EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Diverse and flexible behavioral strategies arise in recurrent neural networks trained on multisensory decision making

Thomas S Wierda, Shirin Dora iD, Cyriel M A Pennartz, Jorge F Mejias iD

PLOS Computational Biology · 2025

DOI: 10.1371/journal.pcbi.1013559 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Computational Biology, 2025. Verfasst von Thomas S Wierda und 3 weiteren Autor:innen. Beteiligte Einrichtung: Cognitive and Systems Neuroscience Group, Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam, The Netherlands. Hauptthemen laut MEDLINE: Recurrent Neural Networks, Decision Making. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 51 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NWA-ORC NWA, Nederlandse Organisatie voor Wetenschappelijk Onderzoek.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Behavioral variability across individuals leads to substantial performance differences during cognitive tasks, although its neuronal origin and mechanisms remain elusive. Here we use recurrent neural networks trained on a multisensory decision-making task to investigate inter-subject behavioral variability. By uniquely characterizing each network with a random synaptic-weights initialization, we observed a large variability in the level of accuracy, bias and decision speed across these networks, mimicking experimental observations in mice. Performance was generally improved when networks integrated multiple sensory modalities. Additionally, individual neurons developed modality-, choice- or mixed-selectivity, these preferences were different for excitatory and inhibitory neurons, and the concrete composition of each network reflected its preferred behavioral strategy: fast networks contained more choice- and mixed-selective units, while accurate networks had relatively less choice-selective units. External modulatory signals shifted the preferred behavioral strategies of networks, suggesting an explanation for the recently observed within-session strategy alternations in mice.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Recurrent Neural NetworksDecision MakingTask Performance and AnalysisAnimalsMicePerceptionBehavior, Animal

Autor:innen und Institutionen

  • Cognitive and Systems Neuroscience Group, Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam, The Netherlands.
  • Department of Computer Science, School of Science, Loughborough University, Loughborough, United Kingdom.

ORCID-Profile: Shirin Dora, Jorge F Mejias

Förderung

  • NWA-ORC NWA
  • Nederlandse Organisatie voor Wetenschappelijk Onderzoek

Zitieren

APA

Wierda, T. S., Dora, S., Pennartz, C. M. A., & Mejias, J. F. (2025). Diverse and flexible behavioral strategies arise in recurrent neural networks trained on multisensory decision making. PLOS Computational Biology. https://doi.org/10.1371/journal.pcbi.1013559

Vancouver

Wierda TS, Dora S, Pennartz CMA, Mejias JF. Diverse and flexible behavioral strategies arise in recurrent neural networks trained on multisensory decision making. PLOS Computational Biology. 2025. doi:10.1371/journal.pcbi.1013559

BibTeX
@article{Wierda2025,
  author = {Wierda, Thomas S and Dora, Shirin and Pennartz, Cyriel M A and Mejias, Jorge F},
  title = {Diverse and flexible behavioral strategies arise in recurrent neural networks trained on multisensory decision making},
  journal = {PLOS Computational Biology},
  year = {2025},
  doi = {10.1371/journal.pcbi.1013559}
}
Kostenlos starten →