Differences in centre of mass measurements between markerless and marker-based motion capture systems during balance and mobility assessments in individuals with chronic and sub-acute stroke

Centre of mass (COM) is a key measurement used to assess balance and mobility. Marker-based motion capture systems have traditionally been used to measure COM, but they are time-consuming and prone to marker error. Markerless motion capture systems offer a potential alternative, reducing setup time while maintaining accuracy. The ease of collecting markerless data may be particularly beneficial when study participants have limited mobility, such as those with stroke. This study aimed to determine the differences in COM measurements between marker-based and markerless motion capture systems during balance and mobility tasks in individuals with sub-acute stroke. Seventeen participants completed the following tasks: walking, quiet standing, sit-to-stand, rise on toes, and backward reactive stepping. COM data were analyzed using two markerless models, a default with 17 segments and a fit model with 11 segments to match the marker-based model to be compared as the reference. The results showed high correlations (R2 = 0.75 to 0.999) and low root-mean-square differences (< 2 cm) in the anterior-posterior and medial-lateral directions. Larger differences (> 4 cm) were observed in the superior-inferior direction, particularly with the default model. These findings suggest that markerless motion capture can be used to measure COM in people with stroke, and that model selection plays an important role in COM estimates.