During wakefulness, cerebral blood flow (CBF) is closely coupled to regional cerebral metabolism; however, CBF is also strongly modulated by breathing, increasing in response to both hypercapnia and hypoxia. During slow wave sleep (SWS), cerebral metabolism, CBF and arterial blood pressure decrease whilst the partial pressure of arterial CO₂ increases due to a reduction in alveolar ventilation. The reduction in CBF during SWS therefore occurs despite a relative state of hypercapnia and could be explained, at least in part, by changes in the cerebral vascular response to CO₂. As sleep is a state in which pathophysiological challenges to the cardio-respiratory system are most marked, it is important to understand the regulation of cerebral blood flow during sleep. In a series of studies in our laboratory, we have determined changes in CBF and the cerebral vascular responses to hypercapnia (CVR-CO₂) and hypoxia using transcranial Doppler ultrasound to measure middle cerebral artery blood flow velocity (MVAV) as a non-invasive index of cerebral perfusion in humans. We confirmed the observation of others that CBF during SWS is reduced; further, we determined that CVR-CO₂ is reduced by 70% compared to the waking state – a change that would permissively allow CBF to fall at a time when arterial PCO2 is increased (Meadows et al. 2003). Strikingly, cerebral vasodilation to mild hypoxia (arterial oxygen saturation -10%) is completely abolished during slow wave sleep (Meadows et al. 2004); indeed these observations are consistent with a cerebral vasoconstriction in response to hypoxia during SWS. Rapid eye movement sleep (REM) is associated with notable fluctuations in cardiovascular and respiratory variables; overall, CBF and cerebral metabolism are reported to be similar to, or slightly raised, compared to the waking state. To test the hypothesis that CVR-CO₂ would also be reduced during REM sleep, MCAV was measured in 12 normal healthy male subjects (age 26.2 ± 4.2 years, body mass index 23.8 ± 3.0 kg.m^-2). Minute ventilation decreased (Wake: 8.2 ± 2.4; REM: 6.0 ± 1.7 l.min^-1; p<0.05) and PETCO₂ increased during REM compared to Wake (41 ± 3 vs. 44 ± 3 mmHg; p<0.05). Overall, MCAV did not change (53.3 ± 13.9; 54.2 ± 15.1 cm.sec^-1, p>0.05). However, MCAV during phasic REM was significantly higher than tonic REM (56.5 ± 13.3 vs. 51.7 ± 9.2 cm.sec^-1, p<0.05). REM sleep may be divided into 2 phases – tonic (absence of eye movements) and phasic (bursts of eye movements) REM. The proportion of tonic and phasic REM was not changed by hypercapnia and CVR-CO₂ during REM was not significantly different from that during wake (Wake: 2.2 ± 0.9; REM: 1.7 ± 1.0 cm.sec^-1.mmHg^-1, p>0.05). Upon waking in the morning, CVR-CO₂ is reduced compared to measurements made during wakefulness in the previous evening (e.g. Meadows et al. 2005); this observation has been linked to the increased risk of stroke in the early morning. The mechanism for this change is uncertain but might, in part, reflect the sleep-related changes in CVR-CO₂ reported above. It is of interest that CVR to hypoxia, upon waking in the morning, is no different from that measured on the previous evening (Meadows et al. 2005) and suggests that state-related changes in CVR to hypercapnia and hypoxia may not be mediated by the same mechanisms. In conclusion, the changes in CVR to hypercapnia and hypoxia, that we report across the sleep-wake cycle, indicate that the regulation of these factors is modulated by changes in neural state. As these changes in CVR-CO₂ mirror the changes in CBF and cerebral metabolism across these states, the regulation of both CBF and CVR-CO₂ may be directly linked to cerebral metabolism. The mechanism for any such linkage remains to be established. These changes in cerebral vascular regulation during sleep may make the brain particularly susceptible to vascular insults associated with sleep-related breathing disorders and stroke.