The purpose of this study was to determine whether exercise modulates the responses of middle cerebral artery blood velocity (MCA V(mean)) and cerebrovascular conductance to sympathetic stimulation (i.e. cold pressor test--CPT). To accomplish this, MCA V(mean) responses were assessed during CPT, static handgrip exercise (HG) at 30% of maximum voluntary contraction and combined condition (HG + CPT), assigned in a counterbalanced order, in eight healthy subjects. Blood pressure (BP), cardiac output (CO) and end-tidal partial pressure of carbon dioxide (PETCO(2)) were also measured non-invasively, and an index of vascular conductance was calculated for MCA (CVCi). BP increased from rest (P < 0·05) during CPT and HG and was additionally augmented during HG + CPT (P < 0·05 versus rest, CPT and HG). Despite the greater augmentation in BP during HG + CPT, MCA V(mean) was similarly increased during both HG (18·5 ± 2%, P < 0·05 versus rest) and combined condition (19·6 ± 2%, P < 0·05 versus rest). MCA V(mean) remained unchanged from rest during CPT only. CVCi was slightly reduced (P < 0·05) from rest during HG but was greatly reduced by CPT (P < 0·05 versus rest). The reduction in CVCi evoked by CPT at rest (-15 ± 2%, P < 0·05 versus rest) was significantly attenuated during HG (-8 ± 2%, P < 0·05 versus CPT). Increases in CO were similar in all trials, and PETCO(2) was unchanged from rest throughout the experiments. In summary, the cerebral conductance index decreases during the cold pressure test while that reduction is smaller when the CPT is conducted during the HG. This was critical for the maintenance of MCA V(mean) during combined condition.