Pith. sign in

REVIEW 3 major objections 4 minor 33 references

Dimension reshapes tissue mechanics into different diffusion laws

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · glm-5.2

2026-07-09 21:53 UTC pith:HIMSGL6Q

load-bearing objection Dimension-dependent continuum limits in tissue mechanics the 3 major comments →

arxiv 2607.07000 v1 pith:HIMSGL6Q submitted 2026-07-08 nlin.CG physics.bio-phphysics.comp-phq-bio.TO

Dimension-dependent continuum limits in tissue mechanics

Matthew J Simpson , Pascal R Buenzli This is my paper
classification nlin.CG physics.bio-phphysics.comp-phq-bio.TO
keywords mechanicstransportcontinuumdimensionalitylinkmacroscopicnonlinearphenomena
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper argues that when you coarse-grain a discrete model of interacting cells into a continuum transport equation, the resulting nonlinear diffusivity D(ρ) ∝ ρ^γ acquires an exponent γ that depends on spatial dimension d through the relation γ = (β − 2)/d, where β is set by the microscopic force law between cells. The same cell-level interaction, placed in one versus two versus three dimensions, produces qualitatively different macroscopic transport. For example, Hooke's-law springs between cells give γ = −2 in one dimension (strongly density-dependent, decreasing diffusivity) but γ ≈ −1 in two dimensions (weaker dependence), and the exponent trends toward zero—linear diffusion—as dimension grows. The authors establish this by a relaxation-rate matching procedure: they compute the slowest exponential decay rate of the discrete mechanical model near its uniform steady state, compute the corresponding decay rate of a continuum nonlinear diffusion PDE, and equate them to read off D(ρ). In one dimension this yields an exact closed-form result, D(ρ) = (k/η) a^(1−ν) ρ^(−(1+ν)), for a family of power-law force laws parameterized by ν. The dimensional scaling γ = (β − 2)/d then extends this to higher dimensions, and two-dimensional simulations confirm the predicted reduction in |γ|. The central claim is that dimensionality is not a geometric afterthought in tissue mechanics—it fundamentally alters the effective macroscopic law that emerges from a fixed microscopic interaction.

Core claim

The exponent γ in the emergent nonlinear diffusivity D(ρ) ∝ ρ^γ is not determined by the microscopic force law alone but is jointly fixed by the force law and the spatial dimension through γ = (β − 2)/d, where β encodes the microscopic force scaling. This means the same cell-interaction law produces different continuum transport laws in different dimensions, and linear diffusion becomes increasingly relevant as dimension grows. The one-dimensional case is solved exactly by matching late-time relaxation rates between the discrete chain and the continuum PDE, giving D(ρ) = (k/η) a^(1−ν) ρ^(−(1+ν)) for a power-law force family parameterized by ν.

What carries the argument

The central mechanism is relaxation-rate matching. The discrete cell model, linearized about its uniform steady state, relaxes exponentially with a rate set by the smallest eigenvalue of a weighted graph Laplacian. The continuum nonlinear diffusion PDE, linearized about its own uniform steady state, also relaxes exponentially with a rate set by D(ρ̄) times the smallest Laplacian eigenvalue of the domain. Equating these two rates across simulations with different steady-state densities ρ̄ reveals D(ρ). The dimensional scaling argument then enters: because density has dimensions L^(−d) and diffusivity has dimensions L²T⁻¹, and because D(ρ) is assumed to depend only on local length scales (the胞

Load-bearing premise

The dimensional scaling γ = (β − 2)/d is derived from the assumption that the macroscopic diffusivity depends only on local length scales—the cell rest length a and the local density ρ—and receives no contribution from macroscopic length scales such as the domain size or initial-condition geometry. This is a dimensional-analysis ansatz, not a rigorous coarse-graining derivation from the discrete equations in dimensions higher than one.

What would settle it

If two-dimensional simulations with systematically varied domain sizes, initial-condition length scales, or network topologies produced diffusivity exponents γ that deviate from (β − 2)/d in a way correlated with a macroscopic length scale, the central scaling relation would be undermined. The current 2D data (γ ≈ −1.10 for ν = 1, predicted −1.0) is consistent but not exact.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This manuscript studies how dimensionality affects the continuum coarse-graining of discrete cell-mechanics models. In 1D, the authors consider an overdamped chain of mechanical links with a parameterized force law f(ℓ) (Eq. 2). By linearizing about the equally-spaced steady state, they obtain the exact late-time relaxation rate (Eq. 4). Matching this to the relaxation rate of a nonlinear diffusion PDE near steady state (Eq. 11) yields a closed-form nonlinear diffusivity D(ρ) = (k/η)a^{1-ν}ρ^{-(1+ν)} (Eq. 12) with no free parameters. The authors then propose a dimensional scaling ansatz D(ρ) = a^β ρ^γ / τ (Eq. 7), which gives γ = (β-2)/d (Eq. 8), and test this in 2D via simulations on triangulated networks. The central claim is that the same microscopic force law produces dimension-dependent macroscopic transport exponents.

Significance. The 1D relaxation-rate matching method is a clean and elegant technique for coarse-graining discrete mechanics into nonlinear diffusion laws without requiring formal continuum limits. The exact 1D result (Eq. 12) is parameter-free and non-circular: the discrete relaxation rate (Eq. 4) and the PDE relaxation rate (Eq. 11) are derived independently, and their matching yields D(ρ) directly. The dimensional scaling relation γ = (β-2)/d (Eq. 8) is a falsifiable, parameter-free prediction that connects microscopic force-law exponents to macroscopic transport exponents across dimensions. The observation that linear diffusion (γ=0) becomes increasingly relevant in higher dimensions is physically interesting. The open-source code and reproducible data are a strength.

major comments (3)
  1. The 2D simulation results in Fig. 3b show substantial deviations from the predicted scaling γ = (β-2)/d (Eq. 8) that are not discussed. With β = 1-ν (from Eq. 12) and d = 2, the predicted γ values are -1.0, 0.0, and 1.0 for ν = 1, -1, -3, respectively. The measured values are -1.10, -0.22, and 0.47. The ν = -1 case is particularly concerning: the prediction is γ = 0 (linear diffusion), but the measured γ ≈ -0.22 represents a qualitative, not merely quantitative, discrepancy. The ν = -3 case shows a 53% deviation. The paper should (i) explicitly tabulate predicted versus measured exponents, (ii) discuss possible sources of deviation (finite-size effects, irregular triangulation spectrum, breakdown of the locality ansatz), and (iii) perform finite-size scaling by varying the number of cells Q to determine whether deviations vanish as Q → ∞. Currently, Q is held fixed and the paper only声称in
  2. No uncertainty estimates or error bars are reported on the fitted exponents in Fig. 3. The 2D relaxation rates are estimated by averaging slopes from 20 sampled nodes, but no sample variance or confidence interval is reported. Given the deviations from prediction discussed above, understanding the statistical uncertainty is essential for assessing whether the scaling relation holds in d > 1.
  3. The abstract states that 'exponents in nonlinear diffusivities are fixed by microscopic mechanics and dimensionality.' This is exact in 1D (derived via Eq. 12) but in d > 1 rests on the dimensional-analysis ansatz of Eq. (7), which assumes D(ρ) depends only on local length scales a and ρ with no macroscopic length-scale contribution. The 2D data validates this approximately, not exactly. The abstract and conclusions should accurately reflect this distinction between the exact 1D result and the approximate higher-dimensional scaling.
minor comments (4)
  1. The paper states 'Simulation results in Figure 3 are independent of L and H (not shown)' but does not mention varying the number of cells Q. Since Q controls the discretization resolution in 2D, this should be clarified or tested.
  2. In the 2D simulation methodology, relaxation rates are estimated by tracking horizontal positions of 20 randomly selected cell nodes. The sensitivity of the fitted exponent to this node selection (e.g., dependence on which 20 nodes are chosen) should be briefly discussed.
  3. Figure 3c is described as showing that β and γ are related by Eq. (7), but the specific predicted vs. measured (β, γ) pairs are not tabulated, making it difficult to assess the quality of agreement independently of the figure.
  4. The phrase 'Simulation data confirms this' (after proposing Eq. 7) could be misread as confirming the exact scaling relation, when it appears to refer only to independence from L and H. Clarifying the scope of what is confirmed would improve precision.

Circularity Check

0 steps flagged

No significant circularity: 1D derivation matches two independently obtained relaxation rates; dimensional scaling is a stated ansatz tested against independent simulations.

full rationale

The paper's 1D result (Eq. 12) is derived by matching the discrete-model relaxation rate (Eq. 4, from eigenvalue analysis of the linearized ODE system) with the PDE relaxation rate (Eq. 11, from linearization of the nonlinear diffusion equation). These are independently derived, and their equating yields D(ρ) without fitting. The dimensional scaling relation γ=(β-2)/d (Eq. 8) is explicitly presented as a proposal under a stated locality assumption, not as a derivation from the discrete equations in d>1. This makes it a falsifiable ansatz, not a circular result. The 2D simulations generate relaxation data independently and compare fitted exponents to the dimensional prediction; the data is not constructed to satisfy the prediction. The self-citation [19] (co-authored by Simpson) provides the steady-state result but is not load-bearing for the central claim, as the paper re-derives all relaxation rates from first principles. The 2D exponent deviations (e.g., γ≈0.47 vs predicted 1.0 for ν=-3) are a correctness/assumption concern—the locality ansatz may be approximate in d>1—but this is not circularity. The derivation chain does not reduce to its inputs by construction at any step.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 0 invented entities

The paper introduces no new physical entities, particles, or forces. The force law family (Eq. 2) is a parameterization of known interaction types. The dimensional scaling relation is a mathematical consequence of the stated assumptions, not a new postulated object. The axiom count is low and the assumptions are standard within the tissue mechanics continuum-limit program, except for the locality assumption which is the key vulnerability.

free parameters (3)
  • ν (force law exponent) = varied: 1, -1, -3 in simulations
    Parameterizes the family of force laws f(ℓ)=k(ℓ^ν-a^ν)/(νa^(ν-1)). Not fitted to data; chosen as test cases spanning concave-up, linear, and concave-down force laws.
  • k, a, η (spring constant, rest length, drag) = k=a=η=1 in all simulations
    Standard mechanical parameters set to unity. Not fitted.
  • β (microscopic scaling exponent) = β=ν-2 (implied by Eq. 12 and Eq. 8)
    Relates the force-law parameter ν to the dimensional scaling. In 1D, β=-(1+ν) from Eq. 12, and γ=(β-2)/d gives γ=-(1+ν) for d=1, which is consistent. The relationship β=ν-2 is not explicitly stated but is implied.
axioms (4)
  • domain assumption D(ρ) depends only on local length scales a and ρ, with no macroscopic length-scale contribution.
    Stated explicitly: 'We propose Equation (7) under the standard assumption that no macroscopic length-scale contributes to D(ρ), such as length-scales associated with the domain or initial condition.' This is the load-bearing assumption for the dimensional scaling relation (Eq. 8).
  • standard math Late-time relaxation is dominated by the leading eigenmode of the linearized system.
    Standard result from linear stability theory. Used in both 1D (Eq. 4) and 2D (Eq. 6) to extract relaxation rates. Justified by the spectral gap of the Laplacian.
  • domain assumption In 2D, the leading relaxation mode is the horizontal mode with eigenvalue (π/L)^2.
    Stated: 'for higher-dimensional relaxation dominated by the horizontal mode of the initial condition, the decay rate is -D(ρ̄)(π/L)^2.' This holds for the chosen initial conditions (density varies only in x) but is not general.
  • domain assumption The discrete model's late-time behavior near steady state is accurately described by a nonlinear diffusion PDE.
    The matching procedure assumes the PDE form (Eq. 9) describes the coarse-grained dynamics. In 1D this is justified by prior work [7,9,19]; in 2D it is assumed but not independently derived.

pith-pipeline@v1.1.0-glm · 13321 in / 2916 out tokens · 271660 ms · 2026-07-09T21:53:07.693241+00:00 · methodology

0 comments
read the original abstract

Continuum descriptions of epithelial tissue mechanics can replace expensive individual-based simulations with tractable macroscopic models, yet the link between cell-scale forces and tissue-scale transport remains poorly understood. We show that dimensionality controls this link: long-time mechanical relaxation rates reveal generalized porous-media-type nonlinear transport phenomena, $D(\rho)\propto\rho^\gamma$. Exponents in nonlinear diffusivities are fixed by microscopic mechanics and dimensionality, providing a novel physical mechanism for emergent macroscopic transport phenomena.

Figures

Figures reproduced from arXiv: 2607.07000 by Matthew J Simpson, Pascal R Buenzli.

Figure 1
Figure 1. Figure 1: FIG. 1. Discrete models and force laws. (a) 1D chain with fixed endpoints, interval lengths [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. 1D continuum–discrete comparisons. (a)–(c) Continuum densities [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Coarse-graining microscopic dynamics into [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

33 extracted references · 33 canonical work pages

  1. [1]

    G. M. Odell, G. Oster, P. Alberch, and B. Burnside, The mechanical basis of morphogenesis. I. epithelial folding and invagination, Developmental Biology85, 446 (1981)

    work page 1981

  2. [2]

    G. F. Oster, J. D. Murray, and A. K. Harris, Mechanical aspects of mesenchymal morphogen- esis, Journal of Embryology and Experimental Morphology78, 83 (1983)

    work page 1983

  3. [3]

    Pitt-Francis, P

    J. Pitt-Francis, P. Pathmanathan, M. O. Bernabeu, R. Bordas, J. Cooper, A. G. Fletcher, G. R. Mirams, P. Murray, J. M. Osborne, A. Walter, S. J. Chapman, A. Garny, I. M. M. van Leeuwen, P. K. Maini, B. Rodr´ ıguez, S. L. Waters, J. P. Whiteley, H. M. Byrne, and D. J. Gavaghan, Chaste: A test-driven approach to software development for biological modelling...

    work page 2009

  4. [4]

    G. R. Mirams, C. J. Arthurs, M. O. Bernabeu, R. Bordas, J. Cooper, A. Corrias, Y. Davit, S.-J. Dunn, A. G. Fletcher, D. G. Harvey, M. E. Marsh, J. M. Osborne, P. Pathmanathan, J. Pitt-Francis, J. Southern, N. Zemzemi, and D. J. Gavaghan, Chaste: An open source C++ library for computational physiology and biology, PLoS Compututational Biology9, e1002970 (2013)

    work page 2013

  5. [5]

    M. J. Simpson and M. J. Plank, Inference and prediction for stochastic models of biological populations undergoing migration and proliferation, Journal of the Royal Society Interface 22, 20250536 (2025)

    work page 2025

  6. [6]

    J. A. Fozard, H. M. Byrne, O. E. Jensen, and J. R. King, Continuum approximations of individual-based models for epithelial monolayers, Mathematical Medicine and Biology: A Journal of the IMA27, 39 (2009)

    work page 2009

  7. [7]

    P. J. Murray, C. M. Edwards, M. J. Tindall, and P. K. Maini, From a discrete to a continuum model of cell dynamics in one dimension, Physical Review E80, 031912 (2009)

    work page 2009

  8. [8]

    P. J. Murray, A. Walter, A. G. Fletcher, C. M. Edwards, M. J. Tindall, and P. K. Maini, Comparing a discrete and continuum model of the intestinal crypt, Physical Biology8, 026011 (2011)

    work page 2011

  9. [9]

    P. J. Murray, C. M. Edwards, M. J. Tindall, and P. K. Maini, Classifying general nonlinear force laws in cell-based models via the continuum limit, Physical Review E85, 021921 (2012)

    work page 2012

  10. [10]

    R. E. Baker, A. Parker, and M. J. Simpson, A free boundary model of epithelial dynamics, Journal of Theoretical Biology481, 61 (2019). 14

    work page 2019

  11. [11]

    R. J. Murphy, P. R. Buenzli, R. E. Baker, and M. J. Simpson, A one-dimensional individual- based mechanical model of cell movement in heterogeneous tissues and its coarse-grained approximation, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences475, 20180838 (2019)

    work page 2019

  12. [12]

    R. J. Murphy, P. R. Buenzli, R. E. Baker, and M. J. Simpson, Mechanical cell competition in heterogeneous epithelial tissues, Bulletin of Mathematical Biology475, 130 (2019)

    work page 2019

  13. [13]

    R. J. Murphy, P. R. Buenzli, T. A. Tambyah, E. W. Thompson, H. J. Hugo, R. E. Baker, and M. J. Simpson, The role of mechanical interactions in emt, Physical Biology18, 046001 (2021)

    work page 2021

  14. [14]

    B. D. Hughes,Random Walks and Random Environments(OUP, Oxford, 1995)

    work page 1995

  15. [15]

    Redner,A Guide to First-Passage Processes(CUP, Cambridge, 2001)

    S. Redner,A Guide to First-Passage Processes(CUP, Cambridge, 2001)

    work page 2001

  16. [16]

    Goldenfeld, O

    N. Goldenfeld, O. Martin, Y. Oono, and F. Liu, Anomalous dimensions and the renormaliza- tion group in a nonlinear diffusion process, Physical Review Letters64, 1361 (1990)

    work page 1990

  17. [17]

    Bricmont and A

    J. Bricmont and A. Kupiainen, Renormalization group for diffusion in a random medium, Physical Review Letters66, 1689 (1991)

    work page 1991

  18. [18]

    J. L. V´ azquez,The Porous Medium Equation: Mathematical Theory(OUP, Oxford, 2007)

    work page 2007

  19. [19]

    P. R. Buenzli, S. Kuba, R. J. Murphy, and M. J. Simpson, Mechanical cell interactions on curved interfaces, Bulletin of Mathematical Biology87, 28 (2025)

    work page 2025

  20. [20]

    Bear,Dynamics of Fluids in Porous Media(American Elsevier, New York, 1972)

    J. Bear,Dynamics of Fluids in Porous Media(American Elsevier, New York, 1972)

    work page 1972

  21. [21]

    G. I. Barenblatt,Scaling, Self-similarity, and Intermediate Asymptotics(CUP, Cambridge, 1996)

    work page 1996

  22. [22]

    M. J. Simpson, F. Jazaei, and T. P. Clement, How long does it take for aquifer recharge or aquifer discharge processes to reach steady state?, Journal of Hydrology501, 241 (2013)

    work page 2013

  23. [23]

    Okubo and S

    A. Okubo and S. A. Levin,Diffusion and Ecological Problems: Modern Perspectives(Springer, New York, 2001)

    work page 2001

  24. [24]

    J. D. Murray,Mathematical Biology I: An Introduction, 3rd ed., Vol. 17 (Springer, New York, 2002)

    work page 2002

  25. [25]

    S. W. McCue, W. Jin, T. J. Moroney, K.-Y. Lo, S.-E. Chou, and M. J. Simpson, Hole-closing model reveals exponents for nonlinear degenerate diffusivity functions in cell biology, Physica D: Nonlinear Phenomena398, 130 (2019)

    work page 2019

  26. [26]

    Kouachi, Eigenvalues and eigenvectors of tridiagonal matrices, Electronic Journal of Linear 15 Algebra15, 115 (2006)

    S. Kouachi, Eigenvalues and eigenvectors of tridiagonal matrices, Electronic Journal of Linear 15 Algebra15, 115 (2006)

    work page 2006

  27. [27]

    Haberman,Applied Partial Differential Equations with Fourier Series and Boundary Value Problems(Pearson, 2013)

    R. Haberman,Applied Partial Differential Equations with Fourier Series and Boundary Value Problems(Pearson, 2013)

    work page 2013

  28. [28]

    M. J. Simpson and R. E. Baker, Exact calculations of survival probability for diffusion on growing lines, disks, and spheres: The role of dimension, Journal of Chemical Physics143, 094109 (2015)

    work page 2015

  29. [29]

    Newman,Networks: An Introduction(OUP, Oxford, 2018)

    M. Newman,Networks: An Introduction(OUP, Oxford, 2018)

    work page 2018

  30. [30]

    Mohar, Some applications of laplace eigenvalues of graphs, inGraph Symmetry: Algebraic Methods and Applications, NATO ASI Series C: Mathematical and Physical Sciences, Vol

    B. Mohar, Some applications of laplace eigenvalues of graphs, inGraph Symmetry: Algebraic Methods and Applications, NATO ASI Series C: Mathematical and Physical Sciences, Vol. 497, edited by G. Hahn and G. Sabidussi (Springer, 1997) pp. 227–275

    work page 1997

  31. [31]

    Code and data,https://github.com/ProfMJSimpson/tissue(2026)

    work page 2026

  32. [32]

    M. J. Simpson and S. W. McCue, Fisher–KPP-type models of biological invasion: open source computational tools, key concepts and analysis, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences480, 20240186 (2024)

    work page 2024

  33. [33]

    Rackauckas and Q

    C. Rackauckas and Q. Nie, DifferentialEquations.jl – a performant and feature-rich ecosystem for solving differential equations in Julia, Journal of Open Research Software5, 15 (2017). 16

    work page 2017