Effects of different interventions based on virtual reality on cognitive and motor function in elderly people with mild cognitive impairment or dementia: a systematic review and network meta-analysis.

Virtual reality (VR)-based interventions are recognized as promising non-pharmacological therapies for enhancing cognitive and motor function in elderly patients with cognitive impairment. However, the comparative effectiveness of different VR-based interventions remains to be established. To address this research gap, a frequentist framework was utilized to conduct a network meta-analysis that integrated both direct and indirect comparative evidence. A comprehensive electronic literature search was conducted in six English databases and three Chinese databases to identify relevant randomized controlled trials (RCTs) published up to October 2025. A random effects model was selected for traditional meta-analysis to directly examine the efficacy of different non-pharmacological interventions. A consistency model was chosen to perform a network meta-analysis to evaluate the relative effects and rank probabilities of other non-pharmacological interventions. A total of 42 randomized controlled trials (RCTs) were included, encompassing four distinct non-pharmacological interventions. Risk of bias was assessed with the Cochrane Risk of Bias 2 tool, and certainty of evidence with CINeMA (Confidence in NMA). Preliminary evidence suggests that virtual reality sport training (VR-SPT) may positively affect cognitive outcomes in patients with cognitive impairment, particularly those with mild cognitive impairment (MCI). However, the evidence supporting the executive function benefits is extremely limited. Specifically, the potential advantages of VR-SPT for executive function rest on only two studies, and the confidence intervals reported were very wide, rendering these findings highly uncertain. Concurrently, the available evidence suggests that VR comprehensive cognitive game combined with sport training (VR-CCG+SPT) may offer a potential advantage in improving motor function in patients with cognitive impairment. Furthermore, in patients with dementia, VR-CCG+SPT also demonstrated more favorable effects on global cognitive function. This network meta-analysis provides a comprehensive comparison of four different virtual reality (VR) interventions on cognitive and motor function in patients with cognitive impairment, offering preliminary evidence for clinical practice and future research. The selection of VR interventions should be guided by the patient's living environment, personal preferences, and the aforementioned preliminary findings, to delay cognitive and motor function decline. However, due to limitations such as a relatively weak evidence base for key comparisons, overly wide confidence intervals for some results, and the use of overlapping MoCA domains in different outcome networks, the results of this study should be interpreted with caution, and the above conclusions require further research for validation. Future research should prioritize conducting well-designed, multi-arm randomized controlled trials to provide more robust evidence concerning the relative efficacy of various VR interventions. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251034623, identifier CRD420251034623.