Abstract

BACKGROUND: The accurate assessment of step variability remains problematic. OBJECTIVE: To determine the minimum time required for assessing spatiotemporal variability during continuous running. METHODS: Seventeen endurance runners performed a running protocol on a treadmill, with a 3-min recording period at 12 km/h. Spatiotemporal parameters (contact and flight times, step length and step frequency) were measured using the OptoGait system and step variability was considered for each parameter, in terms of within-participants standard deviation (SD) and coefficient of variation (CV%). Step variability was calculated over 6 different durations: 0-10s, 0-20s, 0-30s, 0-60s, 0-120s and 0-180s. RESULTS: The repeated measures ANOVA revealed no significant differences between measurements in mean spatiotemporal gait parameters (p≥0.396, ICC≥0.90 in all parameters). The post-hoc analysis confirmed no significant differences in step variability (of each spatiotemporal parameter) between measurements. The Bland-Altman limits of agreement method showed that longer recording intervals yield smaller systematic bias, random errors, and narrower limits of agreement. CONCLUSIONS: The duration of the recording interval plays an important role in the accuracy of the measurement (i.e. variability in spatiotemporal gait parameters), with longer intervals (180s) showing smaller systematic bias and narrower limits of agreement than shorter intervals (10s, 20s, 30s, 60s or 120s).
Galley Proof 30/11/2018; 14:23 File: ies–1-ies181197.tex; BOKCTP/duyan p. 1
Isokinetics and Exercise Science -1 (2018) 1–5 1
DOI 10.3233/IES-181197
IOS Press
How long is required to undertake step
variability analysis during running? A pilot
study
Felipe García-Pinillosa,, Luis E. Roche-Seruendob, Amador García-Ramosc,d,
Rodrigo Ramírez-Campilloeand Pedro Á. Latorre-RomÁnf
aDepartment of Physical Education, Sports and Recreation, Universidad de La Frontera, Temuco, Chile
bUniversidad San Jorge, Campus Universitario, Villanueva de GÁllego, Zaragoza, Spain
cDepartment of Physical Education and Sport, Faculty of Sport Sciences, University of Granada, Granada, Spain
dFaculty of Education, Catholic University of the Most Holy Conception, Concepción, Chile
eDepartment of Physical Activity Sciences, Research Nucleus in Health, Physical Activity and Sport, Universidad
de Los Lagos, Osorno, Chile
fDepartment of Corporal Expression, University of Jaen, Jaen, Spain
Received 5 October 2018
Accepted 2 November 2018
Abstract.
BACKGROUND: The accurate assessment of step variability remains problematic.
OBJECTIVE: To determine the minimum time required for assessing spatiotemporal variability during continuous running.
METHODS: Seventeen endurance runners performed a running protocol on a treadmill, with a 3-min recording period at
12 km/h. Spatiotemporal parameters (contact and flight times, step length and step frequency) were measured using the OptoGait
system and step variability was considered for each parameter, in terms of within-participants standard deviation (SD) and coef-
ficient of variation (CV%). Step variability was calculated over 6 different durations: 0–10 s, 0–20 s, 0–30 s, 0–60 s, 0–120 s and
0–180 s.
RESULTS: The repeated measures ANOVA revealed no significant differences between measurements in mean spatiotemporal
gait parameters (p>0.396, ICC >0.90 in all parameters). The post-hoc analysis confirmed no significant differences in step
variability (of each spatiotemporal parameter) between measurements. The Bland-Altman limits of agreement method showed
that longer recording intervals yield smaller systematic bias, random errors, and narrower limits of agreement.
CONCLUSIONS: The duration of the recording interval plays an important role in the accuracy of the measurement (i.e. vari-
ability in spatiotemporal gait parameters), with longer intervals (180 s) showing smaller systematic bias and narrower limits of
agreement than shorter intervals (10 s, 20 s, 30 s, 60 s or 120 s).
Keywords: Biomechanics, endurance runners, gait variability, movement variability
Corresponding author: Felipe García-Pinillos, Department of
Physical Education, Sports and Recreation. Universidad de La Fron-
tera, Calle Uruguay, Temuco 1980, Chile. Tel.: +34 660062066;
E-mail: fegarpi@gmail.com.
1. Introduction 1
Step variability seems to be related to both in- 2
juries [1,2] and endurance performance [3]. Never- 3
theless, the accurate assessment of step variability 4
remains problematic. In 1995, Belli et al. [4] indi- 5
cated that step variability during running was diffi- 6
cult to estimate due to the lack of measurement de- 7
ISSN 0959-3020/18/$35.00 c
2018 – IOS Press and the authors. All rights reserved
Galley Proof 30/11/2018; 14:23 File: ies–1-ies181197.tex; BOKCTP/duyan p. 2
2F. García-Pinillos et al. / How long is required to undertake step variability analysis during running?
vices. Today, many devices provide real-time feed-8
back on spatiotemporal parameters while running (e.g.,9
OptoGaitTM, StrydTM or MyotestTM ). Therefore, the10
limitation is not how to collect the data but how long11
should last the data collection to obtain accurate as-12
sessments of step variability.13
Belli et al. [4] suggested that 32–64 consecutive14
steps are required to assess step variability, which rep-15
resents approximately 15–20 s when running at sub-16
maximal velocities. To the best of the authors’ knowl-17
edge, the evidence available about how many steps or18
how long should last the data collection to obtain ac-19
curate assessments of step variability is limited, and20
no more studies have reconsidered this topic adapted21
to the new devices. However, some studies have ad-22
dressed this analysis during walking [5–7]. A previous23
work examined the minimum number of steps required24
to accurately estimate spatial and temporal step kine-25
matic variability of subjects walking on a treadmill [5],26
concluding that at least 400 steps are required. To fur-27
ther examine the step variability during running, the28
aim of this study was to determine the minimum time29
required for assessing spatiotemporal variability dur-30
ing continuous running on an instrumented treadmill.31
The authors hypothesised that the variability in spa-32
tiotemporal gait parameters during running would be33
very similar in short recording intervals (e.g. 10 s, 20 s,34
30 s, 60 s or 120 s) compared to a longer recording35
interval (180 s).36
2. Methods37
2.1. Participants38
Seventeen trained male endurance runners (age: 3439
±7 years; height: 1.74 ±0.04 m; body mass: 71.2 ±40
4.3 kg) participated in this study. Participants met the41
inclusion criteria: (i) older than 18 years old, (ii) able to42
run 10 km in <40 min (36.1 ±1.9 min), (iii) training43
on a treadmill at least once per week, (iv) free from in-44
jury (points 3 and 4 refer to the 6 months preceding the45
study). After receiving information on the objectives46
and procedures of the study, participants signed an in-47
formed consent form, which complied with the ethi-48
cal standards of the World Medical Association’s Dec-49
laration of Helsinki (2013). The study was approved50
by the local ethics committee (San Jorge University,51
Zaragoza, Spain).52
2.2. Procedures53
Participants were tested on a motorized treadmill54
(HP cosmos Pulsar 4P, HP cosmos Sports and Medi-55
cal, Gmbh, Germany). A standardized 10-min warm- 56
up (running at 10 km/h) was performed since previous 57
studies on human locomotion have shown that accom- 58
modation to a new condition occurs in 6–8 min [8]. 59
After warming-up, running velocity was increased 1 60
km/h every min until a speed of 12 km/h was reached. 61
The participants ran at 12 km/h for 3 min with a com- 62
plete recording period. Note that 12 km/h is a normal 63
pace for these athletes and is consistent with previous 64
studies [9]. All participants verbally reported feeling 65
comfortable running at the set speed. 66
2.3. Measures 67
Spatiotemporal parameters were measured using the 68
OptoGaitTM system (Optogait; Microgate, Bolzano, 69
Italy), which was previously validated for the assess- 70
ment of spatiotemporal parameters of the gait of young 71
adults [10]. The OptoGaitTM system is able to mea- 72
sure both contact time (CT) and flight time (FT) at 73
1000 Hz. The two parallel bars of the OptoGaitTM sys- 74
tem were placed on the side edges of the treadmill 75
at the same level of the contact surface. CT, FT, step 76
length (SL) and step frequency (SF) were measured for 77
every step [11]. 78
Step variability was assessed for each spatiotem- 79
poral parameter through the within-participant stan- 80
dard deviation (SD) and the coefficient of variation 81
(CV%). Since previous studies have used indistinctly 82
the SD [12] or CV% [3], we incorporated both mea- 83
sures to make comparisons easier. Step variability was 84
examined over 6 recording intervals within the 3-min 85
recording period: 0–10 s, 0–20 s, 0–30 s, 0–60 s, 0– 86
120 s and 0–180 s. 87
2.4. Statistical analysis 88
Descriptive statistics are represented as mean (SD). 89
Tests of normal distribution and homogeneity (Kolmo- 90
gorov-Smirnov and Levene’s test, respectively) were 91
conducted on all data before analysis. One-way re- 92
peated measures ANOVA with Bonferroni post-hoc 93
corrections were conducted on the magnitude of each 94
spatiotemporal parameter as well as on variability out- 95
comes (i.e., SD and CV%) to examine possible differ- 96
ences between the recording intervals (0–10 s, 0–20 s, 97
0–30 s, 0–60 s, 0–120 s, 0–180 s). Effect sizes were 98
calculated using partial eta squared (Eta2) [13]. The as- 99
sociation of the magnitude and variability of the spa- 100
tiotemporal parameters between the recording inter- 101
vals was quantified through the intraclass correlation 102
Galley Proof 30/11/2018; 14:23 File: ies–1-ies181197.tex; BOKCTP/duyan p. 3
F. García-Pinillos et al. / How long is required to undertake step variability analysis during running? 3
Table 1
Descriptive values and association between the magnitude of the spatiotemporal parameters obtained from six time intervals
Mean (standard deviation) ICC (0–10 s ICC (0–20 s ICC (0–30 s ICC (0–60 s ICC (0–120 s
0–10 s 0–20 s 0–30 s 0–60 s 0–120 s 0–180 s vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s)
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
CT (s) 0.26 (0.02) 0.26 (0.02) 0.26 (0.02) 0.27 (0.02) 0.27 (0.02) 0.27 (0.02) 0.981 (0.951–0.993) 0.987 (0.965–0.995) 0.991 (0.976–0.997) 0.997 (0.992–0.999) 0.998 (0.997–0.999)
FT (s) 0.09 (0.03) 0.09 (0.02) 0.09 (0.02) 0.09 (0.02) 0.09 (0.02) 0.09 (0.02) 0.961 (0.900–0.985) 0.973 (0.931–0.990) 0.980 (0.950–0.952) 0.993 (0.981–0.997) 0.999 (0.998–1.00)
SL (cm) 118 (5) 118 (5) 118 (5) 118 (5) 118 (5) 118 (5) 0.996 (0.913–0.988) 0.980 (0.947–0.992) 0.984 (0.960–0.994) 0.994 (0.984–0.998) 0.999 (0.998–0.999)
SF (step/min) 170 (8) 170 (8) 170 (8) 170 (7) 170 (7) 170 (7) 0.968 (0.917–0.988) 0.979 (0.942–0.996) 0.985 (0.961–0.994) 0.994 (0.984–0.998) 0.999 (0.998–1.00)
CT, contact time; FT, flight time; SL, step length; SF, step frequency; ICC, intraclass correlation coefficient; 95% CI, 95% confidence interval.
Table 2
Descriptive values and association observed for the within-participants standard deviation of the spatiotemporal parameters obtained from six time intervals
Mean (standard deviation) ICC (0–10 s ICC (0–20 s ICC (0–30 s ICC (0–60 s ICC (0–120 s
0–10 s 0–20 s 0–30 s 0–60 s 0–120 s 0–180 s vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s)
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
CT (s) 0.006 (0.002) 0.006 (0.002) 0.007 (0.002) 0.007 (0.002) 0.007 (0.002) 0.007 (0.002) 0.905 (0.845–0.937) 0.907 (0.853–0.962) 0.962 (0.903–0.985) 0.985 (0.962–0.994) 0.996 (0.990–0.998)
FT (s) 0.008 (0.002) 0.008 (0.002) 0.009 (0.002) 0.009 (0.002) 0.009 (0.001) 0.009 (0.002) 0.904 (0.808–0.907) 0.901 (0.766–0.922) 0.909 (0.873–0.950) 0.936 (0.833–0.975) 0.991 (0.976–0.996)
SL (cm) 3.52 (1.18) 3.49 (0.98) 3.66 (0.96) 3.63 (0.95) 3.56 (0.96) 3.56 (0.91) 0.909 (0.850–0.929) 0.940 (0.847–0.977) 0.957 (0.890–0.983) 0.963 (0.907–0.986) 0.988 (0.969–0.995)
SF (step/min) 4.65 (1.54) 4.59 (1.44) 4.88 (1.38) 4.77 (1.29) 4.76 (1.30) 4.72 (1.27) 0.909 (0.861–0.941) 0.933 (0.829–0.974) 0.967 (0.915–0.987) 0.980 (0.948–0.992) 0.995 (0.987–0.998)
CT, contact time; FT, flight time; SL, step length; SF, step frequency; ICC, intraclass correlation coefficient; 95% CI, 95% confidence interval.
Table 3
Descriptive values and association observed for the coefficient of variation (%) of the spatiotemporal parameters obtained from six time intervals
Mean (standard deviation) ICC (0–10 s ICC (0–20 s ICC (0–30 s ICC (0–60 s ICC (0–120 s
0–10 s 0–20 s 0–30 s 0–60 s 0–120 s 0–180 s vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s) vs 0–180 s)
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
CT (%) 2.4 (0.8) 2.3 (0.8) 2.5 (0.8) 2.6 (0.8) 2.6 (0.8) 2.7 (0.8) 0.901 (0.681–0.946) 0.904 (0.697–0.949) 0.964 (0.907–0.986) 0.987 (0.966–0.995) 0.996 (0.990–0.998)
FT (%) 10.3 (5.0) 10.1 (4.2) 10.9 (3.8) 10.2 (3.6) 11.4 (4.1) 11.4 (4.0) 0.905 (0.689–0.944) 0.901 (0.685–0.964) 0.959 (0.894–0.984) 0.982 (0.953–0.993) 0.998 (0.995–0.999)
SL (%) 3.0 (1.1) 3.0 (0.9) 3.1 (0.9) 3.1 (0.9) 3.0 (0.9) 3.0 (0.8) 0.901 (0.657–0.943) 0.948 (0.866–0.980) 0.961 (0.901–0.985) 0.968 (0.919–0.988) 0.990 (0.973–0.996)
SF (%) 2.7 (0.8) 2.7 (0.8) 2.9 (0.8) 2.8 (0.7) 2.8 (0.7) 2.8 (0.7) 0.901 (0.732–0.951) 0.927 (0.815–0.972) 0.964 (0.906–0.986) 0.977 (0.940–0.991) 0.994 (0.986–0.998)
CT, contact time; FT, flight time; SL, step length; SF, step frequency; ICC, intraclass correlation coefficient; 95% CI, 95% confidence interval.
Galley Proof 30/11/2018; 14:23 File: ies–1-ies181197.tex; BOKCTP/duyan p. 4
4F. García-Pinillos et al. / How long is required to undertake step variability analysis during running?
coefficient (ICC). Based on the characteristics of this103
experimental design and following the guidelines re-104
ported by Koo and Li [14], the authors decided to con-105
duct a ‘two-way random-effects’ model (ICC [2, k]),106
‘mean of measurements’ type, and ‘absolute’ defini-107
tion for the ICC measurement. The interpretation of the108
ICC was based on the benchmarks reported by a pre-109
vious study [15]: ICC <0 reflects ‘poor’ reliability, 0–110
0.20 ‘slight’, 0.21–0.40 ‘fair’, 0.41–0.60 ‘moderate’,111
0.61–0.80 ‘substantial’, and >0.81 ‘almost perfect’ re-112
liability. The Bland-Altman [16] limits of agreement113
method (mean difference ±1.96 SD) was used to ex-114
amine differences in step variability (i.e., CV%) be-115
tween the shorter recording intervals (0–10 s, 0–20 s,116
0–30 s, 0–60 s and 0–120 s) and the longest interval117
(0–180 s) for each spatiotemporal parameter (CT, FT,118
SL and SF). Heteroscedasticity of error was defined as119
an r2>0.1 [17]. The level of significance used was120
p < 0.05. Data analysis was performed using SPSS121
(version 21, SPSS Inc., Chicago, IL, USA).122
3. Results123
The ANOVAs showed no significant differences in124
the magnitude of the spatiotemporal parameters be-125
tween the recording intervals (CT: p=0.454, F=126
0.998, partial Eta2=0.263; FT: p=0.396, F=127
1.117, partial Eta2=0.285; SL: p=0.433, F=128
1.039, partial Eta2=0.271; SF: p=0.671, F=129
0.643, partial Eta2=0.187). An almost perfect asso-130
ciation was also observed in the magnitude of the four131
spatiotemporal parameters between the recording in-132
tervals (ICC >0.90) (Table 1).133
The ANOVAs conducted on the within-participants134
SD revealed significant differences for FT (p=0.007,135
F=5.153, partial Eta2=0.648) and SF (p=0.035,136
F=3.320, partial Eta2=0.543), while no signif-137
icant differences were observed for CT (p=0.172,138
F=1.825, partial Eta2=0.395) and SL (p= 0.167,139
F=1.851, partial Eta2=0.398). Regarding the140
ANOVAs conducted on CV% values, significant dif-141
ferences were observed for FT (p=0.025, F=3.680,142
partial Eta2=0.568) and SF (p=0.036, F=3.283,143
partial Eta2=0.540), but not for CT (p=0.155,144
F=1.916, partial Eta2=0.406) and SL (p=0.172,145
F=1.825, partial Eta2=0.395). Bonferroni pairwise146
comparisons never reached statistical significance. An147
almost perfect association between the recording inter-148
vals was observed for both the within-participants SD149
(ICC >0.90; Table 2) and the CV% (ICC >0.90; Ta-150
ble 3).151
Bland-Altman plots revealed heteroscedasticity of 152
error for the FT (0–10 s vs. 0–180 s, r2=0.222; and 153
0–60 s vs. 0–180 s, r2=0.199) and for the SL (0–10 s 154
vs. 0–180 s, r2=0.115), while heteroscedasticity of 155
error was not observed for the rest of the recording in- 156
tervals of those spatiotemporal parameters, and for CT 157
and SF (r2<0.1). Longer recording intervals yield 158
smaller systematic bias, random errors, and narrower 159
limits of agreement for each spatiotemporal parameter. 160
4. Discussion 161
The results demonstrated that the duration of the 162
recording interval plays an important role in the accu- 163
racy of the measurement (i.e. variability in spatiotem- 164
poral gait parameters), with longer intervals (180 s) 165
showing smaller systematic bias and narrower limits of 166
agreement than shorter intervals (10 s, 20 s, 30 s, 60 s 167
or 120 s). However, based on the analysis of the mag- 168
nitude of the differences (i.e. ICC) between shorter and 169
longer recording intervals, the authors suggest that in 170
some contexts where the accuracy requirements are not 171
maximum and time-efficient methods are needed (e.g. 172
clinical setting or big groups of athletes), shorter inter- 173
vals allow sufficient accuracy to assess step variabil- 174
ity. Additionally, to correctly interpret these results it 175
is important to note that might be restricted to trained 176
endurance runners [3] and to a running protocol per- 177
formed on a treadmill at a fixed submaximal veloc- 178
ity [11,18]. 179
Despite the importance of step variability, very few 180
studies have focused on determining how many steps 181
are required to accurately estimate spatial and tem- 182
poral step variability during running. Indeed, the au- 183
thors found just one study from 1995 [4], which ex- 184
amined step variability in terms of step duration and 185
vertical body displacement. Belli et al. [4] concluded 186
that 32–64 consecutive steps (20–25 s) are required 187
to accurately estimate step variability during running 188
at submaximal velocities. Since our findings are ex- 189
pressed in time intervals (s) and in order to make com- 190
parisons easier, it is worth noting that 10 s of record- 191
ing interval equals 27–30 step/min, 20 s equals 54– 192
60 step/min, 30 s equals 81–90 step/min, 60 s equals 193
163–177 step/min, 120 s equals 326–354 step/min and 194
180 s equals 489–531 step/min (i.e. on average, partic- 195
ipants run at 12 km/h with a SF of 170 ±7 step/min). 196
Despite methodological differences, the current re- 197
sults are partially in line with those reported by Belli 198
et al. [4], highlighting that mean spatiotemporal gait 199
Galley Proof 30/11/2018; 14:23 File: ies–1-ies181197.tex; BOKCTP/duyan p. 5
F. García-Pinillos et al. / How long is required to undertake step variability analysis during running? 5
parameters during running, and variability in those pa-200
rameters, might be assessed through the data collected201
over a time period of only 10 s (i.e., 30 steps) or, as202
long as the accuracy requirements are not maximum.203
Nevertheless, the Bland-Altman method conducted in204
the current study showed that longer recording inter-205
vals yield smaller systematic bias, random errors, and206
narrower limits of agreement so, if maximum accuracy207
is required (e.g., scientific approach), longer recording208
periods (e.g. 180 s), with a greater number of steps (e.g.209
500), must be used. These results seem to be con-210
sistent with the results reported by previous studies fo-211
cused on the analysis of step variability during walk-212
ing. Owings and Grabiner [5] indicated that at least213
400 steps are required to accurately estimate step vari-214
ability, König et al. [6] proposed the collection of at215
least 50 step cycles, whereas Bruijn et al. [7] concluded216
that longer data series (>300 steps) led to more pre-217
cise estimates of variables referred to step variability.218
Finally, some limitations need to be considered.219
First, the footwear was not standardized, but all run-220
ners wore their own footwear to increase the ecolog-221
ical validity of the study. Second, the protocol itself,222
with participants running on a treadmill at a fixed sub-223
maximal velocity. Notwithstanding these limitations,224
the authors consider that this study provides an answer225
to the initial question and it leaves some unanswered226
questions about the influence of the running velocity227
or level of exhaustion on step variability and the time228
required to accurately measures it.229
In summary, the data suggest that the duration of the230
recording interval plays an important role in the accu-231
racy of the measurement (i.e. variability in spatiotem-232
poral gait parameters during running at 12 km/h), with233
longer intervals (180 s) showing smaller systematic234
bias and narrower limits of agreement than shorter in-235
tervals (10 s, 20 s, 30 s, 60 s or 120 s).236
Therefore, from a practical standpoint, if maximum237
accuracy is required (e.g., scientific approach) longer238
recording periods must be used, but shorter record-239
ing intervals might be a time-efficient option for clin-240
icians or coaches working with big groups of athletes,241
or where logistical issues difficult long-lasting assess-242
ment protocols (e.g., athletes with pain during run-243
ning).244
Conflict of interest245
The authors declare no conflict of interest.246
References 247
[1] Meardon SA, Hamill J, Derrick TRTR. Running injury and 248
stride time variability over a prolonged run. Gait Posture 2011 249
Jan 1; 33(1): 36-40. 250
[2] Mann R, Malisoux L, Nührenbörger C, Urhausen A, Meijer 251
K, Theisen D. Association of previous injury and speed with 252
running style and stride-to-stride fluctuations. Scand J Med 253
Sci Sports 2015 Dec; 25(6): e638-45. 254
[3] Nakayama Y, Kudo K, Ohtsuki T. Variability and fluctuation 255
in running gait cycle of trained runners and non-runners. Gait 256
Posture 2010; 31(3): 331-5. 257
[4] Belli A, Lacour JR, Komi PV, Candau R, Denis C. Mechani- 258
cal step variability during treadmill running. Eur J Appl Phys- 259
iol Occup Physiol 1995; 70(6): 510-7. 260
[5] Owings TM, Grabiner MD. Measuring step kinematic vari- 261
ability on an instrumented treadmill: How many steps are 262
enough? J Biomech 2003; 36(8): 1215-8. 263
[6] König N, Singh NB, von Beckerath J, Janke L, Taylor WR. 264
Is gait variability reliable? An assessment of spatio-temporal 265
parameters of gait variability during continuous overground 266
walking. Gait Posture 2014 Jan; 39(1): 615-7. 267
[7] Bruijn SM, van Dieën JH, Meijer OG, Beek PJ. Statistical 268
precision and sensitivity of measures of dynamic gait stability. 269
J Neurosci Methods 2009; 178(2): 327-33. 270
[8] Schieb DA. Kinematic accommodation of novice treadmill 271
runners. Res Q Exerc Sport 1986 Mar; 57(1): 1-7. 272
[9] Morin JB, Samozino P, Zameziati K, Belli A. Effects of al- 273
tered stride frequency and contact time on leg-spring behavior 274
in human running. J Biomech 2007 Jan; 40(15): 3341-8. 275
[10] Lee MM, Song CH, Lee KJ, Jung SW, Shin DC, Shin SH. 276
Concurrent validity and test-retest reliability of the OPTO- 277
Gait photoelectric cell system for the assessment of spatio- 278
temporal parameters of the gait of young adults. J Phys Ther 279
Sci 2014; 26(1): 81-5. 280
[11] Roche-Seruendo LE, García-Pinillos F, Haicaguerre J, 281
Bataller-Cervero AV, Soto-Hermoso VM, Latorre-Román PÁ. 282
Lack of influence of muscular performance parameters on 283
spatio-temporal adaptations with increased running velocity. 284
J Strength Cond Res 2018 Feb; 32(2): 409-15. 285
[12] Paquette MR, Milner CE, Melcher DA. Foot contact angle 286
variability during a prolonged run with relation to injury his- 287
tory and habitual foot strike pattern. Scand J Med Sci Sports 288
2017; 27(2): 217-22. 289
[13] Levine TR, Hullett CR. Eta squared, partial eta squared, and 290
misreporting of effect size in communication research. Hum 291
Commun Res 2002; 28(4): 612-25. 292
[14] Koo TK, Li MY. A guideline of selecting and reporting intra- 293
class correlation coefficients for reliability research. J Chiropr 294
Med 2016 Jun; 15(2): 155-63. 295
[15] Landis JR, Koch GG. The measurement of observer agree- 296
ment for categorical data. Biometrics 1977; 33(1): 159-74. 297
[16] Bland JM, Altman DG. Comparing methods of measurement: 298
Why plotting difference against standard method is mislead- 299
ing. Lancet 1995; 346(8982): 1085-7. 300
[17] Atkinson G, Nevill AM. Statistical methods for assessing 301
measurement error (reliability) in variables relevant to sports 302
medicine. Sport Med 1998; 26(4): 217-38. 303
[18] Riley PO, Dicharry J, Franz J, Della Croce U, Wilder RP, 304
Kerrigan DC. A kinematics and kinetic comparison of over- 305
ground and treadmill running. Med Sci Sports Exerc [Inter- 306
net] 2008 Jun [cited 2014 Jul 10]; 40(6): 1093-100. Available 307
from: http://www.ncbi.nlm.nih.gov/pubmed/18460996. 308
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
This study aimed to analyse the influence of muscular performance parameters on spatio-temporal gait characteristics during running when gradually increasing speed. 51 recreationally trained male endurance runners (age: 28 ± 8 years) voluntarily participated in this study. Subjects performed a battery of jumping tests (squat jump, countermovement jump, and 20 cm drop jump), and after that, the subjects performed an incremental running test (10 to 20 km/h) on a motorized treadmill. Spatio-temporal parameters were measured using the OptoGait system. Cluster k-means analysis grouped subjects according to the jumping test performance, by obtaining a group of good jumpers (GJ, n = 19) and a group of bad jumpers (BJ, n = 32). With increased running velocity, contact time was shorter, flight time and step length longer, whereas cadence and stride angle were greater (p < 0.001). No significant differences between groups (p ≥ 0.05) were found at any running speed. The results obtained indicate that increased running velocity produced no differences in spatio-temporal adaptations between those runners with good jumping ability and those with poor jumping ability. Based on that, it seems that muscular performance parameters do not play a key role in spatio-temporal adaptations experienced by recreational endurance runners with increased velocity. However, taken into consideration the well-known relationship between running performance and neuromuscular performance, the authors suggest that muscular performance parameters would be much more determinant in the presence of fatigue (exhausted condition), or in the case of considering other variables such as running economy or kinetic.
Article
Full-text available
Running-related injuries remain problematic among recreational runners. We evaluated the association between having sustained a recent running-related injury and speed, and the strike index (a measure of footstrike pattern, SI) and spatiotemporal parameters of running. Forty-four previously injured and 46 previously uninjured runners underwent treadmill running at 80%, 90%, 100%, 110%, and 120% of their preferred running speed. Participants wore a pressure insole device to measure SI, temporal parameters, and stride length (Slength) and stride frequency (Sfrequency) over 2-min intervals. Coefficient of variation and detrended fluctuation analysis provided information on stride-to-stride variability and correlative patterns. Linear mixed models were used to compare differences between groups and changes with speed. Previously injured runners displayed significantly higher stride-to-stride correlations of SI than controls (P = 0.046). As speed increased, SI, contact time (Tcontact), stride time (Tstride), and duty factor (DF) decreased (P < 0.001), whereas flight time (Tflight), Slength, and Sfrequency increased (P < 0.001). Stride-to-stride variability decreased significantly for SI, Tcontact, Tflight, and DF (P ≤ 0.005), as did correlative patterns for Tcontact, Tstride, DF, Slength, and Sfrequency (P ≤ 0.044). Previous running-related injury was associated with less stride-to-stride randomness of footstrike pattern. Overall, runners became more pronounced rearfoot strikers as running speed increased.
Article
Full-text available
[Purpose] The purpose of this study was to investigate the concurrent validity and test-retest reliability of the recently introduced OPTOGait Photoelectric Cell System for the assessment of spatio-temporal parameters of gait. [Subjects] Twenty healthy young adults (mean age = 27.35, SD = 7.4) were asked to walk 3 times on walkway at a comfortable speed. [Methods] Concurrent validity was assessed by comparing data obtained using the OPTOGait and GAITRite systems, and reliability was assessed by comparing data from the first and third OPTOGait sessions. [Results] Concurrent validity, as identified by intra-class correlation coefficients (ICC (2, 1) = 0.929-0.998), coefficients of variation (CVME = 0.32-11.30%), and 95% limits of agreement, showed high levels of correlation. In addition, the test-retest reliability of the OPTOGait Photoelectric Cell System was demonstrated as showing a high level of correlation with all spatio-temporal parameters by intra-class correlation coefficients (ICC (3, 1) = 0.785-0.952), coefficients of variation (CVME = 1.66-4.06%), 95% limits of agreement, standard error of measurement (SEM = 2.17-5.96%), and minimum detectable change (MDC95% = 6.01-16.52%). [Conclusion] The OPTOGait Photoelectric Cell System has strong concurrent validity along with relative and absolute test-retest reliabilities. This portable system with easy-to-use features can be used for clinical assessments or research purposes as an objective means of assessing gait.
Article
Full-text available
Communication researchers, along with social scientists from a variety of disciplines, are increasingly recognizing the importance of reporting effect sizes to augment significance tests. Serious errors in the reporting of effect sizes, however, have appeared in recently published articles. This article calls for accurate reporting of estimates of effect size. Eta squared (η2) is the most commonly reported estimate of effect sized for the ANOVA. The classical formulation of eta squared (Pearson, 1911; Fisher, 1928) is distinguished from the lesser known partial eta squared (Cohen, 1973), and a mislabeling problem in the statistical software SPSS (1998) is identified. What SPSS reports as eta squared is really partial eta squared. Hence, researchers obtaining estimates of eta squared from SPSS are at risk of reporting incorrect values. Several simulations are reported to demonstrate critical issues. The strengths and limitations of several estimates of effect size used in ANOVA are discussed, as are the implications of the reporting errors. A list of suggestions for researchers is then offered.
Article
Full-text available
The current study examined variability and fluctuation in the running gait cycle, focusing on differences between trained distance runners and non-runners. The two groups of participants performed treadmill running at 80%, 100%, and 120% of their preferred speed for 10 min. Stride-interval time-series were recorded during running using footswitches. The average preferred speed was significantly higher for the trained runners than for the non-runners. The trained runners showed significantly smaller variability of stride interval than did the non-runners, and at the same time the scaling exponent alpha evaluated by detrended fluctuation analysis tended to be smaller for the trained runners. These results suggest that expert runners can reduce variability in the trained movement without loosing dynamical degrees of freedom for spatiotemporal organization of the gait pattern.
Article
Objective: Intraclass correlation coefficient (ICC) is a widely used reliability index in test-retest, intrarater, and interrater reliability analyses. This article introduces the basic concept of ICC in the content of reliability analysis. Discussion for researchers: There are 10 forms of ICCs. Because each form involves distinct assumptions in their calculation and will lead to different interpretations, researchers should explicitly specify the ICC form they used in their calculation. A thorough review of the research design is needed in selecting the appropriate form of ICC to evaluate reliability. The best practice of reporting ICC should include software information, "model," "type," and "definition" selections. Discussion for readers: When coming across an article that includes ICC, readers should first check whether information about the ICC form has been reported and if an appropriate ICC form was used. Based on the 95% confident interval of the ICC estimate, values less than 0.5, between 0.5 and 0.75, between 0.75 and 0.9, and greater than 0.90 are indicative of poor, moderate, good, and excellent reliability, respectively. Conclusion: This article provides a practical guideline for clinical researchers to choose the correct form of ICC and suggests the best practice of reporting ICC parameters in scientific publications. This article also gives readers an appreciation for what to look for when coming across ICC while reading an article.
Article
Foot strike pattern and movement variability have each been associated with running injuries. Foot contact angle (FCA) is a common measure of strike pattern. Thus, variability in FCA could be an important running injury risk factor. The purposes of this study were to compare 1) foot contact angle (FCA) and its variability between runners with and without injury history and, 2) FCA variability between habitual rearfoot strike (RFS) and non-RFS runners during a prolonged run. 23 runners with and 21 without injury history participated. Motion capture was used to collect kinematic data during a 40 minute treadmill run. Average FCA and its variability were compared between injury groups and among four time points. FCA and its variability were not different between runners with and without injury history or among time points during the run. FCA variability was lower in non-RFS compared to RFS runners (p < 0.001). Lower FCA variability in non-RFS runners may have implications for higher injury risks due to repeated localized tissue loading. Prospective analyses on the effects of lower FCA variability on injury risk are needed.
Article
The purpose of this study was to determine the effects of treadmill training on the kinematic accommodation and habituation process of novice treadmill runners. Six experienced male college distance runners, but novice treadmill runners, trained on a treadmill operating at 4.0 m[mdot]s−1, 15 min daily for 10 days. Subjects were filmed three times each day in the frontal and sagittal planes, at Minutes 1, 8, and 14 of the run. Stride length, temporal data, and vertical and lateral horizontal displacements of the center of gravity were determined with a computer digitizer system. Analysis of variance revealed that significant alterations occurred in treadmill running kinematics between Days 1 and 2 of the 10-day treadmill training period. Further, for Days 1 through 3, significant within-day stride changes occurred between Minutes 1 and 8, but not between Minutes 8 and 14. These results suggest that minimal amounts of treadmill training are necessary for a subject to fully accommodate to the treadmill.
Article
The assessment of gait variability has become an important indicator for quantifying motor performance. However, the use of treadmills is known to influence the temporal rhythm of gait, while non-continuous (i.e. stop-start) overground walking alters gait variability, leading to erroneous results. Through establishing the "8-walk", an overground walking protocol that allows the collection of a high number of consecutive gait cycles, the aim of this study was to determine the conditions under which gait variability can be assessed reliably. Twelve healthy subjects performed continuous barefoot walking at their preferred speed in a path shaped as an "8". Kinematic data of the dominant foot was collected while subjects walked along the straight 10m sections of the 8-walk during sessions on two different days. Mean spatio-temporal parameters of gait and gait variability were computed for 10, 20, 30, 40, 50 and 60 consecutive cycles. All mean parameters of gait showed excellent reliability (ICC: 0.88-0.98) with only 10 cycles included in the analysis. However, the reliability of spatial and temporal parameters of gait variability improved with increasing number of cycles (ICC: 0.60-0.90) but levelled-off after 50 consecutive cycles, revealing an inter-day test-retest variability of ∼13%. To reliably assess gait variability and evaluate human motor performance, we propose the collection of at least 50 cycles and the use of an 8-walk protocol, which avoids the limitations of treadmill and non-consecutive walking protocols.
Article
Locomotor variability is inherent to movement and, in healthy systems, contains a predictable structure. In this study, detrended fluctuation analysis (DFA) was used to quantify the structure of variability in locomotion. Using DFA, long-range correlations (α) are calculated in over ground running and the influence of injury and fatigue on α is examined. An accelerometer was mounted to the tibia of 18 runners (9 with a history of injury) to quantify stride time. Participants ran at their preferred 5k pace±5% on an indoor track to fatigue. The complete time series data were divided into three consecutive intervals (beginning, middle, and end). Mean, standard deviation (SD), coefficient of variation (CV) and α of stride times were calculated for each interval. Averages for all variables were calculated per group for statistical analysis. No significant interval, group or interval×group effects were found for mean, SD or CV of stride time. A significant linear trend in α for interval occurred with a reduction in α over the course of the run (p=0.01) indicating that over the run, stride times of runners became more unpredictable. This was likely due to movement errors associated with fatigue necessitating frequent corrections. The injured group exhibited lower α (M=0.79, CI(95)=0.70, 0.88) than the non-injured group (p=0.01) (M=0.96, CI(95)=0.88, 1.05); a reduction hypothesized to be associated with altered complexity. Overall, these findings suggest injury and fatigue influence neuromuscular output during running.