Optimal selection of independent components for event-related brain electrical potential enhancement

Author

Xiaorong, Gao ; Neng, Xu ; Bo, Hong ; Gao Shanghai ; Fusheng, Yang

Author_Institution

Dept. of Biomed. Eng., Tsinghua Univ., Beijing, China

fYear

2004

fDate

1-3 Dec. 2004

Abstract

A method based on independent component analysis (ICA) is proposed for the enhancement of multi-channel event-related evoked potential (EP). P300-related independent components are selected according to the a priori knowledge of P300 spatio-temporal pattern, and clear P300 peak is reconstructed by back projection of ICA. It is used for P300 detection in P3 speller brain computer interface and for P100 detection in clinical applications. Encouraging performance is obtained by the method. The number of trials needed is reduced, but the estimated EP signal is clearer than that obtained by conventional ensemble averaging. Applied to the dataset IIb of BCI Competition 2003, the algorithm achieved an accuracy of 100% in P300 detection within 5 repetitions.

Keywords

bioelectric potentials; brain; handicapped aids; independent component analysis; medical signal processing; signal reconstruction; spatiotemporal phenomena; P3 speller brain computer interface; back projection; ensemble averaging; independent component analysis; multi-channel event-related brain electrical potential enhancement; signal reconstruction; Biomedical engineering; Brain computer interfaces; Delay; Electric potential; Electrocardiography; Electroencephalography; Independent component analysis; Medical diagnostic imaging; Signal processing; Signal to noise ratio;

fLanguage

English

Publisher

ieee

Conference_Titel

Biomedical Circuits and Systems, 2004 IEEE International Workshop on

Print_ISBN

0-7803-8665-5

Type

conf

DOI

10.1109/BIOCAS.2004.1454152

Filename

1454152