Nonlinear dynamics quantifies gait variability patterns, which can be useful in evaluating functional ability. A commonly used nonlinear technique is detrended fluctuation analysis (DFA). Safety support structures have previously been shown to alter DFA during gait. However, the effect of a nonweight-supporting treadmill harness on DFA during gait has yet to be determined. The purpose of this study was to determine whether a nonweight-supporting harness influenced the DFA alpha metric (DFA α) of variables typically used to examine gait function. Twenty participants (10 young adults and 10 older adults) were recruited for this study. Each participant completed one testing session on a treadmill consisting of three conditions: (1) no harness, (2) harnessed, but not attached to the support frame, and (3) harnessed and attached to the support frame. Participants walked for 15 min at the same self-selected speed for each condition. The gait variables of stride time, stride length, and step width for each condition were analyzed using DFA α to examine gait function. There were no significant interactions between age group and condition for DFA α of each variable. Additionally, there were no main effects for DFA α for age group or condition. These data indicate that a nonweight-supporting harness can be used for safety without impeding the emergence of natural gait dynamics when stride time, stride length, and step width are the primary variables of interest.

Adults (N = 54, 80.78 ± 6.08 years) who reported falling during the previous 12 months participated in a 12-week wobble board training program with internal focus or external focus (EF) instructions. Verbal manipulation checks were performed after training sessions as a self-report of the attentional foci used. The percentage of sessions in which participants reported using an EF (EFSR) was subsequently calculated. Mean velocity and mean power frequency in the anterior–posterior (MVELOAP and MPFAP) and medial–lateral (MVELOML and MPFML) direction were assessed during a 35-s wobble board task at Weeks 0, 6, 12, 13, 16, and 20, with the latter three as retention tests. Piecewise linear growth models estimated treatment effects on individual growth trajectories of MVELOAP and ML and MPFAP and ML during intervention and retention periods. Regardless of condition, MVELOML significantly decreased (π = −.0019, p = .005) and MPFML increased (π = .025, p < .02) during the intervention period. In analyses including interaction terms, participants in the EF group who reported greater EFSR had superior progression of MPFAP during the intervention (π = .0013, p = .025). Verbal manipulation checks suggest a preference for and advantage of EF for facilitating postural control performance and automaticity.