IMPACT OF WINDOW LENGTH AND DECORRELATION STEP ON ICA ALGORITHMS FOR EEG BLIND SOURCE SEPARATION

Gundars Korats, Steven Le Cam, Radu Ranta

2012

Abstract

Blind Source Separation (BSS) approaches for multi-channel EEG processing are popular, and in particular Independant Component Analysis (ICA) algorithms have proven their ability for artifacts removal and source extraction for this very specific class of signals. However, the blind aspect of these techniques implies well-known drawbacks. As these methods are based on estimated statistics from the data and rely on an hypothesis of signal stationarity, the length of the window is crucial and has to be chosen carefully: large enough to get reliable estimation and short enough to respect the rather non-stationary nature of the EEG signals. In addition, another issue concerns the plausibility of the resulting separated sources. Indeed, some authors suggested that ICA algorithms give more physiologically plausible results depending on the chosen whitening/sphering step. In this paper, we address both issues by comparing three popular ICA algorithms (namely FastICA, Extended InfoMax and JADER) on EEG-like simulated data and assessing their performance by using an original correlation matrices distance measure and a separation performance index. The results are consistent and lead us to a precise idea of minimal sample size that guarantees statistically robust results regarding the number of channels.

References

  1. Bell, A. and Sejnowsky, T. (1995). An informationmaximization approach to blind separation and blind deconvolution. Neural Computation, 7(6):1129-1159.
  2. Cichocki, A. and Amari, S.-I. (2002). Adaptive Blind Signal and Image Processing. Wiley-Interscience.
  3. Delorme, A. and Makeig, S. (2004). EEGLab: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1):9-21. http://sccn.ucsd.edu/wiki/EEGLAB.
  4. Hyvärinen, A. (1999). Fast and robust fixed-point algorithms for independent component analysis. IEEE Transactions on Neural Networks, 10(3):626-634.
  5. Ma, J., Gao, D., Ge, F., and Amari, S.-i. (2006). A one-bitmatching learning algorithm for independent component analysis. In Rosca, J., Erdogmus, D., Príncipe, J., and Haykin, S., editors, Independent Component Analysis and Blind Signal Separation, volume 3889 of Lecture Notes in Computer Science, pages 173-180. Springer Berlin / Heidelberg.
  6. Onton, J. and Makeig, S. (2006). Information-based modeling of event-related brain dynamics. In Neuper, C. and Klimesch, W., editors, Event-Related Dynamics of Brain Oscillations, volume 159 of Progress in Brain Research, pages 99 - 120. Elsevier.
  7. Palmer, J. e. a. (2010). Independent component analysis of high-density scalp EEG recordings. In Independent Component Analysis of High-density Scalp EEG Recordings, 10th EEGLAB Workshop, Jyvaskyla, Finland.
  8. Sanei, S. and Chambers, J. (2007). EEG signal processing. Wiley-Interscience.
  9. Schomer, D. and Lopes da Silva, F. (2011). Niedermeyers's Electroenephalography: Basic Principles, Clinical Applications and Related Fields. Wolters Kluwer, Lippincott Willimas & Wilkins, 6th edition.
  10. Smekhov, A. and Yeredor, A. (2004). Optimization of JADE using a novel optimally weighted joint diagonalization approach. In Proc. EUSIPCO, pages 221- 224.
  11. Särelä, J. and Vigario, R. (2003). Overlearning in marginal distribution-based ICA: Analysis and solutions. Journal of Machine Learning Research, 4:1447-1469.
  12. Vaseghi, S. and Jetelová, H. (2006). Principal and independent component analysis in image processing. In Proceedings of the 14th ACM International Conference on Mobile Computing and Networking (MOBICOM'06), pages 1-5. Citeseer.
  13. Wu, Y., Wu, B., Liu, J., and Lu, H. (2008). Probabilistic tracking on Riemannian manifolds. In Pattern Recognition, 2008. ICPR 2008. 19th International Conference on, pages 1-4. IEEE.
Download


Paper Citation


in Harvard Style

Korats G., Le Cam S. and Ranta R. (2012). IMPACT OF WINDOW LENGTH AND DECORRELATION STEP ON ICA ALGORITHMS FOR EEG BLIND SOURCE SEPARATION . In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing - Volume 1: BIOSIGNALS, (BIOSTEC 2012) ISBN 978-989-8425-89-8, pages 55-60. DOI: 10.5220/0003780000550060


in Bibtex Style

@conference{biosignals12,
author={Gundars Korats and Steven Le Cam and Radu Ranta},
title={IMPACT OF WINDOW LENGTH AND DECORRELATION STEP ON ICA ALGORITHMS FOR EEG BLIND SOURCE SEPARATION},
booktitle={Proceedings of the International Conference on Bio-inspired Systems and Signal Processing - Volume 1: BIOSIGNALS, (BIOSTEC 2012)},
year={2012},
pages={55-60},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0003780000550060},
isbn={978-989-8425-89-8},
}


in EndNote Style

TY - CONF
JO - Proceedings of the International Conference on Bio-inspired Systems and Signal Processing - Volume 1: BIOSIGNALS, (BIOSTEC 2012)
TI - IMPACT OF WINDOW LENGTH AND DECORRELATION STEP ON ICA ALGORITHMS FOR EEG BLIND SOURCE SEPARATION
SN - 978-989-8425-89-8
AU - Korats G.
AU - Le Cam S.
AU - Ranta R.
PY - 2012
SP - 55
EP - 60
DO - 10.5220/0003780000550060