Methods for classifying depression in single channel EEG using linear and nonlinear signal analysis

•EEG based measures for detection of depression.•Comparing linear and nonlinear EEG analysis methods’ sensitivity to detect depression.•No single superior measure to detect depression.•High depression detection sensitivity by combining measures from single EEG channel. Depressive disorder is one of...

Full description

Saved in:

Bibliographic Details
Published in:	Computer methods and programs in biomedicine 2018-03, Vol.155, p.11-17
Main Authors:	Bachmann, Maie, Päeske, Laura, Kalev, Kaia, Aarma, Katrin, Lehtmets, Andres, Ööpik, Pille, Lass, Jaanus, Hinrikus, Hiie
Format:	Article
Language:	eng
Subjects:	Alpha power variability Depression Depression - classification Depression - physiopathology EEG Electroencephalography - methods Female Humans Male Middle Aged Nonlinear Dynamics Nonlinear signal processing Relative gamma power Sensitivity and Specificity Signal Processing, Computer-Assisted Spectral asymmetry index
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	•EEG based measures for detection of depression.•Comparing linear and nonlinear EEG analysis methods’ sensitivity to detect depression.•No single superior measure to detect depression.•High depression detection sensitivity by combining measures from single EEG channel. Depressive disorder is one of the leading causes of burden of disease today and it is presumed to take the first place in the world in 2030. Early detection of depression requires a patient-friendly inexpensive method based on easily measurable objective indicators. This study aims to compare various single-channel electroencephalographic (EEG) measures in application for detection of depression. The EEG recordings were performed on a group of 13 medication-free depressive outpatients and 13 gender and age matched controls. The recorded 30-channel EEG signal was analysed using linear methods spectral asymmetry index, alpha power variability and relative gamma power and nonlinear methods Higuchi's fractal dimension, detrended fluctuation analysis and Lempel-Ziv complexity. Classification accuracy between depressive and control subjects was calculated using logistic regression analysis with leave-one-out cross-validation. Calculations were performed separately for each EEG channel. All calculated measures indicated increase with depression. Maximal testing accuracy using a single measure was 81% for linear and 77% for nonlinear measures. Combination of two linear measures provides the accuracy of 88% and two nonlinear measures of 85%. Maximal classification accuracy of 92% was indicated using mixed combination of three linear and three nonlinear measures. The results of this preliminary study confirm that single-channel EEG analysis, employing the combination of measures, can provide discrimination of depression at the level of multichannel EEG analysis. The performed study shows that there is no single superior measure for detection of depression.
ISSN:	0169-2607 1872-7565