A neural network model of when to retrieve and encode episodic memories
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
Article and author information
Author details
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
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)
Further reading
-
- Genetics and Genomics
- Neuroscience
Thermal nociception in Caenorhabditis elegans is regulated by the Ca²+/calmodulin-dependent protein kinase CMK-1, but its downstream effectors have remained unclear. Here, we combined in vitro kinase assays with mass-spectrometry-based phosphoproteomics to identify hundreds of CMK-1 substrates, including the calcineurin A subunit TAX-6, phosphorylated within its conserved regulatory domain. Genetic and pharmacological analyses reveal multiple antagonistic interactions between CMK-1 and calcineurin signaling in modulating both naive thermal responsiveness and adaptation to repeated noxious stimuli. Cell-specific manipulations indicate that CMK-1 acts in AFD and ASER thermo-sensory neurons, while TAX-6 functions in FLP thermo-sensory neurons and downstream interneurons. Since CMK-1 and TAX-6 act in distinct cell types, the phosphorylation observed in vitro might not directly underlie the behavioral phenotype. Instead, the opposing effects seem to arise from their distributed roles within the sensory circuit. Overall, our study provides (1) a resource of candidate CMK-1 targets for further dissecting CaM kinase signaling and (2) evidence of a previously unrecognized, circuit-level antagonism between CMK-1 and calcineurin pathways. These findings highlight a complex interplay of signaling modules that modulate thermal nociception and adaptation, offering new insights into potentially conserved mechanisms that shape nociceptive plasticity and pain (de)sensitization in more complex nervous systems.
-
- Immunology and Inflammation
- Neuroscience
During aging, microglia – the resident macrophages of the brain – exhibit altered phenotypes and contribute to age-related neuroinflammation. While numerous hallmarks of age-related microglia have been elucidated, the progression from homeostasis to dysfunction during the aging process remains unresolved. To bridge this gap in knowledge, we undertook complementary cellular and molecular analyses of microglia in the mouse hippocampus across the adult lifespan and in the experimental aging model of heterochronic parabiosis. Single-cell RNA-Seq and pseudotime analysis revealed age-related transcriptional heterogeneity in hippocampal microglia and identified intermediate states of microglial aging that also emerge following heterochronic parabiosis. We tested the functionality of intermediate stress response states via TGFβ1 and translational states using pharmacological approaches in vitro to reveal their modulation of the progression to an activated state. Furthermore, we utilized single-cell RNA-Seq in conjunction with in vivo adult microglia-specific Tgfb1 conditional genetic knockout mouse models to demonstrate that microglia advancement through intermediate aging states drives transcriptional inflammatory activation and hippocampal-dependent cognitive decline.