Multi-domain feature analysis for depression: A study of N170 in time, frequency and spatial domains simultaneously

Author

Qiangfeng Zhao ; Junfeng Sun ; Fengyu Cong ; Shan Chen ; Yingying Tang ; Shanbao Tong

Author_Institution

Sch. of Biomed. Eng., Shanghai Jiao Tong Univ., Shanghai, China

fYear

2013

fDate

3-7 July 2013

Firstpage

5986

Lastpage

5989

Abstract

This study investigated the differences in event-related potentials (ERPs) between depression and normal control groups by using the cue-target paradigm with facial expressions as stimuli. Conventional ERPs analysis did not show a significant difference in the N170 amplitude or latency between the two groups. However, the multi-domain feature analysis of N170 by nonnegative tensor factorization (NTF) indicated that N170 in depression group had lower power compared with the normal control group for all three different emotional (i.e., happy, neutral, and sad) facial stimuli (Q ≤ 0.05). The results revealed the perceptual early-stage dysfunction in face processing for depression.

Keywords

electroencephalography; emotion recognition; face recognition; feature extraction; medical disorders; medical image processing; psychology; ERP; N170; NTF; cue-target paradigm; depression; emotional facial stimuli; event-related potentials; face processing; facial expressions; frequency domain; multidomain feature analysis; nonnegative tensor factorization; spatial domain; time domain; Analysis of variance; Educational institutions; Electroencephalography; Face; Feature extraction; Tensile stress; Time-frequency analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE

Conference_Location

Osaka

ISSN

1557-170X

Type

conf

DOI

10.1109/EMBC.2013.6610916

Filename

6610916