A prospective study was conducted on a group of 160 students from 9 to 14 years of age with learning disabilities to predict neuropsychological outcome using perinatal information as predictors. Perinatal information was obtained from the Maternal Perinatal Scale (MPS) (Dean & Gray, 1985). Subjects' neuropsychological functioning was assessed by the Short Neuropsychological Screening Device (SNSD) (Reitan & Herring, 1985). Information concerning subjects' intelligence was obtained from the Wechsler Intelligence Scale for Children-III (WISC-III) administered within the past two years. Hollingshead's Four Factor Index of Social Status was employed to determine subjects' socioeconomic status. A stepwise multiple regression analysis yielded a regression model that contained a subset of 7 perinatal risk factors, involving: (1) Obstetric History; (2) Gestational Age; (3) Psychosocial Events; (4) Delivery; (5) Intrauterine Stress; (6) Teratogenic Stress; and (7) Fetal Oxygenation. A hierarchical regression analysis was further performed to examine if adding socioeconomic and intellectual information to the regression model could increase the prediction of neuropsychological outcome. Results showed that up to 82% of the variability in the neuropsychological outcome was explained by the linear composite of the 7 risk factors. When socioeconomic and intellectual information were added to the regression model, the prediction of neuropsychological outcome was significantly improved. About 201 of the students with learning disabilities in the present study were found to display symptoms similar to minimal brain damage (MBD) relating to poor visual-motor integration, underdeveloped language skills, and aphasic conditions. The results support the theory of a "continuum of reproductive casualty" proposed by Pasamanick et al. (1956). The importance of detecting early indicators of neuropsychological deficits in at risk children was further suggested by the present study. / Department of Educational Psychology