Importance of consistent datasets in musculoskeletal modelling: a study of the hand and wrist

Abstract

Hand musculoskeletal models provide a valuable insight into the loads withstood by the upper limb; however, their development remains challenging because there are few datasets describing both the musculoskeletal geometry and muscle morphology from the elbow to the finger tips. Clinical imaging, optical motion capture and microscopy were used to create a dataset from a single specimen. Subsequently, a musculoskeletal model of the wrist was developed based on these data to estimate muscle tensions and to demonstrate the potential of the provided parameters. Tendon excursions and moment arms predicted by this model were in agreement with previously reported experimental data. When simulating a flexion-extension motion, muscle forces reached 90 N among extensors and a co-contraction of flexors, amounting to 62.6 N, was estimated by the model. Two alternative musculoskeletal models were also created based on anatomical data available in the literature to illustrate the effect of combining incomplete datasets. Compared to the initial model, the intensities and load sharing of the muscles estimated by the two alternative models differed by up to 180% for a single muscle. This confirms the importance of using a single source of anatomical data when developing such models.