Postural sway parameters in seated balancing; their reliability and relationship with balancing performance

This study investigated a representative set of 39 parameters characterizing center of pressure movements (sway) in seated balancing, with the aims to determine test-retest reliability, to clarify the interrelations between these parameters, and to determine which parameters were related to balance loss in seated balancing. 331 subjects volunteered to perform three 30-s seated balancing trials in a single session. Ten subjects lost balance on all three trials, 34 lost balance on one or two trials. The test-retest reliability of postural sway parameters was poor with all intra-class correlations below 0.7 and below 0.4 for 9 parameters. Sway parameters were strongly intercorrelated and many parameters thus provide little added value. Parameters that had no intercorrelations above 0.7 comprised three conventional summary statistics of center of pressure (CoP) movements and 3 parameters reflecting the temporal structure of the CoP trajectories. None of the parameters was related with balance loss in univariate analyses, while multivariate models revealed that higher sway velocity and a lower short-term diffusion coefficient were related with less balance loss. This indicates that a multivariate assessment of CoP trajectories is necessary to characterize balancing performance.