Assessment of Anatomical Representations of the Trunk Muscles in EMG-Assisted Spinal Load Models

Abstract

This study evaluated the differences in biomechanical model performance and predicted spinal loading on L5/S1 for controlled sagittally symmetric lifting exertions, utilizing three different EMG-assisted biomechanical models comprising different torso anatomical inputs. The first model was based on muscle lines-of-action and prediction of cross-sectional areas from an historical biomechanical model. The second model was based upon observations from an MRI study on torso geometry. The third model consisted of a hybrid approach, where geometric inputs from MRI observations were supplemented with lines-of-action observed from cadaver studies, where necessary. To evaluate the models, spinal loading and model performance was evaluated for thirty-five males performing sagittally symmetric lifting tasks starting from 55 deg trunk flexion to upright neutral. All three models resulted in comparable and acceptable model performance. However, differences were observed in predicted spinal loading. The hybrid model resulted in smaller compression force and sagittal plane shear force when compared to the MRI-only model, and smaller sagittal plane shear forces when compared to the historical biomechanical model. These results indicate that the geometric representation of the torso anatomy needs to be considered when evaluating spinal loading predicted from different biomechanical models for comparison to tolerance estimates in attempts to evaluate risk of injury.