Abstract
The neuroimaging literature on structural brain abnormalities in the major psychoses is quantitatively reviewed. The mean effect size for studies of lateral ventriculomegaly in schizophrenia (d = .70) corresponded to 43% nonoverlap between the distributions of schizophrenics and control Ss. Planimetry yielded larger effects than linear methods of ventricular size estimation. Although enlargement of the third ventricle was comparable to that of lateral ventriculomegaly (d = .66), it was found to be significantly greater after differences in measurement method were taken into account. The average cumulative length of hospitalization, adjusted for patients' age and duration of illness, predicted ventriculomegaly in schizophrenia. Studies on schizophrenia and affective disorder differed neither in the extent of reported ventriculomegaly nor in the amount of "cortical atrophy."