Detection of K-complexes in sleep EEG with support vector machines

Sleep is a state that can be characterized by the electrical oscillations of nerve cells, where brain activity is more stable than waking. Transient waveforms observed in sleep electroencephalography are structures with specific amplitude and frequency characteristics that can occur in some stages o...

Full description

Saved in:
Bibliographic Details
Published in2017 25th Signal Processing and Communications Applications Conference (SIU) pp. 1 - 4
Main Authors Kantar, Tugce, Erdamar, Aykut
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2017
Subjects
automatic detection
Brain modeling
EEG
Electroencephalography
Feature extraction
k-complex
Kernel
Medical services
Sleep
Support vector machines
SVM
Online AccessGet full text

Cover

More Information
Summary:Sleep is a state that can be characterized by the electrical oscillations of nerve cells, where brain activity is more stable than waking. Transient waveforms observed in sleep electroencephalography are structures with specific amplitude and frequency characteristics that can occur in some stages of sleep. The determination of the k-complex, which is one of these structures, is performed by visual scoring of all night sleep recordings by expert physicians. For this reason, a decision support system that allows automatic detection of the k-complex can give physicians more objective results in diagnosis. In this study, sleep EEG records scored by a physician were analyzed in different methods from the literature. Three features have been determined that express the k-complex presence and k-complexes were detected using these features and support vector machines. As a result, the performance of the algorithm was evaluated and sensitivity and specificity were determined as 70.83 % and 85.29%, respectively.
DOI:10.1109/SIU.2017.7960311