Feasibility and validation of SageMotion, an IMU-based biofeedback system for gait retraining and modification: a comparative analysis with OpenCap on the execution of gait retraining techniques

This master’s thesis work is the result of six months of research conducted alongside the Human Movement Biomechanics Research Group at the KU Leuven University (Belgium), within a collaboration with the Human Movement Bioengineering laboratory at the University of Padova (Italy). The aim of this study is to test and validate SageMotion, a new wearable IMU-based haptic feedback system for real-time movement assessment and training; the main goal is to establish SageMotion’s suitability as a gait retraining and modification tool for patients affected by knee osteoarthritis. Knee osteoarthritis (OA) is a chronic degenerative joint disease that affects millions of people worldwide, causing joint stiffness, pain during gait and movement restrictions. Many studies have highlighted a correlation between a higher knee adduction moment (KAM) and the presence, severity and progression of medial compartment knee OA; for this reason, the KAM is an often investigated biomechanical variable for assessing the disease occurrence and progression. Among the possible treatments for knee OA, gait retraining has recently emerged as a noninvasive conservative strategy aiming at decreasing joint loading, thanks to the implementation of gait retraining techniques which, through the modification of the patient’s gait kinematics, allow a reduction of the KAM. For this study, a group of four healthy individuals took part to the data collection. Each participant performed three different gait retraining strategies (trunk leaning, toe-in walking and toe-out walking), which were collected simultaneously by SageMotion and OpenCap, the reference system chosen for assessing SageMotion’s validity. OpenCap is a markerless video-based system that enables 3D kinematic and dynamic analysis of human movement using videos captured with iOS devices, which already proved its reliability and suitability for rehabilitation purposes. After the acquisition, for each analyzed technique the data underwent filtering, resampling and synchronization, in order to make them comparable. The last step into the data processing consisted of a statistical analysis, which included the root mean square errors (RMSE) and coefficient of multiple correlation (MCM) calculation, a normality assessment for the groups under comparison and the computation of statistical tests, followed by a false discovery rate (FDR) correction for minimizing the rate of false positives. The analysis covered both a inter-subject comparison and a comprehensive overview concerning the whole system’s validity, regardless of the distinction between subjects. The analysis revealed comparable results between the two systems for the trunk leaning strategy, suggesting the system’s suitability as a gait retraining tool within this technique. The most controversial results occurred from the toe-in and toe-out techniques, resulting in highly significant discrepancies between the two systems for the toe-out walking, while suggesting a much better comparability between SageMotion and OpenCap for the toe-in. This contradictory outcome, together with the lack of an explained calibration procedure, recurrent connection issues during the data collection and imprecise structure of some user apps, led to the conclusion that the SageMotion system, despite its strengths and novelty, still needs to be improved and investigated before being used as a gait retraining and modification tool for patients affected by knee OA. However, this study provided a first insight into SageMotion’s capabilities, highlighting some of its limitations, that could be further analyzed in the future, ideally through a comparison against a gold standard marker-based system.