Efficient feature selection and linear discrimination of EEG signals

• Published in: Neurocomputing (Amsterdam) Vol. 115; pp. 161 - 165
• Main Authors: Rodríguez-Bermúdez, Germán, García-Laencina, Pedro J., Roca-González, Joaquín, Roca-Dorda, Joaquín
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 04.09.2013; Elsevier
• Subjects: Applied sciences; Autoregressive (AR) modelling; Biological and medical sciences; Brain–Computer Interface (BCI); Cognition. Intelligence; Computer science; control theory; systems; Computer systems and distributed systems. User interface; Continuous Wavelet Transform (CWT); Data processing. List processing. Character string processing; Electrodiagnosis. Electric activity recording; Electroencephalogram (EEG); Exact sciences and technology; Fisher Linear Discriminant (FLD); Fundamental and applied biological sciences. Psychology; Hjorth parameters; Investigative techniques, diagnostic techniques (general aspects); Least Angle Regression (LAR); Leave-One-Out (LOO); Medical sciences; Memory organisation. Data processing; Mental imagery. Mental representation; Nervous system; Power Spectral Density (PSD); PRESS statistic; Psychology. Psychoanalysis. Psychiatry; Psychology. Psychophysiology; Software; Leave-One-Out (LOO); PRESS statistic; Fisher Linear Discriminant (FLD); Continuous Wavelet Transform (CWT); Hjorth parameters; Electroencephalogram (EEG); Brain–Computer Interface (BCI); Power Spectral Density (PSD); Autoregressive (AR) modelling; Least Angle Regression (LAR); Brain; Spectral energy distribution; Central nervous system; Activity; Electroencephalography; Autoregressive model; Cross validation; Computer control; Modeling; Encephalon; User interface; Classification; Selection criterion; Pattern extraction; Discriminant analysis; Spectral method; Statistical analysis; Probabilistic approach; Brain―Computer Interface (BCI); Pattern recognition; Fisher information; Experimental result; Discrimination; Wavelet transformation; Electroencephalogram (EEC); Signal processing; Power spectrum
• ISSN: 0925-2312; 1872-8286
• DOI: 10.1016/j.neucom.2013.01.001

Brain–Computer Interface systems (BCIs) based on Electroencephalogram (EEG) signal processing allow us to translate the subject's brain activities into control commands for computer devices. This paper presents an efficient embedded approach for feature selection and linear discrimination of EEG signals. In the first stage, four well-known feature extraction methods are used: Power spectral features, Hjorth parameters, Autoregressive modelling and Wavelet transform. From all the obtained features, the proposed method efficiently selects and combines the most useful features for classification with less computational requirements. Least Angle Regression (LARS) is used for properly ranking each feature and, then, an efficient Leave-One-Out (LOO) estimation based on the PRESS statistic is used to choose the most relevant features. Experimental results on motor-imagery BCIs problems are provided to illustrate the competitive performance of the proposed approach against other conventional methods.