Centre of pressure (CoP) is the most common variable measured when investigating the mechanisms underlying the control of postural sway. While the CoP excursion has been thought to be a stationary signal, more recent evidence suggests that the CoP exhibits random-like, time-varying behaviour (Collins & DeLuca, 1993; Schumann et al. 1995). As these studies have assessed postural sway in two-footed stance for relatively short periods (e.g., 30 secs), it is unknown if these random, time-varying properties of postural sway occur when standing in more difficult stance positions, although such random characteristics have been found for stabilisation tasks of up to 30 minutes in duration (Treffner & Kelso, 1999).
The purpose of this study was to assess the degree of random, time-varying behaviour of the CoP signal during ten minutes of Romberg stance, using random walk and time frequency analyses. The algorithm developed by Collins and DeLuca (1993) was used for the random walk analysis, and power spectrums over 30 s intervals were used to construct the time-frequency analysis plots. Four neurologically normal, young adult male subjects volunteered to participate in this study. and were required to stand in Romberg (heel to toe) stance for one ten-minute trial. Subjects were free to select their foot placement, as long as the heel of the anterior foot was just touching, and directly anterior to, the toes of the posterior foot.
Random walk analyses were performed on the anterior-posterior (AP), medial-lateral
(ML) and planar (resultant) CoP signals. Results suggested that while the CoP
profile of the prolonged Romberg stance exhibited random-like properties, differences
were apparent between the current results of the current study and those of
Collins and DeLuca (1993) which were for shorter duration two-foot stance. A
time-frequency analysis was also performed, and showed that changes in the frequency
and power content of the CoP signal occurred over time. These results were similar
to those reported for two-footed stance over shorter time periods (Schumann
et al., 1995). Differences in the frequency parameters (frequency at maximum
power, frequency bandwidth, and mean and median frequency) were also assessed
for ten, one-minute intervals. Regardless of the within- and between-subject
variability in these parameters, all subjects tended to oscillate within a relatively
narrow frequency band (< 3 Hz). Results of the random-walk and time-frequency
analyses suggest that the random properties of the CoP signal during prolonged
stance appear to change as a function of stance duration.