Despite the wide-spread use of musculoskeletal simulations and its use in estimating spinal loads, much is not known about how to best collect experimental data for modelling purposes. The primary purposes in this study were to determine the effects of tracking of running motion capture data to a model (1) with and without coupling of lumbar spine segments, and (2) with varying combinations of spinal markers. Running trials were collected from 7 participants, with each at three different speeds. The motion data was fit to the Full-Body Lumbar Spine Model (FBLS) with coupling of the lumbar spine enabled (CS) and disabled and therefore rigid (RS) in OpenSim through the Inverse Kinematics tool (IK). Different combinations of markers were chosen as tracking inputs for IK to represent experimental data collection with different marker sets. Root-mean-square (RMS) marker errors of all 13 markers along the spine for each gait cycle were calculated. The CS model resulted in 23.7% lower errors than the RS model (p < 0.001). The marker subset analysis showed that increasing the number of markers in the experimental data collection decreases the error, with the four marker tracking subsets with the highest number of markers tracked having the lowest errors. The location of the marker and timing in the gait cycle did not affect marker error. When spinal mechanics are of interest, the inclusion of a coupled lumbar spine in the model and a larger spinal marker set help better track experimental kinematics when fitting to a model.
© Copyright 2021 Journal of Biomechanics. Elsevier. Kaikki oikeudet pidätetään. / Kääntänyt https://www.DeepL.com/Translator