Classifying event-related desynchronization in EEG, ECoG and MEG signals

Authors:
N. Jeremy Hill;Thomas Navin Lal;Michael Schröder;Thilo Hinterberger;Guido Widman;Christian E. Elger;Bernhard Schölkopf;Niels Birbaumer
Affiliations:
MPI for Biological Cybernetics, Tübingen;MPI for Biological Cybernetics, Tübingen;Fraunhofer FIRST IDA group, Berlin;Inst. of Medical Psychology and Behavioral Neurobiology, University of Tübingen;Department of Epileptology, University of Bonn;Department of Epileptology, University of Bonn;MPI for Biological Cybernetics, Tübingen;Inst. of Medical Psychology and Behavioral Neurobiology, University of Tübingen
Venue:
DAGM'06 Proceedings of the 28th conference on Pattern Recognition
Year:
2006

Citing 2
Cited 3

Gene Selection for Cancer Classification using Support Vector Machines

Machine Learning
A brain computer interface with online feedback based on magnetoencephalography

ICML '05 Proceedings of the 22nd international conference on Machine learning

Detecting semantic category in simultaneous EEG/MEG recordings

CN '10 Proceedings of the NAACL HLT 2010 First Workshop on Computational Neurolinguistics
Importance weighted extreme energy ratio for EEG classification

ICONIP'10 Proceedings of the 17th international conference on Neural information processing: models and applications - Volume Part II
A subject transfer framework for EEG classification

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We employed three different brain signal recording methods to perform Brain-Computer Interface studies on untrained subjects. In all cases, we aim to develop a system that could be used for fast, reliable preliminary screening in clinical BCI application, and we are interested in knowing how long screening sessions need to be. Good performance could be achieved, on average, after the first 200 trials in EEG, 75–100 trials in MEG, or 25–50 trials in ECoG. We compare the performance of Independent Component Analysis and the Common Spatial Pattern algorithm in each of the three sensor types, finding that spatial filtering does not help in MEG, helps a little in ECoG, and improves performance a great deal in EEG. In all cases the unsupervised ICA algorithm performed at least as well as the supervised CSP algorithm, which can suffer from poor generalization performance due to overfitting, particularly in ECoG and MEG.