вход по аккаунту



код для вставкиСкачать
Footwear Science
ISSN: 1942-4280 (Print) 1942-4299 (Online) Journal homepage:
Influence of footwear comfort on the variability of
running kinematics
Christian Meyer, Maurice Mohr, Mathieu Falbriard, Sandro R. Nigg & Benno
M. Nigg
To cite this article: Christian Meyer, Maurice Mohr, Mathieu Falbriard, Sandro R. Nigg & Benno
M. Nigg (2017): Influence of footwear comfort on the variability of running kinematics, Footwear
Science, DOI: 10.1080/19424280.2017.1388296
To link to this article:
Published online: 24 Oct 2017.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
Date: 25 October 2017, At: 05:36
Footwear Science, 2017
Influence of footwear comfort on the variability of running kinematics
Christian Meyera,b*, Maurice Mohra, Mathieu Falbriardc, Sandro R. Nigga and Benno M. Nigga
Human Performance Laboratory, Faculty of Kinesiology, University of Calgary, Calgary, Canada; bSpinal Cord Injury Center,
Balgrist University Hospital, Zurich, Switzerland; cLaboratory of Movement Analysis and Measurement, School of Engineering, EPFL,
Lausanne, Switzerland
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
(Received 5 September 2017; accepted 2 October 2017)
Footwear comfort is an important factor in design, purchase and use of running shoes but current measures require multiple
subjective assessments. Therefore, an objective and more reliable surrogate measure of footwear comfort would be of high
relevance. In other research fields, perceived comfort was found to influence the variability of movement execution.
Consequently, the purpose of this study was to investigate the influence of perceived footwear comfort on the variability of
running kinematics as a potential surrogate measure of comfort. Thirty-six recreational athletes ran in five different
running shoes on an indoor track while their running kinematics were recorded using a foot-mounted tri-axial inertial
measurement unit (IMU). Footwear comfort was measured through multiple subjective assessments. The relative
variability of IMU data was determined across the swing phase of 45 gait cycles and compared between the most and least
comfortable shoes. Lower footwear comfort was associated with lower kinematic variability especially in the second half
of the swing phase but only for variables that are not directly linked to the forward propulsion during running and mainly
describe frontal and transverse joint rotations. The constraints of an uncomfortable shoe may lead to the adaptation of a
more monotonous running style with the goal to stay in the least uncomfortable movement path. This finding may partially
explain a previously described higher injury risk when exercising in footwear of lower comfort, as more repetitive forces
could increase the risk of overuse injury. The results of this study implicate the possible use of IMU-based kinematic
variability as a surrogate measure of footwear comfort, which could complement subjective measures.
Keywords: running shoe comfort; comfort assessment; inertial measurement unit (IMU); kinematic variability;
foot sensor; comfort and kinematics
Footwear comfort is one of the key design features of running shoes (Nigg, 2010). This is due to the observed benefits of high footwear comfort to improve running
performance (Luo, Stergiou, Worobets, Nigg, & Stefanyshyn, 2009) and/or to possibly reduce movement-related
injuries as was shown for several sporting activities
(Kinchington, Ball, & Naughton, 2011; M€
Stefanyshyn, & Nigg, 2001). Furthermore, footwear comfort is the deciding factor when customers purchase new
footwear (Martınez-Martınez et al., 2016). These findings
highlight the importance of comfort in the design, purchase, and use of running shoes and emphasize the demand
for reliable comfort measures.
Subjective measures are widely used to assess comfort
in different contexts (Schiele, 2009; Vitacca et al., 2004;
Zhang, Helander, & Drury, 1996). However, such subjective methods to assess footwear comfort were only found
to be reliable if performed multiple times on different
days and averaged over multiple sessions (Hoerzer,
Trudeau, Edwards, & Nigg, 2016; Mills, Blanch, &
*Corresponding author. Email:
Ó 2017 Informa UK Limited, trading as Taylor & Francis Group
Vicenzino, 2010; M€undermann, Nigg, Stefanyshyn, &
Humble, 2002). Multiple subjective comfort assessments
on different days are not feasible in a running shoe store
or if one wants to measure footwear comfort instantaneously or continuously during exercise. Consequently,
it would be beneficial to develop objective surrogate
measures of footwear comfort that can easily be obtained
during one testing session to complement or even replace
subjective measures.
Most previous studies investigating footwear comfort
focused on benefits regarding injury risk and performance
but not on effects of comfort on the underlying running
kinematics. However, studies in the field of ergonomics
observed effects of comfort on the variability of movement executions or postural control (Madeleine &
Madsen, 2009; Søndergaard, Olesen, Søndergaard, de
Zee, & Madeleine, 2010). Furthermore, different studies
investigating patients with chronic pain, which can be
considered as in a state of high discomfort, reported lower
movement variability during different tasks (Hamill, van
Emmerik, Heiderscheit, & Li, 1999; Madeleine,
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
C. Meyer et al.
Mathiassen, & Arendt-Nielsen, 2008). These previous
findings lead to the assumption that shoe comfort might
affect the execution of the running movement in a similar
way. Specifically, it was hypothesized that footwear comfort influences the kinematic variability of running.
Kinematic variability can be characterized using footmounted inertial measurement units (IMUs) (Rebula,
Ojeda, Adamczyk, & Kuo, 2013). Recent advances in
IMU development have enabled small and light sensors
that are easy to mount on various body segments of a runner (Fong & Chan, 2010). Therefore, a main advantage of
IMUs compared to other devices characterizing kinematic
variability (e.g. motion capture systems) is the easy application outside of the laboratory setting. Thus, they allow a
mobile data collection for the use in a shoe store or in the
field to analyse running kinematic variability and potentially footwear comfort in a natural environment.
The objective of this study was to compare the variability of running kinematics between running shoes of
varying levels of perceived comfort using IMUs.
Thirty-six healthy and active adults (18 females and 18
males, (mean § SD) age: 25.4 § 3.5 years, height:
174.4 § 8.4 cm, weight: 68.4 § 9.2 kg) participated in
this study. The study sample incorporated a wide range of
running experience and training status from novice to
competitive runners to mimic the heterogeneous group of
runners that would participate in in-store running shoe
assessments. All participants gave written informed consent. Ethical approval for this research study involving
human participants was obtained from the University of
Calgary’s Conjoint Health Research Ethics Board, in
spirit of the Helsinki Declaration. Exclusion criteria were
lower extremity injuries within the last three months or a
low physical activity level (<two vigorous physical activities a week (>6 METs) (Ainsworth et al., 2011; Haskell
et al., 2007)). All but one participant reported using a consistent shoe model when running. Of these, seven participants reported using a shoe model that was tested in this
study. Kinematic variability may change after a runner
has adapted to a certain shoe model, however, the
influence of this on the study results might be minor as
only a small amount of the participants were familiarized
to a shoe that was tested. All participants were heel to toe
runners and maintained their striking pattern in all shoes
tested according to data from pressure insoles.
Shoe conditions
Five different running shoe conditions, currently on the
market, were tested by each subject in a randomized
order. The study included stability, neutral and minimalist
shoes. The weight ranged from 225 g for the lightest shoe
(Adipure) to 323.8 g for the heaviest shoe (Brooks).
Besides this, the shoes also differed in midsole properties,
heel-to-toe drop and pronation support (Table 1). For each
shoe, a minimalist index was calculated according to
Esculier, Dubois, Dionne, Leblond, and Roy (2015). It
ranged between 12% for the most cushioned shoe
(Brooks) to 54% for shoe model with the thinnest midsole
(Adipure). Since perceived shoe comfort is expected to
vary between individuals, these shoe differences ensured
that more and less comfortable footwear was provided for
each individual runner.
Each participant was equipped with one IMU (adidas RunGenie), which was placed on the dorsum of the right foot
(Figure 1(a)). The sensor was clipped to the laces of the
shoe and further fixed with tape to minimize relative movement between sensor and shoe. The placement of the sensor
on each shoe was standardized between footwear conditions by clipping it on the same lace. The sensor incorporated an accelerometer and a gyroscope, which acquired
data with reference to a local sensor coordinate system at a
sampling frequency of 250 Hz (Figure 1(c) and 1(d)).
Testing procedure
Each participant completed two testing sessions during
the course of this study. In the first session, a blinded comfort assessment was carried out while the second session
included the measurement of IMU-based running kinematics as well as a second comfort assessment. This study
Table 1. Running shoe characteristics for the shoes tested.
Heel-to-toe drop (mm)
Heel (mm)
Weight (US 9) (g)
Minimalist index
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
Footwear Science
Figure 1. Sensor placement on the dorsum of the right foot (a) and fixation with tape (b). The left-handed local coordinate system of the
IMUs (c) and the directions of the axis respective to the shoe with the x-axis pointing into the page (d). The coordinate system with
respect to the foot after calibration of the IMU signal (e). Met D metatarsal, cal D calcaneus.
design was selected to establish average comfort ratings
for all participants with respect to each of the five running
shoes. This procedure has been shown to ensure reliable
assessments of perceived footwear comfort (Hoerzer
et al., 2016; Mills et al., 2010; M€
undermann et al., 2002).
During the first testing session, participants ran a
200 m lap on an indoor track at their preferred running
speed in each of the five running shoes. During each running trial, participants were blinded to the footwear conditions by a dark fabric covering the shoes. In addition,
participants were blindfolded when putting on the shoes.
A member of the study team assisted during the procedure
but the participants adjusted the tightness of the laces
themselves. Immediately after each lap, participants rated
the perceived comfort of the respective shoe on a 10 cm
visual analogue scale (VAS). After running in all five
shoe conditions, participants ranked the shoes from most
to least comfortable (rank 1–5). Ranking scales and VAS
have been found to be the most reliable tools for assessing
subjective footwear comfort (Mills et al., 2010;
undermann et al., 2002). The second testing session
took place 2–3 days later. For each shoe, participants first
completed a warm up round of 200 m at their preferred
running speed in order to familiarize with a specific footwear condition. Next, IMU data were recorded during a
30 second static standing trial for subsequent sensor calibration procedures. Then, IMU-based running kinematics
were recorded while participants completed three trials of
running 60 m along a straight, flat indoor track. The running speed during the data collection was predetermined
at 3.5 m/s and controlled using timing lights at intervals
of 20 m. For each shoe, a comfort assessment and final
comfort ranking was carried out identical to the first
session with the only difference that participants were not
blinded to the footwear conditions.
IMU data processing
For each condition, all heel strikes were detected, and gait
cycles were identified between two consecutive heel
strikes. The heel strike detection identified the point
before the onset of high frequency accelerations due to
the collision of the foot with the ground during heel strike.
For this purpose, the acceleration data along the sensor zaxis was used. First, the signal was low-pass filtered with
a cut-off frequency of 20 Hz. Peaks in the low-pass filtered signal were present right before the impact and represent smooth leg and foot movements during the swing
phase before heel strike. The amplitude and time of these
peaks were determined. Afterwards, the acceleration
along the sensor z-axis was differentiated. Following each
peak in the low-pass filtered signal, a time point was identified when the differentiated signal was first higher than
the low-pass filtered signal. This time point was considered as the heel strike (Figure 2(a) and 2(b)). The heel
strike detection algorithm was validated in one runner
against a force plate (10 N threshold, 2400 Hz) as the gold
standard. For five steps during overground running, there
was a mean absolute error of 5.9 § 6.4 ms between the
IMU and force plate-based heel strike detection. Given
the IMU sampling frequency of 250 Hz, which corresponds to a sample time of 4 ms, these mean errors are
within two samples and were considered acceptable.
From each 60 m running trial, the first three strides were
omitted and the next 15 strides were used for the analysis
to obtain kinematics during constant speed running.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
C. Meyer et al.
Figure 2. (a and b) Heel strike detection in the data of the IMU placed on the dorsum using the signal of the acceleration in z direction.
(a) Filtering of the signal. (b) Detecting heel strike using the filtered and the differentiated signal of the acceleration along the z-axis. (c
and d) Data of all strides for one condition of one participant (c) and corresponding mean and standard deviation (d). Variability analysis
conducted between dashed vertical lines (50%–95% of gait cycle).
Consequently, a total of 45 strides were available for each
shoe/participant condition (Figure 2(c)).
Next, a sensor calibration procedure was performed.
Using the static standing periods and the running movement itself, the technical frame of each IMU was aligned
with the functional frame of the foot they were affixed to.
The static standing periods were used to find the foot’s
vertical axis and the azimuth was set by hypothesising
that most of the foot rotations in running occur along the
medio-lateral axis of the foot. As such, the influence of
the IMU orientation on the foot is reduced. Only the calibrated data was used for the subsequent data analysis and
all presented results refer to the functional reference frame
of the foot (Figure 1(e)).
For each gait cycle, all kinematic data (three acceleration signals and three angular velocity signals) were timenormalized to the gait cycle duration from 0% to 100%.
The signal of the IMUs around heel strike and toe off
exhibited a substantial amount of noise, which would artificially influence the variability. Furthermore, the foot
with the mounted IMU is relatively stationary during the
stance phase resulting in low information content with
respect to variability. Based on these two reasons, only
the time interval between 50% and 95% of the gait cycle,
representing a portion of the swing phase of running, was
analysed (dotted, vertical lines, Figure 2(c) and 2(d)).
ratings and comfort ranks were averaged across the first
and second testing session. Next, the most and the least
comfortable shoes were determined for each participant
according to the lowest and highest average comfort
ranks. If two shoes exhibited the same average rank, the
respective average VAS scores were used to select the
most and least comfortable shoes. A paired t-test between
average VAS scores of the most and least comfortable
shoes was performed in order to confirm the presence of a
significant difference in perceived comfort between the
selected shoes.
Data analysis: kinematic variability
For each participant, the mean and standard deviation for
each time point within the used portion of the swing phase
were calculated across the 45 time-normalized steps for
each shoe condition (Figure 2(d)). Then, the root mean
square (RMS) of the mean and the standard deviation was
used to determine a relative measure for variability
according to Equation (1) (Enders, Maurer, Baltich, &
Nigg, 2013) and one value was obtained per participant
and condition. This normalized measure was chosen as it
allows comparisons between shoe conditions, different
sensor outputs and participants:
Relative Variability ðRVÞ D
Data analysis: perceived comfort
All statistical tests were performed using IBM SPSS Statistics 22 at a significance level of a D 0.05. Both VAS
RMSstandard deviation
The effect of footwear comfort on relative variability
was evaluated using a repeated measures MANOVA with
Footwear Science
Table 2. Distribution of shoe preference.
Most comfortable
Least comfortable
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
the within-subject factor ‘comfort’ (most vs. least comfortable) and the six acceleration and angular velocity
channels as dependent variables. In the presence of a significant main effect of comfort, paired t-tests were used to
determine effects of comfort on relative variability for
each individual IMU channel. In addition, effects of shoe
type on kinematic variability were investigated using
repeated measures MANOVA with the within-subject factor ‘shoe type’ and the six acceleration and angular velocity channels as dependent variables, followed by
individual repeated measures ANOVAs and pairwise
for 89% of the subjects. Identically, 89% of the participants assigned the lowest two ranks to the selected least
comfortable shoe in both sessions.
Effect of perceived comfort on kinematic variability
Overall, there was a significant main effect of comfort on
relative variability during the swing phase of running (F
(6,30) D 2.593, p D 0.038, Wilk’s L D 0.658, partial h2 D
0.342). On average, all six IMU channels showed a reduction in relative variability in the least comfortable compared to the most comfortable shoe (Figure 4(a)). For the
acceleration along the z-axis, the variability was most
reduced by 2.8% (p D 0.067, Cohen’s d D 0.31) from the
most to the least comfortable shoe. As the relative variability is a normalized measure, it is reported in percentage units. Nevertheless, the changes presented here
describe absolute differences. Similarly, the relative
variability of the angular velocity about the y-axis
was reduced by an average of 1.7% (p D 0.072, Cohen’s
Perceived comfort
Each of the five shoes was ranked as most or least comfortable by at least four participants (Table 2), which demonstrates a wide spread in how footwear comfort is
perceived among different individuals. On average, the
VAS comfort ratings within a subject for the most comfortable shoe were significantly higher by 3.6 cm (p <
0.001) compared to the least comfortable shoe condition
(Figure 3). Most and least comfortable shoes were determined using mean ranks and as a tie breaker mean VAS
ratings of the two testing session. The detected most comfortable shoe was ranked first or second in both sessions
Figure 3. Mean VAS ratings § SD of the most and least comfortable shoe condition for n D 36 subjects.
Figure 4. Mean absolute differences in relative variability §
standard error between most and least comfortable shoe (least –
most comfortable) for all six sensor outputs for (a) the swing
phase of running (50%–95% of the stride) and (b) the second
half of the swing phase from 75% to 95% of the stride. P-values
<0.1 are indicated and significant differences (p < 0.05) are
marked with an asterisk. N D 36 subjects.
C. Meyer et al.
present for all acceleration channels as well as for the
angular velocity about the sensor x-axis. The pairwise
comparisons in relative variability between shoes were
generally unsystematic with a trend towards slightly
higher relative variability in the shoes Brooks and
Sequence in a subset of the variables.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
Figure 5. Comparison of the course of the angular velocity in
the y-direction for most and least comfortable shoe (45 steps) for
two exemplary participants.
d D 0.31). However, when analysing the entire swing
phase, none of the individual channels showed a statistically significant reduction in relative variability between
comfort levels.
Figure 5 shows two exemplary subjects who demonstrate the reduction in variability from the most to the least
comfortable shoe condition with respect to the angular
velocity about the sensor y-axis. Visual inspection of
Figure 5 indicates that for these subjects the variability of
the angular velocity was highest during mid to late swing
phase. Furthermore, the difference in variability between
the most and least comfortable shoe also seems to be highest during mid to late swing phase. Consequently, the statistical comparisons between the kinematic variability of
the most and least comfortable shoe conditions were reevaluated considering only the second half of the swing
phase (75%–95% of the gait cycle).
For the second half of the swing phase, the effect of
perceived comfort was stronger and the average relative
variability of the acceleration along the sensor z-axis and
the angular velocity about the sensor y-axis was significantly reduced by 6.1% (p D 0.041, Cohen’s d D 0.35)
and 3.1% (p D 0.046, Cohen’s d D 0.34), respectively
(Figure 4(b)).
Effect of shoe type on kinematic variability
There was a significant main effect of shoe type on relative variability (F(24,12) D 3.825, p D 0.01, Wilk’s L D
0.116, partial h2 D 0.884 (Figure 6). Subsequent univariate tests indicated that the significant shoe effect was
The purpose of this study was to investigate if perceived
comfort of running footwear affects the variability of running kinematics. The findings of this study indicate that
for certain variables, kinematic variability is reduced during the swing phase when running in a less comfortable
shoe compared to a more comfortable shoe. Therefore,
the initially formulated hypothesis was supported, but is
variable and gait cycle dependent.
The two variables with the highest reduction in kinematic variability from the most to least comfortable shoes
were the angular velocity about the y-axis and the acceleration in z-direction. In the absence of any other lower
extremity joint rotations, these variables would approximate a rotation of the shoe in the transverse plane, i.e.
ankle internal/external rotation, and a medio-lateral translation of the shoe, respectively (Figure 1(e)). This is not
the case during the swing phase of running when all lower
extremity joints are performing three-dimensional rotations simultaneously. Consequently, the angular velocity
about the y-axis and acceleration in z-direction during the
swing phase likely represent combinations of frontal and
transverse plane rotations about the ankle, knee and hip
joint, and possibly even movements of the upper body. In
contrast, variables that predominantly correspond to sagittal plane rotations and translations, i.e. the angular velocity about the z-axis and the acceleration in x-direction,
showed no difference in variability between the most and
least comfortable shoes. Furthermore, these variables generally exhibit the lowest kinematic variability among all
sensor outputs (Figure 6).
Therefore, one may speculate that a change in footwear comfort predominantly affects the variability of joint
movements that are outside of the sagittal plane and not
directly linked to the forward motion during running. On
the other hand, sagittal plane joint movements are essential to execute the running movement, naturally exhibit
lower variability, and thus may not be easily affected by
changing external conditions such as footwear comfort.
This would be similar to previous findings in cycling,
where less essential muscles for performing the tasks
were found to show higher changes in variability of the
activity between effort levels than essential muscles
(Enders et al., 2013). Future investigations should examine which sensor locations, e.g. foot vs. tibia, are most
sensitive to a change in footwear comfort to further
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
Footwear Science
Figure 6. Mean relative variability § SD for all six sensor outputs in the different shoe conditions. P-values represent the results of the
subsequent repeated-measures ANOVAs, significant pairwise differences are further indicated with brackets and asterisks. N D 36.
RV D Relative variability, Gli D Glide, ADZ D Adizero, Bro D Brooks, Seq D Sequence, ADP D Adipure, Ang. velocity D angular
explore the relationship between footwear comfort and
variability of lower extremity kinematics during running.
The kinematic variability of the lower extremity motion
seemed to be highest during the mid to late swing phase,
i.e. before the heel strike and the transition to the stance
phase. Furthermore, the difference in kinematic variability
between running in the most and least comfortable shoe
was most apparent during the late swing phase (Figures 4
and 5). These results are in agreement with a previous study
investigating the variability of the relative movement of
lower extremity joints during running and comparing individuals with and without patellofemoral pain (Hamill et al.,
1999). Similar to this study, it was demonstrated that the
variability of lower extremity coupling in normal runners
was highest during the transitions from stance to swing
(40%–50% of the stride) and from swing to stance (80%–
100% of the stride). This observation is consistent with
dynamical systems theory, where increased variability is a
sign of disruptions in coordinative movement patterns that
are typical for transition phases (Haken, Kelso, & Bunz,
1985). Furthermore, Hamill and colleagues showed that the
differences in the coordinate variability between pain-free
and symptomatic runners were strongest during these transition phases with symptomatic runners showing lower variability. Based on this and subsequent studies, the authors
concluded that the pain experienced by the symptomatic
runners led to a reduction of the possible combinations for
inter-segment coordination and subsequently a reduction in
variability (Hamill et al., 1999; Hamill, Palmer, & Van
Emmerik, 2012a; Heiderscheit, Hamill, & Van Emmerik,
2002). The angular velocities and accelerations measured
by the foot-mounted IMU in this study also represent a
combination of segment movements of the entire lower
limb. Consequently, a similar concept may be used to
explain the differences in kinematic variability between
running in more or less comfortable footwear. Specifically,
it is speculated that a less comfortable shoe reduces the
number of ‘comfortable’ solutions to execute the running
stride and restricts the runner to a narrower range of movement patterns. Since the differences in kinematic variability
were most apparent during the late swing phase when the
runners prepare for heel strike, we hypothesize that a more
comfortable shoe offers more possible solutions to accomplish the landing phase and the associated impact with minimal discomfort. A second intriguing assumption that
follows is that a lower number of solutions to execute the
landing phase in a less comfortable shoe could cause internal forces to be distributed across a smaller surface area of
the load-bearing tissues, possibly resulting in overuse injuries and explaining the link between footwear comfort and
injury risk (Hamill, Palmer, & Van Emmerik, 2012b; Kinchington et al., 2011; M€undermann et al., 2001).
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
C. Meyer et al.
However, the noise in the IMU signals associated with
heel strike and toe off did not allow to evaluate this
assumption in more detail. Future studies should determine if other sensor positions along the leg show a similar
relationship between variability and comfort but less
impact-related noise and could be used to investigate
these speculations. If that is the case, a similar methodology as used in this study could be employed in prospective
studies to explore a potential link between footwear discomfort, kinematic variability and injury risk.
The comparison of the kinematic variability between
the five different running shoes indicates that there was an
effect of shoe type. This broadens the results of a previous
study and adds that the variability does not only differ
between shod and barefoot running but also between different shoes (Kurz & Stergiou, 2003). In this study, the
Brooks and Sequence shoe exhibited a trend towards
higher relative variability in some variables compared to
other shoe conditions. The reason for this effect is currently unknown. It could be that compared to the other
shoe conditions, the high weight of the Brooks and
Sequence may have increased the variability of kinematics during the swing phase (Table 1).
The observed differences in variability between shoes
warrant the question whether the effect of shoe type influenced the results related to perceived comfort. However,
there are several examples demonstrating that these shoe
effects did not bias our findings. For example, a significant
effect of shoe comfort on the relative variability along the yaxis was found while no significant shoe effect was present
for this axis (Figure 4). In addition, for the acceleration in zdirection, the variable with the highest change from most to
least comfortable shoe, the Brooks and the Sequence showed
slightly higher relative variability. However, the two shoes
together were equally chosen as most and least comfortable
shoe condition (Table 2). Therefore, it is unlikely that shoe
effects influenced the within-subject comparisons related to
footwear comfort in the present study.
The subjective comfort questionnaire used in this
study showed a relevant difference between the most and
the least comfortable shoe condition (Mills et al., 2010). It
can therefore be assumed that the differences of the construction features between the shoes were big enough to
establish different comfort levels. Furthermore, the selection of most and least comfortable shoes showed good
agreement (89%) between two testing days, which supports the findings of previous studies, that averaged subjective measures of different sessions provide a reliable
footwear comfort assessment (Hoerzer et al., 2016;
undermann et al., 2002). However, multiple comfort
assessments on different days are not feasible in a shoe
store. This study demonstrates a relationship between
footwear comfort and kinematic variability and therefore
provides a potential objective measure for shoe comfort.
These findings may simplify future running shoe
recommendations, as footwear comfort could be objectively assessed in a single session. Nevertheless, further
research for the development of a full assessment tool is
needed, e.g. a combination of kinematic variability measurements with assessments of foot anthropometry and sensitivity (M€undermann et al., 2001).
A limitation of the present study is that the running
speed was pre-selected. Although other studies have used
similar average speeds (Kurz & Stergiou, 2003: 3.24 m/s;
Sasimontonkul, Bay, & Pavol, 2007: 3.5–4.0 m/s; Sinclair,
Greenhalgh, Brooks, Edmundson, & Hobbs, 2013: 4.0 m/s;
Ueda et al., 2015: 3.5 m/s), the participants probably have
different preferred running speeds. Running at a speed different from the preferred speed (higher or lower) may influence the variability of running kinematics. As a withinsubject design was used, this factor did likely not influence
the within-subject comparisons of kinematic variability
between various shoes. It is assumed that differences in
kinematic variability between different shoes of varying
perceived comfort may become even more pronounced if
runners are exercising at their preferred running speed. A
further limitation is the heterogeneity of the study population in terms of running experience. It is possible that novice runners might exhibit a running style with more
kinematic variability than experienced runners. Again, as a
within-subject analysis was carried out, the influence of
this limitation on the results of the study might be minor.
In summary, kinematic variability during running
seems to be affected by shoes and perceived footwear comfort. Specifically, lower perceived comfort is associated
with lower kinematic variability, mainly for the late swing
phase and mainly for variables that describe frontal and
transverse joint rotations. These findings give new insight
into how running shoes interact with an individual’s running style and provide new approaches to study the unexplained benefits of high footwear comfort with respect to
injury risk and performance. In this study, kinematic variability was characterized with the use of IMUs. Their convenient application is suitable for the use in a running store
and for a continuous assessment of kinematic variability
during running in the field. Therefore, the use of IMUbased kinematic variability may be a potential surrogate
measure of footwear comfort that could complement or
even replace subjective measures in the future to improve
personalized footwear recommendations.
The authors would like to thank adidas (Herzogenaurach,
Germany) for providing the testing shoes.
Disclosure statement
Adidas (Herzogenaurach, Germany) provided the shoes that
were used for this study. However, the results presented in this
Footwear Science
article do not in any way represent a bias toward adidas products
over other brands. Furthermore, the results of the present study
do not constitute endorsement of the product by the authors. The
authors declare no conflict of interest and that all results of the
study are presented clearly, honestly, and without fabrication,
falsification, or inappropriate data manipulation.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
Ainsworth, B.E., Haskell, W.L., Herrmann, S.D., Meckes, N., Bassett, D.R., Tudor-Locke, C., … Leon, A.S. (2011). 2011 compendium of physical activities: A second update of codes and
MET values. Medicine and Science in Sports and Exercise, 43
(8), 1575–1581. doi: 10.1249/MSS.0b013e31821ece12
Enders, H., Maurer, C., Baltich, J., & Nigg, B.M. (2013). Task-oriented control of muscle coordination during cycling. Medicine
and Science in Sports and Exercise, 45(12), 2298–2305.
Retrieved from
Esculier, J.-F., Dubois, B., Dionne, C.E., Leblond, J., & Roy, J.S. (2015). A consensus definition and rating scale for minimalist shoes. Journal of Foot and Ankle Research, 8, 42.
Retrieved from
Fong, D T.P., & Chan, Y.Y. (2010). The use of wearable inertial
motion sensors in human lower limb biomechanics studies:
A systematic review. Sensors (Switzerland). Retrieved from
Haken, H., Kelso, J.A.S., & Bunz, H. (1985). A theoretical
model of phase transitions in human hand movements. Biological Cybernetics. Retrieved from
Hamill, J., Palmer, C., & Van Emmerik, R.E.A. (2012a). Coordinative variability and overuse injury. Sports Medicine,
Arthroscopy, Rehabilitation, Therapy & Technology :
SMARTT, 4(1), 45. Retrieved from
Hamill, J., Palmer, C., & Van Emmerik, R.E.A. (2012b). Coordinative variability and overuse injury. Sports Medicine,
Arthroscopy, Rehabilitation, Therapy & Technology, 4(1),
45. Retrieved from
Hamill, J., van Emmerik, R.E.A., Heiderscheit, B.C., & Li, L.
(1999). A dynamical systems approach to lower extremity
running injuries. Clinical Biomechanics, 14(5), 297–308.
Retrieved from
Haskell, W.L., Lee, I.-M.M., Pate, R.R., Powell, K.E., Blair, S.N.,
Franklin, B.A., … Bauman, A. (2007). Physical activity and
public health: Updated recommendation for adults from the
American College of Sports Medicine and the American
Heart Association. Circulation, 116(9), 1081–1093. Retrieved
Heiderscheit, B.C., Hamill, J., & Van Emmerik, R.E.A. (2002).
Variability of stride characteristics and joint coordination
among individuals with unilateral patellofemoral pain. Journal of Applied Biomechanics, 18(2), 110–121. Retrieved
Hoerzer, S., Trudeau, M.B., Edwards, W.B., & Nigg, B.M.
(2016). Intra-rater reliability of footwear-related comfort
assessments. Footwear Science, 8(3), 155–163. Retrieved
Kinchington, M.A, Ball, K.A, & Naughton, G. (2011). Effects of
footwear on comfort and injury in professional rugby league.
Journal of Sports Sciences, 29(13), 1407–1415. Retrieved
Kurz, M.J., & Stergiou, N. (2003). The spanning set indicates that
variability during the stance period of running is affected by
footwear. Gait and Posture, 17(2), 132–135. Retrieved from
Luo, G., Stergiou, P., Worobets, J., Nigg, B., & Stefanyshyn, D.
(2009). Improved footwear comfort reduces oxygen consumption during running. Footwear Science, 1(1), 25–29.
Retrieved from
Madeleine, P., & Madsen, T.M.T. (2009). Changes in the
amount and structure of motor variability during a deboning
process are associated with work experience and neck-shoulder discomfort. Applied Ergonomics, 40(5), 887–894.
Retrieved from
Madeleine, P., Mathiassen, S.E., & Arendt-Nielsen, L. (2008).
Changes in the degree of motor variability associated with
experimental and chronic neck–shoulder pain during a
standardised repetitive arm movement. Experimental Brain
Research, 185(4), 689–698. Retrieved from
Martınez-Martınez, J.M., Martın-Guerrero, J.D., Soria-Olivas, E.,
Bernabeu, J.A., Escandell-Montero, P., Stark, R.H., … Montiel, E. (2017). Use of SOMs for footwear comfort evaluation.
Neural Computing and Applications, 28(7), 1763–1773.
Retrieved from
Mills, K., Blanch, P., & Vicenzino, B. (2010). Identifying clinically meaningful tools for measuring comfort perception of
footwear. Medicine and Science in Sports and Exercise, 42
(10), 1966–1971. Retrieved from
undermann, A., Nigg, B.M., Stefanyshyn, D.J., & Humble, R.
N. (2002). Development of a reliable method to assess footwear comfort during running. Gait and Posture, 16(1), 38–
45. Retrieved from
undermann, A., Stefanyshyn, D.J., & Nigg, B.M. (2001). Relationship between footwear comfort of shoe inserts and
anthropometric and sensory factors. Medicine and Science
in Sports and Exercise, 33(11), 1939–1945. Retrieved from
Nigg, B.M. (2010). Biomechanics of sport shoes. Calgary: University of Calgary.
Rebula, J.R., Ojeda, L.V., Adamczyk, P.G., & Kuo, A.D. (2013).
Measurement of foot placement and its variability with inertial sensors. Gait and Posture, 38(4), 974–980. Retrieved
Sasimontonkul, S., Bay, B.K., & Pavol, M.J. (2007). Bone contact
forces on the distal tibia during the stance phase of running.
Journal of Biomechanics, 40(15), 3503–3509. Retrieved from
Schiele, A. (2009). Ergonomics of exoskeletons: Subjective performance metrics. In Third Joint EuroHaptics conference
and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems (Ed.), 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems,
IROS 2009 (pp. 480–485). Salt Lake City, UT: World
Haptics 2009. Retrieved from
Sinclair, J., Greenhalgh, A., Brooks, D., Edmundson, C.J., &
Hobbs, S.J. (2013). The influence of barefoot and barefootinspired footwear on the kinetics and kinematics of running
in comparison to conventional running shoes. Footwear Science, 5(1), 45–53. Retrieved from
C. Meyer et al.
Downloaded by [UNIVERSITY OF ADELAIDE LIBRARIES] at 05:36 25 October 2017
Søndergaard, K.H.E., Olesen, C.G., Søndergaard, E.K., de Zee,
M., & Madeleine, P. (2010). The variability and complexity
of sitting postural control are associated with discomfort.
Journal of Biomechanics, 43(10), 1997–2001. Retrieved from
Ueda, T., Hobara, H., Kobayashi, Y., Heldoorn, T.A., Mochimaru, M., & Mizoguchi, H. (2015). Wearing inappropriate
footwear increases the loading rate during running. Footwear Science, 7(Suppl. 1), S55–S56. Retrieved from https://
Vitacca, M., Bianchi, L., Zanotti, E., Vianello, A., Barbano, L.,
Porta, R., & Clini, E. (2004). Assessment of physiologic variables and subjective comfort under different levels of pressure support ventilation. Chest, 126(3), 851–859. Retrieved
Zhang, L., Helander, M.G., & Drury, C.G. (1996). Identifying
factors of comfort and discomfort in sitting. Human Factors,
38(3), 377–389. Retrieved from
Без категории
Размер файла
675 Кб
2017, 19424280, 1388296
Пожаловаться на содержимое документа