Quantification of the global dynamics of gait through the application of chaos theory and nonlinear dynamics




Protsman, Lisa

Journal Title

Journal ISSN

Volume Title



The purpose of this study was to test the reliability and validity of two measures derived from the nonlinear analysis of a time series. The data were derived from interstride intervals of a 10 minute continuous walk. This research approach offers methods that may enhance our understanding of the complex interactive dynamics of functional gait. The reliability of the measures was assessed through an ICC analysis. The validity was assessed by comparison to published research and to surrogate data. Two measures of complexity were used. The Hurst exponent was calculated as a measure of the long range correlations of the time series. The Hurst exponent was derived using two related methodologies, relative dispersion (RD), and detrended fluctuation analysis (DFA). The Lyapunov exponent was calculated as a measure of the divergence and unpredictability of the time series. The reliability of these measures as “fingerprints” of an individual's gait patterns was investigated through a repeated measures design. Clinical application of the a neurological impairment and to discriminate the effect of an intervention. A comparison was made between dynamic and traditional orthotics. The results suggest that the measures were valid. The results also suggested that the measures were extremely sensitive which lead to inconsistency on repeated trials. ICC analysis of the measures indicated that they were not reliable. The results also indicate that both the Hurst and the Lyapunov exponents were able to identify the presence of a neurological impairment. The measures were not able to discriminate the effect of a dynamic orthotic, as compared to a traditional orthotic, on the gait of person with neurological impairments.



Rehabilitation, Therapy, Chaos theory, Nonlinear systems, Time series