The DOI matches its Crossref record; 38 of the 38 references it cites were found in a registry or an index; 3 of the 8 sources its claims cite were checked for support.
Unfolding the Black Box of Recurrent Neural Networks for Path Integration
Open accessAbstract
Abstract Path integration is essential for spatial navigation. Experimental studies have identified neural correlates for path integration, but exactly how the neural system accomplishes this computation remains unresolved. Here, we adopt recurrent neural networks (RNNs) trained to perform a path integration task to explore this issue. After training, we borrow neuroscience prior knowledge and methods to unfold the black box of the trained model, including: clarifying neuron types based on their receptive fields, dissecting information flows between neuron groups by pruning their connections, and analyzing internal dynamics of neuron groups using the attractor framework. Intriguingly, we uncover a hierarchical information processing pathway embedded in the RNN model, along which velocity information of an agent is first forwarded to band cells, band and grid cells then coordinate to carry out path integration, and finally grid cells output the agent location. Inspired by the RNN-based study, we construct a neural circuit model, in which band cells form one-dimensional (1D) continuous attractor neural networks (CANNs) and serve as upstream neurons to support downstream grid cells to carry out path integration in the 2D space. Our study challenges the conventional view of considering grid cells as the principal velocity integrator, and supports a neural circuit model with the hierarchy of band and grid cells.
Overview
CoChat Summary
TL;DR
The authors train recurrent neural networks on a path integration task and then apply neuroscience-style analyses (receptive-field classification, connection pruning, attractor dynamics) to reveal a hierarchical pathway in which velocity input goes to band cells, band and grid cells interact reciprocally, and grid cells output position. Inspired by this, they build a hand-designed circuit model of 1D band-cell CANNs feeding a 2D grid-cell CANN, arguing against the conventional view that grid cells alone are the principal velocity integrator.
Key findings
- In one trained RNN with N_R = 4096 units, the authors identify 764 grid cells, 764 band cells, and 2568 undefined cells; all three types emerge robustly across seeds.
- Pruning velocity inputs to band cells produces the largest error increase (e.g., 2.32 ± 0.43% vs 0.11 ± 0.07% for grid cells and 0.14 ± 0.09% for undefined cells when 750 connections are pruned), while pruning read-out connections from grid cells has the largest effect (9.15 ± 0.71% vs 2.67 ± 0.50% for band cells).
- Pruning either band→grid or grid→band connections substantially degrades performance, indicating bidirectional cooperation is critical; a new RNN built with the hierarchical pathway as a prior matches the original's performance and develops band and grid cells in the first and second modules respectively, while a location-reconstruction task yields grid cells but no band cells.
- Band cells cluster into four groups by spacing and orientation, show direction tuning with preferred directions at θ or θ + π, form ring-like Isomap embeddings consistent with 1D CANNs, and have center-surround-like connectivity; the proposed circuit requires 2N pure band + 4N conjunctive band + N² grid cells versus 5N² for grid-only models.
Methods
- Vanilla RNN (N_R = 4096, ReLU, r_{t+1} = σ(Jr_t + Mv_t), read-out p_t = Wr_t) trained by BPTT with cross-entropy loss on a path integration task in a 2.2 m × 2.2 m arena, initialized from 512 place cells (difference-of-softmax tuning), 100 epochs, RMSProp, learning rate 10⁻⁴.
- Neuron types classified from spatial rate maps using spectral band score (>5.8) and grid score (>0.88) criteria; four step-by-step pruning experiments (velocity inputs, read-out connections, cell types, band–grid connections) with error bars over n = 30 repetitions.
- Population dynamics analyzed via PCA (15 components) followed by Isomap (3D, k = 8); a hand-designed neural circuit model with 1D band-cell CANNs and a 2D grid-cell CANN was simulated over 100,000 steps (Δt = 0.05) for activity mapping and 1,000 steps for trajectory decoding.
Limitations
- Stated by the authors: it is unclear whether undefined cells have spatial tuning (e.g., head direction or boundary vector properties) or contribute to path integration; the nature of band–grid interactions in the trained RNN is not fully solved and may not map directly onto the proposed circuit's connectivity; and the training paradigm is restricted to physical space, leaving generalization to abstract spaces untested.
- Evident in the text: the main circuit-model simulation is idealized with no noise in inputs or model (noise-induced drift is deferred to SI), and the authors report no compute resources, stating such specifications were deemed nonessential.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
3 Valid · 0 Not valid · 4 Not checked · 59 Own results and interpretations
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We train this new model with the same path integration task and have two key observations as expected, which are: 1) the new model achieves a performance comparable to the original model;”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.3 A hierarchical path integration pathway embedded in the RNN
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“2) band and grid cells emerge naturally as the dominant neuron types in the first and second modules, respectively (Fig. 3c).”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.3 A hierarchical path integration pathway embedded in the RNN
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We have identified three mutually exclusive types based on neuronal receptive fields (Fig. 1c, right panel), which are: 1) Grid cell, which exhibits the hexagonally patterned firing field; 2) Band cell, whose receptive field consists of multiple equally spaced parallel lines (the so-called bands) (see more details below); and 3) Undefined cell, which displays neither grid-like nor band-like spatia”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.2.1 Clarifying neuron types
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For example, in one trained RNN with N R = 4096 units, we identify 764 grid cells, 764 band cells, and 2568 undefined cells).”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.2.1 Clarifying neuron types
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“All three cell types emerge robustly in the trained RNNs (see SI E).”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.2.1 Clarifying neuron types
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
36 more claims
Create a free account to see all 66 claims and every source verdict.
- It may state two separate claims as one. 15
- Probably not a substantial claim. 12
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 3.1.2 The module of grid cells
“Grid cells are aligned on a 2D toroidal manifold with a phase vectorϕ = (ϕ 1 , ϕ 1 ).”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 3. Grid cells
“where u g (ϕ, t) and r g (ϕ, t) are the synaptic input and firing rate of a grid cell at phase ϕ = (ϕ 1 , ϕ 2 ).”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 2.3.1 Continuous attractor dynamics of band cells
“Denote ∆ ∈ (-λ/2, λ/2) the offset between the adjacent bands of two band cells, their phase difference is given by ϕ = (∆/λ) × π ∈ (-π, π).”
§ Unfolding the Black Box of Recurrent Neural Networks for Path Integration > 15. Institutional review board (IRB) approvals or equivalent for research with human subjects
“whose value is given by ϕ = (∆/λ) × 2π, with ∆ the offset between the adjacent bands of two band cells.”
Citations
Cites
Simultaneous localization and mapping: part I
Durrant-Whyte, H., & Bailey, T. (2006) · IEEE Robotics & Automation Magazine · doi:10.1109/mra.2006.1638022 · Cited by 4,199
Microstructure of a spatial map in the entorhinal cortex
Hafting, T., Fyhn, M., Molden, S., Moser, M.-B., & Moser, E. I. (2005) · Nature · doi:10.1038/nature03721 · Cited by 4,068
Path integration and the neural basis of the 'cognitive map'
McNaughton, B. L., Battaglia, F. P., Jensen, O., Moser, E. I., & Moser, M.-B. (2006) · Nature Reviews Neuroscience · doi:10.1038/nrn1932 · Cited by 2,024
Taube, J., Muller, R., & Ranck, J. (1990) · The Journal of Neuroscience · doi:10.1523/jneurosci.10-02-00420.1990 · Cited by 1,809
Conjunctive Representation of Position, Direction, and Velocity in Entorhinal Cortex
Sargolini, F., Fyhn, M., Hafting, T., McNaughton, B. L., Witter, M. P., Moser, M.-B., & Moser, E. I. (2006) · Science · doi:10.1126/science.1125572 · Cited by 1,485
The Head Direction Signal: Origins and Sensory-Motor Integration
Taube, J. S. (2007) · Annual Review of Neuroscience · doi:10.1146/annurev.neuro.29.051605.112854 · Cited by 927
Accurate Path Integration in Continuous Attractor Network Models of Grid Cells
Burak, Y., & Fiete, I. R. (2009) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1000291 · Cited by 779
Boundary Vector Cells in the Subiculum of the Hippocampal Formation
Lever, C., Burton, S., Jeewajee, A., O'Keefe, J., & Burgess, N. (2009) · The Journal of Neuroscience · doi:10.1523/jneurosci.1319-09.2009 · Cited by 753
Self-citationThe entorhinal grid map is discretized
Stensola, H., Stensola, T., Solstad, T., Frøland, K., Moser, M.-B., & Moser, E. I. (2012) · Nature · doi:10.1038/nature11649 · Cited by 701
An oscillatory interference model of grid cell firing
Burgess, N., Barry, C., & O'Keefe, J. (2007) · Hippocampus · doi:10.1002/hipo.20327 · Cited by 686
Self-citation
Research works printed but not linked: 1
- [29]29. Zilong Ji, Tianhao Chu, Xingsi Dong, Changmin Yu, Daniel Bush, Neil Burgess, and Si Wu. Dynamical modulation of hippocampal replay sequences through firing rate adaptation. BioRxiv, pages 2024-09, 2024.
Cited by
Zuo, J., He, Y., Zhang, W., Fang, F., & Wu, S. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.05.02.722453 · Cited by 1
Self-citationIf Grid Cells are the Answer, What is the Question? A Review of Normative Grid Cell TheoryPreprint
Dorrell, W., & Whittington, J. C. R. (2026) · eLife · doi:10.7554/elife.111058
If Grid Cells are the Answer, What is the Question? A Review of Normative Grid Cell TheoryPreprint
Dorrell, W., & Whittington, J. C. R. (2026) · eLife · doi:10.7554/elife.111058.1
Metadata
- Authors
- Chu, T., Wu, Y., Burgess, N., Ji, Z., & Wu, S.
- Affiliations
- Peking University (CN)
- Peking University
- University College London (GB)
- Corresponding author
- Wu, Si
- Venue
- Advances in Neural Information Processing Systems 38
- Published
- 2025
- Pages
- 113045-113078
- ISSN
- 1049-5258 (ISSN-L)
- OpenAlex
- W4415559134
- Cited by
- 3
- References
- 29
- Citations by year
- 2026: 3
- FWCI
- 6.72
- Type
- conference paper
- Language
- en
- Open access
- green
- doi.org
- Licence
- cc-by
- Locations
- neural information processing systems (publishedVersion)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- Funders
- National Natural Science Foundation of China (T2421004, 2021ZD0200204)
- National Key Research and Development Program of China (2021ZD0200204, 2024YFF1206500)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Memory and Neural Mechanisms
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Spatial Cognition and Navigation
- Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
- Keywords
- path integration, recurrent neural networks, spatial navigation, grid cells, continuous attractor networks, neural circuit model, hierarchical information processing, velocity integration, attractor dynamics, receptive fields, connection pruning
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Unfolding the Black Box of Recurrent Neural Networks for Path Integration. https://app.cochat.ai/papers/10.52202/085713-3412
- The paper
- Chu, T., Wu, Y., Burgess, N., Ji, Z., & Wu, S. (2025). Unfolding the Black Box of Recurrent Neural Networks for Path Integration. Advances in Neural Information Processing Systems 38. https://doi.org/10.52202/085713-3412
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.