Patients with Parkinson's disease experience motor and perceptual timing difficulties, which are ameliorated by dopaminergic medication. We investigated the neural correlates of motor timing in Parkinson's disease, including the effects of dopaminergic medication on patterns of brain activation. Eight patients with Parkinson's disease and eight healthy controls were scanned with H(15)(2)O positron emission tomography while engaged in three tasks: synchronization (right index finger tapping in synchrony with a tone presented at 1 Hz), continuation (tapping at 1 Hz in the absence of a tone), and a control simple reaction time task. During the first 6 scans, the patients were assessed after overnight withdrawal of medication. Scans 7-12 were completed with the patients in the 'ON' state, after injections of apomorphine, a dopamine receptor agonist. For the healthy controls, relative to the control reaction time task, motor timing (synchronization + continuation) was associated with significantly greater activation in left medial prefrontal cortex (Brodmann area 10, 32), right hippocampus, bilateral angular gyrus (Brodmann area 39), left posterior cingulate (Brodmann area 31) and left nucleus accumbens/caudate. This pattern of brain activation during motor timing was not observed for patients, who showed significantly greater activation in bilateral cerebellum, right thalamus and left midbrain/substantia nigra compared to the control participants. Relative to the externally-paced synchronization task, the internally controlled continuation task was associated with greater activation in the dorsolateral prefrontal cortex (Brodmann area 46/9) in both the control and Parkinson's disease groups. Analysis of medication-related effects indicated that cortical activation was significantly more predominant during motor timing when the patients were 'ON' medication, whereas pallidal and cerebellar activations were evident 'OFF' medication. Effective connectivity analysis established that activity in the left caudate nucleus was associated with increased activity in the right lentiform nucleus and cerebellum 'OFF' medication, and with increased activity in the prefrontal cortex 'ON' medication. These results suggest that in Parkinson's disease, in the 'OFF' medication state, excessive inhibitory pallidal outflow is associated with a lack of adequate frontal activation and reliance on the cerebellum for motor timing. In contrast, our results establish for the first time that administration of dopaminergic medication increases striatal-frontal connectivity between the caudate nucleus and prefrontal cortex during motor timing.