{shortcode-31b81eec33fcd266bd1aaf7dc20e2d7e58b172c8}

Harvard researchers created an algorithm that allows a wearable robot to adapt to and assist the arm movements of stroke and Lou Gehrig’s disease patients in real-time.

The researchers’ machine learning model can allow the robot to intuit the movements of patients with upper arm disabilities due to stroke or Lou Gehrig’s disease – potentially supporting patients’ daily activities like combing their hair, brushing their teeth, and eating.

One of the aims of the project was to design an algorithm that could respond to “the variability in how many different people move their arms,” James Arnold, a lead author on the paper and researcher at the Harvard Biodesign Lab, said.

“We really gained an appreciation for just how differently having a stroke can affect many different people, and the same thing with people living with ALS,” he said. “The effect on their ability to move is so variable that it’s very challenging to create any kind of generalized approach.”

The scientists incorporated new sensors into the wearable robot that could better respond to arm movement. The sensors also recorded data on patients’ movement that allowed the researchers to train a machine learning algorithm to recognize patients’ intended actions.

“It’s trying to guess: do I think the person wants to lift their arm, lower their arm, or do they want a constant amount of support?” Arnold said. “And then, based on that information, the robot tries to assist them as they’re moving their arms.”

The study, published last month in the science journal Nature Communications, was a collaboration between the researchers at Harvard Biodesign Lab and clinical researchers treating ALS patients at Harvard Medical School and Massachusetts General Hospital .

Technology-based solutions for individuals with neuromuscular diseases are “urgent,” co-author Sabrina Paganoni, an associate professor at HMS and scientist at the Healey Center for ALS, said.

“It is important to collaborate across different teams – specifically engineers collaborating with physicians like myself – who have a sense of the clinical needs of the patient population,” Paganoni said. “It’s critical for the patient voice to be front and center during the entire development process.”

The researchers also sought to ensure the wearable robot improved day-to-day functionality.

Prabhat Pathak, the study’s co-lead author and researcher in the Harvard Biodesign Lab, said that it’s important for a patient to carry out the activities of daily living as much as possible beyond just recovery.

The ultimate goal of the project is to simplify the device so that it is more accessible to patients.

“Hopefully in the next one year or couple of years, we’ll have a product that someone can use at home or even in the clinic,” Pathak said.

—Staff writer Danielle J. Im can be reached at danielle.im@thecrimson.com.

— Staff writer Neeraja S. Kumar can be reached at neeraja.kumar@thecrimson.com. Follow her on X @neerajasrikumar.