Pharmacological specificity of the caffeine discriminative stimulus in humans: effects of theophylline, methylphenidate and buspirone.

The present study examined further the pharmacological specificity of the methylxanthine CNS stimulant caffeine as a discriminative stimulus in humans. Nine normal healthy volunteers (ages 19-39) were trained to discriminate between caffeine (320mg/70kg, p.o.) and placebo, using monetary reinforcement of correct letter code identification. After four training sessions, subjects were tested with the training conditions until they were >80% correct on four consecutive sessions. Then dose-effect curves were determined for caffeine (56-320mg/70kg), theophylline (56-320mg/70kg), methylphenidate (10-56mg/70kg), and buspirone (1-32mg/70kg). Seven of nine subjects met the discrimination criterion within four to nine sessions. During dose-effect curve determinations, caffeine and methylphenidate each produced dose-related increases in caffeine-appropriate responding. Theophylline produced caffeine-appropriate responding that was not dose related in a consistent manner across subjects, occasioning an average of 50% caffeine-appropriate responding at most doses tested. Buspirone produced predominantly placebo-appropriate responding. Caffeine-appropriate responding tended to be directly related to ARCI LSD scores, self-reported "bad" effects, "high", and stimulant-bad effects and inversely related to ARCI PCAG scores and sedative ratings. These results agree with non-human data and suggest that the caffeine discriminative stimulus has pharmacological specificity, in that caffeine-appropriate responding generalizes to other stimulants such as theophylline or methylphenidate, but not to non-stimulant compounds such as buspirone.