Fluoxetine, a selective inhibitor of serotonin (5-HT) uptake, was compared with placebo in a randomized double-blind longitudinal trial in 12 healthy volunteers. Sleep polygraphic recordings were performed at home twice before and once after 6 days of medication. After 6 days fluoxetine significantly decreased the amount of rapid eye movement (REM) sleep. The sleep-onset latency and REM latency were increased, but there was no significant increase in the amount of awakenings during night. The relative proportion of stages 2 and 3 increased after fluoxetine administration, although there was no significant change with regard to total amount of slow-wave sleep. Fluoxetine did not induce prominent eye movements during non-rapid eye movement (NREM) sleep in this study. Results of the subjective assessment revealed tendencies of improved sleep and well-being in the fluoxetine group. It is concluded that a comparatively small dose of fluoxetine (20 mg/day) causes the same type of changes in REM sleep which are characteristic of most antidepressive drugs.