Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Efficient reinforcement learning: computational theories, neuroscience and robotics.

Mitsuo Kawato1, Kazuyuki Samejima

  • 1ATR Computational Neuroscience Laboratories, 2-2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-0288, Japan. kawato@atr.jp

Current Opinion in Neurobiology
|March 22, 2007
PubMed
Summary

Reinforcement learning models behavior based on rewards and penalties. This study explores computational challenges in brain modeling, including speed, temporal-difference error computation, and a unified framework for brain areas.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Consensus Paper: Models of Cerebellar Functions.

Cerebellum (London, England)·2026
Same author

Extraction of robust functional connectivity patterns across psychiatric disorders using principal component analysis-based feature selection.

Imaging neuroscience (Cambridge, Mass.)·2026
Same author

Generalizable stratification based on thalamo-somatomotor functional connectivity predicts responses to antidepressants in patients with depression.

Molecular psychiatry·2025
Same author

Computational mechanisms of neuroimaging biomarkers uncovered by multicenter resting-state fMRI connectivity variation profile.

Molecular psychiatry·2025
Same author

Enhancement of the left frontoparietal network through real-time functional magnetic resonance imaging functional connectivity-informed neurofeedback and its impact on working memory in schizophrenia: A pilot study.

Psychiatry and clinical neurosciences·2025
Same author

Success-efficient/failure-safe strategy for hierarchical reinforcement motor learning.

PLoS computational biology·2025

Area of Science:

  • Computational neuroscience
  • Behavioral learning theory

Background:

  • Reinforcement learning (RL) algorithms are influential computational theories for behavior.
  • RL is based on reward and penalty mechanisms.
  • Existing RL models face challenges in biological plausibility.

Purpose of the Study:

  • To address three key theoretical issues in reinforcement learning for brain modeling.
  • To explore computational solutions for slow learning, temporal-difference error calculation, and integrating diverse brain areas.
  • To review computational studies and experimental data related to these challenges.

Main Methods:

  • Review of experimental data supporting RL theories.
  • Analysis of computational studies focusing on meta-parameters, hierarchy, modularity, and supervised learning.

Related Experiment Videos

  • Examination of how these computational approaches address RL limitations.

    • Main Results:

    • Plain reinforcement learning is too slow for brain modeling.
    • The computation of temporal-difference error remains a theoretical challenge.
    • A new computational framework is needed for integrating various brain areas.

    Conclusions:

    • Computational approaches like meta-parameters, hierarchy, modularity, and supervised learning offer potential solutions.
    • These advanced computational strategies are crucial for developing more biologically plausible reinforcement learning models.
    • Further research integrating computational and experimental findings is necessary to advance understanding of brain function.