Heightened bilateral foot coupling in older adults under cognitive load: beyond sway variability and complexity

Postural control under cognitive load is a sensitive marker of aging, yet most dual-task posturography studies analyze the global center of pressure as a single time series, overlooking how the two feet coordinate. Whether the predicted loss of adaptive complexity manifests in individual-limb dynamics or in bilateral coupling remains unclear. Forty-four community-dwelling adults (22 young, mean age 23.27 ± 1.93 years, 12 male; 22 older, mean age 73.05 ± 4.28 years, 7 male) completed standing single-task and two dual-task conditions (visual and auditory n-back) on dual force plates. Linear measures were low-pass filtered, whereas nonlinear measures (multiscale entropy, detrended fluctuation analysis, multifractal DFA, recurrence quantification, and Higuchi fractal dimension) were computed unfiltered because low-pass filtering can distort these measures. Univariate nonlinear analyses revealed widespread anteroposterior Group main effects (DFA, multifractal DFA, recurrence quantification, and Higuchi fractal dimension; F = 6–43, p ≤ 0.018) that were detected in the unfiltered pipeline but not under conventional low-pass filtering, although dual-task-specific Group × Condition interactions were smaller and consistently mediolateral (largest partial η 2 = 0.095). Bilateral anteroposterior coherence revealed a robust Group × Condition interaction ( F (1.83, 73.28) = 9.703, p < 0.001, partial η 2 = 0.195). Older adults exhibited high baseline bilateral anteroposterior synchronization (0.750) that decreased under dual-task (0.716–0.738), whereas young adults showed low baseline synchronization (0.474) that remained stable or slightly increased (0.520–0.534). Additional coupling measures showed significant Group main effects (cross-sample entropy in both directions and mediolateral multiscale cross-entropy, all p < 0.01). In exploratory, FDR-uncorrected analyses, the mediolateral multifractal DTC–cognition correlation present under filtering ( r = −0.41) attenuated in the unfiltered pipeline ( r = −0.26), whereas an anteroposterior association emerged ( r = −0.37) and requires confirmation. These findings suggest that older-group nonlinear differences are concentrated in the anteroposterior direction as Group main effects and were not detected under conventional filtering, whereas the dual-task-specific signature manifests in bilateral foot coordination rather than individual complexity. Bilateral coupling measures, computed with unfiltered nonlinear processing, are positioned as a candidate tool for assessing cognitive-motor integrity in older adults.