Evaluation of depression severity in speech

Zhenyu Liu, Bin Hu*, Fei Liu, Huanyu Kang, Xiaoyu Li, Lihua Yan, Tianyang Wang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

Depression is a frequent affective disorder, leading to a high impact on patients, their families and society. Depression diagnosis is limited by assessment methods that rely on patient-reported or clinician judgments of symptom severity. Recently, many researches showed that voice is an objective indicator for depressive diagnosis. In this paper, we investigate a sample of 111 subjects (38 healthy controls, 36 mild depressed patients and 37 severe depressed patients) through comparative analysis to explore the correlation between acoustic features and depression severity. We extract features as many as possible according to previous researches to create a large voice feature set. Then we employ some feature selection methods to form compact subsets on different tasks. Finally, we evaluate depressive disorder severity by these acoustic feature subsets. Results show that interview is a better choice than reading and picture description for depression assessment. Meanwhile, speech signal correlate to depression severity in a medium-level with statistically significant (p < 0.01).

Original languageEnglish
Title of host publicationBrain Informatics and Health - International Conference, BIH 2016, Proceedings
EditorsHesham Ali, Yong Shi, Giorgio A. Ascoli, Deepak Khazanchi, Michael Hawrylycz
PublisherSpringer Verlag
Pages312-321
Number of pages10
ISBN (Print)9783319471020
DOIs
Publication statusPublished - 2016
Externally publishedYes
EventInternational Conference on Brain Informatics and Health, BIH 2016 - Omaha, United States
Duration: 13 Oct 201616 Oct 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9919 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceInternational Conference on Brain Informatics and Health, BIH 2016
Country/TerritoryUnited States
CityOmaha
Period13/10/1616/10/16

Keywords

  • Acoustic feature
  • Depression severity
  • Feature selection
  • PHQ-9
  • Speech

Fingerprint

Dive into the research topics of 'Evaluation of depression severity in speech'. Together they form a unique fingerprint.

Cite this