ICA去除EEG中眼动伪差和工频干扰方法研究万柏坤.pdfVIP

10 Vol. 31 No. 10 2003 10 ACTA ELECTRONICA SINICA Oct. 2003 ICA EEG 万柏坤, 朱 欣, 杨春梅, 高 扬 ( , 300072) : ( EEG) , . ( Independent Component Analysis, ICA) EEG . ( Alzheimer disease, AD) EEG ( , 0dB) ICA , ( In omax ) ( Extended In omax) ICA , , , ; ICA , ; ( approximate entropy, ApEn) ICA EEG . ICA , . : ( EEG) ; ; ; ( ICA) ; ( In omax) : TH772+ 12 : A : 0372- 2112 ( 2003) 10- 1571- 04 Study on Applying Independent Component Analysis to Remove Blink Artifacts and Power Noise in EEG WAN Ba-i kun, ZHU Xin , YANG Chun-mei, GAO- Yang ( Dep artment of Biomedical Engineering , Ti anj in Universi ty , Tianj in 300072, China ) Abstract Blink arti acts and pow er noise are constantly ound: to strongly iu luence the acquisition and analysis o EEG sig- nals . In this paper, by comparing the e iciencies o tw o ICA algorithms ) In omax- ICA and Extended- In omax- ICA methods in extract- ing blink arti acts and pow er noise in the EEG signals, it w as shown that ICA algorithms were insensitive to disturbance in the condi- tions o low signa-l noise- ratio, and ICA algorithms demonstrated a strong robustness in processing non-stationary signals. Though blink slow w aves could be extracted by in omax algorithm , but power noise w as unlikely to be removed by it . There ore, Extended- In omax ICA algorithm should be used. By applying Extended- In omax algorithms, blink arti acts and pow er noise contained in the 16- channel EEG signals o Alzheimer- disease patients were removed su