Algorithm tailors walking rehab to individual needs

Results of the breakthrough from scientists at the National Centre of Competence in Research (NCCR) Robotics at EPFL and Lausanne University Hospital (CHUV) are published in Science Translational Medicine.

In a clinical study with over 30 patients, the scientists demonstrated that those wearing the smart walking assist immediately improved their locomotor abilities, enabling them to perform activities not possible without the support.

In rehabilitation involving neurological disorders or injury, teaching the nervous system to adopt the correct movements is a major challenge. The loss of muscle mass prevents people from walking correctly, as does the neurological wiring that needs to be trained to relearn proper posture and walking movements. As long as the patient repeats unnatural movements, the nervous system will continue remembering the flawed motion.

The idea of the smart walking assist is to promote natural walking in patients so that the nervous system learns how to walk normally again. Body-weight support systems are already used in rehabilitation but this study is said to be the first time such a support system operates in conjunction with an algorithm that tailors assistance to the needs of each patient.

According to EPFL, the algorithm is based on careful monitoring of the patient as he or she moves, including parameters like leg movement, length of stride and muscle activity.

Based on these observations, the algorithm determines the forces to be applied to the trunk of the body, via the smart walking assist, in order to enable natural walking patterns. This translates into either relieving the patient of his or her own weight, pushing the patient forwards or backwards, to one side or the other, or a combination of these, for a more natural posture.

“I expect that this platform will play a critical role in the rehabilitation of walking for people with neurological disorders,” said Grégoire Courtine, a neuroscientist at EPFL and CHUV.