A large body of data indicates that (+/-)3,4-methylenedioxymethamphetamine (MDMA, 'ecstasy') can damage brain serotonin neurons in animals. However, the relevance of these preclinical data to humans is uncertain, because doses and routes of administration used in animals have generally differed from those used by humans. Here, we examined the pharmacokinetic profile of MDMA in squirrel monkeys after different routes of administration, and explored the relationship between acute plasma MDMA concentrations after repeated oral dosing and subsequent brain serotonin deficits. Oral MDMA administration engendered a plasma profile of MDMA in squirrel monkeys resembling that seen in humans, although the half-life of MDMA in monkeys is shorter (3 vs 6-9 h). MDMA was biotransformed into MDA, and the plasma ratio of MDA to MDMA was 3-5 / 100, similar to that in humans. MDMA accumulation in squirrel monkeys was nonlinear, and plasma levels were highly correlated with regional brain serotonin deficits observed 2 weeks later. The present results indicate that plasma concentrations of MDMA shown here to produce lasting serotonergic deficits in squirrel monkeys overlap those reported by other laboratories in some recreational 'ecstasy' consumers, and are two to three times higher than those found in humans administered a single 100-150 mg dose of MDMA in a controlled setting. Additional studies are needed on the relative sensitivity of brain serotonin neurons to MDMA toxicity in humans and non-human primates, the pharmacokinetic parameter(s) of MDMA most closely linked to the neurotoxic process, and metabolites other than MDA that may play a role.