Evaluating the validity and reliability of inertial measurement units for determining knee and trunk kinematics during athletic landing and cutting movements

Inertial Measurement Units (IMUs) are promising alternatives to laboratory-based motion capture methods in biomechanical assessment of athletic movements. The aim of this study was to investigate the validity of an IMU system for determining knee and trunk kinematics during landing and cutting tasks for clinical and research applications in sporting populations. Twenty-seven participants performed five cutting and landing tasks while being recorded using a gold-standard optoelectronic motion capture system and an IMU system. Intra-class co-efficients, Pearson's r, root-mean-square error (RMSE), bias, and Bland-Altman limits of agreements between the motion capture and IMU systems were quantified for knee and trunk sagittal-and frontal-plane range-of-motion (ROM) and peak angles. Our results indicate that IMU validity was task-, joint-, and plane-dependent. Based on good-to-excellent (ICC) correlation, reasonable accuracy (RMSE < 5 degrees), bias within 2 degrees, and limits of agreements within 10 degrees, we recommend the use of this IMU system for knee sagittal-plane ROM estimations during cutting, trunk sagittal-plane peak angle estimation during the double-leg landing task, trunk sagittal-plane ROM esti-mation for almost all tasks, and trunk frontal-plane peak angle estimation for the right single-leg landing task. Due to poor comparisons with the optoelectronic system, we do not recommend this IMU system for knee frontal -plane kinematic estimations.

Automated summary

The study found that the validity of IMU system varies in different tasks, joints, and planes. The IMU system is recommended for estimating knee sagittal-plane ROM during cutting, trunk sagittal-plane peak angle during the double-leg landing task, trunk sagittal-plane ROM for almost all tasks, and trunk frontal-plane peak angle for the right single-leg landing task with good-to-excellent correlation and reasonable accuracy.