Assessing the validity and reliability of computational phenotyping of motivation
PLOS Mental Health · 16.11.2026 · 3 (11) · e0000730
DOI: 10.1371/journal.pmen.0000730 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS Mental Health, online erschienen am 16.11.2026. Verfasst von Claire Jaffré und 7 weiteren Autor:innen. Die Arbeit nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Fondation pour la Recherche Médicale, Institut National de la Santé et de la Recherche Médicale.
Abstract (Originalsprache)
Motivational disorders are a core clinical feature in neuropsychiatric disorders. The conventional method to assess motivation relies on clinical scales, however, an alternative approach is to use behavioral tasks, that can be combined with computational modelling. This is the so-called “computational phenotyping of motivation” approach. Nevertheless, if one wants to use computational phenotyping of motivation in longitudinal studies, it is crucial to assess the internal validity and the reliability of this approach. We employed two types of reward/effort trade-off tasks: preference tasks, where participants rated or selected among different options involving effort, punishment, and/or reward, and a performance task, where participants exerted motor effort to obtain rewards. Using a simulation approach, we verified the robustness of the computational models used to analyze these tasks. To assess test-retest effects, a sample of 72 healthy volunteers performed the tasks twice, 4 weeks apart. We also assessed the effect of demographic variables such as age and gender. In silico analyses demonstrated robust recovery of computational parameters for most aspects, even under strong data degradation conditions. Similarly, the test-retest reliability of our approach between visits was satisfactory. With the exception of a minor increase of sensitivity to reward in the performance task, no systematic differences were observed between the first and the second visits. Moreover, for most parameters the correlation between the two visits (across participants) was significant and above 0.5. Computational phenotyping of motivation seems reliable and largely unaltered by test-retest procedures, at least when evaluations are performed four weeks apart. These findings indicate that any disparities observed between two assessments can be readily interpreted, hence providing a solid methodological foundation for future longitudinal studies.
Förderung
- Fondation pour la Recherche Médicale
- Institut National de la Santé et de la Recherche Médicale
Zitieren
APA
Jaffré, C., Carrillo, P., Pigeon, E., Daunizeau, J., Bouc, R. L., Dhôte, J., Pessiglione, M., & Vinckier, F. (2026). Assessing the validity and reliability of computational phenotyping of motivation. PLOS Mental Health, 3(11), e0000730. https://doi.org/10.1371/journal.pmen.0000730
Vancouver
Jaffré C, Carrillo P, Pigeon E, Daunizeau J, Bouc RL, Dhôte J, et al. Assessing the validity and reliability of computational phenotyping of motivation. PLOS Mental Health. 2026;3(11):e0000730. doi:10.1371/journal.pmen.0000730
BibTeX
@article{Jaffr2026,
author = {Jaffré, Claire and Carrillo, Pablo and Pigeon, Etienne and Daunizeau, Jean and Bouc, Raphaël Le and Dhôte, Jacob and Pessiglione, Mathias and Vinckier, Fabien},
title = {Assessing the validity and reliability of computational phenotyping of motivation},
journal = {PLOS Mental Health},
year = {2026},
volume = {3},
number = {11},
pages = {e0000730},
doi = {10.1371/journal.pmen.0000730}
}
EUVIMED