Training Ecosystems: A Computational Approach to Uncovering Learning Behavior in Unconventional Contexts

Adrita Samanta; Hananel Hazan; Michael Levin

arxiv: 2605.30109 · v1 · pith:B7FNRRABnew · submitted 2026-05-28 · 🧬 q-bio.PE · q-bio.QM

Training Ecosystems: A Computational Approach to Uncovering Learning Behavior in Unconventional Contexts

Adrita Samanta , Hananel Hazan , Michael Levin This is my paper

Pith reviewed 2026-06-28 23:34 UTC · model grok-4.3

classification 🧬 q-bio.PE q-bio.QM

keywords predator-prey modelhabituationsensitizationrecovery timeecological dynamicsbasal cognitionscale invariancelearning without memory

0 comments

The pith

Simple predator-prey dynamics suffice to produce habituation, sensitization, and discrete number learning in recovery times.

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

The paper establishes that a minimal in silico predator-prey model generates multiple learning behaviors purely through its dynamical interactions, with no added memory mechanism. Across more than 220,000 parameter combinations, recovery time after repeated stimuli displays habituation, sensitization, and scale-invariant discrete number learning, while response magnitude shows a strong opposing pattern in nearly all cases. These capacities cluster by interaction strengths, enabling accurate prediction of which parameter sets produce learning. The results indicate that such learning capacities emerge at the ecosystem level from the equations themselves. This points to a substrate for learning that depends on mathematics rather than specific biological hardware.

Core claim

In a classic predator-prey model, recovery time after perturbations exhibits habituation, sensitization, and a form of discrete number learning in a scale-invariant manner. Response magnitude and recovery time display a pronounced asymmetry, with 90.6 percent of parameter sets showing recovery-time sensitization paired with magnitude habituation. Incidence of these capacities is governed primarily by ecological interaction strengths and allows high-accuracy prediction for unseen parameter combinations.

What carries the argument

The recovery-time metric applied to a perturbed Lotka-Volterra predator-prey system, which tracks changes in both response magnitude and return-to-equilibrium duration as assays of learning.

If this is right

Habituation and sensitization of recovery time persist under stochastic perturbations, while discrete number learning is disrupted even by low noise.
Learning incidence is determined mainly by interaction strengths and can be predicted with high accuracy from parameter space clustering.
The observed asymmetry between response magnitude and recovery time holds across the large majority of parameter combinations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Analogous learning behaviors could appear in any system governed by similar coupled differential equations, including chemical networks or certain economic models.
Field measurements of recovery times in actual ecosystems could test whether natural interaction strengths fall into the learning-prone regions identified here.
Designers of synthetic or engineered systems might exploit these pre-existing dynamical propensities rather than adding separate memory components.

Load-bearing premise

The chosen perturbations and recovery-time metric constitute valid assays of learning that generalize beyond the specific predator-prey equations and the sampled parameter combinations.

What would settle it

Direct observation that recovery times in a real predator-prey population fail to exhibit habituation, sensitization, or number learning after repeated stimuli of the same form would falsify the claim.

read the original abstract

Recent progress in diverse intelligence has shown simple learning capacities below the organism level - single cells and even molecular networks. However, there are still many knowledge gaps around learning capacity above the organism level, and about memory implemented purely by dynamical interactions without explicit memory media. We demonstrate that minimal ecological dynamics (in silico) are sufficient for several kinds of learning, assayed as changes in both, magnitude of response, and of recovery time. Systematic exploration of over 220,000 parameter combinations in a simulated classic predator-prey model revealed that, when perturbed by stimuli, recovery time exhibits habituation, sensitization, and a form of discrete number learning in a scale-invariant manner. Robustness analysis revealed that habituation and sensitization persist under stochastic perturbations, while discrete number learning is disrupted even at low noise levels. Dimensionality reduction revealed that the incidence of learning capacity is primarily determined by ecological interaction strengths. Clear, unique clustering patterns in parameter space allow high prediction accuracy for novel parameter combinations that enable learning. Response magnitude revealed a striking asymmetry: 90.6% of parameter combinations exhibited recovery time sensitization paired with habituation of response magnitude, while the opposite pattern was extremely rare. These findings highlight a set of phenomena at the intersection of ecology, basal cognition, and mathematics with many implications for a wide range of systems describable by similar kinds of equations. These properties provide numerous efforts in biology and engineering with a substrate that has considerable, pre-patterned, propensity for learning, which ultimately arises from mathematics, not depending on the details of physics or biology.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Predator-prey simulations produce habituation-like patterns in recovery times across many parameters, but the claim that this counts as learning rests on unvalidated metrics.

read the letter

The core observation is that a standard predator-prey ODE, when repeatedly perturbed, yields changes in response magnitude and recovery time that match the signatures of habituation, sensitization, and scale-invariant discrete number learning in a large fraction of the 220k parameter sets explored. Interaction strengths drive most of the variation, and the asymmetry (recovery sensitization paired with magnitude habituation in over 90% of cases) is consistent enough to allow decent out-of-sample prediction after dimensionality reduction.

The work is mainly a broad forward simulation sweep on an established model. It does a solid job mapping where these patterns appear and showing that stochastic noise preserves the first two behaviors but breaks the number-learning one. The clustering result is reproducible from the reported protocol and gives a concrete handle on which ecological regimes support the effect.

The soft spot is the interpretation step. Recovery time is a direct readout of the system's linearization or return map, so altered relaxation after repeated stimuli can arise from any shift near a bifurcation or from simple accumulation of state changes; the paper does not supply an external benchmark (such as the Groves-Thompson criteria or controls that separate true decrement from drift) to show these patterns are more than generic relaxation. Without that, the learning label stays provisional.

This is for readers already working on dynamical substrates for basal cognition or on parameter landscapes in Lotka-Volterra-type systems. The simulation volume and the clean parameter clustering make it worth a referee's time to check the exact perturbation protocol, recovery definition, and robustness tests, even if the learning framing will need tightening.

Referee Report

2 major / 1 minor

Summary. The manuscript claims that simulations of a classic predator-prey model across more than 220,000 parameter combinations demonstrate learning-like behaviors: recovery times after repeated perturbations exhibit habituation, sensitization, and a form of discrete number learning in a scale-invariant manner, while response magnitude shows a pronounced asymmetry (90.6% of cases pair magnitude habituation with recovery-time sensitization). Incidence of these capacities is determined primarily by interaction strengths, with clear parameter-space clustering enabling high-accuracy prediction for novel combinations; habituation and sensitization persist under stochastic noise but discrete number learning does not.

Significance. If the recovery-time and magnitude metrics are shown to assay genuine learning rather than generic relaxation, the result would establish that standard ecological ODEs inherently support non-associative and number-like behaviors arising from dynamics alone. This would link population ecology to basal cognition research and supply a mathematically pre-patterned substrate for engineering applications. The scale of the parameter sweep and the reported asymmetry are notable strengths if the interpretation is substantiated.

major comments (2)

[Abstract] Abstract: the central claim that observed changes in recovery time constitute 'habituation, sensitization, and a form of discrete number learning' is load-bearing yet rests on an unvalidated metric; no comparison is supplied to literature criteria (e.g., Groves & Thompson 1970) or controls that would distinguish true response decrement from cumulative drift or eigenvalue-driven relaxation near bifurcations.
[Methods] The manuscript does not detail the precise definition of recovery time, the perturbation protocol, or the quantitative criterion for 'discrete number learning'; without these, the 220,000-run exploration and the robustness claims cannot be independently assessed or reproduced.

minor comments (1)

[Abstract] Abstract contains a minor grammatical issue ('changes in both, magnitude of response, and of recovery time').

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The two major comments correctly identify areas where additional rigor and transparency are required. We have prepared a revised manuscript that directly addresses both points by expanding the Methods section and adding explicit validation against established criteria. Below we respond point by point.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim that observed changes in recovery time constitute 'habituation, sensitization, and a form of discrete number learning' is load-bearing yet rests on an unvalidated metric; no comparison is supplied to literature criteria (e.g., Groves & Thompson 1970) or controls that would distinguish true response decrement from cumulative drift or eigenvalue-driven relaxation near bifurcations.

Authors: We accept that the original manuscript did not supply an explicit mapping to canonical criteria. In the revision we add a new subsection (Methods, §2.4) that (i) quotes the Groves & Thompson (1970) operational definitions of habituation and sensitization, (ii) states the precise quantitative thresholds we apply to recovery-time trajectories, and (iii) reports a control analysis restricted to parameter sets whose dominant eigenvalues lie at least 0.2 units inside the stable half-plane. The same subsection also includes a supplementary figure showing that the reported patterns remain statistically significant after removal of any trajectories whose relaxation time-scale is within 10 % of the unperturbed eigenvalue. These additions make the metric validation explicit without altering the core empirical findings. revision: yes
Referee: [Methods] The manuscript does not detail the precise definition of recovery time, the perturbation protocol, or the quantitative criterion for 'discrete number learning'; without these, the 220,000-run exploration and the robustness claims cannot be independently assessed or reproduced.

Authors: We agree that the original text omitted these operational details. The revised Methods section now contains: (1) the exact recovery-time definition (time to re-enter and remain within 5 % of the pre-perturbation equilibrium for at least three consecutive integration steps); (2) the full perturbation protocol (additive rectangular pulses of amplitude 0.1–2.0 applied to the prey equation for 0.5 time units, with inter-stimulus intervals of 5–50 time units); and (3) the discrete-number-learning criterion (two-sided Wilcoxon rank-sum test on recovery times after n versus n+1 stimuli, with Bonferroni-corrected p < 0.01 and effect size > 0.3). All 220 000 parameter combinations and the stochastic-noise robustness tests can now be reproduced from the updated supplementary code repository. revision: yes

Circularity Check

0 steps flagged

No significant circularity; behaviors extracted from forward simulation of established model

full rationale

The paper conducts exhaustive forward integration of a classic predator-prey ODE system across 220k parameter sets and directly measures response magnitude and recovery time from the resulting trajectories. No learning metric is fitted to data, no parameter is defined in terms of the target behavior, and no uniqueness theorem or ansatz is imported via self-citation. The reported clustering and cross-prediction on held-out parameter combinations is a standard post-hoc classifier trained on the simulation outputs themselves; it does not reduce the core claim (that the ODE dynamics produce the observed patterns) to a tautology. The derivation chain therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the predator-prey ODEs being an adequate minimal model, on the chosen stimulus protocol constituting a valid learning assay, and on the sampled parameter space being representative; no new entities are postulated.

free parameters (1)

interaction strengths and other model parameters
220,000 combinations explored; specific values not given but treated as free variables whose ranges determine learning incidence.

axioms (2)

domain assumption The classic predator-prey equations capture the relevant dynamical interactions for the learning assays.
Invoked by the choice of model and the claim that minimal ecological dynamics suffice.
domain assumption Recovery time after perturbation is a meaningful proxy for learning capacity.
Central to the assay definitions of habituation, sensitization, and number learning.

pith-pipeline@v0.9.1-grok · 5817 in / 1449 out tokens · 18970 ms · 2026-06-28T23:34:29.461447+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

62 extracted references · 3 canonical work pages · 1 internal anchor

[1]

Thompson, R. F. & Spencer, W. A. Habituation: A model phenomenon for the study of neuronal substrates of behavior. Psychol. Rev. 73, 16–43 (1966)

1966
[2]

Rankin, C. H. et al. Habituation revisited: an updated and revised description of the behavioral characteristics of habituation. Neu- robiol. Learn. Mem. 92, 135–138 (2009)

2009
[3]

& Ulijn, R

Kaygisiz, K. & Ulijn, R. V. Can Molecular Systems Learn? https://doi.org/10.1002/syst.202400075 doi:10.1002/syst.202400075

work page doi:10.1002/syst.202400075
[4]

Reber, A. S. & Baluška, F. Cognition in some surprising places. Biochem. Biophys. Res. Commun. 564, 150–157 (2021)

2021
[5]

& Levin, M

Baluška, F. & Levin, M. On Having No Head: Cognition throughout Biological Systems. Front. Psychol. 7, 902 (2016)

2016
[6]

The biogenic approach to cognition

Lyon, P. The biogenic approach to cognition. Cogn. Process. 7, 11–29 (2006)

2006
[7]

& Fontana, W

Katz, Y. & Fontana, W. Probabilistic Inference with Polymerizing Biochemical Circuits. Entropy 24, 629 (2022)

2022
[9]

Cragg, B. G. & Temperley, H. N. Memory: the analogy with ferromagnetic hysteresis. Brain J. Neurol. 78, 304–316 (1955)

1955
[10]

Keresztes, D. et al. Cancer drug resistance as learning of signaling networks. Biomed. Pharmacother. Biomedecine Pharmacother. 183, 117880 (2025)

2025
[11]

Csermely, P. et al. Learning of Signaling Networks: Molecular Mechanisms. Trends Biochem. Sci. 45, 284–294 (2020)

2020
[12]

& Jablonka, E

Ginsburg, S. & Jablonka, E. Epigenetic learning in non-neural organisms. J. Biosci. 34, 633–646 (2009). 25

2009
[13]

V., Carn ey, R

Kukushkin, N. V., Carn ey, R. E., Tabassum, T. & Carew, T. J. The massed -spaced learning effect in non- neural human cells. Nat. Commun. 15, 9635 (2024)

2024
[14]

& Gershman, S

Doan, N., Theroux, A., Ramdas, T. & Gershman, S. J. Associative learning in the protozoan Stentor coeruleus
[15]

Rajan, D. et al. Single-cell analysis of habituation in Stentor coeruleus. Curr. Biol. CB 33, 241-251.e4 (2023)

2023
[16]

& Levin, M

Fields, C. & Levin, M. Competency in Navigating Arbitrary Spaces as an Invariant for Analyzing Cognition in Diverse Embodimen ts. Entropy 24, 819 (2022)

2022
[17]

& Levin, M

Pezzulo, G. & Levin, M. Top-down models in biology: explanation and control of complex living systems above the molecular level. J. R. Soc. Interface 13, 20160555 (2016)

2016
[18]

The Computational Boundary of a “Self”: Developmental Bioelectricity Drives Multic ellularity and Scale -Free Cognition

Levin, M. The Computational Boundary of a “Self”: Developmental Bioelectricity Drives Multic ellularity and Scale -Free Cognition. Front. Psychol. 10, 2688 (2019)

2019
[19]

Pashine, N., Hexner, D., Liu, A. J. & Nagel, S. R. Directed aging, memory, and nature’s greed. Sci. Adv. 5, eaax4215 (2019)

2019
[20]

& Rubinstein, S

Lahini, Y., Gottesman, O., Amir, A. & Rubinstein, S. M. Nonmo notonic Aging and Memory Retention in Disordered Mechanical Sys- tems. Phys. Rev. Lett. 118, 085501 (2017)

2017
[21]

Baulin, V. A. et al. Intelligent soft matter: towards embodied intelligence. Soft Matter 21, 4129–4145 (2025)

2025
[22]

Loeffler, A. et al. Neuromorphic learning, working memory, and metaplasticity in nanowire networks. Sci. Adv. 9, eadg3289 (2023)

2023
[23]

J., van der Wiel, W

Kaspar, C., Ravoo, B. J., van der Wiel, W. G., Wegner, S. V. & Pernice, W. H. P. The rise of intelligent matter. Nature 594, 345–355 (2021)

2021
[24]

Scott, H. L. et al. Evidence for long-term potentiation in phospholipid membranes. Proc. Natl. Acad. Sci. U. S. A. 119, e2212195119 (2022)

2022
[25]

Abramson, C. I. & Levin, M. Behaviorist approaches to investigating memory and learning: A primer for synthetic biology and bioen- gineering. Commun. Integr. Biol. 14, 230–247 (2021)

2021
[26]

& Watson, R

Kouvaris, K., Clune, J., Kounios, L., Brede, M. & Watson, R. A. How evolution learns to generalise: Using the principles of l earning theory to understand the evolution of developmental organisation. PLoS Comput. Biol. 13, e1005358 (2017)

2017
[27]

Watson, R. A. & Szathmáry, E. How Can Evolution Learn? Trends Ecol. Evol. 31, 147–157 (2016)

2016
[28]

Power, D. A. et al. What can ecosystems learn? Expanding evolutionary ecology with learning theory. Biol. Direct 10, 69 (2015)

2015
[29]

Baldwin, J. M. A New Factor in Evolution. Am. Nat. 30, 441–451 (1896)
[30]

Hinton, G. E. & Nowlan, S. J. How Learning Can Guide Evolution. Complex Syst. 1,
[31]

Smith, J. M. When learning guides evolution. Nature 329, 761–762 (1987)

1987
[32]

Sridhar, V. H. et al. The geometry of decision-making in individuals and collectives. Proc. Natl. Acad. Sci. U. S. A. 118, e2102157118 (2021)

2021
[33]

R., Couzin, I

Strandburg-Peshkin, A., Farine, D. R., Couzin, I. D. & Crofoot, M. C. Shared decision -making drives collective movement in wild ba- boons. Science 348, 1358–1361 (2015)

2015
[34]

Couzin, I. D. Collective cognition in animal groups. Trends Cogn. Sci. 13, 36–43 (2009)

2009
[35]

Ward, A. J. W., Sumpter, D. J. T., Couzin, I. D., Hart, P. J. B. & Krause, J. Quorum decision -making facilitates information transfer in fish shoals. Proc. Natl. Acad. Sci. U. S. A. 105, 6948–6953 (2008)

2008
[36]

Bazazi, S. et al. Collective motion and cannibalism in locust migratory bands. Curr. Biol. CB 18, 735–739 (2008)

2008
[37]

Collective minds

Couzin, I. Collective minds. Nature 445, 715 (2007)

2007
[38]

D., Krause, J., James, R., Ruxton, G

Couzin, I. D., Krause, J., James, R., Ruxton, G. D. & Franks, N. R. Collective memory and spatial sorting in animal groups. J. Theor. Biol. 218, 1–11 (2002)

2002
[39]

Dreyer, T. et al. Comparing cooperative geometric puzzle solving in ants versus humans. Proc. Natl . Acad. Sci. 122, e2414274121 (2025)

2025
[40]

& Ikegami, T

Witkowski, O. & Ikegami, T. How to Make Swarms Open-Ended? Evolving Collective Intelligence Through a Constricted Exploration of Adjacent Possibles. Artif. Life 25, 178–197 (2019)

2019
[41]

& Solé, R

Piñero, J. & Solé, R. Statistical physics of liquid brains. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 374, 20180376 (2019)

2019
[42]

Turner, J. S. Termites as models of swarm cognition. Swarm Intell. 5, 19–43 (2011)

2011
[43]

Sumpter, D. J. T. The principles of collective animal behaviour. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 361, 5–22 (2006)

2006
[44]

Lotka, A. J. Elements of Physical Biology. Sci. Prog. Twent. Century 1919-1933 21, 341–343 (1926)

1919
[45]

Fluctuations in the Abundance of a Species considered Mathematically1

Volterra, V. Fluctuations in the Abundance of a Species considered Mathematically1. Nature 118, 558–560 (1926)

1926
[46]

& Levin, M

Biswas, S., Clawson, W. & Levin, M. Learning in Transcriptional Network Models: Computational Discovery of Pathway-Level Memory and Effective Interventions. Int. J. Mol. Sci. 24, 285 (2022)

2022
[47]

& Levin, M

Biswas, S., Manicka, S., Hoel, E. & Levin, M. Gene regulatory networ ks exhibit several kinds of memory: Quantification of memory in biological and random transcriptional networks. iScience 24, 102131 (2021)

2021
[48]

& Levin, M

Pigozzi, F., Goldstein, A. & Levin, M. Associative conditioning in gene regulatory network models increases integrative causal emer- gence. Commun. Biol. 8, 1027 (2025)

2025
[49]

& Solé, R

Pla-Mauri, J. & Solé, R. Engineering Basal Cognition: Minimal Genetic Circuits for Habituation, Sensitization, and Massed –Spaced Learning. ACS Synth. Biol. 15, 716–727
[50]

Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020)

2020
[51]

& Großberger, L

McInnes, L., Healy, J., Saul, N. & Großberger, L. UMAP: Uniform Manifold Approximation and Projection. J. Open Source Softw. 3, 861 (2018)

2018
[52]

Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. J Mach Learn Res 12, 2825–2830 (2011)

2011
[53]

Ester, M., Kriegel, H.-P., Sander, J. & Xu, X. A density-based algorithm for discovering clusters in large spatial databases with noise. in Proceedings of the Second Internat ional Conference on Knowledge Discovery and Data Mining 226–231 (AAAI Press, Portland, Oregon, 1996)

1996
[54]

& Nicholson, M

Elton, C. & Nicholson, M. The Ten-Year Cycle in Numbers of the Lynx in Canada. J. Anim. Ecol. 11, 215 (1942)

1942
[55]

2020 , edition =

McElreath, R. Statistical Rethinking: A Baye sian Course with Examples in R and STAN . (Chapman and Hall/CRC, New York, 2020). doi:10.1201/9780429029608

work page doi:10.1201/9780429029608 2020
[56]

Stein, R. R. et al. Ecological Modeling from Time-Series Inference: Insight into Dynamics and Stability of Intestinal Microbiota. PLOS Comput. Biol. 9, e1003388 (2013)8. Katz, Y., Springer, M. & Fontana, W. Embodying probabilistic inference in biochemical circuits. Preprint at https://doi.org/10.48550/arXiv.1806.10161 (2018)

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1806.10161 2013
[57]

Anthropic bound on the cosmological constant

Weinberg, S. Anthropic bound on the cosmological constant. Phys. Rev. Lett. 59, 2607–2610 (1987)

1987
[58]

Barrow, J. D. Cosmology, life, and the anthropic principle. Ann. N. Y. Acad. Sci. 950, 139–153 (2001)

2001
[59]

Teleological cosmology: the anthropic cosmological principle

Silk, J. Teleological cosmology: the anthropic cosmological principle. Science 232, 1036–1037 (1986)

1986
[60]

Large Number Coincidences and the Anthropic Principle in Cosmology

Carter, B. Large Number Coincidences and the Anthropic Principle in Cosmology. Symp. - Int. Astron. Union 63, 291–298 (1974)

1974
[61]

& Moors, A

De Houwer, J., Barnes -Holmes, D. & Moors, A. What is learning? On the nature and merits of a functional definition of learning. Psychon. Bull. Rev. 20, 631–642 (2013)

2013
[62]

Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds

Levin, M. Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds. Front. Syst. Neurosci. 16, 768201 (2022)

2022
[63]

& Levin, M

McMillen, P. & Levin, M. Collective intelligence: A unifying concept fo r integrating biology across scales and substrates. Commun. Biol. 7, 378 (2024)

2024

[1] [1]

Thompson, R. F. & Spencer, W. A. Habituation: A model phenomenon for the study of neuronal substrates of behavior. Psychol. Rev. 73, 16–43 (1966)

1966

[2] [2]

Rankin, C. H. et al. Habituation revisited: an updated and revised description of the behavioral characteristics of habituation. Neu- robiol. Learn. Mem. 92, 135–138 (2009)

2009

[3] [3]

& Ulijn, R

Kaygisiz, K. & Ulijn, R. V. Can Molecular Systems Learn? https://doi.org/10.1002/syst.202400075 doi:10.1002/syst.202400075

work page doi:10.1002/syst.202400075

[4] [4]

Reber, A. S. & Baluška, F. Cognition in some surprising places. Biochem. Biophys. Res. Commun. 564, 150–157 (2021)

2021

[5] [5]

& Levin, M

Baluška, F. & Levin, M. On Having No Head: Cognition throughout Biological Systems. Front. Psychol. 7, 902 (2016)

2016

[6] [6]

The biogenic approach to cognition

Lyon, P. The biogenic approach to cognition. Cogn. Process. 7, 11–29 (2006)

2006

[7] [7]

& Fontana, W

Katz, Y. & Fontana, W. Probabilistic Inference with Polymerizing Biochemical Circuits. Entropy 24, 629 (2022)

2022

[8] [9]

Cragg, B. G. & Temperley, H. N. Memory: the analogy with ferromagnetic hysteresis. Brain J. Neurol. 78, 304–316 (1955)

1955

[9] [10]

Keresztes, D. et al. Cancer drug resistance as learning of signaling networks. Biomed. Pharmacother. Biomedecine Pharmacother. 183, 117880 (2025)

2025

[10] [11]

Csermely, P. et al. Learning of Signaling Networks: Molecular Mechanisms. Trends Biochem. Sci. 45, 284–294 (2020)

2020

[11] [12]

& Jablonka, E

Ginsburg, S. & Jablonka, E. Epigenetic learning in non-neural organisms. J. Biosci. 34, 633–646 (2009). 25

2009

[12] [13]

V., Carn ey, R

Kukushkin, N. V., Carn ey, R. E., Tabassum, T. & Carew, T. J. The massed -spaced learning effect in non- neural human cells. Nat. Commun. 15, 9635 (2024)

2024

[13] [14]

& Gershman, S

Doan, N., Theroux, A., Ramdas, T. & Gershman, S. J. Associative learning in the protozoan Stentor coeruleus

[14] [15]

Rajan, D. et al. Single-cell analysis of habituation in Stentor coeruleus. Curr. Biol. CB 33, 241-251.e4 (2023)

2023

[15] [16]

& Levin, M

Fields, C. & Levin, M. Competency in Navigating Arbitrary Spaces as an Invariant for Analyzing Cognition in Diverse Embodimen ts. Entropy 24, 819 (2022)

2022

[16] [17]

& Levin, M

Pezzulo, G. & Levin, M. Top-down models in biology: explanation and control of complex living systems above the molecular level. J. R. Soc. Interface 13, 20160555 (2016)

2016

[17] [18]

The Computational Boundary of a “Self”: Developmental Bioelectricity Drives Multic ellularity and Scale -Free Cognition

Levin, M. The Computational Boundary of a “Self”: Developmental Bioelectricity Drives Multic ellularity and Scale -Free Cognition. Front. Psychol. 10, 2688 (2019)

2019

[18] [19]

Pashine, N., Hexner, D., Liu, A. J. & Nagel, S. R. Directed aging, memory, and nature’s greed. Sci. Adv. 5, eaax4215 (2019)

2019

[19] [20]

& Rubinstein, S

Lahini, Y., Gottesman, O., Amir, A. & Rubinstein, S. M. Nonmo notonic Aging and Memory Retention in Disordered Mechanical Sys- tems. Phys. Rev. Lett. 118, 085501 (2017)

2017

[20] [21]

Baulin, V. A. et al. Intelligent soft matter: towards embodied intelligence. Soft Matter 21, 4129–4145 (2025)

2025

[21] [22]

Loeffler, A. et al. Neuromorphic learning, working memory, and metaplasticity in nanowire networks. Sci. Adv. 9, eadg3289 (2023)

2023

[22] [23]

J., van der Wiel, W

Kaspar, C., Ravoo, B. J., van der Wiel, W. G., Wegner, S. V. & Pernice, W. H. P. The rise of intelligent matter. Nature 594, 345–355 (2021)

2021

[23] [24]

Scott, H. L. et al. Evidence for long-term potentiation in phospholipid membranes. Proc. Natl. Acad. Sci. U. S. A. 119, e2212195119 (2022)

2022

[24] [25]

Abramson, C. I. & Levin, M. Behaviorist approaches to investigating memory and learning: A primer for synthetic biology and bioen- gineering. Commun. Integr. Biol. 14, 230–247 (2021)

2021

[25] [26]

& Watson, R

Kouvaris, K., Clune, J., Kounios, L., Brede, M. & Watson, R. A. How evolution learns to generalise: Using the principles of l earning theory to understand the evolution of developmental organisation. PLoS Comput. Biol. 13, e1005358 (2017)

2017

[26] [27]

Watson, R. A. & Szathmáry, E. How Can Evolution Learn? Trends Ecol. Evol. 31, 147–157 (2016)

2016

[27] [28]

Power, D. A. et al. What can ecosystems learn? Expanding evolutionary ecology with learning theory. Biol. Direct 10, 69 (2015)

2015

[28] [29]

Baldwin, J. M. A New Factor in Evolution. Am. Nat. 30, 441–451 (1896)

[29] [30]

Hinton, G. E. & Nowlan, S. J. How Learning Can Guide Evolution. Complex Syst. 1,

[30] [31]

Smith, J. M. When learning guides evolution. Nature 329, 761–762 (1987)

1987

[31] [32]

Sridhar, V. H. et al. The geometry of decision-making in individuals and collectives. Proc. Natl. Acad. Sci. U. S. A. 118, e2102157118 (2021)

2021

[32] [33]

R., Couzin, I

Strandburg-Peshkin, A., Farine, D. R., Couzin, I. D. & Crofoot, M. C. Shared decision -making drives collective movement in wild ba- boons. Science 348, 1358–1361 (2015)

2015

[33] [34]

Couzin, I. D. Collective cognition in animal groups. Trends Cogn. Sci. 13, 36–43 (2009)

2009

[34] [35]

Ward, A. J. W., Sumpter, D. J. T., Couzin, I. D., Hart, P. J. B. & Krause, J. Quorum decision -making facilitates information transfer in fish shoals. Proc. Natl. Acad. Sci. U. S. A. 105, 6948–6953 (2008)

2008

[35] [36]

Bazazi, S. et al. Collective motion and cannibalism in locust migratory bands. Curr. Biol. CB 18, 735–739 (2008)

2008

[36] [37]

Collective minds

Couzin, I. Collective minds. Nature 445, 715 (2007)

2007

[37] [38]

D., Krause, J., James, R., Ruxton, G

Couzin, I. D., Krause, J., James, R., Ruxton, G. D. & Franks, N. R. Collective memory and spatial sorting in animal groups. J. Theor. Biol. 218, 1–11 (2002)

2002

[38] [39]

Dreyer, T. et al. Comparing cooperative geometric puzzle solving in ants versus humans. Proc. Natl . Acad. Sci. 122, e2414274121 (2025)

2025

[39] [40]

& Ikegami, T

Witkowski, O. & Ikegami, T. How to Make Swarms Open-Ended? Evolving Collective Intelligence Through a Constricted Exploration of Adjacent Possibles. Artif. Life 25, 178–197 (2019)

2019

[40] [41]

& Solé, R

Piñero, J. & Solé, R. Statistical physics of liquid brains. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 374, 20180376 (2019)

2019

[41] [42]

Turner, J. S. Termites as models of swarm cognition. Swarm Intell. 5, 19–43 (2011)

2011

[42] [43]

Sumpter, D. J. T. The principles of collective animal behaviour. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 361, 5–22 (2006)

2006

[43] [44]

Lotka, A. J. Elements of Physical Biology. Sci. Prog. Twent. Century 1919-1933 21, 341–343 (1926)

1919

[44] [45]

Fluctuations in the Abundance of a Species considered Mathematically1

Volterra, V. Fluctuations in the Abundance of a Species considered Mathematically1. Nature 118, 558–560 (1926)

1926

[45] [46]

& Levin, M

Biswas, S., Clawson, W. & Levin, M. Learning in Transcriptional Network Models: Computational Discovery of Pathway-Level Memory and Effective Interventions. Int. J. Mol. Sci. 24, 285 (2022)

2022

[46] [47]

& Levin, M

Biswas, S., Manicka, S., Hoel, E. & Levin, M. Gene regulatory networ ks exhibit several kinds of memory: Quantification of memory in biological and random transcriptional networks. iScience 24, 102131 (2021)

2021

[47] [48]

& Levin, M

Pigozzi, F., Goldstein, A. & Levin, M. Associative conditioning in gene regulatory network models increases integrative causal emer- gence. Commun. Biol. 8, 1027 (2025)

2025

[48] [49]

& Solé, R

Pla-Mauri, J. & Solé, R. Engineering Basal Cognition: Minimal Genetic Circuits for Habituation, Sensitization, and Massed –Spaced Learning. ACS Synth. Biol. 15, 716–727

[49] [50]

Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020)

2020

[50] [51]

& Großberger, L

McInnes, L., Healy, J., Saul, N. & Großberger, L. UMAP: Uniform Manifold Approximation and Projection. J. Open Source Softw. 3, 861 (2018)

2018

[51] [52]

Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. J Mach Learn Res 12, 2825–2830 (2011)

2011

[52] [53]

Ester, M., Kriegel, H.-P., Sander, J. & Xu, X. A density-based algorithm for discovering clusters in large spatial databases with noise. in Proceedings of the Second Internat ional Conference on Knowledge Discovery and Data Mining 226–231 (AAAI Press, Portland, Oregon, 1996)

1996

[53] [54]

& Nicholson, M

Elton, C. & Nicholson, M. The Ten-Year Cycle in Numbers of the Lynx in Canada. J. Anim. Ecol. 11, 215 (1942)

1942

[54] [55]

2020 , edition =

McElreath, R. Statistical Rethinking: A Baye sian Course with Examples in R and STAN . (Chapman and Hall/CRC, New York, 2020). doi:10.1201/9780429029608

work page doi:10.1201/9780429029608 2020

[55] [56]

Stein, R. R. et al. Ecological Modeling from Time-Series Inference: Insight into Dynamics and Stability of Intestinal Microbiota. PLOS Comput. Biol. 9, e1003388 (2013)8. Katz, Y., Springer, M. & Fontana, W. Embodying probabilistic inference in biochemical circuits. Preprint at https://doi.org/10.48550/arXiv.1806.10161 (2018)

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1806.10161 2013

[56] [57]

Anthropic bound on the cosmological constant

Weinberg, S. Anthropic bound on the cosmological constant. Phys. Rev. Lett. 59, 2607–2610 (1987)

1987

[57] [58]

Barrow, J. D. Cosmology, life, and the anthropic principle. Ann. N. Y. Acad. Sci. 950, 139–153 (2001)

2001

[58] [59]

Teleological cosmology: the anthropic cosmological principle

Silk, J. Teleological cosmology: the anthropic cosmological principle. Science 232, 1036–1037 (1986)

1986

[59] [60]

Large Number Coincidences and the Anthropic Principle in Cosmology

Carter, B. Large Number Coincidences and the Anthropic Principle in Cosmology. Symp. - Int. Astron. Union 63, 291–298 (1974)

1974

[60] [61]

& Moors, A

De Houwer, J., Barnes -Holmes, D. & Moors, A. What is learning? On the nature and merits of a functional definition of learning. Psychon. Bull. Rev. 20, 631–642 (2013)

2013

[61] [62]

Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds

Levin, M. Technological Approach to Mind Everywhere: An Experimentally-Grounded Framework for Understanding Diverse Bodies and Minds. Front. Syst. Neurosci. 16, 768201 (2022)

2022

[62] [63]

& Levin, M

McMillen, P. & Levin, M. Collective intelligence: A unifying concept fo r integrating biology across scales and substrates. Commun. Biol. 7, 378 (2024)

2024