The pattern of brain metabolic changes produced by olanzapine has yet to be described, despite the theoretical and clinical interest of this new antipsychotic. We studied a group of 17 schizophrenic patients who underwent two fluoro-deoxyglucose-positron emission tomography (FDG-PET) studies under two different conditions: a baseline scan during treatment with either conventional antipsychotics (n=15) or risperidone (n=2) and a second scan performed 17-24 weeks after switching to olanzapine. PET scans were obtained while performing a standard cognitive paradigm (Continuous Performance Test) and analysed by means of Statistical Parametric Mapping. No significant metabolic changes were found in the comparison between pre- and post-olanzapine conditions. A brain map of the statistical power of our design showed that changes up to 3% in the frontal and up to 8% in the occipital region were not likely to exist (1-beta=0.8). The degree of improvement in positive symptoms was related to the amount of activity decrease in the right orbital region and to the amount of activity increase in the primary visual area. Improvement in negative symptoms was associated with an activity increase in the dorsal prefrontal cortex, and a higher baseline activity in both temporal poles. These correlation patterns suggest that the functional mechanism of action of olanzapine may share traits from both typical and atypical neuroleptics.