Abnormalities in visuo-motor control have repeatedly been reported in Parkinson's disease (PD) patients. In the more recent studies, tasks measuring visuo-motor performance are usually computerised tasks requiring the use of a mouse-like manipulandum. In healthy subjects, previous computer mouse experience can influence performance in computerised visuo-motor tasks. We, therefore, investigated the potential confounding effect of mouse experience in a visuo-motor task used in PD patients, the visuo-motor testing (VMT) system, and its concurrent usefulness in the diagnosis of PD. Our study population included 49 early PD patients and 31 controls. The VMT system involves moving a pointer over a sinusoidal path displayed on a computer screen by moving an unseen digitiser mouse over a tablet. In both PD patients and controls, subjects with low mouse experience scored worse than subjects with high mouse experience on variables measuring movement accuracy, direction and speed. After correction for mouse experience the aforementioned deficits were still present in PD patients with low mouse experience. However, PD patients with high mouse experience only showed deficits in movement accuracy. It would seem that previous mouse experience can at least partly compensate for deficits in directional control and speed in PD patients, implying a possible role for training in the treatment of functional motor impairments in PD. Future studies using computerised psychomotor tasks should take sufficient precautions to avoid potential confounding effects of mouse experience. The present data further suggest that only accuracy of visuo-motor control might be a useful parameter in the (early) diagnosis of PD. © 2002 Elsevier Science Ltd. All rights reserved.