This study evaluates the effects of ketamine on healthy and schizophrenic volunteers (SVs) in an effort to define the detailed behavioral effects of the drug in a psychosis model. We compared the effects of ketamine on normal and SVs to establish the comparability of their responses and the extent to which normal subjects might be used experimentally as a model. Eighteen normal volunteers (NVs) and 17 SVs participated in ketamine interviews. Some (n = 7 NVs; n = 9 SVs) had four sessions with a 0.1-0.5 mg/kg of ketamine and a placebo; others (n = 11 NVs; n = 8 SVs) had two sessions with one dose of ketamine (0.3 mg/kg) and a placebo. Experienced research clinicians used the BPRS to assess any change in mental status over time and documented the specifics in a timely way. In both volunteer groups, ketamine induced a dose-related, short (<30 min) increase in psychotic symptoms. The scores of NVs increased on both the Brief Psychiatric Rating Scale (BPRS) psychosis subscale (p =.0001) and the BPRS withdrawal subscale (p =.0001), whereas SVs experienced an increase only in positive symptoms (p =.0001). Seventy percent of the patients reported an increase (i.e., exacerbation) of previously experienced positive symptoms. Normal and schizophrenic groups differed only on the BPRS withdrawal score. The magnitude of ketamine-induced changes in positive symptoms was similar, although the psychosis baseline differed, and the dose-response profiles over time were superimposable across the two populations. The similarity between ketamine-induced symptoms in SVs and their own positive symptoms suggests that ketamine provides a unique model of psychosis in human volunteers. The data suggest that the phencyclidine (PCP) model of schizophrenia maybe a more valid human psychosis/schizophrenia drug model than the amphetamine model, with a broader range of psychotic symptoms. This study indicates that NVs could be used for many informative experimental psychosis studies involving ketamine interviews.