A three-dimensional regression model of the shoulder rhythm

Abstract

Objective: The development of a thorax-fixed regression model for the shoulder which statistically predicts the orientation of the clavicle and the scapula from the humerus orientation.

Background: The application of three-dimensional position recording systems to the shoulder mechanism is limited to laboratory conditions. Studies in the field of ergonomics and sports require a method that can be applied in situ. It was found that the relation between the scapular and the humeral motions is consistent. In order to facilitate the biomechanical research on the shoulder a descriptive statistical model of the shoulder rhythm was developed.

Methods: The orientation of the shoulder bones of 10 subjects was determined in a large range of 23 humerus positions. The elbow was flexed in a splint and the arm was fully supported. The subjects exerted a 20 N external abduction and adduction force at the elbow in a plane perpendicular to the humerus. Other forces and moments were mechanically minimized. During the task the postures of the shoulder were recorded. The linear regression equations for the clavicular and scapular orientations were determined by means of a repeated measurements multi-variate analysis of variance for the co-variates: humerus orientation, initial orientation of the clavicle and the scapula, external force direction, gender and morphological characteristics of the subjects.

Results: The orientation of the clavicle and the scapula was predicted by five linear regression equations, including the co-variates: humerus orientation, external force direction and initial position. Morphology and gender did not significantly contribute to the clavicular and scapular orientation predictions.

Conclusions: A statistical model is developed for the prediction of clavicular and scapular orientations, based on the humerus position, the initial posture and the direction of the external force. The model fitted well on an independent set of recorded position data for a different group of subjects.

Relevance: Morphological data of the shoulder girdle and gender did not significantly contribute to the model structure.