|本期目录/Table of Contents|

[1]安嘉琳,张屾,王玲,等.功能近红外光谱与功能核磁共振运动想象脑机接口双模态比较研究*[J].生物医学工程研究,2018,03:265-270.
 AN Jialin,ZHANG Shen,WANG Ling,et al.Multimodal comparison of functional near-infrared spectroscopy and functional magnetic resonance imaging for motor imagery ?based brain machine interface[J].Journal of Biomedical Engineering Research,2018,03:265-270.
点击复制

功能近红外光谱与功能核磁共振运动想象脑机接口双模态比较研究*(PDF)

《生物医学工程研究》[ISSN:1006-6977/CN:61-1281/TN]

期数:
2018年03期
页码:
265-270
栏目:
出版日期:
2018-09-25

文章信息/Info

Title:
Multimodal comparison of functional near-infrared spectroscopy and functional magnetic resonance imaging for motor imagery ?based brain machine interface
文章编号:
1672-6278 (2018)03-0265-06
作者:
安嘉琳12张屾12王玲12田一竹12李德玉12△汪待发12△
1.北京航空航天大学生物与医学工程学院,北京 100083;2.生物医学工程高精尖创新中心,北京航空航天大学,北京 100083
Author(s):
AN Jialin12ZHANG Shen12WANG Ling12TIAN Yizhu12LI Deyu12WANG Daifa12
1.School of Biological Science and Medical Engineering, Beihang University, Beijing 100083,China;2.Advanced Innovation Center for Biomedical Engineering, Beihang University, Beijing 100083
关键词:
功能近红外光谱运动想象功能核磁共振共空间模式脑机接口
Keywords:
Functional near-infrared spectroscopy Motor imagery Functional magnetic resonance imaging Common spatial pattern Brain machine interface
分类号:
R318
DOI:
10.19529/j.cnki.1672-6278.2018.03.05
文献标识码:
A
摘要:
基于功能近红外光谱(fNIRS)的运动想象脑机接口(BCI)在中风患者神经康复训练发挥着不可或缺和日益重要的作用,但分类率较低。以功能核磁共振(fMRI)作为参考金标准,针对运动想象范式,本研究开展了fMRI-BCI与fNIRS-BCI的对照研究,以探究限制fNIRS-BCI分类率的主要因素。实验选取了12名被试,分别获取两种模态下左右手运动想象数据,并利用共空间模式算法提取空间特异性信息进行分类。从血液动力学响应来看,由于受到浅表血流信号干扰,fNIRS获取的运动想象在双侧运动区脑响应的对比度明显低于fMRI。同时,fMRI-BCI的分类准确率显著高于fNIRS-BCI,超过10个百分点,分别为80.3%±16.0%与67.3%±10.2%。这些结果表明,fNIRS及目前常用的拓扑式排布在获取脑皮层响应信息上存在不足,是运动想象fNIRS-BCI分类率较低的主要限制因素。
Abstract:
The brain machine interface (BCI), based on the functional near-infrared spectroscopy (fNIRS), plays an indispensable and increasingly important role in the training of neurological rehabilitation for stroke patients, while it has the limitation of low classification rate.With functional magnetic resonance (fMRI) as the reference gold standard, for motor imagery paradigm, a controlled study of fMRI-BCI and fNIRS-BCI was carried out to explore the main limit factors of the classification rate of fNIRS-BCI. In this experiment, motor imagery data, left and right hand movements in two modes,respectively, from 12 subjects, were obtained and spatial specific information were extracted with the co-spatial pattern algorithm, and then were classified. From the hemodynamic response, due to the interference of superficial blood flow signals, the brain activities response of motion imagery obtained by fNIRS was significantly lower than that of fMRI in the bilateral motor region. Also, the classification accuracy of fMRI-BCI was significantly higher than that of fNIRS-BCI, with more than 10 percentage points, 80.3%±16.0% and 67.3%±10.2%,respectively. These results indicate that fNIRS and the commonly used method, topological typesetting, are insufficient in obtaining the response information of cerebral cortex, which is the main limiting factor of the low classification rate of fNIRS-BCI.

参考文献/References

-

备注/Memo

备注/Memo:
(收稿日期:2018-03-08)国家自然科学基金资助项目(61675013,61761166007,31771071);国家重点研发计划项目(2016YFF0201002)。△通信作者Email:deyuli@buaa.edu.cn;daifa.wang@buaa.edu.cn
更新日期/Last Update: 2018-09-29