Eight male frequent cocaine smokers participated in a 44- to 47-day inpatient and outpatient study to assess the effects of the noncompetitive N-methyl-D-aspartate (NMDA) antagonist, memantine, on cocaine self-administration, subjective effects, and psychomotor performance. Participants were maintained on memantine (0 and 20 mg daily) for 7-10 days prior to laboratory testing, using a double-blind crossover design. Under each medication condition, participants smoked four doses of cocaine base (0, 12, 25 and 50 mg), and were subsequently given five opportunities, 14 min apart, to self-administer that dose of cocaine or receive a merchandise voucher ($5.00). Each cocaine dose was tested twice under each medication condition, and the order of medication condition and cocaine dose was systematically varied. Vital signs were recorded every 2 min, and subjective effects were assessed at baseline and after each cocaine or voucher delivery. In addition, psychomotor performance was assessed before and after each self-administration session. Memantine maintenance was not associated with changes in psychomotor performance or the number of cocaine doses chosen each session. Memantine maintenance was, however, associated with significant increases in some subjective effects of cocaine, including ratings of 'good drug effect', 'high', 'potency', 'quality', and street value. These data suggest that NMDA antagonists may have limited usefulness as treatment medications for cocaine abuse.