Abstract: OBJECTIVE: Upper limb dysfunction after stroke is a common complication that seriously affects the quality of life and daily activity of patients. Virtual reality technology, as an emerging rehabilitation method, can effectively promote neural remodeling and functional recovery. This study will systematically evaluate the therapeutic effects of different virtual reality devices on upper limb motor dysfunction in patients with stroke.
METHODS: The China National Knowledge Infrastructure (CNKI), WanFang Database, VIP website, PubMed, Web of Science, Embase, and the Cochrane Library were searched to retrieve relevant literature. Relevant data were extracted, and their quality was assessed. The control group received conventional rehabilitation treatment, while the experimental group received virtual reality rehabilitation training in addition to the treatment provided to the control group. Statistical analysis was performed using RevMan 5.4 and Stata 18.0 software.
RESULTS: (1) A total of 12 articles and 571 patients were included in the final analysis. (2) Meta-analysis results showed that the Fugl-Meyer Assessment-Upper Extremity score of the experimental group was 7.29 times higher than that of the control group [mean difference (MD) = 7.29, 95% confidence interval (CI) 5.60–8.98, P < 0.05]. The Action Research Arm Test score of the experimental group was 10.69 times higher than that of the control group (MD = 10.69, 95% CI 4.96–16.43, P < 0.05). Additionally, the Modified Barthel Index score of the experimental group was 8.25 times higher than that of the control group (MD = 8.25, 95% CI 3.38–13.12, P < 0.05). (3) The subgroup analysis results showed that patients aged 50-59 years had better improvements in Fugl-Meyer Assessment-Upper Extremity score. Patients with a disease duration of less than 3 months also showed greater improvements in Fugl-Meyer Assessment-Upper Extremity score. Furthermore, when the intervention period was ≥ 4 weeks, the improvement in Fugl-Meyer Assessment-Upper Extremity score was the most significant. (4) Network meta-analysis demonstrated that the smart glove intervention [MD = -1.05, 95% CI (-1.85, -0.24), P < 0.05] was the most effective for improving upper limb motor function. The Armeo Spring intervention [MD = -1.19, 95% CI (-1.87, -0.51), P < 0.05] was the most effective for enhancing upper limb coordination, while the Kinect intervention [MD = -0.59, 95% CI (-1.13, -0.06), P < 0.05] was most effective for improving hand dexterity. The VREX intervention [MD = -0.76, 95% CI (-1.28, -0.23), P < 0.05] was the most effective for improving activities of daily living. 
CONCLUSION: The smart glove system is the preferred choice for improving upper limb motor function. The Armeo Spring system is the preferred choice for enhancing upper limb coordination. The Kinect system is the preferred choice for improving hand dexterity. The VREX system is the preferred choice for enhancing daily living activities. This study has certain limitations, and the conclusions drawn should be considered as rational references.

Key words: virtual reality devices, daily living ability, upper limb coordination ability, individual differences, system evaluation