A unifying approach to diffusive transport in heterogeneous media
Yann Lanoiselée, Denis S. Grebenkov, Gianni Pagnini · 2026
Read A unifying approach to diffusive transport in heterogeneous media by Yann Lanoiselée, Denis S. Grebenkov, Gianni Pagnini (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-12775.txt.
Introduction
Yann Lanoiselée and co-authors bring many models of odd diffusion under one frame. The media in view are mixed and change over time, so steps vary in size and link to past steps. Their frame joins those two traits in one Gaussian process with random modulation. Known models like Lévy flights and random walks with rests fit inside it. The authors derive moments and tests for when diffusion turns anomalous. Lab tracks of single particles can then be sorted with more care.
Catalog record
| CREATOR | Yann Lanoiselée, Denis S. Grebenkov, Gianni Pagnini |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55801 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.12775
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A unifying approach to diffusive transport in heterogeneous media Yann Lanoiselée, Denis S. Grebenkov, Gianni Pagnini Abstract We introduce the concept of Randomly Modulated Gaussian Processes as a unifying framework for modeling, analyzing and classifying anomalous diffusion models in heterogeneous media. This formulation incorporates correlations in the displacements together with correlated fluctuations of their amplitudes. Most known models of anomalous diffusion (including Continuous-Time Random Walk and fractional Brownian motion) and random diffusivity can be described and generalized within this framework. Moreover, the unified view identifies the main statistical properties to be probed experimentally for a reliable classification of diffusive dynamics. The proposed matrix formulation facilitates the computation of the first four moments and allows for a systematic statistical characterization of the considered processes. The necessary and sufficient conditions are provided for the emergence of anomalous diffusion. General expressions for the non-Gaussian parameter, ergodicity breaking parameter and covariance of squared increments are derived. Biological applications of this framework for systematic analysis and biophysical interpretations of experimental single-particle trajectories are discussed. Diffusion in heterogeneous media is a generic feature of mesoscopic physical systems leading to deviations from classical Brownian motion. Spatial and temporal disorder in the local environment couples to thermal fluctuations, giving rise to anomalous scaling laws and non-Gaussian displacement statistics. Such behavior has been observed across a wide range of systems, including glassy and supercooled materials 52 ; 63 , systems close to the jamming transition 13 , cold-atom experiments 56 , complex fluids such as worm-like micellar solutions 28 , active suspensions 38 , and colloidal particles in structured environments 12 . Closely related phenomena also arise in biological systems, including search and transport inside living cells 10 ; 71 ; 55 ; 66 and on their membranes 70 ; 64 ; 22 . Moreover, anomalous and non-Gaussian diffusion can emerge even in the absence of active driving, solely due to environmental heterogeneity. Understanding diffusion under these conditions is essential for describing transport-controlled processes such as reaction kinetics, equilibration, and molecular signaling, with implications ranging from soft and condensed matter to biological physics. Two major aspects of diffusion in heterogeneous media, which may occur independently, are the anomalous scaling of the mean-squared displacement (MSD) \langle X^{2}(t)\rangle\propto t^{\alpha} , and non-Gaussian displacement statistics. The scaling exponent \alpha distinguishes subdiffusion ( 0<\alpha<1 ), normal diffusion ( \alpha=1 ), and superdiffusion ( 1<\alpha<2 ). Subdiffusion can originate from long stalling periods between particle displacements, as exemplified by the Continuous-Time Random Walk (CTRW) 44 ; 29 . When the stalling periods are power-law distributed (CTRW-Pow) anomalous diffusion emerges in close relation to time-fractional diffusion equation 72 ; 24 ; 40 . Anomalous diffusion can also originate from the viscoelastic properties of the medium, such as the cytoplasm of a living cell, where long-range correlations of particle displacements are captured by fractional Brownian motion (fBm) 41 ; 8 and the generalized Langevin equation (GLE) 74 ; 53 ; 9 ; 19 . Non-Gaussianity is another key aspect of diffusion in heterogeneous media which indicates variability of at least one intensive quantity (e.g., diffusivity) describing the ensemble. This is the basic idea of superstatistics 6 ; 7 , which encompasses heterogeneity described as a continuous stochastic process in diffusing diffusivity (DD) models 15 ; 26 ; 25 ; 14 ; 34 ; 62 ; 69 , a piecewise-constant function in switching diffusivity models 27 ; 17 ; 21 ; 23 ; 45 ; 46 ; 3 ; 49 , or random yet trajectory-wise constant diffusivity 57 ; 58 ; 48 . Non-Gaussianity appears as a generic property of diffusion in heterogeneous media as discussed for CTRW with exponential waiting times (CTRW-Exp) 4 ; 60 . Over the years a large variety of models have been introduced to describe these statistical features. These models employ different mathematical tools such as Fokker-Planck equations, renewal theory, subordination, fractional derivatives, or Gaussian statistics that are difficult to combine to establish a unified framework for general correlated random walks in heterogeneous media. In this Letter, we fill this gap and introduce a theoretical framework for diffusive motion in heterogeneous media that captures both anomalous scaling and non-Gaussian fluctuations within a unified description. The proposed framework naturally incorporates correlations in the displacements, non-Gaussianity due to medium heterogeneity, anomalous power-law scaling of the MSD, deterministic or random changes in statistical properties, and other features. In the present general setting, we introduce the concept of “randomly modulated Gaussian process” (RMGP), in which the amplitudes of Gaussian increments are randomly rescaled by a stochastic process, that extends the existing terminology “randomly scaled Gaussian process” 61 . This framework produces a sequence of successive random positions at discrete time steps that is the most natural, from a pragmatic point of view, for the analysis and interpretation of experimentally acquired trajectories and for computer-based simulations. In other words, denoting by \delta a time step (fixed by the acquisition technique or chosen in simulations), we represent a random trajectory X(t) of the diffusing particle by an N -dimensional vector X , whose components X_{i} are the successive positions X(i\delta) at times i\delta , with i=1,2,\ldots,N . A matrix representation of Gaussian processes is used to incorporate correlations of displacements through a covariance matrix. In addition, the Gaussian jumps are rescaled to mimic heterogeneity as may be described by CTRW or random diffusivity models 48 ; 15 . For clarity of presentation, we focus on one-dimensional settings but its extension to higher dimensions is immediate when the process coordinates are independent. The construction of a random trajectory involves two sets of N random variables: (i) the standard independent identically distributed Gaussian variables \{\xi_{1},\cdots,\xi_{N}\} to represent thermal noises, and (ii) the positive random variables J_{i} to mimic the effect of heterogeneity. The random variables \{J_{1},\cdots,J_{N}\} are independent of \{\xi_{1},\cdots,\xi_{N}\} such that \langle\sqrt{J_{i}}\xi_{j}\rangle=\langle\sqrt{J_{i}}\rangle\langle\xi_{j}\rangle=0 for any i and j ; in turn, eventual correlations between the random modulations J_{i} are allowed. At this stage, the modulated Gaussian variables \sqrt{J_{i}}\xi_{i} are first-order uncorrelated, \langle(\sqrt{J_{i}}\xi_{i})(\sqrt{J_{j}}\xi_{j})\rangle=\langle J_{i}\rangle\delta_{ij} but second-order correlated: \langle(\sqrt{J_{i}}\xi_{i})^{2}(\sqrt{J_{j}}\xi_{j})^{2}\rangle=\langle J_{i}J_{j}\rangle for i\neq j and \langle(\sqrt{J_{i}}\xi_{i})^{4}\rangle=3\langle J_{i}^{2}\rangle reflecting the kurtosis of the Gaussian distribution \langle\xi_{i}^{4}\rangle=3 . Next, we define the n -th increment as Y_{n}=\sum_{k=1}^{n}(\sqrt{C})_{nk}\sqrt{J_{k}}\xi_{k} . In the classical case without modulations (i.e., J_{i}=1 ), one has \langle Y_{j}Y_{k}\rangle=C_{j,k} , i.e., C plays the role of the covariance matrix of Gaussian increments. In the presence of modulations, this is not true anymore, and C is used to induce first-order correlations of increments (in the following we still keep calling it ‘covariance matrix’). The positions are finally obtained by summing the increments: X_{n}=\sum_{i=1}^{n}Y_{i} . First-order correlations control the behavior of the MSD, while second-order correlations determine the fourth moment and thus control non-Gaussianity, as detailed below. The model can be represented in a matrix form as X=L\,\sqrt{C}\,\sqrt{J}\,\xi, (1) where L is a lower triangular matrix with entries 1 performing integration, C is the covariance matrix, J is the diagonal matrix containing random modulations, and \xi is the vector of standard IID Gaussian random variables. As the positions in X should be in length units, we set \xi and J to be unitless, while C has units of a squared length. The matrix representation ( 1 ), which is the key point of the paper,