Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers

J.F. Plested, Tom Gedeon, X.Y. Zhu, Abhinav DHALL, Roland GOECKE

Research output: A Conference proceeding or a Chapter in BookConference contribution

1 Citation (Scopus)

Abstract

We conducted a pilot study experimenting with neural network techniques to use the physiological signals of untrained observers to classify the depression levels of variously depressed people in videos speaking a language the observers did not understand. As the dataset was highly imbalanced, noisy and thus extremely sensitive to relative class sizes, we developed a technique for dynamically oversampling the smaller classes both prior to and during training to approximately align training prediction rates for each class with knowledge of the prevalence of different levels of depression. In predicting the depression levels to a final accuracy of 57.9% over four classes and 78.9% over three classes we demonstrate the likelihood that universal cross-cultural indicators of depression exist. In addition, that some people’s automatic physiological responses to these indicators are strong enough that they can be used to predict depression categories of people to a significant degree of accuracy even when the observer does not understand the language the person is speaking. The final accuracy rate is significantly better than the diagnosis rates of doctors speaking to patients in their own language. The results show the potential these techniques have to improve diagnosis of depression, especially in areas with limited access to mental health professionals. This innovative approach demonstrates the importance of further experimentation in this area and research into universal cross-cultural depression indicators.
Original languageEnglish
Title of host publicationProceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017)
Place of PublicationSan Antonio, TX, USA
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages185-192
Number of pages8
ISBN (Electronic)9781538606803
ISBN (Print)9781538606810
DOIs
Publication statusPublished - 20 Oct 2017
Event1st Workshop on Tools and Algorithms for Mental Health and Wellbeing, Pain, and Distress: 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) - San Antonio, San Antonio, United States
Duration: 23 Oct 201723 Oct 2017
http://mhw.media.mit.edu/

Conference

Conference1st Workshop on Tools and Algorithms for Mental Health and Wellbeing, Pain, and Distress
Abbreviated titleMHWPD
CountryUnited States
CitySan Antonio
Period23/10/1723/10/17
Internet address

Fingerprint

Depression
Language
Mental Health
Research

Cite this

Plested, J. F., Gedeon, T., Zhu, X. Y., DHALL, A., & GOECKE, R. (2017). Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers. In Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017) (pp. 185-192). San Antonio, TX, USA: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ACIIW.2017.8272612
Plested, J.F. ; Gedeon, Tom ; Zhu, X.Y. ; DHALL, Abhinav ; GOECKE, Roland. / Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers. Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017). San Antonio, TX, USA : IEEE, Institute of Electrical and Electronics Engineers, 2017. pp. 185-192
@inproceedings{12fc2948b99142a09393e3c5e7f72695,
title = "Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers",
abstract = "We conducted a pilot study experimenting with neural network techniques to use the physiological signals of untrained observers to classify the depression levels of variously depressed people in videos speaking a language the observers did not understand. As the dataset was highly imbalanced, noisy and thus extremely sensitive to relative class sizes, we developed a technique for dynamically oversampling the smaller classes both prior to and during training to approximately align training prediction rates for each class with knowledge of the prevalence of different levels of depression. In predicting the depression levels to a final accuracy of 57.9{\%} over four classes and 78.9{\%} over three classes we demonstrate the likelihood that universal cross-cultural indicators of depression exist. In addition, that some people’s automatic physiological responses to these indicators are strong enough that they can be used to predict depression categories of people to a significant degree of accuracy even when the observer does not understand the language the person is speaking. The final accuracy rate is significantly better than the diagnosis rates of doctors speaking to patients in their own language. The results show the potential these techniques have to improve diagnosis of depression, especially in areas with limited access to mental health professionals. This innovative approach demonstrates the importance of further experimentation in this area and research into universal cross-cultural depression indicators.",
keywords = "Depression analysis, Physiological signals, Depression indicators, Observer analysis",
author = "J.F. Plested and Tom Gedeon and X.Y. Zhu and Abhinav DHALL and Roland GOECKE",
year = "2017",
month = "10",
day = "20",
doi = "10.1109/ACIIW.2017.8272612",
language = "English",
isbn = "9781538606810",
pages = "185--192",
booktitle = "Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017)",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Plested, JF, Gedeon, T, Zhu, XY, DHALL, A & GOECKE, R 2017, Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers. in Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017). IEEE, Institute of Electrical and Electronics Engineers, San Antonio, TX, USA, pp. 185-192, 1st Workshop on Tools and Algorithms for Mental Health and Wellbeing, Pain, and Distress, San Antonio, United States, 23/10/17. https://doi.org/10.1109/ACIIW.2017.8272612

Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers. / Plested, J.F.; Gedeon, Tom; Zhu, X.Y.; DHALL, Abhinav; GOECKE, Roland.

Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017). San Antonio, TX, USA : IEEE, Institute of Electrical and Electronics Engineers, 2017. p. 185-192.

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers

AU - Plested, J.F.

AU - Gedeon, Tom

AU - Zhu, X.Y.

AU - DHALL, Abhinav

AU - GOECKE, Roland

PY - 2017/10/20

Y1 - 2017/10/20

N2 - We conducted a pilot study experimenting with neural network techniques to use the physiological signals of untrained observers to classify the depression levels of variously depressed people in videos speaking a language the observers did not understand. As the dataset was highly imbalanced, noisy and thus extremely sensitive to relative class sizes, we developed a technique for dynamically oversampling the smaller classes both prior to and during training to approximately align training prediction rates for each class with knowledge of the prevalence of different levels of depression. In predicting the depression levels to a final accuracy of 57.9% over four classes and 78.9% over three classes we demonstrate the likelihood that universal cross-cultural indicators of depression exist. In addition, that some people’s automatic physiological responses to these indicators are strong enough that they can be used to predict depression categories of people to a significant degree of accuracy even when the observer does not understand the language the person is speaking. The final accuracy rate is significantly better than the diagnosis rates of doctors speaking to patients in their own language. The results show the potential these techniques have to improve diagnosis of depression, especially in areas with limited access to mental health professionals. This innovative approach demonstrates the importance of further experimentation in this area and research into universal cross-cultural depression indicators.

AB - We conducted a pilot study experimenting with neural network techniques to use the physiological signals of untrained observers to classify the depression levels of variously depressed people in videos speaking a language the observers did not understand. As the dataset was highly imbalanced, noisy and thus extremely sensitive to relative class sizes, we developed a technique for dynamically oversampling the smaller classes both prior to and during training to approximately align training prediction rates for each class with knowledge of the prevalence of different levels of depression. In predicting the depression levels to a final accuracy of 57.9% over four classes and 78.9% over three classes we demonstrate the likelihood that universal cross-cultural indicators of depression exist. In addition, that some people’s automatic physiological responses to these indicators are strong enough that they can be used to predict depression categories of people to a significant degree of accuracy even when the observer does not understand the language the person is speaking. The final accuracy rate is significantly better than the diagnosis rates of doctors speaking to patients in their own language. The results show the potential these techniques have to improve diagnosis of depression, especially in areas with limited access to mental health professionals. This innovative approach demonstrates the importance of further experimentation in this area and research into universal cross-cultural depression indicators.

KW - Depression analysis

KW - Physiological signals

KW - Depression indicators

KW - Observer analysis

UR - https://ieeexplore.ieee.org/document/8272612/

U2 - 10.1109/ACIIW.2017.8272612

DO - 10.1109/ACIIW.2017.8272612

M3 - Conference contribution

SN - 9781538606810

SP - 185

EP - 192

BT - Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017)

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - San Antonio, TX, USA

ER -

Plested JF, Gedeon T, Zhu XY, DHALL A, GOECKE R. Detection of Universal Cross-Cultural Depression Indicators from the Physiological Signals of Observers. In Proceedings of the 2017 Seventh International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2017). San Antonio, TX, USA: IEEE, Institute of Electrical and Electronics Engineers. 2017. p. 185-192 https://doi.org/10.1109/ACIIW.2017.8272612