Dose-Response Relationship of Virtual Reality Training in Parkinson's Disease: A Meta-Analysis of Overall Motor Function, Functional Mobility, and Balance Confidence.

Abstract

To investigate the dose-response relationship of virtual reality training (VRT) on overall motor function (UPDRS-III), functional mobility (TUG), and balance confidence (ABC) in Parkinson's disease (PD) to optimize rehabilitation protocols. A comprehensive literature search was performed across six electronic databases-PubMed, Embase, Cochrane Library, Web of Science, EBSCOhost, and CNKI-from their inception to August 3, 2025. A systematic search identified 31 randomized controlled trials (RCTs) (n=1,110) evaluating the effects of VRT on balance capabilities and functional gait in patients with PD. Data were extracted into three categories: bibliographic details, participant characteristics, and intervention specifics, alongside quantitative data for TUG, UPDRS-III, and ABC scores. Data extraction was independently conducted by two reviewers, followed by cross-verification and adjudication. Weighted mean differences (WMD) and univariate meta-regression were analyzed following PRISMA guidelines. VRT significantly improved TUG (WMD = -3.69) and ABC (WMD = 11.80), both achieving the minimal clinically important difference (MCID). VRT alone improved TUG but missed the MCID. No significant improvement was observed for UPDRS-III overall. Optimal TUG parameters were identified as 21-40 minutes/session, 4-7 sessions/week, for 6-8 weeks. High-frequency training (≥4 sessions/week) was required for UPDRS-III benefits. Commercial and specialized platforms showed comparable efficacy. VRT significantly enhances TUG and ABC, particularly achieving clinical utility (MCID) when combined with physical therapy. While UPDRS-III effects are limited, high-frequency protocols show potential. A frequency of 4-7 sessions per week was associated with clinically important improvements in TUG. Additionally, patients aged 70 years and older showed no statistical heterogeneity in their response to the intervention (I²= 0%). However, this finding should be interpreted cautiously, as it is based on a small number of studies and may reflect limited statistical power rather than true clinical predictability.