Speaker
Description
The research community is rapidly advancing in the development of cardiac biophysical models for clinical applications, due to their predictive capabilities. Nevertheless, the computational expense associated with high-resolution multi-physics models and their personalisation poses challenges for their translation into clinical settings.
This work introduces a novel methodology integrating physics-informed neural networks (PINNs)~\cite{caforio2024physics ,HOFLER2026106603} with 3D, time-dependent nonlinear biomechanical models of cardiac tissue to reconstruct displacement fields and estimate heterogeneous patient-specific biophysical properties. The proposed learning algorithm utilises information from a limited dataset of displacement and, in certain cases, strain data that can be routinely acquired in clinical settings, and it incorporates residual-based attention, Fourier features, and tailored regularisation strategies, enabling robust reconstruction of parameter fields under noisy conditions and at high spatial resolution. A Pareto front analysis is also conducted to assess the influence of loss weight selection on parameter estimation.
Several benchmarks are presented to highlight the accuracy and robustness of the proposed method. Notably, the study demonstrates the capability of PINNs to detect the presence, location, and severity of scar tissue, thus potentially greatly enhancing diagnostics and treatment plans for cardiac diseases.
Bibliography
@article{caforio2024physics,
title={Physics-informed neural network estimation of material properties in soft tissue nonlinear biomechanical models},
author={Caforio, Federica and Regazzoni, Francesco and Pagani, Stefano and Karabelas, Elias and Augustin, Christoph and Haase, Gundolf and Plank, Gernot and Quarteroni, Alfio},
journal={Computational Mechanics},
pages={1--27},
year={2024},
publisher={Springer},
doi = {https://doi.org/10.1007/s00466-024-02516-x}
}
@article{HOFLER2026106603,
title = {Physics-informed neural network estimation of active material properties in time-dependent cardiac biomechanical models},
journal = {Journal of the Mechanics and Physics of Solids},
pages = {106603},
year = {2026},
issn = {0022-5096},
doi = {https://doi.org/10.1016/j.jmps.2026.106603},
author = {Matthias Höfler and Francesco Regazzoni and Stefano Pagani and Elias Karabelas and Christoph Augustin and Gundolf Haase and Gernot Plank and Federica Caforio},
}
