运动意识脑电的动态独立分量分析  

Dynamic Independent Component Analysis and its Application to EEG of Motor Imagination

作　　者：吴小培[1] 叶中付[2] 郭晓静[1] 张道信[1] 唐希雯[1] 

机构地区：[1]安徽大学教育部计算智能与信号处理实验室,合肥230039 [2]中国科学技术大学信号统计处理研究室,合肥230027

出　　处：《中国生物医学工程学报》2007年第6期818-824,共7页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(60271024);(60771033);安徽省自然科学基金资助(070412038);安徽省人才基金(2004Z028)。

摘　　要：研究了用独立分量分析方法进行运动意识脑电信号特征分析的可行性。提出了用峭度极大动态独立分量分析方法进行μ节律提取的新思想。通过对批处理ICA算法和动态ICA算法在运动意识脑电特征分析的结果比较,得出了动态ICA算法更适合于运动意识脑电特征分析和提取。研究中发现,动态ICA混合矩阵系数的时间波形能准确即时地反映受试者进行左右手运动想象时运动神经皮层的μ节律变化,这一结果对脑认知和脑—机接口研究具有较大的实际意义,为独立分量分析方法在事件相关电位(ERP)特征提取中的应用提供了新的思路。This paper investigated the possibility of using independent component analysis to get the pattern of EEG data while the subjects performing the motor imagination tasks.Considering the non-stationary characteristics of the motor imagery EEG,the dynamic independent component analysis（DICA） based on kurtosis maximization was proposed to detect the dynamic changes of μ rhythm during left and right hand movement imagination.The experimental results showed that the DICA separated the μ rhythm from EEG to one output channel,but the batch ICA fails to do that.The study in this paper also showed that the elements of dynamic mixing matrix are more sensitive to μ rhythm changes than the energy and kurtosis do,which means the dynamic mixing matrix of DICA can be used as the new features to discriminate the motor tasks and has the value for the study of brain cognition and brain—computer interface.The results also demonstrated that the DICA may be a promising tool for the analysis of motor imagery EEG and other event related potentials（ERP）.

关 键 词：运动想象 脑电信号 动态独立分量分析 脑—机接口 

分 类 号：R318[医药卫生—生物医学工程]