Connectivity based feature-level filtering for single-trial EEG BCIs

Heger, D.; Terziyska, E.; Schultz, T.

doi:10.1109/ICASSP.2014.6853962

by D. Heger, E. Terziyska, T. Schultz

Abstract:

EEG-based Brain Computer interfaces (BCIs) often rely on power spectral density features to represent relevant aspects of brain activity. The information flow within human brain networks and the corresponding connectivity patterns may contain useful information to improve BCI performance, however they are typically not leveraged in current systems. In this paper, analyzes of information flow between independent sources of brain activity have been incorporated into the feature extraction stage of a BCI. For this purpose, connectivity measures based on multivariate autoregressive models have been estimated and are applied as filters to power spectral density based features. Two publicly available data sets have been used to evaluate the proposed feature extraction method: a two-back task and a motor imagery task. The results demonstrate significant performance improvements of the proposed method over band-power features and indicate that connectivity in brain networks can be used as powerful feature-level filters for BCIs.

Reference:

Connectivity based feature-level filtering for single-trial EEG BCIs (D. Heger, E. Terziyska, T. Schultz), In Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on, 2014.

Bibtex Entry:

@INPROCEEDINGS{6853962,
author={Heger, D. and Terziyska, E. and Schultz, T.},
booktitle={Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on},
title={Connectivity based feature-level filtering for single-trial EEG BCIs},
year={2014},
pages={2064-2068},
abstract={EEG-based Brain Computer interfaces (BCIs) often rely on power spectral density features to represent relevant aspects of brain activity. The information flow within human brain networks and the corresponding connectivity patterns may contain useful information to improve BCI performance, however they are typically not leveraged in current systems. In this paper, analyzes of information flow between independent sources of brain activity have been incorporated into the feature extraction stage of a BCI. For this purpose, connectivity measures based on multivariate autoregressive models have been estimated and are applied as filters to power spectral density based features. Two publicly available data sets have been used to evaluate the proposed feature extraction method: a two-back task and a motor imagery task. The results demonstrate significant performance improvements of the proposed method over band-power features and indicate that connectivity in brain networks can be used as powerful feature-level filters for BCIs.},
keywords={autoregressive processes;brain;brain-computer interfaces;electroencephalography;feature extraction;medical signal processing;EEG-based brain computer interfaces;band-power feature;brain activity;connectivity measure;connectivity pattern;electroencephalography;feature extraction method;feature-level filtering;human brain network;information flow;motor imagery task;multivariate autoregressive model;power spectral density feature;publicly available data set;single-trial EEG BCIs;two-back task;Brain models;Electroencephalography;Feature extraction;Time series analysis;Transfer functions;Connectivity;Granger causality;brain-computer interfaces;direct directed transfer function;electroencephalography},
doi={10.1109/ICASSP.2014.6853962},

month={May},}