Wepilet, optimal orthogonal wavelets for epileptic seizure prediction with one single surface channel

Wepilet is a series of novel orthogonal wavelets optimized for Electroencephalogram (EEG) signals, specialized for epileptic seizure prediction. The main idea is to design a mother wavelet that when applied to EEG signal to create the feature space, should enable a better classification of the brain state. Wepilet is developed by an iterative optimization process, employing Genetic Algorithm (GA). Frequency sub-band features are first extracted using wepilet under design for the EEG signal captured by one single surface channel. These features are then fed to Support Vector Machines (SVMs) that classify the cerebral state in preictal and inter-ictal classes. The results of the classification are then used to compute the Probability of Error Rate (PER), which in turn is the GA objective function to be minimized. Results in a group of four patients, indicate the efficiency of optimized mother wavelet compared to the well-known Daubechies wavelet in EEG processing.