Biomechanical Analysis of the Sensitivity of Brain Tissue Responses to FE Head Models in the Study of Impact-Induced TBI

Impact-induced traumatic brain injuries (iTBI) occur when the head undergoes a dynamic loading, such as external forces exerted during accident, contacts in sports, and falls. These impacts will lead to acceleration and deceleration of the brain inside the cranium which depending on the severity of the impact, can cause mild to severe TBIs. Computational methods such as finite element (FE) methods have found great attention in the last few decades for studying physical phenomena which are hard to experiment because of moral and technical limitations. Brain injury study of animals and specially humans are major topics which can significantly benefit from computational analysis. A successful FE study requires accurate replicate of the physical model, use of accurate material models, implementation of real-life boundary conditions, and validation of the results. Many studies have employed numerical methods to measure the linear and angular accelerations of the head in order to develop injury criteria. Many studies have also employed FE modeling to assess the dynamic responses of the brain in terms intracranial pressure (ICP), shear stress, and shear strains. These studies have used a single FE head model for their analyses. However, the head kinematics and brain dynamic responses significantly depend on the geometry, number of components, and material models used for each component. To this end, our study investigated the influence of the FE head model on the pressure and shear stress development and distribution using two different FE head models, namely, North Dakota State University Head Model (NDSUHM) and Dartmouth Head Injury Model (DHIM). The head models differed mainly in the number of constituting components, mesh sizes, and geometry. Both FE head models were impacted by a rigid cylindrical impactor with a mass of 5.6 kg at a velocity of 10 m/s at 45º from Frankfort plane. ICP and shear stresses were measured in the brain tissue for both head models. Moreover, to account for the effect of brain material, linear elastic and viscoelastic material models were incorporated in our analyses for both head models. Our results for the cases studies in this work showed that both FE heads predicted similar ICP peak values and patterns, and the results were not affected by the mesh size and the shear stiffness of the brain. The peak coup and countercoup ICPs for both head models were about 300 kPa and -200 kPa, respectively. ICP decreased linearly from the surface of brain (maximum at coup site), towards the center of brain where it became zero and all the way to the back of the brain (countercoup site) where negative pressures developed. However, the shear responses of the brain were significantly different between the two head models and showed strong dependence on head geometry and material models. Unlike ICP, shear stress showed a phase lag and travelled inside the brain much slower due to the brain viscous behavior. At a same point in brain, NDSUHM predicted the peak shear stress around 0.4 kPa while the DHIM predicted it about four times higher at about 1.5 kPa. We concluded that it is essential that a FE model be validated against both pressure and shear responses before being considered for TBI studies.