The Delft Shoulder and Elbow Model (DSEM), a musculoskeletal model of the shoulder and elbow has been extensively developed since its introduction in 1994. Extensions cover both model structures and anatomical data focusing on the addition of an elbow part and muscle architecture parameters. The model was also extended with a new inverse-dynamics optimization cost function and combined inverse-forward-dynamics models. This study is an update on the developments of the model over the last decade including a qualitative validation of the different simulation architectures available in the DSEM. To validate the model, a dynamic forward flexion motion was performed by one subject, of which the motion data and surface EMG-signals of 12 superficial muscles were measured. Patterns of the model-predicted relative muscle forces were compared with their normalized EMG-signals. Results showed relatively good agreement between forces and EMG (mean correlation coefficient of 0.66). However, for some cases, no force was predicted while EMG activity had been measured (false-negatives). The DSEM has been used and has the potential to be used in a variety of clinical and biomechanical applications.

A finite element musculoskeletal model of the shoulder mechanism consisting of the thorax, clavicula, scapula and humerus has been used for analysis of the kinematic and dynamic behavior. The model includes 16 muscles, three joints, three extracapsular ligaments and the motion constraints of the scapulothoracic gliding plane which turns the shoulder girdle into a closed-chain mechanism. Simulations are inverse dynamic. Input variables are the positions of the shoulder girdle and humerus which have been recorded in 10 subjects during unloaded and loaded humeral abduction and anteflexion. Comparisons of muscle force predictions and EMG recordings are hampered by the unknown force-length relationship and the length dependency of EMG amplitude. It is concluded that EMG amplitude cannot be used for validation of complex musculoskeletal models. Muscle function is analyzed with help of a force and moment balance of the three joints. The moment balance includes the contributions of ligaments and the reaction forces at the scapulothoracic gliding plane. The scapulothoracic gliding plane is very important for the motions and the stabilization of the shoulder girdle. The direction and magnitude of joint reaction forces are calculated as well. It is concluded that the model provides good insight into the mechanics of the shoulder mechanism and that it enables an analysis of the function of morphological structures.

The finite element method described in this study provides an easy method to simulate the kinetics of multibody mechanisms. It is used in order to develop a musculoskeletal model of the shoulder mechanism. Each relevant morphological structure has been represented by an appropriate element. For the shoulder mechanism two special-purpose elements have been developed: a SURFACE element representing the scapulothoracic gliding plane and a CURVED-TRUSS element to represent muscles which are wrapped around bony contours. The model contains four bones, three joints, three extracapsular ligaments, the scapulothoracic gliding plane and 20 muscles and muscle parts. In the model, input variables are the positions of the shoulder girdle and humerus and the external load on the humerus. Output variables are muscles forces subject to an optimization procedure in which the mechanical stability of the glenohumeral joint is one of the constraints. Four different optimization criteria are compared. For 12 muscles, surface EMG is used to verify the model. Since the optimum muscle length and force-length relationship are unknown, and since maximal EMG amplitude is length dependent, verification is only possible in a qualitative sense. Nevertheless, it is concluded that a detailed model of the shoulder mechanism has been developed which provides good insight into the function of morphological structures.