Creating a nonparametric brain-computer interface with neural time-series prediction preprocessing

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

The issue of subject-specific parameter selection in an electroencephalogram (EEG)-based brain-computer interface (BCI) is tackled in this paper. Hjorth- and Barlow-based feature extraction procedures (FEPs) are investigated along with linear discriminant analysis (LDA) for classification. These are well-known nonparametric FEPs but their simplicity prevents them from matching the performance of more complex FEPs. Neural time-series prediction preprocessing (NTSPP), which has been shown to enhance the separability of both time- and frequency-based features, is used to improve the applicability of these FEPs. NTSPP uses a number of prediction modules (PMs) to perform m-step ahead prediction of EEG time-series recorded whilst subjects perform left/right motor imagery-based mental tasks. Depending on the PMs, the NTSPP framework normally requires subject-specific parameters to be predefined. In this work each PM is a self-organizing fuzzy neural network (SOFNN). The SOFNN has a self-organizing structure and augmented nonlinear approximation capabilities however, a number of parameters must be defined prior to training. This is problematic therefore the practicality of a general set of parameters, previously selected via a sensitivity analysis (SA), is analyzed. The results indicate that a general set of NTSPP parameters may provide the best results and therefore a fully nonparametric BCI may be realizable.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages2183-2186
Number of pages4
Publication statusPublished - Sep 2006
Eventthe 28th International IEEE Engineering in Medicine and Biology Conference - New York, USA
Duration: 1 Sep 2006 → …

Conference

Conferencethe 28th International IEEE Engineering in Medicine and Biology Conference
Period1/09/06 → …

Fingerprint

Brain computer interface
Time series
Feature extraction
Fuzzy neural networks
Electroencephalography
Discriminant analysis
Sensitivity analysis

Cite this

@inproceedings{506a0dc0ae074962b23d540a4be97811,
title = "Creating a nonparametric brain-computer interface with neural time-series prediction preprocessing",
abstract = "The issue of subject-specific parameter selection in an electroencephalogram (EEG)-based brain-computer interface (BCI) is tackled in this paper. Hjorth- and Barlow-based feature extraction procedures (FEPs) are investigated along with linear discriminant analysis (LDA) for classification. These are well-known nonparametric FEPs but their simplicity prevents them from matching the performance of more complex FEPs. Neural time-series prediction preprocessing (NTSPP), which has been shown to enhance the separability of both time- and frequency-based features, is used to improve the applicability of these FEPs. NTSPP uses a number of prediction modules (PMs) to perform m-step ahead prediction of EEG time-series recorded whilst subjects perform left/right motor imagery-based mental tasks. Depending on the PMs, the NTSPP framework normally requires subject-specific parameters to be predefined. In this work each PM is a self-organizing fuzzy neural network (SOFNN). The SOFNN has a self-organizing structure and augmented nonlinear approximation capabilities however, a number of parameters must be defined prior to training. This is problematic therefore the practicality of a general set of parameters, previously selected via a sensitivity analysis (SA), is analyzed. The results indicate that a general set of NTSPP parameters may provide the best results and therefore a fully nonparametric BCI may be realizable.",
author = "DH Coyle and G Prasad and TM McGinnity",
year = "2006",
month = "9",
language = "English",
pages = "2183--2186",
booktitle = "Unknown Host Publication",

}

Coyle, DH, Prasad, G & McGinnity, TM 2006, Creating a nonparametric brain-computer interface with neural time-series prediction preprocessing. in Unknown Host Publication. pp. 2183-2186, the 28th International IEEE Engineering in Medicine and Biology Conference, 1/09/06.

Creating a nonparametric brain-computer interface with neural time-series prediction preprocessing. / Coyle, DH; Prasad, G; McGinnity, TM.

Unknown Host Publication. 2006. p. 2183-2186.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Creating a nonparametric brain-computer interface with neural time-series prediction preprocessing

AU - Coyle, DH

AU - Prasad, G

AU - McGinnity, TM

PY - 2006/9

Y1 - 2006/9

N2 - The issue of subject-specific parameter selection in an electroencephalogram (EEG)-based brain-computer interface (BCI) is tackled in this paper. Hjorth- and Barlow-based feature extraction procedures (FEPs) are investigated along with linear discriminant analysis (LDA) for classification. These are well-known nonparametric FEPs but their simplicity prevents them from matching the performance of more complex FEPs. Neural time-series prediction preprocessing (NTSPP), which has been shown to enhance the separability of both time- and frequency-based features, is used to improve the applicability of these FEPs. NTSPP uses a number of prediction modules (PMs) to perform m-step ahead prediction of EEG time-series recorded whilst subjects perform left/right motor imagery-based mental tasks. Depending on the PMs, the NTSPP framework normally requires subject-specific parameters to be predefined. In this work each PM is a self-organizing fuzzy neural network (SOFNN). The SOFNN has a self-organizing structure and augmented nonlinear approximation capabilities however, a number of parameters must be defined prior to training. This is problematic therefore the practicality of a general set of parameters, previously selected via a sensitivity analysis (SA), is analyzed. The results indicate that a general set of NTSPP parameters may provide the best results and therefore a fully nonparametric BCI may be realizable.

AB - The issue of subject-specific parameter selection in an electroencephalogram (EEG)-based brain-computer interface (BCI) is tackled in this paper. Hjorth- and Barlow-based feature extraction procedures (FEPs) are investigated along with linear discriminant analysis (LDA) for classification. These are well-known nonparametric FEPs but their simplicity prevents them from matching the performance of more complex FEPs. Neural time-series prediction preprocessing (NTSPP), which has been shown to enhance the separability of both time- and frequency-based features, is used to improve the applicability of these FEPs. NTSPP uses a number of prediction modules (PMs) to perform m-step ahead prediction of EEG time-series recorded whilst subjects perform left/right motor imagery-based mental tasks. Depending on the PMs, the NTSPP framework normally requires subject-specific parameters to be predefined. In this work each PM is a self-organizing fuzzy neural network (SOFNN). The SOFNN has a self-organizing structure and augmented nonlinear approximation capabilities however, a number of parameters must be defined prior to training. This is problematic therefore the practicality of a general set of parameters, previously selected via a sensitivity analysis (SA), is analyzed. The results indicate that a general set of NTSPP parameters may provide the best results and therefore a fully nonparametric BCI may be realizable.

M3 - Conference contribution

SP - 2183

EP - 2186

BT - Unknown Host Publication

ER -