EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An Explainable Deep Learning Approach for Stress Detection in Wearable Sensor Measurements

Martin Karl Moser iD, Maximilian Ehrhart iD, Bernd Resch iD

Sensors · 06.08.2024 · 24 (16) · 5085

DOI: 10.3390/s24165085 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 06.08.2024. Verfasst von Martin Karl Moser und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Geoinformatics, University of Salzburg, 5020 Salzburg, Austria. Hauptthemen laut MEDLINE: Deep Learning, Wearable Electronic Devices, Stress, Psychological, Algorithms. Die Arbeit wurde bisher 50-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: FWF Austrian Science Fund, German Federal Ministry for Digital and Transport (BMDV), Austrian Science Fund (FWF).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Stress has various impacts on the health of human beings. Recent success in wearable sensor development, combined with advancements in deep learning to automatically detect features from raw data, opens several interesting applications related to detecting emotional states. Being able to accurately detect stress-related emotional arousal in an acute setting can positively impact the imminent health status of humans, i.e., through avoiding dangerous locations in an urban traffic setting. This work proposes an explainable deep learning methodology for the automatic detection of stress in physiological sensor data, recorded through a non-invasive wearable sensor device, the Empatica E4 wristband. We propose a Long-Short Term-Memory (LSTM) network, extended through a Deep Generative Ensemble of conditional GANs (LSTM DGE), to deal with the low data regime of sparsely labeled sensor measurements. As explainability is often a main concern of deep learning models, we leverage Integrated Gradients (IG) to highlight the most essential features used by the model for prediction and to compare the results to state-of-the-art expert-based stress-detection methodologies in terms of precision, recall, and interpretability. The results show that our LSTM DGE outperforms the state-of-the-art algorithm by 3 percentage points in terms of recall, and 7.18 percentage points in terms of precision. More importantly, through the use of Integrated Gradients as a layer of explainability, we show that there is a strong overlap between model-derived stress features for electrodermal activity and existing literature, which current state-of-the-art stress detection systems in medical research and psychology are based on.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningWearable Electronic DevicesStress, PsychologicalAlgorithmsHumansNeural Networks, ComputerAdultMaleFemale

Stichwörter der Autor:innen

Deep Generative EnsembleLSTMdeep learningexplainable AIgenerative adversarial networkintegrated gradientsphysiological sensor datastress detectionwearable sensors

Autor:innen und Institutionen

  • Department of Geoinformatics, University of Salzburg, 5020 Salzburg, Austria.
  • Center for Geographic Analysis, Harvard University, Cambridge, MA 02138, USA.

ORCID-Profile: Martin Karl Moser, Maximilian Ehrhart, Bernd Resch

Förderung

  • FWF Austrian Science Fund
  • German Federal Ministry for Digital and Transport (BMDV)
  • Austrian Science Fund (FWF)

Zitieren

APA

Moser, M. K., Ehrhart, M., & Resch, B. (2024). An Explainable Deep Learning Approach for Stress Detection in Wearable Sensor Measurements. Sensors, 24(16), 5085. https://doi.org/10.3390/s24165085

Vancouver

Moser MK, Ehrhart M, Resch B. An Explainable Deep Learning Approach for Stress Detection in Wearable Sensor Measurements. Sensors. 2024;24(16):5085. doi:10.3390/s24165085

BibTeX
@article{Moser2024,
  author = {Moser, Martin Karl and Ehrhart, Maximilian and Resch, Bernd},
  title = {An Explainable Deep Learning Approach for Stress Detection in Wearable Sensor Measurements},
  journal = {Sensors},
  year = {2024},
  volume = {24},
  number = {16},
  pages = {5085},
  doi = {10.3390/s24165085}
}
Kostenlos starten →