I24 - Modeling the Biomechanics of Human Concussion in Swine

Introduction: : Mild Traumatic Brain Injury (mTBI), also known as concussion, affects millions of people every year. Uncovering the link between the biomechanics of impact and tissue injury holds major promise for mTBI prevention. However, detailed investigations of mTBI in living humans remains elusive due to the scarcity of high-resolution tissue-level diagnostics, such as pathological staining. Large animal models offer a unique opportunity to emulate human concussion in-vivo, however, we must ensure that the experiments are able to accurately capture human concussion biomechanics in the animal models before utilizing them in the study of mTBI.
In this study, we investigate the strain patterns in the simulations of human brain resulting from 23 concussive and 3028 impact data collected from athletes to identify commonalities in the distribution of strains in concussive impacts. We then compare this with similar data from two swine models of concussion: an established inertial loading model of concussion [1], and a newly developed model based on direct impacts to the cranium. We demonstrate that the two models can exceed concussion-level brain strains seen in human impacts.

Materials and

Methods: : We used instrumented mouthguards (iMG) to record translational and rotational head kinematics in six degrees-of-freedom (DOF) from 3051 sports impacts, among which 23 resulted in concussion [2]. The kinematics data were used to perform simulations of the impacts using a Finite Element (FE) model of the human head [3], and obtain strain maps. We then used the 95th percentile maximum principal strain (MPS95) to create a classification algorithm to detect concussion using logistic regression (Fig. 2A). The MPS95 value that corresponds to 50% probability of concussion was then used as a threshold in evaluating the swine models of concussion.
In parallel, we developed a new large animal model of concussion that uses direct impacts to the cranium to model concussion in swine. The swine receive impacts near the snout using a pneumatic linear impactor (while anesthetized), inducing rapid rotations in the sagittal plane. The 6 DOF kinematics of the impact are recorded using accelerometers and gyroscopes mounted on to the zygoma bone of the animal. We performed a total of 5 impact experiments using this model. Similar to the human data, the kinematics data is used with a swine FE model (harmonized with the human model) [4] to obtain strain maps in the brain.
In addition to this, 42 impacts performed previously using the inertial loading swine model were obtained for benchmarking the direct impact swine model.
We then used the MPS95 in animal impact simulations to evaluate their ability in reaching concussion-level strains seen in humans impacts.

Results, Conclusions, and Discussions:: A visual inspection of the human and swine impact simulations shows that the swine models can reach the same MPS95 values as seen in the human concussive impacts (Fig. 1 shows an example from the direct impact model compared to an example human concussive impact). In both models, regions of high MPS95 are distributed in two distinct areas: a narrow cortical ring on the outer surface of the brain, and around the corpus callosum.
The logistic regression analysis on the human impacts revealed that an MPS95 value of ~0.2 corresponds to 50% probability of concussion. Using this as a threshold, it can be seen that both swine models have the ability to reach concussion-level strains (Fig. 2B).
Among others, these results show that the cranial impact model enables accurate reproduction of human concussive biomechanics in the swine. The direct-impact nature of this model allows us to recreate realistic human impact situations in swine, thereby enabling us to further our understanding of impact biomechanics, as well as aiding in the development of novel preventative technologies such as helmets.

Acknowledgements and/or References (Optional):: The authors would like to acknowledge the Wu Tsai Human Performance Alliance for supporting this study.

References:
[1] R. Raghupathi and S. S. Margulies, “Traumatic Axonal Injury after Closed Head Injury in the Neonatal Pig,” Journal of Neurotrauma, vol. 19, no. 7, pp. 843–853, Jul. 2002, doi: 10.1089/08977150260190438.
[2] L. C. Wu et al., “In Vivo Evaluation of Wearable Head Impact Sensors,” Ann Biomed Eng, vol. 44, no. 4, pp. 1234–1245, Apr. 2016, doi: 10.1007/s10439-015-1423-3.
[3] D. Lyu, R. Zhou, C. Lin, P. Prasad, and L. Zhang, “Development and Validation of a New Anisotropic Visco-Hyperelastic Human Head Finite Element Model Capable of Predicting Multiple Brain Injuries,” Front. Bioeng. Biotechnol., vol. 10, p. 831595, Mar. 2022, doi: 10.3389/fbioe.2022.831595.
[4] M. Hajiaghamemar and S. S. Margulies, “Multi-Scale White Matter Tract Embedded Brain Finite Element Model Predicts the Location of Traumatic Diffuse Axonal Injury,” Journal of Neurotrauma, vol. 38, no. 1, pp. 144–157, Jan. 2021, doi: 10.1089/neu.2019.6791.