A Transform-Based Feature Extraction Approach for Motor Imagery Tasks Classification

Hamza Baali, Aida Khorshidtalab, Mostefa Mesbah, Momoh J E Salami

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

In this paper, we present a new motor imagery classification method in the context of electroencephalography (EEG)-based brain-computer interface (BCI). This method uses a signal-dependent orthogonal transform, referred to as linear prediction singular value decomposition (LP-SVD), for feature extraction. The transform defines the mapping as the left singular vectors of the LP coefficient filter impulse response matrix. Using a logistic tree-based model classifier; the extracted features are classified into one of four motor imagery movements. The proposed approach was first benchmarked against two related state-of-the-art feature extraction approaches, namely, discrete cosine transform (DCT) and adaptive autoregressive (AAR)-based methods. By achieving an accuracy of 67.35%, the LP-SVD approach outperformed the other approaches by large margins (25% compared with DCT and 6 % compared with AAR-based methods). To further improve the discriminatory capability of the extracted features and reduce the computational complexity, we enlarged the extracted feature subset by incorporating two extra features, namely, Q- and the Hotelling's T2 statistics of the transformed EEG and introduced a new EEG channel selection method. The performance of the EEG classification based on the expanded feature set and channel selection method was compared with that of a number of the state-of-the-art classification methods previously reported with the BCI IIIa competition data set. Our method came second with an average accuracy of 81.38%.

Original languageEnglish
Article number2100108
JournalIEEE Journal of Translational Engineering in Health and Medicine
Volume3
DOIs
Publication statusPublished - 2015

Fingerprint

Imagery (Psychotherapy)
Electroencephalography
Feature extraction
Mathematical transformations
Brain computer interface
Discrete cosine transforms
Singular value decomposition
Brain-Computer Interfaces
Impulse response
Logistics
Computational complexity
Classifiers
Statistics

Keywords

  • Brain-computer interface
  • channel selection
  • feature extraction
  • linear prediction
  • orthogonal transform

ASJC Scopus subject areas

  • Biomedical Engineering
  • Medicine(all)

Cite this

A Transform-Based Feature Extraction Approach for Motor Imagery Tasks Classification. / Baali, Hamza; Khorshidtalab, Aida; Mesbah, Mostefa; Salami, Momoh J E.

In: IEEE Journal of Translational Engineering in Health and Medicine, Vol. 3, 2100108, 2015.

Research output: Contribution to journalArticle

@article{70241ea8ff4240329dd0704474068bc6,
title = "A Transform-Based Feature Extraction Approach for Motor Imagery Tasks Classification",
abstract = "In this paper, we present a new motor imagery classification method in the context of electroencephalography (EEG)-based brain-computer interface (BCI). This method uses a signal-dependent orthogonal transform, referred to as linear prediction singular value decomposition (LP-SVD), for feature extraction. The transform defines the mapping as the left singular vectors of the LP coefficient filter impulse response matrix. Using a logistic tree-based model classifier; the extracted features are classified into one of four motor imagery movements. The proposed approach was first benchmarked against two related state-of-the-art feature extraction approaches, namely, discrete cosine transform (DCT) and adaptive autoregressive (AAR)-based methods. By achieving an accuracy of 67.35{\%}, the LP-SVD approach outperformed the other approaches by large margins (25{\%} compared with DCT and 6 {\%} compared with AAR-based methods). To further improve the discriminatory capability of the extracted features and reduce the computational complexity, we enlarged the extracted feature subset by incorporating two extra features, namely, Q- and the Hotelling's T2 statistics of the transformed EEG and introduced a new EEG channel selection method. The performance of the EEG classification based on the expanded feature set and channel selection method was compared with that of a number of the state-of-the-art classification methods previously reported with the BCI IIIa competition data set. Our method came second with an average accuracy of 81.38{\%}.",
keywords = "Brain-computer interface, channel selection, feature extraction, linear prediction, orthogonal transform",
author = "Hamza Baali and Aida Khorshidtalab and Mostefa Mesbah and Salami, {Momoh J E}",
year = "2015",
doi = "10.1109/JTEHM.2015.2485261",
language = "English",
volume = "3",
journal = "IEEE Journal of Translational Engineering in Health and Medicine",
issn = "2168-2372",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - A Transform-Based Feature Extraction Approach for Motor Imagery Tasks Classification

AU - Baali, Hamza

AU - Khorshidtalab, Aida

AU - Mesbah, Mostefa

AU - Salami, Momoh J E

PY - 2015

Y1 - 2015

N2 - In this paper, we present a new motor imagery classification method in the context of electroencephalography (EEG)-based brain-computer interface (BCI). This method uses a signal-dependent orthogonal transform, referred to as linear prediction singular value decomposition (LP-SVD), for feature extraction. The transform defines the mapping as the left singular vectors of the LP coefficient filter impulse response matrix. Using a logistic tree-based model classifier; the extracted features are classified into one of four motor imagery movements. The proposed approach was first benchmarked against two related state-of-the-art feature extraction approaches, namely, discrete cosine transform (DCT) and adaptive autoregressive (AAR)-based methods. By achieving an accuracy of 67.35%, the LP-SVD approach outperformed the other approaches by large margins (25% compared with DCT and 6 % compared with AAR-based methods). To further improve the discriminatory capability of the extracted features and reduce the computational complexity, we enlarged the extracted feature subset by incorporating two extra features, namely, Q- and the Hotelling's T2 statistics of the transformed EEG and introduced a new EEG channel selection method. The performance of the EEG classification based on the expanded feature set and channel selection method was compared with that of a number of the state-of-the-art classification methods previously reported with the BCI IIIa competition data set. Our method came second with an average accuracy of 81.38%.

AB - In this paper, we present a new motor imagery classification method in the context of electroencephalography (EEG)-based brain-computer interface (BCI). This method uses a signal-dependent orthogonal transform, referred to as linear prediction singular value decomposition (LP-SVD), for feature extraction. The transform defines the mapping as the left singular vectors of the LP coefficient filter impulse response matrix. Using a logistic tree-based model classifier; the extracted features are classified into one of four motor imagery movements. The proposed approach was first benchmarked against two related state-of-the-art feature extraction approaches, namely, discrete cosine transform (DCT) and adaptive autoregressive (AAR)-based methods. By achieving an accuracy of 67.35%, the LP-SVD approach outperformed the other approaches by large margins (25% compared with DCT and 6 % compared with AAR-based methods). To further improve the discriminatory capability of the extracted features and reduce the computational complexity, we enlarged the extracted feature subset by incorporating two extra features, namely, Q- and the Hotelling's T2 statistics of the transformed EEG and introduced a new EEG channel selection method. The performance of the EEG classification based on the expanded feature set and channel selection method was compared with that of a number of the state-of-the-art classification methods previously reported with the BCI IIIa competition data set. Our method came second with an average accuracy of 81.38%.

KW - Brain-computer interface

KW - channel selection

KW - feature extraction

KW - linear prediction

KW - orthogonal transform

UR - http://www.scopus.com/inward/record.url?scp=84946888103&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946888103&partnerID=8YFLogxK

U2 - 10.1109/JTEHM.2015.2485261

DO - 10.1109/JTEHM.2015.2485261

M3 - Article

AN - SCOPUS:84946888103

VL - 3

JO - IEEE Journal of Translational Engineering in Health and Medicine

JF - IEEE Journal of Translational Engineering in Health and Medicine

SN - 2168-2372

M1 - 2100108

ER -