pith. sign in

arxiv: 2604.13657 · v1 · submitted 2026-04-15 · ❄️ cond-mat.soft · cond-mat.mes-hall· physics.bio-ph· physics.comp-ph

Hierarchical Bayesian calibration of mesoscopic models for ultrasound contrast agents from force spectroscopy data

Brieuc Benvegnen , Nikolaos Ntarakas , Tilen Potisk , Ignacio Pagonabarraga , Matej Praprotnik This is my paper

Pith reviewed 2026-05-10 12:13 UTC · model grok-4.3

classification ❄️ cond-mat.soft cond-mat.mes-hallphysics.bio-phphysics.comp-ph
keywords Bayesian calibrationdissipative particle dynamicsultrasound contrast agentsencapsulated microbubblesforce spectroscopyhierarchical inferenceneural network surrogatescapsid mechanics
0
0 comments X

The pith

A hierarchical Bayesian workflow with neural surrogates calibrates mesoscopic DPD models of microbubble shells to force spectroscopy data across multiple diameters.

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

The paper introduces a surrogate-accelerated Bayesian method to calibrate dissipative particle dynamics models of encapsulated microbubble capsids without running prohibitively expensive direct simulations for every parameter set. Deep neural network approximations stand in for the force spectroscopy simulations, transitional Markov chain Monte Carlo explores the posterior, and hierarchical regularization shares information across three different particle diameters for each commercial agent. The resulting posteriors for Definity and SonoVue show that stretching stiffness and bending modulus become well constrained by the published compression and indentation measurements. If the approach holds, it supplies ready-to-use particle-based models whose mechanical parameters are tied directly to experimental force curves rather than chosen by hand.

Core claim

We develop a surrogate-accelerated Bayesian calibration workflow that combines deep neural network surrogates, transitional Markov chain Monte Carlo sampling, and hierarchical regularization across diameters. Applied to published compression data for Definity and indentation data for SonoVue, each covering three distinct diameters, the workflow produces data-informed DPD models in which key force-field parameters such as stretching stiffness and bending modulus are consistently constrained by the measurements.

What carries the argument

The surrogate-accelerated hierarchical Bayesian calibration workflow, in which DNNs approximate DPD force spectroscopy simulations so that transitional MCMC can efficiently sample posteriors while hierarchical regularization couples inference across different microbubble diameters.

If this is right

  • The calibrated DPD models for Definity and SonoVue supply particle-level mechanics that can be inserted directly into larger-scale simulations of ultrasound-driven drug delivery.
  • Key shell parameters remain identifiable when data from multiple diameters are combined, reducing the risk that single-size experiments leave parameters under-determined.
  • The same workflow can generate bespoke models for other ultrasound contrast agents whose capsids are made of lipids, proteins, or polymers.
  • Hierarchical regularization across sizes improves robustness of the inferred posteriors compared with independent calibration of each diameter.

Where Pith is reading between the lines

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

  • The method lowers the barrier to building predictive mesoscopic models for any soft-matter particle whose expensive simulations can be approximated by a neural surrogate.
  • If the inferred parameters prove transferable to dynamic ultrasound conditions, the models could guide the design of agents with tailored resonance or stability properties.
  • Extending the hierarchy to include additional experimental modalities, such as acoustic scattering data, would further tighten the mechanical constraints.

Load-bearing premise

The chosen form of the DPD force field can represent the actual mechanical response of the microbubble capsids and the neural-network surrogate approximates the true simulation outputs accurately enough for reliable posterior sampling.

What would settle it

New compression or indentation experiments on the same Definity or SonoVue particles that produce force curves incompatible with the inferred posterior ranges, or direct DPD runs with the posterior-mean parameters that fail to reproduce the measured force-distance relations within experimental uncertainty.

Figures

Figures reproduced from arXiv: 2604.13657 by Brieuc Benvegnen, Ignacio Pagonabarraga, Matej Praprotnik, Nikolaos Ntarakas, Tilen Potisk.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8 [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9 [PITH_FULL_IMAGE:figures/full_fig_p013_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10 [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
read the original abstract

Ultrasound-guided drug and gene delivery (USDG) is a promising non-invasive approach for targeted therapeutic applications. Mechanical properties of encapsulated microbubbles (EMBs), which serve as contrast agents, strongly affect their specific interactions with ultrasound and are thus critical to the success and efficiency of USDG. Accurate calibration of high-fidelity particle-based models of EMB capsid mechanics is computationally challenging because direct Bayesian inference with dissipative particle dynamics (DPD) is prohibitively expensive. We employ a surrogate-accelerated Bayesian calibration workflow that combines deep neural network (DNN) surrogates, transitional Markov chain Monte Carlo sampling, and hierarchical regularization across EMB diameters. Using this framework, we develop two data-informed DPD models of commercial EMB agents, i.e., Definity and SonoVue, and perform inference of force field parameters based on published compression experiments for Definity and indentation experiments for SonoVue, each spanning three distinct diameters. The inferred posteriors show that key model parameters, such as the stretching stiffness and bending modulus, are consistently constrained by the available data. The presented methodology can be used to derive bespoke, data-informed models for a wide range of ultrasound contrast agents, including encapsulated gas vesicles, EMBs with diverse capsids consisting of lipids, proteins, or polymers, and functionalized with ligands.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript presents a surrogate-accelerated hierarchical Bayesian calibration workflow that uses deep neural network (DNN) surrogates to enable efficient inference of dissipative particle dynamics (DPD) force-field parameters for mesoscopic models of commercial ultrasound contrast agents (Definity and SonoVue). Inference is performed on published compression and indentation force-spectroscopy data spanning three diameters per agent; the resulting posteriors are reported to constrain key parameters such as stretching stiffness and bending modulus, with hierarchical regularization enforcing consistency across diameters.

Significance. If the DNN surrogates are shown to be accurate in the relevant posterior regions, the work supplies a practical route to data-informed DPD models of encapsulated microbubbles. Such models are directly relevant to ultrasound-guided drug delivery, where capsid mechanics govern agent response. The hierarchical structure across diameters is a methodological strength that promotes physically consistent parameter sets without ad-hoc tuning.

major comments (2)
  1. The central claim that the inferred posteriors on stretching stiffness and bending modulus are data-driven (rather than surrogate artifacts) requires that the DNN accurately reproduces DPD force-displacement curves near the posterior means for each of the three diameters. No test-set MAE, maximum error, or direct DPD re-runs at posterior samples are reported; without these, systematic surrogate bias could artificially tighten or shift the posteriors. This validation step is load-bearing for the workflow's validity.
  2. The hierarchical regularization across diameters is presented as a strength, yet the manuscript does not quantify how much the joint posterior tightens relative to independent per-diameter inferences, nor does it show that the regularization does not mask surrogate errors that are correlated across diameters.
minor comments (1)
  1. The abstract would be strengthened by inclusion of at least one quantitative result (e.g., posterior credible-interval widths or surrogate error bounds) so readers can immediately gauge the degree of constraint achieved.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments and positive assessment of the significance of our work. We address each major comment below and will revise the manuscript accordingly to strengthen the validation and presentation of the hierarchical approach.

read point-by-point responses

  1. Referee: The central claim that the inferred posteriors on stretching stiffness and bending modulus are data-driven (rather than surrogate artifacts) requires that the DNN accurately reproduces DPD force-displacement curves near the posterior means for each of the three diameters. No test-set MAE, maximum error, or direct DPD re-runs at posterior samples are reported; without these, systematic surrogate bias could artificially tighten or shift the posteriors. This validation step is load-bearing for the workflow's validity.

    Authors: We agree that explicit validation of the DNN surrogates within the posterior support is essential to rule out systematic bias. In the revised manuscript we will add a new subsection on surrogate fidelity that reports (i) test-set MAE and maximum absolute error over the full training range, and (ii) direct DPD re-simulations at 20–30 posterior samples drawn from each diameter-specific marginal. These comparisons will be shown as overlaid force-displacement curves together with quantitative error metrics, confirming that surrogate discrepancies remain well below the experimental noise level. revision: yes

  2. Referee: The hierarchical regularization across diameters is presented as a strength, yet the manuscript does not quantify how much the joint posterior tightens relative to independent per-diameter inferences, nor does it show that the regularization does not mask surrogate errors that are correlated across diameters.

    Authors: We acknowledge the value of a quantitative comparison. The revised manuscript will include an additional figure and table that repeat the inference independently for each diameter (using identical DNN surrogates and priors) and directly compare the resulting marginal posteriors to the hierarchical joint posterior. We will report the reduction in posterior variance and the shift in credible intervals. To address possible masking of correlated surrogate errors, we will also provide a supplementary analysis of the surrogate residual fields across the three diameters; because the DNNs were trained on independent DPD runs for each diameter, the residuals are uncorrelated by construction, and we will state this assumption explicitly. revision: yes

Circularity Check

0 steps flagged

No significant circularity; inference uses independent experimental data

full rationale

The paper's derivation chain consists of training a DNN surrogate on DPD simulations over a parameter grid, then using that surrogate inside transitional MCMC to infer force-field parameters (stretching stiffness, bending modulus) from published, independent compression/indentation experiments on Definity and SonoVue across three diameters. No step equates a model output to its own input by construction, renames a fitted quantity as a prediction, or relies on a self-citation chain for a uniqueness theorem. The hierarchical regularization across diameters is a standard Bayesian device and does not create self-definition. The central claim that posteriors are data-constrained therefore remains logically independent of the fitted values themselves. Surrogate accuracy is an unverified modeling assumption rather than a circular reduction.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim depends on the domain assumption that DPD is suitable for this system and that the surrogate approximation holds. The parameters are not free but inferred; no new physical entities are introduced.

free parameters (2)
  • stretching stiffness
    Bayesian-inferred parameter from experimental data in the DPD model.
  • bending modulus
    Bayesian-inferred parameter from experimental data in the DPD model.
axioms (2)
  • domain assumption The chosen DPD force field is an appropriate representation for the mechanics of lipid or polymer capsids in EMBs
    Invoked when developing the models for Definity and SonoVue.
  • domain assumption DNN surrogates can be trained to faithfully reproduce DPD simulation outputs for force spectroscopy
    Necessary for the surrogate-accelerated workflow to be valid.

pith-pipeline@v0.9.0 · 5562 in / 1474 out tokens · 78988 ms · 2026-05-10T12:13:59.904646+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

80 extracted references · 80 canonical work pages · 1 internal anchor

  1. [1]

    DISCUSSION The central result of this study is that we develop and apply a surrogate-accelerated Bayesian workflow to a high- fidelity numericalembmodel with quantified uncertainty, and show that, in the present quasi-static regime, the data constraink a andk b and support a reduced force field description. Quantitatively, grouped-holdout surrogate errors...

    work page

  2. [2]

    Indpd, each particle represents a coarse-grained cluster of molecules and evolves according to Newton’s equations of motion

    METHODS dpdsimulations All mesoscale simulations were performed usingdpd, a particle-based method well suited for modeling soft matter systems with explicit hydrodynamics [20]. Indpd, each particle represents a coarse-grained cluster of molecules and evolves according to Newton’s equations of motion. The total force acting on the particleiis given by Fi =...

    work page

  3. [3]

    Stride, Physical Principles of Microbubbles for Ultrasound Imaging and Therapy., Cerebrovasc

    E. Stride, Physical Principles of Microbubbles for Ultrasound Imaging and Therapy., Cerebrovasc. Dis.27, 1 (2009)

    work page 2009

  4. [4]

    Ferrara, R

    K. Ferrara, R. Pollard, and M. Borden, Ultrasound Microbubble Contrast Agents: Fundamentals and Application to Gene and Drug Delivery, Annu. Rev. Biomed. Eng.9, 415 (2007)

    work page 2007

  5. [5]

    A. L. Klibanov, Microbubble Contrast Agents: Targeted Ultrasound Imaging and Ultrasound-Assisted Drug-Delivery Applications, Investig. Radiol.41, 354 (2006)

    work page 2006

  6. [6]

    Datta, C.-C

    S. Datta, C.-C. Coussios, A. Y. Ammi, T. D. Mast, G. M. de Courten-Myers, and C. K. Holland, Ultrasound-enhanced thrombolysis using definity®as a cavitation nucleation agent, Ultrasound Med. Biol.34, 1421 (2008)

    work page 2008

  7. [7]

    S. R. Sirsi and M. A. Borden, State-of-the-art materials for ultrasound-triggered drug delivery, Adv. Drug Deliv. Rev.72, 3 (2014)

    work page 2014

  8. [8]

    Stride and C

    E. Stride and C. Coussios, Cavitation and contrast: the use of bubbles in ultrasound imaging and therapy, Proc. Inst. Mech. Eng. H224, 171 (2010)

    work page 2010

  9. [9]

    C. C. Coussios and R. A. Roy, Applications of acoustics and cavitation to noninvasive therapy and drug delivery, Annu. Rev. Fluid Mech.40, 395 (2008)

    work page 2008

  10. [10]

    X. Guo, Q. Li, Z. Zhang, D. Zhang, and J. Tu, Investigation on the inertial cavitation threshold and shell properties of commercialized ultrasound contrast agent microbubbles, J. Acoust. Soc. Am.134, 1622 (2013)

    work page 2013

  11. [11]

    E. E. Konofagou, Optimization of the ultrasound-induced blood-brain barrier opening, Theranostics2, 1223 (2012)

    work page 2012

  12. [12]

    Tinkov, R

    S. Tinkov, R. Bekeredjian, G. Winter, and C. Coester, Microbubbles as Ultrasound Triggered Drug Carriers, J. Pharm. Sci.98, 1935 (2009)

    work page 1935

  13. [13]

    Tinkov, C

    S. Tinkov, C. Coester, S. Serba, N. A. Geis, H. A. Katus, G. Winter, and R. Bekeredjian, New doxorubicin-loaded phospholipid microbubbles for targeted tumor therapy: in-vivo characterization, J. Control. Release148, 368 (2010)

    work page 2010

  14. [14]

    Hynynen, N

    K. Hynynen, N. McDannold, N. Vykhodtseva, and F. A. Jolesz, Noninvasive MR Imaging–guided Focal Opening of the Blood-Brain Barrier in Rabbits, Radiology220, 640 (2001)

    work page 2001

  15. [15]

    Glynos, V

    E. Glynos, V. Sboros, and V. Koutsos, Polymeric thin shells: Measurement of elastic properties at the nanometer scale using atomic force microscopy, Mater. Sci. Eng. B165, 231 (2009)

    work page 2009

  16. [16]

    Glynos, V

    E. Glynos, V. Koutsos, W. N. McDicken, C. M. Moran, S. D. Pye, J. A. Ross, and V. Sboros, Nanomechanics of biocom- patible hollow thin-shell polymer microspheres, Langmuir25, 7514 (2009)

    work page 2009

  17. [17]

    Evans and A

    E. Evans and A. Yeung, Apparent viscosity and cortical tension of blood granulocytes determined by micropipet aspiration, Biophys. J.56, 151 (1989)

    work page 1989

  18. [18]

    C. C. Church, The effects of an elastic solid surface layer on the radial pulsations of gas bubbles, J. Acoust. Soc. Am.97, 1510 (1995)

    work page 1995

  19. [19]

    J. Tu, J. Guan, Y. Qiu, and T. J. Matula, Estimating the shell parameters of SonoVue®microbubbles using light scattering, J. Acoust. Soc. Am.126, 2954 (2009)

    work page 2009

  20. [20]

    C.-L. Cai, J. Yu, J. Tu, X.-S. Guo, P.-T. Huang, and D. Zhang, Interaction between encapsulated microbubbles: A finite element modelling study, Chin. Phys. B27, 084302 (2018)

    work page 2018

  21. [21]

    Y. Guo, H. Lee, C. Kim, C. Park, A. Yamamichi, P. Chuntova, M. Gallus, M. O. Bernabeu, H. Okada, H. Jo,et al., Ultrasound frequency-controlled microbubble dynamics in brain vessels regulate the enrichment of inflammatory pathways in the blood-brain barrier, Nat. Commun.15, 8021 (2024)

    work page 2024

  22. [22]

    R. D. Groot and P. B. Warren, Dissipative particle dynamics: Bridging the gap between atomistic and mesoscopic simu- lation, J. Chem. Phys.107, 4423 (1997)

    work page 1997

  23. [23]

    Espanol and P

    P. Espanol and P. Warren, Statistical Mechanics of Dissipative Particle Dynamics, EPL30, 191 (1995)

    work page 1995

  24. [24]

    Ntarakas, M

    N. Ntarakas, M. Lah, D. Svenšek, T. Potisk, and M. Praprotnik, Dissipative Particle Dynamics Models of Encapsulated Microbubbles and Nanoscale Gas Vesicles for Biomedical Ultrasound Simulations, ACS Appl. Nano Mater.8, 16053 (2025)

    work page 2025

  25. [25]

    Angelikopoulos, C

    P. Angelikopoulos, C. Papadimitriou, and P. Koumoutsakos, Bayesian uncertainty quantification and propagation in molec- ular dynamics simulations: A high performance computing framework, J. Chem. Phys.137, 144103 (2012)

    work page 2012

  26. [26]

    Gelman and C

    A. Gelman and C. R. Shalizi, Philosophy and the practice of Bayesian statistics, Br. J. Math. Stat. Psychol.66, 8 (2013)

    work page 2013

  27. [27]

    Gregory,Bayesian Logical Data Analysis for the Physical Sciences: A Comparative Approach With Mathematica® Support(Cambridge University Press, 2005)

    P. Gregory,Bayesian Logical Data Analysis for the Physical Sciences: A Comparative Approach With Mathematica® Support(Cambridge University Press, 2005)

    work page 2005

  28. [28]

    A. T. Nordman, S. Engblom, and D. Van der Spoel, Bayesian three-point water models, npj Comput. Mater.11, 366 (2025)

    work page 2025

  29. [29]

    Amoudruz, S

    L. Amoudruz, S. Litvinov, C. Papadimitriou, and P. Koumoutsakos, Bayesian Inference for PDE-based Inverse Problems using the Optimization of a Discrete Loss, arXiv preprint arXiv:2510.15664 (2025)

    work page arXiv 2025

  30. [30]

    S. Wu, P. Angelikopoulos, G. Tauriello, C. Papadimitriou, and P. Koumoutsakos, Fusing heterogeneous data for the calibration of molecular dynamics force fields using hierarchical Bayesian models, J. Chem. Phys.145(2016)

    work page 2016

  31. [31]

    Cailliez, A

    F. Cailliez, A. Bourasseau, and P. Pernot, Calibration of Forcefields for Molecular Simulation: Sequential Design of Computer Experiments for Building Cost-Efficient Kriging Metamodels, J. Comput. Chem.35, 130 (2014). 19

    work page 2014

  32. [32]

    Zhang, L

    Z. Zhang, L. Wang, Z. Wang, X. He, Y. Chen, F. Müller-Plathe, and M. C. Böhm, A coarse-grained molecular dynamics– reactive Monte Carlo approach to simulate hyperbranched polycondensation, RSC Adv.4, 56625 (2014)

    work page 2014

  33. [33]

    C. K. Williams and C. E. Rasmussen,Gaussian Processes for Machine Learning, Vol. 2 (MIT Press, Cambridge, MA, 2006)

    work page 2006

  34. [34]

    D. Wu, M. Chinazzi, A. Vespignani, Y.-A. Ma, and R. Yu, Multi-fidelity Hierarchical Neural Processes, inProc. of the 28th ACM SIGKDD conference on knowledge discovery and data mining(2022) pp. 2029–2038

    work page 2022

  35. [35]

    Raissi, P

    M. Raissi, P. Perdikaris, and G. E. Karniadakis, Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, J. Comput. Phys.378, 686 (2019)

    work page 2019

  36. [36]

    L. Lu, X. Meng, Z. Mao, and G. E. Karniadakis, DeepXDE: A deep learning library for solving differential equations, SIAM Rev.63, 208 (2021)

    work page 2021

  37. [37]

    P. Ray, A. P. Generale, N. Vankireddy, Y. Asoma, M. Nakauchi, H. Lee, K. Yoshida, Y. Okuno, and S. R. Kalidindi, Refining coarse-grained molecular topologies: a Bayesian optimization approach, npj Comput. Mater.11, 234 (2025)

    work page 2025

  38. [38]

    Ching and Y.-C

    J. Ching and Y.-C. Chen, Transitional Markov chain Monte Carlo method for Bayesian model updating, model class selection, and model averaging, J. Eng. Mech.133, 816 (2007)

    work page 2007

  39. [39]

    J. L. Beck and S.-K. Au, Bayesian Updating of Structural Models and Reliability using Markov Chain Monte Carlo Simulation, J. Eng. Mech.128, 380 (2002)

    work page 2002

  40. [40]

    Gelman, Multilevel (Hierarchical) Modeling: What It Can and Cannot Do, Technometrics48, 432 (2006)

    A. Gelman, Multilevel (Hierarchical) Modeling: What It Can and Cannot Do, Technometrics48, 432 (2006)

    work page 2006

  41. [41]

    Economides, G

    A. Economides, G. Arampatzis, D. Alexeev, S. Litvinov, L. Amoudruz, L. Kulakova, C. Papadimitriou, and P. Koumout- sakos, Hierarchical Bayesian Uncertainty Quantification for a Model of the Red Blood Cell, Phys. Rev. Appl.15, 034062 (2021)

    work page 2021

  42. [42]

    J. R. Lindner, J. Song, A. R. Jayaweera, J. Sklenar, and S. Kaul, Microvascular Rheology of Definity Microbubbles after Intra-arterial and Intravenous Administration, J. Am. Soc. Echocardiogr.15, 396 (2002)

    work page 2002

  43. [43]

    Baseri, J

    B. Baseri, J. J. Choi, Y.-S. Tung, and E. E. Konofagou, Multi-Modality Safety Assessment of Blood-Brain Barrier Opening Using Focused Ultrasound and Definity Microbubbles: A Short-Term Study, Ultrasound Med. Biol.36, 1445 (2010)

    work page 2010

  44. [44]

    Schneider, SonoVue, a new ultrasound contrast agent, Eur

    M. Schneider, SonoVue, a new ultrasound contrast agent, Eur. Radiol.9, S347 (1999)

    work page 1999

  45. [45]

    Demené, J

    C. Demené, J. Robin, A. Dizeux, B. Heiles, M. Pernot, M. Tanter, and F. Perren, Transcranial ultrafast ultrasound localization microscopy of brain vasculature in patients, Nat. Biomed. Eng.5, 219 (2021)

    work page 2021

  46. [46]

    De Jong, M

    N. De Jong, M. Emmer, C. T. Chin, A. Bouakaz, F. Mastik, D. Lohse, and M. Versluis, Compression-only behavior of phospholipid-coated contrast bubbles, Ultrasound Med. Biol.33, 653 (2007)

    work page 2007

  47. [47]

    Alexeev, L

    D. Alexeev, L. Amoudruz, S. Litvinov, and P. Koumoutsakos, Mirheo: High-performance mesoscale simulations for mi- crofluidics, Comput. Phys. Commun.254, 107298 (2020)

    work page 2020

  48. [48]

    J. S. Ma, M. A. Gómez Maureira, and J. N. Van Rijn, Eating Sound Dataset for 20 Food Types and Sound Classification Using Convolutional Neural Networks, inCompanion Publication of the 2020 International Conference on Multimodal Interaction(2020) pp. 348–351

    work page 2020

  49. [49]

    Herman and W

    J. Herman and W. Usher, SALib: An open-source Python library for Sensitivity Analysis, J. Open Source Softw.2, 97 (2017)

    work page 2017

  50. [50]

    S. M. Martin, D. Wälchli, G. Arampatzis, A. E. Economides, P. Karnakov, and P. Koumoutsakos, Korali: Efficient and scalable software framework for Bayesian uncertainty quantification and stochastic optimization, Comput. Methods Appl. Mech. Eng.389, 114264 (2022)

    work page 2022

  51. [51]

    J. E. Matheson and R. L. Winkler, Scoring rules for continuous probability distributions, Manag. Sci.22, 1087 (1976)

    work page 1976

  52. [52]

    Hersbach, Decomposition of the Continuous Ranked Probability Score for Ensemble Prediction Systems, Weather Forecast.15, 559 (2000)

    H. Hersbach, Decomposition of the Continuous Ranked Probability Score for Ensemble Prediction Systems, Weather Forecast.15, 559 (2000)

    work page 2000

  53. [53]

    Wälchli, S

    D. Wälchli, S. M. Martin, A. Economides, L. Amoudruz, G. Arampatzis, X. Bian, and P. Koumoutsakos, Load balancing in large scale Bayesian inference, inProc. of the Platform for Advanced Scientific Computing Conference(2020) pp. 1–12

    work page 2020

  54. [54]

    Amoudruz,Simulations and control of artificial microswimmers in blood, PhD thesis, ETH Zurich (2022)

    L. Amoudruz,Simulations and control of artificial microswimmers in blood, PhD thesis, ETH Zurich (2022)

    work page 2022

  55. [55]

    Amoudruz, A

    L. Amoudruz, A. Economides, G. Arampatzis, and P. Koumoutsakos, The stress-free state of human erythrocytes: Data- driven inference of a transferable RBC model, Biophys. J.122, 1517 (2023)

    work page 2023

  56. [56]

    Amoudruz, A

    L. Amoudruz, A. Economides, and P. Koumoutsakos, The volume of healthy red blood cells is optimal for advective oxygen transport in arterioles, Biophys. J.123, 1289 (2024)

    work page 2024

  57. [57]

    Amoudruz, S

    L. Amoudruz, S. Litvinov, R. Murri, V. Eyrich, J. Zudrop, C. Bekas, and P. Koumoutsakos, Scalable, cloud-based simula- tions of blood flow and targeted drug delivery in retinal capillaries, Comput. Phys. Commun.320, 109967 (2025)

    work page 2025

  58. [58]

    Y. Liu, D. Wang, X. Sun, N. Dinh, and R. Hu, Uncertainty quantification for Multiphase-CFD simulations of bubbly flows: a machine learning-based Bayesian approach supported by high-resolution experiments, Reliab. Eng. Syst. Saf.212, 107636 (2021)

    work page 2021

  59. [59]

    M. Lah, N. Ntarakas, T. Potisk, P. Papež, and M. Praprotnik, Open-boundary molecular dynamics of ultrasound using supramolecular water models, J. Chem. Phys.162(2025)

    work page 2025

  60. [60]

    Maresca, A

    D. Maresca, A. Lakshmanan, M. Abedi, A. Bar-Zion, A. Farhadi, G. J. Lu, J. O. Szablowski, D. Wu, S. Yoo, and M. G. Shapiro, Biomolecular Ultrasound and Sonogenetics, Annu. Rev. Chem. Biomol. Eng.9, 229 (2018)

    work page 2018

  61. [61]

    Salahshoor, Y

    H. Salahshoor, Y. Yao, P. Dutka, N. N. Nyström, Z. Jin, E. Min, D. Malounda, G. J. Jensen, M. Ortiz, and M. G. Shapiro, Geometric effects in gas vesicle buckling under ultrasound, Biophys. J.121, 4221 (2022)

    work page 2022

  62. [62]

    Dutka, L

    P. Dutka, L. A. Metskas, R. C. Hurt, H. Salahshoor, T.-Y. Wang, D. Malounda, G. J. Lu, T.-F. Chou, M. G. Shapiro, and G. J. Jensen, Structure of Anabaena flos-aquae gas vesicles revealed by cryo-ET, Structure31, 518 (2023)

    work page 2023

  63. [63]

    C. A. Smith, A. Bar-Zion, Q. Wu, D. Malounda, L. Bau, E. Stride, M. G. Shapiro, and C. C. Coussios, Ultrafast optical and passive acoustic mapping characterization of nanoscale cavitation nuclei based on gas vesicle proteins, AIP Adv.15 20 (2025)

    work page 2025

  64. [64]

    B. A. Kaufmann, K. Wei, and J. R. Lindner, Contrast Echocardiography, Curr. Probl. Cardiol.32, 51 (2007)

    work page 2007

  65. [65]

    Kogan, R

    P. Kogan, R. C. Gessner, and P. A. Dayton, Microbubbles in Imaging: Applications Beyond Ultrasound, Bubble Sci. Eng. Technol.2, 3 (2010)

    work page 2010

  66. [66]

    Velroyen, M

    A. Velroyen, M. Bech, A. Malecki, A. Tapfer, A. Yaroshenko, M. Ingrisch, C. Cyran, S. Auweter, K. Nikolaou, M. Reiser, et al., Microbubbles as a scattering contrast agent for grating-based x-ray dark-field imaging, Phys. Med. Biol.58, N37 (2013)

    work page 2013

  67. [67]

    H. Lim, M. Wortis, and R. Mukhopadhyay, Stomatocyte–discocyte–echinocyte sequence of the human red blood cell: Evidence for the bilayer–couple hypothesis from membrane mechanics, Proc. Natl. Acad. Sci. USA99, 16766 (2002)

    work page 2002

  68. [68]

    H. Lim, M. Wortis, and R. Mukhopadhyay, Red Blood Cell Shapes and Shape Transformations: Newtonian Mechanics of A Composite Membrane: Sections 2.1–2.4, inSoft Matter(Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany,

    work page

  69. [69]

    Božič and G

    B. Božič and G. Gomišček, Role of red blood cell elastic properties in capillary occlusions, Phys. Rev. E86, 051902 (2012)

    work page 2012

  70. [70]

    Revenga, I

    M. Revenga, I. Zuniga, and P. Espanol, Boundary conditions in dissipative particle dynamics, Comput. Phys. Commun. 121, 309 (1999)

    work page 1999

  71. [71]

    D. A. Fedosov, B. Caswell, and G. E. Karniadakis, A Multiscale Red Blood Cell Model with Accurate Mechanics, Rheology, and Dynamics, Biophys. J.98, 2215 (2010)

    work page 2010

  72. [72]

    Paulose, G

    J. Paulose, G. A. Vliegenthart, G. Gompper, and D. R. Nelson, Fluctuating shells under pressure, Proc. Natl. Acad. Sci. USA109, 19551 (2012)

    work page 2012

  73. [73]

    Gneiting and A

    T. Gneiting and A. E. Raftery, Strictly Proper Scoring Rules, Prediction, and Estimation, J. Am. Stat. Assoc.102, 359 (2007)

    work page 2007

  74. [74]

    Gneiting, F

    T. Gneiting, F. Balabdaoui, and A. E. Raftery, Probabilistic forecasts, calibration and sharpness, J. R. Stat. Soc. Ser. B Stat. Methodol.69, 243 (2007)

    work page 2007

  75. [75]

    Buchner Santos, J

    E. Buchner Santos, J. K. Morris, E. Glynos, V. Sboros, and V. Koutsos, Nanomechanical Properties of Phospholipid Microbubbles, Langmuir28, 5753 (2012)

    work page 2012

  76. [76]

    J. K. Morris,The Mechanical Properties of Phospholipid Coated Microbubbles, PhD thesis, University of Edinburgh (2014)

    work page 2014

  77. [77]

    V. R. Joseph, E. Gul, and S. Ba, Designing Computer Experiments with Multiple Types of Factors: The MaxPro Approach, J. Qual. Technol.52, 343 (2020)

    work page 2020

  78. [78]

    Glorot and Y

    X. Glorot and Y. Bengio, Understanding the difficulty of training deep feedforward neural networks, inProc. 13th Inter- national Conference on Artificial Intelligence and Statistics(JMLR, 2010) pp. 249–256

    work page 2010

  79. [79]

    Ansel, E

    J. Ansel, E. Yang, H. He, N. Gimelshein, A. Jain, M. Voznesensky, B. Bao, P. Bell, D. Berard, E. Burovski,et al., Pytorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation, inProc. of the 29th ACM international conference on architectural support for programming languages and operating systems(2024) pp. 929–947

    work page 2024

  80. [80]

    Adam: A Method for Stochastic Optimization

    D. P. Kingma and J. Ba, Adam: A method for stochastic optimization, arXiv preprint arXiv:1412.6980 (2014). 21 SUPPLEMENTARY INFORMATION S1.DPDMODEL IMPLEMENTATION DETAILS Each microbubble shell is represented as a triangulated surface whose vertex positions, connectivity, and reference geometric quantities are generated before thedpdrun starts [22]. Relev...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1021/la204801u 2014