Influence of Shoulder Kinematic Estimate on Joint and Muscle Mechanics Predicted by Musculoskeletal Model

International audienceGoal: Little information is available in the existing literature regarding the influence of the scapular kinematic estimate method on musculoskeletal analysis. This study aimed to assess the propagation of errors due to the method used for scapular kinematics reconstruction in the workflow of musculoskeletal modeling (joint kinematics, joint torques, muscle force and joint reaction force) in shoulder and upper-limb movements. Methods: Two participants performed functional (arm elevation and rotation), daily life (eating and reaching pants pockets) and sports movements (a simulated throwing maneuver). Shoulder kinematics were obtained with five multibody kinematics methods: intra-cortical pins (Pins, reference method), International Society of Biomechanics (ISB), Jackson (Jack), Projection (Proj) and Ellipsoid (Ell) methods. For the five kinematics methods, joint torques, muscle forces and glenohumeral joint reaction forces were computed with the Delft Shoulder and Elbow musculoskeletal model. Results: Differences up to 30° in glenohumeral joint kinematics, compared to the Pins method, resulted in differences less than 3 N.m in joint torque estimation. However, these also resulted in differences up to 50 N and 831 N in the muscle force and joint reaction force estimate, respectively, in comparison to the reference method (Pins). No method yielded the worst or best results in comparison to the Pins method but the differences were task-specific. Conclusion: We concluded that shoulder biomechanical studies based on skin markers should be completed with caution when assessing joint angles, muscle forces and glenohumeral joint reaction forces, while researchers may be more confident with the evaluation of shoulder joint torques