The time courses of the effects of acute doses of amphetamine (5 and 10 mg/70 kg), alcohol (0.3 and 0.6 g/kg), diazepam (5 and 10 mg/70 kg), and marijuana (2.0% and 3.5% delta 9-THC) on performance engendered by each of four computerized behavioral tasks were evaluated in six human subjects. These performance-based tasks have potential commercial utility for drug-use detection in the workplace. Alcohol and marijuana effects were reliably detected for up to three hours following dose administration with most procedures. Amphetamine and diazepam effects were also detected, but the dose effects and time courses were variable. The profile of behavioral effects varied across drugs, suggesting that performance-based testing procedures might be useful in discriminating which drug was administered and the time course of the drug's effects. Results indicate that repeated measurement with performance-based drug detection procedures can provide immediate indications of performance impairment in a cost-effective and noninvasive manner and, as such, would be a useful supplement to biological sample testing for drug-use detection.