Variability of human gait: effect of backward walking and dual-tasking on the presence of long-range autocorrelations.

Information from the central and peripheral nervous systems is continuously integrated to produce a stable gait pattern. However, stride duration fluctuates in a complex manner in healthy subjects, exhibiting long-range autocorrelations that can span over hundreds of consecutive strides. The present study was conducted to explore the mechanisms controlling the long-term fluctuation dynamics of gait. In the first part of the study, stride duration variability was evaluated on a treadmill during forward (FW) and backward walking (BW). Despite the modification of the biomechanical constraints imposed on the locomotor system, the characteristics of the long-range autocorrelations remained unchanged in both modes of locomotion (FW: H = 0.79 ± 0.04 and α = 0.58 ± 0.13; BW: H = 0.79 ± 0.11 and α = 0.53 ± 0.25). In the second part of the study, stride duration variability was assessed while the subjects were performing a dual-task paradigm that combined gait and mental calculation. The long-term variability of stride duration was similar during usual walking (H = 0.80 ± 0.06 and α = 0.57 ± 0.13) and in dual-tasking (H = 0.77 ± 0.06 and α = 0.52 ± 0.16), whereas walking altered the performance of the cognitive task. Hence, the biomechanical and cognitive interferences imposed in the present study were not sufficient to induce a modification of the long-range autocorrelations highlighted in walking variability. These observations underline the robustness of the long-range autocorrelations.