Virtual reality training enhances gait poststroke: a systematic review and meta-analysis

Virtual reality (VR)-based interventions are gaining widespread attention for managing neurological disorders such as stroke. A metastatistical consensus regarding the intervention is strongly warranted. In this study, we attempt to address this gap in the literature and provide the current state of evidence for the effects of VR on gait performance. We conducted both between- and within-group meta-analyses to provide a state of evidence for VR. Moreover, we conducted a search adhering to PRISMA guidelines on nine databases. Out of 1866 records, 32 studies involving a total of 809 individuals were included in this review. Considering all included studies, significant enhancements in gait parameters were observed with VR-based interventions compared with conventional therapy. A between-group meta-analysis reported beneficial significant medium effects of VR training on cadence (Hedge's g = 0.55), stride length ((STrL; Hedge's g = 0.46), and gait speed (Hedge's g = 0.30). Similarly, a within-group meta-analysis further revealed positive medium effects of VR on cadence (Hedge's g = 0.76), STrL (Hedge's g = 0.61), and gait speed (Hedge's g = 0.69). Additional subgroup analyses revealed beneficial effects of joint application of VR and robot-assisted gait training on gait speed (Hedge's g = 0.50). Collectively, findings from this review provide evidence for the effectiveness of VR-based gait training for stroke survivors.