Several recent neuroimaging studies in humans and in monkeys using different radiotracers have reported widely differing values of dopamine transporter (DAT) occupancy by doses of cocaine which are perceived as reinforcing by humans. Here we tested the hypothesis that the measurement of DAT occupancies by drugs with fast pharmacokinetics such as cocaine requires a radioligand with similar kinetics in order to effectively compete with the drug. We measured DAT occupancy by four different doses of cocaine (1.0, 0.5, 0.25, and 0.1 mg/kg) using [11C]d-threo-methylphenidate (a radiotracer which binds rapidly to the DAT in vivo) and compared them to estimates reported previously using [11C]cocaine in the same two baboons and with the same four doses of cocaine [Volkow et al. (1996b) Synapse 24:399-402). Cocaine reduced [11C]d-threo-methylphenidate binding in striatum in a dose-dependent manner, and values were significantly correlated with those obtained previously with [11C]cocaine (r = 0.9, F = 37, P < 0.001). The ED50s (50% occupancy of DAT by cocaine) were 0.27 and 0.17 mg/kg for [11C]d-threo-methylphenidate and [11C]cocaine, respectively. This is significantly lower than values obtained with labeled beta-CIT and other similar radiotracers with a slow uptake and clearance (ED50s: 3-7 mg/kg). We conclude that in vivo measurements of DAT occupancy by rapidly clearing drugs like cocaine requires the use of radiotracers having similar kinetics to the drug itself.