Head Impact Sensors: An Examination of their Performance, Validity and Utility in Boxing Sparring
Concussions represent a major challenge for sports medicine, as they are frequent and may lead to long-term health impairments. To better understand concussions and the risks associated with participation in contact sports, wearable technology has been developed and used with increasing frequency. This technology aims to measure head kinematics upon impact, based on the assumption that the sudden change of head velocity is related to brain damage. While understanding this relationship would allow the development of injury prevention strategies, it has remained elusive, and issues relative to the validity and accuracy of the sensors when used in-vivo may have contributed to this apparent lack of association.
The current thesis aimed to assess the performance of head impact sensors in the laboratory and in-vivo, and improve our understanding of how these sensors can be soundly used. Specific aims were to: 1. Appraise via literature review the achievements and limitations of head impact research. 2. Assess the validity of a select instrumented mouthguard under controlled impacts in the laboratory. 3. Evaluate and compare the performance and validity of an instrumented mouthguard, a skin patch and a patch attached to the headgear during boxing sparring. 4. Explore the feasibility of individual-specific approaches to associate head impact kinematics with acute concussion symptoms resulting from sparring.
Our findings highlighted that head impact sensors have been used to better understand the consequences of head impacts at various levels: from a single concussive impact to a lifetime of contact sports participation. However, the progression of head impact research is limited by large variability in technologies and methodologies, and by the inclusion of false positive impacts. The laboratory study found that the instrumented mouthguard under evaluation measured linear acceleration and angular velocity signals that adequately represented headform motion when the mouthguard was tightly coupled to the headform, although the accuracy was moderate. The in-vivo study highlighted several limitations for the use of the instrumented mouthguard, skin patch, and headgear patch during boxing sparring: • Large proportions of events recorded by the sensors (50-80%) did not meet a set of pre-defined quality criteria: the kinematic signals suggested the sensors moved independently of the skull. Such events were associated with higher peak accelerations that might be spurious. • Impacts landing close to the sensors were associated with more events being triggered and more events suggesting skull/sensor decoupling, for all three sensors. • The skin and headgear patches recorded twice as many video-verified acceleration events as the mouthguard, and there was little to no association in peak accelerations between the patches and the mouthguard.
We also observed that some boxers regularly experience concussion symptoms from sparring, and our exploration suggested that individual-specific approaches are beneficial to understanding the association between head impact kinematics and symptoms. Due to our small sample size, there was limited evidence that the accumulation of head impacts or the highest-magnitude impact sustained during a session were associated with changes in self-reported symptoms. Based on our findings, we recommend that future researchers systematically assess the quality of the raw kinematic signals as part of the data cleaning process. We also recommend that further work be done to explore the use of individualized approaches to better understand the association between head impact exposure and signs and symptoms of concussion. We reason that our findings and recommendations will advance research and sound application of head impact technology going forward.