Sleep-wake homeostatic and internal circadian timedependent brain processes interact to regulate human brain function so that alert wakefulness is promoted during the daytime and consolidated sleep is promoted at nighttime. The consequence of chronically altering the normal relationship between these processes for human brain function is largely unknown. We tested cognitive and vigilance performance while subjects lived in the laboratory for over a month. The subjects lived on either 24.0- or 24.6-hr day lengths. Half of the subjects tested maintained a normal relationship between sleep-wakefulness and internal circadian time (synchronized group), whereas the other half did not (nonsynchronized group). Levels of the sleep-promoting hormone melatonin were high during scheduled sleep in the synchronized group, whereas melatonin levels were high during scheduled wakefulness in the nonsynchronized group. Failure to adapt to the scheduled day length was dependent upon individual differences in intrinsic circadian period. Total sleep time was reduced, sleep latency and Rapid Eye Movement (REM) latency were shortened, and wakefulness after sleep onset was increased in the nonsynchronized group. Cognitive performance improved (i.e., learning) in the synchronized group, whereas learning was significantly impaired in the nonsynchronized group. Attention progressively declined in both groups, suggesting that 8 hr of scheduled sleep per night is insufficient to maintain brain vigilance even when sleep occurs at an appropriate biological time. Our results establish that proper alignment between sleep-wakefulness and internal circadian time is crucial for enhancement of cognitive performance. In addition, our results demonstrate that exposure to dim light (approximately 25 lx) is sufficient to expand the range of entrainment in humans.