基于小波分析方法的脑电诱发电位单导少次提取  

A Wavelet Analysis Method for Single Channel Evoked Potential Extraction with a Few Sweeps

作　　者：王永轩[1] 邱天爽[1] 刘蓉[1] 

机构地区：[1]大连理工大学电子信息与电气工程学部,大连116024

出　　处：《中国生物医学工程学报》2011年第1期34-39,共6页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金(30570475;60872122;60940023)

摘　　要：脑电诱发电位(EP)的单导少次提取具有重要的临床意义。本研究提出一种基于小波分析的方法,用于EP信号的单导少次提取。该算法首先将检测得到的信号经过白化滤波器进行滤波,将自发脑电预白化为白噪声,目的是使噪声的小波系数均匀分布在所有尺度和位移上,而期望信号的小波系数只在少数尺度上为非零值。进而对滤波后的信号的小波系数进行加权处理,再通过反变换得到期望信号,恢复出EP信号。在不同初始信噪比条件下仿真,进行算法测试和性能分析。结果表明,该算法能够抑制自发脑电的干扰,使信噪比提高13 dB,相当于20次叠加平均效果,从而可以实现EP信号的单导快速提取。The signal estimation of single channel brain evoked potential（EP） with a few sweeps is important in clinical applications.A wavelet analysis based method was proposed in this paper to estimate EP signals from the observable noisy ones.The underlying principle is to filter the noisy signal and turn the spontaneous EEG into white noise.Then,through wavelet transform,the coefficients of the noise were distributed in all scales and shifts while those of the expected EP signal concentrated only in several scales.The expected signal was estimated by weighting the coefficients and inverse transformation.EP signal was recovered at last.Simulation experiments were carried out to test the algorithm under different SNRs.The results demonstrated that the interference from spontaneous EEG was eliminated effectively with increased SNR up to 13 dB,equal to 20 times ensemble averaging.

关 键 词：诱发电位 预白化 正交小波变换 小波阈值 

分 类 号：R318[医药卫生—生物医学工程]