Comparative detection based on feature extraction of epileptic spike waves in EEG

Epilepsy is one of the most common neurological disorders that greatly disturb patients´ daily lives. Automatic spike detection of EEG of epileptic patients have a great influence on assisting doctors to quickly and easily diagnose whether the patient is epileptic. Nowadays, there are several existing methods and software that can recognize seizure-related EEG signals and epileptic spikes to help doctors to release their burdensome work, but hardly can we find the research of accuracy of software for automatic epileptic spikes detection. Therefore, we propose to study a comparison on the number of omission, false positives, accuracy and other norms between automatic detection and highly trained clinicians detection. From our comparative analysis, we conclude that automatic epileptic spike detection software often make an inaccurate detection and then we analyze the reasons. As a result, our research gives basis to optimize clinical diagnosis and automatic spike detection methods.