基于运动想象的脑机接口系统联合不同末端效应器对慢性期脑卒中患者上肢运动功能改善的初步研究

Improving upper limb motor function in chronic stroke patients using a brain-computer interface system based on motor imagery combined with different end effectors:a preliminary study

目的:在接受基于运动想象的脑机接口康复训练的患者中,初步探究不同末端效应器对慢性期脑卒中患者上肢运动功能改善的有效性。方法:本研究纳入在我院接受基于运动想象的脑机接口康复训练的慢性期脑卒中患者32例,根据其所接受的训练不同,分为脑机接口联合虚拟现实反馈和脑机接口联合腕关节机器人反馈两组。在干预前后评估患者的上肢运动功能(简化Fugl-Meyer量表上肢部分)、偏瘫侧腕关节主动关节活动范围、偏瘫侧伸腕肌群主动收缩时的表面肌电信号和偏瘫侧运动诱发电位,比较虚拟现实反馈和腕关节机器人反馈对慢性期脑卒中患者上肢运动功能恢复的影响。结果:干预前,两组受试者的人口学资料、上肢运动功能基线数据无显著性差异(P=0.160),干预后,两组受试者的上肢运动功能较干预前显著改善(P<0.001),与虚拟现实反馈组相比,腕关节机器人反馈组干预后的Fugl-Meyer评分和干预前后评分的变化无显著性差异(P=0.06);在偏瘫侧腕关节主动关节活动范围上,干预前,两组受试者都无主动屈-伸腕活动,干预后腕关节机器人组改善为33.13°±24.96°,虚拟现实组改善为27.81°±37.17°,两组比较无显著性差异(P=0.638);在偏瘫侧伸腕肌群表面肌电图评定上,干预后两组的平均波幅和曲线下面积都较干预前显著改善(P<0.001),组间比较提示腕关节机器人组改善更为显著(P<0.001);在运动诱发电位评估上,干预后的引出率在腕关节机器人组为56.25%,在虚拟现实组为18.75%。结论:与BCI-虚拟现实反馈训练相比,BCI-腕关节机器人训练或能更好地改善慢性期脑卒中患者上肢运动功能,这或许是由于整合了感觉-运动信息的反馈方式对上肢运动功能改善的效果更为明显,本研究的结果或有助于优化基于运动想象的BCI训练系统的反馈策略。

Objective:To preliminarily compare the effectiveness of different end-effectors on upper limb motor function improvement in chronic stroke patients undergoing motor imagery-based brain-computer interface(BCI)rehabilitation.Method:Thirty-two chronic stroke patients who received motor imagery-based BCI rehabilitation were included.Based on the BCI training received,patients were divided into two groups:BCI combined with virtual reality feedback and BCI combined with wrist joint robot feedback.Assessment of upper limb motor function(upper limb Fugl-Meyer assessment),active range of motion of the paralyzed wrist,surface electromyography reflecting active contraction of wrist extensor muscles,and motor-evoked potentials(MEP)were conducted be fore and after intervention.Thus,to compare the impact of virtual reality feedback and wrist joint robot feedback on upper limb motor function recovery in chronic stroke patients.Result:Prior to intervention,baseline results were comparable between the two groups(P=0.160).After intervention,significant improvements in upper limb motor function were observed in both groups(P<0.001).The wrist joint robot feedback group showed a trend significant improvement in Fugl-Meyer scores and the changes in Fugl-Meyer scores compared to the virtual reality feedback group(P=0.06).Regarding active range of motion of the paralyzed wrist,both groups initially lacked active flexion-extension movements,and after intervention,the improvement was 33.13°±24.96°in the wrist joint robot group and 27.81°±37.17°in the virtual reality group,with no significant difference between the two groups(P>0.638).In the assessment of surface electromyography for wrist extensor muscle group on the paralyzed side,both groups showed significant improvement in average amplitude and area under the curve after intervention(P<0.001),with the wrist joint robot group demonstrating a more significant improvement in average amplitude(P<0.001),while area under curve showed a trend of improvement.In the MEP evaluation,the elicitation rate after intervention was 56.25%in the wrist joint robot group and 18.75%in the virtual reality group.Conclusion:Motor imagery-based BCI rehabilitation systems effectively improve upper limb motor function in chronic stroke patients.Wrist joint robot feedback appears to be more significant in improvement compared to virtual reality feedback.Integrating sensory and motor information might better affect upper limb motor recovery.These might help to optimize the feedback strategy of BCI training.