Detecting dynamics of human gait using improved rescaled range analysis
Zhuang Jian-Jun,Ning Xin-Bao,Yang Xi,Hou Feng-Zhen,Huo Cheng-Yu
DOI: https://doi.org/10.3321/j.issn:0469-5097.2008.01.005
2008-01-01
Abstract:The stride interval is a measure of the gait rhythm and is typically defined as the time from the initial heel strike to the next heel strike of the same foot.At first inspection,human walking appears to be a periodic and regular process.However,closer examination reveals small fluctuations in the gait pattern,even under constant conditions.One possible explanation for these stride-to-stride variations in the walking rhythm is that they simply represent uncorrelated noise superimposed on a basically regular process.A second possibility is that there are finite-range correlations: the current value is only influenced by the most recent stride intervals,but over the long term,fluctuations are random.A third and less intuitive possibility is that the fluctuations in the stride interval exhibit long-range correlations,as seen in a wide class of scale-free phenomena.In this case,the stride interval at any instant would depend on the interval at relatively remote times,and this dependence would decay in a scale-free(fractal-like),power-law fashion.Actually,human gait,as one of the physiological signals,normally exhibits a fractal structure and has an obviously nonlinear characteristic.How to accurately detect dynamics of human locomotor control system is of much practical significance.An improved rescaled range analysis method in the paper was used to precisely obtain the fractal dynamics of human gait at different free walking speeds,whereas the new time series resulting from the original gait fluctuation sequence randomly shuffled became total random.Long-range correlations of healthy human gait under free condition at all three walking rates were observed,while correlations degraded or even disappeared during three corresponding metronomically-paced walks.The method can be still effective in probing the fractal dynamics of human gait with data length shortened.These findings indicate that the fractal dynamics of walking rhythm are normally quite robust and intrinsic to human locomotor system.It further implies that supra-spinal control of locomotion is critical in generating and maintaining these long-range correlations.Due to analyzing short-term data sequences rather than long datasets required by most nonlinear methods,the algorithm has the characteristics of simplicity and sensitivity,and most importantly,fast calculation as well as powerful anti-noise capacities.These findings have implications for modeling locomotor control and also for quantifying gait dynamics in varying physiologic states.