This study examines the effect of apomorphine (APO), a nonselective D(1)- and D(2)-dopamine receptor agonist, on the firing activity of neurons in the subthalamic nucleus (STN) and internal segment of the globus pallidus (GPi) in patients with Parkinson's disease (PD). Single-unit microelectrode recordings were conducted in 13 patients undergoing implantation of deep brain stimulation electrodes in STN and 6 patients undergoing a pallidotomy. Doses of APO (2.5-8 mg) were sufficient to produce an ON state, but not intended to induce dyskinetic movements. Following baseline recordings from a single neuron, APO was administered and the activity of the neuron followed for an average of 15 min. The spontaneous discharge of neurons encountered before (n = 309), during (n = 146, 10-60 min), and after the effect of APO had waned (n = 127, >60 min) was also sampled, and the response to passive joint movements was noted. In both nuclei, APO increased the overall proportion of spikes in burst discharges (as detected with Poisson "surprise" analysis), and a greater proportion of cells with an irregular discharge pattern was observed. APO significantly decreased the overall firing rates of GPi neurons (P < 0.01), but there was no change in the overall firing rate of neurons in the STN (P = 0.68). However, the mean firing rates of STN neurons during APO-induced movements (choreic or dystonic dyskinesias) that occurred in four patients were significantly lower than OFF-period baseline values (P < 0.05). Concurrent with a reduction in limb tremor, the percentage of cells with tremor-related activity (TCs) was found to be significantly reduced from 19 to 6% in the STN and 14 to 0% in the GPi following APO administration. APO also decreased the firing rate of STN TCs (P < 0.05). During the OFF state, more than 15% of neurons tested (STN = 93, GPi = 63) responded to passive movement of two or more joints. After APO, this proportion decreased significantly to 7% of STN cells and 4% of GPi cells (STN = 28, GPi = 26). These findings suggest that the APO-induced amelioration of parkinsonian symptoms is not solely due to a decrease in overall activity in the GPi or STN as predicted by the current model of basal ganglia function in PD.