It has been reported that the cerebellar vermis is equally involved in both motor imagery about axial movement and the actual execution of postural balance in healthy human subjects, but this finding is yet to be explored in Parkinson's disease (PD). We therefore investigated the neuronal responses during observation of standing posture, imagination of standing and the assumption of an upright posture in ten drug-naïve PD patients using positron emission tomography (PET) with [15O]H2O and evaluated dopamine dysfunction by measuring the level of dopamine transporter binding of [11C]CFT. Within-group statistical parametric mapping (SPM) analysis showed similar cerebellar activation during imagination of standing and its real execution between the PD and control groups (12 healthy subjects); i.e., increases in regional cerebral blood flow (rCBF) were observed in the anterior cerebellar vermis during motor imagination and the posterior vermis during standing. A comparison between the groups showed that the motor execution of standing significantly activated the superior part of the posterior vermis (declive VI) and the paracentral sulcus region in the PD patients, while the prefrontal cortices were deactivated during standing (p<0.001 uncorrected). Correlation analysis within the PD group revealed that the postural rCBF increases in the cerebellar vermis (pyramis) were negatively correlated with putaminal [11C]CFT binding (p<0.01, r = 0.94) and that the postural rCBF reductions in the orbitofrontal cortex were positively correlated with caudate [11C]CFT binding (p<0.05, r = 0.70). These results suggest that while the neural circuits for postural imagery and execution are intact in PD, standing performance, which requires more recruitment of dopaminergic control, may result in compensatory overstimulation of the cerebellar vermis and paracentral foot area in PD patients. Hyperactivity in these areas along with mesocortical hypofunction may be pathophysiological aspects of postural control in PD patients. Hence, our findings would help understand the modifications observed within the neural networks in relationship with postural performance, and possible compensatory mechanisms in PD.