Characterising Shoulder Joint Mechanics in a Functional Posture Using System Identification

Abstract

Shoulder injury is the most common cause of shoulder related pain and disability. A wide range of clinical tests are currently used to assess shoulder injuries. However, these tests are unreliable and ineffective. One potential alternative for assessing the shoulder is to use system identification techniques to interrogate the function of the glenohumeral joint by perturbing the shoulder and measuring its torque and angle. System identification is defined as the process of estimating a mathematical model that describes a particular system when both the input and the output are known (angle and torque). A shoulder perturbation robot has been designed and built to measure the dynamics of the shoulder joint in two degrees of freedom (internal/external rotation and abduction/adduction). It generates perturbations and measures the torque and the angle around the joint. The advantage of this perturbation robot is that it can accurately quantify the shoulder mechanical properties with minimal interference from soft tissues. It can also be reconfigured to measure shoulder mechanics in any posture. This perturbation robot was validated by conducting experiments on a phantom of known mechanical properties to estimate its parameters. In addition, perturbation experiments on a human participant showed that the perturbation robot can produce repeatable measurements. Perturbation experiments were then conducted on 15 healthy participants to measure the mechanics of their shoulders in apprehension posture (where the shoulder is vulnerable to injuries) at various levels of submaximal contraction. Our results showed that the mechanical stiffness and damping parameters increased with contraction torques in all directions. There were no differences in the mechanical properties between the dominant and non-dominant shoulders. In addition, there was a mechanical coupling between the degrees of freedom. The role of shoulder muscle activation on its mechanical properties has been also studied by measuring the electromyography (EMG) of the shoulder superficial muscles during perturbation and submaximal contraction. No reflex activity was observed in the shoulder muscles because of the perturbation. Muscle synergies were also extracted from the EMG measurements. It has been demonstrated that each participant used a different synergy to perform contraction tasks. These differences in the synergies are associated with the differences in the shoulder mechanical properties in abduction/adduction degree of freedom. The shoulder perturbation robot can be used to measure the mechanical properties of the shoulder joint. It can also be used to assess the mechanics of injured shoulders (shoulder instability for instance) to explore how the mechanics might be altered due to shoulder injuries. Therefore, the shoulder perturbation robot has the potential to be used in clinic to monitor the shoulder recovery during rehabilitation.