Pith. sign in

REVIEW 3 major objections 5 minor 99 references

Evaluating Gaussianity of heterogeneous fractional Brownian motion

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper derives an exact kurtosis formula for switching fractional Brownian motion and uses it to show which dwell-time statistics keep the process non-Gaussian.

desk verdict Exact kurtosis for Markovian switching FBM is real and the numerics back it up; the paper's reach exceeds its grasp when it treats the Riemann-Liouville construction as the model rather than a choice. read the letter →

arxiv 2501.10472 v2 pith:DHHMUPO2 submitted 2025-01-16 cond-mat.stat-mech math.PR

Micha{l} Balcerek , Adrian Pacheco-Pozo , Agnieszka Wy{l}omańska , Diego Krapf This is my paper · ORCID
classification cond-mat.stat-mechmath.PR MSC 60G2260J2762G10
keywords fractionalBrownianmotionheterogeneousdiffusionswitchingkurtosisnon-GaussianityHellingerdistanceanomalousdwell-timedistributions
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Heterogeneous diffusion is often diagnosed by kurtosis: a value above 3 says the displacements are not Gaussian. This paper asks whether kurtosis can be trusted for fractional Brownian motion whose diffusivity switches between two values, and it answers with an exact formula rather than a heuristic. The formula shows that Markovian (exponential-exponential) switching produces a transient excess kurtosis that decays to 3 on the switching correlation time, so long-time measurements can miss the heterogeneity. By contrast, if at least one dwell-time distribution has a power-law tail with infinite mean, the kurtosis stays clearly above 3 over all simulated time scales, and the Hellinger distance gives the same verdict. A sympathetic reader should care because kurtosis is the cheapest and most widely used non-Gaussianity statistic in single-particle tracking, and this paper identifies exactly when it works for this model class.

What carries the argument

The load-bearing object is the generalized diffusivity process $D(t)$ entering the fractional integral $X(t)=\sqrt{4H}\int_0^t \sqrt{D(t')}(t-t')^{H-1/2}\xi(t')\,dt'$. Because the noise is Gaussian conditionally on $D$, Isserlis' theorem collapses the fourth moment to a double integral of the diffusivity covariance $\langle D(t')D(t'')\rangle$, yielding the general kurtosis formula Eq. (13). For dichotomous switching, the covariance is computed in Laplace domain via renewal transition probabilities; for exponential dwell times it simplifies to $\langle D(t)D(0)\rangle = p_+ D_+^2 + p_- D_-^2 - p_+ p_- (D_+ - D_-)^2 (1 - e^{-t/t_c})$, with $1/t_c = 1/\tau_+ + 1/\tau_-$. Substituting this into Eq. (13) and expanding the exponential gives the exact series Eq. (26) in powers of $t/t_c$.

What would settle it

Simulate two SFBM ensembles with identical $D_\pm$, $\tau_\pm$, $H$, and initial conditions, one using Eq. (3) and one where each complete FBM increment is scaled by $\sqrt{D(t)}$ at the increment level, and compare their kurtosis curves to Eq. (26); if the increment-scaled version does not follow Eq. (26), the formula is specific to the integral construction rather than to switching FBM generally.

Watch

Extended reading notes

Core claim

Switching fractional Brownian motion (SFBM) is defined by taking Lévy's Riemann-Liouville representation of FBM and letting the generalized diffusion coefficient $D(t)$ switch between $D_+$ and $D_-$ while the memory kernel $(t-t')^{H-1/2}$ remains fixed. The paper's central result is the exact kurtosis $K(t)=\langle X^4(t)\rangle/\langle X^2(t)\rangle^2$. For Markovian switching, Eq. (26) gives $K(t)$ as 3 times a constant term built from the diffusivity-weighted moments plus a series in $t/t_c$ whose coefficients involve $H$ and the diffusivity contrast $(D_+-D_-)^2$, so that $K(t)$ decays to 3 as $t/t_c$ grows. The authors conclude that Markovian switching is statistically homogenized at long times and that detecting it requires resolving times shorter than the correlation time $t_c$, while switching with power-law dwell times (PL-Exp or PL-PL) never converges to Gaussianity on the accessible time scales. They further show that the Hellinger distance and other distribution-distance measures reproduce the kurtosis conclusion, so kurtosis is sufficient for this model.

Load-bearing premise

The load-bearing premise is the construction in Eq. (3): $D(t)$ enters as an independent factor inside the fractional integral, before the Gaussian noise is integrated; if the intended switching FBM randomizes diffusivity at the level of complete FBM increments instead, the covariance and the resulting kurtosis would be different.

Editorial extensions

If this is right

  • For Markovian switching, kurtosis returns to 3 once $t$ is several correlation times $t_c$, so detecting heterogeneity with kurtosis requires time resolution finer than the switching timescale.
  • For either PL-Exp or PL-PL switching, kurtosis stays above the Gaussian 95% confidence band over trajectory lengths of 2048 time steps, so a single kurtosis measurement can flag heterogeneity in such systems.
  • The maximum kurtosis for any two-state SFBM is $K_{\max} = 3(c+1)^2/(4c)$, set by the diffusivity ratio $c = D_+/D_-$; observing kurtosis above this value rules out the two-state SFBM construction.
  • Since Hellinger distance and other distribution-based metrics reproduce the kurtosis verdict, kurtosis alone is sufficient as a non-Gaussianity detector for this model.
  • The CTRW limit ($H=1/2$, $D_-\to 0$) inherits the non-Gaussian behavior, connecting the result to established heavy-tailed continuous-time random walks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The general formula Eq. (13) should apply to any diffusivity process with known mean and covariance, so kurtosis curves could be used to fingerprint other heterogeneous models, such as continuous or superstatistical diffusivity, without new derivations.
  • Because the Markovian kurtosis decays on the scale $t_c$, the paper implies a resolution threshold: experiments with sampling interval larger than roughly $t_c$ will see Gaussian kurtosis even though the system is heterogeneous.
  • The construction dependence flagged by Eq. (3) suggests that comparing empirical kurtosis from single-particle trajectories against Eq. (26) could discriminate between diffusivity entering inside the fractional kernel and diffusivity multiplying complete FBM increments.
  • If one repeats the PL-PL simulations with dwell-time tail exponent $\alpha>1$ (finite mean), the paper's logic predicts eventual Gaussian convergence; testing this would isolate infinite mean as the cause of persistent non-Gaussianity.
Share X Bluesky LinkedIn Reddit HN

Formalized claims in Lean

  1. Claim #1: Switching fractional Brownian motion (SFBM) is defined by taking Lévy's Riemann-Liouville representation of FBM and letting the generalized diffusion coefficient $D(t)$ switch between $D_+$ and $D_-$ while the memory kernel $(t-t')^{H-1/2}$ remains fixed. The paper's central result is the exact kurtosis $K(t)=\langle X^4(t)\rangle/\langle X^2(t)\rangle^2$. For Markovian switching, Eq. (26) gives $

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. Section 2 introduces switching fractional Brownian motion (SFBM) as X(t)=√(4H)∫_0^t √(D(t'))(t−t')^{H−1/2}ξ(t')dt', where D(t) is a two-state stochastic process. Section 3 derives the exact kurtosis formula Eq. (13) in terms of the mean and covariance of D(t). Section 4 specializes to exponential (Markovian) dwell times, yielding the explicit series Eq. (26), and defines power-law (non-Markovian) cases. Section 5 validates Eq. (26) by Monte Carlo simulation and reports kurtosis for Exp-Exp, PL-Exp, and PL-PL switching. Section 6 compares kurtosis with Hellinger distance and other distribution-based measures. The authors conclude that Markovian SFBM converges to Gaussian statistics at long times, that power-law switching does not, and that kurtosis is a robust metric for detecting heterogeneity in FBM-like processes.

Significance. The core derivation is algebraically consistent and is a genuine contribution: Eq. (13) and Eq. (26) give a parameter-free exact expression whose inputs are the covariance of the dichotomous process and the MSD, both taken from prior work, and the Monte Carlo simulations match the formula. This is a strength. The comparison with Hellinger distance and the Gaussian-mixture bound in Appendix A are also useful. The main limitation is the breadth of interpretation: the exact results are tied to one specific multiplicative Riemann-Liouville construction, and the claimed long-time non-Gaussianity for PL-Exp is not established by the finite-time simulations shown.

major comments (3)
  1. [Section 7 and Fig. 5] The conclusion that convergence to Gaussianity is absent whenever at least one state has a scale-free dwell-time distribution is too strong for the PL-Exp case. In Fig. 5 the PL-Exp kurtosis is seen to decay toward 3 over the simulated window, and in a two-state renewal process with one exponential state and one infinite-mean power-law state the fraction of time spent in the exponential state tends to zero, so the Gaussian limit is expected at long times. The paper should either limit the non-Gaussian asymptotics claim to PL-PL, or provide an analytic or much longer-time numerical argument for PL-Exp. This issue is load-bearing for the paper's central Markovian/non-Markovian dichotomy.
  2. [Section 2, Eq. (3), and Section 7] All exact results are derived from the specific construction in Eq. (3), in which D(t') multiplies the white noise under a one-sided Riemann-Liouville kernel. The manuscript's broad claims about kurtosis as a robust metric for heterogeneous FBM, in the abstract and conclusions, are not tested against other natural definitions of switching FBM, such as a stationary-increment Volterra representation with the noise amplitude rescaled by sqrt(D(s)), or increment-level randomization of FBM increments. These alternatives would generally give different covariance and fourth-moment structure. Please either explicitly scope the conclusions to the model in Eq. (3) or add a sensitivity analysis.
  3. [Section 5, PL-PL] For PL-PL, the reported long-time kurtosis plateaus (3.75 for H=0.3 and 3.60 for H=0.7 in Fig. 5) are empirical values from simulations up to T=2048 Delta, with no convergence check or uncertainty quantification. Calling these the values to which kurtosis converges is not supported by the data shown. Please add an analysis of the convergence, for example simulations at several longer T with error bars or a scaling argument, or soften the wording to 'does not approach 3 within the observed time window'.
minor comments (5)
  1. [Fig. 5 caption] The caption says the shaded area is based on 1,000 samples of length 100,000, while the text refers to 100,000 Gaussian processes of length 2,048; please make the description consistent.
  2. [Appendix A, Eq. (A.2)] The notation 'p3σ_1^4' should read 'p·3σ_1^4' (and similarly for the second term), as the current notation is ambiguous.
  3. [Section 4.2] The asymptotic form in Eq. (15) uses a normalization with |Γ(−α)|, while the Pareto density in Eq. (27) uses α t_0^α; the relationship between these two forms should be stated.
  4. [Eq. (26)] The Pochhammer symbol (x)_(n) is called a 'falling factorial'; with the given definition it is a rising factorial (Pochhammer symbol), so the wording should be corrected.
  5. [Section 6] The conclusion that the Hellinger distance 'does not provide additional information' is based on one kernel-density-estimation implementation; a brief note on the choice of bandwidth and its influence would be helpful.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the kurtosis is derived from the stated model and validated by simulation; self-citations are model lineage only.

full rationale

The central result, Eq. (26), is obtained by a direct calculation from the model definition Eq. (3): the fourth moment follows from Isserlis' theorem and the covariance of the dichotomous diffusivity taken from the independent external result Ref. [84], while the denominator uses the mean diffusivity convolution. No parameter is fitted to the kurtosis target and no prediction is re-imported as an input; the only self-citations (Refs. [42] and [44]) supply the model definition and the elementary MSD formula, both of which are also stated explicitly in the text and could be derived within the paper. The Markovian convergence to Gaussianity is a consequence of the exponential covariance, and the non-Markovian persistence is read from simulation-based kurtosis values, not from a fitting step. Thus there is no step in which an output is equivalent to an input by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central formula rests on a specific model definition (multiplicative independent diffusivity), standard Gaussian noise calculations, an equilibrium assumption for Markovian switching, and cited covariance formulas. The only fitted quantities are simulation-derived long-time kurtosis plateaus for the non-Markovian case.

free parameters (2)
  • long-time kurtosis plateau for PL-PL with H=0.3 = 3.75
    Extracted from simulation in Fig. 5; no analytical derivation or reported uncertainty.
  • long-time kurtosis plateau for PL-PL with H=0.7 = 3.60
    Extracted from simulation in Fig. 5; no analytical derivation or reported uncertainty.
assumptions (5)
  • domain assumption The one-sided Riemann-Liouville integral in Eq. (3), with D(t) multiplying the fractional kernel, defines heterogeneous FBM.
    The paper inherits this definition from Refs [42,44] and does not compare it with the stationary-increment FBM representation; the kurtosis formula depends on this choice.
  • domain assumption The white noise xi and the diffusivity process D are independent, so Isserlis theorem can be applied conditionally on D.
    Used in Section 3 to compute the fourth moment as a double integral over the covariance of D; if D and xi were correlated, Eq. (13) would fail.
  • domain assumption For Markovian Exp-Exp switching, D(t) is in equilibrium at t=0, so the covariance depends only on time difference and Eq. (24) applies.
    Section 4.1 and Appendix B assume this equilibrated start; the exact short-time kurtosis depends on it.
  • standard math Isserlis theorem for zero-mean Gaussian white noise with delta-correlation Eq. (2) is valid.
    Used in Eq. (8) to reduce the fourth-order noise correlation; this is a standard Gaussian-moment identity.
  • standard math The Laplace-domain covariance formulas from Ref. [84] for the two-state Markov process are correct.
    Eqs. (20) to (24) rely on the transition-probability Laplace transform result of Miyaguchi et al.; the paper treats this as an established external result.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Evaluating Gaussianity of heterogeneous fractional Brownian motion." pith.science (2026). https://pith.science/paper/DHHMUPO2

@misc{pith2026250110472,
  author       = {Pith},
  title        = {Pith review of: Evaluating Gaussianity of heterogeneous fractional Brownian motion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHHMUPO2}},
  note         = {Machine review of arXiv:2501.10472}
}
read the original abstract

Heterogeneous diffusion processes are prevalent in various fields, including the motion of proteins in living cells, the migratory movement of birds and mammals, and finance. These processes are often characterized by time-varying dynamics, where interactions with the environment evolve, and the system undergoes fluctuations in diffusivity. Moreover, in many complex systems anomalous diffusion is observed, where the mean square displacement (MSD) exhibits non-linear scaling with time. Among the models used to describe this phenomenon, fractional Brownian motion (FBM) is a widely applied stochastic process, particularly for systems exhibiting long-range temporal correlations. Although FBM is characterized by Gaussian increments, heterogeneous processes with FBM-like characteristics may deviate from Gaussianity. In this article, we study the non-Gaussian behavior of switching fractional Brownian motion (SFBM), a model in which the diffusivity of the FBM process varies while temporal correlations are maintained. To characterize non-Gaussianity, we evaluate the kurtosis, a common tool used to quantify deviations from the normal distribution. We derive exact expressions for the kurtosis of the considered heterogeneous anomalous diffusion process and investigate how it can identify non-Gaussian behavior. We also compare the kurtosis results with those obtained using the Hellinger distance, a classical measure of divergence between probability density functions. Through both analytical and numerical methods, we demonstrate the potential of kurtosis as a metric for detecting non-Gaussianity in heterogeneous anomalous diffusion processes.

Figures

Figures reproduced from arXiv: 2501.10472 by the authors.

Figure 1
Figure 1. Evolution of the PDFs of the normalized position in a dichotomous SFBM with exponential dwell times. Each line corresponds to a specific time, with yellow lines representing shorter times and blue lines representing longer times. The plot PDFs correspond to times equal to t/∆ = 10, 20, 50, 100, 200, 400, and 800. (a) Subdiffusive case with Hurst exponent H = 0.3. (b) Superdiffusive case with H = 0.7. On both panels,… view at source ↗
Figure 2
Figure 2. PDFs of the position at time t = 3∆ in a dichotomous SFBM with exponential dwell times (blue line). (a) Subdiffusive case with Hurst exponent H = 0.3. (b) Superdiffusive case with H = 0.7. On both panels, the dashed yellow line depicts the mixture of Gaussian PDFs described, while dash-dotted green line corresponds to a Gaussian with variance 2D−t 2H, and dotted orange line represents a Gaussian with variance 2D+t 2… view at source ↗
Figure 3
Figure 3. Evolution of the PDF of the normalized position in a dichotomous SFBM with one state having a power-law dwell time and the second state, exponential dwell times. Each line corresponds to a specific time, with yellow lines representing shorter times and blue lines representing longer times. The plot PDFs correspond to times equal to t/∆ = 10, 20, 50, 100, 200, 400, and 800. (a) Subdiffusive case with Hurst exponent H… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Kurtosis for the SFBM with Markovian switching (Exp-Exp) with H = 0.3. 5th order (dashed yellow line), 10th order (dash-dotted green line), and 100th order (dashed orange line) are shown, where the order describes the number of terms in Eq. (26). The empirical kurtosis…
Figure 5
Figure 5. Figure 5: Empirical kurtosis for two cases. a) H = 0.3. b) H = 0.7. The lines are calculated based on 100 000 trajectories. The shaded area corresponds to the 95% confidence interval based on 1 000 samples of length 100 000 [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Empirical kurtosis for power-law and exponential dwell times (PL-Exp) of SFBM with H = 0.3. Different lines correspond to different initial conditions (p is the probability of starting in the higher state). The lines are calculated based on 100 000 trajectories. The in…
Figure 7
Figure 7. Figure 7: Empirical Hellinger distance for two cases of SFBM with different types of dwell times. Panel a) corresponds to H = 0.3, b) to H = 0.7. On both panels, the blue lines present the Hellinger distance for the Markovian case, yellow for the dwell times from power-law (lowe…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

99 extracted references · 79 canonical work pages

  1. [1]

    Manzo C, Torreno-Pina J A, Massignan P, Lapeyre Jr G J, Lew enstein M and Garcia Parajo M F 2015 Physical Review X 5 011021

    2015

  2. [2]

    Akimoto T, Barkai E and Saito K 2016 Physical Review Letters 117 180602

    2016

  3. [3]

    Weron A, Burnecki K, Akin E J, Solé L, Balcerek M, Tamkun M M and Krapf D 2017 Scientific Reports 7 5404

    2017

  4. [4]

    Sikora G, Wyłomańska A, Gajda J, Solé L, Akin E J, Tamkun M M and Krapf D 2017 Physical Review E 96 062404

    2017

  5. [5]

    Edelhoff H, Signer J and Balkenhol N 2016 Movement Ecology 4 1–21

    2016

  6. [6]

    Vilk O, Orchan Y, Charter M, Ganot N, Toledo S, Nathan R and Assaf M 2022 Physical Review X 12 031005

    2022

  7. [7]

    Berkowitz B, Klafter J, Metzler R and Scher H 2002 Water Resources Research 38 9–1

    2002

  8. [8]

    Alfarano S, Lux T and Wagner F 2008 Journal of Economic Dynamics and Control 32 101–136

    2008

Show all 99 references
  1. [9]

    Janczura J and Weron R 2013 AStA Advances in Statistical Analysis 97 239–270

    2013

  2. [10]

    Torreno-Pina J A, Manzo C and Garcia-Parajo M F 2016 Journal of Physics D: Applied Physics 49 104002

    2016

  3. [11]

    Muñoz-Gil G, Bachimanchi H, Pineda J, Midtvedt B, Lewen stein M, Metzler R, Krapf D, Volpe G and Manzo C 2023 arXiv preprint arXiv:2311.18100

    2023 arXiv

  4. [12]

    Dieball C, Krapf D, Weiss M and Godec A 2022 New Journal of Physics 24 023004

    2022

  5. [13]

    Jeon J H, Chechkin A V and Metzler R 2014 Physical Chemistry Chemical Physics 16 15811–15817

    2014

  6. [14]

    Miyaguchi T, Uneyama T and Akimoto T 2019 Physical Review E 100 012116

    2019

  7. [15]

    Hidalgo-Soria M, Barkai E and Burov S 2021 Entropy 23 231

    2021

  8. [16]

    Pacheco-Pozo A, Balcerek M, Wyłomanska A, Burnecki K, S okolov I M and Krapf D 2024 Physical Review Letters 133 067102

    2024

  9. [17]

    Metzler R, Jeon J H, Cherstvy A G and Barkai E 2014 Physical Chemistry Chemical Physics 16 24128–24164

    2014

  10. [18]

    Krapf D 2015 Current Topics in Membranes 75 167–207

    2015

  11. [19]

    Norregaard K, Metzler R, Ritter C M, Berg-Sørensen K and Oddershede L B 2017 Chemical Reviews 117 4342–4375

    2017

  12. [20]

    Krapf D and Metzler R 2019 Physics Today 72 48–54 Evaluating Gaussianity of heterogeneous fractional Brown ian motion 21

    2019

  13. [21]

    2022 Communications Physics 5 305

    Sposini V, Krapf D, Marinari E, Sunyer R, Ritort F, Taher i F, Selhuber-Unkel C, Benelli R, Weiss M, Metzler R et al. 2022 Communications Physics 5 305

    2022

  14. [22]

    Kolmogorov A N 1940 Acad. Sci. URSS (NS) 26 115–118

    1940

  15. [23]

    Mandelbrot B B and Van Ness J W 1968 SIAM Review 10 422–437

    1968

  16. [24]

    Deng W and Barkai E 2009 Physical Review E 79 011112

    2009

  17. [25]

    Sabri A, Xu X, Krapf D and Weiss M 2020 Physical Review Letters 125 058101

    2020

  18. [26]

    Janczura J, Balcerek M, Burnecki K, Sabri A, Weiss M and K rapf D 2021 New Journal of Physics 23 053018

    2021

  19. [27]

    Szymanski J and Weiss M 2009 Physical Review Letters 103 038102

    2009

  20. [28]

    Weber S C, Spakowitz A J and Theriot J A 2010 Physical Review Letters 104 238102

    2010

  21. [29]

    Sarfati R and Schwartz D K 2020 ACS Nano 14 3041–3047

    2020

  22. [30]

    Fox Z R, Barkai E and Krapf D 2021 Nature Communications 12 6162

    2021

  23. [31]

    Burnecki K, Kepten E, Janczura J, Bronshtein I, Garini Y and Weron A 2012 Biophysical Journal 103 1839–1847

    2012

  24. [32]

    Kepten E, Bronshtein I and Garini Y 2011 Physical Review E 83 041919

    2011

  25. [33]

    Jeon J H, Tejedor V, Burov S, Barkai E, Selhuber-Unkel C, Berg-Sørensen K, Oddershede L and Metzler R 2011 Physical Review Letters 106 048103

    2011

  26. [34]

    Reverey J F, Jeon J H, Bao H, Leippe M, Metzler R and Selhub er-Unkel C 2015 Scientific Reports 5 11690

    2015

  27. [35]

    Krapf D, Lukat N, Marinari E, Metzler R, Oshanin G, Selhu ber-Unkel C, Squarcini A, Stadler L, Weiss M and Xu X 2019 Physical Review X 9 011019

    2019

  28. [36]

    Afek G, Davidson N, Kessler D A and Barkai E 2023 Reviews of Modern Physics 95 031003

    2023

  29. [37]

    Plerou V, Gopikrishnan P, Amaral L A N, Gabaix X and Stanl ey H E 2000 Physical Review E 62 R3023

    2000

  30. [38]

    Beran J 2013 Long-memory processes (Springer)

    2013

  31. [39]

    Balcerek M, Burnecki K, Thapa S, Wyłomańska A and Chechk in A 2022 Chaos: An Interdisciplinary Journal of Nonlinear Science 32 093114 ISSN 1089-7682

    2022

  32. [40]

    Grzesiek A, Gajda J, Thapa S and Wyłomańska A 2024 Chaos: An Interdisciplinary Journal of Nonlinear Science 34 043154

    2024

  33. [41]

    Woszczek H, Wyłomańska A and Chechkin A 2025 Chaos: An Interdisciplinary Journal of Nonlinear Science 35 023145

    2025

  34. [42]

    Balcerek M, Wyłomańska A, Burnecki K, Metzler R and Krap f D 2023 New Journal of Physics 25 103031

    2023

  35. [43]

    Wang W, Balcerek M, Burnecki K, Chechkin A V, Janušonis S , Ślęzak J, Vojta T, Wyłomańska A and Metzler R 2023 Physical Review Research 5(3) L032025

    2023

  36. [44]

    Pacheco-Pozo A and Krapf D 2024 Physical Review E 110(1) 014105

    2024

  37. [45]

    Chubynsky M V and Slater G W 2014 Physical Review Letters 113 098302

    2014

  38. [46]

    Wang B, Anthony S M, Bae S C and Granick S 2009 Proceedings of the National Academy of Sciences 106 15160–15164

    2009

  39. [47]

    Wang B, Kuo J, Bae S C and Granick S 2012 Nature Materials 11 481–485

    2012

  40. [48]

    Chechkin A V, Seno F, Metzler R and Sokolov I M 2017 Physical Review X 7 021002

    2017

  41. [49]

    Postnikov E B, Chechkin A and Sokolov I M 2020 New Journal of Physics 22 063046

    2020

  42. [50]

    Pastore R, Ciarlo A, Pesce G, Greco F and Sasso A 2021 Physical Review Letters 126 158003

    2021

  43. [51]

    Pacheco-Pozo A and Sokolov I M 2023 The European Physical Journal B 96 152 ISSN 1434-6036

    2023

  44. [52]

    Ribeiro H V, Tateishi A A, Lenzi E K, Magin R L and Perc M 202 3 Communications Physics 6 244

  45. [53]

    Stanislavsky A, Weron K and Weron A 2014 The Journal of Chemical Physics 140

    2014

  46. [54]

    Chechkin A and Sokolov I M 2021 Physical Review E 103 032133

    2021

  47. [55]

    Scher H and Lax M 1973 Physical Review B 7 4491

    1973

  48. [56]

    Scher H and Montroll E W 1975 Physical Review B 12 2455

    1975

  49. [57]

    Dybiec B and Gudowska-Nowak E 2010 Chaos: An Interdisciplinary Journal of Nonlinear Science Evaluating Gaussianity of heterogeneous fractional Brown ian motion 22 20 043129

    2010

  50. [58]

    Painter S 1996 Water Resources Research 32 1183–1195

    1996

  51. [59]

    He W, Song H, Su Y, Geng L, Ackerson B J, Peng H and Tong P 201 6 Nature Communications 7 11701

  52. [60]

    Lanoiselée Y and Grebenkov D S 2018 Journal of Physics A: Mathematical and Theoretical 51 145602

    2018

  53. [61]

    Chakraborty I and Roichman Y 2020 Physical Review Research 2 022020

    2020

  54. [62]

    Shapiro S S and Wilk M B 1965 Biometrika 52 591–611

    1965

  55. [63]

    Jarque C M and Bera A K 1987 International Statistical Review/Revue Internationale de Statistique 55 163–172

    1987

  56. [64]

    D’Agostino R and Pearson E S 1973 Biometrika 60 613–622

    1973

  57. [65]

    Gnedenko BV K A 1954 Limit Distributions of Sums of Independent Random Variables (Cambridge: Addison-Wesley)

    1954

  58. [66]

    Anderson T W 1962 The Annals of Mathematical Statistics 33 1148 – 1159

    1962

  59. [67]

    Kuiper N H 1960 Indagationes Mathematicae (Proceedings) 63 38–47

    1960

  60. [68]

    Watson G S 1961 Biometrika 48 109–114

    1961

  61. [69]

    Anderson T W and Darling D A 1952 The Annals of Mathematical Statistics 23 193 – 212

    1952

  62. [70]

    Lilliefors H W 1967 Journal of the American Statistical Association 62 399–402

    1967

  63. [71]

    Wyłomańska A, Iskander D R and Burnecki K 2020 PLOS One 15 1–36

    2020

  64. [72]

    Yazici B and Yolacan S 2007 Journal of Statistical Computation and Simulation 77 175–183

    2007

  65. [73]

    Das K R and Imon A 2016 American Journal of Theoretical and Applied Statistics 5 5–12

    2016

  66. [74]

    2021 Open Journal of Statistics 11 113

    Khatun N et al. 2021 Open Journal of Statistics 11 113

    2021

  67. [75]

    2011 Journal of Statistical Modeling and Analytics 2 21–33

    Razali N M, Wah Y B et al. 2011 Journal of Statistical Modeling and Analytics 2 21–33

    2011

  68. [76]

    Thadewald T and Büning H 2007 Journal of Applied Statistics 34 87–105

    2007

  69. [77]

    Antoni J 2006 Mechanical Systems and Signal Processing 20 282–307

    2006

  70. [78]

    Hellinger E 1909 Journal Für Die Reine Und Angewandte Mathematik 1909 210–271

    1909

  71. [79]

    Lévy P 1953 Random Functions: General Theory with Special Reference to L aplacian Random Functions University of California Publications in Statistics (Univ ersity of California Press)

    1953

  72. [80]

    Woszczek H, Wyłomańska A and Chechkin A 2024 arXiv preprint arXiv:2410.11546

    2024 arXiv

  73. [81]

    Meroz Y and Sokolov I M 2015 Physics Reports 573 1–29

    2015

  74. [82]

    Vignat C 2012 Statistics & Probability Letters 82 67–71

    2012

  75. [83]

    Graham R L 1994 Concrete Mathematics: a Foundation for Computer Science (Pearson Education India)

    1994

  76. [84]

    Miyaguchi T, Akimoto T and Yamamoto E 2016 Physical Review E 94(1) 012109

    2016

  77. [85]

    Robertson C A and Fryer J G 1969 Scandinavian Actuarial Journal 1969 137–146

    1969

  78. [86]

    Behboodian J 1970 Technometrics 12 131–139

    1970

  79. [87]

    Weigel A V, Tamkun M M and Krapf D 2013 Proceedings of the National Academy of Sciences 110 E4591–E4600

    2013

  80. [88]

    Sadegh S, Barkai E and Krapf D 2014 New Journal of Physics 16 113054

    2014

  81. [89]

    Kurilovich A A, Mantsevich V N, Stevenson K J, Chechkin A V and Palyulin V V 2020 Physical Chemistry Chemical Physics 22 24686–24696

    2020

  82. [90]

    Kurilovich A A, Mantsevich V N, Mardoukhi Y, Stevenson K J, Chechkin A V and Palyulin V V 2022 Physical Chemistry Chemical Physics 24 13941–13950

    2022

  83. [91]

    Margolin G and Barkai E 2004 The Journal of Chemical Physics 121 1566–1577

    2004

  84. [92]

    Stefani F D, Hoogenboom J P and Barkai E 2009 Physics Today 62 34–39

    2009

  85. [93]

    Thiel F, Schimansky-Geier L and Sokolov I M 2012 Physical Review E 86 021117

    2012

  86. [94]

    Beran R 1977 The Annals of Statistics 445–463

    1977

  87. [95]

    Csiszar I 1975 The Annals of Probability 3 146 – 158

    1975

  88. [96]

    Bhattacharyya A 1946 Sankhy¯ a: The Indian Journal of Statistics (1933-1960) 7 401–406

    1946

  89. [97]

    Hidalgo-Soria M, Barkai E and Burov S 2021 Entropy 23

    2021

  90. [98]

    Ghosh S K, Cherstvy A G, Grebenkov D S and Metzler R 2016 New Journal of Physics 18 013027 Evaluating Gaussianity of heterogeneous fractional Brown ian motion 23

    2016

  91. [99]

    Pacheco-Pozo A and Sokolov I M 2021 Phys. Rev. Lett. 127(12) 120601

    2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.