SATURDAY,, MAY 21 BIOL PSYCHIATRY 723 1994; 35:615-747 ties in the hippocnmpal/amygdala complex, parahippoeampal gyms, and superior temporal gyms, mostly on the left side of the brain. MRI pro- rides the opportunity to noninvasively obtain images of the brain of living subjects, circumventing many of the problems with post-mortem studies, but foregoing the histological detail found in those studies, it is often dif- ficult to reliably separate the hippocampal and amygdalar regions on con- ventional MR! scans. We here report preliminary efforts to develop a method for obtaining high resolution MR! scans of the hippocampal re- gion in schizophrenic and control subjects. MR images were acquired on a 1.5 Tesla MR General Electric SIGNA System. The protocol consists of 15, 3ram coronal oblique slices with .35mm in-plane resolution through the temporal lobe region (TE-23, TR-3000, field of view- 18cm, phase encoding steps - 192, NEX-4). Only a few subjects have been run through this new protocol, but the resulting images have shown clear sep- aration between the amygdala and the hippocampal formation on each slice. 388. FRONTAL AND TEMPORAL LOBE BRAIN VOLUMES IN SCHIZOPHRENIA: RELATIONSHIP TO SYMPTOMATOLOGY AND CLINICAL SUBTYPE B. Turetsky, P.E. Cowell, R.C. Gur, R.I. Grossman, & R.E. Gur University of Pennsylvania, Philadelphia, PA Clinical features of schizophrenia have been related to neuroanatomic measures with inconclusive results. Some MR! studies report volume def- icits of the temporal lobes and specific temporal lobe structures. Other studies have failed to fred localized deficits, or have found diffuse abnor- malities inconsistent with a specific localized structural change, in a large sample of patients and controls, we found abnormalities in whole-brain measures that differed according to the relative prominence of negative, Sclmeiderian, and paranoid symptoms. We now examine left and right frontal and temporal lobe volumes, and their relationship to clinical symp- tomatology. Subjects were 71 schizophrenics and 77 controls. Scores on the SANS, SAPS and BPRS were used to classify patients as Negative, Paranoid, and Sclmeiderian subtypes. Patients were also divided into Def. icit vs. Non-Deficit. Transaxial spin-echo MR! images (TR-3000, TE-30 and 80 msec) were obtained. Region-of-Interest (ROD boundaries, repre- senting left and right frontal and temporal lobes were identified in collab- oration with a neuroradiologist. Brain and CSF volumes were obtained separately for each ROI, using a semi-antomated segmentation algorithm. Reliability of measurement across two independent investigators ranged from 0.96 to 0.98 for brain, and from 0.88 to 0.99 for CSF. Schizophrenics had insignificantly smaller brain volumes across all regions. There was a significant diagnosis by region by hemisphere interaction. Schizophrenics exhibited abnormalities of brain asymmetry characterized by relatively greater decreases in size of the left temporal and right frontal lobes, inde- pendent of age, sex and overall brain size. Temporal lobe asymmetry was associated with severity of negative symptoms, while abnormal frontal lobe asymmetry was associated with duration of illness. Deficit patients had a region-specific increase in left temporal CSF. Similarly, Negative schizophrenics had greater left temporal CSF, while Schneiderian patients had more CSF in both left and right temporal lobes. Paranoid patients were not different from controls on these measures. These results suggest that regional patterns of neuroanatomic abnormalities are not uniform across schizophrenic subgroups. They thus provide clues to differences in underlying pathophysiology ~n different patient subgroups. 389. PSYCHIATRIC SEQUELAE OF LOW BIRTH WEIGHT AT 6 YEARS OF AGE Naomi Breslau Henry Ford Health Sciences Center, Detroit, MI 48202; University of Michigan School of Medicine, Ann Arbor, MI Recent studies have reported an increased prevalence of behavior prob- lems in very low birthweight children (VLBW) (_<l,500g). There is evi- dence also that the association might be specific to hyperactivity and inat- tention. Little is known about the association with ADHD proper, espe- cially in low birth weight (LBW) children above 1,500g. This study fo- cused on the entire LBW range (<2,500g) and evaluated 5 common DSM- lll-R psychiatric disorders at 6 years of age, using the NIMHDISC (Par- ent). LBW (n-473) and normal birth weight children (NBW) (n-350) were randomly selected from the ! 983-1985 newborn lists of 2 major hos- pitals in Southeast Michigan, one serving an urban di~dvantaged popula- tion and the other serving a suburban middle-class population, The odds ratio for ADHD associated with LBW was 2.5 (95% Cl !.5-4.0) in the urban and I.! (95% C1.6-2.0) in the suburban population. Other disorders (Oppositional-Defiant, Phobia, Overanxious, Separation Anxiety) were unrelated to LBW in either population. The differential effect of LBW in the urban vs. suburban population was in the subset of LBW children with below average but above subnormal IQ (80-99). LBW children with sub- normal IQ (<80) had a 4-to 6-fold increased odds for ADHD irrespective of site, and those with IQ _>100had no increased odds in either site. Subur- ban LBW children with milder cognitive deficits might be protected from adverse behavioral sequelae by the advantages provided to children in middle-class households, a protection that might not extend to LBW chil- dren with more severe cognitive deficits. 390. SCHIZOPHRENIA: HEAD SIZE CORRELATES WITH 2nd TRIMESTER MARKERS H. S. Bracha I, J.W. Gilger I, D.S. McCray t, E.F. Torrey 2, & C.N. Karson t tPsychiatry, University of Arkansas for Medical Sciences, Slot O 554,4301 West Markham, Little Rock, AR 72205; 2NIMH We examined the relationship between markers of second-trimester-pre- natal insult (inutem, non-shared environment) and head circumference in twin pairs discordant for schizophrenia. Twenty-one pairs of monozygotic (MZ) twins, where one member of each pair was schizophrenic, served as subjects. The two major variables of this study were: the "percent abso- lute within-pair difference in total finger ridge count" (%ad-TFRC), and the "percent signed within-pair difference in head circumference" (%sd- HC). A significant relationship between %ad-TFRC and %ad-HC was found: Pearson r - -.89 (p<.001). These data are in line with our predic- tion: that there is a significant tendency for the psychotic twins to have smaller head circumferences than their non-psychotic cotwins as mem- ber's of the twin pairs vary increasingly in finger ridge counts i.e., as there are increasing differential second-trimester effects [in-utero, non-shared environment] on members of the same twin pair. The results suggest that the within-pair differences in head circumference in MZ twins discordant for schizophrenia may in part be a consequence of differential within-pair prenatal insult occurring around the second trimester since dermal ridges examined in this study are known to develop at that time. - Supported by a DVA grant to H.S. Bracha, M.D.