A Time-Series Prediction Approach for Feature Extraction in a Brain-Computer Interface

D Coyle; TM McGinnity; G Prasad

doi:10.1109/TNSRE.2005.857690

A Time-Series Prediction Approach for Feature Extraction in a Brain-Computer Interface

Research output: Contribution to journal › Article › peer-review

Abstract

This paper presents a feature extraction procedure (FEP) for a brain-computer interface (BCI) application where features are extracted from the electroencephalogram (EEG) recorded from subjects performing right and left motor imagery. Two neural networks (NNs) are trained to perform one-step-ahead predictions for the EEG time-series data, where one NN is trained on right motor imagery and the other on left motor imagery. Features are derived from the power (mean squared) of the prediction error or the power of the predicted signals. All features are calculated from a window through which all predicted signals pass. Separability of features is achieved due to the morphological differences of the EEG signals and each NNs specialization to the type of data on which it is trained. Linear discriminant analysis (LDA) is used for classification. This FEP is tested on three subjects off-line and classification accuracy (CA) rates range between 88% and 98%. The approach compares favorably to a well-known adaptive autoregressive (AAR) FEP and also a linear AAR model based prediction approach.

Original language	English
Pages (from-to)	461-467
Journal	IEEE Transactions on Neural Systems and Rehabilitation Engineering
Volume	13
Issue number	4
DOIs	https://doi.org/10.1109/TNSRE.2005.857690
Publication status	Published (in print/issue) - 1 Dec 2005

Bibliographical note

This paper presents a time-series prediction based signal processing concept for a brain-computer interface (BCI) to enhance signal separability, and improve accuracy and information transfer rate. The work is the basis of a PhD thesis awarded the IEEE Computational Intelligence Society Outstanding Dissertation Award 2008 and is now extensively used as the basis for ongoing research to develop an advanced BCI system. Through the paper, the authors have joined an EU FP7 proposal consortium, Advanced Multimodal Technology for Tetraplegic-patient Enablement in Communication and Control in the Home, involving Philips Research, Shadow Robot Company, and the National Rehabilitation Hospital, Ireland.

Access to Document

10.1109/TNSRE.2005.857690

Cite this

@article{2a31f39e55c041b4848ccda41940b560,

title = "A Time-Series Prediction Approach for Feature Extraction in a Brain-Computer Interface",

abstract = "This paper presents a feature extraction procedure (FEP) for a brain-computer interface (BCI) application where features are extracted from the electroencephalogram (EEG) recorded from subjects performing right and left motor imagery. Two neural networks (NNs) are trained to perform one-step-ahead predictions for the EEG time-series data, where one NN is trained on right motor imagery and the other on left motor imagery. Features are derived from the power (mean squared) of the prediction error or the power of the predicted signals. All features are calculated from a window through which all predicted signals pass. Separability of features is achieved due to the morphological differences of the EEG signals and each NNs specialization to the type of data on which it is trained. Linear discriminant analysis (LDA) is used for classification. This FEP is tested on three subjects off-line and classification accuracy (CA) rates range between 88\% and 98\%. The approach compares favorably to a well-known adaptive autoregressive (AAR) FEP and also a linear AAR model based prediction approach. ",

author = "D Coyle and TM McGinnity and G Prasad",

note = "This paper presents a time-series prediction based signal processing concept for a brain-computer interface (BCI) to enhance signal separability, and improve accuracy and information transfer rate. The work is the basis of a PhD thesis awarded the IEEE Computational Intelligence Society Outstanding Dissertation Award 2008 and is now extensively used as the basis for ongoing research to develop an advanced BCI system. Through the paper, the authors have joined an EU FP7 proposal consortium, Advanced Multimodal Technology for Tetraplegic-patient Enablement in Communication and Control in the Home, involving Philips Research, Shadow Robot Company, and the National Rehabilitation Hospital, Ireland.",

year = "2005",

month = dec,

day = "1",

doi = "10.1109/TNSRE.2005.857690",

language = "English",

volume = "13",

pages = "461--467",

journal = "IEEE Transactions on Neural Systems and Rehabilitation Engineering",

publisher = "IEEE",

number = "4",

}

TY - JOUR

T1 - A Time-Series Prediction Approach for Feature Extraction in a Brain-Computer Interface

AU - Coyle, D

AU - McGinnity, TM

AU - Prasad, G

N1 - This paper presents a time-series prediction based signal processing concept for a brain-computer interface (BCI) to enhance signal separability, and improve accuracy and information transfer rate. The work is the basis of a PhD thesis awarded the IEEE Computational Intelligence Society Outstanding Dissertation Award 2008 and is now extensively used as the basis for ongoing research to develop an advanced BCI system. Through the paper, the authors have joined an EU FP7 proposal consortium, Advanced Multimodal Technology for Tetraplegic-patient Enablement in Communication and Control in the Home, involving Philips Research, Shadow Robot Company, and the National Rehabilitation Hospital, Ireland.

PY - 2005/12/1

Y1 - 2005/12/1

N2 - This paper presents a feature extraction procedure (FEP) for a brain-computer interface (BCI) application where features are extracted from the electroencephalogram (EEG) recorded from subjects performing right and left motor imagery. Two neural networks (NNs) are trained to perform one-step-ahead predictions for the EEG time-series data, where one NN is trained on right motor imagery and the other on left motor imagery. Features are derived from the power (mean squared) of the prediction error or the power of the predicted signals. All features are calculated from a window through which all predicted signals pass. Separability of features is achieved due to the morphological differences of the EEG signals and each NNs specialization to the type of data on which it is trained. Linear discriminant analysis (LDA) is used for classification. This FEP is tested on three subjects off-line and classification accuracy (CA) rates range between 88% and 98%. The approach compares favorably to a well-known adaptive autoregressive (AAR) FEP and also a linear AAR model based prediction approach.

AB - This paper presents a feature extraction procedure (FEP) for a brain-computer interface (BCI) application where features are extracted from the electroencephalogram (EEG) recorded from subjects performing right and left motor imagery. Two neural networks (NNs) are trained to perform one-step-ahead predictions for the EEG time-series data, where one NN is trained on right motor imagery and the other on left motor imagery. Features are derived from the power (mean squared) of the prediction error or the power of the predicted signals. All features are calculated from a window through which all predicted signals pass. Separability of features is achieved due to the morphological differences of the EEG signals and each NNs specialization to the type of data on which it is trained. Linear discriminant analysis (LDA) is used for classification. This FEP is tested on three subjects off-line and classification accuracy (CA) rates range between 88% and 98%. The approach compares favorably to a well-known adaptive autoregressive (AAR) FEP and also a linear AAR model based prediction approach.

U2 - 10.1109/TNSRE.2005.857690

DO - 10.1109/TNSRE.2005.857690

M3 - Article

VL - 13

SP - 461

EP - 467

JO - IEEE Transactions on Neural Systems and Rehabilitation Engineering

JF - IEEE Transactions on Neural Systems and Rehabilitation Engineering

IS - 4

ER -

A Time-Series Prediction Approach for Feature Extraction in a Brain-Computer Interface

Abstract

Bibliographical note

Access to Document

Fingerprint

Cite this