Neuromusculoskeletal human multibody models for the gait of healthy and spinal-cord-injured subjects

Abstract

Spinal-cord-injured (SCI) subjects with some level of hip actuation but with no or very weak actuation at knee and ankle levels can often walk with the assistance of knee-ankle-foot orthoses (KAFO) and crutches. However, since knees are kept extended at all times, even during the swing phase, gait becomes very uncomfortable as the hip must be raised for the swing, thus leading to a high energetic cost which makes fatigue appear quickly. Moreover, the use of crutches leads to high loads in some joints, particularly shoulders, which may produce injuries in the long term. The result is that many patients prefer to use the wheelchair, thus losing the bene ts of walking for rehabilitation and for their general health state. To alleviate these problems, active orthoses have been proposed that detect swing intention and launch a knee exion-extension cycle, thus approaching healthy gait, which can make gait more comfortable and increase the actual chances of walking for users. Some magnitudes, as joint reactions or energetic cost, may serve as indicators to evaluate the e ect of assistive devices in a speci c user and, hence, the likeliness of their actual use in the long term. The combined use of motion-force-EMG capture and personalized neuromusculoskeletal models enables to estimate the mentioned magnitudes, without the need for long and tiring experimental tests which could hardly be undergone by these subjects. Therefore, the present work is devoted, on the one hand, to review, select, adapt, improve and/or develop, and validate, all the necessary methods and tools which are required for this purpose: (i) personalized musculoskeletal models; (ii) methods for analysis at skeletal and musculoskeletal levels; (iii) methods for energetic cost estimation. And, on the other hand, to apply them to two SCI subjects, bilateral and unilateral, respectively, and to compare a new active KAFO with a conventional passive one. Results show that the evaluation of device impact through the mentioned indicators is e ective, and that the new active system objectively streamlines walking to users