Advertisement

Selection of EEG Frequency Bands for Detection of Depression

  • K. Kalev
  • M. Bachmann
Part of the IFMBE Proceedings book series (IFMBE, volume 48)

Abstract

Major depression affects more than 18 million people in the United States every year. Early diagnosis is essential for appropriate treatment of depression by promoting remission and by preventing relapses. With the aim of improving the early diagnosis of depression we analyzed the power spectrum of electroencephalographic (EEG) signal of female depressive subjects (17) and female control subjects (17). Earlier studies have found significant differences in depressive patient EEG power spectrum compared to healthy control subjects. These studies have used traditional EEG frequency bands, which have not been selected for diagnostic purposes of depression. In current study we evaluated EEG relative power in different frequency bands from frequency range 1 - 40 Hz in channel P3 - Cz to find the frequency bands best differentiating depressive subjects from controls. In addition, the linear discriminant analysis through leave-one-out cross-validation was applied. Best results, differentiating depressive subjects from controls, were obtained from frequency band 5 - 7 Hz which is the narrow subset of traditional theta frequency band (4 - 8 Hz). Depression specific relative beta and relative gamma frequency power yielded also smaller classification errors than traditional beta and gamma frequency bands. However, the classification error of traditional alpha band was large and the systematic analyze of different frequency band powers in alpha range did not improve the classification error. In future, it would be interesting to analyze the modified relative band power excluding the alpha frequencies.

Keywords

EEG depression power spectrum 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • K. Kalev
    • 1
  • M. Bachmann
    • 1
  1. 1.Department of Biomedical Engineering, TechnomedicumTallinn University of TechnologyTallinnEstonia

Personalised recommendations