Abstract
OBJECTIVE: Cognitive deficits are prominent in schizophrenia. Patients have an average score one standard deviation below normal on a broad spectrum of cognitive tests. It has been repeatedly noted, however, that 20%-25% of patients differ from this general pattern and score close to normal on neuropsychological testing. This study used brain morphometry to 1) identify brain abnormalities associated with more severe cognitive deficits and 2) help determine whether cognitively relatively intact patients perform better because they have less severe illness or because they have a different illness. METHOD: Patients were assigned to a neuropsychologically near normal (N=21) subgroup if they scored within 0.5 standard deviation of healthy comparison subjects (N=30) on four tests of attention and verbal and nonverbal working memory, and to a neuropsychologically impaired (N=54) group if they scored at least 1.0 standard deviation below that of comparison subjects. Subgroup assignments were confirmed with the California Verbal Learning Test and degraded-stimulus Continuous Performance Test. Volumes of ventricular compartments, hippocampus, amygdala, thalamus, cerebellum, and regional cortical gray and white matter were dependent variables. Differences among groups were evaluated by using linear mixed-model multivariate analyses with gender, age, and height as covariates. RESULTS: Both neuropsychologically near normal and neuropsychologically impaired patients had markedly smaller gray matter and larger third ventricle volumes than healthy comparison subjects. Only neuropsychologically impaired patients, however, had significantly smaller white matter and larger lateral ventricle volumes than healthy comparison subjects. CONCLUSIONS: Although both neuropsychologically impaired and neuropsychologically near normal patients have marked neuropathology in their gray matter, the relative absence of white matter pathology in the neuropsychologically near normal group suggests the possibility of differences in the disease process.