The aim of this study was to examine the potential of serial rCBF studies to directly characterize the regional effects and dynamic time course of the centrally active drug ketamine. The value of a broader application of this technique to other neurally active drugs to characterize the pharmacodynamics of CNS compounds is suggested by these data. Thirteen normal subjects received a 0.3 mg/kg intravenous dose of ketamine over 60 seconds; ten other individuals received placebo in the same manner. For each subject, three baseline PET rCBF scans and seven sequential post-ketamine scans at 10-minute intervals were obtained using H(2)(15)O water. SPM techniques were employed to identify the maxima of any cluster significant by spatial extent analysis at any post-ketamine time point between 0 and 36 min. These extremes from the ketamine group, were identified in placebo scans similarly and grown to a 6x6x12 mm voxel set. The average rCBF values of the ketamine-defined clusters were determined in the drug and placebo conditions at all time points. rCBF across time was plotted for each cluster and compared between drug and placebo. Area under the curve (AUC) was calculated between baseline and 36 minutes. The kinetic characteristics of the ketamine-induced rCBF curves were compared to induced behaviors in each maxima. Ketamine produced distinct patterns of rCBF change over time in different brain regions; maxima within an anatomically defined region responded similarly. Ketamine induced rCBF activations in anterior cingulate, medial frontal and inferior frontal cortices. All maxima with a relative flow reduction with ketamine were in the cerebellum. The pattern of all activations and suppressions was monophasic with the peak changes at 6-16 minutes. In preliminary analysis, individual C(max) and AUC of maxima in the anterior cingulate/medial frontal region tended to correlate with the mild psychotomimetic action of ketamine; whereas, there was no tendency toward correlation with this psychological change in cerebellar maxima. The direct action of a centrally active drug can be assessed regionally and dynamically in brain using rCBF and a scan sequence optimally timed to complement the drug's time course. Ketamine pharmacodynamic response can be related to concurrent behavioral changes, tending to link the behavior with a brain region. This experimental design provides direct characterization of drug action in the CNS in ways heretofore unavailable.