Researchers in the WSU Elson S. Floyd College of Medicine’s Department of Nutrition and Exercise Physiology are using artificial intelligence (AI) technology in clinical exercise testing. Their novel approach could open new doors in a research area that is traditionally difficult to measure.
Identification of ventilatory constraint is a key objective in clinical exercise testing, and expiratory flow-limitation (EFL) is a well-known type of ventilatory constraint. However, EFL is difficult to measure. WSU researchers think deep machine learning, an application of AI, might provide improved results.
“The current methods for assessment of ventilatory constraints and ventilatory limitations to exercise are often imprecise and nuanced,” said Hans Haverkamp, PhD, associate professor of exercise physiology. “We built a convolutional neural network (an application of deep machine learning) to identify exercise breaths that are constrained in adults during high-intensity exercise.”
The results of Haverkamp’s recent testing and research have been published in the journal Scientific Reports. His team’s work demonstrated an accuracy of 90 percent at identifying exercise breaths that were constrained.
“Based on these initial and promising findings, this technology could become an effective tool for measuring exercise ventilatory limitation in many clinical populations,” Haverkamp said. “Our approach holds promise for significantly improving the accuracy and ease of identifying such constraints.”
Haverkamp’s team is already conducting a follow-up project with a larger sample size and a wider age range.