Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Assessing parameter identifiability of a hemodynamics PDE model using spectral surrogates and dimension reduction

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that polynomial chaos surrogates can supply both Sobol sensitivity rankings and profile-likelihood identifiability tests for a PDE-based pulmonary hemodynamics model, and that these identifiability conclusions match the…

desk verdict A plausible proof-of-concept for PCE-based profile-likelihood identifiability, but the simulator check never re-optimizes nuisance parameters, so the agreement claim is under-supported. read the letter →

arxiv 2506.04538 v1 pith:A7W2AEAV submitted 2025-06-05 q-bio.TO stat.APstat.CO

Mitchel J. Colebank This is my paper
classification q-bio.TOstat.APstat.CO MSC 62F1062H2565M0692C35
keywords polynomialchaosexpansionprofilelikelihoodparameteridentifiabilitySobolsensitivityanalysisprincipalcomponenthemodynamicspulmonarycirculationexperimentaldesign
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

The paper argues that a single polynomial chaos expansion (PCE) surrogate, built on a principal-component reduction of time-series outputs, can serve two purposes for expensive PDE-based blood-flow simulators: it ranks parameter influence through Sobol' indices and it supports formal parameter identifiability tests through profile-likelihood confidence intervals. Applied to a one-dimensional pulmonary hemodynamics model with Windkessel and structured-tree boundary conditions, the approach identifies which parameters can be inferred from three realistic measurement designs. The central finding is that identifiability conclusions drawn from the surrogate agree with those of the true PDE simulator, so the expensive simulator need not be run repeatedly for profile-likelihood analysis. This matters because profile-likelihood is the standard frequentist way to detect non-identifiability, but it has been computationally out of reach for PDE models.

What carries the argument

The load-bearing object is the PCE-PCA spectral surrogate: a polynomial chaos expansion (a spectral polynomial approximation orthogonal with respect to the parameter prior, here Legendre polynomials on uniform priors) fit by regression to the scores of the first five principal components of each output time series. Sobol' indices are read off the PCE coefficients, and the profile-likelihood negative log-likelihood is computed from the inverse-transformed surrogate output, with 95% confidence intervals set by the chi-square threshold on the profile-likelihood difference.

What would settle it

Run the true PDE simulator's profile likelihood for the structured-tree three-parameter set ($k_3$, $\alpha$, $\ell_{rr}$) under design D1, re-optimizing the remaining parameters at every profiled value; if the true simulator's finite confidence bounds appear where the surrogate's are flat, or vice versa, then the surrogate-based identifiability claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that PCE surrogates, after dimension reduction via PCA, enable profile-likelihood confidence intervals for PDE model parameters and that these intervals classify parameters as identifiable or non-identifiable in the same way the full simulator would. For the Windkessel model, the stiffness parameters $k_1$ and $k_2$ are practically non-identifiable in all designs, while the design that adds flow data (D3) makes $k_3$, the proximal and distal resistances, and $C_{T,1}$ identifiable. For the structured tree model, inference requires progressive parameter fixing, with $k_3$, $\alpha$, and $\ell_{rr}$ identifiable under D3 and $\beta$ and $r_{\min}$ harder to identify. The paper supports the surrogate's reliability by evaluating the true simulator at the profile points and reporting similar confidence bounds, despite some differences in likelihood shape.

Load-bearing premise

The whole analysis rests on the surrogate reproducing the true simulator's likelihood surface closely enough that identifiability classifications agree; the most exposed point is the structured-tree pressure output, where emulation errors reach outliers of 9-15%.

Editorial extensions

If this is right

  • For PDE simulators too expensive for repeated optimization, profile-likelihood identifiability becomes computable from a one-time surrogate training set.
  • Measurement design directly controls practical identifiability: pressure data alone leaves most parameters non-identifiable, while adding flow data renders most Windkessel and structured-tree parameters practically identifiable.
  • Parameters that are non-influential according to Sobol' indices ($k_1$ and $k_2$) can be fixed first, and the profile-likelihood then shows which remaining parameters are inferable.
  • The surrogate-derived confidence bounds agree in shape and classification with simulator evaluations at the profile points, supporting the use of emulator-based identifiability screening before costly inference.

Reading between the lines

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

  • One extension the paper leaves implicit is reusing the same PCE training set under different measurement-noise covariances, which would turn the profile-likelihood test into a fast experimental-design optimization loop.
  • Because the simulator check evaluates the emulator's profile points without re-optimizing nuisance parameters on the true simulator, the agreement claim would be strengthened by a full simulator re-optimization at each profiled value; the structured-tree results suggest this could matter in rough landscapes.
  • For models with strong correlation across output channels, a multi-output emulator that models cross-output covariance could reduce the emulator error that currently appears as pressure outliers in the structured-tree case.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This paper proposes a PCE-PCA surrogate pipeline for a one-dimensional pulmonary hemodynamics PDE model, using the surrogate both for global sensitivity analysis (Sobol indices) and for profile-likelihood-based practical identifiability assessment across three experimental designs: MPA pressure only (D1), pressure plus daughter-branch area (D2), and pressure plus daughter-branch flow (D3). Windkessel and structured-tree boundary-condition versions are treated separately. The surrogate accuracy is assessed on 50 held-out simulator runs, with median MSRE around 2% and larger structured-tree pressure outliers (9--15%). The identifiability conclusions from the surrogate are then compared with simulator evaluations along the surrogate's profile-likelihood paths in Figures 11 and 12.

Significance. If the proposed equivalence between surrogate and simulator likelihood geometry holds, the paper offers a computationally practical route to formal identifiability analysis for expensive PDE simulators and a useful comparison of experimental designs for pulmonary hemodynamics. The paper's strengths include the public code, the explicit test-set emulator accuracy assessment, the use of both full and reduced parameter sets in the profile-likelihood analysis, and the attempt to check surrogate-based conclusions against the true simulator on selected parameter subsets. However, the validation is weaker than claimed because the simulator check does not recompute the true profile likelihood, and the sensitivity-analysis formulas contain a normalization/index-set error that may affect the pointwise Sobol results.

major comments (3)
  1. [Profile-likelihood: Simulator, Figures 11-12] The simulator comparison does not test the central claim that identifiability conclusions agree. Each simulator curve is generated by passing the profile-likelihood parameters to the simulator and computing the same weighted likelihood, but this does not re-optimize over nuisance parameters for each fixed profiled parameter. The resulting curve is the joint negative log-likelihood along the surrogate's solution path, not the simulator's profile likelihood. The discrepancies the paper itself notes, such as the biased minimum for R_p,2 in D2 and the rougher landscape for l_rr in D1, are exactly where re-optimization could change whether a confidence bound is finite. To support the statement that 'the conclusions about identifiability agree between the surrogate framework and the true simulator,' the authors should compute true profile likelihoods on the simulator, with re-minimization over nuisance parameters, for at least representative profiled parameters and designs, or alternatively soften the claim to agreement of the joint likelihood along the surrogate path.
  2. [Global Sensitivity Analysis, Eqs. (26)-(27)] The covariance formulas used for pointwise Sobol indices are incorrect as written. Equation (26) omits the polynomial normalization factor gamma_k that appears in Eq. (19), and Eq. (27) sums over A_{ST_i}, the set of coefficients that include theta_i, whereas the conditional expectation E[c_q | theta_{~i}] involves basis functions independent of theta_i, i.e., the complement of A_{ST_i}. Consequently Eqs. (24)-(25) and the pointwise Sobol indices in Figures 5 and 6 may be wrong. The authors should correct these formulas to include gamma_k and the appropriate index sets, and rerun or justify the pointwise sensitivity results.
  3. [Profile-likelihood Confidence Intervals and Limitations] The statistical interpretation of the confidence intervals needs clarification. Equation (28) treats Sigma as a measurement noise covariance, but the data are described as noise-free and Sigma is set to a diagonal matrix of mean data values; the Limitations section then says the inherent emulator error is used as the measurement variance. These statements are inconsistent. The chi-square threshold in Eq. (30) is only a valid 95% confidence bound if Sigma is the true covariance of the noise process. As written, the reported intervals are thresholds of a weighted SSE with application-specific scaling. The authors should either specify a coherent noise model with calibrated Sigma, simulate noisy data with known variance, or explicitly qualify the 'formal' identifiability language.
minor comments (4)
  1. [Methods, Eq. (6)] The Windkessel parameter set lists R_{p,2} twice; the second occurrence should be R_{d,2}.
  2. [Profile-likelihood: Windkessel Model] The text references 'Figure 3(a)', 'Figure 3(b)', and 'Figure 3(c)' when discussing the profile-likelihood results; these should be 'Figure 7(a)', 'Figure 7(b)', and 'Figure 7(c)'.
  3. [Typos and notation] There are several typographical errors: 'infleutial' should be 'influential', 'evlauted' should be 'evaluated', 'structed' should be 'structured', 'id enitifiable' should be 'identifiable', and 'Figure 4(b,d,F)' should use a lowercase 'f'.
  4. [Global Sensitivity Analysis, Figure 5] The text notes that total-order Sobol indices S_{Ti}^Y(t) exceed 1.0 in the MPA and attributes this to numerical approximation error in the PCA decomposition. Since total-order indices should lie in [0,1], this behavior should be explained more precisely, or the affected output should be clearly marked as showing an artifact.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the surrogate-based profile-likelihood analysis is self-contained, and the simulator comparison, while not a full re-optimization, is an external check rather than a reduction to the fit.

full rationale

The profile-likelihood intervals are computed from a PCE-PCA emulator fitted to PDE simulator outputs via ordinary least squares (Eq. 11), and the Sobol indices are obtained from the same PCE coefficients by standard spectral formulas (Eqs. 19-25); neither quantity is defined in terms of the other, and the identifiability classification (bounded/unbounded confidence intervals via Eq. 30) is a post-processing rule applied to the emulator likelihood, not a fitted parameter renamed as a prediction. The paper's check against the true simulator (Figs. 11-12) passes the surrogate's profiled parameter values to the simulator and evaluates the same weighted likelihood, which is an external evaluation rather than a circular reduction; the fact that this does not re-minimize over nuisance parameters is a limitation of the validation, but not an instance of the conclusion being equivalent to its inputs by construction. The only noticeable self-citation is the choice of a degree-five PCE and PCA compression, justified by Paun et al. [17], but the paper independently reports test-set MSRE in Figure 3 and variance capture in Table 1, so the self-citation is not load-bearing. No step in the derivation defines a predicted quantity in terms of the target result, fits a parameter and then calls a closely related output a prediction, or imports a uniqueness conclusion from the author's own prior work.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The central claim rests on standard PCE, Sobol, and profile-likelihood mathematics; on the hemodynamics PDE model and boundary condition assumptions taken from prior literature; and on an ad hoc measurement covariance choice. The paper introduces no new physical entities or parameters beyond the surrogate construction choices.

free parameters (4)
  • PCE polynomial order = 5
    Chosen by hand, based on prior work [17]; controls surrogate accuracy.
  • Number of PCA components retained = 5
    Chosen to capture at least ~99% of output variance (Table 1); truncation affects surrogate fidelity.
  • Noise covariance scaling = diagonal of average pressure/flow/area values per design
    Ad hoc nondimensionalization; no actual measurement noise is modeled, and no sensitivity to this choice is tested.
  • Uniform prior parameter ranges = not specified in manuscript
    The ranges for k1, k2, k3, Rp, Rd, CT (or structured tree parameters) are not listed; Sobol indices and profile-likelihood bounds depend on these ranges.
assumptions (7)
  • domain assumption The 1D PDE hemodynamics model assumes laminar, Newtonian, incompressible, axially dominant flow in cylindrical impermeable vessels, with a power-law velocity profile (γ=9).
    Standard 1D hemodynamics assumptions from [18,21] define the simulator used throughout.
  • domain assumption The pressure-area relation is linear with exponential stiffness Eh/r0 = k1 exp(-k2 r0) + k3.
    Constitutive model from [18]; parameters k1,k2,k3 are among those assessed for identifiability.
  • domain assumption Boundary conditions are either 3-element Windkessel or structured tree impedance models.
    Standard distal vasculature models from [18,24].
  • standard math Polynomial chaos expansion with uniform priors and Legendre polynomials; parameters are assumed independent in their prior space.
    Foundations of PCE and Sobol decomposition; the independence assumption enables the hierarchical variance decomposition.
  • standard math Profile-likelihood confidence intervals use a chi-square threshold with one degree of freedom at 95% significance.
    Standard frequentist method for practical identifiability [12].
  • ad hoc to paper Measurement noise covariance is set to a diagonal matrix of average data values, and data are treated as noise-free.
    The paper acknowledges this is application-specific and not a true measurement error model (Limitations section).
  • domain assumption Structural identifiability of the model itself is assumed, and only practical identifiability is assessed.
    The paper does not analyze structural identifiability, so any structural non-identifiability would confound the practical identifiability conclusions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Assessing parameter identifiability of a hemodynamics PDE model using spectral surrogates and dimension reduction." pith.science (2026). https://pith.science/paper/A7W2AEAV

@misc{pith2026250604538,
  author       = {Pith},
  title        = {Pith review of: Assessing parameter identifiability of a hemodynamics PDE model using spectral surrogates and dimension reduction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A7W2AEAV}},
  note         = {Machine review of arXiv:2506.04538}
}
read the original abstract

Computational inverse problems for biomedical simulators suffer from limited data and relatively high parameter dimensionality. This often requires sensitivity analysis, where parameters of the model are ranked based on their influence on the specific quantities of interest. This is especially important for simulators used to build medical digital twins, as the amount of data is typically limited. For expensive models, such as blood flow models, emulation is employed to expedite the simulation time. Parameter ranking and fixing using sensitivity analysis are often heuristic, though, and vary with the specific application or simulator used. The present study provides an innovative solution to this problem by leveraging polynomial chaos expansions (PCEs) for both multioutput global sensitivity analysis and formal parameter identifiability. For the former, we use dimension reduction to efficiently quantify time-series sensitivity of a one-dimensional pulmonary hemodynamics model. We consider both Windkessel and structured tree boundary conditions. We then use PCEs to construct profile-likelihood confidence intervals to formally assess parameter identifiability, and show how changes in experimental design improve identifiability. Our work presents a novel approach to determining parameter identifiability and leverages a common emulation strategy for enabling profile-likelihood analysis in problems governed by partial differential equations.

Figures

Figures reproduced from arXiv: 2506.04538 by the authors.

Figure 1
Figure 1. Schematic of the mathematical models, including three large pulmonary vessels attached to either (a) Windkessel boundary conditions or (b) structured tree boundary conditions. The variables in red denote the parameters to be inferred, as well as their mathematical representations in the boundary conditions. Variables in purple are state variables of the model. vascular junction. At the distal end of the large vessel… view at source ↗
Figure 2
Figure 2. Example one-dimensional profile-likelihood results and upper bounds for identifiability classification. An example identifiability cutoff given by the maximum likelihood estimate, −LL(θMLE), and chi-squared statistics, ∆(a)/2 are provided in blue. A flat profile-likelihood (a) suggests structural or practical identifiability issues, whereas a profile-likelihood that is bounded on one side (b) by the likelihood thres… view at source ↗
Figure 3
Figure 3. Emulator accuracy on 50 test data sets for pressure in the MPA (P1), flows in the LPA and RPA (Q2 and Q3), and areas in the LPA and RPA (A2 and A3). (a) Windkessel boundary conditions. (b) Structured tree boundary conditions. component, with the compliance parameters being minimally influentially as well. The second and third principal components are more sensitive to k3, Rp,1 and Rp,2, with the high principal compo… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Sobol’ indices for the first three principal components (PCs) using Windkessel boundary conditions (a,c,e) and structured tree boundary conditions (b,d,f). Light bars indicate the median first-order index, S q i , across the three vessels while dark grey represent the …
Figure 5
Figure 5. Figure 5: Point-wise Sobol’ indices for the Windkessel boundary conditions calculated from principal component Sobol’ metrics S q i and S q Ti via eq (24). Results are presented for the (a) MPA, (b) LPA, and (c) RPA in the pulmonary tree. appears moderately influential as was th…
Figure 6
Figure 6. Figure 6: Point-wise Sobol’ indices for the Structured tree boundary conditions calculated from principal component Sobol’ metrics S q i and S q Ti via eq (24). Results are presented for the (a) MPA, (b) LPA, and (c) RPA in the pulmonary tree. practically identifiable due to its…
Figure 7
Figure 7. Figure 7: Profile-likelihood results using the PCA-PCE spectral surrogate. (a) Profile likelihood calculated using the three different experimental designs defined previously. The pointwise confidence intervals (defined in eq (29)) define whether parameters are considered identi…
Figure 8
Figure 8. Figure 8: Model evaluations along the profile likelihood presented in [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: Profile-likelihood results using the PCA-PCE spectral surrogate. (a) Profile likelihood calculated using the three different experimental designs defined in previously. The pointwise confidence intervals (defined in eq (29)) define whether parameters are considered ide…
Figure 10
Figure 10. Figure 10: Model evaluations along the profile likelihood presented in [PITH_FULL_IMAGE:figures/full_fig_p020_10.png]
Figure 11
Figure 11. Figure 11: Comparison between the spectral surrogate (red) and simulator evlauted (cyan) profile-likelihood confidence intervals for the Windkessel boundary conditions using the free parameter set including k3, Rp,2, Rd,2, and CT ,2. Results are shown for designs D1 (a), D2 (b),…
Figure 12
Figure 12. Figure 12: Comparison between the spectral surrogate (red) and simulator evaluated (cyan) profile-likelihood confidence intervals for the structured tree boundary conditions using the free parameter set including k3, α, and ℓrr. Results are shown for designs D1 (a), D2 (b), and …

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Think before you fit: parameter identifiability, sensitivity and uncertainty in systems biology models

    stat.ME 2025-08 conditional novelty 3.0 of 10

    A review arguing that identifiability analysis is a prerequisite for reliable parameter estimation and prediction in systems biology, and summarizing available computational tools.

Reference graph

Works this paper leans on

38 extracted references · 34 canonical work pages · cited by 1 Pith paper

  1. [1]

    Challenges and opportunities in uncertainty quantification for healthcare and biological systems

    Kimpton LM, Paun LM, Colebank MJ, Volodina V. Challenges and opportunities in uncertainty quantification for healthcare and biological systems. Philosophical Transactions A. 2025;383(2292):20240232

    work page 2025

  2. [2]

    Guidelines for mechanistic modeling and analysis in cardiovascular research

    Colebank MJ, Oomen PA, Witzenburg CM, Grosberg A, Beard DA, Husmeier D, et al. Guidelines for mechanistic modeling and analysis in cardiovascular research. American Journal of Physiology-Heart and Circulatory Physiology. 2024;327(2):H473–H503

    work page 2024

  3. [3]

    Foundational Research Gaps and Future Directions for Digital Twins

    National Academies for Science, Engineering, and Medicine. Foundational Research Gaps and Future Directions for Digital Twins. National Academies Press; 2024

    work page 2024

  4. [4]

    A guide to uncertainty quantification and sensitivity analysis for cardiovascular applications

    Eck VG, Donders WP, Sturdy J, Feinberg J, Delhaas T, Hellevik LR, et al. A guide to uncertainty quantification and sensitivity analysis for cardiovascular applications. International Journal for Numerical Methods in Biomedical Engineering. 2016;32

    work page 2016

  5. [5]

    Uncertainty quantification: theory, implementation, and applications

    Smith RC. Uncertainty quantification: theory, implementation, and applications. SIAM; 2024

    work page 2024

  6. [6]

    Sensitivity analysis: A review of recent advances

    Borgonovo E, Plischke E. Sensitivity analysis: A review of recent advances. European Journal of Operational Research. 2016;248(3):869–887

    work page 2016

  7. [7]

    Variance-based sensitivity analysis for time-dependent processes

    Alexanderian A, Gremaud PA, Smith RC. Variance-based sensitivity analysis for time-dependent processes. Reliability Engineering and System Safety. 2020;196:106722

    work page 2020

  8. [8]

    Principal component analysis and sparse polynomial chaos expansions for global sensitivity analysis and model calibration: Application to urban drainage simulation

    Nagel JB, Rieckermann J, Sudret B. Principal component analysis and sparse polynomial chaos expansions for global sensitivity analysis and model calibration: Application to urban drainage simulation. Reliability Engineering and System Safety. 2020;195. doi:10.1016/j.ress.2019.106737

    arXiv 2020

Show all 38 references
  1. [9]

    Effects of arterial wall models and measurement uncertainties on cardiovascular model predictions

    Eck VG, Sturdy J, Hellevik LR. Effects of arterial wall models and measurement uncertainties on cardiovascular model predictions. Journal of Biomechanics. 2017;50:188–194

    2017

  2. [10]

    Applicability of the polynomial chaos expansion method for personalization of a cardiovascular pulse wave propagation model

    Huberts W, Donders WP, Delhaas T, van de Vosse FN. Applicability of the polynomial chaos expansion method for personalization of a cardiovascular pulse wave propagation model. International Journal for Numerical Methods in Biomedical Engineering. 2014;30:1679–1704

    2014

  3. [11]

    Efficient uncertainty quantification in a spatially multiscale model of pulmonary arterial and venous hemodynamics

    Colebank MJ, Chesler NC. Efficient uncertainty quantification in a spatially multiscale model of pulmonary arterial and venous hemodynamics. Biomechanics and Modeling in Mechanobiology. 2024;23:1909–1931. doi:10.1007/s10237-024-01875-x

    2024 doi

  4. [12]

    On structural and practical identifiability

    Wieland FG, Hauber AL, Rosenblatt M, T¨ onsing C, Timmer J. On structural and practical identifiability. Current Opinion in Systems Biology. 2021;25:60–69. doi:10.1016/j.coisb.2021.03.005

    2021 doi

  5. [13]

    An in-silico analysis of experimental designs to study ventricular function: A focus on the right ventricle

    Colebank MJ, Chesler NC. An in-silico analysis of experimental designs to study ventricular function: A focus on the right ventricle. PLOS Computational Biology. 2022;18:e1010017. doi:10.1371/journal.pcbi.1010017. September 23, 2025 28/30

    2022 doi

  6. [14]

    Structural identifiability analysis of age-structured PDE epidemic models

    Renardy M, Kirschner D, Eisenberg M. Structural identifiability analysis of age-structured PDE epidemic models. Journal of Mathematical Biology. 2022;84(1):9

    2022

  7. [15]

    Murphy RJ, Maclaren OJ, Simpson MJ. Implementing measurement error models with mechanistic mathematical models in a likelihood-based framework for estimation, identifiability analysis and prediction in the life sciences. Journal of the Royal Society Interface. 2024;21(210):20230402

    2024

  8. [16]

    Parameter identifiability in PDE models of fluorescence recovery after photobleaching

    Ciocanel MV, Ding L, Mastromatteo L, Reichheld S, Cabral S, Mowry K, et al. Parameter identifiability in PDE models of fluorescence recovery after photobleaching. Bulletin of Mathematical Biology. 2024;86(4):36

    2024

  9. [17]

    A comparison of Gaussian processes and polynomial chaos emulators in the context of haemodynamic pulse–wave propagation modelling

    Paun LM, Colebank MJ, Husmeier D. A comparison of Gaussian processes and polynomial chaos emulators in the context of haemodynamic pulse–wave propagation modelling. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences. 2025;383:202...

    2025

  10. [18]

    Numerical simulation and experimental validation of blood flow in arteries with structured-tree outflow conditions

    Olufsen MS, Peskin CS, Kim WY, Pedersen EM, Nadim A, Larsen J. Numerical simulation and experimental validation of blood flow in arteries with structured-tree outflow conditions. Annals of Biomedical Engineering. 2000;28:1281–1299

    2000

  11. [19]

    Sensitivity analysis and uncertainty quantification of 1-D models of pulmonary hemodynamics in mice under control and hypertensive conditions

    Colebank MJ, Qureshi MU, Olufsen MS. Sensitivity analysis and uncertainty quantification of 1-D models of pulmonary hemodynamics in mice under control and hypertensive conditions. International Journal for Numerical Methods in Biomedical Engineering. 2021;37:1–29

    2021

  12. [20]

    Assessing model mismatch and model selection in a Bayesian uncertainty quantification analysis of a fluid-dynamics model of pulmonary blood circulation

    Paun LM, Colebank MJ, Olufsen MS, Hill NA, Husmeier D. Assessing model mismatch and model selection in a Bayesian uncertainty quantification analysis of a fluid-dynamics model of pulmonary blood circulation. Journal of The Royal Society Interface. 2020;17:20200886

    2020

  13. [21]

    Pulse Wave Propagation in the Arterial Tree

    van de Vosse FN, Stergiopulos N. Pulse Wave Propagation in the Arterial Tree. Annual Review of Fluid Mechanics. 2011;43:467–499

    2011

  14. [22]

    Numerical simulation of blood flow and pressure drop in the pulmonary arterial and venous circulation

    Qureshi MU, Vaughan GDA, Sainsbury C, Johnson M, Peskin CS, Olufsen MS, et al. Numerical simulation of blood flow and pressure drop in the pulmonary arterial and venous circulation. Biomechanics and Modeling in Mechanobiology. 2014;13:1137–1154. doi:10.1007/s10237-014-0563-y

    2014 doi

  15. [23]

    Evolution of hemodynamic forces in the pulmonary tree with progressively worsening pulmonary arterial hypertension in pediatric patients

    Yang W, Dong M, Rabinovitch M, Chan FP, Marsden AL, Feinstein JA. Evolution of hemodynamic forces in the pulmonary tree with progressively worsening pulmonary arterial hypertension in pediatric patients. Biomechanics and Modeling in Mechanobiology. 2019;18:779–796. doi:10.1007...

    2019 doi

  16. [24]

    The arterial Windkessel

    Westerhof N, Lankhaar JW, Westerhof BE. The arterial Windkessel. Medical & Biological Engineering & Computing. 2009;47:131–141

    2009

  17. [25]

    Structural and hemodynamic properties of murine pulmonary arterial networks under hypoxia-induced pulmonary hypertension

    Chambers MJ, Colebank MJ, Qureshi MU, Clipp R, Olufsen MS. Structural and hemodynamic properties of murine pulmonary arterial networks under hypoxia-induced pulmonary hypertension. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medici...

    2020

  18. [26]

    UQLab: A framework for uncertainty quantification in Matlab

    Marelli S, Sudret B. UQLab: A framework for uncertainty quantification in Matlab. In: Vulnerability, uncertainty, and risk: quantification, mitigation, and management. ETH Zurich; 2014. p. 2554–2563

    2014

  19. [27]

    Computer model calibration using high-dimensional output

    Higdon D, Gattiker J, Williams B, Rightley M. Computer model calibration using high-dimensional output. Journal of the American Statistical Association. 2008;103(482):570–583

    2008

  20. [28]

    Integration based profile likelihood calculation for PDE constrained parameter estimation problems

    Boiger R, Hasenauer J, Hroß S, Kaltenbacher B. Integration based profile likelihood calculation for PDE constrained parameter estimation problems. Inverse Problems. 2016;32:125009. doi:10.1088/0266-5611/32/12/125009

    2016 doi

  21. [29]

    Personalization of models with many model parameters: an efficient sensitivity analysis approach

    Donders WP, Huberts W, van de Vosse FN, Delhaas T. Personalization of models with many model parameters: an efficient sensitivity analysis approach. International Journal for Numerical Methods in Biomedical Engineering. 2015;31

    2015

  22. [30]

    Bayesian sensitivity analysis of a 1D vascular model with Gaussian process emulators

    Melis A, Clayton RH, Marzo A. Bayesian sensitivity analysis of a 1D vascular model with Gaussian process emulators. International Journal for Numerical Methods in Biomedical Engineering. 2017;33:1–11

    2017

  23. [31]

    An Effective Fractal-Tree Closure Model for Simulating Blood Flow in Large Arterial Networks

    Perdikaris P, Grinberg L, Karniadakis GE. An Effective Fractal-Tree Closure Model for Simulating Blood Flow in Large Arterial Networks. Annals of Biomedical Engineering. 2015

    2015

  24. [32]

    Rarefaction and blood pressure in systemic and pulmonary arteries

    Olufsen MS, Hill NA, Vaughan GDA, Sainsbury C, Johnson M. Rarefaction and blood pressure in systemic and pulmonary arteries. Journal of Fluid Mechanics. 2012;705:280–305. doi:10.1017/jfm.2012.220

    2012 doi

  25. [33]

    Parameter selection and optimization of a computational network model of blood flow in single-ventricle patients

    Taylor-LaPole AM, Paun LM, Lior D, Weigand JD, Puelz C, Olufsen MS. Parameter selection and optimization of a computational network model of blood flow in single-ventricle patients. Journal of the Royal Society Interface. 2025;22. doi:10.1098/rsif.2024.0663

    2025

  26. [34]

    To Sobol or not to Sobol? The effects of sampling schemes in systems biology applications

    Renardy M, Joslyn LR, Millar JA, Kirschner DE. To Sobol or not to Sobol? The effects of sampling schemes in systems biology applications. Mathematical biosciences. 2021; p. 108593

    2021

  27. [35]

    Hemodynamic assessment of pulmonary hypertension in mice: a model-based analysis of the disease mechanism

    Qureshi MU, Colebank MJ, Paun LM, Fix LE, Chesler N, Haider MA, et al. Hemodynamic assessment of pulmonary hypertension in mice: a model-based analysis of the disease mechanism. Biomechanics and Modeling in Mechanobiology. 2019;18:219–243

    2019

  28. [36]

    Emulation-accelerated Hamiltonian Monte Carlo algorithms for parameter estimation and uncertainty quantification in differential equation models

    Paun LM, Husmeier D. Emulation-accelerated Hamiltonian Monte Carlo algorithms for parameter estimation and uncertainty quantification in differential equation models. Statistics and Computing. 2022;32:1–25

    2022

  29. [37]

    Bayesian emulation of complex multi-output and dynamic computer models

    Conti S, O’Hagan A. Bayesian emulation of complex multi-output and dynamic computer models. Journal of Statistical Planning and Inference. 2010;140(3):640 – 651

    2010

  30. [38]

    Optimal experimental design for infinite-dimensional Bayesian inverse problems governed by PDEs: A review

    Alexanderian A. Optimal experimental design for infinite-dimensional Bayesian inverse problems governed by PDEs: A review. Inverse Problems. 2021;37(4):043001. September 23, 2025 30/30

    2021

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.