The electroencephalogram is a powerful tool for understanding the electrical activities of the brain. The automatic and accurate classification of extracranial and intracranial electroencephalogram signals are significant for the evaluation of epilepsy. Electroencephalogram signals contain significant characteristic information about epileptic brain waves. However, the electroencephalogram signals are easily disrupted by the artifacts polluting. This study proposed a clinical decision support system to extract significant epilepsy-related spectral features from the electroencephalogram signal. The artifact-free electroencephalogram signals features were obtained from the Kaiser window based on Finite Impulse Filter. The extracted features were modeled by the Artificial Neural Networks Back Propagation training algorithms which are Levenberg-Marquardt, Bayesian Regularization, and Scaled Conjugate Gradient. The algorithms' classification performances were compared by the accuracy rates. The experiment results show that compared with the Artificial Neural Networks Back Propagation training algorithms, the performance of the Levenberg-Marquardt is better from the point of accuracy rate which achieves a satisfying classification accuracy of 83.01% for extracranial and intracranial electroencephalogram signals.
Extracranial and Intracranial Electroencephalogram Signals Classification Finite Impulse Response Filter Kaiser Window Artificial Neural Networks Training Algorithms
The Scientific Technological Research Council of Turkey (TÜBİTAK)
118E682
Elektroansefalogram beyinin elektriksel aktivitelerini anlamak için güçlü bir araçtır. Ekstrakranial ve intrakranial elektroansefalogram sinyallerinin otomatik ve doğru sınıflandırılması epilepsinin değerlendirilmesi için önemlidir. Elektroansefalogram sinyali, epileptik beyin dalgası hakkında önemli karakteristik bilgi içermektedir. Fakat elektroansefalogram sinyali artefakt kirleticiler tarafından kolaylıkla bozulmaktadır. Bu çalışma, elektroansefalogram sinyalinden epilepsi hakkında önemli spektral özellikleri çıkarmak amacıyla klinik bir karar destek sistemi önermektedir. Artefakttan arındırılmış elektroansefalogram sinyal özellikleri, Kaiser penceresi tabanlı Sonlu Dürtü Yanıtı filtresinden elde edilmiştir. Yapay Sinir Ağları Geri Yayılım eğitim algoritmalarından Levenberg-Marquardt, Bayesian Düzenlenmesi ve Ölçekli Konjugat Gradyan algoritmalarına çıkarılan özellikler uygulanarak modellenmiştir. Algoritmaların sınıflandırma performansları doğruluk oranlarına göre karşılaştırılmıştır. Deneysel sonuçlar, Yapay Sinir Ağları Geri Yayılma eğitim algoritmaları ile yapılan deneyler karşılaştırıldığında, Levenberg-Marquardt algoritması ekstrakranial ve intrakranial elektroansefalogram sinyali için %83,01'lik tatmin edici bir sınıflandırma doğruluğu ile diğer algoritmalara göre daha iyi doğruluk oranı verdiğini gösterir.
Ekstrakranial ve İntrakranial Elektroansefalogram Sinyal Sınıflandırması Sonlu Dürtü Yanıtı Filtresi Kaiser Penceresi Yapay Sinir Ağları Eğitim Algoritmaları
118E682
Primary Language | English |
---|---|
Subjects | Artificial Intelligence |
Journal Section | Research Articles |
Authors | |
Project Number | 118E682 |
Publication Date | December 31, 2020 |
Submission Date | June 18, 2020 |
Acceptance Date | September 20, 2020 |
Published in Issue | Year 2020 Volume: 25 Issue: 3 |
Announcements:
30.03.2021-Beginning with our April 2021 (26/1) issue, in accordance with the new criteria of TR-Dizin, the Declaration of Conflict of Interest and the Declaration of Author Contribution forms fulfilled and signed by all authors are required as well as the Copyright form during the initial submission of the manuscript. Furthermore two new sections, i.e. ‘Conflict of Interest’ and ‘Author Contribution’, should be added to the manuscript. Links of those forms that should be submitted with the initial manuscript can be found in our 'Author Guidelines' and 'Submission Procedure' pages. The manuscript template is also updated. For articles reviewed and accepted for publication in our 2021 and ongoing issues and for articles currently under review process, those forms should also be fulfilled, signed and uploaded to the system by authors.