Eighty-three patients with schizophrenia and 47 healthy controls received positron emission tomography (PET) with 18F-2-deoxyglucose uptake while they were executing the Continuous Performance Test (CPT). The entire cortex was divided into 16 regions of interest in each hemisphere, four in each lobe of the brain, and data from corresponding right and left hemispheric regions were averaged. Data from the schizophrenic patients were subjected to a factor analysis, which revealed five factors that explained 80% of the common variance. According to their content, the factors were identified and labelled 'parietal cortex and motor strip', 'associative areas', 'temporal cortex', 'hypofrontality' (which included midfrontal and occipital areas) and 'frontal cortex'. Hemispheric asymmetry was only confirmed for the temporal cortex. Factor weights obtained in the schizophrenic group were applied to the metabolic data of the healthy controls and factors scales computed. Schizophrenics were significantly more hypofrontal than the controls, with higher values on the 'parietal cortex and motor strip' factor and a trend towards higher values in the temporal cortex. A canonical discriminant analysis confirmed that the 'hypofrontality' and 'parietal cortex and motor strip' factors accurately separated the schizophrenic group from the healthy controls. Hemispheric asymmetry was only confirmed for the temporal lobe. Significantly higher factor scores for the left temporal lobe in schizophrenics than in normals were obtained when calculated for the right and left hemisphere separately. Taken together, our results confirm the importance of hypofrontality as a pattern of cortical metabolic rate and point to the potential importance of parietal and motor strip function in schizophrenia.