Regional cerebral blood flow (rCBF) data from two PET-15O water schizophrenia studies were analyzed using individually placed, magnetic resonance (MR)-guided hippocampal volumes of interest (VOI). In one study, normal (N = 10) and schizophrenic (N = 18) volunteers performed an overlearned auditory discrimination task in rest, control, and decision conditions. In the other study, schizophrenic and normal volunteers received the noncompetitive NMDA receptor antagonist ketamine and placebo and had sequential rCBF evaluations. Moreover, the schizophrenic volunteers were off drug in one study and on antipsychotic drug in the second study, allowing an additional comparison of medication status. VOIs were placed on anterior, middle, and posterior hippocampal areas in each PET image from both studies, redirected from an MR scan, and individually adjusted. While no hippocampal activation was apparent in either the normal or schizophrenic group in the task vs. condition comparison, rCBF was higher in the schizophrenic than in the normal hippocampus in both task and control conditions, independently. In addition, at rest rCBF was significantly higher in the unmedicated group of schizophrenics than in the group of medicated patient volunteers and higher than in the normal comparison group. This suggests that schizophrenia is associated with elevated rCBF in the hippocampus, which "normalizes" with antipsychotic drug treatment. Ketamine, the noncompetitive NMDA receptor antagonist, was more potent in reducing rCBF in the schizophrenic group compared to the normal volunteer group. These data are consistent with a previous report from our laboratory of reduced NMDA receptor NR1 subunit expression and possible abnormal NMDA receptor composition in schizophrenia. These data show an abnormality of hippocampal function in schizophrenia and suggest that this abnormality may be associated with the pathophysiology of the illness.