Classification of focal EEG signals using FBSE based flexible time-frequency coverage wavelet transform

Epilepsy is a neurological disorder which involves the whole range of age from child to elderly people. Focal (FO) epilepsy is a kind of drug resistance epilepsy in which neurosurgical resection provides an opportunity for this life-threatening problem. A most common technique used to identify FO epileptic brain area has visually classified the electroencephalogram (EEG) signals related to FO epilepsy. These EEG signals can also be classified with an automated method based on advanced signal processing techniques to overcome the errors produced during visual observation of EEG signals. In this research work, an automated FO EEG signals classification method has been developed with the help of Fourier-Bessel series expansion (FBSE) based flexible time-frequency coverage wavelet transform. In this method, the features such as mixture correntropy (MC) and exponential energy (EE) have been involved for the classification of FO EEG signals. The classification task has been performed involving 10-fold cross-validation with least-squares support vector machine (LS-SVM) classifier. The developed automated method has also been optimized with probability (p)-value based feature ranking method. The achieved highest classification performance parameters like as accuracy (ACC), sensitivity (SEN), and specificity (SPE) are 95.85%, 95.47%, and 96.24% by this developed automated method. The developed automated method has also been tested at different signal to noise ratio (SNR) levels to check the robustness against noisy environments.