Analysis of Epileptic Seizure Detection Using Deep Learning Algorithms

Epileptic seizures impact a patient's physical function and can cause irreversible damage to the brain. Timely detection of these seizures is crucial for administering appropriate antiepileptic treatment in the medical field. Recently, there has been significant interest in utilizing Deep Learning (DL) algorithms for detecting brain disorders. Deeper algorithms offer enhanced computational efficiency, accuracy, optimization, and reduced loss, thus improving the effectiveness of seizure detection. This study analyzed the various DL algorithms used in recent times for detecting epileptic seizures. The recent feature extraction techniques used in epileptic seizure are Convolutional Neural Network (CNN), Deep Neural Network (DNN), and Residual Network Long Short-Term Memory (ResNet LSTM). The recent classification techniques used in epileptic seizure are Convolutional Neural Network (CNN), CNN-Long Short-Term Memory (CNN-LSTM), and Deep Convolutional Autoencoder-Bidirectional Long Short Memory (DCAR-BiLSTM). The performance metrics used to evaluate the existing methods are accuracy, specificity, sensitivity, Positive Predictive Value (PPV), False Predictive Value (FPV), and Matthews's Correlation Coefficient (MCC).