Z Zhang, Yi Zhou
Parkinson's disease (PD) is the second most common neurodegenerative disorder, featuring progressive dopaminergic neuronal loss, Lewy body (LB) pathology, and a broad spectrum of motor and non-motor symptoms. Although dopaminergic pharmacotherapy remains the cornerstone of symptomatic treatment, long-term medication use is frequently complicated by motor fluctuations, dyskinesia, and limited efficacy for several non-motor symptoms. Exercise has therefore gained increasing attention as a non-pharmacological intervention for PD. This narrative review summarizes current evidence on how exercise intensity influences motor symptoms, non-motor symptoms, and putative biological mechanisms in PD. Low-intensity exercise appears suitable for improving tolerance, mobility, and adherence, particularly in frail or advanced patients. Moderate-intensity exercise may provide a favorable balance between safety and functional benefit, whereas high-intensity aerobic exercise (AE) and high-intensity interval training may improve cardiorespiratory fitness and selected motor outcomes in carefully screened and supervised patients. For non-motor symptoms, exercise effects appear domain-specific: Aerobic and socially engaging exercise may be more relevant for mood symptoms; cognitively demanding and dual-task exercise may benefit executive function; and lower-intensity or timing-adjusted protocols may be preferable for patients with fatigue, sleep disturbance, or autonomic dysfunction. Potential mechanisms include modulation of neurotrophic signaling, neuroplasticity, neuroinflammation, oxidative stress, and cerebral perfusion; however, much of the mechanistic evidence remains indirect or derived from preclinical models. Current clinical evidence is limited by heterogeneous definitions of exercise intensity, small samples, short intervention periods, limited long-term follow-up, and underrepresentation of advanced PD. Future studies should standardize intensity reporting, assess clinically meaningful changes rather than statistical significance alone, and integrate clinical outcomes with validated biomarkers to support individualized exercise prescription.