Analysis of musculo-skeletal loading using electromyography and biomechanical modelling: Dissertation

Musculoskeletal disorders of chronic nature are a serious problem in modern society. Long- and short- term effects of mechanical loads imposed on the structures of the musculo-skeletal system are a major risk factor in the development of these disorders. Two important methods for studying the operation of the musculoskeletal system and indirectly estimating the muscle forces are biomechanical modelling and electromyography (EMG). In this study, tools and methods for the assessment of the musculoskeletal load on the neck-shoulder and low back were developed. The developed tools included signal processing methods for characterizing the load on the basis of the myoelectric signal (MES), practical hardware and software tools for EMG measurements and load analysis and a biomechanical shoulder model suitable for the analysis of endurance-type activities. Also the synergistic operation of shoulder muscles during static work tasks was studied. The developed MES processing methods aim especially for the characterization of the static loading during work. A temporal analysis method based on nonlinear filtering was proposed, where the durations and frequency of the load can be analyzed as a function of the load level. The nature of the signal characteristics of the MES was studied. The myoelectric signals were found to be nonlinear and to have a structure statistically distinguishable from random noise. The dimensionality of the myoelectric signal was found to decrease during local muscular fatigue. The effect of optimization criteria and model constraints on the predicted muscle forces were studied. The developed model includes a load sharing principle, in which the time elapsed from the start of the activity decreases the allowable muscle stress levels on the basis of the stress-endurance curves of individual muscles. The stiffness of the shoulder can be constrained to produce more cocontraction in tasks where the demand for precision is high. The model predictions were compared to the myoelectric recordings from the shoulder muscles during flexion tasks. The results were plausible. Measures describing the relative orientation of the bones had an important effect on the model- predicted load-sharing pattern. The synergic operation and fatigability of the shoulder muscles during a static holding task was analyzed using both the myoelectric signals and the developed shoulder model. During the test the deltoid and the infraspinatus were the first muscles to show EMG signs of fatigue. The predictions of the biomechanical model on the order of fatigue of the shoulder muscles corresponded to the EMG results, which supports the validity of the model.