Title :
Extraction of single-trial evoked potentials with Extended Infomax ICA algorithm and its applications to BCI systems
Author :
Yuan Chengqiang ; Zhang Jianhua
Author_Institution :
Dept. of Autom., East China Univ. of Sci. & Technol., Shanghai, China
Abstract :
Usually evoked potentials were submerged in strong background noise, so we need special methods to extract evoked potentials. In this paper, the method to extract P300 (a kind of evoked potentials) is Extended Infomax ICA. Also this method is applied to character recognition of BCI system based on P300. In order to improve the rate of recognition, the positive peak area, the latency and the peak value of P300 are all taken into account. The result shows that the method works very well.
Keywords :
bioelectric potentials; brain-computer interfaces; character recognition; electroencephalography; feature extraction; independent component analysis; medical signal processing; BCI systems; EEG; P300 extraction; P300 latency; P300 peak value; P300 positive peak area; character recognition; extended Infomax ICA algorithm; independent component analysis; single-trial evoked potential extraction; Abstracts; Automation; Conferences; Educational institutions; Electroencephalography; Electronic mail; Independent component analysis; Extended Infomax ICA; Independent component analysis; P300; evoked potentials; single-trial extraction;
Conference_Titel :
Control Conference (CCC), 2014 33rd Chinese
Conference_Location :
Nanjing
DOI :
10.1109/ChiCC.2014.6896179
