An Index Finger Musculoskeletal Dynamic Model

Abstract

Neuromusculoskeletal models provide a mathematical tool for understanding and simulating human neuromechanics and motor control. The aim of this work is to extend our previously developed models and propose an index finger musculoskeletal dynamic model that serves as a tool in studying and replicating human behaviour. In particular, the focus of this work is to develop a skeletal dynamic model, a musculotendon dynamic model (Hill-type muscle model), and an activation estimation model. The Hill-type muscle model estimates musculotendon forces for given musculotendon lengths, length change rates, and muscle activations. The parameters of the Hill-type model were estimated so that the resulting normalised muscle length is within the operating muscle length and the resulting forces/torques are comparable to experimental data from the literature. In the activation estimation model, muscle activations are optimised by minimising the difference between the resulting torque from the musculotendon dynamic model and the skeletal dynamic model. In the skeletal dynamic model, the torques due to the passive joint properties and gravitational and inertial forces are modelled. Using the estimated Hill-type muscle model parameters, the resulting normalised length for all index muscles ranged between 0.97 and 1.03 in resting posture and between 0.5 and 1.5 in flexion/extension task. The resulting muscle activations ranged between 0 and 1 and related to the activation/deactivation of muscles during the motion task. Finally, the overall consistency between the proposed models is demonstrated and underlines the quality of the developed models.