Comparison of body sway measured by a markerless low-cost motion sensor and by a force plate

Purpose The measurement of body sway is the standard method in the assessment of postural balance. Although the usual technique involves the record of center of pressure (CoP), it may also be done by the measurement of center of mass (CoM). This study compared the CoM estimated from joint positions reported by the Microsoft Kinect v2 (CoMk) to the CoP tracked with a force plate and CoM obtained from CoP (CoMp). Methods The signals of CoP and CoM of forty-six subjects in medial–lateral and anterior–posterior direction were acquired during quiet standing with eyes open or closed, on stable (floor) or unstable (foam pad) surface. Stabilometric parameters (sway area, velocity, range, RMS value, and the frequency band which contain 80% of the signal energy) based on CoP and CoM were calculated to evaluate the performance among conditions. The Spearman correlation coefficient ( r s ) between parameters based on all measurements was also computed. Results Five parameters of CoP were able to distinguish differences among all four conditions, four parameters of CoMp, and three of CoMk. Sway area, range, and RMS values from CoP and CoMk presented high correlation ( r s > 0.7), velocity showed high or moderate correlation (0.3 < r s < 0.9), and frequency parameters presented moderate or low correlation ( r s < 0.7). CoMk and CoMp parameters showed high correlation, except velocity which presented moderate correlation. Conclusion Although the CoM measurements via Kinect v2 have limitations as compared to traditional force plate measurement, these results support the use of the Kinect device for research and clinical assessment of body balance.