Evaluation of the effect of scapular motion on musculoskeletal modeling of the shoulder and upper extremity
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Musculoskeletal modeling possesses the ability to provide information on physiological parameters that cannot be directly measured. However, the validity of the results must be assessed to ensure that the model is sufficiently robust to recreate the mechanics occurring in vivo. The most widely used model of the upper extremity, MoBL ARMS, possesses limitations in representing natural scapular kinematics. Motion capture offers reliable quantification of humerothoracic (HT) motion, but accurate measurements of scapulothoracic (ST) and glenohumeral (GH) contributions to dynamic HT motion are difficult to obtain. In an effort to circumvent this issue, the MoBL ARMS model prescribes ST kinematics that relate to HT motion via regression equations, creating a relationship referred to as scapular rhythm. It is unknown how well this model replicates natural scapular kinematics for motions that generally follow the rhythm, such as shoulder abduction, and those that do not, such as forward reach. Furthermore, it is unknown how scapular kinematic differences affect the torque demands on the GH joint and ultimately influence model-predicted muscle activations. The purpose of this study was to evaluate the validity of the MoBL ARMS model, and specifically, the validity of using prescribed scapular kinematics to replicate natural, unconstrained scapular motion. The expected outcomes include a clear understanding of 1) how well the scapular rhythm model can recreate natural ST motion during shoulder abduction and forward reach, 2) how kinematic differences affect GH torque demands, and 3) how kinematic differences influence model-predicted muscle activations. The orientations of the trunk, scapula, and upper extremity segments of five healthy subjects were measured with motion capture during two motions—shoulder abduction and forward reach. Simultaneously, fine-wire and surface electromyography (EMG) was collected for 17 muscles of the shoulder. Simulations were run on the published and publicly available MoBL ARMS model that prescribes ST kinematics, and a modified version of the model that allows for natural, unprescribed scapular kinematics. Differences between the two versions of the model indicated how well prescribed kinematics followed natural scapular motion and how ST kinematic errors influenced GH joint torques. Finally, model-predicted muscle activations were compared with EMG. Results demonstrated that the scapular rhythm of the MoBL ARMS model is not capable of replicating ST kinematics to a clinical meaningful degree (within 10 degrees). Despite the substantial kinematic errors, only modest errors in GH torque were observed. Model-predicted muscle activations displayed moderate agreement with EMG for on/off timing; however, correlations between the model’s activations and EMG were poor. Agreement between the model and EMG was greater for prime movers than for stabilizing and inactive muscles. These results show that the model may be able to recreate the activations of muscles which drive a motion, but is not sensitive enough to reliably predict the activities of those with subtler functions. Agreement between the model-predicted muscle activations and EMG did not appear to be related to ST kinematic errors. Considering the lack of a clear link between ST kinematic error magnitude and improved activation agreement, it can be inferred that other factors play a substantial role in determining model results. These findings can aid researchers in choosing which future applications are suitable for investigation with the MoBL ARMS model and assist in interpretation of model results.