The predictive and functional calibration method in 3D gait analysis using Human Body Model-II produce different 3D joint angles

Abstract

Predictive and functional calibration methods can be used to estimate joint centre and axis localisation in 3D motion analysis (1-6). The method of Harrington and the geometric sphere fit method are implemented in Human Body Model (HBM-II) as they are the most accurate predictive and functional calibration method respectively (1-6). The effect of calibration methods on kinematics is less researched although relevant for clinical interpretations. Does the Harrington predictive and the combined functional knee and hip calibration method in 3D gait analysis produce comparable 3D joint kinematics? Gait of 12 healthy subjects (11 F, mean(SD) age 26.4 (9.3)years, BMI 24.6 (2.8)kg/m2) was measured at Computer Assisted Rehabilitation ENvironment using HBM-II. Subjects started with a 6 minutes familiarisation period. Afterwards, a static model initialization was done (5 s standing in Tpose) using the predictive method of Harrington (1) followed by a measurement of three minutes walking at 1.1 m/s. Next, the system was reset and a combined functional knee (performing knee extension/flexion movements) and hip (performing starARc movement (6)) calibration was done using the geometric sphere fit method (2). A similar gait measurement was done. Data of 3D joint angles were extrapolated to strides (0-100%). For each subject, the difference in joint angle between the methods was calculated for each instant of the gait cycle. Mean differences were calculated and statistical parametric mapping (paired t-test) was used for group comparisons. Although the waveform patterns were comparable for the methods (Fig. 1A), significant differences in amplitude were observed for sagittal hip, knee and ankle angles and transverse hip angle (Fig. 1C), with maximum mean differences ranging from 3.6° to 7.4° (Fig. 1B). Mean differences in sagittal trunk and pelvis angles and frontal plane angles were smaller (range 0.0°–1.1°) and non-significant. The kinematic differences between methods varied among subjects (e.g. maximum knee flexion difference range: 1.9°-12.5°, Fig. 1D). Download : Download high-res image (457KB)Download : Download full-size image 3D gait analysis using the Harrington predictive or combined functional knee and hip calibration method results in different sagittal hip, knee, ankle angles and transverse hip angle. Differences are clinically relevant as they exceed 5°, corresponding to the measurement error for 3D gait kinematics (7). The difference of 1° in other joint angles indicates no critically interfere of the calibration method. The choice for a calibration method should be consistent in a lab and should be based on the context (4, 6). The functional method is more reliable as it is independent on marker placement, but is sensitive for measurement artefacts and quality of movements (6). This reduces repeatability and limits its use in patients having restricted range of motion. The predictive method is sensitive for marker placement and anthropometric measures (8), therefore requiring experienced operators. However this method is practical and feasible for patients and therefore widely used in the Netherlands.