In this PET study, we have investigated the human brain activity evoked by a visual motion paradigm commonly used to measure motion-related visual evoked potentials (VEPs). Because standard PET activation studies have been performed with motion along four axes, we first determined the pattern of brain activation when motion was restricted to a single axis. Motion back and forward along a single horizontal axis compared with a static condition revealed weak differential activations in the cuneus and the parietal cortex. Human area MT/V5 (middle temporal area) was hardly activated at all in this subtraction. Additional functional MRI experiments proved that MT/V5 activity is significantly higher for motion along four axes than for motion along a single axis. Secondly, we attempted to isolate the pattern of brain activity related to the reversal of motion direction and to the onset of motion, i.e. two transient motion components commonly used in measuring motion-related VEPs. To that end, we added a continuous linear contrast modulation, that reached maximum contrast at reversal or onset of motion, and compared both conditions with a contrast-modulated static or continuous motion condition. Subtraction of the static random dot pattern condition from the single-axis motion reversal condition, both contrast-modulated, revealed three significant activations: the anterior parieto-occipital sulcus, the lateral sulcus and the anterior claustrum. Additional analysis showed that these activations were not due to motion appearance or disappearance, but were due to the combination of motion reversal and contrast modulation. Hence, these activations do not reflect the motion reversal transient per se. In order to isolate a metabolic response to the reversal transient per se, we used a conjunction analysis, which suggests that activity in human MT/V5, the cuneus and a parietal insular region could underlie the motion reversal VEP in our experiments. Subtraction of the static random dot pattern condition from the single-axis motion onset condition, both contrast-modulated, revealed a single significant activation in the posterior cingulate cortex. Although the significance of this activation is unclear, it adds further evidence for the visual function of this region.