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

Hippocampal replay contributes to within session learning in a temporal difference reinforcement learning model.

Adam Johnson1, A David Redish

  • 1Center for Cognitive Sciences and Graduate Program in Neuroscience, University of Minnesota, MN 55455, USA. john5726@umn.edu

Neural Networks : the Official Journal of the International Neural Network Society
|October 4, 2005
PubMed
Summary

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

The vicarious nature of hippocampal theta sequences.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same author

Just how goal-directed are hippocampal theta sweeps, anyway?

Nature neuroscience·2026
Same author

Noncanonical NF-κB pathway driven inflammation across multiple cellular compartments identifies NIK as a therapeutic target for inflammatory bowel disease.

Frontiers in immunology·2026
Same author

V-Cheiloplasty, Step-Down Technique, and Primary Rhinoplasty in Bilateral Cleft Lip Repair: Effective Strategies for Achieving Nasolabial Harmony.

Annals of plastic surgery·2026
Same author

Evaluation and Management of Syndesmotic Injuries: A Review of the Literature.

Current sports medicine reports·2026
Same author

Coordinated Representational Drift Across the Mouse Cortex.

bioRxiv : the preprint server for biology·2026

Hippocampal replay may enhance reinforcement learning speed by providing practice sequences, similar to how Dyna-Q algorithms learn faster. This study models how replay impacts learning rates in a rat task, offering testable predictions.

Area of Science:

  • Computational neuroscience
  • Cognitive neuroscience
  • Reinforcement learning

Background:

  • Temporal difference reinforcement learning (TDRL) algorithms model basal ganglia function but learn slower than animals.
  • Modified TDRL algorithms like Dyna-Q improve learning speed through offline practice.
  • The hippocampus exhibits replay of neural sequences during rest and sleep.

Purpose of the Study:

  • To investigate if hippocampal replay can accelerate TDRL learning.
  • To model the impact of replay on learning dynamics in a multiple-T choice task.
  • To generate testable predictions regarding hippocampal function in learning.

Main Methods:

  • Developed a computational model incorporating developing replay.
  • Simulated a multiple-T choice task with rats' learning characteristics.

Related Experiment Videos

  • Analyzed the effects of simulated replay on error reduction and path stereotyping.

    • Main Results:

    • Simulated replay accelerated the learning of the correct path.
    • Simulated replay slowed the development of stereotyped paths.
    • The model produced testable predictions for hippocampal inactivation and replay effects.

    Conclusions:

    • Hippocampal replay is a plausible mechanism for enhancing reinforcement learning speed.
    • Replay may differentially affect different aspects of learning, such as rapid adaptation versus strategy consolidation.
    • Computational models integrating neural phenomena like replay can yield valuable insights into brain function.