Clinical Biomechanics · Nov 9, 2026

Markerless motion capture–derived sway and sit-to-stand kinematics during routine clinical balance tasks in outpatient rehabilitation: a cross-sectional study

By Azhdar M, Loyola-Sanchez A, Timmermans D et al.

Background Markerless motion capture (MMC) may offer a practical way to quantify movement during routine clinical balance tasks, but descriptive data from real-world outpatient settings remain limited. This study described MMC-derived sway and sit-to-stand kinematics during Romberg and five-times sit-to-stand (FTSTS) testing in outpatient rehabilitation. Methods In this exploratory cross-sectional observational study, adults receiving outpatient rehabilitation for musculoskeletal or neurological conditions (n = 61; mean age 67.8 ± 10.6 years) completed 30-s Romberg testing with eyes open and closed and FTSTS. A single depth-camera MMC system exported time-series kinematics for offline processing. Romberg outcomes included mediolateral (ML) and anteroposterior (AP) sway amplitude and sway speed; FTSTS outcomes included peak rise, sit, lean-forward, and lean-backward speeds. Findings ML sway speed was higher with eyes closed than eyes open (Z = 3.089, p = 0.002; r ≈ 0.40) and remained significant after Bonferroni correction. No clear eyes-open/eyes-closed differences were observed for ML sway amplitude, AP sway amplitude, or AP sway speed. FTSTS peak speeds showed between-participant variability, with lean-forward and lean-backward speeds more dispersed than rise and sit speeds. Mean (SD) peak rise speed was 0.309 (0.088) m/s, peak sit speed was −0.285 (0.090) m/s, peak lean-forward speed was 0.040 (0.033) m/s, and peak lean-backward speed was −0.043 (0.034) m/s. Interpretation Single-camera MMC generated quantitative sway and sit-to-stand kinematic descriptors during routine outpatient tasks. These exploratory findings provide preliminary real-world information on MMC-derived measures and support further investigation using external reference systems.

doi.org/10.1016/j.clinbiomech.2026.107021 PubMed 43019875 Biomarkers