Single channel EEG analysis for detection of depression

• Published in: Biomedical signal processing and control Vol. 31; pp. 391 - 397
• Main Authors: Bachmann, Maie, Lass, Jaanus, Hinrikus, Hiie
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2017
• Subjects: Classification accuracy; Detrended fluctuation analysis (DFA); Electroencephalography (EEG); Single channel; Spectral asymmetry index (SASI); Single channel; Spectral asymmetry index (SASI); Electroencephalography (EEG); Classification accuracy; Detrended fluctuation analysis (DFA)
• ISSN: 1746-8094
• DOI: 10.1016/j.bspc.2016.09.010

•Effectiveness of EEG spectral asymmetry index SASI to detect depression is confirmed.•Classification accuracy of linear SASI is comparable with that of nonlinear DFA.•Combination of SASI and DFA in single channel EEG provides accuracy of 91%. This study is aimed at finding a simple method for detection of depression based on the analysis of single channel short-term EEG signals. The accuracy of linear, spectral asymmetry index (SASI), and nonlinear, detrended fluctuation analysis (DFA), methods for differentiating depressive and healthy subjects was compared. The eyes closed EEG was recorded from 18 common reference (Cz) channels for 34 subjects (17 depressive and 17 control). The signals were stored at 400Hz sampling frequency and digitally filtered with cutoff frequencies at 0.5Hz and at 40Hz. The first 5min of each recording was selected for further analysis. The experiments indicated maximum difference for SASI values in channel Pz and for DFA values in channels Pz and O2. Therefore, channel Pz was selected for comparison of two methods. The results of statistical analysis show that SASI values are significantly higher for depressive than for control group (p=3.577e–05), while DFA values are significantly lower for depressive group (p=0.033). SASI has superior discrimination ability with classification accuracy of 76.5%, while the classification accuracy of DFA was 70.6%. Linear combination of SASI and DFA resulted in 91.2% classification accuracy. Our results demonstrate that the analysis of single channel signal can provide high accuracy of differentiation depression EEG.