Feature extraction for on-line EEG classification using principal components and linear discriminants
The study focuses on the problems of dimensionality reduction by means of principal component analysis (PCA) in the context of single-trial EEG data classification (i.e. discriminating between imagined left- and right-hand movement). The principal components with the highest variance, however, do not necessarily carry the greatest information to enable a discrimination between classes. An EEG data set is presented where principal components with high variance cannot be used for discrimination. In addition, a method based on linear discriminant analysis (LDA), is introduced that detects principal components which can be used for discrimination, leading to data sets of reduced dimensionality but similar classification accuracy.
KeywordsPrincipal component analysis Feature extraction Linear discriminant EEG classification Brain-computer interface Learning vector quantisation Adaptive autoregressive model
Bishop, C. M.
(1995): ‘Neural networks for pattern recognition; (Clarendon Press, Oxford)zbMATHGoogle Scholar
Duda, R. O.
andHart, P. E.
(1973): ‘Pattern classification and scene analysis’ (John Wiley & Sons)Google Scholar
Flotzinger, D., Kalcher, J.
, andPfurtscheller, G.
, (1992): ‘EEG classification by learning vector quantization’,Biomedizinische Technik
, pp. 303–309CrossRefGoogle Scholar
Flotzinger, D., Pfurtscheller, G., Neuper, Ch., Berger, J.
(1994): ‘Classification of non-averaged EEG data by learning vector quanTization and the influence of signal preprocessing’,Med. Biol. Eng. Comput.
, (5), pp. 571–576CrossRefGoogle Scholar
, (1995): ‘Self-organizing maps’ (Springer, Berlin)Google Scholar
Kalcher, J., Flotzinger, D., Neuper, Ch., Gölly, S.
(1996): ‘Graz Brain-Computer Interface II— towards communication between humans and computers based on online classification of three different EEG patterns’,Med. Biol. Eng. Comput.
, pp. 382–388CrossRefGoogle Scholar
McFarland, D. J., Neat, G. W., Read, R. R.
, andWolpaw, J. R.
(1993): ‘An EEG-based method for graded cursor control’,Psychobiol.
, pp. 77–81Google Scholar
Michie, D., Spiegelhalter, D. J.
, andTaylor, C. C.
(1994): ‘Machine Learning, Neural and Statistical Classification’, (Ellis Horwood Limited, Englewood Cliffs, N.J.)zbMATHGoogle Scholar
, andPfurtscheller, G.
(1994): ‘Neural network based classification of non-averaged event-related EEG responses’,Med. Biol. Eng. Comput.
, pp. 189–196CrossRefGoogle Scholar
Pfurtscheller, G., Flotzinger, D.
, andKalcher, J.
(1993): ‘Brain-computer interface—a new communication device for handicapped persons’,J. Microcomp. Appl.
, p. 293–299CrossRefGoogle Scholar
Pfurtscheller, G., Kalcher, J., Neuper, Ch., Flotzinger, D.
(1996): ‘On-line EEG classification during externally-paced hand movements using a neural net-work-based classifier’,Electroenceph. Clin. Neurophy.
, pp. 416–425CrossRefGoogle Scholar
Pfurtscheller, G., Neuper, Ch., Flotzinger, D.
(1997): ‘EEG-based discrimination between imagination of right and left hand movement’,Electroenceph. Clin. Neurophysiol.
(in press)Google Scholar
Pregenzer, M., Pfurtscheller, G.
, andFlotzinger, D.
(1996): ‘Automated feature selection with a distinction sensitive learning vector quantizer’,Neurocomput.
, pp. 19–29CrossRefGoogle Scholar
Ripley, B. D.
(1996): ‘Pattern recognition and neural networks’ (Cambridge University Press)Google Scholar
Schlögl, A. Flotzinger, D.
(1997): ‘Using adaptive autoregressive parameters for a brain-computer-interface experiment’, Proc. 19th Int Conf. IEEE/EMBS, Chicago, pp. 1533–1535Google Scholar
Wolpaw, J. R., McFarland, D. J., Neat, G. W.
, andForneris, C.
(1991): ‘An EEG-based brain-computer interface for cursor control’,Electroenceph. Clin. Neurophysiol.
, pp. 252–259CrossRefGoogle Scholar
Wolpaw, J. R.
, andMcFarland, D. J.
(1994): ‘Multichannel EEG-based brain-computer communication’,Electroenceph. Clin. Neurophysiol.
, pp. 444–449CrossRefGoogle Scholar