Nicotine produces interoceptive stimulus effects in humans, which may be critical in understanding tobacco use. It has not yet clearly been demonstrated that discrimination of nicotine, or any drug, in humans is due to its central effects. We compared effects of mecamylamine (10 mg p.o.), a central and peripheral nicotine antagonist, on nicotine discrimination with those of trimethaphan (10-40 microg/kg per min i.v.), a peripheral nicotine antagonist only, and placebo. Smokers (n = 6) were first trained to reliably discriminate 0 versus 20 microg/kg nicotine by nasal spray and then tested on generalization of this discrimination across a range of nicotine doses (0, 3, 6, 12, 20 microg/kg) following antagonist/placebo pretreatment. Nicotine self-administration was also assessed after generalization testing by having participants intermittently choose between nicotine versus placebo spray. Compared with responding following placebo pre-treatment, discrimination of the highest dose of nicotine was significantly attenuated following mecamylamine but not trimethaphan. Similar results were observed for some subjective responses to nicotine. Mecamylamine also tended to increase nicotine self-administration. Consistent with previous animal studies, these results suggest that discriminative stimulus effects of nicotine in humans are mediated at least in part by its central effects.