Effectiveness of Virtual Reality Balance Training in Post-Stroke Rehabilitation

Thesis Abstract

Post-stroke individuals often experience persistent balance impairments that limit independence and increase fall risk, yet conventional rehabilitation methods may insufficiently address dynamic balance and real-world functional tasks. This study investigates the effectiveness of a structured virtual reality balance training (VRBT) program as an adjunct to standard physiotherapy in improving postural control, gait stability, and activities of daily living among adults with chronic stroke. The aim is to determine whether VRBT yields superior balance outcomes compared with conventional therapy alone and to explore the mechanisms underpinning any observed effects through biomechanical and neurophysiological measures. The study adopts a parallel-group randomized controlled trial design with blinded assessors. Participants will be 120 adults (ages 40–75) at least six months post-stroke, recruited from three urban rehabilitation centers. Eligible participants will have residual balance deficits (Berg Balance Scale 20–45) and adequate cognitive function (Montreal Cognitive Assessment score ? 23). Participants will be randomly assigned (11) to either the VRBT group or the control group receiving conventional balance-focused rehabilitation, for 12 weeks, with sessions three times per week, each lasting 60 minutes. Primary outcomes will include dynamic balance performance measured by the Timed Up and Go with dual task (TUG-DT) and computerized dynamic posturography (CDP) composite scores. Secondary outcomes will encompass gait speed (10-meter walk test), functional mobility (Fugl-Meyer Assessment for the lower extremity), activities of daily living (Barthel Index), falls self-efficacy (ABC Scale), and quality of life ( Stroke-Specific Quality of Life Scale). Secondary biomechanical outcomes will incorporate ground reaction force metrics and center of pressure metrics obtained via instrumented treadmill gait analysis. Neurophysiological correlates will be examined through event-related potentials (P300 latency) and motor-evoked potentials (MEP) elicited by repetitive transcranial magnetic stimulation to probe cortico-motor excitability changes. Instruments include validated VRBT hardware and software delivering progressively challenging balance tasks, a standardized physiotherapy protocol for the control condition, a 6-minute walk test for endurance, and safety monitoring logs for adverse events. Data will be analyzed using intention-to-treat principles. Primary analysis will employ mixed-effects linear models to compare post-intervention outcomes between groups, adjusting for baseline values and covariates (age, time since stroke, lesion side). Mediation analyses will test whether improvements in postural sway and reactive balance mediate gains in functional mobility. Repeated-measures ANOVA will assess within-group changes over time (baseline, mid-intervention, post-intervention, and 3-month follow-up). Secondary biomechanical data will be analyzed with repeated-measures MANOVA to examine interaction effects between group and time on center of pressure velocity and mediolateral sway parameters. Neurophysiological data will be analyzed using repeated-measures ANOVA to detect changes in P300 latency and MEP amplitude, complemented by correlational analyses linking neurophysiological changes with functional outcomes. Multiple imputation will address missing data, and sensitivity analyses will test the robustness of the findings. A preregistered analytic plan will guide hypothesis testing to mitigate selective reporting. Anticipated findings include superior improvements in dynamic balance (TUG-DT and CDP scores) and functional mobility (Fugl-Meyer lower extremity, Barthel Index) in the VRBT group compared with controls, along with enhanced gait stability metrics. It is expected that VRBT will produce greater improvements in balance confidence and quality of life, with concurrent neurophysiological indicators of enhanced cortico-motor connectivity (reduced P300 latency, increased MEP amplitude). The study aims to elucidate whether VRBT fosters neuroplastic adaptations that translate to functional gains, through enhanced sensorimotor integration and task-specific practice. The research contributes to knowledge by providing robust evidence on the additive value of immersive VR-based balance training in post-stroke rehabilitation, clarifying dose–response relationships, identifying potential moderators (age, time since stroke, baseline balance), and delineating neurophysiological mechanisms underlying functional recovery. Practical implications include informing rehabilitation guidelines, optimizing resource allocation for VR-enabled therapies, and guiding integration of VRBT into multidisciplinary stroke recovery programs. The study concludes with policy-relevant recommendations for implementing scalable VR balance interventions in clinical practice, strategies to maximize adherence, and directions for future research, including long-term maintenance of gains and customization for comorbidities.