焦学军, 姜劲, 潘津津, 王春慧, 张朕, 杨涵钧, 徐凤刚
AuthorsHTML:焦学军, 姜劲, 潘津津, 王春慧, 张朕, 杨涵钧, 徐凤刚
AuthorsListE:Jiao Xuejun, Jiang Jin, Pan Jinjin, Wang Chunhui, Zhang Zhen, Yang Hanjun, Xu Fenggang
AuthorsHTMLE:Jiao Xuejun, Jiang Jin, Pan Jinjin, Wang Chunhui, Zhang Zhen, Yang Hanjun, Xu Fenggang
Unit:中国航天员科研训练中心人因工程国防科技重点实验室，北京100094
Unit_EngLish:National Key Laboratory of Human Factors Engineering, China Astronaut Research and Training Center, Beijing 100094, China
Abstract_Chinese:功能性近红外光谱技术(functional near-infrared spectroscopy, fNIRS)作为一种无损光学脑成像技术, 可用于构建脑机接口(brain-computer interface, BCI)以识别操作者肢体控制意图．利用fNIRS技术测量11位参试者的手臂伸展、腿部伸展和手指敲击的前额皮层(prefrontal cortex, PFC)和运动功能皮层(motor cortex, MC)的血氧变化信号, 并利用fNIRS的生理特征和支持向量机建立fNIRS-BCI分类器．结果表明手臂伸展、手指敲击和腿部伸展的四分类fNIRS-BCI平均正确率分别为89.32% 、88.66% 和91.35% ; fNIRS-BCI的运动想象动作的分类正确率不低于运动执行动作; 3种任务范式的混淆矩阵分析结果表明：运动想象诱发的脑功能活动与同侧肢体的运动执行、对侧肢体的运动想象活动产生混淆, 3种任务范式的同侧运动想象和运动执行的血氧数据检验结果存在显著差异．因此, fNIRS-BCI能有效识别运动想象和运动执行活动, 且运动想象和运动执行活动的血氧数据变化具有可分性．
Abstract_English:Functional near-infrared spectroscopy(fNIRS)is a non-invasive optical brain-imaging technology which could be applied in brain-computer interface(BCI)to recognize motor intention in brain. The study used fNIRS to measure hemodynamic variation in prefrontal cortex(PFC)and motor cortex(MC)while 11 participants were completing three different movement tasks，that is，hand movement，leg movement and finger tapping. In order to discriminate between motor execution(ME)and motor imagery(MI)，three fNIRS-BCI classifiers were established using fNIRS feature and support vector machine. The result shows that three four-class fNIRS-BCI classifiers，corresponding to hand movement，finger tapping and leg movement，achieve average classification accuracy of 89.32% ，88.66% and 91.35% respectively. The classification accuracy of MI by fNIRS-BCI is not lower than that of ME. The analysis of confusion matrix for three experiment paradigms shows that cerebration induced by MI confuses with ME of ipsilateral limb and MI of offside limb. The concentration of hemodynamic changes of MI and ME for the same limb in the three paradigms shows significant difference. Hence，the present study verifies the feasibility of fNIRS-BCI to differentiate between MI and ME in brain，and hemodynamic changes in response to MI and ME are separable.
Keyword_Chinese:功能性近红外光谱技术; 脑机接口; 运动想象; 运动执行
Keywords_English:functional near-infrared spectroscopy; brain-computer interface(BCI); motor imagery; motor execution

PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=5830