Effective Seizure Recognition from EEG Data by Asymmetrical Back Propagation Neural Network (ABPN) Classification

A significant task within neurology and epilepsy management is the differentiation of seizures from the Electroencephalogram (EEG signal). The current work of research focuses on suggesting an idea to enhance the detection of seizures by applying the Asymmetrical Back Propagation Neural Network (ABPN) classification. This dynamic aspect of ABPN is ideal for learning the details of the features and patterns of EEG since seizures are usually unilateral. We also compare our ABPN classifier with the standard neural networks to confirm that our proposed classifier perform better in sensitivity, specificity, and the overall assessment of the classifier in the global environment. According to the experiment, the authors are convinced that the ABPN model can solve the problem of lowering the false-positive rate along with the increase in the speed of detection to make it suitable for real-time seizure observation. This research will be useful in the improvement of better equipment to detect the seizures with a positive implication on patient health.