Epileptic Seizure Diagnosis Using Wavelet Transform and Approximate Entropy in EEG Signals

Determining epileptic events in EEG signals poses a considerable challenge as it necessitates a detailed study of comprehensive long-term EEG data acquired from different channels. In this study, author conceived an unexplored approach to automatically recognize seizures from continuous EEG monitoring using a attribute named approximate entropy (ApEn) obtained from the operation of wavelet transforms. They examined the data using a machine learning algorithm, applying the approach to different sets of EEG data gathered by various healthcare organizations and research institutions. The results showed that the suggested approach acquired an outstanding 96.34
^%
accuracy in noticing seizures in a detailed database and displayed an excellent 100% precision when determining data from healthy individuals and people with epilepsy. The ApEn approach revealed excellent accuracy and efficiency resembled to other entropy methods. This study underlines prospects of ApEn as a practical tool for determining and evaluating practices of brain activity associated with epilepsy. This approach could prove to be a useful implement in determining epileptic seizures, finally enhancing the diagnosis and practice of remedies for people with epilepsy.