A neural network model of when to retrieve and encode episodic memories

  1. Qihong Lu  Is a corresponding author
  2. Uri Hasson
  3. Kenneth A Norman
  1. Princeton University, United States

Abstract

Recent human behavioral and neuroimaging results suggest that people are selective in when they encode and retrieve episodic memories. To explain these findings, we trained a memory-augmented neural network to use its episodic memory to support prediction of upcoming states in an environment where past situations sometimes reoccur. We found that the network learned to retrieve selectively as a function of several factors, including its uncertainty about the upcoming state. Additionally, we found that selectively encoding episodic memories at the end of an event (but not mid-event) led to better subsequent prediction performance. In all of these cases, the benefits of selective retrieval and encoding can be explained in terms of reducing the risk of retrieving irrelevant memories. Overall, these modeling results provide a resource-rational account of why episodic retrieval and encoding should be selective and lead to several testable predictions.

Data availability

The code is made publicly available here in a git repo: https://github.com/qihongl/learn-hippoUsers can also use this code ocean capsule to play with one example model to qualitatively replicate some results: https://codeocean.com/capsule/3639589/tree

The following data sets were generated

Article and author information

Author details

  1. Qihong Lu

    Department of Psychology, Princeton University, Princeton, United States
    For correspondence
    qlu@princeton.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0730-5240
  2. Uri Hasson

    Princeton Neuroscience Institute, Princeton University, Princeton, United States
    Competing interests
    The authors declare that no competing interests exist.
  3. Kenneth A Norman

    Princeton Neuroscience Institute, Princeton University, Princeton, United States
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5887-9682

Funding

Office of Naval Research (Multi-University Research Initiative Grant,ONR/DoD N00014-17-1-2961)

  • Uri Hasson
  • Kenneth A Norman

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Copyright

© 2022, Lu et al.

This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 6,347
    views
  • 1,140
    downloads
  • 33
    citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Qihong Lu
  2. Uri Hasson
  3. Kenneth A Norman
(2022)
A neural network model of when to retrieve and encode episodic memories
eLife 11:e74445.
https://doi.org/10.7554/eLife.74445

Share this article

https://doi.org/10.7554/eLife.74445

Further reading

    1. Neuroscience
    Patricia E Phelps, Sung Min Ha ... Xia Yang
    Research Article

    Olfactory ensheathing cells (OECs) are unique glial cells found in both central and peripheral nervous systems where they support continuous axonal outgrowth of olfactory sensory neurons to their targets. Previously, we reported that following severe spinal cord injury, OECs transplanted near the injury site modify the inhibitory glial scar and facilitate axon regeneration past the scar border and into the lesion. To better understand the mechanisms underlying the reparative properties of OECs, we used single-cell RNA-sequencing of OECs from adult rats to study their gene expression programs. Our analyses revealed five diverse OEC subtypes, each expressing novel marker genes and pathways indicative of progenitor, axonal regeneration, secreted molecules, or microglia-like functions. We found substantial overlap of OEC genes with those of Schwann cells, but also with microglia, astrocytes, and oligodendrocytes. We confirmed established markers on cultured OECs, and localized select top genes of OEC subtypes in olfactory bulb tissue. We also show that OECs secrete Reelin and Connective tissue growth factor, extracellular matrix molecules which are important for neural repair and axonal outgrowth. Our results support that OECs are a unique hybrid glia, some with progenitor characteristics, and that their gene expression patterns indicate functions related to wound healing, injury repair, and axonal regeneration.

    1. Neuroscience
    Aida Bareghamyan, Changfeng Deng ... Don B Arnold
    Tools and Resources

    Recombinant optogenetic and chemogenetic proteins are potent tools for manipulating neuronal activity and controlling neural circuit function. However, there are few analogous tools for manipulating the structure of neural circuits. Here, we introduce three rationally designed genetically encoded tools that use E3 ligase-dependent mechanisms to trigger the degradation of synaptic scaffolding proteins, leading to functional ablation of synapses. First, we developed a constitutive excitatory synapse ablator, PFE3, analogous to the inhibitory synapse ablator GFE3. PFE3 targets the RING domain of the E3 ligase Mdm2 and the proteasome-interacting region of Protocadherin 10 to the scaffolding protein PSD-95, leading to efficient ablation of excitatory synapses. In addition, we developed a light-inducible version of GFE3, paGFE3, using a novel photoactivatable complex based on the photocleavable protein PhoCl2c. paGFE3 degrades Gephyrin and ablates inhibitory synapses in response to 400 nm light. Finally, we developed a chemically inducible version of GFE3, chGFE3, which degrades inhibitory synapses when combined with the bio-orthogonal dimerizer HaloTag ligand-trimethoprim. Each tool is specific, reversible, and capable of breaking neural circuits at precise locations.