Local dynamic stability of treadmill walking: Intrasession and week-to-week repeatability
Introduction
Falls are a major health issue in older adults (Stevens et al., 2006). Moreover, repetitive falls degrade the quality of life of patients suffering of various neurological disorders (Finlayson et al., 2006, Kerr et al., 2010, Ramnemark et al., 2000). It has been shown that stride-to-stride variability of gait kinematics is related to fall risk (Brach et al., 2007, Verghese et al., 2009). This increased variability may be the result of a decreased ability to optimally control gait. Several analytical methods have been developed to better take into account the nonlinear features of gait variability (Hamacher et al., 2011, Hausdorff, 2007, Stergiou and Decker, 2011). Originally developed to detect deterministic chaos in nonlinear dynamical systems, the maximal Lyapunov exponent has been advocated as a relevant method to assess the degree of sensitivity of gait to small perturbations, or in other words the local dynamic stability (LDS). Computed from various continuously measured kinematic parameters (speed, acceleration, and joint angles), the LDS represents the rate of divergence between neighbor trajectories in a reconstructed state space that describes the dynamics of the system (Dingwell, 2006, Dingwell and Cusumano, 2000, Terrier and Dériaz, 2013).
Some recent experimental and clinical findings support the hypothesis that LDS could be used to predict fall risk (Bruijn et al., 2011, Lockhart and Liu, 2008, McAndrew et al., 2011, Roos and Dingwell, 2010, Toebes et al., 2012). The validity of LDS compared to other indicators has been recently discussed and LDS was found to be one of the best stability indices (Bruijn et al., 2013). However, several issues must be solved before LDS can be routinely used as an early fall-risk predictor in clinical settings. In the first place, information regarding the reliability of LDS measurements is still sparse (Bruijn et al., 2010a, Kang and Dingwell, 2006). Recently, a good intrassession repeatability (ICC between 0.79 and 0.92) has been described by analyzing long duration outdoor walking (200 strides) in young healthy subjects (N=20) (van Schooten et al., 2013). The same study reports lower intersession (between nonconsecutive days) repeatability lying between 0.38 and 0.63. However, more results are needed about the LDS reliability in short indoor walking tests, which are more adapted to patients with diminished walking capabilities (Hilfiker et al., 2013, Terrier et al., 2013).
Other research questions need to be addressed in order to increase the usability of LDS in clinical settings. Two different methods are used to characterize short-term LDS, which is assumed to be the more relevant time scale for assessing gait stability (Bruijn et al., 2011, Roos and Dingwell, 2010): one method computes divergence over the duration of one stride (Manor et al., 2009, McAndrew Young and Dingwell, 2012), and the other one over the duration of one step (Bruijn et al., 2009, Toebes et al., 2012). It is still not known whether one method offers more precision than the other one. In addition, it is known that a minimal length of continuous signal is necessary to assess the maximal Lyapunov exponents (Kang and Dingwell, 2006, Rosenstein et al., 1993). However, the averaging of several estimates of LDS obtained from distinct short-duration signals produces reliable results, at least at group level (Sloot et al., 2011). Thereby two possibilities should be distinguished: (1) Inherent to the calculation method, a too short-duration signal induces very large error on LDS estimates: long-duration walking tests are therefore mandatory; (2) LDS can be precisely assessed from a short-duration signal, but it substantially varies from strides to strides: as a result, the precision of estimate can be increased by averaging the results obtained from several short-duration walking tests.
To address the aforementioned issues, the LDS of 95 healthy adults walking on a treadmill was assessed from trunk acceleration signals. We aimed to answer the following research questions: (1) what is the intrasession (within a 5 min continuous measurement) and intersession (week-to-week) repeatability (i.e. absolute agreement among repetitions) of LDS? (2) How reliable is an estimate of LDS obtained from short-duration walking tests? (3) Does the repeatability of LDS increase with measurement length as expected? (4) Do the two methods that assess short-term LDS exhibit the same repeatability?
Section snippets
Subjects
One hundred healthy subjects (50 males, 50 females) were recruited to participate in the study. They were selected according to their age and sex. Ten males and 10 females for each decade, between 20 and 69 years old, were included. The data of five subjects had to be discarded due to technical issues (age of the discarded participants (yr): 58, 41, 22, 23, 30; 2 males, 3 females). Therefore, the final sample contained 95 participants (48 males, 47 females) whose characteristics were (mean
Results
Regarding the descriptive statistics of LDS (Fig. 2), the results are normally distributed among individuals (Lilliefors test, p>0.05), with some outliers. A substantial difference exists between the estimates obtained from 35 strides and from 70 strides, with a more marked effect in long-term LDS; the relative differences (70 strides−35 strides/35 strides×100) were on average +40% for λ4–10, +6% for λ1, and +8% for λ0.5.
The results of the intrasession repeatability (ICC and SEM) are shown in
Discussion
Referring to the research questions presented in the introduction, the results can be summarized as follows: (1) the intrasession repeatability of gait LDS (ICC, 70-strides estimates) was around 0.50 for long-term LDS and 0.85 for short-term LDS, the intersession repeatability was around 0.6 for both short- and long-term LDS estimated from 210 strides (3×70). (2) Long-term LDS estimated from short-duration measurements (35 strides) exhibited particularly low repeatability (ICC: 0.20), while
Conclusions
To conclude, the following advices should be given for future gait LDS studies. (1) It is recommended to normalize sample length before computing LDS to thwart the trend to higher LDS estimates with longer measurements. (2) Regarding long-term LDS, its low reliability when few strides are analyzed makes the use of a treadmill highly recommended in order to record long duration walking tests. Furthermore, given the limited reliability, its use should be restricted to group-level assessment with
Conflict of interest statement
There are no known conflicts of interest.
Acknowledgments
The authors would like to thank Olivier Dériaz for his valuable support and thoughtful advice. The study was supported by the Swiss accident insurance company SUVA, which is an independent, non-profit company under public law, and by the clinique romande de réadaptation. The IRR (Institute for Research in Rehabilitation) is supported by the State of Valais and the City of Sion.
References (46)
- et al.
The reliability and validity of measures of gait variability in community-dwelling older adults
Arch. Phys. Med. Rehabil.
(2008) - et al.
Statistical precision and sensitivity of measures of dynamic gait stability
J. Neurosci. Methods
(2009) - et al.
Kinematic variability and local dynamic stability of upper body motions when walking at different speeds
J. Biomech.
(2006) - et al.
Risk factors for falling among people aged 45–90 years with multiple sclerosis
Arch. Phys. Med. Rehabil.
(2006) Gait dynamics, fractals and falls: finding meaning in the stride-to-stride fluctuations of human walking
Hum. Mov. Sci.
(2007)- et al.
Number of strides required for reliable measurements of pace, rhythm and variability parameters of gait during normal and dual task walking in older individuals
Gait Posture
(2010) - et al.
Intra-session reliability of local dynamic stability of walking
Gait Posture
(2006) - et al.
Dynamic stability of superior vs. inferior segments during walking in young and older adults
Gait Posture
(2009) - et al.
Differential effects of plantar desensitization on locomotion dynamics
J Electromyogr. Kinesiol
(2009) - et al.
Dynamic stability of human walking in visually and mechanically destabilizing environments
J. Biomech.
(2011)
Voluntarily changing step length or step width affects dynamic stability of human walking
Gait Posture
A new method for evaluating motor control in gait under real-life environmental conditions. Part 1: The instrument
Clin. Biomech.
Penny-wise and pound-foolish: the impact of measurement error on sample size requirements in clinical trials
Biol. Psychiatry
Influence of simulated neuromuscular noise on movement variability and fall risk in a 3D dynamic walking model
J. Biomech.
A practical method for calculating largest Lyapunov exponents from small data sets
Phys. D Nonlinear Phenom.
Human movement variability, nonlinear dynamics, and pathology: is there a connection?
Hum. Mov. Sci.
Local dynamic stability and variability of gait are associated with fall history in elderly subjects
Gait Posture
Assessing gait stability: the influence of state space reconstruction on inter- and intra-day reliability of local dynamic stability during over-ground walking
J. Biomech.
Sensitivity of trunk variability and stability measures to balance impairments induced by galvanic vestibular stimulation during gait
Gait Posture
Gait variability and the risk of incident mobility disability in community-dwelling older adults
J. Gerontol. Ser. A Biol. Sci. Med. Sci.
Some experimental results in the correlation of mental abilities
Br. J. Psychol.
Maximum Lyapunov exponents as predictors of global gait stability: a modelling approach
Med. Eng. Phys.
The effects of arm swing on human gait stability
J. Exp. Biol.
Cited by (39)
Tranquilizer effect on the Lyapunov exponents of lame horses
2020, HeliyonCitation Excerpt :The variability of trial-to-trial largest Lyapunov exponent is still large. In regard to the calculation of the largest Lyapunov exponent, there is no consensus on which time scale should be used (Reynard and Terrier, 2014). Further studies are needed to improve the reliability of this method.
Comparison of selected measures of gait stability derived from center of pressure displacement signal during single and dual-task treadmill walking
2019, Medical Engineering and PhysicsCitation Excerpt :Forty seconds of data was used after discarding data related to approximately the first 4 strides. The analysis was based on at least 30 consecutive strides, which is consistent with previous works [18,35,38,39] during WO trials and when performing two visuomotor cognitive tasks of increasing difficulty [37]. The following gait measures were calculated:
Determinants of gait stability while walking on a treadmill: A machine learning approach
2017, Journal of Biomechanics