Subject-specific time-frequency selection for multi-class motor imagery-based BCIs using few Laplacian EEG channels

• Published in: Biomedical signal processing and control Vol. 38; pp. 302 - 311
• Main Authors: Yang, Yuan, Chevallier, Sylvain, Wiart, Joe, Bloch, Isabelle
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2017; Elsevier
• Subjects: Bioengineering; Brain–computer interfaces; Computer Science; FDA-type F-score; Human-Computer Interaction; Life Sciences; Machine Learning; Motor imagery; Multi-class classification; Signal and Image Processing; Time-frequency selection; Multi-class classification; Motor imagery; FDA-type F-score; Brain–computer interfaces; Time-frequency selection
• ISSN: 1746-8094; 1746-8108
• DOI: 10.1016/j.bspc.2017.06.016

•A novel method is proposed to select optimal time-frequency areas for feature extraction.•We extend the application of FDA-type F-score to multi-class cases using only three EEG channels.•Our method yields better results than the state-of-the-art methods using fewer EEG channels.•The performance of our method is stable over subjects and robust to artifacts. The essential task of a motor imagery brain–computer interface (BCI) is to extract the motor imagery-related features from electroencephalogram (EEG) signals for classifying motor intentions. However, the optimal frequency band and time segment for extracting such features differ from subject to subject. In this work, we aim to improve the multi-class classification and to reduce the required EEG channel in motor imagery-based BCI by subject-specific time-frequency selection. Our method is based on a criterion namely Fisher discriminant analysis-type F-score to simultaneously select the optimal frequency band and time segment for multi-class classification. The proposed method uses only few Laplacian EEG channels (C3, Cz and C4) located around the sensorimotor area for classification. Applied to a standard multi-class BCI dataset (BCI competition III dataset IIIa), our method leads to better classification performance and smaller standard deviation across subjects compared to the state-of-art methods. Moreover, adding artifacts contaminated trials to the training dataset does not necessarily deteriorate our classification results, indicating that our method is tolerant to artifacts.