The effects of ritanserin, a 5-hydroxytryptamine-2 (5-HT2) receptor antagonist, and chlordiazepoxide, a benzodiazepine agonist, on sleep-wakefulness disturbances in rats after acute administration of cocaine and after discontinuation of chronic cocaine treatment were examined. Intraperitoneal (IP) injection of chlordiazepoxide (10 mg/kg) but not ritanserin (0.63 mg/kg) prevented the increase of wakefulness (W) and the reduction of light slow wave sleep (SWS1) and deep slow wave sleep (SWS2) induced by an acute injection of cocaine (20 mg/kg IP). Daily injection of cocaine (20 mg/kg for 5 days, then 30 mg/kg for 5 days IP) at the onset of the light phase elicited an increase of W and a concomitant decrease of SWS1, SWS2 and paradoxical sleep (PS) in the light phase, followed by a rebound in SWS2 and PS in the subsequent dark phase. Following cocaine discontinuation, the circadian distribution of sleep-wakefulness states remained disturbed in saline-treated rats for at least 5 days. Both ritanserin (0.63 mg/kg IP/day) and chlordiazepoxide (10 mg/kg IP/day) reduced the alteration in the distribution of W and SWS2 throughout the light-dark cycle from the first day of administration on, but failed to prevent PS alterations. The mechanisms by which both compounds exert their effect are probably quite different. For chlordiazepoxide sedative and sleep-inducing properties probably play a major role. In contrast, for ritanserin SWS2-increasing properties and its ability to reverse preference for drugs of abuse without inducing aversion might be key factors.