目的：系统评价虚拟现实技术对脑卒中患者认知功能的影响。方法：计算机检索国内外多个数据库中关于虚拟现实技术干预脑卒中后认知障碍患者的随机对照试验，检索时限为建库至2025年1月7日。使用Cochrane协作网中的偏倚风险评估工具和物理治疗证据数据库（PEDro）量表评估文献质量，采用RevMan5.4.1软件进行数据分析。结果：最终纳入14篇文献，涉及723例患者。纳入的文献均为高质量文献，PEDro量表评分6～10分。Meta分析结果显示，虚拟现实技术可以改善脑卒中认知障碍患者蒙特利尔认知评估（MoCA）评分[MD=2.53, 95%CI (0.82, 4.24), P=0.004]、简易智能精神状态量表（MMSE）评分[MD=1.68, 95%CI (1.12, 2.24)，P＜0.00001]、P300潜伏期[MD=－25.48, 95%CI (－38.99, －11.96), P=0.0002]、P300波幅[MD=0.72, 95%CI (0.14, 1.30), P=0.02]、Barthel指数或改良Barthel指数（BI或MBI）评分[MD=5.14, 95%CI (3.25, 7.03), P＜0.00001]；但对连线测试TMT-A评分[MD =－12.47, 95%CI (－27.16, 2.23), P=0.10]、TMT-B评分效果不显著[MD = －44.35, 95%CI (－117.45, 28.74), P=0.23]。结论：虚拟现实技术可以改善脑卒中患者的认知功能和日常生活活动能力。

Objective: To systematically evaluate the impact of virtual reality (VR) technology on cognitive function in stroke patients. Methods: A computer-based search was conducted across multiple domestic and international databases for randomized controlled trials (RCTs) investigating VR interventions in patients with post-stroke cognitive impairment. The search period spanned from the inception of the databases to January 7, 2025. The Cochrane Collaboration's Risk of Bias tool and the Physiotherapy Evidence Database (PEDro) scale were employed to assess the quality of the included studies. Data analysis was performed using RevMan 5.4.1 software. Results: A total of 14 studies, involving 723 patients, were ultimately included. All included studies were of high quality, with PEDro scale scores ranging from 6 to 10. Meta-analysis results indicated that VR technology significantly improved the Montreal Cognitive Assessment (MoCA) score [MD=2.53, 95% CI (0.82, 4.24), P=0.004], the Mini-Mental State Examination (MMSE) score [MD=1.68, 95% CI (1.12, 2.24), P<0.00001], the P300 latency [MD=－25.48, 95% CI (－38.99, －11.96), P=0.0002], the P300 amplitude [MD=0.72, 95% CI (0.14, 1.30), P=0.02], and the Barthel Index or Modified Barthel Index (BI or MBI) score [MD=5.14, 95% CI (3.25, 7.03), P<0.00001] in stroke patients with cognitive impairment. However, no significant effects were observed on the Trail Making Test Part A (TMT-A) score [MD=－12.47, 95% CI (－27.16, 2.23), P=0.10] or the TMT-B score [MD=－44.35, 95% CI (－117.45, 28.74), P=0.23]. Conclusion: VR technology can improve cognitive function and activities of daily living (ADL) abilities in stroke patients.