Center of Mass Affect Joint Kinetics Estimated by Markerless and Marker-based Systems During Running
Hui Tang,Jiahao Pan,Barry Munkasy,Li
DOI: https://doi.org/10.1249/01.mss.0000980288.27945.eb
2023-01-01
Medicine & Science in Sports & Exercise
Abstract:BACKGROUND: Markerless motion capture (ML) is gradually transforming biomechanics data collection. It has been reported that compared to the marker-based system (MB), differences in kinematic and kinetics variables estimated by the ML might be induced by pose estimations. PURPOSE: To investigate how differences in joint centers (JCs) and segment center of mass (COM) affect the lower extremity joint kinetics estimated by the two systems during treadmill running. METHODS: 15 healthy recreational active young adults were recruited. Participants run on a treadmill at 3.58 m/s for 2 minutes. The kinematic data were recorded simultaneously by eight infrared and eight high-resolution video cameras. An instrumented treadmill recorded the force data. Lower extremity JCs and segments COM were extracted from the last ten strides from both MB and ML. Distances between foot COM and ankle JC, combined foot and leg COM and knee JC, and combined foot, leg & thigh COM and hip JC were calculated, along with sagittal plane joint moments and powers at the right hip, knee, and ankle joint using data collected by the two systems. A simple linear regression was conducted using Spatial Parametric Mapping (SPM) analysis with an alpha level of .05. RESULTS: The SPM analysis showed significant associations between the distances COMs to JCs COMs and their corresponding joint kinetics. Particularly, regions of interest (ROIs) for ankle joint moments were during the 35%-72%, and 88%-96% (p < .001), whereas the ROIs for powers were during 10%-15%, and 52%-62% (p < .001) of the stride cycle. Similarly, the significant ROIs for the knee joint moments were during 22%-28%, 38%-68%, and 82%-100% (p < .001), whereas for powers were 12%-16%, 42%-60%, and 68%-95% (p < .001) of the stride cycle. For the hip joint, the ROIs of the moments were during the 40%-65%, and 85%-100% (p < .001), whereas for the powers, ROIs were during 3%-5%, 13%-20%, 42%-68% (p < .001) of the stride cycle. CONCLUSIONS: Compared to the MB system, ML posted longer distances between the COM and JC at the ankle and knee but shorter at the hip joint when capturing the motion during treadmill running, which may affect the differences in estimated kinetics between the two systems.