Tardive dyskinesia (TD) occurs in approximately 20% of patients treated chronically with antipsychotic drugs and constitutes a major public health problem. The cause of this disorder remains unknown, and no effective treatment has yet been found. The major etiologic theory (dopamine [DA] supersensitivity hypothesis) suggests that TD is the pharmacologic opposite of Parkinson's disease and implies that all patients with TD should respond uniformly to specific pharmacologic agents. Clinical research, however, has not borne this out. To evaluate pharmacologic response in TD syndromes, 15 patients underwent single dose acute administration of four different drugs: a DA agonist (bromocriptine 5 mg orally), a DA antagonist (haloperidol 5 mg intravenously), a cholinergic agonist (physostigmine 2 mg intravenously) and a cholinergic antagonist (benztropine 4 mg intravenously), individually in separate procedures at weekly intervals for four consecutive weeks in randomized order and under controlled double-blind conditions. Patients were evaluated for their clinical and endocrine responses. Pre- and post-drug administration TD exams were blindly rated. Results were not consistent with the DA supersensitivity theory; instead they demonstrated marked inter- and intrasubject variability in pharmacologic responses. Greatest uniformity in response was found among the tardive dystonic subjects, although this also was not consistent with a DA supersensitivity hypothesis. TD appears to be a pharmacologically heterogeneous condition, which may reflect the neurochemical complexity of the basal ganglia.