Implementing acromioclavicular and coracoclavicular ligaments with a novel sternoclavicular posterior rotation estimation approach improves agreement between simulated muscle activation and surface electromyography in shoulder musculoskeletal simulation.

Acromioclavicular (AC) and coracoclavicular (CC) ligaments stabilize the shoulder, but they are often omitted from musculoskeletal (MSK) simulations. Because ligament force depends on clavicle-scapula kinematics, sternoclavicular posterior rotation (SCpr) estimation may affect simulated muscle activation when AC-CC ligaments are implemented. This study aimed to determine how AC-CC ligament implementation and SCpr estimation affect the agreement between simulated muscle activation and surface electromyography (sEMG) during arm elevation-lowering. Eight healthy males performed arm elevation-lowering in coronal, scapular, and sagittal planes under unloaded and 2-kg loaded. We recorded three-dimensional marker trajectories of upper-extremities and thorax and sEMG from muscles around the glenohumeral, scapulothoracic, and SC joints. Inverse dynamic MSK simulations were performed under four conditions that combined ligament implementation (with or without ligaments) and SCpr estimation. The negligible-AC approach assumed that AC rotation did not contribute to scapulothoracic rotation and estimated SCpr by optimization to minimize residual AC rotation, whereas the non-negligible-AC approach estimated SCpr by optimization after accounting for the literature-based AC contribution to scapulothoracic rotation and removing it. Agreement was quantified using Pearson's correlation coefficient r after constrained dynamic time warping. The condition combining ligament implementation with the non-negligible AC approach (Lig-NonNegAC) showed r ≥ 0.5 in 26/36 muscle-plane combinations in unloaded tasks and 23/36 in loaded tasks, corresponding to 12 muscles across three elevation planes. Ligament implementation alone under the negligible-AC assumption did not consistently improve agreement. These findings suggest that combining AC-CC ligament implementation with non-negligible-AC approach improves waveform agreement between simulated muscle activation and sEMG during the overhead arm elevation-lowering.