基于ICA算法的事件相关电位的消噪与提取

Denoising and Extraction of ERP Based on ICA Algorithm

文章介绍了独立分量分析(IndependentComponentAnalysis,ICA)的基本理论及扩展信息最大(ExtendedInfo-max)ICA算法,并将该算法应用于多道单次事件相关电位(Single-trialEventRelatedPotential,SERP)的消噪中。实验结果表明,采用该算法成功地消除了SERP中的眼电(Electrooculographic,EOG)、肌电、α波、μ波等噪声,随后结合平均技术,实现了微弱的多道事件相关电位(EventRelatedPotential,ERP)在强噪声中的有效提取.

In this paper,the basic theory and algorithm of ICA are introduced,and then the extended imfomax ICA al-gorithm is applied to the denoising of multichannel Single-trial event related potential data.The experimental results show that ICA can successfully remove the artifacts such as EOG,muscle artifacts,αwaves,μwaves,subsequently,the artifact-free weak ERP signals can efficiently be extracted from strong artifacts activities with averaging method.