Surface-based versus voxel-based finite element head models: comparative analyses of strain responses

Area of Science:

• Biomechanics
• Computational Mechanics
• Neuroscience

Background:

• Finite element (FE) models are vital for human head injury assessment.
• Developing accurate, hexahedral-meshed FE head models, especially capturing intricate brain structures, remains challenging.
• Limited understanding exists regarding strain response similarities between surface-based and voxel-based FE head models.

Purpose of the Study:

• To compare the strain responses of surface-based and voxel-based FE head models derived from the same imaging data.
• To investigate the impact of mesh smoothing on strain response discrepancies.
• To provide quantitative insights into the general similarity of strain responses between different FE head model meshing techniques.

Main Methods:

• Developed three anatomically detailed FE head models from a single imaging dataset: one surface-based with conforming meshes and two voxel-based (with and without mesh smoothing).
• Ensured identical numerical settings across all models, differing only in mesh type.
• Simulated head impacts using these three models to analyze strain responses.

Main Results:

• Common injury metrics, such as 99 percentile strain and the volume of brain elements exceeding strain thresholds, showed virtually identical responses across the three models.
• Significant strain differences (>0.1) were observed at interfacial boundaries (e.g., cortical folds, falx, tentorium) between surface- and voxel-based models.
• Non-interfacial regions exhibited remarkable strain similarity.
• Mesh smoothing in voxel-based models marginally reduced strain discrepancies compared to the surface-based model.

Conclusions:

• Surface- and voxel-based FE head models demonstrate general similarity in strain responses when assessing common injury metrics.
• Quantitative insights confirm the comparability of these meshing approaches for head impact simulations.
• Caution is advised when utilizing interfacial strain data for injury prediction due to observed discrepancies.