Structural brain deviations associated with schizophrenia and psychotic bipolar disorder assessed using computational morphometry

13. Neuroimaging, Structural

201

TE=6.9ms, TR=17.7 ms, 1.5ram slices no gap and volumes were determined with the MEASURE image analysis (Barta, 1997) by a rater blind to the subject diagnosis. The hippocampal region measured included the subicular complex, hippocampus proper, dentate gyrus, alveus, and fimbria, and was traced anterior to posterior on coronal slices according to anatomic boundaries described by Watson et al. (Watson, 1992). Likewise, the amygdala volumes were determined utilizing Watson criteria. Automated k-means-based segmentation was used to determine whole brain volumes (Petropoulos, 1999). Chronic patients on haloperidol had significantly smaller left hippocampi and right amygdala to brain ratio than early SZ patients when controlling for age. The haloperidol treated patients also had smaller right amygdala to brain volume than the clozapine patients when controlling for age but no differences in hippocampal volumes. There were no significant differences between the early SZ and clozapine treated patients. Thus, the consistently reported reduction of hippocampal volumes in chronic SZ may not be a trait of the illness, but may represent volume loss associated with the course of disease and/or a medication effect (i.e. current treatment with potent D2 blocking agents). Follow-up data of the treated early SZ patients may help to elucidate the course of disease from medication effect. Supported by: Mental Illness and Neuroscience Discovery (MIND) Institute,and NARSAD

CEREBRAL TISSUE ALTERATIONS AND DAILY LIFE STRESS EXPERIENCE IN PSYCHOSIS M. C. Marcelis,* I. M y i n - G e r m e y s , J. Suckling, R Woodruff', R H o f m a n , E. Bullmore, R Delespaul, J. van Os

Dept. of Psychiatry & Neuropsychology - European Graduate School of Neuroscience, Maastricht University, Maastricht, Netherlands Objective: To examine whether the total volumes of cerebrospinal fluid (CSF), cerebral grey matter and white matter were correlated with the experience of environmental stress in daily life situations. Method: Twenty-seven patients with psychosis underwent MRI scanning and a random time-sampling self-assessment technique (Experience Sampling Method (ESM)) to determine subjective daily life stress experiences. Total cerebral tissue volumes were derived from an automated segmentation procedure. Results: CSF volume was positively associated with daily life event-related stress (~=0.016, p=0.002), while the association with total white matter was negative (~=-0.013, p=0.005). The effects were independent of each other and of total cerebral volume and other confounders. No large or significant association was found with grey matter volume. Conclusion: Subjective stress experience in daily life is associated with increased CSF and reduced white matter volumes in patients with psychosis, suggesting functional significance of these cerebral measures. Supported by the Dutch Brain Society and the Dutch Prevention Fund.

AMYGDALA VOLUME IN PATIENTS WITH SCHIZOPHRENIA FROM MULTIPLY AFFECTED FAMILIES AND THEIR UNAFFECTED RELATIVES N. K. Marshall,* C. M c D o n a l d , K. Schulze, R. M. M u r r a y

Department of Psychological Medicine, Institute of Psychiatry, London, United Kingdom Several morphometric studies on schizophrenia have reported reduced volume of the hippoeampal-amygdala complex. However these structures have distinctive functions: the hippocampus has a

central role in the encoding and retxieval of episodic memory, and the amygdala is involved in emotional memory, emotional facial recognition and fear conditioning. There is some evidence for specific amygdala volume reduction in schizophrenia and the cause of this remains unclear. We sought to explore amygdala volume in schizophrenia further and its relationship to two potential aetiological factors: susceptibility genes for the illness and a putative environmental risk factor previously linked to hippocampal volume reduction, obstetric complications (OCs). We obtained MRI scans from patients with schizophrenia from multiply affected families (n=22), their unaffected relatives (n=22) and controls (n=26). The unaffected relatives included a group of subjects (n= 10) who appeared to be transmitting genetic risk to their offspring by virtue of themselves having a sibling and/or parent affected. Amygdala volume was measured using DispImage and data were analysed with multiple regression controlling for age, gender, height, whole brain volume and clustering subjects within families to account for non-independence of measures. Patients with schizophrenia had significantly reduced volume of the left amygdala (p= 0.001; B = -165; 95% CI [-262 - -69]) and right amygdala (p= 0.006; B = -176; 95% CI [-299 - -53]). There were no significant differences in amygdala volume between the relatives groups and the control group. Data on OCs were available for 33 of the subjects. There was a trend for subjects with OCs to have smaller left amygdala volume (p= 0.07; B = -165; 95% CI [-344 15]). There were no significant interactions between obstetric complications and subject group. We conclude that amygdala volume is reduced bilaterally in schizophrenia. We failed to find any evidence that this reduction is related to susceptibility genes for schizophrenia, since even those relatives most likely to carry susceptibility genes did not significantly differ from controls. However, we did find suggestive evidence that a history of obstetric complications is related to reduced left amygdala volume, a finding which echoes our previous report in an independent sample that left hippocampal volume reduction is related to obstetric complications. This study was supported by NARSAD

STRUCTURAL BRAIN DEVIATIONS ASSOCIATED WITH SCHIZOPHRENIA AND PSYCHOTIC BIPOLAR DISORDER ASSESSED USING COMPUTATIONAL MORPHOMETRY C. M c D o n a l d , * X. A. Chitnis, E. T. Bullmore, J. Suckling, R. M. Murray

Division of Psychological Medicine, Institute of Psychiatry, King, London, United Kingdom There is evidence that an overlap exists between schizophi'enia and bipolar disorder in clinical features and certain risk factors, but it is not clear whether this extends to brain morphology. We used an automated whole brain analysis technique to explore regional tissue volume changes in patients with schizophrenia and bipolar 1 disorder with a history of psychotic symptoms. MRI scans were obtained from 41 patients with schizophrenia and 37 patients with bipolar 1 disorder, who had experienced psychotic symptoms at some stage of their illness, from multiply affected families, as well as 52 controls who reflected the patient groups in terms of age, gender and parental social class. Images were optimally normalised into Talairach space for grey and white matter segmentation, using study specific tissue templates, and segmented into grey, white and CSF tissue maps using SPM99. Global tissue volume differences were analysed using clustered regression analysis. Between-group differences in the rel-

International Congress on Schizophrenia Research 2003

202 ative volume of grey matter and white matter were calculated at each voxel, and spatial statistics and permutation tests were used for inference. Patients with schizophrenia had significantly increased global CSF volumes compared to controls (B = 22.58, p = 0.02). The global tissue volumes of the bipolar patients did not differ from controls. Patients with schizophrenia had voxel clusters of grey matter volume deficit (p=0.003) involving the left prefrontal and orbitofrontal cortex, bilateral medial temporal lobe, bilateral insula, cerebellum, thalamic and right STG regions compared to controls, as well as white matter volume deficit in the left frontal and right parietal regions (p=0.006). Patients with bipolar disorder had no grey matter changes, but displayed white matter volume deficits in the right frontal region as well as bilateral parietal regions (p=0.006). This study provides evidence for marked differences in the morphological characteristics of schizophrenia and psychotic bipolar disorder. Schizophrenia is characterized by prominent deficits in subcortical and cortical grey matter volume predominantly in frontal and temporal lobes, as well as certain associated white matter regions, whereas grey matter volume is preserved in psychotic bipolar disorder, which is characterized instead by regional white matter volume deficits. Dr. McDonald is supported by the Wellcome Trust.

MRI VOLUMETRIC BRAIN MEASUREMENTS NEAR ONSET OF ILLNESS AND OUTCOME IN SCHIZOPHRENIA: A PROSPECTIVE LONGITUDINAL STUDY WITH FIVE-YEAR FOLLOW-UP E Milev,* B. C. Ho, S. Arndt, E Nopoulos, N. C. A n d r e a s e n

Psychiatry, MHCRC, UIHC, Iowa City, IA, USA A number of demographic and phenomenological variables have been identified as predictors of outcome in schizophrenia. Far fewer studies have examined the relationships between neuromorphological variables obtained at illness onset and subsequent outcome; the results of these studies have been contradictory. 123 subjects with schizophrenia received morphometric MRI of the brain close to illness onset, and were followed for an average of five years. The relationships between regional brain volumes at onset and outcome five years later were studied using multivariate analysis of covariance and correlation analysis. Outcome measures included psychosocial functioning, weeks per year receiving inpatient treatment, and persistence of severe psychotic, disorganized and negative symptoms. Temporal lobe tissue volume at onset was found to be predictive of outcome. Smaller temporal lobe gray matter volume (both left and right) near illness onset was associated with greater persistence of hallucinations during subsequent follow-up. There were no significant associations between hallucinations and any of the temporal white matter volumes (total, right, left), or between delusions and any of the temporal white or gray matter volumes. None of the other volumetric brain measures were found to be predictive. The association between initial temporal lobe gray matter volume and subsequent persistent hallucinations may help identify individuals who are at higher risk for persistent hallucinations and help guide their treatment planning. However, regional brain volumes assessed near illness onset, in general, do not appear to be indicative of subsequent outcome in schizophrenia.

13. Neuroimaging, Structural POOR INSIGHT IN THE PSYCHOSES: NEUROCOGNITIVE IMPAIRMENT OR PSYCHOLOGICAL DEFENCE? AN ANALYSIS OF COMPETING MODELS K. Morgan,* E Dazzan, C. Morgan, X. Chitnis, J. Suckling, R. Mallett, J. Left, R. M. Murray, A. David Division of Psychological Medicine, Institute of Psychiatry, Denmark Hill, London, United Kingdom The foremost explanatory models of poor insight in the psychoses focus on either neurocognitive impairment or mechanisms of psychological defence. We investigated which of these models could best explain poor illness awareness in a group of 74 first onset psychosis patients (mean age 27, 61% male, DSM-1V schizophrenia n=34, other psychosis n=40). The patients were recruited over a three-year period from an inner city area of South London as part of the AESOP first-onset psychosis study. Illness awareness was rated with David's insight schedule (SAI-E). Neurocognitive impairment was based on neuropsychological test performance and an analysis of structural MRI data. Dual echo MRI images were acquired at 1.5T and differences in grey and white matter volume between 'low' and 'high' illness awareness groups were estimated at each intracerebral voxel after registration of the images in standard space. Mechanisms of psychological defence were examined in relation to measures of the patients' self concept and their ratings for depression. Symptom data was acquired with the Schedules for Clinical Assessment in Neuropsychiatry. Socio-economic data (e.g. employment history) was also collected and included in the analysis. A stepwise multiple linear regression analysis controlling for age, sex, symptom severity and education showed bilateral grey matter deficits located at the posterior cingulate gyms (PCG)/superior parietal lobule (SPL), higher self-esteem and long term unemployment were predictive of poor illness awareness. Together they accounted for 41% of the variance in illness awareness (R Square = 0.41, p<0.001). A further examination of the data revealed that increased grey matter volume in the PCG/SPL cluster correlated with increased depression scores (r=.25, p=.030). Scores for increased depressive symptoms also correlated with reduced self-esteem (r=-.41, p<00 l ). It would appear therefore, that depression scores may mediate the link between grey matter volume in the PCG/SPL cluster and illness awareness and the link between self-esteem and illness awareness. Rather than being 'competing' models of explanation, theories of insight pertaining to neuropsychology and psychological defence are areas of inquiry that warrant an integrated approach. The study is funded by the U K Medical Research Council. We would also like to thank the Stanley Medical Research Centre for their support.

REGIONAL EXTRA-CORTICAL CSF INCREASES DESPITE ABSENCE OF SIGNIFICANT GRAY MATTER LOSSES IN SCHIZOPHRENIA K. L. Narr,* T. S h a r m a , R. E Woods, E M. T h o m p s o n , E. R. Sowell, D. Rex, S. Kim, D. A s u n c t i o n , S. Jang, J. C. Mazziotta, A. W. Toga

Neurology, UCLA School of Medicine, Los Angeles, CA, USA Increases in cortical and subcorticat CSF are widely observed in schizophrenia, although cortical gray matter reductions appear less consistently, as documented in reviews of the imaging literature. The modulating effects of biological sex and age, linked with both normal brain variation and disease processes, however, may account for

International Congress on Schizophrenia Research 2003