Abstract: We examined the extent to which apnoea-induced extremes of oxygen demand/carbon dioxide production impact redox regulation of cerebral bioenergetic function. Ten ultra-elite apnoeists (six men and four women) performed two maximal dry apnoeas preceded by normoxic normoventilation, resulting in severe end-apnoea hypoxaemic hypercapnia, and hyperoxic hyperventilation designed to ablate hypoxaemia, resulting in hyperoxaemic hypercapnia. Transcerebral exchange of ascorbate radicals (by electron paramagnetic resonance spectroscopy) and nitric oxide metabolites (by tri-iodide chemiluminescence) were calculated as the product of global cerebral blood flow (by duplex ultrasound) and radial arterial (a) to internal jugular venous (v) concentration gradients. Apnoea duration increased from 306 ± 62 s during hypoxaemic hypercapnia to 959 ± 201 s in hyperoxaemic hypercapnia (P ≤ 0.001). Apnoea generally increased global cerebral blood flow (all P ≤ 0.001) but was insufficient to prevent a reduction in the cerebral metabolic rates of oxygen and glucose (P = 0.015–0.044). This was associated with a general net cerebral output (v > a) of ascorbate radicals that was greater in hypoxaemic hypercapnia (P = 0.046 vs. hyperoxaemic hypercapnia) and coincided with a selective suppression in plasma nitrite uptake (a > v) and global cerebral blood flow (P = 0.034 to <0.001 vs. hyperoxaemic hypercapnia), implying reduced consumption and delivery of nitric oxide consistent with elevated cerebral oxidative–nitrosative stress. In contrast, we failed to observe equidirectional gradients consistent with S-nitrosohaemoglobin consumption and plasma S-nitrosothiol delivery during apnoea (all P ≥ 0.05). Collectively, these findings highlight a key catalytic role for hypoxaemic hypercapnia in cerebral oxidative–nitrosative stress. (Figure presented.). Key points: Local sampling of blood across the cerebral circulation in ultra-elite apnoeists determined the extent to which severe end-apnoea hypoxaemic hypercapnia (prior normoxic normoventilation) and hyperoxaemic hypercapnia (prior hyperoxic hyperventilation) impact free radical-mediated nitric oxide bioavailability and global cerebral bioenergetic function. Apnoea generally increased the net cerebral output of free radicals and suppressed plasma nitrite consumption, thereby reducing delivery of nitric oxide consistent with elevated oxidative–nitrosative stress. The apnoea-induced elevation in global cerebral blood flow was insufficient to prevent a reduction in the cerebral metabolic rates of oxygen and glucose. Cerebral oxidative–nitrosative stress was greater during hypoxaemic hypercapnia compared with hyperoxaemic hypercapnia and coincided with a lower apnoea-induced elevation in global cerebral blood flow, highlighting a key catalytic role for hypoxaemia. This applied model of voluntary human asphyxia might have broader implications for the management and treatment of neurological diseases characterized by extremes of oxygen demand and carbon dioxide production.