Options
The measurement of three-dimensional body segment parameters using dual energy X-ray absorptiometry and skin geometry : an application in Gait analysis
Loading...
Type
Thesis
Author
Lee, Mei Kay
Supervisor
Balasekaran, Govindasamy
Abstract
The purpose of this study was to present an in vivo approach that could be used to estimate the internal mass distribution of an object for the computation of three-dimensional (3D) body segment parameters (BSP). The method was validated using criterion values derived from experiments and based on a magnetic resonance imaging (MRI) scan of a single thigh segment. This study also examined the influence of differences in segment inertial parameters derived from different BSP estimation models on joint kinetics through the use of inverse dynamics solutions during gait for four participants with different body types.
BSP are essential inputs for the computations in kinetics of motion applied in the field of biomechanics. These data are usually obtained from population-specific predictive equations which present limitations in being representative of the population understudy. These limitations motivated the development of direct BSP measurements on living humans which, however, in turn posed other limitations such as exposure to radiation, high cost and sheer lack of facilities that could be readily available. In addition, some methods provide only two-dimensional (2D) measurements while others require intensive tomographic images for 3D reconstruction.
Therefore, the proposed method used X-ray imaging and 3D exterior geometry to acquire 3D BSP. The accuracy of the estimated and computed thickness of material using reams of paper, as well as a calibration device was evaluated. The average percentage errors between the measured and computed thickness were 2.97% and 0.96%, respectively. Criterion values BSP using magnetic resonance imaging (MRI) on a single thigh segment were also computed. Percentage errors for all BSP values derived from the proposed method compared to the criterion values were less than 2%. Percentage error between the total body mass measurements and total body mass derived from the proposed method were less than 1% for all four participants. There were significant differences (p < 0.01) between the proposed method, as compared to cadaver-based and living-based models for segment mass, centre of mass and moments of inertia. Significant phase effects (p < 0.05) were observed between the stance and swing phase based on the root mean squared errors (RMSE), ranging from 0.0177 to 0.0234 Nm.kg-1, and 0.0234 to 0.0970 Nm.kg-1 for the knee and hip joints, respectively. Peak knee flexor moments were significantly less than that derived from living-based model by 0.1400 Nm.kg-1 (34.16%) during swing phase. Greater differences were also observed in the peak joint powers, ranging between 0.0024 to 0.6123 W.kg-1 during the swing phase.
The findings suggest that variations in joint kinetics occurred during movements involving high limb acceleration. This may imply that the joint kinetics during normal gait at stance phase were relatively insensitive to BSP variations. Considering the accuracy of the method, it could be used as an in vivo method to obtain direct 3D BSP measurements. Further research can include efforts in the development of regression equations for use by those who do not have access to dual-energy X-ray absorptiometry (DEXA) machine or a 3D scanner.
BSP are essential inputs for the computations in kinetics of motion applied in the field of biomechanics. These data are usually obtained from population-specific predictive equations which present limitations in being representative of the population understudy. These limitations motivated the development of direct BSP measurements on living humans which, however, in turn posed other limitations such as exposure to radiation, high cost and sheer lack of facilities that could be readily available. In addition, some methods provide only two-dimensional (2D) measurements while others require intensive tomographic images for 3D reconstruction.
Therefore, the proposed method used X-ray imaging and 3D exterior geometry to acquire 3D BSP. The accuracy of the estimated and computed thickness of material using reams of paper, as well as a calibration device was evaluated. The average percentage errors between the measured and computed thickness were 2.97% and 0.96%, respectively. Criterion values BSP using magnetic resonance imaging (MRI) on a single thigh segment were also computed. Percentage errors for all BSP values derived from the proposed method compared to the criterion values were less than 2%. Percentage error between the total body mass measurements and total body mass derived from the proposed method were less than 1% for all four participants. There were significant differences (p < 0.01) between the proposed method, as compared to cadaver-based and living-based models for segment mass, centre of mass and moments of inertia. Significant phase effects (p < 0.05) were observed between the stance and swing phase based on the root mean squared errors (RMSE), ranging from 0.0177 to 0.0234 Nm.kg-1, and 0.0234 to 0.0970 Nm.kg-1 for the knee and hip joints, respectively. Peak knee flexor moments were significantly less than that derived from living-based model by 0.1400 Nm.kg-1 (34.16%) during swing phase. Greater differences were also observed in the peak joint powers, ranging between 0.0024 to 0.6123 W.kg-1 during the swing phase.
The findings suggest that variations in joint kinetics occurred during movements involving high limb acceleration. This may imply that the joint kinetics during normal gait at stance phase were relatively insensitive to BSP variations. Considering the accuracy of the method, it could be used as an in vivo method to obtain direct 3D BSP measurements. Further research can include efforts in the development of regression equations for use by those who do not have access to dual-energy X-ray absorptiometry (DEXA) machine or a 3D scanner.
Date Issued
2010
Call Number
QP310.W3 Lee
Date Submitted
2010