Beyond parameter estimation: extending biomechanical modeling by the explicit exploration of model topology

Area of Science:

• Biomechanics
• Computational Biology
• Systems Biology

Background:

• Current biomechanical modeling relies on parameter estimation within assumed topologies, limiting realistic function prediction.
• Complex biological systems, such as finger tendon networks, present significant challenges for existing modeling approaches.

Purpose of the Study:

• To present advances in a novel modeling paradigm that simultaneously explores both model topology and parameter values.
• To address the challenge of selecting realistic model topologies for predicting biomechanical function.

Main Methods:

• Developed a novel computational environment for quasi-static simulations of arbitrary elastic structure topologies under large deformations.
• Applied simulation methods to investigate the role of finger tendon network topology in tension propagation to joints.
• Introduced a novel inference algorithm for simultaneous exploration of topology and parameter values in synthetic networks.

Main Results:

• Demonstrated that assumed model topology significantly influences tension propagation in finger tendon networks.
• Successfully showed the simultaneous inference of topology and parameters for synthetic tendon networks.
• Highlighted critical issues in inferring topological features from input/output data.

Conclusions:

• The proposed extended modeling paradigm is crucial for complex biomechanical systems.
• Further research is needed to address challenges in observability, separability, and uniqueness for real anatomical systems.
• This work paves the way for extracting causal biomechanical models from experimental data.