Title :
Feature Extraction and Subset Selection for Classifying Single-Trial ECoG during Motor Imagery
Author :
Wei, Qingguo ; Gao, Xiaorong ; Gao, Shangkai
Author_Institution :
Dept. of Electron. Eng., Nanchang Univ.
fDate :
Aug. 30 2006-Sept. 3 2006
Abstract :
The electrocorticogram (ECoG) recorded from subdural electrodes is a kind of BCI signal source that has the potential to achieve good classification results. The feature extraction and its subset selection are crucial for increasing classification accuracy rate. This paper proposes a new algorithm for classifying single-trial ECoG during motor imagery. The nonlinear regressive coefficients between signals on 10 leads are extracted in two frequency bands 0-3 Hz and 8-30 Hz as classification features. A genetic algorithm is used for the selection of the optimal feature subset and a support vector machine for their evaluation. The generalization error of 7% is achieved on data set I of BCI Competition III
Keywords :
bioelectric phenomena; biomedical electrodes; feature extraction; genetic algorithms; learning (artificial intelligence); medical signal processing; regression analysis; support vector machines; user interfaces; 0 to 3 Hz; 8 to 30 Hz; BCI signal; brain-computer interface; electrocorticogram; feature extraction; genetic algorithm; motor imagery; nonlinear regressive coefficient; single-trial ECoG classification; subdural electrodes; subset selection; support vector machine; Biomedical engineering; Data mining; Feature extraction; Fingers; Frequency; Genetic algorithms; Materials requirements planning; Support vector machine classification; Support vector machines; Tongue; brain-computer interface; electrocorticogram; genetic algorithm; nonlinear regression;
Conference_Titel :
Engineering in Medicine and Biology Society, 2006. EMBS '06. 28th Annual International Conference of the IEEE
Conference_Location :
New York, NY
Print_ISBN :
1-4244-0032-5
Electronic_ISBN :
1557-170X
DOI :
10.1109/IEMBS.2006.260561
