Verbal and symbol concentration, learning, problem solving, clear thinking, manual skills, and memory were tested in 42 surgical residents to assess the effects of acute sleep deprivation on specific neuropsychological parameters. A series of eight neuropsychological tests--digit symbols, digit vigilance, story memory, trail making, PASAT, Raven matrices, delayed story, and pegboard--and a questionnaire on mood states were completed by the residents both when fatigued (less than 4 hours of sleep: mean, 2.0 +/- 1.5 hours) and when rested (more than 4 hours of sleep: mean, 6.5 +/- 1.0 hours), with at least 7 days between tests. In order to eliminate the effects of learning from the first test series, randomization of residents was performed so that one half were first evaluated when rested and one half when fatigued. ANOVA, multiple regression analysis, and the Student t test were used to assess differences. In the acute sleep-deprived state, residents were less vigorous and more fatigued, depressed, tense, confused, and angry (p less than 0.05) than they were in rested state. Despite these changes in mood, however, the responses on all of the functional tests were no different statistically in those who were rested and those who were fatigued (even in those with less than 2 hours' sleep). We conclude that acute sleep deprivation of less than 4 hours alters mood state but does not change performance in test situations in which concentration, clear thinking, and problem solving are important.