Researchers at the University of Michigan have developed a machine learning model that uses data from wearable sensors to evaluate patients’ performance during balance training exercises at home. The system, which employs four inertial measurement units (IMUs) worn on the body, aims to provide feedback similar to what physical therapists offer in clinical settings.
The research team believes this technology could help patients make faster progress in physical therapy and maintain their abilities after completing prescribed sessions. It may also assist physical therapists in making informed health care decisions.
“Our machine learning model used data from wearable sensors to predict how physical therapists would rate patients’ performance on balance exercises, providing a basis to make recommendations about the most appropriate set of exercises to perform next. This type of AI-based support would be helpful in between appointments or after people complete their insurance-reimbursed sessions with a clinician,” said Kathleen Sienko, Arthur F. Thurnau Professor of mechanical engineering at U-M and senior author of the study published in the Journal of NeuroEngineering and Rehabilitation.
The model was created using sensor data combined with assessments from physical therapists. Funding for the project came from the National Science Foundation and U-M AI & Digital Health Innovation.
Balance training is important for reducing fall risk, especially among older adults and those with sensory or motor impairments. Physical therapists typically observe patients during clinic visits to assess difficulty levels and recommend challenging but safe exercises.
The researchers are also considering ways to use this technology for remote care, which could benefit rural patients by reducing travel while keeping clinicians involved.
“Understanding what the patient and the therapist need has to be part of the algorithms we put together. I’m excited to merge different types of data to create a decision support system for both parties,” said Leia Stirling, co-author and professor of industrial and operations engineering and robotics at U-M.
To develop their model, participants performed standing balance exercises while wearing 13 IMUs attached with velcro straps. Physical therapist participants watched videos of these sessions and rated each exercise’s difficulty on a scale from 1 to 5.
The team trained convolutional neural networks—a type of machine learning model—to predict balance difficulty based on sensor data. They found that using just four sensors placed on each thigh, lower back, and upper back was sufficient for accurate predictions. The model achieved nearly 90% accuracy compared with expert ratings.
“It is very important to understand both the strengths and potential failure modes of machine learning in physical therapy, where people’s well-being is directly at stake. For example, an overfitted model may perform poorly with new patients, leading to mispredictions and unsafe exercise recommendations. To protect patients, these systems should be validated on real-world data and used with therapist oversight so unexpected or risky suggestions can be caught before harm,” said Xun Huan, associate professor of mechanical engineering at U-M.
A related study published in IEEE Transactions on Neural Systems and Rehabilitation Engineering involved physical therapists wearing eye tracking glasses while assessing participants’ balance exercises. This helped researchers learn more about how therapists make decisions as tasks become more difficult.
“It was interesting to see how complicated the physical therapists’ balance assessments are and to consider how best to capture the factors they consider in our models,” said Emma Nigrelli, doctoral student in mechanical engineering at U-M and lead author of the eye tracking study.
Looking ahead, the team hopes their work will lay groundwork for wider adoption of machine-learning-assisted balance training tools.
“In some regions, access to physical therapists specializing in balance rehabilitation may not be possible,” Sienko said. “I was excited by the possibility of developing something that could expand access to services like balance training—not only for people in rural areas across the U.S. who may lack regular access to physical therapists, but also for individuals globally.”
Safa Jabri, Jeremiah Hauth, Lauro Ojeda and Jenna Wiens from U-M College of Engineering as well as Wendy Carender from Michigan Medicine Department of Otolaryngology contributed to this research.
Funding was provided by the National Science Foundation (CMMI-2125256; 2125256) and University of Michigan. The team is seeking partners interested in bringing this technology into practical use.
