Enhancing Seizure Detection from EEG Signals-Optimization Driven Feature Selection and Classification using Artificial Neural Networks

EEG seizure detection is critical in enabling prompt intervention and therapy for people with epilepsy. Using Artificial Bee Colony (ABC) optimization with an Artificial Neural Network (ANN) to serve as classifier, a novel seizure detection method is given in this research. The objective is to increase the efficacy of EEG signal-based seizure detection. First, significant elements from the EEG signals are retrieved, which include statistical metrics, frequency-domain properties, and time-frequency representations. The ABC optimization approach is then used to pick the most informative and discriminative features, with the goal of reducing dimensionality and improving classification performance. The selected characteristics are then sent into the ANN classifier to identify seizures. Extensive tests were carried out using a publicly accessible EEG dataset to assess the efficacy of our suggested technique. The proposed technique attained a high accuracy of 90% in seizure detection, according to the data. The combination of ABC optimization for feature selection and ANN classification showed successful in distinguishing seizure and non-seizure EEG segments.