Employment and Comparison of Different Artificial Neural Networks for Epilepsy Diagnosis from EEG Signals

In this study, it has been intended to analyze Electroencephalography (EEG) signals by Wavelet Transform (WT) for diagnosis of epilepsy, to employ various Artificial Neural Networks (ANNs) for the signals’ automatic classification. Furthermore, carrying out a performance comparison has been aimed. T...

Full description

Saved in:
Bibliographic Details
Published in:Journal of medical systems 2012-02, Vol.36 (1), p.347-362
Main Authors: Sezer, Esma, Işik, Hakan, Saracoğlu, Esra
Format: Article
Language:English
Subjects:
Electroencephalography
Electroencephalography - methods
Epilepsy
Epilepsy - diagnosis
Health Informatics
Health Sciences
Humans
Medical diagnosis
Medicine
Medicine & Public Health
Neural networks
Neural Networks (Computer)
Original Paper
Principal Component Analysis
Principal components analysis
ROC Curve
Signal Processing, Computer-Assisted
Statistics for Life Sciences
Wavelet Analysis
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this study, it has been intended to analyze Electroencephalography (EEG) signals by Wavelet Transform (WT) for diagnosis of epilepsy, to employ various Artificial Neural Networks (ANNs) for the signals’ automatic classification. Furthermore, carrying out a performance comparison has been aimed. Three EEG signals have been decomposed into frequency sub bands by WT and the feature vectors have been extracted from these sub bands. In order to reduce the sizes of the extracted feature vectors, Principal Component Analysis (PCA) method has been applied when necessary and these feature vectors have been classified by five different ANNs as either epileptic or healthy. The performance evaluation has been carried out by conducting ROC analysis for the used ANN models that and their comparisons have also been included.
ISSN:0148-5598
1573-689X
DOI:10.1007/s10916-010-9480-5