EUVIMEDEUVIMED entdecken
ValidierungsstudieOpen Access

Automated Speech Analysis for Risk Detection of Depression, Anxiety, Insomnia, and Fatigue: Algorithm Development and Validation Study

Automatisierte Sprachanalyse zur Risikoerkennung von Depression, Angst, Insomnie und Fatigue: Studie zur Algorithmusentwicklung und -validierung

Rachid Riad iD, Martin Denais iD, Marc de Gennes iD, Adrien Lesage iD, Vincent Oustric iD, Xuan Nga Cao iD, Stéphane Mouchabac iD, Alexis Bourla iD

Journal of Medical Internet Research · 31.10.2024 · 26 · e58572

DOI: 10.2196/58572 ↗

Auf einen Blick

Bei 865 gesunden Erwachsenen sagte mobile Sprachanalyse Depression, Angst, Insomnie und Fatigue mit moderater bis guter Güte voraus. Die Modellleistung unterschied sich nach Demografie.

Fragestellung

Wie gut können mobile Sprachdaten Depression, Angst, Insomnie und Fatigue in der Allgemeinbevölkerung erkennen und deren Ausprägung schätzen?

Methode

Entwicklungs- und Validierungsstudie mit train-valid-test-Split; mobile Sprachaufnahmen, Deep-Learning-Modelle und Fairness-/Unsicherheitsanalyse.

Ergebnisse

Bestes Modell: Whisper M mit max pooling und Oversampling. Depression: AUC 0,76 bzw. 0,78; Angst: AUC 0,77; Insomnie: AUC 0,73; Fatigue: AUC 0,68. Scores korrelierten mit r=0,31 bis 0,49.

Einschränkungen

Fairness über Altersgruppen schwächer; Unsicherheits- und Fairnessaspekte wurden zwar geprüft, aber die Leistung war nicht in allen Gruppen gleich.

Schlussfolgerung der Autor:innen

Die Studie zeigt das Potenzial sprachbasierter Systeme für eine mehrdimensionale psychische Gesundheitsbeurteilung in der Allgemeinbevölkerung. Fairness und Unsicherheitsabschätzung sind wichtige Beiträge für die klinische Nutzung.

Kernaussagen

  • 865 gesunde Erwachsene wurden einbezogen.
  • Sprachanalyse erkannte Depression, Angst, Insomnie und Fatigue.
  • Leistung variierte nach Geschlecht, Bildung und Alter.

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

Kurz gefasst

Validierungsstudie in der Zeitschrift Journal of Medical Internet Research, online erschienen am 31.10.2024. Verfasst von Rachid Riad und 7 weiteren Autor:innen. Beteiligte Einrichtung: Callyope, Paris, France. Hauptthemen laut MEDLINE: Sleep Initiation and Maintenance Disorders, Depression, Fatigue, Anxiety. Die Arbeit wurde bisher 41-mal zitiert (Crossref) und nennt 70 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

While speech analysis holds promise for mental health assessment, research often focuses on single symptoms, despite symptom co-occurrences and interactions. In addition, predictive models in mental health do not properly assess the limitations of speech-based systems, such as uncertainty, or fairness for a safe clinical deployment.

Objective

We investigated the predictive potential of mobile-collected speech data for detecting and estimating depression, anxiety, fatigue, and insomnia, focusing on other factors than mere accuracy, in the general population.

Methods

We included 865 healthy adults and recorded their answers regarding their perceived mental and sleep states. We asked how they felt and if they had slept well lately. Clinically validated questionnaires measuring depression, anxiety, insomnia, and fatigue severity were also used. We developed a novel speech and machine learning pipeline involving voice activity detection, feature extraction, and model training. We automatically modeled speech with pretrained deep learning models that were pretrained on a large, open, and free database, and we selected the best one on the validation set. Based on the best speech modeling approach, clinical threshold detection, individual score prediction, model uncertainty estimation, and performance fairness across demographics (age, sex, and education) were evaluated. We used a train-validation-test split for all evaluations: to develop our models, select the best ones, and assess the generalizability of held-out data.

Results

The best model was Whisper M with a max pooling and oversampling method. Our methods achieved good detection performance for all symptoms, depression (Patient Health Questionnaire-9: area under the curve [AUC]=0.76; F1-score=0.49 and Beck Depression Inventory: AUC=0.78; F1-score=0.65), anxiety (Generalized Anxiety Disorder 7-item scale: AUC=0.77; F1-score=0.50), insomnia (Athens Insomnia Scale: AUC=0.73; F1-score=0.62), and fatigue (Multidimensional Fatigue Inventory total score: AUC=0.68; F1-score=0.88). The system performed well when it needed to abstain from making predictions, as demonstrated by low abstention rates in depression detection with the Beck Depression Inventory and fatigue, with risk-coverage AUCs below 0.4. Individual symptom scores were accurately predicted (correlations were all significant with Pearson strengths between 0.31 and 0.49). Fairness analysis revealed that models were consistent for sex (average disparity ratio [DR] 0.86, SD 0.13), to a lesser extent for education level (average DR 0.47, SD 0.30), and worse for age groups (average DR 0.33, SD 0.30).

Conclusions

This study demonstrates the potential of speech-based systems for multifaceted mental health assessment in the general population, not only for detecting clinical thresholds but also for estimating their severity. Addressing fairness and incorporating uncertainty estimation with selective classification are key contributions that can enhance the clinical utility and responsible implementation of such systems.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Sleep Initiation and Maintenance DisordersDepressionFatigueAnxietyHumansAdultMaleFemaleMiddle AgedAlgorithmsSpeechSurveys and QuestionnairesYoung Adult

Stichwörter der Autor:innen

anxietycomputer-aided diagnosisdepressionfatiguemachine learningmental healthmental health symptom detectionspeech analysisspeech biomarkersspeech-based systemsvoice analysisvoice detection

Autor:innen und Institutionen

  • Callyope, Paris, France.
  • Department of Psychiatry, Saint-Antoine Hospital, Sorbonne University, Assistance publique - Hôpitaux de Paris, Paris, France.
  • Infrastructure for Clinical Research in Neurosciences, Paris Brain Institute, Paris, France.
  • Medical Strategy and Innovation Department, Clariane, Paris, France.
  • NeuroStim Psychiatry Practice, Paris, France.

ORCID-Profile: Rachid Riad, Martin Denais, Marc de Gennes, Adrien Lesage, Vincent Oustric, Xuan Nga Cao, Stéphane Mouchabac, Alexis Bourla

Zitieren

APA

Riad, R., Denais, M., de Gennes, M., Lesage, A., Oustric, V., Cao, X. N., Mouchabac, S., & Bourla, A. (2024). Automated Speech Analysis for Risk Detection of Depression, Anxiety, Insomnia, and Fatigue: Algorithm Development and Validation Study. Journal of Medical Internet Research, 26, e58572. https://doi.org/10.2196/58572

Vancouver

Riad R, Denais M, de Gennes M, Lesage A, Oustric V, Cao XN, et al. Automated Speech Analysis for Risk Detection of Depression, Anxiety, Insomnia, and Fatigue: Algorithm Development and Validation Study. Journal of Medical Internet Research. 2024;26:e58572. doi:10.2196/58572

BibTeX
@article{Riad2024,
  author = {Riad, Rachid and Denais, Martin and de Gennes, Marc and Lesage, Adrien and Oustric, Vincent and Cao, Xuan Nga and Mouchabac, Stéphane and Bourla, Alexis},
  title = {Automated Speech Analysis for Risk Detection of Depression, Anxiety, Insomnia, and Fatigue: Algorithm Development and Validation Study},
  journal = {Journal of Medical Internet Research},
  year = {2024},
  volume = {26},
  pages = {e58572},
  doi = {10.2196/58572}
}
Kostenlos starten →