Impact of series length on statistical precision and sensitivity of autocorrelation assessment in human locomotion

Warlop, Thibault; Bollens, Benjamin; Detrembleur, Christine; Stoquart, Gaëtan; Lejeune, Thierry; Crevecoeur, Frédéric

doi:10.1016/j.humov.2017.07.003

Cited by 18 publications

(28 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The optimization of parameters is very important as clearly shown in Figure 1, where huge biases were found when using the worst-case parameters. For DFA and BC, these errors could be caused by an insufficient number of box sizes, an unbalanced density of points along the x-coordinate, or deviations from linearity occurred for smaller and larger box sizes (Peng et al, 1995;Hu et al, 2001;Damouras et al, 2010;Warlop et al, 2017). However, even when the optimal parameters were implemented, some biases were seen from the plots of theoretical H and mean estimatedĤ for DFA and BC (Figure 1).…”

Section: Robust Implementationmentioning

confidence: 99%

“…Finally, it is worth noting that additional techniques may be able to improve the accuracy and reduce the variability of the original methods for short stride interval time series, such as using overlapping windows or evenly spaced versions of DFA (Almurad and Delignières, 2016;Warlop et al, 2017).…”

Section: Limitations and Future Studiesmentioning

confidence: 99%

“…Some previous studies have investigated these practical aspects, but have typically studied them in isolation (e.g., only impact of series length, or only one of the DFA parameters such as box size range). In addition, Warlop et al (2017) found that using more robust methods (such as evenly-spaced box size increments) could provide reliable results for shorter time series (256 instead of 512 data points). To the best of our knowledge, however, a systematic and comprehensive investigation of the length of stride interval time series and all the DFA parameters has never been completed.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Fractal Analysis of Human Gait Variability via Stride Interval Time Series

2020

View full text Add to dashboard Cite

Fractal analysis of stride interval time series is a useful tool in human gait research which could be used as a marker for gait adaptability, gait disorder, and fall risk among patients with movement disorders. This study is designed to systematically and comprehensively investigate two practical aspects of fractal analysis which significantly affect the outcome: the series length and the parameters used in the algorithm. The Hurst exponent, scaling exponent, and/or fractal dimension are computed from both simulated and experimental data using three fractal methods, namely detrended fluctuation analysis, box-counting dimension, and Higuchi's fractal dimension. The advantages and drawbacks of each method are discussed, in terms of biases and variability. The results demonstrate that a careful selection of fractal analysis methods and their parameters is required, which is dependent on the aim of study (either analyzing differences between experimental groups or estimating an accurate determination of fractal features). A set of guidelines for the selection of the fractal methods and the length of stride interval time series is provided, along with the optimal parameters for a robust implementation for each method.

show abstract

Section: Robust Implementationmentioning

confidence: 99%

Section: Limitations and Future Studiesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fractal Analysis of Human Gait Variability via Stride Interval Time Series

2020

View full text Add to dashboard Cite

show abstract

“…Third, Hurst exponents were only computed by using the Detrended Fluctuation Analysis for the sake of homogeneity. However, Rescaled Range Analysis has also been used previously [34] and may have been included also. Finally, other nonlinear analysis tools such as Lyapounov exponents [35] could have been computed in order to consider their added value.…”

Section: Discussionmentioning

confidence: 99%

“…The time series gathered are of variable length, in particular because of the small walking speed of patients suffering from neurodegenerative diseases. As shown in [20], the computed value of variability indices may significantly differ from their exact, asymptotic, value when time series are shorter than 256 points. We therefore choose to truncate time series longer than 256 SI to their last 256 points and to keep unchanged the shorter time series in order to reduce bias associated to time series lengths.…”

Section: Methodsmentioning

confidence: 99%

Added Value and Clinical Significance of Nonlinear Variability Indices of Walking Stride Interval in Neurodegenerative Diseases

Dierick¹,

Vandevoorde²,

Chantraine³

et al. 2020

Preprint

View full text Add to dashboard Cite

Though self-paced walking is highly stereotyped, the stride interval fluctuates from one stride to the next around an average value with a measurable statistical variability. In clinical gait analysis, this variability is usually assessed with indices such the standard deviation or the coefficient of variation (CV). The aim of this study is to understand the added value that nonlinear indices of walking stride interval variability, such as Hurst exponent (H) and Minkowski fractal dimension (D), can provide in a clinical context and to suggest a clinical significance of these indices in the most common neurodegenerative diseases: Parkinson, Huntington, and amyotrophic lateral sclerosis. Although evidence have been accumulated that the stride interval organization at long range displays a more random, less autocorrelated, gait pattern in neurodegenerative diseases compared with young healthy individuals, it is today necessary to recompute CV, H, and D indices in a unified way and to take into account aging impact on these indices. In fact, computed nonlinear indices of variability are mainly dependent on stride interval time series length and algorithms used, making quantitative comparisons between different studies difficult or even impossible. Here, we recompute these indices from available stride interval time series, either coming from our lab or from online databases, or made available to us by the authors of previously conducted research. We confirm that both linear and nonlinear variability indices are relevant indicators of aging process and neurodegenerative diseases.

show abstract