Independent components for EEG signal classification

This paper addresses movement imagery detection via electroencephalogram (EEG) signal classification. Independent component analysis (ICA) is employed to factorise the time domain EEG signal. A three-layer Neural Network (NN) frame work is constructed to classify the movement imageries using the power spectrum features of independent components (ICs). The main contributions of the paper are reflected in the following points: (1) unlike existing methods, the ICA is not used to reject artifacts but considered as the source for extracting features to train the Neural Network classifiers; (2) a voting NN classification framework is proposed. The experiment results is based on the data obtained from the 2008 Berlin BCI Competition database and shows the proposed method has a high classification capacity.

[1]  M. Teshnehlab,et al.  An Evolutionary Artifact Rejection Method For Brain Computer Interface Using ICA , 2013 .

[2]  Bogdan M. Wilamowski.  Comparison of training algorithms and network architectures , 2013, 2013 IEEE 17th International Conference on Intelligent Engineering Systems (INES).

[3]  R. Palaniappan,et al.  Brain Computer Interface Design Using Band Powers Extracted During Mental Tasks , 2005, Conference Proceedings. 2nd International IEEE EMBS Conference on Neural Engineering, 2005..

[4]  Hao Yu,et al.  Selection of Proper Neural Network Sizes and Architectures—A Comparative Study , 2012, IEEE Transactions on Industrial Informatics.

[5]  Pierre Comon,et al.  Independent component analysis, A new concept? , 1994, Signal Process..

[6]  Wolfgang Rosenstiel,et al.  Online Artifact Removal for Brain-Computer Interfaces Using Support Vector Machines and Blind Source Separation , 2007, Comput. Intell. Neurosci..

[7]  Samy Bengio,et al.  HMM and IOHMM modeling of EEG rhythms for asynchronous BCI systems , 2004, ESANN.

[8]  R. Stott,et al.  The World Bank , 2008, Annals of tropical medicine and parasitology.

[9]  Clemens Brunner,et al.  Spatial filtering and selection of optimized components in four class motor imagery EEG data using independent components analysis , 2007, Pattern Recognit. Lett..

[10]  G Pfurtscheller,et al.  Using time-dependent neural networks for EEG classification. , 2000, IEEE transactions on rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society.

[11]  G. Schott.  Penfield's homunculus: a note on cerebral cartography. , 1993, Journal of neurology, neurosurgery, and psychiatry.

[12]  Charles W. Anderson,et al.  Classification of EEG Signals from Four Subjects During Five Mental Tasks , 2007 .

[13]  Pontus Forslund,et al.  A Neural Network Based Brain-Computer Interface for Classification of Movement Related EEG , 2003 .

[14]  Grega Repovš,et al.  Dealing with Noise in EEG Recording and Data Analysis , 2010 .

[15]  Tzyy-Ping Jung,et al.  A Collaborative Brain-Computer Interface for Improving Human Performance , 2011, PloS one.