Epileptic Seizure Detection Using EEG Signals: A Review

Abstract

Epilepsy is chronic neurological disorder of repeated seizures of very brief or large time
periods. It has a considerably negative impact on both the quality and the expectancy of life of the
patient. Electroencephalogram or EEG are test to evaluate the electrical signals of brain neurons and
hence to detect epilepsy. The living standard of such patients can be improved drastically if occurrence
of seizures can be predicted which is the main aim of EEG. Analysis of seizures from long-term EEG
recordings for neurologists is complicated and time-consuming. The use of machine training algorithms
for the recognition and classification of Epileptic EEG signal has led to the automated detection of
epilepsy that has gained considerable popularity amongst scientists because it can prove very useful for
treatment. Advanced EEG machines are now available which provides high quality signals but the
main challenge in automated epilepsy detection is to collect precise data and to develop detection
algorithm with minimum computation. Feature extraction and classification are the two key steps
involved in machine learning. The extraction feature reduces the space of the input pattern by keeping
information and assigns the classifier to a class label. We present various methods of extraction and
grading algorithms to automatically detect epilepsy in this paper.