独立分量分析技术在EEG工频干扰分离中的应用

工频干扰是脑电图(EEG)中常见噪声，严重影响EEG-信号的提取和分析。通过比较Fastica、Extended Infomax、EGLD、Pearson—ICA等四种独立分量分析(ICA)算法和奇异值分解(SVD)技术用于分离EEG中工频干扰的效果，确证ICA方法有很好的抗干扰性，而常用的SVD技术则难以奏效；其中推广的最大熵(Extended Info—max)ICA算法有较好的收敛性，文中使用该算法成功地从16导联早老性痴呆症患者EEG信号中(含混入的工频干扰，最低信噪比约为0dB)分离出工频干扰。ICA在生物医学信号处理特别是临床医学工程中潜在着重要应用前景和研究价值。

Apply ICA (Independent Component Analysis) to Removing Power Noise from EEG

Power noise is constantly found in EEG signals; thus the acquisition and analysis of EEG signals can be strongly influenced. Comparison of the efficiencies of four ICA algorithms (Fastica,Extended Infomax,EGLD,Pearson ICA) and SVD methods in extracting power noise in the EEG signals showed that ICA algorithms appear insensitive to the noise disturbance, whereas the commonly used SVD method does not. By applying the Extended Infomax ICA with better convergence in this paper, it was demonstrated that the power noise contained in the 16 channel EEG signals of one Alzheimer disease patient were removed successfully(the lowest signal noise ratio for power noise is 0 dB). ICA has a possible important value and prospect in biomedical signal processing, especially in clinical medical engineering.