To localize a central nervous feed-forward mechanism involved in cardiovascular regulation during exercise, brain activation patterns were measured in eight subjects by employing positron emission tomography and oxygen-15-labeled water. Scans were performed at rest and during rhythmic handgrip before and after axillary blockade with bupivacaine. After the blockade, handgrip strength was reduced to 25% (range 0-50%) of control values, whereas handgrip-induced heart rate and blood pressure increases were unaffected (13 +/- 3 beats/min and 12 +/- 5 mmHg, respectively; means +/- SE). Before regional anesthesia, handgrip caused increased activation in the contralateral sensory motor area, the supplementary motor area, and the ipsilateral cerebellum. We found no evidence for changes in the activation pattern due to an interaction between handgrip and regional anesthesia. This was true for both the blocked and unblocked arm. It remains unclear whether the activated areas are responsible for the increase in cardiovascular variables, but neural feedback from the contracting muscles was not necessary for the activation in the mentioned areas during rhythmic handgrip.