Abstract
Background
Treadmills provide a safe and convenient way to study the gait of people with Parkinson’s disease (PD), but outcome measures derived from treadmill gait may differ from overground walking.
Objective
To investigate how the relationships between gait metrics and walking speed vary between overground and treadmill walking in people with PD and healthy controls.
Methods
We compared 29 healthy controls to 27 people with PD in the OFF-medication state. Subjects first walked overground on an instrumented gait walkway, then on an instrumented treadmill at 85, 100 and 115% of their overground walking speed. Average stride length and cadence were computed for each subject in both overground and treadmill walking.
Results
Stride length and cadence both differed between overground and treadmill walking. Regressions of stride length and cadence on gait speed showed a log-log relationship for both overground and treadmill gait in both PD and control groups. The difference between the PD and control groups during overground gait was maintained for treadmill gait, not only when treadmill speed matched overground speed, but also with ±15% variation in treadmill speed from that value.
Significance
These results show that the impact of PD on stride length and cadence and their relationship to gait speed is preserved in treadmill as compared to overground walking. We conclude that a treadmill protocol is suitable for laboratory use in studies of PD gait therapeutics.
Keywords: overground, treadmill, Parkinson’s disease, gait disorders, speed
Introduction
Gait disorders are among the most disabling motor symptoms of Parkinson’s disease (PD) and are often refractory to pharmacological and neuromodulatory interventions. Abnormalities such as reduced gait speed and shortened steps or strides can markedly reduce mobility and are linked with increased fall risk and reduced quality of life in PD [1–3].
Developing improved treatments for gait impairment will require robust gait metrics which are practical to implement. Among the most practical metrics are the triad of gait speed, cadence, and stride (or step) length, any two of which determine the third. Though gait speed is related to the ability for community ambulation in elderly [4,5] and PD [6] populations, using gait speed alone as the primary outcome measure has disadvantages. Gait speed reduction in PD [7] appears to be due to impaired regulation of stride length, i.e. shortened stride length [9,10] while regulation of cadence is preserved [8,9]. To achieve a less abnormal gait speed, a person with PD can modulate their cadence to partially compensate for their decreased stride length, thus, uncoupling gait speed from the underlying Parkinsonian deficit.
Treadmills provide a convenient method to evaluate gait or locomotor training [10,11]. Two practical advantages of treadmills are: allowing uninterrupted collection of more gait cycles within limited space and facilitating use of a safety harness in fall-prone populations. Two important and unanswered questions are: 1) can gait measures obtained from treadmills be used as a proxy for overground gait in the PD population, and 2) are the differences in gait measures between PD and controls comparable between treadmills versus overground? The present study undertook to answer these questions.
Comparison of overground to treadmill gait has an extensive literature, with mixed results. For example, treadmill walking, compared to overground walking, is reported to have increased double support and stance times in healthy young and older adults [12] and reduced stride length and increased cadence in older adults [13]. In contrast, other studies found no significant differences [14–16] or differences in gait variability but not the primary spatiotemporal measures [14,17]. The differences may also depend on the population studied [compare 16,21]. Three studies have compared gait metrics between treadmill and overground walking in PD, with equivocal results. Two small studies reported increased step length and reduced cadence during treadmill gait [19,20]. Another study with 36 subjects showed no significant changes in stride length but a decrease in variability of swing and stride time during treadmill walking [17]. In these studies, PD patients were tested in the ON-medication state, so it is not known if gait measures from treadmills are comparable to overground walking in the OFF-medication state. In addition, no study has examined if the relationship of gait metrics to gait speed is maintained from overground to treadmill. There is an exponential relationship between gait speed and both cadence and stride during overground walking [21], allowing it to be linearized by a log transform, thus simplifying statistical analysis. However, it is unclear if the log-log relationship holds for treadmill walking in PD.
The goal of our study is to investigate the relationships among gait metrics (stride length, cadence, and walking speed) during overground and treadmill walking in people with PD and healthy controls. In this study, we compared PD and control subjects in overground and treadmill gait with treadmill speed set to match the overground speed, aiming to best approximate the overground condition. To assess the sensitivity to accuracy in matching overground speed, we also tested treadmill gait at ±15% of overground speed. We hypothesized that: 1) stride length and cadence would differ systematically between overground and treadmill gait but that 2) differences between PD and controls would be at least as great for treadmill as for overground and that 3) the log-log relationship of stride length and cadence to gait speed would hold for PD in both treadmill and overground gait, and that 4) differences between PD and controls during treadmill walking would not depend critically on matching overground speed.
Materials and Methods
Subjects
Fifty-six individuals (27 PD, 29 controls; demographics in Table 1[22]) were included in the study. All subjects with PD included in the study had a clinical diagnosis of idiopathic PD (Hoehn and Yahr scale of II–IV). Subjects with PD were tested in the morning after 12-hour withdrawal from antiparkinsonian medications (practically defined OFF-medication state). Subjects with deep brain stimulation (DBS; n = 8) were switched to the off state at least 1 hour before testing. For inclusion in the study, all individuals needed to be able to ambulate independently in the OFF-medication state for 50 feet and for 20 minutes (with rest periods every 2–3 minutes). Exclusion criteria for all subjects included medical conditions other than PD which impaired walking independently and a diagnosis of dementia or a score < 25 on the Mini-Mental State Exam [23]. All subjects gave written informed consent. The protocol was approved by the University of Minnesota Institutional Review Board.
Table 1.
Summary of participant demographics
| Group | |||
|---|---|---|---|
| PD | Control | Control (Age > 40) | |
| Sex (Male/Female) | 16/11 | 15/14 | 8/12 |
| Age (years)* | 64.3 ± 9.6 | 51.6 ± 19.4 | 63.0 ± 10.0 |
| Disease duration (years)* | 8.7± 5.7 | NA | NA |
| MDS-UPDRS motor score*#$ | 32.2 ± 17.0 (range: 7 – 68) | NA | NA |
| MDS-UPDRS motor factor 1 subscore*#^ | 8.2 ± 4.3 | NA | NA |
| Overground walking speed (m/s)* | 1.07 ± 0.23 | 1.25 ± 0.16 | 1.27 ± 0.16 |
| Leg length (m)* | 0.90 ± 0.05 | 0.92 ± 0.04 | 0.92 ± 0.04 |
| DBS (Yes/No) | 8/19 | NA | NA |
| Side (Left/Right/Bilateral) | 1/2/5 | NA | NA |
| Target (STN/GPi) | 7/1 | NA | NA |
Values are presented as mean ± SD.
MDS-UPDRS stands for Movement Disorder Society-Sponsored Unified Parkinson’s Disease Rating Scale Revision.
Two participants were excluded due to missing scored items.
One participant was excluded due to missing factor 1 scored items.
NA, not applicable; DBS, deep brain stimulation; STN, subthalamic nucleus; GPi, globus pallidus internus.
Protocol
Subjects walked continuously, at a self-selected speed, around an oval course, which included a straight section with a pressure sensitive walkway (GAITRite [24], CIR systems, Franklin, NJ) until at least 30 valid steps were collected [25]. During the overground protocol, they wore the safety harness used later for treadmill walking. After a rest break, subjects first did 2-min trials on a treadmill (Model: C-Mill [26], Motekforce Link, Culemburg, The Netherlands) with speed matched to the overground walking speed (measured with the gait mat). Four practice trials were completed, and results from the 5th trial are reported here. Subjects then completed two trials at fast (115% overground walking speed) and slow (85% overground speed) speeds in random order. Seated rest (average: 1.5 min) was provided before each trial. Following this, subjects completed a virtual obstacle avoidance task which we have previously reported [27].
Data analysis
For overground walking, gait speed, stride length, and cadence were computed and exported directly from the GAITRite software. For treadmill walking, the time and locations of the heel-strike and toe-off events were computed from the center of pressure timeseries (sampled at 500 Hz) by the treadmill software and exported to custom Python and MATLAB scripts for computation of gait metrics (stride length and cadence). Mean value was computed for each trial in each subject.
Group difference in demographics were examined with independent t-test for continuous variables and chi-square for categorical variables. To investigate how gait metrics varied between overground and treadmill walking in PD and controls (our first and second hypotheses), linear mixed effects regression models were used with gait metrics as dependent variables; subject was the random-intercept term; modality (overground vs treadmill), group (PD vs control) and group X modality (X indicates interaction term) were fixed factors.
Most previous research on gait has compared overground vs treadmill without accounting for the relationship between gait speed and gait metrics [21]. Therefore, we did two regressions, one without gait speed and the second, with gait speed included (third hypothesis). For the second regression, we first log-transformed (base 10) gait speed, stride length and cadence and verified graphically that the resulting relationships were linear (see Fig 3). Then, we performed the regression with (log-) gait metrics as dependent variables and (log-) speed, group, and modality as fixed effects including interaction terms.
Figure 3.
Log-log plot of the relationship between gait metrics and 3 different gait speeds (85, 100 and 115% overground walking speeds) during treadmill walking. Three points from each subject, connected by solid lines. Orange is for control and blue is for PD. Regression lines for control group are represented in solid line, whereas dashed line is for PD.
To assess the dependence of these results on matching treadmill to overground speed (fourth hypothesis), we repeated this regression with the 85% and 115% treadmill trials. To compare the within-subject relationship of gait metrics to gait speed to the between-subjects relationship, we first computed the regression slope of stride length or cadence vs. gait speed, over the 85%, 100%, and 115% treadmill trials for each subject. We next computed the difference of this slope from the between-subjects slope. Finally, we examined whether this difference differed significantly from zero using a one sample t-test.
For all mixed effects analyses, age, sex and leg length were included as covariates. Significance was tested with F- tests (Satterthwaite’s degrees of freedom method). Two-tailed p-value threshold was set at 0.05 with Bonferroni correction for multiple tests. All statistical analyses were performed in R [28].
Results
Demographics
Mean age was 12.9 years lower in controls than PD subjects (Table 1), due to the inclusion of a group of 9 young controls (t (54) = −3.12, p < 0.01). When subjects under 40 were excluded, the age difference was 2.5 years and was not significant (t (44) = −0.91, p = 0.73). There were no significant group differences in leg length (t(54) = 1.75, p = 0.17; t(44) = 1.57, p = 0.25 (age > 40)) and sex (χ2 = 0.31, p = 1.00 ; χ2 = 1.57, p = 0.50 (age > 40)). Additionally, PD gait speed was about 0.2 (m/s) less than controls for overground gait speed (t(54) = 3.35, p = 0.001, see Supplementary Material and Fig. S1 for more detailed analysis).
Overground vs. treadmill, PD vs. control: differences in gait metrics (mixed model without gait speed)
Stride length decreased by an average of 0.07 (± 0.09) m while cadence increased by 0.1 (± 0.14) steps/s in treadmill compared to overground walking (F(1,54) = 34.62, p < 0.001; F(1,54) = 31.97, p < 0.001, Figs 1, S1). There was also a group effect with shorter stride length (0.20 m, F(1,51) = 12.36, p < 0.01) in PD compared to controls, but no significant group difference was found in cadence (F(1,51) = 2.01, p = 0.33). The group X modality interaction was not significant for stride length (F(1,54) = 2.27, p = 0.28) and cadence (F(1,54) = 4.50, p = 0.08). For significance tests of covariates and more details of the mixed models, see Supplementary Material.
Figure 1.
The change in average stride length and cadence between overground and treadmill walking for individual subject. Open circle represents individual average value while solid lines represents an increasing trend and dashed lines represents a decreasing trend from overground to treadmill respectively. Orange is for control and blue is for PD.
Overground vs. treadmill, PD vs. control: differences in gait metrics (mixed model with gait speed)
Stride length was significantly related to gait speed (F(1,49) = 103.37, p < 0.001), and modality (F(1,52) = 15.21, p = 0.001) (Fig. 2), borderline related to group (F(1,49) = 5.37, p =0.05) with all interaction terms nonsignificant (p value range: 0.13 to 0.34) including, importantly, the group X modality interaction (p = 0.33). Similarly, cadence was significantly related to gait speed (F(1,49) = 10.68, p < 0.01), group (F(1,49) = 12.19, p <0.01) and modality (F(1,52) = 13.34, p < 0.01) with one significant interaction term, speed X group (F(1,49) = 10.62, p < 0.01). For significance tests of covariates, see Supplementary Material.
Figure 2.
Log-log plot of the relationship between gait metrics and gait speed during overground (left column) and treadmill (right column) walking. Stride length (top row) and cadence (bottom row). Open circle represents individual average value (orange = controls, blue = PD). Regression lines for all subjects in each group are represented as solid lines of the corresponding color.
In summary, the regression analyses with and without the gait speed variable were similar, i.e. stride length was shorter, and cadence was greater, for treadmill than for overground gait, with a significant group effect showing faster cadence in PD compared to controls. Since the group effect remained significant in the model which included gait speed, PD was associated with faster cadence at a given gait speed in comparison to controls. For both stride length and cadence, the group X modality interaction was nonsignificant, supporting that treadmill and overground gait metrics distinguished equally well between PD and controls. For cadence, the speed by group interaction was significant, i.e. the PD vs. control group disparity was significantly greater with slower gait speed. Since slower gait speed was associated with more severe PD (see Supplementary Figs S1, S2), this suggests control vs PD disparity was greater in more severely affected PD subjects, which is foreseeable.
Effect of treadmill gait speed
Although treadmill and overground gait metrics distinguished equally well between Parkinson’s and control subjects, this finding used a treadmill speed set to match the self-selected overground speed. To determine whether this result was sensitive to variations in treadmill speed, we compared three levels of treadmill speed (“pace”) corresponding to 85%, 100%, and 115% of overground speed. The group X pace interaction effect was not significant for both stride length (F(2, 108) = 2.22, p = 0.23) and cadence (F(2,108) = 0.18, p = 1.00), suggesting that the results were robust to deviations of treadmill speed from overground speed.
As shown in Fig. 3, the within-subjects relationship of gait metrics to gait speed was the same as the between-subjects relationship in most subjects. The t-test confirmed that the difference between each within-subject slope and the between-subject slope (for that subject’s group) was not significantly different from zero for both controls (t (28) = −0.63, p = 1.00) and PD (t (26) = −1.06, p = 1.00) in stride length. Similarly, no significant differences from zero were seen for cadence (control: t (28) = 0.67, p = 1.00; PD: t (26) = 1.15, p = 1.00). The results were unchanged with analyses that included only similar-age control subjects (see Supplementary Material).
Discussion
This study compared overground and treadmill walking in people with PD and healthy controls, using gait speed, stride length, and cadence as primary metrics. We found consistent but small decreases in stride length and increases in cadence on the treadmill compared to overground. Most importantly, we showed that gait assessed with the treadmill distinguished the PD cohort from control subjects equally well compared to overground gait, suggesting that treadmill walking is a suitable proxy for overground walking for measures of stride length and cadence (see Supplementary Fig S3 for receiver-operating characteristic analysis).
These results differ from Bello et al. [19]. In our data, the group X modality (treadmill vs. overground) interaction did not reach significance for either stride length or cadence, whereas Bello et al. reported significant interaction effects for both step length and cadence. This discrepancy might be explained by differences in protocol and analysis. For example, gait speed was not included in their model. Their study was conducted in the ON-medication state, the sample size was smaller (n = 8) and their subjects were instructed to hold the handrails during treadmill walking, which can affect the spatiotemporal features of gait [29,30]. Importantly, we conducted our tests in the OFF-medication state, to avoid confounding effects of treatment such as dyskinesias, or ON/OFF pharmacokinetic fluctuations. In addition, we analyzed treadmill data across three paces (85%, 100%, 115% of overground speed) and found no significant group X pace interaction, i.e. our finding does not depend on a perfect match between treadmill speed and overground speed.
We also replicated the log-log relationship of stride length to speed in controls [31], and showed this relationship was preserved in PD (Fig. 3). On the other hand, the relationship between cadence and gait speed was different (significant group X speed interaction) between the PD and control groups during both overground and treadmill walking. We suggest that a compensation for shortened stride by increased cadence which is greater in individuals with more severe Parkinsonism, distorts the normal speed-cadence relationship. Interestingly, we also found the within-subject slope of the speed-stride length, and speed-cadence relationships to be similar to the between-subjects slopes for both PD and control, and comparable to previous reports of this relationship in PD in the ON-medication state [8,20].
Strengths and Limitations
In this study, the PD subjects were tested in the OFF-medication state, using the standard “practically defined off” procedure. Our sample also included individuals with DBS tested in the OFF-stimulation condition. This procedure was intended to avoid ON/OFF fluctuations and treatment induced dyskinesia since the goal was to assess how PD affects gait in a stable state without confounding medication-induced variation or involuntary movements. Furthermore, the PD subjects tested in the study were with moderate disease severity, and that our findings may not generalize outside this sample population.
Gait metrics may depend on the treadmill features. For example, our treadmill walking surface was 3 meters long; consequently, subjects may have felt less constrained to match the treadmill speed exactly (without fluctuations) compared to a typical treadmill (~ 1.6 m), where tighter control of gait speed is required to avoid reaching the end of the treadmill. Our treadmill had servocontrolled belt speed, possibly resulting in less belt speed fluctuation than devices without this feature. Taken together, these characteristics may have made walking on our treadmill a better match for overground walking in comparison to other types of treadmill. Finally, our treadmill gait metrics were derived from the center of pressure trajectory of the embedded force plate. This could have introduced differences between treadmill gait metrics and overground gait metrics computed from the pressure sensor mat. However, the differences between gait metrics for treadmill and overground agree with other reports in the literature [12,13,32], using different measurement technology. Lastly, treadmill gait requires practice, so that unfamiliarity with treadmill walking might have affected the difference between PD and control subjects. However, we allowed 4 practice trials, before the first trial used in the analysis. A small practice effect (amounting to 0.03 m increase in stride length) was observed over the 5 trials, but with nonsignificant group X trial number interaction.
This study was not a comprehensive evaluation of PD gait. Our goal was to validate, against a “gold standard” overground gait protocol [25], a limited set of basic, widely used gait metrics for treadmill use, to quantify parkinsonism. We note that stride (or step) length is one of the most commonly reported gait metrics in studies of DBS [33] and is strongly associated with fall risk in the PD population [3]. Gait speed may surpass it slightly in both respects, but requires either overground gait, or a very specialized treadmill to measure. Nonetheless, future studies are warranted to examine gait speed with other spatiotemporal gait metrics, including additional sagittal plane metrics (e.g. double support time), frontal plane metrics (e.g. stride width) and metrics of variability (e.g. coefficients of variation).
Conclusion
Stride length and cadence differed between overground and treadmill gait under self-selected speed, in both controls or PD patients, but treadmill gait was just as effective as overground gait when comparing performance between people with PD and controls when the treadmill speed was matched or within a tolerance of ±15% of the individual’s selected overground speed.
Supplementary Material
Highlights.
Stride length decreased and cadence increased in treadmill compared to overground.
PD subjects’ overground gait was slower than controls’.
The overground relationship of gait metrics to speed was preserved with treadmill.
Treadmill distinguished PD from control gait equally well compared to overground.
Acknowledgement:
This work was supported by the National Institutes of Health [grant numbers P50 NS123109, RO1 NS088679]; and the University of Minnesota Neuromodulation Innovations (MnDrive)
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest statement
The authors have no conflict of interest to report.
Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- [1].Wood BH, Bilclough JA, Bowron A, Walker RW, Incidence and prediction of falls in Parkinson’s disease: a prospective multidisciplinary study., J. Neurol. Neurosurg. Psychiatry 72 (2002) 721–725. 10.1371/journal.pone.0005521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Nemanich ST, Duncan RP, Dibble LE, Cavanaugh JT, Ellis TD, Ford MP, Foreman KB, Earhart GM, Predictors of gait speeds and the relationship of gait speeds to falls in men and women with parkinson disease, Parkinsons. Dis 10 (2013) 283. 10.1155/2013/141720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Creaby MW, Cole MH, Gait characteristics and falls in Parkinson’s disease: A systematic review and meta-analysis, Park. Relat. Disord 57 (2018) 1–8. 10.1016/j.parkreldis.2018.07.008. [DOI] [PubMed] [Google Scholar]
- [4].Cohen JJ, Sveen JD, Walker JM, Brummel-Smith K, Establishing criteria for community ambulation, Top. Geriatr. Rehabil 3 (1987) 71–78. 10.1097/00013614-198710000-00012. [DOI] [Google Scholar]
- [5].Andrews AW, Chinworth SA, Bourassa M, Garvin M, Benton D, Tanner S, Update on distance and velocity requirements for community ambulation, J. Geriatr. Phys. Ther 33 (2010) 128–134. 10.1097/JPT.0b013e3181eda321. [DOI] [PubMed] [Google Scholar]
- [6].Elbers RG, Van Wegen EEH, Verhoef J, Kwakkel G, Is gait speed a valid measure to predict community ambulation in patients with Parkinson’s disease?, J. Rehabil. Med 45 (2013) 370–375. 10.2340/16501977-1123. [DOI] [PubMed] [Google Scholar]
- [7].Knutsson E, An analysis of parkinsonian gait, Brain. 95 (1972) 475–486. 10.1093/brain/95.3.475. [DOI] [PubMed] [Google Scholar]
- [8].Morris ME, Iansek R, Matyas TA, Summers JJ, The pathogenesis of gait hypokinesia in parkinson’s disease, Brain. 117 (1994) 1169–1181. 10.1093/brain/117.5.1169. [DOI] [PubMed] [Google Scholar]
- [9].Morris ME, Iansek R, Matyas TA, Summers JJ, Stride length regulation in Parkinson’s disease: Normalization strategies and underlying mechanisms, Brain. 119 (1996) 551–568. 10.1093/brain/119.2.551. [DOI] [PubMed] [Google Scholar]
- [10].Herman T, Giladi N, Hausdorff JM, Treadmill training for the treatment of gait disturbances in people with Parkinson’s disease: A mini-review, J. Neural Transm 116 (2009) 307–318. 10.1007/s00702-008-0139-z. [DOI] [PubMed] [Google Scholar]
- [11].Earhart GM, Williams AJ, Treadmill Training for Individuals With Parkinson Disease, Phys. Ther 92 (2012) 893–897. 10.2522/ptj.20110471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Malatesta D, Canepa M, Menendez Fernandez A, The effect of treadmill and overground walking on preferred walking speed and gait kinematics in healthy, physically active older adults, Eur. J. Appl. Physiol 117 (2017) 1833–1843. 10.1007/s00421-017-3672-3. [DOI] [PubMed] [Google Scholar]
- [13].Watt JR, Franz JR, Jackson K, Dicharry J, Riley PO, Kerrigan DC, A three-dimensional kinematic and kinetic comparison of overground and treadmill walking in healthy elderly subjects, Clin. Biomech 25 (2010) 444–449. 10.1016/j.clinbiomech.2009.09.002. [DOI] [PubMed] [Google Scholar]
- [14].Hollman JH, Watkins MK, Imhoff AC, Braun CE, Akervik KA, Ness DK, A comparison of variability in spatiotemporal gait parameters between treadmill and overground walking conditions, Gait Posture. 43 (2016) 204–209. 10.1016/j.gaitpost.2015.09.024. [DOI] [PubMed] [Google Scholar]
- [15].Dingwell JB, Cusumano JP, Cavanagh PR, Sternad D, Local dynamic stability versus kinematic variability of continuous overground and treadmill walking, J. Biomech. Eng 123 (2001) 27–32. 10.1115/1.1336798. [DOI] [PubMed] [Google Scholar]
- [16].Parvataneni K, Ploeg L, Olney SJ, Brouwer B, Kinematic, kinetic and metabolic parameters of treadmill versus overground walking in healthy older adults, Clin. Biomech 24 (2008) 95–100. 10.1016/j.clinbiomech.2008.07.002. [DOI] [PubMed] [Google Scholar]
- [17].Frenkel-Toledo S, Giladi N, Peretz C, Herman T, Gruendlinger L, Hausdorff JM, Treadmill walking as an external pacemaker to improve gait rhythm and stability in Parkinson’s disease, Mov. Disord 20 (2005) 1109–1114. 10.1002/mds.20507. [DOI] [PubMed] [Google Scholar]
- [18].Riley PO, Paolini G, Della Croce U, Paylo KW, Kerrigan DC, A kinematic and kinetic comparison of overground and treadmill walking in healthy subjects, Gait Posture. 26 (2007) 17–24. 10.1016/j.gaitpost.2006.07.003. [DOI] [PubMed] [Google Scholar]
- [19].Bello O, Sanchez JA, Vazquez-Santos C, Fernandez-Del-Olmo M, Spatiotemporal parameters of gait during treadmill and overground walking in Parkinson’s disease, J. Parkinsons. Dis 4 (2014) 33–36. 10.3233/JPD-130251. [DOI] [PubMed] [Google Scholar]
- [20].Ambrus M, Sanchez JA, Fernandez-del-Olmo M, Walking on a treadmill improves the stride length-cadence relationship in individuals with Parkinson’s disease, Gait Posture 68 (2019) 136–140. 10.1016/j.gaitpost.2018.11.025. [DOI] [PubMed] [Google Scholar]
- [21].Grieve DW, Gait patterns and the speed of walking, Biomed Eng. 3 (1968) 119–122. http://ci.nii.ac.jp/naid/10004740984/en/ (accessed November 20, 2019). [Google Scholar]
- [22].Goetz CG, Tilley BC, Shaftman SR, Stebbins GT, Fahn S, Martinez-Martin P, Poewe W, Sampaio C, Stern MB, Dodel R, Dubois B, Holloway R, Jankovic J, Kulisevsky J, Lang AE, Lees A, Leurgans S, LeWitt PA, Nyenhuis D, Olanow CW, Rascol O, Schrag A, Teresi JA, van Hilten JJ, LaPelle N, for the M.D.S.U.R.T. Force, Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results, Mov. Disord 23 (2008) 2129–2170. 10.1002/mds.22340. [DOI] [PubMed] [Google Scholar]
- [23].Folstein MF, Folstein SE, McHugh PR, “Mini-mental state”: A practical method for grading the cognitive state of patients for the clinician., J. Psychiatr. Res 12 (1975) 189–198. 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
- [24].McDonough AL, Batavia M, Chen FC, Kwon S, Ziai J, The validity and reliability of the GAITRite system’s measurements: A preliminary evaluation, Arch. Phys. Med. Rehabil 82 (2001) 419–425. 10.1053/apmr.2001.19778. [DOI] [PubMed] [Google Scholar]
- [25].Galna B, Lord S, Rochester L, Is gait variability reliable in older adults and Parkinson’s disease? Towards an optimal testing protocol, Gait Posture. 37 (2013) 580–585. 10.1016/j.gaitpost.2012.09.025. [DOI] [PubMed] [Google Scholar]
- [26].van der Veen SM, Hammerbeck U, Baker RJ, Hollands KL, Validation of gait event detection by centre of pressure during target stepping in healthy and paretic gait, J. Biomech 79 (2018) 218–222. 10.1016/j.jbiomech.2018.07.039. [DOI] [PubMed] [Google Scholar]
- [27].Lu C, Twedell E, Elbasher R, McCabe M, MacKinnon CD, Cooper SE, Avoiding virtual obstacles during treadmill gait in Parkinson’s disease, Front. Aging Neurosci 11 (2019) 76. 10.3389/fnagi.2019.00076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].R Development Core Team, R: A language and environment for statistical computing, R Found. Stat. Comput (2016). 10.1017/CBO9781107415324.004. [DOI] [Google Scholar]
- [29].Ijmker T, Lamoth CJ, Houdijk H, Tolsma M, Van Der Woude LHV, Daffertshofer A, Beek PJ, Effects of handrail hold and light touch on energetics, step parameters, and neuromuscular activity during walking after stroke, J. Neuroeng. Rehabil 12 (2015). 10.1186/s12984-015-0051-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Buurke TJW, Lamoth CJC, van der Woude LHV, den Otter R, Handrail Holding During Treadmill Walking Reduces Locomotor Learning in Able-Bodied Persons, IEEE Trans. Neural Syst. Rehabil. Eng 27 (2019) 1753–1759. 10.1109/TNSRE.2019.2935242. [DOI] [PubMed] [Google Scholar]
- [31].Kuo AD, A Simple Model of Bipedal Walking Predicts the Preferred Speed–Step Length Relationship, J. Biomech. Eng 123 (2001) 264. 10.1115/1.1372322. [DOI] [PubMed] [Google Scholar]
- [32].Bayat R, Barbeau H, Lamontagne A, Speed and temporal-distance adaptations during treadmill and overground walking following stroke, Neurorehabil. Neural Repair 19 (2005) 115–124. 10.1177/1545968305275286. [DOI] [PubMed] [Google Scholar]
- [33].Schlenstedt C, Shalash A, Muthuraman M, Falk D, Witt K, Deuschl G, Effect of high-frequency subthalamic neurostimulation on gait and freezing of gait in Parkinson’s disease: a systematic review and meta-analysis, Eur. J. Neurol 24 (2017) 18–26. 10.1111/ene.13167. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.