Validation of Machine Learning-Based Assessment of Major Depressive Disorder from Paralinguistic Speech Characteristics in Routine Care - Research Article

Aug 3, 2026

Upload by Violet

Machine learning speech analysis can enable long-term monitoring of major depressive disorder (MDD) during and after treatment. To evaluate this, 550 telephone interview speech samples from 267 individuals in routine care were collected. A machine learning system was trained and tested to determine MDD absence/presence using paralinguistic speech characteristics, compared against MDD diagnosis from the Structured Clinical Interview for DSM-IV. The system achieved 66% accuracy (70% sensitivity, 62% specificity) and improved over chance, yet it did not outperform established depression scales, and no substantial validity gains were achieved.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Depression and Anxiety

Hindawi Depression and Anxiety Volume 2024, Article ID 9667377, 12 pages [https://doi.org/10.1155/2024/9667377](https://doi.org/10.1155/2024/9667377) Research Article Validation of Machine Learning-Based Assessment of Major Depressive Disorder from Paralinguistic Speech Characteristics in Routine Care Jonathan F. Bauer , 1 Maurice Gerczuk ,2 Lena Schindler-Gmelch , 1 Shahin Amiriparian ,2 David Daniel Ebert ,3 Jarek Krajewski ,4 Björn Schuller ,2,5 and Matthias Berking 1 1Department for Clinical Psychology and Psychotherapy, Friedrich-Alexander-Universität Erlangen-Nürnberg, 91052 Erlangen, Germany 2Chair of Embedded Intelligence for Health Care & Wellbeing, University of Augsburg, 86159 Augsburg, Germany 3Department for Sport and Health Sciences, Technical University Munich, 80992 Munich, Germany 4Rhenish University of Applied Science Cologne, 50676 Cologne, Germany 5Group on Language, Audio, & Music, Imperial College London, London SW7 2AZ, UK Correspondence should be addressed to Jonathan F. Bauer; [jonathan.f.bauer@fau.de](jonathan.f.bauer@fau.de) Received 24 March 2023; Revised 14 March 2024; Accepted 22 March 2024; Published 9 April 2024 Academic Editor: Giulia Landi Copyright © 2024 Jonathan F. Bauer et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. New developments in machine learning-based analysis of speech can be hypothesized to facilitate the long-term monitoring of major depressive disorder (MDD) during and after treatment. To test this hypothesis, we collected 550 speech samples from telephone-based clinical interviews with 267 individuals in routine care. With this data, we trained and evaluated a machine learning system to identify the absence/presence of a MDD diagnosis (as assessed with the Structured Clinical Interview for DSM-IV) from paralinguistic speech characteristics. Our system classiﬁed diagnostic status of MDD with an accuracy of 66%(sensitivity: 70%, speciﬁcity: 62%) . Permutation tests indicated that the machine learning system classiﬁed MDD signiﬁcantly better than chance. However, deriving diagnoses from cut-oﬀ scores of common depression scales was superior to the machine learning system with an accuracy of 73% for the Hamilton Rating Scale for Depression (HRSD), 74% for the Quick Inventory of Depressive Symptomatology–Clinician version (QIDS-C), and 73% for the depression module of the Patient Health Questionnaire (PHQ-9) . Moreover, training a machine learning system that incorporated both speech analysis and depression scales resulted in accuracies between 73 and 76% . Thus, while ﬁndings of the present study demonstrate that automated speech analysis shows the potential of identifying patterns of depressed speech, it does not substantially improve the validity ofclassiﬁcations from common depression scales. In conclusion, speech analysis may not yet be able to replace common depression scales in clinical practice, since it cannot yet provide the necessary accuracy in depression detection. This trial is registered with DRKS00023670 . 1. Introduction Major depressive disorder (MDD) is a leading cause of disability worldwide [1] characterized by symptoms of depressed mood, loss of motivation, and behavioral alterations such as reduced activity and disturbed sleep [2] . Psychotherapeutic and pharmacotherapeutic interventions as well as their combination have been shown to be eﬀective treatments for MDD (e.g., [3]) . However, various studies have found high relapse rates after these treatments (e.g.,[4]) . Thus, there is a signiﬁcant need to closely monitor health status after acute treatment and to respond quickly with follow-up interventions if sustained remission is not achieved [5–7] . 2 Depression and Anxiety Such monitoring is not only likely to improve patients’health status b

FAQs

What data and clinical basis were used to validate the machine learning assessment of MDD?

The study used 550 speech samples from telephone-based clinical interviews with 267 individuals in routine care. MDD diagnosis for labeling was assessed using the Structured Clinical Interview for DSM-IV.

How accurate was the machine learning system based on paralinguistic speech characteristics?

The system classified MDD diagnostic status with 66% accuracy, with 70% sensitivity and 62% specificity, and performed significantly better than chance based on permutation tests.

Did speech-based machine learning outperform common depression scales in this study?

No. Classification from common depression scale cut-off scores was superior, with accuracies around 73–74% depending on the scale, while combining speech analysis with scales produced accuracies between 73 and 76%. The overall conclusion is that speech analysis does not yet replace standard depression scales in clinical practice.

Aug 3, 2026

Upload by Violet