独立分量分析结合小波去噪算法提取诱发电位信号的仿真实验  

Simulation experiment of extracting evoked potential based on the combination of independent component analysis and wavelet analysis

作　　者：邹凌[1,2,3] 王新光[1] 马正华[1] 杨长春[1] 

机构地区：[1]江苏工业学院信息学院,江苏省常州市213164 [2]北京师范大学认知神经科学与学习国家重点实验室,北京市100875 [3]北京师范大学应用实验心理北京市重点实验室,北京市100875

出　　处：《中国组织工程研究与临床康复》2009年第43期8503-8505,共3页Journal of Clinical Rehabilitative Tissue Engineering Research

基　　金：北京师范大学认知神经科学与学习国家重点实验室开放课题资助项目;北京师范大学应用实验心理北京市重点实验室开放研究基金题资助项目~~

摘　　要：采用结合独立分量分析和小波去噪算法的方法提取诱发电位信号。首先使用扩展信息最大的独立分量分析算法分析仿真的脑电信号,分离出诱发电位信号,自发脑电信号,肌电干扰与高斯噪声,然后使用小波阀值收缩算法滤除诱发电位信号中残留的一些高频噪声。仿真实验表明,基于独立分量分析的算法可以将混合在诱发电位信号中的干扰信号分离开来,而结合方法的提取结果在波形、相关系数指标等方面均优于单独使用独立分量分析算法的提取结果,这为实际临床应用中诱发电位的有效提取提供了思路。In this paper we present a method combined independent component analysis （ICA） algorithm and wavelet de-noising algorithm to extract the evoked potentials （EPs）. First, the extended Infomax-ICA algorithm was used to analyze the simulated electroencephalograph signals and separate the simulated EPs, spontaneous electroencephalograph, muscle artifacts and Gaussian noise. Then, the wave shrink method was applied to de-noise some high frequency noise let~ in EPs. The simulation results showed that ICA-based method can extract the weak EPs efficiently from strong background artifacts, and the combined method used for the EPs extraction has good results with wave and relative coefficient than ICA algorithm alone, which provides a good thought for extracting EPs in the clinical application.

关 键 词：扩展信息最大化独立分量分析 小波阀值收缩去噪 诱发电位信号 提取 

分 类 号：R318[医药卫生—生物医学工程]