Ph.D Thesis

Ph.D StudentTrainin Ehud
SubjectThe Joint Action of Biomechanical and Neural Control in
Determining Motor Behavior: A Computational Study
DepartmentDepartment of Electrical and Computer Engineering
Supervisor PROF. Ron Meir
Full Thesis textFull thesis text - English Version


The vertebrate motor system consists of the biomechanical system, the peripheral nervous system, the spinal cord and the brain. Each of these sub-systems is divided into further components. Much research has been done on each component in isolation. The approach underlying this research is that many properties of a given component may be understood only within the context of the constraints and control objectives of the overall motor system.

The first part of the study demonstrates this approach by explaining motor neural activity in the broader context of the controlled system dynamics and control objectives. The neural activity in the caudal part of monkey primary motor cortex is explained based on a model of spinal‑biomechanical system dynamics combined with control objectives. The model explains a variety of non‑trivial patterns of neural activity reported in the literature and shows how limb mechanics, muscle unidirectionality and the finite time response of the muscles and the spinal cord are responsible for basic features of the observed neural activity.

The second part of the study is targeted at modeling correctly the influence of muscle length and shortening velocity on muscle force production. We show that widely used muscle models do not always model the influence of length and velocity correctly, which sometimes lead to a significant qualitative deviation from the observed force under normal physiological conditions. We further show it is necessary to add to the cross-bridge muscle model a property of detachment rate dependency on muscle length,. 

The third part of the study is aimed explaining the biomechanical feedback achieved through the influence of muscle length and shortening velocity on muscle force production. We first show that the cat Soleus muscle is activated together with the other ankle extending muscles, which constitute together a muscle synergy, even during tasks, in which Soleus is inefficient.  Thus it seems that the muscle synergy imposes a constraint on the problem of muscle load sharing. Yet, it is shown that a variety of biomechanical properties are adapted to significantly reduce a potentially much greater problem, arising from the activation of slow muscle during fast movement, by contributing to fast force decline of a slow muscle. We further use an optimal control model to show that the control role of one of one of these properties is possible due to the non-overlap of the length ranges corresponding to the force rise and fall phases of the locomotion cycle.