Sports science at the frontier of elite athletic performance

Dr. Adam Kiefer highlights the importance of physiological, emotional and perceptual-cognitive behavior in elite sports performance training.

Dr. Adam Kiefer is an assistant professor of exercise and sport science at the University of North Carolina at Chapel Hill, USA, and a co-director of the Simulation, Training, Analytics and Rehabilitation (STAR) Heel Performance Laboratory. His research takes a complex systems approach to both performance enhancement and injury prevention in sports, by integrating portable technologies with artificial intelligence (AI) to improve training and clinical outcomes.

What is the overarching vision of your research?

I co-direct the Simulation, Training, Analytics and Rehabilitation (STAR) Heel Performance Laboratory. Our mission is to conduct research and innovate real-world applications at the intersection of behavioral and movement science, immersive technologies, and artificial intelligence to improve health and performance outcomes.

The overarching vision of our research is the digital transformation of sports training to provide data-driven insights relative to what athletes are seeing and thinking as they perform. Ultimately, we want to make these insights actionable for coaches and athletes to develop personalized training interventions that maximize performance enhancement for each athlete.

Our initial research focuses on integrating mobile eye tracking with standard high-definition video cameras to develop a vision-in-action digital solution. With this, we investigate questions around how physical and mental stress influence visual attention and motor performance in dynamic, sport-specific tasks.

What has inspired you to embark on this journey, and what keeps you motivated to carry on?

First, as sports and sports training become more and more developed and professionalized, the physical capabilities of athletes as a differentiator become less and less.

The development of expertise in sports requires moving beyond a focus on the physical domains such as strength, power and endurance and toward a new frontier of performance enhancement focused on psychological, emotional and perceptual-cognitive behaviors.

However, unlike the physical domains, which are relatively easy to measure and apply to objective training applications, these other behavioral domains are not as easily understood or readily translatable. Thus, there is a need to identify new behavioral markers that index modifiable psychological and perceptual-cognitive processes that can inform the design and implementation of training interventions.

Second, while mobile eye tracking hardware has made it easier than ever to capture data in sports training contexts, there are currently no solutions for expediting data processing or contextualizing the data relative to meaningful events during a given training session. This has led to a data processing bottleneck that, in turn, has severely limited the adoption of eye tracking for research and training. Solving this problem will accelerate research applications in these areas, while also making eye tracking data more accessible to more people by lowering the technical barrier associated with data coding and post-processing.

What would you highlight as the main finding of your work so far?

We have spent the last year developing a digital platform to automate the data collection, analysis, and reporting of eye gaze and movement data to minimize the human workload required for data post-processing and analysis. Further, we have built this platform in the cloud, which has expedited the development and training of machine learning models necessary for data processing automation. Our main highlight is that we have successfully completed version 1.0 of this platform and have applied for a patent in partnership with the University of North Carolina at Chapel Hill’s Office of Technology and Commercialization. This fall, we completed the licensing of this intellectual property to Elipsys LLC, a sports technology start-up company co-founded by our team.

We have also deployed portions of this platform for research in both basketball and football to examine how the eye tracking variable Quiet Eye is affected by performance stress, including task pressure and fatigue. Quiet Eye is an index of the athlete’s ability to quiet their focus, assessed through their eye gaze fixation data, during specific phases of task performance. Importantly, it is a true vision-in-action performance measure. While our data collection is ongoing, we will soon have preliminary profiles of this measure in both club and elite athletes across both basketball and football.

How has your work benefited from employing eye tracking in your experiments?

Eye tracking has been integral to our research, and our team is very optimistic about this technology, along with a supportive digital infrastructure, unlocking the next frontier of sports performance assessment and training. The ease of use of Tobii Pro Glasses 3 eye tracking hardware along with the accessibility of the data streams for construction of custom data pipelines and post-processing, has contributed to our successful prototyping and deployment of our digital platform.

From your current perspective, and extensive experience with eye tracking, what would you advise those considering to adopt it in their research?

When considering mobile eye tracking, it is easy to collect one’s desired data. That capability is rarely the issue. However, what most overlook is how easy or difficult it will be to contextualize that data relative to a specific research question. This problem is compounded when the experimental environment is more dynamic in nature. So, it is important to consider how data will be time synced (when the technology is used with other peripheral hardware devices) and event coded to provide meaningful structure to the data. This can take quite a bit of time, even with a small hand-coding team. Therefore, these are important considerations that need to be discussed with related problems needing to be solved prior to the start of any experiment.

Are you able to share your future ambitions with your work?

We are pursuing several exciting avenues as part of our future work. Related to digital transformation, we will continue to develop our automated processing digital platform across more sports and sport-specific contexts.

We are also exploring action recognition models to further contextualize the eye gaze and movement data while allowing our research to scale to more complex, competitive environments. This development will, in turn, continue to accelerate our research program by alleviating the bottleneck of data coding and post-processing so that we can more quickly ask and answer important research questions relevant to athlete performance.

Related information

To get the newest from Dr. Kiefer’s lab, follow him on ResearchGate and the lab’s Twitter account.