Abstract
A comprehensive image-based computational modelling pipeline is required for high-fidelity patient-specific cardiac simulations. However, traditional simulation methods are a limitation in these approaches due to their prohibitively slow speeds. We developed a physics-based training scheme using differentiable finite elements to compute the residual force vector of the governing PDE, which is then minimized to find the optimal network parameters. We used neural networks for their representation power, and finite elements for defining the problem domain, specifying the boundary conditions, and performing numerical integrations. We incorporated spatially varying fiber structures into a prolate spheroidal model of the left ventricle. A Fung-type material model including active contraction was used. We developed two versions of our model, one was trained on a reduced basis of the solution space, and one was trained on the full solution space. The models were trained against two pressure-volume loops and validated on a third loop (Fig. 1). We validated our implementation against conventional FEM simulation using FEniCS. While the reduced order model was trained faster than the full-order model, we achieved mean and standard deviation of the nodal error between the NNFE solution and the FE solution with 10–3 cm, with both models, where the characteristic length was 1 cm (Fig. 2a). The NNFE model predicted each solution within 0.6 ms whereas the FE models took up to 500 ms for each state. The NNFE method can be simultaneously trained over the entire range of physiological boundary conditions. The trained NNFE can predict stress–strain responses for any physiological boundary condition without retraining.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Liu, H., Narang, H., Gorman, R., Gorman, J., Sacks, M.S.: On the interrelationship between left ventricle infarction geometry and ischemic mitral regurgitation grade. In: Ennis, D.B., Perotti, L.E., Wang, V.Y. (eds.) FIMH 2021. LNCS, vol. 12738, pp. 425–434. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-78710-3_41
Liu, H., Soares, J.S., Walmsley, J., Li, D.S., Raut, S., Avazmohammadi, R., et al.: The impact of myocardial compressibility on organ-level simulations of the normal and infarcted heart. Sci. Rep-uk. 11(1), 13466 (2021)
Soares, J.S., Li, D.S., Lai, E., III J.H.G., Gorman, R.C., Sacks, M.S.: Functional imaging and modelling of the heart. In: 9th International Conference, FIMH 2017, Toronto, ON, Canada, 11–13 June 2017, Proceedings. Lect. Notes. Comput. Sci. 10263, 493–501 (2017)
Sacks, M.S., Motiwale, S., Goodbrake, C., Zhang, W.: Neural network approaches for soft biological tissue and organ simulations. J. Biomech. Eng. 144(12), 121010 (2022)
Zhang, W., Li, D.S., Bui-Thanh, T., Sacks, M.S.: Simulation of the 3D hyperelastic behavior of ventricular myocardium using a finite-element based neural-network approach. Comput. Meth. Appl. M. 394, 114871 (2022)
Scroggs, M.W., Dokken, J.S., Richardson, C.N., Wells, G.N.: Construction of arbitrary order finite element degree-of-freedom maps on polygonal and polyhedral cell meshes. ACM T Math Softw. 48(2), 1–23 (2022)
Scroggs, M.W., Baratta, I.A., Richardson, C.N., Wells, G.N.: Basix: a runtime finite element basis evaluation library. J. Open Source Softw. 7(73), 3982 (2022)
Alnæs, M.S., Logg, A., Ølgaard, K.B., Rognes, M.E., Wells, G.N.: Unified form language: a domain-specific language for weak formulations of partial differential equations. ACM Trans. Math Softw. Toms. 40(2), 9 (2014)
Kingma, D.P., Adam, B.J.: A method for stochastic optimization. Arxiv. (2014)
Kirby, R.C.: Algorithm 839: FIAT, a new paradigm for computing finite element basis functions. ACM Trans. Math Softw. Toms. 30(4), 502–516 (2004)
Roberts, A., et al.: Scaling Up Models and Data with $\texttt{t5x}$ and $\texttt{seqio}$. arxiv. (2022)
Acknowledgements
NIH R01 HL073021 and Platform for Advanced Scientific Computing (Swiss Federation).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Motiwale, S., Zhang, W., Sacks, M.S. (2023). High-Speed High-Fidelity Cardiac Simulations Using a Neural Network Finite Element Approach. In: Bernard, O., Clarysse, P., Duchateau, N., Ohayon, J., Viallon, M. (eds) Functional Imaging and Modeling of the Heart. FIMH 2023. Lecture Notes in Computer Science, vol 13958. Springer, Cham. https://doi.org/10.1007/978-3-031-35302-4_55
Download citation
DOI: https://doi.org/10.1007/978-3-031-35302-4_55
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-35301-7
Online ISBN: 978-3-031-35302-4
eBook Packages: Computer ScienceComputer Science (R0)