Cadence and age interact to determine long‑range correlations in human walking.

INTRODUCTION: Consecutive strides during four-minute walking trials are not fixed but show temporal variability that typically deteriorates with age. This exploratory study investigated if deviations from average overground walking speed influence gait variability, including age as a covariate. In addition, we investigated whether pairs of young and older adults recruited with similar height, weight, and gait kinematics would have different gait variability. METHODS: Of sixty-nine healthy adults, the trials of fifty adults were matched based on their propensity scores using gait characteristics and anthropometrics. A Bayesian multilevel linear mixed effect model was then applied to Hurst exponents, a measure of long-range correlations (LRCs), obtained from stride intervals and lengths. RESULTS: Model fitting identified cross-level interactions between age group and cadence and between walking direction (clockwise or counterclockwise) and side (left or right leg). Very small effects suggest that stride interval LRCs are influenced by walking direction, side, and their interaction, as well as the interaction between age group and cadence. Only the age group main effect influenced stride length LRCs. SIGNIFICANCE: The results suggest that differences in overground gait variability between age groups are unlikely to be attributed to their gait speed, height, or weight. The results also point towards a needed future experiment that resolves the finding that LRCs are partially dependent on walking direction, walking path curvature, and the significance of LRC asymmetry within the context of aging, falls, and disease prevention.