Highlights • A state-of-the-art FE head model enhanced with DTI data (fractional anisotropy and fiber orientation) and new brain material law is utilized to develop a model based brain injury criterion. • A total of 109 real world head trauma cases are simulated using the advanced head model. • Different intra-cerebral parameters (Von Mises stress, first principal stress, first principal strain, Von Mises equivalent strain, axonal strain, axonal strain rate, CSDM (0.1), CSDM (0.15) and CSDM (0.25)) are extracted to develop a model based criterion. • In-depth Statistical analysis was conducted over different intra-cerebral parameters to find the best suitable parameters to predict diffuse axonal injury (DAI). • Axonal strain provides the best statistical correlation among the other intracerebral parameters and the 50% risk of DAI has been established at 15% of axonal strain.

Abstract Traumatic brain injury (TBI) is the leading cause of death and permanent impairment over the last decades. In both the severe and mild TBIs, diffuse axonal injury (DAI) is the most common pathology and leads to axonal degeneration. Computation of axonal strain by using finite element head model in numerical simulation can enlighten the DAI mechanism and help to establish advanced head injury criteria. The main objective of this study is to develop a brain injury criterion based on computation of axonal strain. To achieve the objective a state-of-the-art finite element head model with enhanced brain and skull material laws, was used for numerical computation of real world head trauma. The implementation of new medical imaging data such as, fractional anisotropy and axonal fiber orientation from Diffusion Tensor Imaging (DTI) of 12 healthy patients into the finite element brain model was performed to improve the brain constitutive material law with more efficient heterogeneous anisotropic visco hyper-elastic material law. The brain behavior has been validated in terms of brain deformation against Hardy et al. (2001), Hardy et al. (2007), and in terms of brain pressure against Nahum et al. (1977) and Trosseille et al. (1992) experiments. Verification of model stability has been conducted as well. Further, 109 well-documented TBI cases were simulated and axonal strain computed to derive brain injury tolerance curve. Based on an in-depth statistical analysis of different intra-cerebral parameters (brain axonal strain rate, axonal strain, first principal strain, Von Mises strain, first principal stress, Von Mises stress, CSDM (0.10), CSDM (0.15) and CSDM (0.25)), it was shown that axonal strain was the most appropriate candidate parameter to predict DAI. The proposed brain injury tolerance limit for a 50% risk of DAI has been established at 14.65% of axonal strain. This study provides a key step for a realistic novel injury metric for DAI.