradykinin-induced mechanical hyperalgesic behavior. These

changes have a slow time course which is in the range of 7-14 days. This novel finding could principally imply that the sensitivity of nociceptors are under neuroendocrine control and that nociceptor sensitivity can be regulated from remote body domains and by the brain via this neuroendocrine pathway. The second mechanism is novel and has several implications. In connection with the vagotomy several interesting questions and problems are raised: 0

The vagal afferents which are involved in modulation of hyperalgesic behavior project

I

A

NTS

.cortex medulla

I1 A

I

/

I Brain Stem

F&oeiiac

ganglion

I

@ vagotomy @ Denmation of the adrenal medulla

@ Removal of the

adrenal medulla

Fig. 5. Summary scheme of afferent, central and efferent pathways which may be involved in decrease of baseline threshold and of bradykinin-induced paw-withdrawal threshold following subdiaphragmatic vagotomy. Activity in vagal afferents which innervate small and large intestines and project to the nucleus of the solitary tract (NTS)centrally inhibit the pathway to preganglionic neurons innervating the adrenal medullae and neurons of the nociceptive system, e.g. in the dorsal horn. Interruption of these afferents leads to disinhibition of the central pathway to the preganglionic neurons innervating the adrenal medulla and of the central nociceptive system: (1) subdiaphragmatic vagotomy and (2) removal or (3) denervation of the adrenal medullae. Modified from Khasar et al. (1998b).

283

0

0

through the celiac branches of the abdominal vagus nerves and supply the small and large intestine but not liver and stomach. These vagal afferents are capsaicin-sensitive whereas most of those vagal afferents which innervate stomach or liver are not (Berthoud and Neuhuber, 1994; Berthoud et al., 1997). Do these vagal afferents monitor toxic and other events at the inner defense line of the body (the ‘gut associated lymphoid tissue’, GALT)? What are the physiological stimuli to activate these vagal afferents? Is it possible to increase mechanical pawwithdrawal threshold (baseline, to intradermally injected bradykinin) by physiological stimulation of these vagal afferents? The changes following vagotomy (decreased mechanical baseline threshold and enhanced bradykinin-induced mechanical hyperalgesic behavior; Khasar et al., 1998a, b) are generated by the interruption of vagal afferents. Therefore, the vagal afferents involved must be tonically active (Schwartz and Moran, 1994, 1996). This conclusion is fully supported by further experiments in which we studied the influence of vagal afferents on a nociceptive-neuroendocrine negative feedback pathway controlling neurogenic inflammation of the synovia of the rat knee joint (bradykinin-induced synovial plasma extravasation which is dependent on the sympathetic innervation of the synovia (Miao et al., 1996a, b; Green et al., 1997)). Subdiaphragmatic vagotomy or cutting the celiac branches of the abdominal vagus nerves potentiates this nociceptive-neuroendocrine negative feedback control (Miao et al., 1997a, b; Janig et al., 1998). Maier, Watkins and coworkers have shown, using the thermal tail-flick reflex, that behavioral hyperalgesia in rats produced by intraperitoneal injection of the illness-inducing bacterial cell wall endotoxin lipopolysaccharide (LPS) is mediated by activity in subdiaphragmatic vagal afferents projecting through the hepatic branch (Watkins et al., 1994a, 1995d). The results of these experiments imply that stimulation of vagal afferents generates thermal hyperalgesic behavior. These results are apparently at variance with our results showing that removal of activity in

0

vagal afferents leads to hyperalgesic behavior. However, it must be kept in mind that: (a) two different types of nociceptive behavior (mechanical, thermal) have been tested; (b) two different groups of vagal afferents (innervating smalVlarge intestine and liver, respectively) are involved; (c) induction and depression of the two hyperalgesic behaviors probably occur at different time courses; (d) different (central and peripheral) mechanisms are probably involved in the modulation of both nociceptive behaviors from the visceral domain. The hormonal signal released from the adrenal medullae has not yet been identified. The most likely candidate is epinephrine, but an enkephalin or an enkephalin-containing neuropeptide cannot be excluded. These substances are released on impulse activity in preganglionic sympathetic neurons innervating the adrenal medullae (Jany et al., 1986; Engeland et al., 1986). The decrease of paw-withdrawal threshold (baseline as well as after intradermal injection of bradykinin) following vagotomy takes several days to reach peak effect and the recovery following additional denervation of the adrenal medullae also takes several days. This hormonal signal probably acts at the nociceptor resulting in its sensitization. It is unlikely to act centrally otherwise it would be expected that mechanical hyperalgesic behavior induced by intradermal injection of prostaglandin E, (PGE,) would also be enhanced after vagotomy. However, this possibility cannot entirely be ruled out if one assumes that PGE,-induced hyperalgesic behavior and bradykinin-induced hyperalgesic behavior involve significantly non-overlapping peripheral afferent and central pathways, etc. Can the change of the putative hormonal signal quantitatively be measured in the blood in animals after vagotomy and correlated with the nociceptive behavior? Can the change in nociceptive behavior be mimicked by continuous infusion of this putative hormonal signal? The changes of paw-withdrawal threshold have a slow time course. The mechanism of this slow time-course is not clear at the moment. The hormonal signal has obviously to act over a long time to induce changes in the micromilieu of the

284

nociceptor population which in turn leads to their sensitization. This signal may act not directly on the nociceptors but on other cells (e.g. macrophages, mast cells, keratinocytes) which then release substances that generate the sensitization. The slow time course implies that it is unlikely that the changes of paw-withdrawal threshold can acutely be prevented or induced by pharmacological interventions (e.g. by local or systemic injection of an adrenoceptor blocker or agonist). The basis of the decrease in paw-withdrawal threshold to mechanical stimulation of skin may be a decrease in threshold to mechanical stimulation of all or of a subpopulation of cutaneous nociceptors or the recruitment of normally silent (very high threshold) cutaneous nociceptive aferents (see Kress et al., 1992, Janig and Koltzenburg, 1993, Michaelis et al., 1996) or both. The change of sensitivity of a population of cutaneous nociceptors generated by a hormonal signal which is regulated by the brain would be a novel mechanism of sensitization. This novel mechanism of sensitization of the nociceptor population by a signal from the sympathoadrenal system would be different from mechanisms which lead to activation and/or sensitization of nociceptors by sympathetic-afferent coupling under pathophysiological conditions (Jiinig and McLachlan, 1994; Jiinig, 1996a; Janig et al., 1996). Which central pathways are involved leading to activation of preganglionic sympathetic neurons that innervate the adrenal medullae after subdiaphragmatic vagotomy? Are only sympathetic neurons which innervate the adrenal medullae activated after vagotomy or also other functional types of sympathetic neurons (Janig, 1985, 1996b; Jiinig and McLachlan, 1992, 1999; see Janig and Habler, Chapter 25, this volume)? Is the vago-sympathetic pathway to the adrenal medullae used by the brain in the modulation of the sensitivity of the nociceptor population? Do the same changes, related to the abdominal vagal afferents and the adrenal medulla, also occur in other behavioral pain models? For example, do the changes, probably induced by the hormonal signal in the cutaneous nociceptor population, also occur in deep somatic and

visceral nociceptive afferents? Finally, is it possible that these mechanisms operate in such ill-defined pain syndromes as irritable bowel syndrome, functional dyspepsia, fibromyalgia etc. (Wolfe et al., 1990; Mayer and Raybould 1993; Goebell et al., 1998)?

Summary We have shown that activity in subdiaphragmatic vagal afferents modulates mechanical hyperalgesic behavior in the rat. Subdiaphragmatic vagotomy decreases paw-withdrawal threshold to mechanical stimulation (baseline and after intradermal injection of bradykinin), thus enhancing mechanical hyperalgesic behavior. Most of this decrease is generated by an endocrine signal released by the adrenal medullae because denervation or removal of the adrenal medullae prevents or reverses these changes. This novel mechanism may imply that: (a) the brain is able to regulate sensitivity of nociceptors all over the body by a neuroendocrine mechanism, (b) sensitivity of nociceptors can be influenced by changes in parts of the body which are remote from the location of the sensitized nociceptors and (c) circulating catecholamines can influence nociceptors in a way which is different from those reported so far (see Janig and McLachlan, 1994; Janig, 1996a; Janig et al., 1996).

Acknowledgement This work was supported by NIH grant AR32634.

References Bernstein, I.L. (1996) Neural mediation of food aversions and anorexia induced by tumor necrosis factor and tumors. Neurosci. Biobehav. Rev., 20: 177-1 8 1. Berthoud, H.-R., Carlson, N.R. and Powley, T.L. (1991) Topography of efferent vagal innervation of the rat gastrointestinal tract. Am. J. Physiol., 260: R206R207. Berthoud H.-R. and Neuhuber W.L. (1994) Distribution and morphology of vagal afferents and efferents supplying the digestive system. In: Y. TachC, D.L. Wingate and T.F.Burks (Eds), Innervation of the Gut. Pathophysiological Implications, CRC Press, Boca Raton, pp. 43-66. Berthoud, H.-R., Kressel, M., Raybould, H.E. and Neuhuber, W.L. (1995) Vagal sensors in the rat duodenal mucosa: distribution and structure as revealed by in vivo DiI-tracing. Anat. Embryol., 191: 203-212. Berthoud, H.-R., Patterson, L.M., Willing, A.E., Mueller, K. and Neuhuber, W.L. (1 997) Capsaicin-resistant vagal afferent

285 fibres in the rat gastrointestinal tract: anatomical identification and functional integrity. Brain Res., 746: 195-206. Blatteis, C.M. and Sehic, E. (1997) Fever: How may circulating pyrogens signal to the brain? News Physiol. Sci., 12: 1-9. Bret-Dibat, J.L., Bluthe, R.M., Kent, S., Kelley, K.W. and Dantzer, R. (1995) Lipopolysaccharide and interleukin-1 depress food-motivated behavior in mice by a vagal-mediated mechanism. Brain Behav. Immunol., 9: 242-246. Engeland, W.C., Bereiter, D.F. and Gann, D.S. (1986) Sympathetic control of adrenal secretion of enkephalins after hemorrhage in awake dogs. Am. J. Physiol., 251: 341-348. Ericsson, A., Kovacs, K.J. and Sawchenko, P.E. (1994) A functional anatomical analysis of central pathways subserving the effects of interleukin-1 on stress-related neuroendocrine neurons. J. Neurosci., 14: 897-913. Fields, H.L. and Basbaum, A.I. (1999) Central nervous systemmechanisms of pain modulation. In: P.D. Wall and R. Melzack (Eds), Textbook of Pain, 4th Edn., Churchill Livingstone, Edinburgh London, pp. 309-329. Foreman, R.D. (1989) Organization of the spinothalamic tract as a relay for cardiopulmonary sympathetic afferent fiber activity. Pmg. Sens. Physiol., 9: 1-51. Gebhart, G. F. and Randich, A. (1992) Vagal modulation of nociception. Am. Pain. SOC.J., 1: 26-32. Goebell, H., Holtmann, G. and Talley, N. (Eds) (1998) Functional Dyspepsia and Irritable Bowel Syndrome: Concepts and Controversies, Kluwer Academic Publishers, Dordrecht, Boston, London. Gonzales, R., Goldyne, M.E., Taiwo, Y.O. and Levine, J.D. (1989) Production of hyperalgesic prostaglandins by sympathetic postganglionic neurons. J. Neurochem., 53: 1595-1 598. Green, P.G., Jtinig, W. and Levine, J.D. (1997) Negative feedback neuroendocrine control of inflammatory response in the rat is dependent on the sympathetic postganglionic neuron. J. Neumsci., 17: 3234-3238. Grundy, D. (1992) Vagal afferent mechanisms of mechano- and chemoreception. In: S. Ritter, R.C. Ritter and C.D. Barnes (Eds), Neuroanatomy and Physiology of Abdominal Afferents, CRC Press, Boca Raton, pp. 181-191. Grundy, D. and Scratcherd, T. (1989) Sensory afferents from the gastrointestinal tract. In: J.D. Wood (Ed.), Handbook of Physiology, Section 6, Vol. 1, Part 1. Am. Physiol. SOC., Bethesda, Maryland, pp. 593-620. Jhig, W. (1985) Organization of the lumbar sympathetic outflow to skeletal muscle and skin of the cat hindlimb and tail. Rev. Physiol. Biochem. Phrmacol., 102: 119-213. Jiinig, W. (1996a) The puzzle of ‘reflex sympathetic dystrophy’: mechanisms, hypotheses, open questions. In: W. Janig and M. Stanton-Hicks (Eds), Reflex Sympathetic Dystrophy - A Reappraisal, IASP Press, Seattle, pp. 1-24. Jiinig, W. (1996b) Spinal cord reflex organization of sympathetic systems. In: R. Bandler, G. Holstege and C.B. Saper (Eds), Pmg. Brain Res, 107: 43-77. Jiinig, W. (1996~)Neurobiology of visceral afferent neurons: neuroanatomy, functions, organ regulations and sensations. Biol. Psychol., 42: 29-35.

Janig, W. and Koltzenburg, M. (1993) Pain arising from the urogenital tract. In: G. Burnstock (Ed.), The Autonomic Nervous System, Vol. 2 (Maggi, G. (Ed.), Nervous Control of the Urogenital System), Hanvood Academic Publishers, Chur, Switzerland, pp. 523-576. Janig, W. and McLachlan, E.M. (1992). Characteristics of function-specific pathways in the sympathetic nervous system. Trends Neurosci., 15: 4 7 5 4 8 1. Janig, W. and McLachlan, E. M. (1994) The role of modifications in noradrenergic peripheral pathways after nerve lesions in the generation of pain. In: H.L. Fields and J.C. Liebeskind (Eds), Pharmacological Approaches to the Treatment of Pain: New Concepts and Critical Issues, Progress in Pain Research and Management, Vol. 1, IASP Press, Seattle, pp. 101-128. Janig, W. and McLachlan, E.M. (1999) Neurobiology of the autonomic nervous system. In: R. Bannister and C.J. Mathias (Eds),Autonomic Failure, 4th Edn., Oxford University Press, Oxford, pp. 3-15. Jhig, W., Levine, J. D. and Michaelis, M. (1996) Interaction of sympathetic and primary afferent neurons following nerve injury and tissue trauma. Prog,: Brain Res., 112: 161-184. Jtinig, W., Green, P.G., Khasar, S.G., Miao, F.J.-P., Strausbaugh, H. and Levine, J.D. (1998) Role of vagal afferents in neuroendocrine control of joint inflammation and cutaneous mechanical hyperalgesia. In: H. Goebell, G. Holtmann and N. Talley (Eds), Functional Dyspepsia and Irritable Bowel Syndrome: Concepts and Controversies, Kluwer Academic Publishers, Dordrecht, Boston, London, pp. 41-54. Jany, H., Duker, E.M. and Wuttke, W. (1986) Adrenal release of catecholamines and Met-enkephalin before and after stress as measured by a novel in vivo dialysis method in the rat. Neurosci. Lett., 60: 273-278. Khasar, S.G., Green, P.G. and Levine, J.D. (1993) Comparison of intradermal and subcutaneous hyperalgesic effects of inflammatory mediators in the rat. Neurosci. Lett., 153: 2 15-2 18. Khasar, S.G., Miao, F.J.-P. and Levine, J.D. (1995) Inflammation modulates the contribution of receptor-subtypes to bradykinin-induced hyperalgesia in the rat. Neuroscience, 69: 685-690. Khasar, S.G., Miao, F. J.-P., Janig, W. and Levine, J.D. (1998a) Modulation of bradykinin-induced mechanical hyperalgesia in the rat skin by activity in the abdominal vagal afferents. Europ. . I Neurosci., . 10: 435-444. Khasar, S.G., Miao, F. J.-P., Janig, W. and Levine, J.D. (1998b) Vagotomy-induced enhancement of mechanical hyperalgesia in the rat is sympathoadrenal-mediated. J. Neurosci., 18: 3043-3049. Kress, M., Koltzenburg M., Reeh, P.W. and Handwerker, H.O. (1992). Responsiveness and functional attributes of electrically localized terminals of cutaneous C-fibers in vivo and in vitro. J. Neurophysiol., 68: 581-595. Kwiat, G.C. and Basbaum, A.I. (1992) The origin of brainstem noradrenergic and serotonergic projections to the spinal cord dorsal horn in the rat. Somatosens. Mot. Res., 9: 157-173.

286 Lay& S., BluthC, RM., Kent, S., Combe, C, MBdina, C., Pamet, P., Kelley, K. and Dantzer, R. (1995) Subdiaphragmatic vagotomy blocks induction of IL-1 f3 mRNA in mice brain in response to peripheral LPS. Am. J. Physiol., 268: R13271331. Loewy, A. and Spyer, K.M. (Eds) (1990) Central Regulation of Autonomic Functions, Oxford University Press, New York. Mayer, E.A. and Raybould, H.E. (Eds) (1993) Basic and Clinical Aspects of Chronic Abdominal Pain, Pain Research and Clinical Management, Vol. 9, Elsevier, Amsterdam. McDonald, D.M. (1990) The ultrastructure and permeability of tracheobronchial blood vessels in health and disease. Eux Respir: J., 12 : 572s-585s. Mehler, S.T. and Gebhart, G.F. (1992) A critical review of the afferent pathways and the potential chemical mediators involved in cardiac pain. Neuroscience, 48: 501-524. Mense, S. and Meyer, H. (1988) Bradykinin-induced modulation of the response behaviour of different types of feline group I11 and IV muscle receptors. J. Physiol. (Lond.), 398: 49-63. Miao, F. J.-P., Janig, W. and Levine, J.D. (1996a) Role of sympathetic postganglionic neurons in synovial plasma extravasation induced by bradykinin. J. Neurophysiol., 75: 7 15-724. Miao, F. J.-P., Green, P., Coderre, T.J., J h g , W. and Levine, J.D. ( 1996b) Sympathetic-dependence in bradykinin-induced synovial plasma extravasation is dose-related. Neurosci. Lett., 205: 165-168. Miao, F. J.-P., Janig, W. and Levine, J.D. (1997a) Vagal branches involved in inhibition of bradykinin-induced synovial plasma extravasation produced by intrathecal nicotine and noxious stimulation in the rat. J. Physiol. (Lond.), 498: 473484. Miao, F. J.-P., Jhig, W., Green, P.G. and Levine, J.D. (1997b) Inhibition of bradykinin-induced plasma extravasation produced by noxious cutaneous and visceral stimuli and its modulation by vagal activity. J. Neurophysiol., 78: 1285-1292. Michaelis, M., Hlbler, H.-J. and Janig, W. (1996) Silent afferents: a further class of nociceptors? Clin. Exp. Pharmacol. Physiol., 23: 14-20. Neugebauer, V., Schaible, H.G. and Schmidt, R.F. (1989) Sensitization of articular afferents to mechanical stimuli by bradykinin. Pfriigers Arch., 415: 330-335. Niijima, A. (1992) The afferent discharges from sensors for interleukin-l beta in hepatoportal system in the anesthetized rat. J. Physiol. (Lond.j, 446: 236P. Niijima, A. (1996) The afferent discharges from sensors for interleukine l p in the hepatoportal system in the anesthetized rat. J. Auton. New. Syst., 61: 287-291. Precht, J.C. and Powley, T.L. (1990) The fibre composition of the abdominal vagus in the rat. Anat. and Embryol., 181: 101-1 15. Randich, A. and Gebhart, G . F. (1992) Vagal afferent modulation of nociception. Brain Res. Rev., 17: 77-99. Richards, W., Hillsley, K., Eastwood, C. and Grundy, D. (1996) Sensitivity of vagal mucosal afferents to cholecystokinin and

its role in afferent signal transduction in the rat. J. Physiol. (Lond.j, 497: 473-48 1. Ritter, S., Ritter, R.C. and Barnes, C.D. (Eds) (1992) Neuroanatomy and Physiology of Abdominal Afferents. CRC Press, Boca Raton. Schwartz, G.J. and Moran, T.H. (1994) CCK elicits vagal afferent activity arising from gastric and duodenal sites. In: J.R. Reeve, V. Eysselein and T.E. Solomon (Eds), Cholecystokinin, Vol 713, Ann. of NY Acad. Sci., New York, pp. 121-128. Schwartz, G.J. and Moran, T.H. (1996) Sub-diaphragmatic vagal afferent integration of meal-related gastrointestinal signals. Neurosc. Biobehav. Rev., 20: 47-56. Sehic, E. and Blatteis, C.M. (1996) Blockade of lipopolysaccharide-induced fever by subdiaphragmatic vagotomy in guinea-pigs. Brain Res., 726: 160-166. Shanahan, F. (1994) The intestinal immune system. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, Raven Press, New York, pp. 643-684. Taiwo, Y.O. and Levine, J.D. (1988) Characterization of the arachidonic acid metabolites mediating bradykinin and noradrenaline hyperalgesia. Brain Res., 458: 402406. Treede, R.-D., Meyer, R.A., Raja, S.N. and Campbell, J.N. (1992) Peripheral and central mechanisms of cutaneous hyperalgesia. Prog. Neurobiol., 38: 397421. Walls, E.K., Wang, F.B., Holst, M.-C., Phillips, R.J., Voreis, J.S., Perkins, A.R., Pollard, L.E. and Powley, T.R. (1995a) Selective vagal rhizotomies: a new dorsal surgical approach used for intestinal deafferentation. Am. J. Physiol., 269: R 1279-R 1288. Walls, E.K., Phillips, R.J., Wang, F.B., Holst, M.-C. and Powley, T.R. (1995b) Suppression of meal size by intestinal nutrients is eliminated by celiac vagal deafferentation. Am. J. Physiol., 269: R 1410-Rl4 19. Wan, W., Wetmore, L., Sorensen, C.M., Greenberg, A.H. and Nance, D.M. (1994) Neural and biochemical mediators of endotoxin- and stress-induced c-fos expression in the rat brain. Brain Res. Bull., 34: 7-15. Watkins, L.R., Wiertelak, E.P., Goehler, L.E., Mooney-Heiberger, K., Martinez, J., Furness, L., Smith, K.P. and Maier, S.F. (l994a) Neurocircuitry of illness-induced hyperalgesia. Brain Res., 639: 283-299. Watkins, L.R., Wiertelak, E.P., Goehler, L.E., Smith, K.P., Martin, D. and Maier, S.F. (1994b) Characterization of cytokine-induced hyperalgesia. Brain Res., 654: 15-26. Watkins, L.R., Goehler, L.E., Relton, J., Brewer, M.T. and Maier, S.F. (1995a) Mechanisms of tumor necrosis factor-a (TNF-alpha) hyperalgesia. Brain Rex, 692: 244-250. Watkins, L.R., Goehler, L.E., Relton, J., Tartaglia, N., Silbert, L., Martin, D. and Maier, S.F. (1995b) Blockade of interleukin-I induced hypertherrnia by subdiaphragmatic vagotomy: evidenced for vagal mediation of immune-brain communication. Neurosci. Lett., 183: 27-31. Watkins, L.R., Maier, S.F. and Goehler, L.E. (1995~)Cytokineto-brain communication: a review and analysis of alternative mechanisms. Life Sci., 57: 101 1-1026.

287 Watkins, L.R., Maier, S.F. and Goehler, L.E. (1995d) Immune activation: the role of pro-inflammatory cytokines in inflammation, illness responses and pathological pain states. Pain, 63: 289-302. Wolfe, F., Smythe, H.A., Yunus, M.B. et a1 (1990) The American College of Rheumtology 1990 criteria for the

classification of fibromyalgia: report of the multicenter criteria committee. Arthritis Rheum., 33: 160-172. Yeomans, D.C., Clark, EM., Paice, J.A. and Proudfit, H.K. (1992) Antinociception induced by electrical stimulation of spinally projecting noradrenergic neurons in the A7 catecholamine cell group of the rat. Pain, 48:449461.

This Page Intentionally Left Blank

SECTION VII

The influence of brain and mind on the body

This Page Intentionally Left Blank

E.A. Mayer and C.B. Saper (Eds.) Progress in BrQbI Research,Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAPTER 21

Affect, cognition, the immune system and health Margaret E. Kemeny'i29* and Tara L.Gruenewald'

'Depament of Psychology, University of California,Los Angeles, CA 90095, USA Department of Psychiatry and Biobehavioral Sciences, University of California, LQS Angeles, CA 90095, USA

Affect, cognition, the immune system and health

Stress and the immune system Naturalistic stressors

A growing body of evidence in the field of psychoneuroimmunology documents the linkages between the mind, the brain and the immune system (Ader et al., 1991). The central and autonomic nervous systems can engage in bidirectional communication with immunologic cells in blood, tissue and immune organs (see Chapter 27 by Felton and Chapter 4 by Sternberg in this volume for a discussion of this literature). This chapter will discuss the linkages between the mind and the immune system, focusing on the newer evidence that cognitive states have immunological correlates (Kemeny et al., 1992; Cole and Kemeny, 1997). It begins by reviewing research linking exposure to stressful life experiences to immune system functioning, primarily in humans. We then review the research on affect, the regulation of affect and the immune system including studies of affective disorders, such as major depression, mood states (such as depressed and anxious mood) and emotions such as sadness. We then turn to the research indicating that specific cognitive states, including beliefs, have immunological and health correlates.

*Correspondingauthor. Tel.: (3 10) 206-4954; Fax: (3 10) 206-5895; e-mail: [email protected]

A large body of research has documented the ability of naturalistic stress to affect various immune parameters measured in vitro. Stressors as diverse as medical school examinations, major life events (e.g. bereavement, separatioddivorce), and chronic care-giving have been shown to be associated with both enumerative and functional effects on the immune system. Exposure to naturalistic stressors has been associated with decreases in the total number of circulating leukocytes, as well as alterations in percent composition of different cell subtypes. 'Ifrpical enumerative effects include decreases in the percent of CD4 (helperhducer) T cells (Kiecolt-Glaser et al., 1986, 1987a), the number and percent of CD8 (cytotoxic) T cells (Schaeffer et al., 1985; McKinnon et al., 1989), and natural killer (NK) cell number and percent (Glaser et al., 1986; Herbert and Cohen, 1993). Naturalistic stressors can be associated with a decrease in immune functioning. A wide variety of stressors are associated with a decrease in the proliferative response of lymphocytes to mitogens (Dorian et al., 1982; Glaser et al., 1985; Schaeffer et al., 1985; Kiecolt-Glaser et al., 1987b), as well as decreased natural killer cell activity against tumor targets (NKCA; Glaser et al., 1986; Irwin et al., 1986, 1988). Reduced production of the cytokines interferon-g (IFN-g, Glaser et al., 1986) and interleukin-1P (IL-1p; Kiecolt-Glaser et al., 1995)

292

by stimulated lymphocytes and macrophages, respectively, in stressed individuals may be an important component of this decrement in immune functioning. Stress has also been associated with increased antibody titers to the latent herpes viruses Epstein-Barr virus (EBV) and herpes simplex virus (HSV; Glaser et al., 1987, 1993, 1994; KiecoltGlaser et al., 1987b), suggesting a decreased ability of the immune system to control viral latency. In total, these effects suggest that exposure to both major and minor life events can impact a variety of indicators of immune system functioning. It does not appear that these relationships are due to health behaviors as most of the analyses of stressors and immune processes have controlled for behaviors such as drug and alcohol use, exercise, sleep, nutrition, and others. It remains unclear whether immunologic changes observed in conjunction with stressors are responsible for observed relationships between stress and immunologically mediated or resisted diseases. Chronic stress, for example, has been shown to increase susceptibility to a upper respiratory infection following inoculation with rhinovirus (Cohen, S . et al., 1998), to predict HSV recrudescence (Cohen, F. et al., 1999a), and to slow wound healing (Kiecolt-Glaser et al., 1995). These clinical effects may result from stress-induced immune alterations but this has not been confirmed (Kemeny, 1994). There are animal models that support a stressor-immune system-health link. For example, Ben-Eliyahu and colleagues (1991) have shown that alterations in natural killer cell activity may underlie the effects of stress on metastases to the lung in rats with a breast tumor. In addition Bonneau and colleagues (1991, 1994) have shown that alterations in Herpes Simplex Virus (HSV) pathogenesis following stressor exposure is related to the suppression of HSV-specific cytotoxic T cells and natural killer cell activity in rodents. Experimentally induced stress

Complementing these studies of relatively longerterm naturalistic stressors are efforts to document the immune effects of brief stress induced in an experimental laboratory setting. Typical paradigms monitor changes in autonomic nervous system and

immune functioning while exposing participants to acute psychological stressors of brief (e.g. 20 min) duration, such as mental arithmetic, puzzle solving, and Stroop color-naming tasks. Physical exercise stressors and creative field approaches, such as parachute jumping, have also been employed. Consistent with studies of naturalistic stress, brief induced stress is also associated with a variety of immune changes. The direction of the changes are consistent with studies of naturalistic stressors for some immune parameters. For example, experimentally induced stress leads to a decrease in the proliferative response of lymphocytes to mitogens as observed with naturalistic stressors (Bachen et al., 1992; Herbert et al., 1994; Sgoutas-Emch et al., 1994). However, in contrast to the decrease in circulating NK cells and decreased NK function observed in response to naturalistic stress, brief induced psychological and physical stress is associated with increases in NK cell number (Bachen et al., 1992; Schedlowski et al., 1993; Sgoutas-Emch et al., 1994). A rise in NK cell cytotoxicity, tested in vitro, has also been observed (Naliboff et al., 1991; Murray et al., 1992; Schedlowski et al., 1993). Sympathetic nervous system (SNS) responses appear to mediate the increase in NK cell number and activity. Enumerative and functional effects vary with the intensity of sympathetic arousal (Herbert et al., 1994, Sgoutas-Emch et al., 1994; Benschop et al., 1995), are induced by injection of epinephrine and epinephrine agonists (Van Tits et al., 1990), and can be abolished with administration of P-adrenergic antagonists (Murray et al., 1992; Benschop et al., 1994). It is unclear whether increases in NK activity are due purely to SNS effects on NK cell number and redistribution (Bachen et al., 1995), or are only partially due to such enumerative changes (Schedlowski et al., 1993). Increases in CD8 cells also follow acute stress (Manuck et al., 1991; Nabiloff et al., 1991; Sgoutas-Emch et al., 1994), but it is unclear whether this increase is a result of increased numbers of NK cells which also express the CD8 + cell phenotype (Brosschot et al., 1992; Benschop et al., 1996). Overall, these immune effects are transitory, occurring within five minutes of stressor onset, and diminishing soon after with complete

293

return to baseline levels or below within thirty minutes to one hour The transitory increases in NK cell number and function found after acute stress may represent an integral part of the body’s adaptive ‘fight-flight’ reaction (Futterman et al., 1994). NK cells are lymphocytes capable of non-specific killing of virally infected cells and some tumor cells in vitro. They are considered first line of defense cells since they engage in rapid non-specific killing without prior exposure to antigen. It is known that stressrelated SNS activity stimulates the release of epinephrine and induces a variety of adaptive physiological effects on the cardiovascular and other systems in order to ready the host for a physical threat. SNS products also induce changes in the number and distribution of lymphocyte subsets including NK cells via their effects on leukocyte trafficking and modification of adhesion molecule expression (Ottaway and Husband, 1992). These stress effects on first line of defense cells may be a part of the body’s preparatory response to deal with physical threat and potential injury and infection that is also activated in response to psychological threats.

depressed mood). We will also review research on affective disorders, such a major depression, which involve the presentation of mood disturbance as well as vegetative signs and symptoms, such as sleep disturbance. Depression

Affect and the immune system There is a great deal of difference across individuals and within individuals across time in psychological responses to stressful life events. Individual differences in psychological responses to stress and in psychological characteristics that are brought to the stressful context may play a critical role in shaping the physiological changes associated with stressors (Kemeny and Laudenslager, 1999; Segerstrom et al., 1999). Affect has been proposed as a key mediator of the effects of stressful experience on physiology. In this review, we distinguish among classes of affective responses. The term emotion is used to refer to very short affective experiences with distinctive facial expressions that can be observed across cultures (e.g. happiness, sadness, anger, fear; Ekman, 1994). Moods are defined here as longer-term affective states (e.g. depressed mood, hostility, anxiety), that are accompanied by negative cognitions (e.g. negative thoughts about the self and future in

\

Major depression has been studied extensively with regard to immune system functioning. Patients with major depression have been found to show alterations in both enumerative and functional immune system parameters. A meta-analytic review of research evaluating the relationship between depression and immunity found that depression was reliably associated with an increase in the number of neutrophils and total leukocytes, but a decrease in total lymphocytes including B cells, NK cells, CD4 T cells, CD8 T cells, and total T cells (Herbert and Cohen, 1993b). Functional decrements are associated with depression, including decreased lymphocyte responsiveness to mitogens (e.g. Zisook et al., 1994; Castle et al., 1995), reduced NK cell activity (Irwin et al., 1986, 1987, 1990; Maes et al., 1992a; Zisook et al., 1994; Bauer et al., 1995), and blunted delayed-type hypersensitivity skin response to antigen (Hickie et al., 1993, 1995). These enumerative and functional effects appear to be more pronounced with increasing severity of depression (Irwin et al, 1987; Bauer et al., 1995) and age (Herbert and Cohen, 1993; Bauer et al., 1995). In addition to these immunosuppressive effects, Maes and colleagues (1994, 1995a) have suggested that depression is also associated with immune activation. Individuals with major depression have been shown to exhibit elevated levels of markers of immune activation, including elevated levels of activated T cells (HLADR, CD25+; Maes et al., 1992b) and B cells (HLA-DR+, CD19+, Maes et al., 1992c), and higher levels of g-FN (Maes et al., 1994), serum neopterin (Duch et al., 1984; Maes et al., 1994), and interleukins-1, 2, and 6 (Maes et al., 1995b; Seidel et al., 1995, 1996), in comparison to nondepressed individuals. While the mechanisms underlying concurrent activation of cells with deficiencies in their functional capacity have not been clearly established, Maes and associates

294

(1995a) have suggested that immune suppression effects may be the result of homeostatic feedback systems initiated during immune activation. Decreased lymphocyte response to mitogens and other immunosuppressive effects associated with depression could occur through a variety of pathways including hypothalamic-pituitary adrenal hyperactivation (Maes et al., 1991, 1996), which may be influenced by increases in IL- 1 (Maes et al, 1993), lymphocyte exhaustion as a result of overstimulation (Maier and Watkins, 1998), and decrements in responsiveness caused by increased concentrations of IFN-g (Fuchs et al., 1989). The immuno-enhancing effects of increases in IL- 1 may also be attenuated by concomitant increases in IL-1 receptor antagonist concentrations during depression (Maes et al., 1995c, 1997). Experimentally induced affect

Experimental studies have assessed the immunologic effects of induced mood states. These studies help to reconcile whether or not correlations between affect and immune status result from neurophysiological effects of affect on immunologic cells or the effects of immune cell products on the brain and mood (Maier and Watkins, 1998). Induction of both positive and negative mood states for brief periods have been shown to lead to increases in the number of CD8+ cells and NK cells, as well as natural killer cell activity (Futterman et al., 1994). Similar to effects witnessed in acute stress studies, induced emotion effects on immune parameters are short-term and appeared to be mediated by sympathetic activity, as immunologic changes are associated with increases in heart rate (Knapp et al., 1992; Futterman et al., 1994), and blood pressure (Knapp et al., 1992). Biological distinctions across emotions

An important and clinically relevant question is whether or not different affective states have distinctive biological correlates. A few studies have shown discriminable neurophysiological correlates across different emotional states suggesting that there may be a neurophysiological substrate that would allow for other biological differences. For

example, Ekman et al. (1983) found that different induced emotional states had distinctive A N S correlates. These researchers were able to discriminate between positive and negative emotions as well as between different negative emotions. In terms of the CNS, Lane and colleagues (1997) have used positron emission tomographic (PET) measurements of regional blood flow to show activity in different regions of the brain during pleasant and unpleasant affective states (Lane et al., 1997). Specifically, unpleasant affect was uniquely associated with activation of the bilateral occipito-temporal cortex and cerebellum, and left parahippocampal gyrus, hippocampus, and amygdala. In a separate preliminary investigation using PET technology, differences in brain activity were found for happiness, sadness and disgust (Lane et al., 1997). For example, increases in activity in the anterior insular cortex were found during recallinduced sadness but not disgust. Thus, the experience of specific emotions may be associated with activity in distinctive regions of the brain. If this is the case, it is then possible that distinctive patterns of ‘downstream’ endocrine and immunologic changes would be found with different affective states. Our laboratory has been interested in possible immunologic and health differences associated with distinct affective states. Using an experimental paradigm involving mood-specific improvisational monologues, we found that induced positive mood led to increased lymphocyte responsivity to PHA and induced negative mood led to a decreased response (see Fig. 1; Futterman et al., 1994). However, a decreased response to PHA occurred in response to recall of both positive and negative emotions in another study (although effects appear to be stronger for negative recall; Knapp et al., 1992). Differential antigen specific IgA antibody responses have been found to be associated with positive and negative daily mood ratings (Stone et al., 1987, 1994). We have also been interested in discriminating between two affective responses to loss, grief and depressed mood. In samples of HIV positive individuals, we found that the death of a partner (Kemeny et al., 1995) or a close friend (Kemeny and Dean, 1995) to AIDS was associated with

295

< r a

1 4.7

TIME 1

TIME

TIME 2

TIME 3

OF BLOOD D R A W

Fig. 1. Mean level of proliferative response to PHA-High dose (log transformed, counts per minx 10’) for the positive and negative mood states after a 20-min baseline period (Time l), after the 20-min ‘mood-induction’period (Time 2), and again after a 20-minrecovery period (Time 3). Standard error bars are indicated on one side of the mean only. (Futterman et al., 1994).

immunologic changes over time consistent with more rapid HIV progression. Bereaved individuals who reported elevated levels of depressed mood were more likely to show this immune decline. However, we found no evidence of negative immunologic changes over time associated with individual differences in level of grief in response to the loss (Kemeny and Dean, 1995). In fact, in one study, HIV positive individuals with high levels of grief (without depression) showed more positive immunological changes over time (in terms of CD4 levels and serum activation markers) than individuals with low levels of grief (Kemeny et al., 1999). It is possible that grief is a normal and ‘adaptive’ response to loss without the negative physiological implications of depression. This kind of distinction between grief and depression is supported by a study showing that HIV positive individuals with the trait of negative affectivity, which predisposes towards negative affect, experience more depression but less grief following a loss event (Wayment and Kemeny, 1999). We believe that the different patterns of immune associations with grief and depression may be attributable in part to the cognitive differences between these two affective states. While depression is often accompanied by negative beliefs about the self and the future, grief is not defined in such terms. In support of this notion, we have found that

the self-reproach aspect of depressed mood was a significant predictor of CD4 decline among a bereaved, HIV positive sample, while other subcomponents of depressed mood (e.g. sad affect, confusion, sleep disturbance) were not. Also, we have found that negative beliefs about the future are predictive of immune decline and accelerated mortality in bereaved HIV positive individuals, as will be described below. All studies of psychosocial factors predicting HIV progression control for baseline health and immune status, health behaviors such as drug and alcohol use, medication regimen, etc., demographics, and other potential confounding factors (Cole and Kemeny, 1999).

Affect regulation and the immune system There are large individual differences in the way in which individuals manage or regulate their affective experience. One important dimension is the extent to which individuals experience or inhibit their affective responses. For example, while some individuals show a correspondence between their self-reports of anxiety and physiological indicators of anxiety, individuals with a ‘repressive style’ report low levels of anxiety coupled with high behavioral and physiological indications of distress during a laboratory stressor (Weinberger et al., 1979). In fact, individuals with a repressive style show more autonomic reactivity to a stressor than those with other affective styles. It is believed that individuals who demonstrate a repressive style minimize or inhibit negative emotional responses (Jamner et al., 1988; Barger et al., 1997). Based on a variety of studies, this emotional inhibition is presumed to be automatic (however, there is some evidence for effortful processing; Weinberger et al., 1979). Individuals with this style of affect regulation have been found to have higher antibody titers to EBV (Esterling et al., 1990), suggesting inadequate control over viral latency. Inhibition of emotional expression has also been shown to have autonomic correlates. Gross and Levenson (1993) found that those instructed to inhibit their emotional expression in an experimental context showed increased electrodermal activity and some decrease in heart rate. Behavioral inhibition in other contexts has also been shown to

296

be associated with increases in electrodermal activity (Pennebaker and Chew, 1985) and overactive delayed-type hypersensitivity responses during social engagement with strangers (Cole et al., 1998). Pennebaker and his colleagues have found that brief experimentally induced emotional disclosure/ expression that involves writing about one’s deepest thoughts and feelings regarding a traumatic event promotes immediate decreases in skin conductance (Pennebaker et al., 1987) as well as increases in negative mood (although improvements in mood have been found over time; Pennebaker and Beall, 1986; Donnelly and Murray, 1991). In addition, this disclosure paradigm is associated with reduced health center visits over a six month follow-up period (Pennebaker and Beall, 1986), increased lymphocyte proliferation in response to mitogens (Pennebaker et al., 1988), increased hepatitis B antibody to a vaccination (Petrie et al., 1995), and decreased antibody to EBV-VCA, suggesting greater immunologic control of this latent virus (Esterling et al., 1994). There is some indication that the positive effects of disclosure are due to increases in emotional experience and cognitive processing, since the effects were linked to greater evidence of negative emotion and cognitive insight and understanding of the problem during the disclosure task (e.g. Esterling et al., 1994). In studies of a related construct, Davidson (1998) has examined individual differences in stable prefrontal activation asymmetry using recordings of brain electrical activity from the scalp. He and his colleagues have shown that individuals with greater relative right prefrontal activation report less positive and more negative dispositional mood and more behavioral inhibition than those with greater left prefrontal activation. They also respond more strongly to negative affective challenges (Davidson, 1998). These researchers have also shown that right frontally activated female college students had lower NK cell activity levels than leftactivated participants (Kang et al., 1991). Greater relative right-sided prefrontal activation in Rhesus monkeys has been associated with higher basal levels of cortisol (Kalin et al., 1998). Thus, the CNS correlates of negative mood coupled with

behavioral inhibition may have effects on both the endocrine and immune systems. A more comprehensive conceptualization of affect regulation has been formulated by Lane and colleagues. They describe ‘levels of emotional awareness’ as individual differences in the capacity to experience emotion in a differentiated and complex way (Lane et al., 1990). A person’s emotional responses to standardized scenes are rated in terms of levels of emotional awareness, including emotional experiences described as bodily sensations, action tendencies, single emotions, blends of emotion and combinations of blends. Lane and his colleagues have found that level of emotional awareness is associated with neuronal activity in certain brain regions, specifically the anterior cingulate cortex (Lane et al., 1998). These regions may be more involved in response to emotion cues in individuals better at detecting complex cues. In a separate study they have shown that attending to a subjective emotional response when viewing emotional pictures is associated with activity in different brain regions than is attending to spatial aspects of the same stimulus (Lane et al., 1997). Specifically, they found increased activity in the anterior cingulate cortex during attention to the emotional aspects of the stimuli. These data suggest to the authors that this region is related to emotional processing and representation, a notion that they note is consistent with Papez’ suggestion that this region is “the seat of emotional experience” (Papez, 1937). The data in these areas suggest that affect regulation style may moderate the relationship between affective experience and physiology. Studies of affect regulation suggest that the notion that negative emotions have ‘negative’ physiological correlates and that positive emotions have positive correlates is, at the very least, simplistic. Instead, the experience of emotions, whether positive or negative, may be more adaptive than their automatic or willful inhibition. In contrast, the experience of negative moods (e.g. depressed mood, hostility, anxiety), particularly persistent negative moods, may have negative biological and health consequences. Moreover, our studies suggest that it may be the negative cognitions (e.g. self-blame, negative expectancies)

297

accompanying certain negative moods that account for some of their negative biological effects (see below).

Cognitive states and the immune system In the field of psychoneuroimmunology, the vast majority of human research focuses on exposure to stressors and affective responses. Relatively recently, investigators have become interested in the possible biological correlates of cognitive states. While the importance of cognitive responses to stress is a central tenet of stress and coping theory (Lazarus and Folkman, 1984), it has been argued that the effects of cognitive states on health are mediated by the extent to which these states modulate ‘distress’. There are two problems with this presumption. First, it does not consider the possibility that different affective states may have different biological correlates. This kind of psychobiological specificity would be obscured by a single general distress measure or a composite negative mood measure. In fact, reliance on global distress measures may have resulted in the frequently observed failure to detect correlations between levels of distress and immunological parameters in studies of individuals exposed to a stressor. Secondly, the presumption that distress or affect mediates all effects of cognitive states on the immune system and health excludes the possibility that cognitive states themselves have biological correlates with health consequences. Below we review immunologic and health studies of cognitive states, including studies of beliefs about the future, beliefs about control and beliefs about the self. Beliefs about the ficture Generalized expectation styles such as dispositional optimism and pessimism have been studied as predictors of physical health outcomes. Dispositional optimism is the tendency to expect positive future outcomes and pessimism is the tendency to expect negative outcomes. Optimism has been found to be associated with positive health outcomes (Scheier and Carver, 1992) such as reduced complications during and faster recovery from coronary artery bypass surgery (Scheier et al.,

1989). However, some studies have found no health effects of dispositional optimism (e.g. Reed et al., 1994, 1999; Segerstrom et al., 1998) and one found that pregnant women who were more dispositionally optimistic were more likely to have a late pregnancy and birth complications. Dispositional pessimism has been found to be associated with negative health outcomes, such as increased risk of cancer mortality (Schulz et al., 1996), and lower self-rated health (Robinson-Whelen et al., 1997).A related constuct, general hopelessness, predicted all-cause mortality, cardiovascular and non-cardiovascular mortality, and incidence of cancer, myocardial infarction, and progression of carotid atherosclerosis in males enrolled in a longitudinal heart disease study (Everson et al., 1996, 1997). In one study, dispositional optimism buffered the effects of acute stress on specific lymphocyte subsets. However, optimists were more vulnerable to stress-related immune changes, including a decrease in NK cell activity, if the stressful experiences persisted (Cohen et al., 1999b). Thus, some of the above contradictions in findings may relate to differences in stressor parameters, such as chronicity. This conclusion is supported by an experimental study of stressor controllability which found that more dispositionally optimistic participants showed the greatest decrease in NK cell activity after exposure to an uncontrollable laboratory stressor of 20 min duration but the effects did not continue once the stressor was over (Sieber et al., 1992). Thus, optimists may rebound quickly in response to acute stressors but may have more difficulty when a stressor persists. Our laboratory group has conducted a series of studies on situation-specificand dispositional optimism in HIV positive individuals and consistently found that situation-specific beliefs about the future, but not stable optimistic dispositions, predicted positive immunological and health outcomes. In each case we evaluated the individual’s expectations about their future health as a predictor of HIV progression. In the first study, Reed and colleagues (1994) found that negative HIV-specific expectations among men with AIDS predicted a shorter survival time, controlling for medication regimen, health behaviors, health and immune status at the time of assessment of

298

expectations, and other possible confounders (see Fig. 2). These relationships were also not due to more global psychological states such as depression, social isolation, passive coping, and negative affectivity. Thus, the relationship between this cognitive state and survival time did not appear to be related to affective state, as we measured it. The shortest survival time was observed in the men with high levels of negative expectancies about future health and exposure to an AIDS related loss of a close individual over the past year. In a second study, Reed and colleagues (1998) set out to determine whether these two factors would also predict disease progression among men at the opposite end of the health continuum, those who were asymptomatic. In HIV positive men who had never had HIV-related symptoms and therefore little information on which to predict future health, we found that negative expectations about one’s future health coupled with exposure to a recent AIDS-related bereavement event predicted an

increased likelihood of the development of HIVrelated symptoms over the next two and a half to three and a half years, controlling for potential confounders as above (Reed et al., 1999). We have also determined if expectancies and bereavement have immunological correlates that might explain the observed health relationships. We selected a group of highly optimistic HIV positive men and a group of highly pessimistic HIV positive men stratified on CD4 T cell count at baseline. We found that negative HIV-specific expectancies plus exposure to a bereavement event predicted a number of immunological indicators of HIV progression over a two to three year follow-up period (Kemeny et al., 1999). Specifically, this group showed a more rapid loss of CD4 T cells and T cell function than the other groups. They also showed increased serum level of markers of immune activation (e.g. neopterin, p2 microglobulin) and cell surface markers of immune activation than those without these factors. Increases in these

0

Fig. 2. * Low negative expectancies group n = 50. + High negative expectancies group n = 24. Product-limit estimator survival probability curves for high and low Realistic Acceptance groups. Vertical axis represent value of survival estimator function. Curves are significantly different from one another; Wilcoxon X;.,,95= 7.453, p < .01. Median estimated survival time: low negative expectancies = 18 months; high negative expectancies = 9 months. (Reed et al., 1994)

299

markers have been shown to predict a more rapid rate of HIV progression. Again these effects were not due to confounders or more general psychological states. These data suggest that beliefs about future health may in fact have positive biological correlates that confer a health benefit in the context of HIV infection (although causal direction cannot be confirmed in these longitudinal studies). Similar findings that document a link between beliefs about a negative prognosis (termed ‘stoic acceptance’) and health have been found in studies of cancer patients (Greer et al., 1979). We have shown that these associations may generalize outside the context of HIV in one study of healthy law students. Segerstrom et al. (1998) found that situation-specific optimism (in this case, optimism about succeeding in law school) among healthy law students predicted higher CD4 T cell levels and higher NK cell activity during the stress of first semester law school.

Beliefs about control Controllability of a stressor has been investigated in animal and human studies. In rats, exposure to uncontrollable shocks is associated with decreases in CD4 T cells and the CD4/CD8 T cell ratio (Nakata et al. 1996), decreased lymphocyte response to the mitogens PHA and ConA (Laudenslager et al., 1983), and enhanced tumor growth, reduced tumor rejection rates, and decreased cancer survival time (Sklar and Anisman, 1979; Visintainer et al., 1982; Lewis et al., 1983). These effects are attenuated or not present in rats receiving controllable shock. However, these effects are not always reliably replicated. Maier and Laudenslager ( 1988) report that lymphocyte proliferation responses vary widely across identical replication studies employing animals. Inconsistencies in human studies occur, as well. One study with humans found a reversal effect with participants exposed to controllable mild shock and loud noise exhibiting a reduction in lymphocyte proliferation to C o d , while individuals in the uncontrollable group did not (Weiss et al., 1990). One possible explanation for these unstable results may be the measurement of in vitro lymphocyte

proliferation to antigen (Maier and Laundenslager, 1988). More consistent immune responses to uncontrollable shock in rats have been found when assessing in vivo antibody-specific responses to a novel antigen (e.g. Laudenslager et al., 1988). In addition to studies of actual controllability, there is one study of beliefs in control and their immunologic correlates in humans. Sieber and colleagues (1992) examined the acute effects of four conditions on NK cell activity: brief uncontrollable loud noise stress, loud noise with a key that the participant could use to switch off the noise temporarily (controllable stress), loud noise with a key that was not related to the noise (perceived control) and no loud noise (the neutral condition). Exposure to the uncontrollable noise stressor was associated with reduced NK cell activity. The individuals in the other conditions showed no immune decrement. In other words, those who in fact had control over the stress as well as those who believed they did but did not showed no stress related immune decrements. These data support the notion that not only does actual controllability buffer the effects of stress on immune processes in some contexts, but beliefs about control may be as protective (Sieber et al., 1992). These data are consistent with research on ‘illusions’ documenting that illusions of control have positive psychological correlates and may represent adaptive responses to stressful life experiences (Taylor, 1983).

Beliefs about the self We have been interested in whether beliefs about the self predict immune processes in HIV positive individuals. In the first study of HIV positive men in New York we found that only the self-reproach aspect of depressed mood predicted CD4 T cell decline over time in men who had been exposed to an AIDS-related bereavement event (Kemeny and Dean, 1995). We have also examined coping interviews of men with HIV infection and coded them for style of attribution of negative events. One type of attribution of negative events involves the assumption that the cause of the event has to do with a negative quality that is a stable aspect of oneself. We found that those men who attributed negative events to such characterological aspects of

300

themselves were more likely to demonstrate CD4 decline over time (Segerstrom et al., 1996). This attributional style in healthy adults at age 25 was associated with worse physician-rated physical health at ages 45-60 (Peterson et al., 1988). These self-blaming attributions were also associated with decreased CD4 + /CD8 + cell ratios and reduced lymphocyte response to mitogens among a sample of elderly adults (Kamen-Siege1et al., 1991). In a separate study designed to evaluate the ability of psychological inhibition to predict HIV progression, Cole et al. (1996a) used concealment of homosexuality as a model of psychological inhibition. Cole found that HIV positive gay men ‘in the closet’ showed accelerated time to a critically low CD4 level, to an AIDS diagnosis and to mortality over a nine year follow-up period compared to men out of the closet. Among the HIV negative men in the study, being in the closet was associated with a greater risk of infectious illness and skin cancer (Cole et al., 1996b). In order to determine the reasons why being in the closet predicted more rapid HIV progression additional analyses were undertaken. Results indicated that ‘closeted’ individuals were particularly sensitive to social rejection and that rejection sensitivity was an even better predictor of CD4 decline, AIDS onset and mortality than degree of concealed homosexuality (Cole et al., 1997; see Fig. 3). In these studies, three distinct measures of negative beliefs about the self-predicted immune and clinical indicators of HIV progression. These effects were not due to health behavior and other confounders and were also not due to depression or other more global psychological states, such as general self-esteem. The common psychological element in these studies was self-blame and sensitivity to blame by others. The relationship between beliefs about the self and the immune system is also supported by an experimental study documenting that experimental enhancement of negative self-evaluation decreased NK cell activity in healthy adults (Strauman et al., 1993). Low selfesteem has also predicted a failure to show habituation of adrenocortical stress responses to repeated psychological stress, which may explain some of these immunological effects (Kirschbaum et al., 1995).

1.0 1

0’ , 0

1

2

4 5 8 7 YIpn.IlEOMACSentry

3

8

9

10

Fig. 3. Times to AIDS onset and HIV-related mortality for gay men at the 75th percentile (closed points) and 25th percentile (open points) on homosexuality-specific rejection sensitivity. Values come from Cox proportional hazards regressions controlling for all biobehavioral covariates, and estimates are truncated at median event time to facilitate comparison and to avoid extrapolation beyond observed data. MACS = Multicenter AIDS Cohort Study. (Cole et al., 1997).

The relationships among beliefs about the future, control and self and immune processes and health are supported by clinical trials which have demonstrated that cognitive-behavioral interventions specifically designed to alter negative cognitions such as self-blame can improve immune parameters such as NK cell activity (see Fig. 4;Fawzy et al., 1990) and cancer prognosis and survival (Fawzy et al., 1993). These types of interventions have also impacted relevant immune parameters in HIV positive individuals in some studies (Antoni et al., 1991) but not others (Kemeny and Miller, 1999). In our studies of beliefs about the future and beliefs about the self we have not found that relationships with health outcomes were mediated by generalized distress measures or negative mood states, such as depression or anxiety. In other words, negative expectancies did not predict HIV progression because negative expectancies led to or was accompanied by depressed mood. These findings have led us to suspect that cognitive states may have biological correlates apart from their associa-

30 1

Fig. 4. Interferon alfa-augmented natural killer cell (NK) activity (percent lysis of target cells at 25 : 1 effector target cell ratio) at baseline, six weeks, and six months in the interventiongroup (light bars) (n= 17) and control (dark bars) (n= 16) patients, with the unshaded portion of bars indicating SE. (Fawzy et al., 1990).

tion with affect. However, it is also possible that we are looking in the wrong place for affect mediators. In a recent study, we found that persistent feelings of HIV-related shame and guilt predicted accelerated CD4 decline, AIDS onset and death among a sample of HIV positive men (Weitzman et al., 1999). It is possible that our observations that selfblame and analogous concepts predicted HIV progression were due to self-blame related shame and guilt. Thus, while depression and other mood states did not mediate the relationship of self-blame on HIV progression, the emotional reactions of shame and guilt may have. Future studies of cognitive states will help to clarify the mediators of these effects. Defeat, disengagement and giving up

Could a specific psychobiological state underlie the relationships between many of the psychological predictors we have evaluated and immune processes and health? We propose that a generalized state of engagemenddisengagementwith important life goals may be a critical pathway. A variety of cognitive and affective states can induce disengagement with important life goals. We propose that three key inducers of disengagement are negative expectations about the future, persistent self-blame, and psychological inhibition. Maintenance of life goal engagement may require a more positive view

of the future and oneself as well as an ability to experience and express one’s thoughts and feelings in regard to traumatic or threatening experiences. In support of this theory, our research group has found that the relationship between negative expectancies and immune decline in bereaved HIV positive individuals is mediated by disengagement with life goals (Kemeny et al., 1999). In other words, bereaved individuals who reported negative expectations about their future health only show decrements in immune parameters if they are more disengaged with their life goals. Moreover, disengagement with life goals itself was a significant predictor of immune decline. To examine this issue in another way we selected from our cohort of HIV positive men those who met the criteria for ‘longterm non-progression’, defined as no AIDS diagnosis or significant CD4 decrement after 10 years of HIV infection. We compared this small group to a carefully matched group of more rapid progressors and found that engagement with life goals at the baseline assessment point was a significant predictor of long-term non-progression (Kemeny et al., 1999). In both studies, effects were found only for ‘intrinsic goals’ that involved internally generated motivations such as the search for meaning and the importance of relationships. A measure of ‘extrinsic’ life goals that reflected a preoccupation with others’ view of themselves (such as being popular or attractive, having possessions, etc.) did not have immunologic associations. A third study from our research group also supports the importance of engagement with intrinsic life goals. Bower and colleagues (1998) specifically focused on finding meaning, an intrinsic life goals, among bereaved HIV positive men. While there is substantial evidence that traumatic events can result in significant dysphoria, there is a growing body of data indicating that some individuals who have confronted a trauma show psychological benefits such as finding meaning from the experience (Taylor, 1983; Affleck and Tennen, 1996). Finding meaning can involve a greater appreciation of life and relationships, a move towards a more spiritual viewpoint, and a more positive view of the self. Finding meaning has been associated with positive psychological changes as well as, in one study, lower morbidity

302

over an eight-year follow-up among individuals who had suffered a myocardial infarction (Affleck et al. 1987). In the Bower study (1998), those who found meaning from the loss event (based on intensive bereavement interviews) were significantly less likely to show CD4 decline and less likely to die from AIDS over the next nine years. Finding meaning was most likely to occur in those bereaved individuals who engaged in ‘cognitive processing’ regarding the loss event. Rather than inhibiting their thoughts and feelings about the loss event, these individuals attempted to deliberately think about the death and its implications for their lives. Among those who engaged in cognitive processing, about half found meaning from the bereavement experience. Cognitive processing without finding meaning did not predict health benefits. Thus, cognitive and emotional processing of traumatic experiences may be a critical component in the maintenance of intrinsic life goals, such as finding meaning, following a traumatic event. These findings are supported by a recent study from our laboratory demonstrating that psychological inhibition, the opposite of cognitive and emotional processing, was associated with lower CD4 cell levels in HIV positive women, irrespective of level of depression and other psychological and behavioral factors (Eisenberger et al., 1999). In this engagemenvdisengagement model, we propose that for those exposed to a significant trauma, such as bereavement or the threat of mortality, psychological factors that reduce one’s engagement with intrinsic life goals will be associated with negative biological effects. Specifically, negative expectations, self-blame and inhibition of thoughts and feelings regarding the event will be capable of leading to disengagement with life goals and consequent health decrements. Consistent with this theory, we have shown that the impact of negative beliefs about one’s future health on HIVrelated immune parameters is a result of the effects of negative expectancies on disengagement with intrinsic life goals. The notion that negative expectations can lead to disengagement is consistent with the theoretical position of Scheier and Carver (1992) who argue that people who believe that their goals are attainable will continue to strive towards

them while those who do not will disengage from the goals. We have also shown that psychological inhibition, or the failure to emotionally and cognitively confront a traumatic experience, can lead to a failure to find meaning, an intrinsic life goal, and an increased risk of death to AIDS (Bower et al., 1998). Goal disengagement may explain some of the effects of chronic stress and exposure to traumatic events on health. Chronic stress or severe trauma can alter one’s expectations about the future, damage one’s sense of self or induce self-blame, and cause the inhibition of painful emotional responses. These cognitive responses to stress could then cause the individual to disengage from what is important to them, or give up, which may have physiological effects. There is a long history of focus on concepts closely allied with disengagement. For example, Victor Frank1 (1963) vividly described the loss of the will to live that can occur with severe trauma such as confinement in a concentration and the central role that meaning played in withstanding such dire circumstances. The will to live and engagement with life goals are closely allied concepts. In another formulation, Engel and Schmale (1972) argued that conservation-withdrawal is a behavioral and biological response in which survival is supported by disengagement, inactivity and conservation of resources. ‘Giving up’ in response to an event can be followed or accompanied by conservation-withdrawal,according to these authors. The concept of vital exhaustion has close conceptual overlap with disengagement. The measure of vital exhaustion includes physical exhaustion (e.g. excess tiredness) as well as items such as ‘giving up trying’, ‘dead end’, and others that reflect a kind of disengagement. Vital exhaustion has been shown to precede cardiovascular events such as myocardial infarctions in some studies (Appels and Mulder, 1989). There are animal models that may provide support for the importance of goal engagement as well as an understanding of the physiological mechanisms that may underlie these effects. One relevant model is the defeat model studied primarily in rodents. In this paradigm, an adult male rat is introduced into the home cage of another

303

male rat. The intruder is attacked by the resident and the intruder eventually develops a set of submissive postures and behaviors (e.g. lying motionless on the back, freezing, decreased activity, suppressed food intake). ‘Defeat’ has been associated with opioid mediated analgesia (Mizcek et al., 1982) as well as elevations in ACTH and cortisol, increased sympathetic activity, alterations in body temperature, reduced body weight, and release of dopamine in the prefrontal cortext (e.g. Tidey and Miczek, 1996). In addition, defeat behaviors have been associated with lowered immune functioning (Fleshner et al., 1989; Stefanski, 1998) and increased tumor metastases (Stefanski and Ben-Eliyahu, 1996). Some of these effects are long term despite removal of the rodent from exposure to the home-cage animal immediately after the defeat posture is induced (Koolhaas et al., 1997). We are interested in whether or not there is a human individual difference factor with biological correlates that could represent a human analog to defeat. (Other animal models of individual differences in coping with stress may also be relevant; see Bandler et al., Chapter 24, this volume.) In sum, there is extensive evidence that both naturalistic stressors and experimentally induced stress can affect functional and enumerative aspects of the immune system. However, individuals vary dramatically in their psychological responses to stressful events and a growing literature supports the link between specific cognitive and affective responses and immune processes. While the role that cognition plays in psychoneuroimmunology is a relatively new area of investigation, our own studies suggest that specific cognitive factors (i.e. negative expectancies, self-blame, psychological inhibition) may be related to immunological processes relevant to health. Moreover, some of these cognitive states may affect physiological parameters independent of their affective accompaniments. Many of the cognitive states that we and others have found to be associated with physiological processes and health may be important to the extent that they influence a major motivational domain, engagement or disengagement with life goals. There may be at least two evolution-based

responses to stressors, the fighdflight reaction, which involves activity designed to deal with the threat, and the defeat reaction, which involves disengagement (Henry and Grim, 1990). Each may involve coordinated behavioral and biological responses to the stressor designed to serve distinct adaptational goals (Weiner, 1992). Therefore, the biological correlates of each may be different and have unique health consequences if the response is prolonged. The notion that giving up or disengagement may have important health implications is not a new one, and has its origins in psychosomatic medicine as well as some ancient philosophies. We resurrect this concept and encourage research using modem methodologies to determine its importance in health and disease.

Acknowledgements The research reviewed in this chapter was supported by NIMH grant MH42918, a Research Scientist Development Award from the NIMH (MH00820), the UCLA Psychoneuroimmunology Program and the Bing Foundation.

References Ader, R., Felten, D.L. and Cohen, N. (1991) Psychoneuroimmunology, 2nd Edn., Academic Press, Orlando. Affleck, G., Croog, S. and Tennen, H. (1987) Causal attributions, perceived benefits, and morbidity after a heart attack: an 8-year study. J. Consult. Clin.Psychol., 55: 29-35. Affleck, G. and Tennen, H. (1996) Construing benefits from adversity: adaptional significance and dispositional underpinnings. J. Personality, 64: 899-922. Antoni, M.H., Baggett, L., Ironson, G., LaPierre, A., August, S., Klimas, N., Schneiderman, N. and Fletcher, M.A. (1991) Cognitive-behavioralstress management intervention buffers distress responses and immunologic changes following notification of HIV-1 seropositivity. J. Consult. Clin. Psychol., 59: 906-915. Appels, A. and Mulder, P.(1989) Fatigue and heart disease. The association between ‘viral extraction’ and past, present and future coronary heart disease. J. Psychosom. Res., 33: 727-738. Bachen, E.A., Manuck, S.B., Cohen, S., Muldoon, M.F., Raible, R., Herbert, T.B. and Rabin, B.S. (1995) Adrenergic blockade ameliorates cellular immune responses to mental stress in humans. Psychosom. Med., 57: 366-372. Bachen, E.A., Manuck, S.B., Marsland, A.L., Cohen, S., Malkoff, S.B., Muldoon, M.F. and Rabin, B.S. (1992) Lymphocyte subset and cellular immune responses to a brief experimental stressor. Psychosom. Med., 54: 673-679.

304 Barger, S.D., Kircher, J.C. and Crooyle, R.T. (1997) The effects of social context and defensiveness on the physiological responses of repressive copers. J. Personality SOC. Psychol., 73: 1118-1128. Bauer, M.E., Gauer, G.J., Luz, C., Silveira, R.O., Nardi, N.B. and von Muhlen, C.A. (1995) Evaluation of immune parameters in depressed patients. Life Sciences, 57: 665-674. Ben-Eliyahu, S., Y d y a , R., Liebeskind, J.C., Taylor A.N. and Gale, R.P. (1991) Stress increases metastatic spread of a mammary tumor in rats: evidence for mediation by the immune system. Brain, Behav. Immun.,5 : 193-205. Benschop, R.J., Godaert, G.L.R., Geenen, R., Brosschot, J.F., DeSmet, M.B.M., Olff, M., Heijnen, C.J. and Ballieux, R.E. (1995) Relationships between cardiovascular and immunological changes in an experimental stress model. Psychol. Med., 25: 323-327. Benschop, R.J., Nieuwenhuis, E., Tromp, E., Godaert, G., Ballieux, R.E. and van Doornen, L. (1994) Effects of padrenergic blockade on immunologic and cardiovascular changes induced by mental stress. Circulation, 89: 762-769. Benschop, R.J., Rodriguez-Feuerhahn, M. and Schedlowski, M. (1996) Catecholamine-induced leukocytosis: early observations, current research, and future directions. Brain, Behav. Immun., 10: 77-91. Bonneau, R.H. (1994) Experimental approaches to identify mechanisms of stress-induced modulation of immunity to herpes simplex virus infection. In: R. Glaser and J. KiecoltGlaser (Eds), Handbook of Human Stress and Immunity, Academic Press, New York. Bonneau, R.H., Sheridan, J.F., Feng, N. and Glaser, R. (1991) Stress-induced suppression of herpes simplex virus (HSV)specific cytotoxic T lymphocytes and natural killer cell activity and enhancement of acute pathgenesis following local HSV infection. Brain, Behav. Immun.,5: 170-192. Bower, J., Kemeny, M.E., Taylor, S.E. and Fahey, J.L. (1998) Cognitive processing, discovery of meaning, CD4 decline, and AIDS-related mortality among bereaved HIV seropositive men. J. Consult. Clin. Psychol., 66: 979-986. Brosschot, J.F., Benschop, R.J., Godaert, G.L., DeSmet, M.B.M., Olff, M., Heijnen, C.J. and Ballieux, R.E. (1992) Effects of experimental psychological stress on distribution and function of peripheral blood cells. Psychosom. Med., 54: 394-406. Castle, S., Wilkins, S., Heck, E., Tanzy, K. and Fahey, J. (1995) Depression in caregivers of demented patients is associated with altered immunity: impaired proliferative capacity, increased CD8 + , and a decline in lymphocytes with surface signal transduction molecules (CD38 + ) and a cytotoxicity marker (CD56+ CD8+). Clin. Expel: Immun., 101: 487-193. Cohen, S., Frank, E., Doyle, W.J., Skoner, D.P., Rabin, B.S. and Gawaltney, J.M. Jr. (1998) Qpes of stressors that increase susceptibility to the common cold in healthy adults. Healrh Psychol., 17: 214-223. Cohen, F., Kearney, K., Zegans, L., Kemeny, M.E., Neuhaus, J. and Stites, D. (1999b) Differential immune system changes

with acute and persistent stress for optimists vs. pessimists. Brain, Behav. Immun., 13: 155-174. Cohen, F., Kemeny, M.E., Keamey, K., Zegans, L., Neuhaus, J., Conant, M. and Stites, D. (1999a) Persistent stress as a predictor of genital herpes recurrence. Arch. Int. Med., 159: 2430-2436. Cole, S.W. and Kemeny, M.E. (1997) The psychobiology of AIDS. Crit. Rev. Neurobiol., 11: 289-321. Cole, S.W. and Kemeny, M.E. (2000) Psychosocial influences on the progression of HIV infection. In R. Ader, D.L. Felten and N. Cohen (Eds), Psychoneuroimmunology, 3rd Edn., Academic Press, New York, in press. Cole, S.W., Kemeny, M.E., Taylor, S.E. and Visscher, B.R. (1996b) Elevated physical health risk among gay men who conceal their homosexual identity. Health Psychol., 15: 243-25 1. Cole, S.W., Kemeny, M.E., Taylor, S.E. and Visscher, B.R. (1997) Social identity and physical health: accelerated HTV progression in rejection-sensitive gay men. J. Personal. SOC. Psychol., 72: 320-335. Cole, S.W., Kemeny, M.E., Taylor, S.E., Visscher, B.R. and Fahey, J.L. (1996a) Accelerated course of human immunodeficiency virus infection in gay men who conceal their homosexual identity. Psychosom. Med., 58: 2 19-231. Cole, S.W., Kemeny, M.E., Weitzman, O.B. and Anton, P. (1999) Socially inhibited individuals show heightened DTH response during intense social engagement. Brain, Behav. Immun.,13: 187-200. Davidson, R.J. (1998) Affective style and affective disorders: perspectives from affective neuroscience. Cogni?. Emor., 12: 307-330. Donnelly, D.A. and Murray, E.J. (1991) Cognitive and emotional changes in written essays and therapy interviews. J. SOC. Clin. Psychol., 10: 334-350. Dorian, B.J., Garfinkel, P., Brown, G., Shore, A., Gladman, D. and Keystone, E. (1982) Aberrations in lymphocyte subpopulations and function during psychological stress. Clin. Exp. Immunol., 50: 132-138. Duch, D.S., Woolf, J.H., Nichol, C.A., Davidson, J.R. and Garbutt, J.C. (1984) Urinary excretion of biopterin and neopterin in psychiatric disorders. Psychiatry Res., 11: 83-89. Eisenberger, N., Kemeny, M.E. and Wyatt, G. (1999) Psychological inhibition and CD4 T space cell level in HIV positive women. Submitted. Ekman, P. (1994) Moods, emotions, and traits. In: R.J. Davidson, P. Ekman and K. Scherer (Eds), The Nature of Emotion: Fundamental Questions, Oxford University Press, New York, pp. 56-58. Ekman, P., Levenson, R.W. and Friesen, W.V. (1983) Autonomic nervous system activity distinguishes among emotions. Science, 4616: 1208-10. Engel, G.L. and Schmale, A.H. (1972) Conservation-withdrawal: a primary regulatory process for organismic homeostasis. Ciba Foundat. Symp., 8: 57-75. Esterling, B.A., Antoni, M.H., Kumar, M. and Schneiderman, N. (1990) Emotional repression, stress disclosure responses,

305 and Epstein-Burr viral capsid antigen titers. Psychosom. Med., 52: 397410. Esterling, B.A., Antoni, M.H., Fletcher, M.A., Margulies and Schneiderman, N. (1994) Emotional disclosure through writing or speaking modulates latent Epstein-Ban virus antibody titers. J. Consult. Clin. Psychol., 62: 130-140. Everson, S., Goldberg, D.E., Kaplan, G.A., Cohen, R.D., Pukkala, E., Tuomilehto, J. and Salonen, J.T. (1996) Hopelessness and risk of mortality and incidence of myocardial infarction and cancer. Psychosom. Med., 58: 113-121. Everson, S.A., Kaplan, G.A., Goldberg, D.E., Salonen, R. and Salonen, J.T. (1997) Hopelessness and 4-year progression of carotid atherosclerosis.Arterioscles Thmmb. Vasc. Biol., 17: 1490-1 495. Fawzy, F.1, Fawzy, N.W., Hyun, C.S., Elashoff, R., Guthrie, D., Fahey, J.L. and Morton, D.L. (1993) Malignant melanoma: effects of an early structured psychiatric intervention, coping, and affective state on recurrence and survival 6 years later. Arch. Gen. Psychiatry, 50: 681. Fawzy, F., Kemeny, M., Fawzy, N., Elashof, R., Morton, D., Cousins, N. and Fahey, J.L. (1990) A structured psychiatric intervention for cancer patients: 11. Changes over time in immunological measures. Arch. Gen. Psychiatry, 47: 729-735. Fleshner, M., Laudenslager, M.L., Simons, L. and Maier, S.F. (1989) Reduced serum antibodies associated with social defeat in rats. Physiol. Behav., 45: 1183-7. Frankl, V.E. (1963) Man’s Searchfor Meaning: An Introduction to Logotherapy, Washington Square Press, New York. Fuchs, D., Malkovsky, M., Reibnegger, G., Werner, E.R., Forni, G. and Wachter, H. (1989) Endogenous release of interferongamma and diminished response of peripheral blood mononuclear cells to antigenic stimulation. Immunol. Lett., 23: 103-108. Futterman, A.D., Kemeny, M.E., Shapiro, D. and Fahey, J. L. (1994) Immunological and physiological changes associated with induced positive and negative mood. Psychosom. Med., 56: 599-5 11. Glaser, R., Kiecolt-Glaser, J.K., Stout, J.C., Tam, K.L., Speicher, C.E. and Holliday, J.E. (1985) Stress-related impairments in cellular immunity. Psychiatry Res., 16: 233-239. Glaser, R., Pearl, D.K., Kiecolt-Glaser,J.K. and Malarkey, W.B. (1994) Plasma cortisol levels and reactivation of latent epstein-ban virus in response to examination stress. Psychoneuroendocrinology, 19: 765-772. Glaser, R., Pearson, G.R., Bonneau, R.H., Esterling, C.A. and Kiecolt-Glaser, J.K. (1993) Stress and the memory t-cell response to the epstein-barr virus in healthy medical students. Health Psychol., 12: 435-442. Glaser, R., Rice, J., Sheridan, J., Fertel, R., Stout, J., Speicher, C., Pinsky, D., Kotur, M., Post, A., Beck, M. and KiecoltGlaser, J. (1987) Stress-related immune suppression: health implications. Brain, Behav. Immun., 1: 7-20. Glaser, R., Rice, J., Speicher, C.E., Stout, J.C. and KiecoltGlaser, J.K. (1986) Stress depresses interferon production by

leukocytes concomitant with a decrease in natural killer cell activity. Behav. Neurosci., 100: 675-678. Greer, S., Morris, T. and Pettingale, K.W. (1979) Psychological response to breast cancer: effect on outcome. Lancet, 8146: 785-7. Gross, J.J. and Levenson, R.W. (1993) Emotional suppression: Physiology, self-report, and expressive behavior. J. Personality SOC. Psychol., 64: 970-986. Henry, J.P. and Grim, C.E. (1990) Psychosocial mechanisms of primary hypertension. Editorial review. J. Hypertens., 8: 783-793. Herbert, T.B. and Cohen, S. (1993b) Depression and immunity: a meta-analytic review. Psychol. Bull., 113: 472-486. Herbert, T.B. and Cohen, S. (1993a) Stress and immunity in humans: a meta-analytic review. Psychosom. Med., 55: 364-379. Herbert, T.B., Cohen, S., Marsland, A.L., Bachen, E.A., Rabin, B.S., Muldoon, M.F. and Manuck, S.B. (1994) Cardiovascular reactivity and the course of immune response to an acute psychological stressor. Psychosom. Med., 56: 337-344. Hickie, I., Hickie, C., Bennett, B., Wakefield, D., Silove, D., Mitchell, P. and Lloyd, A. (1995) Biochemical correlates of in vivo cell-mediated immune dysfunction in patients with depression: a preliminary report. Int. J. Immunophamacol., 17: 685-690. Hickie, I., Hickie, C., Lloyd, A., Silove, D. and Wakefield, D. (1993) Impaired in vivo immune responses in patients with melancholia. Bs J. Psychiatry, 162: 651-657. Irwin, M., Daniels, M., Bloom, E. T., and Weiner, H. (1986) Life events, depression, and natural killer cell activity. Psychophannacol. Bull., 22: 1093-1096. Irwin, M., Daniels, M., Risch, S.C., Bloom, E. and Weiner, H. (1988) Plasma cortisol and natural killer cell activity during bereavement. Biol. Psychiatry, 24: 173-178. Irwin, M., Daniels, M., Smith, T.L., Bloom, E. and Weiner, H. (1987) Impaired natural killer cell activity during bereavement. Brain, Behav. Immun., 1: 98-104. Irwin, M., Patterson, T., Smith, T.L., Caldwell, C., Brown, S.A., Gillin, J.C. and Grant, I. (1990) Reduction of immune function in life stress and depression. Biol. Psychiatry, 27: 22-30. Jamner, I.D., S c h w a , G.E. and Leigh, H. (1988) The relationship between repressive and defensive coping styles and monocyte, eosinophile, and serum glucose levels: support for the opioid peptide hypothesis of repression. Psychosom. Med., 50: 567-575. Kalin, N.H., Larson, C., Shelton, S.E. and Davidson, R.J. (1998) Asymmetric frontal brain activity cortisol, and behavior associated with fearful temperament in Rhesus monkeys. Behav. Neumsci., 112: 286-292. Kamen-Siegel, L., Rodin, J., Seligman, M.E. and Dwyer, J. (1991) Explanatory style and cell-mediated immunity in elderly men and women. Health Psychol., 10: 229-235. Kang, D.H., Davidson, R.J., Coe, C.L., Wheeler, R.W., Tomarken, A.J. and Ershler, W.B. (1991) Frontal brain

306 asymmetry and immune function. Behav. Neurosci., 105: 860-869. Kemeny, M.E. (1994) Stressful events, psychological responses and progression of HIV infection. In: R. Glaser and J. Kiecolt-Glaser (Eds), Handbook of Human Stress and Immunity, Academic Press, New York, pp. 245-266. Kemeny, M.E. and Dean, L. (1995) Effects of AIDS-related bereavement on HIV progression among gay men in New York City. AIDS Educ. Prev., 7 36-47. Kemeny, M.E. and Laudenslager, M. (1999) Beyond stress: The role of individual differences factors in psychoneuroimmunologic relationships. Brain, Behav. Immun., in press. Kemeny, M.E. and Miller, G. (1999) Effects of psychosocial interventions on the immune system. In: M. Schedlowski and U. Tewes (Eds), Pscyhoneuroimmunology: An Interdisciplinary Introduction, Plenum Publishing, New York. Kemeny, M., Reed, G., Taylor, S., Visscher, B. and Fahey, J.L. (1999) Negative HIV-specific expectancies predict immunologic evidence of HIV progression. Submitted. Kemeny, M.E.. Solomon, G. and Fahey, J.L. (1999) A psychological predictor of long-term non-progression with HIV infection. In preparation. Kemeny, M.E., Solomon, G., Morley, J. and Herbert, T. (1992) Psychoneuroimmunology. In C. Nemeroff (Ed.), Neuroendocrinology, CRC Press, Boca Raton, pp. 563-594. Kemeny, M.E. and Wayment, H. (1999) How are grief and depression different? Predictors of grief and depression among gay men following an AIDS-related loss, submitted. Kemeny, M.E., Weiner, H., Duran, R., Taylor, S.E., Visscher, B. and Fahey, J.L. (1995) Immune system changes following the death of a partner in HIV seropositive gay men. Psychosom. Med., 57: 547-554. Kemeny, M., Weiner, H., Duran, R., Visscher, B. and Fahey, J. (1999) Grief and depressed mood as differential predictors of immune change in bereaved HIV positive men. In preparation. Kiecolt-Glaser, J.K., Fisher, L.D., Ogrocki, P., Stout, J.C., Speicher, C.E. and Glaser, R. (1987b) Marital quality, marital disruption, and immune function. Psychosom. Med., 49: 13-34. Kiecolt-Glaser, J.K., Glaser, R., Shuttleworth, E.C., Dyer, C.S., Ogrocki, P. and Speicher, C.E. (1987a) Chronic stress and immunity in family caregivers of Alzheimer's disease victims. Psychosom. Med., 49: 523-535. Kiecolt-Glaser, J.K., Glaser, R., Strain, E.C., Stout, J.C., Tan, K.L., Holliday, J.E. and Speicher, C.E. (1986) Modulation of cellular immunity in medical students. J. Behav. Med., 9: 5-21. Kiecolt-Glaser, J.K., Marucha, P.T., Malarkey, W.B., Mercado, A.M. and Glaser, R. (1995) Slowing of wound healing by psychological stress. Lancet, 346: 1194-1 196. Kirschbaum, C., Prussner, J.C., Stone, A.A., et. al. (1995) Persistent high cortisol responses to repeated psychological stress in a subpopulation of healthy men. Psychosom. Med., 57: 468-474.

Knapp, P.H., Levy, E.M., Giorgi, R.G., Black, P.H., Fox, B.H. and Heeren, T.C. (1992) Short-term immunological effects of induced emotion. Psychosom. Med., 54: 133-148. Koolhass, J.M., Meerlo, P., De Boer, S.F., Strubbe, J.H. and Bohus, B. (1997) The temporal dynamics of the stress response. Neurosci. Biobehav. Rev., 21: 775-782. Lane, R.D., Fink, G.R., Chau, P.M.L. and Dolan, R.J. (1997) Neural activation during selective attention to subjective emotional responses. NeuroReport, 8: 3969-3972. Lane, R.D., Reiman, E.M., Ahem, G.L., Schwartz, G.E. and Davidson, R.J. (1997) Neuroanatomical correlates of happiness, sadness, and disgust. Am. J. Psychiatry,, 154: 926-933. Lane, R.D., Reiman, E.M., Axelrod, B., Yun, L., Holmes, A. and Schwartz, G.E. (1998) Neural correlates of levels of emotional awareness: Evidence of an interaction between emotion and attention in the anterior cingulate cortex. J. Cog. Neurosci., 10: 525-535. Lane, R.D., Reiman, E.M., Bradley, M.M., Lang, P.J., Ahern, G.L., Davidson, R.J. and Schwartz, G.E. (1997) Neuroanatomical correlates of pleasant and unpleasant emotion. Neuropsychologia, 35: 1437-1444. Lane, R.D., Quinlan, D.M., Schwartz, G.E., Walker, P.A. and Zeitlin, S.B. (1990) The levels of emotional awareness scale: A cognitive-developmental measure of emotion. J. Personality Assess., 55: 124-134. Laudenslager, M.L., Fleshner, M., Hofstadter, P., Held. P.E., Simons, L. and Maier, S.F. (1988) Suppression of specific antibody production by inescapable shock stability under varying conditions. Brain, Behav. Immun., 2: 92-101. Laudenslager, M.L., Ryan, S.M., Drugan, R.C., Hyson, R.L. and Maier, S.F. (1983) Coping and immunosuppression: Inescapable but not escapable shock suppresses lymphocyte proliferation. Science, 221: 568-570. Lazarus, R.S. and Folkman, S . (1984) Stress, Appraisal, and Coping. Springer, New York. Lewis, J.W., Shavit, Y., Terman, G.W., Gale, R.P. and Liebeskind, J.C. (1983-1984) Stress and morphine affect survival of rats challenged with a mammary ascites tumor (MAT13762B). Nut. Immun. Cell Growth Reg., 3: 43-50. Maes, M., Bosmans, E., DeJongh, R., Kenis, G., Vandoolaeghe, E. and Neels, H. (1997) Increased serum IL-6 and LL-1 receptor antagonist concentrations in major depression and treatment resistant depression. Cytokine, 9: 853-858. Maes, M., Bosmans, E., Meltzer, H., Scharpe, S. and Suy, E. (1993) Interleukin-1B: A putative mediator of HPA axis hyperactivity in major depression? Am. J. Psychiatry, 150: 1189-1 193. Maes, M., Bosmans, E., Suy, E., Minner, B. and Raus, J. (1991) A further exploration of the relationship between immune parameters and the HPA-axis activity in depressed patients. Psychol. Med., 21: 313-320. Maes, M., Lambrechts, J., Bosmans, E., Jacobs, J., Suy, E., Vandervorst, C., DeJonckheere, C., Minner, B. and Raus, J. (1992b) Evidence for a systematic immune activation during depression: results of leukocyte enumeration by flow cytometry in conjunction with monoclonal antibody staining. Psychol. Med., 22: 45-53.

307 Maes, M., Meltzer, H.Y., Bosmans, E., Bergmans, R., Vandoolaeghe, E., Ranjan, R. and Desnyder, R. (1995b) Increased plasma concentrations of interleukin-6, soluble interleukin6, soluble interleukin-2 and transfemn receptor in major depression. J.Affect. Dis., 34: 301-309. Maes, M., Scharpe, S., Meltzer, H.Y., Okayli, G., Bosmans, E., D'Hondt, P., Bossche, B.V. and Cosyns, P. (1994) Increased neopterin and interferon-gamma secretion and lower availability of L-tryptophan in major depression: Further evidence for an immune response. psychiatry Res., 54: 143-160. Maes, M., Smith, R. and Scharpe, S. (1995a) The monocyte-tlymphocyte hypothesis of major depression. Psychoneuroendocrinology, 20: 111-1 16. Maes, M., Stevens, W.J., DeClerck, L.S., Bridts, C.H., Peeters, D., Schotte, C. and Cosyns, P. (1992~)A significantly increased number and percentage of B cells in depressed subjects: Results of flow cytometric measurements. J. Afect. Dis., 24: 127-134. Maes, M., Stevens, W., Peeters, D., DeClerck, L., Scharpe, S., Bridts, C., Schotte, C. and Cosyns, P. (1992a) A study on the blunted natural killer cell activity in severely depressed patients. Life Sciences, 50: 505-513. Maes, M., Vandoolaeghe, E., Ranjan, R., Bosmans, E., Bergmans, R. and Desnyder, R. (199%) Increased serum interleukin-1-receptor-antagonist concentrations in major depression. J. Affect. Dis., 36: 29-36. Maes, M., Vandoolaeghe, E., Ranjan, R., Bosmans, E., Gastel, A.V., Bergmans, R. and Desnyder, R. (1996) Increased serum soluble CD8 or suppressor/cytotoxic antigen concentrations in depression: Suppressive effects of glucocorticoids. Biol. Psychiatry, 4 0 1273-1281. Maier, S.F. and Laudenslager, M.L. (1988) Inescapable shock, shock controllability, and mitogen stimulated lymphocyte proliferation. Brain, Behav. Immun., 2: 87-91. Maier, S.F. and Watkins, L.R. (1998) Cytokines for pscyhologists: Implications of bidirectional immune-to-brain communication for understanding behavior, mood, and cognition. Psychol. Rev., 105: 83-107. Manuck, S.B., Cohen, S., Rabin, B.S., Muldoon, M.F. and Bachen, E.A. (1991) Individual differences in cellular immune response to stress. Psychol. Sci., 2: 111-115. McKinnon, W., Weisse, C.S., Reynolds, C.P., Bowles, C.A. and Baum, A. (1989) Chronic stress, leukocyte subpopulations and humoral response to latent viruses. Health Psychol., 8: 389402. Miczek, K.A., Thompson, M.L. and Shuster, L (1982) Opioidlike analgesia in defeated mice. Science, 215: 1520-1522. Murray, D.R., Irwin, M., Rearden, C.A., Ziegler, M., Motulsky, H. and Maisel, A.S. (1992) Sympathetic and immune interactions during dynamic exercise. Circulation, 86: 203-21 3. Nakata, A,, Araki, S., Tanigawa, T., Sakurai, S. and Yokoyama, M. (1996) Effects of uncontrollable and controllable electric shocks on t lymphocyte subpopulations in the peripheral blood, spleen, and thymus of rats. Neuroimmunomodulation, 3: 336-341.

Naliboff, B.D., Benton, D., Solomon, G.F., Morley, J.E., Fahey, J.L., Bloom, E.T., Makinodan, T. and Gilmore, S.L. (1991) Immunological changes in young and old adults during brief laboratory stress. Psychosom. Med., 53: 121-132. Ottaway, C.A. and Husband, A.J (1992) Central nervous system influences on lymphocyte migration. Brain, Behav. Immun., 6: 97. Papez, J.W. (1937) A proposed mechanism of emotion. Arch. Neuml. Psychiatry, 38: 725-134. Pennebaker, J.W. and Beall, S.K. (1986) Confronting a traumatic event: Toward an understanding of inhibition and disease. J. Abnorm. Psychol., 95: 274281. Pennebaker, J.W. and Chew, C.H. (1985) Deception, electrodermal activity, and the inhibition of behavior. J. Personal. SOC.Pvchol., 49: 1427-1433. Pennebaker,J.W., Hughes, C.F. and O'Heeron, R.C. (1987) The psychophysiology of confession: linking inhibitory and psychosomatic processes. J. Personal. SOC. Pscyhol., 58: 528-537. Pennebaker, J.W., Kiecolt-Glaser, J.K. and Glaser, R. (1988) Disclosure of traumas and immune function: health implication for psychotherapy. J. Consult. Clin. Psychol., 56: 239-245. Peterson, C., Seligman, M.E.P. and Vaillant, G.E. (1988) Pessimistic explanatory-style is a risk factor for physical illness: a thirty-five-yearlongitudinal study. J. Personal. SOC. Psychol., 55: 23-27. Petrie, K.L., Booth, R.J., Pennebaker, J.W., Davison, K.P. and Thomas, M.G. (1995) Disclosure of trauma and immune response to a hepatitis B vaccination program. J. Consult. Clin. Psychol., 63: 787-192. Reed, G., Kemeny, M.E., Taylor, S.E. and Visscher, B. (1999) Negative HIV-specific expectancies and AIDS-related bereavement as predictors of symptom onset in asymptomatic HIV seropositive gay men. Health Psychol.., 18: 354-363. Reed, G.M., Kemeny, M.E., Taylor, S.E., Wang, H.Y. and Visscher, B.R. (1994) Realistic acceptance as a predictor of decreased survival time in gay men with AIDS. Health Psychol., 13: 299-307. Robinson-Whelen, S., Kim, C., MacCallum, R.C. and KiecoltGlaser, J.K. (1997) Distinguishing optimism from pessimism in older adults: is it more important to be optimistic or not to be pessimistic? J. Personal. SOC.Psychol., 73: 1345-1353. Schaeffer, M.A., McKinnon, W., Baum, A., Reynolds, C.P., Rikli, P., Davidson, L.M. and Fleming, I. (1985) Immune status as a function of chronic stress at Three Mile Island. Psychosom. Med., 47: 85. Schedlowski,M., Jacobs, R., Alker, J., Prohl, F., Stratmann, G., Richter, S., Hadicke, A., Wagner, T.O.F., Schmidt, R.E. and Tewes, U. ( 1993) Psycholophysiological, neuroendocrine and cellular immune reactions under psychological stress. Neumpsychobiology, 28: 87-90. Scheier, M.F. and Carver, C. (1992) Effects of optimism on psychological and physical well-being: theoretical overview and empirical update. Cogn. Thel: Res., 16: 201-228.

308 Scheier, M.F., Matthews, K.A., Owens, J.F., Magovern Sr., G.J., Lefebvre, R.C., Abbott, R.A. and Carver, C.S. (1989) Dispositional optimism and recovery from coronary artery bypass surgergy: the beneficial effects on physical and psychological well-being. J. Personal. SOC. Psychol., 57: 1024-1040. Schulz, R., Bookwala, J., Knapp, J.E., Scheier, M. and Williamson, G.M. (1996) Pessimism, age, and cancer mortality. Psychol. Aging, 11: 304-309. Segerstrom, S., Kemeny, M.E. and Laudenslager, M. (1999) Individual differences in immunologic reactivity. In: R. Ader, D.L. Felten and N. Cohen (Eds), Psychoneumimmunology, 3rd Edn., Academic Press, New York. Segerstrom, S.C., Taylor, S.E.,Kemeny, M.E. and Fahey, J.L. (1998) Optimism and coping with stress: psychological and immune effects. J. Personal. SOC.Psychol., 74: 1646-1655. Segerstrom, S.C., Taylor, S.E., Kemeny, M.E., Reed, G.M. and Visscher, B.R. (1996) Causal attributions predict rate of immune decline in HIV seropositive gay men. Health Psychol., 15: 485-493. Seidel, A., Arolt, V., Hunstiger, M., Rink, L., Behnisch, A. and Kirchner, H. (1 995) Cytokine production and serum proteins in depression. Scand. J. Immunol., 41: 534-538. Seidel, A., Arolt, V., Hunstiger, M., Rink, L., Behnisch, A. and Kirchner,H. (1996) Increased CD56 + natural killer cells and related cytokines in major depression. Clin. Immunol. Immunopathol., 78: 83-85. Sgoutas-Emch, S.A., Cacioppo, J.T., Uchino, B.N., Malarkey, W., Pearl, D., Kiecolt-Glaser, J.K. and Glaser, R. (1994) The effects of an acute psychological stressor on cardiovascular, endocrine, and cellular immune response: a prospective study of individuals high and low in heart rate reactivity. Psychophysiology, 3 1: 264-27 1. Sieber, W.J., Rodin, J., Larson, L., Ortega, S. and Cummings, N. (1992) Modulation of human natural killer cell activity by exposure to uncontrollable stress. Brain, Behnv. Immun., 6: 141-156. Sklar, L.W. and Anisman, H.(1979) Stress and coping factors influence tumor growth. Science, 205: 513-515. Stefanski, V. (1998) Social stress in loser rats: opposite immunolocical effects in submissive and subdominant males. Physiol. Behav., 63: 605-613. Stefanski, V. and Ben-Eliyahu, S. (1996) Social confrontation and tumor metastasis in rats: Defeat and B-adrenergic mechanisms. Physiol. Behav., 60: 277-283.

Stone, A.A., Cox, D.S., Valdimarsdottir, H., Jandorf, L. and Neale, J.M. (1987) Evidence that secretory IgA antibody is associated with daily mood. J. Personal. SOC. Psychol., 52: 988-993. Stone, A.A., Neale, J.M., Cox D.S., Napoli, A,, Valdimarsdottir, H., Kennedy-Moore, E. (1994) Daily events are associated with a secretory immune response to an oral antigen in men. Health Psychol., 13: 440-6. Strauman, T.J., Lemieux, A.M. and Coe, C.L. (1993) Selfdiscrepancy and natural killer cell activity: immunological consequences of negative self-evaluation. J. Personal. SOC. P ~ c h o l .64: , 1042-1052. Taylor, S.E. (1983) Adjustment to threatening events: a theory of cognitive adaption. Am. Psychol., 41: 1161-1173. Tidey, J.W. and Klaus, A.M. (1996) Social defeat stress selectively alters mesocorticolimbic dopamine release: an in vivo micordialysis study. Brain Res., 721: 140-149. Van Tits, L.J.H., Michel, M.C., Grosse-Wilde, H., Happel, M., Eigler, F.W., Soliman, A. and Brodde, O.E. (1990) Catecholamines increase lymphocyte P2-adrenergic receptors via a P2-andrenergic, spleen-dependent process. Am. J. Physiol., 258: E191-E202. Visintainer, M.A., Volpicelli, J.R. and Seligman, M.E.P. (1982) Tumor rejection in rats after inescapable or escapable shock. Science, 216: 437-739. Wayment, H. and Kemeny, M. (1999) Are grief and depression different? Psychosocial predictors of grief and depression among bereaved gay men. Submitted. Weinberger, D.A., Schwartz, G.E. and Davidson, R.J. (1979) Low-anxious, high-anxious, and repressive coping styles: psychosomatic patterns and behavioral and physiological responses to stress. J. Abnorm. Psychol., 8: 369-380. Weiner, H. (1992) Perturbing the organism. The biology of stressful experiences. The University of Chicago Press, Chicago. Weisse, C.S., Pato, C.N., McAllister. C.G., Littman, R., Breier, A., Paul, S.M. and Baum, A. (1990) Differential effects of controllable and uncontrollable acute stress on lymphocyte proliferation and leukocyte percentages in humans. Brain, Behav. Immun., 4: 339-35 1. Weitzman, O., Kemeny, M.E. and Fahey, J.L. (1999). Persistent shame and guilt predicts CD4 decline, AIDS onset and mortality, submitted. Zisook, S., Chuchter, S. R., Irwin, M., Darko, D. F., Sledge, P. and Resovsky, K. (1994) Bereavement, depression, and immune function. Psychiatry Res., 52: 1-10.

E.A. Mayer and C.B.Saper (ELIS.) Progress in Brain Research, Vol 122 0 2000 Elsevier Science BV. All rights reserved.

CHAPTER 22

Memory networks in the prefrontal cortex Joaquin M. Fuster Neuropsyrhiatric Institute, University of California, Los Angeles, CA 90024, USA

Introduction In the primate, the cortex of the frontal lobe appears devoted in its entirety to the representation and execution of actions. The frontal cortex as a whole can therefore be considered ‘motor cortex’ in the broadest sense of the word. It coordinates actions in practically all the domains of adaptation of the organism to its environment: skeletal and ocular motility, logical reasoning, communication and the spoken language. Even visceral actions and emotional behavior are regulated by certain orbital and medial areas of the frontal cortex. In this chapter I will outline the rationale for the role of the dorsolateral prefrontal cortex in the temporal organization of action, as well as some of the mechanisms that support it. I will begin with certain basic assumptions about the cortex in general and the frontal cortex in particular. The cognitive functions of the cortex of the frontal lobe, as those of any other part of the neocortex, consist in the activation and processing within and between networks of representation, or memory networks. Those networks are widely distributed and highly specific, defined by their synaptic structure and connectivity. Thus the memory code is a relational code, and all memory is associative. The cortical networks of memory extend across modules and areas by any anatomical

*Corresponding author. Tel.: 310 8250247; Fax: 310 8256792; e-mail:[email protected]

definition. Memory networks overlap and are profusely interconnected with one another. Thus, one neuron or group of neurons anywhere in the cortex can be part of many networks and thus many memories, This is why it is virtually impossible, by any method, to localize a memory. The networks of executive or motor memory are distributed in the cortex of the frontal lobe, and like the perceptual networks of posterior - post-rolandic - cortex, are hierarchically organized. The base of the executive hierarchy consists of the motoneurons and anterior roots of the spinal cord. Above that, in ascending order, are the motor nuclei of the mesencephalon, the cerebellum, and portions of the diencephalon, including certain nuclei of the hypothalamus, the thalamus, and the basal ganglia. Above the basal ganglia is the frontal cortex, which itself is hierarchically organized. At the base of the cortical motor hierarchy is the primary motor cortex, for the representation and execution of elementary skeletal movements. Above it is the premotor cortex, serving more complex movements defined by goal and trajectory, including certain premotor areas involved in speech. At the summit is the prefrontal cortex. We can safely infer that this cortex represents - we do not yet know how - the broad schemas or plans of action in the skeletal and speech domains, and in addition is critically involved in the enactment of those schemas plans. Indeed, one of the most consistent and specific components of the frontal-lobe syndrome is the inability to formulate and to enact programs of behavioral, linguistic or cognitive action, such as logical reasoning.

310

Temporal structuring of action From those general assumptions, especially from the distributed nature of cortical networks, it follows that we cannot rightfully consider the cognitive functions of the prefrontal cortex in isolation from those of the rest of the frontal cortex or, for that matter, from the totality of the neocortex and the subjacent anatomical stages of the executive hierarchy. Pursuing methodological neatness, we have often been misled to the localization of cognitive functions that are not localizable. In my opinion, this is true for the so-called working memory, for the so-called ‘central executive’, for spatial memory and for various forms or aspects of attention. All these are indeed cognitive functions within the physiological purview of the frontal lobe, but none of them is localized there. What appears localized there, to some degree, is the distributed representational substrate, the content of those functions, in other words, the networks of executive memory. It is by transactions within those networks, and between those networks and others elsewhere in the neocortex, that the prefrontal cortex most probably exerts its role in the temporal organization of behavior, speech, and logical reasoning. In this chapter I will attempt to outline some of the physiological transactions in cortical memory networks that appear to support the role of the prefrontal cortex in the enactment of temporal structures of action, and thus in establishing temporal order in behavior, in reasoning, and in speech. Basically, at the root of that order is a set of cognitive operations - I should like to argue physiological operations - that implement a fundamental principle: the mediation of cross-temporal contingencies between events, between words, between stimuli, between particular stimuli and particular acts. That principle can be expressed by two simple and complementary logical statements with a temporal dimension: “If now this, then later that action; if earlier that, then now this action”. The first proposition is temporally prospective, the second temporalIy retrospective. The dorsolateral prefrontal cortex plays a critical role in the cortical dynamics that implement the mediation of crosstemporal contingencies. This has now been

substantiated in the human and non-human primate by several behavioral and functional methods. At the behavioral level there is apparently no better way to study temporal structuring and the mediation of cross-temporal contingencies than the use of delay-task paradigms (e.g. delayed response, delayed matching to sample). The single trial in a delay task is the epitome of the temporal structure or gestalt of behavior. Because it separates by time two events that are mutually contingent, and because these events are for all practical purposes novel inasmuch as they change at random from one trial to the next, the delay task is also the epitome of the cross-temporal contingency. Thus the delay task is one of the most practical methods to investigate the neuropsychology and the neurophysiology of executive networks. Here I will briefly summarize some of our past and current work with delay tasks in support of the following: (a) the critical importance of dorsolateral prefrontal cortex for the mediation of cross-temporal contingencies; (b) the role of prefrontal networks in short-term memory, also called working memory, which constitutes the retrospective aspect of cross-temporal contingencies; (c) the role of prefrontal networks in short-term attentive set, the prospective aspect of crosstemporal contingencies; and (d) the cortical mechanisms of short-term active memory, and the importance of reentry through recurrent circuits as one of those mechanisms. The cooling of dorsolateral prefrontal cortex large portions of areas 9 and 46 of Brodmann - in the monkey have been found to induce reversible deficits in the performance of visual, auditory and tactile delay tasks (Fuster, 1997a). These are tasks in which motor acts are contingent on prior sensory stimuli that the animal must retain for a few seconds or minutes. The deficit with visual stimuli can be observed whether those stimuli are, for behavioral purposes, spatially defined or not. Ostensibly, the animal with inactivated or depressed dorsolateral prefrontal cortex cannot suitably mediate the cross-temporal contingency between a stimulus of any of those three modalities and the consequent response to it, especially if the time between the two is relatively long ( > 5 sec). If the task is a visual non-spatial delay task (delayed

31 1

matching to sample), a comparable deficit can be obtained by cooling inferotemporal cortex. The latter procedure does not affect a visuospatial delay task. However, I should emphasize, the evidence of multimodal deficit from cooling a large prefrontal region does not exclude the regional stimulusspecificity within it. In fact, there is abundant evidence from anatomical and functional studies for regional specificity within dorsolateral prefrontal cortex, with regard to stimuli as well as contingency tasks (Fuster et al., 1982; Petrides et al, 1993; Petrides and Pandya, 1994; GoldmanRakic, 1995). Here I mention the multimodal prefrontal deficits of the monkey in delay tasks simply to highlight the importance of the dorsolateral prefrontal cortex as a whole for the mediation of cross-temporal contingencies. To further highlight the cross-temporal role of the prefrontal cortex, I can refer to the numerous functional imaging studies of the human performing delay tasks (Paulesu et al., 1993; Petrides et al., 1993; Courtney et al., 1996; Owen et al., 1996; Smith et al., 1996; Aguirre et al., 1998; Gabrieli et al., 1998). They all demonstrate the activation of various prefrontal areas during one form or another of cross-temporal integration. We contributed to this literature with a PET-fluordeoxyglucose study of human subjects performing a visual delayed matching-to-sample task (Swartz et al., 1995). We were able to verify the activation of not only dorsolateral prefrontal cortex but, postrolandic cortex - including visual cortex - during the mnemonic retention of abstract visual images. This points to the functional cooperation of frontal and posterior cortical areas in that temporal process, an issue to which I shall return later. The cortical mechanisms behind temporal integration, however, can best be studied in real time at the cellular level. At that level, neuronal phenomena reveal that the translation of perception into action, across time, depends on the cooperation of at least two temporally complementary cognitive functions: (a) active short-term or working memory for sensory stimuli, and (b) short-term attentive set, also conceptually understandable as the activation of prospective motor memory. Further, what neuronal physiology shows is that the integrative work of those two functions depends on the close

cooperation between widely separated cortical areas, and that in that coordination the prefrontal cortex plays a fundamental role (Fuster 1997b). Let us briefly examine those two temporally converse but complementary functions on which the structuring of action appears to depend.

Short-term memory There is now conclusive evidence, from several methodologies, that the frontal cortex as a whole, especially the dorsolateral prefrontal cortex, is critically involved in all forms of active (‘working’) memory toward a goal, in other words, toward the completion of a gestalt of action, whether that is in the domain of behavior, reasoning or speech (Fuster, 1997a,). Some of that evidence has been cited above in the context of temporal organization. What defines frontal memory in the active state is precisely the teleological quality of a memory that has been mobilized in the construction of future action. From that teleological quality of frontal memory derive all the apparent physiological attributes of individual areas and cell groups in the dorsolateral frontal cortex. The sensory cell groupings and sensory memory subdivisions of that cortex are probably the stepping stones or pathways of access to that executive, teleological memory, and thus the paths to the action. Consequently, sensory working memory is probably a servant of the so-called central executive, rather than the other way around. Furthermore, certain areas of medial or orbital prefrontal cortex are the recipients of visceral input and of information related to reward or emotion. These areas have also prominent inputs from the amygdala and other regions of the limbic system. They may be in the pathways to emotional behavior or visceral action, or both. The participation of prefrontal neurons in shortterm memory has been amply documented by numerous descriptions of single units that respond with activation of firing frequency to the presence of a memorandum for prospective action (e.g. Fuster, 1973; Funahashi et al., 1989). Many such units have two characteristics that unmistakably define them as so-called ‘memory cells’: (a) the specific response to one or more memoranda, and (b) the temporal decay of their firing in the course

312

of the memorization period; this phenomenon is most conspicuous in manual memory tasks with long delays (10 sec or longer). These cells are probably the constituents of motor memory networks that are activated by the recall or retrieval of a memorandum for the retention of that memorandum toward the correct action. Indeed their discharge has been related not only to the memorandum but to the efficiency of performance of the task. In any event, different sensory memoranda seem to have different prefrontal distributions in accord with the provenance of fibers from different parts of posterior, perceptual cortex, that arrive to the prefrontal cortex. Recently we have been studying the distribution of auditory memory cells in monkeys that perform a cross-modal audiovisual task (Bodner et al., 1996). Using a task with variable probabilities of association between a visual stimulus and a later manual response, we were able to examine the temporal characteristics of memory cells in the dorsolateral prefrontal cortex of the monkey (Quintana and Fuster, 1992). The task combines delayed matching-to-sample with delayed conditional discrimination. A color is the initial cue at the beginning of each trial. After a delay of 12 sec, a second visual cue is given, and the animal is required then to perform a manual choice that depends on both cues, the first and the second, thus on the combination of two visual stimuli separated by time. Each combination of the two stimuli double contingency - determines whether the response after the delay will be to the left or to the right. That combination changes at random between trials. However, two of the trial-initiating cue colors, blue and yellow, are always followed by white (second cue) and require response respectively to the left (after blue) and right (after yellow). The other two, red and green, are followed by either white or a side-by-side display of green and red (relative position of the two colors changing at random between left and right). If after red or green the second cue is white, the monkey must choose left or right response, respectively; if the second cue is red-green or green-red the monkey must choose the color that matches the first cue, which may be on the left or the right. Because of the randomicity with which the stimulus combinations

are presented, and because of the above design of contingencies, blue and yellow at the start of a trial predict the response side with 100% probability, whereas red and green do it with only 75% probability. The exploration of the prefrontal cortex during performance of that double-contingency task revealed two broad categories of ‘memory cells’, that is, cells with elevated discharge during the delay period. Some cells responded with a different level of activation depending on the color of the initial cue, in other words, they seemed to prefer certain colors, and their discharge in the course of the delay tended to descend toward baseline level. They appeared to distinguish and remember the colors and, during the delay period, to ‘look back’ to the color of the first cue. Their discharge during that period seemed to reflect the temporal decline of short-term memory. The second category of cells in dorsal prefrontal cortex behaved in the opposite way. Their discharge during the delay reflected the direction of the manual response before that response was prompted by the second cue, as if anticipating that response and preparing for it. These cells seemed to ‘look forward in time’. Their discharge accelerated as the second cue and the motor response approached. Furthermore, the degree of acceleration varied in proportion to the certainty with which the animal could predict the direction of that response. Most probably the neurons of this second type were involved in the setting of the neuromotor apparatus for the response. This is why we attributed these cells to the attentive set or active motor memory, the second of the two temporally integrative prefrontal functions postulated (next section). A remarkable finding is that the cells of the two types, sensory-coupled or motor-coupled cells, appear anatomically intermixed with each other. It is not possible to discern a separate topography for them, overlapping as they do with each other in the prefrontal cortex around the sulcus principalis. It would appear from this fact alone that, during the delay, there is a direct transfer of information from sensory memory cells to motor set cells. More probable, however, is the temporal transfer from one prefrontal network to another, both with ties to

313

the visual networks of posterior origin (inferotemporal) and the motor networks downstream in the executive hierarchy.

Short-term attentive set The second class of prefrontal neurons, those that seem to anticipate the action and to prime the motor apparatus for it, may in fact be the substrate for the short-term activation of motor memory, in other words, the converse of sensory short-term memory. Motor memory, in this context, would be the socalled ‘memory of the future’ (Ingvar, 1985). Thus, in principle, there seems to be a sensory and a motor short-term memory, one retrospective and the other prospective, the two complementing each other at the service of the ‘frontal executive’ in the mediation of cross-temporal contingencies. That function of active prospective memory can also be viewed as the motor counterpart of sensory attention; we may call it motor attention or attentive set. Its cellular manifestations, of course, acquire special significance in light of the well-known neuropsychological significance of the dorsolateral prefrontal cortex in the formulation and execution of action plans. The motor-set cells would be the microcosm of the planning functions of the frontal lobe. In general, the sustained reactions of prefrontal cells in delay tasks are probably related to the wellknown field potentials that can be observed on the human frontal cortex between mutually contingent but temporally separate events, notably the ‘contingent negative variation’ or CNV. The accelerating cell reactions of ‘set cells’ are especially reminiscent of the Bereitschuftspotential, or ‘readiness potential’, which is another negative potential that takes place as a continuation of the CNS, right before a pre-instructed motor act. Those surface negative potentials, which appear over frontal cortex after a sensory stimulus and before an action contingent on it, most probably reflect the underlying activation of large neuronal populations engaged in active short-term memory and attentive set. In conclusion, the ramping-up cells and the surface negative field potentials, especially the ‘readiness potential’, seem to be the electrophysio-

logical manifestations of attentive set, or motor attention. This is the prospective function of the dorsolateral prefrontal cortex. It is attention directed to the action in preparation. This kind of attention is focused in the representation of the action and, at the same time, in the components of long-term motor memory which are activated ad hoc for the execution of every part of the sequence of behavior in progress, from its initiation to its goal. Those cells that seem to predict future actions - albeit only for the short term - indicate that there are mechanisms in the dorsolateral prefrontal cortex not only for evoking the prospective motor act but for preparing the motor apparatus for it. Perhaps those mechanisms include the priming of structures in lower stages of the motor hierarchy for the impending movement (e.g. premotor cortex, basal ganglia, pyramidal system).

Cortical dynamics of the perception-action cycle Given that the dorsolateral prefrontal cortex mediates cross-temporal contingencies between perception and movement, and given that this mediation is apparently the result of the coordination of two cognitive functions of that cortex, active memory and attentive set, now the question is, what are the neural mechanisms of that functional coordination. How does the prefrontal cortex mediate the transfer of information from the past to the future, from the perception to the action? Clearly, these mechanisms must involve both, local processes, as suggested by our two types of neighboring cells (memory and set cells), and remote or transcortical processes as some of our other cellular evidence and some of the imaging literature suggests. We postulate that in both these kinds of processes, local and remote, reentry or reverberation through recurrent circuits plays an important role. In our effort to test the reentry hypothesis, we used several analytical methods. The first was to develop an artificial network with an architecture that was essentially recurrent (Zipser et al., 1993). We trained that model network to ‘sample’ an external input, to retain it in short-term memory, and to produce an output that was a specified function of that input. One of our purposes was to

3 14

find out if the ‘cells’ of such a recurrent model would behave like real cortical cells in active shortterm memory. To train our model we used a variation of the backpropagation method (Rumelhart et al., 1986). That is an error-reducing procedure that allows the network, through many iterations, to adjust synaptic weights to keep a stable relationship between input and output despite variations of the input. After training, those weights stay fixed. When the network has been fully trained, a memory trial can be simulated by loading an input, i.e. the memorandum, and by holding a gate open through the memory period or delay until the recall, when another load signal closes the gate and emits the output. Some of the units in the network, under these conditions, behave like real cortical cells in a delay task. The output cells behave unremarkably, since they simply reflect the input-to-output function that is defined by the modeler. What is remarkable is the behavior of the internal or ‘hidden’ units of the model, which with adequate scaling behave like real cells in the memory task. We see cells ramping up, cells ramping down, and cells that appear to be a mixture of the two. I should note that those patterns of network unit discharge, which so much resemble the discharge patterns of real cells, are part of a repertoire of many patterns obtained by multiple repetitions of the sample-andhold function of the model, and are a product of the internal architecture of the model. We can conclude from these findings that the firing patterns of cortical cells are understandable as a result of the sustained activation of fully trained recurrent networks with preestablished synaptic weights. Backpropagation, which is the training mechanism that was used by us to train the network and to establish its weights, is irrelevant to that conclusion. The most important point is that, after the weights have been established, the shortterm activation of the artificial, recurrent memory network, elicits in some of its units patterns of firing that are practically identical to those of real cortical cells in short-term memory. Thus, the role of reentry in the cortical dynamics of short-term memory receives support from the behavior of the units in a network model in which recurrence is an essential feature of the functional architecture.

Here it is appropriate to refer to some experiments in which the reentry between cortical areas in short-term memory was tested more directly (Fuster et al., 1985). We first trained monkeys to perform a visual delayed matching task with colors. Then we implanted cooling probes and microelectrode carriers, bilaterally, on two cortical areas that we know from both cooling and single-unit studies are involved in that task, in other words, areas that contain components of the cortical networks activated in visual short-term memory: the dorsolateral prefrontal cortex and the inferotemporal cortex. Then, while the animal performed the task, we proceeded to cool one cortex, prefrontal or temporal, while at the same time we recorded units from the other. Cooling either cortex to 20°C reversibly impairs the monkey’s performance of the task, while modifying in various ways the spontaneous and memory - delay period - discharge of the cells in the other cortex. In a small but distinct contingent of prefrontal and inferotemporal cells, the discharge during memory is subtly but characteristically modified. These are cells that are differentially activated during the memorization of the sample color: they seem to prefer one color in short-term memory over the others. Under cooling of the distant cortex, inferotemporal or prefrontal, these cells show lesser differences in their memory discharge; they differentiate colors less than while that cortex is at normal temperature. In no case have we observed the opposite effect, that is, increased color differentiation by cooling. It would appear therefore that, in the absence of input from the remote cortex under cooling, some cells in both prefrontal and inferotemporal cortex become less active in the memorization of their preferred color. Concomitantly, the monkey’s capacity to retain colors diminishes. These data would support the notion that short-term visual memory is maintained by tonic reentrant, i.e. reverberating, excitability between inferotemporal and prefrontal cortices. That presumptive mechanism of recurrent excitation between the prefrontal - executive - cortex and the sensory association cortex can be conceptualized as a mechanism of temporal closure at the summit of the perception-action cycle (Fig. 1). This cycle is a basic principle of biological cybernetics.

315

It is the circular flow of neural information that links an organism to its environment (Fig. 1). The neuroanatomy of the cycle essentially consists of two parallel hierarchies of neural structures, one sensory and the other motor, that extend through the entire length of the nerve axis, from the spinal cord to the highest cortex of association and the prefrontal cortex. All structures are interlinked at all levels by reciprocal connections; feedforward and feedback operate between stages and between sensory and motor structures at all levels. During the performance of new or recently acquired behavior, sensory information is processed along the sensory hierarchy. That information is thus translated into action, which is processed down the motor hierarchy to produce changes in the environment. These changes lead to sensory

SENSORY HIERARCHY

changes, which are processed in the sensory hierarchy and then modulate further action, and so on and so forth. The posterior cortex of association and the dorsolateral prefrontal cortex are part of the cycle if the behavior contains novelty or uncertainty, and has to bridge time spans with short-term memory and attentive set, in other words, while it has to mediate cross-temporal contingencies.When those requirements disappear and the behavior becomes automatic (as in walking or performance of learned routines), the actions are integrated in lower structures of the cycle (for example, premotor cortex, basal ganglia) and the processing of sensory inputs is shunted at lower levels of the cycle. Thus, the so-called central executive, the dorsolateral prefrontal cortex, at the top of the motor

MOTOR HIERARCHY

Fig. 1. Cortical anatomy of the perception-action cycle, human brain in the inset. Unlabeled regions represent subareas of labeled regions or intermediate areas between those that are labeled; arrows represent aggregates of connections demonstratedin the monkey. All connections between areas are bidirectional, permitting feedforward as well as feedback.

316

hierarchy and the perception-action cycle, integrates actions with perceptions especially in the presence of novelty and complexity. It does so in close cooperation with remote areas of the neocortex and with structures lower in the executive hierarchy. There is topographic specificity within the prefrontal cortex with regard to the nature or modality of sensory input as well as the nature of the action that the input calls for at any point in time and in any given context. This topographic specificity, however, should not obscure the overarching role of that cortex as a whole in bridging temporal gaps and organizing new actions in all domains of behavior, reasoning and language.

References Aguirre, G.K., Zarahn, E. and D’Esposito, M. (1998) Neural components of topographical representation. Proc. Natl. Acad. Sci. USA, 95: 839-846. Bodner, M., Kroger, J. and Fuster, J.M. (1996) Auditory memory cells in dorsolateral prefrontal cortex. NeuroReport, 7: 1905-1908. Courtney, S.M., Ungerleider, L.G., Keg, K. and Haxby, J.V. (1996) Object and spatial visual working memory activate separate neural systems in human cortex. Cerebr: Cort., 6: 39-49. Funahashi, S., Bruce, C.J. and Goldman-Rakic, P.S. (1989) Mnemonic coding of visual space in the monkey’s dorsolateral prefrontal cortex. J. Neurophysiol., 61: 331-349. Fuster, J.M. (1973) Unit activity in prefrontal cortex during delayed-response performance: neuronal correlates of transient memory. J. Neurophysiol., 36: 61-78. Fuster, J.M. (1997a) The Prefrontal Cortex Anatomy: Physiology and Neuropsychology of the Frontal Lobe, Lippincott-Raven,Philadelphia, PA. Fuster, J.M. (1997b) Network memory. Trends Neurosci., 20: 451459. Fuster, J.M., Bauer, R.H. and Jervey, J.P. (1982) Cellular discharge in the dorsolateral prefrontal cortex of the monkey in cognitive tasks. Exp. Neurol., 77: 679-694.

Fuster, J.M., Bauer, R.H. and Jervey, J.P. (1985) Functional interactions between inferotemporal and prefrontal cortex in a cognitive task. Brain Rex, 330: 299-307. Gabrieli, J.D.E., Poldrack, R.A. and Desmond, J.E. (1998) The role of left prefrontal cortex in language and memory. Proc. Natl. Acad. Sci. USA, 95: 906-913. Goldman-Rakic, P.S. (1995) Architecture of the prefrontal cortex and the central executive. Proc. Natl. Acad. Sci., 769: 71-83. Ingvar, D.H. (1985) ‘Memory of the future’. An essay on the temporal organization of conscious awareness. Hum. Neurobiol., 4: 127-136. Owen, A.M., Moms, G., Sahakian, B.J., Polkey, C.E. and Robbins, T.W. (1996) Double dissociations of memory and executive functions in working memory tasks following frontal lobe excisions, temporal lobe excisions or amygdalohippocampectomy in man. Brain, 119: 1597-1615. Paulesu, E., Frith, C.D. and Frackowiak, R.S.J. (1993) The neural correlates of the verbal component of working memory. Nature, 362: 342-344. Petrides, M., Alivisatos, B., Evans, A.C. and Meyer, E. (1993) Dissociation of human mid-dorsolateral from posterior dorsolateral frontal cortex in memory processing. Proc. Natl. Acad. Sci. USA, 90: 873-877. Petrides, M. and Pandya, D.N. (1994) Comparative architectonic analysis of the human and the macaque frontal cortex. In: F. Boller and J. Grafman (Eds), Handbook of Neuropsychology, Elsevier, Amsterdam, pp. 17-58. Quintana, J. and Fuster, J.M. (1992) Mnemonic and predictive functions of cortical neurons in a memory task. NeuroReport, 3: 721-724. Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986) Learning representations by back-propagating errors. Nature, 323: 533-536. Smith, E.E., Jonides, J. and Koeppe, R.A. (1996) Dissociating verbal and spatial working memory using PET. Cerebr: Cortex, 6: 11-20. Swartz, B.E., Halgren, E., Fuster, J.M., Simpkins, F., Gee, M. and Mandelkern, M. (1995) Cortical metabolic activation in humans during a visual memory task. Cerebr: Cort., 3: 205-214. Zipser, D., Kehoe, B., Littlewort, G . and Fuster, J.A. (1993) Spiking network model of short-term active memory. J. Neurosci., 13: 3406-3420.

E.A. Mayer and C.B. Saper (Eds.) Progress in Brain Research,Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAPTER 23

Non-conscious brain processing indexed by psychophysiological measures Daniel Tranel” Psychophysiology Laboratory9 Division of Cognitive Neuroscience, Department of Neurology, University of Iowa College of Medicine, Iowa City, IA 52242

Introduction Few would argue with the notion that much of our neural processing, and the cognitive operations it subserves, takes place beneath the level of conscious awareness. In defining consciousness, in fact, one is inevitably faced with the complementary challenge of defining unconsciousness; the two go hand in hand, and together comprise what we know as mental life (e.g. Churchland, 1986; Dennett, 1991; Churchland and Sejnowski, 1992; Kosslyn and Koenig, 1992; Milner and Rugg, 1992; Crick, 1994; Damasio, 1999). One of the daunting challenges for science, and for the science of cognitive psychology in particular, has been to design and implement rigorous experimental paradigms that would permit a peek into the contents, operations, and products of non-conscious processing.I *Corresponding author. Tel.: 319-356-2671; Fax: 319-356-4505 ‘In the context of this chapter, what I mean by ‘nonconscious’ is the type of information processing (cognitive and neural) that takes place outside of awareness; ‘conscious’, on the other hand, covers everything of which one is aware. This distinction does not correspond to a ‘verbal-non-verbal’ dichotomy, and in fact, there is plenty of non-verbal processing (e.g. by right hemisphere) that is very much within the scope of ‘conscious’, even if one cannot put into words the contents or results of that processing. In experimental paradigms, the verbal self-report of the patient may be used as an index of conscious processing, but this is a matter of convenience, and should not be taken to imply that ‘conscious’ processing is necessarily dependent upon or is in any way tantamount to verbal report.

Scientific exploration of non-conscious processing has been actively pursued in recent years, and has yielded several important breakthroughs. One approach that has been helpful in this area of inquiry is known as the ‘lesion method’. In this method, the experimental preparation is the human adult brain, and in particular, a brain that has been damaged in a focal, discrete fashion, causing a circumscribeddeficit in some aspect of cognition or behavior, e.g. an inability to recognize familiar faces, or to learn new facts, or to retrieve emotional knowledge. My colleagues and I have used this method to explore a variety of questions regarding brain-behavior relationships. A key technique in our work has been the utilization of a psychophysiological probe - namely, the electrodermal skin conductance response - as a dependent variable to study cognition and behavior. The skin conductance response (SCR), which provides an index of autonomic nervous system activation, is exquisitely sensitive to a host of stimulus and processing features, such as importance, novelty, familiarity, relevance - features than can be grouped under the broad concept of ‘signal value’. We have applied this technique in experiments with brain-damaged subjects, and some of the highlights of this work are summarized here. A final introductory comment about the nature of our subject population is in order. The effects of brain injury on various cognitive and behavioral functions are often profound; for example, in the condition known as prosopagnosia, patients may be rendered incapable of recognizing any familiar faces, even their own (e.g. as seen in the mirror or in a photograph); in global amnesia, patients may

318

be rendered incapable of learning any new factual knowledge. Such deficits are not subtle, and they provide a fertile ground for exploring potential preserved operations that may be unfolding beneath the level of conscious awareness. In normal subjects, it is often quite difficult to rig an experimental paradigm in such a way that one can be certain that conscious processing did not contribute in any way to the behavioral performance. Designing such paradigms is obviously possible (Kihlstrom, 1987; Reber, 1989; Roediger, 1990; Lewicki et al., 1992; Kolb and Braun, 1995; Luck et al., 1996; Berns et al., 1997; Leuthold and Kopp, 1998; Ohman and Soares, 1998), but the point here is that the status of brain-damaged individuals often provides a highly compelling paradigm for exploring non-conscious processing. No less importantly, the investigation of such patients, when coupled with careful modem-era neuroanatomical analysis such as that which has become available in recent years from fine-grained neuroimaging techniques (Damasio and Frank, 1992; Damasio and Damasio, 1997; Frank et al., 1997), can permit inferences about putative neural substrates for non-conscious processes.

I. Non-consciousdiscriminationof familiar stimuli in agnosia Several decades ago, it was demonstrated that normal subjects can produce evidence of detection and recognition of stimuli which had been degraded or camouflaged in such a manner so as to preclude conscious awareness (Lazarus and McCleary, 1951; Adams, 1957; Reiser and Block, 1965; Rousey and Holzman, 1967; Corteen and Wood, 1972). In some of the most convincing of these studies, psychophysiological indices, such as the SCR, provided the most sensitive index of nonconscious ‘recognition’. In the mid 1980s, our laboratory and others (Bauer, 1984; Tranel and Damasio, 1985; Bauer and Verfaellie, 1988; Tranel and Damasio, 1988) picked up on this line of work, and began applying such paradigms to the investigation of brain-damaged subjects. Covert face recognition

Non-conscious recognition of familiar faces in prosopagnosia

Neurological patients with the condition known as prosopagnosia, which is caused by bilateral occipitotemporal lesions, lose the ability to recognize familiar faces. The patients cannot recognize faces of family members, close friends, and even their own; when they look at those faces, they have no inkling that the faces are familiar, and no sense that they ought to know the persons. When asked directly, and even when prompted with careful questioning, the patients insist that familiar faces look like complete strangers. This occurs despite the fact that the patients have normal visual perceptual abilities, that is, the problem is not due to faulty perception, or some basic inability to see normally. We used a psychophysiological index (SCR) to explore whether prosopagnosic subjects, despite their profound inability to recognize familiar faces at conscious level, might nonetheless produce some evidence that they can discriminate well-known faces from faces of strangers (Tranel and Damasio, 1985, 1988). Using a standard method described elsewhere (Tranel and Damasio, 1989, 1994), skin conductance was recorded while subjects sat and viewed a series of face stimuli. The stimulus sets included some faces that ought to have been well known to the subjects ( e g family members, themselves, famous persons; we refer to these as targets), mixed in random order with faces the subjects had never seen before (non-targets).Each face was presented for 2 s, one at a time, at intervals of 20-25 s. Subjects were instructed to view each face carefully, but no verbal or motor response was called for. For each face, the amplitude of the largest SCR that began within 1 to 5 s after stimulus onset was recorded. We then calculated the average SCR to the target faces, and the average SCR to the non-target faces, and compared these statistically. A summary of some of the data from these experiments is presented in Table 1. In six different subjects with prosopagnosia, all of them produced significantly larger-amplitude SCRs to target faces, compared to non-targets. This occurred in three separate experiments, one in which target faces were family members and friends, one in which targets were famous individuals (movie stars, politicians), and one in which the targets were faces

3 19

TABLE

1

Non-conscious face recognition in prosopagnosia. Skin conductance response magnitudes (in US)* Subject

EH-034 LDV-692 psD-868 AG-1052 DH-1358 RE1478

Retrograd+famiIy Target Non-target 0.93 1.66 0.17 0.53 1.34 0.29

Retrogradefamous Target Non-target

0.05 0.15 0.04 0.01 0.17 0.01

0.73 1.08 0.06 0.12 1.54 0.18

0.01 0.02 0.02 0.00 0.04 0.03

Anterograde Target Non-target -

-

0.35 0.02 0.29 1.18 0.19

0.02 0.01 0.00 0.04 0.03

* We record skin conductance in such a manner that an increase in response magnitude reflects, in general, an increase in autonomic nervous system activation. of persons with whom the subject had had considerable exposure after the onset of their condition, but not before. In sum, six prosopagnosic subjects produced clear evidence of non-conscious discrimination of facial stimuli they could not otherwise recognize, and for which even a remote sense of familiarity was lacking. These findings suggest that some part of the physiological process of face recognition remains intact in the subjects, although the results of this process are unavailable to consciousness. The fact that the subjects were able to produce SCR discrimination of faces to which they had been exposed only after the onset of their condition is particularly intriguing, as it suggests that the neural operations responsible for the formation and maintenance of new ‘face records’ can proceed independently from conscious influence (Greve and Bauer, 1990). Recently, we had an opportunity to explore nonconscious face recognition in a subject with a developmental form of prosopagnosia (Jones and Tranel, 1998). The subject was a 5-year-old boy who had never learned to recognize faces normally, despite superior intellectual abilities and no detectable neuroanatomical abnormalities (hence the label ‘developmental’). In a psychophysiological experiment designed according to the protocol described above, we found that the boy was able to generate discriminatory SCRs to target faces such as family members, himself, and various professionals he had worked with extensively over the past few years. In keeping with his prosopagnosia, none of these faces were recognized at conscious (overt) level. To our knowledge, this is the first

demonstration of non-conscious visual recognition in the setting of a developmental condition, and it extends to an early developmental phase the findings we have reported in adult subjects. Other paradigms have also yielded evidence of non-conscious or ‘covert’ face recognition in prosopagnosic subjects. Bauer (Bauer, 1984; Bauer and Verfaellie, 1988), for example, presented prosopagnosics with either correct or incorrect face-name pairs, and found that subjects produced larger-amplitude SCRs to the correct pairs (an effect that also obtains in normal individuals). Rizzo et al. (1987) showed that prosopagnosic subjects produced different scanpath patterns for familiar faces, compared to unfamiliar ones. de Haan (de Haan et al., 1987a, 1987b), using a reaction time paradigm in which a prosopagnosic subject had to decide whether two photographs ‘matched’ (were of the same individual) or did not match (were of different individuals), found that reaction time was systematically faster for familiar faces, compared to unfamiliar ones. Other examples of dissociations between visual perception and consciousness as a consequence of brain damage have been reviewed recently (Viggiano, 1996; Farah and Feinberg, 1997). Figure 1 shows in schematic fashion the neural basis we have hypothesized to account for the face recognition profile of prosopagnosic subjects with bilateral occipitotemporallesions. A familiar face is exposed to the subject, and the subject can perceive that face normally, with intact early visual association cortices. However, the lesions preclude normal activation of neural processing stations in higher-

320

Fig. 1. Diagram of possible route of neural activation subserving nonconscious face discrimination in subjects with occipitotemporaldamage. The damage (cross-hatch) precludes activation (by faces) of ventral occipital and temporal association cortices. Dorsal cortices can operate, however, allowing signals to reach ventromedial prefrontal cortices and, in turn, activate autonomic control nuclei such as the amygdala.

order visual association cortices along the ventral occipital and temporal regions. The deficient activation of these stations in turn prevents the retrieval of information pertinent to the face (and to the person to whom that face belongs), such as the name, occupation, and personality characteristics. In other words, knowledge on which conscious recognition depends fails to be triggered. However, neural processing stations located more dorsally and laterally in the occipitotemporal and occipitoparietal regions are intact. Thus, critical processing signals triggered by perception of structural face features, and perhaps by scanpaths, can proceed successfully along a dorsal route, eventually reaching frontal cortices and in particular, the ventromedial prefrontal region. The latter, in turn, can signal intact autonomic control nuclei, such as the amgydala and hypothalamus, which would then trigger the discriminatory skin conductance response.

A double dissociation between conscious and nonconscious face recognition As alluded to earlier, stimuli with strong affective valence and ‘signal value’ produce large-amplitude

SCRs in normal subjects (e.g. Tranel et al., 1985; Bradley et al., 1992). In a series of studies of braindamaged subjects with bilateral damage to ventromedial prefrontal cortices, however, we found that the subjects were remarkably impaired in their ability to generate SCRs to highly potent ‘signal’ stimuli, such as nudes and mutilation scenes (Damasio et al., 1990, 1991). We have interpreted this outcome as reflecting an impairment of somatic marker activation (Bechara et al., 1999, in press), by which we mean defective activation of bodily states that would normally accompany the perception of emotionally arousing stimuli. The somatic marker framework has been elaborated in detail elsewhere (Damasio, 1994). In brief, the theory posits that ‘marker’ signals arising from bioregulatory processes (including those which express themselves in emotions and feelings) are key influences guiding behavior, especially in social situations. The markers, which can be either overt or covert, are critical for normal reasoning and decision making. The somatic marker hypothesis predicts that somatic marker activation (indexed by SCRs) should occur in relationship to perceiving highly familiar faces, given that such stimuli have a high degree of personal relevance, familiarity, and overall signal value (Tranel et al., 1985; Tranel and Damasio, 1985, 1988). Following this rationale, we predicted that subjects with ventromedial prefrontal lesions would show defective electrodermal responses to familiar faces, as they do for other emotionally laden visual stimuli. This prediction has received empirical support (Tranel et al., 1995). Specifically, we found that four subjects with bilateral ventromedial prefrontal lesions failed to generate SCRs to pictures of familiar faces derived from either the retrograde or anterograde compartments. This occurred despite the fact that the ventromedial subjects have normal overt (conscious) recognition of the faces. Together with previous findings in prosopagnosic subjects, who demonstrate impaired conscious face recognition but normal non-conscious recognition, these results constitute a behavioral and anatomical ‘double dissociation’ between overt and covert face recognition. The findings indicate that the neural systems that process the somatic-based valence, or what

321

could be termed the ‘emotional significance’ of stimuli, are separate from the systems that process the factual information associated with those same stimuli. Covert recognition of auditory stimuli

We have also used our psychophysiology paradigm to investigate non-conscious ‘recognition’ in a patient with a special form of auditory agnosia, thereby extending our work on non-conscious processing to the auditory modality. Subject X-1012 is a 51-year-old, right-handed man who worked as a professional opera singer and professor of voice. He suffered a right temporoparietal hemorrhagic infarction a few years prior to the psychophysiological studies described below, which produced a mild left hemiparesis and hernihypoesthesia. Subsequently, he could no longer interpret music correctly. He had trouble recognizing musical pieces and performers, and in judging the quality of his own voice. Neuroanatomical analysis showed a right posterior lesion, involving the insular cortex and white matter of the superior temporal lobe, undercutting the primary auditory cortex in Heschl’s gyrus. Auditory recognition of familiar stimuli We administered to X-1012 a series of special experiments designed to test his ability to recognize familiar musical pieces. A variety of opera segments, well known to X-1012 from prior to the onset of his brain injury, were selected. Some of the segments were instrumental, and some included a singer. We tested his ability to recognize the pieces and the singers. Compared to a control subject matched for musical knowledge of the specialized type required for this task, X-1012 was severely impaired in the recognition of both instrumental and singing operatic pieces; in the most striking example, he did not even recognize his own singing voice. In sum, X-1012 demonstrated a marked recognition impairment for familiar unique musical information, and since he had no impairments in

basic auditory perception, the profile conforms to a subtype of auditory agnosia. This set the stage for psychophysiological experiments aimed at determining whether the subject could demonstrate ‘covert’ discrimination of the familiar music segments that he could not recognize consciously. Psychophysiological studies Subjects and procedures. Three subjects were studied: X-1012; an ‘expert’ control matched to X1012 on musical knowledge; and a ‘naive’ control who had no special experience with or knowledge of opera music (a Montana rancher brother of the author). 7 b o experiments were conducted. In each, there were six target stimuli (music segments with which the two expert subjects were highly familiar), and 14 non-target stimuli (music segments selected to match the targets in general sound characteristics, but with which the subjects had no familiarity). In the first experiment (Music without Voice), the music segments were instrumental; in the second (Music with Voice), the segments involved singing. Each segment was 8-10s in duration, and they were presented one at a time, in random order, while the subject sat quietly and listened. Skin conductance was recorded. An average target SCR and non-target SCR were calculated for each subject and for each experiment, and these cotlstituted the dependent measures. The target and non-target SCRs were compared statistically.

Results The results are presented in Table 2. The naive control, as expected since he knew none of the stimuli, did not produce discriminatory SCRs to the target stimuli in either experiment. The expert control, in contrast, produced discriminatory largeamplitude SCRs to the target segments in both the Music without Voice and Music with Voice experiments. This indicated that the target stimuli had the intended effect of serving as significant ‘signal stimuli’ to someone expert in this music domain. In X-1012, there were also clear discriminatory SCRs to the target stimuli. This outcome obtained for both the Music without Voice and Music with Voice experiments.

322

TABLE 2 Non-conscious music recognition in auditory agnosia. Skin conductance response magnitudes (in US) Subject x-1012 Expert control Naive control

Music without voices Target Non-target 0.70 1.10 0.01

0.15 0.05 0.04

The target music segments in these experiments were unrecognized by X-1012 at conscious level, and hence the outcome from the psychophysiological index reflects a form of non-conscious recognition in the auditory domain. This extends the phenomenon described previously in regard to visual stimuli. The results suggest that as holds for the visual modality, it is possible for sensory association cortices to activate autonomic effectors in response to a ‘signal’ stimulus, even if that stimulus is not processed in such a manner as to permit conscious recognition. The autonomic responses (SCRs) again provided a sensitive index of non-conscious recognition of familiar stimuli. 11. Non-consciouslearning and retrieval of

affective valence For about two decades, we have been studying a patient known as Boswell, who became severely amnesic following herpes simplex encephalitis (Damasio et al., 1985; Damasio et al., 1989). The disease left him with bilateral lesions that destroyed nearly all cortical and subcortical components of the limbic system, including the mesial temporal lobes (amygdala, entorhinal cortex, hippocampus), the temporal poles and the lateral and inferior aspects of the temporal lobes, the insular cortices, the posterior orbital cortices, and the basal forebrain. Boswell has severe anterograde amnesia, and cannot acquire any new declarative information, such as facts, events, names, objects, or faces. He also has a profound retrograde amnesia, and outside of a couple of sheds of knowledge about his early life, he cannot remember any unique material from his past. As part of his amnesia, he is severely prosopagnosic: he cannot recognize faces of family

Target 1.15 1.43

0.07

Music with voices Non-target 0.10

0.32 0.03

members, friends, or even his own face, and he has never been able to learn faces of persons to whom he has had extensive exposure since his illness. In fact, he has shown no hint of recognition for faces of physicians, psychologists, or caregivers with whom he has had numerous and extensive contacts over the past couple of decades. Some years ago, we noticed that in Boswell’s daily environment (a care facility), he appeared to have a consistent preference for one particular caregiver, a nursing aide to whom Boswell would frequently go when he desired treasured items such as cigarettes or gum. In other words, Boswell appeared to gravitate systematically towards this person, even though it was clear from our studies that Boswell had never learned any factual information about the person, including the person’s name, face, hours of work, and personality characteristics. But Boswell’s behavior hinted at the possibility that he had learned, perhaps at non-conscious level, something about this aide, and we designed a set of experiments to test this idea formally (Tranel and Damasio, 1993). Over the course of several days in our laboratory, Boswell was exposed to three different stimulus persons: (a) a ‘Good Guy’, who treated Boswell very kindly and granted requests for treats and rewards, (b) a ‘Bad Guy’, who never gave Boswell treats and who was always responsible for having Boswell perform tedious neuropsychological experiments, (c) a ‘Neutral Guy’, who approached Boswell in a completely neutral fashion. At the end of a week of exposure, Boswell was tested for declarative knowledge of the three persons, of which he had none. Even with extensive prompting, he could not produce any information about the persons, about his experience with them, or about how they looked. When shown pictures of the

323

persons, he rated them as complete strangers, and denied any inkling of recognition. We then assessed Boswell’s non-conscious or ‘covert’ knowledge for the three stimulus persons. One procedure involved a two-alternative forced choice paradigm. Each of the faces of the Good Guy, Bad Guy, and Neutral Guy was paired with an unfamiliar (but visually similar) face, and the pair was shown to Boswell with the instruction to “choose the person you would go to for a reward”. The pairings were repeated 18 times for each target stimulus person. The results, shown in Fig. 2, indicated that on 15 of 18 trials, Boswell systematically selected the Good Guy over unfamiliar foils. By contrast, he chose the Bad Guy on only 4

of 18 trials, indicating a systematic bias against the Bad Guy. The Neutral Guy was picked at about random level (10 of 18 trials). We also tested Boswell’s non-conscious learning in this experiment with a psychophysiological paradigm. SCRs were recorded while Boswell was shown the faces of the Good Guy, Bad Guy, and Neutral Guy, mixed randomly with unfamiliar foils. The results showed that Boswell produced discriminatory, large-amplitudeSCRs to the Good Guy and Bad Guy, but not to the Neutral Guy or to the unfamiliar non-target faces. These findings confirm the forced choice results, and indicate that Boswell had learned something about faces that were associated with strong affective valence, even

-

‘GOOD GUY BAD GUY” EXPERIMENT

Fig. 2. The graph shows Boswell’s selection frequency in a two-alternative forced choice procedure for the three stimulus persons. The zero-line denotes chance (50%);the bars show the extent to which Boswell’s selections differed from chance.

324

though there was no evidence of this learning at conscious level. This finding is important for several reasons. Because of what we know about Boswell’s lesions, we can conclude that the entorhinal and perirhinal cortices, hippocampus, amygdala, and higher-order neocortices in the anterior temporal region are not required to support this form of covert learning and retrieval. Also, this demonstration is possible only in a patient such as Boswell, because in individuals with normal or only partially impaired factual learning, fact memory will almost inevitably contribute to the performance. We have suggested that Boswell acquired a bond between some aspect of the processing of a face, and some aspect of the processing of a good or bad somatic state that corresponded to the valence associated with the face. The acquisition might proceed as follows. Initially, Boswell perceives a face, e.g. the face of the Good Guy. Directly thereafter, the Good Guy does something pleasant, causing a positive somatic state involving a particular set of changes in visceral and musculoskeletal parameters which are mapped neurally. Then a bond develops between components of the neural processing of the face, and the somatic state. Given the representation of the target face (the Good Guy’s face), this bond regenerates at least part of the original processing set, and on the basis of this reactivated set of signals, Boswell can make systematic choices between faces. It is our impression that the neostriatum, and in particular the caudate, which is completely intact in Boswell, may be a key part of the neural system which subserves such learning (cf. Saper, 1996).

In. Non-consciousinfluences of somatic states on decision-making In a series of experiments designed to explore the neural basis of decision making, we developed a card game known as the Gambling Task, in which the goal is to maximize profit on a loan of play money, with response selection being guided by various schedules of immediate reward and delayed punishment (Bechara et al., 1994). In the Gambling Task, subjects sit in front of four decks of cards equal in appearance and size, and are given a $2000 loan of play money. The subjects

are told that the game requires a long series of card selections (there are 100 in all, but this is never told to the subjects), one card at a time, from any of the four decks, until they are told to stop. After turning each card, the subjects receive money (the amount, which varies from deck to deck, is announced after the card is turned). After turning certain cards, the subjects are both given money and asked to pay a penalty (again, the amount is announced only after the card is turned, and it varies from deck to deck and from position to position within a given deck, according to a schedule unknown to the subjects). The way we have the task set up, turning any card from deck A or deck B yields $100, whereas turning any card from deck C or deck D yields $50. However, the ultimate future yields of the decks vary, because the penalty amounts are higher in the high paying decks (A and B), and lower in the low paying decks (C and D). Thus, it turns out that it is better to select from decks C and D, because over the long haul, the subject will come out with a higher net gain. In sum, decks A and B are ‘disadvantageous’ because they cost the subject in the long run; decks C and D are ‘advantageous’ because they result in an overall gain to the subject in the long run. In order to assess somatic state activation during the Gambling Task (Bechara et al., 1996), we measured skin conductance responses while subjects played the game. We focused in particular on anticipatory SCRs, defined as the SCRs generated immediately prior to the point at which the subject turned a card from any given deck, i.e. during the time period the subject was thinking about which deck to choose. It turned out that normal subjects, as they became experienced with the task, began to generate SCRs prior to the selection of some cards. A statistical analysis indicated that anticipatory SCRs in relation to decks A and B (the ‘bad’ decks) were significantly higher that those generated in relation to decks C and D (the ‘good’ decks). Also, the results indicated that the anticipatory SCRs developed over time, that is, the subjects began to respond electrodermally, and responded more systematically, after selecting several cards from each deck, and thereby encountering several instances of reward and punishment. The anticipatory SCRs gradually became more pronounced prior to the

325

selection of cards from the disadvantageous decks (A and B). We then explored the notion that the overt reasoning used to decide advantageously in a complex situation (the Gambling Task, in this instance) is actually preceded by a non-conscious biasing step, one whose results would not be available to consciousness, and that may use neural systems other than those that support declarative knowledge (Bechara et al., 1997). In this study, 10 normal control subjects performed the Gambling Task. Behavioral (task performance), ‘psychophysiological (SCRs), and self-report measures were obtained. The self-report data were used to judge whether subjects had developed a conscious notion of how the game worked, and on the basis of this, we divided the task into four ‘knowledge periods’, defined below. To do this, we interrupted the task briefly after each subject had made 20 card selections, and asked the subject to respond to the following prompts: (a) “Tell me all you know about what is going on in this game”; and (b) “Tell me how you feel about this game”. The prompts were repeated at 10-card intervals for the duration of the task. After sampling from all four decks, and before they had encountered any punishments, subjects preferred decks A and B, and no anticipatory SCRs were generated. We defined this as the prepunishment period. After encountering some punishments in decks A and B, subjects began to generate anticipatory SCRs to these decks. During this period, though, none of the subjects had any notion as to what was happening in the task, based on their self-reports. We defined this as the prehunch period. Roughly half-way through the task (by about card 50), subjects began to express a ‘hunch’ that decks A and B were less favorable and more risky, and they generated anticipatory SCRs whenever pondering whether to select a card from the A or B decks. We termed this the hunch period. The balance of the task we termed the conceptual period, and during this phase, most of the subjects (7 of 10) reported some knowledge as to the fact that decks A and B were ‘bad’ in the long run, and that decks C and D were ‘good’ in the long run. We scored the behavioral data (card selections) and anticipatory SCR data according to the four

‘knowledge periods’, and obtained several important findings. During the pre-hunch period, there was a significant increase in the magnitude of anticipatory SCRs, i.e. subjects began to develop anticipatory SCRs before they had any clue as to what was happening in the game, and before their behavior changed clearly in favor of the good decks. During the next two periods (hunch and conceptual), the subjects sustained anticipatory SCR activity to the bad decks, but it began to wane for the good decks. Also, the three subjects who failed to reach the conceptual period still continued to avoid the bad decks, and to generate anticipatory SCRs whenever they did opt for a bad deck. We interpreted these results as suggesting that the sensory representations of a situation requiring a decision led to two non-exclusive, interacting chains of events. (a) The sensory representation of a situation, andlor the facts evoked by it, activate neural systems that hold nondeclarative dispositional knowledge related to one’s previous emotional experience of similar situations. We believe the ventromedial prefrontal cortices are an important part of the structures holding such dispositional knowledge. Their activation produces consequent activation of autonomic and neurotransmitter nuclei, along with other brain regions (cf. Mayer, 1995). The ensuing signals, which remain non-conscious, act as covert biases on circuits that support processes of cognitive evaluation and reasoning. (b) The representation of a situation generates (i) overt recall of pertinent facts (e.g. potential response options and probable future outcomes), and (ii) application of conscious, cognoscible reasoning strategies to facts and options. In sum, the results suggest that non-conscious biases guide reasoning and decision-making behavior before conscious knowledge does, and without the help of such biases, overt knowledge may be insufficient to ensure advantageous behavior. Interestingly, studies in social psychology have led to a similar notion, namely, that individuals can learn, and make decisions, with information that is not available to conscious awareness (e.g. Lewicki et al., 1992). We believe that the autonomic responses detected in our experiment (especially those evident in the pre-hunch period) are evidence for a process of non-conscious signaling. We believe

326

this, in turn, refects access of records of previous experience shaped by reward, punishment, and related emotional states.

IV. Conditioning without awaremess As indicated earlier, patients with amnesia, commonly caused by damage to mesial temporal structures (including hippocampus), basal forebrain, or diencephalic nuclei, cannot acquire new information such as facts, words, and faces. However, it has been demonstrated repeatedly that such patients often remain capable of acquiring new sensorimotor skills (e.g. Cohen and Squire, 1980; Eslinger and Damasio, 1986; Gabrieli et al., 1993; Tranel et al., 1994). The key distinction here is between two different types of memorable material (Squire, 1992): ‘Declarative’ information, which refers to representations of facts and events that can be brought to mind in image form, inspected with the ‘mind’s eye’, and testifed about at conscious level. “on-declarative’ information, which refers to knowledge that is not amenable to imagetic representation or conscious inspection. In fact, another way to conceptualize this distinction is along a dimension of ‘explicit’ and ‘implicit’ (e.g. Schacter, 1987, 1992). The acquisition of non-declarative information by amnesic patients who are unable to acquire factual knowledge pertaining to the learning situation can be conceived of as a type of non-conscious learning, in the sense that the patients are unable to report at conscious level anything about what they have learned. Several forms of conditioning, which involves the learning of a contingent association between two stimuli, have been classified as a type of non-declarative memory (Squire, 1992). Immediately following, some particularly striking examples of non-conscious learning, as investigated with conditioning paradigms, are presented. Emotional conditioning without declarative knowledge

We conducted a study in which we investigated the learning of declarative knowledge and the acquisi-

tion of conditioned emotional responses in a subject with severe amnesia caused by focal bilateral damage to the hippocampus (Bechara et al., 1995). Subjects The brain-damaged subject was WC-1606, a 47-year-old, right-handed man who, four years prior to the conditioning studies described below, sustained bilateral damage to the hippocampus (specifically, CA1 neurons) as a consequence of ischemia-anoxia. WC- 1606 has severe anterograde amnesia for both verbal and non-verbal material, but his basic intellectual abilities are normal, and he has normal attention, speech and language, and perception, We also studied four normal control subjects, who were of comparable age and education to the brain-damaged subject. Procedures Tho conditioning experiments were conducted, one visual-auditory and one auditory-auditory. In the visual-auditory experiment, four monochrome slides (green, blue, yellow, blue) served as the stimuli; in the auditory-auditory experiment, four computer-generated tones of different frequencies served as the stimuli. In both experiments, a sudden loud noise served as the unconditioned stimulus (UCS). The dependent measure was the skin conductance response (SCR), recorded with our standard method. The conditioning protocol was comprised by three phases. (a) habituation phase: subjects were presented the stimuli (slides or tones) repeatedly, in random order, until the stimuli no longer elicited orienting SCRs. (b) conditioning phase: the four color slides (or tones) were presented in irregular order, one at a time. There were 26 presentations, of the following nature: six were blue slides followed by the UCS (paired CS); six were blue slides without the UCS (unpaired CS); 14 were red, green, or yellow slides, none of which was ever followed by the UCS (non-CSs). Thus, the blue slide was the paired conditioned stimulus (CS). (This arrangement was the same for the tones; i.e. one tone served as the paired CS, and the other three were non-CSs.) (c) extinction phase: the CS

327

was presented repeatedly until it no longer elicited SCRs from the subjects. Five minutes after completion of the conditioning experiment, the subjects were administered a questionnaire to measure acquisition of declarative knowledge of the experiment. Specifically, subjects were asked about the nature of the CS, and about the nature of the CS-UCS relationship. There were four questions per experiment.

Results The four control subjects produced large-amplitude SCRs to the CS in both the visual-auditory and auditory-auditory experiments (Fig. 3). These SCRs were significantly larger than the SCRs produced to the unpaired stimuli (non-CSs). Subject WC-1606 also acquired conditioned autonomic responses, as evidenced by his large-amplitude

A

Fig. 3. The graphs show SCR magnitudes in the habituation, conditioning, and extinction phases of the visual-auditory (A) and auditory-auditory (B) experiments, for the brain-damaged patient WC-1606 (red lines) and four normal control subjects (blue lines). Each habituation score depicts the mean SCR magnitude for the last three slides or sounds preceding the conditioning phase; each conditioning score depicts the mean SCR magnitude for the six presentations of the unpaired CS (not followed by the UCS); each extinction score depicts the mean SCR magnitude for the first three repeated presentations of the CS during the extinction phase.

328

SCRs to the CS in both experiments, and the relatively smaller SCRs to the unpaired stimuli (Fig. 3). On the factual learning questionnaires, the four control subjects attained nearly perfect scores; by contrast, WC- 1606 was severely impaired in his factual learning, obtaining scores at or close to zero for both the visual-auditory and auditory-auditory experiments. Comment This study demonstrates a striking example of preserved emotional conditioning in a subject who has severe amnesia for declarative knowledge. WC1606 produced evidence, based on the SCR probe, that he had acquired an emotional connection to the CSs, despite the fact that careful questioning revealed that he had learned nothing about the nature of the conditioning experiment, about any of the stimuli involved, and in particular, anything about whether any particular stimuli had special significance. In short, he acquired a conditioned emotional response despite being entirely unable to learn anything factual about the circumstances surrounding that acquisition. It is interesting to note that we obtained exactly the reverse outcome in a patient with bilateral amygdala damage (but without damage to the hippocampus). The patient failed to acquire conditioned SCRs, but was entirely normal in the acquisition of declarative knowledge about the conditioning situation. We have interpreted these results as indicating that the human amygdala is critical for emotional conditioning, and for the coupling of exteroceptive sensory information with interoceptive information concerning somatic states (Nahm et al., 1993; Damasio, 1994; Bechara et al., 1995; Furmark et al., 1997). The hippocampus, on the other hand, is critical for learning declarative knowledge, but appears to be unnecessary for the acquisition of conditioned autonomic responses. Aversive conditioning without awareness

An elegant series of experiments from Ohman’s laboratory has also demonstrated robustly the phenomenon of conditioning without awareness

(Ohman and Soares, 1993, 1994, 1998; Soares and Ohman, 1993a, 1993b; Esteves et al., 1994a,b; Mlot, 1998). In one study of this type, Ohman and Soares (1994) showed pictures of snakes, spiders, flowers, and mushrooms to subjects who were preselected because they were phobic of snakes and spiders. An extremely short (30 ms) stimulus exposure was used, immediately followed by a masking stimulus, so that perceptual information was blocked from entering conscious awareness. The investigators found that the subjects produced high-amplitude SCRs to the snakes and spiders, despite being unaware at conscious level of the content of the stimulus material. Ohman’s laboratory has reported similar results with fear-conditioningparadigms, using snakes, spiders, or angry faces as the target stimuli. In another recent experiment along these same lines, Ohman and Soares (1998) showed pictures of fear-relevant (snakes, spiders) or fear-irrelevent (flowers, mushrooms) stimuli to subjects who were not snake- or spider-phobic. Stimulus exposure was very brief (30 ms), and a masking stimulus was presented immediately after each stimulus, to block conscious perception of the target stimuli. A conditioning paradigm was used, in which certain stimuli (CS) were paired with an electric shock (UCS). SCRs were utilized as the dependent measure of conditioning. The investigators obtained clear and reliable conditioning effects for the fear-relevant stimuli (snakes and spiders), despite the fact that subjects were completely unable to identify consciously the conditioned stimuli. Interestingly, fear-irrelevant stimuli did not show conditioning effects. In a follow-up experiment, Ohman and Soares (1998) replicated this outcome, and also showed that subjects were completely unable to discriminate conditioned stimuli on the basis of a forced choice procedure. In sum, the experiments show reliable aversive conditioning to pictures of snakes and spiders, in the absence of any conscious recognition of the conditioned stimuli. Recent functional imaging studies, using the masking paradigm developed by Ohman, have confirmed findings from the conditioning experiments, and have also demonstrated that selective activation of the amygdala occurs in subjects who

329

are exposed to or conditioned to angry faces that the subjects cannot perceive consciously. This result was obtained in a functional magnetic resonance imaging study that used a simple passive viewing procedure (Whalen et al., 1998), and in a positron emission tomography study that used a conditioning paradigm (Morris et al., 1998). In the latter study, activation of the right amygdala (along with skin conductance increases) was especially prominent when subjects were completely unaware of having perceived the conditioned stimuli (angry faces); when subjects were able to report awareness of the angry faces, there was relatively greater activity in the left amygdala. The authors interpreted this result as evidence that the left and right amygdala may play different roles in conscious and non-conscious processing, with the left amygdala being relatively more important for conscious operations, and the right amygdala being relatively more important for non-conscious operations. Together, the results from these functional imaging studies are quite consistent with findings from lesion work summarized earlier.

V. Summary and concluding comments Our program of research into neural and cognitive aspects of non-conscious processing began nearly a decade and a half ago, with the discovery that a psychophysiological index - the skin conductance response - could reveal evidence of preserved face recognition in brain-damaged patients with prosopagnosia. Since that time, we and others have used this index successfully in a variety of paradigms, to uncover various forms of preserved learning and recognition in patients with other types of agnosia, amnesia, and disturbed emotional processing. Findings from our experiments, and from other laboratories, indicate that in many situations, nonconscious processing can proceed quite independently from conscious operations, and in fact, the two modes appear to have considerably different neural substrates. We have also found that non-conscious processing triggers important somatic markers (emotions) that provide key influences in reasoning and decision making. Together, the findings comprise an impressive array of new

breakthroughs shedding light on a wide range of neural and cognitive operations that support the large part of our mental life that takes place beneath the level of conscious awareness. In addition, evidence of preserved non-conscious processing in brain-damaged patients provides encouraging hints for potential rehabilitation and treatment programs. The discovery of preserved learning of non-declarative information in patients with impaired declarative learning, for example, has already spurred the development and implementation of treatment programs for managing anxiety and other behavioral maladies in such patients (e.g. Suhr et al., 1999). We have also used the knowledge derived from our experiments on non-conscious face recognition to help address the challenges that agnosic patients face in their dayto-day lives. It is likely that applications such as this will find an increasingly important role in the management of cognitive and behavioral symptoms in neurological patients (e.g. Anderson, 1996), and this line of work also holds considerable promise for the treatment of patients with psychiatric disease. The importance of pursuing careful scientific investigation of non-conscious processing, from both neural and cognitive perspectives, can hardly be overemphasized. Had we not secured a means to explore non-conscious learning and recognition in brain-damaged patients, through the application of a psychophysiological technique, we would have missed a number of key opportunities to learn how important, and how extensive, non-conscious processing actually is in the human brain. This would not only have prevented us from gaining important new scientific knowledge, but it would also have narrowed substantially the scope of theory-building regarding brain-behavior relationships (cf. Damasio, 1989). Finally, it is clear that meaningful progress in our understanding of non-conscious processing requires a multidisciplinary approach to the topic, as evident in the proceedings summarized in this volume. Insofar as the neural basis of nonconscious processing is concerned, it is exciting indeed to ponder the potentialities for further breakthroughs based on new neuroimaging techniques, including positron emission tomography and functional magnetic resonance imaging.

330

Acknowledgements I thank my colleagues, Antonio Damasio, Hanna Damasio, Antoine Bechara, and Ralph Adolphs, for their collaboration in the studies reported in this chapter. I am also grateful to Ken Manzel for his help with the figures. Our work is supported by NINDS Program Project Grant NS19632.

References Adams, J.K. (1957) Laboratory studies of behavior without awareness. Psychol. Bull., 54: 383405. Anderson, S.W. (1996) Cognitive rehabilitation in closed head injury. In: M. Rizzo and D. Tranel (Eds), Head Injury and Postconcussive Syndrome, Churchill Livingstone, New York, pp. 457468. Bauer, R.M. (1984) Autonomic recognition of names and faces in prosopagnosia: A neuropsychological application of the Guilty Knowledge Test. Neumpsychologia, 22: 457469. Bauer, R.M. and Verfaellie, M. (1988) Electrodermal discrimination of familiar but not unfamiliar faces in prosopagnosia. Brain Cogn., 8: 240-252. Bechara, A,, Damasio, A.R., Damasio, H. and Anderson, S.W. (1994) Insensitivity to future consequences following damage to human prefrontal cortex. Cognition, 50: 7-12. Bechara, A., Damasio, H., Tranel, D. and Damasi0,A.R. (1997) Deciding advantageously before knowing the advantageous strategy. Science, 275: 1293-1295. Bechara, A,, Tranel, D., Damasio, H.,Adolphs, R., Rockland, C. and Damasio, A.R. (1995) Double dissociation of conditioning and declarative knowledge relative to the amygdala and hippocampus in humans. Science, 269: I 1 15-1 118. Bechara, A., Tranel, D., Damasio, H. and Damasio, A.R. (1996) Failure to respond autonomically to anticipated future outcomes following damage to prefrontal cortex. Cerebl: Cort., 6: 2 15-225. Berns, G.S., Cohen, J.D. and Mintun, M.A. (1997) Brain regions responsive to novelty in the absence of awareness. Science, 276: 1272-1275. Bradley, M.M., Greenwald, M.K., Petry, M.C. and Lang, P.J. (1992) Remembering pictures: Pleasure and arousal in memory. J. Expel: Psychol.: Learn., Mem. Cogn., 18: 379-390. Churchland, P.S. (1986) Neumphilosophy: Toward a Unijied Science of the Mind/Brain, MIT Press, Cambridge, MA. Churchland, P.S. and Sejnowski, T.J. (1992) The Computational Brain, h4IT Press, Cambridge, MA. Cohen, N.J. and Squire, L.R. (1980) Preserved learning and retention of pattern-analyzing skill in amnesia: Dissociation of knowing how and knowing that. Science, 210: 207-210. Corteen, R.S. and Wood, B. (1972) Autonomic responses to shock-associated words in an unattended channel. J. Exper: Psychol., 94: 308-313.

Crick, F. (1994) The Astonishing Hypothesis: The Scientijic Search for the Soul, Scribners, New York. Damasio, A.R. (1989) Time-locked multiregional retroactivation: A systems-level proposal for the neural substrates of recall and recognition. Cognition, 33: 25-62. Damasio, A.R. (1994) Descartes’ Error: Emotion, Reason and the Human Bruin, Grosset/Putnam, New York. Damasio, A.R. (1999) The Feeling of What Happens: Body and Emotion in the Making of Consciousness, Harcourt Brace, New York. Damasio, A.R., Eslinger, P., Damasio, H., Van Hoesen, G.W. and Cornell, S. (1985) Multimodal amnesic syndrome following bilateral temporal and basal forebrain damage. Arch. Neuml., 42: 252-259. Damasio, A.R., Tranel, D. and Damasio, H. (1989) Amnesia caused by herpes simplex encephalitis, infarctions in basal forebrain, Alzheimer’s disease and anoxia. In: F. Boller and J. Grafman (Eds), Handbook of Neumpsychology, Vol. 3, Elsevier, Amsterdam, pp. 149-166. Damasio, A.R., Tranel, D. and Damasio, H. (1990) Individuals with sociopathic behavior caused by frontal damage fail to respond autonomically to social stimuli. Behav. Brain Res.. 41: 81-94. Damasio, A.R., Tranel, D. and Damasio, H. (1991) Somatic markers and the guidance of behavior: Theory and preliminary testing. In: H.S. Levin, H.M. Eisenberg and A.L. Benton (Eds), Frontal Lobe Function and Dysfunction, Oxford University Press, New York, pp. 217-229. Damasio, H. and Damasio, A.R. (1997) The lesion method in behavioral neurology and neuropsychology. In: T.E. Feinberg and M.J. Farah (Eds), Behavioral Neurology and Neumpsychology, McGraw-Hill, New York, pp. 69-82. Damasio, H. and Frank, R. (1992) Three-dimensional in vivo mapping of brain lesions in humans. Arch. Neurol., 49: 137-143. de Haan, E.H.F., Young, A. and Newcombe, F. (1987a) Faces interfere with name classification in a prosopagnosic patient. Cortex, 23: 309-316. de Haan, E.H.F., Young, A. and Newcombe, F. (1987b) Face recognition without awareness. Cogn. Neumpsychol., 4: 385-4 15. Dennett, D.C. (199 1) Consciousness Explained, Little, Brown and Company, Boston. Eslinger, P.J. and Damasio, A.R. (1986) Preserved motor learning in Alzheimer’s disease: Implications for anatomy and behavior. J. Neumsci., 6: 3006-3009. Esteves, F., Dimberg, U. and Ohman, A. (1994a) Automatically elicited fear: Conditioned skin conductance responses to masked facial expressions. Cogn. Emot., 8: 393413. Esteves, F., Parra, C., Dimberg, U. and Ohman, A. (1994b) Non-conscious associative learning: Pavlovian conditioning of skin conductance responses to masked fear-relevant facial stimuli. Psychophysiology, 31: 375-385. Farah, M.J. and Feinberg, T.E. (1997) Consciousness of perception after brain damage. Sem. Neuml., 17: 145-152.

33 1 Frank, R.J., Damasio, H. and Grabowski, T.J. (1997) Brainvox: An interactive, multimodal, visualization and analysis system for neuroanatomical imaging. NeuroImage, 5: 13-30. Furmark, T., Fischer, H., Wik, G., Larsson, M. and Fredrikson, M. (1997) The amygdala and individual differences in human fear conditioning. NeuroReport, 8: 3957-3960. Gabrieli. J.D.E., Corkin, S., Mickel, S.F. and Growdon, J.H. (1993) Intact acquisition and long-term retention of mirrortracing skill in Alzheimer’s disease and global amnesia. Behav. Neurosci., 107: 899-910. Greve, K.W. and Bauer, R.M. (1990) Implicit learning of new faces in prosopagnosia: An application of the mere-exposure paradigm. Neuropsychologia, 28: 1035-1041. Jones, R. and Tranel, D. (1998) Severe ‘associative’ developmental prosopagnosia in a child with superior intellect. J. Int. Neuropsychol. SOC.,4: 60. Kihlstrom, J. (1987) The cognitive unconscious. Science, 237: 145-1452. Kolb, F.C. and Braun, J. (1995) Blindsight in normal observers. Nature, 377: 336-338. Kosslyn, S.M. and Koenig, 0. (1992) Wet Mind: The New Cognitive Neuroscience, The Free Press, New York. Lazarus, R.S. and McCleary, R.A. (1951) Autonomic discrimination without awareness: A study of subception. Psychol. Rev., 58: 113-122. Leuthold, H. and Kopp, B. (1998) Mechanisms of priming by masked stimuli: Inferences from event-related brain potentials. Psycholo. Sci., 9: 263-269. Lewicki, P., Hill, T. and Czyzewska, M. (1992) Non-conscious acquisition of information. Am. Psychol., 47: 796-801. Luck, S.J., Vogel, E.K. and Shapiro, K.L. (1996) Word meanings can be accessed but not reported during the attentional blink. Nature, 383: 616-618. Mayer, E.A. (1995) Gut feelings: what turns them on? Gastroenterology, 108: 927-93 1. Milner, A.D. and Rugg, M.D. (Eds.) (1992) The Neuropsychology of Consciousness, Academic Press, New York. Mlot, C. (1998) Unmasking the emotional unconscious. Science, 280: 1006. Moms, J.S., Ohman, A. and Dolan, R.J. (1998) Conscious and unconscious emotional learning in the human amygdala. Nature, 393: 4674‘0. Nahm, F.K.D., Tranel, D., Damasio, H. and Damasio, A.R. ( 1993) Cross-modal associations and the human amygdala. Neuropsychologia, 3 1: 727-744. Ohman, A. and Soares, J.J.F. (1993) On the automaticity of phobic fear: Conditioned skin conductance responses to masked phobic stimuli. J. A b n o n . Psychol., 102: 121-132. Ohman, A. and Soares, J.J.F. (1994) Unconscious anxiety: Phobic responses to masked stimuli. J. A b n o n . Psychol., 103: 231-240. Ohman, A. and Soares, J.J.F. (1998) Emotional conditioning to masked stimuli: Expectancies for aversive outcomes following non-recognized fear-relevant stimuli. J. Exper: Psychol.: Gen., 127: 69-82. Reber, AS. (1989) Implicit learning and tacit knowledge. J. Exper: Psychol.: Gen., 118: 219-235.

Reiser, M.F. and Block, J.D. (1965) Discrimination and recognition of weak stimuli. 111. Further experiments on interaction of cognitive and autonomic-feedback mechanisms. Psychosom. Med., 27: 274-285. Rizzo, M., Hurtig, R. and Damasio, A.R. (1987) The role of scanpaths in facial learning and recognition. Ann. Neurol., 22: 41-45. Roediger, H.L. (1990) Implicit memory: Retention without remembering. Am. Psychol., 45: 1043-1056. Rousey, C. and Holzman, P.S. (1967) Recognition of one’s own voice. J. Personality SOC.Psychol., 6: 464466. Saper, C.B. (1996) Role of the cerebral cortex and striatum in emotional motor response. Prog. Brain Res., 107: 537-550. Schacter, D.L. (1987) Implicit memory: History and current status. J. Exper: Psychol.: Learn., Mem. C o p , 13: 501-518. Schacter, D.L. (1992) Implicit knowledge: New perspectives on unconscious processes. Proc. Natl. Acad. Sci. USA, 89: 1113-1 117. Soares, J.J.F. and Ohman, A. (1993a) Backward masking and skin conductance responses after conditioning to nonfeared but fear-relevant stimuli in fearful subjects. Psychophysiology, 30: 460466. Soares, J.J.F. and Ohman, A. (1993b) Preattentive processing, preparedness and phobias: Effects of instruction on conditioned electrodermal responses to masked and non-masked fear-relevant stimuli. Behav. Res. Thec, 31: 87-95. Squire, L.R. (1992) Memory and the hippocampus: A synthesis from findings with rats, monkeys and humans. Psychol. Rev., 99: 195-231. Suhr, J., Anderson, S. and Tranel, D. (1999) Progressive muscle relaxation in the management of behavioral disturbance in Alzheimer’s disease. Neuropsychol. Rehab., 9: 3 1 4 . Tranel, D. and Damasio, A.R. (1985) Knowledge without awareness: An autonomic index of facial recognition by prosopagnosics. Science, 228: 1453-1454. Tranel, D. and Damasio, A.R. (1988) Non-conscious face recognition in patients with face agnosia. Behav. Brain Res., 30: 235-249. Tranel, D. and Damasio, A.R. (1993) The covert learning of affective valence does not require structures in hippocampal system or amygdala. J. Cogn. Neurosci., 5: 79-88. Tranel, D., Bechara, A. and Damasio, A.R. (1999) Decisionmaking and the somatic marker hypothesis. In: M. Gazzaniga (Ed.), The Cognitive Neurosciences, 2nd a n . , MIT Press, Cambridge, MA. pp. 1047-1061. Tranel, D., Damasio, A.R., Damasio, H. and Brandt, J.P. (1994) Sensorimotor skill learning in amnesia: Additional evidence for the neural basis of nondeclarative memory. Learn. Mem., 1: 165-179. Tranel, D. and Damasio, H. (1989) Intact electrodermal skin conductance responses after bilateral amygdala damage. Neuropsychologia, 27: 381-390. Tranel, D. and Damasio, H. (1994) Neuroanatomical correlates of electrodermal skin conductance responses. Psychophysiology, 31: 427-438.

332 Tranel, D., Damasio, H. and Damasio, A.R. (1995) Double dissociation between overt and covert face recognition. J. Cogn. Neurosci., I: 425-432. Tranel, D., Fowles, D.C. and Damasio, A.R. (1985) Electrodermal discrimination of familiar and unfamiliar faces: A methodology. Psychophysiology, 22: 403408.

Viggiano, M.P. (1996) Event-related potentials in brain-injured patients with neuropsychological disorders: A review. J. Clin. Exper. Neuropsychol., 18: 63 1-647. Whalen, P.J., Rauch, S.L., Etcoff, N.L., McInerney, S.C., Lee, M.B. and Jenike, M.A. (1998) Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge. J. Neumsci., 18: 41 1-418.

E.A. Mayer and C.B. Saper @is.) Progress in Brain Research,Vol 122 8 2000 Elsevier Science BV. All rights reserved.

CHAPTER 24

Brain mediation of active and passive emotional coping Richard Bandler',2.*,Joseph L. Price3 and Kevin A. Keay'92

' Department of Anatomy and Histology, The University of Sydney, Sydney, NSW Australia 2006

'Pain Management and Research Centre, Royal North Shore Hospital, St Leonards, NSW Australia 2065 Department of Anatomy and Neurobiology, Washington University, School of Medicine, St. Louis, MO 63110, USA

Introduction Animals, including humans, react with distinct emotional coping strategies to different sets of environmental demands. Although the hypothalamus has been usually assigned the key role in coordinating emotional responses, it is difficult to reconcile this view with findings that integrated and directed emotional reactions are readily evoked in animals in which the hypothalamus has been either surgically isolated from the rest of the brain (Ellison and Flynn, 1968; Gellen et al., 1972) or extensively lesioned (Kelly et al., 1946; Fernandez DeMolina and Hunsperger, 1962). An emerging body of evidence suggests that longitudinal columns of midbrain neurons located lateral and ventrolateral to the aqueduct, within the periaqueductal gray region (PAG), play special roles in coordinating distinct emotional strategies for coping with different classes of stress, threat or pain. Such studies fit well with earlier findings that extensive lesions involving the PAG eliminated or attenuated emotional coping reactions (i.e. threat, freezing, escape, vocalization) evoked by confrontation with another animal, forced restraint, electric shock, aversive noise, or placement in an open field (Bandler, 1988). This chapter will consider functional-anatomical evidence that the lateral and ventrolateral columns *Corresponding author. Tel.: (61) (2) 9351.2500; Fax: (61) (2) 9351.2813; e-mail: [email protected]

of the PAG coordinate fundamentally opposite modes of emotional coping. It will be proposed that the PAG offers a useful point of entry for defining neural circuits which mediate different emotional coping strategies. As an example, recent studies of the connections of primate orbital and medial prefrontal cortical (PFC) fields with specific longitudinal neuronal columns within the PAG will be reviewed. The findings of discrete orbital and medial PFC projections to columns of the PAG, and related PFC and PAG columnar connections with specific subregions of hypothalamus and amygdala, offer an initial anatomical delineation of neural circuits mediating different emotional coping strategies characterized by either 'engagement with' or 'disengagement from' the external environment. The practices of meditation and yogic breathing, which are suggested alternative methods for the management of pain and stress, likely engage the same neural circuits via which particular emotional coping strategies are expressed.

Active and passive emotional coping Emotional coping includes the capacity to affect appropriate responses to both 'escapable' and 'inescapable' threatening or stressful situation, and to facilitate recovery and healing, once a threat or stress passes. Active emotional coping strategies such as confrontation, fight or flight are adaptive for coping with classes of threat or stress from which escape is possible. Autonomic changes associated with such coping strategies are domi-

334

nated by sympatho-excitation, e.g. hypertension and tachycardia. In contrast, passive emotional coping (also known as conservation-withdrawal), which comprises quiescence, immobility and decreased responsiveness to the environment, is adaptive for coping with ‘inescapable’ stressors, e.g. consequences of severe injury - blood loss, deep pain, and for facilitating recovery and healing. The autonomic components associated with passive coping are characterized usually by sympathoinhibition, e.g. hypotension and bradycardia.

Neural substrates of active vs. passive emotional coping: functional studies Lateral column of the PAG organizes active emotional coping When an excitatory amino acid (EAA) is microinjected into the region of the PAG immediately lateral to the aqueduct, a calm and placid animal

responds as if its environment has become suddenly, very threatening (Bandler, 1982; Bandler and Depaulis, 1991; Bandler et al., 1991; Depaulis et al., 1992; Bandler and Shipley, 1994; Zhang et al., 1994; Bandler and Keay, 1996). Further, the region of the lateral column excited determines which active coping strategy is utilized (see Fig. 1). Excitation of neurons within the rostral half of the lateral column evokes a confrontational style of response - the animal faces and threatens whatever the stressor. In the cat, for example, this response is char