Using Statistics, Computational Modelling and Artificial Intelligence Methods to Study and Strengthen the Link between Kinematic Impacts and mTBIs
Master of Engineering Science
Mechanical and Materials Engineering
Mild traumatic brain injuries (mTBIs) are frequently occurring, yet poorly understood, injuries in sports (e.g., ice hockey) and other physical recreation activities where head impacts occur. Helmets are essential pieces of equipment used to protect participants’ heads from mTBIs. Evaluating the performance of helmets to prevent mTBIs using simulations on anatomically accurate computational head finite element models is critically important for advancing the development of safer helmets. Advancing the level of detail in, and access to, such models, and their continued validation through state-of-the-art brain imaging methods and traditional head injury assessment procedures, is also essential to improve safety. The significant research contributions in this thesis involve evaluating the decrease in blunt impact-induced brain axon fiber tract strains that various helmets provide by studying outputs of existing finite element brain models and implementing open-source artificial intelligence technology to create a novel pipeline for predicting such strains.
Summary for Lay Audience
Instances of concussions in sports such as ice hockey are often underreported. The mechanisms of how concussions occur are also not fully understood. Concussions are currently thought to be caused by direct or indirect impacts to the head or upper body that cause damage to the tissues that make up the brain. Using laboratory experiments and simulating impacts with detailed computer models to evaluate the effect that an impact has on the brain are important steps in learning how concussions are caused from different types of impacts and how they can be treated most effectively. Improving the quality and level of detail of such models through methods such as brain imaging and concussion metric baseline testing is of high importance for the field of brain injury research. The best way to prevent concussions in sports and other physical recreation activities where head and upper body impacts may occur is by wearing a helmet. Developing helmets that protect not just against focal injuries such as fractures or gashes, but also against tissue injuries such as concussions, is imperative to increase the safety of activities where head or upper body impacts may occur. In this research, laboratory tests were paired with computer model simulations to evaluate the relationship between impact type and severity and concussion metrics. These computer model simulations were also used to evaluate the ability of different helmets to protect against concussions. Lastly, with the goal of increasing the speed of solving individual head impact computer model simulations, artificial intelligence was implemented to use the results from existing simulations to be able to predict the output of new head impact cases. This artificial intelligence concussion prediction pipeline presents many advantages over running individual computer models for each head impact case, but most notably everything in the pipeline is open source. Being open source allows anyone with an interest in concussions and other brain injuries to explore the relation between impacts and concussions and brain injuries without the need to purchase specific software licences or have access to high performance research computers.
Patterson, Andrew Luke McConnell, "Using Statistics, Computational Modelling and Artificial Intelligence Methods to Study and Strengthen the Link between Kinematic Impacts and mTBIs" (2023). Electronic Thesis and Dissertation Repository. 9176.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Biomechanical Engineering Commons, Biomechanics and Biotransport Commons, Sports Sciences Commons