Concussion Prevention

Toward Reliable Assessment of Concussion: A Study on Head Finite Element Models Considering Human Variability

The objectives of this pilot study are to develop parametric head FE models accounting for the geometric variations in the population, and conduct a feasibility study using population-based simulations to evaluate biomechanical factors and impact exposure factors associated with the higher rates of sports-related concussions in females than males. We will pursue the following three specific aims.

AIM 1: To develop statistical skull and brain geometry models accounting for morphological variations among the sampled population (>=14 year-old). CT scans from at least 100 subjects will be collected from the Department of Radiology at the University of Michigan under an existing IRB. A sequence of process, including image segmentation, landmark identification, template mesh mapping, generalized procrustes alignment, principal component analysis, and multi-variant regression analysis will be conducted to understand the variance among skull and brain geometry and predict skull and brain geometry with the most significant human parameters.

AIM 2: To link the statistical skull and brain geometry models to a State-of-the-Art midsize male FE head model, so that personalized FE head models can be rapidly generated. In this aim, we plan to use the Uniform Latin Hypercube method to sample 25 male and 25 female subjects ranging from 14 to 25 years old accounting for a large range of head sizes and shapes, and use the mesh morphing method to rapidly generate 50 FE head models corresponding to the sampled subjects.

AIM 3: To conduct a set of pair-comparison parametric simulations between sampled male and female subjects to evaluate the effects of head geometric factors on brain tissue responses under varying impact conditions. We will apply head kinematics data collected previously from high-school ice-hockey players to all the 50 FE head models with varied impact magnitudes and directions. The model-predicted brain tissue responses will be collected for each model. Statistical analysis (e.g. ANOVA and ANCOVA) will be conducted to quantify the sensitivity of head geometric factors and impact exposure factors between males and females.