目的：系统评价非侵入性神经调控联合虚拟现实技术（NINT-VR）对脑卒中患者上肢运动功能的临 床疗效。方法：检索PubMed、Embase、Web of Science、Cochrane Library、Medline和中国知网、万方、维普 数据库中关于 NINT-VR 治疗脑卒中后上肢运动功能障碍的临床随机对照试验，检索时间自建库起至 2024年9月。使用物理治疗证据数据库量表（PEDro）进行文献质量评估，RevMan 5.4.1软件进行常规Meta分析，Stata 17.0软件进行网状Meta分析、敏感性分析和发表偏倚评价。结果：纳入12篇文献，共538例 受试者。Meta分析显示，NINT-VR可改善脑卒中患者的Fugl-Meyer上肢运动功能评定量表（FMA-UE）评 分（MD=8.32，95% CI 6.18～10.45，P＜0.00001），上肢动作研究量表（ARAT）评分（MD=9.58，95% CI 6.24～12.91，P＜0.00001）和改良 Barthel 指数（MBI）评分（MD=11.08，95%CI 8.52～13.65，P＜0.00001）。 亚组分析显示，NINT-VR干预方式和单次治疗时间是显著的异质性来源，相比对照组，单次治疗时间≤ 30 min（MD=4.68，95%CI 1.93～7.43，P=0.0008）或＞30 min（MD=10.22，95%CI 8.55～11.89，P＜0.00001）， FMA-UE 评分均有显著改善。对FMA-UE评分进行网状Meta分析，不同NINT-VR干预方式效果的累积 概率排序为经颅直流电刺激联合VR非同时治疗＞经颅磁刺激联合VR非同时治疗＞经颅直流电刺激联 合VR同时治疗＞单独VR训练。结论：NINT-VR在脑卒中后改善患者上肢运动功能和日常生活活动能 力方面具有显著优势，以经颅直流电刺激联合VR非同时治疗的干预方式效果最好。

To systematically evaluate the clinical efficacy of non-invasive neuromodulation techniques combined with virtual reality (NINT-VR) for upper limbs motor dysfunction in stroke patients. Methods: A comprehensive search in databases of PubMed, Embase, Web of Science, Cochrane Library, Medline, CNKI, Wanfang, VIP Database was performed for randomized controlled trials (RCTs) of NINT-VR for upper limbs motor dysfunction in stroke patients. The search time was from databases inception to September 2024. Quality assessment was performed using Physiotherapy Evidence Database (PEDro). Meta-analysis was performed using RevMan 5.4.1. Network meta-analysis was performed and funnel plots were drawn to test for publication bias of the outcome indicators using Stata 17.0 software. Results: A total of 12 RCTs involving 538 patients were included. The meta-analysis results showed that, NINT-VR improved fugl-meyer assessment upper extremities scale (FMA-UE) (MD=8.32, 95% CI 6.18～10.45, P<0.00001) in stroke patients. Action research arm test (ARAT) score (MD=9.58, 95% CI 6.24～12.91, P<0.00001) and modified barthel index (MBI) score (MD=11.08, 95% CI 8.52～13.65, P<0.00001) also had been improved compared to the control group. Subgroup analysis showed that NINT-VR intervention plan and duration of single treatment were significant sources of heterogeneity. FMA-UE score could be improved by NINT-VR for both less than 30 minutes at a time (MD=4.68, 95%CI 1.93～7.43, P=0.0008) and more than 30 minutes at a time (MD=10.22, 95% CI 8.55～11.89, P<0.00001). For the effect of NINT-VR intervention plan, the cumulative probability ranking of network meta-analysis were transcranial direct current stimulation (tDCS) and VR (non-simultaneous)>transcranial magnetic stimulation (TMS) and VR (non-simultaneous)>tDCS and VR (simultaneous) >VR treatment. Conclusion: NINT-VR demonstrates significant benefits for improving upper limbs motor dysfunction and activity of daily living (ADL) in stroke patients, and tDCS and VR (non-simultaneous) was one of the most effective intervention plan.