K.F. Berman

226. Beyond hypofrontality in functional brain imaging of schizophrenia

patients are usually unable to perform as well as normal, healthy subjects (NV). In order to better characterize and represent this cognitivemaladaptationwe have studiedNV and SZ subjectsfollowing extensive(3 -5 thousandtrials) tonerecognitiontraining.The task consistedof a tone presentedevery 2 seconds;the subjectrecognized the tone as relativelyhighor lowin frequency.Thereferencestimulus was fixed but the test stimulusvariedduringthe trainingphase of the study. Sufficient disparity between the tones was used in order to generate an 80%accuracy score. A subgroupof SZ (n=12) was able to perform as well as the NV (n= 12) with respect to accuracy and response time (RT). Using positronemissiontomography(PET) and 15-Oxygen,water, blood flow methodology,we studied subjects recognizinglow disparity tones. The contrasting conditions showed similar shifts in brain activity when switchingfrom one conditionto another for NV and SZ. The correlationmaps for blood flow (rCBF) and RT, accurate trials only, differed significantlybetween groups. WhereasNVexhibitedgreaterrCBFfor faster RT in auditorycortices and cerebellum,SZ subjectsexhibitedrobustcorrelationsfor fast RT in frontal, cuneus and occipital regions. Multiple scan conditions permit analyses of brain changes between and within conditions, between and within groups.

A method for template-based volumetry: Investigation of gender and age-related structural differences in schizophrenic patients, their siblings and normal controls

Methods Ninety patients diagnosed with DSM-IIIR schizophrenia, 146 of their first-degree relatives and 113 normal controls were scanned using a 1.5 Tesla MRI scanner. Each Tl-weighted structural MR image of the head comprising 124 sagittal slices (voxel size 1.5mm x .9375mm x .9375mm) was intensity-normalized, passed through a skull-stripping routine (using Freesurfer software [2]) and then segmented into gray matter, white matter and cerebrospinal fluid (CSF) maps within SPM99 [3]. Gray matter volumes were obtained from the segmented images by counting voxels above a threshold of 0.5. The segmented maps were normalized to a standard Talairach template using SPM99. Ventricular volumes were determined by applying over-inclusive masks of the ventricles to the normalized CSF maps and counting the number of voxels enclosed witbin these regions. The resulting volumes were divided by the determinant of the affine transformation matrix to obtain the corresponding volumes in the original MR images.