EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Generalised machine learning models outperform personalised models for cognitive load classification in real-life settings

Christoph Anders, Ipsita Bhaduri, Bert Arnrich

Frontiers in Digital Health · 06.10.2025 · 7

DOI: 10.3389/fdgth.2025.1650085 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Digital Health, online erschienen am 06.10.2025. Verfasst von Christoph Anders und 2 weiteren Autor:innen. Beteiligte Einrichtung: Digital Health - Connected Healthcare, Hasso Plattner Institute, University of Potsdam, Potsdam, Germany. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

By issuing work-break reminders, for example, personal assistants for cognitive load could be beneficial in maintaining health and life satisfaction in society. Wearable sensors facilitate the necessary real-time collection of physiological data. Still, publicly available real-life data sets obtained with wearable sensors are scarce, especially considering multi-modal recordings. Furthermore, data is usually recorded in either completely controlled or uncontrolled environments, missing the opportunity to study participants across optimal laboratory and realistic real-life settings.

Methods

This work collected data from ten university students during given and self-chosen cognitive load tasks, resembling typical working environments from over 40% of the OECD population, and investigated if commercially available sensors suffice for building cognitive load assistants. The study design accounted for a balanced distribution of eight working hours per participant, split between controlled and uncontrolled environments.

Results

Across participants, no single feature correlated significantly with cognitive load, but differences in smartwatch indices and biomarkers were identified between low- and high-load scenarios. Generalised machine learning models like Logistic Regression achieved F1 scores of up to 0.91, 0.77, and 0.54 for two, three, and five-class classification, respectively.

Discussion

The presented study design marks a step towards real-life mental state assistants, and the anonymised dataset was made publicly available.

Schlagwörter

Stichwörter der Autor:innen

cognitive load experimentshuman-centered computingmachine learningpersonal assistanttime-series classificationuncontrolled environmentwavelet decompositionwearable sensors

Autor:innen und Institutionen

  • Digital Health - Connected Healthcare, Hasso Plattner Institute, University of Potsdam, Potsdam, Germany.

Zitieren

APA

Anders, C., Bhaduri, I., & Arnrich, B. (2025). Generalised machine learning models outperform personalised models for cognitive load classification in real-life settings. Frontiers in Digital Health, 7. https://doi.org/10.3389/fdgth.2025.1650085

Vancouver

Anders C, Bhaduri I, Arnrich B. Generalised machine learning models outperform personalised models for cognitive load classification in real-life settings. Frontiers in Digital Health. 2025;7. doi:10.3389/fdgth.2025.1650085

BibTeX
@article{Anders2025,
  author = {Anders, Christoph and Bhaduri, Ipsita and Arnrich, Bert},
  title = {Generalised machine learning models outperform personalised models for cognitive load classification in real-life settings},
  journal = {Frontiers in Digital Health},
  year = {2025},
  volume = {7},
  doi = {10.3389/fdgth.2025.1650085}
}
Kostenlos starten →