The preictal state is very useful for seizure prediction, as it starts few minutes before the seizure. The aim of this project is to predict epileptic seizure by detecting the start of preictal state’s sufficient time before the ictal state or onset of seizure starts. Early prediction helps patients, as medication can be done by the doctors to prevent the seizure. This project aims at providing cost-effective solution for predicting and monitoring Seizure Onset with Electroencephalogram data.
The concept of Machine Learning is implemented fore predicting the onset of seizure before in hand. A neural network is implemented to predict the EEG signals received by the sensors. It has two hidden layers for prediction. This neural network has an accuracy up to 97%, which is good, effective neural network to predict an occurring seizure. The Arduino Mega 2560 is used for process the signals received by the sensors. EEG sensors, pulse sensor, and temperature sensor are used to measure EEG signals, heart rate and the body temperature of the patient.
Analog signals are received by the sensors and given to Arduino Mega 2560 for digitization using built-in A-D converter. These digital signals are recorded for future use. They are now given to the neural network for prediction of the seizure of the patient. If there is a prediction of seizure, the patient can refrain themselves from doing any work that might cause harm to them such as swimming or driving. This system gives almost accurate results in predicting seizures and is patient specific results based on their history. The reports are analyzed to keep the patient safe and healthy.