Comparative Study on Epilepsy Prediction Using Modern Techniques

Epilepsy is a persistent, contagious brain condition that affects about 50 million individuals worldwide. It is characterized by repeated epilepsy, which are short lived bursts of jerking movements that can either affects the whole body or just a tiny proportion of it. The way a seizure presents itself may differ greatly among epileptics. They may differ depending on the human characteristics. If we are aware of the four stages of a seizure, we can better understand what happens in the brain during seizures. 20% of epileptics experience preictal stages, and thus, this stage may serve as initial seizure warning signs. People who went through the prodromal phase may even feel empowered during this time since they may be informed of an impending seizure. Since the embedded information in the preictal phase follows the ictal phase, it is possible to predict epilepsy in earlier from the brain waves. It alerts the doctors to the impending seizure. Seizures may be anticipated, according to multiple sources in the fields of engineering, medicine, and patent publications. Some of the seizure prediction methods include frequency-based approaches, statistical EEG signal analysis, nonlinear dynamics (chaos), and intelligently designed systems. For epilepsy prediction, there are both machine and deep learning techniques. The methods are usually categorized on the basis of which type of classifier or which type of classification technique are used for prediction. The various preictal prediction techniques and their outcomes were thoroughly discussed in this review.