From the perceptron to the cerebellum

Jean-Pierre Nadal; Nicolas Brunel; Vincent Hakim

From the perceptron to the cerebellum

T0 review · 0 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper argues that Marr-Albus-Ito theory remains the organizing framework for cerebellar motor learning.

desk verdict A solid, honest review of Marr-Albus-Ito theory that consolidates the field and flags its own soft spots; nothing new, but it deserves a serious referee and probably acceptance after minor fixes. read the letter →

arxiv 2505.14355 v1 pith:7QRWWQ6P submitted 2025-05-20 q-bio.NC cond-mat.dis-nnphysics.bio-ph

Nicolas Brunel , Vincent Hakim , Jean-Pierre Nadal This is my paper · ORCID

classification q-bio.NCcond-mat.dis-nnphysics.bio-ph

keywords cerebellumperceptronMarr-Albus-ItotheoryPurkinjecellgranulecellsclimbingfibersynapticplasticitysupervisedlearning

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This review argues that the Marr-Albus-Ito theory remains the organizing framework for cerebellar motor learning. The core idea is that a Purkinje cell acts as a perceptron: granule cells recode mossy-fiber inputs in a much higher-dimensional, sparser space, and a climbing fiber from the inferior olive supplies the teacher that depresses active parallel-fiber synapses. The paper traces this proposal from its origin through statistical-physics calculations of perceptron capacity, showing how the theory predicts the measured distribution of silent versus active synapses and how it extends to correlated inputs, temporal basis functions, and a perturbation-based solution to credit assignment. A sympathetic reader comes away with the cerebellum as a concrete biological example of a supervised learning machine.

What carries the argument

The central object is the sign-constrained perceptron model of a Purkinje cell, with $N$ parallel-fiber inputs, non-negative synaptic weights, and a threshold on the summed input; the learning rule is depression at active parallel-fiber synapses when the climbing fiber fires. Two quantities carry the argument: the expansion ratio of the granule-cell representation, which makes overlapping mossy-fiber patterns nearly orthogonal and linearly separable, and the critical capacity $\alpha_c = p_{\max}/N$ from statistical-physics analyses, which predicts that at maximal capacity a large fraction of synapses should be silent, matching the distribution seen in paired recordings.

What would settle it

A falsifying observation would be a motor-learning experiment in which complex spikes are recorded during a well-controlled task and shown to encode reward expectation or movement rather than error, or in which blocking complex spikes does not prevent learning; either would break the teacher role. Alternatively, large-scale imaging showing dense, low-dimensional granule-cell activity during natural behavior would undercut the expansion argument.

Watch

Extended reading notes

Core claim

The paper's central claim is that the cerebellar cortex should be understood as a supervised learning machine of the perceptron type: Purkinje cells are the output units, granule cells provide an expanded representation that makes input patterns easier to separate, and each climbing fiber acts as a teacher that reports error and depresses active parallel-fiber synapses. It consolidates the evidence by combining anatomical numbers (about 100,000 parallel-fiber inputs per Purkinje cell), the combinatorial recoding argument for expansion, classical linear-separability results, statistical-physics capacity calculations for sign-constrained weights, and quantitative fits to paired electrophysiological recordings. The review also presents later extensions: learning correlated input-output sequences, analog perceptrons, temporal basis functions for properly timed responses, and a perturbation-based algorithm in which climbing fibers serve both as exploratory noise and as an error feedback signal. The contribution is not a new experiment but a defense of Marr-Albus-Ito theory as the organizing account of cerebellar motor learning.

Load-bearing premise

The load-bearing premise is that each climbing fiber delivers a dedicated per-cell error signal and that the associations a Purkinje cell must learn are statistically independent; if real climbing fibers mainly carry reward or behavioral information, or if mossy-fiber inputs are strongly correlated, the capacity and credit-assignment arguments shift.

Editorial extensions

If this is right

The high-dimensional, sparse recoding by granule cells should make mossy-fiber patterns linearly separable, giving a functional reason for the large expansion between mossy fibers and parallel fibers.
At maximal capacity with sign-constrained weights, the perceptron predicts a large fraction of silent parallel-fiber synapses, matching the roughly 80 percent silent contacts reported in paired recordings.
Learning statistically correlated input-output sequences can raise capacity, and bistability helps further when output correlations exceed input correlations.
Temporal basis functions from unipolar brush cells, diverse mossy-fiber time scales, and short-term plasticity allow Purkinje cells to produce correctly timed outputs even after the sensory input has ended.
A perturbation-based algorithm in which climbing fibers both perturb movements and signal success reproduces observed plasticity rules and offers a candidate solution to the cerebellar credit-assignment problem.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If climbing fibers turn out to carry reward-related signals as often as error signals, the perceptron metaphor would need to be broadened to a reinforcement-learning rule, preserving Marr-Albus-Ito as a family of supervised and perturbation learners rather than a strict error-teacher account.
The expansion principle may be a general circuit motif: the same expansion-and-sparsification design appears in hippocampal and insect olfactory relays, so the optimality result for four to five inputs per expansion cell suggests a testable common design rule across systems.
The independence limitation points to a natural experiment: record natural mossy-fiber activity during a learned behavior and measure how much capacity a Purkinje cell retains when input correlations match those of the behavior.
The perturbation-based algorithm implies that spontaneous complex spikes before learning act as exploratory motor perturbations, a prediction that could be tested by tracking trial-to-trial movement variability against complex-spike timing.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This is a review, not a research paper, and it is a good review. It traces the perceptron-to-cerebellum arc from Marr and Albus through Gardner's capacity calculations to modern work on expansion recoding, temporal basis functions, and credit assignment. If you want one readable account of where Marr-Albus-Ito theory stands and where it wobbles, this is it.

What the paper does well: it is accurate on the biology and the theory. The geometric argument for granule-cell expansion is clearly explained, the perceptron capacity results are correctly summarized, and the comparison of predicted silent-synapse fractions with Isope and Barbour's data is presented fairly. It also gives real credit to alternative views: climbing fibers carry reward and behavioral signals, not just error; the perturbation-based credit-assignment scheme is explicitly flagged as awaiting in vivo demonstration; and the independence assumption in standard capacity calculations is acknowledged as unrealistic, with correlated-input extensions cited. The conclusion is appropriately modest about going beyond Marr-Albus-Ito, which is the right note for a review that is also a tribute.

The soft spots are minor. Figure 1's legend says "molecular layer interneurons (UBC)" where it should say "MLI" or "basket/stellate cells"—an obvious typo, but the kind that confuses readers. "Thomson's lab" is a typo for Thompson's lab. The paper is also visibly authored by people who did much of the work being reviewed, so self-citation is frequent; here that is appropriate rather than inflationary, since the cited papers are the ones that actually established the results.

The only real concern, if you want one, is that the paper does not critically discriminate between the strong claim that the cerebellum is a supervised learning machine and the weaker claim that Marr-Albus-Ito theory is the best current framework. But the authors themselves use the weaker framing throughout, so this is not a flaw in execution.

Who is this for? A reader who wants a compact, reliable entry point into the theory of cerebellar learning, or a physicist curious how perceptron ideas migrated into neuroscience. It is not a research contribution and makes no pretense of being one. I would send it to peer review and accept after the typos are fixed.

Referee Report

0 major / 7 minor

Summary. This review article, written in honor of Gérard Toulouse, traces the intellectual path from the perceptron to the cerebellum. It recounts how Marr and Albus cast Purkinje cells as perceptrons supervised by climbing-fiber teaching signals, with granule cells providing a high-dimensional expansion of mossy-fiber inputs. The paper reviews the statistical-physics and Gardner-style capacity calculations for perceptrons with sign-constrained weights, the comparison of the predicted synaptic weight distribution to experimental Purkinje-cell data, the extension to temporal basis functions for delayed and timed responses, and the modern debate about whether climbing fibers carry a dedicated per-cell error signal or also reward and behavioral information. It closes by identifying open questions, including credit assignment in complex movements, the functional role of cerebellar rhythms, and the extent to which cerebellar learning goes beyond the classical Marr-Albus-Ito framework.

Significance. As a review for a special issue commemorating Gérard Toulouse, the paper serves a clear and appropriate purpose: it explains how the perceptron, a central object in the statistical physics of neural networks, became a foundational model for cerebellar motor learning and how this perspective remains influential. Its significance lies in synthesis rather than new results, but the synthesis is accurate, balanced, and current. The authors explicitly acknowledge the main limitations of the classical theory, including the unrealistic statistical-independence assumption in capacity calculations and the evidence that climbing fibers are not purely error signals. The review is particularly valuable for readers outside the cerebellum field who want a condensed, authoritative account of the Marr-Albus-Ito theory and its modern variants, including the perturbation-based credit-assignment proposal of Bouvier et al. It also gives appropriate credit to the historical role of Toulouse and his collaborators in importing statistical physics ideas into neuroscience.

minor comments (7)

[Fig. 1 caption] The caption of Fig. 1A labels 'molecular layer interneurons (UBC)', which conflicts with the standard abbreviation for unipolar brush cells and with the text that treats molecular layer interneurons and UBCs as distinct cell types; the caption should be corrected to 'molecular layer interneurons (MLI)'.
[Section II] The sentence 'Numerous experiments, notably from Thomson's lab [22]' contains a typo: the laboratory is that of Richard Thompson, so 'Thomson's lab' should be 'Thompson's lab'.
[Section VI] The phrase 'the presence of absence of a second complex spike' should read 'the presence or absence of a second complex spike'.
[Throughout] Several typographical errors should be corrected: 'representions' in Section III, 'paralell' in Section IV, 'Purkjinje' in Section VI, and 'seeked' in Section I.
[Section III and reference list] The name 'Cayco Gajic' is written inconsistently (without a hyphen in the text, with a hyphen in the reference list); it should be unified to 'Cayco-Gajic'.
[Section II] The phrase 'Electrophysiological experiments inin vitro preparations' contains a doubled 'in'; it should read 'in in vitro preparations' or 'in vitro preparations'.
[Reference list] Several references contain character-encoding artifacts (e.g., 'p. 960ˆ a€“962', '38–55' garbled as 'p. 38ˆ a€“55'): these should be cleaned in the final typeset version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is a review that makes no derivation, fit, or prediction reducing to its own inputs.

full rationale

This paper is a review of Marr-Albus-Ito theory and its modern developments; it does not present a new derivation or a prediction that is forced by its own assumptions. The closest quantitative comparison is in Section IV, where the perceptron weight distribution is confronted with experimental recordings, but the authors explicitly say the model parameters are obtained by 'Fitting the model to the empirical data', so the comparison is presented as a fit rather than as an independent prediction. The capacity calculations are attributed to Gardner's statistical-physics framework and to previous published work (e.g., Brunel et al. 2004), and those results are externally checkable rather than defined in terms of the review's claims. The paper's self-citations, including Clopath et al. and Bouvier et al., are used to report published, experimentally confronted findings, not to forbid alternative interpretations or to import uniqueness conclusions. The authors also explicitly acknowledge open limitations: 'one unrealistic assumption is that learned associations are statistically independent' and, regarding the perturbation credit-assignment algorithm, 'it remains to be demonstrated that this algorithm operates in vivo in the cerebellum during learning of any motor behavior.' These are honest caveats, not circular moves. No load-bearing step reduces, by the paper's own equations or by definition, to its inputs; therefore the appropriate circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

This review introduces no free parameters or invented entities. The axioms listed are modeling assumptions from the reviewed literature, not new postulates. For example, the perceptron analogy for Purkinje cells and the climbing fiber error signal are central to Marr-Albus-Ito theory and are explicitly discussed as assumptions in Sections IV and VI.

assumptions (3)

domain assumption Purkinje cells can be modeled as binary or analog perceptrons with non-negative weights.
The review's central narrative rests on this modeling analogy, introduced in Section IV.
domain assumption Climbing fibers provide a teaching or error signal that drives plasticity at parallel fiber-Purkinje cell synapses.
Core to Marr-Albus-Ito theory as reviewed in Section VI.
domain assumption Granule cell expansion into high dimensions improves pattern separation.
Reviewed in Section III as the functional role of granule cells.

how reviews work

0 comments

Cite this review

Pith. "Pith review of From the perceptron to the cerebellum." pith.science (2026). https://pith.science/paper/7QRWWQ6P

@misc{pith2026250514355,
  author       = {Pith},
  title        = {Pith review of: From the perceptron to the cerebellum},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7QRWWQ6P}},
  note         = {Machine review of arXiv:2505.14355}
}

read the original abstract

The perceptron has served as a prototypical neuronal learning machine in the physics community interested in neural networks and artificial intelligence, which included G\'erard Toulouse as one of its prominent figures. It has also been used as a model of Purkinje cells of the cerebellum, a brain structure involved in motor learning, in the early influential theories of David Marr and James Albus. We review these theories, more recent developments in the field, and highlight questions of current interest.

Figures

Figures reproduced from arXiv: 2505.14355 by the authors.

**Figure 1.** FIG. 1: Anatomy of the cerebellum. A: Cell types and connectivity showing Purkinje cells (PC), granule cells (GrC), unipolar [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** FIG. 2: The Purkinje cell as a Perceptron. A: Sketch of a perceptron with granule cell inputs, tasked with learning input-output [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** FIG. 3: Simulation of a Granular network with or without UBCs, that are characterized by a broad diversity of synaptic time [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: FIG. 4: A. Plasticity rule predicted by a stochastic gradient descent algorithm using climbing fibers as both the perturbation [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

97 extracted references · 78 canonical work pages

[1]

Solvable model of a spin-glass,

D. Sherrington and S. Kirkpatrick, “Solvable model of a spin-glass,”Physical review letters, vol. 35, no. 26, p. 1792, 1975

1975
[2]

Coexistence of spin-glass and ferromagnetic orderings,

M. Gabay and G. Toulouse, “Coexistence of spin-glass and ferromagnetic orderings,”Physical Review Letters, vol. 47, no. 3, p. 201, 1981. 10

1981
[3]

Theory of the frustration effect in spin glasses: I,

G. Toulouse, “Theory of the frustration effect in spin glasses: I,”Comm. Phys., vol. 2, pp. 115–19, 1977

1977
[4]

Mezard, G

M. Mezard, G. Parisi, and M. Virasoro,Spin Glass Theory and Beyond. WORLD SCIENTIFIC, 1986

1986
[5]

Neural networks and physical systems with emergent collective computational abilities.,

J. J. Hopfield, “Neural networks and physical systems with emergent collective computational abilities.,”Proceedings of the national academy of sciences, vol. 79, no. 8, pp. 2554–2558, 1982

1982
[6]

Spin glass model of learning by selection.,

G. Toulouse, S. Dehaene, and J.-P. Changeux, “Spin glass model of learning by selection.,”Proceedings of the National Academy of Sciences, vol. 83, no. 6, pp. 1695–1698, 1986

1986
[7]

Networks of formal neurons and memory palimpsests,

J.-P. Nadal, G. Toulouse, J.-P. Changeux, and S. Dehaene, “Networks of formal neurons and memory palimpsests,” Europhysics Letters, vol. 1, no. 10, p. 535, 1986

1986
[8]

Information storage in sparsely coded memory nets,

J.-P. Nadal and G. Toulouse, “Information storage in sparsely coded memory nets,”Network: Computation in Neural Systems, vol. 1, no. 1, pp. 61–74, 1990

1990

Show all 97 references

[9]

Information capacity of a perceptron,

N. Brunel, J.-P. Nadal, and G. Toulouse, “Information capacity of a perceptron,”Journal of Physics A: Mathematical and General, vol. 25, no. 19, p. 5017, 1992

1992
[10]

G. M. Shepherd,The synaptic organization of the brain. Oxford university press, 2003

2003
[11]

Computational principles of supervised learning in the cerebellum,

J. L. Raymond and J. F. Medina, “Computational principles of supervised learning in the cerebellum,”Annual review of neuroscience, vol. 41, no. 1, pp. 233–253, 2018

2018
[12]

The theory and neuroscience of cerebellar cognition,

J. D. Schmahmann, X. Guell, C. J. Stoodley, and M. A. Halko, “The theory and neuroscience of cerebellar cognition,” Annual review of neuroscience, vol. 42, no. 1, pp. 337–364, 2019

2019
[13]

The cerebellar cortex,

C. Hull and W. G. Regehr, “The cerebellar cortex,”Annual review of neuroscience, vol. 45, no. 1, pp. 151–175, 2022

2022
[14]

Quantitative study of granule and Purkinje cells in the cerebellar cortex of the rat,

R. J. Harvey and R. M. Napper, “Quantitative study of granule and Purkinje cells in the cerebellar cortex of the rat,”J Comp Neurol, vol. 274, pp. 151–157, 1988

1988
[15]

Quantitative studies on the mammalian cerebellum,

R. J. Harvey and R. M. Napper, “Quantitative studies on the mammalian cerebellum,”Prog Neurobiol, vol. 36, pp. 437–463, 1991

1991
[16]

The entire trajectories of single olivocerebellar axons in the cerebellar cortex and their contribution to cerebellar compartmentalization,

I. Sugihara, H.-S. Wu, and Y. Shinoda, “The entire trajectories of single olivocerebellar axons in the cerebellar cortex and their contribution to cerebellar compartmentalization,”Journal of Neuroscience, vol. 21, no. 19, pp. 7715–7723, 2001

2001
[17]

J. C. Eccles, M. Ito, and J. Szent´ agothai,The cerebellum as a neuronal machine. Springer-Verlag, 1967

1967
[18]

A theory of cerebellar cortex,

D. Marr, “A theory of cerebellar cortex,”J. Physiol., vol. 202, pp. 437–470, 1969

1969
[19]

A theory of cerebellar function,

J. S. Albus, “A theory of cerebellar function,”Mathematical Biosciences, vol. 10, pp. 26–51, 1971

1971
[20]

Climbing fibre induced depression of both mossy fibre responsiveness and glutamate sensitivity of cerebellar Purkinje cells,

M. Ito, M. Sakurai, and P. Tongroach, “Climbing fibre induced depression of both mossy fibre responsiveness and glutamate sensitivity of cerebellar Purkinje cells,”J. Physiol., vol. 324, pp. 113–134, 1982

1982
[21]

Purkinje-cell plasticity and cerebellar motor learning are graded by complex-spike duration,

Y. Yang and S. G. Lisberger, “Purkinje-cell plasticity and cerebellar motor learning are graded by complex-spike duration,” Nature, vol. 510, pp. 529–532, 2014

2014
[22]

The nature of reinforcement in cerebellar learning,

R. F. Thompson, J. K. Thompson, J. J. Kim, D. J. Krupa, and P. G. Shinkman, “The nature of reinforcement in cerebellar learning,”Neurobiology of learning and memory, vol. 70, no. 1-2, pp. 150–176, 1998

1998
[23]

Neural population dynamics during reaching,

M. M. Churchland, J. P. Cunningham, M. T. Kaufman, J. D. Foster, P. Nuyujukian, S. I. Ryu, and K. V. Shenoy, “Neural population dynamics during reaching,”Nature, vol. 487, no. 7405, pp. 51–56, 2012

2012
[24]

On simplicity and complexity in the brave new world of large-scale neuroscience,

P. Gao and S. Ganguli, “On simplicity and complexity in the brave new world of large-scale neuroscience,”Curr. Opin. Neurobiol., vol. 32, pp. 148–155, 2015

2015
[25]

Neural manifolds for the control of movement,

J. A. Gallego, M. G. Perich, L. E. Miller, and S. A. Solla, “Neural manifolds for the control of movement,”Neuron, vol. 94, no. 5, pp. 978–984, 2017

2017
[26]

Neural population geometry: An approach for understanding biological and artificial neural networks,

S. Chung and L. F. Abbott, “Neural population geometry: An approach for understanding biological and artificial neural networks,”Curr. Opin. Neurobiol., vol. 70, pp. 137–144, 2021

2021
[27]

Geometrical and statistical properties of systems of linear inequalities with applications in pattern recogni- tion.,

T. M. Cover, “Geometrical and statistical properties of systems of linear inequalities with applications in pattern recogni- tion.,”IEEE Trans, vol. EC-14, p. 326, 1965

1965
[28]

Sch¨ olkopf and A

B. Sch¨ olkopf and A. J. Smola,Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond. The MIT Press, 12 2001

2001
[29]

Attention is all you need,

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin, “Attention is all you need,” inAdvances in Neural Information Processing Systems(I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett,...

2017
[30]

Linking connectivity, dynamics, and computations in low-rank recurrent neural net- works,

F. Mastrogiuseppe and S. Ostojic, “Linking connectivity, dynamics, and computations in low-rank recurrent neural net- works,”Neuron, vol. 99, no. 3, pp. 609–623, 2018

2018
[31]

Opening the black box: low-dimensional dynamics in high-dimensional recurrent neural net- works,

D. Sussillo and O. Barak, “Opening the black box: low-dimensional dynamics in high-dimensional recurrent neural net- works,”Neural computation, vol. 25, no. 3, pp. 626–649, 2013

2013
[32]

Re-evaluating circuit mechanisms underlying pattern separation,

N. A. Cayco-Gajic and R. A. Silver, “Re-evaluating circuit mechanisms underlying pattern separation,”Neuron, vol. 101, no. 4, pp. 584–602, 2019

2019
[33]

Structure within the cerebellar input layer enables lossless sparse encoding,

G. Billings, E. Piasini, A. L ˚A‘rincz, Z. Nusser, and R. Silver, “Structure within the cerebellar input layer enables lossless sparse encoding,”Network, vol. 83, pp. 960–74, 2014

2014
[34]

Sparseness and expansion in sensory representations,

B. Babadi and H. Sompolinsky, “Sparseness and expansion in sensory representations,”Neuron, vol. 83, no. 5, pp. 1213– 1226, 2014

2014
[35]

Optimal degrees of synaptic connectivity,

A. Litwin-Kumar, K. D. Harris, R. Axel, H. Sompolinsky, and L. Abbott, “Optimal degrees of synaptic connectivity,” Neuron, vol. 93, no. 5, pp. 1153–1164, 2017

2017
[36]

Sparse synaptic connectivity is required for decorrelation and pattern separation in feedforward networks,

N. A. Cayco-Gajic, C. Clopath, and R. A. Silver, “Sparse synaptic connectivity is required for decorrelation and pattern separation in feedforward networks,”Nature communications, vol. 8, no. 1, p. 1116, 2017

2017
[37]

Cerebellar granule cells acquire a widespread predictive feedback signal during motor learning,

A. Giovannucci, A. Badura, B. Deverett, F. Najafi, T. D. Pereira, Z. Gao, I. Ozden, A. D. Kloth, E. Pnevmatikakis, L. Paninski,et al., “Cerebellar granule cells acquire a widespread predictive feedback signal during motor learning,” 11 Nature neuroscience, vol. 20, no. 5, pp. ...

2017
[38]

Cerebellar granule cells encode the expectation of reward,

M. J. Wagner, T. H. Kim, J. Savall, M. J. Schnitzer, and L. Luo, “Cerebellar granule cells encode the expectation of reward,”Nature, vol. 544, no. 7648, pp. 96–100, 2017

2017
[39]

Sensorimotor representations in cerebellar granule cells in larval zebrafish are dense, spatially organized, and non-temporally patterned,

L. D. Knogler, D. A. Markov, E. I. Dragomir, V. ˇStih, and R. Portugues, “Sensorimotor representations in cerebellar granule cells in larval zebrafish are dense, spatially organized, and non-temporally patterned,”Current Biology, vol. 27, no. 9, pp. 1288–1302, 2017

2017
[40]

Cerebellar granule cell axons support high- dimensional representations,

F. Lanore, N. A. Cayco-Gajic, H. Gurnani, D. Coyle, and R. A. Silver, “Cerebellar granule cell axons support high- dimensional representations,”Nature neuroscience, vol. 24, no. 8, pp. 1142–1150, 2021

2021
[41]

Local synaptic inhibition mediates cerebellar granule cell pattern separation and enables learned sensorimotor associations,

E. A. Fleming, G. D. Field, M. R. Tadross, and C. Hull, “Local synaptic inhibition mediates cerebellar granule cell pattern separation and enables learned sensorimotor associations,”Nat. Neurosci., vol. 27, pp. 689–701, 2024

2024
[42]

Optimal information storage and the distribution of synaptic weights: perceptron versus Purkinje cell.,

N. Brunel, V. Hakim, P. Isope, J.-P. Nadal, and B. Barbour, “Optimal information storage and the distribution of synaptic weights: perceptron versus Purkinje cell.,”Neuron, vol. 43, pp. 745–57, 2004

2004
[43]

The perceptron: a probabilistic model for information storage and organization in the brain,

F. Rosenblatt, “The perceptron: a probabilistic model for information storage and organization in the brain,”Psych. Review, vol. 65, pp. 386–408, 1958

1958
[44]

Rosenblatt,Principles of neurodynamics

F. Rosenblatt,Principles of neurodynamics. Washington: Spartan books, 1962

1962
[45]

V. N. Vapnik,The Nature of Statistical Learning Theory. Springer-Verlag, 1999

1999
[46]

Non-holographic associative memory,

D. Willshaw, O. Buneman, and H. Longuet-Higgins, “Non-holographic associative memory,”Nature, vol. 222, p. 960ˆ a€“962, 1969

1969
[47]

The phase space of interactions in neural network models.,

E. Gardner, “The phase space of interactions in neural network models.,”J. Phys. A, vol. 21, pp. 257–270, 1988

1988
[48]

Optimal storage properties of neural network models,

E. Gardner and B. Derrida, “Optimal storage properties of neural network models,”Journal of Physics A: Mathematical and General, vol. 21, p. 271, jan 1988

1988
[49]

What can we learn from synaptic weight distributions?,

B. Barbour, N. Brunel, V. Hakim, and J.-P. Nadal, “What can we learn from synaptic weight distributions?,”Trends Neurosci., vol. 30, pp. 622–629, 2007

2007
[50]

Properties of unitary granule cell→Purkinje cell synapses in adult rat cerebellar slices.,

P. Isope and B. Barbour, “Properties of unitary granule cell→Purkinje cell synapses in adult rat cerebellar slices.,”J Neurosci, vol. 22, pp. 9668–78, 2002

2002
[51]

Storage of correlated patterns in standard and bistable Purkinje cell models,

C. Clopath, J.-P. Nadal, and N. Brunel, “Storage of correlated patterns in standard and bistable Purkinje cell models,” PLoS Comput. Biol., vol. 8, p. e1002448, 2012

2012
[52]

Optimal properties of analog perceptrons with excitatory weights,

C. Clopath and N. Brunel, “Optimal properties of analog perceptrons with excitatory weights,”PLoS Comput. Biol., vol. 9, p. e1002919, 2013

2013
[53]

Cerebellar ltd and pattern recognition by purkinje cells,

V. Steuber, W. Mittmann, F. E. Hoebeek, R. A. Silver, C. I. De Zeeuw, M. H ˜A¤usser, and E. De Schutter, “Cerebellar ltd and pattern recognition by purkinje cells,”Neuron, vol. 54, pp. 121–136, 2007

2007
[54]

Nonspecific synaptic plasticity improves the recognition of sparse patterns degraded by local noise,

K. Safaryan, R. Maex, N. Davey, R. Adams, and V. Steuber, “Nonspecific synaptic plasticity improves the recognition of sparse patterns degraded by local noise,”Sci Rep, vol. 7, p. 46550, 2017

2017
[55]

Mechanisms and functional roles of glutamatergic synapse diversity in a cerebellar circuit,

V. Zampini, J. K. Liu, M. A. Diana, P. P. Maldonado, N. Brunel, and S. Dieudonn ˜A©, “Mechanisms and functional roles of glutamatergic synapse diversity in a cerebellar circuit,”Elife, vol. 5, 2016

2016
[56]

Adaptive filter model of the cerebellum,

M. Fujita, “Adaptive filter model of the cerebellum,”Biol. Cybern., vol. 45, pp. 190–206, 1982

1982
[57]

Neural network model of the cerebellum: Temporal discrimination and the timing of motor responses,

D. V. Buonomano and M. Mauk, “Neural network model of the cerebellum: Temporal discrimination and the timing of motor responses,”Neural Computation, vol. 6, p. 38ˆ a€“55, 1994

1994
[58]

Timing mechanisms in the cerebellum: Testing predictions of a large-scale computer simulation,

J. F. Medina, K. S. Garcia, W. L. Nores, N. M. Taylor, and M. D. Mauk, “Timing mechanisms in the cerebellum: Testing predictions of a large-scale computer simulation,”Journal of Neuroscience, vol. 20, no. 14, pp. 5516–5525, 2000

2000
[59]

Cerebellar motor learning: When is cortical plasticity not enough?,

J. Porrill and P. Dean, “Cerebellar motor learning: When is cortical plasticity not enough?,”PLOS Comp. Biol., vol. 3, p. e197, 2007

2007
[60]

A temporal basis for predicting the sensory consequences of motor commands in an electric fish,

A. Kennedy, G. Wayne, P. Kaifosh, K. Alvi˜ na, L. F. Abbott, and N. B. Sawtell, “A temporal basis for predicting the sensory consequences of motor commands in an electric fish,”Nat. Neurosci., vol. 17, pp. 416–422, 2014

2014
[61]

Graded heterogeneity of metabotropic signaling underlies a continuum of cell-intrinsic temporal responses in unipolar brush cells,

C. Guo, V. Huson, E. Z. Macosko, and W. G. Regehr, “Graded heterogeneity of metabotropic signaling underlies a continuum of cell-intrinsic temporal responses in unipolar brush cells,”Nat. Commun., vol. 12, p. 5491, 2021

2021
[62]

Synaptic diversity enables temporal coding of coincident multisensory inputs in single neurons,

F. P. Chabrol, A. Arenz, M. T. Wiechert, T. W. Margrie, and D. A. DiGregorio, “Synaptic diversity enables temporal coding of coincident multisensory inputs in single neurons,”Nat. Neurosci., vol. 18, pp. 718–727, 2015

2015
[63]

Synaptic basis of a sub-second representation of time in a neural circuit model,

A. Barri, M. T. Wiechert, M. Jazayeri, and D. A. DiGregorio, “Synaptic basis of a sub-second representation of time in a neural circuit model,”Nat. Commun., vol. 13, p. 7902, 2022

2022
[64]

The cerebellar microcircuit as an adaptive filter: experimental and computational evidence,

P. Dean, J. Porrill, C.-F. Ekerot, and H. J¨ orntell, “The cerebellar microcircuit as an adaptive filter: experimental and computational evidence,”Nat. Rev. Neurosci., vol. 11, pp. 30–43, 2010

2010
[65]

At the edge of chaos: How cerebellar granular layer network dynamics can provide the basis for temporal filters,

C. R¨ ossert, P. Dean, and J. Porrill, “At the edge of chaos: How cerebellar granular layer network dynamics can provide the basis for temporal filters,”PLoS Comput. Biol., vol. 11, p. e1004515, 2015

2015
[66]

Metabotropic glutamate receptor activation in cerebellar purkinje cells as substrate for adaptive timing of the classically conditioned eye-blink response,

J. C. Fiala, S. Grossberg, and D. Bullock, “Metabotropic glutamate receptor activation in cerebellar purkinje cells as substrate for adaptive timing of the classically conditioned eye-blink response,”Journal of Neuroscience, vol. 16, no. 11, pp. 3760–3774, 1996

1996
[67]

A biophysical model of synaptic delay learning and temporal pattern recognition in a cerebellar purkinje cell,

V. Steuber and D. Willshaw, “A biophysical model of synaptic delay learning and temporal pattern recognition in a cerebellar purkinje cell,”J Comput Neurosci, vol. 17, pp. 149–64, 2004

2004
[68]

Cerebellar learning using perturbations,

G. Bouvier, J. Aljadeff, C. Clopath, C. Bimbard, J. Ranft, A. Blot, J.-P. Nadal, N. Brunel, V. Hakim, and B. Barbour, “Cerebellar learning using perturbations,”eLife, vol. 7, p. e31599, 2018

2018
[69]

Encoding of error and learning to correct that error by the purkinje cells of the cerebellum,

D. J. Herzfeld, Y. Kojima, R. Soetedjo, and R. Shadmehr, “Encoding of error and learning to correct that error by the purkinje cells of the cerebellum,”Nat. Neurosci., vol. 21, pp. 736–743, 2018

2018
[70]

Prediction signals in the cerebellum: beyond supervised motor learning,

C. Hull, “Prediction signals in the cerebellum: beyond supervised motor learning,”Elife, vol. 9, p. e54073, 2020. 12

2020
[71]

Predictive and reactive reward signals conveyed by climbing fiber inputs to cerebellar purkinje cells,

D. Kostadinov, M. Beau, M. Blanco-Pozo, and M. H ˜A¤usser, “Predictive and reactive reward signals conveyed by climbing fiber inputs to cerebellar purkinje cells,”Nat Neurosci, vol. 22, pp. 950–962, 2019

2019
[72]

Cerebellar climbing fibers multiplex movement and reward signals during a voluntary movement task in mice,

K. Ikezoe, N. Hidaka, S. Manita, M. Murakami, S. Tsutsumi, Y. Isomura, M. Kano, and K. Kitamura, “Cerebellar climbing fibers multiplex movement and reward signals during a voluntary movement task in mice,”Comm Biol, vol. 6, p. 924, 2023

2023
[73]

Adaptive gain control of vestibuloocular reflex by the cerebellum,

D. A. Robinson, “Adaptive gain control of vestibuloocular reflex by the cerebellum,”Journal of Neurophysiology, vol. 39, no. 5, pp. 954–969, 1976

1976
[74]

Visual influence on rabbit horizontal vestibulo-ocular reflex presumably effected via the cerebellar flocculus,

M. Ito, T. Shiida, N. Yagi, and M. Yamamoto, “Visual influence on rabbit horizontal vestibulo-ocular reflex presumably effected via the cerebellar flocculus,”Brain Research, vol. 65, no. 1, pp. 170–174, 1974

1974
[75]

The vestibulo-ocular reflex as a model system for motor learning: what is the role of the cerebellum?,

P. Blazquez, Y. Hirata, and S. Highstein, “The vestibulo-ocular reflex as a model system for motor learning: what is the role of the cerebellum?,”Cerebellum, vol. 3, pp. 188–92, 2004

2004
[76]

Cerebellar-dependent adaptive control of primate saccadic system,

L. M. Optican and D. A. Robinson, “Cerebellar-dependent adaptive control of primate saccadic system,”Journal of Neurophysiology, vol. 44, no. 6, pp. 1058–1076, 1980

1980
[77]

Chapter 6 - cerebellum-dependent motor learning: Lessons from adaptation of eye movements in primates,

S. Dash and P. Thier, “Chapter 6 - cerebellum-dependent motor learning: Lessons from adaptation of eye movements in primates,” inCerebellar Learning(N. Ramnani, ed.), vol. 210 ofProgress in Brain Research, pp. 121–155, Elsevier, 2014

2014
[78]

Complex spike activity in the oculomotor vermis of the cerebellum: A vectorial error signal for saccade motor learning?,

R. Soetedjo, Y. Kojima, and A. F. Fuchs, “Complex spike activity in the oculomotor vermis of the cerebellum: A vectorial error signal for saccade motor learning?,”Journal of Neurophysiology, vol. 100, no. 4, pp. 1949–1966, 2008

1949
[79]

Simple statistical gradient-following algorithms for connectionist reinforcement learning,

R. J. Williams, “Simple statistical gradient-following algorithms for connectionist reinforcement learning,”Machine learn- ing, vol. 8, pp. 229–256, 1992

1992
[80]

Learning in neural networks by reinforcement of irregular spiking,

X. Xie and H. S. Seung, “Learning in neural networks by reinforcement of irregular spiking,”Physical Review Eˆ a€”Statistical, Nonlinear, and Soft Matter Physics, vol. 69, no. 4, p. 041909, 2004

2004
[81]

A cerebellar learning model of vestibulo-ocular reflex adaptation in wild-type and mutant mice,

C. Clopath, A. Badura, C. I. De Zeeuw, and N. Brunel, “A cerebellar learning model of vestibulo-ocular reflex adaptation in wild-type and mutant mice,”Journal of Neuroscience, vol. 34, no. 21, pp. 7203–7215, 2014

2014
[82]

Beyond parallel fiber ltd: the diversity of synaptic and non-synaptic plasticity in the cerebellum,

C. Hansel, D. J. Linden, and E. D’Angelo, “Beyond parallel fiber ltd: the diversity of synaptic and non-synaptic plasticity in the cerebellum,”Nature neuroscience, vol. 4, no. 5, pp. 467–475, 2001

2001
[83]

Distributed synergistic plasticity and cerebellar learning,

Z. Gao and C. van Beugen, B.and De Zeeuw, “Distributed synergistic plasticity and cerebellar learning,”Nat Rev Neurosci, vol. 13, p. 619ˆ a€“635, 2012

2012
[84]

A transcriptomic atlas of the mouse cerebellum reveals regional specializations and novel cell types,

V. Kozareva, C. Martin, T. Osorno, S. Rudolph, C. Guo, C. Vanderburg, N. Nadaf, A. Regev, W. Regehr, and E. Macosko, “A transcriptomic atlas of the mouse cerebellum reveals regional specializations and novel cell types,”bioRxiv, 2020

2020
[85]

Structured cerebellar connectivity supports resilient pattern separation,

T. M. Nguyen, L. A. Thomas, J. L. Rhoades, I. Ricchi, X. C. Yuan, A. Sheridan, D. G. Hildebrand, J. Funke, W. G. Regehr, and W.-C. A. Lee, “Structured cerebellar connectivity supports resilient pattern separation,”Nature, vol. 613, no. 7944, pp. 543–549, 2023

2023
[86]

A deep-learning strategy to identify cell types across species from high-density extra- cellular recordings,

M. Beau, D. J. Herzfeld, F. Naveros, M. E. Hemelt, F. Dˆ a€™Agostino, M. Oostland, A. S´ anchez-L´ opez, Y. Y. Chung, M. Maibach, S. Kyranakis,et al., “A deep-learning strategy to identify cell types across species from high-density extra- cellular recordings,”bioRxiv, 2024

2024
[87]

Control of mental activities by internal models in the cerebellum,

M. Ito, “Control of mental activities by internal models in the cerebellum,”Nature Reviews Neuroscience, vol. 9, no. 4, pp. 304–313, 2008

2008
[88]

Cerebellum and Learning

N. A. Cayco GajicCommunication at Les Treilles workshop “Cerebellum and Learning”’, 2024

2024
[89]

Discharge frequencies in the cerebral and cerebellar cortex.,

E. D. Adrian, “Discharge frequencies in the cerebral and cerebellar cortex.,”Proc Phys Soc, vol. 83, pp. 32–33, 1935

1935
[90]

High- frequency organization and synchrony of activity in the purkinje cell layer of the cerebellum,

C. de Solages, G. Szapiro, N. Brunel, V. Hakim, P. Isope, P. Buisseret, C. Rousseau, B. Barbour, and C. Lena, “High- frequency organization and synchrony of activity in the purkinje cell layer of the cerebellum,”Neuron, vol. 58, pp. 775–788, 2008

2008
[91]

Electrical coupling mediates tunable low-frequency oscillations and resonance in the cerebellar Golgi cell network,

G. P. Dugue, N. Brunel, V. Hakim, E. Schwartz, M. Chat, M. Levesque, R. Courtemanche, C. Lena, and S. Dieudonne, “Electrical coupling mediates tunable low-frequency oscillations and resonance in the cerebellar Golgi cell network,”Neuron, vol. 61, pp. 126–139, 2009

2009
[92]

Backpropagation and the brain,

T. P. Lillicrap, A. Santoro, L. Marris, C. J. Akerman, and G. Hinton, “Backpropagation and the brain,”Nature Reviews Neuroscience, vol. 21, no. 6, pp. 335–346, 2020

2020
[93]

Monoamine pathways to the cerebellum and cerebral cortex,

N. And ˜A©n, K. Fuxe, and U. Ungerstedt, “Monoamine pathways to the cerebellum and cerebral cortex,”Experientia, vol. 23, pp. 838–9, Oct 1967

1967
[94]

Serotonergic input into the cerebellar cortex modulates anxiety-like behavior,

P. W. Chin and G. J. Augustine, “Serotonergic input into the cerebellar cortex modulates anxiety-like behavior,”Journal of Neuroscience, vol. 45, no. 14, 2025

2025
[95]

Chapter 5 cholinergic innervation and receptors in the cerebellum,

D. Jaarsma, T. J. Ruigrok, R. Caff ˜A©, C. Cozzari, A. I. Levey, E. Mugnaini, and J. Voogd, “Chapter 5 cholinergic innervation and receptors in the cerebellum,” inThe Cerebellum: From Structure to Control(C. De Zeeuw, P. Strata, and J. Voogd, eds.), vol. 114 ofProgress in Brai...

1997
[96]

Acetylcholine modulates cerebellar granule cell spiking by regulating the balance of synaptic excitation and inhibition,

T. R. Fore, B. N. Taylor, N. Brunel, and C. Hull, “Acetylcholine modulates cerebellar granule cell spiking by regulating the balance of synaptic excitation and inhibition,”Journal of Neuroscience, vol. 40, no. 14, pp. 2882–2894, 2020

2020
[97]

The cerebellum also receives monoaminergic [93, 94] and cholinergic [95, 96] projections

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Solvable model of a spin-glass,

D. Sherrington and S. Kirkpatrick, “Solvable model of a spin-glass,”Physical review letters, vol. 35, no. 26, p. 1792, 1975

1975

[2] [2]

Coexistence of spin-glass and ferromagnetic orderings,

M. Gabay and G. Toulouse, “Coexistence of spin-glass and ferromagnetic orderings,”Physical Review Letters, vol. 47, no. 3, p. 201, 1981. 10

1981

[3] [3]

Theory of the frustration effect in spin glasses: I,

G. Toulouse, “Theory of the frustration effect in spin glasses: I,”Comm. Phys., vol. 2, pp. 115–19, 1977

1977

[4] [4]

Mezard, G

M. Mezard, G. Parisi, and M. Virasoro,Spin Glass Theory and Beyond. WORLD SCIENTIFIC, 1986

1986

[5] [5]

Neural networks and physical systems with emergent collective computational abilities.,

J. J. Hopfield, “Neural networks and physical systems with emergent collective computational abilities.,”Proceedings of the national academy of sciences, vol. 79, no. 8, pp. 2554–2558, 1982

1982

[6] [6]

Spin glass model of learning by selection.,

G. Toulouse, S. Dehaene, and J.-P. Changeux, “Spin glass model of learning by selection.,”Proceedings of the National Academy of Sciences, vol. 83, no. 6, pp. 1695–1698, 1986

1986

[7] [7]

Networks of formal neurons and memory palimpsests,

J.-P. Nadal, G. Toulouse, J.-P. Changeux, and S. Dehaene, “Networks of formal neurons and memory palimpsests,” Europhysics Letters, vol. 1, no. 10, p. 535, 1986

1986

[8] [8]

Information storage in sparsely coded memory nets,

J.-P. Nadal and G. Toulouse, “Information storage in sparsely coded memory nets,”Network: Computation in Neural Systems, vol. 1, no. 1, pp. 61–74, 1990

1990