Pith. sign in

REVIEW 4 major objections 5 minor 33 references

Three-body oscillator coupling is just pairwise links with hidden channel memory

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 03:04 UTC pith:I3C6MMIO

load-bearing objection Sound analytical core with a genuinely new decomposition; dynamical findings are preliminary but honestly reported the 4 major comments →

arxiv 2607.07662 v1 pith:I3C6MMIO submitted 2026-07-08 nlin.AO physics.app-phphysics.comp-ph

Impact of Channel Dynamics on Higher-order Interactions of Oscillators

Gabriel Marghoti , Isabela A. Martins , Giovana S. Spezzatto , Emanuel B. S. A. Cambraia , Sergio R. Lopes , Thiago L. Prado This is my paper
classification nlin.AO physics.app-phphysics.comp-ph
keywords topologieshigher-orderlatenttransmissionvariableschannelchannelsdynamics
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 the standard three-body Kuramoto interaction term sin(θ_j + θ_k − 2θ_i), widely used to model higher-order interactions in oscillator networks, is not a fundamental coupling but an artifact that emerges only when one assumes transmission channels respond instantaneously and symmetrically. The authors build a model where each pairwise connection carries a latent transmission variable with finite inertia (a relaxation timescale τ). When that timescale is fast (the adiabatic limit) and the environmental modulation tensor B_ijk is symmetric under swapping j and k, the familiar three-body term drops out. But when either assumption is relaxed — finite channel memory or asymmetric modulation — the interaction takes a more general form, Σ B_ijk sin(θ_j − θ_i) cos(θ_k − θ_i), which does not map onto any static hyperedge. The authors show that symmetric tensors produce bistability and cluster states, while antisymmetric tensors combined with finite channel inertia produce anti-phase cluster synchronization — a dynamical regime with no static-topology counterpart. However, they find that this anti-phase clustering is a finite-size effect that washes out in all-to-all networks of 100 or more oscillators, as mean-field averaging erases the localized non-reciprocal feedback that sustains it. The central claim is that static higher-order topologies are a restrictive special case of a richer, asymmetric, memory-laden transmission dynamics.

Core claim

The standard (1,1,−2) three-body Kuramoto hyperedge emerges only under the joint conditions of adiabatic channel elimination and strict permutation symmetry B_ijk = B_ikj. When the symmetry is broken, the interaction decomposes into a symmetric part C_ijk that recovers the standard three-body term and an antisymmetric part D_ijk that produces a fundamentally different coupling sin(θ_j − θ_k), which has no static-hyperedge representation. When channel inertia is finite, the transmission variable retains a memory kernel, and the antisymmetric component can drive anti-phase cluster synchronization — but only in small networks, because mean-field homogenization destroys the effect for N ≥ 100 in

What carries the argument

A latent transmission variable u_ij(t) with first-order relaxation timescale τ, modulated by a third-party environmental tensor B_ijk. The variable sits between the phase dynamics and the coupling, so that the effective interaction is a convolution of past phase differences weighted by an exponential kernel. Decomposing B_ijk into symmetric (C_ijk) and antisymmetric (D_ijk) components cleanly separates the standard three-body term from the novel non-reciprocal term.

Load-bearing premise

The paper's most novel dynamical result — anti-phase cluster synchronization from antisymmetric tensors with finite channel inertia — is shown to vanish for networks of 100 or more oscillators in all-to-all coupling. The authors hypothesize that sparse networks may preserve it, but this is untested.

What would settle it

Construct a sparse-network oscillator system with antisymmetric B_ijk and finite channel inertia at large N. If anti-phase clustering still vanishes, the paper's novel dynamical regime has no macroscopic relevance in any topology.

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

If this is right

  • Models that assign static (1,1,−2) hyperedges to oscillator networks are implicitly assuming both instantaneous transmission and perfect reciprocity of environmental modulation — assumptions that are physically implausible in neural, social, or directed transport networks.
  • Antisymmetric interaction tensors, which produce no standard three-body coupling at all in the adiabatic limit, become dynamically consequential once channel memory is included, suggesting that a large class of interactions invisible to static topology analysis may be active in real systems.
  • The finite-size collapse of anti-phase clustering in all-to-all networks implies that sparse or spatially embedded architectures may be the natural setting where non-reciprocal channel dynamics produce macroscopic effects.
  • Experimental oscillator systems with controllable transmission delays (e.g., coupled lasers, electronic circuits with band-limited coupling) could test whether antisymmetric modulation and finite channel inertia produce the predicted two-cluster states.

Where Pith is reading between the lines

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

  • If sparse networks do preserve the anti-phase clustering (as the authors hypothesize but do not test), then the interaction between network topology and channel timescale becomes a design parameter: one could tune clustering behavior by adjusting either the sparsity or the transmission delay, not just the coupling strength.
  • The decomposition into symmetric and antisymmetric tensor components suggests a natural classification scheme for higher-order interactions: symmetric components map to known static topologies, while antisymmetric components represent a qualitatively distinct, dynamically generated interaction class that current hypergraph frameworks cannot represent.
  • The memory kernel structure implies that the effective interaction at time t depends on the full trajectory of phases, not just their instantaneous values — meaning that systems with the same instantaneous phase configuration but different histories could evolve differently, a property no static topology can capture.

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

4 major / 5 minor

Summary. This manuscript derives a mathematical equivalence between standard 3-body Kuramoto couplings and pairwise interactions mediated by latent transmission variables with finite inertia. The core analytical result (§IV) shows that adiabatic elimination of the transmission variable $u_{ij}$ yields the standard $sin(θ_j + θ_k - 2θ_i)$ term only under the symmetry condition $B_{ijk} = B_{ikj}$ (Eq. 16); when this symmetry is broken, a more general interaction $Σ B_{ijk} sin(θ_j - θ_i)cos(θ_k - θ_i)$ emerges (Eq. 9). Numerical experiments on small networks (N=10, 30) explore dynamical consequences: symmetric tensors produce bistability, while antisymmetric tensors with finite channel inertia produce anti-phase cluster synchronization. The authors honestly report that the antisymmetric clustering vanishes for N≥100 in all-to-all networks (Fig. 5) and hypothesize that sparse topologies may preserve it.

Significance. The analytical derivation in §IV is the paper's principal contribution: it is parameter-free, proceeds by straightforward adiabatic elimination and trigonometric decomposition, and provides a legitimate mechanistic grounding for why the standard (1,1,−2) hyperedge is a special case rather than a universal rule. The decomposition into symmetric and antisymmetric tensor components (Eqs. 12–15) is clean and illuminating. The dynamical findings (bistability for symmetric tensors, clustering for antisymmetric ones) are genuine dynamical predictions rather than fitted results, and the honest acknowledgment that clustering is a finite-size effect in all-to-all networks is commendable. Reproducible code is provided via Zenodo. The paper bridges the higher-order network literature with the physical-transmission-medium perspective in a substantive way.

major comments (4)
  1. §V, Fig. 2: The phase diagrams are constructed for N=10 only, with no error bars or statistical uncertainty estimates across the 100 trials mentioned in the text. Given that the central dynamical claims (bistability regimes, partial synchronization boundaries) are drawn from these diagrams, the absence of uncertainty quantification makes it difficult to assess whether the regime boundaries are robust or artifacts of finite-size fluctuations. At minimum, confidence intervals on the order-parameter thresholds used to classify regimes should be reported, or the diagrams should be labeled explicitly as schematic.
  2. §V, Fig. 5 and surrounding text: The most novel dynamical finding — anti-phase clustering driven by antisymmetric tensors with finite τ — is shown to vanish for N≥100. The authors hypothesize that sparse networks may preserve it but provide no test. This is the load-bearing concern for the paper's practical significance. A single sparse-network experiment (even at moderate N with a ring or degree-regular topology) would substantially strengthen or falsify the central claim. Without it, the paper's novel dynamical regime rests on an unverified conjecture.
  3. §IV, Eq. (15): The paper does not explicitly discuss the structural reason for the clustering's fragility. In the adiabatic limit with uniform all-to-all antisymmetric $D_{ijk}$, the antisymmetric term reduces to $Σ_{j,k} D_{ijk} sin(θ_j - θ_k)$, which is independent of the target index $i$. This means the antisymmetric contribution acts as a uniform frequency shift in the adiabatic limit and cannot produce cluster structure — clustering requires finite-τ memory effects that break this $i$-independence. Making this mechanism explicit would clarify why the phenomenon is inherently non-adiabatic and finite-size, and would help frame the sparse-network hypothesis more precisely.
  4. §V: Only uniform tensor entries ($C_{ijk}=1.0$ or $D_{ijk}=1.0$ for all indices) were tested. Since the paper's central analytical point is that symmetry-breaking in $B_{ijk}$ produces qualitatively new dynamics, testing at least one heterogeneous antisymmetric tensor (e.g., random or structured $D_{ijk}$) would demonstrate that the findings are not specific to the uniform case and would strengthen the connection between the analytical and numerical contributions.
minor comments (5)
  1. §II, Eq. (2): The 1/N normalization on the $K_2$ term, combined with the 1/N in Eq. (1), produces a 1/N² prefactor on the higher-order term (Eq. 7). This is the normalization causing the HOI term to vanish as N→∞ for fixed K₂, which is directly relevant to the finite-size discussion in §V. This connection should be stated explicitly rather than left implicit.
  2. Fig. 3 caption: The 'asymmetric case' (panels c1–c4, where $B_{ijk}$ has unit entries strictly for $j<k$) is introduced only in the figure caption and not in the main text of §V. This case should be described in the body, including how it relates to the symmetric/antisymmetric decomposition of §IV.
  3. §IV.B, Eq. (7): The transition from Eq. (6) to Eq. (7) involves substituting the adiabatic solution into Eq. (1), but the 1/N prefactor from Eq. (1) combined with the 1/N from Eq. (2) yields the 1/N² factor. This compound normalization should be made explicit for the reader.
  4. §V: The O(N³) scaling is cited as the reason larger phase diagrams are computationally unfeasible. A brief comment on whether sparse-tensor representations or subset sampling could mitigate this would be helpful, especially given that the paper itself advocates sparse topologies for future work.
  5. References [4]–[7], [9], [14], [15], [17] include several 2026 arXiv preprints. Where published versions exist, they should be referenced; where only preprints are available, this is fine but the access dates could be noted.

Simulated Author's Rebuttal

4 responses · 0 unresolved

We thank the referee for a careful and constructive reading of our manuscript. The referee correctly identifies the analytical derivation in §IV as the principal contribution and raises four major comments concerning: (1) the absence of uncertainty quantification in the N=10 phase diagrams, (2) the untested sparse-network hypothesis for antisymmetric clustering, (3) the need to make explicit the structural mechanism behind clustering fragility, and (4) the restriction to uniform tensor entries. We agree with all four points and will revise the manuscript accordingly. Comments 1, 3, and 4 can be fully addressed in revision. Comment 2 requires new numerical experiments that we will conduct and report, though we are candid about computational constraints on what can be completed within a revision cycle.

read point-by-point responses

  1. Referee: §V, Fig. 2: The phase diagrams are constructed for N=10 only, with no error bars or statistical uncertainty estimates across the 100 trials mentioned in the text. Given that the central dynamical claims (bistability regimes, partial synchronization boundaries) are drawn from these diagrams, the absence of uncertainty quantification makes it difficult to assess whether the regime boundaries are robust or artifacts of finite-size fluctuations. At minimum, confidence intervals on the order-parameter thresholds used to classify regimes should be reported, or the diagrams should be labeled explicitly as schematic.

    Authors: The referee is correct. The phase diagrams in Fig. 2 currently lack uncertainty quantification, and this is a genuine gap. We will address this in two ways. First, we will recompute the regime boundaries with confidence intervals on the order parameters (R1 and R2) across the 100 trials, reporting the standard error or bootstrap intervals at each (K1, K2) grid point. Second, we will relabel Fig. 2 explicitly as schematic in the caption, clarifying that the boundaries are illustrative rather than precise transition lines. We note that the figure caption already uses the word 'Schematic,' but the main text does not adequately emphasize this, and no uncertainty estimates are provided anywhere. Both will be corrected. revision: yes

  2. Referee: §V, Fig. 5 and surrounding text: The most novel dynamical finding — anti-phase clustering driven by antisymmetric tensors with finite τ — is shown to vanish for N≥100. The authors hypothesize that sparse networks may preserve it but provide no test. This is the load-bearing concern for the paper's practical significance. A single sparse-network experiment (even at moderate N with a ring or degree-regular topology) would substantially strengthen or falsify the central claim. Without it, the paper's novel dynamical regime rests on an unverified conjecture.

    Authors: We agree that this is the most consequential concern. The sparse-network hypothesis is currently unsupported by any numerical evidence, and we will conduct the requested experiment. Specifically, we will run simulations on degree-regular (ring-like) topologies at moderate N (e.g., N=30, 50, 100) with the antisymmetric tensor D_ijk=1.0 and finite τ, measuring R1 and R2 over 100 trials as in the existing Fig. 5 protocol. We will report whether anti-phase clustering persists, is enhanced, or still vanishes. We are candid that the O(N^3) scaling of the higher-order interactions limits how large N can be pushed, but degree-regular topologies at N≤100 are computationally feasible. If the clustering persists on sparse topologies, this substantially strengthens the paper; if it does not, we will report this honestly and reframe the contribution accordingly. In either case, the conjecture will be replaced by evidence. revision: yes

  3. Referee: §IV, Eq. (15): The paper does not explicitly discuss the structural reason for the clustering's fragility. In the adiabatic limit with uniform all-to-all antisymmetric D_ijk, the antisymmetric term reduces to Σ_{j,k} D_ijk sin(θ_j − θ_k), which is independent of the target index i. This means the antisymmetric contribution acts as a uniform frequency shift in the adiabatic limit and cannot produce cluster structure — clustering requires finite-τ memory effects that break this i-independence. Making this mechanism explicit would clarify why the phenomenon is inherently non-adiabatic and finite-size, and would help frame the sparse-network hypothesis more precisely.

    Authors: This is an excellent observation that we should have made explicit. The referee's argument is correct: in the adiabatic limit with uniform all-to-all antisymmetric D_ijk, the term Σ_{j,k} D_ijk sin(θ_j − θ_k) is indeed independent of the target index i, reducing to a uniform frequency shift that cannot produce cluster structure. This is precisely why clustering requires finite τ — the memory kernel breaks the i-independence by introducing temporally delayed, non-reciprocal feedback. We will add a paragraph in §IV (after Eq. 15) making this mechanism explicit, and will connect it to the finite-size fragility discussed in §V: in the thermodynamic limit, the mean-field homogenization further suppresses the localized temporal correlations that finite τ introduces. This will also sharpen the motivation for the sparse-network experiments, since sparsity limits the number of interacting channels and may preserve the i-dependence of the delayed feedback. revision: yes

  4. Referee: §V: Only uniform tensor entries (C_ijk=1.0 or D_ijk=1.0 for all indices) were tested. Since the paper's central analytical point is that symmetry-breaking in B_ijk produces qualitatively new dynamics, testing at least one heterogeneous antisymmetric tensor (e.g., random or structured D_ijk) would demonstrate that the findings are not specific to the uniform case and would strengthen the connection between the analytical and numerical contributions.

    Authors: The referee is right that testing only uniform tensors leaves open whether the dynamical findings are artifacts of the uniform case. We will add simulations with at least one heterogeneous antisymmetric tensor — specifically, random D_ijk drawn from a uniform distribution on [0, 1] (with D_ikj = -D_ijk to preserve antisymmetry) — at the same parameter values used in Fig. 4 (N=30, K1=0.2, K2=20, varying τ). If the clustering phenomenon persists with heterogeneous tensors, this confirms that the result is not an artifact of uniformity; if it does not, we will report this and discuss what it implies about the robustness of the antisymmetric mechanism. We will also test a structured heterogeneous case (e.g., D_ijk proportional to a distance-dependent kernel) if space permits. This directly connects the analytical generality of the tensor decomposition to the numerical experiments. revision: yes

Circularity Check

0 steps flagged

No circularity found. The derivation chain is parameter-free and self-contained.

full rationale

The paper's central derivation proceeds as follows: (1) Define a Kuramoto model with latent transmission variables u_ij (Eqs. 1–2), with model parameters K_1, K_2, τ, A_ij, B_ijk as inputs. (2) Apply adiabatic elimination (u̇_ij ≈ 0) to obtain the quasi-steady state (Eq. 6) — a standard quasi-steady-state approximation, not a fit. (3) Substitute Eq. 6 into Eq. 1 to obtain the emergent HOI term Σ B_ijk sin(θ_j − θ_i)cos(θ_k − θ_i) (Eqs. 7, 9, 10) — straightforward algebraic substitution. (4) Bisect the sum, swap dummy indices (Eq. 11), decompose B_ijk into symmetric C_ijk and antisymmetric D_ijk (Eqs. 12–13), and apply the trigonometric identity sin(x)cos(y) ± cos(x)sin(y) = sin(x ± y) to obtain Eq. 15. (5) Show that under D_ijk = 0 (i.e., B_ijk = B_ikj), the standard 3-body coupling sin(θ_j + θ_k − 2θ_i) is recovered (Eq. 16). Every step in this chain is a parameter-free mathematical identity or standard approximation. No parameter is fitted to a target and then 'predicted.' The dynamical results (bistability, anti-phase clustering, finite-size collapse) are numerical simulations of the defined model with stated parameter values, not restatements of fitted quantities. The paper cites ref [17] (Kuehn and Murphy) for the general adiabatic-elimination approach, but this is an external citation by different authors and is not load-bearing for the specific derivation — the algebra in Eqs. 6–16 is self-contained. No self-citations by the present authors appear in the reference list. The derivation is clean.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 1 invented entities

legacy field

free parameters (4)
  • K1 = varied (parameter space scan)
    Pairwise coupling strength, treated as an input parameter scanned over a range.
  • K2 = varied (parameter space scan)
    Higher-order coupling strength, treated as an input parameter scanned over a range.
  • τ = varied (0.001 to ~1000)
    Transmission channel timescale, treated as an input parameter.
  • C_ijk / D_ijk entries = 1.0 or 0.0 (uniform)
    The symmetric and antisymmetric tensor components are set to uniform values (1.0 or 0.0) for all entries excluding the null diagonal. This is a modeling choice, not fitted to data.
axioms (3)
  • domain assumption The transmission variable u_ij obeys first-order linear relaxation (Eq. 2).
    §II, Eq. 2. The choice of first-order relaxation is a modeling assumption; higher-order or nonlinear channel dynamics could yield different emergent behavior.
  • domain assumption The environmental modulation enters as cos(θ_k - θ_i) in the u_ij equation.
    §II, Eq. 2. The specific functional form of the modulation is postulated, not derived from an underlying physical model.
  • domain assumption Natural frequencies are drawn from N(0, 0.1).
    §II. Standard Kuramoto assumption; the specific variance is a modeling choice.
invented entities (1)
  • Latent transmission variable u_ij(t) independent evidence
    purpose: Represents the propagated signal or effective interaction current from j to i, mediating pairwise interactions.
    The variable is a modeling construct, but the paper provides falsifiable predictions: the equivalence to 3-body couplings in the adiabatic limit (Eq. 16) and the finite-size scaling of clustering (Fig. 5) are testable against simulations and potentially against experimental oscillator systems.

pith-pipeline@v1.1.0-glm · 13091 in / 3582 out tokens · 195138 ms · 2026-07-09T03:04:49.320425+00:00 · methodology

0 comments
read the original abstract

Modeling higher-order interactions (HOIs) in nonlinear networks with static topologies is often physically restrictive. We demonstrate that standard 3-body Kuramoto couplings are mathematically equivalent to pairwise connections modulated by latent variables of transmission channels. While standard HOI topologies emerge in the adiabatic limit of these variables, relaxing this constraint reveals that latent channel timescales dictate collective macroscopic states. Specifically, transmission inertia drives bistability for symmetric interaction tensors and anti-phase cluster synchronization for antisymmetric ones. Furthermore, dynamically induced clustering in global topologies emerges as a finite-size effect of the dynamics of the local channels. Ultimately, we show that relying exclusively on static topologies restricts interaction modeling. Integrating latent variables captures the transient inertia and fundamental asymmetry of physical networks, bridging the analytical utility of higher-order functions with the reality of the underlying transmission medium.

Figures

Figures reproduced from arXiv: 2607.07662 by Emanuel B. S. A. Cambraia, Gabriel Marghoti, Giovana S. Spezzatto, Isabela A. Martins, Sergio R. Lopes, Thiago L. Prado.

Figure 1
Figure 1. Figure 1: FIG. 1. Model reductionism, from physical transmission [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Schematic phase diagrams of the synchronization [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Snapshots of the phases of a network of [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Order parameter distributions for (a) symmetric and [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Influence of network size [PITH_FULL_IMAGE:figures/full_fig_p007_5.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 · 4 internal anchors

  1. [1]

    Parameter space analysis for larger networks was computationally unfeasible due to theO(N 3) scaling complexity of the higher-order interactions

    in favor of a non-reciprocal flow force. Parameter space analysis for larger networks was computationally unfeasible due to theO(N 3) scaling complexity of the higher-order interactions. Evaluating the Kuramoto (R 1) and Daido (R 2, Eq.3 form= 2) order parameters from both fully synchro- nized and uniformly spaced initial conditions under an 5 FIG. 2. Sch...

    work page arXiv 2025

  2. [2]

    Kuramoto, Chemical turbulence, inChemical Oscilla- tions, Waves, and Turbulence(Springer, 1984) pp

    Y. Kuramoto, Chemical turbulence, inChemical Oscilla- tions, Waves, and Turbulence(Springer, 1984) pp. 111– 140

    work page 1984

  3. [3]

    A. P. Mill´ an, J. J. Torres, and G. Bianconi, Explo- sive higher-order Kuramoto dynamics on simplicial com- plexes, Phys. Rev. Lett.124, 218301 (2020)

    work page 2020

  4. [4]

    Battiston, E

    F. Battiston, E. Amico, A. Barrat, G. Bianconi, G. de Ferraz de Arruda, B. Franceschiello, I. Iacopini, S. K´ efi, V. Latora, Y. Moreno, T. Murray, T. P. Peixoto, G. Vaccario, and G. Petri, The physics of higher-order networks in complex systems, Nat. Phys.17, 1093 (2021)

    work page 2021

  5. [5]

    T. P. Peixoto, L. Peel, T. Gross, and M. De Domenico, Graphs are maximally expressive for higher-order inter- actions, arXiv:2602.16937 (2026)

    work page arXiv 2026

  6. [6]

    Lucas, L

    M. Lucas, L. Gallo, A. Ghavasieh, F. Battiston, and M. De Domenico, Reducibility of higher-order networks from dynamics, Nat. Commun.17, 1551 (2026)

    work page 2026

  7. [7]

    Kirkley, H

    A. Kirkley, H. Felippe, and F. Battiston, Structural re- ducibility of hypergraphs, Phys. Rev. Lett.135, 247401 (2025)

    work page 2025

  8. [8]

    Llabr´ es, R

    J. Llabr´ es, R. Toral, M. San Miguel, and F. V´ azquez, Reducibility of higher-order to pairwise interactions: So- cial impact models on hypergraphs, Phys. Rev. Res.8, 10.1103/wy1x-3px8 (2026)

    work page doi:10.1103/wy1x-3px8 2026

  9. [9]

    A. D. Letten and D. B. Stouffer, The mechanistic basis for higher-order interactions and non-additivity in com- petitive communities, Ecol. Lett.22, 423 (2019)

    work page 2019

  10. [10]

    Higher-order interactions in ecology can be hidden in plain sight

    V. Calleja-Solanas, S. Lamata-Ot´ ın, C. G´ omez-Ambrosi, J. G´ omez-Garde˜ nes, and S. Meloni, Higher-order in- teractions in ecology can be hidden in plain sight, arXiv:2605.06301 (2026)

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  11. [11]

    Kuramoto and H

    Y. Kuramoto and H. Nakao, On the concept of dynamical reduction: the case of coupled oscillators, Philos. Trans. R. Soc. A377, 20190041 (2019)

    work page 2019

  12. [12]

    Le´ on and D

    I. Le´ on and D. Paz´ o, Phase reduction beyond the first order: The case of the mean-field complex Ginzburg- Landau equation, Phys. Rev. E100, 012211 (2019)

    work page 2019

  13. [13]

    Nijholt, J

    E. Nijholt, J. L. Ocampo-Espindola, D. Eroglu, I. Z. Kiss, and T. Pereira, Emergent hypernetworks in weakly cou- pled oscillators, Nat. Commun.13, 4849 (2022). 9

    work page 2022

  14. [14]

    C. Bick, T. B¨ ohle, and C. Kuehn, Higher-order net- work interactions through phase reduction for oscillators with phase-dependent amplitude, J. Nonlinear Sci.34, 10.1007/s00332-024-10053-3 (2024)

    work page doi:10.1007/s00332-024-10053-3 2024

  15. [15]

    Emergence of higher-order interactions in systems of coupled Kuramoto oscillators with time delay

    N. Fujii, K. Taga, R. Muolo, B. Rink, and H. Nakao, Emergence of higher-order interactions in systems of coupled Kuramoto oscillators with time delay, arXiv:2512.16193 (2026)

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  16. [16]

    Composite-Operator Scaling on Triadic Hypergraphs: Formation Transitions in Multi-Agent Architectures with Three-Body Coupling

    E. Salazar, Composite-operator scaling on triadic hy- pergraphs: Formation transitions in multi-agent ar- chitectures with three-body coupling, arXiv:2604.27038 (2026)

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  17. [17]

    Zhang, P

    Y. Zhang, P. S. Skardal, F. Battiston, G. Petri, and M. Lucas, Deeper but smaller: Higher-order interactions increase linear stability but shrink basins, Sci. Adv.10, eado8049 (2024)

    work page 2024

  18. [18]

    Kuehn and F

    C. Kuehn and F. Murphy, Emergent higher-order struc- ture from fast adaptive networks, arXiv:2603.19382 (2026)

    work page arXiv 2026

  19. [19]

    L. Peel, T. P. Peixoto, and M. De Domenico, Statistical inference links data and theory in network science, Nat. Commun.13, 6794 (2022)

    work page 2022

  20. [20]

    Lambiotte, M

    R. Lambiotte, M. Rosvall, and I. Scholtes, From networks to optimal higher-order models of complex systems, Nat. Phys.15, 313 (2019)

    work page 2019

  21. [21]

    Zhang, M

    Y. Zhang, M. Lucas, and F. Battiston, Higher-order in- teractions shape collective dynamics differently in hyper- graphs and simplicial complexes, Nat. Commun.14, 1605 (2023)

    work page 2023

  22. [22]

    Muolo, H

    R. Muolo, H. Nakao, and M. Coraggio, When higher- order interactions enhance synchronization: the case of the Kuramoto model, arXiv:2508.10992 (2025)

    work page arXiv 2025

  23. [23]

    L. V. Gambuzza, F. Di Patti, L. Gallo, S. Lepri, M. Ro- mance, R. Criado, M. Frasca, V. Latora, and S. Boc- caletti, Stability of synchronization in simplicial com- plexes, Nat. Commun.12, 1255 (2021)

    work page 2021

  24. [24]

    P. S. Skardal and A. Arenas, Abrupt desynchronization and extensive multistability in globally coupled oscillator simplexes, Phys. Rev. Lett.122, 248301 (2019)

    work page 2019

  25. [25]

    Le´ on, R

    I. Le´ on, R. Muolo, S. Hata, and H. Nakao, Theory of phase reduction from hypergraphs to simplicial com- plexes: A general route to higher-order Kuramoto mod- els, Physica D482, 134858 (2025)

    work page 2025

  26. [26]

    Ji, C.-Y

    B.-Y. Ji, C.-Y. Wang, and L. L¨ u, Cluster synchronization on sparse simplicial complexes with higher-order repul- sive coupling, New J. Phys.27, 043012 (2025)

    work page 2025

  27. [27]

    Symmetry-based selection rules for higher-order interactions in coupled oscillators

    I. L´ eon, R. Muolo, Y. Zhang, and M. Lucas, Symmetry- based selection rules for higher-order interactions in cou- pled oscillators, arXiv:2606.04904 (2026)

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  28. [28]

    Berner, E

    R. Berner, E. Sch¨ oll, and S. Yanchuk, Multiclusters in networks of adaptively coupled phase oscillators, SIAM J. Appl. Dyn. Syst.18, 2227 (2019)

    work page 2019

  29. [29]

    Seliger, S

    P. Seliger, S. C. Young, and L. S. Tsimring, Plasticity and learning in a network of coupled phase oscillators, Phys. Rev. E65, 041906 (2002)

    work page 2002

  30. [30]

    Aoki and T

    T. Aoki and T. Aoyagi, Co-evolution of phases and con- nection strengths in a network of phase oscillators, Phys. Rev. Lett.102, 034101 (2009)

    work page 2009

  31. [31]

    Daido, Multibranch entrainment and scaling in large populations of coupled oscillators, Phys

    H. Daido, Multibranch entrainment and scaling in large populations of coupled oscillators, Phys. Rev. Lett.77, 1406 (1996)

    work page 1996

  32. [32]

    J. A. Acebr´ on, L. L. Bonilla, C. J. P´ erez Vicente, F. Ri- tort, and R. Spigler, The Kuramoto model: A simple paradigm for synchronization phenomena, Rev. Mod. Phys.77, 137 (2005)

    work page 2005

  33. [33]

    Marghoti, Channeldynamicsonhigherorderinteraction, https://doi.org/10.5281/zenodo.21220174(2026)

    G. Marghoti, Channeldynamicsonhigherorderinteraction, https://doi.org/10.5281/zenodo.21220174(2026)

    work page doi:10.5281/zenodo.21220174(2026 2026