The Maudsley Family Study of Psychosis: A Quest for Intermediate Phenotypes (Maudsley Monographs)

THE MAUDSLEY Maudsley Monographs

MAUDSLEY MONOGRAPHS HENRY MAUDSLEY, from whom the series of monographs takes its name, was the founder of The Maudsley Hospital and the most prominent English psychiatrist of his generation. The Maudsley Hospital was united with the Bethlem Royal Hospital in 1948 and its medical school, renamed the Institute of Psychiatry at the same time, became a constituent part of the British Postgraduate Medical Federation. It is now a school of King's College, London, and entrusted with the duty of advancing psychiatry by teaching and research. The South London & Maudsley NHS Trust, together with the Institute of Psychiatry, are jointly known as The Maudsley. The monograph series reports high quality empirical work on a single topic of relevance to mental health, carried out at the Maudsley. This can be by single or multiple authors. Some of the monographs are directly concerned with clinical problems; others, are in scienti®c ®elds of direct or indirect relevance to mental health and that are cultivated for the furtherance of psychiatry.

Editor Professor A. S. David MPhil MSc FRCP MRCPsych MD Assistant Editor Professor T. Wykes BSc PhD MPhil

1955±1962 1962±1966 1966±1970 1970±1979 1979±1981 1981±1983 1983±1989 1989±1993 1993±1999

Previous Editors Professor Sir Aubrey Lewis LLD DSc MD FRCP and Professor G. W. Harris MA MD DSc FRS Professor Sir Aubrey Lewis LLD DSc MD FRCP Professor Sir Denis Hill MB FRCP FRCPsych DPM and Professor J. T. Eayrs PhD DSc Professor Sir Denis Hill MB FRCP FRCPsych DPM and Professor G. S. Brindley Professor G. S. Brindley MD FRCP FRS and Professor G. F. M. Russell MD FRCP FRC(ED) FRCPsych Professor G. F. M. Russell MD FRCP FRCP(ED) FRCPsych Professor G. F. M. Russell MD FRCP FRCP(ED) FRCPsych and Professor E. Marley MA MD DSc FRCP FRCPsych DPM Professor G. F. M. Russell MD FRCP FRCP(ED) FRCPsych and Professor B. H. Anderton BSc PhD Professor Sir David Goldberg MA DM MSc FRCP FRCPsych DPM

Maudsley Monographs number ®fty

The Maudsley Family Study of Psychosis A Quest for Intermediate Phenotypes Edited by Colm McDonald

First published 2008 by Psychology Press 27 Church Road, Hove, East Sussex BN3 2FA Simultaneously published in the USA and Canada by Psychology Press 270 Madison Avenue, New York, NY 10016 Psychology Press is an imprint of the Taylor & Francis Group, an Informa business

This edition published in the Taylor & Francis e-Library, 2008. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” Ø 2008 Psychology Press All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. The publisher makes no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for any errors or omissions that may be made. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data The Maudsley family study of psychosis : a quest for intermediate phenotypes / edited by Colm McDonald. p. ; cm. ± (Maudsley monograph, ISSN 0076-5465 ; no. 50) Includes bibliographical references and index. ISBN 978-1-84169-734-5 (hardback) 1. Schizophrenia±Genetic aspects. 2. Manic-depressive illness± Genetic aspects. 3. Phenotype. 4. Genetic markers. I. McDonald, Colm. II. Series: Maudsley monographs ; no. 50. [DNLM: 1. Schizophrenia±genetics. 2. Family. 3. Genetic Predisposition to Disease±genetics. 4. Psychotic Disorders±genetics. W1 MA997 no.50 2008 / WM 203 M448 2008] RC514.M375 2008 616.89©8042±dc22 2008001507

ISBN 0-203-88525-2 Master e-book ISBN

ISBN: 978-1-84169-734-5 ISSN: 0076-5465

Contents

List of contributors List of ®gures List of colour plates List of tables Preface 1.

Exploring intermediate phenotypes of psychosis Colm McDonald, Robin Murray The contribution of genetic liability to schizophrenia Heritability estimates for schizophrenia Mode of genetic transmission Molecular genetics Overlap in genetic liability for schizophrenia and bipolar disorder Endophenotypes The Maudsley Family Study of Psychosis References

2.

The Maudsley Family Study of Psychosis ± overview of clinical methodology and characteristics Colm McDonald Participant recruitment Study phases Inclusion and exclusion criteria Clinical assessments

ix xi xiii xv xix 1 1 3 3 4 6 10 12 14 21 21 22 22 23

v

vi

CONTENTS

Sociodemographic characteristics of the complete sample Clinical characteristics of the sample Genetic liability scale References 3.

Auditory evoked potentials as genetic trait markers of schizophrenia I. Williams, S. Frangou and E. Bramon Introduction How severe are neurophysiological de®cits in schizophrenia? Meta-analyses of the published literature The acquisition and analysis of EEG/ERP data in the Maudsley Family Study of Psychosis Results Discussion Conclusion References

4.

Are eye movement abnormalities related to susceptibility genes for schizophrenia? James MacCabe, Jolanta Zanelli Introduction Methodology Results Discussion Conclusions References

5.

Neuropsychological impairments in patients with schizophrenia and their unaffected relatives Timothea Toulopoulou, Francesca Filbey and Eugenia Kravariti Background Study 1: Episodic memory Study 2: Executive function Study 3: Sustained and selective attention Study 4: Intellectual asymmetry Overall summary of the neuropsychological ®ndings and conclusions References

6.

Neurological abnormalities in patients with schizophrenia from singly and multiply affected families and their relatives Paola Dazzan, Timothy D. Grif®ths Introduction Methods

25 27 29 38 41 41 44 49 51 54 63 64

71 71 75 78 82 88 88

93 93 94 105 114 120 124 126

133 133 135

CONTENTS

Results Discussion Conclusion Acknowledgements References 7.

Structural brain deviations in schizophrenia and bipolar disorder ± to what extent are they genetically mediated? Colm McDonald Introduction Study 1: Region-of-interest analyses of patients with schizophrenia or bipolar disorder and their unaffected relatives Study 2: Structural brain deviations associated with schizophrenia and bipolar disorder assessed using computational morphometry Study 3: Structural brain deviations associated with genetic liability to schizophrenia and bipolar disorder assessed using computational morphometry Summary References

8.

Summary and implications Colm McDonald Summary of key ®ndings Implication of ®ndings The next stages

Appendix 1 Index

vii

138 146 151 151 151 155 155 157 167 179 188 189 197 197 203 207 209 213

List of contributors

Elvira Bramon, Senior Lecturer, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Paola Dazzan, Senior Lecturer, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Francesca Filbey, Research Scientist, The MIND Research Network, Adjunct Assistant Professor, Psychology Department, University of New Mexico, Albuquerque, NM 87131 USA Sophia Frangou, Reader, Head, Section of Neurobiology of Psychosis, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London

ix

x

LIST OF CONTRIBUTORS

Timothy Grif®ths, Professor of Cognitive Neurology, Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH Eugenia Kravariti, Lecturer, Section of Epidemiology and Social Psychiatry, Institute of Psychiatry, King's College London, London James MacCabe, Clinical Lecturer, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Colm McDonald, Professor of Psychiatry, College of Medicine, Nursing and Health Sciences, National University of Ireland Galway, Galway, Ireland Robin M. Murray, Professor of Psychiatry, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Timothea Toulopoulou, Research Scientist/Lecturer, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Ian Williams, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London Jolanta Zanelli, Division of Psychological Medicine and Psychiatry, Institute of Psychiatry, King's College London, London

List of ®gures

2.1 2.2 3.1 3.2 3.3 3.4 4.1 5.1 6.1 7.1 7.2

Step-by-step demonstration of genetic liability score calculation in a single family multiply affected with bipolar disorder 31±34 Genetic liability scores in families with differing density of illness 35±37 P300 amplitude and P300 latency 46 Forest plots of primary studies in unaffected relatives of patients with schizophrenia and main meta-analysis ®ndings 48 Meta-analysis of the P50 ratio 50 Mismatch negativity group average waveforms 55 Composition of subject groups and contrasts 76 Scatterplot of genetic liability scores and Verbal±Spatial Contrast IQ (unadjusted values) 123 Distributions of primary and integrative abnormalities between groups 144 Mean total lateral ventricular volume, adjusted for confounds, in each participant group in entire sample 162 Mean total hippocampal volume, adjusted for confounds, in each participant group in the entire sample 163

xi

List of colour plates

Plate 1 P300 group average waves from Phases 2 and 3 of the Maudsley Family Study of Psychosis. Plate 2 Example of Stroop colour±word stimuli. Plate 3 Map of grey matter volume de®cits when comparing patients with schizophrenia to healthy comparison participants, superimposed onto a single brain in standard stereotactic space. Plate 4 Map of grey matter volume de®cits when comparing patients with schizophrenia to patients with bipolar disorder, superimposed onto a single brain in standard stereotactic space. Plate 5 Map of white matter volume de®cits when comparing patients with schizophrenia to healthy comparison participants and patients with bipolar disorder to healthy comparison participants, superimposed onto a single brain in standard stereotactic space. Plate 6 Grey matter endophenotypes. Plate 7 Figures demonstrating similar linear associations between systemic tissue volume de®cits and genetic risk estimated separately for patients and their non-psychotic relatives for schizophrenia or bipolar disorder. Plate 8 White matter endophenotypes. Plate 9 Linear associations, demonstrating disorder-speci®c grey matter endophenotypes and a disorder-generic white matter endophenotype, between systemic tissue de®cits de®ned as endophenotypic for schizophrenia or bipolar disorder and genetic liability scores estimated separately for non-psychotic relatives of schizophrenia patients and non-psychotic relatives of bipolar patients.

xiii

List of tables

2.1 2.2 3.1 3.2 3.3 3.4 3.5 3.6 3.7 4.1 4.2 4.3 4.4 4.5 4.6 5.1 5.2

Sociodemographic characteristics of each participant group Clinical details on premorbid traits and course of illness in patients Main results of the meta-analysis of P300 case±control studies Main results of the meta-analysis of P300 family studies Main results of meta-analyses of P50 case±control studies The 1997 (Phase 1) and 2005 (Phases 2 and 3) Maudsley Family Study of Psychosis samples. Results for P300 amplitude and latency by group and region Mismatch negativity amplitudes for three groups at F3, F4 and FZ Main results from the model at F3 and at FZ Summary of the most plausible neurophysiological endophenotypes for psychosis Lifetime DSM-IV diagnoses and demographic details in the ®ve groups Overall mean values with standard deviation at 30 degrees/second Overall mean values with standard deviation at 15 degrees/second Comparisons of group means for the smooth pursuit variables Means for saccadic scores Comparisons of group means for saccadic distractibility Demographic characteristics of schizophrenia patients, relatives and normal control participants for Study 1 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in verbal memory; patients and relatives versus controls

26 28 45 47 49 52 54 54 59 79 80 80 81 83 83 98 99

xv

xvi

5.3

5.4 5.5

5.6 5.7 5.8

5.9 5.10

5.11 5.12 5.13 6.1 6.2 6.3 6.4 7.1 7.2 7.3 7.4

LIST OF TABLES

Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in verbal memory; familial patients and relatives, non-familial patients and relatives, and presumed obligates versus controls 100 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in visual memory; schizophrenia patients and relatives versus controls 101 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in visual memory; familial patients and relatives, non-familial patients and relatives, and presumed obligates versus controls 102 Demographic characteristics of schizophrenia patients, relatives and normal control participants for Study 2 108 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in planning ability; patients and relatives versus controls 109 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in planning ability; familial schizophrenia patients and relatives, non-familial schizophrenia patients and relatives, non-familial patients and relatives and presumed obligates versus controls patients and relatives, and presumed obligates versus controls 110 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in spatial working memory and strategy formation; patients and relatives versus controls 111 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in spatial working memory and strategy formation; familial patients and relatives, non-familial patients and relatives, and presumed obligates versus controls 112 Sociodemographic means of subjects per group classi®cation for Study 3 116 Attention measures per group when compared to the normal controls with age, gender and IQ covariates 118 Sociodemographic characteristics for Study 4 122 Neurological measures used 136 Group demographics 139 Group scores for individual measures 141±142 Group primary and integrative measures 145 Mean adjusted regional brain volumes in each participant group 159 Analyses of regional brain volumes for each participant group compared to the control group 160 Talaraich co-ordinates and Brodmann areas for regions of grey matter volume de®cit in schizophrenia compared to normal comparison participants 170 Talaraich co-ordinates and Brodmann areas for regions of grey matter volume de®cit in schizophrenia compared to bipolar disorder 172

LIST OF TABLES

7.5 7.6

7.7 8.1

xvii

Talaraich co-ordinates and Brodmann areas for regions of white matter volume de®cit in patients (schizophrenia and bipolar disorder) compared to normal comparison participants 173 Anatomical location, approximate Brodmann areas, cluster size and loading scores on ®rst principal components (PC) analyses for endophenotypic regions of grey and white matter signi®cantly associated with genetic liability (GL) for schizophrenia and bipolar disorder 181±182 Genetic±phenotypic (G-P) associations between genetic liability scores and grey or white matter endophenotypic systems 183 Summary of main neurobiological abnormalities supported as intermediate phenotypes of schizophrenia by the Maudsley Family Study of Psychosis 204

Preface Colm McDonald and Robin Murray

Psychotic illnesses such as schizophrenia have devastating personal consequences for those af¯icted, and for their families. Despite some advances in recent years, our understanding of the neurobiological underpinnings of psychosis is limited. Schizophrenia has a strong hereditary component but the central question remains: what precisely is inherited? It is likely that inherited gene variants in concert with environmental risk factors impair critical aspects of brain function and brain structure, and that the psychotic symptoms emerge as a manifestation of these neurobiological dysfunctions. For decades various research groups have teased out neurobiological abnormalities associated with schizophrenia. A recurring caveat to these endeavours is that it is dif®cult in case±control studies to ascertain the extent to which any abnormalities are a manifestation of the disease process or of factors correlating with the illness (eg. medication, smoking, substance abuse, poor diet, institutionalization). One way to examine the neurobiological impact of gene action without the risk of such confounds is to study unaffected ®rst-degree relatives of patients. First-degree relatives share 50% of a patient's gene sequence on average and (in the case of polygenic disorders, as schizophrenia most probably is) are likely to carry some quantity of susceptibility genes for illness to a greater extent than the general population, which can produce neurobiological abnormalities on the pathway to psychosis even though insuf®cient to cause the fullblown illness. As such abnormalities cannot be related to medication or any health effects of having a psychotic illness, xix

xx

PREFACE

they can be considered a manifestation of gene action, i.e. intermediate phenotypes or endophenotypes. The question therefore in pursuing this line of investigation becomes `What neurobiological abnormalities do the unaffected relatives of patients who are at high genetic liability for psychosis display?' This is the core theme that the Maudsley Family Study of Psychosis was developed to address. In this monograph we comprehensively describe the background and implementation of this major study and the key results from the principal domains investigated, with separate chapters devoted to structural neuroimaging, cognition, auditory evoked potentials, eye tracking and neurological signs. Throughout these chapters, we demonstrate how, in concert with similar studies elsewhere, this line of research can advance our understanding of the genetic underpinnings of psychotic illness. This multifaceted study was designed and implemented under the overall and ongoing supervision of Robin Murray. In addition to the authors of these chapters, many other researchers have contributed enormously to various aspects of this study since its inception in 1993, including the design, subject recruitment, data collection, data management, statistical analysis and dissemination of results. These researchers include Tonmoy Sharma, Thordur Sigmundsson, Eric Lancaster, Paul Birkett, Trevor Crawford, Mark Taylor, Heather King, Darren Mockler, Jessica Yakeley, Harvey Wickham, Anton Grech, Katja Schulze, Nicky Marshall, Ben Chapple, Helen Simon, Ed Bullmore, Pak Sham, Xavier Chitnis, Seema Quraishi, Dimitris Dikeos, Emma Dempster, David Collier and Muriel Walshe. We are grateful to the funding agencies which supported various aspects of the study including the Wellcome Trust, the Medical Research Council, the Psychiatry Research Trust, the Stanley Medical Research Institute and the National Alliance for Research on Schizophrenia and Depression. We are also particularly grateful to all the patients and their families who generously contributed their time and frequently travelled considerable distances to participate. It is their effort, more than any other, that has made this research possible and we hope that the improved understanding of the pathways from genetic risk to psychosis contributed to by this study will help in the quest to ®nd better ways to treat and prevent these destructive illnesses in the years ahead. This monograph will be of interest to academics within the ®elds of psychiatry, psychology and neuroscience as well as mental health professionals and lay readers who are interested in research advances into psychotic illness. The editor is very grateful to the other authors for their manuscripts and hopes that the reader will be informed and enthused by the contents of this book.

CHAPTER ONE

Exploring intermediate phenotypes of psychosis Colm McDonald, Robin Murray

THE CONTRIBUTION OF GENETIC LIABILITY TO SCHIZOPHRENIA The tendency for schizophrenia to run in families has long been recognized. A century ago, Emil Kraepelin, who described the syndrome of `dementia praecox', the forerunner of schizophrenia, noted: `I had myself found formerly in Heidelberg general hereditary predisposition to dementia praecox in about 70 per cent of the cases in which about this point reliable statements were to hand' (Kraepelin, 1919). Since Kraepelin's initial observations, overwhelming evidence has emerged from family, twin and adoption studies that schizophrenia has a strong genetic component.

Family studies In a seminal study pooling the results of about forty European family studies performed between 1920 and 1987, Gottesman (1991) demonstrated that the lifetime morbid risk of developing schizophrenia among the relatives of patients with schizophrenia increased with the degree of genetic relatedness to the affected individual. The risk to third-degree relatives was 2%, to second-degree relatives around 4±6%, to siblings or children around 9±13% and the risk to identical twins or the offspring of dual matings was 46±48%. Furthermore, the risk increased if more than one relative was affected, for example, the risk if both a sibling and parent are affected was 17%. The majority of the studies summarized by Gottesman (1991) were 1

2

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

performed before the advent of operationalized criteria for diagnosing schizophrenia. However, several studies that did employ operationalized diagnostic criteria have essentially con®rmed the results of these earlier studies (Gershon et al., 1988; Kendler et al., 1993a; Maier et al., 1993).

Twin studies Although family studies show clearly that schizophrenia is familial, they do not con®rm that this relationship is genetic, as family members also share much of their environment in common. The relative contributions of genetic and environmental risk factors can be disentangled by studying rates of concordance for the disorder in twins. If both members of a twin pair have schizophrenia, they are classi®ed as concordant, whereas if only one member of a twin pair is affected, they are classi®ed as discordant. A disorder is likely to be under genetic in¯uence if concordance rates are higher in monozygotic (MZ) twins, who share 100% of their genes, than dizygotic (DZ) twins, who share 50% of their genes on average. The results of all the major twin studies across several different countries ®nd the concordance rate for MZ twins to be substantially higher than that found for DZ twins. Kendler (1983) pooled the results of twelve such studies and found the probandwise concordance rate for MZ twins to be 53%, whereas the rate for DZ twins is 15%. Again, more recent studies using operationalized diagnostic criteria for schizophrenia con®rmed the markedly higher MZ than DZ probandwise concordance rates (McGuf®n et al., 1984; Onstad et al., 1991). Although these studies provide substantial evidence for a genetic contribution to schizophrenia, the lack of 100% concordance among monozygotic twins also provides compelling support for the importance of environmental factors in contributing to schizophrenia.

Adoption studies The relative contributions of genetic and environmental risk factors to the aetiology of schizophrenia can also be dissected by the use of adoption studies. In addition to shared genes, relatives (and in particular identical twins) also share an extensive environment in common, including social and cultural behaviours, biological hazards and psychological stresses, which could include common risk factors for schizophrenia. A series of naturalistic adoption studies have been performed in an attempt to minimize the effect of such common environmental risk factors in family studies of schizophrenia by comparing rates in biological and rates in non-biological relatives of patients with schizophrenia. These studies have found higher rates of schizophrenia in the adopted-away offspring of mothers of patients with schizophrenia compared to adopted-away offspring of controls (Heston, 1966; Rosenthal et al., 1975) and even compared to the rate in a small group

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

3

of subjects who were adopted by parents one of whom later developed schizophrenia (Wender et al., 1974). Furthermore, the rates of schizophrenia or schizophrenia-spectrum disorders among the biological relatives of adoptees with schizophrenia is higher than the rates in adoptive relatives and the relatives of control adoptees (Kety, 1983; Kety et al., 1994), including a higher rate of schizophrenia in paternal half siblings of adoptees with schizophrenia than in paternal half siblings of adoptees without schizophrenia, indicating that high rates of schizophrenia in the offspring of patients with schizophrenia are not related to prenatal or perinatal in¯uences (Kety, 1988). As with family and twin studies, the results of these early adoption studies have been con®rmed when more rigorous operationally de®ned diagnostic criteria are used to diagnose schizophrenia and its related disorders (Kendler et al., 1994; Tienari et al., 1994). Thus, the combined ®ndings from these adoption studies provide important support for the conclusions, derived from family and twin studies, that familial clustering of schizophrenia is an expression of shared genetic factors rather than shared environmental factors.

HERITABILITY ESTIMATES FOR SCHIZOPHRENIA Statistical models can be applied to data derived from twin studies to estimate the likely heritability of an illness ± the proportion of the variance in liability contributed to by genes. Estimates of heritability vary across samples and methods of ascertainment but usually involve a model that includes estimates of genetic effects, common environmental effects and non-shared environmental effects. Such estimates have ranged between 41% and 87% for schizophrenia (Cardno et al., 1999; Kendler, 1983), with the heritability estimates using operationally de®ned diagnoses tending to be in the upper end of this range (Farmer et al., 1987; Onstad et al., 1991). It has also been reported that the common environmental component can be removed from the model without weakening the ®t but increasing the heritability to 87%, with the remainder of the liability explained by nonshared environmental effects (Cardno et al., 1999; McGuf®n et al., 1994). In further support of these ®ndings, Cannon et al. (1998) used structural equation modelling in a population cohort of Finnish twins, thus excluding any bias associated with estimates of liability based on index twins, and demonstrated that 83% of the variance in liability was due to additive genetic factors, with the remaining 17% due to unique environmental factors.

MODE OF GENETIC TRANSMISSION Although results from family, twin and adoption studies provide evidence that genetic factors play a part in the aetiology of schizophrenia, the exact mechanisms of genetic transmission remain unidenti®ed. The simplest

4

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

model of genetic transmission is one in which a single gene is responsible for the illness. If such a model were true, genetic phenomena such as incomplete penetrance (where genetic variation is not always expressed in the clinical phenotype) and pleiotropy (where a single genetic alteration can result in variable phenotypic expression) are required to explain the apparent non-Mendelian pattern of transmission and heterogenous clinical presentation of these illnesses. However, studies on the recurrence risk from twin and family studies using statistical modelling have demonstrated that schizophrenia is very unlikely to be the product of a single gene or a collection of single gene disorders, even taking into account incomplete penetrance (McGue et al., 1985; O'Rourke et al., 1982). More complex models of genetic transmission involving multiple genes and environmental risk factors are more likely to be responsible for the patterns of inheritance observed. One such model is the liability/threshold model, which was ®rst applied to schizophrenia by Gottesman and Shields (1967). In this model, the liability to develop schizophrenia is normally distributed in the population and is due to multiple genes of small effect acting additively and in combination with environmental risk factors, but only those individuals whose liability exceeds a certain critical threshold manifest the illness. Relatives have an increased liability compared to the general population, and thus a higher proportion of relatives also lie beyond the threshold (re¯ected in the higher prevalence of schizophrenia in the relatives of patients). The number of genes involved in such a model is unpredictable and inheritance could be oligogenic ± a small number of genes of moderate effect (e.g. fewer than ten) or polygenic ± multiple genes of small effect (e.g. more than 100).

MOLECULAR GENETICS Linkage studies Linkage occurs when a genetic marker and a disease gene lie close to each other on the same chromosome. In this case, the marker and the disease gene will be found to occur together more often in affected family members than would be expected by chance. Linkage studies require families that contain several affected members and are most appropriately employed to detect a small number of genes of relatively large effect, i.e. mono/oligogenic inheritance rather than polygenic inheritance. Historically, the results of linkage studies in schizophrenia were disappointing and characterized by multiple failed replications, most likely contributed to by a combination of weak genetic effects and small sample sizes. In recent years, there has been some progress with successful replications of linkage to several chromosomal regions. This was facilitated by the development of highly polymorphic

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

5

genetic markers evenly spaced throughout the genome, which enabled genomewide scans for susceptibility loci, and by large-scale international collaborations to achieve greater statistical power through combining large numbers of subjects. Two meta-analyses in recent years (with different methodologies) of genomewide schizophrenia linkage studies have supported the existence of susceptibility genes on chromosomes 8p and 22q (Badner & Gershon, 2002; Lewis et al., 2003) with further strong support for loci at 13q (Badner & Gershon, 2002) and 2q (Lewis et al., 2003), and weaker support for loci at 1q, 3p, 5q, 6p, 11q, 14p and 20q (Lewis et al., 2003). More recent genomewide linkage studies in various populations continue to emerge at a rapid pace and implicate other chromosomal regions including 2q37 (Wijsman et al., 2003), 10q22 (Faraone et al., 2006), 11p (Suarez et al., 2006) and 18p11 (Faraone et al., 2005). Some of the chromosomal loci reported may well prove to be false positives in due course, but others are likely to harbour susceptibility genes.

Association studies Association studies require no major assumption other than the existence of a genetic contribution to the disorder. In contrast to linkage analysis, the aim is to examine the frequency of marker alleles in a sample of unrelated patients compared to a sample of ethnically matched controls. Association studies are more appropriately employed in the detection of polygenic inheritance, in which a large number of genes have relatively minor effects. A higher frequency of the marker allele in the patient group suggests that the marker allele is itself related to susceptibility to the disease or else is closely linked to the disease allele. A disadvantage is that the marker must be very tightly linked to the disease gene (<1 cM). Consequently, such studies have largely been con®ned to testing candidate genes thought to have functional signi®cance in the illness or to ®ne mapping of chromosomal loci previously identi®ed as showing linkage. There have been many negative studies and failed replications using the former technique, but some support has emerged for weak associations between schizophrenia and polymorphisms of certain genes controlling neurotransmission pathways, including the dopamine receptor DRD2 (Glatt et al., 2003b), dopamine receptor DRD3 (Jonsson et al., 2003), serotonergic receptor 5-HT2A (Abdolmaleky et al., 2004), metabotropic glutamate receptor GRM3 (Chen et al., 2005; Egan et al., 2004) and protein kinase AKT1 (Emamian et al., 2004; Schwab et al., 2005). Stronger support has been provided in recent years for genes conferring susceptibility to schizophrenia through ®ne mapping of loci identi®ed through linkage studies. The ®rst of these, neuregulin 1, was identi®ed after systematic study of the 8p22-p11 linkage site (Stefansson et al., 2002)

6

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

revealed that a multimarker haplotype in the 5© end of this gene was associated with schizophrenia; this ®nding has been strengthened by replications in different populations since (Tosato et al., 2005). Neuregulin 1 has an important role in neurodevelopment, synaptic plasticity, regulation of glutamate and other neurotransmitter receptor expression and so is a plausible candidate gene for schizophrenia. An at-risk haplotype within the gene dysbindin was identi®ed from ®ne mapping of the 6p22.3 locus (Straub et al., 2002) and the likelihood of the association of this gene with schizophrenia has been strengthened by successful replications (Williams et al., 2005). Dysbindin has many biological actions, but of particular relevance to schizophrenia is its expression in glutamatergic neurons and synapses of the hippocampus. Although neuregulin and dysbindin have the most powerful support as schizophrenia susceptibility genes from molecular genetics studies to date (Norton et al., 2006; Owen et al., 2004), several other genes are strongly implicated by association studies, including D-amino-acid oxidase inhibitor (DAOA) at 13q34 (Chumakov et al., 2002; Norton et al., 2006), regulator of G protein signalling 4 (RGS4 ) at 1q22 (Chowdari et al., 2002; Talkowski et al., 2006), disrupted in schizophrenia 1 (DISC1) at 1q42 (Callicott et al., 2005; Hennah et al., 2003) and proline dehydrogenase (PRODH ) at 22q11 (Li et al., 2004; Liu et al., 2002). The 22q11 site is of particular interest in schizophrenia genetics as, in addition to a locus of replicated linkage ®ndings, it is the site of a chromosomal deletion causing velocardiofacial syndrome, which is associated with schizophrenia in 24% of affected patients (Murphy et al., 1999). Considerable effort has also been devoted to the study of catechol-Omethyltransferase (COMT ) as a susceptibility gene for schizophrenia as it, too, is located at 22q11, has a key role in dopamine catabolism in the prefrontal cortex and has a functional variant ± a single nucleotide polymorphism that causes valine to methionine substitution at codon 158, the valine allele of which is known to increase enzyme activity. The valine allele has been associated with reduced performance in tests of frontal lobe function (Egan et al., 2001). In contrast to some early positive ®ndings, recent meta-analyses have not found signi®cant association between schizophrenia and this polymorphism (Glatt et al., 2003a; Munafo et al., 2005), although it remains possible that it may modify the phenotype.

OVERLAP IN GENETIC LIABILITY FOR SCHIZOPHRENIA AND BIPOLAR DISORDER The current distinction between schizophrenia and bipolar disorder ®rst emerged when Kraepelin (1899) divided psychotic illness into two diagnostic categories ± dementia praecox and manic depressive insanity ± on the basis of symptoms, course and outcome. This division has persisted for a century

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

7

and remains embedded in the major diagnostic systems in current use: the International Classi®cation of Diseases (ICD-10) and the Diagnostic and Statistical Manual of Mental Disorders version 4 (DSM-IV). Although the two illnesses appear clinically distinguishable in their pure forms, the boundary between these illnesses has always been blurred and there are no pathognomic symptoms on which clinicians can rely to differentiate them. Many patients with schizophrenia have symptoms of depression and mania; patients with bipolar disorder can have Schneiderian `®rst rank' symptoms of schizophrenia during illness exacerbation; and the diagnosis of `schizoaffective disorder' is necessary to categorize the substantial fraction [around 8% (Brockington & Leff, 1979)] of psychotic patients who cannot be classi®ed into either of the major psychotic branches. Another of Kraepelin's major distinctions between the illnesses regarding outcome (that schizophrenia ends in dementia and manic depression recovers) also turned out to be untrue. Schizophrenia has a relatively favourable outcome in up to half the cases diagnosed (Moller et al., 1982) and many cases of manic depression end in chronicity or have persistent residual affective symptoms between episodes (Angst, 2002). Like schizophrenia, bipolar disorder is a highly heritable syndrome. Heritability is estimated at between 73% and 87% from twin studies employing operationalized criteria for bipolar disorder, with the best-®tting model consisting of added genetic and non-shared environmental factors (Kendler et al., 1993c; McGuf®n et al., 2003). There is also accumulating evidence for an overlap in genetic predisposition to schizophrenia and bipolar disorder.

Family studies Family studies have consistently found higher rates of schizoaffective disorder, as well as unipolar depression, both in ®rst-degree relatives of patients with schizophrenia (Gershon et al., 1988; Maier et al., 1993) and in ®rst-degree relatives of patients with bipolar disorder (Gershon et al., 1982; Weissman et al., 1984). There is less evidence of direct co-aggregation of schizophrenia and bipolar disorder in families; for example, some studies ®nd an increased rate of schizophrenia in the relatives of bipolar-disorder patients (Angst et al., 1980; Tsuang et al., 1980) and others fail to ®nd such an increase (Gershon et al., 1982; Maier et al., 1993). However, this failure to ®nd signi®cant co-aggregation may be due to lack of statistical power in some studies. In a review of three independent genetic studies on schizophrenia (the Danish Adoption Study, Iowa 500 Family Study and Roscommon Family Study), Kendler and Gardner (1997) found that the odds ratio of unipolar depression and bipolar disorder in the relatives of patients with schizophrenia were 1.3 and 1.9, respectively, although the latter

8

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

failed to reach statistical signi®cance. In a large population-based study of 2299 individuals with bipolar disorder, Mortensen and colleagues (Mortensen et al., 2003) reported that having a parent or sibling with schizophrenia (and excluding schizoaffective disorder) increased the risk for bipolar disorder by a factor of 3.7±5.7. Maier and colleagues (Maier et al., 2002) examined rates of psychotic and affective disorder in extended pedigrees and found that the prevalence rate of bipolar disorder in ®rstdegree relatives of ninety-one patients with schizophrenia was signi®cantly increased (at 3.9%) compared with equivalent relatives of controls (0.9%). Furthermore, the evidence for direct co-aggregation becomes more substantial when data are con®ned to the more severe and psychotic forms of bipolar disorder rather than mild bipolar disorder. Hence in the Roscommon Family Study, Kendler and colleagues found an increased rate of psychotic affective illness among the relatives of patients with schizophrenia, but no increased rate of non-psychotic bipolar disorder (Kendler et al., 1993a, 1993b). In the New York High Risk project, there was an increased risk of schizophrenia-related psychoses in the offspring of affective disorder subjects; and offspring of parents with schizophrenia had an increased risk of psychotic but not non-psychotic bipolar disorder (Erlenmeyer-Kimling et al., 1997). Valles and colleagues report an increased risk of schizophrenia as well as bipolar disorder in the relatives of a sample of patients with severe bipolar disorder and a high rate of psychotic symptoms (Valles et al., 2000). Interestingly, within bipolar disorder families there is evidence for aggregation of psychotic symptoms (Potash et al., 2001, 2003a; Schurhoff et al., 2003), which is also consistent with the hypothesis that a tendency to psychosis is genetically transmitted, and Potash and colleagues (Potash et al., 2001, 2003a) suggest that it is in such psychotic bipolar-disorder families that researchers should seek genes predisposing to both schizophrenia and bipolar disorder.

Twin studies There have been case reports of identical twins (Dalby et al., 1986) and even triplets (McGuf®n et al., 1982) containing individuals affected with both schizophrenia and bipolar disorder. In an interesting re-analysis of the Maudsley twin series, Cardno and colleagues (Cardno et al., 2002) also found evidence for shared genetic susceptibility to schizophrenia, mania and schizoaffective disorder. Other quantitative genetic studies had used a hierarchical approach to diagnosis, such that if a subject had ever ful®lled criteria for schizophrenia, this `trumped' any previous diagnosis of mania and they were given schizophrenia as their lifetime diagnosis. In this study, the authors abandoned the hierarchical approach and allowed subjects to ful®ll criteria for more than one diagnosis in their lifetime. There was

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

9

signi®cant co-morbidity for both schizophrenic and manic syndromes within individual probands and the within-pair correlations for these syndromes were more than twice as great in MZ than in same-sex DZ twin pairs, indicating a genetic contribution to this co-morbidity. Subsequent independent pathway model-®tting demonstrated signi®cant genetic correlations between the manic and schizophrenic syndromes, which also both shared genetic liability with schizoaffective disorder. Of note, 96% of the manic syndrome patients in the study had had psychotic symptoms, so this study could not answer the question as to whether less severe forms of mania or hypomania share genetic liability with schizophrenia; but, again, it underlines a likely overlap in genetic liability between schizophrenia and those bipolar subjects with severe mania and psychotic symptoms. The authors conclude that some genes predispose to psychosis in general, and that more speci®c genes contribute to the exact form of psychosis.

Molecular genetics The meta-analysis of whole genome linkage studies by Badner and colleagues (Badner & Gershon, 2002) examined bipolar disorder as well as schizophrenia and identi®ed 13q32 and 22q11 as sites containing susceptibility genes for both schizophrenia and bipolar disorder. Berrettini (2000, 2003) examined overlapping linkage sites for both disorders by investigating whether, for con®rmed linkage sites for schizophrenia, there has been at least one linkage ®nding reported in bipolar disorder, and vice versa. Using this approach, he identi®ed ®ve linkage regions that might harbour genes for both disorders: 18p11.2, 13q32, 22q11, 8p22 and 10p14. Furthermore, a linkage study that con®ned analysis to subjects from severe psychotic bipolar-disorder families found evidence for linkage at 13q31 and 22q12 (Potash et al., 2003b) and a linkage study that examined families on the basis of having schizoaffective disorder, bipolar subtype (Hamshere et al., 2005), demonstrated linkage at 1q42, a strong schizophrenia site, and suggestive linkage at 22q11 and 19p13. Several association studies of bipolar disorder have examined genes implicated in schizophrenia, in particular where linkage studies suggest that genes at these loci confer susceptibility for both syndromes. The most replicated of these is the reported association between haplotypes of the DAOA/G30 locus at 13q and bipolar disorder (Detera-Wadleigh & McMahon, 2006; Hattori et al., 2003; Williams et al., 2006). Other association studies have linked bipolar disorder or psychotic subtypes of bipolar disorder to neuregulin 1, dysbindin and DISC1 (Breen et al., 2006; Craddock et al., 2006). The evidence to date from linkage and association studies is consistent with genetic epidemiological studies in indicating that schizophrenia and

10

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

bipolar disorder are caused by several genes of small effect, some of which may be common to the two illnesses and some of which confer speci®c susceptibility. Although clear progress has been made in the molecular genetics of psychosis in recent years, it remains incremental ± unambiguous functional variants that could drive the neurobiological processes towards psychosis have not yet been identi®ed for the proposed susceptibility genes. Exploring the impact of allelic variation in postulated susceptibility genes on neurobiological processes and identifying alternative phenotypes that might be intermediate between genetic variation and the classic clinical phenotype are now key areas of research to illuminate the complex mechanisms by which genetic variation confers liability for psychotic illness.

ENDOPHENOTYPES The pathway from genetic sequence variation to the behavioural phenotype of schizophrenia is lengthy and complex. Genes encode for protein molecules and it is most likely that genetic sequence variation underlying schizophrenia is associated with alteration of these gene products, which then alter cell biology. These cellular changes interfere with the development and functioning of neural circuits at a systems level and from such dysfunctions the behavioural symptoms and signs by which we diagnose the schizophrenic syndrome emerge. These pathways are further complicated by genetic and allelic heterogeneity, interactions of susceptibility genes with other gene variants (epistatic effects) and environmental risk factors (gene± environment interactions), and by alterations in gene expression due to other heritable factors such as DNA methylation (epigenetic effects). Unlike other polygenic multifactorial diseases, like diabetes and coronary heart disease, genetic research into schizophrenia suffers from the added complexities of having a phenotype that is entirely clinically based, with as yet no validating biological test (even at post-mortem), and environmental risk factors that are unknown or of minor effect. Furthermore, psychotic illness involves the complex organ of the human brain: unlike organs with a small number of common cellular types, neurons are frequently distinct from one another in their cellular processes and in their local and regional networks. Given these complexities, and the probability of continuous genetic liability within the population, the use of the qualitative diagnostic phenotype (affected/unaffected) is likely to be underpowered as the sole phenotype in genetic studies. Considerable interest has therefore focused on identifying genetically valid traits that are more homogenous and can be measured quantitatively. Candidates for such traits include subsets of the clinical syndrome, for example, selected by symptom clusters or course (`subphenotypes'), and markers of neurobiological or cognitive dysfunction (`intermediate phenotypes' or `endophenotypes').

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

11

Endophenotypes are objectively measured biological abnormalities that represent more proximal effects of susceptibility genes than the clinical phenotype (Gottesman & Gould, 2003; Wickham & Murray, 1997). It is assumed that because they are more proximal to genetic sequence variation, such traits will be modelled by a less complex genetic architecture than the diagnostic phenotype, and some may even follow simple Mendelian patterns of inheritance. The exploitation of endophenotypes has helped to elucidate genetically driven biological processes for a number of medical conditions. For example, haemochromatosis is a polygenic disorder of iron metabolism that can lead to iron overload and liver cirrhosis. Elevated levels of serum transferrin saturation (the iron-binding protein in the blood) were identi®ed as endophenotypic for haemochromatosis (Pietrangelo, 2004). Allelic variation in the haemochromatosis gene (HFE ) was linked to transferrin saturation. However, not all those with the susceptibility allele, manifest by high serum transferrin saturation, develop clinical haemachromatosis, indicating that it is a necessary but not suf®cient cause of the illness (Beutler, 2003). The more simple genetic architecture linking HFE allelic variation with the endophenotype of serum transferrin saturation enabled the dissection of one genetically driven biological mechanism from other genetic and environmental factors contributing to the illness (Beutler, 2003). Endophenotypes should ideally ful®ll six criteria to be useful in genetic studies (Gershon & Goldin, 1986; Leboyer et al., 1998). They should: 1. 2. 3. 4. 5. 6.

be heritable themselves be associated with the illness in the general population be state independent, i.e. be manifest whether or not the illness is active co-segregate with the illness within families, i.e. the illness is more prevalent among relatives who manifest the marker than among those relatives who do not be measurable in both affected and unaffected subjects be found more frequently among the biological relatives of patients than in healthy controls.

Some of these criteria may be overly stringent for psychotic disorders, hence the high heritability criterion for an endophenotype can be interpreted less rigorously in the light of known environmental factors contributing to its variation (Berrettini, 2005). For example, abnormalities of the P50 auditory evoked potential wave are potentially endophenotypic for schizophrenia but are also know to be normalized by smoking and a higher proportion of patients with schizophrenia smoke. Also, the diagnostic speci®city of endophenotypes is less relevant for psychotic syndromes de®ned by their clinical symptoms than for biological tests, since, as

12

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

pointed out above for schizophrenia and bipolar disorder, there may be considerable overlap between these clinical syndromes and thus overlapping endophenotypes underlying them (Berrettini, 2005). A key criterion for an endophenotype is that it must be found to a greater extent among individuals at higher genetic liability for the illness than control populations. The standard study design for eliciting this is to assess ®rst-degree relatives of patients with the illness. Those siblings who are unaffected twins in MZ pairs discordant for schizophrenia represent the highest genetic risk for schizophrenia than any other class of sibling as they share their entire gene sequence with the affected patient, and some study designs exploring endophenotypes focus on such individuals, comparing biological measures in such individuals to unaffected twin pairs or along a gradient of genetic liability including DZ twins, i.e. unaffected twin from discordant MZ pairs > unaffected twin from discordant DZ twin pairs > healthy twins. Given the dif®culties in ascertaining large samples of twins, a more widely applied design is to examine the unaffected ®rst-degree relatives of patients. These can be siblings or parents, who would largely have lived through the risk period for illness, or younger relatives (children, adolescent siblings of patients), a proportion of whom will develop the illness and may demonstrate neurobiological abnormalities characteristic of the prodromal or premorbid phase of illness on the pathway to frank psychosis. Putative endophenotypic markers for schizophrenia, which are reported to be present to a greater extent in unaffected ®rst-degree relatives of patients than in controls, include abnormalities of the auditory evoked response such as the P300 and P50 waves (Heinrichs, 2004; Weisbrod et al., 1999); oculomotor dysfunction, such as on smooth pursuit and antisaccade tasks (Holzman, 2000); neuropsychological de®cits (Cannon et al., 1994; Kremen et al., 1994) and abnormalities of brain structure, such as ventricular enlargement and medial temporal lobe volume de®cits (Seidman et al., 1999; Staal et al., 2000).

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS The Maudsley Family Study of Psychosis was initiated in 1993 to implement the endophenotypic approach to schizophrenia by examining neurobiological and cognitive abnormalities not just in patients but also in unaffected adult relatives (siblings and parents) who carry increased genetic liability for illness. The major domains of investigation are neuroimaging, neuropsychology, auditory evoked potentials, eye tracking and clinical assessment of neurological signs. Some other studies searching for endophenotypes of schizophrenia have considered all unaffected relatives, regardless of family history, as a single

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

13

group of participants at similar likelihood of expressing genetically driven neurobiological abnormalities. In the Maudsley Family Study of Psychosis, we considered that those families multiply affected with schizophrenia would be more likely to have a higher load of schizophrenia susceptibility genes, and would be more likely to demonstrate intermediate phenotypes associated with such genes, than families in which only one member had developed the illness. We thus attempted to broaden the likely genetic load by recruiting families with a varying density of illness. Hence some families were multiply affected (`familial'), i.e. had at least two members among ®rst- and/or second-degree relatives affected, whereas others were singly affected (`non-familial'), i.e. only one known member of the family was affected with schizophrenia. The familial/non-familial (or sporadic) distinction among patients has been proposed as a study design to separate out genetic and environmental risk factors for schizophrenia (Murray et al., 1985). It has had variable success, as such patients often display similar abnormalities (Roy & Crowe, 1994). It is important to stress that this was not the intention of the present study of unaffected relatives of patients, as relatives of patients from singly affected families were still expected to demonstrate biological abnormalities resulting from susceptibility genes for illness; rather, we presumed that any such endophenotypic effects would be more prominent among relatives from multiply affected families. Among a small number of the `familial' families, further variation in genetic risk was sought by speci®cally seeking families where transmission appeared unilinear, i.e. a family history of schizophrenia was present in at least one ®rst-degree relative of only one of the parents of the index patient, while the other parent's family was unaffected by psychotic illness as far as could be ascertained. In these families, the parent who appeared to be transmitting genetic liability from sibling/ parent to child was hypothesized to be at particularly high likelihood of carrying susceptibility genes for illness and described as a `presumed obligate carrier' of genetic liability for schizophrenia. The number of parents recruited for different sections of the study who were presumed obligate carriers of genetic liability was small and, in general, analyses were performed on larger samples of relatives in the ®rst instance. Even with this varying level of presumed genetic loading, considerable variation still existed in the density of illness among familial families and some analyses were also performed modelling this likely variation quantitatively using a continuous `genetic liability scale'. The derivation of this scale is described in detail in Chapter 2 and analyses of its relationship with cognitive function and brain structure presented in later chapters. Given the apparent clinical and genetic overlap between schizophrenia and bipolar disorder described above, the clinical phenotype has also been broadened in recent years to include families multiply affected with psychotic bipolar 1 disorder.

14

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

The purpose of this monograph is to: (1) bring together key ®ndings, selected for highlighting by the authors, from each of the main arms of the Maudsley Family Study of Psychosis into a single volume; (2) provide a comprehensive description of the background and clinical details of the entire sample of participants; (3) present major unpublished data analyses, including smooth pursuit eye tracking results, memory function and regionof-interest neuroimaging data combined across the entire sample of families affected with schizophrenia; and (4) draw together ®ndings from the different arms of the study. Given the size and duration of this project and the large number of analyses performed, the monograph does not attempt to cover all the published ®ndings (a list of the forty-three data-based peer-reviewed publications to date from the study is provided in Appendix 1 for interested readers). Instead, key analyses and results from the arms of the study are highlighted, with a synthesis of these ®ndings in the closing chapter. In each chapter, neurobiological abnormalities detected in patients with schizophrenia when compared with controls are presented. For the neuroimaging analyses, abnormalities detected in patients with psychotic bipolar disorder compared with controls and compared with schizophrenia are also presented with a discussion of how such imaging abnormalities relate to the Kraepelinian distinction between these syndromes. In each chapter, analyses of neurobiological abnormalities in unaffected relatives of patients with schizophrenia (and for neuroimaging with bipolar disorder) are presented, with discussion of which abnormalities are supported by the study as potential intermediate phenotypes of illness. Researchers operating in the various arms of the study have used slightly differing study designs to explore abnormalities in unaffected relatives of patients, while taking into consideration the impact of higher presumed genetic loading on neurobiological abnormalities by virtue of a stronger family history. Hence, each section has considered the familial/non-familial distinction, and some subsections (cognition, neuroimaging) have also speci®cally examined abnormalities in those parents of patients who are presumed obligate carriers of genetic risk for illness or have utilized the quantitative genetic liability scale to model varying genetic liability. The last chapter ®nishes with a summary of key ®ndings from the monograph and a description of future analyses and studies that will be performed to further advance the quest for intermediate phenotypes of psychosis.

REFERENCES Abdolmaleky, H.M., Faraone, S.V., Glatt, S.J., Tsuang, M.T. (2004). Meta-analysis of association between the T102C polymorphism of the 5HT2a receptor gene and schizophrenia. Schizophrenia Research, 67, 53±62. Angst, J. (2002). Historical aspects of the dichotomy between manic-depressive disorders and schizophrenia. Schizophrenia Research, 57, 5±13.

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

15

Angst, J., Frey, R., Lohmeyer, B., Zerbin-Rudin, E. (1980). Bipolar manic-depressive psychoses: results of a genetic investigation. Human Genetics, 55, 237±254. Badner, J.A., Gershon, E.S. (2002). Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia. Molecular Psychiatry, 7, 405±411. Berrettini, W.H. (2000). Are schizophrenic and bipolar disorders related? A review of family and molecular studies. Biological Psychiatry, 48, 531±538. Berrettini, W.H. (2003). Bipolar disorder and schizophrenia: not so distant relatives? World Psychiatry, 2, 68±72. Berrettini, W.H. (2005). Genetic bases for endophenotypes in psychiatric disorders. Dialogues in Clinical Neuroscience, 7, 95±101. Beutler, E. (2003). The HFE Cys282Tyr mutation as a necessary but not suf®cient cause of clinical hereditary hemochromatosis. Blood, 101, 3347±3350. Breen, G., Prata, D., Osborne, S., Munro, J., Sinclair, M., Li, T., et al. (2006). Association of the dysbindin gene with bipolar affective disorder. American Journal of Psychiatry, 163, 1636±1638. Brockington, I.F., Leff, J.P. (1979). Schizo-affective psychosis: De®nitions and incidence. Psychological Medicine, 9, 91±99. Callicott, J.H., Straub, R.E., Pezawas, L., Egan, M.F., Mattay, V.S., Hariri, A.R., et al. (2005). Variation in DISC1 affects hippocampal structure and function and increases risk for schizophrenia. Proceedings of the National Academy of Sciences USA, 102, 8627±8632. Cannon, T.D., Zorrilla, L.E., Shtasel, D., Gur, R.E., Gur, R.C., Marco, E.J., et al. (1994). Neuropsychological functioning in siblings discordant for schizophrenia and healthy volunteers. Archives of General Psychiatry, 51, 651±661. Cannon, T.D., Kaprio, J., Lonnqvist, J., Huttunen, M., Koskenvuo, M. (1998). The genetic epidemiology of schizophrenia in a Finnish twin cohort. A population-based modeling study. Archives of General Psychiatry, 55, 67±74. Cardno, A.G., Marshall, E.J., Coid, B., Macdonald, A.M., Ribchester, T.R., Davies, N.J., et al. (1999). Heritability estimates for psychotic disorders: the Maudsley twin psychosis series. Archives of General Psychiatry, 56, 162±168. Cardno, A.G., Rijsdijk, F.V., Sham, P.C., Murray, R.M., McGuf®n, P. (2002). A twin study of genetic relationships between psychotic symptoms. American Journal of Psychiatry, 159, 539±545. Chen, Q., He, G., Chen, Q., Wu, S., Xu, Y., Feng, G., et al. (2005). A case±control study of the relationship between the metabotropic glutamate receptor 3 gene and schizophrenia in the Chinese population. Schizophrenia Research, 73, 21±26. Chowdari, K.V., Mirnics, K., Semwal, P., Wood, J., Lawrence, E., Bhatia, T., et al. (2002). Association and linkage analyses of RGS4 polymorphisms in schizophrenia. Human and Molecular Genetics, 11, 1373±1380. Chumakov, I., Blumenfeld, M., Guerassimenko, O., Cavarec, L., Palicio, M., Abderrahim, H., et al. (2002). Genetic and physiological data implicating the new human gene G72 and the gene for D-amino acid oxidase in schizophrenia. Proceedings of the National Academy of Sciences USA, 99, 13675±13680. Craddock, N., O'Donovan, M.C., Owen, M.J. (2006). Genes for schizophrenia and bipolar disorder? Implications for psychiatric nosology. Schizophrenia Bulletin, 9±16. Dalby, J.T., Morgan, D., Lee, M.L. (1986). Schizophrenia and mania in identical twin brothers. Journal of Nervous and Mental Diseases, 174, 304±308. Detera-Wadleigh, S.D., McMahon, F.J. (2006). G72/G30 in schizophrenia and bipolar disorder: Review and meta-analysis. Biological Psychiatry, 60, 106±114. Egan, M.F., Goldberg, T.E., Kolachana, B.S., Callicott, J.H., Mazzanti, C.M., Straub, R.E., et al. (2001). Effect of COMT Val108/158Met genotype on frontal lobe function and risk for schizophrenia. Proceedings of the National Academy of Sciences USA, 98, 6917±6922.

16

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Egan, M.F., Straub, R.E., Goldberg, T.E., Yakub, I., Callicott, J.H., Hariri, A.R., et al. (2004). Variation in GRM3 affects cognition, prefrontal glutamate, and risk for schizophrenia. Proceedings of the National Academy of Sciences USA, 101, 12604±12609. Emamian, E.S., Hall, D., Birnbaum, M.J., Karayiorgou, M., Gogos, J.A. (2004). Convergent evidence for impaired AKT1-GSK3beta signaling in schizophrenia. Natural Genetics, 36, 131±137. Erlenmeyer-Kimling, L., Adamo, U.H., Rock, D., Roberts, S.A., Bassett, A.S., SquiresWheeler, E., et al. (1997). The New York High-Risk Project. Prevalence and comorbidity of axis I disorders in offspring of schizophrenic parents at 25-year follow-up. Archives of General Psychiatry, 54, 1096±1102. Faraone, S.V., Skol, A.D., Tsuang, D.W., Young, K.A., Haverstock, S.L., Prabhudesai, S., et al. (2005). Genome scan of schizophrenia families in a large Veterans Affairs Cooperative Study sample: evidence for linkage to 18p11.32 and for racial heterogeneity on chromosomes 6 and 14. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics, 139, 91±100. Faraone, S.V., Hwu, H.G., Liu, C.M., Chen, W.J., Tsuang, M.M., Liu, S.K., et al. (2006). Genome scan of Han Chinese schizophrenia families from Taiwan: Con®rmation of linkage to 10q22.3. American Journal of Psychiatry, 163, 1760±1766. Farmer, A.E., McGuf®n, P., Gottesman, I.I. (1987). Twin concordance for DSM-III schizophrenia. Scrutinizing the validity of the de®nition. Archives of General Psychiatry, 44, 634±641. Gershon, E.S., Hamovit, J., Guroff, J.J., Dibble, E., Leckman, J.F., Sceery, W., et al. (1982). A family study of schizoaffective, bipolar I, bipolar II, unipolar, and normal control probands. Archives of General Psychiatry, 39, 1157±1167. Gershon, E.S., Goldin, L.R. (1986). Clinical methods in psychiatric genetics. I. Robustness of genetic marker investigative strategies. Acta Psychiatrica Scandinavica, 74, 113±118. Gershon, E.S., DeLisi, L.E., Hamovit, J., Nurnberger, J.I. Jr., Maxwell, M.E., Schreiber, J., et al. (1988). A controlled family study of chronic psychoses. Schizophrenia and schizoaffective disorder. Archives of General Psychiatry, 45, 328±336. Glatt, S.J., Faraone, S.V., Tsuang, M.T. (2003a). Association between a functional catechol Omethyltransferase gene polymorphism and schizophrenia: meta-analysis of case±control and family-based studies. American Journal of Psychiatry, 160, 469±476. Glatt, S.J., Faraone, S.V., Tsuang, M.T. (2003b). Meta-analysis identi®es an association between the dopamine D2 receptor gene and schizophrenia. Molecular Psychiatry, 8, 911±915. Gottesman, I.I. (1991). Schizophrenia Genesis: The Origins of Madness. New York: H. Freeman & Co. Gottesman, I.I., Gould, T.D. (2003). The endophenotype concept in psychiatry: Etymology and strategic intentions. American Journal of Psychiatry, 160, 636±645. Gottesman, I.I., Shields, J. (1967). A polygenic theory of schizophrenia. Proceedings of the National Academy of Sciences USA, 58, 199±205. Hamshere, M.L., Bennett, P., Williams, N., Segurado, R., Cardno, A., Norton, N., et al. (2005). Genomewide linkage scan in schizoaffective disorder: Signi®cant evidence for linkage at 1q42 close to DISC1, and suggestive evidence at 22q11 and 19p13. Archives of General Psychiatry, 62, 1081±1088. Hattori, E., Liu, C., Badner, J.A., Bonner, T.I., Christian, S.L., Maheshwari, M., et al. (2003). Polymorphisms at the G72/G30 gene locus, on 13q33, are associated with bipolar disorder in two independent pedigree series. American Journal of Human Genetics, 72, 1131±1140. Heinrichs, R.W. (2004). Meta-analysis and the science of schizophrenia: Variant evidence or evidence of variants? Neuroscience and Behavioral Reviews, 28, 379±394. Hennah, W., Varilo, T., Kestila, M., Paunio, T., Arajarvi, R., Haukka, J., et al. (2003).

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

17

Haplotype transmission analysis provides evidence of association for DISC1 to schizophrenia and suggests sex-dependent effects. Human Molecular Genetics, 12, 3151±3159. Heston, L.L. (1966). Psychiatric disorders in foster home reared children of schizophrenic mothers. British Journal of Psychiatry, 112, 819±825. Holzman, P.S. (2000). Eye movements and the search for the essence of schizophrenia. Brain Research, Brain Research Reviews, 31, 350±356. Jonsson, E.G., Flyckt, L., Burgert, E., Crocq, M.A., Forslund, K., Mattila-Evenden, M., et al. (2003). Dopamine D3 receptor gene Ser9Gly variant and schizophrenia: Association study and meta-analysis. Psychiatric Genetics, 13, 1±12. Kendler, K.S. (1983). Overview: A current perspective on twin studies of schizophrenia. American Journal of Psychiatry, 140, 1413±1425. Kendler, K.S., Gardner, C.O. (1997). The risk for psychiatric disorders in relatives of schizophrenic and control probands: A comparison of three independent studies. Psychological Medicine, 27, 411±419. Kendler, K.S., McGuire, M., Gruenberg, A.M., O'Hare, A., Spellman, M., Walsh, D. (1993a). The Roscommon Family Study I. Methods, diagnosis of probands, and risk of schizophrenia in relatives. Archives of General Psychiatry, 50, 527±540. Kendler, K.S., McGuire, M., Gruenberg, A.M., O'Hare, A., Spellman, M., Walsh, D. (1993b). The Roscommon Family Study IV. Affective illness, anxiety disorders, and alcoholism in relatives. Archives of General Psychiatry, 50, 952±960. Kendler, K.S., Pedersen, N., Johnson, L., Neale, M.C., Mathe, A.A. (1993c). A pilot Swedish twin study of affective illness, including hospital- and population-ascertained subsamples. Archives of General Psychiatry, 50, 699±700. Kendler, K.S., Gruenberg, A.M., Kinney, D.K. (1994). Independent diagnoses of adoptees and relatives as de®ned by DSM-III in the provincial and national samples of the Danish Adoption Study of Schizophrenia. Archives of General Psychiatry, 51, 456±468. Kety, S.S. (1983). Mental illness in the biological and adoptive relatives of schizophrenic adoptees: Findings relevant to genetic and environmental factors in etiology. American Journal of Psychiatry, 140, 720±727. Kety, S.S. (1988). Schizophrenic illness in the families of schizophrenic adoptees: Findings from the Danish national sample. Schizophrenia Bulletin, 14, 217±222. Kety, S.S., Wender, P.H., Jacobsen, B., Ingraham, L.J., Jansson, L., Faber, B., et al. (1994). Mental illness in the biological and adoptive relatives of schizophrenic adoptees. Replication of the Copenhagen Study in the rest of Denmark. Archives of General Psychiatry, 51, 442±455. Kraepelin, E. (1899). Psychiatrie, ein Lehrbuch fuÈr Studierende und AÈrtze., 6th ed. Leipzig: Barth. Kraepelin, E. (1919). Dementia praecox and paraphrenia (R.M. Barclay, transl) (p. 232). Edinburgh: E&S Livingstone. Kremen, W.S., Seidman, L.J., Pepple, J.R., Lyons, M.J., Tsuang, M.T., Faraone, S.V. (1994). Neuropsychological risk indicators for schizophrenia: A review of family studies. Schizophrenia Bulletin, 20, 103±119. Leboyer, M., Bellivier, F., Nosten-Bertrand, M., Jouvent, R., Pauls, D., Mallet, J. (1998). Psychiatric genetics: Search for phenotypes. Trends in Neuroscience, 21, 102±105. Lewis, C.M., Levinson, D.F., Wise, L.H., DeLisi, L.E., Straub, R.E., Hovatta, I., et al. (2003). Genome scan meta-analysis of schizophrenia and bipolar disorder, part II: Schizophrenia. American Journal of Human Genetics, 73, 34-48. Li, T., Ma, X., Sham, P.C., Sun, X., Hu, X., Wang, Q., et al. (2004). Evidence for association between novel polymorphisms in the PRODH gene and schizophrenia in a Chinese

18

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

population. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics, 129, 13±15. Liu, H., Heath, S.C., Sobin, C., Roos, J.L., Galke, B.L., Blundell, M.L., et al. (2002). Genetic variation at the 22q11 PRODH2/DGCR6 locus presents an unusual pattern and increases susceptibility to schizophrenia. Proceedings of the National Academy of Sciences USA, 99, 3717±3722. Maier, W., Lichtermann, D., Minges, J., Hallmayer, J., Heun, R., Benkert, O., et al. (1993). Continuity and discontinuity of affective disorders and schizophrenia. Results of a controlled family study. Archives of General Psychiatry, 50, 871±883. Maier, W., Lichtermann, D., Franke, P., Heun, R., Falkai, P., Rietschel, M. (2002). The dichotomy of schizophrenia and affective disorders in extended pedigrees. Schizophrenia Research, 57, 259±266. McGue, M., Gottesman, I.I., Rao, D.C. (1985). Resolving genetic models for the transmission of schizophrenia. Genetic Epidemiology, 2, 99±110. McGuf®n, P., Reveley, A., Holland, A. (1982). Identical triplets: Non-identical psychosis? British Journal of Psychiatry, 140, 1±6. McGuf®n, P., Farmer, A.E., Gottesman, I.I., Murray, R.M., Reveley, A.M., et al. (1984). Twin concordance for operationally de®ned schizophrenia. Con®rmation of familiality and heritability. Archives of General Psychiatry, 41, 541±545. McGuf®n, P., Asherson, P., Owen, M., Farmer, A. (1994). The strength of the genetic effect. Is there room for an environmental in¯uence in the aetiology of schizophrenia? British Journal of Psychiatry 164, 593±599. McGuf®n, P., Rijsdijk, F., Andrew, M., Sham, P., Katz, R., Cardno, A. (2003). The heritability of bipolar affective disorder and the genetic relationship to unipolar depression. Archives of General Psychiatry, 60, 497±502. Moller, H.J., von Zerssen, D., Werner-Eilert, K., Wuschner-Stockheim, M. (1982). Outcome in schizophrenic and similar paranoid psychoses. Schizophrenia Bulletin, 8, 99±108. Mortensen, P.B., Pedersen, C.B., Melbye, M., Mors, O., Ewald, H. (2003). Individual and familial risk factors for bipolar affective disorders in Denmark. Archives of General Psychiatry, 60, 1209±1215. Munafo, M.R., Bowes, L., Clark, T.G., Flint, J. (2005). Lack of association of the COMT (Val158/108 Met) gene and schizophrenia: A meta-analysis of case±control studies. Molecular Psychiatry, 10, 765±770. Murphy, K.C., Jones, L.A., Owen, M.J. (1999). High rates of schizophrenia in adults with velo-cardio-facial syndrome. Archives of General Psychiatry, 56, 940±945. Murray, R.M., Lewis, S.W., Reveley, A.M. (1985). Towards an aetiological classi®cation of schizophrenia. Lancet, 1, 1023±1026. Norton, N., Williams, H.J., Owen, M.J. (2006). An update on the genetics of schizophrenia. Current Opinion in Psychiatry, 19, 158±164. O'Rourke, D.H., Gottesman, I.I., Suarez, B.K. (1982). Refutation of the general single-locus model for the etiology of schizophrenia. American Journal of Human Genetics, 34, 630±649. Onstad, S., Skre, I., Torgersen, S., Kringlen, E. (1991). Twin concordance for DSM-III-R schizophrenia. Acta Psychiatrica Scandinavica, 83, 395±401. Owen, M.J., Williams, N.M., O'Donovan, M.C. (2004). The molecular genetics of schizophrenia: new ®ndings promise new insights. Molecular Psychiatry, 9, 14±27. Pietrangelo, A. (2004). Hereditary hemochromatosis ± a new look at an old disease. New England Journal of Medicine, 350, 2383±2397. Potash, J.B., Willour, V.L., Chiu, Y.F., Simpson, S.G., MacKinnan, D.F., Pearlson, G.D., et al. (2001). The familial aggregation of psychotic symptoms in bipolar disorder pedigrees. American Journal of Psychiatry, 158, 1258±1264. Potash, J.B., Chiu, Y.F., MacKinnon, D.F., Miller, E.B., Simpson, S.G., McMahon, F.J., et

1. INTERMEDIATE PHENOTYPES OF PSYCHOSIS

19

al. (2003a). Familial aggregation of psychotic symptoms in a replication set of 69 bipolar disorder pedigrees. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics, 116, 90±97. Potash, J.B., Zandi, P.P., Willour, V.L., Lan, T.H., Huo, Y., Avramopoulos, D., et al. (2003b). Suggestive linkage to chromosomal regions 13q31 and 22q12 in families with psychotic bipolar disorder. American Journal of Psychiatry, 160, 680±686. Rosenthal, D., Wender, P.H., Kety, S.S., Schulsinger, S.S., Welner, J., Rieder, R.O. (1975). Parent±child relationships and psychopathological disorder in the child. Archives of General Psychiatry, 32, 466±476. Roy, M.A., Crowe, R. (1994). Validity of the familial and sporadic subtypes of schizophrenia. American Journal of Psychiatry, 151, 805±814. Schurhoff, F., Szoke, A., Meary, A., Bellivier, F., Rouillon, F., Pauls, D., et al. (2003). Familial aggregation of delusional proneness in schizophrenia and bipolar pedigrees. Americal Journal of Psychiatry, 160, 1313±1319. Schwab, S.G., Hoefgen, B., Hanses, C., Hassenbach, M.B., Albus, M., Lerer, B., et al. (2005). Further evidence for association of variants in the AKT1 gene with schizophrenia in a sample of European sib-pair families. Biological Psychiatry, 58, 446±450. Seidman, L.J., Faraone, S.V., Goldstein, J.M., Goodman, J.M., Kremen, W.S., Toomey, R., et al. (1999). Thalamic and amygdala-hippocampal volume reductions in ®rst-degree relatives of patients with schizophrenia: An MRI-based morphometric analysis. Biological Psychiatry, 46, 941±954. Staal, W.G., Pol, H.E.H., Schnack, H.G., Hoogendoorn, M.L.C., Jellema, K., Kahn, R.S. (2000). Structural brain abnormalities in patients with schizophrenia and their healthy siblings. American Journal of Psychiatry, 157, 416±421. Stefansson, H., Sigurdsson, E., Steinthorsdottir, V., Bjornsdottir, S., Sigmundsson, T., Ghosh, S., et al. (2002). Neuregulin 1 and susceptibility to schizophrenia. American Journal of Human Genetics, 71, 877±892. Straub, R.E., Jiang, Y., MacLean, C.J., Ma, Y., Webb, B.T., Myakishev, M.V., et al. (2002). Genetic variation in the 6p22.3 gene DTNBP1, the human ortholog of the mouse dysbindin gene, is associated with schizophrenia. American Journal of Human Genetics, 71, 337±348. Suarez, B.K., Duan, J., Sanders, A.R., Hinrichs, A.L., Jin, C.H., Hou, C., et al. (2006). Genomewide linkage scan of 409 European-ancestry and African±American families with schizophrenia: Suggestive evidence of linkage at 8p23.3-p21.2 and 11p13.1-q14.1 in the combined sample. American Journal of Human Genetics, 78, 315±333. Talkowski, M.E., Seltman, H., Bassett, A.S., Brzustowicz, L.M., Chen, X., Chowdari, K.V., et al. (2006). Evaluation of a susceptibility gene for schizophrenia: genotype based metaanalysis of RGS4 polymorphisms from thirteen independent samples. Biological Psychiatry, 60, 152±162. Tienari, P., Wynne, L.C., Moring, J., Lahti, I., Naarala, M., Sorri, A., et al. (1994). The Finnish adoptive family study of schizophrenia: Implications for family research. British Journal of Psychiatry, 164, 20±26. Tosato, S., Dazzan, P., Collier, D. (2005). Association between the neuregulin 1 gene and schizophrenia: a systematic review. Schizophrenia Bulletin, 31, 613±617. Tsuang, M.T., Winokur, G., Crowe, R.R. (1980). Morbidity risks of schizophrenia and affective disorders among ®rst degree relatives of patients with schizophrenia, mania, depression and surgical conditions. British Journal of Psychiatry, 137, 497±504. Valles, V., van Os, J., Guillamat, R., Gutierrez, B., Campillo, M., Gento, P., et al. (2000). Increased morbid risk for schizophrenia in families of in-patients with bipolar illness. Schizophrenia Research, 42, 83±90. Weisbrod, M., Hill, H., Niethammer, R., Sauer, H. (1999). Genetic in¯uence on auditory

20

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

information processing in schizophrenia: P300 in monozygotic twins. Biological Psychiatry, 46, 721±725. Weissman, M.M., Gershon, E.S., Kidd, K.K., Prusoff, B.A., Leckman, J.F., Dibble, E., et al. (1984). Psychiatric disorders in the relatives of probands with affective disorders. The Yale University ± National Institute of Mental Health Collaborative Study. Archives of General Psychiatry, 41, 13±21. Wender, P.H., Rosenthal, D., Kety, S.S., Schulsinger, F., Welner, J. (1974). Crossfostering. A research strategy for clarifying the role of genetic and experiential factors in the etiology of schizophrenia. Archives of General Psychiatry, 30, 121±128. Wickham, H., Murray, R.M. (1997). Can biological markers identify endophenotypes predisposing to schizophrenia? International Review of Psychiatry, 9, 355±364. Wijsman, E.M., Rosenthal, E.A., Hall, D., Blundell, M.L., Sobin, C., Heath, S.C., et al. (2003). Genome-wide scan in a large complex pedigree with predominantly male schizophrenics from the island of Kosrae: evidence for linkage to chromosome 2q. Molecular Psychiatry, 8, 695±705. Williams, N.M., O'Donovan, M.C., Owen, M.J. (2005). Is the dysbindin gene (DTNBP1) a susceptibility gene for schizophrenia? Schizophrenia Bulletin, 31, 800±805. Williams, N.M., Green, E.K., Macgregor, S., Dwyer, S., Norton, N., Williams, H., et al. (2006). Variation at the DAOA/G30 locus in¯uences susceptibility to major mood episodes but not psychosis in schizophrenia and bipolar disorder. Archives of General Psychiatry, 63, 366±373.

CHAPTER TWO

The Maudsley Family Study of Psychosis ± overview of clinical methodology and characteristics Colm McDonald

PARTICIPANT RECRUITMENT Families affected with schizophrenia, or more recently bipolar disorder, were recruited through voluntary support groups or by direct referral from mental health clinicians throughout the UK. The main voluntary support groups who assisted with recruitment were the National Schizophrenia Fellowship (Rethink) for schizophrenia and the Manic Depressive Fellowship for bipolar disorder. These groups provide support and advice for patients and their carers, and run regional meetings at which researchers from the study presented; they also produce regular newsletters in which the study was advertised. Letters were sent to consultant psychiatrists within the South London and Maudsley NHS Trust, as well as other Trusts around the UK, providing information about the study and requesting clinicians to refer families who might be interested in taking part. Most of the families recruited lived in south-east England. Controls were recruited from the community via advertisements in local newspapers and from local staff. These controls were chosen to match the characteristics of the combined group of patients and relatives on the basis of age, gender and parental social class, and were compensated for the time and inconvenience associated with having the assessments. In the case of schizophrenia, the variation of likely genetic risk in the family members was increased by recruiting participants who were either `familial', i.e. had other relatives among their ®rst- and/or second-degree 21

22

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

relatives affected with a similar psychotic disorder, or `non-familial', i.e. had no known relatives affected with a psychotic disorder. In the case of bipolar disorder, all participants recruited were `familial' in that they had a family history of bipolar disorder or another functional psychotic disorder among their ®rst- and/or second-degree relatives. As bipolar disorder ranges considerably in severity and some genetic epidemiology studies suggest that bipolar disorder with psychotic symptoms is most likely to share genetic liability with schizophrenia (see Chapter 1), participants were speci®cally recruited who had experienced manic episodes with psychotic symptoms at some point during their illness exacerbation.

STUDY PHASES There have been four waves of recruitment and data collection since the study's inception. Phase 1 of the study commenced in 1993 and consisted of families multiply affected with schizophrenia who also demonstrated unilinear transmission of genetic risk, i.e. contained presumed obligate carriers, and a sample of controls (Crawford et al., 1998; Frangou et al., 1997a, 1997b; Sharma et al., 1997, 1998). Phase 2 of the study focused predominantly on non-familial schizophrenia or schizoaffective disorder with some additional `familial' families and controls included (Grif®ths et al., 1998; McDonald et al., 2002; Toulopoulou et al., 2003a, 2003b). Phase 3 incorporated another wave of recruitment of both familial and non-familial families with schizophrenia or schizoaffective disorder and controls (Bramon et al., 2004, 2005; MacCabe et al., 2005; McDonald et al., 2004, 2006). Phase 4 consisted of families multiply affected with psychotic bipolar 1 disorder and some controls (McDonald et al., 2004, 2006). In total, 550 research participants participated to some extent in the Maudsley Family Study of Psychosis, consisting of 138 patients with schizophrenia or schizoaffective disorder (n=11), 187 of their unaffected ®rst-degree relatives, 40 patients with bipolar 1 disorder, 55 of their unaffected ®rst-degree relatives and 130 healthy volunteers.

INCLUSION AND EXCLUSION CRITERIA All participants were aged 16±69 years, and their ®rst language was English. Participants were excluded if they:

· · ·

had a history of organic brain disease had experienced head trauma resulting in loss of consciousness for more than 5 minutes ful®lled Diagnostic and Statistical Manual of Mental Disorders version 3R or 4 (DSM-IIIR or DSM-IV) criteria for substance or alcohol dependence in the 12 months prior to assessment.

2. CLINICAL METHODOLOGY AND CHARACTERISTICS

23

The studies were approved by the South London and Maudsley NHS Trust/Institute of Psychiatry Ethical Committee (Research) and all participants gave written informed consent to participate. The schizophrenia families were also chosen on the basis of the following inclusion criteria:

· · ·

The index patient had a lifetime DSM diagnosis of schizophrenia or schizoaffective disorder. The index patient had ®rst-degree relatives who were also willing to participate in the study. The patient had either: (1) another individual among his or her ®rstand/or second-degree relatives affected with schizophrenia or another functional psychotic disorder (`familial'); or (2) no known relative with a functional psychotic disorder as far as their third-degree relatives (`non-familial').

The bipolar disorder families were chosen on the basis of the following inclusion criteria:

· · · ·

The index patient had a lifetime DSM-IV diagnosis of bipolar 1 disorder (i.e. had experienced manic episodes rather than hypomanic episodes). The index patient had experienced delusions and/or hallucinations at some point during his or her symptomatic exacerbations. The patient had ®rst-degree relatives who were also willing to participate in the study. The patient had another individual among his or her ®rst- and/or second-degree relatives affected with bipolar disorder or another functional psychotic disorder.

None of the controls had a personal history of a psychotic, bipolar or schizophrenia spectrum disorder, nor any known family history of functional psychosis. The presence of other axis 1 psychiatric disorders was not an exclusion factor for controls. Had this been used as an exclusion criterion, when it was not an exclusion criterion for unaffected relatives, there would be a risk of recruiting `supernormal' controls (Kendler, 2003), with subsequent bias in detecting neurobiological abnormalities in unaffected relatives which were potentially related to factors other than having a ®rst-degree relative with the disorder.

CLINICAL ASSESSMENTS After recruitment and screening to ensure inclusion and exclusion criteria were met, patients, relatives, and controls were assessed using the same

24

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

BOX 2.1 Core structured clinical assessments performed with participants

· · · · · ·

Schedule for Affective Disorders and Schizophrenia±Lifetime (SADS±L) Annett Handedness scale Family History Research Diagnostic Criteria (FHRDC) or the Family Interview for Genetic Studies (FIGS) Schedule for Schizotypal Personality (SSP) Premorbid Adjustment Scale (PAS) Premorbid and Schizoid-Schizotypal Traits scale (PSST)

clinical scales in face-to-face interviews (Box 2.1). Sociodemographic details were collected and socio-economic status based on best ever occupation was derived from Of®ce of Population Censuses and Surveys Standard Occupational Classi®cation (HMSO, 1991). The parental occupation at the time of the participant's birth was also chosen as a measure of socioeconomic status in order to account for any social drift associated with the illnesses. Where both parents were in occupation, the higher social class was taken. Handedness was assessed using the Annett scale (Annett, 1970). Structured diagnostic interviews were performed on all participants using the Schedule for Affective Disorders and Schizophrenia±Lifetime version (Spitzer & Endicott, 1978) and additional information regarding the timing and nature of psychopathology was collected to enable lifetime DSM-IIIR (Phases 1 and 2) or DSM-IV (Phases 3 and 4) diagnoses to be made. Interviews were supplemented with information collected from relatives to con®rm the diagnoses and with medical notes when they were available and any doubt remained about the diagnosis. For relatives not assessed directly, information regarding psychiatric diagnoses was obtained from the most reliable informants (usually including the mother of the index patient in each family) using the Family History Research Diagnostic Criteria (Endicott et al., 1975) (Phases 1 and 2) or the Family Interview for Genetic Studies (Maxwell, 1992; Nurnberger et al., 1994) (Phases 3 and 4) and supplemented by medical notes where they were available. The Schedule for Schizotypal Personality (Baron et al., 1981) was used to assess nonpsychotic relatives and controls for schizotypal traits. DSM diagnoses of schizotypal personality disorder were made if the individual was considered to portray a `threshold or true' level of at least ®ve of the nine symptom clusters: (1) ideas of reference; (2) excessive social anxiety; (3) odd beliefs or magical thinking, which in¯uenced behaviour and were inconsistent with subcultural norms; (4) unusual perceptual experiences; (5) odd, eccentric, or peculiar behaviour or appearance; (6) no close friends or con®dants outside of ®rst-degree relatives; (7) odd speech; (8) inappropriate or constricted affect; (9) suspiciousness or paranoid ideation.

2. CLINICAL METHODOLOGY AND CHARACTERISTICS

25

Most patients also had assessments of premorbid function using the Premorbid Adjustment Scale (PAS) and the Premorbid and Schizoid± Schizotypal Traits scale (PSST) (Foerster et al., 1991) rated by maternal interview. The PAS scales were completed for two periods, childhood (5±11 years) and adolescence (12±16 years) and consisted of ratings of the following areas on a 0±7-point scale: social isolation, peer relations, scholastic performance, adaptation to school, interests; as well as a rating of sociosexual adjustment from age 16±20. The PSST scale consisted of pre-morbid ratings of the following areas on a 0±3-point scale: social isolation, affect, suspiciousness/sensitivity, thought content/beliefs, speech, antisocial behaviour, asocial behaviour, other abnormalities. Summing the individual scores produced total PAS and PSST scores and higher scores indicate worse premorbid social adjustment and higher levels of pre-morbid dysfunction.

SOCIODEMOGRAPHIC CHARACTERISTICS OF THE COMPLETE SAMPLE The sociodemographic characteristics of the entire sample of participants are displayed in Table 2.1. Each of the studies investigating cognitive or neurobiological endophenotypes was completed on a proportion of these participants. All participants were Caucasian. Differences between these participant groups on sociodemographic characteristics were analysed using one-way analysis of variance, with Bonferroni post-hoc tests, and Pearson's 2 tests. There was a signi®cant age difference between the groups, and post-hoc analysis revealed that the relatives of patients with schizophrenia were signi®cantly older than the patients and also signi®cantly older than controls. This was because the relatives groups contained parents of patients as well as siblings. The signi®cant gender difference between the groups was attributable to a larger proportion of men in the schizophrenia patient groups. There was also a signi®cant height difference between the groups, which post-hoc analysis demonstrated was due to non-familial schizophrenia patients being taller than their relatives. This in turn was due to the higher proportion of men in the schizophrenia patient group, as this effect disappeared when gender was controlled for. There were no signi®cant differences between the participant groups in years of education, proportion of left handers or proportion of participants whose parents were from social classes I or II. However, a smaller proportion of the schizophrenia patient groups had ever attained an occupational level of social class I or II themselves, indicative of underperformance or downward social drift associated with schizophrenia, given the equivalent levels of parental social class between the groups and the social class of their ®rst-degree relatives. By contrast, there was no evidence of loss of social class level among the bipolar patients.

Male gender Left handed Best social class (I or II) Parental social class (I or II)

Age (years) Age range Height (cm) Years of education

63.9 9.2 13.7

35.4

46 6 10

25

39

40 7 55

39.1

40.4 7.4 55.4

%

n

n

%

46.1 14.3 16±69 169.9 9.9 13.6 2.8

Mean SD

Mean SD

37.4 9.8 19±62 173.3 9.3 13.0 3.1

Relatives of patients with familial schizophrenia (n=99)

Patients with familial schizophrenia (n=72)

38

46 11 13

n

57.7

69.7 20.8 19.6

%

32.3 8.2 17±62 175.4 8.3 13.2 2.5

Mean SD

Patients with non-familial schizophrenia (n=66)

31

37 11 57

n

35.3

42.0 15.7 64.6

%

50.8 13.4 17±69 168.3 9.7 13.8 2.9

Mean SD

Relatives of patients with non-familial schizophrenia (n=88)

19

15 3 18

n

47.5

37.5 7.5 45.0

%

40.7 11.5 22±64 170.8 10.1 14.1 3.2

Mean SD

Patients with familial bipolar disorder (n=40)

23

27 10 34

n

41.8

49.1 18.2 61.8

%

43.8 15.3 17±68 170.0 10.2 14.4 3.6

Mean SD

44

59 14 55

n

33.7

45.4 11.6 42.2

%

37.8 13.3 18±69 171.0 9.5 14.3 3.4

Mean SD

Relatives of Controls patients with (n=130) familial bipolar disorder (n=55)

TABLE 2.1 Sociodemographic characteristics of each participant group (n=550)

p

Chi 2

10.66

0.10

<0.001 0.15 <0.001

0.001 0.06

3.70 2.07

25.12 9.46 51.74

<0.001

p 19.16

F

Statistic

2. CLINICAL METHODOLOGY AND CHARACTERISTICS

27

CLINICAL CHARACTERISTICS OF THE SAMPLE Of the 138 patients in the `schizophrenia' groups, 127 ful®lled DSM criteria for a diagnosis of schizophrenia and 11 ful®lled DSM criteria for a diagnosis of schizoaffective disorder (7 from the familial subsample and 4 from the non-familial subsample). All 40 patients with bipolar disorder ful®lled DSM-IV criteria for bipolar 1 disorder and all bar one had experienced psychotic symptoms, i.e. delusions or hallucinations, at some stage during episodes of illness exacerbation. The patient who did not report psychotic symptoms was considered likely to express the same phenotype as she had two other relatives with bipolar 1 disorder who participated in the study and had experienced psychotic symptoms. All participants were outpatients at the time of assessment. Almost all of the patients with schizophrenia were taking antipsychotic medication at the times of the assessments, mostly atypical antipsychotic medications. Of patients with bipolar disorder, most were taking lithium or another mood stabilizer, but four patients were medication free at the time of assessment. Of the 187 unaffected relatives of patients with schizophrenia, 47 had ful®lled criteria for another axis 1 disorder at some point in their lives: 36 for major depressive disorder, 4 for panic disorder, 2 for generalized anxiety disorder, 1 for obsessive compulsive disorder, 1 for bulimia nervosa, 1 for social phobia, 1 for bipolar II disorder and 1 for alcohol dependence syndrome. Fourteen relatives of patients with schizophrenia ful®lled criteria for schizoid or schizotypal personality disorders, 7 of whom had also experienced an axis 1 disorder. The remaining 133 relatives of schizophrenia patients had never ful®lled criteria for any DSM disorder. Of the 55 relatives of patients with bipolar disorder, 13 had ful®lled criteria for another DSM-IV axis 1 disorder at some point in their lives: 10 for major depressive disorder, 2 for panic disorder and 1 for alcohol dependence syndrome. Eight of the control sample participants ful®lled criteria for a DSM axis 1 disorder at some point in their lives: 7 for major depressive disorder and 1 for alcohol dependence syndrome. None of the bipolar relatives or controls ful®lled criteria for schizotypal personality disorder. Of those relatives and controls who had ever ful®lled criteria for an axis 1 psychiatric disorder, the vast majority were not taking any medication at the time of the assessments. Some further clinical information on the patients is provided in Table 2.2. Age of onset of symptoms was de®ned as age of ®rst psychotic symptoms for schizophrenia and age of ®rst discrete episode of mood disorder for bipolar participants. There was no signi®cant difference between patient groups (F=2.34, p=0.10). Patients with bipolar disorder had a higher mean age of ®rst hospitalization than the schizophrenia patient groups (F=3.12, p=0.047). There were no signi®cant differences in the number of admissions

3

2

1

20.8 23.8 5.6 6.2 9.9 2.8

4.9 5.6 5.5 4.0 5.4 2.3

8±35 15±42 0±201 0±15 1±22 0±11

20.1 22.5 4.3 6.7 10.4 3.4

4.6 4.7 4.0 5.3 6.9 3.4

13±37 13±37 0±201 0±19 0±28 0±15

Range 22.3 25.6 5.3 2.9 4.2 1.3

Mean

5.5 7.8 5.4 2.5 3.3 1.5

SD

13±36 15±52 0±201 0±10 0±13 0±5

Range

Patients with bipolar disorder (n=40)

Number of admissions recorded as 20 if there were 20 or more. PAS, Premorbid Adjustment Scale (1, childhood; 2, adolescence), rated by maternal interview and available on 46 patients with familial schizophrenia, 41 patients with non-familial schizophrenia and 19 patients with bipolar disorder. PSST, Premorbid Schizoid±Schizotypal Traits, rated by maternal interview available on 44 patients with familial schizophrenia, 34 patients with nonfamilial schizophrenia and 19 patients with bipolar disorder.

Age of ®rst symptoms Age of ®rst hospitalization No. of hospital admissions PAS 12 PAS 22 PSST3

SD

Mean

Range

Mean

SD

Patients with non-familial schizophrenia (n=99)

Patients with familial schizophrenia (n=72)

TABLE 2.2 Clinical details on premorbid traits and course of illness in patients

2. CLINICAL METHODOLOGY AND CHARACTERISTICS

29

to hospital (F=0.83, p=0.44). Schizophrenia patients had higher maternally rated scores of pre-morbid social dysfunction during childhood (F=5.14, p=0.007) and adolescence (F=8.40, p<0.001) and of pre-morbid schizoid± schizotypal traits (F=3.85, p=0.02) than patients with bipolar disorder. This is consistent with data from several cohort studies, which indicate that schizophrenia is a more `neurodevelopmental' disorder and associated with more substantial early de®cits than bipolar disorder (Murray et al., 2004).

GENETIC LIABILITY SCALE An additional measure of continuous genetic liability was incorporated into some of the endophenotypic analyses of brain imaging in the sample, and is described in detail here. As the family size and number of affected individuals within families who participated in the study varied considerably, the likely genetic liability in such families is also likely to vary. This is ignored by a simple relatives versus controls comparison, which treats all unaffected relatives as a homogenous group. The continuous genetic liability model assumed that families most densely affected with illness would manifest the action of susceptibility genes to a greater extent and would be more likely to display the neurobiological abnormalities linked to such genes. To take advantage of this likely variation in genetic susceptibility in the study sample, a quantitative measure of genetic liability for each participant was calculated. This scale was devised by Professor Pak Sham and implemented using a program written by Dr Harvey Wickham in S-Plus 6.1 (Insightful Corporation, Seattle). The derivation of a similar `genetic liability score' has been described previously (Lawrie et al., 2001). For the purposes of calculating the genetic liability scale, a polygenic multifactorial liability-threshold model of illness was adopted (Gottesman, 1991), whereby liability was assumed to be continuous within the population with a Gaussian distribution and separate scales were derived for schizophrenia and bipolar disorder. Patients were initially assumed to have an expected liability above a threshold, which was based on the population prevalence rates of the illnesses. These were derived from the 1-month prevalence rates identi®ed in the Epidemiological Catchment Area study (Regier et al., 1993), taken as 0.7% for schizophrenia and 0.5% for bipolar disorder ± slightly lower than the prevalence rate quoted for mania (0.6%) ± in order to better re¯ect that subset of bipolar patients who experience psychotic symptoms, reported to be around 80% (Cassidy et al., 1998). Given these assumptions, initial imputed liabilities were 2.78 for patients with schizophrenia and 2.89 for patients with bipolar disorder. Other participants with functional psychotic disorders within the schizophrenia or bipolar families were assumed to be expressing the same phenotype as the

30

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

index patient and were allotted the same initial liability. Empirical evidence to support this assumption comes from the study by Potash and colleagues (Potash et al., 2001) of multiply affected bipolar disorder pedigrees. These authors report familial aggregation of psychotic disorders within such families, suggesting that similar susceptibility genes underlie the production of different psychotic syndromes in these families. A second threshold was included in the schizophrenia families to categorize those participants with personality disorders related to schizophrenia (schizoid, paranoid, schizotypal), assumed to have a population prevalence of 3.3% from the pooled rates of ten community studies (Torgersen et al., 2001), which produced an initial expected liability of 2.08 for such individuals with personality disorder. Other relatives were considered unaffected and had an initial expected liability of ±0.08 within schizophrenia families and ±0.07 within bipolar disorder families. A vector of liabilities, L, initially imputed to each family member was thus derived for each family. These scores were then adjusted for each participant to account for family size and affection distribution. First, a correlation matrix for each family, R, was constructed, describing the genetic inter-relationships of all individuals over the age of 16 years in each family as far as second-degree from the index patient, i.e. self = 1, ®rst-degree relatives = 0.5, seconddegree relatives = 0.25, spouse = 0. Assuming that genes are the only source of familial resemblance (as has been demonstrated by twin studies; Cannon et al., 1998; Cardno et al., 1999), a second correlation matrix of liabilities to illness in each family, V, was produced by multiplying the off-diagonal elements of R by an estimate of heritability (h2), taken to be 0.7 for both schizophrenia and bipolar disorder. A vector of expected genetic risks, G, for each family is then given by the formula: G = h2RV ±1L under the assumptions of normal distribution theory (Mardia et al., 1979). A step-by-step demonstration of the calculation of genetic liability scores for a single family multiply affected with bipolar disorder is given in Figure 2.1a±e. The genetic liability scale is optimally employed for the calculation of genetic risk in those families in which the patient has a family history of illness and therefore positive information is employed (i.e. which particular participants are affected and the genetic relatedness and number of affected individuals). When no other family members are affected, as in those families classi®ed as non-familial in which there were no schizotypal relatives, or when no members are affected, as in the control families, the principal variable that in¯uences the genetic liability score is family size.

(b) Vector of initially imputed liabilities for each of the 16 members in the family (L) 1 ±0.07 2 ±0.07 3 2.89 4 ±0.07 5 ±0.07 6 ±0.07 7 ±0.07 8 ±0.07 9 ±0.07 10 ±0.07 11 ±0.07 12 ±0.07 13 2.89 14 2.89 15 ±0.07 16 ±0.07

(a) Pedigree diagram with individuals numbered

continues overleaf

Figure 2.1 Step-by-step demonstration of genetic liability score calculation in a single family multiply affected with bipolar disorder.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 1 0 0 0 0.5 0.5 0.5 0.5 0.5 0 0 0 0 0.25 0.25 0.25

2 0 1 0 0 0.5 0.5 0.5 0.5 0.5 0 0 0 0 0.25 0.25 0.25

3 0 0 1 0 0 0 0 0 0 0.5 0.5 0.5 0.5 0.25 0.25 0.25

4 0 0 0 1 0 0 0 0 0 0.5 0.5 0.5 0.5 0.25 0.25 0.25

5 0.5 0.5 0 0 1 0.5 0.5 0.5 0.5 0 0 0 0 0.25 0.25 0.25

6 0.5 0.5 0 0 0.5 1 0.5 0.5 0.5 0 0 0 0 0.25 0.25 0.25

7 0.5 0.5 0 0 0.5 0.5 1 0.5 0.5 0 0 0 0 0.25 0.25 0.25

8 0.5 0.5 0 0 0.5 0.5 0.5 1 0.5 0 0 0 0 0.25 0.25 0.25

(c) Correlation matrix of genetic inter-relationships between individuals (R)

Figure 2.1 (continued )

9 0.5 0.5 0 0 0.5 0.5 0.5 0.5 1 0 0 0 0 0.5 0.5 0.5

10 0 0 0.5 0.5 0 0 0 0 0 1 0.5 0.5 0.5 0.5 0.5 0.5

11 0 0 0.5 0.5 0 0 0 0 0 0.5 1 0.5 0.5 0.25 0.25 0.25

12 0 0 0.5 0.5 0 0 0 0 0 0.5 0.5 1 0.5 0.25 0.25 0.25

13 0 0 0.5 0.5 0 0 0 0 0 0.5 0.5 0.5 1 0.25 0.25 0.25

14 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.5 0.5 0.25 0.25 0.25 1 0.5 0.5

15 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.5 0.5 0.25 0.25 0.25 0.5 1 0.5

16 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.5 0.5 0.25 0.25 0.25 0.5 0.5 1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 1 0 0 0 0.35 0.35 0.35 0.35 0.35 0 0 0 0 0.175 0.175 0.175

2 0 1 0 0 0.35 0.35 0.35 0.35 0.35 0 0 0 0 0.175 0.175 0.175

3 0 0 1 0 0 0 0 0 0 0.35 0.35 0.35 0.35 0.175 0.175 0.175

4 0 0 0 1 0 0 0 0 0 0.35 0.35 0.35 0.35 0.175 0.175 0.175

5 0.35 0.35 0 0 1 0.35 0.35 0.35 0.35 0 0 0 0 0.175 0.175 0.175

(d) Correlation matrix of liabilities to illness (V) 6 0.35 0.35 0 0 0.35 1 0.35 0.35 0.35 0 0 0 0 0.175 0.175 0.175

7 0.35 0.35 0 0 0.35 0.35 1 0.35 0.35 0 0 0 0 0.175 0.175 0.175

8 0.35 0.35 0 0 0.35 0.35 0.35 1 0.35 0 0 0 0 0.175 0.175 0.175

9 0.35 0.35 0 0 0.35 0.35 0.35 0.35 1 0 0 0 0 0.35 0.35 0.35

10 0 0 0.35 0.35 0 0 0 0 0 1 0.35 0.35 0.35 0.35 0.35 0.35

11 0 0 0.35 0.35 0 0 0 0 0 0.35 1 0.35 0.35 0.175 0.175 0.175

12 0 0 0.35 0.35 0 0 0 0 0 0.35 0.35 1 0.35 0.175 0.175 0.175

13 0 0 0.35 0.35 0 0 0 0 0 0.35 0.35 0.35 1 0.175 0.175 0.175

14 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.35 0.35 0.175 0.175 0.175 1 0.35 0.35

16 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.35 0.35 0.175 0.175 0.175 0.35 0.35 1

continues overleaf

15 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.175 0.35 0.35 0.175 0.175 0.175 0.35 1 0.35

34

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Figure 2.1 (continued) (e) Genetic liability vector G (= h2RV-1L) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

0.03 0.03 1.95 ±0.08 0.01 0.01 0.01 0.01 0.21 0.63 0.42 0.42 1.99 1.75 0.19 0.19

Scores derived from such families are based on less information and are more prone to error than those from multiply affected families. Therefore only families in which the index patient had a family history of illness were included in analyses that utilized the genetic liability scale. Examples of genetic liability scores for individuals within families of differing illness density are provided in Figure 2.2 (Figure 2.2(b) shows the pedigree described in Figure 2.1). Although continuous, the distribution of the genetic liability score was bimodal with peaks for relatives and for patients. As the main contribution to liability was presence of illness, the mean score of patients was higher than that of relatives. Although the scale does successfully demonstrate variation in genetic liability among participants, this bimodal distribution is inevitable because the scale is based on limited information. If all information were available (e.g. through accurate diagnoses of very extensive pedigrees), the point of rarity between relatives and patients would close. Because of this bimodal distribution, any direct analysis of a variable against the genetic liability score would largely be comparing the level of that variable in patients to that in relatives. To utilize the information regarding variation in likely genetic liability provided by this quantitative measure therefore, patients and relatives were combined in the analyses, but participant group (patient versus relative) was controlled for. Subsequent interaction analyses were then employed to examine whether the relationship between genetic liability and tissue volume differed for patients and relatives. The results of such analyses with brain morphometric variables are presented in detail in Chapter 3.

continues overleaf

(a) Genetic liability scores in high density family: unaffected individual with brother, mother and grandmother affected has very high genetic liability (0.51)

Figure 2.2 Genetic liability scores in families with differing density of illness

(b) Genetic liability scores in high density family: unaffected parent who appears to be transmitting genetic susceptibility (i.e. `presumed obligate carrier'), with sister, father and son affected, has relatively high genetic liability (0.63); this is the pedigree described in Figure 2.1

Figure 2.2 (continued )

(c) Genetic liability scores in low density family: although still `familial', the large number of unaffected individuals are re¯ected relatively by lower genetic liability scores for an unaffected sibling (0.10)

38

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

REFERENCES Annett, M. (1970). A classi®cation of hand preference by association analysis. British Journal of Psychology, 61, 303±321. Baron, M., Asnis, L., Greun, R. (1981). The schedule for schizotypal personalities (SSP). A diagnostic interview for schizotypal features. Psychiatry Research, 4, 213±228. Bramon, E., Croft, R.J., McDonald, C., Virdi, G.K., Gruzelier, J.G., Baldeweg, T., et al. (2004). Mismatch negativity in schizophrenia: A family study. Schizophrenia Research, 67, 1±10. Bramon, E., McDonald, C., Croft, R.J., Landau, S., Filbey, F., Gruzelier, J., et al. (2005). Is the P300 wave an endophenotype for schizophrenia? A meta-analysis and a family study. Neuroimage, 27, 960±968. Cannon, T.D., Kaprio, J., Lonnqvist, J., Huttunen, M., Koskenvuo, M. (1998). The genetic epidemiology of schizophrenia in a Finnish twin cohort. A population-based modeling study. Archives of General Psychiatry, 55, 67±74. Cardno, A.G., Marshall, E.J., Coid, B., Macdonald, A.M., Ribchester, T.R., Davies, N.J., et al. (1999). Heritability estimates for psychotic disorders: The Maudsley twin psychosis series. Archives of General Psychiatry, 56, 162±168. Cassidy, F., Murry, E., Forest, K., Carroll, B.J. (1998). Signs and symptoms of mania in pure and mixed episodes. Journal of Affective Disorders, 50, 187±201. Crawford, T.J., Sharma, T., Puri, B.K., Murray, R.M., Berridge, D.M., Lewis, S.W. (1998). Saccadic eye movements in families multiply affected with schizophrenia: The Maudsley Family Study. American Journal of Psychiatry, 155, 1703±1710. Endicott, J., Andreasen, N.C., Spitzer, R.L. (1975). Family History Research Diagnostic Criteria. New York: New York State Psychiatric Institute, Biometrics Research Division. Foerster, A., Lewis, S.W., Owen, M.J., Murray, R.M. (1991). Pre-morbid adjustment and personality in psychosis. Effect of sex and diagnosis. British Journal of Psychiatry, 158, 171±176. Frangou, S., Sharma, T., Alarcon, G., Sigmundsson, T., Takei, N., Binnie, C., et al. (1997a). The Maudsley Family Study II: Endogenous event-related potentials in familial schizophrenia. Schizophrenia Research, 23, 45±53. Frangou, S., Sharma, T., Sigmundsson, T., Barta, P., Pearlson, G., Murray, R.M. (1997b). The Maudsley Family Study IV. Normal planum temporale asymmetry in familial schizophrenia ± A volumetric MRI study. British Journal of Psychiatry, 170, 328±333. Gottesman, I.I. (1991). Schizophrenia genesis: The origins of madness. New York: H. Freeman & Co. Grif®ths, T.D., Sigmundsson, T., Takei, N., Rowe, D., Murray, R.M. (1998). Neurological abnormalities in familial and sporadic schizophrenia. Brain, 121, 191±203. HMSO (1991). Of®ce of Population Censuses and Surveys. Standard Occupational Classi®cation. Volume 3. London: HMSO. Kendler, K.S. (2003). The genetics of schizophrenia: Chromosomal deletions, attentional disturbances, and spectrum boundaries. American Journal of Psychiatry, 160, 1549±1553. Lawrie, S.M., Whalley, H.C., Abukmeil, S.S., Kestelman, J.N., Donnelly, L., Miller, P., et al. (2001). Brain structure, genetic liability, and psychotic symptoms in subjects at high risk of developing schizophrenia. Biological Psychiatry, 49, 811±823. Maccabe, J.H., Simon, H., Zanelli, J.W., Walwyn, R., McDonald, C.D., Murray, R.M. (2005). Saccadic distractibility is elevated in schizophrenia patients, but not in their unaffected relatives. Psychological Medicine, 35, 1727±1736. Mardia, K.V., Kent, J.T., Bibby, J.M. (1979). Multivariate Analysis. New York: Academic Press. Maxwell, M.E. (1992). Family interview for genetic studies. Bethesda, Maryland: Clinical

2. CLINICAL METHODOLOGY AND CHARACTERISTICS

39

Neurogenetics Branch, Intramural Research Program, National Institute of Mental Health (NIMH). McDonald, C., Grech, A., Toulopoulou, T., Schulze, K., Chapple, B., Sham, P.C., et al. (2002). Brain volumes in familial and non-familial schizophrenic probands and their unaffected relatives. American Journal of Medical Genetics, 114, 616±625. McDonald, C., Bullmore, E.T., Sham, P.C., Chitnis, X., Wickham, H., Bramon, E., et al. (2004). Association of genetic risks for schizophrenia and bipolar disorder with speci®c and generic brain structural endophenotypes. Archives of General Psychiatry, 61, 974±984. McDonald, C., Marshall, N., Sham, P.C., Bullmore, E.T., Schulze, K., Chapple, B., et al. (2006). Regional brain morphometry in patients with schizophrenia or bipolar disorder and their unaffected relatives. American Journal of Psychiatry, 163, 478±487. Murray, R.M., Sham, P., van Os, J., Zanelli, J., Cannon, M., McDonald, C. (2004). A developmental model for similarities and dissimilarities between schizophrenia and bipolar disorder. Schizophrenia Research, 71, 405±416. Nurnberger, J.I., Blehar, M.C., Kaufmann, C.A., York-Cooler, C., Simpson, S.G., HarkavyFriedman, J., et al. (1994). Diagnostic interview for genetic studies. Rationale, unique features, and training. NIMH Genetics Initiative. Archives of General Psychiatry, 51, 849±859. Potash, J.B., Willour, V.L., Chiu, Y.F., Simpson, S.G., MacKinnon, D.F., Pearlson, G.D., et al. (2001). The familial aggregation of psychotic symptoms in bipolar disorder pedigrees. American Journal of Psychiatry, 158, 1258±1264. Regier, D.A., Narrow, W.E., Rae, D.S., Manderscheid, R.W., Locke, B.Z., Goodwin, F.K. (1993). The de facto US mental and addictive disorders service system. Epidemiologic catchment area prospective 1-year prevalence rates of disorders and services. Archives of General Psychiatry, 50, 85±94. Sharma, T., DuBoulay, G., Lewis, S., Sigmundsson, T., Gurling, H., Murray, R. (1997). The Maudsley Family Study 1. Structural brain changes on magnetic resonance imaging in familial schizophrenia. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 21, 1297±1315. Sharma, T., Lancaster, E., Lee, D., Lewis, S., Sigmundsson, T., Takei, N., et al. (1998). Brain changes in schizophrenia. Volumetric MRI study of families multiply affected with schizophrenia ± The Maudsley Family Study 5. British Journal of Psychiatry, 173, 132±138. Spitzer, R.L., Endicott, J. (1978). Schedule for affective disorders and schizophrenia ± Lifetime version. New York: New York State Psychiatric Institute. Torgersen, S., Kringlen, E., Cramer, V. (2001). The prevalence of personality disorders in a community sample. Archives of General Psychiatry, 58, 590±596. Toulopoulou, T., Morris, R.G., Rabe-Hesketh, S., Murray, R.M. (2003a). Selectivity of verbal memory de®cit in schizophrenic patients and their relatives. American Journal of Medical Genetics, 116, 1±7. Toulopoulou, T., Rabe-Hesketh, S., King, H., Murray, R.M., Morris, R.G. (2003b). Episodic memory in schizophrenic patients and their relatives. Schizophrenia Research, 63, 261±271.

CHAPTER THREE

Auditory evoked potentials as genetic trait markers of schizophrenia I. Williams, S. Frangou and E. Bramon

INTRODUCTION De®cits in various domains of cognitive functioning have increasingly been recognized over the last decade as a core feature of schizophrenia (Heinrichs, 2004). These not only include de®cient performance in higher cognitive domains but also extend to information processing at the sensory and pre-attentive level. Such cognitive abnormalities can be studied through neurophysiological techniques. Event-related potentials (ERPs) are changes in the human electroencephalogram (EEG) that occur in response to stimuli. They allow measurement of cortical physiology while the subject is performing cognitive tasks and thus provide a non-invasive technique of evaluating neural activation during cognitive processes. Importantly, ERPs are inexpensive and easily obtained from large numbers of subjects, and are stable quantitative measures. The main limitation of neurophysiological techniques to examine cortical activity is their low spatial resolution and thus the precise anatomical sources of the signals are not well identi®ed. However, this is compensated for by an extremely high temporal resolution that enables measurement of changes in the human EEG during cognition and perception in vivo, in real time. Neurophysiological experiments used in psychosis research range from basic pre-pulse inhibition tests (Kumari et al., 2005) to high-order N400 language tasks (Matsumoto et al., 2005). However, it is the P300, mismatch negativity (MMN) and P50 waveforms that have been best characterized as schizophrenia endophenotypes. In this 41

42

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

chapter, we describe each of these waves and present a meta-analysis of existing data on P300 and P50 wave abnormalities in patients with schizophrenia and their unaffected relatives. We then describe data from the Maudsley Family Study of Psychosis, which explored abnormalities of the P300 and MMN waves in patients with schizophrenia and in their unaffected ®rst-degree relatives.

The P300 wave The P300 waveform is a measure of cortical activity during tasks of stimuli discrimination in which the participant must use memory and sustained attention abilities. It is typically elicited using an auditory `oddball' paradigm, when participants actively attend to the task of distinguishing `oddball' or target stimuli from standard non-target stimuli. They are usually asked to give a response to the targets only, and their accuracy and speed can be measured. The amplitude of the P300 wave is maximal at central± parietal scalp regions and has been conceptualized as the physiological correlate of a working memory update of changes in the environment (Donchin & Coles, 1988). The P300 latency (the time it takes for the wave to peak) corresponds to stimulus evaluation time (Kutas et al., 1977). It has been found that shorter P300 latency is correlated with superior cognitive performance (Johnson et al., 1985). Amplitude reduction of the P300 wave in schizophrenia was ®rst reported in 1972 (Roth & Cannon, 1972). The value of the P300 wave in schizophrenia research has recently received further support from meta-analysis work, which con®rms that patients with schizophrenia have severe P300 amplitude reductions as well as moderate latency delays (Jeon & Polich, 2001, 2003). Over the last 20 years, several studies have explored the P300 wave in non-psychotic relatives of schizophrenia patients with somewhat con¯icting conclusions. The P300 amplitude was reported to be reduced in relatives and therefore a promising endophenotype (Kidogami et al., 1991; Roxborough et al., 1993; Schreiber et al., 1992) as was its latency (Blackwood et al., 1991) However, evidence to the contrary has also emerged (Blackwood et al., 1991, 2001; Kidogami et al., 1991; Winterer et al., 2003). Some studies may well have been underpowered. In addition, P300 deviances may only be present in a proportion of relatives, while potential publication biases may also have contributed to the existing confusion. A segregation analysis of P300 latency phenotypes in Scottish families suggested that the P300 latency could be a useful measure of genetic predisposition to schizophrenia in asymptomatic relatives that could increase the chances of linkage studies detecting susceptibility genes (Sham et al., 1994). Interest in the genetics of the P300 wave has recently been enhanced by its association with two genes that are believed to play a role in

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

43

schizophrenia: disrupted in schizophrenia-1 (DISC1) (Blackwood & Muir, 2004) and the catechol-O-methyltransferase gene (COMT ) (Shifman et al., 2002). Although the issue is far from clear, the case for the use of P300 alternative phenotypes in psychiatric genetic research is growing.

Mismatch negativity First described in 1978, the mismatch negativity (MMN) wave examines the early pre-attentive stages of auditory perception. It has been proposed to re¯ect activation of neural structures within primary auditory cortex or adjacent supratemporal auditory regions (Sams et al., 1991). In response to a repetitive auditory stimulus, the brain automatically develops an accurate neuronal trace or `echoic memory' that represents the physical features of the stimulus. The MMN may form part of an alerting system in humans and some animals, which enables them to detect unusual and possibly dangerous events in the environment (Tiitinen et al., 1994). MMN is elicited in a typical `oddball' paradigm in which a sequence of repetitive standard sounds is interrupted by an infrequently occurring deviant or `oddball' stimulus. However, unlike the P300, the MMN paradigm does not require the participant to attend or respond to the stimuli. The MMN occurs rapidly following deviant stimuli and is most clearly seen by subtraction of the wave elicited by the standard stimulus from that elicited by the deviant stimuli. It is a wave of negative voltage, with maximum amplitude in frontocentral areas, and which peaks as soon as 50±200 ms after stimulus onset. Shelley and colleagues (1991) were the ®rst to report reduced MMN amplitudes in patients with schizophrenia. It is now thought that this re¯ects a de®cit in the window of temporal integration of auditory information during the early phases of auditory processing (Michie, 2001). Amplitude reduction of MMN has been replicated in several small groups of patients with schizophrenia (Alain et al., 1998; Catts et al., 1995; Javitt et al., 1995; Oades et al., 1997; Shelley et al., 1999) although there have also been a number of negative reports (Jessen et al., 2001; Kathmann et al., 1995; Kirino & Inoue, 1999; O'Donnell et al., 1994). Two small studies indicate that the MMN amplitude may also be impaired amongst the unaffected relatives of patients with schizophrenia (Jessen et al., 2001; Michie, 2001). Further interest in studying MMN in schizophrenia has stemmed from its possible relationship to the function of the neurotransmitter N-methylD-aspartate (NMDA). Animals and humans given phencyclidine and other NMDA antagonists show a selective decrease in MMN amplitudes, while other ERPs remain unaltered (Javitt, 2000; Javitt et al., 1996; Umbricht et al., 2000). Furthermore, phencyclidine can induce transient psychotic states in healthy participants and can aggravate pre-existing symptoms of the

44

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

illness (Olney & Farber, 1995). If NMDA receptors are involved in the generation of MMN, the wave could be used as a non-invasive method to measure NMDA function in vivo (Umbricht & Krljes, 2005). Abnormal MMN amplitude is a feature of chronic schizophrenia but its suitability as endophenotype for the disease is not established and the genetic basis of this de®cit has not yet been explored.

P50 gating Patients with schizophrenia suffer from dysfunction in the allocation of attention (McGhie & Chapman, 1961). For these individuals a perceptual impairment ± an inability to ignore, or ®lter out, irrelevant sensory information ± is paralleled by a neurophysiological abnormality in sensory gating (Cullum et al., 1993; Erwin et al., 1998). Typically, sensory gating is assessed by considering the P50 ERP obtained through what is known as the conditioning±testing paradigm. P50 ERP waves are generated by pairs of identical clicks presented 500 milliseconds apart; where the ®rst and second clicks are referred to as conditioning and testing respectively. Healthy individuals exhibit a substantial reduction or `wave suppression' of the second (testing P50) wave relative to the ®rst (conditioning P50) as electroencephalographic activity evoked by subsequent stimuli is attenuated (i.e. `gated'). The diminished test P50 wave is probably due to activation of inhibitory neural circuitry by the conditioning P50 stimuli (Freedman et al., 1997). Interestingly, individuals with schizophrenia (Adler et al., 1982; Boutros et al., 1991) and many of their unaffected ®rst-degree relatives (Clementz et al., 1998; Jessen et al., 2001; Michie et al., 2002) exhibit de®cits in sensory gating and thus a less ef®cient P50 suppression. This P50 suppression de®cit has been reported to be linked to a chromosome 15 locus (15q14) harbouring the -7 nicotinic acid receptor gene, which is a putative candidate gene for schizophrenia (Freedman et al., 1997). This same locus has subsequently been linked to a broader-phenotype-diagnosis of schizophrenia (Riley et al., 2000) that is consistent with the notion that sensory gating impairment, as measured electrophysiologically, is an inherited trait in families with a history of schizophrenia. As such, it is a most plausible endophenotype for schizophrenia candidate gene studies.

HOW SEVERE ARE NEUROPHYSIOLOGICAL DEFICITS IN SCHIZOPHRENIA? META-ANALYSES OF THE PUBLISHED LITERATURE First, we set out to address the question of whether P300 and P50 are distributed differently between healthy and affected individuals using metaanalysis (Bramon et al., 2004b). Meta-analysis is a powerful technique for

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

45

TABLE 3.1 Main results of the meta-analysis of P300 case±control studies P300 amplitude

P300 latency

Number of studies Number of participants

43 44 1339 patients / 1153 controls 1348 patients / 1174 controls

Main meta-analysis pooled standardized effect size

PSES=0.85 95% CI: 0.65 to 1.05 p<0.001

PSES=±0.57 95% CI: ±0.75 to ±0.38 p<0.001

Meta-analysis of 11 drug-free studies, 244 patients and 272 controls Heterogeneity test Egger's test of publication bias

PSES=1.23 95% CI: 0.86 to 1.60 p<0.001

PSES=±0.48 95% CI: ±0.80 to ±0.15 p=0.004

p<0.001 p=0.03 Evidence of publication bias leading to overestimation of de®cits.

p<0.001 p=0.98 No evidence of bias

PSES, pooled standardized effect size (weighted mean of standardized effect size SES for primary studies). PSES and SES are de®ned as Cohen's d (Cohen, 1969): (control mean ± patient mean) / pooled standard deviation or (relative mean ± patient mean / pooled standard deviation). Reprinted from Schizophrenia Research, 70, Bramon et al. Meta-analysis of the P300 and P50 waveforms in schizophrenia, pp. 315±329. Copyright (2004), with permission from Elsevier.

integrating quantitative data from several studies with the resultant increase in sample size providing more statistical power to detect subtle deviations. We identi®ed 46 studies on the P300 wave that were suitable for analysis. These included a total of 1443 patients with schizophrenia and 1251 controls. Compared to healthy volunteers, patients with schizophrenia had signi®cantly smaller P300 amplitudes. The pooled effect size for P300 amplitude was 0.85 (95% con®dence interval (CI): 0.65 to 1.05, p<0.001). The P300 latency was signi®cantly delayed in the patient group compared to healthy volunteers. The pooled effect size was ±0.57 (95% CI ±0.75 to ±0.38, p<0.001). In addition, a meta-analysis of a subset of eleven P300 drug-free studies clearly con®rmed signi®cant amplitude reductions and latency delays in unmedicated patients. Therefore, the P300 deviances described in schizophrenia are not caused by antipsychotics but are genuinely associated with the disease. Further details can be seen in Table 3.1 and Figure 3.1. Are these P300 amplitude and latency de®cits in patients with schizophrenia also found in their unaffected relatives? To address this question we conducted an additional meta-analysis of published family studies. We identi®ed 11 studies suitable for analysis, containing a total of 472 relatives

Blackwood (1994) Glabus (1994) Strik (1994b) Strik (1994a) Anderson (1995) Javitt (1995) Jyoti Rao (1995) O’Donnell (1995b) O’Donnell (1995a) Souza (1995) Bougerol (1996) Juckel (1996) Stefansson (1996) Tretsman (1996) Boutros (1997) Frangou (1997) Pallanti (1997) Shajahan (1997) Wagner (1997*) Coburn (1998*) Umbricht (1998) D’Amato (1999) Ford (1999) Frodl-Bauch (1999) Hill (1999) Laurent (1999*) Pallanti (1999) Salisbury (1999) Weisbrod (1999) Brown (2000) Karoumi (2000) Mathalon (2000*) Turetsky (2000) Williams (2000) Alain (2001) Blackwood (2001) Martin-Loeches (2001) Winterer (2001*) Higashima (2002) Iwanami (2002) Nieman (2002) Price (2002*) Gonul (2003*) Combined

–2

0

–4

–2

P300 Amplitude

2

4

0

2

Blackwood (1994*) Glabus (1994) Strik (1994b) Strik (1994a) Javitt (1995) Jyoti Rao (1995) O’Donnell (1995b) O’Donnell (1995a) Souza (1995) Bobes (1996) Bougerol (1996) Stefansson (1996) Tretsman (1996) Boutros (1997) Frangou (1997) Pallanti (1997) Shajahan (1997) Wagner (1997*) Coburn (1998*) Turetsky (1998) Umbricht (1998) Weir (1998) D’Amato (1999) Ford (1999) Frodl-Bauch (1999) Hill (1999) Laurent (1999*) Pallanti (1999) Salisbury (1999) Weisbrod (1999) Brown (2000) Karoumi (2000) Mathalon (2000*) Turetsky (2000) Williams (2000) Alain (2001) Blackwood (2001) Martin-Loeches (2001) Winterer (2001*) Higashima (2002) Iwanami (2002) Nieman (2002) Price (2002*) Gonul (2003*) Combined

P300 Latency

Figure 3.1 P300 amplitude and P300 latency. Graph of standardized effect sizes for individual studies and pooled effect size obtained by meta-analysis. This forest plot shows the standardized effect sizes (SES) of each study on the P300 amplitude and 300 latency included in the meta-analysis. Studies conducted with drug-free patients are marked with *. The horizontal lines represent 95% con®dence interval (CI) for the SES in each individual study. The size of the squares represents the weights given to studies. The diamond shows the pooled standardized effect size (PSES) of all studies using meta-analysis tools in STATA 7. The PSES was 0.85 (95% CI: 0.65 to 1.05; p<0.001) for the P300 amplitude and ±0.57 (95% CI: ±0.75 to ±0.38; p<0.001) for the P300 latency. Thus, this meta-analysis shows that, compared to controls, patients have a statistically signi®cant and marked P300 amplitude reduction and P300 latency delay. (Reprinted from Schizophrenia Research, 70, Bramon et al. Meta-analysis of the P300 and P50 waveforms in schizophrenia, pp. 315±329. Copyright (2004), with permission from Elsevier.)

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

47

TABLE 3.2 Main results of the meta-analysis of P300 family studies P300 amplitude

P300 latency

Number of studies Number of participants

11 472 relatives / 513 controls

9 342 relatives / 409 controls

Meta-analysis pooled standardized effect size

PSES=0.61 95% CI: 0.30 to 0.91 p<0.001

PSES=±0.50 95% CI: ±0.88 to ±0.13 p=0.009

Heterogeneity test Egger's test of publication bias

p<0.001 p=0.008 Evidence of bias leading to overestimation of de®cits

p=0.02 p=0.08 Questionable evidence of bias leading to underestimation of de®cits.

PSES, pooled standardized effect size (weighted mean of standardized effect size SES for primary studies). PSES and SES are de®ned as Cohen's d: (control mean ± patient mean) / pooled standard deviation or (relative mean ± patient mean / pooled standard deviation). Reprinted from Neuroimage, 27, Bramon et al. Is the P300 wave an endophenotype for schizophrenia, pp. 960±968. Copyright (2005), with permission from Elsevier.

and 513 controls. As shown in Table 3.2 and Figure 3.2, compared to controls, the non-psychotic relatives have signi®cant reductions in P300 amplitude with a pooled effect size of 0.61 (95% CI 0.30 to 0.91, p<0.001). The P300 latency was signi®cantly delayed in relatives compared to controls with a pooled effect size of ±0.50 (95% CI: ±0.88 to ±0.13, p=0.009) (Bramon et al., 2005). This meta-analysis shows that the P300 de®cits exist in relatives and on the whole are of moderate severity. Unaffected relatives' performance is between that of patients and controls. It remains possible that only some relatives have P300 deviances. Further details are provided in Figure 3.2 and Table 3.2. De®cits in the P50 wave suppression response were also considered in a meta-analysis of case±control studies (Bramon et al., 2004b). Twenty studies were found to be suitable for analysis, which included 421 patients and 401 healthy volunteers. The P50 ratio was signi®cantly larger in the patients compared to healthy volunteers; the pooled effect size being 1.56 (95% CI ±2.05 to ±1.06, p<0.001). There were no signi®cant differences in the P50 latency. Three of the studies reported P50 ratio abnormalities in unmedicated patients (Jin et al., 1997, 1998; Myles-Worsley, 2002): thus, a confounding effect of antipsychotics seems unlikely and P50 deviances can be attributed to the disease. Further details are shown in Table 3.3 and Figure 3.3. A meta-analysis of family studies considering the P50 wave suppression could not be performed as there are not enough published studies. The limited existing data, however, indicates that the unaffected relatives of

B l a c kwood (1991) Kidogami (1991) Schreiber (1992) Roxborough (1993) Frangou (1997) Kimble (1999) Weisbrod (1999) Karoumi (2000) Turetsky (2000) B l a c kwood (2001) Winterer (2003)

Combined

–1

0

–2

–1

1

P300 Amplitude

2

Blackwood (1991) Kidogami (1991) Schreiber (1992) Roxborough (1993) Frangou (1997) Weisbrod (1999) Karoumi (2000) Turetsky (2000) Blackwood (2001)

Combined

P300 Latency

0

1

Figure 3.2 Forest plots of primary studies in unaffected relatives of patients with schizophrenia and main meta-analysis ®ndings. These forest plots show the standardized effect sizes (SES) of each primary study on the P300 amplitude or latency included in the meta-analysis. The horizontal lines represent 95% con®dence intervals (CI) for the SES in each individual study. The size of the squares represents the weights given to studies. The diamond shows the pooled standardized effect size (PSES) of all studies using meta-analysis tools. P300 amplitude PSES = 0.61 (95% CI: 0.30 to 0.91; p<0.001). P300 latency PSES = ±0.50 (95% CI: ±0.88 to ±0.13; p=0.009). Thus, this meta-analysis shows that, compared to controls, relatives have statistically signi®cant and moderate P300 amplitude reductions and P300 latency delays. (Reprinted from Schizophrenia Research, 70, Bramon et al. Metaanalysis of the P300 and P50 waveforms in schizophrenia, pp. 315±329. Copyright (2004), with permission from Elsevier.)

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

49

TABLE 3.3 Main results of meta-analyses of P50 case±control studies P50 ratio

P50 latency

Number of studies Number of participants

17 359 patients / 340 controls

12 316 patients / 281 controls

Main meta-analysis pooled standardized effect size

PSES=±1.56 95% CI: ±2.05 to ±1.06 p<0.001

PSES=0.08 95% CI: ±0.09 to 0.25 p=0.34

Heterogeneity test

Clear heterogeneity between studies in the P50 ratio p<0.001 No evidence of bias p=0.22

No evidence of heterogeneity for the P50 latency p=0.24 No evidence of bias p=0.70

Egger's test of publication bias

PSES, pooled standardized effect size (weighted mean of standardized effect size SES for primary studies). PSES and SES are de®ned as Cohen's d: (control mean ± patient mean) / pooled standard deviation or (relative mean ± patient mean / pooled standard deviation) Reprinted from Schizophrenia Research, 70, Bramon et al. Meta-analysis of the P300 and P50 waveforms in schizophrenia, pp. 315±329. Copyright (2004), with permission from Elsevier.

patients with schizophrenia also show reduced P50 suppression (Clementz et al., 1998; Myles-Worsley, 2002).

THE ACQUISITION AND ANALYSIS OF EEG/ERP DATA IN THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS Electroencephalograms (EEG) were collected from seventeen scalp sites according to the 10/20 International System (FP1, FP2, F7, F8, F3, F4, C3, C4, P3, P4, FZ, CZ, PZ, T3, T4, T5, T6) and grounded at FPZ using silver/ silver-chloride electrodes (Jasper, 1958). The left ear lobe served as reference and vertical, horizontal and radial electro-oculographs monitored eye movements. Data were continuously digitized at 500 Hz with a 0.03- to 120Hz band-pass ®lter (24 dB/octave roll-off ). Impedances were kept below 5 k . All tasks described were in the auditory modality and stimuli were delivered through bilateral intra-aural earphones. Data collection and analysis were conducted using Neuroscan software. P300 was assessed using an auditory oddball paradigm. Stimuli were 400 80-dB tones, with a 2-second inter-stimulus interval; 80% of the tones were `non-targets' of 1000 Hz and 20% were `targets' of 1500 Hz in a random sequence. Participants were instructed to press a button in response to targets only. EEG data were analysed at midline sites FZ, CZ and PZ. Standard procedures were employed to minimize ocular artefacts (Semlitsch

50

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Guterman (1994) Waldo (1994) Griffith (1995a) Griffith (1995b) Clementz (1997b) Jin (1997*) Clementz (1998a) Clementz (1998b) Jin (1998) final* Yee (1998) Boutros (1999) R oss (1999a) R oss (1999b) Olincy (2000) Clementz (2001) M yles-Worsley (2002) Kisley (2003)

C ombined

–6

–4

–2 P50 Ratio

0

2

Figure 3.3 Meta-analysis of the P50 ratio. Graph of standardized effect sizes for individual studies and pooled effect size obtained by meta-analysis. This forest plot shows the standardized effect sizes (SES) of each study on the P50 ratio included in the meta-analysis. Studies conducted with drug-free patients are marked with *. The horizontal lines represent 95% con®dence interval (CI) for the SES in each individual study. The size of the squares represents the weights given to studies. The diamond shows the pooled standardized effect size (PSES) of all studies using meta-analysis tools in STATA 7. The PSES for the P50 ratio was ±1.56 (95% CI: ±2.05 to ±1.06; p<0.001). Thus, this meta-analysis shows that, compared to controls, patients have a statistically signi®cant P50 ratio increase re¯ecting a severe brain-gating de®cit. (Reprinted from Schizophrenia Research, 70, Bramon et al. Metaanalysis of the P300 and P50 waveforms in schizophrenia, pp. 315±329. Copyright (2004), with permission from Elsevier.)

et al., 1986). Only epochs with correctly detected targets were included and these were further band-pass ®ltered 0.03±45 Hz. The P300 was identi®ed blind to affectedness group, via a computer algorithm, as the largest positive peak in the range 250±650 ms post-stimulus (Bramon et al., 2005; Frangou et al., 1997). The MMN wave was obtained via a two-tone duration task. Stimuli were 1200 tones of 80 dB intensity and 1000 Hz frequency, with a 0.3second inter-stimulus interval. Eighty-®ve per cent of the tones were `standards' of 25-ms duration and 15% were `deviants' of 50-ms duration. Participants were sitting comfortably in an armchair and were instructed to keep their eyes open and disregard the sounds presented to them.

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

51

RESULTS P300 wave abnormalities in the Maudsley Family Study sample The early sample of the Family Study (Phase 1; see Chapter 2 for description of study phases) compared auditory ERPs between thirty-three patients with schizophrenia from multiply affected families, ®fty-seven of their nonschizophrenic ®rst-degree relatives, and thirty-two unrelated normal controls. Data are shown in Table 3.4. Compared to controls, patients with schizophrenia showed signi®cant reduction in amplitude of the P300 wave at the CZ and PZ sites (p=0.0l and p<0.00l, respectively). The latency of the P300 waveform was also signi®cantly increased (p<0.00l at FZ, CZ and PZ). This con®rmed previous ®ndings of increased latency and reduced amplitude of the P300 wave in patients with schizophrenia (Roth & Cannon, 1972; St Clair et al., 1989). The group of unaffected relatives showed similar abnormalities to patients with schizophrenia. At PZ amplitude was reduced, but only at trend level, while the P300 latency was signi®cantly increased in all recording sites (p<0.001 at all sites). A strong positive correlation was also found between age and P300 latency in the schizophrenia group. The absence of a similar pattern in normal controls and relatives indicated that this age-related increase in P300 latency may be associated more with illness duration than ageing per se. This study therefore provided further evidence that abnormalities in the P300 response may serve as markers of genetic predisposition to schizophrenia. More recently, we collected new data from Phases 2 and 3 of the study, including both multiply-affected as well as single-proband families. This included thirty patients with a DSM-IV diagnosis of schizophrenia or schizoaffective disorder, forty of their non-psychotic ®rst-degree relatives and forty unrelated controls with no personal or family history of psychosis. For the purpose of these analyses, relatives from families with differing family histories were combined to maximize statistical power. Linear mixed models were used to estimate P300 amplitude or latency group differences adjusted by age and sex. The P300 amplitude was smaller in the patient group than in controls for all electrodes; however, this was only at trend level at FZ and CZ sites and was not signi®cant at PZ. Compared to controls, the P300 amplitude was not signi®cantly reduced in the relatives. Compared to controls, the P300 latency was delayed in patients as well as in their relatives in all three sites examined. Patient latency delays were signi®cant at FZ [Dif. ms=±32.0 (95% CI ±51.7 to ±12.1), p<0.01], CZ [Dif. ms=±38.1 (95% CI ±60.2 to ±16.0), p<0.01], and PZ [Dif. ms=±33.6 (95% CI ±58.7 to ±8.5), p=0.01]. Similarly, the latency delays for the relative

52

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

TABLE 3.4 The 1997 (Phase 1) and 2005 (Phases 2 and 3) Maudsley Family Study of Psychosis samples. Results for P300 amplitude and latency by group and region P300 amplitude mean (SD) V

P300 latency mean (SD) ms

Phase 1 P300 Family Study Results (Frangou et al., 1997) Site

Controls (n=32)

Relatives (n=57)

Patients (n=33)

Controls (n=32)

FZ CZ PZ

6.5 (7.9) 7.6 (5.6) 12.1 (3.2)

5.1 (7.1) 4.6 (5.3) 8.8 (3.7)*

2.9 (7.00)** 364 (39) 3.7 (3.6)** 363 (31) 5.4 (3.7)** 356 (22)

Relatives (n=57)

Patients (n=33)

413 (52)** 411 (49)** 417 (56)**

425 (53)** 423 (53)** 414 (36)**

Patients (n=30)

Phases 2 and 3 P300 Family Study Results (Bramon et al., 2005) Site

Controls (n=40)

Relatives (n=40)

Patients (n=30)

Controls (n=40)

Relatives (n=40)

FZ CZ PZ

6.9 (4.4) 10.6 (6.7) 13.9 (8.2)

7.0 (5.6) 8.8 (8.0) 13.2 (6.7)

4.7 (6.1) 7.7 (6.5) 11.4 (7.1)

328 (27) 332 (36) 347 (51)

369 (52) ** 348 (46) ** 372 (55) ** 360 (53) ** 387 (60) * 364 (58) **

The raw data are reported (mean of peak amplitude or latency). Statistical group comparisons reported are controls ± patients or controls ± relatives. * Signi®cant at 5% level. ** Signi®cant at 1% level. Adapted from data previously published in two papers: Reprinted from Neuroimage, 27, Bramon et al. Is the P300 wave an endophenotype for schizophrenia, pp. 960±968. Copyright (2005), and Schizophrenia Research, 23, Frangou et al. The Maudsley family study 2. Endogenous event-related potentials in familial schizophrenia, pp. 45±53. Copyright (1997) with permission from Elsevier.

group were signi®cant at PZ [Dif. ms=±24.8 (95% CI ±49.1 to ±0.45), p=0.05], FZ [Dif. ms=±29.3 (95% CI ±47.5 to ±11.1), p<0.01] and CZ [Dif. ms=±26.3 (95% CI ±46.7 to ±6.0), p=0.01]. The group averaged P300 waves are shown in Plate 1.

Does variation in the catechol-O-methyl transferase gene influence P300 endophenotypes? Two recent studies have reported an association between COMT genotype and P300 phenotypes, and constitute the ®rst descriptions of a precise genetic in¯uence on the P300 wave (Gallinat et al., 2003; Tsai et al., 2003). We therefore set out to replicate such an association in the Maudsley Family Study of Psychosis (Bramon et al., 2006). Participants were 189 Caucasian individuals, including 62 patients (of whom 56, 5 and 1 had DSM-IV diagnosis of schizophrenia, schizoaffective disorder and psychotic

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

53

disorder not otherwise speci®ed respectively), as well as 94 of their nonpsychotic ®rst-degree relatives and 33 unrelated healthy controls. The COMT Val158Met polymorphism was genotyped using the allelespeci®c assay described by Hoda et al. (1996), with failed and ambiguous calls repeated using a Single Nucleotide Primer Extension assay (Amersham International, Amersham, UK). Genotype frequencies were 53 Val/Val, 79 Val/ Met, and 57 Met/Met and, although it deviates from Hardy±Weinberg equilibrium (2 1df=5.1; p=0.024), it needs to be remembered this is a family study design, where the sample comprises non-independent observations (Gail et al., 2001). Unrelated healthy controls from our family study are in Hardy±Weinberg equilibrium (2 1df=0.61; p=0.43). Multilevel regression modelling was used to examine the effect of COMT genotype on P300 amplitude, adjusting by sex, age, scalp region (frontal or parietal) and group (patient, relative or control). No signi®cant interactions of genotype, group and region were found. There was no signi®cant main effect of COMT genotype on P300 amplitude [Dif.=±0.41 V (95% CI ±2.62 to 1.80); p=0.72]. Using equivalent analysis, we found no signi®cant main effect of COMT genotype on P300 latency [Dif.=7.83 ms (95% CI ±10.07 to 25.73); p=0.39]. Logistic regression analyses (adjusted for the family clusters in our data) showed that there was no signi®cant association between COMT genotype and schizophrenia [odds ratio (OR) Val/Met=1.4 (95% CI 0.7 to 2.8); p=0.32; OR Met/Met=0.9 (95% CI 0.5 to 1.8); p=0.76]. Thus, analysis of these data does not support the hypothesis that allelic variation at this functional polymorphism of the COMT gene is linked to an endophenotype of P300 abnormalities in schizophrenia.

Mismatch negativity abnormalities We investigated whether MMN is a potential marker of genetic vulnerability to schizophrenia by comparing this wave in a group of twenty-®ve patients with schizophrenia, thirty-seven of their unaffected relatives and twenty controls (Bramon et al., 2004a). As shown in Tables 3.5 and 3.6, at FZ the MMN amplitude was reduced by 19% in the patients compared to the unaffected group (which comprised the relatives and controls) and this reduction was statistically signi®cant (p<0.01). At FZ the unaffected relatives did not differ signi®cantly from controls. At F3 the mean MMN amplitude was reduced by 20.7% in the patients compared to the unaffected group and this difference was statistically signi®cant (p=0.01). At F3, the unaffected relatives did not differ signi®cantly from the controls. Finally, at the F4 location, linear regression analysis showed no signi®cant effect of group on the MMN amplitude. No correlation between MMN amplitude and dose of antipsychotic medication was found. As we had no data on drug-free patients, a medication effect in

54

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS TABLE 3.5 Mismatch negativity (MMN) amplitudes for three groups at F3, F4 and FZ

Mismatch negativity amplitude in V (SD) Electrode

Patients

Relatives

Controls

Patient's % decrease1

F3 FZ F4

±1.47 (0.76) ±1.61 (0.98) ±1.77 (0.96)

±1.86 (0.75) ±2.01 (0.94) ±1.93 (0.94)

±1.85 (0.72) ±1.98 (0.77) ±1.86 (0.71)

20.7* 19.3* 6.6

1

Percentage reduction in patients versus the unaffected participants (including both relatives and controls). * Signi®cant at 1% level.

Reprinted from Schizophrenia Research, 67, Bramon et al. Mismatch negativity in schizophrenia: A family study, pp. 1±10. Copyright (2004), with permission from Elsevier.

TABLE 3.6 Main results from the model at F3 and at FZ Electrode

Comparisons

Coef®cient

95% CI

F3

Patient versus unaffected Relatives versus controls Patients versus unaffected Relatives versus controls

±0.44 ±0.01 ±0.60 0.23

±0.79 ±0.54 ±1.01 ±0.26

FZ

to to to to

p-value ±0.10 0.51 ±0.20 0.72

0.01 0.95 <0.01 0.35

The table shows the results from the linear regression analyses comparing the MMN amplitude across groups. The analyses have accounted for the family clusters in the sample. `Unaffected' includes both relatives and controls. Data used here with kind permission from Elsevier. Reprinted from Schizophrenia Research, 67, Bramon et al. Mismatch negativity in schizophrenia: A family study, pp. 1±10. Copyright (2004), with permission from Elsevier.

this sample could not be ruled out. MMN group average waveforms are presented graphically in Figure 3.4.

DISCUSSION Are ERPs abnormalities potential endophenotypes for schizophrenia? For ERP abnormalities to represent potentially useful endophenotypes, they must meet the endophenotypic criteria set out in Chapter 1. Only reliable, inherited traits that occur among patients and their asymptomatic relatives represent useful endophenotypes.

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

55

+

0-

Controls Relatives

0Patients Controls

1.000 mV

50.00 ms – –100.00

–50.00

0.00

50.00

100.00

150.00

200.00

250.00

Figure 3.4 Mismatch negativity (MMN) group average waveforms. (Reprinted from Neuroimage, 27, Bramon et al. Is the P300 wave an endophenotype for schizophrenia, pp. 960±968. Copyright (2005), with permission from Elsevier.)

P300 The case for the use of P300 wave abnormalities as endophenotypes for psychosis continues to grow as an ever increasing number of papers are published. Promisingly, our group has observed high reliabilities, with an intra-class correlation co-ef®cient (ICC) of 0.86 for P300 amplitude and 0.87 for P300 latency (Hall et al., 2006b). A recent meta-analysis of twin studies reported a heritability of 60% for P300 amplitude and 51% for its latency (van Beijsterveldt & van Baal, 2002). From model ®tting in the Maudsley twin series the P300 amplitude heritability has been estimated to be 69%. For P300 latency, our results showed evidence for the presence of familiality (65%), but lack of power to distinguish between genetic and shared environmental in¯uences (Hall et al., 2006b). P300 amplitude and latency also had a signi®cant negative correlation (r=±0.40), which is entirely mediated by genetic factors. Thus, about half of the genes contributing to P300 amplitude also affect its latency (Hall et al., 2006b). Our meta-analyses of the existing literature con®rm that patients with schizophrenia have clear P300 abnormalities compared to healthy volunteers. The magnitude of these differences is comparable with the most robust ®ndings reported for brain morphometric and neuropsychological

56

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

abnormalities in schizophrenia (Bramon et al., 2004b). By means of a metaanalysis of family studies, we were also able to establish that P300 de®cits in both amplitude and latency exist in unaffected relatives of patients with schizophrenia, a fact that is crucial if these are to be used as endophenotypes in genetic studies of the disease (Bramon et al., 2005). The early sample of the Maudsley Family Study (with multiply-affected families only) demonstrated signi®cant de®cits in patients, and that the clinically unaffected ®rst-degree relatives of patients show a level of deviance that is intermediate between patients and controls, a fact that is consistent with a genetic relationship with P300 abnormalities (Frangou et al., 1997). A strong positive correlation was also found between age and P300 latency in the schizophrenia group and similar age-related prolongation of P300 latency had previously been reported in a group of male patients with schizophrenia (O'Donnell et al., 1995). It was ®rst thought that perhaps the age-related increase in P300 latency might re¯ect progressive cognitive deterioration either due to schizophrenia or to factors associated with chronicity or treatment. However, this is now felt not to be the case. Such age-related effects were not observed in a large meta-analysis of P300 traits in patients with schizophrenia and controls; in addition, it was reported that P300 latency in healthy participants increases naturally with age (Jeon & Polich, 2003). Our more recent sample of families with multiple and single cases of schizophrenia further con®rmed that the P300 latency was delayed in patients and in their relatives. Delayed P300 latency in unaffected relatives was not dependent on the presence of a strong family history in this sample, as it was identi®ed to a signi®cant degree in the mixed sample of relatives from singly- and multiply-affected families comprising Phases 2 and 3 of the study, as well as from the multiply-affected families in Phase 1 of the study. However, although amplitude reduction was observed in both patients and relatives, this was not found to be statistically signi®cant. Given the totality of existing evidence, we conclude that abnormalities of P300 amplitude and in particular latency constitute promising endophenotypes for schizophrenia.

Mismatch negativity The Maudsley twin series provided the ®rst twin study of MMN. This work showed that the MMN wave has moderate±high reliability, with intra-class correlation coef®cients of 0.67 and 0.66 for peak amplitude and mean amplitude respectively (Hall et al., 2006b). Furthermore, the MMN amplitude proved to be signi®cantly heritable with estimates of 63% and 68% for MMN peak amplitude and mean amplitude respectively (Hall et al., 2006b). In our family study, frontal and left frontal MMN was impaired in patients with schizophrenia compared to both their relatives and controls.

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

57

This ®nding is in agreement with several previous studies reporting similar abnormalities (Hirayasu et al., 1998; Javitt et al., 1995; Shelley et al., 1999) and a recent large-scale meta-analysis further supporting the view that MMN de®cits represent a trait marker of schizophrenia (Umbricht & Krljes, 2005). However, in contrast with the two previous studies (Jessen et al., 2001; Michie, 2001), we did not ®nd MMN de®cits in unaffected relatives of patients with schizophrenia. Our failure to replicate this ®nding in our larger sample perhaps suggests that any MMN amplitude reduction in relatives is subtle or only present in a subgroup. A more recent study agrees with our conclusion, reporting that MMN amplitude is reduced in relatives of patients with schizophrenia compared with controls, but only at trend level (Price et al., 2006). MMN de®cits do not seem to be in¯uenced by medication. Two groups (Catts et al., 1995; Javitt et al., 1995) have reported reduced MMN amplitudes in small groups of neuroleptic-free patients, whose MMN was indistinguishable from that of patients on medication. A larger and more recent study of thirty-one neuroleptic-free patients also found signi®cantly reduced MMN amplitudes compared with controls (Korostenskaja et al., 2005). This is in agreement with MMN studies with other atypical antipsychotics such as clozapine and risperidone in schizophrenia patients, which found no effects on MMN latencies and amplitudes (Umbricht et al., 1998, 1999). These ®ndings support the hypothesis that a decrease in MMN amplitude is likely to be associated with schizophrenia as an intrinsic feature of the disorder, rather than a medication effect. MMN generation is thought to index the functional state of N-methyl-Daspartate receptor (NMDAR)-mediated neurotransmission. When healthy volunteers are challenged with a NMDAR antagonist such as ketamine, not only do they develop transient psychotic symptoms but they also demonstrate reductions in MMN amplitude that correlate with the severity of the psychopathology induced (Umbricht et al., 2002). Given that glutamatergic neurotransmission plays a crucial role in the pathogenesis of schizophrenia (Hirsch et al., 1997) some researchers hypothesize that MMN amplitude de®cits may re¯ect vulnerability towards the disease (Umbricht et al., 2002). Our alternative interpretation is that MMN de®cits may correlate with the presence of symptoms of psychosis. As mentioned above, the MMN amplitude reductions reported in unaffected relatives of patients are controversial and we did not replicate them. Although Brockhaus-Dumke and colleagues (2005) found MMN amplitude reductions amongst people with prodromal symptoms of psychosis, these did not reach statistical signi®cance (Brockhaus-Dumke et al., 2005). Finally, it has also been observed that a fair number of patients in their ®rst episode of psychosis do not have a signi®cantly reduced MMN amplitude with the conventional two-tone duration or intensity paradigms

58

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

(Umbricht & Krljes, 2005). The available evidence suggests that MMN de®cits develop with illness progression and may index ongoing neuropathological changes in the auditory cortex in chronic schizophrenia. MMN amplitude reductions seem to be a `state' rather than a `trait' marker of psychosis. Thus, although further research is needed in longitudinal designs and in larger samples at risk, we conclude that MMN does not constitute a useful endophenotype for schizophrenia genetic research.

P50 Patients with schizophrenia and their ®rst-degree relatives show de®cits in the gating of the P50 electrophysiological response (Freedman et al., 1983; Waldo et al., 2000). Similar gating problems also appear in adolescents showing symptoms consistent with a heightened risk for imminent onset of psychosis (Myles-Worsley et al., 2004). Although no original data is presented in this chapter, our meta-analysis of the literature strengthens the case for the use of P50 wave suppression as a potential endophenotype (Bramon et al., 2004b). Furthermore, the P50 T/C ratio is reliable (intraclass correlation co-ef®cient of 0.66) and has a high heritability (estimated as 68%), all of which supports the use of this measure as an endophenotype (Hall et al., 2006b). Heinrichs (2004) conducted a meta-analysis comparing an extensive number of biological markers put forward in schizophrenia research, including brain structural changes, memory, word generation, eye tracking and intellectual ability. Heinrichs concluded that most neurobiological markers show modest differences between cases and controls and cognitive de®cits and in particular the P50 wave are the strongest and most promising endophenotypes for the disease. Table 3.7 summarizes the evidence for how well each of the P300, MMN and P50 wave abnormalities ful®ll endophenotypic criteria for schizophrenia.

Genetic influences on neurophysiological endophenotypes Although the search for genetic factors implicated in schizophrenia has been a slow and painstaking process, progress continues to be made. Linkages to several chromosomal regions are being replicated and a number of candidate genes have been put forward, including catechol-O-methyltransferase (COMT ), disrupted in schizophrenia-1 (DISC1), alpha-7 nicotinic cholinergic receptor (CHRNA7 ), neuregulin (NRG1) and dysbindin (DTNBP1). While these are promising, with some plausible pathophysiological mechanisms, they have not received clear replication and con®rmation (Craddock et al., 2006). Using alternative phenotypes has been proposed as a potential

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

59

TABLE 3.7 Summary of the most plausible neurophysiologic endophenotypes for psychosis

Heritability estimates Reliability (ICC) De®cits in patients

De®cits in relatives

P300

MMN

*Amplitude 60% *Latency 51% (van Beijsterveldt & van Baal, 2002) Very high: Amplitude 0.86 Latency 0.88 (Hall et al., 2006b) *Severe amplitude reductions and moderate latency delays (Bramon et al., 2004b) *Amplitude reductions and latency delays of moderate severity (Bramon et al., 2004b)

Peak amplitude 63% T/C ratio 68% Mean amplitude 68% (Hall et al., 2006b) (Hall et al., 2006b) Moderately high: Peak amplitude 0.67 Mean amplitude 0.66 (Hall et al., 2006b) *Severe amplitude reductions (Umbricht & Krljes, 2005)

P50

Moderately high: T/C ratio 0.66 (Hall et al., 2006b) *Severe P50 ratio increases (Bramon et al., 2004b)

De®cits reported Con¯icting evidence (Myles-Worsley, 2002) (Bramon et al., 2004a; Jessen et al., 2001; Michie, 2001)

ICC, intra-class correlation coef®cient. *The evidence is supported by a meta-analysis of the literature.

solution; however, to date, only a few studies have examined the relationship between genetics and ERP endophenotypes for schizophrenia. Blackwood and colleagues (2001) identi®ed a large Scottish family carrying a balanced translocation affecting the DISC1 gene and with a very high prevalence of psychiatric morbidity. The P300 amplitude was reduced in patients and relatives who carried the translocation compared to relatives with a normal karyotype. Furthermore, the amplitude reduction was independent of the presence or absence of symptoms because asymptomatic translocation carriers showed similar P300 amplitude reduction as was found in translocation carriers who were diagnosed with schizophrenia, bipolar disorder or unipolar depression. According to Blackwood and colleagues, the genes DISC1 and DISC2, which are directly disrupted by the breakpoint on chromosome 1, may have a role in the aetiology of psychosis spectrum disorders. The effect of COMT genotype on P300 traits has been investigated in two recent studies. In a sample of healthy Chinese females, COMT Met/Met carriers had earlier P300 latency (signi®cant in all three sites FZ, CZ and PZ) and increased P300 amplitude (non-signi®cant). This agrees with the general hypothesis that the Met/Met COMT genotype results in enhanced cognition, when measured with neurophysiological methods (Tsai et al., 2003). Conversely, another study with European participants reported that

60

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

reductions in frontal P300 amplitude were associated with inheriting the Met/Met COMT genotype; this association was signi®cant in a group of 49 patients with schizophrenia and showed a similar direction (although not signi®cant) in 170 healthy controls. However, inheriting the Met/Met genotype did not in¯uence P300 amplitudes at central or parietal sites. The smaller P300 wave in frontal regions was interpreted as neuronal noise reductions resulting from enhanced dopaminergic transmission in Met/Met carriers. The authors also argued that central and parietal P300 re¯ect stable signals (not noise) hence explaining their different ®ndings by region (Gallinat et al., 2003). Our family study failed to replicate this previous work and we found no association between COMT genotype and P300 amplitude or latency performance (Bramon et al., 2006). Of course, nonreplications raise concerns about power. Based on the data of Gallinat et al. and Tsai et al., the effect size for the COMT in¯uence on frontal P300 amplitude is approximately 1.55 for patients and 0.32 for controls (one-way ANOVA, three groups,  = 5%). Thus, our sample provided 79% power to replicate associations in controls and over 95% power in patients or in the pooled sample. P300 paradigms in schizophrenia research are not standardized and this study, as well as the two preceding ones, had some variations in neurophysiological procedures. There were also clear ethnic, clinical and demographic study differences, which are likely to modulate any possible association between COMT genotype and P300. If such an association exists it is likely to be very subtle indeed. Freedman and colleagues were the ®rst to report that markers in the -7 nicotinic receptor gene (CHRNA7 ) at 15q13-14 are strongly linked with the P50 gating de®cits. They also found associations between several polymorphisms here, including variants in the promoter region, and these sensory gating de®cits (Freedman et al., 1997; Houy et al., 2004; Raux et al., 2002). The CHRNA7 receptor has been a plausible candidate for involvement in these P50 gating de®cits for some time. Nicotine in high doses transiently normalizes the abnormality in P50 inhibition in patients with schizophrenia and in their relatives, much as it normalizes inhibition in rat models (Adler et al., 1992, 1993; Bickford & Wear, 1995). Patients with schizophrenia are particularly heavy tobacco smokers, even when compared with other psychiatric patients, and this heavy nicotine use may re¯ect an attempt at self-medication of an endogenous neuronal gating de®cit (Goff et al., 1992). The above ®ndings support the role of the -7-nicotinic receptor in the pathophysiology of sensory and attentional disturbance in schizophrenia. While such associations are of course encouraging, there are many questions still to be answered in the dissection of the genetics of psychosis. One point to consider is whether the genes in¯uencing each ERP endophenotype are at least partially distinct from each other. A recent twin

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

61

study has used multivariate analyses to investigate the genetic and environmental relationship between P300, MMN and P50 suppression (Hall et al., 2006a). This study provided no evidence of genetic overlap between these three ERPs. Consequently, it is felt that each paradigm serves to evaluate different brain information processing functions that may be mediated by distinct neurobiological mechanisms, which in turn are in¯uenced by different sets of genes.

Are ERP endophenotypes specific to schizophrenia? A common criticism of neurophysiological endophenotypes is their lack of speci®city. Indeed, abnormalities in the P300 can be found in depression (Pfefferbaum et al., 1984), alcoholism (Pfefferbaum et al., 1991) and dementia (Pfefferbaum et al., 1989). In one of the largest studies of auditory P300 in affective disorders, both P300 latency and P300 amplitude were found to be impaired in patients with schizophrenia or bipolar depression compared to patients with unipolar affective disorders and normal controls (Muir et al., 1991). Decreased P300 amplitude and delayed P300 latency has also been reported in relatives of bipolar patients (Pierson et al., 2000), indicating that these may also serve as endophenotypes for this disease. MMN has also been observed to be reduced in other clinical conditions, such as in children with learning disabilities (Kraus et al., 1996), adults with dyslexia (Baldeweg et al., 1999; Kujala et al., 2000) and patients with Alzheimer's disease (Pekkonen et al., 1994). However, among the major psychiatric illnesses, MMN seems to be relatively speci®c to schizophrenia. A recent study demonstrated that among patients with schizophrenia, schizoaffective disorder, bipolar disorder, and major depression, de®cits in MMN generation were observed exclusively in patients with schizophrenia or schizoaffective disorder (Umbricht et al., 2003). Although abnormal P50 suppression has most frequently been related to psychosis, it has also been reported in patients with migraine (Ambrosini et al., 2001), bipolar disorder with psychosis (Olincy & Martin, 2005) and cocaine abuse (Boutros et al., 2000). Acute psychological stress has also been shown to disrupt P50 suppression in healthy participants (White & Yee, 1997; Yee & White, 2001), while pain has also been associated with transient loss of suppression of the P50 response (Johnson & Adler, 1993). Variation in the CHRNA7 gene is also relevant to epilepsy (Riley et al., 2002). While we know that ERP de®cits are not speci®c to schizophrenia, it is also true that a number of the candidate genes potentially involved may not be unique to schizophrenia. Initial evidence indicates that a number of the most promising loci, including DISC1, NRG1, RSG4, G72, also show

62

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

evidence of linkage and/or association with bipolar disorder (Badner & Gershon, 2002; Berrettini, 2004; Ophoff et al., 2002; Potash et al., 2003), although it remains unclear whether these associations are limited to psychotic forms of bipolar illness. Certainly, from a genetic point of view, the two illnesses may be far from independent (Bramon & Sham, 2004). To the extent that genes causing event related potential abnormalities in schizophrenia are shared with other neuropsychiatric disorders, some degree of overlap in endophenotypes is to be expected. Thus the non-speci®city of neurophysiological de®cits is not simply a limitation but a re¯ection of the common genetic architecture underlying the functional psychoses.

Is the move towards multivariate endophenotypes the future? Cannon and Keller (2006) proposed a model where a number of potential schizophrenia endophenotypes may in¯uence one another. The pathways that lead from genes through different levels of phenotypes are described using a watershed analogy. Much like the numerous tributaries that eventually coalesce into a major river, many upstream microbiological processes (e.g. dopaminergic regulation in the prefrontal cortex) in¯uence further downstream macrobiological processes (e.g. working memory). A slightly harmful mutation that dysregulates dopamine in the prefrontal cortex may not affect brain function generally but will probably undermine speci®c downstream processes such as working memory and, through this, increase the risk for schizophrenia. Cannon and Keller further argue that there is convergent evidence of genetic mediation across multiple levels of analysis, including anatomical, physiological, functional, and behavioural. A substantial degree of overlap appears likely for a number of the most promising genes associated with schizophrenia, including neuregulin, dysbindin, DISC1, G72, and RGS4, given these genes impinge on common cellular signalling pathways associated with glutamatergic and GABAergic neurotransmission (Cannon & Keller, 2006). Although the complex cascading model proposed by Cannon and Keller seems plausible, the neurophysiological endophenotypes examined to date appear to be independent and governed by distinct underlying mechanisms. As mentioned above, our group found no genetic or environmental correlations between P300, MMN and P50 traits (Hall et al., 2006a). Similarly, no correlation was found between sensory gating de®cits in schizophrenia patients at different stages of information processing including pre-attentive (P50 suppression), early attentive (N100), later attentive (P200), and attention-dependent P300 components (Boutros et al., 2004). Also, studies aiming to investigate the relationship between various inhibition tasks such as between pre-pulse inhibition and antisaccade performance (Cadenhead et

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

63

al., 2002; Kumari et al., 2005); between eye tracking (smooth pursuit and antisaccade) and P50 inhibition (Louchart-de la Chapelle et al., 2005); and between P50 suppression and pre-pulse inhibition (Oranje et al., 2006; Schwarzkopf et al., 1993) have shown little or no evidence of a relationship between these measures. Naturally, failing to ®nd correlations between endophenotypes does not prove they re¯ect independent processes. Indeed, the studies available may be underpowered to ®nd weak, yet clinically meaningful, correlations. Price et al. (2006) considered four ERP measures ± P300, MMN, P50 and antisaccade (AS) ± in a family study design. Previous ®ndings for each of the endophenotypes were replicated, with patients with schizophrenia differing from controls, in the expected direction, for each ERP measure. The relative group means lay between those of the control and patient groups but were closer to the latter. Again the P300, MMN and P50 measures were not correlated, which agrees with the twin work from our group (Hall et al., 2006a). With a weighted combination of the above four electrophysiological markers, Price and colleagues (2006) generated a multivariate endophenotype that could discriminate patient versus control groups more effectively than any individual measure (Price et al., 2006). Re-assigning participants on the basis of a multivariate endophenotype seemed to result in signi®cantly greater group homogeneity. The clinical validity of such analysis depends, in exactly the same way as for single phenotypic traits, on the unknown links between the endophenotypes and schizophrenia candidate genes. However, if these links are real, then a multivariate electrophysiological endophenotype may allow more powerful analysis of smaller data sets than diagnosisbased or single endophenotype-based studies. There is virtually no research examining the relationship between endophenotypes obtained through different techniques; for example, neuropsychology or structural MRI and EEG. Each individual endophenotype may tackle a speci®c aspect of cognitive processing or neurobiological development. Furthermore, they may even be governed by distinct sets of genes. However, when several endophenotypes are studied concomitantly, highrisk individuals (i.e. patients and relatives) who exhibit abnormalities on multiple measures may carry a higher genetic loading than individuals who display fewer or no de®cits. Therefore, despite the lack of genetic overlap, collecting various physiological measurements from the same individuals may increase the power of linkage and association studies and thus lead to a greater understanding of the genetics of psychotic disorders.

CONCLUSION Neurophysiological traits are amongst the most plausible endophenotypes for psychosis. They are non-invasive and easily obtained measures which

64

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

are both highly heritable and reliable. The de®cits we observed amongst patients with schizophrenia and their relatives in our sample strengthen the case for their use in this way. More promising still are the ®rst reports of positive association with polymorphisms in schizophrenia candidate genes. The P300 amplitude reductions and latency delays as well as the impairments in P50 gating appear to be the most informative measures for genetic studies in the future. It is our hope that an endophenotype-based approach such as this has the potential to assist in the genetic dissection of schizophrenia and ultimately increase the understanding of how cognitive functioning is affected in patients with psychosis.

REFERENCES Adler, L.E., Pachtman, E., Franks, R.D., Pecevich, M., Waldo, M.C., Freedman, R. (1982). Neurophysiological evidence for a defect in neuronal mechanisms involved in sensory gating in schizophrenia. Biological Psychiatry, 17(6), 639±654. Adler, L.E., Hoffer, L.J., Grif®th, J., Waldo, M.C., Freedman, R. (1992). Normalization by nicotine of de®cient auditory sensory gating in the relatives of schizophrenics. Biological Psychiatry, 32(7), 607±616. Adler, L., Hoffer, L., Wiser, A., Freedman, R. (1993). Normalisation of auditory physiology by cigarette smoking in schizophrenic patients. American Journal of Psychiatry, 150, 1856±1861. Alain, C., Hargrave, R., Woods, D. (1998). Processing of auditory stimuli during visual attention in patients with schizophrenia. Biological Psychiatry 44, 1151±1159. Ambrosini, A., De Pasqua, V., Afra, J., Sandor, P.S., Schoenen, J. (2001). Reduced gating of middle-latency auditory evoked potentials (P50) in migraine patients: another indication of abnormal sensory processing? Neuroscience Letters, 306(1±2), 132±134. Badner, J.A., Gershon, E.S. (2002). Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia. Molecular Psychiatry, 7(4), 405±411. Baldeweg, T., Richardson, A., Watkins, S., Foale, C., Gruzelier, J. (1999). Impaired auditory frequency discrimination in dyslexia detected with mismatch evoked potentials. Annals of Neurology, 45(4), 495±503. Berrettini, W. (2004). Bipolar disorder and schizophrenia ± Convergent molecular data. Neuromolecular Medicine, 5(1), 109-117. Bickford, P.C., Wear, K.D. (1995). Restoration of sensory gating of auditory evoked response by nicotine in ®mbria-fornix lesioned rats. Brain Research, 705(1±2), 235±240. Blackwood, D.H.R., Muir, W.J. (2004). Clinical phenotypes associated with DISC1, a candidate gene for schizophrenia. Neurotoxicity Research, 6(1), 35±41. Blackwood, D., St Clair, D., Muir, W., Duffy, J. (1991). Auditory P300 and eye tracking dysfunction in schizophrenic pedigrees. Archives of General Psychiatry, 48, 899±909. Blackwood, D.H.R., Fordyce, A., Walker, M.T., St Clair, D.M., Porteous, D.J., Muir, W.J. (2001). Schizophrenia and affective disorders ± Cosegregation with a translocation at chromosome 1q42 that directly disrupts brain-expressed genes: Clinical and P300 ®ndings in a family. American Journal of Human Genetics, 69(2), 428±433. Boutros, N.N., Zouridakis, G., Overall, J. (1991). Replication and extension of P50 ®ndings in schizophrenia. Clinical Electroencephalography, 22(1), 40±45. Boutros, N., Campbell, D., Petrakis, I., Krystal, J., Caporale, M., Kosten, T. (2000). Cocaine use and the mid-latency auditory evoked responses. Psychiatry Research, 96(2), 117±126. Boutros, N.N., Korzyukov, O., Jansen, B., Feingold, A., Bell, M. (2004). Sensory gating

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

65

de®cits during the mid-latency phase of information processing in medicated schizophrenia patients. Psychiatry Research, 126(3), 203±215. Bramon, E., Sham, P. (2004). The shared genetic architecture which underlies schizophrenia and bipolar disorder. In: C. McDonald, K. Schulze, R. Murray, P. Wright (Eds.), Schizophrenia, Challenging the Orthodox. (pp 173±181). London: Taylor & Francis. Bramon, E., Rabe-Hesketh, S., Murray, R., Frangou, S. (2001). Meta-analysis of the P300 and P50 waveforms in schizophrenia. Schizophrenia Research, 59(1±2, suppl), 198. Bramon, E., Croft, R., McDonald, C., Virdi, G., Gruzelier, J., Baldeweg, T., et al. (2004a). Mismatch negativity in schizophrenia: A family study. Schizophrenia Research, 67, 1±10. Bramon, E., Rabe-Hesketh, S., Sham, P., Murray, R., Frangou, S. (2004). Meta-analysis of the P300 and P50 waveforms in schizophrenia. Schizophrenia Research, 70(2±3), 315±329. Bramon, E., McDonald, C., Croft, R.J., Landau, S., Filbey, F., Gruzelier, J.H., et al. (2005). Is the P300 wave an endophenotype for schizophrenia? A meta-analysis and a family study. Neuroimage, 27(4), 960±968. Bramon, E., Dempster, E., Frangou, S., McDonald, C., Schoenberg, P., MacCabe, J.H., et al. (2006). Is there an association between the COMT gene and P300 endophenotypes? European Psychiatry, 21(1), 70±73. Brockhaus-Dumke, A., Tendolkar, I., Pukrop, R., Schultze-Lutter, F., Klosterkotter, J., Ruhrmann, S. (2005). Impaired mismatch negativity generation in prodromal subjects and patients with schizophrenia. Schizophrenia Research, 73(2±3), 297±310. Cannon, T.D., Keller, M.C. (2006). Endophenotypes in the genetic analyses of mental disorders. Annual Review of Clinical Psychology, 2, 267±290. Cadenhead, K.S., Light, G.A., Geyer, M.A., McDowell, J.E., Braff, D.L. (2002). Neurobiological measures of schizotypal personality disorder: De®ning an inhibitory endophenotype? American Journal of Psychiatry, 159(5), 869±871. Catts, S.V., Shelley, A.-M., Ward, P.B., Liebert, B., McConaghy, N., Andrews, S., Michie, P.T. (1995). Brain potential evidence for an auditory sensory memory de®cit in schizophrenia. American Journal of Psychiatry, 152(2), 213±219. Clementz, B., Geyer, M., Braff, D. (1998). Poor P50 suppression among schizophrenia patients and their ®rst-degree biological relatives. American Journal of Psychiatry, 155(12), 1691±1694. Cohen, J. (1969). Statistical Power Analysis for the Behavioural Sciences. New York: Academic Press. Craddock, N., O'Donovan, M.C., Owen, M.J. (2006). Genes for schizophrenia and bipolar disorder? Implications for psychiatric nosology. Schizophrenia Bulletin, 32(1), 9±16. Cullum, C., Harris, J., Waldo, M., Smernoff, E., Madison, A., Nagamoto, H., et al. (1993). Neurophysiological and neuropsychological evidence for attentional dysfunction in schizophrenia. Schizophrenia Research, 10 August(2), 131±141. Donchin, E., Coles, M.G.H. (1988). Is the P300 component a manifestation of context updating? Behavioral and Brain Sciences, 11(3), 357±374. Erwin, R.J., Turetsky, B.I., Moberg, P., Gur, R.C., Gur, R.E. (1998). P50 abnormalities in schizophrenia: relationship to clinical and neuropsychological indices of attention. Schizophrenia Research, 33, 157±167. Frangou, S., Sharma, T., Alarcon, G., Sigmundsson, T., Takei, N., Binnie, C., Murray, R.M. (1997). The Maudsley Family Study, II: Endogenous event-related potentials in familial schizophrenia. Schizophrenia Research, 23, 45±53. Freedman, R., Adler, L., Waldo, M., Pachtman, E., Franks, R. (1983). Neurophysiological evidence for a defect in inhibitory pathways in schizophrenia: Comparison of medicated and drug-free patients. Biological Psychiatry, 18, 537±551. Freedman, R., Coon, H., Myles-Worsley, M., Orr-Urtreger, A., Olincy, A., Davis, A., et al.

66

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

(1997). Linkage of a neurophysiologtical de®cit in schizophrenia to a chromosome 15 locus. Procedures of the National Academy of Science, 94, 587±592. Gail, M.H., Pee, D., Carroll, R. (2001). Effects of violations of assumptions on likelihood methods for estimating the penetrance of an autosomal dominant mutation from kin-cohort studies. Journal of Statistical Planning and Inference, 96(1), 167±177. Gallinat, J., Bajbouj, M., Sander, T., Schlattmann, P., Xu, K., Ferro, E.F., et al. (2003). Association of the G1947A COMT (Val(108/158)Met) gene polymorphism with prefrontal P300 during information processing. Biological Psychiatry, 54(1), 40±48. Goff, D.C., Henderson, D.C., Amico, E. (1992). Cigarette smoking in schizophrenia: relationship to psychopathology and medication side effects. American Journal of Psychiatry, 149(9), 1189±1194. Hall, M., Schulze, K., Bramon, E., Murray, R.M., Sham, P., Rijsdijk, F. (2006). Genetic overlap between P300, P50 and duration mismatch negativity. American Journal of Medical Genetics (Neuropsychiatric Genetics), 141B, 336±343. Hall, M.H., Schulze, K., Rijsdijk, F., Picchioni, M., Ettinger, U., Bramon, E., et al. (2006). Heritability and reliability of P300, P50 and duration mismatch negativity. Behavioral Genetics, 36(6), 845±857. Heinrichs, R.W. (2004). Meta-analysis, and the science of schizophrenia: variant evidence or evidence of variants? Neuroscience and Biobehavioral Reviews, 28(4), 379±394. Hirayasu, Y., Asato, N., Ohta, H., Hokama, H., Arakaki, H., Ogura, C. (1998). Abnormalities of auditory event-related potentials in schizophrenia prior to treatment. Biological Psychiatry, 43, 244±253. Hirsch, S.R., Das, I., Garey, L.J., deBelleroche, J. (1997). A pivotal role for glutamate in the pathogenesis of schizophrenia, and its cognitive dysfunction. Pharmacology Biochemistry and Behavior, 56(4), 797±802. Hoda, F., Nicholl, D., Bennett, P., Arranz, M., Aitchison, K.J., AlChalabi, A., et al. (1996). No association between Parkinson's disease and low-activity alleles of catechol Omethyltransferase. Biochemical and Biophysical Research Communications, 228(3), 780±784. Houy, E., Raux, G., Thibaut, F., Belmont, A., Demily, C., Allio, G., et al. (2004). The promoter-194 C polymorphism of the nicotinic alpha 7 receptor gene has a protective effect against the P50 sensory gating de®cit. Molecular Psychiatry, 9(3), 320±322. Jasper, H. (1958). Report to the committee on methods of clinical examination in electroencephalography. Electroencephalography and Clinical Neurophysiology, 10, 371±375. Javitt, D.C. (2000). Intracortical mechanisms of mismatch negativity dysfunction in schizophrenia. Audiology and Neuro-Otology, 5(3±4), 207±215. Javitt, D.C., Doneshka, P., Grochowski, S., Ritter, W. (1995). Impaired mismatch negativity generation re¯ects widespread dysfunction of working memory in schizophrenia. Archives of General Psychiatry, 52, 550±558. Javitt, D., Steinschneider, M., Schroeder, C., Arezzo, J. (1996). Role of cortical N-methyl-Daspartate receptors in auditory sensory memory and mismatch negativity generation: Implications for schizophrenia. Procedures of the National Academy of Science, 93, 11962± 11967. Jeon, Y.W., Polich, J. (2001). P300 asymmetry in schizophrenia: A meta-analysis. Psychiatry Research, 104(1), 61±74. Jeon, Y.W., Polich, J. (2003). Meta-analysis of P300 and schizophrenia: Patients, paradigms, and practical implications. Psychophysiology, 40(5), 684±701. Jessen, F., Fries, T., Kucharski, C., Nishimura, T., Hoenig, K., Maier, W., et al. (2001). Amplitude reduction of the mismatch negativity in ®rst-degree relatives of patients with schizophrenia. Neuroscience Letters, 309(3), 185±188. Jin, Y., Potkin, S.G., Patterson, J.V., Sandman, C.A., Hetrick, W.P., Bunney, J.W.E. (1997).

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

67

Effects of P50 temporal variability on sensory gating in schizophrenia. Psychiatry Research, 70, 71±81. Jin, Y., Bunney, J.W.E., Sandman, C.A., Patterson, J.V., Fleming, K., Moenter, J.R., Kalali, A.H. (1998). Is P50 suppression a measure of sensory gating in schizophrenia? Biological Psychiatry, 43, 873±878. Johnson, M.R., Adler, L.E. (1993). Transient impairment in P50 auditory sensory gating induced by a cold-pressor test. Biological Psychiatry, 33(5), 380±387. Johnson, R., Jr, Pfefferbaum, A., Kopell, B.S. (1985). P300 and long-term memory: Latency predicts recognition performance. Psychophysiology, 22(5), 497±507. Kathmann, N., Wagner, M., Rentsorff, N., Engel, R. (1995). Delayed peak latency of the mismatch negativity in schizophrenics and alcoholics. Biological Psychiatry, 37, 754±757. Kidogami, Y., Yoneda, H., Asaba, H., Sakai, T. (1991). P300 in ®rst degree relatives of schizophrenics. Schizophrenia Research, 6(1), 9±13. Kirino, E., Inoue, R. (1999). The relationship of mismatch negativity to quantitative EEG and morphological ®ndings in schzophrenia. Journal of Psychiatric Research, 33, 445±456. Korostenskaja, M., Dapsys, K., Siurkute, A., Maciulis, V., Ruksenas, O., Kahkonen, S. (2005). Effects of olanzapine on auditory P300 and mismatch negativity (MMN) in schizophrenia spectrum disorders. Progress in Neuropsychopharmacology and Biological Psychiatry, 29(4), 543±548. Kraus, N., McGee, T.J., Carrell, T.D., Zecker, S.G., Nicol, T.G., Koch, D.B. (1996). Auditory neurophysiologic responses and discrimination de®cits in children with learning problems. Science, 273, 971±973. Kujala, T., Myllyviita, K., Tervaniemi, M., Alho, K., Kallio, J., Naatanen, R. (2000). Basic auditory dysfunction in dyslexia as demonstrated by brain activity measurements. Psychophysiology, 37(2), 262±266. Kumari, V., Ettinger, U., Crawford, T.J., Zachariah, E., Sharma, T. (2005). Lack of association between prepulse inhibition and antisaccadic de®cits in chronic schizophrenia: implications for identi®cation of schizophrenia endophenotypes. Journal of Psychiatric Research, 39(3), 227±240. Kutas, M., McCarthy, G., Donchin, E. (1977). Augmenting mental chronometry: The P300 as a measure of stimulus evaluation time. Science, 197(4305), 792±795. Louchart-de la Chapelle, S., Nkam, I., Houy, E., Belmont, A., Menard, J.F., Roussignol, A.C., et al. (2005). A concordance study of three electrophysiological measures in schizophrenia. American Journal of Psychiatry, 162(3), 466±474. Matsumoto, K., Yamazaki, H., Nakamura, M., Sakai, H., Miura, N., Kato, T., (2005). Reduced word-repetition effect in the event-related potentials of thought-disordered patients with schizophrenia. Psychiatry Research, 134(3), 225±231. McGhie, A., Chapman, J. (1961). Disorders of attention and perception in early schizophrenia. British Journal of Medical Psychology, 34, 103±116. Michie, P.T. (2001). What has MMN revealed about the auditory system in schizophrenia? International Journal of Psychophysiology, 42(2), 177±194. Michie, P.T., Innes-Brown, H., Todd, J., Jablensky, A. (2002). Duration of MMN in biological relatives of patients with schizophrenia spectrum disorders. Biological Psychiatry, 52(7), 749±758. Muir, W., St Clair, D., Blackwood, D. (1991). Long-latency auditory event-related potentials in schizophrenia and in bipolar and unipolar affective disorder. Psychological Medicine, 21, 867±879. Myles-Worsley, M. (2002). P50 sensory gating in multiplex schizophrenia families from a Paci®c island isolate. American Journal of Psychiatry, 159(12), 2007±2012. Myles-Worsley, M., Ord, L., Blailes, F., Ngiralmau, H., Freedman, R. (2004). P50 sensory

68

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

gating in adolescents from a Paci®c island isolate with elevated risk for schizophrenia. Biological Psychiatry, 55(7), 663±667. O'Donnell, B., Hokama, H., McCarley, R., Smith, R., Salisbury, D., Mondrow, E., et al. (1994). Auditory ERPs to non-target stimuli in schizophrenia: Relationship to probability, task demands, and target ERPs. International Journal of Psychophysiology, 17, 219±231. O'Donnell, B.F., Faux, S.F., McCarley, R.W., Kimble, M.O., Salisbury, D.F., Nestor, P.G., et al. (1995). Increased rate of P300 latency prolongation with age in schizophrenia. Archives of General Psychiatry, 52, 544±549. Oades, R., Dittmann-Balcar, A., Zerbin, D., Grzella, I. (1997). Impaired attention-dependent augmentation of MMN in nonparanoid vs paranoid schizophrenic patients: A comparison with obsessive-compulsive disorder and healthy subjects. Biological Psychiatry, 41, 1196±1210. Olincy, A., Martin, L. (2005). Diminished suppression of the P50 auditory evoked potential in bipolar disorder subjects with a history of psychosis. American Journal of Psychiatry, 162(1), 43±49. Olney, J.W., Farber, N.B. (1995). Glutamate receptor dysfunction and schizophrenia. Archives of General Psychiatry, 52(12), 998±1007. Ophoff, R.A., Escamilla, M.A., Service, S.K., Spesny, M., Meshi, D.B., Poon, W., et al. (2002). Genomewide linkage disequilibrium mapping of severe bipolar disorder in a population isolate. American Journal of Human Genetics, 71(3), 565±574. Oranje, B., Geyer, M.A., Bocker, K.B., Leon Kenemans, J., Verbaten, M.N. (2006). Prepulse inhibition and P50 suppression: Commonalities and dissociations. Psychiatry Research, 143(2±3), 147±158. Pekkonen, E., Jousmaki, V., Kononen, M., Reinikainen, K., Partanen, J. (1994). Auditory sensory memory impairment in Alzheimer's disease: An event-related potential study. Neuroreport, 5(18), 2537±2540. Pfefferbaum, A., Wenegrat, B., Ford, J., Roth, W., Kopell, B. (1984). Clinical applications of the P3 component of event-related potentials II. Dementia, depression and schizophrenia. Electroencephalography and Clinical Neurosphysiology, 59, 104±124. Pfefferbaum, A., Ford, J.M., White, P.M. (1989). P3 in schizophrenia is affected by stimulus modality, response requirements medication status and negative symptoms. Archives of General Psychiatry, 46, 1935±1044. Pfefferbaum, A., Ford, J., White, P., Mathalon, D. (1991). Event-related potentials in alcoholic men: P3 amplitude re¯ects family history but not alcoholic consumption. Alcoholism: Clinical and Experimental Research, 15, 839±850. Pierson, A., Jouvent, R., Quintin, P., Perez-Diaz, F., Leboyer, M. (2000). Information processing de®cits in relatives of manic depressive patients. Psychological Medicine, 30(3), 545±555. Potash, J.B., Zandi, P.P., Willour, V.L., Lan, T.H., Huo, Y., Avramopoulos, D., et al. (2003). Suggestive linkage to chromosomal regions 13q31 and 22q12 in families with psychotic bipolar disorder. American Journal of Psychiatry, 160(4), 680±686. Price, G.W., Michie, P.T., Johnston, J., Innes-Brown, H., Kent, A., Clissa, P., Jablensky, A.V. (2006). A multivariate electrophysiological endophenotype, from a unitary cohort, shows greater research utility than any single feature in the Western Australian Family Study of Schizophrenia. Biological Psychiatry, 60(1), 1±10. Raux, G., Bonnet-Brilhault, F., Louchart, S., Houy, E., Gantier, R., Levillain, D., et al. (2002). The 2 bp deletion in exon 6 of the `alpha 7-like' nicotinic receptor subunit gene is a risk factor for the P50 sensory gating de®cit. Molecular Psychiatry, 7(9), 1006±1011. Riley, B.P., Makoff, A., Mogudi-Carter, M., Jenkins, T., Williamson, R., Collier, D., Murray, R. (2000). Haplotype transmission disequilibrium and evidence for linkage of the CHRNA7

3. AUDITORY EVOKED POTENTIAL ABNORMALITIES

69

gene region to schizophrenia in Southern African Bantu families. American Journal of Medical Genetics, 96(2), 196±201. Riley, B., Williamson, M., Collier, D., Wilkie, H., Makoff, A. (2002). A 3-Mb map of a large segmental duplication overlapping the alpha 7-nicotinic acetylcholine receptor gene (CHRNA7) at human 15q13-q14. Genomics, 79(2), 197±209. Roth, W.T., Cannon, E.H. (1972). Some features of auditory evoked-response in schizophrenics. Archives of General Psychiatry, 27(4), 466. Roxborough, H., Muir, W.J., Blackwood, D.H.R., Walker, M.T., Blackburn, I.M. (1993). Neuropsychiological and P300 abnormalities in schizophrenics and their relatives. Psychological Medicine, 23, 305±314. Sams, M., Kaukoranta, E., Hamalainen, M. (1991). Cortical activity elicited by changes in auditory stimuli: Different sources for the magnetic N100m and mismatch responses. Psychophysiology, 28, 21±29. Schreiber, H., Stolzborn, G., Kornhuber, H.H., Born, J. (1992). Event-related potential correlates of impaired selective attention in children at high-risk for schizophrenia. Biological Psychiatry, 32(8), 634±651. Schwarzkopf, S.B., Lamberti, J.S., Smith, D.A. (1993). Concurrent assessment of acoustic startle and auditory P50 evoked potential measures of sensory inhibition. Biological Psychiatry, 33(11±12), 815±828. Semlitsch, H.V., Anderer, P., Schuster, P., Presslich, O. (1986). A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP. Psychophysiology, 23(6), 695±703. Sham, P.C., Morton, N.E., Muir, W.J., Walker, M., Collins, A., Shields, D.C., et al. (1994). Segregation analysis of complex phenotypes: an application to schizophrenia and auditory P300 latency. Psychiatric Genetics, 4, 29±38. Shelley, A., Ward, P., Catts, S., Michie, P., Andrews, S., McConaghy, N. (1991). Mismatch negativity: An index of preattentive processing de®cit in schizophrenia. Biological Psychiatry, 30, 1059±1062. Shelley, A., Silipo, G., Javitt, D. (1999). Diminished responsiveness of ERPs in schizophrenic subjects to changes in auditory stimulation parameters: Implications for theories of cortical dysfunction. Schizophrenia Research, 37, 65±79. Shifman, S., Bronstein, M., Sternfeld, M., Pisante-Shalom, A., Lev-Lehman, E., Weizman, A., et al. (2002). A highly signi®cant association between a COMT haplotype and schizophrenia. American Journal of Human Genetics, 71(6), 1296±1302. St Clair, D., Blackwood, D., Muir, W. (1989). P300 abnormality in schizophrenic subtypes. Journal of Psychiatric Research, 23(1), 49±55. Tiitinen, H., May, P., Reinikainen, K., Naatanen, R. (1994). Attentive novelty detection in humans is governed by pre-attentive sensory memory. Nature, 372, 90±92. Tsai, S.J., Yu, Y.W.Y., Chen, T.J., Chen, J.Y., Liou, Y.J., Chen, M.C., Hong, C.J. (2003). Association study of a functional catechol-O-methyltransferase-gene polymorphism and cognitive function in healthy females. Neuroscience Letters, 338(2), 123±126. Umbricht, D., Krljes, S. (2005). Mismatch negativity in schizophrenia: A meta-analysis. Schizophrenia Research, 76(1), 1±23. Umbricht, D., Javitt, D., Novak, G., Bates, J., Pollack, S., Lieberman, J., Kane, J. (1998). Effects of clozapine on auditory event-related potentials in schizophrenia. Biological Psychiatry, 44(8), 716±725. Umbricht, D., Javitt, D., Novak, G., Bates, J., Pollack, S., Lieberman, J., Kane, J. (1999). Effects of risperidone on auditory event-related potentials in schizophrenia. International Journal of Neuropsychopharmcology, 2(4), 299±304. Umbricht, D., Schmid, L., Koller, R., Vollenweider, F., Hell, D., Javitt, D. (2000). Ketamine-

70

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

induced de®cits in auditory and visual context-dependent processing in healthy volunteers. Archives of General Psychiatry, 57, 1139±1147. Umbricht, D., Koller, R., Vollenweider, F.X., Schmid, L. (2002). Mismatch negativity predicts psychotic experiences induced by NMDA receptor antagonist in healthy volunteers. Biological Psychiatry, 51, 400±406. Umbricht, D., Koller, R., Schmid, L., Skrabo, A., Grubel, C,. Huber, T., Stassen, H. (2003). How speci®c are de®cits in mismatch negativity generation to schizophrenia? Biological Psychiatry, 53(12), 1120±1131. van Beijsterveldt, C.E.M., van Baal, G.C.M. (2002). Twin and family studies of the human electroencephalogram: A review and a meta-analysis. Biological Psychology, 61(1±2), 111±138. Waldo, M.C., Adler, L.E., Leonard, S., Olincy, A., Ross, R.G., Harris, J.G., Freedman, R. (2000). Familial transmission of risk factors in the ®rst-degree relatives of schizophrenic people. Biological Psychiatry, 47(3), 231±239. White, P.M., Yee, C.M. (1997). Effects of attentional and stressor manipulations on the P50 gating response. Psychophysiology, 34, 703±711. Winterer, G., Egan, M.F., Raedler, T., Sanchez, C., Jones, D.W., Coppola, R., Weinberger, D.R. (2003). P300 and genetic risk for schizophrenia. Archives of General Psychiatry, 60(11), 1158±1167. Yee, C.M., White, P.M. (2001). Experimental modi®cation of P50 suppression. Psychophysiology, 38(3), 531±539.

CHAPTER FOUR

Are eye-movement abnormalities related to susceptibility genes for schizophrenia? James MacCabe, Jolanta Zanelli

INTRODUCTION A century ago, Diefendorf and Dodge (1908) ®rst described the phenomenon that patients with schizophrenia have dif®culty accurately maintaining visual gaze on a predictably moving target. Modern research into eye-tracking abnormalities as a neurobiological abnormality representing a possible genetic marker of schizophrenia dates back to the 1970s, when not only patients but also their well relatives were reported to display abnormal eye tracking (Holzman et al., 1974). This chapter describes the major abnormalities of eye tracking reported in patients with schizophrenia and presents data from the Maudsley Family Study of Psychosis, which assessed the utility of eye-tracking dysfunction as an intermediate phenotype by virtue of its presence in unaffected relatives at presumed differing genetic liability.

Assessment of eye movements Eye movements in humans can be divided into two types: smooth pursuit and saccades. In smooth pursuit movements, the direction of gaze changes smoothly, to maintain a target centred on the fovea, when either the target or the head is in motion. The vestibulo-ocular re¯ex compensates for head movements, and will not be considered further. Smooth pursuit tracking requires the brain to process visual inputs on target position and velocity, continuously compare this with eye position, and generate motor outputs to allow a smooth, accurate movement of both eyes at a constant velocity. 71

72

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Saccades are rapid eye movements that serve to redirect the gaze rapidly from one point to another. Saccades can occur from rest or during smooth pursuit. Eye movements normally occur without conscious intervention, and can be guided by visual, vestibular or auditory inputs. Saccades can also be inhibited or generated voluntarily in normal individuals.

Smooth pursuit Smooth pursuit eye movements are assessed by asking participants to follow a moving target while the head is immobilized. The direction of gaze in real time is measured by means of an infrared scalar re¯ection device or other apparatus, and this is compared with the position of the target. The accuracy of smooth pursuit depends on correct matching of the gaze velocity and target velocity, the generation of appropriate saccades to correct any such inaccuracies, and the absence of intrusive saccades. The accuracy of smooth pursuit can be measured using either gain or root mean squared error (RMSE). Gain is simply a ratio of mean eye position or velocity to mean target position or velocity. A gain of 1, therefore, corresponds to perfect tracking. A value for gain greater than 1 indicates that gaze velocity is greater than target velocity, whereas a value of less than 1 shows that the participant's gaze is `lagging behind' the target. Thus, gain is a measure not only of the accuracy of pursuit, but also the direction of any error. As its name implies, RMSE is calculated by summing the squared position errors at multiple time points, and taking the square root of the result. The RMSE is thus the mean distance of the participant's gaze from the target position, but with no indication of the direction of any error. Both gain and RMSE are usually measured on a segment of uninterrupted smooth pursuit, using specialized software. However, it should be noted that the choice of which segment to measure is generally taken by the operator, so there is a possibility for observer bias if the observer is not blinded to the individual's status. Purely subjective ratings of smooth pursuit accuracy are also possible, in which the experimenter compares the trace with a series of sample traces of differing accuracy.

Saccades during smooth pursuit Saccades that occur during smooth pursuit comprise three types: (1) the optokinetic re¯ex re-aligns the gaze from one moving target to another, such as when observing moving scenery out of a train window; (2) compensatory saccades re-align the gaze with the target to correct inaccuracies in smooth pursuit, or a change in velocity of the target; and (3) intrusive saccades are interruptions to smooth pursuit, which move the gaze away from the target.

4. EYE MOVEMENT ABNORMALITIES

73

Compensatory saccades are of two main types: catch-up saccades (CUS) and back-up saccades (BUS). When the eye lags behind the target during a period of low gain pursuit, and position error accumulates, a CUS (in the same direction as the target) is required to bring the gaze back on to the target. Conversely, when the gaze moves ahead of the target, a BUS is required. There are several types of intrusive saccade. The most studied in schizophrenia are square wave jerks (SWJ; pairs of small-amplitude saccades, the ®rst saccade taking the gaze away from the target and the second returning it to the target) and anticipatory saccades (in which the gaze moves to a point ahead of the target and remains there until the target catches up) (Friedman et al., 1992a; Levy et al., 1994). The detection of saccades from a trace is usually performed using a computer. As this involves some subjective input from the observer, the possibility of bias is introduced if observers are not blinded.

Saccades generated from rest Re¯exive saccades are generated when a novel visual stimulus appears in the periphery of the visual ®eld. It requires an assessment of the difference in position between the gaze and the new stimulus, and the generation of a rapid and accurate saccade to bring the target onto the fovea. Re¯exive saccades are relatively simple to detect using software or by visual inspection of the eye movement trace. The latency and accuracy of the initial saccade can be measured. Re¯exive saccades can usually be inhibited voluntarily, and saccades can also be initiated voluntarily. The ability to inhibit re¯exive saccades and to generate voluntary saccades is commonly assessed using the antisaccade task (Hallett, 1978). The task begins with the participant ®xating on a target directly ahead. The target then moves rapidly to a peripheral location, and the participant is required to inhibit his or her re¯exive saccade (in pursuit of the target) and to generate, instead, a saccade in the opposite direction, which is termed an antisaccade. A saccadic distractibility error is recorded if a participant fails to inhibit the re¯exive saccade in the direction of the target. Saccadic distractibility is usually expressed as the percentage of trials in which the initial eye movement was towards the stimulus. Variations of the task include a gap or overlap (McDowell & Clementz, 1997) between the offset of the central stimulus and the onset of the peripheral stimulus. Other variables that can be measured during the antisaccade task include the latency and the spatial accuracy of the antisaccades, although these have received less attention as putative biological markers in schizophrenia (Everling & Fischer, 1998).

74

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Smooth pursuit eye movements are generally assessed using an electronically generated stimulus, moving in the horizontal plane. Head movement must be minimized, so that the vestibulo-ocular re¯ex is not activated. Various techniques are available for measuring eye position, of which probably the most accurate is the infrared oculograph, a device that detects the centre of the pupil by re¯ecting an infrared light source off the eye. The resulting eye movement traces are recorded and compared with the stimulus using specialised software, and/or with the naked eye.

Previous findings on eye movements in schizophrenia Smooth pursuit abnormalities The most consistent ®nding in schizophrenia is that smooth pursuit velocity in schizophrenia patients is low (gain is less than 1), with the result that the gaze progressively lags behind a moving target, necessitating frequent catch-up saccades to catch up with the target. This abnormality can be measured in two ways: by calculating the gain (eye velocity divided by target velocity) or by counting the frequency and amplitude of catch-up saccades. Several studies have found this abnormality in the unaffected relatives of schizophrenia patients. In patients, they do not appear to be related to antipsychotic medication, mental state or other potential confounders. These observations suggest that smooth pursuit abnormalities are a marker for genetic liability to schizophrenia. Other smooth pursuit abnormalities that have been observed in schizophrenia patients include an excess of intrusive eye movements, such as anticipatory saccades and square wave jerks, and a reduced accuracy of pursuit as assessed by qualititatively rating traces of eye movements.

Saccades Compared with the smooth pursuit system, abnormalities in the saccadic eye-tracking system have been studied only recently in schizophrenia. Most studies have not shown any disturbance in the ability to generate visually guided re¯exive saccades in schizophrenia, but many groups have now demonstrated a reduced ability to inhibit visually guided saccades and/or to generate saccades voluntarily. This chapter describes analyses of smooth pursuit and antisaccade eye movements assessed on participants from Phases 2 and 3 of the Maudsley Family Study of Psychosis. The results on the antisaccade task presented here, but not those for smooth pursuit abnormalities, have been published previously (Maccabe et al., 2005). Results on antisaccade abnormalities from Phase 1 of the study have also been published previously (Crawford et al.,

4. EYE MOVEMENT ABNORMALITIES

75

1998). In the present series of analyses, we sought to conduct a methodologically robust assessment of the validity of eye movement abnormalities as putative endophenotypic markers for schizophrenia. To maximize statistical power, we initially assessed eye movement abnormalities in the entire sample of patients with schizophrenia and the entire sample of relatives of patients with schizophrenia. We next took account of family history by assessing whether any abnormalities were more prominent in patients from multiply affected families compared with patients from singly-affected families and in unaffected relatives from multiply-affected families compared with unaffected relatives from singly-affected families. We considered that endophenotypic abnormalities would be more prominent in those participants from multiply-affected families.

METHODOLOGY Participant recruitment, inclusion and exclusion criteria and clinical assessments are described in Chapter 2. The study groups are represented diagrammatically in Figure 4.1. A total of ®fty-three patients ful®lling DSMIV (American Psychiatric Association (APA), 1994) criteria for schizophrenia (n=47), schizoaffective disorder (n=5) or schizophreniform disorder (n=1) were included in the study, along with eighty unaffected relatives and forty-one controls. The eye-movement tasks were conducted using the Amtech ET3 eyetracking system (AmTech GmbH, Weinheim, Germany). A Dell Optiplex 560/L computer was used to control the apparatus, and to record eye movments, using AmTech ET3 software. Eye movements were detected by means of an infrared re¯ection oculograph, with eye position sampled at 200 Hz. The participants were seated in a darkened room and were asked not to move their head, which rested on an adjustable frame. Each participant completed a set of twelve practice trials to ensure that they understood the test instructions. All participants appeared to understand the smooth pursuit task. Participants with an error rate of over 50% on the antisaccade task were given a further explanation of the task, followed by a second practice battery, with a maximum of two practice batteries.

Smooth pursuit task The stimulus was provided by a red neon±helium laser spot of 0.5 degree, projected on a black screen 180 cm from the participant's eyes. The target moved horizontally Ô15 degrees of visual angle, and had a trapezoidal waveform pro®le. There were ®ve complete cycles at 15 degrees per second and ®ve cycles at 30 degrees per second. The rater was blind to participant group. Saccadic eye movements were identi®ed based on the criteria of Abel and Friedman (Abel et al., 1991;

76

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Total sample

Schizophrenic

Non-schizophrenic

Relatives

Multiply-affected family

Controls

Singly-affected family Multiply-affected family

Singly-affected family

Figure 4.1 Composition of subject groups and contrasts

Friedman et al., 1992a). The de®nition of a saccade was based on two requirements: (1) the difference between the velocity of the saccade and the velocity immediately before the saccade had to be >30 degrees/second; and (2) the amplitude of the fast eye movement had to be 0.5±5.0 degrees (Arolt et al., 1996). After calibration of the original data, saccades in both directions were detected and counted. The ®rst cycle and the 125 milliseconds before and after the turning points were discounted. Blinks and other artefacts were detected manually and removed. The saccades were classi®ed as follows: (1) catch-up saccade (CUS), de®ned as a forward saccade that reduced the disparity between the gaze and target positions (catch-up saccades with the amplitude greater than 5 degrees of visual angle were subclassi®ed as `large CUS'); (2) back-up saccade (BUS), similarly de®ned but in the opposite direction; (3) square wave jerk (SWJ), a pair of saccades in opposite directions separated by 200±400 ms of smooth pursuit; (4) anticipatory, a forward saccade that takes the eye away from the target; and (5) unclassi®ed, any saccade not ®tting the above classi®cation. The mean amplitude of catch-up saccades was also calculated. For each half cycle of target movement, the gain during the longest and uninterrupted period of smooth pursuit was calculated. This resulted in ten gain values for each participant (®ve rightward and ®ve leftward) for each target speed, from which the mean was taken. The recordings were also rated qualitatively, again blind to participant group, using the scale of Shagass (Shagass et al., 1974).

4. EYE MOVEMENT ABNORMALITIES

77

Antisaccade task For the antisaccade task, the stimulus consisted of a central light-emitting diode (LED), with two peripheral LEDs at 15 degrees eccentricity from the central LED. All LEDs were mounted on a board, 180 cm away from the participant's head. At the start of each trial, a central LED was illuminated. After 800 milliseconds, the central LED was extinguished and, simultaneously, one of the peripheral targets (Ô15 degrees eccentricity) was illuminated for 3 seconds (accompanied by an audible signal) and then returned to the centre. The participants were instructed to ®xate on the central target until it moved to the peripheral position, and then to direct their gaze as quickly and accurately as possible to its mirror position (a position of equal distance from, but in the opposite direction to, the peripheral target). Eye movement recordings were later analysed using AmTech Eyemap v2.0, by a trained rater who was blind to diagnosis or participant group. Blinks and other artefacts were identi®ed by inspection of the trace, and removed from the analysis. A saccade was de®ned as a de¯ection of 2.5 degrees. A correct antisaccade was recorded if a target to the left of the central LED initiated a saccade to the right, and vice versa. Any initial saccade (discounting the ®rst 80 milliseconds) toward the peripheral target was scored as a distractibility error. Where this was followed by a saccade in the correct direction, the latter was termed a corrective saccade. The number of corrected distractibility errors divided by the number of analysable trials gave the distractibility error score. As well as the distractibility error score, the mean latency of each type of saccade (antisaccade, corrective saccade, and distractibility error) was calculated for each participant. Participants completed two sets of twelve experimental trials, during which the position of the stimulus was varied pseudo-randomly throughout the set to prevent predictive saccades, in the order L R R R L L R L L R R L.

Blinding Raters were not involved in the design or recruitment for the study, or the eye movement testing. They did not come into contact with any of the patients and had no information about them other than their day and month of birth and initials. Patients, relatives and controls were analysed in a random order to prevent bias due to rater drift.

Statistical analysis Statistical analyses were conducted using SPSS v10.1 (SPSS Inc, Illinois, USA) and STATA v7.0 (Stata Corp, Texas, USA). Age and years of

78

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

education were compared between groups using analysis of variance, and gender using the 2 statistic. As members of a family are likely to share both genes and environment, observations from within the same family should not be assumed to be independent, particularly for a putative genetic marker. We therefore used robust standard errors to take account of the dependence of observations within families using the `robust cluster' option in STATA. The rationale for its use in studies of this type has been discussed more fully by RabeHesketh and others (Rabe-Hesketh et al., 2001). Age and gender were entered as co-variates. To allow the calculation of adjusted means, we used a linear regression equation without a constant, and centred the age variable around the mean age of the sample. Gender was coded such that female was the reference category. The regression coef®cients for the dummy group variables therefore represented the group means, adjusted to the mean age of the sample and to female gender. The following dependent variables were entered into the regression analysis: the frequency of each saccade category (total number of saccades, CUS, BUS), gain, amplitude of total number of saccades, CUS amplitude and qualitative rating scores using the Shagass scale. Family ID number was used as the clustering variable. We made comparisons between combinations of groups in an orthogonal design, according to Figure 4.1. The comparisons were as follows: test 1, patients with schizophrenia versus all other groups; test 2, all unaffected relatives versus controls; test 3, patients with schizophrenia from multiplyaffected versus singly-affected families; test 4, unaffected relatives from multiply-affected versus singly-affected families.

RESULTS Demographic details The demographic details of the groups are shown in Table 4.1. The schizophrenia patients had a mean length of illness of 12.5 years (SD 10.6). The groups were well matched for years of education, but there were signi®cant group differences for age (F=9.9 (5df ), p<0.001) and gender (2=18.77 (5 df ), p<0.01), and these variables were therefore included in the regression equation.

Smooth pursuit Tables 4.2 and 4.3 show the crude and adjusted mean values for smooth pursuit variables at 30 and 15 degrees per second. The results of the contrasts are presented in Table 4.4. Intrusive saccades, SWJ and antisaccades

22

Sz, MA Fam

1 3

2

Schizophreniform Schizoaffective disorder disorder

6 4

7 1

3

Major Other depression DSM-IV diagnosis

42 37

21

No disorder

36.6

9.2

14.2 14.7

8.6 12.9

11.7

<0.001

51.4 41.9

32.6 47.0

14.1

13.8 13.7

13.6 14.2

NS

0.2

2.8 3.5

2.2 3.2

3.9

SD

Mean

Mean

SD

Education (years)

Age (years)

MA Fam, multiply-affected family; NS, not signi®cant; Rel, relative; SA Fam, singly-affected family; Sz, schizophrenia.

Rel, SA Fam Control F Chi square p

Sz, SA Fam 25 Rel, MA Fam

Schizophrenia

Group

TABLE 4.1 Lifetime DSM-IV diagnoses and demographic details in the ®ve groups

14.6 <0.05

30.8 41.5

67.9 41.9

68.0

Gender (% males)

95.20 28.75 25.54 1.92 2.33 2.10 3.13

(7.55) (7.76) (8.32) (1.69) (0.93) (0.51) (0.90)

(SD)

94.62 29.01 26.30 2.02 2.40 2.15 3.21

M*

Sz, MA Fam

(9.64) (7.83) (8.90) (1.43) (0.62) (0.42) (0.95)

(SD) 87.36 30.02 27.23 1.59 2.61 2.32 3.37

M*

Sz, SA Fam

88.23 29.45 26.31 1.41 2.50 2.25 3.24

M 96.51 25.73 22.21 2.51 2.15 2.00 2.82

(6.33) (6.52) (6.45) (2.61) (0.70) (0.44) (0.85)

(SD) 98.21 26.12 24.52 1.85 1.97 1.95 2.60

M*

Rel, MA Fam M

M*

93.04 (11.29) 95.31 28.05 (10.20) 28.68 21.77 (9.16) 24.30 3.28 (3.81) 2.43 2.36 (0.86) 2.12 2.04 (0.42) 1.96 3.13 (0.98) 2.84

(SD)

Rel, SA Fam M

(SD)

Controls M*

93.55 (12.08) 94.42 27.93 (7.74) 28.03 23.07 (7.50) 25.07 2.40 (2.46) 2.00 2.32 (0.82) 2.24 2.06 (0.32) 2.05 3.00 (0.96) 2.90

M

M*

95.55 (11.57) 94.86 32.46 (12.06) 34.21 27.50 (11.34) 28.18 3.33 (4.16) 4.06 1.51 (0.85) 1.54 1.20 (0.35) 1.23 2.92 (1.10) 3.00

(SD)

(SD)

M*

93.43 (8.78) 92.46 38.10 (16.83) 40.59 31.53 (15.79) 32.40 3.93 (4.86) 5.04 1.54 (0.72) 1.60 1.27 (0.35) 1.32 3.10 (0.92) 3.24

M

Sz, SA Fam (SD)

M*

98.41 (7.75) 98.12 29.51 (12.40) 29.65 24.36 (9.66) 26.19 3.85 (4.40) 2.67 1.21 (0.46) 1.10 1.10 (0.23) 1.04 2.85 (0.97) 2.60

M

Rel, MA Fam

(SD)

M* 96.18 (6.91) 96.08 34.58 (14.29) 34.12 26.15 (10.22) 28.10 5.21 (6.22) 3.55 1.41 (0.64) 1.26 1.18 (0.31) 1.10 3.15 (0.98) 2.83

M

Rel, SA Fam

(SD)

M* 97.42 (17.87) 96.86 33.00 (12.36) 34.01 26.69 (10.08) 28.31 3.71 (4.32) 3.29 1.40 (0.61) 1.34 1.16 (0.27) 1.14 2.98 (0.95) 2.85

M

Controls

M*, means are age and sex corrected; BUS, back-up saccades; CUS, catch-up saccades; MA Fam, multiply-affected family; Rel, relative; SA Fam, singly-affected family; Sz, schizophrenia.

Gain Total saccades CUS BUS Total amplitude CUS amplitude Rating scale

M

Sz, MA Fam

TABLE 4.3 Overall mean values with standard deviation at 15 degrees/second

M*, means are age and sex corrected; BUS, back-up saccades; CUS, catch-up saccades; MA Fam, multiply-affected family; Rel, relative; SA Fam, singly-affected family; Sz, schizophrenia.

Gain Total saccades CUS BUS Total amplitude CUS amplitude Rating scale

M

TABLE 4.2 Overall mean values with standard deviation at 30 degrees/second

2.91 1.95 1.56 0.87 2.66 3.5 2.87

<0.005 0.054 NS NS <0.01 <0.001 <0.005

30 deg/s F p

1.71 2.0 1.29 1.55 2.61 3.17 2.09

NS <0.05 NS NS <0.01 <0.005 <0.05

15 deg/s F p NS NS NS NS NS NS NS

15 deg/s F p

3.1 <0.005* 0.82 0.51 NS 1.59 0.39 NS 1.14 0.95 NS 0.77 0.88 NS 0.24 1.33 NS 0.89 0.62 NS 0.84

30 deg/s F p

Sz, MA Fam vs. Sz, SA Fam

1.15 1.05 0.46 0.3 1.29 1.34 1.01

NS NS NS NS NS NS NS

30 deg/s F p 0.1 0.9 0.57 0.22 1.49 1.19 0.77

NS NS NS NS NS NS NS

15 deg/s F p

Relatives vs. controls

1.74 1.21 0.14 0.88 1.08 0.1 1.3

NS NS NS NS NS NS NS

30 deg/s F p

1.18 1.81 0.89 0.87 1.42 0.98 1.2

NS NS NS NS NS NS NS

15 deg/s F p

Familial relative vs. non-familial

*The difference between the groups is in the opposite direction to that predicted. BUS, back-up saccades; CUS, catch-up saccades; deg/s, degrees per second; MA, multiply-affected family; NS, not signi®cant; SA Fam, singly-affected family; Sz, schizophrenia.

Gain Total saccades CUS BUS Total amplitude CUS amplitude Shagass scale

Contrast

Sz vs. all others

TABLE 4.4 Comparisons of group means for the smooth pursuit variables

82

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

occurred very rarely during pursuit, so they were not included in contrast analyses. Participants with schizophrenia had higher overall saccade amplitude, higher catch-up saccade amplitudes and lower scores on the Shagass scale than those without schizophrenia, at both target velocities. The total number of saccades also differed at 30 degrees per second, although the signi®cance was borderline at the 0.05 level. In addition, at 30 degrees per second only, patients with schizophrenia had lower gain. All of these differences were as expected. However, there were no signi®cant differences between relatives and controls, nor between multiply- and singly-affected families, except that patients from singly-affected families had a lower gain than patients from multiply-affected families, contrary to what we hypothesized for an endophenotypic measure.

Antisaccade task The results for antisaccade distractibility are shown in Tables 4.5 and 4.6. Patients with schizophrenia performed much worse than all other groups. However, there were no differences between relatives and controls, nor between multiply- and singly-affected families. There were no signi®cant differences between any groups for latency of either correct or incorrect saccades, nor in the proportion of antisaccade errors that were corrected.

DISCUSSION Our study replicated all previous studies by demonstrating that schizophrenia patients performed signi®cantly worse on eye-tracking tasks than relatives and healthy volunteers. However, there was no evidence that such performance indicates a genetic vulnerability to schizophrenia: unaffected relatives performed no worse than healthy volunteers, and there was no signi®cant difference between members of singly- or multiply-affected families, other than a small difference in gain between patients from singlyand multiply-affected families, which was in the opposite direction to that hypothesized. Whereas these results do not preclude a genetic in¯uence on eye movement tasks, they do not support a genetic overlap with schizophrenia. Our results raise two important questions: 1. 2.

Why are our results not in agreement with some previous studies, which showed differences between relatives and healthy volunteers? If we are correct in concluding that eye-movement abnormalities are not genetic markers for schizophrenia, then why are they present in schizophrenia?

4. EYE MOVEMENT ABNORMALITIES

83

TABLE 4.5 Means for saccadic scores Group

Saccadic distractibility score (SDS)

SDS adj

% of ADEs corrected

ADE latency (msec)

Antisaccade latency (msec)

Sz, MA Fam Sz, SA Fam Rel, MA Fam Rel, SA Fam Control

41.7 48.5 22.0 29.4 26.6

44.3 52.5 18.8 26.8 27.3

87.2 93.1 90.3 96.8 95.2

305 298 278 317 302

417 395 407 445 400

(29.0) (28.2) (15.1) (19.4) (19.4)

(24.6) (17.0) (30.1) (14.6) (16.7)

(54) (35) (86) (52) (84)

(171) (133) (56) (80) (85)

The adjusted means for SDS are based on the mean age of the entire sample, and female gender. ADE, antisaccade distractibility error; adj, adjusted; MA Fam, multiply-affected family; Rel, relative; SA Fam, singly-affected family; Sz, schizophrenia.

TABLE 4.6 Comparisons of group means for saccadic distractibility Test group(s)

ADE-score

Comparison group(s)

ADE-score

F (1,98)

p

Sz All Rel Sz, MA Fam Rel, MA Fam

45.3 26.6 41.7 22.0

All others Controls Sz, SA Fam Rel, SA Fam

26.6 26.6 48.5 29.4

32.3 0.51 1.03 1.48

<0.00001 NS NS NS

ADE, antisaccade distractibility error; MA Fam, multiply-affected family; NS, not signi®cant; Rel, relative(s); SA Fam, singly-affected family; Sz, schizophrenia.

1. Why are our results not in agreement with previous studies? Before answering this question, it should be noted that not all studies have demonstrated differences in eye-movement task performance between relatives and controls (Levy et al., 2004).

Task parameters Smooth pursuit task The absence of a difference between the groups on gain at 15 degrees/ second in the present study is consistent with previous reports (Oepen et al., 1990; Clementz et al., 1992). It may be that differences between groups would be greater at higher velocities than 30 degrees.

84

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Antisaccade task McDowell and others have investigated the effects of manipulations of stimulus parameters on antisaccade performance. In one study (McDowell et al., 1999), the eccentricity of the stimulus was set at 8 or 16 degrees. Both patients and relatives could be distinguished from controls better in the 16degree condition. At 15 degrees eccentricity, our study should therefore be well placed to detect group differences. The same research group (McDowell & Clementz, 1997) has also manipulated the overlap between the illumination of the central cue and the peripheral target, and assessed the effects of such manipulations on task performance. They found that a better separation between groups was achieved in the `overlap' condition than in the standard version of the task. It is thus possible that, had we used the overlap condition, a difference would then have emerged between relatives and controls. Our study used twenty-four trials per participant, whereas some have used more (Karoumi et al., 2001; McDowell & Clementz, 1997; Thaker et al., 1996, 2000). However, the meta-analysis by Levy and colleagues (Levy et al., 2004) demonstrated that the number of trials did not in¯uence the effect size.

Statistical analysis Our analysis differed from most previous studies in two main respects. First, we used robust standard errors, taking into account non-independence of observations from within the same family. Second, we controlled for age and gender. Several studies have demonstrated a worsening with age on the smooth pursuit (Ross et al., 1999) and antisaccade task (Butler et al., 1999; Olincy et al., 1997; Sha®q-Antonacci et al., 1999; Sweeney et al., 2001), indicating that age should be controlled for in analyses of saccadic distractibility. Gender may also have an effect on saccadic distractibility: Crawford et al. found that females performed worse than males in a similar sample of schizophrenia participants, their healthy ®rst-degree relatives, and normal controls from Phase 1 of the study (Crawford et al., 1998).

Selection bias Eye-movement tasks have a poor speci®city for schizophrenia. Smooth pursuit abnormalities (Blackwood et al., 1996; Flechtner et al., 1997; Sweeney et al., 1999) and increased saccadic distractibility rates (Katsanis et al., 1997; Sereno & Holzman, 1995; Sweeney et al., 1998; Tien et al., 1996) have been found in both unipolar and bipolar affective disorders. It follows that, if they are to be comparable, relative and control groups must be subject to the same exclusion criteria with respect to psychiatric morbidity and differ only with regard to family history of schizophrenia.

4. EYE MOVEMENT ABNORMALITIES

85

However, in some previous studies, controls were included only if they had a negative personal and family history of any psychiatric disorder, whereas this restriction was not applied to relatives. As the lifetime population prevalence of psychiatric disorders is estimated at over 1 in 4 (Goldberg 1991) such `normal' controls are not typical of the general population, and should not be compared with relatives who have signi®cant levels of morbidity. This issue has recently been explored by Levy and others in a meta-analysis (Levy et al., 2004), which demonstrated that whereas studies with asymmetrical exclusion criteria tended to demonstrate large signi®cant differences between relatives and controls, those with symmetrical criteria showed smaller, usually non-signi®cant, effects. In the present study, psychosis and schizophrenia-spectrum disorders were the only exclusion criteria in healthy volunteers, ensuring that our controls were representative of the population. The only selection criterion that distinguished our relative and control group was the presence of psychosis in a ®rst-degree relative. (In actuality, a higher proportion of relatives than controls had experienced a non-psychotic disorder such as depression at some point in their lives as described in Chapter 2, since depression is over-represented among relatives of patients with schizophrenia.)

Observer bias Although the detection of saccades during smooth pursuit is usually performed on a computer, there is considerable scope for bias to be introduced. Most saccade-detection systems are designed to be very sensitive so that no saccades are missed. When a possible saccade is identi®ed by the computer, the operator then makes a judgement as to whether a saccade is present and may classify it into various types. Unless clearly de®ned parameters are used to de®ne a saccade, the presence or absence of a saccade is a subjective judgement. Similarly, the measurement of gain can be in¯uenced by the rater choosing which sections of the trace are representative and should therefore be included. Clearly, subjective scales such as the Shagass scale are also subject to observer bias. To minimize observer bias in this study, we used trained raters who were not involved in the recruitment of participants, employing objective de®nitions of the various types of saccade, and blind to diagnosis or participant group.

Publication bias The tendency of authors, journals and reviewers to favour positive results has been well documented. Furthermore, with some notable exceptions (Brownstein et al., 2003), the published studies on eye movements and schizophrenia that have failed to show statistically signi®cant differences

86

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

between relatives and healthy volunteers have not always made this clear. We are also aware of other negative ®ndings that have not been published (P. Clissa, A. Jablensky, personal communication).

Type II error Our sample size was adequate to demonstrate signi®cant differences between patients with schizophrenia and non-schizophrenic participants. We acknowledge that the contrast between patients with schizophrenia and non-schizophrenic participants had somewhat greater power than the contrast between unaffected relatives and controls. Nevertheless, our data indicate that if the ®rst-degree relatives of schizophrenia patients do have a de®cit in eye-movement tasks, it is likely to be very small compared to the de®cit in patients.

2. If eye-movment tasks are not a genetic markers for schizophrenia, then why are they present in schizophrenia? Marker of environmental risk factors A number of early and late environmental risk factors for schizophrenia have been identi®ed, including pregnancy and birth complications, winter birth and drug misuse (Murray & Fearon, 1999). Eye-movement disorders could be a marker for one or more of these risk factors. We intend to conduct further analyses to explore this possibility.

Marker of frontal cortical pathology Using positron emission tomography, O'Driscoll and colleagues have demonstrated that poor smooth pursuit eye tracking is associated with impaired frontal lobe functioning in schizophrenia (O'Driscoll et al., 1999). Everling and Fischer (1998) recently reviewed the functional imaging, brain lesion and electrophysiological studies of the antisaccade task, and concluded that the most promising candidate region is the prefrontal cortex. Functional MRI studies suggest that prefrontal activity is increased in controls, but not in patients with schizophrenia during antisaccade tasks (McDowell et al., 2002). Furthermore, several studies have found correlations between antisaccade distractibility error score and the Wisconsin Card Sorting Test, an established test of frontal lobe dysfunction (Milner & Petrides, 1996), in both controls and patients with schizophrenia (Crawford et al., 1995a, 1995b, 1996; Karoumi et al., 1998; Radant et al., 1997; Rosse et al., 1993; Tien et al., 1996).

4. EYE MOVEMENT ABNORMALITIES

87

Direct effect of antipsychotic drugs: Although eye-movement abnormalities have been observed in neurolepticnaõÈve schizophrenia patients (Hutton et al., 1998), other studies have found poorer performance in drug-treated patients in smooth pursuit (Friedman et al., 1992b) and antisaccade tasks (Crawford et al., 1995). Karoumi and others (Karoumi et al., 2001) noted a positive correlation between antipsychotic dose (measured in chlorpromazine dose equivalents) and distractibility error rate (Spearman's rho = 0.48, p<0.05). Although we have not yet had an opportunity to investigate the effects of medication on eyemovement performance in smooth pursuit abnormalities, there was no correlation between saccadic distractibility and current antipsychotic dose, measured in chlorpromazine equivalents (r=±.153, p=NS).

Chronic effect of medication usage Antipsychotic medications may have long-term effects on brain function. Thaker and others demonstrated a two-fold excess of saccadic distractibility errors in schizophrenia patients with tardive dyskinesia (TD), compared with those without TD (Thaker et al., 1989). In our own sample, saccadic distractibility was signi®cantly correlated with length of illness (r=.37, p<0.01) but not with age (r=.18, p=NS) in patients with schizophrenia, a ®nding that is compatible with an effect of chronic medication or disease progression.

Marker of neurodegeneration or neuroplasticity Although there is no conclusive evidence at present of neurodegeneration in schizophrenia (Allin & Murray, 2002), there have recently been some suggestions of a progression of brain abnormalities in schizophrenia; such changes are not incompatible with the neurodevelopmental hypothesis, and may be a result of neuroplasticity (Weinberger & McClure, 2002). Our ®nding of a correlation between saccadic distractibility and length of illness is consistent with such an interpretation.

Poor understanding of the task All patients appeared to have no dif®culty understanding the smooth pursuit task. However, the schizophrenia patients may have had a worse comprehension of the antisaccade task than other groups. However, distractibility errors were followed by a corrective saccade in around 90% of instances, with no signi®cant difference between patients with schizophrenia and other groups. Furthermore, we took steps to minimize any such effect by re-explaining the task if the proportion of errors exceeded 50% in the practice trial.

88

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Part of a generalised neuropsychological deficit In a separate study using an overlapping sample, we have found substantial correlations between ADE score and both premorbid and current IQ, verbal memory and associative learning, most of which were not present in relatives or healthy volunteers. However, smooth pursuit measures appeared unrelated to neuropsychological function (Zanelli et al., unpublished data).

CONCLUSIONS This study replicates previous ®ndings that schizophrenia patients perform poorly on smooth pursuit and antisaccade tasks, compared to their unaffected relatives and controls. However, our data do not provide support for the hypotheses that the unaffected relatives of patients show impaired performance, nor that performance in either task is associated with genetic loading for schizophrenia. Some previous studies have found abnormalities in unaffected relatives. However, we have argued that many of these previous studies may have been biased. Many used non-identical inclusion and exclusion criteria for relatives and control samples, most used non-blinded raters to assess eye tracking function, and very few took intra-familial dependence into account in the analysis. These results con®rm that eye-tracking abnormalities exist in schizophrenia patients, but suggest that smooth pursuit and antisaccade tasks are unlikely to prove useful as genetic markers of schizophrenia.

REFERENCES Abel, L.A., Friedman, L., Jesberger, J., Malki, A., Meltzer, H.Y. (1991). Quantitative assessment of smooth pursuit gain and catch-up saccades in schizophrenia and affective disorders. Biological Psychiatry, 29, 1063±1072. Allin, M., Murray, R.M. (2002). Schizophrenia: A neurodevelopmental or neurodegenerative disorder? Current Opinion in Psychiatry, 15, 9±15. American Psychiatric Association (APA) (1994). Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC: APA. Arolt, V., Lencer, R., Nolte, A., Muller-Myhsok, B., Purmann, S., Schumann, M., et al. (1996). Eye tracking dysfunction is a putative phenotypic susceptibility marker of schizophrenia and maps to a locus on chromosome 6p in families with multiple occurrence of the disease. American Journal of Medical Genetics, 67, 564±579. Blackwood, D.H., Sharp, C.W., Walker, M.T., Doody, G.A., Glabus, M.F., Muir, W.J. (1996). Implications of comorbidity for genetic studies of bipolar disorder: P300 and eye tracking as biological markers for illness. British Journal of Psychiatry, Supplement, 85±92. Brownstein, J., Krastoshevsky, O., McCollum, C., Kundamal, S., Matthysse, S., Holzmann, P.S., et al. (2003). Antisaccade performance is abnormal in schizophrenia patients but not in their biological relatives. Schizophrenia Research, 63, 13±25. Butler, K.M., Zacks, R.T., Henderson, J.M. (1999). Suppression of re¯exive saccades in younger and older adults: Age comparisons on an antisaccade task. Memory and Cognition, 27, 584±591. Clementz, B.A., Grove, M.W., Iocono, W.G. (1992). Smooth-pursuit eye movement

4. EYE MOVEMENT ABNORMALITIES

89

dysfunction and liability for schizophrenia: Implications for genetic modelling. Journal of Abnormal Psychology, 101, 117±129. Crawford, T.J., Haeger, B., Kennard, C., Reveley, M.A., Henderson, L. (1995). Saccadic abnormalities in psychotic patients. I. Neuroleptic-free psychotic patients. Psychological Medicine, 25, 461±471. Crawford, T.J., Puri, B.K., Nijran, K.S., Jones, B., Kennard, C., Lewis, S.W. (1996). Abnormal saccadic distractibility in patients with schizophrenia: A 99mTc-HMPAO SPET study. Psychological Medicine, 26, 265±277. Crawford, T.J., Sharma, T., Puri, B.K., Murray, R.M., Berridge, D.M., Lewis, S.W. (1998). Saccadic eye movements in families multiply affected with schizophrenia: The Maudsley Family Study. American Journal of Psychiatry, 155, 1703±1710. Diefendorf, A.R., Dodge, R. (1908). An experimental study of the ocular reaction on the insane from photographic records. Brain, 31, 451±489. Everling, S., Fischer, B. (1998). The antisaccade: A review of basic research and clinical studies. Neuropsychologia, 36, 885±899. Flechtner, K.M., Steinacher, B., Sauer, R., Mackert, A. (1997). Smooth pursuit eye movements in schizophrenia and affective disorder. Psychological Medicine, 27, 1411±1419. Friedman, L., Abel, L.A., Jesberger, J.A., Malki, A., Meltzer, H.Y. (1992). Saccadic intrusions into smooth pursuit in patients with schizophrenia or affective disorder and normal controls. Biological Psychiatry, 31, 1110±1118. Friedman, L., Jesberger, J.A., Meltzer, H.Y. (1992b). Effect of typical antipsychotic medications and clozapine on smooth pursuit performance in patients with schizophrenia. Psychiatry Research, 41, 25±36. Goldberg, D. (1991). Filters to care. In Jenkins, R., Grif®ths, S. (eds), Indicators for Mental Health in the Population, a series of two workshops organised by the Faculty of Public Health Medicine, the Royal College of Psychiatrists and the Department of Health. London: HMSO. Hallett, P.E. (1978). Primary and secondary saccades to goals de®ned by instructions. Vision Research, 18, 1279±1296. Holzman, P.S., Proctor, L.R., Levy, D.L., Yasillo, N.J., Meltzer, H.Y., Hurt, S.W. (1974). Eye-tracking dysfunctions in schizophrenic patients and their relatives. Archives of General Psychiatry, 31, 143±151. Hutton, S.B., Crawford, T.J., Puri, B.K., Duncan, L.J., Chapman, M., Kennard, C., et al. (1998). Smooth pursuit and saccadic abnormalities in ®rst-episode schizophrenia. Psychological Medicine, 28, 685±692. Karoumi, B., Ventre-Dominey, J., Vighetto, A., Dalery, J., d'Amato, T. (1998). Saccadic eye movements in schizophrenic patients. Psychiatry Research, 77, 9±19. Karoumi, B., Saoud, M., d'Amato, T., Rosenfeld, F., Denise, P., Gutknecht, C., et al. (2001). Poor performance in smooth pursuit and antisaccadic eye-movement tasks in healthy siblings of patients with schizophrenia. Psychiatry Research, 101, 209±219. Katsanis, J., Kortenkamp, S., Iacono, W.G., Grove, W.M. (1997). Antisaccade performance in patients with schizophrenia and affective disorder. Journal of Abnormal Psychology, 106, 468±472. Levy, D.L., Holzman, P.S., Matthysse, S., Mendell, N.R. (1994). Eye tracking and schizophrenia: A selective review. Schizophrenia Bulletin, 20, 47±62. Levy, D.L., O'Driscoll, G.A., Matthysse, S., Cook, S., Holzman, P.S., Mendell, N.R. (2004). Antisaccade performance in biological relatives of schizophrenia patients: A meta-analysis. Schizophrenia Research, 71, 113±125. Maccabe, J.H., Simon, H., Zanelli, J.W., Walwyn, R., McDonald, C.D., Murray, R.M. (2005). Saccadic distractibility is elevated in schizophrenia patients, but not in their unaffected relatives. Psychological Medicine, 35, 1727±1736. McDowell, J.E., Clementz, B.A. (1997). The effect of ®xation condition manipulations on

90

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

antisaccade performance in schizophrenia: Studies of diagnostic speci®city. Experimental Brain Research, 115, 333±344. McDowell, J.E., Myles-Worsley, M., Coon, H., Byerley, W., Clementz, B.A. (1999). Measuring liability for schizophrenia using optimized antisaccade stimulus parameters. Psychophysiology, 36, 138±141. McDowell, J.E., Brown, G.G., Paulus, M., Martinez, A., Stewart, S.E., Dubowitz, D.J., et al. (2002). Neural correlates of re®xation saccades and antisaccades in normal and schizophrenia subjects. Biological Psychiatry, 51, 216±223. Milner, B., Petrides, M. (1996). Behavioural effects of frontal lobe lesions in man. Trends in Neuroscience, 11, 403±407. Murray, R.M., Fearon, P. (1999). The developmental `risk factor' model of schizophrenia. Journal of Psychiatric Research, 33, 497±499. O'Driscoll, G.A., Benkelfat, C., Florencio, P.S., Wolff, A.-L.V.G., Joober, R., Lal, S., Evans, A.C. (1999). Neural correlates of eye tracking de®cits in ®rst-degree relatives of schizophrenic patients: A positron emission tomography study. Archives of General Psychiatry, 56, 1127±1134. Oepen, G., Thoden, U., Warmke, C. (1990). Association of tardive dyskinesia with increased frequency of eye movement disturbances in chronic schizophrenia patients. European Archives of Psychiatry and Neurological Science, 239, 241±245. Olincy, A., Ross, R.G., Young, D.A., Freedman, R. (1997). Age diminishes performance on an antisaccade eye movement task. Neurobiology of Aging, 18, 483±489. Rabe-Hesketh, S., Toulopoulou, T., Murray, R.M. (2001). Multilevel modeling of cognitive function in schizophrenic patients and their ®rst degree relatives. Multivariate Behavioral Research, 36, 279±298. Radant, A.D., Claypoole, K., Wingerson, D.K., Cowley, D.S., Roy-Byrne, P.P. (1997). Relationships between neuropsychological and oculomotor measures in schizophrenia patients and normal controls. Biological Psychiatry, 42, 797±805. Ross, R.G., Olincy, A., Harris, J.G., Radant, A., Adler, L.E., Compagnon, N., et al. (1999). The effects of age on a smooth pursuit tracking task in adults with schizophrenia and normal subjects. Biological Psychiatry, 46, 383±391. Rosse, R.B., Schwartz, B.L., Kim, S.Y., Deutsch, S.I. (1993). Correlation between antisaccade and Wisconsin Card Sorting Test performance in schizophrenia. American Journal of Psychiatry, 150, 333±335. Sereno, A.B., Holzman, P.S. (1995). Antisaccades and smooth pursuit eye movements in schizophrenia. Biological Psychiatry, 37, 394±401. Sha®q-Antonacci, R., Maruff, P., Whyte, S., Tyler, P., Dudgeon, P., Currie, J. (1999). The effects of age and mood on saccadic function in older individuals. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 54, P361ÐP368. Shagass, C., Amadeo, M., Overton, D.A. (1974). Eye-tracking performance in psychiatric patients. Biological Psychiatry, 9, 245±260. Sweeney, J.A., Strojwas, M.H., Mann, J.J., Thase, M.E. (1998). Prefrontal and cerebellar abnormalities in major depression: evidence from oculomotor studies. Biological Psychiatry, 43, 584±594. Sweeney, J.A., Luna, B., Haas, G.L., Keshavan, M.S., Mann, J.J., Thase, M.E. (1999). Pursuit tracking impairments in schizophrenia and mood disorders: Step-ramp studies with unmedicated patients. Biological Psychiatry, 46, 671±680. Sweeney, J.A., Rosano, C., Berman, R.A., Luna, B. (2001). Inhibitory control of attention declines more than working memory during normal aging. Neurobiology of Aging, 22, 39±47. Thaker, G.K., Nguyen, J.A., Tamminga, C.A. (1989). Increased saccadic distractibility in

4. EYE MOVEMENT ABNORMALITIES

91

tardive dyskinesia: Functional evidence for subcortical GABA dysfunction. Biological Psychiatry, 25, 49±59. Thaker, G.K., Cassady, S., Adami, H., Moran, M., Ross, D.E. (1996). Eye movements in spectrum personality disorders: Comparison of community subjects and relatives of schizophrenic patients. American Journal of Psychiatry, 153, 362±368. Thaker, G.K., Ross, D.E., Cassady, S.L., Adami, H.M., Medoff, D.R., Sherr, J. (2000). Saccadic eye movement abnormalities in relatives of patients with schizophrenia. Schizophrenia Research, 45, 235±244. Tien, A.Y., Ross, D.E., Pearlson, G., Strauss, M.E. (1996). Eye movements and psychopathology in schizophrenia and bipolar disorder. Journal of Nervous and Mental Diseases, 184, 331±338. Weinberger, D.R., McClure, R.K. (2002). Neurotoxicity, neuroplasticity, and magnetic resonance imaging morphometry: What is happening in the schizophrenic brain? Archives of General Psychiatry, 59, 553±558.

CHAPTER FIVE

Neuropsychological impairments in patients with schizophrenia and their unaffected relatives Timothea Toulopoulou, Francesca Filbey and Eugenia Kravariti

BACKGROUND In the early 1990s, when the neuropsychological component of the Maudsley Family Study of Psychosis (MFSP) began, it was thought that genes modulating cognitive function might partially overlap with those that increase liability to schizophrenia. This is not a surprising idea given that schizophrenia manifests itself, through its symptoms, to some extent at the cognition level. However, for a long time the de®cits seen in cognition were viewed as a result of the disease, rather than being part of the mechanism that gives rise to it. The shift in thinking occurred for some following a number of reports. First, there were studies suggesting that cognitive impairment is present not only in patients with chronic schizophrenia but even in those who experience their ®rst episode (Gur et al., 1991; Saykin et al., 1991), implying that the cognitive de®cit seen in patients with schizophrenia is not only a result of chronic illness. Similarly, reports of cognitive de®cits among unmedicated patients (Saykin et al., 1994) meant that researchers could no longer dismiss cognitive impairment as a consequence of medication usage. Furthermore, longitudinal studies showed that the cognitive de®cits remain relatively stable throughout the course of illness (Censits et al., 1997; Rund, 1998), reinforcing the idea that the cognitive impairment is a stable trait that does not disappear when the symptoms fade away. Finally, ®ndings of cognitive abnormalities in pre-schizophrenic children or adults (David et al., 1995; Jones et al., 1994) indicated that the 93

94

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

impairment predates the onset of psychosis, and cannot simply be an epiphenomenon of the disease process. Since the early 1990s, studies on ®rst-episode or unmedicated patients, longitudinal studies, investigations of high-risk groups and populationbased prospective studies have all con®rmed that cognitive dysfunction is a core feature of schizophrenia, an integral part of the disorder's pathophysiology (Daban et al., 2005; Hoff et al., 2005; Johnstone et al., 2005; Reichenberg et al., 2006). At the same time, there was evidence that cognitive impairment is present in an attenuated form in some of the healthy relatives of patients with schizophrenia [see Kremen et al. (1994) for a review]. The neuropsychological arm of the MFSP was set out to explore the idea that speci®c cognitive dysfunctions in schizophrenia represent a familial trait of presumed genetic origin. This chapter presents key ®ndings from four major domains of cognitive function explored in the study: memory, executive function, attention and intellectual asymmetry. Study 1 focuses on episodic memory, including new analyses of memory function across Phases 1 to 3 of the study. Study 2 deals with executive function, assessed in Phases 2 and 3 of the study. These studies explore abnormalities associated with schizophrenia in these cognitive domains as well as abnormalities detectable in the unaffected relatives of patients with schizophrenia compared with controls. Further analyses within the relatives group divided by family history and speci®cally in those relatives who were presumed obligate carriers of genetic risk are also presented. Study 3 focuses on attention function, which was assessed in Phase 3 of the MFSP. This study explores abnormalities in patients with schizophrenia compared with controls and in the whole sample of their unaffected relatives, without subdivision into smaller groups based on family history. Study 4 utilizes the novel study design of examining the association between a putative endophenotype of schizophrenia, intellectual asymmetry (verbal± spatial contrast IQ), with the quantitative scale of genetic liability described in Chapter 2 in the unaffected relatives of patients with schizophrenia across Phases 1 to 3 of the study.

STUDY 1: EPISODIC MEMORY Memory can be subdivided into semantic and episodic memory (Tulving, 1989). The former relates to memory for general knowledge whereas the latter refers to memory for events taking place at speci®c times in particular places (Tulving, 2002). Memory dysfunction in schizophrenia, especially episodic memory, is well documented and considered among the most impaired domains of cognitive function in this disorder (Cirillo & Seidman, 2003; Tyson et al., 2005). It is also thought to be affected in at least some of the relatives of patients with schizophrenia.

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

95

Memory function in the relatives of patients with schizophrenia Several studies have reported that the relatives of patients with schizophrenia have de®cits in memory function (Kremen et al., 1994; Sitskoom et al., 2004; Whyte et al., 2005). It has been suggested that verbal memory meets the criteria of a risk indicator for schizophrenia, along with executive functioning and auditory attention (Faraone et al., 1995) Furthermore, there is evidence that memory dysfunction remains stable over time in the relatives of schizophrenia patients (Faraone et al., 1999), supporting the hypothesis that it is a stable trait caused by genes that also increase liability for schizophrenia. In the most recent systematic review and meta-analyses of declarative memory in the unaffected adult relatives of patients with schizophrenia, Whyte and colleagues (2005) reported memory impairments of small to moderate effect sizes with the greatest de®cits on the immediate list recall (d=0.65), immediate story recall (d=0.53) and delayed story recall (d=0.52). This review supported a de®cit during the early stages of memory processing (encoding) as the underlying cause of the manifest memory impairment. The memory component of the MFSP incorporated the following aims in its design (Toulopoulou et al., 2003a, 2003b): (1) regression equations were used to eliminate the effects of intellectual functioning, thus exploring the level of memory function relative to the general intelligence level; (2) the memory tests were chosen to evaluate long-term episodic memory. In other words, tests measuring memory for recently presented information, but not engaging the working memory system; (3) distinctions were made between verbal and visuospatial episodic memory, where the modality distinction refers to the presumed encoding of the information; (4) both immediate and delayed recall was assessed in order to investigate the relative integrity of encoding and consolidation processes in episodic memory; and (5) a further distinction was made between episodic recall and learning in order to explore the retention rates of information accumulated over a single trial or alternatively, over several exposures.

Method Participants A total of 346 individuals took part in this component of the MFSP, this included 226 participants contributing to analyses published previously (Toulopoulou et al., 2003a, 2003b) and a further 120 participants recruited for Phase 3 of the study. The participants comprised 105 patients meeting DSM-IV criteria for schizophrenia (n=100) or schizoaffective disorder (n=5), 150 of their ®rst-degree non-psychotic relatives and 91 healthy volunteers.

96

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Fourteen of the non-psychotic relatives had a previous history of major depressive disorder, and a further four had experienced a single episode of major depression, minor depressive disorder, generalized anxiety disorder or social phobia. All relatives and controls were symptom free at the time of testing.

Neuropsychological measures Russell's version (Russell, 1975) of the Wechsler Memory Scale (WMS; Wechsler, 1945) was used to assess material-speci®c immediate and delayed recall. The procedure encompasses two of the seven WMS subtests ± logical memory and visual reproduction ± testing, respectively, verbal and visual recall. In addition, a third subtest of the WMS ± the associate learning ± and a visual analogue of it ± the abstract paired associates ± were used to measure verbal and visual learning (Goldstein, 1988). Current general intellectual function was assessed with a short-form of the WAIS-R (Wechsler, 1981).

Data analyses For every single test, a series of separate analyses was conducted to compare performance between three groups (patients with schizophrenia, relatives and controls), ®ve groups (familial schizophrenia patients, familial relatives, non-familial schizophrenia patients, non-familial relatives and controls) and between presumed obligate carriers and controls. To adjust for the effect of covariates, a regression model was employed with each test score as the dependent variable and group, sex and age as the predictor factors. In addition, as we wished to identify dysfunctional cognitive systems that exist even when differences in general intellectual function are accounted for, all results presented were adjusted for current IQ. As with any family study, observations of family members who belong to the same family are not likely to be independent (i.e. they share a similar value of the variable). Ignoring within-family correlations would in¯ate the signi®cance of between-group differences. To account for this, the regression equation had a component for clustered observations using a robust estimator for the variances of the regression co-ef®cient estimates.

Results Results from the published work For ease of reference in the discussion, before proceeding to the results of the current analyses, we present here brie¯y the ®ndings of the earlier analyses based on a subsample (n=226) of the current sample and published in Toulopoulou et al. (2003a,b).

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

97

Verbal memory and learning The patients with schizophrenia recalled less information than controls in immediate (z=6.1, p<0.001) and delayed recall (z=6.15, p<0.001). Their relatives also performed less well than controls in both conditions (immediate recall: z=4.6, p<0.001; delayed recall: z=3.8, p<0.001). The difference in immediate recall remained statistically signi®cant, even after excluding those relatives with Axis I diagnosis (z=2.37, p=0.018) and with schizotypal personality disorder (z=3.44, p=0.001). Similarly, the difference in delayed recall between relatives and controls was maintained after excluding those with schizotypal personality disorder (z=2.73, p=0.006). The patients with schizophrenia retained less information over time than controls (z=3.6, p<0.001) and learned less well than controls over several trials (z=3.7, p<0.001). Their relatives did not differ signi®cantly from controls on these measures. Visual memory and learning The patients recalled less visual material in immediate recall than controls (z=3.3, p<0.01). Similarly, their relatives performed worse than controls (z=2.28, p<0.05), although this difference was nulli®ed when considering only the performances of those relatives who were free of psychiatric diagnosis. Impairments were also found in the delayed recall for the patient group (z=3.1, p<0.01), but the relatives performed at a level comparable to that of controls. There were no signi®cant differences in the amount of visual information retained in the delayed condition in relation to the amount of material that was acquired during the ®rst presentation. In terms of visual learning, patients again performed less well than controls (z=4.8, p<0.001). The relatives also performed worse than controls (z=2.2, p<0.05). However, the difference between the relatives and controls disappeared after exclusion of those with an Axis I diagnosis or with schizotypal personality disorder. Further analyses suggested that the patient group showed trends for a selective impairment in immediate verbal recall (t=±1.93, p=0.058) and a selective de®cit in delayed verbal recall (t=±2.05, p<0.05) suggesting that the latter domain was disproportionately more impaired than any other of the measures considered. Their relatives showed a selective impairment in verbal immediate recall (t=±2.29, p<0.05) compared to the other measures.

Results from the current analyses Table 5.1 gives the demographic characteristics of the sample (n=346). There were no signi®cant differences in the level of educational attainment between the groups (2=25.584, p<0.001) but there were signi®cant differences in terms of age (F=52.962, p<.001), gender (2=25.584, p<0.001) and

98

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

TABLE 5.1 Demographic characteristics of schizophrenia patients, relatives and normal control participants for Study 1

Number of subjects Age (range) Age (mean/SD) Sex (female/male) Education (mean/SD) Current IQ

Schizophrenia patients

Relatives

Controls

Signi®cance

105 17±74 34.76 (9.8) 33/72 13.27 (2.819) 95.31 (16.5)

150 16±74 51.23 (14.4) 94/56 13.56 (2.793) 110.07 (16.6)

91 18±77 39.30 (14.4) 52/39 13.93 (3.300) 111.4 (15.8)

F=52.962, p<0.001 2=25.584, p<0.001 F=1.178, p=0.309 F=30.66 p<0.001

current IQ (F=30.66, p<0.001). Age differences were mainly because the relative group included parents who, by de®nition, have to be older than their offspring. Controls roughly re¯ected the diverse age ranges of the two experimental groups. The gender distribution was unequal across groups because the schizophrenia sample contained more men whereas the relative sample contained more women. All results presented below are adjusted for age, sex and IQ differences. Verbal memory and learning Tables 5.2 and 5.3 show the mean scores, standard deviations and the age, sex and current IQ adjusted t and p values for verbal memory in patients with schizophrenia, their relatives and control participants. The patients with schizophrenia recalled less information than controls in immediate and delayed recall and retained less information over time compared to controls. In addition, they found it harder to learn associating words over time. Their relatives, however, performed at comparative level to that of controls on all these measures. The analyses looking separately at those with or without a family history of schizophrenia indicated that both groups of patients had dif®culties with verbal memory and associate learning. In addition, some of their relatives ± those from the non-familial sample ± also performed signi®cantly worse than controls on the immediate and delayed recall of verbal memory (Table 5.3). In contrast, the familial relatives, including the presumed obligate carrier group of relatives, did not show any such de®cits. Visual memory and learning The results for the visual memory and learning are given in Tables 5.4 and 5.5. The patients recalled less information than controls in the delayed recall of visual memory and in the abstract associate learning task. As with the verbal memory and learning, there were no signi®cant differences between the total sample of relatives and controls. Irrespective of whether

8.1Ô3.9 t=±4.84, p<0.0001 5.6Ô3.98 t=±4.85, p<0.0001 65.2Ô29.2 t=±3.89, p<0.0001 13.8Ô4.0 t=±3.51, p<0.001

Logical memory (immediate recall)

Logical memory (delayed recall)

Logical memory (% retained)

Associate learning

Patients versus controls

15.2Ô3.8 t=±0.29, p=0.769

76.2Ô19.3 t=±0.89, p=0.373

8.4Ô3.7 t=±1.27, p=0.206

10.8Ô3.5 t=±1.11, p=0.269

Relatives versus controls

16.6Ô3.3

81.1Ô14.8

10.05Ô3.6

12.2Ô3.4

Controls

TABLE 5.2 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in verbal memory; patients and relatives versus controls

7.9Ô4.1 t=±3.93 p<0.0001

5.6Ô4.1 t=±3.81 p<0.0001

67.8Ô30.5 t=±2.42 p=0.02

13.3Ô4.1 t=2.90 p=0.004

Logical memory (immediate recall)

Logical memory (delayed recall)

Logical memory (% retained)

Associate learning

15.3Ô4.1 t=0.51 p=0.61

78.5Ô19.2 t=0.25 p=0.80

9.1Ô3.9 t=0.32 p=0.753

11.3Ô3.6 t=0.47 p=0.639

74

Familial relatives2

14.2Ô3.9 t=±4.02 p<0.0001

62.3Ô27.8 t=±3.55 p=0.001

5.6Ô3.8 t=±4.26 p<0.0001

8.3Ô3.8 t=±4.00 p<0.0001

51

Non-familial patients3

2

Familial schizophrenia patients versus controls. Familial relatives (including presumed obligate carriers) versus controls. 3 Non-familial schizophrenia patients versus controls. 4 Non-familial relatives versus controls. 5 Presumed obligate carriers versus controls.

1

54

Number of subjects

Familial patients1

15.1Ô3.6 t=±0.56 p=0.578

74.1Ô19.2 t=±1.67 p=0.096

7.8Ô3.4 t=±2.39 p=0.018

10.2Ô3.3 t=±2.26 p=0.025

76

Non-familial relatives4

14.5Ô4.9 t=±0.04 p=0.96

71.2Ô21.0 t=±1.03 p=0.30

8.6Ô3.7 t=±0.06 p=0.951

11.9Ô3.8 t=0.76 p=0.452

17

Presumed obligates5

16.6Ô3.3

81.1Ô14.8

10.05Ô3.6

12.2Ô3.4

91

Controls

TABLE 5.3 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in verbal memory; familial patients and relatives; non-familial patients and relatives, and presumed obligates versus controls

8.01Ô4.01 t=±1.90, p<0.059 6.1Ô4.1 t=±2.21, p=0.028 77.05Ô41.4 t=±0.65, p=0.513 11.6Ô3.6 t=±3.51, p=0.001

Visual reproduction (immediate recall)

Visual reproduction (delayed recall)

Visual reproduction (% retained)

Associate learning

Schizophrenia patients

13.3Ô3.2 t=±0.29, p=0.769

79.0Ô30.7 t=0.17, p=0.863

7.1Ô4.0 t=0.25, p=0.800

8.8Ô3.7 t=±0.24, p=0.813

Relatives

14.7Ô3.2

83.7Ô21.5

8.7Ô3.5

10.3Ô3.0

Controls

TABLE 5.4 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in visual memory; schizophrenia patients and relatives versus controls

7.6Ô4.3 t=±1.54 p=0.126

6.2Ô4.1 t=±1.18 p=0.238

83.4Ô47.3 t=0.64 p=0.525

11.6Ô3.8 t=±2.90 p=0.004

Visual reproduction (immediate recall)

Visual reproduction (delayed recall)

Visual reproduction (% retained)

Abstract paired associates

83.7Ô21.5

14.7Ô3.2 12.0Ô3.1 t=±0.60 p=0.551 13.2Ô3.2 t=±0.56 p=0.578

11.7Ô3.3 t=±4.02 p<0.0001

13.5Ô3.2 t=0.51 p=0.612

2

8.7Ô3.5

10.3Ô3.0

91

Controls

75.3Ô22.2 t=±0.04 p=0.967

5.9Ô3.1 t=±0.47 p=0.640

8.2Ô4.0 t=±0.04 p=0.965

17

Presumed obligates5

79.8Ô32.8 t=0.31 p=0.756

7.1Ô4.0 t=±0.20 p=0.839

8.8Ô3.6 t=±0.65 p= 0.519

76

Non-familial relatives4

70.5Ô33.5 t=±1.95 p=0.052

6.0Ô4.1 t=±2.41 p=0.017

8.3Ô3.6 t=±1.43 p=0.155

51

Non-familial patients3

78.2Ô28.6 t=0.18 p=0.858

7.2Ô3.9 t=0.89 p=0.374

8.7Ô3.8 t=0.26 p= 0.795

74

Familial relatives2

Familial schizophrenia patients versus controls. Familial relatives (including presumed obligate carriers) versus controls. 3 Non-familial schizophrenia patients versus controls. 4 Non-familial relatives versus controls. 5 Presumed obligate carriers versus controls.

1

54

Number of subjects

Familial patients1

TABLE 5.5 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in visual memory; familial patients and relatives; non-familial patients and relatives, and presumed obligates versus controls

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

103

patients had a family history of schizophrenia or not, they performed worse than controls in visual learning. In addition, the non-familial patients showed impairments in both measures of the delayed recall. Neither the familial relatives (including the presumed obligate carriers) nor their nonfamilial counterparts showed any impairment in this domain.

Discussion This work con®rms what numerous studies have previously demonstrated: a memory dysfunction in schizophrenia (Hill et al., 2004; Nuyen et al., 2005). The impairment was seen in both the verbal and visual modalities of episodic memory. De®cits were present despite adjusting for differences in the current general intellectual level, suggesting that memory function in this group of patients was disproportionately lower than would have been expected given the level of their general intellectual function. Compared to controls, the familial and non-familial probands performed similarly in verbal memory and visual learning but differed in visual memory. A general lack of differences between these two groups is compatible with much of the literature (Roy & Crowe, 1994). In agreement with most of the published literature (Faraone et al., 1999; Goldberg et al., 1995), including a recent systematic review and metaanalysis of declarative memory in unaffected adult relatives of patients with schizophrenia (Whyte et al., 2005), evidence of subtle impairments in episodic memory were identi®ed in some of the relatives (non-familial) of the patients with schizophrenia. The decreased performances were present in both the immediate and delayed recall conditions, whereas the percentage retention score ± an index of forgetting ± was unimpaired, suggesting that the memory de®cit was possibly the consequence of an inability to acquire the information in the ®rst place rather than the result of diminishing memory trace. Story recall relies on the integrity of short-term and long-term processes, which are thought to be mediated by a network of cortical and subcortical brain structures. Associate learning was not impaired, consistent with other published data (Goldberg et al., 1993, 1995). Associate learning assesses learning or retention of verbal material that has been accumulated over a number of trials. Normal levels of associate learning in combination with appropriate levels of retention scores on episodic memory in the relatives of patients with schizophrenia gives further strength to the idea that the observed de®cit in verbal memory among the relatives is not underlined by an impairment related to consolidation processes. Consistent with some studies, neither visual immediate recall (Goldberg et al., 1993, 1995) nor delayed recall (Cannon et al., 1994) nor the abstract paired associates test was affected in this sample of relatives. This pattern of impaired verbal memory with intact spatial components supports laterality

104

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

effects, implicating greater left hemisphere dysfunction. The ®nding is compatible with some previous work on the relatives of patients with schizophrenia (Cannon et al., 1994; Kremen et al., 1994). Furthermore, the results are not inconsistent with the recent meta-analyses on declarative memory mentioned above, which found much smaller random effects mean weighted effect sizes in the immediate visual recall (d = 0.32) and the delayed visual recall (d=0.32) compared to the immediate (d=0.53) and the delayed story recall (d=0.52). As suggested earlier, verbal memory impairment has been viewed by some (but see Cannon et al. (2000)) as a promising indicator of disorderrelated genotype (Faraone et al., 1995). This potential endophenotype appears to be transmitted, in this sample, within singly-affected families only. The familial relatives were unimpaired, including those classi®ed as presumed obligate carriers, who were hypothesized to be the most impaired with regard to any potential endophenotype. It is intriguing that these familial relatives failed to show any de®cit in memory function. It is possible that, because the present presumed obligate carrier sample is relatively small (n=17), it lacks suf®cient statistical power to detect signi®cant differences even if in reality these were present. In a paper examining the cognitive performance of presumed obligate carriers for psychosis that included relatives from multiply-affected bipolar families and incorporated a portion of the sample included here, we reported de®cits in verbal memory and visuospatial manipulations among the presumed obligate carrier sample. We suggested that these individuals seem to transmit vulnerability for psychosis to their offspring that incorporated an impairment to recall verbal material and inability in perceiving spatial relations (Toulopoulou et al., 2005). Similarly, an investigation concurrent with the present study examining the structural brain abnormalities in a partly overlapping sample of presumed obligate carriers found that 27% of them had brain abnormalities compared to 16% of the nonobligate carrier familial relatives and 11% of unrelated controls (Sharma et al., 1997) and that these relatives were most likely to display lateral ventricular enlargement (McDonald et al., 2002). These results had strengthened the hypothesis that presumed obligate carriers have brain abnormalities and had increased our expectation of ®nding similar abnormalities at the memory function level in the current study. However, our ®ndings in relation to family history are not unique in the neuropsychological literature. Although, Faraone et al. (2000) found more impairments in relatives from multiplex families than in those from simplex families, other studies failed to show increased levels of cognitive dysfunction with enhanced genetic loading for schizophrenia (Byrne et al., 2003). It is of interest that the memory de®cit was not signi®cant when the analyses were done on the total sample of the relatives of patients with

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

105

schizophrenia. Our previous published work, based on a smaller subsample (n=226) of the present population, suggested that the relatives were impaired in the immediate and delayed recall of verbal memory as well as in the immediate recall of visual reproduction and the abstract paired associates test (Toulopoulou et al., 2003a, 2003b). In fact, the difference between the relatives and controls in verbal memory had remained signi®cant even after excluding those relatives with a lifetime psychiatric diagnosis. Furthermore, our previous work suggested that the immediate verbal recall de®cits were disproportionately more impaired compared to the other cognitive measures and this survived adjustments for age, gender and current IQ. It was expected, therefore, that impairments ± at least in the immediate recall of verbal memory ± would also be present in this sample; however, the current analyses instead demonstrated that these initial ®ndings were diluted in the enlarged sample. To summarize, we found that patients with schizophrenia had de®cits in most aspects of memory functioning. Non-familial relatives also had de®cits in the immediate and delayed verbal recall. Overall, these results suggest that episodic memory impairments may be a familial trait, which indicates increased risk for schizophrenia.

STUDY 2: EXECUTIVE FUNCTION Executive processing refers to the ability to control and sequence cognition and behaviour. It is an umbrella term encompassing separable components of cognitive function, including the generation and implementation of action plans with the aim to achieve a goal state, the monitoring and updating of information held for brief moments in time, the formation of strategic plans to enhance performance, and the diversion or shifting of attention in a controlled way. In the MFSP, executive function was assessed by administering tests that evaluated: planning ability, working memory, strategy formation and rapid mental ¯exibility. Here, we focus on the ®rst three, as the test we used to assess the last domain was not sensitive enough to detect subtle de®cits [for more information, see Toulopoulou et al. (2003a, 2003b)].

Executive function in patients with schizophrenia and their relatives Schizophrenia involves extensive compromise of executive function (Fioravanti et al., 2005) as characterized by de®cits in higher-level planning (Rushe et al., 1999), working memory (Badcock et al., 2005) strategy formation (Joyce et al., 2002) and attentional set-shifting (Pantelis et al., 1999). Impaired executive processing is present at the early stages of the development of psychosis (Bartok et al., 2005) and in ®rst-episode schizophrenia (Joyce et al., 2002), and appears to be stable across time.

106

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Most studies of executive function in the relatives of patients with schizophrenia primarily focused on measuring the ability to form and shift attentional set and the capacity to demonstrate mental ¯exibility. These processes are frequently assessed by the Wisconsin Card Sorting Test (WCST; Heaton, 1981) and the Trail Making Test (Reitan, 1958). Both tasks had yielded contradictory results, with some studies ®nding differences among relatives of schizophrenia patients and controls and others ®nding no such differences. Nonetheless, recent meta-analyses on the relatives of patients with schizophrenia reported Trails B as having one of the largest effect sizes compared to other cognitive tests (d=0.51; Sitskoorn et al., 2004; Snitz et al., 2006), implying that set shifting is among the most impaired domains of cognitive function among the relatives of patients with schizophrenia. In addition to assessing concept formation and mental ¯exibility, studies have examined another component of executive processing in recent years, that of working memory. One of the ®rst studies examining this domain reported de®cits in 40±50% of the relatives assessed (Park et al., 1995), a ®nding subsequently replicated in other studies. The genetic determination of working memory (in particular spatial working memory) was supported by reports of a higher correlation within monozygotic than within dizygotic pairs (Cannon et al., 2000), making it a good candidate endophenotype for schizophrenia. Overall, impairments in executive function among relatives of patients with schizophrenia have become more recognized, especially after the latest meta-analysis, which suggested that the largest effect sizes were associated with tests that load highly on executive control functions, such as working memory, set shifting and suppression of pre-potent responses (Snitz et al., 2006). The aim of this component of the MFSP was to characterize the nature of executive dysfunction in patients with schizophrenia and their relatives by examining each of the main components of executive processing separately in the same sample of patients with schizophrenia and their nonpsychotic relatives. As with the study on episodic memory, the schizophrenia sample and their respective relatives were further examined after subdividing according to the patient's family history status (familial/nonfamilial sample) to assess whether presumed variation in genetic loading affected executive function.

Method Participants Sixty-three schizophrenia patients, 98 of their non-psychotic relatives (including 10 presumed obligate carriers from 51 families) and 66 normal

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

107

controls participated in this component of the MFSP (227 participants in total).

Neuropsychological measures Executive processing was assessed by administering tests of: (1) planning ability, measured using the three-dimensional computerized Tower of London task; and (2) spatial working memory and strategy formation, tested using the Executive Golf task (Toulopoulou, 2003a, 2003b). Four measures were identi®ed in the three-dimensional computerized Tower of London Test: 1.

2. 3. 4.

Planning time: this refers to time taken to think about the problem prior to initiation of movement. Performance in this variable should be interpreted with caution and within the context of performance in the other Tower of London measures, as taking too little or too much time to plan can be equally disadvantageous. Subsequent execution/thinking time: representing time taken between selection of the ®rst disk and completion of the problem. Accuracy of problem solving: this describes the average number of moves required to solve the problems. Motor execution time (as measured by the control condition): this re¯ects time taken to execute responses independent of planning ability.

The results on planning and subsequent execution/thinking times are given after adjusting for motor performance by subtracting the equivalent motor execution times as measured by the control condition. For ease of reporting, all analyses were carried out after collapsing scores across the levels of dif®culty. Three measures were identi®ed on the Executive Golf task: (1) withinsearch errors, number of errors within a single search; (2) between-search errors, number of errors between searches; and (3) strategy formation, a measure of a self-generated search sequence that could facilitate performance. All analyses were carried out after collapsing scores for level of dif®culty.

Data analyses We used the same approach as for the memory data. Brie¯y, regression equations were used to take in to account the effects of age, sex, IQ and the non-independence of observations for people who belong to the same family.

108

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

TABLE 5.6 Demographic characteristics of schizophrenia patients, relatives and normal control participants for Study 2

Number of subjects Age (range) Age (mean/SD) Sex (female/male) Education (mean/SD) Current IQ

Schizophrenia patients

Relatives

Controls

Signi®cance

63 17±55 34.1 (9.1) 22/41 12.8 (2.6) 95.0 (17.7)

98 18±74 49.5 (15.8) 60/38 13.5 (2.8) 109.0 (19.1)

66 18±77 38.8 (14.2) 33/33 13.5 (3.2) 110.2 (15.9)

F=26.4, p<0.001 2=10.631, p=0.005 F=1.3, p=0.264 F=14.4, p<0.001

Results The demographic characteristics of the schizophrenia patients, relatives and normal controls are shown in Table 5.6. No signi®cant differences were found in the level of education (F=1.3, p=0.264), but a signi®cant variation was found in age distribution across the three groups (F=26.4, p<0.001). The age disparity between the schizophrenia sample and some of their relatives was large (the relative group included parents who ± naturally ± were quite a lot older than their offspring). The average age of the control group was intermediate between that of patients and relatives, re¯ecting the fact that controls were recruited to mirror the age ranges found in the two experimental groups of patients and relatives. The gender distribution was unequal across groups (2=11.8, p=0.003), with the schizophrenia sample containing more men and with the group of relatives including more women. In addition, as expected, there were signi®cant differences in IQ, mainly due to the difference between patients and the other two groups (F=14.4, p<0.001). Analysis throughout was carried out adjusting for age, gender and IQ confounding effects.

Planning ability Table 5.7 shows the mean scores, standard deviations and the age, sex and current IQ adjusted t and p values in planning ability for patients with schizophrenia, relatives and normal controls. Table 5.8 depicts the corresponding values for the familial and non-familial groups. Planning time and subsequent execution/thinking time All groups took similar planning times, with none of the comparisons yielding any signi®cant differences, including those on the familial and nonfamilial sample. Similarly, neither schizophrenia patients nor their relatives

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

109

TABLE 5.7 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in planning ability; patients and relatives versus controls

Planning time Subsequent execution/ thinking time Accuracy of problem-solving (number of moves) Motor execution time

Patients versus controls

Relatives versus controls

Controls

5.42Ô0.46 t=±0.08, p=0.939 13.46Ô1.2 t=±0.02, p=0.988 5.93Ô0.16 t=2.16, p=0.03 4.30Ô0.21 t=3.86, p<0.0001

6.67Ô0.52 t=±0.02, p=0.982 16.14Ô1.70 t=0.92, p=0.361 5.58Ô0.13 t=1.52, p=0.13 3.95Ô0.14 t=1.32, p=0.19

5.78Ô0.41 10.60Ô1.2 5.23Ô0.11 3.21Ô0.10

were signi®cantly different from controls on subsequent execution/thinking times, suggesting that all groups took suf®ciently adequate time to plan and execute their moves. Accuracy of problem-solving Patients with schizophrenia took signi®cantly more moves than controls to solve the series of problems. There were no signi®cant differences between relatives and controls when looking at the sample of relatives in total. Subsequent analysis subdividing the sample into familial/non-familial suggested that, although neither familial patients with schizophrenia nor their relatives differed from controls in this measure, both non-familial groups, i.e. the patients with the non-familial schizophrenia and their relatives differed signi®cantly from controls. Motor execution time Although patients with schizophrenia took longer to execute motor responses than controls, there were no such signi®cant differences between relatives and controls. The separate analyses on the familial/non-familial samples suggested that both groups of patients with schizophrenia performed more poorly than controls but that their relatives performed similarly to controls.

Spatial working memory and strategy formation The results on the spatial working memory and strategy formation analyses are given on Tables 5.9 and 5.10.

6.49Ô0.85 t=1.01 p=0.315

13.06Ô1.34 t=±0.81 p=0.418

5.85Ô0.26 t=0.97 p=0.334

4.52Ô0.31 t=2.88 p<0.005

Planning time

Subsequent execution/thinking time

Accuracy of problem-solving (number of moves)

Motor execution time

3.99Ô0.21 t=1.29 p=0.200

5.25Ô0.12 t=±0.68 p=0.497

13.89Ô2.41 t=±0.25 p=0.801

7.25Ô0.89 t=0.88 p=0.379

52

Familial relatives2

4.12Ô.29 t=2.78 p=0.006

6.01Ô0.21 t=2.18 p=0.031

13.81Ô1.99 t=0.33 p=0.745

4.51Ô0.40 t=±0.96 p=0.341

34

Non-familial schizophrenia patients3

2

Familial schizophrenia patients versus controls. Familial relatives (including presumed obligate carriers) versus controls. 3 Non-familial schizophrenia patients versus controls. 4 Non-familial relatives versus controls. 5 Presumed obligate carriers versus controls.

1

29

Number of subjects

Familial schizophrenia patients1

3.91Ô0.19 t=0.69 p=0.494

5.94Ô0.22 t=2.76 p=0.007

18.68Ô2.3 t=1.55 p=0.123

6.00Ô0.43 t=±1.27 p=0.206

46

Non-familial relatives4

4.84Ô2.79 t=0.94 p=0.349

5.77Ô1.15 t=1.44 p=0.153

16.21Ô10.80 t=1.09 p=0.278

8.56Ô7.31 t=0.69 p=0.492

10

Presumed obligates5

3.22Ô.11

5.23Ô0.11

10.61Ô1.22

5.78Ô0.41

66

Controls

TABLE 5.8 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in planning ability; familial schizophrenia patients and relatives; non-familial schizophrenia patients and relatives, and presumed obligates versus controls

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

111

TABLE 5.9 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in spatial working memory and strategy formation; patients and relatives versus controls

Within-search errors Between-search errors Strategy formation1 1

Patients versus controls

Relatives versus controls

Controls

0.78Ô.15 t=1.98, p=0.05 4.94Ô0.39 t=2.99, p<0.003 13.42Ô0.22 t=2.98, p=0.004

0.58Ô0.08 t=±1.01, p=0.313 3.47Ô0.26 t=±0.13, p=0.900 13.14Ô.18 t=2.16, p=0.033

0.39Ô0.07 2.74Ô0.29 12.00Ô0.28

A higher score denotes lesser use of strategy.

Within-search and between-search errors Patients with schizophrenia made more within-search and between-search errors than controls; there were no signi®cant differences between relatives and controls. The familial and non-familial division did not yield any signi®cant differences, except for the comparison between familial relatives and controls, which suggested that familial relatives were better than controls, making fewer within-search errors. On between-search errors, apart from the non-familial patients, no other group performed signi®cantly worse than controls. Strategy formation Patients with schizophrenia and their relatives made fewer searches starting from the same location than controls. However, a closer inspection suggested that neither familial patients with schizophrenia nor their relatives (including presumed obligate carriers) made signi®cantly less use of strategy than controls. Patients with non-familial schizophrenia and their relatives, however, made less use of strategy than controls.

Discussion Planning ability Patients with schizophrenia took longer to execute motor responses and made more moves above the ideal minimum when compared to controls, suggesting impairment in planning ability; their relatives showed no such de®cit. Subsequent analysis looking for potential differential planning performances of probands with different family-history status (i.e. familial/ non-familial) showed that only patients with non-familial schizophrenia, but not their familial counterparts, made more moves above the ideal

12.84Ô0.29 t=0.98 p=0.328

Strategy formation6 13.54Ô0.27 t=3.07 p=0.003

5.52Ô0.61 t=3.33 p<0.001

0.85Ô0.22 t=1.81 p=0.073

34

Non-familial patients3

2

Familial schizophrenia patients versus controls. Familial relatives (including presumed obligate carriers) versus controls. 3 Non-familial schizophrenia patients versus controls. 4 Non-familial relatives versus controls. 5 Presumed obligate carriers versus controls. 6 A higher score denotes lesser use of strategy.

1

3.14Ô0.35 t=±0.96 p=0.339

Between-search errors 4.23Ô0.45 t=1.18 p=0.240

13.27Ô0.35 t=1.87 p=0.064

0.44Ô0.09 t=±2.09 p=0.039

0.69Ô0.18 t=1.00 p=0.321

Within-search errors

52

29

Familial relatives2

Number of subjects

Familial patients1

13.47Ô0.18 t=3.23 p=0.002

3.83Ô0.38 t=0.72 p=0.473

0.74Ô0.12 t=0.14 p=0.889

46

Non-familial relatives4

13.48Ô1.28 t=0.50 p=0.619

5.08Ô3.6 t=0.96 p=0.343

0.91Ô1.36 t=±0.04 p=0.966

10

Presumed obligates5

12.00Ô0.28

2.74Ô0.29

0.39Ô0.07

66

Controls

TABLE 5.10 Mean scores, standard deviations and age, sex and current IQ adjusted t and p values in spatial working memory and strategy formation; familial patients and relatives; non-familial patients and relatives, and presumed obligates versus controls

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

113

minimum when compared to controls; their relatives (i.e. non-familial relatives) showed a similar pattern on the latter measure. In evaluating performance on the Tower of London task, the number of moves required to solve the problems and both the initial and subsequent thinking latencies were assessed. Overall, patients with schizophrenia and their relatives were able to provide solutions to the problems, suggesting that there were no major impairments. However, patients and some of their relatives (non-familial relatives) completed the tasks inef®ciently, taking signi®cantly more moves above the ideal minimum when compared to controls. Patients with schizophrenia also had longer motor response times, as measured by the three-dimensional CTL-Control task, and in calculating thinking latencies all times were adjusted to account for this. Inaccurate overall performance in a context of taking normal times in planning and executing a sequence of moves might indicate that patients with schizophrenia (in particular the non-familial group) and the relatives from the non-familial sample had a tendency to initiate responses before a plan was fully formed. This was despite not taking very short planning times. These premature responses could be due to impulsivity or to the failure of response inhibition mechanisms (Morris et al., 1995).

Spatial working memory and strategy formation Patients with schizophrenia made more within-search and between-search errors than controls, indicating that they have dif®culties with retaining information for brief moments in time. In addition, neither the patients nor their relatives spontaneously came up with the strategy that most controls generate when tackling the task. The familial/non-familial comparison revealed that only the non-familial schizophrenia group made more between-search errors and made inef®cient use of strategy when compared to controls; their relatives (i.e. non-familial relatives) also showed impairment on the latter measure. The Executive Golf task provides a measure of `on-line' processing of spatial information and explores the extent to which the implementation of strategic algorithms facilitates task performance. In evaluating performance, two types of searching errors (within-search and between-search errors) and an index of strategy formation were calculated. Patients with schizophrenia, but not their non-psychotic relatives, made signi®cantly more within-search errors (i.e. re-selecting a location within a single search) than controls, indicating that patients found it dif®cult to retain simple self-generated spatial sequences. Similarly, patients with schizophrenia made signi®cantly more between-search errors (selecting a position that has already been successful in a single game) than controls. The cognitive processes involved in avoiding these types of error are probably more complex than those

114

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

involved in avoiding within-search errors, perhaps placing a greater load on short-term memory systems such as the visuospatial scratch pad. To explore the extent to which strategy impairment might contribute to spatial working memory de®cits in this sample of patients and relatives, the computer was programmed to record their search sequences. Earlier investigations using a virtually identical test suggested that normal control participants, when tackling the spatial working memory task, use a repetitive search pattern that essentially retraces a former search sequence but modi®es it to take into account previously reinforced positions (Robbins, 1998). This self-generated strategy can substantially reduce the load on memory systems, making the task substantially easier to perform. In the present study, patients and their relatives made less use of strategy, and this lack of generating or implementing a strategy could have contributed, at least in part, to the observed spatial working memory de®cits found in patients. Although the relatives were unable to produce the search strategy most normal control participants generate, they showed no signi®cant spatial working memory impairments, possibly suggesting that they relied more heavily on their spatial working memory ability to approach the task. When the analyses distinguished between familial and non-familial patients with schizophrenia, only the latter showed spatial working memory impairments and de®cits in strategy formation. Their relatives (non-familial relatives) also showed dif®culties in coming up with a strategy to solve the problems. As with the ®ndings in the memory data, therefore, when differentiating by family history, only the non-familial group show the de®cits outlined above. To summarize, aspects of executive functioning involving planning ability, spatial working memory and strategy formation were investigated in a group of patients with schizophrenia and their ®rst-degree relatives. Patients with schizophrenia and their relatives showed speci®c de®cits of executive functioning when compared to controls. Patients with schizophrenia showed impairments in planning, spatial working memory and strategy formation. Some of their relatives also showed de®cits in planning ability and strategy formation. Impairments were associated with the nonfamilial relative sample only, and support this measure as an intermediate phenotype in these families.

STUDY 3: SUSTAINED AND SELECTIVE ATTENTION Attention in schizophrenia The ability to maintain goals over a period of time, known as sustained attention, has been widely studied in patients with schizophrenia, and impaired sustained attention, as measured by the Continuous Performance

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

115

Test (CPT), has been well-replicated for decades (Orzack & Kornetsk, 1966; Walker & Shaye, 1982). Some studies have reported that those with a family history of the illness perform more poorly than those without a family history (Walker & Shaye, 1982), and that de®cits as measured by the CPT are trait symptoms independent of the disease process (Nieuwenstein et al., 2001). In addition to sustained attention, selective or focused attention is also dysfunctional in people with schizophrenia. A widely used index of impaired inhibition, the Stroop effect, is commonly impaired in patients with schizophrenia. This phenomenon is observed when the time taken to read a word in the presence of an interfering incongruent colour is much longer than when there is no interference (i.e. colour of ink is congruent to word). Patients with schizophrenia generally show greater Stroop interference than normal participants, which has been attributed to an inability to attend to the task-relevant dimension (Barch et al., 2004; Hepp, 2007).

Impairments in attention in the relatives of people with schizophrenia Impaired attention has received great interest in the last decade because de®cits have also been observed in the unaffected family members of people with schizophrenia and, thus, may serve as an indicator of genetic vulnerability to schizophrenia (Gold & Harvey, 1993). Of the attention domains, sustained attention de®cits may be the most promising endophenotype for schizophrenia (Gooding et al., 2006). Mirsky and colleagues (1992) reported in a cohort of Israeli and Irish families in which sustained attention (along with focused attention) differentiated the groups the most. Using various versions of the CPT, de®cits of sustained attention in healthy relatives of patients schizophrenia have been shown to be similar to that of their respective probands (Chen & Faraone, 2000) to be disease-speci®c (Asarnow et al., 2002) and to increase schizophrenia risk ratios compared with schizophrenia alone (Chen et al., 2004).

Methods Participants Table 5.11 describes the demographic characteristics of the participants included in these experiments.

Tasks of attention A 14-minute computerized test was used to test visual sustained attention. The test involved 360 random single-letter presentations with a 10% target

116

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS TABLE 5.11 Sociodemographic means of subjects per group classi®cation for Study 3

Number of subjects Age (at time of testing) Sex (male/female) WAIS-R full IQ Level of education10

Patients

Relatives

Controls

55 35Ô91,2 37/184,5 95Ô136,7 3.42Ô1.18

99 49Ô151,3 41/584 111Ô126 3.29Ô1.39

61 40Ô152,3 29/325 112Ô127 3.34Ô1.58,9

p<.001, 2,5p<.05, 8,9p<.01, 101, no quali®cations; 2, Certi®cate of secondary education (CSE); 3, O levels; 4, A levels; 5, university degree. 1,3,4,5,7

chance. Each letter was displayed for 250 milliseconds with random interstimulus intervals of 1, 2 or 4 seconds. Participants were instructed to respond to every individual letter on the screen by button pressing, except when the letter was `X'. Variables measured included: 1.

2.

3.

Beta-value ( ) t-score: measure of overall rate of information processing or response tendency; = Y-axis height of normal curve using proportion of commission errors/Y-axis height of normal curve using proportion of hits. t-scores represent the score of the participant taking the test relative to scores obtained by the comparative study group. A tscore of 50 represents the average of the comparison group with a standard deviation of 10. t-scores of <40 represent individuals who respond more frequently than normal; t-scores >60 indicate participants who respond less frequently than normal. d-prime (d©) t-score: measure of overall perceptual sensitivity or how well participants discriminated targets from non-targets; d© = z-score normal curve deviate using proportion of commission errors ± z-score normal curve deviate using proportion of hits. t-scores >60 indicate poor performance. Index: measure of overall CPT performance; weighted sum of all variables measured: number of hits, reaction time to target, number of omission errors, number of commission errors, hit reaction time standard error, variability of standard errors, d© and . A score of 8±11 corresponds to borderline, while a score of >11 would suggest attention problems.

An auditory CPT modi®ed from that used by Faraone and colleagues (1995) was employed to test sustained attention to auditory stimuli. This is a 10-minute test in which participants were instructed to respond to the letter `A' by button pressing. Three hundred single-letter stimuli were

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

117

presented randomly at 2-second intervals with a 10% target chance. Hit reaction time to targets was measured. Two different tasks were performed to assess selective attention in participants. We utilized the widely used Stroop Colour±Word task (Trenerry et al., 1989), which tests the ability to attend only to the colour aspect of a word while ignoring the distractor (i.e. how the word is read; Plate 2). We measured colour±word score (SCORE), which is the number of correct responses minus incorrect responses (cut-off score: 18±49 age group = 99; 50+ age group = 62). We also used the Letter Cancellation task (Diller, 1974), which measures the ability to quickly select targets by performing a timed test wherein two target letters (`E' and `C') are identi®ed within other non-target letters (distractors). Normative data indicate a performance limit of 0±2 errors within a total performance time of 2 minutes. The total reaction time (RT) to complete the test was recorded.

Analyses The same type of analyses as described in Study 1 and 2 were employed. In addition, the analyses were re-run to control for differences in education instead of current IQ. Demographic differences between the patients, relatives and controls were examined using analyses of variance.

Results Comparisons of the demographic variables between groups indicated that the patients were signi®cantly younger than controls (t=2, p<0.05) and their well relatives (t=6, p<0.001) during assessment. The patient group also contained signi®cantly fewer women than the relatives (2=9, p<0.001) and the control group (2=4, p<0.05). With regards to IQ, the patients had signi®cantly lower IQ than the relatives (t=6, p<0.001) and the normal controls (t=4, p<0.001). The relatives were signi®cantly older than controls at assessment (t=3 p<0.001), but were both similar in IQ and gender distribution compared to the normal controls. Both the patients (2=4, p<0.01) and the relatives (2=4, p<0.01) were signi®cantly different from the normal controls on their highest level of education completed (see Table 5.11). Table 5.12 provides the mean performance scores of each group for each attention variable. During the visual CPT, the patients had that indicated a response pattern less frequent than normal (>60). However, none of the groups had d© that indicated poor perceptual sensitivity (>60) or indicated attention problems as measured by the CPT-visual index (>11). As expected, compared to normal controls, the patients performed signi®cantly worse on the visual CPT (t=2.85, p<0.01) and index (t=2.06, p<0.05), Stroop score (t=4.23, p<0.001) and LC RT (t=3.81, p<0.001), despite controlling for age, gender and IQ. A trend-level difference was also

118

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

TABLE 5.12 Attention measures per group when compared to the normal controls with age, gender and IQ covariates

CPT1 d© Index A-CPT2 RT Stroop score LC3 RT

Patients mean Ô SD

Relatives mean Ô SD

Controls mean Ô SD

56Ô14 t=0.01, p=1.0 69Ô23 t=2.85, p<0.01 6.1Ô5.6 t=2.06, p<0.05

49Ô11 t=1.33, p=0.2 59Ô20 t=1.24, p=0.2 3.6Ô4.1 t=1.37, p=0.2

52Ô8

149Ô151 t=1.78, p=0.08

125Ô88 t=1.84, p=0.07

96Ô66

102Ô90 t=4.23, p<0.001

102Ô13 t=1.1, p=0.3

106Ô9

144Ô51 t=3.81, p<0.001

126Ô36 t=.84, p=0.4

118Ô42

54Ô15 3.3Ô4.3

1

CPT, continuous performance test. A-CPT RT, auditory continuous performance test mean reaction time to hits in ms. 3 LC RT, letter cancellation total response time in seconds. 2

found in the A-CPT RT, such that the patients were slower to respond to hits than the normal control group (t=1.78, p=0.08). The relatives were also slower than the normal controls on the A-CPT RT, and this difference was also at trend-level (t=1.84, p=0.07) despite co-varying for age.

Discussion As expected, the patients had worse performance than the normal controls on most of the measures considered and the relatives demonstrated a difference (at the trend-level with IQ as a covariate, and at p<0.05 with level of education instead of IQ as a covariate) compared to the normal controls during the auditory CPT task, but showed normal levels in the visual modality of sustained attention and on selected attention. Slower reaction times during the A-CPT could be explained by a number of reasons. First, participants may be responding cautiously and may tend to focus on minimizing errors. Second, their slower reaction time may be due to decreased perceptual sensitivity causing these individuals to respond less frequently due to de®cits in detecting or processing sensory information. Evidence for perceptual de®cits has been reported in both people with schizophrenia and their healthy relatives, which suggests that this may be an endophenotype

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

119

for schizophrenia (Bredgaard, 2000; Cimmer et al., 2006). This is in line with reports that have illustrated greater decrements in the auditory modality versus the visual modality in schizophrenia. For example, using visual and auditory versions of the CPT, Mussgay and Hertwig (1990) also reported that information-processing de®cits are more pronounced in the auditory modality than the visual modality. Additionally, in an effort to address the inconsistencies of EEG ®ndings in the visual modality, but not in the auditory modality, Wood et al. compared ERPs in both modalities in schizophrenia (Wood et al., 2006). They reported that perceptual processing is impaired in the auditory while remaining intact in the visual modality in people with schizophrenia. Imaging studies have suggested that hypofrontality may underlie this auditory information-processing impairment in schizophrenia (Cohen, 1998). The absence of a greater difference between the relatives and normal controls may re¯ect the lack of sensitivity of our selected measures. In an attempt to ®nd suitable tasks for both the well relatives and the patients with schizophrenia, the relative simplicity of the present study's visual CPT paradigm could have been a factor in the absence of measurable impairment in the well relatives. Previous paradigms to measure attention impairment in the well relatives of patients with schizophrenia have involved more challenging tests of visual sustained attention that involved aspects of working memory or distractors such as the `AX'-task (where `X' preceded by an `A' is the target). It has also been reported that functioning for tests that have been robust discriminators among relatives of patients with schizophrenia may be limited to younger samples and that neuropsychological impairment may not be generalizable in relatives over 60 years of age (Faraone et al., 1996). However, similar to reports from other studies, we found that the relatives' performance on several attention measures was intermediate between patients and normal controls (Laurent, 1999; Mirsky, 1995). The current ®ndings did not support the hypothesis that selective attention is impaired in the ®rst-degree relatives of people with schizophrenia. Negative ®ndings of selective attention de®cits in high-risk populations have also been reported in literature (Asarnow, 1978; Asarnow et al., 1977; Schreiber, 1995). In a study by Asarnow (1978), selective attention in the offspring of mothers with schizophrenia was not found to be impaired using the Stroop colour±word task, but were clearly slower than normal controls on a dif®cult level of another task of selective attention called the Spokes test. The results of the present study also illustrate that attention was subject to the effects of intellectual ability. When IQ was replaced with level of education as a covariate, the difference in the A-CPT reaction time between the relatives and the normal controls reached signi®cance. Previous studies have shown that attention is directly associated with intelligence (Feinberg, 1991). For example, Necka (1996) found in a group of normal participants

120

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

that those with higher IQ had signi®cantly more attentional resources to distribute among competing tasks, and thus performed better on tasks of attention. To summarize, these ®ndings demonstrate that auditory sustained attention discriminates ®rst-degree relatives of people with schizophrenia. The absence of greater disparity in visual sustained and selective attention between the relatives and the normal controls may indicate that the measures selected for this study to ascertain attention dysfunction are not fully sensitive to the possible abnormalities that are associated with the pathogenesis of schizophrenia. However, we cannot exclude the possibility that this may be a function of our sampling methods. The recruitment of the families was mainly through a national organization, the National Schizophrenia Fellowship (NSF), whose membership can be characterized by high-functioning and highly involved relatives and patients. The ®ndings from the current sample, therefore, even though representative of a portion of families who have been affected with schizophrenia, may not be necessarily generalizable to the rest of the population. Nonetheless, despite the fact that they were high functioning, they still demonstrated impairments in auditory sustained attention.

STUDY 4: INTELLECTUAL ASYMMETRY Acknowledgement: The results of this study have been published previously in the British Journal of Psychiatry (Kravariti et al., 2006).

Introduction Schizophrenia is associated with IQ de®cits across the lifespan (Heinrichs & Zakzanis, 1998). Pre-schizophrenic individuals perform lower on tests of intellectual ability than would be predicted from family and environmental variables (Aylward et al., 1984), and affected probands show a de®cit of about ten IQ points compared to population norms (Heinrichs & Zakzanis, 1998; Winder, 1960). Once established, the abnormality remains relatively stable, pointing to a `static encephalopathy' rather than a degenerative process (Hoff et al., 2005; Russell et al., 1997). It is unclear whether the intellectual de®cit in schizophrenia is an epiphenomenon of a neurodevelopmental insult that causes the disorder, an independent, genetically transmitted trait that increases the risk for the disease or just one of the pleiotropic effects of the schizophrenia genotype. The uncertainty surrounding the origin of the abnormality is compounded by dif®culties in de®ning `de®cit' in this context: As a group, preschizophrenic and schizophrenic individuals still perform within the average range of ability, but their total IQ distribution is shifted to the left compared to that of an unselected population. Such a pattern indicates a

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

121

small effect operating throughout the population at risk, and suggests that `de®cit' should be de®ned in relative rather than absolute terms: Even when no intellectual abnormality is apparent by psychometric standards, individuals may still function at a lower level than they would if they did not have increased liability to schizophrenia. Although the effect of the pre-schizophrenic state on overall intellectual function is often dif®cult to detect in a given individual, an apparent discrepancy between a person's verbal and non-verbal intellectual skills is a more obvious pointer to potential (lateralized) brain dysfunction.

Methodology Study participants One-hundred-and-eight ®rst-degree relatives of schizophrenia probands, drawn from sixty-four families with one or more affected members, were included in this component of the analysis. The sociodemographic characteristics of the sample are presented in Table 5.13.

Genetic liability for schizophrenia Genetic liability (GL) for schizophrenia ± modelled using the continuous quantitative measure described in Chapter 2 ± was calculated based on normal distribution theory assuming a polygenic, multifactorial, liabilitythreshold model of schizophrenia, with initially imputed scores based on presence of illness adjusted to account for family size, age, affection status and genetic relatedness as far as second degree from the index patient. Higher scores on the GL scale re¯ect higher presumed genetic liability for schizophrenia.

General intelligence and intellectual asymmetry A short form of the Wechsler Intelligence Scale ± Revised (WAIS-R; Wechsler, 1981) comprising Vocabulary, Comprehension, Similarities, Block Design and Object Assembly was administered to all participants. Employing the formulae provided by Cavanan and Beckmann (1993), we estimated scores on two orthogonal factors, derived from a principal component analysis of the WAIS-R: General Ability IQ, equivalent to Wechsler's FullScale IQ, and Verbal±Spatial Contrast IQ, an index of asymmetry of brain function. Both indices have a population mean of 100 and an SD of 15. Verbal±Spatial Contrast IQ can produce markedly different characterizations of ability compared with Verbal IQ (VIQ) and Performance IQ (PIQ) (Cavanan, 1986): scores below 100 indicate asymmetry in favour of verbal skills, whereas scores above 100 indicate asymmetry in favour of spatial skills

122

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS TABLE 5.13 Sociodemographic characteristics for Study 4 (n=108)

Characteristic

Number of ®rst-degree relatives

Male, female Age in years: range, mean (SD) Professional/managerial, skilled/partly skilled, unskilled/unemployed1 Right-handed, left-handed English as ®rst language

39, 69 16±69, 48 (14) 37, 53, 18 97, 11 108

1

Parental socioeconomic status at birth, as determined using the Standard Occupational Classi®cation (Of®ce of Population Censuses and Surveys, 1991).

(Cavanan & Beckmann, 1993). Short versions of the Wechsler scales based on their factor structure have detected differences between left- and rightsided brain lesions even in small sample sizes, whereas the traditional Wechsler IQ scores have failed to do so in larger samples (Cavanan, 1986; Cavanan & Beckmann, 1993).

Statistical analysis The association of GL with Verbal±Spatial Contrast IQ was examined using multiple regression analysis, adjusting for age, gender, years of education, parental socioeconomic status and General Ability IQ, and incorporating multilevel modelling with a robust estimator for the variances of the regression co-ef®cient estimates in order to account for the non-independence of observations amongst related participants as before. General ability IQ was included as a co-variate in the analysis, because previous research suggested that larger verbal-performance splits are more common at the higher IQ levels (Iverson, 2001). The remaining covariates were selected on the basis of their signi®cant association with Verbal±Spatial Contrast IQ in exploratory simple regression analyses. The analysis was repeated: (1) after excluding participants with any lifetime psychiatric diagnosis; (2) after excluding those at the extremes of the age distribution (19 relatives younger than 20 years or older than 60 years); and (3) separately for men and women. These steps were taken because psychiatric disorders other than schizophrenia may exert independent effects on intellectual function, because the interpretation of the component scores requires some caution at the extremes of the age distribution (Cavanan, 1986) and because gender distribution was skewed in our sample.

Results The means and standard deviations for General Ability IQ and Verbal± Spatial Contrast IQ were 100.1 (14.6) and 95.6 (13.8), respectively. Verbal±

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

123

130 120 110

Verbal–Spatial contrast IQ

100 90 80 70 60 –.2

0.0

.2

.4

.6

.8

1.0

Genetic liability Figure 5.1 Scatterplot of genetic liability scores and Verbal±Spatial Contrast IQ (unadjusted values). (Reproduced with permission of the Royal College of Psychiatrists from Kravariti et al., 2006).

Spatial Contrast IQ was normally distributed (Kolmogorov±Smirnov=0.05, p=0.2). About 65% of the participants obtained Verbal±Spatial Contrast IQs below 100 (verbal skills superior to spatial skills), and the negative distance from the population mean reached or exceeded 1 SD in 23% of the sample. Genetic loading was signi®cantly negatively associated with Verbal±Spatial Contrast IQ (co-ef®cient=±17.95, 95% CI: ±27.84 to ±8.06, p=0.001), indicating that increases in GL co-occurred with increases in asymmetry of brain function, with a relative superiority of verbal to spatial skills. This pattern was consistent and remained signi®cant or nearsigni®cant after excluding participants with psychiatric diagnoses or at the extremes of the age distribution (p<0.05), and after performing the analyses separately for women (p<0.05) and men (p=0.06). Figure 5.1 presents a scatterplot of Verbal±Spatial Contrast IQ and GL in the total sample.

Discussion Our study suggests that a superiority of verbal- to perceptual-motor skills is an indicator of genetic risk for schizophrenia. This conclusion is in line with earlier evidence that the Performance Scale of the Wechsler tests elicits

124

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

more consistent de®cits in schizophrenia patients than the Verbal Scale. In a meta-analysis of thirty-®ve studies by Heinrichs and Zakzanis (1998), the overlaps for WAIS-R Verbal IQ and Performance IQ between schizophrenia patients and normal controls were 45% and 29%, respectively, indicating a more pronounced de®cit of the schizophrenia group in the Performance Scale. The suggested imbalance between Verbal IQ and Performance IQ is also present in children and adolescents with schizophrenia spectrum disorders (Kravariti et al., 2003; McCarthy et al., 2005) and seems to predate the onset of schizophrenia. In a study by Amminger et al. (2000), youths with later outcomes of schizophrenia-related psychoses were impaired in Performance IQ, but not Verbal IQ, compared to those with affective disorder or no psychiatric outcomes. Whereas the above ®ndings suggest that a Verbal IQ>Performance IQ pro®le is a lifelong characteristic of schizophrenia, the present study is the ®rst one to report a relationship between genetic vulnerability for schizophrenia and intellectual asymmetry. Our ®ndings indicate that a superiority of verbal intelligence to spatial intelligence is not just an epiphenomenon of a neurodevelopmental insult, but a putative endophenotype for the disorder. However, as with many cognitive variables measured in patients with schizophrenia (e.g. Full-Scale IQ), there was a `downward shift' (rather than a skewness or bimodality) of the distribution of Verbal±Spatial Contrast IQ in our participants. As a result, a substantial minority showed higher spatial relative to verbal abilities. Lower Verbal±Spatial Contrast IQ may therefore re¯ect the impact of susceptibility genes and help dissect the genetic heterogeneity of schizophrenia.

OVERALL SUMMARY OF THE NEUROPSYCHOLOGICAL FINDINGS AND CONCLUSIONS Four major domains of cognitive function were assessed as part of the neuropsychological arm of the MFSP: episodic memory, executive function, attention and intellectual asymmetry. A number of ®ndings emerged, with the main ones concerning the sample of relatives as follows: First, the relatives of the patients with schizophrenia from the non-familial sample show de®cits in immediate and delayed verbal recall, suggesting impairment of verbal episodic memory. Second, the relatives from the non-familial sample are inef®cient in planning a series of moves and in generating spontaneous strategies when solving problems, suggesting impairment in these aspects of executive function. Third, relatives in general appear to show impairments in sustained auditory attention, but have normal levels of visual sustained and selective attention. Fourth, the discrepancy between a relatively higher verbal intelligence in association with a comparatively

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

125

lower spatial intelligence is accentuated as genetic vulnerability for schizophrenia increases indicating an effect of genetic loading on this cognitive function.

Familial/non-familial distinction Whereas reduced intellectual asymmetry was more prominent in those relatives from more densely affected families ± as re¯ected in higher genetic loading scores ± the neuropsychological assessments of memory and executive function in the MFSP produced the intriguing and counterintuitive ®ndings of more substantial impairments (i.e. which reached statistical signi®cance when compared with controls) in non-familial relatives than familial relatives of patients for episodic memory, accuracy of problemsolving and strategy formation. This is precisely the opposite characteristic to what we predicted for an intermediate phenotype. This ®nding in no way supports a non-genetic or environmental contribution towards such dysfunction, as any abnormalities detected in unaffected relatives of patients could potentially re¯ect the action of susceptibility genes, regardless of the density of family history of illness. Possible interpretations of these ®ndings are: 1.

2.

3.

Familial relatives might still be neuropsychologically impaired but not in the memory or particular aspects of executive processing examined in the present study. If this is the case, then familial relatives might represent a distinct group that can be differentiated in terms of cognitive function from their non-familial counterparts. Sampling variation produced heterogeneity within the familial relatives group, since not all these families carry the neuropsychological endophenotype: some of the familial relatives were impaired, scoring more than 2 SDs below the mean of the control group. Presumed obligate carriers also showed some non-signi®cant impairments on some measures of memory or executive function, which may have reached statistical signi®cance in a larger sample size. It is possible that, due to our recruiting methods whereby only some family members were recruited to participate, the study was biased towards those familial relatives who were the most motivated and the least anxious and suspicious. It is possible that a more epidemiological sample of relatives, and indeed patients, would show more memory and executive function abnormalities. The apparent differences between the familial and non-familial relatives sample may be in that the former comprises more family members that have reached a certain threshold going beyond which the disease is clinically expressed, i.e. if these relatives also had the additional risk

126

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

genes associated with the neuropsychological impairment, they would cross threshold for illness and be categorized as affected. By contrast, relatives from the non-familial sample might have a higher threshold for developing illness or carry other protective factors that enable them to carry the neuropsychological endophenotype and not manifest the disorder. In conclusion, the ®ndings suggest that these cognitive dysfunctions represent endophenotypic traits that may act as risk indicators, which, in conjunction with other factors, can render an individual liable to develop schizophrenia. Utilizing intermediate phenotypes represented by verbal memory, planning ability, strategy formation or auditory sustained attention could potentially assist in identifying genes that are responsible for increasing susceptibility for schizophrenia or in clarifying the role of putative susceptibility genes in the pathophysiology of the illness. As not all relatives manifest the impairment, utilizing these neuropsychological impairments as intermediate phenotypes in gene identi®cation studies are likely to be informative in only a subset of families. Alternatively, examining the performance distributions in each cognitive process separately, and selecting those individuals that are at the extreme ends of the distribution, may be more informative for such studies. Equally, selecting healthy controls based on whether they perform exceptionally well or very poorly on these domains may also prove useful in identifying genes that are associated with these aspects of cognition and, possibly, schizophrenia. The neuropsychological component of the MFSP has shown a number of possible cognitive endophenotypes for schizophrenia, as have several other studies. It is perhaps time, however, to move beyond doing neuropsychological assessments in heterogenous samples of relatives of patients with schizophrenia and take what is already available a step further. By using all the information collected on the different aspects of brain function, statistical models can be developed that would allow for identi®cation of the optimum combination of the measures that will best predict genetic risk. Looking for genes in the relatives of those patients only that are associated with predicted increased genetic risk could be more useful than the alternative approach of indiscriminately including all relatives in genetic analyses.

REFERENCES Amminger, G.P., Edwards, J., Brewer, W.J., McGorry, P.D. (2000). Performance IQ and verbal IQ in psychotic disorders. Schizophrenia Research, 41, 276. Asarnow, R.F. (1978). Residual performance de®cit in clinically remitted schizophrenics: A marker of schizophrenia? Journal of Abnormal Psychology, 87, 597±608. Asarnow, R.F., Steffy, R.A., Maccrimmon, D.J., Cleghorn, J.M. (1977). Attentional

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

127

assessment of foster children at risk for schizophrenia. Journal of Abnormal Psychology, 86, 267±275. Asarnow, R.F., Nuechterlein, K.H., Subotnik, K.L., Fogelson, D.L., Torquato, R.D., Payne, D.L., et al. (2002). Neurocognitive impairments in nonpsychotic parents of children with schizophrenia and attention-de®cit/hyperactivity disorder ± The University of California, Los Angeles family study. Archives of General Psychiatry, 59, 1053±1060. Aylward, E., Walker, E., Bettes, B. (1984). Intelligence in schizophrenia ± meta-analysis of the research. Schizophrenia Bulletin, 10, 430±459. Badcock, J.C., Michie, P.T., Rock, D. (2005). Spatial working memory and planning ability: Contrasts between schizophrenia and bipolar I disorder. Cortex, 41, 753±763. Barch, D.M., Carter, C.S., Cohen, J.D. (2004). Factors in¯uencing Stroop performance in schizophrenia. Neuropsychology, 18, 477±484. Bartok, E., Berecz, R., Glaub, T., Degrell, I. (2005). Cognitive functions in prepsychotic patients. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 29, 621±625. Bredgaard, R. (2000). Information processing and attentional dysfunctions as vulnerability indicators in schizophrenic spectrum disorders. World Journal of Biological Psychiatry, 1, 5±15. Byrne, M., Clafferty, B.A., Cosway, R., Grant, E., Hodges, A., Whalley, H.C., et al. (2003). Neuropsychology, genetic liability, and psychotic symptoms in those at high risk of schizophrenia. Journal of Abnormal Psychology, 112, 38±48. Cannon, T.D., Zorrilla, L.E., Shtasel, D., Gur, R.E., Gur, R.C., Marco, E.J., et al. (1994). Neuropsychological functioning in siblings discordant for schizophrenia and healthyvolunteers. Archives of General Psychiatry, 51, 651±661. Cannon, T.D., Huttunen, M.O., Lonnqvist, J., Tuulio-Henriksson, A., Pirkola, T., Glahn, D., et al. (2000). The inheritance of neuropsychological dysfunction in twins discordant for schizophrenia. American Journal of Human Genetics, 67, 369±382. Cavanan, A., Beckmann, J. (1993). Deriving principal-component IQ scores from the WAIS-R. British Journal of Clinical Psychology, 32, 81±86. Cavanan, A., Dunn, G., McMillan, T.M. (1986). Principal components of the WAIS-R. British Journal of Clinical Psychology, 25, 81±86. Censits, D.M., Ragland, J.D., Gur, R.C., Gur, R.E. (1997). Neuropsychological evidence supporting a neurodevelopmental model of schizophrenia: A longitudinal study. Schizophrenia Research, 24, 289±298. Chen, W.J., Faraone, S.V. (2000). Sustained attention de®cits as markers of genetic susceptibility to schizophrenia. American Journal of Medical Genetics, 97, 52±57. Chen, W.J., Chang, C.H., Liu, S.K., Hwang, T.J., Hwu, H.G. (2004). Sustained attention de®cits in nonpsychotic relatives of schizophrenic patients: A recurrence risk ratio analysis. Biological Psychiatry, 55, 995±1000. Cimmer, C., Szendi, I., Csifcsak, G., Szekeres, G., Kovacs, Z.A., Somogyi, I., et al. (2006). Abnormal neurological signs, visual contrast sensitivity, and the de®cit syndrome of schizophrenia. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 30, 1225±1230. Cirillo, M.A., Seidman, L.J. (2003). Verbal declarative memory dysfunction in schizophrenia: From clinical assessment to genetics and brain mechanisms. Neuropsychology Review, 13, 43±77. Cohen, R.M. (1998). Abnormalities in the distributed network of sustained attention predict neuropleptic treatment response in schizophrenia. Neuropsychopharmacology, 19, 36±47. Daban, C., Amado, I., Bourdel, M.C., Loo, H., Olie, J.P., Poirier, M.F., et al. (2005). Cognitive dysfunctions in medicated and unmedicated patients with recent-onset schizophrenia. Journal of Psychiatric Research, 39, 391±398. David, A.S., Malmberg, A., Lewis, G., Brandt, L., Allebeck, P. (1995). Premorbid

128

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

neuropsychological performance as a risk factor for schizophrenia ± 13-year follow-up of 50,000 conscripts. Schizophrenia Research, 15, 114. Diller, L. (1974). Studies in Cognition and Rehabilitation in Hemiplegia. New York: New York University Medical Center, Institute of Rehabilitation Medicine. Faraone, S.V. (2000). Neuropsychologic functioning among the nonpsychotic relatives of schizophrenic patients: The effect of genetic loading. Biological Psychiatry, 48, 120±126. Faraone, S.V., Seidman, L.J., Lyons, M.J., Kremen, W.S., Pepple, J.R., Tsuang, M.T. (1995). Neuropsychological functioning among the nonpsychotic relatives of schizophrenic patients ± A diagnostic ef®ciency analysis. Journal of Abnormal Psychology, 104, 286±304. Faraone, S.V., Seidman, L.J., Kremen, W.S., Toomey, R., Lyons, M.J., Tsuang, M.T. (1996). Neuropsychological functioning among the elderly nonpsychotic relatives of schizophrenic patients. Schizophrenia Research, 21, 27±31. Faraone, S.V., Seidman, L.J., Kremen, W.S., Toomey, R., Pepple, J.R., Tsuang, M.T. (1999). Neuropsychological functioning among the nonpsychotic relatives of schizophrenic patients: A 4-year follow-up study. Journal of Abnormal Psychology, 108, 176±181. Faraone, S.V., Seidman, L.J., Kremen, W.S., Toomey, R., Pepple, J.R., Tsuang, M.T. (2000). Neuropsychologic funtioning among the nonpsychotic relatives of schizophrenic patients: The effect of genetic loading. Biological Psychiatry, 48, 120±126. Feinberg, J.R. (1991). WAIS-R intrasubtest scatter in a chronic schizophenic population: Is it an attention problem? Journal of Clinical Psychology, 47, 327±335. Fioravanti, M., Carlone, O., Vitale, B., Cinti, M.E., Clare, L. (2005). A meta-analysis of cognitive de®cits in adults with a diagnosis of schizophrenia. Neuropsychology Review, 15, 73±95. Gold, J.M., Harvey, P.D. (1993). Cognitive de®cits in schizophrenia. Psychiatric Clinics of North America, 16, 295±312. Goldberg, T.E., Torrey, E.F., Gold, J.M., Ragland, J.D., Bigelow, L.B., Weinberger, D.R. (1993). Learning and memory in monozygotic twins discordant for schizophrenia. Psychological Medicine, 23, 71±85. Goldberg, T.E., Torrey, E.F., Gold, J.M., Bigelow, L.B., Ragland, R.D., Taylor, E., et al. (1995). Genetic risk of neuropsychological impairment in schizophrenia ± A study of monozygotic twins discordant and concordant for the disorder. Schizophrenia Research, 17, 77±84. Goldstein, L.H. (1988). Verbal and abstract designs paired associate learning after unilateral and temporal lobectomy. Cortex, 24, 41±52. Gooding, D.C., Matts, C.W., Rollmann, E.A. (2006). Sustained attention de®cits in relation to psychometrically identi®ed schizotypy: Evaluating a potential endophenotypic marker. Schizophrenia Research, 82, 27±37. Gur, R.E., Mosley, P.D., Resnick, S.M., Levick, S., Erwin, R., Saykin, A.J., et al. (1991). Relations among clinical scales in schizophrenia. American Journal of Psychiatry, 148, 472±478. Heaton, R.K. (1981). Wisconsin Card Sorting Test (WCST). Odessa, Florida: Psychological Assessment Resources. Heinrichs, R.W., Zakzanis, K.K. (1998). Neurocognitive de®cit in schizophrenia: A quantitative review of the evidence. Neuropsychology, 12, 426±445. Hepp, H.H. (2007). The Stroop effect in schizophrenic patients. Schizophrenia Research, 22, 187±195. Hill, S.K., Beers, S.R., Kmiec, J.A., Keshavan, M.S., Sweeney, J.A. (2004). Impairment of verbal memory and learning in antipsychotic-naive patients with ®rst-episode schizophrenia. Schizophrenia Research, 68, 127±136. Hoff, A.L., Svetina, C., Shields, G., Stewart, J., Delisi, L.E. (2005). Ten year longitudinal study

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

129

of neuropsychological functioning subsequent to a ®rst episode of schizophrenia. Schizophrenia Research, 78, 27±34. Iverson, G. (2001). Base rates of WAIS-R VIQ-PIQ differences in 1593 psychiatric inpatients. Journal of Clinical Psychology, 57, 1579±1587. Johnstone, E.C., Ebmeier, K.P., Miller, P., Owens, D.G.C., Lawrie, S.M. (2005). Predicting schizophrenia: Findings from the Edinburgh high-risk study. British Journal of Psychiatry, 186, 18±25. Jones, P., Rodgers, B., Murray, R., Marmot, M. (1994). Child developmental risk factors for adult schizophrenia in the British 1946 birth cohort. Lancet, 344, 1398±1402. Joyce, E., Hutton, S., Mutsatsa, S., Gibbins, H., Webb, E., Paul, S., et al. (2002). Executive dysfunction in ®rst-episode schizophrenia and relationship to duration of untreated psychosis: The West London Study. British Journal of Psychiatry, 181, S38±S44. Kravariti, E., Morris, R.G., Rabe-Hesketh, S., Murray, R.M., Frangou, S. (2003). The Maudsley Early-Onset Schizophrenia Study: Cognitive function in adolescent-onset schizophrenia. Schizophrenia Research, 65, 95±103. Kravariti, E., Toulopoulou, T., Mapua-Filbey, F., Schulze, K., Walshe, M., Sham, P., et al. (2006). Intellectual asymmetry and genetic liability in ®rst-degree relatives of probands with schizophrenia. British Journal of Psychiatry, 188, 186±187. Kremen, W.S., Seidman, L.J., Pepple, J.R., Lyons, M.J., Tsuang, M.T., Faraone, S.V. (1994). Neuropsychological risk indicators for schizophrenia ± A review of family studies. Schizophrenia Bulletin, 20, 103±119. Laurent, A. (1999). Attentional de®cits in patients with schizophrenia and in their nonpsychotic ®rst-degree relatives. Psychiatry Research, 89, 147±159. McCarthy, J., Kraseski, K., Schvartz, I., Mercado, V., Daisy, N., Tobing, L., et al. (2005). Sustained attention, visual processing speed, and IQ in children and adolescents with schizophrenia spectrum disorder and psychosis not otherwise speci®ed. Perceptual and Motor Skills, 100, 1097±1106. McDonald, C., Grech, A., Toulopoulou, T., et al. (2002). Brain volumes in familial and nonfamilial schizophrenic probands and their unaffected relatives. American Journal of Medical Genetics, 114, 616±625. Mirsky, A.F. (1992). On familial factors in the attention de®cit in schizophrenia: A review and report of two new subject samples. Journal of Psychiatric Research, 26, 383±403. Mirsky, A.F. (1995). Analysis of the attention de®cit in schizophrenia: A study of patients and their relatives in Ireland. Journal of Psychiatric Research, 29, 23±42. Morris, R.G., Rushe, T., Woodruffe, P.W.R., Murray, R.M. (1995). Problem-solving in schizophrenia ± A speci®c de®cit in planning ability. Schizophrenia Research, 14, 235±246. Mussgay, L., Hertwig, R. (1990). Signal-detection indexes in schizophrenics on a visual, auditory, and bimodal continuous performance test. Schizophrenia Research, 3, 303±310. Necka, E. (1996). Attentional resources, working memory capacity, and intelligence: The mediating role of arousal. International Journal of Psychology, 31, 2471. Nieuwenstein, M.R., Aleman, A., de Haan, E.H.F. (2001). Relationship between symptom dimensions and neurocognitive functioning in schizophrenia: A meta-analysis of WCST and CPT studies. Journal of Psychiatric Research, 35, 119±125. Nuyen, J., Sitskoorn, M.M., Cahn, W., Kahn, R.S. (2005). Verbal memory in ®rst-episode schizophrenia: Heterogeneity in performance? Journal of the International Neuropsychological Society, 11, 152±162. Of®ce of Population Censuses and Surveys (OPCS) (1991). Standard Occupational Classi®cation, Volume 3. London: HMSO. Orzack, M.H., Kornetsk, C. (1966). Attention dysfunction in chronic schizophrenia. Archives of General Psychiatry, 14, 323. Pantelis, C., Barber, F.Z., Barnes, T.R.E., Nelson, H.E., Owen, A.M., Robbins, T.W. (1999).

130

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Comparison of set-shifting ability in patients with chronic schizophrenia and frontal lobe damage. Schizophrenia Research, 37, 251±270. Park, S., Holzman, P.S., Goldmanrakic, P.S. (1995). Spatial working-memory de®cits in the relatives of schizophrenic patients. Archives of General Psychiatry, 52, 821±828. Reichenberg, A., Weiser, M., Rapp, M.A., Rabinowitz, J., Caspi, A., Schmeidler, J., et al. (2006). Premorbid intra-individual variability in intellectual performance and risk for schizophrenia: A population-based study. Schizophrenia Research, 85, 49±57. Reitan, R.M. (1958) Validity of the trail-making test as an indicator of organic brain damage. Perceptual and Motor Skills, 8, 271±276. Robbins, T.W. (1998). Dissociating executive functions of the prefrontal cortex. In: The prefrontal cortex. New York: Oxford University Press. Roy, M.A., Crowe, R.R. (1994). Validity of the familial and sporadic subtypes of schizophrenia. American Journal of Psychiatry, 151, 805±814. Rund, B.R. (1998). A review of longitudinal studies of cognitive functions in schizophrenia patients. Schizophrenia Bulletin, 24, 425±435. Rushe, T.M., Morris, R.G., Miotto, E.C., Feigenbaum, J.D., Woodruff, P.W.R., Murray, R.M. (1999). Problem-solving and spatial working memory in patients with schizophrenia and with focal frontal and temporal lobe lesions. Schizophrenia Research, 37, 21±33. Russell, A.J., Munro, J.C., Jones, P.B., Hemsley, D.R., Murray, R.M. (1997). Schizophrenia and the myth of intellectual decline. American Journal of Psychiatry, 154, 635±639. Russell, E.W. (1975). A multiple scoring method for the assessment of complex memory functions. Journal of Consulting and Clinical Psychology, 43, 800±809. Saykin, A.J., Gur, R.C., Gur, R.E., Mosley, P.D., Mozley, L.H., Resnick, S.M., et al. (1991). Neuropsychological function in schizophrenia ± Selective impairment in memory and learning. Archives of General Psychiatry, 48, 618±624. Saykin, A.J., Shtasel, D.L., Gur, R.E., Kester, D.B., Mozley, L.H., Sta®niak, P., et al. (1994). Neuropsychological de®cits in neuroleptic naive patients with ®rst-episode schizophrenia. Archives of General Psychiatry, 51, 124±131. Schreiber, H. (1995). Comparative assessment of saccadic eye movements, psychomotor and cognitive performance in schizophrenics, their ®rst-degree relatives and control subjects. Acta Psychiatrica Scandinavica, 91, 195±201. Sharma, T., DuBoulay, G., Lewis, S., Sigmundsson, T., Gurling, H., Murray, R. (1997). The Maudsley family study. 1. Structural brain changes on magnetic resonance imaging in familial schizophrenia. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 21, 1297±1315. Sitskoom, M.M., Aleman, A., Ebisch, S.J.H., Appels, M.C.M., Kahn, R.S. (2004). Cognitive de®cits in relatives of patients with schizophrenia: A meta-analysis. Schizophrenia Research, 71, 285±295. Snitz, B.E., MacDonald, A.W., Carter, C.S. (2006). Cognitive de®cits in unaffected ®rst-degree relatives of schizophrenia patients: A meta-analytic review of putative endophenotypes. Schizophrenia Bulletin, 32, 179±194. Toulopoulou, T., Mapua-Filbey, F., Quraishi, S., Kravariti, E., Morris, R.G., McDonald, C., et al. (2005). Cognitive performance in presumed obligate carriers for psychosis. British Journal of Psychiatry, 187, 284±285. Toulopoulou, T., Morris, R.G., Rabe-Hesketh, S., Murray, R.M. (2003b). Selectivity of verbal memory de®cit in schizophrenic patients and their relatives. American Journal of Medical Genetics Part B-Neuropsychiatric Genetics 116B, 1±7. Toulopoulou, T., Rabe-Hesketh, S., King, H., Murray, R.M., Morris, R.G. (2003a). Episodic memory in schizophrenic patients and their relatives. Schizophrenia Research, 63, 261±271. Trenerry, M.R., Crosson, B., DeBoe, J., Leber, W.R. (1989). Stroop ± neuropsychological screening test. Odessa, Florida: Psychological Assessment Resources.

5. NEUROPSYCHOLOGICAL IMPAIRMENTS

131

Tulving, E. (1989). Memory: Performance, knowledge, and experience. The European Journal of Cognitive Psychology, 1, 3±26. Tulving, E. (2002). Episodic memory: From mind to brain. Annual Review of Psychololgy, 53, 1±25. Tyson, P.J., Laws, K.R., Roberts, K.H., Mortimer, A.M. (2005). A longitudinal analysis of memory in patients with schizophrenia. Journal of Clinical and Experimental Neuropsychology, 27, 718±734. Walker, E., Shaye, J. (1982). Familial schizophrenia ± A predictor of neuromotor and attentional abnormalities in schizophrenia. Archives of General Psychiatry, 39, 1153±1156. Wechsler, D. (1945). Manual for the Wechsler Memory Scale. New York: Psychological Corporation. Wechsler, D. (1981). Wechsler Adult Intelligence Scale ± Revised Manual. New York: Psychological Corporation. Whyte, M.C., McIntosh, A.M., Johnstone, E.C., Lawrie, S.M. (2005). Declarative memory in unaffected adult relatives of patients with schizophrenia: A systematic review and metaanalysis. Schizophrenia Research, 78, 13±26. Winder, C. (1960). Some psychological studies of schizophrenics. In D. Jackson (ed.). The Etiology of Schizophrenia. New York: Basic Books. Wood, S.M., Potts, G.F., Hall, J.F., Ulanday, J.B., Netsiri, C. (2006). Event-related potentials to auditory and visual selective attention in schizophrenia. International Journal of Psychophysiology, 60, 67±75.

CHAPTER SIX

Neurological abnormalities in patients with schizophrenia from singly- and multiply-affected families and their relatives Paola Dazzan, Timothy D. Grif®ths

INTRODUCTION Neurological soft signs (NSS) include abnormalities in sensory and motor performance that are indicative of non-speci®c cerebral dysfunction (Dazzan & Murray, 2002; Grif®ths et al., 1998; Shibre et al., 2002). Early neurological studies of patients with schizophrenia showed that neurological signs could be demonstrated in the absence of a recognized underlying neurological diagnosis (Quitkin et al., 1976; Tucker et al., 1974). These studies examined a mixture of localizing neurological signs such as the plantar response, as well as tests of motor co-ordination and sensory integration. Signs demonstrated in these studies were designated `soft signs' because of the absence of known accompanying focal structural pathology. In later studies, the term `soft sign' has been used more restrictively for signs that do not suggest primary tract or nucleus pathology, as opposed to `hard signs', which do (Woods et al., 1986). The variation in the signs examined in different studies, and the use of the terms `hard' and `soft' to describe different combinations of signs, makes evaluation of previous work extremely dif®cult. Moreover, the term `soft sign' implies that the signs sought are not acceptable in classical neurology and cannot be de®ned with any rigour; this is not the case. In fact, there is now agreement that these signs are present in excess in patients with schizophrenia at the time of the ®rst psychotic episode (Dazzan & Murray, 2002). They are thought to be indicative of non-speci®c cerebral dysfunction (Dazzan & Murray, 2002; 133

134

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Grif®ths et al., 1998; Shibre et al., 2002). Indeed, their presence in psychosis has been associated, even in ®rst-episode patients, with reductions of brain areas implicated in the pathogenesis of the disorder, like the association cortex and the basal ganglia (Dazzan et al., 2004; Keshavan et al., 2003). Studies conducted on relatives of patients with schizophrenia suggest that neurological signs in schizophrenia are due, at least in part, to a genetic vulnerability to this disorder. In fact, an excess of these signs has also been reported in the relatives and in the offspring of patients with schizophrenia (Schubert & McNeil, 2004). Signs that have been reported in relatives vary and include problems in motor co-ordination, motor sequencing and sensory integration, with differences possibly related to variability in methodology. Interestingly, those relatives who are presumed `carriers' of the genetic risk seem to have higher signs scores than presumed non-carriers (Gourion et al., 2004). Even stronger support for a genetic aetiology of neurological dysfunction in schizophrenia comes from twin studies, reporting that nonpsychotic twins from pairs discordant for schizophrenia have higher neurological signs scores than healthy control twins (Kelly et al., 2004; Picchioni et al., 2006). This evidence, together with evidence that neurological abnormalities are present from early childhood and have trait characteristics, makes them an interesting endophenotype for schizophrenia. As described in Chapter 1, one mechanism of varying likely genetic loading in families affected with schizophrenia is to divide participants with schizophrenia into populations with a presumed high genetic risk by virtue of having other close relatives affected (familial schizophrenia) and those with a low genetic risk by virtue of having no known relatives affected (nonfamilial schizophrenia) (Murray et al., 1985). The purpose of such a division is to produce populations which are enriched or depleted in terms of genetic risk. Patients with non-familial schizophrenia may also be more likely to portray neurobiological abnormalities associated with prenatal environmental factors (Murray & Jones, 1996). These differing risk factors may be re¯ected in phenotypic differences between non-familial and familial participants. A survey of studies comparing familial and non-familial schizophrenia suggests that there are neurological differences between populations de®ned in this way (Roy & Crowe, 1994). In the Maudsley Family Study of Psychosis, we carried out a systematic neurological examination of patients with schizophrenia and their unaffected relatives of presumed differing genetic risk using a schedule in which signs were divided into `primary' and `integrative'. Primary signs are elicited by a standard neurological examination and may re¯ect focal abnormality, whereas integrative signs are likely to depend on integration of, or between, motor and sensory processing, and therefore re¯ect more diffuse brain processing. The evaluation of integrative signs was based on the most rigorously de®ned and validated neurological instrument that has been

6. NEUROLOGICAL ABNORMALITIES

135

previously used in schizophrenia (Buchanan & Heinrichs, 1989). This neurological assessment was used to investigate the pattern of neurological abnormality in familial and non-familial patients with schizophrenia, as a test of the hypothesis that the phenotype is different in these two populations. The same neurological features were examined in the unaffected ®rstdegree relatives of the familial and non-familial patients with schizophrenia, to test the hypothesis that such neurological abnormalities re¯ect genetic liability. We considered that neurological abnormalities over-represented in unaffected relatives from multiply affected families would be most likely to re¯ect endophenotypic measures of schizophrenia.

METHODS The study population comprised 214 participants divided into ®ve groups: (1) participants with schizophrenia from multiply-affected families; (2) their unaffected ®rst-degree relatives; (3) participants with schizophrenia without a family history of psychotic illness in their ®rst- or second-degree relatives (non-familials); (4) unaffected ®rst-degree relatives of non-familial patients with schizophrenia, and (5) healthy volunteers with no personal or family history of schizophrenia. Three participants included in the study had a lifetime history of fewer than three generalized seizures (one participant each in the two relative groups and the familial schizophrenic group).

Neurological assessment Neurological assessments were carried out on the 214 participants by a single neurological assessor (T.D.G.) over a period of 2 years. The assessor was blind to participant diagnosis and family status. A structured schedule was used in a stereotyped order (the measures are listed in Table 6.1). A scoring instrument was used to assign a value of 0, 1 or 2 to each of the measures in the schedule, where 0 represents no abnormality, 2 represents a score at or above a reference criterion regarded as clearly abnormal, and 1 represents an intermediate criterion. The schedule was divided into tests of primary neurological dysfunction and tests of integrative neurological dysfunction. `Primary neurological dysfunction' means dysfunction identi®ed by a standard neurological examination, including cranial nerve and eye movement examinations, lateralizing limb pyramidal signs and frontal release signs. `Integrative neurological dysfunction' means abnormality of function that is likely to depend on integration within the motor and sensory systems or between the motor and sensory systems. Integrative dysfunction was separated in this way to de®ne a group of functions that may depend on distributed processing, and is not used as a synonym for `executive' or presumed frontal lobe dysfunction. The term `soft sign' is not used in this

136

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS TABLE 6.1 Neurological measures used

Primary function

Integrative function

Cranial nerve palsy R Cranial nerve palsy L Smooth pursuit Saccade to target Saccade to command Synkinesis Convergence Tone increase R limb Tone increase L limb Limb hypere¯exia R Limb hypere¯exia L Plantar R Plantar L Romberg Chorea R Chorea L Tremor R Tremor L Mirror movements R Mirror movements L Glabellar re¯ex Snout re¯ex Grasp re¯ex R Grasp re¯ex L Suck re¯ex

Audiovisual integration Stereognosis R Stereognosis L Graphaesthesia R Graphaesthesia L Extinction R/L confusion Tandem walk Rapid alternate movements R Rapid alternate movements L Finger±thumb opposition R Finger±thumb opposition L Finger±thumb opposition R Finger±nose R Finger±nose L Fist ring R Fist ring L Fist edge±palm R Fist edge±palm L Oszeretski

Measures were all assessed using an instrument that divided measures into 0 (completely normal), 1 or 2 (markedly abnormal). For the integrative measures, the same rating scale was applied as in Buchanan and Heinrichs (1989) to allow comparison. This instrument was extended in the present study to give ratings of 0±2 for primary measures not included in the previous instrument. The main analysis in this paper used all measures. The measures shown in bold type were used in a further analysis, which used a more restrictive de®nition of primary and integrative function (see text). L, left; R, right.

chapter, as it implies that the signs are de®ned without rigour and are unacceptable in classical neurology. However, the primary measures used broadly correspond to `hard signs' used in previous studies (e.g. Woods et al., 1991) and the integrative signs include measures previously de®ned as `soft signs'. The same scoring measures as in the instrument of Buchanan and Heinrichs were used for common measures, to allow comparison with previous studies using that rigorously de®ned instrument. However, that instrument does not include the primary assessment used here. Previous neurological studies of schizophrenia have included a smaller number of either primary or integrative signs or a mixture of the two. In the

6. NEUROLOGICAL ABNORMALITIES

137

present study, an attempt was made to classify all signs as either primary or integrative, rather than excluding large numbers of signs. There were, however, some signs that were dif®cult to classify as either primary or integrative. For example, the ®st±ring, ®st-edge±palm test and Oszeretski test involve complex motor sequencing and the use of proprioceptive and visual feedback and therefore ful®ll the de®nition of integrative sign given above. However, these tasks are impaired in damage to the frontal lobes and might also be regarded as focal signs. Similarly, many of the signs involving sensory integration have been localized to the parietal cortices. In the cases of both the frontal-sequencing signs and signs involving sensory integration, the likely occurrence of convergent processing in one particular area of association cortex was not seen to be a reason to exclude these signs from the integrative schedule. Classical `frontal release' signs, on the other hand (snout, grasp and suck), were classi®ed as primary signs on the basis of being phenomena dependent on the release of motor routines, likely to be localized, and not dependent on sensorimotor integration. Other signs that were dif®cult to classify were the tandem walk and ®nger±nose tests, which were included in the integrative part of the schedule. Both may be abnormal in focal cerebellar damage, but can also be affected by conditions in which coordination between sensory input and motor output is affected, as in deafferentation due to certain ganglionopathies. Additional, separate, measures were carried out to assess the extrapyramidal side-effects of drugs. The Abnormal Involuntary Movement in Schizophrenia (AIMS) scale (National Institute of Mental Health, 1976) and the Targeting of Abnormal Kinetic Effects (TAKE) scales (Wojcik et al., 1980) were used as measures of side-effects of antipsychotic drugs. A measure of rating consistency within and between raters was obtained by making video tapes of seventeen of the examinations. Within-rater agreement rates for the individual measures varied from 0.85 to 1.0 and between-rater agreement rates varied from 0.64 to 1.0.

Analysis The frequency of individual abnormalities was compared across the groups, and scores for the primary, integrative and total abnormalities were calculated by adding the individual scores for the measures in Table 6.1. The distribution of the primary, integrative and total neurological scores was assessed overall and for the individual participant groups, and the effects of the variables age, sex, social class, handedness, pre-morbid IQ and drugs was assessed. The distribution of both primary and integrative abnormalities was found to be highly skewed. Logarithmic and other transformations of the data did not yield a normal distribution of the data. Non-parametric

138

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

analysis was therefore employed to compare the frequencies of neurological abnormalities between the groups. The overall effect of group without correction for confounding variables was ®rst assessed using the Kruskal±Wallis test. A detailed analysis was then carried out to assess the differences between the normal control group and each of the test groups using logistic regression to control for confounding variables. The primary and integrative scores for each participant were ®rst dichotomized into normal and abnormal using broad and narrow criteria for neurological abnormality. Narrow criteria represent a conservative estimate of abnormality that is met by a small proportion of the normal group (~2% for the primary and integrative scores), whereas broad criteria are based on a less conservative threshold for abnormality. The different criteria were used to investigate the effect of threshold criteria on the demonstration of differences between the groups. Using these criteria, the frequency of neurological abnormalities was compared between the groups using logistic regression to control for the confounding variables age and social class. Separate regression analyses were carried out for the comparison of each of the participant groups with the control groups, to allow calculation of an odds ratio, and a probability value for the odds ratio being different from 1. The odds ratio represents a measure of the estimated probability of participants with neurological abnormality being in the relevant non-control group in relation to those with no neurological abnormality. This is also expressed in the results as a signi®cance level for neurological abnormality for each comparison carried out. Analysis was also carried out based on a more exclusive de®nition of primary and integrative function using the measures shown in bold type in Table 6.1. Extrapyramidal signs, `frontal release' signs and cerebellar signs were excluded from this analysis. The analysis was also extended to look for a group by sex effect. The purpose of this was to look for variation in differences between the groups with respect to sex. Theories of neurodevelopmental schizophrenia suggest an increased susceptibility in male nonfamilial cases, and a review of structural brain studies using CT suggests that males show the most marked differences between the non-familial and familial groups (Murray & Jones, 1996).

RESULTS Group demographics and group matching Table 6.2 shows the demographic characteristics for the groups. There were no signi®cant differences in sex, years in education, social class or handedness between the groups. The signi®cant difference in age between the groups is due to the relatives of the patients being older, these groups

59.0 11.6 (3.9) 80.6 95.4 71.9 20.7 (4.5) 395 (0±1600)

32 37 (4.5) 42.9 13.1 (3.0) 90.5 103.6 (14.1) 85.7

63 42 (16.5)

Familial schizophrenia relative

57.1 12.7 92.8 103.0 75.0 20.3 400.0 (4.51) (150±3000)

(21.4)

(2.4)

28 32 (8.3)

Sporadic schizophrenia

40.9 12.5 (3.0) 93.2 114.3 (15.8) 79.5

44 53 (12.5)

Sporadic schizophrenia relative

44.7 13.4 (3.1) 90.9 109.8 (9.5) 63.8

47 33.0 (13.5)

Control

0.372 0.091 0.242 0.0011 0.082 0.793 0.264

<0.00011

p value

(i) values are expressed as mean and standard deviation if these are normally distributed and range for non-normally distributed variables; (ii) p value is based on statistical tests appropriate to the nature of variables; (iii) social class is expressed as the proportion of patients in classes 1±3 compared with the total in classes 1±6 (for the schizophrenia patients the social class of the parents was used); (iv) pre-morbid IQ is the estimated global IQ based on the National Adult Reading Test (Nelson, 1982); (v) handedness is the percentage of participants who are strongly right handed [scoring 12 right on the 12-point classi®cation by Annett (1970) assessed by T.D.G. before the neurological assessment)]; (vi) the last row gives the mean drug dose in chlorpromazine equivalents [calculated as in Baldessarini et al. (1984) and Davis (1976)]. L, left; R, right. 1 analysis of variance; 2X 2 test; 3t test; 4Mann±Whitney U exact two-tailed test.

Sex (% male) Years of education Social class 1±3 (%) Premorbid IQ Handedness (% strong R) Age at onset of symptoms Drugs

Number Age

Familial schizophrenia

TABLE 6.2 Group demographics

140

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

containing a high proportion of parents of patients with schizophrenia. Age was found to be a signi®cant confounder for the integrative score and was taken into account in subsequent analysis. Pre-morbid IQ was also signi®cantly different between the groups, but was speci®cally not controlled for in the group comparison in view of the evidence that the pre-morbid IQ of patients with schizophrenia is lower than that of the general population (Watt & Lubensky, 1976).

Prevalence and distribution of individual neurological abnormalities Table 6.3 shows the individual measures in the groups. Among the primary measures, broken smooth pursuit was increased in the schizophrenic groups. Chorea was increased in both schizophrenia groups and also in the group of relatives from families multiply affected with schizophrenia. Postural limb tremor was increased in both groups with schizophrenia. Amongst the classical frontal release signs, glabellar tap was increased in both groups with schizophrenia, and snout in both groups with schizophrenia and in relatives of patients with schizophrenia, although no participants demonstrated a grasp re¯ex, even with distraction. Amongst the integrative scores, abnormalities of motor sequencing (®st±ring test, ®st-edge±palm test and Oszeretski) were increased in both groups of patients with schizophrenia and in relatives of patients with schizophrenia compared with the control group, as was the case for audiovisual integration, graphaesthesia and extinction. Right±left confusion was relatively common, even in the normal participants. As our aim was to look at the differences in overall primary and integrative scores between the groups, signi®cance levels were not calculated for the distribution of individual measures. The distributions of the individual scores and of the composite primary and integrative scores were highly skewed: further analysis was carried out using non-parametric methods rather than analysis of variance.

Effects of age, sex, social class, handedness and drugs on the neurological scores Age had a signi®cant effect on the neurological measures; a linear relationship between age and both integrative and total number of abnormal signs was found. Social class also had a signi®cant effect on the primary and integrative measures. These confounders were taken into account in the group comparisons. Sex and handedness did not have a signi®cant effect on the neurological measures. Drug dose was not a signi®cant confounder for the primary, integrative or total neurological scores in the groups with schizophrenia.

Primary measures1 Cranial palsy R Cranial palsy L Smooth pursuit Saccade target Saccade command Synkinesis Convergence Tone increase R Tone increase L Hypere¯exia R Hypere¯exia L Plantar R Plantar L Romberg Chorea R Chorea L Tremor R Tremor L Mirror movements R Mirror movements L Glabellar tap Snout Grasp R Grasp L Suck AIMS TAKE

Group

3 3 3 3 3 3 3 6 6 6 3 0 0 3 12 0 0 3 27 33

6 3 24 0 0 0

Familial schizophrenia

0 0 9 3

3 5 9 0 0 0 0 0 0 0 0 0 0 0 6 6 2 2 2 3 0 6

Familial schizophrenia relative 0 0 25 0 0 11 7 7 7 0 0 4 4 0 11 7 14 18 7 11 4 4 0 0 0 15 30

Sporadic schizophrenia

5 7 7 0 0 2 0 0 0 0 0 0 0 2 0 0 2 2 9 9 0 2 0 0 0 7 2

Sporadic relative

TABLE 6.3 Group scores for individual measures

0 0 2 2 0 0 0 2 2 0 0 6 4 0 2 0 0 0 12 0

0 0 0 0 0 0

Control

continues overleaf

27 0 3 16 12 12 30 9 0 0 9 9 3 3 24 24 33 30 27

Familial schizophrenia

6 0 0 8 9 2 28 0 0 0 15 15 0 0 14 12 35 35 17

Familial schizophrenia relative 17 0 0 14 14 4 14 0 0 4 0 0 4 4 14 14 21 21 18

Sporadic schizophrenia

8 0 0 9 7 9 18 2 0 0 0 0 0 0 14 14 30 30 11

Sporadic relative

0 2 2 0 2 0 15 0 0 0 0 2 0 0 2 2 9 11 2

Control

L, left; R, right. 1 Figures represent percentage of participants in each group scoring other than zero. AIMS and TAKE are composite extrapyramidal rating scores that do not contribute to the primary score. 2Figures represent percentage of participants in each group scoring other than zero.

Integrative measures2 Audiovisual integration Stereognosis R Stereognosis L Graphaesthesia R Graphaesthesia L Extinction R/L confusion Tandem walk Rapid alternate movements R Rapid alternate movements L Finger±thumb opposition R Finger±thumb opposition L Finger±nose R Finger±nose L Fist±ring R Fist±ring L Fist edge±palm R Fist edge±palm L Oszeretski

Group

TABLE 6.3 continued

6. NEUROLOGICAL ABNORMALITIES

143

Group comparison: primary and integrative measures Figure 6.1 shows the distribution of individual scores in each group and demonstrates the skewed distribution and differences in the proportion of abnormal participants between the groups. Non-parametric analysis using the Kruskal±Wallis test showed a signi®cant effect of group on both the primary (p<0.01) and the integrative (p<0.01) score. This preliminary analysis does not take into account the effect of confounding variables between the groups; further testing was carried out to examine differences in the expression of abnormal signs between the test groups and the normal control group using logistic regression to control for confounding variables. Table 6.4 shows the distribution of primary and integrative abnormalities between the groups expressed as the proportion of each group ful®lling broad and narrow de®nitions of abnormality. Odds ratios are also shown for the comparison between the proportion of abnormal participants in each group with the normal control group. A narrow de®nition of abnormality, which de®ned 2.1% of the control population as abnormal (primary score of 3), was used for the primary abnormalities. Using this de®nition of abnormality, a signi®cant proportion of abnormal participants was demonstrated only in the non-familial schizophrenia group (21%). A broad de®nition of primary abnormality was also used; this de®ned 15% of the normal control group as abnormal (primary score of 1). Using the broader de®nition of abnormality, a signi®cant proportion of participants in both the non-familial and familial schizophrenia groups were de®ned as abnormal (57 and 39%, respectively). For the integrative abnormalities, a narrow de®nition of abnormality de®ned 2.1% of the normal group as abnormal (integrative score of 3). Using this narrow de®nition, both groups of schizophrenia patients and both relative groups showed a signi®cant increase in the proportion of abnormal participants (within the range 25±50%). A broad de®nition of abnormality de®ned 23% of the control group as abnormal (integrative score 2). Using this broad de®nition of abnormality, the familial schizophrenia group and the group of familial schizophrenia relatives showed signi®cant increases in the proportion of abnormal participants (67 and 47%, respectively). The proportion of abnormal participants in the nonfamilial schizophrenia and non-familial relative group was not signi®cantly abnormal using the broad criterion. Analysis using the more restricted de®nition of primary and integrative abnormality based on the measures shown in bold type in Table 6.1 yielded the same result. Logistic regression was carried out based on the same confounding variables. Using a de®nition of abnormality de®ning 2% of the control population as abnormal, primary neurological scores were signi®cantly different from the control

Group

l l l ic ic ro ilia ia ilia ia ad nia ont ad nia r r m m n n o o e fa re C Fa hre sp re Sp phr p e oph ive oph t v o o i a z z at iz el iz hi hi el ch R sch sc sc R s

0

1

2

3

4

5

6

7

8

9 (a)

Group

l l l ic ic ro ilia ia ilia ia ad nia ont ad nia r r m m n n o o e fa re C Fa hre sp re Sp phr p e oph ive oph t v o o i a z z at iz el iz hi hi el ch R sch sc sc R s

20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

Figure 6.1 Distributions of (a) primary and (b) integrative abnormalities between groups

Primary score

10

Integrative score

(b)

3.3** 1.3±8.1 38.1 24.5*** 3.1±192

6.60*** 2.4±18.3 41.7 35.7*** 4.2±304

1.7 0.6±8.8

3.3* 1.1±9.8 49.2

0.5 0.04±7.0 28.6

3.8 0.3±42.4 38.9

61.1

3.1

Familial schizophrenia relative

8.3

Familial schizophrenia

19.0*** 2.2±174

2.5 0.9±7.2 25.9

37.0

7.3*** 2.4±22.1

11.2** 1.2±99.4 57.2

21.0

Sporadic schizophrenia

11.5*** 1.3±99

3.1 1.0±9.9 36.4

50.0

1.16 0.28±4.9

0.8 0.05±13.8 22.7

4.5

Sporadic schizophrenia relative

2.1

23.4

14.9

2.1

Control

Odds ratios are for the comparison of the proportion of abnormal participants in group compared with normal control group and are adjusted for the confounding variables age and social class. *P , 0.05; **P , 0.01; ***P , 0.005.

Integrative score % ful®lling broad criterion abnormality (score >1) Odds ratio for broad criterion 95% CI for odds ratio % ful®lling narrow criterion abnormality (score >2) Odds ratio for narrow criterion 95% CI for odds ratio

Primary score % ful®lling narrow criterion abnormality (score >2) Odds ratio for narrow criterion 95% CI for odds ratio % ful®lling broad criterion abnormality (score >0) Odds ratio for broad criterion 95% CI for odds ratio

Group

TABLE 6.4 Group primary and integrative measures

146

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

group in the non-familial schizophrenia group only (odds ratio 13.8, CI 1.6±122). Using a de®nition of abnormality de®ning ~10% of the control population as abnormal, integrative neurological scores were only signi®cantly different from the control group in the familial schizophrenia group (odds ratio 12.1, CI 1.6±23) and the familial schizophrenia relative group (odds ratio 3.0, CI 1.0±9.3). Further analysis for the presence of group by sex interaction did not show a signi®cant effect on the primary or integrative scores.

DISCUSSION Frequency of individual abnormalities A striking feature of this study compared with previous studies of neurological abnormality in schizophrenia is the overall low frequency of abnormal individual neurological signs in all groups. For example, fewer than 4% of participants in any group were found to have abnormality of ®nger±nose testing compared with values of greater than 18% in one study (Buchanan & Heinrichs, 1989), even in the control group using the same scoring instrument for that measure. Part of the difference may be due to the attribution of equivocal signs; the assessor in this study used a conservative approach and always scored a sign as the lower value if it was between 0 and 1 or between 1 and 2. Interestingly, another study using a similarly conservative approach has also reported rates lower than those commonly found in the literature (Dazzan et al., 2004).

Primary neurological abnormalities in the schizophrenia groups Unfortunately, many of the studies of neurological dysfunction in schizophrenia either looked for only one or two primary abnormalities, such as an upgoing plantar response or hearing loss, or did not include any. A minority of studies have looked at primary neurological abnormalities in schizophrenia patients and found an increase compared with normal controls and patients with affective disorder (Bolton et al., 1998; Egan et al., 2001; Ismail et al., 1998; Kinney et al., 1992, 1993; Picchioni et al., 2006; Woods et al., 1986). In this study, which used a narrow de®nition of primary abnormality, the group of non-familial patients with schizophrenia was shown to contain a signi®cantly increased proportion of abnormal participants. There are good grounds for preferring a narrow de®nition of abnormality for primary signs, on the basis that a standard clinical examination of cranial nerve, eye signs, lateralizing limb pyramidal signs and frontal release signs would be expected to de®ne only a small proportion of control populations as

6. NEUROLOGICAL ABNORMALITIES

147

abnormal. The increase in primary abnormalities in the group of nonfamilial schizophrenia patients is in accordance with studies showing that early neurological events are over-represented in this group (Geddes & Lawrie, 1995; Lewis & Murray, 1987; Lyon et al., 1989). Previous studies of primary abnormality in schizophrenia used a criterion that de®ned up to 15% of the control population as abnormal in one study (Kinney et al., 1991), although the difference between schizophrenia and control populations was also demonstrated in an earlier study, in which 4% of the control population were de®ned as abnormal (Kinney et al., 1986). When a broad criterion of abnormality is used in the present study, as in the earlier studies, both schizophrenia groups with and without a family history of schizophrenia show a signi®cant excess of abnormal participants. This study therefore con®rms previous reports of an increase in primary abnormalities in schizophrenia, though only in the non-familial group using the narrow criterion. The study also agrees with previous ones about the effect of age and sex on primary abnormality, which were not found to be signi®cant confounders.

Integrative neurological abnormalities in the schizophrenic groups There has been considerable interest in the concept of schizophrenia as a disorder of integration of anatomically distinct brain functions, or connectivity (Friston & Frith, 1995). The concept is supported by functional imaging studies (Friston & Frith, 1995), by studies of EEG coherence (Higashima et al., 2006) and, more recently, by diffusion tensor imaging (Kubicki et al., 2005) and structural imaging studies (Dazzan et al., 2004). In the present study, signs likely to depend on integration within and between the motor and sensory systems were analysed separately from primary signs. This was in an effort to de®ne signs re¯ecting distributed cortical processing, although a de®cit in these functions need not necessarily be due to a distributed de®cit; a focal lesion at one point in a cortical network can have the same effect. Most of the previous studies of neurological abnormality in schizophrenia have included integrative signs, often exclusively (Arango et al., 2000; Buchanan et al., 1990; Cox & Ludwig, 1979a, 1979b; Dazzan et al., 2004; Gupta et al., 1995; Keshavan et al., 2003; Kolakowska et al., 1985; Manschreck et al., 1981; Owens & Johnstone, 1980; Rossi et al., 1990; Sanders et al., 1994; Schroder et al., 1991; Torrey 1980; Tucker et al., 1974). The studies are dif®cult to compare directly because of different methodology and patient groups, but some general ®ndings emerge. In the studies in which patients with schizophrenia have been compared with normal control groups, patients often show an increase in integrative neurological signs. Studies of ®rst-episode patients with

148

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

schizophrenia suggest that this integrative de®cit is pre-existing, and that is independent from use of antipsychotic drugs (Dazzan & Murray, 2002). In the present study, which used a broad de®nition of abnormality, the familial schizophrenia group, but not the non-familial schizophrenia group, was shown to be signi®cantly abnormal. We suggest that the broader de®nition of abnormality, unlike the case for primary signs, does have face validity for integrative signs. The range of integrative signs used includes more subtle signs of brain dysfunction, which are abnormal in a larger proportion of non-patient populations than for the primary signs. The increased proportion of abnormal participants in the patients with schizophrenia from multiply-affected families is in accordance with the existence of a familial integrative abnormality. When a narrower de®nition of abnormality was used, both groups of patients with schizophrenia were found to have a signi®cantly increased proportion of abnormal participants. We therefore con®rm previous ®ndings of integrative abnormalities in patients with schizophrenia, although only in the familial schizophrenia group using what we believe to be a reasonable criterion of abnormality.

Neurological abnormalities and familial schizophrenia Primary and integrative signs have been considered separately in this study as correlating with focal damage and disconnection respectively. In terms of underlying mechanism, perinatal or prenatal focal insult in non-familial schizophrenia (Lewis & Murray, 1987) could produce focal damage. A prediction based on this is that primary signs could occur in non-familial schizophrenia, but might be less likely to occur in familial schizophrenia. An underlying genetic brain disconnection predisposing to schizophrenia, however, might lead to an increase in integrative abnormalities in familial individuals with schizophrenia and their relatives. The effect of family history on the presence of primary abnormality in patients with schizophrenia has been the subject of studies (Woods et al., 1987, 1991) reporting an increase in focal signs in schizophrenia patients with, as opposed to those without, a family history of schizophrenia. This is the reverse of the ®nding in the present study. A study by the same group of unaffected relatives of patients with schizophrenia who were selected on the basis of a positive family history also showed an increase in primary signs (Kinney et al., 1986). An increase of primary neurological signs was also shown in unaffected relatives of patients with schizophrenia not selected on the basis of a positive family history (Egan et al., 2001; Ismail et al., 1998; Kinney et al., 1991; Schubert & McNeil, 2004). Two recent twin studies that have evaluated twin pairs concordant and discordant for schizophrenia also did not ®nd signi®cant differences in focal signs between monozygotic twins

6. NEUROLOGICAL ABNORMALITIES

149

concordant and discordant for schizophrenia (Kelly et al., 2004; Picchioni et al., 2006). Kinney and colleagues (Kinney et al., 1991) argue for a familial tendency towards focal neurological abnormalities in both the multiply-affected and non-familial groups, a hypothesis that is not supported by our data. Differences in ®ndings may re¯ect the fact that the sample sizes of patients and their relatives in these studies was much smaller than our sample. Furthermore, some of these studies also used a less conservative criterion for primary abnormality, and different individual measures. With respect to integrative abnormalities, most of the previous work on the relationship between family history of schizophrenia and the presence of abnormalities has been carried out on children at a high genetic risk for schizophrenia. These detailed, repeated studies of small numbers of children have shown the presence of neurological abnormalities that might be compared with the integrative abnormalities studied here, although comparison between such different studies can be made only with extreme caution. It is of considerable interest, however, to note that the `pandevelopmental' disorder described by Fish (1977; meaning a maturational disorder that involves the brain) was present in one study in children who were high risk by virtue of family history, but not in children at high risk due to birth trauma. Moreover, there is a suggestion from some high-risk studies that the early de®cits do correlate with sustained neurological dysfunction (Marcus et al., 1985; Rosso et al., 2000). Studies of integrative dysfunction that have been performed on nonpsychotic relatives of patients with schizophrenia generally report an increase of integrative abnormalities in the non-psychotic relatives (Ismail et al., 1998; Lawrie et al., 2001; Rosso et al., 2000; Yazici et al., 2002). Furthermore, among relatives of patients with schizophrenia, the presumed carriers of genetic loading seem to show more integrative de®cits than the presumed non-carriers (Gourion et al., 2004), suggesting a genetic contribution to the risk of integrative dysfunction. One small study has shown a signi®cant increase in mainly integrative abnormalities in patients with schizophrenia from multiply-affected families, but not non-familial patients with schizophrenia, compared with normal and psychiatric controls (Walker & Shaye, 1982). This is in accordance with the present study. Another study of ®fty-eight patients with schizophrenia not selected on the basis of family history and only thirty-one relatives (Rossi et al., 1990) has shown an increase in predominantly integrative signs in both groups, unlike the case here. It is of note that some authors have reported that, while there was no association between a positive family history of psychosis and neurological abnormalities at the time of ®rst presentation, the patients with a history of psychotic disorder in ®rst-degree relatives show a signi®cant neurological

150

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

deterioration at follow up (Madsen et al., 1999). Therefore, it is possible that, if a positive family history is associated with more progression in neurological dysfunction, studies of patients at varying stages of schizophrenia can provide different results.

Drug effects and blinding in assessment of the groups with schizophrenia We investigated whether differences between the two schizophrenia groups were due to antipsychotic medications. We looked at the prescribed drug doses and we found that these were similar between the two groups with schizophrenia. Similarly, the scores of the scales that can be considered a measure of drug dosage were also comparable (AIMS and TAKE scores). The double dissociation of neurological abnormality in the schizophrenia groups for primary and integrative abnormality is not, therefore, consistent with a simple medication effect. If the primary abnormalities seen in the non-familial group were due to a general effect of drugs in schizophrenia, they should also have occurred in the familial group. Likewise, if integrative neurological abnormalities in the familial schizophrenia group were due to a general drug effect in schizophrenia, then they should have been increased in the non-familial group too. Finally, the possible confounding effect of drug dose was assessed using logistic regression, and was not found to be signi®cant for the primary or integrative scores. The issue of blinding is a second possible concern. Blinding to schizophrenia versus non-schizophrenia status cannot ever be complete in such a study, as the presence of any mannerisms, suggestive appearance or drugrelated extrapyramidal features in psychotic individuals cannot be masked. However, the observer was fully blinded to the familial status of the two schizophrenia groups. The presence of a double dissociation of neurological de®cits shown in the two groups with schizophrenia cannot therefore be attributed to blinding differences between the schizophrenia groups.

The importance of threshold criteria This study suggests that the demonstration of differences in neurological function in groups with non-familial and familial schizophrenia depends on the threshold criteria used to de®ne abnormality. When certain threshold criteria for abnormality are used, the differences between the two schizophrenia groups are not apparent. Previous published studies have not used different thresholds of abnormality to investigate group differences. In view of the demonstrated effect, we suggest that thresholds are as important as the neurological measures themselves, and should be chosen with great care.

6. NEUROLOGICAL ABNORMALITIES

151

CONCLUSION In this study, by using clearly de®ned populations that are enriched or depleted for genetic risk, and a systematic neurological instrument, a different pattern of abnormality has been shown in the two schizophrenia patient groups. Integrative neurological signs appear most prominent in patients with familial schizophrenia, whereas focal neurological signs appear most prominent in patients with non-familial schizophrenia. These data support a hypothesis whereby integrative neurological signs re¯ect distributed impaired brain connectivity linked to genetic risk for schizophrenia, whereas focal neurological signs in schizophrenia re¯ect subtle brain damage of environmental origin. In further support of integrative neurological signs as an endophenotype of schizophrenia, similar integrative neurological abnormalities have been demonstrated in the non-psychotic ®rst-degree relatives of familial schizophrenia patients.

ACKNOWLEDGEMENTS This work was published in full in the journal Brain (Grif®ths et al., 1998)

REFERENCES Annett, M.A. (1970). Classi®cation of hand preference by association analysis. British Journal of Psychology, 61, 303±321. Arango, C., Kirkpatrick, B., Buchanan, R.W. (2000). Neurological signs and the heterogeneity of schizophrenia. American Journal of Psychiatry, 157, 560±565. Baldessarini, R.J., Katz, B., Cotton, P. (1984). Dissimilar dosing with high-potency and lowpotency neuroleptics. American Journal of Psychiatry, 141, 748±752. Bolton, D., Gibb, W., Lees, A., Raven, P., Gray, J.A., Chen, E., et al. (1998). Neurological soft signs in obsessive compulsive disorder: standardised assessment and comparison with schizophrenia. Behavioral Neurology, 11, 197±204. Buchanan, R.W., Heinrichs, D.W. (1989). The neurological evaluation scale (NES): A structured instrument for the assessment of neurological signs in schizophrenia. Psychiatry Research, 27, 335±350. Buchanan, R.W., Kirkpatrick, B., Heinrichs, D.W., Carpenter, W.T., Jr. (1990). Clinical correlates of the de®cit syndrome of schizophrenia. American Journal of Psychiatry, 147, 290±294. Cox, S.M., Ludwig, A.M. (1979a). Neurological soft signs and psychopathology: incidence in diagnostic groups. Canadian Journal of Psychiatry, 24, 668±673. Cox, S.M., Ludwig, A.M. (1979b). Neurological soft signs and psychopathology. I. Findings in schizophrenia. Journal of Nervous and Mental Disease, 167, 161±165. Davis, J.M. (1976). Comparative doses and costs of antipsychotic medication. Archives of General Psychiatry, 33, 858±861. Dazzan, P., Murray, R.M. (2002). Neurological soft signs in ®rst-episode psychosis: A systematic review. British Journal of Psychiatry, Supplement, 43, s50±s57. Dazzan, P., Morgan, K.D., Orr, K.G., Hutchinson, G., Chitnis, X., Suckling, J., et al. (2004). The structural brain correlates of neurological soft signs in AESOP ®rst-episode psychoses study. Brain, 127, 143±153. Egan, M.F., Hyde, T.M., Bonomo, J.B., Mattay, V.S., Bigelow, L.B., Goldberg, T.E., et al.

152

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

(2001). Relative risk of neurological signs in siblings of patients with schizophrenia. American Journal of Psychiatry, 158, 1827±1834. Fish, B. (1977). Neurobiologic antecedents of schizophrenia in children. Evidence for an inherited, congenital neurointegrative defect. Archives of General Psychiatry, 34, 1297±1313. Friston, K.J., Frith, C.D. (1995). Schizophrenia: a disconnection syndrome? Clinical Neuroscience, 3, 89±97. Geddes, J.R., Lawrie, S.M. (1995). Obstetric complications and schizophrenia: A metaanalysis. British Journal of Psychiatry, 167, 786±793. Gourion, D., Goldberger, C., Olie, J.P., Loo, H., Krebs, M.O. (2004). Neurological and morphological anomalies and the genetic liability to schizophrenia: A composite phenotype. Schizophrenia Research, 67, 23±31. Grif®ths, T.D., Sigmundsson, T., Takei, N., Rowe, D., Murray, R.M. (1998). Neurological abnormalities in familial and nonfamilial schizophrenia. Brain, 121, 191±203. Gupta, S., Andreasen, N.C., Arndt, S., Flaum, M., Schultz, S.K., Hubbard, W.C., et al. (1995). Neurological soft signs in neuroleptic-naive and neuroleptic-treated schizophrenic patients and in normal comparison subjects. American Journal of Psychiatry, 152, 191±196. Higashima, M., Takeda, T., Kikuchi, M., Nagasawa, T., Koshino, Y. (2006). Functional connectivity between hemispheres and schizophrenic symptoms: A longitudinal study of interhemispheric EEG coherence in patients with acute exacerbations of schizophrenia. Clinical EEG and Neuroscience, 37, 10±15. Ismail, B.T., Cantor-Graae, E., Cardenal, S., McNeil, T.F. (1998). Neurological abnormalities in schizophrenia: Clinical, etiological and demographic correlates. Schizophrenia Research, 30, 229±238. Kelly, B.D., Cotter, D., Denihan, C., Larkin, D., Murphy, P., Kinsella, A., et al. (2004). Neurological soft signs and dermatoglyphic anomalies in twins with schizophrenia. European Psychiatry, 19, 159±163. Keshavan, M.S., Sanders, R.D., Sweeney, J.A., Diwadkar, V.A., Goldstein, G., Pettegrew, J.W., et al. (2003). Diagnostic speci®city and neuroanatomical validity of neurological abnormalities in ®rst-episode psychoses. American Journal of Psychiatry, 160, 1298±1304. Kinney, D.K., Woods, B.T., Yurgelun-Todd, D. (1986). Neurological abnormalities in schizophrenic patients and their families. II. Neurologic and psychiatric ®ndings in relatives. Archives of General Psychiatry, 43, 665±668. Kinney, D.K., Yurgelun-Todd, D.A., Woods, B.T. (1991). Hard neurologic signs and psychopathology in relatives of schizophrenic patients. Psychiatry Research, 39, 45±53. Kinney, D.K., Yurgelun-Todd, D.A., Woods, B.T. (1992). Neurologic signs in patients with paranoid and nonparanoid schizophrenia. Journal of Neuropsychiatry and Clinical Neuroscience, 4, 447±449. Kinney, D.K., Yurgelun-Todd, D.A., Woods, B.T. (1993). Neurological hard signs in schizophrenia and major mood disorders. Journal of Nervous and Mental Disease, 181, 202± 204. Kolakowska, T., Williams, A.O., Jambor, K., Ardern, M. (1985). Schizophrenia with good and poor outcome. III: Neurological `soft' signs, cognitive impairment and their clinical signi®cance. British Journal of Psychiatry, 146, 348±357. Kubicki, M., McCarley, R., Westin, C.F., Park, H.J., Maier, S., Kikinis, R., et al. (2005). A review of diffusion tensor imaging studies in schizophrenia. Journal of Psychiatric Research, 41, 15±30. Lawrie, S.M., Byrne, M., Miller, P., Hodges, A., Clafferty, R.A., Cunningham Owens, D.G., et al. (2001). Neurodevelopmental indices and the development of psychotic symptoms in subjects at high risk of schizophrenia. British Journal of Psychiatry, 178, 524±530. Lewis, S.W., Murray, R.M. (1987). Obstetrics complications, neurodevelopmental deviance, and risk of schizophrenia. Journal of Psychiatric Research, 21, 413±421.

6. NEUROLOGICAL ABNORMALITIES

153

Lyon, M., Barr, C.E., Cannon, T.D., Mednick, S.A., Shore, D. (1989). Fetal neural development and schizophrenia. Schizophrenia Bulletin, 15, 149±161. Madsen, A.L., Vorstrup, S., Rubin, P., Larsen, J.K., Hemmingsen, R. (1999). Neurological abnormalities in schizophrenic patients: A prospective follow-up study 5 years after ®rst admission. Acta Psychiatrica Scandinavica, 100, 119±125. Manschreck, T.C., Maher, B.A., Rucklos, M.E., Vereen, D.R., Jr., Ader, D.N. (1981). De®cient motor synchrony in schizophrenia. Journal of Abnormal Psychology, 90, 321±328. Marcus, J., Hans, S.L., Mednick, S.A., Schulsinger, F., Michelsen, N. (1985). Neurological dysfunctioning in offspring of schizophrenics in Israel and Denmark. A replication analysis. Archives of General Psychiatry, 42, 753±761. Murray, R.M., Jones, P.B. (1996). Nonfamilial schizophrenia ± another male preserve? European Psychiatry, 11, 286±288. Murray, R.M., Lewis, S.W., Reveley, A.M. (1985). Towards an aetiological classi®cation of schizophrenia. Lancet, 1, 1023±1026. National Institute of Mental Health (1976). Abnormal Involuntary Movement Scale (AIMS). Rockville, Maryland: National Institute of Mental Health, DHEW Publication (ADM). Nelson, H.E. (1982). The National Adult Reading Test. Windson, UK: NFER-Nelson. Owens, D.G., Johnstone, E.C. (1980). The disabilities of chronic schizophrenia ± their nature and the factors contributing to their development. British Journal of Psychiatry, 136, 384±395. Picchioni, M.M., Toulopoulou, T., Landau, S., Davies, N., Ribchester, T., Murray, R.M. (2006). Neurological abnormalities in schizophrenic twins. Biological Psychiatry, 59, 341±348. Quitkin, F., Rifkin, A., Klein, D.F. (1976). Neurologic soft signs in schizophrenia and character disorders. Organicity in schizophrenia with premorbid asociality and emotionally unstable character disorders. Archives of General Psychiatry, 33, 845±853. Rossi, A., De Cataldo, S., Di Michele, V., Manna, V., Ceccoli, S., Stratta, P., et al. (1990). Neurological soft signs in schizophrenia. British Journal of Psychiatry, 157, 735±739. Rosso, I.M., Bearden, C.E., Hollister, J.M., Gasperoni, T.L., Sanchez, L.E., Hadley, T., et al. (2000). Childhood neuromotor dysfunction in schizophrenia patients and their unaffected siblings: A prospective cohort study. Schizophrenia Bulletin, 26, 367±378. Roy, M.A., Crowe, R.R. (1994). Validity of the familial and nonfamilial subtypes of schizophrenia. American Journal of Psychiatry, 151, 805±814. Sanders, R.D., Keshavan, M.S., Schooler, N.R. (1994). Neurological examination abnormalities in neuroleptic-naive patients with ®rst-break schizophrenia: Preliminary results. American Journal of Psychiatry, 151, 1231±1233. Schroder, J., Niethammer, R., Geider, F.J., Reitz, C., Binkert, M., Jauss, M., et al. (1991). Neurological soft signs in schizophrenia. Schizophrenia Research, 6, 25±30. Schubert, E.W., McNeil, T.F. (2004). Prospective study of neurological abnormalities in offspring of women with psychosis: Birth to adulthood. American Journal of Psychiatry, 161, 1030±1037. Shibre, T., Kebede, D., Alem, A., Kebreab, S., Melaku, Z., Deyassa, N., et al. (2002). Neurological soft signs (NSS) in 200 treatment-naive cases with schizophrenia: A community-based study in a rural setting. Nordic Journal of Psychiatry, 56, 425±431. Torrey, E.F. (1980). Neurological abnormalities in schizophrenic patients. Biological Psychiatry, 15, 381±388. Tucker, G.J., Campion, E.W., Kelleher, P.A., Silberfarb, P.M. (1974). The relationship of subtle neurologic impairments to disturbances of thinking. Psychotherapy and Psychosomatics, 24, 165±169. Walker, E., Shaye, J. (1982). Familial schizophrenia. A predictor of neuromotor and attentional abnormalities in schizophrenia. Archives of General Psychiatry, 39, 1153±1156.

154

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Watt, N.F., Lubensky, A.W. (1976). Childhood roots of schizophrenia. Journal of Consulting and Clinical Psychology, 44, 363±375. Wojcik, J.D., Gelenberg, A.J., LaBrie, R.A., Mieske, M. (1980). Prevalence of tardive dyskinesia in an outpatient population. Comprehensive Psychiatry, 21, 370±380. Woods, B.T., Kinney, D.K., Yurgelun-Todd, D. (1986). Neurologic abnormalities in schizophrenic patients and their families. I. Comparison of schizophrenic, bipolar, and substance abuse patients and normal controls. Archives of General Psychiatry, 43, 657±663. Woods, B.T., Yurgelun-Todd, D., Kinney, D.K. (1987). Relationship of neurological abnormalities in schizophrenics to family psychopathology. Biological Psychiatry, 22, 325±331. Woods, B.T., Kinney, D.K., Yurgelun-Todd, D.A. (1991). Neurological `hard' signs and family history of psychosis in schizophrenia. Biological Psychiatry, 30, 806±816. Yazici, A.H., Demir, B., Yazici, K.M., Gogus, A. (2002). Neurological soft signs in schizophrenic patients and their nonpsychotic siblings. Schizophrenia Research, 58, 241±246.

CHAPTER SEVEN

Structural brain deviations in schizophrenia and bipolar disorder ± to what extent are they genetically mediated? Colm McDonald

INTRODUCTION Subtle volumetric deviations of several brain structures have been consistently demonstrated when comparing patients with schizophrenia to controls. Some of these are global, such as a prominent enlargement of lateral ventricles and a small reduction in cortical grey matter, whereas others are more regional, such as reduced volume of medial temporal lobe structures, prefrontal cortex and thalamus (Shenton et al., 2001; Wright et al., 2000). Most studies have utilized hypothesis-based region-of-interest techniques to investigate these deviations, by which a small number of structures are chosen for measurement and delineated manually. These region-of-interest studies are laborious and con®ne the areas explored for deviations to a small fraction of the brain. Other studies have been performed using automated computational neuroanatomy techniques to explore the entire brain for tissue deviation. These computational neuroanatomy studies have identi®ed volume de®cit in portions of the prefrontal, medial temporal, lateral temporal and thalamic regions, as well as in other areas less commonly chosen for region-of-interest measurement, such as the insula and anterior cingulate gyrus (Ananth et al., 2002; Hulshoff Pol et al., 2001; Kuperberg et al., 2003; Sigmundsson et al., 2001; Wright et al., 1999a). It remains unclear whether brain deviations identi®ed in case±control studies are linked to the illness process itself, genetic risk for schizophrenia or environmental risk factors such as obstetric complications. It is also 155

156

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

uncertain whether such deviations are additionally associated with bipolar disorder or genetic risk for bipolar disorder. Some of the morphometric deviations associated with schizophrenia have been reported in unaffected ®rst-degree relatives of patients at high likelihood of carrying susceptibility genes for the illness, including ventricular enlargement (Cannon et al., 1993; Staal et al., 2000), hippocampal volume reduction (Seidman et al., 2002) and prefrontal cortex de®cit (Cannon et al., 2002), suggesting that these brain abnormalities are endophenotypic for schizophrenia. Exposure to pregnancy and birth complications have also been linked to ventricular enlargement and hippocampal volume reduction in schizophrenia (Baare et al., 2001b; McNeil et al., 2000; Van Erp et al., 2002). There is also evidence that schizophrenia is characterized by more extensive brain structural deviations than are found in bipolar disorder. Meta-analyses of methodologically robust case±control MRI studies of schizophrenia have demonstrated the illness to be associated with enlarged lateral ventricles and reduced volume of the cerebrum, medial temporal lobes, lateral temporal cortex and thalamus (Nelson et al., 1998; Wright et al., 2000), whereas the only consistent ®nding to emerge from a metaanalysis of bipolar-disorder studies using similar methodology was lateral ventricular enlargement, with preservation of several other structures including the hippocampus (McDonald et al., 2004b). However, the morphometry of bipolar disorder is relatively under-researched and considerably heterogenous. There is evidence from studies that compared samples of bipolar-disorder patients and schizophrenia patients with the identical control group that ventricular enlargement is also found in male patients with bipolar disorder (Swayze et al., 1990), and in psychotic bipolar disorder (Strasser et al., 2005), but that hippocampal volume de®cit is speci®c to schizophrenia (Altshuler et al., 2000). The present chapter describes ®ndings of the key morphometry studies from the Maudsley Family Study of Psychosis. Some of the results from these studies have been published previously (McDonald et al., 2002, 2004a, 2005, 2006; Schulze et al., 2003). These studies aimed to identify brain deviations associated with schizophrenia and bipolar disorder, and to investigate whether such brain deviations are endophenotypic for schizophrenia or bipolar disorder by virtue of their presence in unaffected relatives at high genetic risk for these psychotic disorders. The techniques of hypothesis-based region-of-interest morphometry of speci®c structures commonly found to deviate morphometrically in psychotic illness (cerebrum, ventricles and hippocampus) and of exploratory computational morphometry assessing regional brain deviations throughout grey and white matter were applied to datasets within the sample to address these aims.

7. STRUCTURAL BRAIN DEVIATIONS

157

STUDY 1: REGION-OF-INTEREST ANALYSES OF PATIENTS WITH SCHIZOPHRENIA OR BIPOLAR DISORDER AND THEIR UNAFFECTED RELATIVES The speci®c hypotheses for the region-of-interest studies included: (1) patients with schizophrenia would have increased lateral and third ventricular volume and decreased hippocampal volume compared to controls; (2) patients with bipolar disorder would have increased lateral ventricular volume, but preserved hippocampal volume; (3) unaffected relatives of patients with schizophrenia would have increased ventricular volume compared to controls, which would be most prominent in those relatives from multiply affected families; (4) unaffected relatives of patients with bipolar disorder would have increased ventricular volume; (5) unaffected relatives of patients with schizophrenia would have reduced hippocampal volume; (6) unaffected relatives of patients with bipolar disorder would have preserved hippocampal volume. Structural MRI brain scans were successfully obtained on 465 participants throughout Phases 1±4 of the study, comprising 107 patients with schizophrenia or schizoaffective disorder (n=9), 150 of their unaffected ®rst-degree relatives, 38 patients with bipolar 1 disorder, 52 of their unaffected ®rst-degree relatives and 118 healthy comparison participants. These analyses were performed on a combination of participants described previously under different arms of the Maudsley Family Study of Psychosis (McDonald et al., 2002, 2006; Schulze et al., 2003; Sharma et al., 1998).

MRI acquisition and processing For each participant, a set of high-resolution, 1.5-mm-thick contiguous coronal T1-weighted MRI images extending through the entire brain was obtained using three-dimensional spoiled gradient recall echo sequences on 1.5 Tesla General Electric Signa System scanners. Data for Phases 1 and 2 were acquired at St. George's Hospital, London, using one of the following protocols: echo time = 5 ms, repetition time = 35 ms, number of excitations = 1, ®eld of view = 20 cm, acquisition matrix = 256Ò256 and ¯ip angle = 35 degrees (n=148) or TE=3.7 ms, TR=14.7 ms, number of excitations = 1, FOV = 20 cm, acquisition matrix = 256Ò256 and ¯ip angle 20 degrees (n=72). Data for Phases 3 and 4 were acquired at the Maudsley Hospital, London, using the following protocol: TE=5.8 ms, TR=13.1 ms, number of excitations = 1, acquisition matrix = 256Ò256Ò128, ¯ip angle = 20 degrees (n=243). Each MRI image was rated blind to group af®liation using MEASURE (version 0.8, Johns Hopkins University, Baltimore, Maryland), an image analysis program that employs stereologically unbiased estimation of volume (Barta et al., 1997; Doherty et al., 2000). In this program, a grid is

158

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

applied over the whole volume of the brain and the rater marks grid points falling into the structure of interest, taking into consideration all three orthogonal views. The MEASURE program calculates the volume of a structure by multiplying the number of marked grid points by the volume of an elementary cuboid. Prior to any measurements, head tilt was corrected by aligning each brain along the anterior±posterior commissure axis in the sagittal plane and along the interhemispheric ®ssure in the coronal and axial planes. The following regions were rated: cerebral volume, right and left lateral ventricles, third ventricle and right and left hippocampus, using established structure boundaries and grid spacings for each of these regions (McDonald et al., 2002; Schulze et al., 2003; Sharma et al., 1998). Inter-rater reliability estimates based on random samples of ten brains calculated using the intraclass correlation co-ef®cient were: cerebral volume 0.91±0.99, left lateral ventricle 0.90±0.99, right lateral ventricle 0.95±0.99, third ventricle 0.90±0.98, right hippocampus 0.86±0.94, left hippocampus 0.87±0.92. Intra-rater reliability based on ten brains rated at least 6 months apart were all between 0.91 and 0.99.

Data analysis Morphometric analyses were performed using STATA (version 8.2; Stata Corporation, Texas, USA). The data were analysed with multivariate linear regression using robust variance estimators for clustered data to account for any intrafamilial correlation of volumetric measurements. Each participant group was compared to the control group in a single analysis for each brain structure with volume as the dependent variable and controlling for age, gender, handedness, height, a quadratic function of age and scanner. Height was included as an independent variable as it is a good predictor of general head size (Andreasen et al., 1994b) and is independent of any diseasespeci®c processes. A quadratic function of age was included to model potential non-linear age effects on brain structure. Cerebral volume was additionally included as a co-variate in the analyses of hippocampal volumes. The distributions of measurements for lateral and third ventricular volumes were skewed, and values were successfully logarithmically transformed prior to analyses. All tests were two-tailed and used a 0.05 level of signi®cance.

Results The mean unadjusted volumes of each regional brain structure in each participant group is shown in Table 7.1 and the results of the multiple regression analyses in Table 7.2.

21.45 12.51 11.51 6.81 9.95 6.02 959.47 98.36 0.89 0.45 2.36 0.27 2.29 0.30

17.63 8.50 9.20 4.66 8.42 4.14 987.54 89.45 0.82 0.40 2.41 0.24 2.35 0.23

SD

Mean

Mean

SD

Relatives of familial schizophrenia patients (n=92)

Familial patients with schizophrenia (n=58) SD

20.50 8.32 10.92 4.56 9.61 4.37 973.38 99.74 0.98 0.43 2.38 0.30 2.35 0.29

Mean

Non-familial patients with schizophrenia (n=49)

16.56 8.52 8.03 1000.02 0.80 2.45 2.49

Mean 8.24 4.35 4.18 79.16 0.31 0.29 0.26

SD

Relatives of non-familial schizophrenia patients (n=58) SD

16.35 6.47 8.49 3.80 7.84 3.05 984.17 77.88 0.75 0.29 2.50 0.24 2.45 0.26

Mean

SD 17.70 9.36 9.38 5.61 8.30 4.08 996.43 72.05 0.83 0.40 2.49 0.30 2.47 0.33

Mean

SD 14.88 7.23 7.66 3.69 7.23 3.79 968.81 77.03 0.73 0.27 2.45 0.24 2.40 0.25

Mean

Familial Relatives of Controls (n=118) bipolar disorder patients with bipolar disorder patients (n=52) (n=38)

*Values are adjusted to the sample mean for age, a quadratic function of age, gender, height, handedness and MRI scanner. Hippocampal volumes were also adjusted for cerebral volume.

Total lateral ventricles Left lateral ventricle Right lateral ventricle Whole brain volume Third ventricle Left hippocampus Right hippocampus

Region

TABLE 7.1 Mean adjusted* regional brain volumes (mL) in each participant group

±0.10

Right hippocampus

p

B

95% CI

0.17, <0.001 0.15 0.01, 0.52 0.28 0.20, <0.001 0.15 0.01, 0.57 0.30 0.11, 0.001 0.14 0.01, 0.47 0.27 ±39.59, 0.46 22.00 ±4.49, 17.96 48.50 0.04, 0.01 0.08 ±0.05, 0.33 0.21 ±0.21, 0.01 ±0.04 ±0.12, ±0.03 0.04 ±0.18, 0.02 ±0.03 ±0.11, ±0.01 0.04

95% CI

Relatives of familial schizophrenia patients

0.37

0.31

0.25

0.10

0.04

0.04

0.03

p

±0.10

±0.07

0.29

0.87

0.32

0.41

0.36

B 0.22, 0.51 0.26, 0.56 0.17, 0.47 ±31.27, 33.00 0.15, 0.43 ±0.16, 0.01 ±0.19, ±0.01

95% CI

Non-familial patients with schizophrenia

0.03

0.14

<0.001

0.96

<0.001

<0.001

<0.001

p

0.01

0.08

0.09

37.51

0.10

0.09

0.09

B ±0.06, 0.24 ±0.07, 0.25 ±0.06, 0.25 8.24, 66.79 ±0.05, 0.23 ±0.01, 0.17 ±0.09, 0.10

95% CI

Relatives of non-familial schizophrenia patients

0.90

0.08

0.21

0.01

0.23

0.27

0.23

p

95% CI

0.10 ±0.06, 0.27 0.10 ±0.08, 0.28 0.11 ±0.05, 0.27 18.65 ±14.13, 51.44 ±0.01 ±0.15, 0.14 0.07 ±0.04, 0.18 0.08 ±0.02, 0.17

B

Patients with bipolar disorder

0.12

0.20

0.93

0.26

0.19

0.29

0.21

p

95% CI 0.13 ±0.05, 0.31 0.14 ±0.06, 0.33 0.12 ±0.07, 0.30 33.67 4.67, 62.68 0.09 ±0.08, 0.25 0.09 ±0.01, 0.18 0.07 ±0.03, 0.17

B

Relatives of bipolar disorder patients

0.15

0.06

0.31

0.02

0.22

0.17

0.17

p

Signi®cant ®ndings are highlighted in bold.

Multiple linear regression analyses were performed with regional brain volume as the dependent variable and age, a quadratic function of age, gender, height and handedness as co-variates, and using multilevel modelling to accommodate intra-familial correlation. Whole brain volume was included as an additional covariate in the analysis of hippocampal volume.

±0.12

0.18

±10.81

0.30

0.38

0.34

B

Familial patients with schizophrenia

Left hippocampus

Total lateral ventricles Left lateral ventricle Right lateral ventricle Whole brain volume Third ventricle

Region

TABLE 7.2 Analyses of regional brain volumes for each participant group compared to the control group

7. STRUCTURAL BRAIN DEVIATIONS

161

Lateral ventricular volume Patients with both familial and non-familial schizophrenia had signi®cantly enlarged lateral ventricular volumes bilaterally when compared to controls. The unaffected relatives of the familial schizophrenia patients also had larger lateral ventricular volume bilaterally when compared to controls. Those twenty relatives of familial schizophrenia patients who were `presumed obligate carriers' of genetic risk for schizophrenia displayed signi®cantly enlarged total lateral ventricular volume compared to controls with an even larger regression co-ef®cient than other relatives (B=0.27; 95% CI=0.01, 0.52; p=0.04). By contrast, relatives of non-familial schizophrenia patients had lateral ventricular volume measurements that did not signi®cantly differ from controls. Similarly, neither patients with bipolar disorder nor their unaffected relatives had lateral ventricular volumes that signi®cantly differed from the control sample. Nor did the ventricular volume measurements of those eleven relatives of bipolar-disorder patients who were `presumed obligate carriers' of genetic risk signi®cantly differ from controls (B=0.11; 95% CI=±0.25, 0.48, p=0.54). However, when compared directly, the lateral ventricular volume measurements of patients with schizophrenia were not signi®cantly larger than those of patients with bipolar disorder (B=0.20; 95% CI=±0.04, 0.43; p=0.10). As other psychiatric illnesses, including major depressive disorder, and schizotypal disorder, have been associated with neuroanatomical changes such as ventricular enlargement, the analyses were repeated after excluding those sixty-one relatives and controls included in the study who had ful®lled criteria for a DSM psychiatric disorder at some stage of their lives. The results were essentially unchanged and the unaffected relatives of the familial schizophrenia patients continued to display signi®cantly larger total lateral ventricular volume (B=0.17; 95% CI=0.01, 0.33; p=0.03) when compared with controls. Figure 7.1 shows the mean lateral ventricular volumes adjusted for confounds in each group, demonstrating the stepwise pattern of increased lateral ventricular volume: greatest in patients with schizophrenia, then in those relatives from multiply-affected families, who are more likely to carry susceptibility genes, and then those relatives from singly-affected families, in whom ventricular size did not differ from the control sample.

Third ventricular volume Patients with both familial and non-familial schizophrenia had signi®cantly enlarged third ventricular volume compared to controls. Neither the familial nor the non-familial schizophrenia relatives group differed signi®cantly from the control group in third ventricular measurements, thus the ®ndings for lateral ventricles were not mirrored in the third ventricles, despite the strong correlation in the volumes of lateral and third ventricles

162

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Lateral ventricular volume (mL)

24

**

22 20

*

18 16 14 12 10 Patients with Relatives of Relatives of schizophrenia familial non-familial (n=107) schizophrenia schizophrenia patients patients (n=92) (n=58)

Controls (n=117)

Bipolar-disorder Relatives of patients bipolar-disorder (n=38) patients (n=53)

Figure 7.1 Mean (95% CI) total lateral ventricular volume (mL), adjusted for confounds, in each participant group in entire sample (n=465). *p<0.05; **p<0.001. Patients with familial and non-familial schizophrenia combined for this ®gure.

in the sample (Pearson correlation co-ef®cient = 0.67, p<0.001). Neither patients with bipolar disorder nor their unaffected relatives had third ventricular volumes that differed from the control sample. The third ventricular volume measurements of schizophrenia or bipolar presumed obligate carrier relatives did not signi®cantly differ from controls. When directly compared, patients with schizophrenia had signi®cantly larger third ventricular volumes than patients with bipolar disorder (B=0.21; 95% CI=0.01, 0.41; p=0.04).

Whole brain volume Whole brain volume did not differ signi®cantly between either group of patients with schizophrenia and the control group, nor between relatives of familial schizophrenia patients and the control group. Patients with bipolar disorder had no signi®cant differences in whole brain volume. However, the groups of relatives of bipolar-disorder patients and non-familial patients with schizophrenia had slightly larger whole brain volume measurements than the control group. The whole brain volume measurements of schizophrenia or bipolar presumed obligate carrier relatives did not signi®cantly differ from controls. There were no signi®cant differences between patients with schizophrenia and patients with bipolar disorder when these patient

7. STRUCTURAL BRAIN DEVIATIONS

163

Hippocampal volume (mL)

5.2

5

4.8

4.6

4.4 Patients with Relatives of Relatives of schizophrenia familial non-familial (n=107) schizophrenia schizophrenia patients patients (n=58) (n=92)

Controls (n=117)

Bipolar-disorder Relatives of patients bipolar-disorder (n=38) patients (n=53)

Figure 7.2 Mean (95% CI) total hippocampal volume (mL), adjusted for confounds, in each participant group in the entire sample (n=464). *p<0.01. Patients with familial and non-familial schizophrenia combined for this ®gure.

groups were compared directly (B=±13.22; 95% CI=±55.34, 28.88; p=0.54).

Hippocampal volume Patients with familial schizophrenia had signi®cantly reduced hippocampal volume bilaterally when compared to controls. Patients with non-familial schizophrenia had signi®cantly reduced right hippocampal volume measurements compared with controls, but left hippocampal volume reductions failed to reach statistical signi®cance. Neither of the groups of unaffected relatives of patients with schizophrenia differed signi®cantly from the control group in measurements of left or right hippocampal volume. Neither patients with bipolar disorder nor their unaffected relatives signi®cantly differed from the control sample in left or right hippocampal volume measurements. The hippocampal volume measurements of schizophrenia or bipolar presumed obligate carrier relatives did not signi®cantly differ from controls. When compared directly, patients with schizophrenia had smaller left hippocampal volumes (B=±0.18; 95% CI=±0.33, ±0.02; p=0.02) and right hippocampal volumes (B=±0.21; 95% CI=±0.33, ±0.08; p=0.002) than patients with bipolar disorder. Figure 7.2 demonstrates mean total hippocampal volume measurements, adjusted for confounds, in each participant group.

164

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Gender interactions There were no signi®cant group by gender interactions for total lateral ventricular volume, whole brain volume, third ventricular volume or left or right hippocampal volume, indicating that the signi®cant ®ndings of the initial analyses were contributed to by participants of both genders.

Discussion Morphometric abnormalities associated with schizophrenia and bipolar disorder This component of the study demonstrates that the morphometric abnormalities most commonly associated with schizophrenia, i.e. lateral and third ventricular enlargement and hippocampal volume reduction (Nelson et al., 1998; Shenton et al., 2001; Wright et al., 2000), are con®ned to schizophrenia and are not also shared by psychotic bipolar disorder. Thus, the ®rst hypothesis was con®rmed and second hypothesis partially con®rmed. When the two patient groups were compared directly, patients with schizophrenia had signi®cantly enlarged third ventricular volumes and signi®cantly reduced hippocampal volumes compared to patients with bipolar disorder, underlining the speci®city of these morphometric deviations to schizophrenia. Previous structural imaging studies have reported mild lateral ventricular enlargement in bipolar disorder. These studies often used heterogeneous samples of patients, which included patients with unipolar depression (Bearden et al., 2001; Elkis et al., 1995; Soares & Mann, 1997). The present study was performed on a relatively homogenous sample of patients who all ful®lled criteria for bipolar 1 disorder and had experienced psychotic symptoms during episodes of illness exacerbation. Although regression coef®cients were positive (in the direction of enlargement), no signi®cant ventricular enlargement was detected in bipolar disorder. There is a possibility of type II error in the present study and it may be that a larger sample size would have detected signi®cant ventricular enlargement. This interpretation is supported by the lack of signi®cant differences in measurements of the lateral ventricles when schizophrenia and bipolar disorder were compared directly, consistent with reports that ventricular volume measurements in patients with bipolar disorder are often intermediate between schizophrenia and controls (McDonald et al., 2004b). Schizophrenia has also been linked to mild cerebral volume de®cit, which did not emerge in the present study. However, this volume de®cit is very subtle, around 2% in meta-analysis (Wright et al., 2000). In keeping with the present study, meta-analyses (Hoge et al., 1999; McDonald et al., 2004b) report preservation of cerebral volume in bipolar disorder.

7. STRUCTURAL BRAIN DEVIATIONS

165

Previous studies that compared hippocampal volume in patients with schizophrenia and bipolar disorder (or `affective psychosis'), either with each other or the same control group, have generally found that medial temporal lobe volume de®cit was speci®c to schizophrenia (Altshuler et al., 2000; Hirayasu et al., 1998; Pearlson et al., 1997), although hippocampal de®cit in both disorders has also been reported (Velakoulis et al., 1999). As is the case with ventricular volume, however, there is stronger evidence in the literature that hippocampal volume is reduced in unipolar depression (Campbell et al., 2004; Sheline et al., 1996) and the present study therefore further supports the existence of morphometric distinctions between the major diagnoses within the psychotic and affective disorders spectrum.

Morphometric endophenotypes associated with schizophrenia and bipolar disorder The third hypothesis was con®rmed: unaffected relatives of familial patients with schizophrenia showed lateral ventricular enlargement, and this was particularly prominent in presumed obligate carriers of genetic risk for schizophrenia, thus supporting the latter as a potential endophenotype of schizophrenia. There was insuf®cient evidence to support third ventricular volume enlargement as an endophenotype for schizophrenia in the present study, in contrast to some other studies of unaffected ®rst-degree relatives (Lawrie et al., 2001; Staal et al., 2000). The utility of lateral ventricular volume as an endophenotype for schizophrenia must be considered in the light of the questionable heritability of the volume of this structure. There is disagreement among normal twin studies regarding the heritability of this structure, with some reporting high (Pfefferbaum et al., 2000; Reveley et al., 1984) and others low heritability estimates (Baare et al., 2001a; Wright et al., 2002). Further studies are needed to clarify this issue, but clearly low heritability would undermine the use of lateral ventricular volume as an endophenotype. Researchers could attempt to overcome this problem, when analysing associations between lateral ventricular volume as an endophenotypic measure and genotypic variation, by controlling for or excluding participants who had experienced putative environmental risk factors for lateral ventricular enlargement, such as hypoxic birth complications. There was no evidence for ventricular volume enlargement in unaffected ®rst-degree relatives of patients with bipolar disorder (or in presumed obligate carriers), failing to support the fourth hypothesis. Again, a type II error is possible given the size and direction of the positive regression coef®cient. The slight increase in cerebral volume in relatives of patients with nonfamilial schizophrenia and in relatives of patients with bipolar disorder compared to the control sample was unexpected. The latter ®nding echoes

166

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

some reports from the neuropsychological literature, where bipolar relatives were reported to perform better than relatives of patients with schizophrenia and controls on measures of IQ (Kremen et al., 1998) and premorbid IQ (Gilvarry et al., 2000), although de®cits in speci®c domains such as declarative memory and executive function are also reported. Given the unexpected direction of these ®ndings and number of statistical tests performed, these results remain tentative and warrant replication. The study refutes the ®fth and con®rms the sixth hypothesis, in that it provided no evidence that hippocampal volume reduction is endophenotypic for schizophrenia or for bipolar disorder. Even those presumed obligate familial relatives most likely to be unaffected gene carriers, and who share enlarged lateral ventricles with their affected relatives, had hippocampal volume measurements that did not signi®cantly differ from controls. This ®nding is consistent with some other studies of relatives of patients with schizophrenia (Harris et al., 2002; Staal et al., 2000); however, it contrasts with other volumetric studies of the hippocampus in relatives of patients with schizophrenia (Seidman et al., 2002; van Erp et al., 2002, 2004) or adolescents/young adults at a high risk of developing schizophrenia (Lawrie et al., 2001). On top of these con¯icting ®ndings, the utility of hippocampal volume as an endophenotype is undermined by its relatively low heritability (0.40) compared to other brain morphometric measures (Sullivan et al., 2001), consistent evidence from twin studies of schizophrenia for non-genetic factors impacting on hippocampal volume (McNeil et al., 2000; van Erp et al., 2004), and the susceptibility of the hippocampus to environmental risk factors such as obstetric complications (McNeil et al., 2000; Nosarti et al., 2002; van Erp et al., 2002) and stress-induced glucocorticoid excess (Sapolsky, 2000). The preservation of hippocampal volume in unaffected relatives of bipolar disorder patients, as well as in patients themselves, provides further evidence that this structure is not impacted by genetic risk for bipolar disorder, as might be hypothesized if, for example, genetically mediated hippocampal de®cit were associated with bipolar disorder but reversed in patient groups due to the neurotrophic effects of treatment with mood stabilizers.

Methodological issues Strengths of the study include the very large numbers of patients and relatives included, high-resolution MRI and direct comparison of families with schizophrenia or bipolar disorder against a common control group. A regression analysis for clustered observations was used, which takes into account the non-independence of individuals within families, a factor ignored by some other family studies, despite evidence for a high degree of genetic control of brain structures (Thompson et al., 2001; White et al., 2002).

7. STRUCTURAL BRAIN DEVIATIONS

167

Regional brain volume was assessed using a stereological method. There is evidence for the reliability and validity of this method of morphometry (Barta et al., 1997; Doherty et al., 2000), which has been used previously to measure small structures such as hippocampal volume (Matsumoto et al., 2001; Sheline et al., 1996). This method also allows simultaneous viewing of the structure in three orthogonal planes, which enables reliable visualization of structures, especially hippocampal boundaries such as the alveus. Several statistical tests were performed on a number of regions of interest and therefore the risk of type I error increases due to multiple comparisons. On the one hand, the main ®ndings from the study ± that patients with schizophrenia and their unaffected relatives from multiply affected families have enlarged ventricular volume and that patients with schizophrenia have reduced hippocampal volume ± were all hypothesized and replicate previous ®ndings, and so the statistical threshold utilized is justi®able for these comparisons. On the other hand, the ®ndings of enlarged cerebral volume in the relatives of patients with bipolar disorder or non-familial schizophrenia fail to reach signi®cance after correction for multiple testing and should be treated with more caution.

Conclusion These results con®rm previous ®ndings of enlarged lateral and third ventricular volume and reduced hippocampal volume in schizophrenia and provide evidence that these ®ndings are speci®c to schizophrenia and are not shared by psychotic bipolar disorder. The study provides evidence that ventricular enlargement is a marker of genetic liability for schizophrenia, by ®nding increased ventricular enlargement associated with increasing genetic risk among the unaffected ®rst-degree relatives of patients with schizophrenia. The study failed to support a relationship between hippocampal volume loss and genetic susceptibility to schizophrenia. Abnormalities of ventricular or hippocampal volume were not linked to genetic liability for bipolar disorder. Thus, in terms of ventricular and hippocampal volume, both the morphometric characteristics and the morphometric endophenotypes of schizophrenia and bipolar disorder are distinct.

STUDY 2: STRUCTURAL BRAIN DEVIATIONS ASSOCIATED WITH SCHIZOPHRENIA AND BIPOLAR DISORDER ASSESSED USING COMPUTATIONAL MORPHOMETRY This study utilized computational morphometry to explore brain deviations throughout grey and white matter characterizing patients with schizophrenia or bipolar disorder. As voxel-based analysis can be confounded by images acquired using different scanners/sequences (Ashburner & Friston,

168

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

2000), this study was con®ned to participants on whom MRI data were acquired utilizing the identical parameters on the same scanner at the Maudsley Hospital. Only those participants with schizophrenia who had a family history of illness were included, as these were considered most comparable to the sample of familial bipolar disorder patients (this patient sample is identical to that included in Study 3, described later in this chapter on the relationship between genetic liability and brain structure). The schizophrenia patients included those who had a history of schizophrenia or another functional psychosis in their ®rst- and/or second-degree relatives or schizotypal disorder among their ®rst-degree relatives, but excluded other non-familial schizophrenia patients and patients with schizoaffective disorder. One-hundred-and-fourteen participants from Phases 3 and 4 were included, comprising 25 patients with schizophrenia, 37 patients with bipolar 1 disorder and 52 normal comparison participants.

MRI preprocessing and analysis Optimized voxel-based morphometry (Ashburner & Friston, 2000; Good et al., 2001) was used to segment MRI data and co-register probabilistic maps of grey matter and white matter volume density for each participant in a standard anatomical space. Each participant's MRI scan was segmented into grey, white and cerebrospinal ¯uid (CSF) tissue classes in native space and global tissue volumes were estimated. Customized study-speci®c grey, white and CSF template images in standard stereotactic space were then created from the control group, in order to minimize any scanner speci®c bias and to provide a template matched to the sample. The original brain scan of each participant included in the study was normalized to the customized grey matter template, thus removing any contribution of nonbrain tissue or other tissue types to this spatial normalization step; these normalization parameters were applied back onto the original brain image of the participant to produce an image `optimally' normalized for grey matter segmentation. The images were then re-segmented, using the customized tissue templates as probability maps, and the grey matter maps retained. This procedure was repeated using parameters derived from normalizing each white matter map to the white matter template and reapplying to the original image to derive white matter tissue maps for each participant. The grey and white matter images were then modulated through multiplying voxel values by the Jacobian determinants from the spatial normalization to correct for volume changes introduced at this step (Good et al., 2001). Finally, all normalized, segmented, modulated grey and white matter tissue maps were smoothed at 4 mm using a full width at half maximum (FWHM) isotropic gaussian kernel for subsequent nonparametric analyses.

7. STRUCTURAL BRAIN DEVIATIONS

169

Global tissue volume differences between patients and controls were calculated using multiple linear regression with global tissue volume as the dependent variable and co-varying for age, gender and height. Case± control differences in regional grey and white matter volume between each patient group and the control group, and case±case differences between the two patient groups, were estimated by ®tting an analysis of covariance (ANCOVA) model at each intracerebral voxel in standard space, with age, gender and global tissue volume as co-variates. Inference was by permutation testing and cluster level statistics, with the signi®cance level set at that which less than one false positive cluster would be expected by chance (Bullmore et al., 1999; Sigmundsson et al., 2001). The mass of each cluster for each individual was transferred to a spreadsheet and, where multiple clusters were present, principal components (PC) analysis without rotation was used to explore the extent of correlation between discrete clusters and to reduce the dimensionality of the data prior to further analyses. Multiple linear regression with PC scores as the dependent variable and age, gender and global tissue volume as co-variates was used to test for a pathoplastic effect of gender on case±control differences in brain structure. The twotailed probability threshold for signi®cance was set at p=0.05.

Results Global tissue volume Patients with schizophrenia had larger global measures of CSF (B=24.00; 95% CI=5.05, 42.95; p=0.01) than controls, whereas patients with bipolar disorder did not differ from controls (B=9.34; 95% CI=±7.15, 25.83; p=0.26). There was no signi®cant difference in global grey matter volume measurements between patients with schizophrenis and controls (B=4.35; 95% CI=±18.97, 27.67; p=0.71) or between bipolar patients and controls (B=8.93; 95% CI=±11.36, 29.22; p=0.39). There was no difference in global white matter volume measurements between schizophrenia patients and controls (B=4.78; 95% CI=±12.43, 23.41; p=0.59) or between bipolar patients and controls (B=±0.07; 95% CI=±15.28, 15.14; p=0.99).

Regional tissue volume Grey matter Patients with schizophrenia had spatially distributed regions of grey matter volume de®cit compared to controls in twelve three-dimensional voxel clusters (Plate 3, Table 7.3). These areas were predominantly bilateral and included the hemispheres and vermis of the cerebellum, orbitofrontal cortex and temporal pole (more prominently on the right) extending to the lateral

± ± ±/20/21/28/ 38 ±/20/28/35/ 36/38 ±/9/10/11/ 32/45/47 ±/2/27/28/ 30/34/40/43

Cerebellar vermis and left cerebellar hemisphere Right cerebellar hemisphere Right superior, middle, inferior temporal gyri, parahippocampal gyrus, hippocampus Left superior temporal gyrus, parahippocampal gyrus, fusiform gyrus, hippocampus Right superior, middle, inferior, medial frontal gyri, anterior cingulate gyrus, anterior insula Bilateral thalamus, lenticular nucleus, amygdala, entorhinal cortex; left caudate head; right hippocampus, parahippocampal gyrus, insula, postcentral gyrus, inferior parietal lobule Left hippocampus, parahippocampal gyrus Left inferior, middle, superior frontal gyri, insula, lenticular nucleus Left medial frontal gyrus, anterior cingulate gyrus Right inferior parietal lobule, precentral gyrus, postcentral gyrus Right precuneus Left precuneus 27/28/30/35 8/9/10/11/ 44/45/46/47 8/9/10/32 2/3/4/6/40 7/31 7/31

Brodmann area

Area

46 ±13 ±48 ±47

±14

2

±10 59 6 ±9

44

34

±28 21

±1

±32

±26 ±40

±74 ±56 12

y

±10 42 44

x

Talairach co-ordinate of centroid voxel

20 35 48 47

±6 15

4

±8

±28

±32 ±32 ±24

z

369 315 568 335

525 2019

2920

1166

379

1995 580 1188

No. of voxels in cluster

TABLE 7.3 Talaraich co-ordinates and Brodmann areas for regions of grey matter volume de®cit in schizophrenia (n=25) compared to normal comparison participants (n=52)

7. STRUCTURAL BRAIN DEVIATIONS

171

temporal cortex, anterior cingulate gyrus, basal ganglia, thalamus, medial temporal lobe, insula, dorsolateral prefrontal cortex (more prominently on the left), right postcentral gyrus and inferior parietal lobule, and the precuneus bilaterally. PC analysis showed that these grey matter de®cits were highly correlated across regions, with all regions of grey matter volume de®cit loading positively on the ®rst PC, which explained 74.0% of the total variance. Schizophrenia was strongly associated with reduced scores on this ®rst PC (B=±1.07; 95% CI=±1.49, ±0.66; p<0.001) and there was no signi®cant interaction between participant group and gender (B=0.27; 95% CI=±0.61, 1.14; p=0.55), indicating that this pattern of grey matter de®cit was contributed to by both male and female schizophrenia patients. Patients with schizophrenia had no signi®cant regions of grey matter excess compared to controls. Patients with bipolar disorder had no signi®cant differences in regional grey matter volume compared to the control sample. Grey matter volume de®cits in spatially distributed regions were also evident when patients with schizophrenia were compared directly to those with bipolar disorder Plate 4, Table 7.4). These de®cits were in several of the regions identi®ed in the schizophrenia versus control comparison and comprised bilateral superior lateral temporal cortex, basal ganglia, insula, prefrontal cortex and precuneus, and right medial temporal lobe and thalamus. There were no signi®cant regions of grey matter excess when comparing schizophrenic patients to bipolar patients. White matter Patients with schizophrenia had de®cits in regional white matter volume compared to the control sample in two spatially extensive threedimensional voxel clusters (Plate 5, Table 7.5). These comprised prefrontal regions, anterior and posterior corpus callosum, left temporal lobe, and bilateral parietal areas, and included regions characteristically occupied by the long white matter tracts of the superior longitudinal fasciculus, occipitofrontal fasciculus bilaterally, left inferior longitudinal fasciculus, as well as interhemispheric tracts within the anterior and posterior sections of the corpus callosum. Volumes of the two voxel cluster regions were highly correlated among individuals (r=0.90, p<0.001) and mean volumes were used to examine gender interactions. Patients with schizophrenia had signi®cantly reduced mean volume scores compared to controls (B=±1.59; 95% CI=±2.46, ±0.72; p<0.001) and there was there was no signi®cant interaction between participant group and gender (B=0.64; 95% CI=±1.18, 2.46; p=0.49). Patients with bipolar disorder had distributed regional de®cits of white matter in four voxel clusters (Plate 5, Table 7.5). These comprised the

Brodmann area

±/21/27/28/ 34/38 21/22 ± 21/22/42 4/6/9/44/45/46 6/22/43/44 9/44/45/46 7/31

Area

Right superior, middle, temporal gyri, amygdala, hippocampus, parahippocampal gyrus, right putamen Left middle, superior temporal gyri Left lenticular nucleus, bilateral caudate nucleus, right thalamus Right middle, superior temporal gyri Left inferior, middle frontal gyri, insula, precentral gyrus Right inferior frontal gyrus, insula, precentral gyrus, postcentral gyrus Right inferior, middle frontal gyri Bilateral precuneus

1 ±17 1 ±14 12 6 7 ±42

±58 ±7 56 ±50 35 56 1

y

31

x

Talairach co-ordinate of centroid voxel

±6 3 4 25 16 20 47

±19

z

280 577 299 955 415 201 938

439

No. of voxels in cluster

TABLE 7.4 Talaraich co-ordinates and Brodmann areas for regions of grey matter volume de®cit in schizophrenia (n=25) compared to bipolar disorder (n=37) (Figure 7.4).

Area

Bipolar vs. controls White matter Bilateral temporoparietal lobes between the superior temporal gyrus, inferior parietal lobule, supramarginal gyrus and angular gyrus laterally and the lateral ventricle, posterior cingulate gyrus and precuneus medially, and including the splenium and posterior body of the corpus callosum Right medial frontal lobe between the anterior cingulate gyrus and medial frontal gyrus medially and the middle frontal gyrus laterally, genu of the corpus callosum Left medial frontal lobe between the anterior cingulate gyrus and medial frontal gyrus medially and the middle frontal gyrus laterally, genu of the corpus callosum Brainstem, involving the pons, midbrain and extending bilaterally to include the cerebral peduncles and posterior limbs of the internal capsule

Schizophrenia vs. controls White matter Bilateral frontal lobe between medial frontal gyrus, anterior cingulate gyrus medially and inferior and middle frontal gyri laterally, extending on the left to the superior frontal gyrus, insula and precentral gyrus; genu of the corpus callosum, right anterior limb of the internal capsule; right temporoparietal lobe between the superior and middle temporal gyri, postcentral gyrus, inferior parietal lobule, supramarginal gyrus and angular gyrus laterally and the hippocampus, lateral ventricle, cuneus, posterior cingulate gyrus and precuneus medially; left splenium Right temporo-parietal area between the lateral ventricle, posterior cingulate gyrus and precuneus medially and the superior temporal gyrus, insula, postcentral gyrus, precentral gyrus and inferior parietal lobule laterally; right splenium

Analysis

±38

29 25 ±23

19 ±19 ±1

±29

27

±4

±2

y

±18

x

±19

30

26

33

35

23

z

Talairach co-ordinate of centroid voxel

1378

620

852

4216

2523

5471

No. of voxels in cluster

TABLE 7.5 Talaraich co-ordinates and Brodmann areas for regions of white matter volume de®cit in patients (schizophrenia n=25, bipolar disorder n=37) compared to normal comparison participants (n=52) (Figure 7.5).

174

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

brainstem, bilateral prefrontal areas, anterior and posterior corpus callosum, bilateral parietal areas, and also included regions normally occupied by the superior longitudinal fasciculus, occipitofrontal fasciculus and major interhemispheric tracts. All regions of white matter volume de®cit loaded positively on the ®rst PC, which explained 72.4% of the total variance. Bipolar disorder was strongly associated with reduced scores on this ®rst PC (B=±0.79; 95% CI=±1.19, ±0.39; p<0.001) and there was there was no signi®cant interaction between participant group and gender (B=±0.37; 95% CI=±1.18, 0.44; p=0.37). There were substantial areas of precise overlap of white matter volume de®cit compared to normal comparison participants among schizophrenic and bipolar-disorder participants (Plate 5). There were no signi®cant differences in regional white matter volume when patients with schizophrenia were compared directly to patients with bipolar disorder.

Discussion Grey matter Mirroring ®ndings from the region-of-interest study (which examined a smaller number of structures in a larger number of participants), this study provides evidence that schizophrenia and bipolar disorder are characterized by quite distinctive deviations of regional brain morphometry within grey matter, and are thus likely to represent discrete disease entities from this neurobiological perspective. Although global grey matter volumes did not signi®cantly deviate from controls, schizophrenia was also associated with an increase in global CSF, consistent with cortical de®cit and with ventricular enlargement, as identi®ed in Study 1. Schizophrenia was associated with substantial distributed grey matter de®cits involving the frontotemporal cortex, medial temporal lobe, insula and medial thalamic regions, which have frequently been reported to display volume de®cit in region-ofinterest and computational morphometric neuroimaging studies of schizophrenia (Cannon et al., 2002; Goldstein et al., 1999; Honea et al., 2005; Hulshoff Pol et al., 2001; Kuperberg et al., 2003; Sigmundsson et al., 2001; Wright et al., 1999a, 2000). Consistency between region-of-interest and computational morphometry ®ndings in the medial temporal lobe is con®rmed by the high correlation (r=0.47, p=0.019) in patients with schizophrenia between total hippocampal volume as measured by regionof-interest morphometry (Study 1) and summed voxel intensities underlying the cluster of grey matter de®cit involving the medial temporal lobes identi®ed by computational morphometry in the same patients who were included in the present study. Grey matter volume was remarkably preserved in this homogenous sample of bipolar-disorder participants, despite

7. STRUCTURAL BRAIN DEVIATIONS

175

that fact that they were chosen to be most akin to schizophrenia participants in coming from the severe end of the bipolar spectrum and had experienced psychotic symptoms during episodes of illness exacerbation. Some groups have reported deviations of grey matter structures in bipolar disorder, including de®cits in sections of the prefrontal cortex (Lochhead et al., 2004; Lopez-Larson et al., 2002; Nugent et al., 2006; Sassi et al., 2004) and excesses in regions such as the amygdala (Altshuler et al., 2000; Lochhead et al., 2004; Strakowski et al., 1999). The possibility that medication could reverse grey matter volume de®cit in bipolar-disorder participants cannot be excluded by the present cross-sectional study. Most patients were taking lithium, which is neurotrophic, and some reports indicate that lithium may increase grey matter volume in vivo (Moore et al., 2000; Sassi et al., 2002, 2004). When compared directly to bipolar patients, schizophrenia patients again demonstrated distributed regions of grey matter volume de®cit which involved most of the regions identi®ed in the schizophrenia-control comparison and included bilateral frontotemporal cortex, insula, basal ganglia, precuneus, right medial temporal lobe and thalamus. These ®ndings in relation to the independent bipolar-patient sample further underline the speci®city of regional grey matter volume de®cits in these areas to schizophrenia. Previous studies that compared patients with schizophrenia and bipolar disorder (or `affective psychosis') either with each other or the same control group have reported con¯icting ®ndings. Some found that grey matter or medial temporal lobe volume de®cit was speci®c to schizophrenia (Altshuler et al., 2000; Hirayasu et al., 2001; Kasai et al., 2003; Kubicki et al., 2002a; Pearlson et al., 1997; Velakoulis et al., 2006), whereas others found evidence for some regional grey matter de®cit in both disorders (Friedman et al., 1999; Kasai et al., 2003; Kubicki et al., 2002a; Lim et al., 1999b; Strasser et al., 2005). However, there are multiple methodological differences between these studies conducted over the course of a decade, including changes in scanner technology and data analysis, as well as variation in sample size and diagnostic inclusion criteria. The safest conclusion in the light of this prior research is that the present study has provided clear evidence for greater salience of grey matter abnormalities in patients with schizophrenia compared to matched patients with bipolar disorder, suggesting that schizophrenia is associated with more severe and extensive disorganization of cortical and subcortical grey matter.

White matter In contrast to the ®ndings in relation to grey matter, distributed white matter volume de®cit within regions characteristically occupied by the major tranverse and longitudinal tracts was found in both disorders. Regional

176

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

white matter volume reduction has been less comprehensively studied in schizophrenia than grey matter. Earlier studies focused on area or shape measurements of the corpus callosum and most found reduced area measurements or distorted shape (Downhill et al., 2000; McCarley et al., 1999). Studies using computational morphometry have identi®ed regional white matter volume de®cit in schizophrenia within frontotemporal and parietal regions, anterior corpus callosum and internal capsule (Shapleske et al., 2002; Sigmundsson et al., 2001; Spalletta et al., 2003; Suzuki et al., 2002; Zhou et al., 2003). This evidence converges with that from the alternative imaging techniques of magnetization transfer imaging and diffusion transfer imaging, which have identi®ed white matter abnormalities in patients with schizophrenia compared to controls predominantly involving frontotemporal regions (Bagary et al., 2003; Buchsbaum et al., 2006; Kubicki et al., 2002b; Lim et al., 1999a), and with evidence from neurocytochemistry, neuropathology and gene expression studies implicating white matter dysfunction in schizophrenia (Davis et al., 2003). Increased rates of qualitatively measured hyperintense white matter lesions in subcortical and periventricular regions are among the most consistently reported anatomical abnormalities in bipolar disorder (Altshuler et al., 1995; Bearden et al., 2001). Some studies found high rates of such lesions in schizophrenia, especially among elderly participants with late onset of psychotic symptoms (Davis et al., 2003). Regional morphometry of white matter has been less commonly studied in bipolar disorder, but some studies have reported reduced white matter volumes in bipolar participants, consistent with the present study (Haznedar et al., 2005; Kieseppa et al., 2003; McIntosh et al., 2005). Furthermore, an expanding number of diffusion tensor imaging studies have detected white matter abnormalities in bipolar disorder, especially in frontal regions (Adler et al., 2004; Beyer et al., 2005; Haznedar et al., 2005). Interestingly, ultrastructural abnormalities of oligodendroglial cells in the prefrontal cortex are reported in both schizophrenia and bipolar disorder (Uranova et al., 2001, 2004). The ®nding from the present study that schizophrenia and bipolar disorder are both characterized by white matter volume de®cit in frontal and parietal regions ascertained by MRI scanning is in keeping with the hypothesis that both major types of psychosis represent a disorder of anatomical connectivity between components of large-scale neurocognitive networks (Bullmore et al., 1997; Wright et al., 1999b).

Methodological issues Strengths of this study include the moderately large numbers and clinical homogeneity of the psychotic disorder groups and the use of the same group of healthy volunteers for both case±control comparisons, which was

7. STRUCTURAL BRAIN DEVIATIONS

177

well-matched on key sociodemographic variables to both patient groups. Contemporary computational tools for fully automated whole brain morphometric analysis and non-parametric cluster level testing were used. The principal advantages of cluster level testing are that it confers greater sensitivity by incorporating information from more than one voxel in the test statistic and it also substantially reduces the search volume or number of tests required for a whole brain analysis, thereby mitigating the multiple comparisons problem. Parametric tests for spatial extent statistics in brain mapping may be over-conservative; hence the preferred use of a relatively assumption-free non-parametric permutation test based on data resampling (Bullmore et al., 1999; Hayasaka & Nichols, 2003). The study also has some limitations besides the general issue of type II error mentioned above. Both groups of patients were recruited on the basis of having other family members also affected with similar illnesses. It is therefore possible that the results of this study may not be generalizable to samples of patients with non-familial forms of psychosis. Although the morphometric analysis of white matter volume de®cit suggested involvement of certain longitudinal and interhemispheric tracts, the anatomical labelling of tracts on the basis of Talairach co-ordinates is heuristic. A more compelling demonstration that speci®c tracts are involved by both psychotic disorders, and that anatomical connectivity between frontal and temporoparietal cortex is compromised as a result, could be provided by future studies incorporating diffusion tensor imaging and tractography techniques.

Conclusion This study provides partial support for the seminal Kraepelinian dichotomy of psychosis because schizophrenia was characterized by a distinctive pattern of distributed grey matter de®cit in frontotemporal, subcortical and cerebellar regions, whereas psychotic bipolar disorder was not associated with signi®cant grey matter abnormality. However, the classic dichotomy is partially subverted by the demonstration of white matter abnormalities in common between the two disorders, suggesting that anatomical disconnectivity between frontal and temporoparietal cortex may be an important substrate for emergence of psychotic syndromes in general.

STUDY 3: STRUCTURAL BRAIN DEVIATIONS ASSOCIATED WITH GENETIC LIABILITY TO SCHIZOPHRENIA AND BIPOLAR DISORDER ASSESSED USING COMPUTATIONAL MORPHOMETRY If the Kraepelinian dichotomy of psychosis is correct, the neuroanatomical endophenotypes associated with genetic risks for schizophrenia and bipolar

178

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

disorder should also be distinct. Unlike the clinical phenotypes, endophenotypes are not subject to the impact of medication or other illness consequences that might vary between the two disorders. Most studies that have attempted to elucidate the impact of susceptibility genes associated with schizophrenia or bipolar disorder on brain anatomy have examined the degree of volumetric deviation compared to a control group of particular brain structures, as measured by region-of-interest, in participants considered likely to carry such susceptibility genes, such as unaffected ®rstdegree relatives, offspring or co-twins. The genetic liability (GL) scale described in Chapter 2 represents a quantitative measure of variable genetic risk, and computational morphometry represents a method of assessing volumetric change throughout the entire brain, rather than a small number of structures. In this study, these two powerful techniques were combined to examine variation of regional grey and white matter with quantitative variation of GL in multiply-affected families in order to identify more comprehensively the neuroanatomical endophenotypes of schizophrenia and bipolar disorder. As the GL scale is most reliably measured on participants from multiply-affected families, controls and participants with no family history of illness were not included in these analyses. The analysis was performed on those patients included in Study 2 of this chapter and their unaffected relatives. One-hundred-and-forty-eight participants were included, comprising twenty-®ve patients with schizophrenia, thirty-six of their unaffected relatives, thirty-seven patients with bipolar disorder and ®fty of their unaffected relatives. The patients with schizophrenia and their relatives were from twentyseven families and in each family the index patient had at least one other family member among their ®rst- and/or second-degree relatives affected with schizophrenia (twenty families), another functional psychotic disorder (three families) or schizotypal disorder (four families). Bipolar-disorder participants and their relatives were from thirty-two families and in each family the patient with bipolar disorder had at least one other family member among their ®rst- and/or second-degree relatives affected with psychotic bipolar disorder (twenty-four families) or another functional psychotic disorder (eight families).

MRI analysis The acquisition, pre-processing and analysis of structural MR images were as described in Study 2. Analyses were performed separately for the schizophrenia and bipolar families, including both patients and relatives, with GL score as the dependent variable and the co-varying for age, gender and affection status (i.e. patients versus relatives). The mass of each cluster produced by the analyses for each individual was transferred to a

7. STRUCTURAL BRAIN DEVIATIONS

179

spreadsheet and, where multiple clusters were present, PC analysis without rotation was performed to explore the extent of correlation between endophenotypic regions and to reduce the dimensionality of data for further analyses at systems level. In general, anatomical variation was strongly correlated between brain regions associated with genetic risk, i.e. the ®rst PC always accounted for more than 70% of total variance. Individual scores on the ®rst PC were therefore used as summary measures of anatomical variation in endophenotypic systems comprising two or more correlated grey or white matter regions associated with genetic risks for schizophrenia or bipolar disorder.

Multilevel modelling of MRI endophenotypes It was anticipated that variation in putative anatomical endophenotypes should be associated to the same extent with variable genetic risk in both patients and relatives, and that endophenotypic variation might be associated speci®cally with genetic risk for one type of psychosis or associated generically with genetic risks for both types of psychosis. To explore these issues, the association between anatomical variation in endophenotypic systems (as de®ned by PC scores) and genetic liability was modelled in different participant groups, employing the same hierarchical observation model as described in Study 1 for the region-of-interest analyses, i.e. robust variance estimators for clustered data to account for any intrafamilial correlation of measurements. The two-tailed probability threshold for signi®cance in these systems level analyses was set at p=0.05. The association between GL and related endophenotypic systems was ®rst explored separately for schizophrenia and bipolar groups of patients and relatives, to test the hypothesis that genetic risk was associated with endophenotypic variation in non-psychotic relatives as well as patients. An interaction term, patient status Ò GL score, was de®ned and entered into these clustered regression analyses to estimate whether the relationship between tissue deviation and GL score differed between patients and relatives. The associations between endophenotypes, de®ned by prior analysis of families with that disorder, and GL in non-psychotic relatives from both types of family, were then explored. The voxel densities under regional clusters identi®ed as endophenotypic for one disorder were extracted from the unaffected relatives of patients with the other disorder. An interaction term, relatives of schizophrenic patient status Ò GL score, was de®ned and entered into clustered regression analyses performed on the combined relatives group to estimate whether the relationship between tissue deviation and GL score differed between the two sets of relatives: disorderspeci®c endophenotypes would be associated with genetic risk only in nonpsychotic relatives of index patients with a diagnosis of that disorder;

180

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

whereas disorder-generic endophenotypes would be associated with genetic risk in non-psychotic relatives of patients with both types of disorder.

Results Grey matter Genetic risk for schizophrenia was associated with distributed grey matter volume de®cits in orbital, prefrontal and premotor parts of frontal cortex, caudate nucleus and thalamus bilaterally, and left insula and lateral temporal cortex. There were seven voxel clusters of regional grey matter volume de®cit; detailed description of the anatomical areas affected, including Brodmann areas where applicable, is given in Table 7.6. A ®gure demonstrating these regions (red voxels) is shown in Plate 6. PC analysis showed that these grey matter de®cits were highly correlated across regions, implying genetically determined effects on volume of a cortico-subcortical network. All regions of grey matter volume de®cit loaded positively on the ®rst PC (see Table 7.6 for factor loadings), which had an eigenvalue of 5.14 and explained 73.5% of the total variance in the group of patients with schizophrenia and their relatives. Scores on this ®rst PC were strongly associated with genetic risk in patients with schizophrenia and their nonpsychotic relatives in a clustered multiple regression analysis controlling for age, gender and affection status and there was no signi®cant interaction between participant group (patient versus relative) and genetic liability score (Table 7.7) indicating that this pattern of grey matter de®cit was not determined solely by abnormality in the patients. A scatterplot demonstrating the similar relationship between grey matter factor scores for each individual and genetic liability scores in patient and relative subgroups is shown in Plate 7(a). The relationship between increased genetic risk and grey matter volume de®cits in this cortico-subcortical system remained signi®cant when the analysis was con®ned to those twenty families where the patient's family history consisted speci®cally of schizophrenia (û=±1.49; 95% CI=±2.63, ±0.31; p=0.016). By contrast, genetic risk for bipolar disorder was associated with grey matter de®cit in an almost completely separate and relatively circumscribed set of regions, principally right anterior cingulate gyrus and ventral striatum. This region (blue voxels) is demonstrated in Plate 6. There was a single voxel cluster of regional grey matter volume de®cit and a detailed description of the anatomical areas affected, including Brodmann areas, is given in Table 7.6. Regional analysis con®rmed that genetic risk was associated with reduced grey matter volume of anterior cingulate and striatum in both patients with bipolar disorder and their relatives, and there was no signi®cant interaction between participant group (patient versus relative) and

White matter

Grey matter

Medial frontal gyrus, orbital gyrus, inferior/middle frontal gyri Inferior/middle frontal gyrus, precentral/postcentral gyrus Middle/superior frontal gyri Middle frontal gyrus Thalamus, anterior cingulate gyrus, caudate nucleus, brainstem Thalamus Superior/middle temporal gyri, transtemporal gyrus, precentral/postcentral gyrus, insula Lateral frontal lobe between middle and inferior frontal gyri and extending to the anterior insula and postcentral gyrus Temporal lobe between middle temporal gyrus and hippocampus/parahippocampal gyrus, extending to the superior temporal gyrus and posterior insula Splenium of corpus callosum

Schizophrenia families

Area1

± ±

L R and L

L

11/47 6/9/40/43/44/45 9/10/46 6/8/9 ±/25 ± 6/21/22/40/ 41/42/43 ±

Brodmann area

R R L L L and R R L

Side

515

645

633

402 602 423 333 474 284 933

Voxels in cluster

0.88

0.91

0.93

0.86 0.88 0.78 0.86 0.91 0.87 0.84

Loading on 1st PC2

TABLE 7.6 Anatomical location, approximate Brodmann areas, cluster size and loading scores on ®rst principal compenents (PC) analyses for endophenotypic regions of grey and white matter signi®cantly associated with genetic liability (GL) for schizophrenia and bipolar disorder

2

1

Medial frontal gyrus, anterior cingulate gyrus, caudate nucleus, anterior putamen Medial frontal lobe between the anterior cingulate/medial frontal gyri and middle frontal gyrus, extending into the genu of the corpus callosum Lateral frontal lobe between the inferior frontal gyrus, anterior insula and the caudate nucleus and anterior cingulate gyrus Temporal lobe between the superior/middle temporal gyri and hippocampus/parahippocampal gyrus, posterior cingulate gyrus Parietal lobe between the lateral ventricle, posterior cingulate gyrus, precuneus and the inferior parietal lobule, supramarginal/angular gyri and extending to the postcentral gyrus

9/11/24/25/ 32 ± ± ± ±

R L L R

R

1014

1041

1112

507

689

0.90

0.89

0.90

0.90

±

Anatomical localization of cluster extent and ascribed Brodmann areas were derived from the two-dimensional centroid voxels and spatial extent of the cluster in each axial slice. Principal components (PC) analysis was used to reduce the dimensionality of data for further analyses at systems level when more than one cluster was present. In each analysis there were strong positive loadings for every cluster on the ®rst PC.

White matter

Grey matter

Bipolar families

±1.85, 1.30 0.04, 1.27

±0.27 0.66

0.04

0.72

<0.001

p

0.009

0.53

0.004

p

0.49

0.11

±1.31



±0.30

±2.33

±1.47



±0.26, 1.25

±1.82, 2.03

±1.98, ±0.63

95% CI

±1.04, 0.44

±5.02, 0.35

±2.80, ±0.14

95% CI

p

0.20

0.92

<0.001

p

0.42

0.09

0.03

White matter endophenotypic systems (see Plate 8)

Overall G-P association results corroborate cluster-level mapping results at systems level; tests for difference in strength of G-P association between patients and relatives con®rm that anatomic variation in these systems is not associated with genetic liability only in patients; tests for difference in strength of G-P association between relatives of patients with schizophrenia and bipolar disorder indicate that grey matter endophenotypes are disorder speci®c, differentially associated with genetic liability for different types of psychosis, whereas white matter endophenotypes are disorder generic. Multiple linear regression analyses were performed with ®rst PC scores as dependent variables controlling for age, gender and participant group, using multilevel modelling to accommodate intra-familial correlation.

±1.93, ±0.77

±1.35

Overall G-P association (pooling patients with bipolar disorder and their relatives) Test for difference in strength of G-P association between patients with bipolar disorder and their relatives Test for difference in strength of G-P association between relatives of patients with schizophrenia and relatives of patients with bipolar disorder

95% CI

±1.65, ±0.25

±0.95



±3.39, 1.79

±0.80

Bipolar endophenotypes

±2.88, ±0.61

±1.75

Overall G-P association (pooling patients with schizophrenia and their relatives) Test for difference in strength of G-P association between patients with schizophrenia and their relatives Test for difference in strength of G-P association between relatives of patients with schizophrenia and relatives of patients with bipolar disorder

95% CI



Schizophrenia endophenotypes

Grey matter endophenotypic systems (see Plate 6)

TABLE 7.7 Genetic±phenotypic (G-P) associations between genetic liability scores and grey or white matter endophenotypic systems.

184

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

GL score (Table 7.7), again indicating that this association was not determined solely by an abnormality in the patients. A scatterplot demonstrating the similar relationship between grey matter cluster scores for each individual and GL scores in patient and relative subgroups is shown in Plate 7(b).

White matter There were also strong associations between genetic risk for each type of psychosis and anatomical variation in white matter. However, the white matter endophenotypes associated with genetic risk in the two disorders were overlapping, unlike their very different grey matter endophenotypes. Risk for schizophrenia was associated with white matter de®cits in posterior corpus callosum and left frontal and temporoparietal regions. Plate 8 demonstrates these regions (red voxels). There were three voxel clusters of regional white matter volume de®cit and a detailed description of the anatomical areas affected is given in Table 7.6. De®cits in these regions were highly correlated and all regions of white matter volume de®cit loaded positively on the ®rst PC (Table 7.6), which had an eigenvalue of 2.48 and explained 82.5% of the total variance in the group of patients with schizophrenia and their relatives. First PC scores were signi®cantly associated with GL in a clustered multiple regression analysis controlling for age, gender and group and the interaction between participant group (patient versus relative) and GL score was not statistically signi®cant (see Table 7.7). A scatterplot demonstrating the similar relationship between white matter factor scores for each individual and genetic liability scores in patient and relative subgroups is shown in Plate 7(c). Genetic risk for bipolar disorder was associated with white matter de®cits in anterior corpus callosum and bilateral frontal, left temporoparietal and right parietal regions. These regions are demonstrated in Plate 8 (blue voxels). There were four voxel clusters of regional white matter volume de®cit and a detailed description of the anatomical areas affected is given in Table 7.6. De®cits in these regions were highly correlated and all regions of white matter volume de®cit loaded positively on the ®rst PC (see Table 7.6), which had an eigenvalue of 3.23 and explained 80.7% of the total variance in the group of patients with bipolar disorder and their relatives. First PC scores were strongly associated with GL in a clustered multiple regression analysis controlling for age, gender and group and there was no signi®cant interaction between participant group (patient versus relative) and GL score (see Table 7.7). A scatterplot demonstrating the similar relationship between white matter factor scores for each individual and GL scores in patient and relative subgroups is shown in Plate 7(d).

7. STRUCTURAL BRAIN DEVIATIONS

185

Disorder specificity of grey and white matter endophenotypes Genetic risk for bipolar disorder was not signi®cantly associated with volume de®cit in the grey matter endophenotype for schizophrenia and there was a signi®cant interaction between the two relative groups (schizophrenia versus bipolar relatives) and GL upon PC scores [see Table 7.7; Plate 9(a)]. These results indicate that grey matter variation in this distributed frontostriatal and temporal system is an endophenotypic marker associated speci®cally with genetic risk for schizophrenia. Likewise, genetic risk for schizophrenia was not signi®cantly associated with volume de®cit in the grey matter endophenotype for bipolar disorder and there was a signi®cant interaction between the two relative groups (schizophrenia versus bipolar relatives) and GL upon PC scores [see Table 7.7, Plate 9(b)]. These results indicate that grey matter variation in this relatively circumscribed cingulate and striatal system is an endophenotypic marker associated speci®cally with genetic risk for bipolar disorder. However, GL for bipolar disorder was associated with anatomic de®cit in the white matter endophenotype de®ned by univariate analysis of the schizophrenia group and similarly GL for schizophrenia was associated with anatomic de®cit in the white matter endophenotype de®ned by analysis of the bipolar group, i.e. there were no signi®cant interactions between relatives groups and genetic loading on the systemic white matter endophenotype associated with either disorder, indicating that the relationship between increased genetic liability and white matter de®cit in these regions was present in both sets of relatives (see Table 7.7). A ®ner-grained analysis of genetic risk and endophenotypic association for each white matter cluster showed that GL score was associated generically with variation in the left hemispheric parts of both schizophrenia and bipolar endophenotypes, but that genetic risk for bipolar disorder was associated speci®cally with the right hemispheric parts of the bipolar white matter endophenotype. Plate 9(c) shows the similar relationship between reducing white matter volume in the left temporoparietal cluster identi®ed as endophenotypic for bipolar disorder and increasing GL for either disorder among unaffected relatives. Taken together, these results indicate that white matter de®cit in left frontal and temporoparietal regions is an endophenotypic marker associated generically with genetic risk for both schizophrenia and bipolar disorder.

Discussion Distinct grey matter endophenotypes These results provide partial support for the Kraepelinian dichotomy of psychosis in so far as there were markedly different grey matter phenotypes

186

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

associated with genetic risks for schizophrenia and for psychotic bipolar disorder. Genetic risk for schizophrenia was associated with a relatively extensive system of frontal, temporal and subcortical grey matter de®cit, comprising the caudate nucleus and thalamus, orbital, prefrontal and premotor parts of frontal cortex, and left insula and lateral temporal cortex. This is compatible both with regional structural de®cits which have been identi®ed by prior case±control studies of schizophrenia (Andreasen et al., 1994a; Honea et al., 2005; Sigmundsson et al., 2001; Wright et al., 2000), including Study 2 of this chapter, and with the clinical evidence that patients with schizophrenia suffer more severely than those with bipolar illness from impairments of language, social cognition and executive functions, which normally depend on integrity of these structures. The ®nding of reduced caudate and thalamus volume by computational morphometry is consistent with the lateral ventricular enlargement identi®ed by region-ofinterest analyses in Study 1, as these structures border the lateral ventricles. Indeed, the correlation between lateral ventricular volume and the summed voxel intensities underlying the endophenotypic grey matter cluster involving the thalamus and caudate in schizophrenia families is very high (r=0.62, p<0.001), indicating that the same individuals contribute to both these ®ndings. Volume de®cit in this grey matter system was associated with increasing genetic risk in non-psychotic relatives as well as patients with schizophrenia (con®rming that anatomic variation in this system is a marker for genetic risk rather than for caseness), but not signi®cantly associated with genetic risk among non-psychotic relatives of patients with bipolar disorder (indicating that this endophenotypic brain system represents genetic risk speci®cally for schizophrenia). Risk for bipolar disorder was associated with more local grey matter de®cits in right anterior cingulate cortex and ventral striatum, both of which are components of brain circuits for emotional processing (Rolls, 1999), and have been identi®ed as abnormal in previous case±control studies of bipolar disorder with structural and functional neuroimaging (Drevets et al., 1997; Hirayasu et al., 1999; Sassi et al., 2004). This study demonstrates that volume de®cit in these regions is a marker for genetic risk even among non-psychotic relatives, not merely a marker for the presence of bipolar disorder in patients; and that this endophenotypic brain system is indicative of genetic risk speci®cally for bipolar disorder. Studies examining unaffected relatives or discordant twins of patients with schizophrenia have previously linked genetic risk for schizophrenia to volumetric reduction of thalamus (Lawrie et al., 1999; McIntosh et al., 2004; Staal et al., 1998) and of temporal and prefrontal cortical grey matter (Cannon et al., 1998, 2002; Lawrie et al., 1999; McIntosh et al., 2004), especially the dorsolateral prefrontal cortex (Cannon et al., 2002). No evidence was found for grey matter reduction with genetic risk in a recent

7. STRUCTURAL BRAIN DEVIATIONS

187

twin study of bipolar disorder (Kieseppa et al., 2003). One other family study reported anterior thalamic de®cit associated with genetic risk for bipolar disorder (McIntosh et al., 2004) but failed to ®nd any signi®cant tissue de®cits in association with a quantitative scale of GL (McIntosh et al., 2006). However, this latter study reported prominent prefrontal grey matter de®cits in association with GL to schizophrenia ± in contrast to bipolar disorder and similar to the present ®ndings (McIntosh et al., 2006).

Common white matter endophenotypes The comparative design of this study also draws attention to aspects of the brain phenotype that are common between the two forms of psychosis. Genetic risk was associated with distributed white matter volume de®cit in both disorders, which overlapped in left frontal and temporoparietal regions. This ®nding echoes the results of Study 2, in which the two disorders demonstrated overlapping white matter de®cits in case±control analyses, and indicates that a likely source of such white matter de®cits resides in common susceptibility genes for both forms of psychosis. Interestingly, there is evidence from gene expression pro®ling of frontal cortical tissue for particular downregulation of genes related to myelination and oligodendrocyte function in schizophrenia and bipolar disorder (Hakak et al., 2001; Tkachev et al., 2003). Studies of discordant twins have reported a genetic impact on global white matter volume reduction in schizophrenia (Hulshoff Pol et al., 2004) and left hemispheric white matter volume reduction in bipolar disorder (Kieseppa et al., 2003), although other studies assessing unaffected relatives of patients with schizophrenia have failed to ®nd a genetic effect on global white matter volume (Cannon et al., 1998; Seidman et al., 1999; Staal et al., 2000) or regional white matter (McIntosh et al., 2005). However, the latter study did detect prefrontal and anterior internal capsule white matter de®cit in association with a similar quantitative measure of genetic liability in families affected by schizophrenia, but failed to detect any white matter de®cits in bipolar families (McIntosh et al., 2006). The current study maps the location of a generic white matter endophenotype for psychosis to territories normally occupied by major intrahemispheric tracts: the left superior longitudinal fasciculus, which connects the frontal lobe to the temporal, parietal and occipital lobes; and the left inferior longitudinal fasciculus, which connects the temporal pole to the occipital lobe. Risk for psychosis in general is therefore associated with a pattern of white matter de®cit that is likely to compromise intrahemispheric anatomical connectivity between left prefrontal and temporoparietal cortex. This is compatible with a substantial body of case±control data and theory implicating cortical dysconnectivity, and particularly frontotemporal dysconnectivity (Andreasen et al., 1998; Bullmore et al., 1997; Friston & Frith,

188

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

1995; Weinberger et al., 1994), as a critical substrate for generation of psychotic symptoms. This phenotypic overlap in relation to white matter de®cit is consistent with common susceptibility genes for psychosis, but is also compatible with a ®nal common pathway for separable genetic effects. The tendency of psychotic symptoms to aggregate in bipolar pedigrees (Potash et al., 2001; Schurhoff et al., 2003) has led to the suggestion that psychotic bipolar disorder represents a genetic subset of bipolar disorder that might share psychosis susceptibility genes with schizophrenia (Potash et al., 2001). The current study assessed precisely such bipolar families and suggests that susceptibility genes underlying both schizophrenia and psychotic bipolar disorders are associated with frontotemporal white matter de®cit. This study therefore provides suggestive evidence that candidate genes common to both forms of psychosis will include those involved in myelination and oligodendrocyte function. The demonstration that patients shared volume de®cits in these grey and white matter systems with unaffected relatives likely to be carrying susceptibility genes, but who had never experienced psychotic illness and were therefore free from any effects of disease progression and medication on brain morphology, supports the validity of these neuroanatomical endophenotypes as genetic trait markers. Treating genetic risk as a continuous variable among non-psychotic relatives, rather than assuming that all relatives shared the same level of risk, is a more realistic assumption in light of the probable variation between families in their exposure to multiple susceptibility genes. This may have conferred greater statistical power to detect brain endophenotypes on this regression analysis of anatomical variation and continuous GL scores than would have been attained by treating patients and relatives as two discrete levels of genetic risk as in Study 1.

Conclusion The major implication from this study is that genetic risks for schizophrenia and bipolar disorder are associated with both speci®c (grey matter) and generic (white matter) brain structural endophenotypes. The anatomically segregated expression of speci®c and generic genetic effects is consistent with morphometric deviations linked to the clinical phenotypes of schizophrenia and bipolar disorder.

SUMMARY The morphometric studies emanating from the Maudsley Family Study of Psychosis are large, in¯uential and utilize complemetary methodologies to enhance their validity. These studies demonstrate that the brain structural deviations and endophenotypes characterizing schizophrenia and bipolar

+

Controls Patients Relatives

FZ

2 V CZ

100 ms

PZ

– –400

–200

0

200

400

600

800

1000

ms

Plate 1 P300 group average waves from Phases 2 and 3 of the Maudsley Family Study of Psychosis. (Reprinted from Schizophrenia Research, 67, Bramon et al. Mismatch negativity in schizophrenia: A family study, pp. 1±10. Copyright (2004), with permission from Elsevier.)

Plate 2 Example of Stroop colour±word stimuli. Response inhibition is measured by one's ability to name the colour (rather than reading the word) of congruent and incongruent colour and word stimuli.

Plate 3 Map of grey matter volume de®cits when comparing patients with schizophrenia to healthy comparison participants, superimposed onto a single brain in standard stereotactic space. The brain slices are orientated in the plane of the Talairach atlas; distance (mm) above or below the intercommissural line is inset in the left corner of each slice; the right side of the brain is depicted by the right side of each slice. Cluster-wise probability of false-positive activation p=0.003; expected number of false positive tests <1 over the whole map. (Figure reproduced, with permission of the Royal College of Psychiatrists, from McDonald et al., 2005.)

Plate 4 Map of grey matter volume de®cits when comparing patients with schizophrenia to patients with bipolar disorder, superimposed onto a single brain in standard stereotactic space. The brain slices are orientated in the plane of the Talairach atlas; distance (mm) above or below the intercommissural line is inset in the left corner of each slice; the right side of the brain is depicted by the right side of each slice. Cluster-wise probability of false-positive activation p=0.002; expected number of false-positive tests <1 over the whole map. (Figure reproduced, with permission of the Royal College of Psychiatrists, from McDonald et al., 2005.)

Plate 5 Map of white matter volume de®cits when comparing patients with schizophrenia to healthy comparison participants (red voxels) and patients with bipolar disorder to healthy comparison participants (blue voxels), superimposed onto a single brain in standard stereotactic space. Green voxels indicate overlapping regions. The brain slices are orientated in the plane of the Talairach atlas; distance (mm) above or below the intercommissural line is inset in the left corner of each slice; the right side of the brain is depicted by the right side of each slice. Cluster-wise probability of false-positive activation p=0.011 for schizophrenia vs. control analysis and p=0.008 for bipolar disorder vs. control analysis; expected number of false-positive tests <1 for each analysis. (Figure reproduced, with permission of the Royal College of Psychiatrists, from McDonald et al., 2005.)

Plate 6 Grey matter endophenotypes. Map of grey matter volume de®cits associated with increasing genetic risks for schizophrenia (red voxels) and bipolar disorder (blue voxels) superimposed onto a single brain in standard stereotactic space. Green voxels indicate overlapping regions. Cluster-wise probability of false-positive activation p0.01 for both families with schizophrenia and bipolar families; expected number of false-positive tests <1 for each analysis. The brain slices are orientated in the plane of the Talairach atlas; distance (mm) above or below the intercommissural line is inset in the left corner of each slice; the right side of the brain is depicted by the right side of each axial slice. (Figure reproduced with permission from McDonald et al., Archives of General Psychiatry, 61(10), 974984. Copyright Ø (2004), American Medical Association. All rights reserved.

(a)

3

2

1.4

Grey matter cluster value

Grey matter factor score

(b)

1.6

1

0

–1

1.2

1.0

0.8

–2

0.6

–3 0.4

0.6

0.8

1.0

1.2

1.4

0.4

0.6

Genetic loading score (adjusted)

1.2

1.4

1.6

1.4

1.6

(d)

3

2

White matter factor score

2

White matter factor score

1.0

Genetic loading score (adjusted)

(c)

3

0.8

1

0

–1

–2

1

0

–1

–2

–3

–3

–4 0.4

0.6

0.8

1.0

Genetic loading score (adjusted)

1.2

1.4

0.4

0.6

0.8

1.0

1.2

Genetic loading score (adjusted)

Plate 7 Figures demonstrating similar linear associations between systemic tissue volume de®cits and genetic risk estimated separately for patients (full squares) and their nonpsychotic relatives (open triangles) for schizophrenia (red) or bipolar disorder (blue). (a) Grey matter in schizophrenia; (b) grey matter in bipolar disorder; (c) white matter in schizophrenia; (d) white matter in bipolar disorder. The ®rst principal components (PC) analysis factor scores summarize correlated tissue de®cit over several brain regions for each individual. Genetic liability (GL) scores are adjusted to the sample mean for age, gender and affection status.

Plate 8 White matter endophenotypes. Map of white matter volume de®cits associated with genetic risks for schizophrenia (red voxels) and bipolar disorder (blue voxels) superimposed onto a single brain in standard stereotactic space. Green voxels indicate overlapping regions. Cluster-wise probability of false-positive activation p0.01 for both schizophrenic families and bipolar families; expected number of false positive tests <1 for each analysis. The brain slices are orientated in the plane of the Talairach atlas; distance (mm) above or below the intercommissural line is inset in the left corner of each slice; the right side of the brain is depicted by the right side of each axial slice. (Figure reproduced with permission from McDonald et al., Archives of General Psychiatry, 61(10), 974±984. Copyright Ø (2004), American Medical Association. All rights reserved.

(a)

(b) 1.6

Relatives of patients with schizophrenia Relatives of patients with bipolar disorder

Bipolar disorder grey matter endophenotype

Schizoophrenia grey matter endophenotype

4

3

2

1

0

–1

–2 –3 0.4

0.6

0.8

1.0

1.2

1.4

Relatives of patients with schizophrenia Relatives of patients with bipolar disorder

1.4

1.2

1.0

.8

.6

1.6

0.4

0.6

0.8

Genetic liability score

1.0

1.2

1.4

1.6

Genetic liability score

(c) Bipolar disorder white matter endophenotype

5.0

Relatives of patients with schizophrenia Relatives of patients with bipolar disorder

4.5

4.0

3.5

3.0

2.5

2.0 1.5 0.4

0.6

0.8

1.0

1.2

1.4

1.6

Genetic liability score

Plate 9 Linear associations, demonstrating disorder-speci®c grey matter endophenotypes and a disorder-generic white matter endophenotype, between systemic tissue de®cits de®ned as endophenotypic for schizophrenia or bipolar disorder and genetic liability (GL) scores estimated separately for non-psychotic relatives of schizophrenia patients (red) and nonpsychotic relatives of bipolar patients (blue). (a) Grey matter de®cits endophenotypic for schizophrenia; (b) grey matter de®cits endophenotypic for bipolar disorder; (c) left temporoparietal white matter de®cit endophenotypic for both disorders. Grey matter endophenotypes are represented by ®rst principal components (PC) analysis factor scores, which summarize correlated grey matter de®cit over distributed brain regions for each individual. GL scores are adjusted to the sample mean for age, gender and affection status.

7. STRUCTURAL BRAIN DEVIATIONS

189

disorder are distinct but overlapping. Schizophrenia is distinctly associated with lateral ventricular enlargement, hippocampal de®cit, increased CSF and grey matter de®cit in frontotemporal cortex, medial temporal lobe, insula, thalamus, parietal and cerebellar regions, whereas bipolar disorder is associated with none of these ®ndings. Both types of psychosis are characterized by distributed white matter de®cit involving major inter- and intrahemispheric tracts, consistent with the concept that psychosis in general is characterized by anatomical dysconnectivity. Unaffected relatives of patients with schizophrenia at highest likelihood of carrying susceptibility genes for illness also demonstrate lateral ventricular enlargement and grey matter de®cit in frontal, temporal and subcortical regions, whereas GL for bipolar disorder is linked to grey matter de®cit in regions known to subserve emotional regulation, the right anterior cingulate and ventral striatum ± indicating distinct grey matter endophenotypes for the two illnesses. Echoing the case±control ®ndings through an entirely separate approach, GL for each disorder is linked to white matter de®cit in left frontal and temporoparietal regions, supporting a generic white matter endophenotype for psychosis in general and suggesting that genes common to the major psychotic illnesses may be associated with myelination and oligodendrocyte function. As these morphometric deviations within grey and white matter were present in those unaffected relatives at high GL as well as in patients, they cannot be ascribed to medication or any other illness effects. The ®ndings from these studies partially support the seminal Kraepelinian dichotomy of psychosis and suggest that common and speci®c susceptibility genes for psychosis may be driving the overlapping and distinct morphometric deviations in schizophrenia and bipolar disorder.

REFERENCES Adler, C.M., Holland, S.K., Schmithorst, V., Wilke, M., Weiss, K.L., Pan, H., et al. (2004). Abnormal frontal white matter tracts in bipolar disorder: A diffusion tensor imaging study. Bipolar Disorders, 6, 197±203. Altshuler, L.L., Curran, J.G., Hauser, P., Mintz, J., Denicoff, K., Post, R. (1995). T2 hyperintensities in bipolar disorder: Magnetic resonance imaging comparison and literature metaanalysis. American Journal of Psychiatry, 152, 1139±1144. Altshuler, L.L., Bartzokis, G., Grieder, T., Curran, J., Jimenez, T., Leight, K., et al. (2000). An MRI study of temporal lobe structures in men with bipolar disorder or schizophrenia. Biological Psychiatry, 48, 147±162. Ananth, H., Popescu, I., Critchley, H.D., Good, C.D., Frackowiak, R.S., Dolan, R.J. (2002). Cortical and subcortical gray matter abnormalities in schizophrenia determined through structural magnetic resonance imaging with optimized volumetric voxel-based morphometry. American Journal of Psychiatry, 159, 1497±1505. Andreasen, N.C., Arndt, S., Swayze, V. 2nd, Cizadlo, T., Flaum, M., O'Leary, D., et al. (1994a). Thalamic abnormalities in schizophrenia visualized through magnetic resonance image averaging. Science, 266, 294±298. Andreasen, N.C., Flashman, L., Flaum, M., Arndt, S., Swayze, V. 2nd, O'Leary, D.S., et al.

190

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

(1994b). Regional brain abnormalities in schizophrenia measured with magnetic resonance imaging. Journal of the American Medical Association, 272, 1763±1769. Andreasen, N.C., Paradiso, S., O'Leary, D.S. (1998). `Cognitive dysmetria' as an integrative theory of schizophrenia: A dysfunction in cortical±subcortical±cerebellar circuitry? Schizophrenia Bulletin, 24, 203±218. Ashburner, J., Friston, K.J. (2000). Voxel-based morphometry ± the methods. Neuroimage, 11, 805±821. Baare, W.F., Hulshoff Pol, H.E., Boomsma, D.I., Posthuma, D., de Geus, E.J., Schnack, H.G., et al. (2001a). Quantitative genetic modeling of variation in human brain morphology. Cerebral Cortex, 11, 816±824. Baare, W.F., van Oel, C.J., Hulshoff Pol, H.E., Schnack, H.G., Durston, S., Sitskoorn, M.M., et al. (2001b). Volumes of brain structures in twins discordant for schizophrenia. Archives of General Psychiatry, 58, 33±40. Bagary, M.S., Symms, M.R., Barker, G.J., Mutsatsa, S.H., Joyce, E.M., Ron, M.A. (2003). Gray and white matter brain abnormalities in ®rst-episode schizophrenia inferred from magnetization transfer imaging. Archives of General Psychiatry, 60, 779±788. Barta, P.E., Dhingra, L., Royall, R., Schwartz, E. (1997). Improving stereological estimates for the volume of the structures identi®ed in three-dimensional arrays of spatial data. Journal of Neuroscience Methods, 75, 111±118. Bearden, C.E., Hoffman, K.M., Cannon, T.D. (2001). The neuropsychology and neuroanatomy of bipolar disorder: A critical review. Bipolar Disorders, 3, 106±150. Beyer, J.L., Taylor, W.D., MacFall, J.R., Kuchibhatla, M., Payne, M.E., Provenzale, J.M., et al. (2005). Cortical white matter microstructural abnormalities in bipolar disorder. Neuropsychopharmacology, 30, 2225±2229. Buchsbaum, M.S., Friedman, J., Buchsbaum, B.R., Chu, K.W., Hazlett, E.A., Newmark, R., et al. (2006). Diffusion tensor imaging in schizophrenia. Biological Psychiatry, 60, 1181±1187. Bullmore, E.T., Frangou, S., Murray, R.M. (1997). The dysplastic net hypothesis: An integration of developmental and dysconnectivity theories of schizophrenia. Schizophrenia Research, 28, 143±156. Bullmore, E.T., Suckling, J., Overmeyer, S., Rabe-Hesketh, S., Taylor, E., Brammer, M.J. (1999). Global, voxel, and cluster tests, by theory and permutation, for a difference between two groups of structural MR images of the brain. IEEE Transactions on Medical Imaging, 18, 32±42. Campbell, S., Marriott, M., Nahmias, C., MacQueen, G.M. (2004). Lower hippocampal volume in patients suffering from depression: A meta-analysis. American Journal of Psychiatry, 161, 598±607. Cannon, T.D., Mednick, S.A., Parnas, J., Schulsinger, F., Praestholm, J., Vestergaard, A. (1993). Developmental brain abnormalities in the offspring of schizophrenic mothers. I. Contributions of genetic and perinatal factors. Archives of General Psychiatry, 50, 551±564. Cannon, T.D., van Erp, T.G.M., Huttunen, M., Lonnqvist, J., Salonen, O., Valanne, L., et al. (1998). Regional gray matter, white matter, and cerebrospinal ¯uid distributions in schizophrenic patients, their siblings, and controls. Archives of General Psychiatry, 55, 1084±1091. Cannon, T.D., Thompson, P.M., van Erp, T.G., Toga, A.W., Poutanen, V.P., Huttunen, M., et al. (2002). Cortex mapping reveals regionally speci®c patterns of genetic and diseasespeci®c gray-matter de®cits in twins discordant for schizophrenia. Proceedings of the National Academy of Sciences USA, 99, 3228±3233. Davis, K.L., Stewart, D.G., Friedman, J.I., Buchsbaum, M., Harvey, P.D., Hof, P.R., et al. (2003). White matter changes in schizophrenia: Evidence for myelin-related dysfunction. Archives of General Psychiatry, 60, 443±456.

7. STRUCTURAL BRAIN DEVIATIONS

191

Doherty, C.P., Fitzsimons, M., Holohan, T., Mohamed, H.B., Farrell, M., Meredith, G.E., et al. (2000). Accuracy and validity of stereology as a quantitative method for assessment of human temporal lobe volumes acquired by magnetic resonance imaging. Magnetic Resonance Imaging, 18, 1017±1025. Downhill, J.E. Jr., Buchsbaum, M.S., Wei, T., Spiegel-Cohen, J., Hazlett, E.A., Haznedar, M.M., et al. (2000). Shape and size of the corpus callosum in schizophrenia and schizotypal personality disorder. Schizophrenia Research, 42, 193±208. Drevets, W.C., Price, J.L., Simpson, J.R. Jr., Todd, R.D., Reich, T., Vannier, M., et al. (1997). Subgenual prefrontal cortex abnormalities in mood disorders. Nature, 386, 824±827. Elkis, H., Friedman, L., Wise, A., Meltzer, H.Y. (1995). Meta-analyses of studies of ventricular enlargement and cortical sulcal prominence in mood disorders. Comparisons with controls or patients with schizophrenia. Archives of General Psychiatry, 52, 735±746. Friedman, L., Findling, R.L., Kenny, J.T., Swales, T.P., Stuve, T.A., Jesberger, J.A., et al. (1999). An MRI study of adolescent patients with either schizophrenia or bipolar disorder as compared to healthy control subjects. Biological Psychiatry, 46, 78±88. Friston, K.J., Frith, C.D. (1995). Schizophrenia: A disconnection syndrome? Clinical Neuroscience, 3, 89±97. Gilvarry, C., Takei, N., Russell, A., Rushe, T., Hemsley, D., Murray, R.M. (2000). Premorbid IQ in patients with functional psychosis and their ®rst-degree relatives. Schizophrenia Research, 41, 417±429. Goldstein, J.M., Goodman, J.M., Seidman, L.J., Kennedy, D.N., Makris, N., Lee, H., et al. (1999). Cortical abnormalities in schizophrenia identi®ed by structural magnetic resonance imaging. Archives of General Psychiatry, 56, 537±547. Good, C.D., Johnsrude, I.S., Ashburner, J., Henson, R.N., Friston, K.J., Frackowiak, R.S. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage, 14, 21±36. Hakak, Y., Walker, J.R., Li, C., Wong, W.H., Davis, K.L., Buxbaum, J.D., et al. (2001). Genome-wide expression analysis reveals dysregulation of myelination-related genes in chronic schizophrenia. Proceedings of the National Academy of Sciences USA, 98, 4746±4751. Harris, J.G., Young, D.A., Rojas, D.C., Cajade-Law, A., Scherzinger, A., Nawroz, S., et al. (2002). Increased hippocampal volume in schizophrenics' parents with ancestral history of schizophrenia. Schizophrenia Research, 55, 11±17. Hayasaka, S., Nichols, T.E. (2003). Validating cluster size inference: Random ®eld and permutation methods. Neuroimage, 20, 2343±2356. Haznedar, M.M., Roversi, F., Pallanti, S., Baldini-Rossi, N., Schnur, D.B., Licalzi, E.M., et al. (2005). Fronto-thalamo-striatal gray and white matter volumes and anisotropy of their connections in bipolar spectrum illnesses. Biological Psychiatry, 57, 733±742. Hirayasu, Y., Shenton, M.E., Salisbury, D.F., Dickey, C.C., Fischer, I.A., Mazzoni, P., et al. (1998). Lower left temporal lobe MRI volumes in patients with ®rst-episode schizophrenia compared with psychotic patients with ®rst-episode affective disorder and normal subjects. American Journal of Psychiatry, 155, 1384±1391. Hirayasu, Y., Shenton, M.E., Salisbury, D.F., Kwon, J.S., Wible, C.G., Fischer, I.A., et al. (1999). Subgenual cingulate cortex volume in ®rst-episode psychosis. American Journal of Psychiatry, 156, 1091±1093. Hirayasu, Y., Tanaka, S., Shenton, M.E., Salisbury, D.F., DeSantis, M.A., Levitt, J.J., et al. (2001). Prefrontal gray matter volume reduction in ®rst episode schizophrenia. Cerebral Cortex, 11, 374±381. Hoge, E.A., Friedman, L., Schulz, S.C. (1999). Meta-analysis of brain size in bipolar disorder. Schizophrenia Research, 37, 177±181. Honea, R., Crow, T.J., Passingham, D., Mackay, C.E. (2005). Regional de®cits in brain

192

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

volume in schizophrenia: A meta-analysis of voxel-based morphometry studies. American Journal of Psychiatry, 162, 2233±2245. Hulshoff Pol, H.E., Schnack, H.G., Mandl, R.C., van Haren, N.E., Koning, H., Collins, D.L., et al. (2001). Focal gray matter density changes in schizophrenia. Archives of General Psychiatry, 58, 1118±1125. Hulshoff Pol, H.E., Brans, R.G., van Haren, N.E., Schnack, H.G., Langen, M., Baare, W.F., et al. (2004). Gray and white matter volume abnormalities in monozygotic and same-gender dizygotic twins discordant for schizophrenia. Biological Psychiatry, 55, 126±130. Kasai, K., Shenton, M.E., Salisbury, D.F., Onitsuka, T., Toner, S.K., Yurgelun-Todd, D., et al. (2003). Differences and similarities in insular and temporal pole MRI gray matter volume abnormalities in ®rst-episode schizophrenia and affective psychosis. Archives of General Psychiatry, 60, 1069±1077. Kieseppa, T., van Erp, T.G., Haukka, J., Partonen, T., Cannon, T.D., Poutanen, V.P., et al. (2003). Reduced left hemispheric white matter volume in twins with bipolar I disorder. Biological Psychiatry, 54, 896±905. Kremen, W.S., Faraone, S.V., Seidman, L.J., Pepple, J.R., Tsuang, M.T. (1998). Neuropsychological risk indicators for schizophrenia: a preliminary study of female relatives of schizophrenic and bipolar probands. Psychiatry Research, 79, 227±240. Kubicki, M., Shenton, M.E., Salisbury, D.F., Hirayasu, Y., Kasai, K., Kikinis, R., et al. (2002a). Voxel-based morphometric analysis of gray matter in ®rst episode schizophrenia. Neuroimage, 17, 1711±1719. Kubicki, M., Westin, C.F., Maier, S.E., Frumin, M., Nestor, P.G., Salisbury, D.F., et al. (2002b). Uncinate fasciculus ®ndings in schizophrenia: A magnetic resonance diffusion tensor imaging study. American Journal of Psychiatry, 159, 813±820. Kuperberg, G.R., Broome, M.R., McGuire, P.K., David, A.S., Eddy, M., Ozawa, F., et al. (2003). Regionally localized thinning of the cerebral cortex in schizophrenia. Archives of General Psychiatry, 60, 878±888. Lawrie, S.M., Whalley, H., Kestelman, J.N., Abukmeil, S.S., Byrne, M., Hodges, A., et al. (1999). Magnetic resonance imaging of brain in people at high risk of developing schizophrenia. Lancet, 353, 30±33. Lawrie, S.M., Whalley, H.C., Abukmeil, S.S., Kestelman, J.N., Donnelly, L., Miller, P., et al. (2001). Brain structure, genetic liability, and psychotic symptoms in subjects at high risk of developing schizophrenia. Biological Psychiatry, 49, 811±823. Lim, K.O., Hedehus, M., Moseley, M., de Crespigny, A., Sullivan, E.V., Pfefferbaum, A. (1999a). Compromised white matter tract integrity in schizophrenia inferred from diffusion tensor imaging. Archives of General Psychiatry, 56, 367±374. Lim, K.O., Rosenbloom, M.J., Faustman, W.O., Sullivan, E.V., Pfefferbaum, A. (1999b). Cortical gray matter de®cit in patients with bipolar disorder. Schizophrenia Research, 40, 219±227. Lochhead, R.A., Parsey, R.V., Oquendo, M.A., Mann, J.J. (2004). Regional brain gray matter volume differences in patients with bipolar disorder as assessed by optimized voxel-based morphometry. Biological Psychiatry, 55, 1154±1162. Lopez-Larson, M.P., DelBello, M.P., Zimmerman, M.E., Schwiers, M.L., Strakowski, S.M. (2002). Regional prefrontal gray and white matter abnormalities in bipolar disorder. Biological Psychiatry, 52, 93±100. Matsumoto, H., Simmons, A., Williams, S., Pipe, R., Murray, R., Frangou, S. (2001). Structural magnetic imaging of the hippocampus in early onset schizophrenia. Biological Psychiatry, 49, 824±831. McCarley, R.W., Wible, C.G., Frumin, M., Hirayasu, Y., Levitt, J.J., Fischer, I.A., et al. (1999). MRI anatomy of schizophrenia. Biological Psychiatry, 45, 1099±1119. McDonald, C., Grech, A., Toulopoulou, T., Schulze, K., Chapple, B., Sham, P.C., et al.

7. STRUCTURAL BRAIN DEVIATIONS

193

(2002). Brain volumes in familial and non-familial schizophrenic probands and their unaffected relatives. American Journal of Medical Genetics, 114, 616±625. McDonald, C., Bullmore, E.T., Sham, P.C., Chitnis, X., Wickham, H., Bramon, E., et al. (2004a). Association of genetic risks for schizophrenia and bipolar disorder with speci®c and generic brain structural endophenotypes. Archives of General Psychiatry, 61, 974±984. McDonald, C., Zanelli, J., Rabe-Hesketh, S., Ellison-Wright, I., Sham, P., Kalidindi, S., et al. (2004b). Meta-analysis of magnetic resonance imaging brain morphometry studies in bipolar disorder. Biological Psychiatry, 56, 411±417. McDonald, C., Bullmore, E., Sham, P., Chitnis, X., Suckling, J., Maccabe, J., et al. (2005). Regional volume deviations of brain structure in schizophrenia and psychotic bipolar disorder: Computational morphometry study. British Journal of Psychiatry, 186, 369±377. McDonald, C., Marshall, N., Sham, P.C., Bullmore, E.T., Schulze, K., Chapple, B., et al. (2006). Regional brain morphometry in patients with schizophrenia or bipolar disorder and their unaffected relatives. American Journal of Psychiatry, 163, 478±487. McIntosh, A.M., Job, D.E., Moorhead, T.W., Harrison, L.K., Forrester, K., Lawrie, S.M., et al. (2004). Voxel-based morphometry of patients with schizophrenia or bipolar disorder and their unaffected relatives. Biological Psychiatry, 56, 544±552. McIntosh, A.M., Job, D.E., Moorhead, T.W., Harrison, L.K., Lawrie, S.M., Johnstone, E.C. (2005). White matter density in patients with schizophrenia, bipolar disorder and their unaffected relatives. Biological Psychiatry, 58, 254±257. McIntosh, A.M., Job, D.E., Moorhead, W.J., Harrison, L.K., Whalley, H.C., Johnstone, E.C., et al. (2006). Genetic liability to schizophrenia or bipolar disorder and its relationship to brain structure. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics, 141, 76±83. McNeil, T.F., Cantor-Graae, E., Weinberger, D.R. (2000). Relationship of obstetric complications and differences in size of brain structures in monozygotic twin pairs discordant for schizophrenia. American Journal of Psychiatry, 157, 203±212. Moore, G.J., Bebchuk, J.M., Wilds, I.B., Chen, G., Manji, H.K. (2000). Lithium-induced increase in human brain grey matter. Lancet, 356, 1241±1242. Nelson, M.D., Saykin, A.J., Flashman, L.A., Riordan, H.J. (1998). Hippocampal volume reduction in schizophrenia as assessed by magnetic resonance imaging ± A meta-analytic study. Archives of General Psychiatry, 55, 433±440. Nosarti, C., Al-Asady, M.H., Frangou, S., Stewart, A.L., Rifkin, L., Murray, R.M. (2002). Adolescents who were born very preterm have decreased brain volumes. Brain, 125, 1616±1623. Nugent, A.C., Milham, M.P., Bain, E.E., Mah, L., Cannon, D.M., Marrett, S., et al. (2006). Cortical abnormalities in bipolar disorder investigated with MRI and voxel-based morphometry. Neuroimage, 30, 485±497. Pearlson, G.D., Barta, P.E., Powers, R.E., Menon, R.R., Richards, S.S., Aylward, E.H., et al. (1997). Ziskind±Somerfeld Research Award 1996. Medial and superior temporal gyral volumes and cerebral asymmetry in schizophrenia versus bipolar disorder. Biological Psychiatry, 41, 1±14. Pfefferbaum, A., Sullivan, E.V., Swan, G.E., Carmelli, D. (2000). Brain structure in men remains highly heritable in the seventh and eighth decades of life. Neurobiology of Aging, 21, 63±74. Potash, J.B., Willour, V.L., Chiu, Y.F., Simpson, S.G., MacKinnon, D.F., Pearlson, G.D., et al. (2001). The familial aggregation of psychotic symptoms in bipolar disorder pedigrees. American Journal of Psychiatry, 158, 1258±1264. Reveley, A.M., Reveley, M.A., Chitkara, B., Clifford, C. (1984). The genetic basis of cerebral ventricular volume. Psychiatry Research, 13, 261±266. Rolls, E.T. (1999). The Brain and Emotion. Oxford: Oxford University Press.

194

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Sapolsky, R.M. (2000). Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders. Archives of General Psychiatry, 57, 925±935. Sassi, R.B., Nicoletti, M., Brambilla, P., Mallinger, A.G., Frank, E., Kupfer, D.J., et al. (2002). Increased gray matter volume in lithium-treated bipolar disorder patients. Neuroscience Letters, 329, 243±245. Sassi, R.B., Brambilla, P., Hatch, J.P., Nicoletti, M.A., Mallinger, A.G., Frank, E., et al. (2004). Reduced left anterior cingulate volumes in untreated bipolar patients. Biological Psychiatry, 56, 467±475. Schulze, K., McDonald, C., Frangou, S., Sham, P., Grech, A., Toulopoulou, T., et al. (2003). Hippocampal volume in familial and nonfamilial schizophrenic probands and their unaffected relatives. Biological Psychiatry, 53, 562±570. Schurhoff, F., Szoke, A., Meary, A., Bellivier, F., Rouillon, F., Pauls, D., et al. (2003). Familial aggregation of delusional proneness in schizophrenia and bipolar pedigrees. American Journal of Psychiatry, 160, 1313±1319. Seidman, L.J., Faraone, S.V., Goldstein, J.M., Goodman, J.M., Kremen, W.S., Toomey, R., et al. (1999). Thalamic and amygdala-hippocampal volume reductions in ®rst-degree relatives of patients with schizophrenia: An MRI-based morphometric analysis. Biological Psychiatry, 46, 941±954. Seidman, L.J., Faraone, S.V., Goldstein, J.M., Kremen, W.S., Horton, N.J., Makris, N., et al. (2002). Left hippocampal volume as a vulnerability indicator for schizophrenia: A magnetic resonance imaging morphometric study of nonpsychotic ®rst-degree relatives. Archives of General Psychiatry, 59, 839±849. Shapleske, J., Rossell, S.L., Chitnis, X.A., Suckling, J., Simmons, A., Bullmore, E.T., et al. (2002). A computational morphometric MRI study of schizophrenia: Effects of hallucinations. Cerebral Cortex, 12, 1331±1341. Sharma, T., Lancaster, E., Lee, D., Lewis, S., Sigmundsson, T., Takei, N., et al. (1998). Brain changes in schizophrenia. Volumetric MRI study of families multiply affected with schizophrenia ± the Maudsley Family Study 5. British Journal of Psychiatry, 173, 132±138. Sheline, Y.I., Wang, P.W., Gado, M.H., Csernansky, J.G., Vannier, M.W. (1996). Hippocampal atrophy in recurrent major depression. Proceedings of the National Academy of Sciences USA, 93, 3908±3913. Shenton, M.E., Dickey, C.C., Frumin, M., McCarley, R.W. (2001). A review of MRI ®ndings in schizophrenia. Schizophrenia Research, 49, 1±52. Sigmundsson, T., Suckling, J., Maier, M., Williams, S., Bullmore, E., Greenwood, K., et al. (2001). Structural abnormalities in frontal, temporal, and limbic regions and interconnecting white matter tracts in schizophrenic patients with prominent negative symptoms. American Journal of Psychiatry, 158, 234±243. Soares, J.C., Mann, J.J. (1997). The anatomy of mood disorders ± review of structural neuroimaging studies. Biological Psychiatry, 41, 86±106. Spalletta, G., Tomaiuolo, F., Marino, V., Bonaviri, G., Trequattrini, A., Caltagirone, C. (2003). Chronic schizophrenia as a brain misconnection syndrome: A white matter voxelbased morphometry study. Schizophrenia Research, 64, 15±23. Staal, W.G., Hulshoff, H.E., Schnack, H., van der Schot, A.C., Kahn, R.S. (1998). Partial volume decrease of the thalamus in relatives of patients with schizophrenia. American Journal of Psychiatry, 155, 1784±1786. Staal, W.G., Pol, H.E.H., Schnack, H.G., Hoogendoorn, M.L.C., Jellema, K., Kahn, R.S. (2000). Structural brain abnormalities in patients with schizophrenia and their healthy siblings. American Journal of Psychiatry, 157, 416±421. Strakowski, S.M., DelBello, M.P., Sax, K.W., Zimmerman, M.E., Shear, P.K., Hawkins, J.M., et al. (1999). Brain magnetic resonance imaging of structural abnormalities in bipolar disorder. Archives of General Psychiatry, 56, 254±260.

7. STRUCTURAL BRAIN DEVIATIONS

195

Strasser, H.C., Lilyestrom, J., Ashby, E.R., Honeycutt, N.A., Schretlen, D.J., Pulver, A.E., et al. (2005). Hippocampal and ventricular volumes in psychotic and nonpsychotic bipolar patients compared with schizophrenia patients and community control subjects: A pilot study. Biological Psychiatry, 57, 633±639. Sullivan, E.V., Pfefferbaum, A., Swan, G.E., Carmelli, D. (2001). Heritability of hippocampal size in elderly twin men: Equivalent in¯uence from genes and environment. Hippocampus, 11, 754±762. Suzuki, M., Nohara, S., Hagino, H., Kurokawa, K., Yotsutsuji, T., Kawasaki, Y., et al. (2002). Regional changes in brain gray and white matter in patients with schizophrenia demonstrated with voxel-based analysis of MRI. Schizophrenia Research, 55, 41±54. Swayze, V.W. 2nd, Andreasen, N.C., Alliger, R.J., Ehrhardt, J.C., Yuh, W.T. (1990). Structural brain abnormalities in bipolar affective disorder. Ventricular enlargement and focal signal hyperintensities. Archives of General Psychiatry, 47, 1054±1059. Thompson, P.M., Cannon, T.D., Narr, K.L., van Erp, T., Poutanen, V.P., Huttunen, M., et al. (2001). Genetic in¯uences on brain structure. Nature Neuroscience, 4, 1253±1258. Tkachev, D., Mimmack, M.L., Ryan, M.M., Wayland, M., Freeman, T., Jones, P.B., et al. (2003). Oligodendrocyte dysfunction in schizophrenia and bipolar disorder. Lancet, 362, 798±805. Uranova, N., Orlovskaya, D., Vikhreva, O., Zimina, I., Kolomeets, N., Vostrikov, V., et al. (2001). Electron microscopy of oligodendroglia in severe mental illness. Brain Research Bulletin, 55, 597±610. Uranova, N., Vostrikov, V., Orlovskaya, D., Rachmanova, V. (2004). Oligodendroglial density in the prefrontal cortex in schizophrenia and mood disorders: A study from the Stanley Neuropathology Consortium. Schizophrenia Research, 67, 269±275. van Erp, T.G., Saleh, P.A., Rosso, I.M., Huttunen, M., Lonnqvist, J., Pirkola, T., et al. (2002). Contributions of genetic risk and fetal hypoxia to hippocampal volume in patients with schizophrenia or schizoaffective disorder, their unaffected siblings, and healthy unrelated volunteers. American Journal of Psychiatry, 159, 1514±1520. van Erp, T.G., Saleh, P.A., Huttunen, M., Lonnqvist, J., Kaprio, J., Salonen, O., et al. (2004). Hippocampal volumes in schizophrenic twins. Archives of General Psychiatry, 61, 346±353. Velakoulis, D., Pantelis, C., McGorry, P.D., Dudgeon, P., Brewer, W., Cook, M., et al. (1999). Hippocampal volume in ®rst-episode psychoses and chronic schizophrenia ± A highresolution magnetic resonance imaging study. Archives of General Psychiatry, 56, 133±141. Velakoulis, D., Wood, S.J., Wong, M.T., McGorry, P.D., Yung, A., Phillips, L., et al. (2006). Hippocampal and amygdala volumes according to psychosis stage and diagnosis: A magnetic resonance imaging study of chronic schizophrenia, ®rst-episode psychosis, and ultra-high-risk individuals. Archives of General Psychiatry, 63, 139±149. Weinberger, D.R., Aloia, M.S., Goldberg, T.E., Berman, K.F. (1994). The frontal lobes and schizophrenia. Journal of Neuropsychiatry and Clinical Neuroscience, 6, 419±427. White, T., Andreasen, N.C., Nopoulos, P. (2002). Brain volumes and surface morphology in monozygotic twins. Cerebral Cortex, 12, 486±493. Wright, I.C., Ellison, Z.R., Sharma, T., Friston, K.J., Murray, R.M., McGuire, P.K. (1999a). Mapping of grey matter changes in schizophrenia. Schizophrenia Research, 35, 1±14. Wright, I.C., Sharma, T., Ellison, Z.R., McGuire, P.K., Friston, K.J., Brammer, M.J., et al. (1999b). Supra-regional brain systems and the neuropathology of schizophrenia. Cerebral Cortex, 9, 366±378. Wright, I.C., Rabe-Hesketh, S., Woodruff, P.W.R., David, A.S., Murray, R.M., Bullmore, E.T. (2000). Meta-analysis of regional brain volumes in schizophrenia. American Journal of Psychiatry, 157, 16±25. Wright, I.C., Sham, P., Murray, R.M., Weinberger, D.R., Bullmore, E.T. (2002). Genetic

196

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

contributions to regional variability in human brain structure: methods and preliminary results. Neuroimage, 17, 256±271. Zhou, S.Y., Suzuki, M., Hagino, H., Takahashi, T., Kawasaki, Y., Nohara, S., et al. (2003). Decreased volume and increased asymmetry of the anterior limb of the internal capsule in patients with schizophrenia. Biological Psychiatry, 54, 427±436.

CHAPTER EIGHT

Summary and implications Colm McDonald

This book describes key ®ndings from the Maudsley Family Study of Psychosis, which was developed to investigate intermediate phenotypes of psychotic illness. As outlined in the preface, a core theme addressed by the studies is `what neurobiological abnormalities do the unaffected relatives of patients who are at high genetic liability for psychosis display?' The different arms of the study have indeed detected certain neurobiological abnormalities in relatives at high genetic liability for the illness and thus rejected the null hypothesis ± that unaffected relatives do not differ from healthy volunteers. As is typical of scienti®c endeavours, the study poses many new questions while putting forward answers to the original ones ± some of these questions are currently being addressed by further phases of the study, which continues to evolve to take advantage of new technological developments and discoveries emanating from related areas of psychosis research such as molecular genetics. This chapter summarizes and integrates the key ®ndings described in preceding chapters and describes how this study and others will attempt to address new questions in the years ahead in order to better understand the aetiopathogenesis of psychotic illness.

SUMMARY OF KEY FINDINGS This monograph describes key ®ndings from over a decade of investigation by many researchers who have worked on the Maudsley Family Study of 197

198

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Psychosis, based at the Institute of Psychiatry, London, under the overall supervision of Professor Robin Murray. Results have emanated from four waves of clinical, cognitive and neurobiological data obtained from families multiply and singly affected with schizophrenia and multiply affected with psychotic bipolar disorder incorporating up to 550 participants. Several of the analyses have been published previously and a full list of peer-reviewed publications from the study is provided in Appendix 1. The chapters of this book provide a comprehensive delineation of the background and clinical characteristics of the entire sample of participants, highlight key ®ndings from the main arms of the study and present major unpublished data analyses, including smooth pursuit eye-tracking results, memory function and region-of-interest neuroimaging data combined across the entire sample of families affected with schizophrenia. The various overlapping datasets have been approached in slightly different ways by researchers working on different arms of the study, but with common strands running through the structure of the analyses. In each chapter of this book, the authors described abnormalities detected in patients with schizophrenia compared with healthy volunteers (and, for neuroimaging, abnormalities detected in psychotic bipolar-disorder patients compared with healthy volunteers and patients with schizophrenia). Additionally, the chapters' authors present results of analyses assessing whether similar abnormalities are detectable in the unaffected relatives of patients. Some analyses were performed on the entire sample of relatives compared with controls to maximize statistical power (evoked potentials, eye tracking, cognition) with subsequent analyses taking into account family history of illness, i.e. adopting the hypothesis that families multiply affected with illness will be most likely to display endophenotypic abnormalities. Other analyses (neurological signs, neuroimaging) divided participants into familial and non-familial subsections from the outset. Some subsections of the study (cognition, neuroimaging) speci®cally assessed abnormalities in the small number of parents who are assumed to be obligate carriers of genetic risk for illness or modelled genetic risk using the quantitative scale of genetic liability. These more sophisticated analyses re¯ect attempts by researchers to increase power for detecting endophenotypes by taking into account the likely increased genetic loading indicated by a stronger family history of illness.

Evoked potentials The evoked potential data from different arms of the study consistently demonstrated that both patients with schizophrenia and their unaffected relatives display signi®cantly delayed latency of the P300 wave, an electrophysiological measure of stimulus processing time in response to auditory

SUMMARY AND IMPLICATIONS

199

odd ball stimuli, strongly supporting this measure as an endophenotype of schizophrenia. There was also suggestive but weaker evidence for P300 amplitude reduction as an endophenotypic marker. In contrast, mismatch negativity evoked potential measurements, which most likely re¯ect the functional state of NMDA-receptor-mediated neurotransmission, were impaired in the frontal regions in patients with schizophrenia only and were unimpaired in their unaffected relatives. These data indicate that MMN de®cits are most likely to re¯ect state-related illness progression rather than a useful endophenotypic marker of schizophrenia for molecular genetics research. Each of the auditory evoked potential measurements appears to evaluate different brain information-processing functions mediated by distinct neurobiological mechanisms, which in turn are likely to be in¯uenced by different sets of genes. Some initial work linking molecular genetics and endophenotypic markers was completed, assessing whether variation in P300 wave in these participants was associated with a functional genetic polymorphism in the COMT gene, which alters metabolism of dopamine in the prefrontal cortex. No association was found, providing evidence against the involvement of this polymorphism in the impaired stimulus processing measured by the P300 wave. Future hypotheses to be addressed include whether there is a relationship between P300 wave latency and genetic variation of other emerging putative susceptibility genes for schizophrenia, as well as whether incorporating multivariate analyses of several electrophysiological markers may provide more statistical power than a single endophenotype used in isolation. Examining in detail the correlation between electrophysiological measures and other neurobiological measurements in the sample is also planned; for example, assessing whether the relatively strongly supported endophenotypes of delayed P300 latency and distributed white matter de®cit are found in the same individuals ± if so, this would support a common process of genetically driven structural and functional dysconnectivity of frontotemporo-parietal circuits in schizophrenia.

Eye tracking Patients with schizophrenia who participated in the study consistently displayed eye-tracking abnormalities. When performing the smooth pursuit task, patients displayed higher overall and catch-up saccade amplitudes and more qualitative abnormalities than other participants. When performing the antisaccade task, patients with schizophrenia displayed substantially higher antisaccade distractibility error scores than other participants. By contrast, no abnormalities were detected in unaffected relatives of patients when compared with controls, even in those relatives from multiply-

200

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

affected families most likely to carry susceptibility genes for illness. Thus, the Maudsley Family Study of Psychosis supports eye-tracking abnormalities as illness-related neurobiological abnormalities (possibly due to medication usage, cognitive dysfunction or neurodegenerative changes), rather than endophenotypic for schizophrenia. The lack of eye-tracking abnormalities in unaffected relatives of patients with schizophrenia in the study is all the more convincing as a true negative ®nding because it cannot be attributed to participant selection, given the many neurobiological abnormalities within other domains detected in the same or overlapping participants, i.e. relatives included in the study do display abnormalities of brain anatomy, cognition, auditory evoked potentials and neurological signs. Future studies will assess whether eye-tracking abnormalities in patients are correlated with other measurements of illness progression and whether more homogenous subsets of relatives, for example, selected by presence of potential susceptibility alleles for schizophrenia, display eye-tracking abnormalities.

Cognitive function Memory dysfunction is a prominent cognitive de®cit underpinning schizophrenia, and the neuropsychological arm of this study again con®rmed that patients with familial and non-familial schizophrenia demonstrate de®cits in both verbal and visual components of episodic memory, over and above impairments in general intellectual level. Relatives of non-familial patients also showed subtle de®cits in immediate and delayed verbal recall, whereas percentage retention score was unimpaired, suggesting that the memory de®cit in these participants was the consequence of an inability adequately to acquire information in the ®rst place rather than the result of impaired storage, retention or retrieval of information. Interestingly, this pattern of de®cit implicates a genetic effect on frontal lobe rather than medial temporal lobe processes. Surprisingly, in this component of the study, unaffected relatives of familial patients had no episodic memory dysfunction when compared with controls, despite the fact that an overlapping sample of individuals displayed neuroanatomical and neurophysiological abnormalities in other arms of the study. Similarly, patients with non-familial schizophrenia displayed executive function de®cits compared with controls, including impairments of planning ability, spatial working memory and strategy formation, and their unaffected relatives also displayed de®cits in spatial working memory and strategy formation. Another executive functioning task, rapid mental ¯exibility was unimpaired in patients and relatives. A smaller number of participants took part in measurements of attention function and patients had signi®cant de®cits in attention when compared

SUMMARY AND IMPLICATIONS

201

with controls, with a pattern indicative of selective attention de®cits. Unaffected relatives had trend-level de®cits in auditory (but not visual) sustained attention suggestive of a possible genetic effect on this measure. Evidence also emerged from the neuropsychological arm of the study that intellectual asymmetry, in the form of superiority of verbal ± to perceptual ± motor skills, is an indicator of genetic liability for schizophrenia, as this measure was correlated with the quantitative genetic loading scale scores of unaffected relatives. These studies from the neuropsychological arm of the study, in concert with other studies of unaffected relatives/co-twins of patients with schizophrenia, do provide evidence that some aspects of cognitive dysfunction are linked to susceptibility genes for schizophrenia. A future area of study will be to further decompose the components underlying complex neuropsychological tasks in order to dissect out those cognitive tasks more closely linked to the underlying genetic architecture. Future studies will also assess the relationship between speci®c susceptibility alleles and performance on neuropsychological tasks and develop statistical models which optimally combine different aspects of cognitive function to best predict genetic liability for schizophrenia.

Neurological signs In common with most other studies examining soft neurological signs in schizophrenia, patients had higher rates of these subtle indicators of neurological dysfunction than controls, albeit at a lower prevalence rate than reported elsewhere ± presumably due to the high threshold applied by raters for diagnosing speci®c signs in the current study. Non-familial patients with schizophrenia displayed signi®cantly increased rates of narrowly de®ned primary neurological signs compared with controls, consistent with the hypothesis that such patients may have been subject to a form of environmental focal brain insult. Integrative neurological signs, when broadly de®ned, were most prevalent in familial patients with schizophrenia compared with controls, indicative of diffuse cerebral dysfunction and impaired connectivity between sensory and motor networks. Furthermore, unaffected relatives of patients with schizophrenia from multiply-affected families also displayed signi®cantly increased rates of broadly de®ned integrative neurological signs. By contrast, those relatives from singly-affected families, who were presumed to be at lower risk of carrying susceptibility genes for the illness, did not signi®cantly differ from controls. This study provides evidence that integrative neurological signs, an apparent indicator of impaired connectivity of neural networks, are endophenotypic for schizophrenia. Future work will examine the association of integrative neurological signs with other indices of dysconnectivity

202

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

in the samples, such as neuroanatomical variables and auditory evoked potentials, and with allelic variation in putative susceptibility genes for schizophrenia.

Structural neuroimaging The structural neuroimaging arm of the study demonstrated that schizophrenia and psychotic bipolar 1 disorder are associated with distinct but overlapping deviations of regional brain volume. Region-of-interest analyses demonstrated that schizophrenia is associated with enlarged lateral ventricular volume and reduced hippocampal volume. Computational morphometry analysis of grey matter con®rmed the hippocampal de®cit associated with schizophrenia and demonstrated additional widespread volume de®cit involving the medial temporal lobe more extensively, the lateral temporal cortex, basal ganglia, thalamus, insula, anterior cingulate gyrus, dorsolateral prefrontal cortex, orbitofrontal cortex, right parietal cortex, precuneus and cerebellum. By contrast, psychotic bipolar 1 disorder was associated with none of these morphometric abnormalities. However, computational morphometry analysis of white matter demonstrated that both schizophrenia and bipolar disorder were associated with substantial volume de®cit of white matter in frontal and temporoparietal regions. The regions of white matter affected involved both intrahemispheric and interhemispheric tracts and these ®ndings support structural dysconnectivity as an underlying substrate of both schizophrenia and bipolar disorder. Echoing the disease-related morphometric deviations, the structural neuroimaging arm of the study also demonstrated that the potential morphometric endophenotypes of schizophrenia and psychotic bipolar 1 disorder are distinct but overlapping. Employing region-of-interest analyses and the standard approach of comparing unaffected relatives to controls, enlarged lateral ventricular volume was identi®ed as potentially endophenotypic for schizophrenia, whereas hippocampal volume de®cit was found to be disease speci®c rather than endophenotypic. In an attempt to maximize statistical power by employing computational morphometry and the quantitative measure of genetic liability as a predictor variable, a distributed network of grey matter de®cit representing a subset of the case± control differences was supported as endophenotypic for schizophrenia. This network comprised the caudate nucleus and thalamus, orbital, prefrontal and premotor parts of frontal cortex, and left insula and lateral temporal cortex. Thus susceptibility genes are implicated as the major risk factor for the frontal, lateral temporal and subcortical volume de®cits associated with schizophrenia, whereas the cause of volume de®cit in the medial temporal lobe, precuneus and cerebellum is presumed non-genetic in origin.

SUMMARY AND IMPLICATIONS

203

In contrast to the distributed grey matter regions supported as endophenotypic for schizophrenia, a distinct and localized region of grey matter de®cit comprising the right medial frontal lobe and ventral striatum, regions known to be involved in emotional processing and previously implicated in the pathophysiology of mood disorders, was implicated as endophenotypic for bipolar disorder. Mirroring the ®ndings of overlapping disease related morphometric deviations, the study demonstrated that distributed white matter de®cit was potentially endophenotypic for both schizophrenia and bipolar disorder with speci®c overlap in the left frontal and temporoparietal regions, indicating that structural dysconnectivity of left fronto-temporo-parietal regions represented a generic manifestation of genetic susceptibility for both forms of psychosis. A further question now arising from these studies is whether different susceptibility genes may be responsible for the distinct grey matter de®cits supported as endophenotypic for schizophrenia and for bipolar disorder, whereas common susceptibility genes for both psychotic disorders may be responsible for the shared white matter de®cits identi®ed. Therefore, potential common candidate genes for psychosis may be those which impact on white matter structure such as through myelination and oligodendrocyte function. However, more complex hypotheses are also consistent with the data. For example, it is equally possible that the white matter de®cits supported as endophenotypic for both disorders represent a ®nal common pathway of differing genetic processes; or that similar genes render participants who are at high genetic liability for either disorder susceptible to grey matter de®cit, but that discrete disease-modifying factors differentiate the impact of these genes towards functionally diverse brain regions. What these studies provide are quantitative endophenotypes with which to test such hypotheses in future studies, and the opportunity to select candidate genes for these potential endophenotypes in a manner that can be informed by knowledge regarding the pathophysiological mechanisms whereby such genes could impact on brain structure.

IMPLICATION OF FINDINGS Thousands of studies, performed over recent decades, have reported a range of abnormalities in patients with schizophrenia or bipolar disorder compared with healthy volunteers. In common with such studies, the Maudsley Family Study of Psychosis identi®ed abnormalities associated with schizophrenia in almost all variables examined utilizing the investigative techniques of evoked potentials, eye tracking, cognitive function, neurological signs and neuroimaging in patients with schizophrenia compared with controls. A recurrent caveat to the interpretation of crosssectional case±control studies is the uncertain extent to which illness-related

204

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS TABLE 8.1 Summary of main neurobiological abnormalities supported as intermediate phenotypes of schizophrenia by the Maudsley Family Study of Psychosis

Neurobiological measure

Increased P300 wave latency Reduced frontal mismatch negativity amplitude Impaired smooth pursuit eye tracking Antisaccade eye tracking abnormalities Verbal episodic memory impairment Visual episodic memory impairment Executive dysfunction in spatial working memory Executive dysfunction in strategy formation Visual sustained attention de®cit Selective attention de®cit Superiority of verbal to perceptual motor skills Increased integrative neurological signs Increased focal neurological signs Enlarged lateral ventricles Medial temporal lobe de®cits Cerebellar, precuneus grey matter de®cits Frontal, lateral temporal, subcortical grey matter de®cits Frontal, temporal±parietal white matter de®cits

Patients with schizophrenia

Relatives of patients with schizophrenia

X X X X X X X X X X X X X X X X X X

X

X* X* X X X X X

Neurobiological abnormalities in bold are supported as potential intermediate phenotypes of schizophrenia by the Maudsley Family Study of Psychosis, because they are identi®able in unaffected relatives as well as patients. Abnormalities found in patients alone, but not relatives, are likely to be illness related or linked to environmental risk factors for illness. * Signi®cant impairments con®ned to non-familial relatives only and not identi®ed in relatives from multiply-affected families compared with controls.

factors such as mental state, antipsychotic medication, cigarette smoking, substance abuse, poor diet or institutionalization have contributed to the abnormalities detected. In concert with a number of other studies assessing unaffected individuals at high genetic risk for psychosis, the Maudsley Family Study of Psychosis has provided evidence that certain neurobiological abnormalities are not produced by illness-related confounds but are rather trait-related features of psychosis most likely linked to underlying genetic susceptibility for illness. A summary of the main abnormalities identi®ed in patients and those supported as potential intermediate phenotypes of schizophrenia, by virtue of being also detectable in unaffected relatives of patients, is provided in Table 8.1. The various modalities of investigation employed in the study touch on different domains of neurobiological function. Nevertheless, some anatomical and physiological convergence of the proposed intermediate phenotypes

SUMMARY AND IMPLICATIONS

205

is apparent. Whereas cognitive functioning as assessed by neuropsychological measures generally depends on the integrity of neural networks rather than focal brain regions, the endophenotypic involvement of executive dysfunction and impaired acquisition of verbal information for later recall is suggestive of frontal dysfunction. The neuroanatomical endophenotypes identi®ed for schizophrenia also centre on de®cits of grey and white matter in frontal and closely connected lateral temporal and subcortical structures. Impaired connectivity (without speci®c neuroanatomical localization) is implicated by de®cits of integrative neurological signs and delayed P300 latency. Taken together, these ®ndings from different arms of the study are consistent with the hypothesis that schizophrenia is a genetic disorder of disturbed connectivity between frontal and temporal/subcortical structures. Molecular processes underpinning connectivity in these regions, e.g. those related to synaptic plasticity and oligodendrocyte function, are therefore supported as suitable targets to explore pathways from genetic variation to neurobiological intermediate phenotypes.

Negative findings in unaffected relatives As outlined in the preceding chapters, and also demonstrated in Table 8.1, a number of the neurobiological abnormalities associated with schizophrenia were not detected in unaffected relatives and are therefore presumed not to represent intermediate phenotypes of illness. These abnormalities can instead be considered to represent direct illness effects, such as due to antipsychotic medication, neurotoxic effects of psychosis or poor physical health or, alternatively, the impact of environmental risk factors on neurobiology. The relatively large sample size and certain strengths of the methodology employed in the study serve to underline the validity of these `negative' results in relatives. In Chapter 4, the lack of any detectable eye-tracking abnormalities in unaffected relatives, in contrast to some other studies, is discussed in the light of the Maudsley Family Study of Psychosis methodology. These points equally support the validity of negative results from other sections of the study (e.g. spatial working memory impairment, medial temporal lobe de®cits), which found that relatives did not signi®cantly differ from controls. One point relates to the observation that some other studies seeking endophenotypes for schizophrenia recruited healthy volunteers with stricter inclusion criteria than were applied to unaffected relatives, e.g. no personal or family history of any psychiatric illness. This may have resulted in the recruitment of `supercontrols' for such studies, which might have overstated abnormalities in unaffected relatives. Following sound epidemiological principles, in order that relative and control groups be comparable without selection bias, they must be subject to the same exclusion

206

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

criteria with respect to psychiatric morbidity and differ only with regard to family history of schizophrenia. Another point is that some other studies ignored the fact that neurobiological variables assessed in relatives and patients ought not be treated as independent measures in statistical analyses. Studies within the Maudsley Family Study of Psychosis generally employed robust standard errors, which took account of family membership, and controlled for demographic confounds such as age and gender. Given these aspects of methodology, which may account for some of the negative ®ndings in unaffected relatives in the Maudsley Family Study of Psychosis, those abnormalities that were successfully detected in measures of evoked potentials, cognitive function, neurological signs and neuroanatomy among the unaffected relatives of patients are more likely to be valid.

Variations with presumed genetic loading The study also differed from some others in the literature by attempting to model likely genetic loading into the sample of unaffected relatives rather than treat these participants as a homogenous sample at an equivalent level of genetic risk. This was largely through employing the familial/non-familial distinction, although some sections of the study speci®cally examined variables in those relatives who were presumed obligate carriers of genetic risk or in association with the quantitative scale of genetic liability. Results demonstrated that certain measures ± increased intellectual asymmetry, increased rates of integrative neurological signs, enlarged lateral ventricles, frontotemporal and subcortical grey matter de®cits, frontotemporal and parietal white matter de®cits ± were indeed more prominent in those unaffected relatives who had presumed higher genetic loading. However, the cognitive measures of verbal episodic memory impairment, executive dysfunction in planning ability and in strategy formation were detected in non-familial relatives only and not those from multiply-affected families. This counterintuitive pattern of de®cit was con®ned to these memory and executive function measures and was not found in other cognitive measures or other neurophysiological, neurological or neuroanatomical domains. This ®nding cannot interpreted as supporting an environmental contribution towards these abnormalities (unlike differences found between patients with and without a family history of illness, whereby the latter could be conceptualized as being more likely to portray the impact of environmental risk factors), as any abnormalities in unaffected relatives of patients, including from singly affected families, can be considered as potentially endophenotypic for illness. One interpretation offered in Chapter 5 for this counterintuitive ®nding is sampling variation ± it may be that familial

SUMMARY AND IMPLICATIONS

207

relatives with higher cognitive functioning were more likely to participate in the study whereas a more epidemiological sample would have also incorporated patients and relatives from multiply-affected families who displayed such cognitive impairments. Another interpretation was that a potential threshold effect for these cognitive measures operates in multiply-affected families, i.e. familial relatives carry increased risk for illness and if these relatives also had the critical additional risk genes associated with these executive function and memory de®cits, they would cross the threshold for illness and be categorized as affected. Relatives from the non-familial sample, by contrast, may have had a higher threshold for developing illness or might carry other protective factors that enabled them to carry the neuropsychological endophenotype without manifesting illness. The apparent heterogeneity of neuropsychological test results among unaffected relatives emphasizes the need to select more homogenous groups of unaffected individuals for genetic studies, de®ned for example by impaired performance on certain neuropsychological tasks.

THE NEXT STAGES Considerable future work will be performed on the present sample and others to further elucidate the implications of the ®ndings. These include neuropsychological assessments and auditory evoked potential assessments of the multiply-affected bipolar families. The extent to which abnormalities detected in the various arms of the study tap into similar or separable neurobiological processes will be assessed by a series of correlation studies. More novel imaging methodologies will be employed to further explore dysfunction of the brain networks underlying both forms of psychosis. For example, diffusion tensor imaging (DTI) techniques are currently being applied to a subset of the Maudsley Family Psychosis sample to more accurately delineate the white matter abnormalities detected by standard structural imaging reported in this monograph. White matter appears homogenous on conventional MRI, which cannot discern white matter tract direction or organization, whereas DTI gives information about the microstructural features of white matter. Extensive functional MRI studies are underway to more clearly identify those abnormally functioning neural networks both in patients and their unaffected relatives using tasks designed to activate more extensive frontal connections incorporating working memory, executive control and facial emotional recognition tasks. Secondary analyses of fMRI data in which connectivity is modelled statistically will be performed to assess higher-level abnormalities of functional and effective connectivity, i.e. disturbance of information ¯ows between predetermined neural networks.

208

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Valuable future work will be to relate the neurobiological variations identi®ed in the study with genotypic variation, i.e. to utilize these endophenotypes as well as the clinical phenotypes in molecular genetic studies of psychosis. Studies are underway linking these neurobiological endophenotypes to initially proposed gene variants associated with psychotic illness described in Chapter 1, including those reported for neuregulin 1, dysbindin, DISC1, RGS4, DAOA and COMT. Several more potential gene variants using the clinical phenotypes will emanate from international research groups in the years ahead, which in turn will be related to these potential endophenotypes in an attempt to better elucidate their biological signi®cance. By pursuing these approaches in the years ahead we hope that the complex pathways whereby genetic variation produces neurobiological dysfunction underlying schizophrenia and bipolar disorder will be clari®ed, and that the resulting improved understanding of these syndromes will provide substantially more scope to intervene in order to better manage and prevent these devastating psychotic disorders.

Appendix 1

List of data based peer reviewed publications from the Maudsley Family Study of Psychosis up to August 2007:

· · · · · ·

Birkett, P., Sigmundsson, T., Sharma, T., Toulopoulou, T., Grif®ths, T.D., Reveley, A., Murray, R. (2007). Reaction time and sustained attention in schizophrenia and its genetic predisposition. Schizophrenia Research, 95(1±3), 76±85. Bramon, E., Croft, R., McDonald, C., Virdi, G., Gruzelier, J., Baldeweg, T., et al. (2004). Mismatch negativity in schizophrenia: A family study. Schizophrenia Research, 67, 1±10. Bramon, E., McDonald, C., Croft, R., Landau, S., Filbey, F., Gruzelier, J., et al. (2005). Is the P300 wave an endophenotype for schizophrenia? A metaanalysis and a family study. Neuroimage, 27, 960±968. Bramon, E., Walshe, M., McDonald, C., Martin, B., Toulopoulou, T., Wickham, H., et al. (2005). Dermatoglyphics and schizophrenia: A metaanalysis and investigation of the impact of obstetric complications upon a-b ridge count. Schizophrenia Research, 75, 399±404. Bramon, E., Dempster, E., Frangou, S., McDonald, C., Schoenberg, P., MacCabe, J., et al. (2006). Is there an association between the COMT gene and P300 endophenotypes? European Psychiatry, 21, 70±73. Chapple, B., Grech, A., Sham, P., Toulopoulou, T., Walshe, M., Schulze, K., et al. (2004). Normal cerebral asymmetry in familial and non-familial schizophrenic probands and their unaffected relatives. Schizophrenia Research, 67, 33±40.

209

210

· · · · · · · · · · · · ·

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Chua, S., Sharma, T., Takei, N., Murray, R., Woodruff, P. (2000). A magnetic resonance imaging study of corpus callosum size in familial schizophrenic subjects, their relatives, and normal controls. Schizophrenia Research, 41, 397± 403. Connor, S., Ng, N., McDonald, C., Schulze, K., Morgan, K., Dazzan, P., Murray, R. (2004). A study of hippocampal shape anomaly in schizophrenia and in families multiply affected by schizophrenia or bipolar disorder. Neuroradiology, 46, 523±534. Crawford, T.J., Sharma, T., Puri, B.K., Murray, R.M., Berridge, D.M., Lewis, S.W. (1998). Saccadic eye movements in families multiply affected With schizophrenia: The Maudsley Family Study. American Journal of Psychiatry, 155, 1703±1710. Dempster, E., Toulopoulou, T., McDonald, C., Bramon, E., Walshe, M., Filbey, F., et al. (2005). Association between BDNF val66met genotype and episodic memory. American Journal of Medical Genetics Part, B. (Neuropsychiatric Genetics), 134B, 73±75. Dempster, E.L., Toulopoulou, T., McDonald, C., Bramon, E., Walshe, M., Wickham, H., et al. (2006). Episodic memory performance predicted by the 2bp deletion in exon 6 of the `alpha 7-like' nicotinic receptor subunit gene. American Journal of Psychiatry, 163, 1832±1834. Dikeos, D.G., Wickham, H., McDonald, C., Walshe, M., Sigmundsson, T., Bramon, E., et al. (2006). Distribution of symptom dimensions across Kraepelinian divisions. British Journal of Psychiatry, 189, 346±353. Frangou, S., Sharma, T., Alarcon, G., Sigmundsson, T., Takei, N., Binnie, C., Murray, R.M. (1997). The Maudsley Family Study, II: Endogenous eventrelated potentials in familial schizophrenia. Schizophrenia Research, 23, 45±53. Frangou, S., Sharma, T., Sigmundsson, T., Barta, P., Pearlson, G., Murray, R.M. (1997). The Maudsley Family Study 4. Normal planum temporale asymmetry in familial schizophrenia. A volumetric MRI study. British Journal of Psychiatry, 170, 328±333. Grif®ths, T., Sigmundsson, T., Takei, N. (1998). Minor physical anomalies in familial and sporadic schizophrenia: The Maudsley Family Study. Journal of Neurology, Neurosurgery and Psychiatry, 64, 56±60. Grif®ths, T., Sigmundsson, T., Takei, N., Rowe, D., Murray, R. (1998). Neurological abnormalities in familial and sporadic schizophrenia. Brain, 121, 191±203. Kravariti, E., Toulopoulou, T., Mapua-Filbey, F., Schulze, K., Walshe, M., Sham P., et al. (2006). Intellectual asymmetry and genetic liability in ®rstdegree relatives of probands with schizophrenia. British Journal of Psychiatry, 188, 186±187. MacCabe, J., Simon, H., Zanelli, J., Walwyn, R., McDonald, C., Murray, R. (2005). Saccadic distractibility is elevated in schizophrenic patients but not in their well relatives. Psychological Medicine, 35, 1727±1736. McDonald, C., Grech, A., Toulopoulou, T., Schulze, K., Chapple, B., Sham, P., et al. (2002). Brain volumes in familial and non-familial schizophrenic probands and their unaffected relatives. American Journal of Medical Genetics (Neuropsychiatric Genetics), 114, 616±625.

APPENDIX 1

· · · · · · · · · · · ·

211

McDonald, C., Bullmore, E., Sham, P., Chitnis, X., Wickham, H., Bramon, E., Murray, R. (2004). Association of genetic risks for schizophrenia and bipolar disorder with speci®c and generic brain structural endophenotypes. Archives of General Psychiatry, 61, 974±984. McDonald, C., Bullmore, E., Sham, P., Chitnis, X., Suckling, J., MacCabe, J., et al. (2005). Regional volume deviations of brain structure in schizophrenia and psychotic bipolar disorder: Computational morphometry study. British Journal of Psychiatry, 186, 369±377. McDonald, C., Marshall, N., Sham, P.C., Bullmore, E.T., Schulze, K., Chapple, B., et al. (2006). Regional brain morphometry in patients with schizophrenia or bipolar disorder and their unaffected relatives. American Journal of Psychiatry, 163, 478±487. Rabe-Hesketh, S., Toulopoulou, T., Murray, R. (2001). Multilevel modelling of cognitive function in schizophrenic patients and their ®rst-degree relatives. Multivariate Behavioral Research, 36, 279±289. Rijsdijk, F.V., vanHaren, N.E.M., Picchioni, M.M., McDonald, C., Toulopoulou, T., Hulshoff Pol, H.E., et al. (2005). Brain MRI abnormalities in schizophrenia: Same genes or same environment? Psychological Medicine,, 35(10), 1399±1409 Schulze, K., McDonald, C., Frangou, S., Sham, P., Grech, A., Toulopoulou, T., et al. (2003). Hippocampal volume in familial and non-familial schizophrenic probands and their unaffected relatives. Biological Psychiatry, 53, 562±570. Schulze, K., MacCabe, J., Rabe-Hesketh, S., Crawford, T., Marshall, N., Zanelli, J., et al. (2006). The relationship between eye movement and brain structural abnormalities in patients with schizophrenia and their unaffected relatives. Journal of Psychiatric Research, 40, 589±598. Schulze, K.K., Hall M.-H., McDonald, C., Marshall, N., Walshe, M., Murray, R.M., Bramon, E. (2007). P50 auditory evoked potential suppression in bipolar disorder patients with psychotic features and their unaffected relatives. Biological Psychiatry, 62, 121±128. Schulze, K.K., Hall, M.-H., McDonald, C., Marshall, N., Walshe, M., Murray, R.M., Bramon, E. (2008). Auditory P300 in patients with bipolar disorder and their unaffected relatives. Bipolar Disorders, 10(3), 377±386. Sharma, T., Du Boulay, G., Lewis, S., Sigmundsson, T., Gurling, H., Murray, R. (1997). The Maudsley family study I: Structural brain changes on magnetic resonance imaging in familial schizophrenia. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 21, 1297±1315. Sharma, T., Lancaster, E., Lee, D., Lewis, S., Sigmundsson, T., Takei, N., et al. (1998). Brain changes in schizophrenia: Volumetric MRI study of families multiply affected with schizophrenia ± The Maudsley Family Study 5. British Journal of Psychiatry, 173, 132±138. Sharma, T., Lancaster, E., Sigmundsson, T., Lewis, S., Takei, N., Gurling, H., et al. (1999). Lack of normal pattern of cerebral asymmetry in familial schizophrenic patients and their relatives ± The Maudsley Family Study. Schizophrenia Research, 40, 111±120.

212

· · · · · · · · · · · ·

THE MAUDSLEY FAMILY STUDY OF PSYCHOSIS

Spence, S.A., Grasby, P.M., Liddle, P.F., Stefan, M.D., Sharma, T., Murray, R.M., et al. (2000). Functional anatomy of verbal ¯uency in people with schizophrenia and those at genetic risk: Focal dysfunction and distributed disconnectivity reappraised. British Journal of Psychiatry, 176, 52±60. Stefanis, N., Frangou, S., Yakeley, J., Sharma, T., O'Connell, P., Morgan, K., et al. (1999). Hippocampal volume reduction in schizophrenia: effects of genetic risk and pregnancy and birth complications. Biological Psychiatry, 46, 697±702. Toulopoulou, T., Morris, R., Rabe-Hesketh, S., Murray, R. (2003). Selectivity of verbal memory de®cit in schizophrenic patients and their relatives. American Journal of Medical Genetics-Neuropsychiatric Genetics B, 116, 1±7 Toulopoulou, T., Rabe-Hesketh, S., King, H., Murray, R., Morris, R. (2003). Episodic memory in schizophrenic patients and their relatives. Schizophrenia Research, 63, 261±271. Toulopoulou, T., Grech, A., Morris, R., Schulze, K., McDonald, C., Chapple, B., et al. (2004). The relationship between volumetric brain changes and cognitive function: a family study on schizophrenia. Biological Psychiatry, 56, 447±453. Toulopoulou, T., Mapua-Filbey, F., Quraishi, S., Kravariti, E., Morris, R.G., McDonald, C., et al. (2005). Cognitive performance in presumed obligate carriers for psychosis. British Journal of Psychiatry, 187, 284±285. Toulopoulou, T., Quraishi, S., McDonald, C., Murray, R. (2006). The Maudsley Family Study: Premorbid and current general intellectual function levels in familial bipolar I disorder and schizophrenia. Journal of Clinical and Experimental Neuropspychology, 28, 1±17. Toulopoulou T., Chua, S.E., Lam, I., Cheung, V., Murray, R.M., David, A.S. (2008). Evidence of normal hearing laterality in familial schizophrenic patients and their relatives. American Journal of Medical Genetics (Neuropsychiatric Genetics) B, 147B, 73±76. van Haren, N., Picchioni, M., McDonald, C., Marshall, N., Davis, N., Ribchester, T., et al. (2004). A controlled study of brain structure in monozygotic twins concordant and discordant for schizophrenia. Biological Psychiatry, 56, 454±461. Walshe, M., McDonald, C., Taylor, M., Zhao, J., Sham, P., Grech, A., et al. (2005). Obstetric complications in patients with schizophrenia and their unaffected siblings. European Psychiatry, 20, 28±34. Walshe, M., Taylor, M., Schulze, K., Bramon, E., Frangou, S., Stahl, D., et al. (2007) Familial liability to schizophrenia and premorbid adjustment. British Journal of Psychiatry, 191, 260±261 Zanelli, J., Simon, H., Rabe-Hesketh, S., Walshe, M., McDonald, C., Murray, R., MacCabe, J. (2005). Eye tracking in schizophrenia: Does the antisaccade task measure anything that the smooth pursuit task does not? Psychiatry Research, 136, 181±188.)

Index

The abbreviation MFSP is used for Maudsley Family Study of Psychosis. Abnormal Involuntary Movement in Schizophrenia (AIMS) scale, 137 Adoption studies, 2±3 Alpha-7 nicotinic cholinergic receptor (CHRNA7), 58, 60, 61 Annett Handedness scale, 24 Antipsychotic drugs, 27, 53±54, 150 effects on eye movements, 87 Antisaccade distractibility error (ADE), 83, 86, 88 Antisaccade task, 73, 77, 82, 84, 86, 199 Associate learning, 103 Association studies, 5±6, 9±10, 63 Attention impairment, 114±120 research methodology (MFSP), 115±117 Auditory evoked potentials acquisition and analysis of EEG/ERP data, 49±50 meta-analyses of published literature, 44±49 mismatch negativity (MMN), 43±44, 53±54, 56±58, 59, 61, 199 P50 gating/wave suppression, 44, 47±49, 50, 58, 59, 60, 61, 64

P300 waveform, 42±43, 45±47, 48, 51±53, 55±56, 59±60, 61, 198±199 Back-up saccades (BUS), 73, 76, 80±81 Bipolar disorder, 6±10 brain structural deviations and, 156, 164±166, 177±189, 202 ERP endophenotypes and, 61±62 morphometric abnormalities associated with, 164±165 morphometric endophenotypes associated with, 165±166, 167 research participant recruitment (MFSP), 21±23 study methodology and background see Research characteristics (MFSP); Research methodology (MFSP) Brain structural deviations areas of previous studies, 155±156 computational morphometry assessment of association with schizophrenia and bipolar disorder, 167±177

213

214

INDEX

computational morphometry assessment of genetic liability to schizophrenia and bipolar disorder, 177±188 disorder speci®city of grey and white matter endophenotypes, 185, 202±203 morphometric abnormalities associated with schizophrenia and bipolar disorder, 164±165 morphometric endophenotypes associated with schizophrenia and bipolar disorder, 165±166 multilevel modelling of MRI endophenotypes, 179±182 region-of-interest study, 155, 157±167, 202 Brain tissue volumes, in schizophrenia and bipolar disorder global tissue, 169 grey matter, 169±171, 172, 174±175, 180±184, 185±187, 202±203 white matter, 171±176, 183, 184, 187±188, 202, 203 Brain volumes, in familial and nonfamilial schizophrenia patients and their relatives hippocampal, 163, 166, 167, 202 lateral ventricular, 161, 165 third ventricular, 161±162, 165 whole brain, 162±163 Catch-up saccades (CUS), 73, 76, 80±81 Catechol-O-methyltranferase (COMT) gene, 6, 43, 52±53, 58, 59±60, 199, 208 CHRNA7 (alpha-7 nicotinic cholinergic receptor), 58, 60, 61 Chromosomal deletion, 6 Clinical assessments, 23±25 Clinical characteristics of MFSP research sample, 27±29 Clinical methodology see Research methodology (MFSP) Cognitive de®cits see Neuropsychological impairments COMT (catechol-O-methyltranferase) gene, 6, 43, 52±53, 58, 59±60, 199, 208 Concordance rates, in twins, 2

Continuous Performance Test (CPT), 114±115, 116±117 D-amino-acid oxidase inhibitor (DAOA) gene, 6, 9, 208 Dementia praecox, 1, 6±7 Diffusion tensor imaging (DTI), 207 Disrupted in schizophrenia genes DISC1, 6, 9, 43, 58, 59, 61, 62, 208 DISC2, 59 Dizygotic (DZ) twins, 2, 9, 12 Drugs, antipsychotic, 27, 53±54, 150 effects on eye movements, 87 DSM (Diagnostic and Statistical Manual of Mental Disorders) diagnoses, 7, 23±24, 27 Dysbindin gene, 6, 9, 58, 62, 208 DZ (dizygotic) twins, 2, 9, 12 Endophenotypes, 10±12 brain structural deviations see Brain structural deviations eye-movement abnormalities see Eyemovement abnormalities grey matter, 185±187, 202±203 see also Brain structural deviations morphometric, 165±166, 167 multilevel modelling of MRI endophenotypes, 179±180 neurobiological abnormalities supported by MFSP as intermediate phenotypes, 204 neurological see Neurological abnormalities neurophysiological see Neurophysiological abnormalities neuropsychological see Neuropsychological impairments white matter, 185, 187±188, 202 see also Brain structural deviations Episodic memory, 94±105, 200 ERPs (event-related potentials), 41 auditory evoked potentials see Auditory evoked potentials as endophenotypes for schizophrenia, 54±62, 63 Event-related potentials see ERPs Evoked potentials see Auditory evoked potentials

INDEX

Exclusion criteria for MFSP research, 22±23 Executive function impairment executive function in patients with schizophrenia and their relatives, 105±114, 200 motor execution time, 109 planning ability, 107, 108±109, 110, 111±113 problem-solving, 109 research methodology (MFSP), 106±107 strategy formation, 111, 112, 113±114 Executive Golf task, 107, 113 Eye-movement abnormalities assessment of eye movements, 71±74 observer bias, 73, 77, 85 publication bias, 85±86 research methodology (MFSP), 75±78, 84±85 schizophrenia and, 74±75, 78±88 statistical analysis, 77±82, 84 type II error, 86 Family History Research Diagnostic Criteria (FHRDC), 24 Family Interview for Genetic Studies (FIGS), 24 Family studies contribution of genetic liability to schizophrenia, 1±2 overlap in genetic liability for schizophrenia and bipolar disorder, 7±8 Frontal lobe pathology, 86 G72 gene, 61, 62 Gain measurement of eye movements, 72 Genetic liability (GL) continuous genetic liability model, 29 contribution to schizophrenia, 1±3 endophenotypic regions of grey and white matter associated with, 181±182 genetic in¯uences on neurophysiological endophenotypes, 58±61 genetic liability scale, 29±37, 178 in intellectual asymmetry, 121

215

overlap for schizophrenia and bipolar disorder, 6±10 structural brain deviations associated with GL to schizophrenia and bipolar disorder, 177±188 Genetic transmission, mode of, 3±4 Genetics, molecular see Molecular genetics Haemochromatosis, 11 Heritability estimates for bipolar disorder, 7 for schizophrenia, 3 see also Genetic liability Inclusion criteria for MFSP research, 23 Incomplete penetrance, 4 Intellectual asymmetry, 120±124 Intermediate phenotypes brain structural deviations and see Brain structural deviations eye-tracking dysfunction see Eyemovement abnormalities meaning of endophenotypes, 10±12 morphometric endophenotypes, 165±166 see also Brain structural deviations neurobiological abnormalities supported by MFSP as, 204 neurological see Neurological abnormalities neurophysiological see Neurophysiological abnormalities neuropsychological see Neuropsychological impairments IQ de®cits, 120±124 Kraepelin, Emil, 1, 6±7 Kraepelinian dichotomy of psychosis, 177±178, 185±186 Learning associate, 103 verbal memory and, 97, 98, 99±100 visual memory and, 97, 98±103 Letter Cancellation task, 117 Liability-threshold model of illness, 4, 29, 121 Linkage studies, 4±5, 9, 42, 63

216

INDEX

Manic Depressive Fellowship, 21 Manic depressive insanity, 6±7 correlations of manic and schizophrenic syndromes, 9 Memory episodic, 94±105, 200 learning and, 97, 98±103 in relatives of patients with schizophrenia, 95±105, 200 research methodology (MFSP), 95±96 semantic, 94 spatial working memory, 111, 112, 113±114 verbal, 97, 98, 99±100, 104 visual, 97, 98±103 Methodology see Research methodology (MFSP) Mismatch negativity (MMN), 43±44, 53±54, 56±58, 59, 61, 199 MMN see Mismatch negativity Molecular genetics contribution of genetic liability to schizophrenia, 4±6 overlap in genetic liability for schizophrenia and bipolar disorder, 9±10 Monozygotic (MZ) twins, 2, 9, 12 Morphometric abnormalities see Brain structural deviations Motor execution time, 109 MRI brain scans, 157±158, 168, 178±179 MRI endophenotypes see Brain structural deviations MZ (monozygotic) twins, 2, 9, 12

neurological soft signs (NSS), 133±134 primary/focal, 136±137, 141, 143±147, 148±149, 151 research methodology (MFSP), 135±138, 150 in schizophrenia patients and their relatives, 133±135, 138±151, 201±202 threshold criteria, 150 Neurological soft signs (NSS), 133±134 Neurophysiological abnormalities genetic in¯uences, 58±61 multivariate endophenotypes, 62±63 neurophysiological experiments, 41±44, 49±54 schizophrenia and ERP abnormalities, 54±62, 63 severity of neurophysiological de®cits in schizophrenia literature, 44±49 Neuropsychological impairments of attention, 114±120, 200±201 of executive function, 105±114, 200 familial/non-familial distinction, 98, 100, 102, 110, 111±113, 114, 125±126 of intellectual asymmetry, 120±124 measures, 96, 107 of memory, 94±105, 200 N-methyl-D-aspartate (NMDA), 43±44 N-methyl-D-aspartate receptor (NMDAR)-mediated nuerotransmission, 57 NSS (neurological soft signs), 133±134

National Schizophrenia Fellowship (Rethink), 21, 120 Neuregulin genes, 6, 62 NRG1 , 5±6, 9, 58, 61, 208 Neurological abnormalities blinding in assessment, 150 drug effects, 150 factors in neurological scores, 140 familial schizophrenia and, 141±146, 148±150, 201 frequency of individual abnormalities, 146 integrative, 136±137, 142, 143±146, 147±148, 149, 151 measures, 136

P50 gating/wave suppression, 44, 47±49, 50, 58, 59, 60, 61, 64 P300 waves, 42±43, 45±47, 48, 51±53, 55±56, 59±60, 61, 198±199 Participant recruitment for MSFP research, 21±22 PAS (Premorbid Adjustment Scale), 24, 25, 28 Penetrance, incomplete, 4 Planning ability, 108±109, 110, 111±113 Pleiotropy, 4 Premorbid Adjustment Scale (PAS), 24, 25, 28 Premorbid and Schizoid-Schizotypal Traits scale (PSST), 24, 25, 28, 203

INDEX

Problem-solving, 109 Proline dehydrogenase (PRODH) gene, 6 PSST (Premorbid and SchizoidSchizotypal Traits scale), 24, 25, 28, 203 Re¯exive saccades, 73, 74 Regulator of G protein signalling 4 (RGS4) gene, 6, 61, 62, 208 Research characteristics (MFSP) clinical characteristics of the sample, 27±29 demographics see Demographics of MFSP research scope and purpose, 12±14 sociodemographics, 24, 25±26, 116, 122 Research methodology (MFSP) attention impairment, 115±117 clinical assessments, 23±25 executive function impairment, 106±107 eye-movement abnormalities, 75±78, 84±85 genetic liability scale, 29±37 inclusion and exclusion criteria, 22±23 intellectual asymmetry, 121±122 memory impairment, 95±96 neurological abnormalities, 135±138, 150 participant recruitment, 21±23 structural brain deviations associated with schizophrenia and bipolar disorder, 157±158, 166±167, 168±169, 176±177, 178±179 study phases, 22 Rethink (National Schizophrenia Fellowship), 21, 120 RGS4 (regulator of G protein signalling 4) gene, 6, 61, 62, 208 Saccades, 72, 74±76, 83 antisaccade distractibility error (ADE), 83, 86, 88 antisaccade task, 73, 77, 82, 84, 86, 199 back-up (BUS), 73, 76, 80±81 catch-up (CUS), 73, 76, 80±81 compensatory, 72, 73, 76, 80±81 generated from rest, 73±74

217

intrusive, 73 observer bias in detection, 73, 77, 85 re¯exive, 73, 74 saccadic distractibility errors, 73, 83, 84, 87 schizophrenia and, 74±78, 80±83, 84±86, 88 during smooth pursuit, 72±73, 78±82, 83 Schedule for Affective Disorders and Schizophrenia-Lifetime (SADS-L), 24 Schedule for Schizotypal Personality (SSP), 24 Selective attention, 115, 117, 119±120 Smooth pursuit eye movements, 72, 74 impaired frontal lobe functioning and, 86 saccades during, 72±73, 78±82, 83 schizophrenia and, 74, 86, 199 task methodology, 75±76 Sociodemographics of MFSP research, 24, 25±26, 116, 122 Spatial working memory, 111, 112, 113±114 Square wave jerks (SWJs), 73, 76 SSP (Schedule for Schizotypal Personality), 24 Strategy formation, 111, 112, 113±114 Stroop Colour±Word task, 117 Stroop effect, 115 Study methodology and background see Research characteristics (MFSP); Research methodology (MFSP) Subphenotypes, 10 Sustained attention impairment, 114±120 Targeting of Abnormal Kinetic Effects (TAKE) scale, 137 Tower of London test, 107, 113 Trail Making Test, 106 Twin studies contribution of genetic liability to schizophrenia, 2 on genetic overlap between ERPs, 60±61 morphometric endophenotypes, 165

218

INDEX

neurological abnormalities, 134, 148±149 overlap in genetic liability for schizophrenia and bipolar disorder, 8±9 thalamic de®cit, 186±187 Velocardiofacial syndrome, 6

Verbal memory, 97, 98, 99±100, 104 associate learing, 103 Visual memory, 97, 98±103 Wechsler Intelligence Scale-Revised (WAIS-R), 96, 116, 121 Wechsler Memory Scale (WMS), 96 Wisconsin Card Sorting Test, 86, 106