Abstract
BACKGROUND: Stroke survivors frequently experience impaired trunk control, affecting sitting balance and daily activities. Virtual reality (VR) balance training presents an engaging alternative to conventional methods, though its comparative effectiveness remains uncertain. PURPOSE: To compare the effects of VR-based versus traditional balance training on trunk control, sitting balance, and ADL in stroke patients. METHODS: Thirty-six participants were randomly assigned to either VR-based training or conventional sitting balance training group. Both groups completed a 4-week intervention. Outcome measures including the Modified Barthel Index (MBI), Trunk Impairment Scale (TIS), and Center of Pressure (COP) parameters under static/dynamic sitting conditions were assessed at baseline and post-intervention. RESULTS: Both groups showed significant time effects in trunk control (TIS total score, p < 0.001), with the VR group demonstrating greater improvement (d = 1.72 vs. 1.07). A significant time×group interaction was found for dynamic balance (TIS-dynamic, p = 0.013) and dynamic mediolateral displacement (p = 0.016), indicating VR training produced superior gains. While both groups improved in activities of daily living (MBI, p < 0.001), no significant time×group interaction was observed. Only the VR group showed significant improvement in static anterior-posterior displacement (p = 0.039), with no time×group effects found for sway velocity or static mediolateral displacement. CONCLUSION: While both interventions similarly improved trunk control, sitting balance, and ADL, VR-based training demonstrated superior efficacy in enhancing lateral trunk control during dynamic tasks, suggesting its potential advantage for facilitating dynamic postural adjustments in stroke rehabilitation. TRIAL REGISTRATION: https://www.chictr.org.cn (No. ChiCTR2500100936); Registration date: 4-17-2025 (retrospectively registered).