SEP 05, 2017

Using AI to Help Some Patients Walk Again

WRITTEN BY: Brenda Kelley Kim

After a stroke or spinal cord injury, many patients cannot walk at all or have trouble with walking and need assistance from canes, walkers or wheelchairs. New technology that involves a mobile harness to support patients and help them with activities that would otherwise be impossible due to their injury, has been developed by scientists from NCCR Robotics at École polytechnique fédérale de Lausanne (EPFL) and at the Lausanne University Hospital (CHUV.)

The harness supports the weight of a patient in an upright position and can adjust, in real time, depending on the size and movement of the patient. In a study with more than 30 volunteers who had mobility issues after a stroke or spinal cord injury, wearing the smart walking assist harness improved their abilities to ambulate and perform activities of daily living. The study was published recently in the journal Science Translational Medicine.

After any kind of brain injury, vascular accident such as a stroke, or spinal cord injury, rehab is crucial to re-train the body to make up for the damage that impacts movement. Because mobility is often severely impaired, muscle mass can be lost and this also contributes to the nervous system deteriorating. A patient needs to learn to move differently, and re-wire the electrical signals to correct the flawed movement the injury or disease causes, but this is not always possible. That’s where the smart walking assist harness comes in. It can support the weight of the body in an upright position, taking the stress off muscles while the patient learns the proper way to move using different nerves and muscles. Algorithms play a part as well and the latest research by the team at EPFL and CHUV are the first to combine the muscle support with algorithms targeted specifically to each patient’s individual need.

The algorithms are developed after monitoring and recording each patient’s movements down the exact length of stride, leg movement and gait. Once that data is established, the algorithms are used to direct the harness to push forward or backward, to rock to one side or another or make any other adjustments based on the patient wearing it at the time.

Grégoire Courtine, neuroscientist at EPFL and at the Lausanne University Hospital explained, "I expect that this platform will play a critical role in the rehabilitation of walking for people with neurological disorders.” The research conducted by Courtine and the rest of the team at EPFL, in partnership with universities across the EU has resulted in a device called RYSEN, which is almost like a robotic physical therapist. While harnesses that hold patients upright for therapy are not uncommon, the RYSEN device is designed to adjust the load bearing in real time, using artificial intelligence (AI) algorithms to gauge what kind of support a patient requires at any given moment. Moving forward, the team hopes to have the device in larger clinical trials. The video below has more information on this exciting development in AI and assistive technology.

Sources: EPFL, Science Translational Medicine, Engineers Australia