Pith. sign in

REVIEW 4 major objections 5 minor 6 references

Combining quasi-static and high frequency experiments for the viscoelastic characterization of brain tissue

T0 review · 4 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a five-parameter fractional Kelvin–Voigt model unifies quasi-static rheometer data and 300–2100 Hz MRE data for three porcine brain regions into one viscoelastic response curve.

desk verdict Useful regional porcine brain data and a sensible fKV fit, but the 0–2100 Hz "unification" is an interpolation across an unmeasured gap anchored by a model-derived 0 Hz point. read the letter →

arxiv 2601.22743 v1 pith:QRYFMGB2 submitted 2026-01-30 physics.med-ph cond-mat.mtrl-sciphysics.bio-ph

Laura Ruhland , Nina Reiter , Silvia Budday , Kai Willner This is my paper · ORCID
classification physics.med-phcond-mat.mtrl-sciphysics.bio-ph MSC 92C1074D10 PACS 87.19.R83.60.Bc
keywords BraintissueViscoelasticityFractionalKelvin-VoigtmodelMagneticresonanceelastographyRheometryPorcineRegionalmechanicsMulti-scalecharacterization
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

Brain-tissue mechanics has long suffered from contradictory results depending on whether the tissue is loaded slowly or by high-frequency waves, and most studies characterize only one scale. This paper measures the corona radiata, putamen, and thalamus of porcine brain in both regimes—multi-modal rheometry (compression, tension, torsion, and relaxation) plus tabletop magnetic resonance elastography at 300–2100 Hz—and calibrates a fractional Kelvin–Voigt model with two springpot elements in parallel to a spring. The central claim is that this five-parameter model, anchored at 0 Hz by a storage modulus computed from the quasi-static data, reproduces the entire frequency response for all three regions, including the switch from elastic-dominated behavior at low frequencies to viscous-dominated behavior at high frequencies. A sympathetic reader cares because a single mechanical law across time scales would let simulations of surgery, injury, and development use one set of parameters instead of choosing between inconsistent measurements.

What carries the argument

The central object is the fractional Kelvin–Voigt model with two fractional elements (springpots) in parallel with an elastic spring: G*(ω)=μ_e + c1(iω)^{β1} + c2(iω)^{β2}, with 0≤β_i≤1 and β1<β2. Each springpot constant is c_i = μ_i^{1-β_i} η_i^{β_i} with η_i fixed to 1 Pa s, so the five free parameters are μ_e, μ_1, β_1, μ_2, β_2. The model is calibrated by adding the rheometer-derived storage modulus at 0 Hz as an extra data point to the MRE frequency sweep, then minimizing a least-squares residuum. The lower-exponent element (β between 0.34 and 0.46) sets the storage modulus and low-frequency behavior; the higher-exponent element (β between 0.93 and 0.96) sets the loss modulus at high fr

What would settle it

Run small-strain oscillatory shear tests on the same porcine brain regions at intermediate frequencies (roughly 0.1–300 Hz) and compare the measured G' and G'' with the fractional Kelvin–Voigt model's prediction. If the measurements deviate systematically from the model's smooth power-law bridge, or if the 0 Hz anchor shifts with strain amplitude or post-mortem time, the unified characterization fails. A quicker check: perturb the 0 Hz G'(0) value by its estimated uncertainty and see whether the 300–2100 Hz MRE fit still holds.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that a single fractional Kelvin–Voigt model, G*(ω)=μ_e + c1(iω)^{β1} + c2(iω)^{β2}, unifies the viscoelastic response of ex vivo porcine brain tissue. The spring fixes the zero-frequency stiffness; one springpot shapes storage modulus and low-frequency behavior, the other the loss modulus at high frequencies. The 0 Hz anchor comes from the storage modulus computed from the rheometer Prony fits; with it, the model reproduces the MRE data (300–2100 Hz) and the low-frequency points for all three regions. The same regional ordering appears in both domains (corona radiata stiffest), and the loss modulus rises faster than the storage modulus, so the loss tan

Load-bearing premise

The unified curve rests on treating the storage modulus computed from large-strain, nonlinear rheometer fits as the true zero-frequency limit of the small-amplitude linear MRE response, across an unmeasured 0.053–300 Hz gap and using samples with different post-mortem times.

Editorial extensions

If this is right

  • A single five-parameter set describes brain tissue from slow surgical loading to MRE wave frequencies, so finite-element simulations can cover both regimes without re-fitting the material law.
  • Regional ordering detectable in both domains—corona radiata stiffest, putamen and thalamus similar—means whole-brain models should use region-specific parameters at all scales.
  • The model quantitatively predicts the loss tangent crossing from below one to above one, locating where brain tissue turns from elastic-dominated to viscous-dominated, which is relevant for impact and injury simulations.
  • The two springpot exponents separate roles (one for storage/elastic response, one for loss/viscous response), giving a compact way to summarize regional mechanical state.
  • The same bridging strategy worked for three brain regions, suggesting it could be applied to other soft tissues with scale-dependent disagreements.

Reading between the lines

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

  • If the same anchoring procedure were applied to human brain data from indentation and in vivo MRE, it might reconcile many contradictory stiffness values reported in the literature; the authors stop short of making that claim.
  • The model's smooth power-law bridge across the unmeasured 0.053–300 Hz gap is a concrete prediction: small-strain oscillatory shear tests in that range would show whether the true response follows the fractional curve or has additional features.
  • The fitted second springpot shear moduli are extremely small (near 10^-16 kPa for corona radiata), so the high-frequency element is nearly a pure dashpot; whether that degeneracy persists in human tissue or at even higher frequencies is a testable question.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper characterizes three porcine brain regions (corona radiata, putamen, thalamus) with two techniques: multi-modal large-strain rheometer tests (compression/tension/shear) analyzed with a nonlinear Ogden-Prony model, and tabletop MRE at 15 frequencies between 300 and 2100 Hz analyzed with a Bessel-function wave solution. The authors then calibrate a two-element fractional Kelvin-Voigt (fKV) model, G*(ω)=μ_e + Σ c_i(iω)^β_i, using the MRE dynamic moduli together with a single 0 Hz storage-modulus value computed from the Prony fit to the quasi-static rheometer data. They report consistent regional differences across both domains, observe a transition from elastic-dominated to viscous-dominated behavior with increasing frequency, and claim that the calibrated fKV model unifies the two responses into one 'comprehensive characterization' over 0–2100 Hz.

Significance. If the unified-model claim were fully supported, the paper would provide a practical bridge between quasi-static and high-frequency measurements, with region-specific parameters for a single fractional constitutive law. The experimental work is careful and the regional comparison is a strength: the corona radiata is consistently stiffest, and the identified μ_e values reproduce the rheometer-derived 0 Hz anchor. The paper is also transparent about the two-stage calibration and the fact that the MRE data contain no low-frequency information. However, the central claim is currently stronger than the evidence: the 0 Hz anchor is model-derived rather than measured, there are no experimental data between 0.053 Hz and 300 Hz, and the low-frequency loss modulus is not used in the calibration. The model therefore interpolates across the entire intermediate band, and the 'comprehensive characterization' conclusion should be substantially tempered.

major comments (4)
  1. [§2.3.2, Tables 3 and 4] The 0 Hz anchor used in the fKV calibration is not a measured quantity. It is computed from the Ogden-Prony parameters identified in §2.2.2 via Eq. (7), i.e., G'(0)=μ_∞ = μ_0(1-Σg_k). Since the fKV model has G*(0)=μ_e, the resulting μ_e values in Table 4 are essentially identical to the Table 3 anchors (0.053 vs 0.057, 0.063 vs 0.063, 0.052 vs 0.052). The low-frequency branch of the model is therefore enforced by construction, not validated by data. The paper itself notes in §3.3 that μ_e correlates with the rheometer storage modulus 'since those values were added to the data set for the model calibration.' The abstract and conclusion claim a 'unified' and 'comprehensive' characterization; this overstates what is demonstrated. At most, the model provides a constrained interpolation anchored at one model-derived point at 0 Hz and MRE data from 300 Hz upward.
  2. [§2.3.2, Fig. 8; §3.3] There are no experimental data between 0.053 Hz and 300 Hz. The model is fit to one 0 Hz point and to 300–2100 Hz data; the excellent visual agreement in Fig. 8 is an interpolation across an unmeasured band that spans more than three orders of magnitude in frequency. Moreover, the low-frequency loss modulus is entirely unconstrained: the rheometer-derived G'' values at 0.027 and 0.053 Hz (shown in Fig. 6 and computed from Eq. 7) are not included in the calibration, and the 0 Hz anchor carries no G'' information. Consequently, the fKV model can match the endpoint storage modulus and the high-frequency response while misrepresenting the intermediate regime. The claim that the model 'very well captures the response over the whole frequency range' (Sec. 3.3) should be restricted to the measured frequencies, or the authors should add the low-frequency G'' points and/or a sensitivity analysis
  3. [Table 4 and Fig. 10] The fitted μ_2 values are 2.23e-16 kPa (CR), 1.21e-13 kPa (P), and 2.13e-10 kPa (T). These are numerically zero and span six orders of magnitude. With β_2 ≈ 0.93–0.96, the contribution c_2 = μ_2^(1-β_2) remains finite, but the parameter μ_2 itself is nonphysical and the model appears overparameterized for the available data. This is a direct symptom of the missing intermediate-frequency constraints: the second fractional element is tasked to bridge a 0 Hz anchor and high-frequency MRE without any data between. The authors should discuss identifiability, fix or constrain μ_2, or report the effective c_2 instead of μ_2.
  4. [§2.1] Rheometer measurements were conducted 1–8 h post mortem, while MRE measurements were performed 2–3.5 h post mortem. The discussion cites Weickenmeier et al. (2018) for post-mortem stiffening, so the two datasets may represent different tissue states. The 0 Hz anchor depends on the rheometer data, and the high-frequency MRE data on the MRE samples; if post-mortem time affects stiffness, the 'unified' model merges responses from different material states. The authors should either quantify the effect of post-mortem time or explicitly list this as a limitation of the combined calibration.
minor comments (5)
  1. [§4] The discussion states that the storage modulus is 'related to the initial shear modulus μ_0'. This is not correct: the storage modulus at 0 Hz from Eq. (7) is the long-term shear modulus μ_∞ = μ_0(1-Σg_k), not μ_0. The initial shear modulus is the zero-frequency limit of the *relaxed* modulus, not the storage modulus. Please correct this wording.
  2. [§2.3.2] The phrase 'we included the storage modulus ... as an additional parameter at 0 Hz' is confusing: the storage modulus is a data point, not a parameter. Rephrase, e.g., 'we added a single data point at 0 Hz'.
  3. [Fig. 9] The upper panels show the model response from 0 to 100 Hz, but there are no data in this range. Plotting the model as a solid line across the unmeasured gap visually implies validation. Consider shading the interpolated region or plotting only measured frequencies.
  4. [Table A.7] The entry for α_1, P vs T, reads '01'; this is likely a typo for '0.1' or '1'. Please check.
  5. [§2.4 / Appendix A] The p-values in the appendix are given without indicating which comparisons remain significant after Holm-Bonferroni correction. The text mentions the correction, but the tables show raw p-values. Please mark corrected significance levels.

Circularity Check

1 steps flagged · score 6.0 of 10

Low-frequency unification is by construction: μ_e is calibrated to the rheometer-derived 0 Hz anchor, so the model's G′(0) is its own input.

  1. self definitional [Sec. 2.3.2 (Eq. 9), Sec. 3.3, Tables 3–4]
    "However, the measured frequency response does not contain any information about the response in the lower frequency regime. Therefore, we included the storage modulus obtained from the calibration of the rheometer measurements with Eq.7 as an additional parameter at 0 Hz to the MRE data. ... The elastic shear moduli μ_e of the three brain regions correlate with the storage moduli obtained from the rheometer parameter identification in Tab. 3, since those values were added to the data set for the model calibration, as described in Sec. 2.3.2."

    In the fractional Kelvin-Voigt model (Eq. 9), G*(0) = μ_e. The 0 Hz storage modulus (Table 3) is not a direct measurement but is computed from the rheometer-fitted Prony parameters via Eq. 7, then appended as the only low-frequency point in the calibration data vector x = {x_rheo, x_1, ..., x_n}. Calibrating μ_e to this exact point makes the model's low-frequency limit equal to the input anchor by construction; Table 4 μ_e (0.053/0.063/0.052 kPa) reproduces Table 3 (0.057/0.063/0.052 kPa). The reported agreement over the 'whole frequency range' is therefore not an independent test at the low-frequency end, and the 0.053–300 Hz interval is unconstrained.

full rationale

The derivation chain is mostly an honest calibration rather than a first-principles derivation. The rheometer data are fitted with an Ogden-Prony model (Eqs. 2, 5–7), the MRE data are fitted with a Bessel-function wave solution, and the fractional Kelvin-Voigt model (Eq. 9) is calibrated to the MRE moduli plus a single 0 Hz storage-modulus point computed from the Prony parameters (Table 3). Because G*(0) = μ_e in Eq. 9, the low-frequency endpoint of the 'unified' curve is the calibration input itself; the paper explicitly acknowledges this in Sec. 3.3, which is why the circularity is partial rather than hidden. The high-frequency agreement is a fit to independent MRE data and is not circular, and no load-bearing self-citation or imported uniqueness theorem appears. The main overreach is presenting the fit as a wide-band 'comprehensive characterization' across a 0.053–300 Hz gap that no data constrain; that gap and the post-mortem time differences are correctness risks, but the low-frequency agreement is also partly by construction, yielding a score of 6.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central contribution is a calibration, so the ledger is dominated by fitted parameters. Ten Ogden-Prony parameters are fitted to the rheometer data; the five fractional Kelvin-Voigt parameters are fitted to the MRE data plus a rheometer-derived 0 Hz anchor; η=1 Pa s and ν=0.49 are chosen by hand. The main structural assumptions are linearity across strain regimes, homogeneity/isotropy, the Bessel solution for MRE, and combining tissue from different post-mortem times. No invented entities appear.

free parameters (5)
  • Ogden-Prony quasi-static parameters (μ1, α1, μ2, α2, g1, g2, k1, k2, τ1, τ2) = CR averaged: 0.14, 12.84, 0.17, -24.26, 0.78, 0.08, 0.67 s, 0.05, 0.6, 67.62 s
    Identified by inverse ABAQUS fitting to rheometer compression/tension/shear data (Sec. 2.2.2, Eq. 8, Tables 1-2).
  • Fractional Kelvin-Voigt parameters (μ_e, μ1, β1, μ2, β2) = CR averaged: 0.053 kPa, 0.689 kPa, 0.463, 2.23e-16 kPa, 0.964
    Calibrated to combined 0 Hz rheometer anchor plus MRE G',G'' (Sec. 2.3.2, Table 4). μ2 is effectively zero for CR.
  • Springpot viscosity η = 1 Pa s
    Set by hand because μ_i and η_i are linearly dependent (Sec. 2.3.2, Eq. 10); this normalization determines the reported μ_i values.
  • Poisson ratio ν = 0.49
    Fixed in Eq. 4 to compute bulk modulus; chosen from Hinrichsen et al. (2023), not identified from data (Sec. 2.2.2).
  • 0 Hz storage modulus anchor G'(0) = CR 0.057, P 0.063, T 0.052 kPa
    Not measured; calculated with Eq. 7 from fitted Prony parameters and used as a data point in the fractional calibration (Sec. 2.3.2, Table 3).
assumptions (6)
  • domain assumption Samples are homogeneous and isotropic at the measured scale; direction-dependent white-matter behavior is neglected.
    Isotropic Ogden model (Eq. 2) is applied to corona radiata despite known axon orientation effects (Reiter et al., 2021, 2025).
  • domain assumption The linear-theory storage modulus computed from nonlinear large-strain rheometer data is the correct low-frequency limit of the linear MRE response.
    Sec. 2.2.2/2.3.2: G'(0) from the Prony Fourier transform (Eq. 7) is spliced to MRE data across a 0.053 Hz to 300 Hz gap.
  • ad hoc to paper A two-element fractional Kelvin-Voigt model with β1<β2 is an adequate representation over 0-2100 Hz.
    Model form and inequality constraint are assumed from Bonfanti et al. (2020), not derived or tested against alternatives (Sec. 2.3.2).
  • domain assumption MRE wave fields are described by the Bessel-function solution for a homogeneous, isotropic, linear-viscoelastic cylinder.
    Sec. 2.3.2: dynamic moduli are obtained by fitting an analytical Bessel solution to wave images (Braun et al., 2018).
  • domain assumption Rheometer (1-8 h post mortem) and MRE (2-3.5 h post mortem) samples from different brains can be combined as one material state.
    Sec. 2.1: post-mortem times differ; brain stiffens post mortem (Weickenmeier et al., 2018), yet the unified model treats both datasets as the same tissue.
  • standard math Fourier/fractional calculus identities used to convert Prony series and fractional elements to G'(ω), G''(ω) are standard.
    Eqs. 7 and 9 rely on conventional Fourier transforms of exponential and fractional power-law kernels; no proof is provided, but the identities are standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Combining quasi-static and high frequency experiments for the viscoelastic characterization of brain tissue." pith.science (2026). https://pith.science/paper/QRYFMGB2

@misc{pith2026260122743,
  author       = {Pith},
  title        = {Pith review of: Combining quasi-static and high frequency experiments for the viscoelastic characterization of brain tissue},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QRYFMGB2}},
  note         = {Machine review of arXiv:2601.22743}
}
read the original abstract

Mechanical models of brain tissue are a beneficial tool to simulate neurosurgical interventions, disease progression, or brain development. However, the accuracy and predictive capacity of such a model relies on a precise experimental characterization of the tissue's mechanical behavior. Such a characterization is yet limited by inconsistent or contradictory experimental responses reported in the literature, particularly when measurements are performed in different time or length scales. Although brain tissue has been extensively investigated in previous studies, the combination of experimental findings from different scales has received limited attention. In this study, we combine ex vivo mechanical responses of porcine brain tissue obtained at different time scales in a mechanical model. We investigated the mechanical behavior of three different brain regions in the quasi-static domain with multi-modal large strain rheometer measurements and at high frequencies with magnetic resonance elastography (MRE). A comparative analysis of the mechanical parameters obtained from both experimental techniques demonstrated consistent regional variations in the viscoelastic behavior across the two domains. However, the mechanical behavior changes from a higher elasticity in the quasi-static and low frequency domain to a dominating viscosity at high frequencies. Based on the quasi-static and the high frequency behavior, we calibrated a fractional Kelvin-Voigt model and consequently unified the two responses in a single mechanical model to obtain a comprehensive characterization of the tissue's mechanical behavior.

Figures

Figures reproduced from arXiv: 2601.22743 by the authors.

Figure 1
Figure 1. Exemplary samples of a) the rheometer experiment on brain, b) the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Comparison of the averaged storage and loss modulus for the corona [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Comparison of the stress response obtained in multi-modality rheometer experiments for the corona radiata (CR), the putamen (P) and the thalamus (T). [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Comparison of the viscoelastic parameters. Prony parameters [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Comparison of the storage and loss modulus calculated from the Ogden-Prony parameters for the corona radiata (CR), the putamen (P) and the thalamus [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Mean storage modulus G’, loss modulus G” and loss tangent [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Fractional Kelvin-Voigt model calibrated for the averaged frequency data from 0 Hz to 2100 Hz for the corona radiata, the putamen and the thalamus. The [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Contribution of the individual fractional elements to the overall storage (G’) and loss modulus (G”) of the putamen. The low frequency response (0 Hz to [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: Comparison of the viscoelastic parameters in the frequency domain. A fractional Kelvin-Voigt model with two fractional elements (shear modulus [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

6 extracted references · 2 canonical work pages

  1. [1192]

    Bayly, P.V ., Taber, L.A., Kroenke, C.D., 2014

    doi:10.1002/mrm.26207. Bayly, P.V ., Taber, L.A., Kroenke, C.D., 2014. Mechani- cal forces in cerebral cortical folding: a review of mea- surements and models. Journal of the mechanical behav- ior of biomedical materials 29, 568–581. doi:10.1016/ j.jmbbm.2013.02.018. BenSaïda, A., 2025. Shapiro-Wilk and Shapiro-Francia normality tests. URL:https://de.math...

    work page doi:10.1002/mrm.26207 2014

  2. [2018]

    Magnetic Reso- nance in Medicine 79, 470–478

    A compact 0.5 T MR elastography device and its ap- plication for studying viscoelasticity changes in biological tissues during progressive formalin fixation. Magnetic Reso- nance in Medicine 79, 470–478. doi:10.1002/mrm.26659. Budday, S., Nay, R., De Rooij, R., Steinmann, P., Wyrobek, T., Ovaert, T.C., Kuhl, E., 2015a. Mechanical properties of gray and wh...

    arXiv 2015

  3. [2019]

    Journal of The Royal Society Interface 16, 20190356

    Prion-like spreading of Alzheimer’s disease within the brain’s connectome. Journal of The Royal Society Interface 16, 20190356. doi:10.1098/rsif.2019.0356. Forte, A.E., Gentleman, S.M., Dini, D., 2017. On the character- ization of the heterogeneous mechanical response of human brain tissue. Biomechanics and Modeling in Mechanobiol- ogy 16, 907–920. doi:10...

    arXiv 2019

  4. [2020]

    Human Brain Mapping 41, 5282–5300

    Standard-space atlas of the viscoelastic properties of the human brain. Human Brain Mapping 41, 5282–5300. doi:10.1002/hbm.25192. Ipek-Ugay, S., Drießle, T., Ledwig, M., Guo, J., Hirsch, S., Sack, I., Braun, J., 2015. Tabletop magnetic resonance elas- tography for the measurement of viscoelastic parameters of small tissue samples. Journal of Magnetic Reso...

    arXiv 2015

  5. [2021]

    Journal of the Mechanical Behav- ior of Biomedical Materials 120, 104587

    Measuring viscoelastic parameters in Magnetic Res- onance Elastography: a comparison at high and low mag- netic field intensity. Journal of the Mechanical Behav- ior of Biomedical Materials 120, 104587. doi:10.1016/ j.jmbbm.2021.104587. Appendix A. Result of the statistical analysis See Tab. A.5 to A.9. Table A.5: Kruskal-Wallisp-value for rheometer test ...

    arXiv 2021

  6. [2022]

    Current protocols 2, e381

    Tissue-Scale Biomechanical Testing of Brain Tissue for the Calibration of Nonlinear Material Models. Current protocols 2, e381. doi:10.1002/cpz1.381. Fornari, S., Schäfer, A., Jucker, M., Goriely, A., Kuhl, E.,

    work page doi:10.1002/cpz1.381

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.