Collaborative Research Investigates Use of Wearable Sensors in Contact Sports
A collaborative investigation was undertaken to test the accuracy of wearable sensors during contact sports. “Video Analysis Verification of Head Impact Events Measured by Wearable Sensors" was published in The American Journal of Sports Medicine. The research team included Nelson Cortes, PhD; Andrew E. Lincoln, ScD; Gregory D. Myer, PhD, ATC; Lisa Hepburn, PhD, MPH; Michael Higgins, PhD, PT, ATC; Margot Putukian, MD; and Shane V. Caswell, PhD, ATC. The research was completed in collaboration with the Sports Medicine Assessment, Research & Testing (SMART) Laboratory at George Mason University in Manassas, VA.
The goal of this study was to describe the frequency and magnitude of head impact events recorded by wearable sensors that were confirmed by video analysis. As wearable sensors are increasingly used to quantify the frequency and magnitude of head impact events in contact sports and with the growing concerns about the potential for long-term effects of multiple head injuries, data are needed to verify the quality and quantity of head impact events recorded by wearable sensors.
The study included a cohort of thirty boys and 35 girls from high school lacrosse teams. Each participant was assigned a unique sensor that was labeled to correspond to his or her jersey number. The girls each had an X-Patch sensor affixed behind the right ear and the boys had a GForce Tracker sensor secured to the inside of the helmet’s shell at the crown. To corroborate the data from the sensors, a professional videographer was stationed at a high vantage point over the midfield line. The sensors and video were calibrated before each game to ensure precise date and time information across the multiple instruments.
The study found that “65% and 32% of all head impacts recorded during boys’ and girls’ lacrosse games were verified as true game-related head impacts by video analysis, respectively.” However, the remaining 35% of boys’ and 68% of girls’ head impacts were not verified by video or were not part of the game action, suggesting a high rate of false-positive impacts and an overestimation of verified head impact events by the wearable sensors.
“Our findings indicate that the wearable sensors are highly sensitive measurement devices that can record and classify head impacts that may not actually be head impacts,” the authors said in the article’s conclusions. The rate of false-positives indicated that the sensors were identifying other bodily impacts (e.g., shoulders, torso, arms).
“These findings have practical implications for how sensors should be deployed and the existing data interpreted. Without meticulous procedures during usage, it is plausible that impacts measured during game days can be erroneously elevated because of the excessive false-positive rate,” the authors concluded.
The American Journal of Sports Medicine, 2017. DOI: 10.1177/0363546517706703