Concurrent validity and within-session reliability of gait kinematics measured using an inertial motion capture system with repeated calibration

Introduction: Wearable inertial measurement units (IMUs) enable gait analysis in the clinic, but require calibrations that may affect subsequent gait measurements. This study assessed concurrent validity and within-session reliability of gait kinematics measured by a frequently calibrated IMU-based system. Calibration pose accuracy and intra-rater repeatability, and IMU orientation tracking accuracy, were additionally quantified. Methods: Calibration poses and gait were recorded in 15 women using IMUs and optical motion capture (OMC) (reference standard) simultaneously. Participants performed six consecutive trials: each comprising a calibration pose and a walk. IMU tracking was assessed separately (once-off) using technical static and dynamic tests. Differences of > 5 degrees constituted clinical significance. Results: Concurrent validity for gait revealed clinically significant between-system differences for sagittal angles (root-mean-square error [RMSE] 6.7 degrees-15.0 degrees; bias -9.3 degrees-3.0 degrees) and hip rotation (RMSE 7.9 degrees; bias -4.2 degrees). After removing modelling offsets, differences for all angles (except hip rotation) were < 5 degrees. Gait curves correlated highly between systems (r > 0.8), except hip rotation, pelvic tilt and -obliquity. Within-session reliability of IMU-measured gait angles was clinically acceptable (standard error of measurement [SEM] < 5 degrees). Calibration poses were repeatable (SEM 0.3 degrees-2.2 degrees). Pose accuracy revealed mean absolute differences (MAD) < 5 degrees for all angles except sagittal ankle, hip and pelvis. IMU tracking accuracy demonstrated RMSE <= 2.0 degrees. Conclusion: A frequently calibrated IMU system provides reliable gait measurements; comparing highly to OMC after removing modelling differences. Calibration poses can be implemented accurately for most angles and consistently. IMU-measured gait data are clinically useful and comparable within participants, but should not be compared to OMC-measured data. (C) 2020 Elsevier Ltd. All rights reserved.