Modelling subcortical bone in finite element analyses: A validation and sensitivity study in the macaque mandible.

Finite element analysis (FEA) is a fundamental method to study stresses and strains in complex structures, with the accuracy of an FEA being reliant on a number of variables, not least the precision and complexity of the model's geometry. Techniques such as computed tomography (CT) allow general geometries to be derived relatively quickly; however, constraints on CT image resolution mean defining subcortical geometries can be problematic. In relation to the overall mechanical response of a complex structure during FEA, the consequence of variable subcortical modelling is not known. Here we test this sensitivity with a series of FE models of a macaque mandible with different subcortical geometries and comparing the FEA strain magnitudes and orientations. The validity of the FE models was tested by carrying out experimental strain measurements on the same mandible. These strain measurements matched the FE predictions, providing confidence that material properties and model geometry were suitably defined. Results of this study show that cortical bone alone is not as effective in resisting bending as it is when coupled with subcortical bone, and as such subcortical geometries must be modelled during an FEA. This study demonstrates that the fine detail of the mandibular subcortical structure can be adequately modelled as a solid when assigned an appropriate Young's modulus value, in this case ranging from 1 to 2 GPa. This is an important and encouraging result for the creation of FE models of materials where CT image resolution or poor preservation prevent the accurate modelling of subcortical bone.