Epilepsy seizure detection using EEG signals

Abstract

Epilepsy is a neurological disease that is referred to as a disorder of the central nervous system characterized by the loss of consciousness and convulsions. Epileptic patients are subject to epileptic seizures caused by abnormal electrical discharges that lead to uncountable movements, convulsions and the loss of consciousness. Approximately 50 million people around the world are diagnosed with epilepsy, children and adults in the age range of 65–70 years old are effected the most. Although the main cause of this disease is unknown, however, most of the symptoms of the epilepsy seizure can be medically treated. Epileptic patients are subject to seizures that cause uncontrollable movements and loss of consciousness which can lead to serious injuries, and sometimes death. As a result, computerized seizure detection techniques are vital solutions for epileptic patients to protect them from dangers at the time of a seizure. In this paper, we propose an epilepsy seizures detecting method that can be implemented in a hardware device to help epileptic patients. The Electroencephalogram (EEG) is widely recognized for diagnosing and assessing brain activities and disorder. Our study utilized an EEG datasets that is used in various research regarding epilepsy detection. We processed the EEG signal in both time and frequency domains and applied a Chebyschev filter for preprocessing the signal, then, by using Wavelet Analysis, we decomposed the filtered signal into five sub-bands in both time and frequency domain. However, we only used the Delta sub-band for further processing. Discrete Wavelet Transform was used for feature extraction, then thresholding was implemented in order to determine the noisy part of the signal. Moreover, we applied some widely used classifiers for epilepsy seizure detection, and compared our results with other approches.