Enhanced Detection of Epileptic Seizure Using Hybrid Framework of Slantlet Transform and Spiking Neural Network Model

Epileptic seizures are sudden disturbances in the brain’s electrical activity, disrupting normal function. Early detection, accurate diagnosis, and timely treatment are essential to manage and reduce seizure frequency. This paper presents an intelligent model for epileptic seizure detection using two standard imbalanced EEG datasets. Electroencephalogram (EEG) signals serve as input for training and testing the model. The Slantlet Transform (SLT), an advanced signal processing technique, is used for effective feature extraction, offering improved localization and representation of EEG signals. These extracted features are utilized in a spiking neural network (SNN) to create a hybrid model, SLTSNN. The model is evaluated using the CHB-MIT and UCI-ESR datasets. Performance metrics such as accuracy, sensitivity, and F1-score are calculated and compared with five machine learning classifiers Logistic Regression (LRC), Support Vector Machine (SVMC), Long Short-Term Memory (LSTM), Random Forest (RFC), and Decision Tree (DTC) and one deep learning model, the Convolutional Neural Network (CNN). SLTSNN outperforms all baseline models, achieving 98% accuracy, 98% sensitivity, and 96% F1-score on the UCI-ESR dataset, and 93% across all three metrics on the CHB-MIT dataset. It also performs better than the results reported in eight existing methods.