A musculoskeletal model of the rat hindlimb: Application to neuroprosthesis development and quantitative gait evaluation

A musculoskeletal model is a valuable tool for investigation of the biomechanics and control of movement, and for development and evaluation of treatments for neural deficits. This dissertation describes the development of a dynamic model of the rat hindlimb and its application. Musculoskeletal morphology and muscle physiology parameters characterizing the dynamic behavior of muscle activation and contraction as well as skeletal articulation were measured and found to be consistent under scaling between rats. These data showed muscle functions and force production at the foot were highly dependent on posture and gait phase. The model was validated with experimentally collected electromyograms, kinematics and ground reaction forces. An inverse dynamics - static optimization technique was used to solve the redundant actuator problem resulting from the high number of muscles relative to the degrees of freedom of the joints. The validated model was used to evaluate two hypotheses. (1) A metric combining physiological and mechanical measures of gait effectiveness can serve as a quantitative measure of locomotion recovery after spinal cord injury. It was found that a metric of recovery composed of muscle activation levels, body weight support and motive force was capable of quantifying several functional indicators of locomotor recovery, and of providing insight into the neural mechanics of recovery, in particular the phased recovery of different features of neural locomotor control. (2) The space of forces producible through artificial excitation of neuroanatomically derived muscle groups approximates the naturally producible force space. It was found that eleven groups corresponding are sufficient to reproduce 80% of the forces producible through individual excitation, implying that idealized excitation of a small number of nerve fascicles may be sufficient to produce forces required for standing and locomotion. This reduction in input dimension has the potential to simplify FES controller development and thereby improve functional recovery from neural injury.