Recent Advances in An Optimized Learning Network based Ictal and Interictal States of Automatic Seizure Detection Using Multi-Channel Scalp EEG: A Systematic Review

Abstract

Automatic seizure detection using multi-channel scalp electroencephalography has emerged as a critical research domain due to its potential to assist clinicians in early diagnosis and continuous monitoring of epilepsy. The differentiation between ictal and interictal states remains a complex challenge due to the nonlinear, non-stationary, and high-dimensional nature of EEG signals. Recent advances in optimized learning networks, including deep learning architectures such as convolutional neural networks, recurrent neural networks, and hybrid models, have significantly improved classification performance. These models leverage temporal and spatial correlations in EEG data while incorporating optimization strategies such as attention mechanisms, evolutionary algorithms, and hyperparameter tuning to enhance detection accuracy. This systematic review explores contemporary developments in optimized learning frameworks for seizure detection, focusing on multi-channel scalp EEG analysis. The study evaluates methodological trends, datasets, feature extraction techniques, and model optimization approaches. Furthermore, it highlights challenges such as data imbalance, generalization, and computational complexity. The review aims to provide a comprehensive understanding of current research directions and identify opportunities for future advancements in automated seizure detection systems.