14. Neuroimaging, functional

14. Neuroimaging, functional

217 14. Neuroimaging, Functional STUDYING VERBAL USING fMR1 SELF-MONITORING F. Ahmad, E. Amaro, S. Williams, V. Curtis, C. Andrew, M. Brammer, R.M...

2MB Sizes 1 Downloads 51 Views

217

14. Neuroimaging, Functional STUDYING VERBAL USING fMR1

SELF-MONITORING

F. Ahmad, E. Amaro, S. Williams, V. Curtis, C. Andrew, M. Brammer, R.M. Murray, P.K. McGuire Department of Psychological London SE5 8AF. UK

Medicine,

Institute

of Psychiatry,

Verbal self-monitoring can be investigated using paradigms in which subjects speak but hear their speech after it has been acoustically distorted. Such tasks are difficult to study using WRI, as scanner noise interferes with the subjects’ ability to hear their own speech. Based on the fact that the peak haemodynamic response occurs S s after stimulus presentation, we developed a method which allows the task to be performed in silence. In the test condition the pitch of the subject’s voice was lowered by 6 semitomes, in the control condition their voice was unaltered. One hundred echoplanar images depicting BOLD contrast over the whole brain were acquired. We used an ABAB blocked design alternating between distorted speech and unaltered speech. A TR of 3 s was divided into a 1.5 s silent period during which the subject responded and 1.5 s for acquisition. Subjects had no difficulty hearing their speech or discriminating between speech of different pitch. Preliminary data indicated that reading in the presence of distortion was associated with temporal, prefrontal, and cerebellar responses. Our data suggests that this method offers a means of overcoming some of the effects of scanner noise during fMR1.

EVENT-RELATED fMR1 WITHOUT SCANNER ACOUSTIC NOISE E. Amaro Jr, M.J. Brammer, S.C.R. Williams, C. Andrew, V. Curtis, F. Ahmad, P. Monks, L.J. Reed, P.K. McGuire Department of Psychological De Crespigny Park, Denmark

Medicine, Institute of Psychiatry, Hill, London SE5 8AF, UK

The noise of image acquisition during fMR1 can interfere with subjects’ ability to hear stimuli and their own speech, and with activation in temporal cortex. As many functions implicated in schizophrenia involve verbal processing, we are developing acquisition paradigms which minimise the impact of scanner noise. Four sequences involving single trials of stimuli were tested in volunteers. Three involved single words: (1) Continuous series of volumes, TR = 2 s, interstimulus interval (1%) = 20 s. (2) Series of 5 volumes, compressed EPI, TR = 2 s, 2 s post stimulus, ISI = 13 s. (3) Single volumes, dual acquisition compressed EPI, TR = 9-l 1 s, IS1 = 20 s. Paradigm 4 used a similar sequence to 3, but the stimulus was scanner noise. During 1 there was continuous scanner noise. In 2 noise was restricted

to the acquisition period. In 3 both the task and image acquisition were performed in silence. l-3 all produced temporal activation, with responses peaking around 5 s and of greatest amplitude with 3. Scanner noise (4) also produced temporal activation peaking at 5 s, but the shape and amplitude of response were different. These data underline the confounding effects of scanner noise, but suggest that they can be reduced using novel acquisition sequences.

AFFECT RECOGNITION IN SCHIZOPHRENIA: AN fMR1 STUDY A.A. Baird, S.A. Gruber, D.A. Yurgelun-Todd Brain Imaging Center, McLean School, Belmont, Massachusetts

Hospital, Harvard 02178. USA

Medical

It has been hypothesized that schizophrenia may be associated with a disruption of the development of normal hemispheric lateralization. Structural and functional measures of the temporal lobe have consistently revealed anomalies of the left hemisphere in schizophrenia subjects. In nonpsychiatric subjects, fearful affect recognition is associated with an increase in left amygdala activation, while recognition of pleasant affective stimuli has been associated with increases in right amygdala activation. To determine whether affective processing is abnormally lateralized in schizophrenia patients, we applied fMRI to study bilateral amygdala response to a facial effect recognition paradigm, which included both fearful and happy stimuli. Ten DSM-IV schizophrenic patients and 10 healthy control subjects were examined. During recognition of fearful affect, control subjects demonstrated a significantly greater increase in left, compared with right, amygdala activation (t = 4.44, p = 0.0003). Schizophrenic subjects, however, produced a reversed pattern of activation with significantly increased right, compared with left, amygdala activation (t = -2.01, p = 0.05) during recognition of fearful affect. These findings suggest significant physiologic differences between schizophrenic patients and normal controls during emotional discrimination and labeling of fearful stimuli in comparison to happy facial affect, and provide further evidence of abnormal lateralization associated with schizophrenia.

PREFRONTAL CORTEX AND CONTEXT PROCESSING IN MEDICATION-NAIVE FIRST-EPISODE PATIENTS WITH SCHIZOPHRENIA D.M. Barth, C.S. Carter, A. Macdonald, F.W. Sabb, D.C. Noll, J.D. Cohen Department of Psychology, Washington University, Campas Box 1125, One Brookings Drive, St Louis, Missouri 63130, USA In previous work, we have hypothesized deficits in a range of tasks can be explained

that schizophrenic by a disturbance

218

in one prefrontal cortex (PFC) function: the representation and maintenance of context (a component of working memory). Although several lines of research support this hypothesis, the majority of this work has been with chronic, usually medicated, patients. Thus, it is not clear whether such deficits are present at the onset of schizophrenia or develop over time. To examine this issue, we are studying first-episode patients with and without schizophrenia. At admission, participants are medication-naive, and cognitive function is assessed 3 times (admission, 4 weeks, 6 months). Participants also complete an fMRI experiment at admission to study PFC activation during cognitive task performance (AX-CPT). Preliminary analyses confirm the presence of deficits in context representation and maintenance among medication-naive first-episode schizophrenia patients. Further, fMR1 analyses suggest that this cognitive deficit is associated with a failure to appropriately activate specific regions of PFC, despite normal stimulus-driven activation of motor and visual regions. As such, these preliminary results provide crucial information regarding the relationships between cognitive deficits and cortical activation early in the course of schizophrenia.

ABNORMAL ANTERIOR CINGULATE AND INFERIOR FRONTAL ACTIVATION DURING COGNITIVE INHIBITION IN SCHIZOPHRENIA: AN EVENT-RELATED fMR1 STUDY A. Belger, L. Fermaglich, D.C. D’Souza, J.C. Gore, J.H. Krystal Department New Haven,

of Psychiatry, Connecticut

Yale University 06510, USA

School

of Medicine,

The anterior cingulate cortex (AC) has been demonstrated to be activated in healthy subjects during selective attention tasks such as the Stroop task. This activation has been associated with response selection and inhibition of interfering incompatible information. We examined the integrity of cingulate and inferior frontal regions in schizophrenic patients during a novel inhibition task, designed to enable trial-by-trial assessment of cingulate activation during incompatible events. Eleven healthy controls and 12 schizophrenic patients were tested during a simulus-counting task. On one-third of the trials the stimuli consisted of stars; on another one-third they were digits which matched the number of stimuli (compatible), and on the other one-third the stimuli were digits which were incongruent (incompatible) with the number of stimuli. Gradient echo, echoplanar images were acquired on a 1.5 T MR system. During each of 10 experimental runs, 128 images (TR = 1.5 s, TE = 45 ms, r = 60”, 7-mm slice thickness) were acquired for each of 8 axial slices extending dorsally from the AC-PC. Segments consisting of the 5 images preceding and the 9 images following each target were excised and the time course of each voxel was convolved with a target average activation time course obtained in a prior study. Behaviorally, healthy subjects showed significant facilitation and interference effects for compatible and incompatible trials, respectively. Schizophrenic

patients showed signihcantly greater interference effects, consistent with previous reports of deficits in inhibitory functions, The fMR1 results revealed selective activation of the AC and inferior frontal/insular regions in healthy subjects during the incompatible trials but not for schizophrenic patients, suggesting that abnormal anterior cingulate and inferior frontal functions may contribute to poor inhibitory processes in schizophrenia. (Supported by NARSDA.)

MODULATION OF FRONTAL AND CINGULATE ACTIVATION TO INFREQUENT VISUAL TARGETS BY KETAMINE: AN EVENT-RELATED FUNCTIONAL MRI STUDY A. Belger, E. Kirino, S. Madonick, C.J. Gatenby, G. McCarthy, J.C. Gore, J.H. Krystal Depurtmemt New Haven,

of Psychiatry, Connecticut

Yale University 06510. USA

School

ofA4edicine.

McCarthy et al. (1997) demonstrated activation of prefrontal cingulate and parietal cortex in response to infrequently presented visual stimuli to which subjects were pressing a button. A follow-up study (Belger et al., 1998) indicated that schizophrenic subjects failed to activate prefrontal and anterior cingulate regions to infrequent targets during an identical oddball task. We propose that this deficit may be linked to abnormal NMDA receptor functions in these regions in schizophrenia. To test this possibility, we examined the effects of NMDA receptor function modulation by ketamine on the prefrontal, cingulate and parietal activation associated with target direction in healthy subjects during an oddball task. Methods: Subjects viewed a series of squares (90-94X of trials) and circles (random 6610% trials per run) presented l/l.5 s during a placebo and a ketamine infusion phase. Subjects responded with a button press to the circles only. Gradient echo, echoplanar images were acquired on a 1.5 T MR system. During each of 10 experimental runs, 128 images (TR = 1.5 s, TE = 45 ms, c( = 60”, 7-mm slice thickness) were acquired for each of 8 axial slices extending dorsally from 8 mm below AC-PC. Segments consisting of the 6 images preceding and the 9 images following each target were excised and the time course of each voxel was convolved with the average activation time course obtained in the prior study. Results: Focal activation of the middle frontal gyri (MFG), anterior cingulate (ACG), and posterior parietal regions were obtained for the targets. Activity in the MFG and ACG only were selectively modulated by ketamine administration. Implications: Thus, ketamine produced focal impairments in ACG and MFG functions, suggesting NMDA receptor deficits could contribute to frontal/cortical activation deficits associated with disorders such as schizophrenia.

219

SINGLE TRIAL ANALYSES OF AUDITORYEVOKED RESPONSE SUPPRESSION IN SCHIZOPHRENIA AND NORMAL SUBJECTS L.D. Blumenfeld,

B.A. Clementz

Department of Psychology, University of California, La Jolla, California 92093-0109, USA

San Diego,

Schizophrenia subjects have worse auditory-evoked response suppression than normal subjects in a paired-stimulus paradigm. This effect is evident for the P50, gamma band, and Ml00 responses. This suppression effect is based, however, on scores from data averaged over many stimulus presentations. One hypothesis that needs to be evaluated is whether averaged data are equally representative of individual trial responses for both schizophrenia and normal subjects. If not, then a poor suppression interpretation may not capture the nuances of group differences during a paired-stimulus paradigm. Twenty schizophrenia and 20 normal subjects were presented 120 pairs of 83 dB clicks (500 ms ISI, 9 s ITI). The Ml00 (2-20 Hz) and the gamma band (24-48 Hz) responses were scored for both Sl and S2. We used two techniques that allowed us to score data from individual trials. First, we measured brain responses using a 148-channel whole-head biomagnetometer. Second, we integrated information from all channels on individual trials by spatially filtering the data using a generalized eigenvector algorithm. Preliminary results suggest that across trials, schizophrenia patients had more variable response latencies than normal subjects to Sl, that normal subjects had more variable response latencies at S2 than at Sl, and that the groups had similar response latencies at S2.

THE INFLUENCE OF ANTIPSYCHOTICS ON THE DISTURBED FRONTAL NETWORK IN SCHIZOPHRENIA: AN fMR1 STUDY D.F. Braus, G. Ende, M. Ruf, S. Stuck, F.A. Henn NMR-Research, Central Institute D-68072 Mannheim. Germany

of Mental

Health

treated with typical neuroleptics, in comparison to atypical neuroleptic medication. From this we conclude that under atypical drugs the prefrontal activation patterns tend towards normal control values. Our results suggest that fMR1 detects an abnormal attentional network that can be more favorably modulated by atypical antipsychotic drug treatment. Further studies with a more specific fMR1 paradigm are needed to clarify this issue.

(ZI),

Recent studies showed that atypical antipsychotics seem to have a more favorable effect on cognitive deficits in schizophrenia than typical neuroleptics. To clarify this finding, we investigated with fMR1 the activation pattern (BOLDresponse) of the prefrontal lobe in neuroleptic-naive schizophrenics and in patients under stable antipsychotic medication using a well-established simple visual (6 Hz checkerboard) and auditory (drum noise) stimulation paradigm applied simultaneously. Our data show that there were no differences in the activation pattern of the predefined sensory cortices (Vl-VS, AllA2) but a significant (p i 0.05) prefrontal hypofunction in first-episode schizophrenic patients in comparison to controls, In schizophrenics under stable medication, there was a significantly (p < 0.05) reduced prefrontal BOLD response in patients

MEG RECORDINGS OF M300 IN CONTROLS WITH SCHIZOPHRENIA J.M. Cafiive, J.C. Edgar, G. Miller,

M.P. Weisend

Department of Psychiatry, Albuquerque VA Medical Center, 1501 San Pedro SE, Albuquerque, New Mexico 87108 Objectives: The P300 response has been extensively studied in schizophrenic subjects, with reduced amplitude of the P300 response being a consistent finding. While the P300 response is generally assumed to be produced by multiple generators, EEG techniques are limited in their ability to localize the generators of the P300 response. Magnetoencephalographic (MEG) source-localization techniques may prove more useful in localizing the 300 msec response. While a few MEG studies have attempted to localize the M300 response, these studies have either used MEG systems with a low number of sensors, have restricted the area over which source localization was conducted, or have reported that they were unnable to record a M300 response. This study sought to examine the feasibility of recording and of localizing the M300 response with a 122-channel system. Methods: A male sample of 5 schizophrenic and 5 normal control subjects participated in the study. One hundred and twenty-two channels of MEG and 19 channels of EEG were simultaneously recorded. MEG data was analyzed 30 msec before, to 30 msec after, the P300 response. Different sourcelocalization algorithms (least squares and R-MUSIC) were used to examine the localization of the generators during the 60 msec period. Discussion: Both chi-squared and singular-value decomposition analyses on the 60msec period confirmed that multiple dipoles were involved in generated the observed magnetic field during the 60 msec period. Preliminary estimates of dipole locations suggest parietal and temporal sources,

ANTERIOR CINGULATE CORTEX AND COGNITIVE DISABILITY IN SCHIZOPHRENIA: AN EVENT-RELATED fMR1 STUDY C.S. Carter, D.B. Barth, A. McDonald, A. Stenger, J.D. Cohen

D. Nell,

Department of Psychiatry, University of Pittsburgh, 3811 O’Hara Street, Pittsburgh, Pennsylvania 15208,

USA

220

Histopathological and functional neuroimaging studies suggest abnormalities in the local circuitry and physiology of the anterior cingulate cortex (ACC) in schizophrenia. Recent cognitive imaging studies suggest that these disturbances are related to cognitive disability in schizophrenia. The ACC has been implicated in executive processes during cognition. Recent event-related fMR1 studies indicate that this region serves an evaluative function during normal congition, detecting task conditions that indicate poor performance is likely, and signalling the need to engage strategic processes such as attention allocation to maintain performance. We studied 12 schizophrenia patients and 12 normal controls using event-related fMRI and a degraded stimulus version of the AX-CPT task. Stimulus degradation was individually titrated so that overall difficulty was comparable for all subjects. Error-related activity in the ACC, an index of the performance-monitoring function of this region, was contrasted between the groups. Controls showed the transient, response-related increase in ACC activity during errors seen in previous studies, while schizophrenia patients failed to show this response. These results suggest that one component of cognitive disability in schizophrenia is a loss of an on-line mechanism for detecting the need to engage strategic processes, reflecting a disturbance in the function of the ACC.

DIFFERENTIAL ENGAGEMENT OF THE PREFRONTAL AND FUSIFORM CORTEX DURING LANGUAGE TASKS IN SCHIZOPHRENIA V.A. Curtis, P.K. McGuire, M.J. Brammer, S.C.R. Williams, R.G. Morris, TX Sharma, R.M. Murray, E.T. Bullmore Department of Psychological De Crespigny Park, London

Medicine. SE.5 8AF.

Institute UK

of Psychology,

Insults to the brain in early development can be compensated for by plastic reorganisation of the cortex. Schizophrenia may have a developmental pathogenesis and has been associated with functional impairment of neural systems for language. Reorganisation to compensate for these deficits might be manifest as differential activation of regions within networks that mediate languages. We used functional MRI to measure cerebral blood oxygenation changes induced by performance of two periodic tasks (paced covert verbal fluency and paced semantic decision) in 5 schizophrenic patients and 5 matched controls. The patients demonstrated task-specific hypofrontality during verbal fluency. They also demonstrated significantly increased functional activation in the right fusiform gyrus during both tasks. Both the left prefrontal and fusiform cortices are part of a network that mediates in language processing. The attenuation of prefrontal and greater activation of fusiform cortex in the patients with schizophrenia might reflect reorganisation within this network consistent with a neurodevelopmental model for the disorder.

EFFECT OF ACUTE METABOLIC STRESS ON CEREBRAL BLOOD FLOW IN PATIENTS WITH SCHIZOPHRENIA I. Elman, G. Duncan, CM. Adler, L. Kestler, D.R. Gastfriend, D. Pickar, A. Breier Department of Psychiatry, Harvard Massachusetts 02114. USA

Medical

School.

Boston,

One of the more consistent findings in the brain imaging studies of schizophrenia is failure to increase cerebral blood flow (CBF) to prefrontal and temporal regions when challenged by cognitive tasks. The interpretation of this phenomenon is, however, problematic because it is unclear if decreased CBF results from a primary cortical lesion or is secondary to unrelated factors such as motivational and attentional deficits. To address this issue, we examined cerebral perfusion during metabolic stress that is free from performance confound. The metabolic stressor used was 2-deoxyglucose (2DG), a glucose analog that blocks glycolysis by competitively inhibiting cerebral hexokinase resulting in a hypoglycemic-like state. An intravenous bolus of 2DG (40 mg/kg) was administered to 17 patients with schizophrenia and 13 healthy comparison subjects, and CBF was measured with HZr50 PET. PET and MRI images were coregistered, and the size of the brain was proportionally adjusted in each axis to Talairach’s atlas. CBF group differences were assessed with the SPM analysis. Schizophrenic patients showed significantly reduced activation in the left superior, middle, and inferior temporal gyri. These findings suggest that schizophrenia is associated with a primary neurobiological pathology in the left temporal lobe. The implications of these data for the pathophysiology of schizophrenia are discussed.

THE PREDICTION

OF HALOPERIDOL

BINDING TO CENTRAL DOPAMINE D, RECEPTORS FROM PLASMA LEVELS: A PET STUDY P.B. Fitzgerald, S. Kapur, R.B. Zipursky, C. Jones, G. Remington Addiction and Mental Health Corporation, 2.50 College Street, Toronto, Ontario, M5T

Clarke Division, IR8, Canada

Central measurement of dopamine Dz receptor occupancy has been shown to predict several clinical effects of antipsychotic medication. However, measurement with PET is not feasible in routine clinical work. We have previously shown that plasma levels of haloperidol and dopamine D, occupancy induced by haloperidol are related to each other in a statistically significant fashion. The purpose of this study was to test the capacity of a previously derived equation (%D, occupancy = plasma level/ED50 + plasma level, where ED50 = 0.40 rig/ml) to predict striatal D, occupancy from plasma level in a new prospective sample of patients. Twenty-one patients participa-

221

ted in the study and were treated with low dose (l-2.5 mg/d) of haloperidol. The patients were scanned at steady state, 12 hours after the last dose. Plasma haloperidol levels, measured using GCMS, varied from 0.19-1.95 rig/ml. D, occupancy (PET measurement) varied from 38-87%. D, occupancy was reliably predicted from measured plasma levels using the above equation (Pearson’s r =0.864, p=O.O03) with minimal bias (prediction bias 3.89%). The study indicates that reliable prediction of Dz occupancy from plasma levels is possible and provides a potentially useful surrogate measure of D, occupancy for research and clinical practice.

EFFECTS OF OLANZAPINE ON REGIONAL CEREBRAL BLOOD FLOW AS ASSESSED BY FUNCTIONAL MAGNETIC RESONANCE IMAGING IN PATIENTS WITH SCHIZOPHRENIA M. Flaum, V. Magnotta, L. Sears,S. Arndt, D.S. O’Leary. D. Miller, C. Andreasen Mental Health Clinical Research Center, Department qf Psychiatry, University of Iowa, Iowa City, Iowa 52242, USA We are in the process of collecting serial fMR1 studies on a group of patients with schizophrenia on and off the atypical antipsychotic olanzapine. Patients are initially studied after at least a 3-week washout period from psychoactive medications, and then again after 3 weeks of treatment with lo-20 mg per day of olanzapine. All scans are performed on the same Siemens 1.5 Tesla Magnetom Vision MR Scanner. Each session consists of 4 fMR1 ‘runs,’ during which the subjects perform sensorimotor tasks of increasing complexity. A block diagram is used with 40-second intervals of control versus experimental task performance, each repeated 12 times. A total of 120 sets of images are acquired (TR = 4 seconds) through the entire brain (16 contiguous slices, 7 mm thick). A normal control group is also undergoing the same scanning procedure twice, 3 weeks apart, At the time of this writing, 10 patients and 4 controls have completed the protocol, with an anticipated sample size of 12 and 8 respectively. A correlational approach will be applied to the individual functional images, followed by r to Z transformations, normalization into Talairach space, and randomization analysis of patient-versus-control differences in location and magnitude of activation across the 2 time points.

AN fMR1 STUDY OF SEX DIFFERENCES IN AUDITORY VERBAL WORKING MEMORY J.M. Goldstein, L.J. Seidman, R. Anagnoson, J.M. Goodman, H. Breiter, M. Jerram, M.F. Ward, M.R. Patti, S.V. Faraone, M.T. Twang, B.R. Rosen, R. Weisskoff Harvard Medical School, Department of Psychiatry at Massachusetts Mental Health Center, Boston, Massachusetts 02115, USA

Brain regions involved in processing attention and working memory (WM) in normals have also been found to be abnormal in schizophrenia. Some of these regions are normally sexually dimorphic. Understanding normal sex differences in the use of these brain regions will contribute to understanding sex differences in schizophrenia. Functional MRI (fMR1) was used to identify sex differences in processing auditory verbal sustained attention and WM, by comparing a simple vigilance ‘AX Continuous Performance Task (CPT) with a WM task. Hypothesized sex differences were in lateral/inferior and posterior parietal cortices, anterior cingulate, hippocampus and amygdala, and lateralization of activations. Sixteen normal controls were well-matched within sex on sociodemographic characteristics and task performance; 8 patients were also evaluated. Fifteen contiguous, whole brain, 7-mm axial images were acquired with an asymmetric spin-echo T2-weighted sequence. Motion-corrected, Talairached data were evaluated using statistical nonparametric mapping. Despite similar performance, men and women differentially activated Broca’s area positively, and differentially activated, positively and negatively, paralimbic and subcortical limbic regions (e.g., anterior cingulate, insula, and amygdala), areas that integrate sensory information with internal drives and affect. Results have implications for understanding sex differences in normals and schizophrenic patients in the affective modulation of arousal when performing cognitive tasks.

A FUNCTIONAL MRI STUDY OF WORKING MEMORY IN FIRST-EPISODE SCHIZOPHRENIC PATIENTS J.M. Goodman, L.J. Seidman, M. Patti, R.D. Strous, M. Strauss, B. Caplan, J.K. Patel, S. Zimmet, B. Jenkins, A.I. Green MGH-NMR Charlestown,

Center, CNY-2, Massachusetts

149 Thirteenth 02129, USA

Street,

Working memory ( WM) impairment is a key cognitive deficit in schizophrenia and presumably reflects dysfunction of prefrontal circuitry. This observation is based largely on data from chronic medicated patients who perform poorly. The present pilot functional MRI study compared brain activation associated with WM in a small sample of first-episode schizophrenics who had received antipsychotic medication for less than 4 months, and healthy demographically matched controls. In a ‘2-Back’ WM task, subjects observed sequences of letters and indicated whether each was the same or different from the letter presented 2 items previously. In a control task, subjects indicated whether or not the letter was an X. MR signal was compared between the 2 tasks using nonparametric statistical mapping. Performance on the 2 tasks was similar between groups. Brain activations were also similar between groups, with some exceptions: schizophrenics showed less extensive activation in dorsolateral prefrontal cortex, and more activation in the inferior prefrontal cortex, frontal operculum, and supplementary motor area/anterior cingulate. Schizophrenics, but not controls, also showed large areas of negative activation

222

(baseline > control) in portions of the cingulate gyrus. These results suggest that although some schizophrenics perform adequately on a WM task, they do so by utilizing different brain regions than normals.

FUNCTIONAL MAGNETIC RESONANCE IMAGING OF SCHIZOPHRENIC AND BIPOLAR PATIENTS DURING VERBAL FLUENCY S.A. Gruber, A.A. Baird, P.F. Renshaw, B.M. Cohen, D.A. Yurgelun-Todd Brain Imaging Center, McLean School, Belmont, Massachusetts

Hospital, Harvard 02178, USA

Medical

Neuroimaging studies of cerebral activation in healthy adult subjects indicate that frontal and temporal cortical regions demonstrate the greatest change in metabolism during word production tasks. We have previously applied fh4RI to schizophrenic patients and control subjects, and found that controls showed greater activation in the left dorsolateral prefrontal cortex than schizophrenic patients, Cognitive deficits of frontal and temporal functions have also been reported to be present in patients with bipolar disorder, although noted to be qualitatively different from those manifested by patients with schizophrenia. Specifically, patients with bipolar disorder exhibit greater total speech output, and faster spontaneous talking, which is reflected in their higher number of words spoken per minute than schizophrenic patients. Conversely, patients with schizophrenia often exhibit paucity of speech and word-finding difficulties. The present study measured signal intensity changes in the dorsolateral prefrontal cortex of 18 clinically stable patients with schizophrenia and bipolar disorder during a verbal fluency task. Results indicated that despite similar performance in both groups, bipolar patients exhibited greater cortical activation of the dorsolateral prefrontal cortex than schizophrenic patients. These findings underscore the distinct differences in the underlying neurophysiology of these 2 psychiatric disorders.

THE SCHIZOPHRENIA PATIENT’S INABILITY TO EFFICIENTLY SWITCH NEURAL RESOURCES BETWEEN FRONTAL AND SENSORY/MOTOR REGIONS MAY CONTRIBUTE TO IMPAIRED WORKING MEMORY H.H. Holcomb, A.C. Lahti, D.R. Medoff, M. Weiler, Ning Ma, C.A. Tatnminga Maryland Psychiatric Research Center, University PO Box 21247, Baltimore, Maryland21228, USA

of Maryland,

The inability to recognize and recall repetitive, ambiguous stimuli is a fundamental aspect of the schizophrenic patient’s

impaired cognition. We have studied 12 normal volunteers (NV) and 18 schizophrenia patients (SZ) in behavioral and functional brain imaging protocols to better define the neural basis of this deficit. Subjects participated in a 2-tone recognition task. Initially all subjects completed 8 tone recognition blocks, 100 forced-choice, hxed intertrial-interval decisions per block. The 8 blocks represented multiple levels of difficulty ranging from extremely easy to exceptionally difficult. After extensive training, subjects provided multiple 15oxygen brain blood flow scans, positron emission tomography (PET), while making repetitive difficult tone recognitions. Analyses of the behavioral data indicate that NV became progressively slower as the difficulty of the task increased; SZ, in marked contrast, showed no tendency to delay their responses in conjunction with greater task difficulty. Brain image data revealed that NV strongly increased blood flow to the anterior cingulate only when challenged by the recognition task; SZ demonstrated a large anterior cingulate activation with a simple sensory-motor task and no additional activation when confronted with the challenging recognition task. NV also exhibited a strong positive correlation between decision response time and frontal blood flow activity; SZ exhibited a strong negative correlation in the same regions. The inability to use frontal regions in conjunction with longer time delays may partially account for SZ cognitive dysfunction.

VARIABILITY OF FRONTAL LOBE FUNCTIONAL NEUROANATOMY IN SCHIZOPHRENIA: IMPLICATIONS FOR GROUP ANALYSES J.L. Holt, J. Van Horn, J. Callicott, G. Esposito, A. Meyer-Lindenberg, M. Egan, D.R. Weinberger, K.F. Berman Unit on Integrative Neuroimaging, Clinical Brain Branch, National Institute of Mental Health/IRP, Bethesda, Maryland 20892, USA

Disorders NIH,

Background: Cortical hypofunction is the most prevalent functional neuroimaging finding in group analyses of schizophrenia. As these analyses compare intensity and spatial extent of activation, hypofunction has been interpreted as impaired recruitment of cortical neurons. Alternatively, group differences could reflect greater variability in functional neuroanatomy across individual patients, We used single-subject PET analyses to address this question. Methods: Eight controls and eight schizophrenic patients were scanned while performing an N-back working-memory task. Data were analyzed with traditional within- and betweengroup SPM. Additional analyses identified intensity and coordinates of local maximums, mean activation, spatial extent, and centroid coordinates. Results: Consistent with previous results, controls, as a group, activated dorsolateral prefrontal cortex more than patients (p < 0.0001). Surprisingly, there was no difference between local maximum z scores of individual subjects. However, the extent of and average over-activated areas were significantly

223

greater in controls (p < 0.002, and p < 0.03, respectively). Analysis of homogeneity of variance of centroid coordinates approached significance (p < 0.08), and no patients activated at the locale of the between-group local maximum, while 6 controls did. Controls consistently activated multiple common-cortical areas, but no more that 4 patients commonly activated any given voxel. Conclusion: Valid group statistics rely upon individual subjects’ signals originating from a common coordinate or, in other words, signal ‘overlap’. While between-group differences clearly exist as previously reported, it is possible that they may be overestimated by analyses that depend upon group ‘common denominators’.

RISPERIDONE RESTORES FRONTOPARIETAL ACTIVATION BY A WORKING MEMORY TASK IN PATIENTS WITH SCHIZOPHRENIA G.D. Honey, E.T. Bullmore, S.C.R. Williams, T. Sharma

W. Soni, M. Varatheesan,

Section of Cognitive Psychopharmacology, Department of Psychological Medicine. London SE5 SAF, UK

Institute

of Psychiatry,

Antipsychotic drug treatment of schizophrenia may be complicated by side-effects of widespread dopaminergic antagonism, including exacerbation of negative and cognitive symptoms due to prefrontal hypodopaminergia. Risperidone increases prefrontal dopamine in animal models. Substitution of risperidone for typical antipsychotic drugs may be associated with enhanced functional activation of prefrontal cortex. We used functional MRI and a verbal working memory task in two groups of schizophrenic patients at baseline and 6 weeks later. One group was treated with typical antipsychotic drugs throughout. Risperidone was substituted for typical antipsychotics after baseline assessment in the second group. A matched group of healthy volunteers were also studied on a single occasion. Working memory activated a network of fronto-parietal brain regions in controls. Activation was relatively attenuated in both patient groups at baseline and in the typically-treated group at 6 weeks. The risperidone-treated group had significantly greater response at 6 weeks in dorsolatera1 prefrontal cortex, supplementary motor area, and parietal cortex (p < 0.005). These data provide the first direct evidence for enhanced prefrontal cortical function in schizophrenic patients following substitution of risperidone for typical antipsychotic drugs and indicate the potential value of fMR1 as a tool for longitudinal assessment of psychopharmacological effects on cerebral physiology.

INVESTIGATION OF THE EFFECT OF TYPICAL VERSUS ATYPICAL ANTIPSYCHOTICS ON MOTOR FUNCTION USING FUNCTIONAL MRI G.D. Honey, E.T. Bullmore, S.C.R. Williams, T. Sharma

W. Soni, M. Varatheesan,

Section of Cognitive Psychopharmacology, Department of Psychological Medicine, London SE5 8AF. UK

Institute

of Psychiatry,

We have previously demonstrated abnormal lateralisations of motor systems in schizophrenic patients treated with typical neuroleptics. This study aimed to investigate the effect of risperidone on motor function. Two groups of 6 male righthanded schizophrenic patients treated with typical neuroleptics were recruited. We used functional MRI and a paced, visually cued joystick task in two groups of schizophrenic patients at baseline and 6 weeks later. One group was treated with typical antipsychotic drugs throughout. Risperidone was substituted for typical antipsychotics after baseline assessment in the second group. A matched group of healthy volunteers were also studied on a single occasion. Control subjects activated bilateral sensorimotor cortex, inferior parietal lobe, SMA, and cerebellum. Typically-treated patients showed an abnormal baseline response; however, at 6 weeks, activation was similar to controls. Risperidone-treated patients demonstrated an abnormal lateralisation of sensorimotor cortex at baseline and 6 weeks; however, there was increased activation of the SMA at 6 weeks. Risperidone did not reverse abnormal lateralisation but did improve SMA activation. This demonstrates an anatomically specific effect of risperidone, consistent with a serotonergic modulation of brain regions innervated by dopamine. The typically-treated group indicate that dynamic changes in brain activation associated with motor stimulation occur with practice.

DYSFUNCTIONAL CORTICALCEREBELLAR-THALAMIC-CORTICAL CIRCUITS IN CHRONIC SCHIZOPHRENIA MEASURED WITH POSITRON EMISSION TOMOGRAPHY J.J. Kim, S. Mohamed, N.C. Andreasen, D.S. O’Leary, G.L. Watkins, L.L. Boles Ponto, R.D. Hichwa Mental Health Clinical Research Iowa City, Iowa 52242, USA

Center,

2911 JPP,

UIHC,

Whether chronicity of illness produces progressive neural change is an important issue in contemporary schizophrenia research. Positron emission tomography (PET) offers an opportunity to visualize and measure blood flow in vivo in order to address this issue. We have previously studied first-episode, neuroleptic-naive schizophrenic patients during random episodic silent thought (REST) and reported that abnormalities

224

in blood flow are present at onset, including decreased flow in prefrontal regions and increased flow in the thalamus and cerebellum, indicating that abnormalities in distributed dysfunctional circuits are not due to chronicity or the effects of medication. In the present study we examine the same cognitive condition in a sample of 30 patients who have been chronically ill (but have been medication-free for at least 3 weeks) in order to determine if their patterns of flow abnormality are similar to first-episode patients. As in our previous study, the chronic patients showed decreased flow in prefrontal areas and increases in thalamic and cerebellar regions when compared with normal controls. These results provide evidence that a similar imbalance in corticallcerebellar-thalamic-cortical circuits (CCTCC) occurs in both first-episode and chronic schizophrenia and suggest that primary neural changes do not show a severe progression as the illness becomes chronic. Further, these results suggest that the primary neural changes may occur in the CCTCC.

ATTENUATED RIGHT TEMPORAL ACTIVATION AND CONTEXTUAL PROCESSING IN PATIENTS WITH FORMAL THOUGHT DISORDER T. Kircher, E. Bullmore, M. Brammer, A. Simmons, S. Williams, R. Murray, P. McGuire Institute of Psychiatry and King’s College School of Medicine and Dentistry, Department of Psychological Medicine, De Crespigny Park, London SE5 8AF. UK

There is evidence that the processing of contextual information is impaired in patients with positive formal thought disorder (PFTD). We used fMR1 to examine the neural correlates of tasks that required processing of sentence context in schizophrenic patients with PFTD. Six dextral male patients with prominent PFTD at time of scanning were compared with 7 healthy volunteers. Echoplanar images were acquired over 5 minutes at 1.5 T, while subjects were visually presented with sentence stems, one word at a time. Subjects read each word aloud and completed the sentence so that it made sense. In a GENERATION condition subjects had to generate, in a BASELINE condition, the final word. Images were analysed using GBAM (Brammer et al., 1997). Controls showed significantly greater activation of the right superior temporal gyrus, whereas the patients showed more signal change in the left inferior temporal gyrus. The right temporal lobe is implicated in drawing inferences from discourse, while the left inferior temporal cortex is engaged in lexico-semantic processing. The data are thus consistent with evidence that thought disorder is amounted with defective processing of linguistic context and suggest that patients process the information at more of a single-word level.

CHANGES IN rCBF AFTER ACUTE CHALLENGE WITH HALOPERIDOL OLANZAPINE IN PATIENTS WITH SCHIZOPHRENIA

. AND

A.C. Lahti, H.H. Holcomb, M.A. Weiler, A. Parwani, T. Michaelidis, D.R. Medoff, C.A. Tamminga Maryland Psychiatric Research Center, University School of Medicine, PO Box 21247, Baltimore, Maryland 21228, USA

of Maryland

In order to characterize the dynamic time course of neuronal activity induced by antipsychotic medication given acutely to schizophrenic patients, serial regional cerebral blood flow (rCBF) measurements (using PET and H,“O) were obtained over time. Schizophrenic patients with a history of partial good response to traditional antipsychotics were included in this study. They remained free of all medications, including antipsychotics, for a period of 2 weeks prior to receiving a series of scans following a single acute oral dose of 10 mg of haloperidol (n = 6) or 15 mg of olanzapine (n = 4). Scans were obtained at baseline (2) and every hour thereafter for a period of up to 6 hours post dose. For both populations, average images were constructed by contrasting baseline scans and scans obtained after medications and analyzed using SPM techniques. Acute haloperidol administration produced a significant increase in rCBF in caudate nucleus and a significant decrease in L middle temporal gyrus, R inferior frontal gyrus, and cerebellum. After acute olanzapine administration, significant increased activation was seen in the superior temporal gyrus, superior frontal medialis, and at a trend level in the cingulate gyrus. Decreased activation was found in lingual gyrus. These preliminary data indicate that the pattern of neuronal activity induced acutely with these 2 antipsychotics is vastly different, with a remarkable lack of activation of the basal ganglia with olanzapine. This technique provides a method to study and compare regional and dynamic functional changes induced by various antipsychotics.

D2 RECEPTOR BINDING OF OLANZAPINE VERSUS RISPERIDONE IN PATIENTS WITH FIRST EPISODE SCHIZOPHRENIA J. Lavalaye, D.H. Linszen, B.P.R. Gersons, E.A. van Royen Department Tafelbergweg

of Psychiatry, Academic Medical Centre, 25. 1 IO5 BC Amsterdam, The Netherlands

Olanzapine and risperidone are atypical antipsychotic drugs with few extrapiramidal side effects. This might be because of lower binding to the dopamine D2 receptor. We examined 36 patients (31 male, 5 female, mean and 21 years [16 to 281) with a first or second psychotic episode according to DSM-IV. Medication dose was stable for 4-6 weeks. SPECT scans were made 120 minutes after injection

225

of 1lOMBq 1231-IBZM. Binding potential was calculated as striatum/occipital cortex. Extrapiramidal side effects (EPS) were rated with the Simpson Angus rating scale (SA). Both patient groups showed a low specific binding of iz31-IBZM on visual inspection. Binding ratios of olanzapine and risperidone were 1.29 (kO.14) and 1.22 (kO.ll), respectively (p = 0.119). A linear correlation between dose and D2 binding was found (p < 0.01) for olanzapine. Ratios of clinically most used doses of olanzapine (15 mg) and risperidone (4 mg) were 1.35 (k0.14) versus 1.19 (kO.12) (p=O.O23). Three patients on risperidone showed moderate EPS. No relation between rZ31-IBZM binding and EPS was found. In conclusion, there proved to be little difference between olanzapine and risperidone as to their binding to the D2 receptor. Therefore, the low emergence of side effects of olanzapine seems not related to a low D2 receptor binding but mediated by a different mechanism.

TEMPORAL LOBE ACTIVATION AUDITORY HALLUCINATIONS: A FUNCTIONAL MRI STUDY B.R.

Lennox,

University Macmillan UK

DURING

S.B.G. Park, P.B. Jones, P.G. Morris

of Nottingham Department House, Porchester Road,

Furthermore, several studies have reported abnormal coordination of activity between frontal and temporal lobes during such tasks (e.g., Frith et al., 1995), though other studies have not found such abnormalities (Spence et al., 1998). To investigate this issue further, we used fMR1 to measure regional cerebral activity during a 2-back working memory task in 8 schizophrenic subjects and 8 healthy controls. Both groups of subjects exhibited activation in right and left lateral frontal cortex, supplementary motor cortex, and in cerebellum. However, the magnitude of this activation was significantly less in schizophrenic subjects in right and left frontal cortex and in cerebellum. Furthermore, the healthy controls, but not the schizophrenic patients, exhibited suppression of activity in left temporal lobe, during the 2-back task relative to rest. These findings confirm previous findings of reduced activation in schizophrenia and, in addition, indicate abnormal coordination of activity between frontal cortex and left temporal cortex.

of Psychiatry, Duncan Nottingham, NG3 6AA.

Imaging studies suggest abnormalities of temporal lobe function in schizophrenia perhaps related to a tendency to misattribute internally generated stimuli. We mapped the areas directly activated during episodes of auditory hallucination in 4 righthanded subjects with schizophrenia using continuous whole brain scanning with high-resolution fMRI at 3.0 T. Areas in the superior temporal gyrus were activated consistently during episodes of hallucination in all 4 subjects, and there was a strict temporal relation between the evolution of this activation and the subjects’ awareness of the hallucination. Activation was laterahzed to the right in 3 subjects, and to the left in 1 subject. The left dorso-lateral prefrontal cortex was also activated in all 4 subjects. These findings support a theory of failure of internal monitoring and abnormal lateralization of language in schizophrenia.

AN fMR1 STUDY OF FRONTO-TEMPORAL COORDINATION DURING WORKING MEMORY IN SCHIZOPHRENIA P.F. Liddle, A. Mendrek, A.M. Smith, K.A. Kiehl Department of Psychiatry, University of British Columbia, 2255 Wesbrook Mall, Vancouver, British Columbia V6T2A1, Canada Many studies have reported that schizophrenic patients exhibit reduced activation of frontal lobes in executive tasks.

SIMILARITIES AND DIFFERENCES BETWEEN SCHIZOPHRENIC SUBJECTS AND PSYCHOPATHS IN THE MECHANISMS OF RESPONSE INHIBITION: AN fMR1 STUDY P.F. Liddle, A.M. Smith, A. Mendrek, K.A. Kiehl, R.D. Hare Department of Psychiatry, University of British Columbia, 2255 Wesbrook Mall, Vancouver, British Columbia V6TZA1, Canada

While schizophrenia and psychopathy differ in many respects, they show some similar abnormalities of the regulation of speech, affect and behaviour. We have previously shown that both schizophrenic subjects and psychopaths produce an excess of commission errors during a Go/No Go task. In addition, both groups exhibit reduction in the negative potential recorded at frontal scalp electrodes during No Go trials (Kiehl et al., 1998). To examine the similarities and differences between the cerebral mechanisms of response inhibition in these two disorders, we employed fMRI to measure regional cerebral activity during a Go/No Go task in 8 schizophrenic subjects, 8 psychopaths and 8 healthy controls. The healthy subjects exhibited significant activation of right lateral-frontal cortex and suppression of activity in medial-frontal cortex, relative to rest. Compared with healthy subjects, the schizophrenic subjects exhibited significantly less activation in lateral-frontal cortex and significantly greater suppression in medial-frontal cortex, whereas psychopaths showed significantly greater activation in the lateral-frontal cortex and significantly less suppression in medial-frontal cortex. Thus, both disorders involve malfunction of the frontal sites engaged in response inhibition, leading to similar dysregulation of behaviour, but the mechanisms are different.

226

DYNAMIC SUSCEPTIBILITY CONTRAST MEASUREMENTS OF CEREBELLAR BLOOD VOLUME IN SCHiZOPHRENIA R.T. Loeber, A.R. Sherwood, P.F. Renshaw, B.M. Cohen, D.A. Yurgelun-Todd Program in Behavioral Neuroscience, Boston University, McLean Hospital Brain Imaging Center, Belmont, Massachusetts 02478, USA Morphometric measurements on the brains of schizophrenic patients have consistently found decreased cerebellar vermal volume and increased fourth ventricular volume. While there are contradictory findings, recent studies utilizing cognitive activation paradigms with PET and SPECT have also identified differences in cerebellar activation between schizophrenic patients and normal controls. This study took advantage of the increased contrast and spatial resolution of blood volume measurements produced by Dynamic Susceptibility Contrast fMRI to examine whether cerebellar blood volume was correlated with tissue volume in schizophrenia. Cerebellar blood volume measures were completed in 3 study groups: patients with schizophrenia, a psychiatric control group of patients with bipolar disorder, and a psychiatrically healthy control group. Patients with schizophrenia had blood volume measures that were higher than healthy controls across all cerebellar regions analyzed. In contrast, the comparison group of patients with bipolar disorder had measures of cerebellar blood volume that were lower than both the schizophrenia group and healthy controls. Morphometric measures were made from slices matched to the blood volume data and analyzed for gray matter to white matter ratios, as well as overall tissue volume. The ratio of vermis to whole cerebellum tissue volume indicated that the patients with schizophrenia had a ratio which was 24% less than healthy controls. The bipolar comparison group had a similar trend with a 19% smaller ratio than that of healthy controls, although this difference was not statistically significant. Contrary to expectation, there were no statistically significant correlations between gray and white matter ratios and cerebellar blood volume for any study group. Thus the changes observed in cerebellar blood volume in both psychiatric groups appear throughout the cerebellum, whereas morphologic changes were limited to reduced vermal volume. Furthermore. the mean differences in cerebellar blood volume measures for the psychiatric subject groups and healthy controls are in opposite direction for the 2 diagnostic groups, suggesting independent pathological processes.

FRONTAL LOBE METABOLISM AND ABERRANT NEURODEVELOPMENT: A COMPARATIVE MRS STUDY AMONG ADOLESCENTS WITH AUTISM, SCHIZOPHRENIA, AND HEALTHY SUBJECTS C. Lys, R. Findling, L. Friedman, T. Ng, P.F. Buckley Department Cleveland,

of Psychiatry, Case Western Ohio 44106, USA

S.C. Schulz, M. Xue, Reserve

University,

Magnetic resonance spectroscopy (MRS), the functional congener of structural MRI, has proved valuable in evaluating in vivo cerebral metabolism in schizophrenia. Evidence from 31P MRS studies in schizophrenia suggests dysregulation of lipid metabolism, which may signify altered synaptic pruning indicative of aberrant neurodevelopment. ‘H MRS studies suggest decrements in N-acetyl-aspartate (NAA), which may reflect cortical hypoplasia rather than neurodegeneration. Adolescent schizophrenia is an advantageous patient sample in which to examine these issues, since the potential confound of illness chronicity is minimized. Moreover, since autism is unequivocally a neurodevelopmental disorder, comparison between patients with schizophrenia and autism may illuminate the relative impact of neurodevelopmental arrest in adolescentonset schizophrenia. Patients with DSM-IFdiagnoses of autism (n = 7; mean age 14.3 years; 7 males) and schizophrenia (n = 11; mean age 16.1 years; 8 male, 3 female), and 9 normal adolescents (mean age 15.7 years; 4 male, 5 female) underwent ‘H MRS of the frontal lobes, using the following parameters on a 1.5 T Siemens magnetom system with a head coil using PRESS type Chemical Shift Imaging (CSI) pulse sequence with 16 x 16 phase encoding steps. A coronal slice was selected in the frontal lobe: volume of interest (VOI) of 72 mm x 48 mm x 12 mm (slice thickness), TR = 1.5 s, TE = 135 ms, NA (number of scans) = 2, 192 mm x 192 mm field of view (FOV), 1.728 cc voxel size. The VOI was positioned symmetrically in the frontal lobe in order to compare the spectra bilaterally. Spectral data were processed and expressed as the ratios of NAA/choline (Cho), NAA/creatine (Cr) and Cho/Cr. NAA/Cr was significantly @ ~0.05, 2-tailed test) reduced in patients compared with controls; there was no difference in metabohtes between patient groups. These results suggest that the NAA cortical decrements observed in adults with schizophrenia are also present in adolescents with schizophrenia and are comparable in magnitude to those observed in autism, a disorder of unequivocal neurodevelopmental origin. (This work was supported by the Stanley Foundation.)

DIFFERENT rCBF SCHIZOPHRENIA

PROFILES

FOR

PATIENTS

GROUPED

DICHOTIC LISTENING (FUSED TASK) BRAIN ASYMMETRIES

BY

WORD

D. Malaspina, G. Bruder, A. Berman, V. Furman, R. Goetz, J. Gorman, R. Van Heertum New York Psychiatric Columbia University,

Institute and Department New York 10032, USA

of Psychiatry,

Schizophrenia patients vary greatly in their brain asymmetries on dichotic listening word tasks, some lacking the expected right ear (left hemisphere) advantage (LHA). We hypothesized that patients subgrouped by asymmetries would differ in symptoms and rCBF. Nine controls and 16 schizophrenia patients (8 with LHA [25.8+ 17.11 matched to 8 without LHA [-4.2 * 2.81) had SPECT rCBF studies with Tc~~-HMPAO(~O Mci) during visual fixation. Images were spatially normalized into the Talairach Atlas coordinate system and then analyzed

221

for group rCBF differences using SPM. Clinically, LHAnegative patients had greater positive symptoms, higher BPRS scores, and lower Mini-Mental scores. Both groups had decreased rCBF in the right frontal lobe (middle, precentral, inferior frontal gyri), insula. Both groups had increased rCBF in bilateral parahippocampal gyri, cerebellum, fusiform. The group without LHA additionally had increased rCBF in right middle and inferior temporal gyri, uncus, amygdala, and substantia nigra, and in bilateral frontal regions (medial, precentral, paracentral) and bilateral parietal postcentral gyrus. Thus both schizophrenia groups showed decreased prefrontal and increased medial temporal blood flow compared to controls. But the group lacking behavioral evidence for a left hemisphere advantage for processing words also showed right-sided hypermetabolism in a cortical-subcortical network.

FUNCTIONAL MAGNETIC RESONANCE IMAGING OF NORMAL AND SCHIZOPHRENIA SUBJECTS DURING SACCADIC TASKS J.E. McDowell, SE. Stewart, A. Galuhn, D.L. Braff Departments California,

of Psychiatry and Psychology, San Diego, La Jolla, California

University of 92093, USA

In an antisaccade task, the subject is asked to look at a visual field and to look in the opposite direction of a stimulus placed to the left or right. On antisaccade tasks, schizophrenia patients and their first-degree relatives generate a decreased proportion of correct responses (glances away from the visual cue) compared with other psychiatric and normal comparisons subjects. This pattern of performance among schizophrenia patients is consistent with dysfunction of the prefrontal cortex and/or its linked subcortical circuitry, and is informative about a schizophrenia-related neuropathology. Functional magnetic resonance imaging (fMR1) allows neural blood oxygenation level-dependent (BOLD) signals to be observed while subjects are engaged in saccadic tasks. Extant NRI saccadic studies of normal subjects have reported increased signal strength in regions of prefrontal and parietal cortices. Here we report signal differences between normal and schizophrenia subjects. All subjects participated in two 5.5-minute functional MRI runs, during which time they were engaged alternately in 30 seconds of saccade and 30 seconds of antisaccade tasks. Functional MRI was performed using a 1.5 Tesla magnet, using echo planar sequences (TE = 60 ms, TR = 3000 ms, FOV 240 mm, 32 3-mm slices). Preliminary analysis of 9 normal and 5 schizophrenia subjects demonstrates antisaccade-related activation of prefrontal and parietal cortices in normal subjects. As predicted, this pattern of antisaccade-related activation of prefrontal and parietal cortices was decreased in schizophrenia patients. Reasons for the differential patterns of neural activation will be discussed.

AN fMR1 STUDY OF PATIENTS WITH SCHIZOPHRENIA BEFORE AND AFTER COGNITIVE REHABILITATION J.D.C. Mellers, M. Brammer, C. Reeder, J. Corner, C. Williams, T. Wykes Institute

of Psychiatry,

London

SE5 SAF,

UK

Cognitive performance in patients with schizophrenia can be improved with rehabilitation. We have investigated whether such improvements are associated with changes in regional brain function. Using multislice, echoplanar fMRI regional brain activity was compared during a control and an experimental condition which differed with respect to the demands placed on verbal working memory (the ‘two back task’). Subjects were 6 healthy volunteers and two groups of patients with chronic schizophrenia. The first group of patients received cognitive rehabilitation, the second group were given a control treatment (intensive occupational therapy). The patients were scanned before and after treatment. The healthy volunteers were also scanned twice to control for practice effects. Analvsisof arouu -. data in the normal controls revealed activation in the anterior and posterior parasagittal cortex, left temperoparietal cortex, and left dorso-lateral prefrontal cortex. The schizophrenic subjects showed an attenuation of response in all brain regions despite adequate levels of task performance. Changes in brain activation following treatment are currently being analysed.

DEFICITS IN ALLOCATION OF ATTENTIONAL RESOURCES IN SCHIZOPHRENIA INVESTIGATED USING AN AUDITORY ODDBALL EVENTRELATED fMR1 PARADIGM V. Menon, R.T. Anagnoson, G.H. Glover, A. Pfefferbaum, J.M. Ford Department University USA

of Psychiatry and Behavioral Sciences, Stanford Medical School, Stanford, California 94305-5550,

The P300 component of the scalp-recorded, event-related potential (ERP) is reduced in amplitude in schizophrenics. This reduction is thought to reflect deficits in allocation of directed attentional resources to target events. The aim of this study is to investigate the neural correlates of this abnormality with functional magnetic-resonance imaging (WRI) using methods developed in (Menon et al., NeuroReport, 1998). Eleven schizophrenics and 13 age and IQ matched normal controls were imaged while performing an auditory oddball task consisting of 80 frequent tones (1000 Hz) and 20 rare target tones (2000 Hz) with an IS1 of 8 sec. Controls showed significant bilateral activation (height threshold: 2 > 3.09, p < 0.001; extent threshold: p < 0.05) in the supramarginal gyrus (SMG), posterior superior temporal gyrus (STG) and dorsolateral prefrontal cortex (DLPFC). Schizophrenic subjects showed

228

activation of the left SMG but this activation was considerably less than that of the controls. Schizophrenic subjects showed no activation of the DLPFC, perhaps reflecting deficits in working memory updating processes. The results suggest that decreased levels of activation in these cortical regions may underlie the P300 deficits in schizophrenia and may reflect the inability to allocate directed attentional resources to salient target events, (Supported by NARSAD, Department of Veterans Affairs, and NIMH (MH30854, MH40052.)

COMPARISON OF THE EFFECT OF HALOPERIDOL AND RISPERIDONE ON REGIONAL CEREBRAL BLOOD FLOW IN SCHIZOPHRENIA D.D. Miller,

D.S. O’Leary,

N.C. Andreasen

Mental Health Clinical Research Center, Department of Psychiatry, The University of Iowa, Iowa City, Iowa S2242. USA Objective:

Recent

MR

and PET studies have found that have disturbances in various regions associated with the cortico-cerebellar-thalamic-cortical circuit (CCTCC). Our previous PET work has consistently shown that persons with schizophrenia have lower blood flow in frontal regions and higher blood flow in the thalamus and cerebellum at rest, compared to normal controls. Our current working hypothesis is that persons with schizophrenia have an anatomic and functional disruption in neuronal connectivity and communication in the CCTCC, leading to impairments in fundamental cognitive process, impairment in one or more second-order cognitive processes, and finally to symptoms of schizophrenia. In the present study we evaluated and compared the effects of haloperidol and risperidone on region cerebral blood flow (rCBF) in persons suffering from schizophrenia, with a particular interest in the various nodes of the CCTCC. Method We measured rCBF using PET with [‘so] water in a cohort of 32 rigorously diagnosed and assessed patients with schizophrenia. Each subject was scanned twice, after having been off antipsychotic medication for at least 3 weeks and while receiving a stable dose of either risperidone (n = 19) or haloperidol (n = 13). We utilized a nonparametric (randomization) analysis (Arndt et al., 1996) to directly compare the difference images (on medication-off medication). Results: We found that risperidone produced different patterns of blood flow compared to haloperidol, in informative regions of the CCTCC such as the frontal lobes, caudate, putamen, insula, and cerebellum at rest. Conclusion: Risperidone and haloperidol have significantly different patterns of blood flow in various key nodes on the CCTCC. Further work is needed to determine whether these differences in rCBFs are related to differing effects on symptoms of schizophrenia and cognitive functioning.

persons with schizophrenia

THE RELATIONSHIP BETWEEN HALOPERIDOL DOSE AND REGIONAL CEREBRAL BLOOD FLOW IN PERSONS WITH SCHIZOPHRENIA D.D. Miller,

D.S. O’Leary,

N.C. Andreasen

Mental Health Clinical Research Center, Department Psychiatry, University of Iowa, Iowa City, Iowa 52242,

of USA

Objective: In previous SPECT and PET studies we have found that typical antipsychotics increase regional cerebral blood flow (rCBF) in the caudate and putamen, and decrease rCBF in frontal regions, anterior cingulate, posterior cingulate, insula, and cerebellum. We have speculated that at least some of these effects on rCBF are related to haloperidol’s potent postsynaptic dopamine D, receptor blockade. In the present study we evaluated the effect of haloperidol on rCBF in persons suffering from schizophrenia by examining correlations between the dose of haloperidol and regional flow, postulating that Sow might reflect receptor upregulation or indirect effects of the medication on cortical function. Method: We measured resting rCBF using PET with [“O] water in a cohort of 18 rigorously diagnosed and assessed patients with schizophrenia treated with haloperidol. Each subject was scanned after receiving a stable dose of haloperidol for 3 weeks. The haloperidol doses ranged from 0.035 to 1.6 mg/kg/d. A correlational analysis between rCBF at rest and haloperidol dose (mg/kg/d) was completed, calculated on a pixel-by-pixel basis. Results: We found that blood flow in temporal lobes, motor strip, and hippocampus was positively correlated with haloperido1 dose. Conversely, blood flow in frontal lobes, thalamus, insula, and cerebellum was negatively correlated with haloperido1 dose. Conclusion: It appears that dopamine D, blockade affects blood flow differently in different brain regions. Surprisingly, blood flow in the basal ganglia, an area rich in dopamine D, receptors, was not correlated with haloperidol dose. Further work is needed to determine whether these correlations are related to changes in specific symptoms, cognitive function, and/or adverse effects.

THALAMIC ACTIVITY AND CLOZAPINE RESPONSE IN TREATMENT-REFRACTORY PATIENTS V. Molina, J.L. Lampreave, R. Monk, J. Sanz, M. Desco, F. Collazos, C. Benito, F. Maldonado, T. Palomo, J.L. Carreras Department 47. 28009,

of Psychiatry, Madrid, Spain

Hospital

Gregorio

Many but not all treatment-refractory respond to clozapine. In previous SPECT found that subcortical, specially thalamic,

Manafidn.

Ibiza,

schizophrenics studies our group perfusion while on

229

classical neuroleptics discriminated between the patients that subsequently responded and those who did not respond to clozapine. In order to improve the reliability of these results, we studied an entirely different sample (n = 30) during classical neuroleptics and after 6 months of clozapine (dose 400 to 600 mg/d) with MRI and F,,-FDG PET. We compared the thalamic glucose metabolism between responders and nonresponders to this drug. The SPECT results were confirmed, in that glucose metabolism while on classical neuroleptics was higher in the responder group. The thalamic metabolic rate decreased in the responders, but did not significantly change in nonresponders. These data suggest the possibility of increasing our knowledge of schizophrenia by looking for a relationship between regional activity profiles and pharmacological response to given drugs.

STUDYING THE CEREBRAL LATERALIZATION OF LANGUAGE PROCESSING IN SCHIZOPHRENIA FUNCTIONAL MRI

A PET STUDY OF SPATIALLY-DIRECTED ATTENTION IN PATIENTS WITH SCHIZOPHRENIA D.S. O’Leary, N.C. Andreasen, G.L. Watkins, R.D. Hichwa University USA

of Iowa Hospitals

L.L. Boles Ponto,

and Clinics,

Iowa

Iowa

52242,

We have previously reported that patients with schizophrenia have rCBF abnormalities while attending to spatially-localized auditory and visual stimuli. We have typically reported data from subtractions using relatively ‘low-level’ comparison tasks; i.e., simple reaction time or eyes-closed rest, and have included data from all subjects demonstrating adequate performance. The present study reports data from a selected subset of patients (n = 8) with performance equivalent to control subjects. This analysis furthermore utilized a more challenging comparison condition that isolated spatial attention effects. Statistical com-

parisons of the between-group-difference

USING

City,

formed separately for auditory Some sensory-modality-specific

and visual differences

images were perattention conditions. were observed; i.e.,

less activation in left auditory cortices, lenticular nuclei, and

P.J. Monks, L.J. Reed, S.C.R. Williams, C. Andrew, M.J. Brammer, E.T. Bullmore, R.M. Murray, P.W. Woodruff. P.K. McGuire

inferior frontal lobe in auditory conditions. In the visual conditions, less activation was observed in the L pulvinar and R inferior frontal lobe. Interestingly, in both auditory and visual conditions, patients had less attention-related activation in the anterior cingulate, R pulvinar nucleus, and the dentate

nucleus of the R cerebellum. These data suggest that patients Department of Psychological London SE5 SAF, UK

Medicine,

Institute

of Psychiatr.y,

Language processing is a major focus of functional imaging research in schizophrenia. Using functional MRI, we have previously demonstrated in schizophrenia-impaired lateralization of temporal cortical activation when listening to speech. The current study examined higher-level language functions using word and reverse-word stimuli (in blocks) to distinguish activation associated with complex phonological processing from higher-level linguistic processing. An alternative design using single word and reverse-word stimuli was employed for comparison with the principal block design experiments. Alternating pairs of stimuli (blocks of spoken words, reversewords, and silence) were presented in three 5-minute experiments. Functional MRI data were acquired using a 1.5 T GE Signa scanner (TE 40 ms, TR = 3 s and analysed using generic brain activation mapping. Single words and reverse-words were presented at inter-stimulus interval of 20 s to detect eventrelated activation (TR = 2 s). Preliminary findings (n = 4) suggest that both words and reverse-words strongly activate primary auditory cortices bilaterally, and similar results emerged with single words. Comparison of words with reverse-words in the block design located activation in phase with words in the left inferior frontal and left posterior superior temporal regions, We are currently analysing the responses to these tasks in a large group of patients and healthy volunteers,

with schizophrenia have deficits in a cross-modal, spatialattention system that includes the cerebellum. A new model of cerebellar function consistent with these findings is discussed.

PET IMAGING OF VULNERABLE COGNITIVE PROCESSES IN SCHIZOPHRENIA: DO EARLY COMPENSATORY CHANGES PRECEDE COGNITIVE DECLINE? B.T. O’Sullivan, G.F. Johnson, B. Woodham, N. Barrett, G. Smith, M. Large, P. Michie, D. Kavanagh, F. Karayanidis, B. Horwitz Departments of Psychiatry and Physiology, Health and University of Sydney, Australia

Central

Sydney

Switching of attention and reloading or updating of working memory (tests of ‘fluid intelligence’) are especially vulnerable in people with schizophrenia (O’Sullivan et al., 1997). The present study was designed to compare the dynamic correlations between prefrontal activation (ArCBF) and task performance

(reaction time and accuracy) in patients recently diagnosed with schizophrenia relative to normal control PET was used while 10 normal control subjects diagnosed and high-functioning schizophrenic formed a battery of visual attention and working

subjects. I50 and 10 recently patients permemory tasks

230

(Smith et al., 1998). The normal control and patient groups were matched for age, gender, years of education, and task performance. All patients satisfied D&W-IV diagnostic criteria for schizophrenia with an onset of illness within the last 5 years. Psychophysical data were obtained for each subject on all tasks. Image data were analysed using SPM and Analyze with MR image coregistration to identify areas of ArCBF for each image subtraction. Correlation analyses (Horwitz et al., 1994) of ArCBF and psychophysical data were also performed in both groups. Both groups activated specific regions in the orbitofrontal cortex and the right dorsolateral prefrontal cortex during switching of attention and reloading of working memory. In the schizophrenic group the left dorsolateral cortex was also activated. In addition, there was a positive correlation between reaction time and prefrontal rCBF change in the patient group that was not present in the normal subjects. We consider these findings to indicate brain compensatory mechanisms that allow patients to maintain a level of task performance equivalent to the normal group. These findings may be developed further as a neuroimaging marker for the diagnosis of early schizophrenia and/or to monitor treatment response.

EFFECT OF WISCONSIN CARD-SORTING TEST ON REGIONAL CEREBRAL BLOOD FLOW IN FIRST-EPISODE SCHIZOPHRENIC PATIENTS, AGED 1O-45 K. Pagsberg, T. Mackeprang, B. Fagerlund, B. Glentheij, R. Hemmingsen Department Copenhagen,

of Psychiatry, Denmark

Bispebjerg

University

Hospital,

In ongoing clinical studies we are investigating a group of first-episode schizophrenic patients (age lo-45 years) who are either drug-naive or newly medicated with antipsychotics. In a subsection of the study we are measuring psychopathology (PANS& SANS/SAPS), age, gender, and functional magnetic resonance imaging (WRI). Subjects are MR-scanned in a 1.5 T Siemens Vision Scanner. A Tl-weighted sequence (MPRAGE) and a T2 and proton density-weighted sequence (double-spin echo) are acquired for the structural scans. fMRI is performed using echo-planar sequences. Activation during the scanning procedure is done by running a computer version of the Wisconsin Card Sorting Test from a pc-program projected to a mirror in the headcoil. The paradigm runs in both an AB-design, where the paradigm shifts between active problem solving and a rest period, and a random-event design, where the signals related to set-shifting are isolated. In order to visualise larger cerebral arteries and veins, a flow-sensitive sequence is applied.

PET STUDIES OF LEXICAL IN SCHIZOPHRENIA

PROCESSING

J.V. Pardo, P.J. Pardo, H.D. Temporini, J. Uecker Cognitive Neuroimaging Unit (IlP), Veterans Aff&irs Medical Center, One Veterans Drive, Minneapolis 55417. USA A cardinal symptom of schizophrenia consists of an abnormality in language which manifests itself clinically as a disorder of the form of thought. We and others studied previously some basic aspects of language processing in normal control volunteers. The simplest paradigm employs processing of single words, commonly referred to as lexical processing. Here, we present our preliminary data concerning the functional pathology of lexical processing in schizophrenia. Controls and patients with schizophrenia (mostly medicated with atypical antipsychotics drugs but still symptomatic) were studied with PET measurements of regional cerebral blood flow while resting with eyes closed; while looking at common, concrete nouns; while reading words aloud; and while generating verbs appropriate to the nouns. Clinical ratings included the BPRS, SANS, and SAPS. Data were analyzed pixelwise by intersubject averaging after stereotactic transformation. Abnormalities in patients with schizophrenia arose even in the contrast between the conditions of viewing nouns and of resting with eyes closed, with aberrant activation in patients with schizophrenia localizing to the ventral inferotemporal cortex. This region was previously implicated in the processing of visual objects (i.e., the ‘what’ pathway) and of lexicosemantic information. Other contrasts revealed widespread abnormalities in the functional architecture of lexical processing in the group with schizophrenia. The ability to visualize abnormal lexical processing in schizophrenia may lead to an understanding of the pathophysiology of language dysfunction in these patients. (Supported in part by the Department of Veterans Affairs, NARSAD, and the Kaul Fund.)

INVESTIGATION OF THE NEURAL CORRELATES OF THREAT PERCEPTION IN PARANOID SCHIZOPHRENIA: AN fMR1 STUDY M.L. Phillips, L. Williams, C. Senior, E.T. Bullmore, M. Brammer, C. Andrew, S.C.R. Williams, AS. David Department of Psychological 103 Denmark Hill, Londoh

Medicine, Institute SE5 8AZ. UK

of Psychiatry,

Functional neuroimaging studies investigating the neural correlates of anger and fear, basic emotions experienced when threatened, have highlighted the role of the amygdala. We employed functional fMR1 to investigate the neural correlates of these two emotions and of a non-threat-related emotion, disgust, in paranoid schizophrenics (PS) (DSM-IV criteria) (n = 5), non-paranoid psychotic patients (NP) (diagnoses: bipolar affective disorder and schizophrenia; n = 5), and normal controls (NC)

231

(a = 6). We aimed to demonstrate greater amygdala activation by fearful and angry faces in PS than controls, supporting theories of increased threat perception in PS, but no amygdala activation by faces depicting disgust in all subjects. In three 5 minute experiments, faces depicting either fear, anger, or disgust were presented in contrast to neutral faces. Subjects decided upon the sex of each face by button-press, In NC and PS, faces depicting fear or anger but not disgust activated the amygdala. In PS, there was increased amygdala activation compared with NC and NP for angry faces. In NP, only faces depicting disgust activated the amygdala, suggesting that these patients have difficulty in recognising threatening expressions. The findings overall support studies implicating the amygdala in threat perception and suggest increased sensitivity to threat in PS.

DISTINGUISHING PREDOMINATELY NEGATIVE SYMPTOM SCHIZOPHRENIA WITH FDG PET S.G. Potkin, G. Alva, D. Keator, K. Fleming, R. Bera, D. Carreon, L. Kranz Department Universit?,

of Psychiatry, of California,

Irvine Hall Room 166, Irvine, California 92697-3960,

USA

The importance of negative symptoms in schizophrenia research has increased as more effective treatments have become available. We have used FDG PET to identify brain areas that distinguish negative-symptom schizophrenia. Fourteen patients with DSM-IV schizophrenia entering clinical trials were PETscanned after a 5-10 day, neuroleptic wash-out period while performing the degraded CPT-activation task. Subjects were divided into predominately negative symptom (PNS) or predominately positive symptom (PPS) groups on the basis of their baseline PANSS scores. PNS subjects (n = 7) had PANSS-negative subscale scores greater than 20 and greater than their positive subscale scores. PNS subjects also had a global negative-symptom rating of at least ‘moderate’ and absence of depression. PPS subjects had PANSS-positive subscale greater than negative substale. The major PET findings revealed decreased metabolism in the frontal and temporal cortex of the PNS patients, compared to the PPS patients. Specifically, the dorso-lateral prefrontal cortex, (p < 0.02), and the lateral temporal (p < 0.03) and the posterior cingulate (p < 0.03) were significantly decreased in the patients with predominately negative symptoms. This study suggests that the prefrontal cortex and related structures play an important role in negative symptoms.

A FUNCTIONAL OF VISUAL- AND TASKS REVEALS MEDIOFRONTAL MEDIOTEMPORAL SCHIZOPHRENIC J. Quak,

A.M.J.

Paans,

Department of Psychiatry, PO Bo.x 30.001, 9700RB

NEUROIMAGING STUDY MEMORY-SEARCH A PATTERN OF HYPOACTIVITY AND HYPERACTIVITY IN PATIENTS R.J.

van

University Groningen,

den

Bosch,

J. Korf

Hospital Groningen The Netherlands

Frontal activation was studied in a group of twenty-two schizophrenic patients and twenty healthy subjects (controls). Previous studies suggest different frontal activation in patients with schizophrenia compared to controls. Findings seem to be related to task difficulty, performance, and patients’ characteristics. With Hai50-PET we studied brain activation while patients and controls performed three visual- and memory-search tasks and varied in difficulty. In all three tasks a similar pattern of relative cerebral mediofrontal hypoactivity and mediotemporal (including hippocampus) hyperactivity of patients compared to controls was observed. As a group, patients performed like the control group on the easiest task. However patients performed worse on the most difficult task. This somewhat reflected less mediofrontal activation in the schizophrenic group compared to the controls. Research is ongoing on the effect of performance within the patient group compared to the healthy subjects and on the effect of symptom profiles on the observed pattern of mediofrontal hypoactivity and mediotemporal hyperactivity

LEFT HEMISPHERE ACTIVATION WITH A SET-SHIFTING TASK IS REDUCED IN MEDICATED SCHIZOPHRENIC PATIENTS, BUT PERFORMANCE IS NOT AFFECTED N.F. Ramsey, H.A.M. Koning, R.S. Kahn Department of Psychiatry, University Heidelberglaan 100. 3584 CX Utrecht,

Hospital Utrecht, The Netherlands

3D fMRI was used to examine brain activity associated with set-shifting in schizophrenic patients. Eight patients on atypical neuroleptics and thirteen healthy controls were scanned during engagement in a set-shifting task that involves selection of one of two visual stimuli (presented together every 3.5 s) which differ on four features (‘rules’). The correct response depends on the active rule, which has to be discovered (shifting of set) and then applied (maintaining set). The rule is changed repeatedly without warning. Stringent statistical analyses focussed on (1) the continuous-performance aspect (task versus rest, reflecting general activity), and (2) the set-shifting aspect (setshift versus maintain, reflecting shift-specific activity). Both general and specific activity was located mainly in frontal and parietal structures. Relative to controls, patients exhibited significantly lower general activity in left-frontal cortex (BA 6, 9/46) and in superior parietal (BA 7) and anterior cingulate (BA 24), bilaterally. Task performance and specific activity were identical for the groups, indicating normal processing during set-shifting. The results indicate that reduced taskrelated, general activity in patients (mainly in left-frontal cortex) does not affect shifting-specific activity nor performance on the task. This may reflect increased brain activity in these patients during rest, in selected structures.

232

MEG EVIDENCE OF REVERSED CEREBRAL ASYMMETRY IN SCHIZOAFFECTIVE DISORDER M. Reite, P. Teale, J. Sheeder, D.C. Rojas, D. Arciniegas, J. Walker Biomagnetic Imaging Laboratory, Department of Psychiatry, University of Colorado Health Sciences Center, Denver, Colorado 80262, USA Abnormal cerebral asymmetry is observed in schizophrenic psychoses and may be a feature of psychotic disorders generally. These studies often group schizoaffective and schizophrenic patients together, despite disagreement in the literature over the extent to which these disorders are biologically similar. Prior to assuming a homogeneity in underlying neurobiology, defining the nature of such findings in schizoaffective disorder alone is necessary. The present study was designed to investigate an MEG measure of cerebral asymmetry of the somatosensory cortex in schizoaffective disorder. Our subject population consisted of 32 subjects, 16 with schizoaffective disorder, and 16 controls. Equivalent current dipole (ECD) generators in both hemispheres were estimated for the 20-ms-latency somatosensory-evoked field (M20) component associated with contralateral median-nerve stimulation. Controls demonstrated asymmetry in M20-ECD location (right anterior to left), and schizoaffective subjects demonstrated a reversal (left anterior to right). This may imply a reversal of the underlying neuroanatomical structure, the post-central gyrus, but this will have to be evaluated directly through morphological analysis. Anomalous cerebra1 asymmetry may be an expression of underlying genetic abnormalities among patients with psychosis, and neurobiological methods that demonstrate such asymmetry may prove to be both scientifically and clinically useful indices of psychosis.

SEX, SYMMETRY, M.

Reite,

D.

Arciniegas,

Department of Psychiatry, Sciences Center, Denver,

AND PSYCHOSIS P. Teale,

D. Rojas,

University of Colorado Colorado 80262, USA

J. Sheeder

The 20-msslatency somatosensory evoked field (M20) component also demonstrates asymmetry in normals, being further anterior in the right hemisphere. Sex-related differences in M20 asymmetry have not yet been described. In patients with schizoaffective disorder, and psychotic (but not nonpsychotic) bipolar disorder, there appears to be a reversal of symmetry, but this reversal appears to be present in both males and females. Anomalous asymmetry may characterize psychosis, but the sex-specific nature of the asymmetry may be differentially expressed in auditory and somatosensory systems. (Supported by USPHS MH47476.)

REDUCED FRONTAL ACTIVATION DURING VERBAL FLUENCY IN SCHIZOPHRENIA: AN fMR1 STUDY M.A. Reveley, M. Horsfield, D. Wilcock, J. Burke, P. Morris Departments of Psychiatry and Radiology, University of Leicester, Leicester LE2 7LX.

UK

Method: Five male dextral-schizophrenic patients and five volunteers matched for demographic variables were given a non-paced silent generation of words task, beginning with a visually presented cue letter. This alternated with non-paced, silent repetition of visually presented neutral words. Echoplanar images were acquired at 1.5 Tesla for six slices. Brain activation maps were constructed by nonparametric hypothesis testing. Results: Compared to controls, schizophrenics showed significantly reduced power of response @ < 0.0001) in the left inferior and middle frontal gyri during word generation. There were no significant differences between schizophrenics and controls during word repetition, or in any other brain areas imaged during either word generation or word repetition. Discussion: These results further confirm the frontal lobe deficit found in schizophrenia using the new technique of WRI. Further research will explore larger samples in different treatment and clinical conditions. The safety and noninvasive nature of flvlR1 enables repetitive functional imaging of large samples.

Health

Schizophreniform disorders specifically, and perhaps psychotic disorders generally, appear to be associated with anomalous brain asymmetry. Sex may also influence the nature of cerebral asymmetries. The source location of the lOO-ms-latency auditory evoked field (MlOO) is normally asymmetrical, with sources further forward in the right hemisphere, and more so in males. Ml00 abnormalities appear to be emphasized in the left hemisphere of males, and in the right hemisphere of females, with schizophrenia. Specifically, the normal Ml00 asymmetry is diminished in males with schizophrenia, and is either increased or no different from normal in females with schizophrenia. The Ml00 is generated in the superior temporal gyrus, whose volume appears reduced in males, but not females, with schizophrenia.

ALTERATIONS IN TONOTOPY IN THE AUDITORY CORTEX IN SCHIZOPHRENIA AND CORRESPONDING CHANGES IN PERCEPTUAL DISTANCE: MAGNETIC SOURCE IMAGING AND PSYCHOACOUSTIC EVIDENCE D.C. Rojas, J.L. Sheeder, P.D. Teale, M.L.

Reite

Department of Psychiatry, Sciences Center, Colorado

Health

University of Colorado 80262, USA

Tonotopy, the spatial mapping of frequency in the cochlea, is the dominant level of organization in the auditory cortex of

233

the mammalian brain. We hypothesize that some of the previous auditory perceptual findings in schizophrenia relate to changes in this fundamental organizational scheme. Using a 37-channel biomagnetometer system, we have studied 11 adult patients with schizophrenia and 10 comparison subjects. A single equivalent current dipole (ECD) was fit to the lOO-ms latency component of the auditory-evoked field to pure tones spanning 5 octaves starting at 250-Hz. Both hemispheres were studied in separate recording sessions, and the mean inter-ECD spacing between adjacent frequency generators was computed. Data on perceptual distances on the same frequencies were also obtained by means of magnitude estimation. In comparison subjects, the ECD locations were linearly and approximately equidistantly spaced along the long axis of the auditory cortex, as determined by MRI. Similarly, perceptual distances were also modeled well by a simple linear regression and correlated with ECD locations. In patients with schizophrenia, however, no such linear relationships were observed, indicating an alteration of their tonotopic and perceptual organization. These results may have implications for fundamental auditory perceptual experiences of people with schizophrenia.

LEFT HEMISPHERE UNDERACTIVATION IN SCHIZOPHRENIA DURING A MENTAL STATE ATTRIBUTION TASK: AN fMR1 STUDY T.A. Russell, K. Rubia, E.T. Bullmore, S.C.R. Williams, T. Sharma Section of Cognitive Psychopharmacology, Psychiatry, De Crespigny Park, London

Institute SE5 8AF,

of UK

It has been suggested that schizophrenics show a deficit in representing the mental states of others. We hypothesized that a performance deficit on a mental state attribution task would manifest as reduced functional neuro-activation. Five dextral schizophrenics and 7 age- and handedness-matched healthy controls were recruited. The paradigm involved a control condition where subjects decided whether pairs of eyes were male or female and an experimental condition where subjects decided which of two complex mental state terms best described the intention or desire of the eyes (Mental State Attribution, MSA). Schizophrenics were significantly impaired in their performance on the MSA task @ < 0.05). In controls, it elicited significant activation in the left middle and inferior frontal gyrus extending into anterior insula and in left middle/superior temporal gyrus. Schizophrenics showed activation in left inferior frontal gyrus and left middle temporal gyrus. The activation in the patient group was significantly reduced in a group difference analysis (ANCOVA, p < 0.01). Normal subjects activated a left hemisphere neural network involving inferior frontal lobe and temporal lobe for MSA. Impaired performance on the MSA task in schizophrenics was associated with reduced activation in this left hemispheric network. This is consistent with left temporal lobe dysfunction in schizophrenia.

SEMANTIC, PHONOLOGICAL, AND EPISODIC MEMORY PROCESSING IN SCHIZOPHRENIA: FUNCTIONAL MRI ACTIVATION PATTERNS INDICATE A NEED FOR NEW MODELS OF DYSFUNCTION A.J. Saykin, L.A. Flashman, T.W. McAllister, S.C. Johnson, S.J. Guerin, C.J. Brown, M.B. Sparling, C.H. Moritz, A.C. Mamourian, R.M. Vidaver Brain Imaging Laboratory, Departments of Psychiatry and Radiology, Dartmouth Medical School, Lebanon NH03756 and New Hampshire Hospital, Concord, New Hampshire, USA To examine the neuroanatomic substrate of memory dysfunction, 10 medicated patients (DSM-IV schizophrenia or schizoaffective) and 10 controls were studied with whole brain EPI at 1.5 Tesla using a local gradient coil (Medical Advances). Auditory activation tasks alternated between rest and one of four decision conditions regarding properties of word pairs: phonological (pseudo-word-match); semantic (category-exemplar and function match); and episodic recognition (old/new). Scans were motion corrected, spatially normalized, and analyzed with a random-effects model (SPM96). Task performance of patients was impaired for semantic and episodic memory. On the phonological task, patients were >95% accurate yet showed extensive fronto-temporal activation that contrasted with bi-temporal foci in controls. During semantic and episodic processing, controls showed focal left IFG, bilateral STG, and cingulate activity. By contrast, patients showed both spatially extended, left fronto-temporal and cingulate foci, and activation of right-cortical regions. BPRS severity correlated with cingulate and medial SFG across tasks. Recognition performance correlated with right lateral-frontal activity in controls only; for both groups performance was related to right anterior insula-activation. Multiple patterns of abnormal memory activation in schizophrenia (increase/decrease, spatial extent, shifted foci, contralateral recruitment) indicate a need for new, more comprehensive models of neural-processing deficits.

CORRELATION OF CEREBRAL BLOOD FLOW AND VERBAL WORKING MEMORY PERFORMANCE USING FUNCTIONAL MRI: IMPLICATIONS FOR SCHIZOPHRENIA RESEARCH T. Sharma, G.D. Honey, E.T. Bullmore Section of Cognitive Psychopharmacology, Department Psychological Medicine, Institute of Psychiatry, London SE5 8AF. UK

oj

Deficient working memory is fundamental to contemporary neurocognitive models of schizophrenia. The functional neuroanatomy underlying working memory has been studied in

234

normal controls and in schizophrenic patients. However, the relationship between fMR1 signal and task performance is not fully understood. We used multislice echoplanar functional MRI to study 22 right-handed, male healthy volunteers. Images were acquired over a S-minute period at 1.5 Tesla. Subjects were required to press a button to indicate the visual presentation of the letter ‘X’ (control condition) and the repetition of any letter presented two previously (working memory condition). Reaction time and target accuracy were monitored during the task. Generic brain activation maps were constructed with reaction time as a covariate. Subjects were able to perform the task to a high level, (mean accuracy = 96%). A strong positive correlation was observed between blood flow and reaction time @ < 0.01) in the posterior parietal cortex. Posterior brain regions are thought to subserve the modality-specific slave stores of working memory. This study indicates that task performance efficiency, as indicated by reaction time, correlates with activation of posterior brain regions underlying verbal maintenance. This has implications of interpretation of future studies in clinical populations.

MONITORING AT DIFFERENT

INNER SPEECH PRODUCED RATES

S.S. Shergill,

Murray,

Department De Crespigny

R.M.

of Psychological Park,

London

P.K. Medicine, SE5 8AZ,

McGuire Institute UK

of Psychiatry,

Background: Positive psychotic symptoms, such as auditory hallucinations, in schizophrenia have been postulated to occur as a result of failure to monitor one’s own thoughts. Production of inner speech may involve concurrent modulation of the temporal cortices. We investigated the neural activity underlying changes in the rate of inner speech production using functional MRI. Methods: Seven normal right-handed male subjects were scanned using the IOP 1.5 Tesla magnet during covert articulation of the single word REST at varying rates in a standard ABAB design. The different rates used were 1 word per second and 1 word per 4 seconds, compared within 5-minutes experiment, each condition alternating every 30 seconds. Analysis was carried out using the standard IOP analysis software. Results: Preliminary analysis suggests that there is increased activity in the left inferior frontal gyrus/insula (BA 45), anterior and middle cingulate gyri (BA 32/24), and bilateral temporal gyri (BA 21/22) during the faster rate. The slower rate demonstrated right lateral inferior parietal lobule (BA 40) and anterior cingulate gyrus (BA 32) response. Conclusions: The most parsimonious conclusion is that there are modulatory changes in the bilateral temporal cortices in response to increased rate of generation of inner speech. The changes in the cingulate may be in response to increased attentional demands in accordance with increased rate; further examination of the phases of response may clarify the nature of modulation further.

NEUROPHYSIOLOGICAL CORRELATES DISTURBED FACIAL AFFECT RECOGNITION IN SCHIZOPHRENIA M.

Streit,

Department DiisseldorJ

W. Wiilwer, of Psychiatry, Germany

OF

W. Gaebel Heinrich-Heine-University

The main objective of the study was to identify neurophysiological correlates of the well-documented facial affect recognition deficit in schizophrenic patients. Electroencephalography and magnetoencephalography were used to measure brain activity in schizophrenic patients and healthy controls by means of event-related potential designs. During the neurophysiological measurements, subjects performed a facial emotion-recognition task and two control tasks (recognition of complex objects and recognition of blurred objects). The EEG results revealed clear differences in amplitude (N2 latency range) recorded over frontotemporal scalp regions between patients (N = 16) and controls (n = 16) during facial affect recognition. Further differences were found in the 300&500 ms latency range. First analyses of the MEG data of five healthy volunteers showed that recognition of facial expressions of emotion activates the inferior prefrontal cortex, amygdala, and different parts of the temporal cortex in a relatively consistent time sequence. In conclusion, our results show that EEG is able to detect neurophysiological correlates of the facial affect recognition deficit in schizophrenia. MEG can identify critically involved anatomical structures and seems to provide first insights into the interplay between these structures, which might be critically relevant for understanding the origin of the deficit.

AUDITORY-VISUAL SPEECH PERCEPTION IS ABNORMAL IN SCHIZOPHRENIA: AN fMR1 STUDY S.A. S.L. S.C.

Surguladze, G.A. Calvert, M.J. Brammer, Rossell, V.P. Giampietro, E.T. Sullmore, Williams, R. Campbell, A.S. David

Institute UK

of Psychiatry,

De Crespigny

Park,

London

SE5 8AF.

The majority of investigators regard speech perception as a cross-modal process integrating meaningful stimuli. Disturbances of this process could have relevance to the understanding of paranoid-hallucinatory states in schizophrenia. In this study of schizophrenia we addressed the issue of functional neuroanatomy of auditory and visual speech perception. Fourteen patients with schizophrenia and 6 normal controls were examined using WRI. Visual stimuli included a videotaped human face with still lips (A-condition), silently pronouncing numbers (B), or producing senseless, closedmouth movements (C). In an auditory-alone condition, Ss listened to numbers through headphones. The results of fMRI scanning showed between-group differences in BOLD response

235

in tasks in which Ss attended to senseless lip movements. In particular, schizophrenia patients activated left superior temporal, auditory association/inferior parietal (Brodmann areas 22, 40/42 respectively) regions, which were also activated during the auditory-alone condition. The results of the study show abnormal cross-modal interactions between auditory and visual stimuli in schizophrenia. This could underlie certain symptoms, including auditory verbal hallucinations and ideas of reference.

FUNCTIONAL ANATOMY OF SCHIZOPHRENIC SYMPTOMS: A NEUROBEHAVIORAL PROBE OF AN EMOTIONAL RESPONSE S.F. Taylor, I. Liberzon, R.A. Koeppe, L.R. Decker, R. Tandon Department of Psychiatry, University Michigan 48109-0118. USA

of Michigan,

Ann Arbor,

In order to identify anatomical regions relevant to schizophrenic pathophysiology, we have developed a neurobehavioral probe with aversive visual stimuli. We hypothesized that negative symptoms would be associated with less activation in response to aversive stimuli. Ten schizophrenic patients and 8 healthy control subjects underwent PET studies with the [0-15]water technique. Subjects viewed images with aversive visual content (facial mutilation, morgue pictures) or nonaversive content. Subjects gave verbal ratings of unpleasantness on a 5-point scale. Simultaneous recording of skin conductance and eye movements permitted evaluation of peripheral emotional responses and assurance of task compliance. Patients and controls gave almost identical ratings of unpleasantness for the aversive images, consistent with prior work showing comparable subjective emotional responses in schizophrenia. For the patients, ratings of unpleasantness correlated inversely with negative symptoms (r = -0.74,~ = 0.02). Viewing aversive stimuli activated the right insula, lingual gyrus and bilateral occipital/temporal areas in both groups. The subjective rating of unpleasantness correlated positively with activation of the lingual gyrus (r = 0.57, p = O.Ol), replicating earlier work in our laboratory. In the patients, greater negative symptoms were associated with less activation of this region (r = -0.63, p = 0.05). The results suggest that negative symptoms may be associated with reduced modulation of visual processing of salient stimuli.

CINGULATE GYRUS ACTIVATION AND ATTENTION IN SCHIZOPHRENIA: A REGIONAL CEREBRAL BLOOD FLOW STUDY E. Tek, A.C. Lahti, J.L. Roffman, H.H. Holcomb, T. Michaelidis, C.A. Tamminga Maryland Psychiatric Research Center, University School of Medicine, PO Box 21247, Baltimore, Maryland21228. USA

of Maryland

Neuronal models of attention implicate the anterior cingulate gyrus in control over motivational elements of attention, specifically in decision-making tasks requiring a pre-existing, conscious plan. Previous studies have indicated pronounced anterior cingulate gyrus activation in normal subjects engaged in the ‘reading versus naming’ task of the Stroop interference test. Schizophrenia, a heterogenous disease characterized by positive and negative constellations of symptoms, has long been associated with impaired attentional performance. We examined the role of the anterior cingulate gyrus in motivational attention among 18 schizophrenic inpatients at the Maryland Psychiatric Research Center. Performance on the Stroop test was correlated with regional cerebral blood flow (as determined by bolus method H2i50 positron emission tomography) assessed while subjects were undergoing another attentionrequiring cognitive task. Unlike previous studies with healthy subjects, performance on the Stroop test was found to correlate negatively with activation in the cingulate gyrus of schizophrenic patients. Moreover, in contrast to current hypotheses linking attention deficits with negative symptoms, no differences were observed between patients with or without enduring negative symptoms. These findings suggest that neural circuits mediating attentional processes are similarly deficient in schizophrenic patients with positive and negative symptoms,

FUNCTIONAL MAGNETIC RESONANCE IMAGING OF THE AUDITORY P300 DEFICIT IN SCHIZOPHRENIA B.I. Turetsky, D. Charbonnier, J. Raz, D. Alsop, R.E. Gur Department Philadelphia.

of Psychiatry, Pennsylvania

University of Pennsylvania, 19104, USA

The P300 ERP is a physiologic index of cognitive processes, typically elicited by infrequent task-relevant sensory stimuli. Abnormal auditory P300 is a robust marker of informationprocessing deficits in patients with schizophrenia and other neuropsychiatric disturbances. Topographic studies have clearly established that the scalp-recorded P300 is not a unitary phenomenon. Rather, it is a composite representation of the activity of multiple anatomically and functionally distinct brain regions. However, electrophysiologic methods offer relatively poor spatial resolution and, consequently, the neuroanatomic sources of the P300 abnormality cannot be clearly delineated, even with extremely dense electrode arrays, Functional magnetic resonance imaging (fMRI), in contrast, offers excellent spatial resolution. However, the delayed onset and extended duration of the hemodynamic response have typically restricted its use to fixed-block experimental designs in which different experimental conditions are segregated and presented separately for relatively long periods, rather than the randomized stimulus presentation and short interstimulus interval required to elicit the auditory P300. We have developed a method for estimating the incremental hemodynamic response in fMR1 experiments in which randomized trials of different experimental conditions are presented in close proximity to each other. We have utilized this approach to delineate both the neuroanatomic sources and

236

the time-course of the hemodynamic response in both auditoryoddball and novelty-P300 experiments in normal individuals. Focal hemodynamic changes, that vary systematically with oddball target probability (corresponding to the P300 response), have been observed in the left superior temporal gyrus, anterior cingulate, and thalamus. We have now applied the same procedures to patients with schizophrenia, allowing us to probe directly the integrity of these physiologic networks and identify specific regional impairments underlying the P300 deficit in schizophrenia. Results of the analysis for both patients and control subjects will be presented.

A ROLE FOR CORTICO-STRIATOTHALAMIC LOOPS IN MODEL PSYCHOSES: A PET STUDY WITH FDG F.X. Vollenweider,

K.L. Leenders

Psychiatric University Hospital Ziirich, Box 68, CH-8029, Ziirich, Switzerland

Research

Department,

Recent human studies into the pathophysiology of psychotomimetic drugs such as hallucinogens (LSD, psilocybin), dissociative anaesthetics (PCP, ketamine), and stimulants (amphetamines) suggest that multiple, interacting neurotransmitter systems are involved in drug-induced and possibly also schizophrenic psychoses. Specifically, it is thought that a dysbalance between serotonin, glutamate, dopamine, and GABA within the limbic cortico-striato-thalamic circuitry (CSTCloops) is critical to psychotic symptom formation. To explore the role of CSTC-loops in psychoses, the effects of d-amphetamine, racemic ketamine, S- and R-ketamine, and psilocybin on cerebral metabolism were investigated in normals using FDG-PET and psychometric ratings. Normalised PET data were subjected to a factor analysis, and factor scores for each individual subject were computed (n = 106). Multiple regression analysis between psychopathology and metabolic factor scores revealed: (1) that ‘depersonalization associated with grandiosity’ relates to metabolic increases in the frontalparietal and temporal cortices and decreases in the occipital cortex; (2) that ‘hallucinations and changed meeting’ is associated with metabolic activation in fronto-parietal, temporal, and striatal regions, and decreases in the occipital cortex; and (3) that ‘anxious ego-dissolution’ is associated with increased thalamic activity. Our data support the view that psychotic symptomatology is associated with a distributed neuronal network including the frontal cortex, basal ganglia, and thalamus.

A FUNCTIONAL MRI STUDY OF WORKING MEMORY IN NON-PSYCHOTIC SIBLINGS OF SCHIZOPHRENIC PATIENTS M.F. Ward, L.J. Seidman, J.M. Goodman, S.V. Faraone, J.M. Goldstein, H.C. Breiter, M. Patti, B.R. Rosen, M.T. Tsuang Department of Psychiatry Center, Harvard Medical USA

at Massachusetts School, Boston,

Mental Health Massachusetts 02115,

The genetic predisposition to schizophrenia is expressed in non-psychotic first-degree relatives (RELs) as subtle neurocognitive phenotypes such as attention and working memory dysfunction. In a previous sample we found that compared to controls, RELs had altered patterns of brain activation during attentionally demanding working-memory tasks, as measured with BOLD fMR1. In this study, a more genetically loaded sample of RELs (n = 13) all siblings in primarily multiplexschizophrenia families, was compared with fMR1 to a new sample of controls (n = 15) using a refined series of auditory CPT tasks. Subjects were matched on age, ethnicity and sex, had estimates of potential IQ in the average range, and were matched on performance on the baseline CPT variant. Wholebrain activation was measured during CPT variants graded in difficulty by varying demands on working memory and interference suppression. Imaging was performed on a 1.5 T GE Signa echo-planar device. Data were motion-corrected, and datasets with z 1.5 mm of detected motion during task were excluded. Preliminary results demonstrate similar differences in activation pattern, as found in our previous sample. Analyses target the prefrontal cortex, anterior cingulate, and thalamus, key brain regions of attention and working-memory networks hypothesized to be impaired in schizophrenia.

REGIONAL BLOOD FLOW EFFECTS OF HALOPERIDOL AND CLOZAPINE M.A. Weiler, A.C. Lahti, H.H. Holcomb, CA. Tamminga Maryland Psychiatric Research Center, University School of Medicine. PO Box 21247, Baltimore, Maryland 21228. USA

of Maryland

Clozapine is observed to be atypical of antipsychotics due to its lower dopamine 2 blocking and effectiveness in a significant proportion of patients that do not respond to typical antipsychotics. In order to begin to determine differences between clozapine and typical D, receptor blocking agents, we studied 6 patients with schizophrenia while off medication and stabilized on haloperidol or clozapine using HZ150 bolus positron emission tomography to measure regional cerebral blood flow. Statistical parametric mapping analysis comparing off and on drug conditions revealed different patterns of change to clozapine. Haloperidol produced increases in the thalamus and hippocampus, while clozapine produced decreases in these areas. Nucleus accumbens activity was also decreased by clozapine. Both drugs produced increases in the motor cortex and decreases in frontal, cingulate and temporal cortices. Lateral frontal lobe decreases were more prominent on haloperidol, while clozapine produced a central frontal decrease. Further analysis will detail drug-behavior interactions. The results of this preliminary analysis will help in formulating hypotheses about how clozapine differs in action from typical antipsychotic agents.

231

BRAIN ACTIVATION DURING COGNITIVE STIMULATION WITH THE STROOP TEST: A FUNCTIONAL MRI STUDY IN HEALTHY VOLUNTEERS AND SCHIZOPHRENIC PATIENTS E. WeiD, A. Hofer, S. Golaszewski, F. Aichner, W.W. Fleischhacker

Ch. Kremser,

Department of Biological Psychiatry. Innsbruck University Clinics, Anichstrasse 35, A-6020 Innsbruck, Austria

Attentional deficits are a prominent feature of cognitive dysfunctions in patients with schizophrenia. The Stroop test stimulates frontal brain regions, especially the anterior cingulate gyms, which is proposed to be an important component of frontal attentional control system. It has been demonstrated by positron emission tomography (PET) that patients with schizophrenia fail to activate the anterior cingulate gyrus during selective attention task. We are currently carrying out a study to more precisely localize frontal activation during psychometric test application using functional magnetic resonance imaging (fMR1). Twenty schizophrenic patients and 20 ageand sex-matched controls will be evaluated. Cortical activation is examined using echo planar MRI with a slice EPI sequence and an in plane resolution of 64 x 64 pixels and 3-mm slice thickness. So far 12 volunteers and 10 patients have been examined. Final results will be presented and discussed.

THE REGISTRATION OF fMR1 AND HIGH RESOLUTION ANATOMICAL MRI SCANS USING MUTUAL INFORMATION C.G. Wible, W.M. Wells, S.S. Yoo, D. Kacher, R. Kikinis, F. Jolesz, R.W. McCarley Department of Psychiatry, Brockton VAMC, Brockton,

Harvard Medical School, Boston; Massachusetts 02401, USA

Mutual information is an information-theoretic algorithm that has been successfully used for registering images from different modalities, such as CT and MRI, or MRI and PET (Wells et al., 1996, Medical Image Analysis, 1( 1): 35). This algorithm was used to register fMR1 activation images, which have low spatial resolution, to higher resolution anatomical images. This registration and a subsequent reformatting of the NRI images allows for: (1) the acquisition of high-resolution anatomical images in a session that is separate from the functional imaging session, and (2) brain activation to be superimposed on a 3-D model of the cortical surface of each individual subject, allowing for a gyral-based identification of the anatomical region showing functional activation. An initial validation study compared results from fMRI activity from a

hand motor task for 2 subjects, where the relative positions of activated regions superimposed on anatomical scans were compared in conditions that required registration using mutual information and those that did not. fMR1 motor mapping for hand movement (resolution 7 mm x 3.75 mm x 3.75 mm3) was performed, and a high-resolution anatomical scan (SPGR) was acquired within the same session (resolution 1.5 x 0.97 x 0.97 mm3). The localization of activation on the SPGR image required minima1 registration (not using mutual information) and a reformatting of the fMR1 data. A second SPGR was acquired in a separate session. The localization of activation relative to this SPGR required registration using mutual information and then reformatting. The resulting anatomical location of activation for the two conditions was found to be comparable, even when the SPGR images were taken in different planes (e.g., coronal versus sagittal). Detailed methods and results will be presented.

FUNCTIONAL MAGNETIC RESONANCE IMAGING STUDIES OF CORTICAL ACTIVATION DURING WORD PRODUCTION: EFFECTS OF PHARMACOLOGIC INTERVENTION D.A. Yurgelun-Todd, A.A. Baird, S.A. Gruber, P.F. Renshaw, B.M. Cohen, D. Goff Brain Imaging Massachusetts

Center, 02178,

McLean USA

Hospital,

Belmont,

Recent functional imaging studies of schizophrenic patients have reported decreased frontal and increased temporal activation in subjects with schizophrenia during word generation. Additionally, the amelioration of both negative symptoms and performance on some cognitive tasks has been demonstrated in schizophrenic patients after treatment with atypical antipsychotics and when d-cycloserine is added to conventional neuroleptics to clarify the effects of pharmacological intervention on cortical activation. We applied fMR1 techniques before and after treatment during a verbal fluency paradigm. We studied 2 groups of schizophrenic patients: the first group include 9 DSM-IV schizophrenic patients who were treated with an atypical antipsychotic agent. The second group included 12 DSM-IV schizophrenic patients enrolled in a double-blind study examining the efficacy of the coadministration of d-cycloserine and conventional neuroleptics. Cortical activation was measured using neuroanatomically defined regions of interest based on both conventional MR and fMR images, and included the frontal and temporal lobes. In both studies, the active medication groups showed an increase in temporal activation after treatment. These results suggest that treatment with atypical antipsychotics and d-cycloserine added to conventional neuroleptics may provide detectable effects through the modulation of the frontotemporal network. These treatments, while producing similar changes in activation, do not appear to normalize the activity of these regions.