一种基于手部精细运动分类的脑机接口方法研究  

Study on a brain computer interface based on decoding of precise hand movements

作　　者：李睿 刘宇琪 刘卫平 兀瑞 白端阳 杨东 张俊峰[3] 赵杰 LI Rui;LIU Yuqi;LIU Weiping;WU Rui;BAI Duanyang;YANG Dong;ZHANG Junfeng;ZHAO Jie(Department of Brain,Xi'an People's Hospital(Fourth Hospital of Xi an),Xi an 710000,China)

机构地区：[1]西安市人民医院(西安市第四医院)脑科病院,西安710000 [2]西安理工大学机械与精密仪器工程学院 [3]西安医学院

出　　处：《临床神经外科杂志》2024年第4期384-388,395,共6页Journal of Clinical Neurosurgery

基　　金：西安市人民医院(西安市第四医院)科研孵化基金立项项目[2022BSH01(BH-1)];陕西省教育厅专项科研计划项目(22JK0471)。

摘　　要：目的 探索一种基于6种手部精细运动解码的脑机接口(BCI)技术。方法 在分析了6种手部常见精细运动脑电图(EEG)产生机理与响应特性的基础上,设计了手部精细运动执行BCI范式,实现了一种基于卷积神经网络的运动相关EEG信号解码模型,并搭建了基于手部精细运动的BCI系统,对8例健康受试者、2例因病变累及顶叶而出现明显运动功能障碍患者的6种精细运动手势EEG信号进行了分类。结果 10例受试者在基于手部精细运动的BCI系统下,EEG信号的分类准确率为(79.20±6.05)%。结论 基于6种手部精细运动解码的BCI方法具有一定有效性和泛化能力。Objective To explore a motor-related brain computer interface(BCI) technology based on the decoding of six precise hand movements.Methods Based on the analysis of the mechanisms and response characteristics of six common types of fine motor electroencephalography(EEG) in the hand,a BCI paradigm for fine motor execution was designed.A motion related EEG signal decoding model based on convolutional neural networks was implemented,and a BCI system based on fine motor was built.Six types of fine motor gesture EEG signals from eight healthy subjects and two patients with significant motor dysfunction due to lesions involving the parietal lobe were classified.Results The classification accuracy of EEG signals in 10 subjects under the BCI system based on fine hand movements was(79.20±6.05)%.Conclusions The BCI method based on decoding six types of fine hand movements has certain effectiveness and generalization ability.

关 键 词：精细运动 脑机接口 脑电信号 解码方法 

分 类 号：R651[医药卫生—外科学]