目的：通过近红外光谱成像技术（fNIRS）实时监测健康受试者和脑卒中患者虚拟现实（VR）环境下 步行时脑区皮质激活情况，初步探讨VR步态康复训练下脑卒中患者大脑活动的变化。方法：招募脑卒中 患者和健康志愿者各17例，分别纳入卒中组和对照组。采用Berg平衡功能量表（BBS）评估受试者的平 衡功能，采用计时起立-行走测（TUGT）评估受试者步行功能；通过fNIRS采集评估2组受试者静息状态、 一般环境踏步和VR环境下踏步时脑区氧合血红蛋白（HbO）浓度数据；分析2组受试者脑区显著激活通 道与Berg、TUGT量表评分的相关性。结果：①静息状态下，与对照组相比，卒中组患者左侧运动感觉皮 质HbO浓度降低（P＜0.05），右侧运动感觉皮质HbO浓度升高（P＜0.05）；②一般踏步任务中，脑卒中患者 右侧运动皮质激活减少（P＜0.05），左侧感觉皮质激活增加（P＜0.05）；VR环境下踏步任务时，脑卒中患者 左侧前额叶激活增加（P＜0.05）。③2组受试者脑区激活有显著差异的通道与Berg总分和TUGT得分进 行Spearman相关性分析结果显示，在VR环境下踏步中，前额叶皮质激活（ch9）和Berg总分呈负相关（r= －0.521，P=0.003），与TUGT得分呈正相关（r=0.434，P=0.017）。结论：VR影响脑卒中患者步行功能的机 制可能与前额叶皮质的激活变化有关。

To monitor the cerebral cortical activation in healthy subjects and stroke patients during walking under virtual reality (VR) environment in real-time using functional near-infrared spectroscopy (fNIRS), and to preliminarily explore the changes in brain activity of stroke patients under VR gait rehabilitation training. Methods: A total of 17 stroke patients and 17 healthy volunteers were recruited and assigned to the stroke group and the control group, respectively. The Berg Balance Scale (BBS) was used to assess the balance function of the subjects, and the Timed Up and Go Test (TUGT) was used to evaluate their walking function. The fNIRS was employed to collect data on the concentration of oxyhemoglobin (HbO) in the brain regions of the two groups of subjects at rest, during stepping in a general environment, and during stepping in a VR environment. The correlation between the significantly activated channels in the brain regions of the two groups and the scores of BBS and TUGT was analyzed. Results: (1) At rest, compared with the control group, the stroke group had a lower HbO concentration in the left sensorimotor cortex (P< 0.05) and a higher HbO concentration in the right sensorimotor cortex (P<0.05). (2) During the general stepping task, the activation of the right motor cortex decreased in stroke patients (P<0.05), while the activation of the left sensory cortex increased (P<0.05). In the VR stepping task, the activation of the left prefrontal cortex increased in stroke patients (P<0.05). (3) Spearman correlation analysis of the channels with significant differences in brain region activation between the two groups and the total scores of Berg and TUGT scores showed that in the VR stepping task, the activation of the prefrontal cortex (ch9) was negatively correlated with the total scores of Berg (r=－0.521, P=0.003), and positively correlated with the TUGT scores (r=0.434, P=0.017). Conclusion: The mechanism by which VR affects the walking function of stroke patients may be related to changes in the activation of the prefrontal cortex.