Multifractality in stride-to-stride variations reveals that walking involves more movement tuning and adjusting than running

Taylor J Wilson; Madhur Mangalam; Nick Stergiou; Aaron D Likens

doi:10.3389/fnetp.2023.1294545

Multifractality in stride-to-stride variations reveals that walking involves more movement tuning and adjusting than running

Front Netw Physiol. 2023 Oct 19:3:1294545. doi: 10.3389/fnetp.2023.1294545. eCollection 2023.

Authors

Taylor J Wilson¹, Madhur Mangalam¹, Nick Stergiou^{1

2}, Aaron D Likens¹

Affiliations

¹ Division of Biomechanics and Research Development, Department of Biomechanics, Center for Research in Human Movement Variability, University of Nebraska at Omaha, Omaha, NE, United States.
² Department of Physical Education and Sport Science, Aristotle University, Thessaloniki, Greece.

Abstract

Introduction: The seemingly periodic human gait exhibits stride-to-stride variations as it adapts to the changing task constraints. The optimal movement variability hypothesis (OMVH) states that healthy stride-to-stride variations exhibit "fractality"-a specific temporal structure in consecutive strides that are ordered, stable but also variable, and adaptable. Previous research has primarily focused on a single fractality measure, "monofractality." However, this measure can vary across time; strideto-stride variations can show "multifractality." Greater multifractality in stride-tostride variations would highlight the ability to tune and adjust movements more. Methods: We investigated monofractality and multifractality in a cohort of eight healthy adults during self-paced walking and running trials, both on a treadmill and overground. Footfall data were collected through force-sensitive sensors positioned on their heels and feet. We examined the effects of self-paced walking vs. running and treadmill vs. overground locomotion on the measure of monofractality, α-DFA, in addition to the multifractal spectrum width, W, and the asymmetry in the multifractal spectrum, W_Asym, of stride interval time series. Results: While the α-DFA was larger than 0.50 for almost all conditions, α-DFA was higher in running and locomoting overground than walking and locomoting on a treadmill. Similarly, W was greater while locomoting overground than on a treadmill, but an opposite trend indicated that W was greater in walking than running. Larger W_Asym values in the negative direction suggest that walking exhibits more variation in the persistence of shorter stride intervals than running. However, the ability to tune and adjust movements does not differ between treadmill and overground, although both exhibit more variation in the persistence of shorter stride intervals. Discussion: Hence, greater heterogeneity in shorter than longer stride intervals contributed to greater multifractality in walking compared to running, indicated by larger negative W_Asym values. Our results highlight the need to incorporate multifractal methods to test the predictions of the OMVH.

Keywords: DFA; complexity; gait; locomotion; multifractal spectrum; multiscale.

Abstract

Grants and funding