Two approaches for detection of abnormalities in EEG signals

In this study, we suggest two interesting approaches which allow good discrimination between normal and abnormal EEG signals. The most interesting advantage of these approaches is that classification relies only on one statistical feature which leads to a great data size reduction and hence a reduced memory space. Discrete wavelets transform (DWT) and smoothed pseudo Wigner-Ville distribution (SPWV) are used along comparison of their performances in exploitation and extraction of features from these signals. By exploiting the independent component analysis algorithm ICA and Support Vector Machines SVM classification, we show the relevance of the statistical variances regardless of the transform: DWT or SPWV. We proved that it is possible to achieve recognition rate of 100% for both DWT and SPWV, but SPWV requires more processing time. Therefore the SPWV is not adaptable to a classification in real-time monitoring of seizure detection.

[1]  Abdulhamit Subasi,et al.  EEG signal classification using PCA, ICA, LDA and support vector machines , 2010, Expert Syst. Appl..

[2]  Elif Derya Übeyli.  Time-varying biomedical signals analysis with multiclass support vector machines employing Lyapunov exponents , 2007, Digit. Signal Process..

[3]  Malika Kedir-Talha,et al.  Neural networks and SVM for heartbeat classification , 2012, 2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA).

[4]  Mostefa Mesbah,et al.  Time-Frequency Detection of EEG Abnormalities , 2003 .

[5]  Mostefa Mesbah,et al.  Detection of newborn EEG seizure using optimal features based on discrete wavelet transform , 2003, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03)..

[6]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[7]  Javad Hashemi,et al.  Automatic detection of epileptic seizure using time-frequency distributions , 2006 .

[8]  Abdulhamit Subasi,et al.  Automatic recognition of alertness level by using wavelet transform and artificial neural network , 2004, Journal of Neuroscience Methods.

[9]  Samir Avdakovic,et al.  Energy Distribution of EEG Signals: EEG Signal Wavelet-Neural Network Classifier , 2010, ArXiv.

[10]  R. Shantha Selva Kumari,et al.  Seizure detection in EEG using time frequency analysis and SVM , 2011, 2011 International Conference on Emerging Trends in Electrical and Computer Technology.

[11]  Yann LeCun,et al.  Comparing SVM and convolutional networks for epileptic seizure prediction from intracranial EEG , 2008, 2008 IEEE Workshop on Machine Learning for Signal Processing.