Skip to content
DraconiticResearch
Search Sources API
201,683 studies31 sourcesUpdated 14 min ago
Peer-reviewedCC BY-NC-SAJournal of Neuroscience

One-shot learning of paired association navigation with biologically plausible schemas

  • Moorthy Ganeshkumar1,2,3,4
  • Cheston Tan5
  • Camilo Libedinsky6,7,2,8,9,4
  • Shih‐Cheng Yen1,2,8,4
  • Andrew Y. Y. Tan10,11,12,8,9,13
13 affiliations
  1. 1Engineering Design and Innovation Centre, College of Design and Engineering, National University of Singapore
  2. 2Integrative Sciences and Engineering Programme, NUS Graduate School, National University of Singapore
  3. 3School of Engineering and Applied Sciences, Harvard University
  4. 4The N.1 Institute for Health, National University of Singapore
  5. 5Institute for Infocomm Research, Agency for Science, Technology and Research
  6. 6Department of Psychology, National University of Singapore
  7. 7Institute of Molecular and Cell Biology, Agency for Science, Technology and Research
  8. 8NUS Artificial Intelligence Institute, National University of Singapore
  9. 9Neurobiology Programme, Life Sciences Institute, National University of Singapore
  10. 10Cardiovascular Disease Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore
  11. 11Department of Physiology, Yong Loo Lin School of Medicine, National University of Singapore
  12. 12Healthy Longevity Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore
  13. 13Neurobiology Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore
Open at sourceDOICite

Abstract

After learning a multiple paired-association (MPA) navigation task, rodents can rapidly learn new cue–location associations after a single exposure, a phenomenon proposed to be facilitated by cognitive schemas. However, it remains unclear what schemas support this behavior and how they could be implemented in biological neural circuits. Here, we identify a set of schemas sufficient for gradual then one-shot paired-association navigation, and implement them in a fully neural agent using biologically plausible plasticity rules. The agent gradually acquires a low-dimensional spatial metric representation using a path-integration based temporal-difference error, enabling self-localization across different environments. It also contains a memory network that forms one-shot associations between sensory cues and goal locations, implemented using reward-modulated exploratory Hebbian (EH) plasticity. A third network performs vector subtraction between the agent’s current and goal location to compute goal-directed actions. Supplementing these schemas with an actor–critic enables one-shot navigation past obstacles. Lastly, learning a working-memory gating mechanism using the temporal-difference error enables one-shot learning even with brief cue and distractor presentation. The model reproduces key behavioral features of rodent learning in the MPA task and generates testable predictions about the neural mechanisms underlying schema-based one-shot learning. Significance Statement Schema-dependent learning is exemplified by better recall of a lecture when one already has a schema for understanding. A rodent navigation task showing a shift from slow to rapid learning has been proposed to reflect such schema use, yet how schemas are neurally implemented and enable faster learning is poorly understood. We clarify this gap in this computational study by describing several schemas and implement them as neural networks with biologically plausible learning rules. Composing them into an agent reproduces schema-dependent rodent learning behavior. Because schemas also shape self beliefs, clarifying their neural basis has implications for psychopathology. At the same time, understanding how biological schemas support rapid learning may inform advances in machine learning.

Related

  1. Recurrent Spiking Neural Networks for Continuous Motor Control: A Biologically Plausible Model of Motor Cortex
    H. Imamura, F. Imamura, J. Igarashi +2
    bioRxivPreprintCC BYFull textNeurosciencebioRxiv
  2. Targeted Analysis of Whole Exome Sequencing Data Identifies Biologically Plausible Causal Variants in Genetically Unresolved Oculocutaneous Albinism Patients From India
    T. Dutta, D. Dey, A. Saha +3
    medRxivPreprintCC0Full textCell biologymedRxiv
  3. A likelihood-based framework for simultaneously learning both noise and growth dynamics using biologically-informed neural networks
    Rebecca M. Crossley, Ruth E. Baker
    arXivPreprintCC BY 4.0Full textBioinformatics & computational biologyEvolutionary biologyarXiv
  4. A redundant encoding algorithm for artificial sensory information speeds learning and improves multisensory-guided navigation
    M. C. Dadarlat, S. Senneka
    bioRxivPreprintCC BY-NC-NDFull textNeurosciencebioRxiv
  5. A Bio-Plausible Visual Neural Network for Locust-Inspired Collision Perception
    Qinbing Fu, Jiani Li, Jiajun Huang +1
    arXivPreprintCC BY-NC-ND 4.0Full textNeurosciencearXiv

Cite

BibTeXRIS
BibTeX
@article{ganeshkumar2026oneshot,
  title = {{One-shot learning of paired association navigation with biologically plausible schemas}},
  author = {Moorthy Ganeshkumar and Cheston Tan and Camilo Libedinsky and Shih‐Cheng Yen and Andrew Y. Y. Tan},
  journal = {Journal of Neuroscience},
  year = {2026},
  month = oct,
  doi = {10.1523/jneurosci.1803-24.2026},
  url = {https://doi.org/10.1523/jneurosci.1803-24.2026},
  note = {CC BY-NC-SA},
}
RIS
TY  - JOUR
TI  - One-shot learning of paired association navigation with biologically plausible schemas
AU  - Moorthy Ganeshkumar
AU  - Cheston Tan
AU  - Camilo Libedinsky
AU  - Shih‐Cheng Yen
AU  - Andrew Y. Y. Tan
PY  - 2026
DA  - 2026/10/26
JO  - Journal of Neuroscience
DO  - 10.1523/jneurosci.1803-24.2026
UR  - https://doi.org/10.1523/jneurosci.1803-24.2026
AB  - After learning a multiple paired-association (MPA) navigation task, rodents can rapidly learn new cue–location associations after a single exposure, a phenomenon proposed to be facilitated by cognitive schemas. However, it remains unclear what schemas support this behavior and how they could be implemented in biological neural circuits. Here, we identify a set of schemas sufficient for gradual then one-shot paired-association navigation, and implement them in a fully neural agent using biologically plausible plasticity rules. The agent gradually acquires a low-dimensional spatial metric representation using a path-integration based temporal-difference error, enabling self-localization across different environments. It also contains a memory network that forms one-shot associations between sensory cues and goal locations, implemented using reward-modulated exploratory Hebbian (EH) plasticity. A third network performs vector subtraction between the agent’s current and goal location to compute goal-directed actions. Supplementing these schemas with an actor–critic enables one-shot navigation past obstacles. Lastly, learning a working-memory gating mechanism using the temporal-difference error enables one-shot learning even with brief cue and distractor presentation. The model reproduces key behavioral features of rodent learning in the MPA task and generates testable predictions about the neural mechanisms underlying schema-based one-shot learning. Significance Statement Schema-dependent learning is exemplified by better recall of a lecture when one already has a schema for understanding. A rodent navigation task showing a shift from slow to rapid learning has been proposed to reflect such schema use, yet how schemas are neurally implemented and enable faster learning is poorly understood. We clarify this gap in this computational study by describing several schemas and implement them as neural networks with biologically plausible learning rules. Composing them into an agent reproduces schema-dependent rodent learning behavior. Because schemas also shape self beliefs, clarifying their neural basis has implications for psychopathology. At the same time, understanding how biological schemas support rapid learning may inform advances in machine learning.
SN  - 0270-6474
SN  - 1529-2401
ER  -