Abstract
Traumatic brain injuries (TBI) are one of the least understood injuries to the body. Finite element (FE) models of the brain are crucial for understanding brain injury and developing injury mitigation countermeasures, and these models have predicted that the magnitude of human brain deformation (and the resulting brain strain and injury risk) is dependent on magnitude, duration, and direction of the angular velocity of the human head. However, this hypothesis has never been demonstrated experimentally. Furthermore, the computational models lack the experimental data necessary to validate the brain’s response to a controlled dynamic rotation that is consistent with exposures in sports and automotive crashes. Therefore, the goal of this dissertation was to improve the experimental understanding of brain deformation under rotational loading and to improve the biofidelity of FE brain modeling capabilities. The goal was achieved using both experimental and computational aims. The experimental aims focused on developing a methodology to measure in-situ human brain motion under rotational loading, and on using the methodology to build a dataset of brain deformation. The computational aims focused on developing a methodology to evaluate FE brain models in comparison to experimental data, and a framework to optimize the material properties of the models to improve their biofidelity.
A new method was developed to collect dynamic brain motion data using sonomicrometry. Small, neutrally-dense ultrasound crystals were embedded into human cadaveric brain tissue, and point-to-point distance measurements between crystal pairs were recorded during head impact. This method provided highly accurate and repeatable data that allowed for the measurement of brain deformation at various locations within the brain and for multiple severities for each specimen. A total of six cadaveric human specimens were tested and combined into a brain deformation dataset, containing approximately 5,000 displacement curves. The dataset was aggregated into average response corridors that represent the variance in brain deformation response among the tested specimens.
To encourage a consistent method of validating FE brain models, two widely used models were evaluated using various techniques of comparing the FE nodal motion to the experimental brain deformation data. A sensitivity analysis of the effect of the models’ material properties on brain motion was conducted to identify parameters that could be calibrated to improve the biofidelity of the model. The sensitivity analysis results were then used to predict improved material properties for the FE brain models. Overall, the computational aims provide an overarching framework for FE model developers to evaluate and optimize models based on the experimental dataset.
This dissertation advances the understanding of human brain deformation through the development of a methodology and dataset quantifying in situ human brain deformation. The contribution of a dataset of brain deformation, including average data corridors, will have a broad impact on the TBI biomechanics field, allowing researchers to develop and evaluate the next generation of FE brain models. An improved experimental understanding and modeling of brain mechanics will be an important step towards mitigating the incidences and consequences of TBI, thereby helping to reduce the societal burden of brain injuries.
