Coupling measures in BCIs

Summary

This project aims to analyze the suitability of novel features for brain-computer interfaces (BCIs). These features describe the relationships or coupling between single EEG signals. In contrast, conventional univariate features such as logarithmic band power totally ignore this information.
The main scientific goal in this project is to study coupling measures as a new feature type for EEG-based brain-computer interfaces (BCIs). Before the start of this project, only a few studies have experimented with BCIs based on coupling measures, and they almost exclusively used the phase-locking value (PLV) as an explicit measure for channel synchronization. The majority of existing BCI studies have relied on features that take only single EEG channels into account, such as logarithmic band power or univariate autoregressive (AR) parameters.
We address this issue by modeling multiple EEG channels simultaneously with a multivariate (or vector) autoregressive (VAR) model, which is a generalization of the univariate AR model. A VAR model incorporates couplings between channels, and many widely used explicit coupling measures (for example the partial directed coherence or the directed transfer function) can be derived from the model parameters. Such a vector AR model is at the basis of a more general framework, which allows analyzing information flow based on the concept of Granger causality. For these reasons, our main approach is based on using the VAR model coefficients directly as features for BCIs, rather than computing a derived explicit coupling measure. However, we will also study the feasibility of the latter approach, which involves using explicit connectivity measures as features.

Peer-reviewed publications

M. Billinger, G. R. Müller-Putz, C. Neuper, C. Brunner. A BCI framework based on single trial functional connectivity. In preparation, 2012.

M. Billinger, I. Daly, V. Kaiser, J. Jin, B. Z. Allison, G. R. Müller-Putz, C. Brunner. Is it significant? Guidelines for reporting BCI performance. In: B. Z. Allison, S. Dunne, R. Leeb, J. del R. Millán, A. Nijholt (editors). Toward Practical BCIs: Bridging the Gap from Research to Real-World Applications, 2012.

C. Brunner, G. Andreoni, L. Bianchi, B. Blankertz, C. Breitwieser, S. Kanoh, C. A. Kothe, A. Lécuyer, S. Makeig, J. Mellinger, P. Perego, Y. Renard, G. Schalk, I. P. Susila, B. Venthur, G. R. Müller-Putz. BCI Software Platforms. In: B. Z. Allison, S. Dunne, R. Leeb, J. del R. Millán, A. Nijholt (editors). Toward Practical BCIs: Bridging the Gap from Research to Real-World Applications, 2012.

C. Brunner, M. Billinger, C. Vidaurre, C. Neuper. A comparison of univariate, vector, bilinear autoregressive and band power features for brain-computer interfaces. Medical & Biological Engineering & Computing, 49:1337-1346, 2011. [DOI] [PDF]

M. Billinger, V. Kaiser, C. Neuper, C. Brunner. Automatic frequency band selection for BCIs with ERDS difference maps. Proceedings of the Fifth International Brain-Computer Interface Conference, Graz, September 22-24, 2011. [PDF]

T. Solis-Escalante, G. R. Müller-Putz, C. Brunner, V. Kaiser, G. Pfurtscheller. Analysis of sensorimotor rhythms for the implementation of a brain switch for healthy subjects. Biomedical Signal Processing and Control, 5:15-20, 2010. [DOI]

C. Brunner, M. Billinger, C. Neuper, G. Pfurtscheller. Influence of spatial filters on the performance of EEG-based BCIs using autoregressive models. Proceedings of the Neuromath Workshop, Leuven, Belgium, 2009.

Poster presentations

M. Billinger, C. Neuper, G. R. Müller-Putz, C. Brunner. User-centric performance estimation in a continuous online BCI. TOBI Workshop III, Würzburg, March 20-22, 2012. [PDF]

M. Billinger, V. Kaiser, C. Neuper, C. Brunner. Automatic frequency band selection for BCIs with ERDS difference maps. Fifth International Brain-Computer Interface Conference, Graz, September 22-24, 2011. [PDF]

M. Billinger, C. Brunner, C. Neuper. Classification of adaptive autoregressive models at different sampling rates in a motor imagery-based BCI. Fourth International BCI Meeting, Pacific Grove, CA, USA , 2010. [PDF]

C. Brunner, M. Billinger, C. Neuper. A comparison of univariate, multivariate, bilinear autoregressive, and bandpower features for brain-computer interfaces. Fourth International BCI Meeting, Pacific Grove, CA, USA, 2010. [PDF]

C. Brunner, M. Billinger, C. Neuper, G. Pfurtscheller. Influence of spatial filters on the performance of EEG-based BCIs using autoregressive models. Neuromath Workshop, Leuven, Belgium, 2009. [PDF]

Project leader
image/svg+xml

Clemens Brunner, Dipl.-Ing. Dr.tech
Technische Universität Graz
Institut für Semantische Datenanalyse

Funding
image/svg+xml

Duration
image/svg+xml

2008 - 2011