We examined the interactive influence of hypoxia and exercise, and hypercapnia and exercise, on regional cerebral perfusion and sympathetic activation. Twenty healthy young adults (seven women) completed study trials including (1) rest in normoxia ( : ∼96%, : ∼36 mmHg), normocapnic hypoxia ( : ∼84%, : ∼36 mmHg), and normoxic hypercapnia ( : ∼98%, : ∼46 mmHg) and (2) unilateral rhythmic handgrip exercise (45% of maximal voluntary contraction at 1 Hz for 3 min) under the same gas conditions. Based on the exercising arm, blood flow in the contralateral internal carotid (ICA) and ipsilateral vertebral (VA) arteries, anterior and posterior cerebral O delivery ( ), and muscle sympathetic nerve activity (MSNA) were measured in each trial. During exercise in hypoxia, ICA, VA, anterior and posterior were significantly lower, whereas total MSNA was significantly greater, than the sum of the responses evoked by either hypoxia or exercise alone. During exercise in hypercapnia, ICA and anterior were significantly greater, whereas MSNA was lower, than the sum of the responses evoked by either hypercapnia or exercise alone. The VA and posterior responses to hypercapnic exercise were not different from the summated responses. These findings suggest that the brain is hypoperfused and sympathetic outflow potentiated during hypoxic exercise, and that the brain is hyperperfused and sympathetic discharge constrained during hypercapnic exercise. The contrasting consequences for cerebral perfusion and sympathetic activation indicate a potential involvement of Cushing's mechanism in the autonomic control during exercise in healthy humans. KEY POINTS: Brain O-demand and -supply are mismatched, and muscle sympathetic nerve activity (MSNA) is enhanced in humans exercising at high altitude; the link between the two phenomena remains elusive. We evaluated the isolated and interactive effects of exercise, hypoxia, and hypercapnia on blood flow in the internal carotid (ICA) and vertebral (VA) arteries, and MSNA. The interaction of hypoxia and exercise was hypo-additive for ICA and VA and anterior and posterior cerebral O delivery ( ), but hyper-additive for MSNA. The interaction of hypercapnia and exercise was hyper-additive for ICA and anterior , additive for VA and posterior , and hypo-additive for MSNA. These observations indicate that a suboptimal brain perfusion during hypoxic exercise coincides with a potentiated sympathetic outflow, while a (supra-)optimal brain perfusion during hypercapnic exercise coincides with a suppressed sympathetic outflow. Our findings suggest that Cushing's mechanism may play a role in the autonomic control in exercising humans.