A model predicting individual shoulder muscle forces based on relationship between electromyographic and 3D external forces in static position.

To study the potentiality for developing an EMG-based model for the human shoulder, mapping of relations between static hand forces and electromyographic (EMG) activity of 13 shoulder muscles, were performed. The procedure was to perform by the hands slowly varying isometric forces up to 20% maximum voluntary force in the three-dimensional space. By combining these data with literature values on muscle physiological cross-sectional area and moment arm data, an EMG-based model was developed for estimating muscle forces in the glenohumeral joint. The model was validated for one standardized position by comparing joint moment, calculated from EMG by using the model, with moments from the external force. The highest correlation between these moments was found assuming a linear EMG/force calibration at low force level (< 20% MVC), giving correlations from 0.65 to 0.95 for the abduction/adduction moment and from 0.70 to 0.93 for the flexion/extension moment, for the six subjects. Moments calculated from EMG were for most subjects somewhat lower than the moments from the external force; the mean residual error ranged from 1.6 to 9.9 Nm. Taking this into account, the results can be used for assessment of muscle forces based on recordings of external forces at the hands during submaximal static work tasks without substantially elevated arms.