1. We measured relative cerebral blood flow (rCBF) changes with positron emission tomography and H(2) 15O in six normal subjects repeatedly performing a spatial stimulus-response compatibility task. Subjects had two motor response conditions. They were instructed to respond with the left hand to a left visual field light stimulus and with the right hand to a right visual field light stimulus (compatible condition), and with the right hand to a left visual field light stimulus and with the left hand to a right visual field light stimulus (incompatible condition). Six rCBF measurements per condition were performed in each subject. 2. Reaction times were faster (P < 0.0005) in the compatible (287 ms) than the incompatible (339 ms) condition (spatial stimulus-response compatibility effect). A bilateral increase (P < 0.05) in rCBF in the superior parietal lobule of the two hemispheres was observed during the incompatible condition when compared with the compatible one. No rCBF decreases were observed. Reaction times correlated (P < 0.0001) with the rCBF in the two activated superior parietal lobule areas. 3. Reaction times decreased with practice according to a linear trend (P < 0.05). Practice-related linear rCBF increases (P < 0.05) were observed in the dorsolateral prefrontal, premotor, and primary motor cortex of the left hemisphere. No significant rCBF decreases were observed. 4. Practice did not affect the spatial stimulus-response compatibility effect. A parallel shortening of reaction times was observed in both compatible and incompatible conditions, in both left and right hand responses, and in both left and right visual fields. Accordingly, when rCBF was analyzed, the spatial stimulus-response compatibility by practice interaction did not show any significant activated area. 5. These findings suggest that the two activated areas in the left and right superior parietal lobules subserve the mapping of the visual stimulus spatial attributes onto the motor response spatial attributes and that the rCBF increases in the incompatible response condition represent the more complex computational remapping required when stimuli and response do not match spatially. 6. The dorsolateral prefrontal, premotor, and motor rCBF linear increases in the left hemisphere seem to reflect the effect of practice on cortical processes common to both compatible and incompatible response conditions. These cortical processes presumably strengthen the links between stimuli and responses under different stimulus-response compatibility conditions. The lateralization of the rCBF increases suggests a left hemisphere superiority in these processes.