pith. sign in

arxiv: 2509.06546 · v1 · submitted 2025-09-08 · ⚛️ physics.bio-ph

A multiscale theory for network advection-reaction-diffusion

Hadrien Oliveri , Emilia Cozzolino , Alain Goriely This is my paper

Pith reviewed 2026-05-18 18:46 UTC · model grok-4.3

classification ⚛️ physics.bio-ph
keywords network transportadvection-reaction-diffusiongraph Laplacianmultiscale homogenizationmacroscale modelfirst principlestransport operator
0
0 comments X

The pith

Transport on networks is obtained from first principles by homogenizing advection-reaction-diffusion along the edges.

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

The authors model exchanges between network nodes using advection-reaction-diffusion processes inside the connecting edges. They apply multiscale analysis to derive an effective transport operator at the level of the entire network. This operator, an effective graph Laplacian, is determined entirely by the microscale parameters and mechanisms. This matters for building mechanistic models of propagation in systems like brain networks or infection spread, rather than relying on assumed forms for the Laplacian. The work also explores the scaling behavior of this operator as edge lengths change.

Core claim

Using advection-reaction-diffusion as a generic mechanism for inter-nodal exchanges, we derive a multiscale network transport model and obtain the corresponding linear transport operator at the macroscale from first principles. This effective graph Laplacian is fully determined by the transport mechanisms along the edges at the microscale.

What carries the argument

The homogenization of the advection-reaction-diffusion PDE defined on each edge to produce a macroscale linear transport operator on the network graph.

If this is right

  • The resulting operator accurately describes transport across the network.
  • Scaling properties of the operator with edge length follow directly from the microscale coefficients.
  • The approach provides a systematic way to incorporate detailed physics of edge transport into network models.

Where Pith is reading between the lines

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

  • Models of protein transport in neurodegenerative diseases could be refined by choosing appropriate reaction and diffusion terms at the microscale.
  • Comparison with agent-based or full PDE simulations on finite networks would test the accuracy for given scale separations.
  • Similar homogenization techniques might extend to other microscale processes like active transport or stochastic jumps.

Load-bearing premise

Microscale transport inside each edge is well described by a standard advection-reaction-diffusion equation and there is a clear separation between edge length and total network size.

What would settle it

Solving the full advection-reaction-diffusion system numerically on a network and comparing the long-term node concentrations to those predicted by the effective operator; mismatch would indicate the homogenization does not hold.

Figures

Figures reproduced from arXiv: 2509.06546 by Alain Goriely, Emilia Cozzolino, Hadrien Oliveri.

Figure 1
Figure 1. Figure 1: Schematic of the model’s structure. At the microscale, the system is a physical network [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A summary of the upscaling process. From the initial microscale data provided by [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Example simulation of the FKPP process on a two-node network (total nodal mass). [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) The 83-node brain connectome and the Braak regions. (b) Example simulation of the [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
read the original abstract

Mathematical network models are extremely useful to capture complex propagation processes between different regions (nodes), for example the spread of an infectious agent between different countries, or the transport and replication of toxic proteins across different brain regions in neurodegenerative diseases. In these models, transport is modeled at the macroscale through an operator, the so-called graph Laplacian, based on the edge properties and topology, capturing the fluxes between different nodes of the network. However, this phenomenological approach fails to take into account the physical processes taking place at the microscale within the edge. A fundamental problem is then to obtain a transport operator from mechanistic principles based on the underlying transport process. Using advection-reaction-diffusion as a generic mechanism for inter-nodal exchanges, we derive a multiscale network transport model and obtain the corresponding linear transport operator at the macroscale from first principles. This effective graph Laplacian is fully determined by the transport mechanisms along the edges at the microscale. We show that this operator correctly captures the transport, and we study its scaling properties with respect to edge length.

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.

Referee Report

1 major / 1 minor

Summary. The paper develops a multiscale theory for transport on networks by modeling inter-nodal exchanges via an advection-reaction-diffusion PDE at the microscale on each edge. It applies homogenization to derive an effective linear transport operator (graph Laplacian) at the macroscale from first principles, asserting that this operator is fully determined by the microscale mechanisms, and examines its scaling properties with edge length.

Significance. If the central derivation holds with appropriate justification, the work would be significant for applications in epidemiology and neuroscience, where network models describe processes such as disease spread or protein propagation. Deriving the macroscale operator mechanistically rather than phenomenologically, with no free parameters, strengthens the link between microscale physics and macroscale predictions and could improve model fidelity in these domains.

major comments (1)
  1. [multiscale homogenization derivation] The homogenization step assumes scale separation between edge length and network diameter to average the microscale advection-reaction-diffusion PDE into an effective operator, but the manuscript provides no explicit error bounds, convergence rates, or remainder estimates (e.g., O(ε) where ε is the scale ratio). This is load-bearing for the claim that the macroscale operator is fully determined by microscale mechanisms without residual terms, especially since advection or reaction can generate boundary layers that violate the averaging assumption.
minor comments (1)
  1. [Abstract] The abstract could more explicitly state the form of the derived effective operator and the precise assumptions on scale separation to aid readers.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback and for acknowledging the potential impact of our work on fields such as epidemiology and neuroscience. We respond to the major comment as follows.

read point-by-point responses

  1. Referee: The homogenization step assumes scale separation between edge length and network diameter to average the microscale advection-reaction-diffusion PDE into an effective operator, but the manuscript provides no explicit error bounds, convergence rates, or remainder estimates (e.g., O(ε) where ε is the scale ratio). This is load-bearing for the claim that the macroscale operator is fully determined by microscale mechanisms without residual terms, especially since advection or reaction can generate boundary layers that violate the averaging assumption.

    Authors: We agree that a more detailed justification of the homogenization procedure would enhance the manuscript. Our derivation is based on a formal multiscale expansion assuming a clear separation of scales, which allows us to average the microscale PDE to obtain the effective macroscale operator without free parameters. To address this comment, we will revise the manuscript to include an explicit discussion of the error associated with the homogenization. Specifically, we will state that the approximation error is of order O(ε), where ε denotes the ratio of the typical edge length to the network diameter, under the assumption that the advection and reaction terms do not induce persistent boundary layers that span the entire edge. We will also add a brief analysis showing that any boundary layers are localized and their contribution to the averaged transport is incorporated into the effective nodal conditions. This addition will be supported by references to established homogenization results for advection-reaction-diffusion systems. revision: yes

Circularity Check

0 steps flagged

Derivation proceeds via standard multiscale homogenization without reduction to inputs by construction.

full rationale

The paper starts from the advection-reaction-diffusion PDE on individual edges and applies a homogenization procedure to obtain an effective macroscale graph Laplacian. This averaging step is a standard first-principles calculation that produces the operator from the microscale transport coefficients and geometry; it does not define the target quantity in terms of itself, fit parameters to macroscale data, or rely on load-bearing self-citations whose validity depends on the present result. No equations in the abstract or described chain exhibit the self-definitional, fitted-prediction, or ansatz-smuggling patterns. The result is therefore self-contained against external mathematical benchmarks for homogenization.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The derivation rests on the standard advection-reaction-diffusion PDE at microscale and scale-separation assumptions typical of homogenization theory; no new free parameters or invented entities are introduced beyond those already present in the microscale model.

axioms (2)
  • domain assumption Advection-reaction-diffusion PDE governs transport inside each edge
    Invoked as the generic mechanism for inter-nodal exchanges in the abstract.
  • domain assumption Scale separation between edge length and network diameter permits homogenization
    Required for the multiscale upscaling step that produces the macroscale operator.

pith-pipeline@v0.9.0 · 5710 in / 1250 out tokens · 36703 ms · 2026-05-18T18:46:08.951065+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

  1. [1]

    , author Voss, H

    author Abdelnour, F. , author Voss, H. , author Raj, A. , year 2014 . title Network diffusion accurately models the relationship between structural and functional brain connectivity networks . journal Neuroimage volume 90 , pages 335--347 . https://www.sciencedirect.com/science/article/abs/pii/S1053811913012597, :10.1016/j.neuroimage.2013.12.039

    work page doi:10.1016/j.neuroimage.2013.12.039 2014

  2. [2]

    , author Thompson, T.B

    author Ahern, A. , author Thompson, T.B. , author Oliveri, H. , author Lorthois, S. , author Goriely, A. , year 2025 . title Modelling the coupling between cerebrovascular pathology and amyloid beta spreading in Alzheimer's disease . journal Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences volume 481 , pages 20240548 . h...

    work page doi:10.1098/rspa.2024.0548 2025

  3. [3]

    , author Durastante, F

    author Benzi, M. , author Durastante, F. , author Zigliotto, F. , year 2025 . title Modeling advection on distance-weighted directed networks . journal Mathematical Models and Methods in Applied Sciences volume 35 , pages 1237--1265 . https://doi.org/10.1142/S0218202525500162, :10.1142/S0218202525500162

    work page doi:10.1142/s0218202525500162 2025

  4. [4]

    , author Cozzolino, E

    author Bertsch, M. , author Cozzolino, E. , author Tora, V. , year 2025 . title Well-posedness of a network transport model . journal Nonlinear Analysis volume 253 , pages 113714 . https://www.sciencedirect.com/science/article/pii/S0362546X24002335, :10.1016/j.na.2024.113714

    work page doi:10.1016/j.na.2024.113714 2025

  5. [5]

    , author Thompson, T.B

    author Brennan, G.S. , author Thompson, T.B. , author Oliveri, H. , author Rognes, M.E. , author Goriely, A. , year 2024 . title The role of clearance in neurodegenerative diseases . journal SIAM Journal on Applied Mathematics volume 84 , pages S172--S198 . https://doi.org/10.1137/22M1487801, :10.1137/22M1487801

    work page doi:10.1137/22m1487801 2024

  6. [6]

    , author Vogel, J

    author Chaggar, P. , author Vogel, J. , author Binette, A.P. , author Thompson, T.B. , author Strandberg, O. , author Mattsson-Carlgren, N. , author Karlsson, L. , author Stomrud, E. , author Jbabdi, S. , author Magon, S. , et al., year 2025 . title Personalised regional modelling predicts tau progression in the human brain . journal PLoS Biology volume 2...

    work page doi:10.1371/journal.pbio.3003241 2025

  7. [7]

    , author Gerhard, S

    author Daducci, A. , author Gerhard, S. , author Thiran, J.P. , et al., year 2012 . title T he C onnectome M apper: A n O pen- S ource P rocessing P ipeline to M ap C onnectomes with MRI . journal PLoS One volume 7 , pages e48121 . https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0048121, :10.1371/journal.pone.0048121

    work page doi:10.1371/journal.pone.0048121 2012

  8. [8]

    , author Mitidieri, E

    author De Figueiredo, D.G. , author Mitidieri, E. , year 1990 . title Maximum principles for linear elliptic systems . journal Rendiconti dell'istituto matematico di Trieste volume 22 , pages 36--66 . https://link.springer.com/chapter/10.1007/978-3-319-02856-9_21, :10.1007/978-3-319-02856-9_21

    work page doi:10.1007/978-3-319-02856-9_21 1990

  9. [9]

    , author Sch \"a fer, A

    author Fornari, S. , author Sch \"a fer, A. , author Jucker, M. , author Goriely, A. , author Kuhl, E. , year 2019 . title Prion-like spreading of Alzheimer's disease within the brain's connectome . journal Journal of the Royal Society Interface volume 16 , pages 20190356 . https://royalsocietypublishing.org/doi/full/10.1098/rsif.2019.0356, :10.1098/rsif....

    work page doi:10.1098/rsif.2019.0356 2019

  10. [10]

    , author Sch \"a fer, A

    author Fornari, S. , author Sch \"a fer, A. , author Kuhl, E. , author Goriely, A. , year 2020 . title Spatially-extended nucleation-aggregation-fragmentation models for the dynamics of prion-like neurodegenerative protein-spreading in the brain and its connectome . journal Journal of Theoretical Biology volume 486 , pages 110102 . https://www.sciencedire...

    work page doi:10.1016/j.jtbi.2019.110102 2020

  11. [11]

    , author Barrat, A

    author Gautreau, A. , author Barrat, A. , author Barth \'e lemy, M. , year 2007 . title Arrival time statistics in global disease spread . journal Journal of Statistical Mechanics: Theory and Experiment volume 2007 , pages L09001 . https://iopscience.iop.org/article/10.1088/1742-5468/2007/09/L09001, :10.1088/1742-5468/2007/09/L09001

    work page doi:10.1088/1742-5468/2007/09/l09001 2007

  12. [12]

    , year 2021

    author Kuhl, E. , year 2021 . title Computational Epidemiology . publisher Springer , address Cham . https://link.springer.com/book/10.1007/978-3-030-82890-5, :10.1007/978-3-030-82890-5

    work page doi:10.1007/978-3-030-82890-5 2021

  13. [13]

    , author Rahman, P

    author Linka, K. , author Rahman, P. , author Goriely, A. , author Kuhl, E. , year 2020 . title Is it safe to lift COVID-19 travel bans? The Newfoundland story. journal Computational Mechanics volume 66 , pages 1081--1092 . https://link.springer.com/article/10.1007/s00466-020-01899-x, :10.1007/s00466-020-01899-x

    work page doi:10.1007/s00466-020-01899-x 2020

  14. [14]

    , author Mezias, C

    author Pandya, S. , author Mezias, C. , author Raj, A. , year 2017 . title Predictive model of spread of progressive supranuclear palsy using directional network diffusion . journal Frontiers in neurology volume 8 , pages 692 . https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2017.00692, :10.3389/fneur.2017.00692

    work page doi:10.3389/fneur.2017.00692 2017

  15. [15]

    , author Zeighami, Y

    author Pandya, S. , author Zeighami, Y. , author Freeze, B. , author Dadar, M. , author Collins, D. , author Raj, A. , year 2019 . title P redictive model of spread of P arkinson's pathology using network diffusion . journal NeuroImage volume 192 , pages 178--194 . https://www.sciencedirect.com/science/article/abs/pii/S1053811919301685, :10.1016/j.neuroim...

    work page doi:10.1016/j.neuroimage.2019.03.001 2019

  16. [16]

    , year 1953

    author Pini, B. , year 1953 . title Sui sistemi di equazioni lineari a derivate parziali del secondo ordine dei tipi ellittico e parabolico . journal Rendiconti del Seminario Matematico della Universit\`a di Padova volume 22 , pages 265--280 . https://www.numdam.org/item/RSMUP_1953__22__265_0/

    work page 1953

  17. [17]

    , author Thompson, T.B

    author Putra, P. , author Thompson, T.B. , author Chaggar, P. , author Goriely, A. , year 2021 . title Braiding Braak and Braak: Staging patterns and model selection in network neurodegeneration . journal Network Neuroscience volume 5 , pages 929--956 . https://direct.mit.edu/netn/article/5/4/929/107175/Braiding-Braak-and-Braak-Staging-patterns-and, :10.1...

    work page doi:10.1162/netn_a_00208 2021

  18. [18]

    , author Kuceyeski, A

    author Raj, A. , author Kuceyeski, A. , author Weiner, M. , year 2012 . title A network diffusion model of disease progression in dementia . journal Neuron volume 73 , pages 1204--1215 . https://www.cell.com/neuron/fulltext/S0896-6273(12)00135-3, :10.1016/j.neuron.2011.12.040

    work page doi:10.1016/j.neuron.2011.12.040 2012

  19. [19]

    , author LoCastro, E

    author Raj, A. , author LoCastro, E. , author Weiner, M. , et al., year 2015 . title Network diffusion model of progression predicts longitudinal patterns of atrophy and metabolism in Alzheimer's disease . journal Cell reports volume 10 , pages 359--369 . https://www.cell.com/cell-reports/fulltext/S2211-1247(14)01063-8, :10.1016/j.celrep.2014.12.034

    work page doi:10.1016/j.celrep.2014.12.034 2015

  20. [20]

    , author Chaggar, P

    author Thompson, T.B. , author Chaggar, P. , author Kuhl, E. , author Goriely, A. , author the Alzheimer's Disease Neuroimaging Initiative , year 2020 . title Protein-protein interactions in neurodegenerative diseases: A conspiracy theory . journal PLoS computational biology volume 16 , pages e1008267 . https://journals.plos.org/ploscompbiol/article?id=10...

    work page doi:10.1371/journal.pcbi.1008267 2020

  21. [21]

    , author Torok, J

    author Tora, V. , author Torok, J. , author Bertsch, M. , author Raj, A. , year 2025 . title A network-level transport model of tau progression in the Alzheimer's brain . journal Mathematical Medicine and Biology: A Journal of the IMA volume 00 , pages 1--27 . https://doi.org/10.1093/imammb/dqaf003, :10.1093/imammb/dqaf003

    work page doi:10.1093/imammb/dqaf003 2025

  22. [22]

    , author Kuhl, E

    author Weickenmeier, J. , author Kuhl, E. , author Goriely, A. , year 2018 . title The multiphysics of prion-like diseases: progression and atrophy . journal Physical Review Letters volume 121 . https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.121.158101, :10.1103/PhysRevLett.121.158101

    work page doi:10.1103/physrevlett.121.158101 2018