When Dr. Jennifer Hogg makes her presentation, “Cognitive Strategies to Increase Performance,” at a National Strength and Conditioning Association conference, she’ll discuss the potential benefits of artificial intelligence. Not merely from the cutting-edge science standpoint, but as a solution to an obstacle faced by the particular population her talk will target: U.S. Navy pilots. The obstacle? Remote assignments.
An associate professor in the Health and Human Performance department at the University of Tennessee at Chattanooga, Hogg speaks in August to NSCA’s Annual Tactical Training conference. NCSA’s Tactical Strength and Conditioning sub-specialty is for those who work with law enforcement, fire and rescue, military, protective services and other emergency personnel to promote wellness, boost performance, and reduce injury risk, the organization says. Hogg has been involved with the NCSA main national conference in the past and, for this workshop, was asked to address how military pilots might mitigate errors in flight and combat fatigue. It’s theme? Military Pilot Optimization.
“I will be presenting primarily on cognitive strategies that can be used for injury prevention and for performance enhancements” Hogg said. “The AI piece of this is, in the field or in combat or maybe in remote locations, to have an onboard kind of an ‘AI clinician,’ we might say. Such as, when somebody is suspected of having a concussion, it could do a quick-and-easy assessment or, if you’re in a training situation, it could take you through a training protocol in a ‘gamified’ manner.
“That would be a training protocol in which the pilots and air crew would engage in some visual-spatial games or tasks, or maybe engage their working memory and then give them a score at the end.”
“Then, just like any other neuromuscular training paradigm, you could have them do it three times a week for six weeks and then see if their in-flight performance improves. In theory, all of that could be done without the need for a human clinician if the AI is a high-fidelity model.”
Such possibilities could be derived from data collected from already-available virtual-reality devices with onboard eye tracking and handheld controllers that have accelerometers and gyroscopes.
“From that you can gather position data, you can gather postural sway, and then eye tracking and head movement,” Hogg said. “Those devices have reached the point that they’re sensitive enough to detect some nuanced changes in vision and eye movement. Trained against a larger database, a model could assess both the performance capability of a pilot and if they’re at risk for injury. Such as, if they sustained a concussion, are they able to return or when would they be good enough to return?”
She is currently collaborating with colleagues at Emory University in Atlanta in pursuing methods of using AI technology to improve injury and performance outcomes.
After highlighting what’s currently known and available to military pilots, Hogg said, she will identify what’s still to come.
“I think wearable sensors will be very big in athletic training, the gathering of bulk data. It’s already made training data sets,” Hogg said and cited as a leader in the field, Catapult, an Australia-based global company specializing in athletic performance technology.
“Of course, there are lots of HIPAA (federal privacy) concerns around that with access to everyone’s data, but there is tremendous potential for insight. Athlete monitoring is easily the biggest application, and I’ve seen other pretty sophisticated applications,” she said. “For instance, the use of MRI scans of hamstring and quadriceps muscles in machine learning algorithms to indicate potential for injury. The idea is to process the image from an MRI scan of the entire lower limb and use this to predict a future hamstring injury.”
Visit UTC Health and Human Performance
Hogg said she feels obliged to prepare her students for the use of AI tools they are apt to find in the workplace.
“The challenge is, I don’t quite know how to prepare them yet because the field is moving so quickly,” she said. “Communication is key because, in so many ways, the technology is a black box. There are so many hyper parameters to tune in a machine-learning model, and how you tune one of them might change the outcome entirely.
“What I can teach them is how to know which questions to ask and how to be more critical assessors of numbers, in particular. They don’t need to know everything about why a number might not be plausible, but if they can know how to spot red flags and what to do next, that would be my goal.
“I’m also teaching our students how AI might be useful for some of what they do, like drafting a rehab protocol for an injury. It might possibly be helpful in diagnosing somebody with an injury. Conversely, so much of what’s involved in this field, the athletic trainer has to be present for and has to interpret a physical exam in which they put their hands on an athlete. I don’t hear my students expressing concern about AI changing that anytime soon.”
Undergraduate studies in UTC Health and Human Performance
Graduate studies in UTC Health and Human Performance
About CHAI–the UTC Artificial Intelligence Initiative
Are you a UTC faculty member or scholar researching an AI field? Email us so that we can share your story: UTC-AI@utc.edu