Objective: To compare wearable sensor-based and marker-less 3D video-based gait analysis in two neurodegenerative disorders associated with gait dysfunction.

Background: Quantitative gait analysis supports diagnosis of neurodegenerative disorders. Technological approaches differ, and there is uncertainty about comparability of derived parameters.

Method: Patients with Parkinson Disease (PwPD, N=9, f:4, MDS-UPDRS III: 23.25 (+-11.9)) and Multiple Sclerosis (PwMS, N=19, f:8, EDSS: 3.2 (+- 1.4)), and healthy subjects (HC, N=24, f:12) were recorded with two different gait analysis systems. Wearable sensors (MoL; Mobility Lab, APDM) recorded subjects waking 7m back and forth in comfortable and maximum speed. With same speed instructions, very short walks (<3.5m) towards a 3D camera using marker-less visual-perceptive motion capture (VPC; Motognosis Labs using Microsoft Kinect V2) were performed. For comfortable speed walk, gait speed, stride length and cadence were derived from both systems, whereas for maximum speed walk, only gait speed was used. Independent t-test and Cohen’s D were used to determine disease effects against HC. Agreement between both systems were described by Spearman’s Rho.

Results: Both systems showed good correlations with lowest in comfortable gait speed in HC (rho=.68) and highest in maximum gait speed in PwMS (rho=.s87). In PwMS, comfortable speed walk, gait speed and cadence were lower than in HC for VPC (gait speed* D: -.96; cadence* D: -1.03), but not for MoL. Differences at maximum gait speed were more pronounced and observed with both systems (CPV**;  D: -1.27; MoL**; D: -1.08).In PwPD, both systems detected lower comfortable walk gait speed and stride length than in HC in (gait speed (VPC* D: -1.5; MoL** D: -1.4); stride length (VPC* D: -1.6; MoL** D: -1.1)) and lower maximum gait speed (VPC** D: -1.7; MoL** D: -2.1). (* p<.01 **p<.001)

Conclusion: Results from marker-free (VPC) and wearable (MoL) systems showed good agreement, despite differences in methodology and performed tasks.  In maximum gait speed for PwMS, VPC showed higher effect sizes, MoL effect sizes were higher in PwPD. Results generally fitted disease effects expected from the literature and in combination with outcomes relating to disease severity, both systems might be beneficial to use in clinical routine.

« Back to 2019 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/wearable-versus-marker-free-technologies-does-it-matter-for-clinical-gait-analysis/