BACKGROUND Recent studies report lengthening of latency and the reduction of amplitude of the P300 component of event-related potentials (ERP) recorded to auditory targets following a night of sleep deprivation. However, no study has examined whether sleep deprivation impacts the circadian effect on P300. In the present study, we examined the impact of 24-40 h of sleep deprivation and of the circadian rhythm on the P300. METHODS There were 18 subjects tested at 08:30, 14:30, and 21:30 on two different days: once following a night of normal nocturnal sleep and once following a night in which subjects did not sleep. Behavioral and electrophysiological measures were recorded during each session. RESULTS P300 amplitude was not significantly affected either by sleep deprivation or by time of day. However, P300 latency was lengthened by 27 ms following sleep deprivation [F (1,14) = 28.55; p < 0.0001]. There was a significant interaction between sleep deprivation and circadian rhythm [F (2,13) = 3.27; p < 0.05]. When subjects were not sleep deprived, P300 latency was decreased during the late evening session. In contrast, when subjects were sleep deprived, P300 latency was significantly increased during the 21:30 session. P300 latency significantly predicted reaction time when subjects were sleep deprived, but not when they were not sleep deprived. CONCLUSIONS Sleep deprivation impacts negatively on the ability to correctly categorize stimuli in discrimination tasks especially during the late evening hours. The impact of sleep loss at critical hours of the day on cognitive processing should be included in any planning of operations requiring speedy and accurate decisions.