Epileptic Seizure Prediction on EEG Data using a Firefly Algorithm trained with Deep Neural Networks

An electroencephalogram (EEG) captures electrical activity in the brain without requiring any intrusive procedures. Electrodes are connected to the scalp to detect aberrant brain waves. To identify seizures and diagnose epilepsy, EEGs are often employed. However, they may also be used to analyse and diagnose sleep problems and brain injuries. Recently advancement of optimization, convolutional neural networks (CNNs) have replaced manual feature extraction in the processing of raw electroencephalogram (EEG) inputs. This method makes it easier to distinguish between interictal, ictal, and preictal intervals, which is necessary for the diagnosis of epileptic seizures. This research presents a unique method for predicting epileptic seizures the Firefly Algorithm with Deep Learning (FADL-ESP). The major area of the FADL-ESP method is to identify different types of seizures. Significant simulations were run using medical datasets to show the FADL-ESP system's enhanced performance. The experimental validation of the FADL-ESP method using an EEG signal dataset results showed that FADL-ESP outperformed the state-of-the-art models already on the market.