基于联合近似对角化算法的脑电图中眼电伪迹消除  

Removal of Ocular Artifact from EEG Based on JADE

作　　者：唐艳[1] 汤井田[1] 

机构地区：[1]中南大学信息物理学院生物医学研究所,长沙410083

出　　处：《生物医学工程学杂志》2008年第3期520-523,共4页Journal of Biomedical Engineering

基　　金：湖南省自然科学基金资助项目(07JJ6045)

摘　　要：利用联合近似对角化(JADE)算法对脑电图中眼电伪迹成分进行剔除。针对JADE算法能够同时分离超高斯和亚高斯信号的特点,将脑电图信号分解成独立分量,利用伪迹脑地形图的特征,将伪迹分量分离,得到不含伪迹的脑电图信号。实验结果表明,该算法具有较强的稳健性和实用性。The method of Joint approximate diagonalization (JADE) is used to remove the ocular artifact from Electroencephalogram(EEG). As the JADE algorithm can be used to separate sub-gaussian and super-gaussian sources, it is possible to extract the signals that are statistically independent, and to identify the spatial map that is associated with the artifactual component. Thus the method can remove artifactual component. Satisfactory results illustrate the robust and practicability of the algorithm.

关 键 词：盲源信号分离 联合近似对角化 脑电图 

分 类 号：R318[医药卫生—生物医学工程]