【摘要】 对结合独立分量分析(independent component analysis,简称ICA)和遗传算法(genetic algorithm,简称GA)的运动想象脑电(motor imagery electroencephalogram,简称MI-EEG)特征检测及其优化方法开展研究.设计了基于ICA的MI-EEG分类算法.在此基础上,针对不同受试个体,用GA算法对运动想象诱发的事件相关去同步(event-related desynchronization,简称ERD)频段进行优化选择,用以改善运动想象脑-机接口(brain-computer interface,简称BCI)系统的识别率.实验结果表明,基于ICA的GA算法特征优化方法具有较好的可靠性和实用性,可用于在线BCI的设计与实现.更多还原

【Abstract】 The paper focused on the research of feature detection and the optimization algorithm of motor imagery electroencephalogram(MI-EEG) by using independent component analysis(ICA)and genetic algorithm(GA),and a MI-EEG classification method based on ICA was proposed in this paper.For different individual,the method used GA to optimize selection of the optimal event-related desynchronization(ERD)band which was induced by motor imagery,and it improved the recognition rate of motor imagery brain-computer interface(BCI)system.Experimental results revealed that the feature optimization of GA method based on ICA,which not only had better reliability and practicability,but could be used for the design and implementation of online BCI.更多还原