EEG Signals Classification Based on Time Frequency Analysis

Abstract

This paper presents a method to characterize, identify and classify some pathological Electroencephalogram (EEG) signals. We use some Time Frequency Distributions (TFDs) to analyze its nonstationarity. The analysis is conducted by the spectrogram (SP), the Choi–Williams Distribution (CWD) and the Smoothed Pseudo Wigner Ville Distribution (SPWVD). The studies are carried on some real EEG signals collected from a known database. The estimation of the best value of parameters for each distribution is achieved using the Rényi entropy (RE). The time-frequency results have permitted to characterize some pathological EEG signals. In addition, the Rényi Marginal Entropy (RME) is used for the purpose of detecting the peak seizures and discriminates between normal and pathological EEG signals. The frequency bands are evaluated using the Marginal Frequency (MF). The EEG signal classification of two sets A and E containing normal and pathologic EEG signals, respectively, is performed using our proposed method based on energy extraction of signals from time-frequency plane. Also, the Moving Average (MA) is used as a tool to obtain better classification results. The results conducted on real-life EEG signals illustrate the effectiveness of the proposed method.

Spectral information of EEG signals with respect to epilepsy classification

“…In study [20], EEG analysis using TFDs and particularly the spectrogram (SP), the Choi-Williams distribution (CWD) and the SPWVD are performed. The purpose of the study was both the identification of the seizure peaks and the classification of the EEG signals.…”

“…Most of the studies for epileptic activity detection/classification using EEG signal processing, formulate methodologies that analyse the EEG signal by extracting informative features from it [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20]. To this end, spectral analysis of the EEG signal is essential, since epileptic activity interrupts normal brain functionality.…”

“…The EEG spectral analysis is based on a set of frequency sub-bands. Researchers have mainly used wavelet transform (WT) [3][4][5][6][7][8][9][10][11][12][13][14][15][16] and time-frequency distributions (TFD) [17][18][19][20] to analyse the EEG spectral patterns. However, although spectral analysis is a well-known approach, with numerus studies including spectral characteristics in the features extracted from the EEG, the importance of the frequency sub-bands that are used to analyse the signal has never been thoroughly investigated in the literature.…”

“…However, although spectral analysis is a well-known approach, with numerus studies including spectral characteristics in the features extracted from the EEG, the importance of the frequency sub-bands that are used to analyse the signal has never been thoroughly investigated in the literature. It is medically established that brainwaves are divided based on their frequency into several sub-bands, being delta (1-4 Hz), theta (4)(5)(6)(7)(8), alpha (8)(9)(10)(11)(12)(13), beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) and gamma [21]. Thus, several researchers roughly focus on these sub-bands [3-14, 17, 18], with the technical limitations that the analysis technique imposes (i.e.…”

Time-Frequency Analysis of EEG Spectrograms using 2-D Gabor Filters for Epileptic Seizure Classification

“…Using these methods for EEG signals, they can be treated as images and divided in order to extract features from subimages. These techniques have shown good results in terms of accuracy for different applications [20]. Specifically, Short Time Fourier Transform (STFT) is an alternative that has previously used in epileptic seizure analysis by [25] and other recent works with competitive results in classification, where statistical, energy, and other features are extracted from spectrograms [7,21,15,14,2].…”