Learning to walk with a wearable robot in 880 simple steps: a pilot study on motor adaptation.

Eight healthy men, mean age 27 years with range 22-41, were a local convenience sample, so the findings may not apply to women, older adults, or people with different body sizes. The study did not test people with stroke, spinal cord injury, or other walking impairments, and the authors state that transfer to impaired users is an open question. Participants walked on a treadmill at a fixed speed of 0.9 m/s and 10 degrees incline, so the results may not apply to overground walking or different speeds and surfaces. There was no randomised allocation or separate control group; each participant followed the same sequence of zero-force and assisted walking. The assistive settings were fixed for all participants and chosen from prior work with one pilot participant, so individual tuning was not tested. The authors note that differences in walking conditions and wearable robots make direct comparison with previous adaptation studies difficult. The study did not include unexpected no-assistance catch trials, speed or surface changes, or dual-task conditions, which would help test the internal model and cognitive load ideas. Mean muscle activity and joint kinematics did not explain all variation in energy expenditure.

The only funding named in the paper is the National Centre of Competence in Research Robotics.

Do not treat the 6.5% energy reduction as proof that the device helps patients walk. It was the largest average reduction in eight healthy men walking uphill at a fixed speed, and the study did not test people with gait disorders.