Content uploaded by Jean-Benoît Morin
Author content
All content in this area was uploaded by Jean-Benoît Morin on Feb 12, 2014
Content may be subject to copyright.
ORIGINAL ARTICLE
Mechanical determinants of 100-m sprint running performance
Jean-Benoı
ˆt Morin •Muriel Bourdin •
Pascal Edouard •Nicolas Peyrot •Pierre Samozino •
Jean-Rene
´Lacour
Received: 29 December 2011 / Accepted: 1 March 2012
ÓSpringer-Verlag 2012
Abstract Sprint mechanics and field 100-m perfor-
mances were tested in 13 subjects including 9 non-spe-
cialists, 3 French national-level sprinters and a world-class
sprinter, to further study the mechanical factors associated
with sprint performance. 6-s sprints performed on an
instrumented treadmill allowed continuous recording of
step kinematics, ground reaction forces (GRF), and belt
velocity and computation of mechanical power output and
linear force–velocity relationships. An index of the force
application technique was computed as the slope of the
linear relationship between the decrease in the ratio of
horizontal-to-resultant GRF and the increase in velocity.
Mechanical power output was positively correlated to
mean 100-m speed (P\0.01), as was the theoretical
maximal velocity production capability (P\0.011),
whereas the theoretical maximal force production capa-
bility was not. The ability to apply the resultant force
backward during acceleration was positively correlated to
100-m performance (r
s
[0.683; P\0.018), but the
magnitude of resultant force was not (P=0.16). Step
frequency, contact and swing time were significantly cor-
related to acceleration and 100-m performance (positively
for the former, negatively for the two latter, all P\0.05),
whereas aerial time and step length were not (all P[0.21).
Last, anthropometric data of body mass index and lower-
limb-to-height ratio showed no significant correlation
with 100-m performance. We concluded that the main
mechanical determinants of 100-m performance were (1) a
‘‘velocity-oriented’’ force–velocity profile, likely explained
by (2) a higher ability to apply the resultant GRF vector
with a forward orientation over the acceleration, and (3) a
higher step frequency resulting from a shorter contact time.
Keywords Performance Force–velocity Power output
Ground reaction force application
Introduction
The 100-m event is the standard measure of the extreme
speed capabilities of human bipedal locomotion and
defines the ‘‘world’s fastest human’’ for a given time per-
iod. The scientific research about the limits of human
locomotion and the determinants of sprint performance has,
therefore, considered record holders and world champions
as examples of the limits of muscular, physiological and
mechanical features of human locomotion. Paradoxically,
Communicated by Guido Ferretti.
J.-B. Morin P. Edouard
University of Lyon, 42023 Saint Etienne, France
J.-B. Morin P. Edouard
Laboratory of Exercise Physiology (EA4338),
42000 Saint-Etienne, France
J.-B. Morin (&)
Laboratoire de Physiologie de l’Exercice (EA4338),
Me
´decine du Sport-Myologie, CHU Bellevue,
42055 Saint-Etienne cedex 2, France
e-mail: jean.benoit.morin@univ-st-etienne.fr
M. Bourdin J.-R. Lacour
University of Lyon, University Lyon 1, IFSTTAR,
UMR_T9406, LBMC, 69921 Oullins, France
N. Peyrot
University of La Re
´union, DIMPS (EA4075),
97430 Le Tampon, France
P. Samozino
Laboratory of Exercise Physiology (EA4338),
University of Savoie, 73376 Le Bourget-du-Lac, France
123
Eur J Appl Physiol
DOI 10.1007/s00421-012-2379-8
elite 100-m sprinters have been the specific focus of very
few experimental studies. To our knowledge, only Weyand
et al. (2000) presented experimental data obtained in the
three 100-m medalists of the 1996 Olympic Games in a
specific study about top speed production, and a more
detailed physiological and biomechanical case study about
the fastest sprinter with leg amputation (Weyand et al.
2009).
Further, data of sprint kinematics obtained during offi-
cial events such as the World Championships in Athletics,
i.e. not during specific experimental studies, have been
published (e.g. Moravec et al. 1988). Other studies have
considered world-class sprinters, but only used their offi-
cial performance data as inputs of mathematical models
(e.g. Arsac and Locatelli 2002; Beneke and Taylor 2010;
Ward-Smith and Radford 2000), and did not perform or
report specific experimental measurements. This lack of
experimental data in elite sprinters contrasts with other
competitive sports for which experimental case studies of
world top-level athletes and world record holders have
been published, for instance in rowing (Lacour et al. 2009),
cycling (Coyle 2005), or middle-distance and marathon
running (Jones 2006; Lucia et al. 2008). This may also be a
limit to a thorough and clear understanding of the deter-
minants of sprint running ability, specifically when the
populations of sprint studies do not include top-level
sprinters. In the present study, we had the unique oppor-
tunity to specifically study a group of subjects including a
young world-class male sprinter, and three French national-
level sprinters.
When considering the physiological correlates of 100-m
performance, except for muscle fibers distribution (e.g.
Baguet et al. 2011; Gollnick and Matoba 1984) and the
capacity for using high-energy phosphates (Hirvonen et al.
1987), no clear consensus has been made from experi-
mental data on the fact that 100-m performance and human
maximal running speed were predominantly determined by
physiological factors/pathways such as for instance lactate
accumulation or clearance (e.g. Bret et al. 2003; Hirvonen
et al. 1987). Consequently, and in light of existing studies
about high-speed running mechanics, we propose that
neurological and mechanical factors are more relevant to
100-m sprint performance and top speed in humans. For
instance, Weyand et al. (2000,2010) related the specific
ability to run at high speed to the production of high
amounts of vertical ground reaction force (GRF) per unit
body weight (BW) (Weyand et al. 2000), and to the time
needed/available to apply these high amounts of force onto
the supporting ground (Weyand et al. 2010) through
experimentally controlled research designs. Other scientists
showed the important role of horizontal GRF and impulses
in animals (e.g. Roberts and Scales 2002) and human
(Hunter et al. 2005) acceleration capability.
The ability of athletes to specifically apply high amounts
of GRF in the horizontal direction at the various speeds
produced over a typical sprint acceleration is well descri-
bed by linear force–velocity (F–V) relationships and 2nd
degree polynomial power–velocity relationships (Jaskolska
et al. 1999; Morin et al. 2010), as it is also the case in
horizontal or incline push-off (e.g. Samozino et al. 2012)or
cycling (e.g. Dorel et al. 2010). In particular, since
mechanical power is the product of force and velocity, the
slope of the linear F–V relationship (Jaskolska et al. 1999;
Morin et al. 2010) may indicate the relative importance of
force and velocity qualities in determining the maximal
power output, and the individual F–V profile of each sub-
ject. Such individual F–V profiles have recently been
studied and related to power output and performance in
jumping exercises (Samozino et al. 2012). These individual
F–V relationships describe the changes in external hori-
zontal force generation with increasing running velocity
and may be summarized through their two theoretical
extrema: the theoretical maximal horizontal force the legs
could produce over one contact phase at null velocity
(F
H
0), and the theoretical maximal velocity of the treadmill
belt the legs could produce during the same phase under
zero load (V0). These integrative parameters characterize
the mechanical limits of the entire neuromuscular system
during sprint running, encompass numerous individual
muscle mechanical properties, morphological, neural and
technical factors (Cormie et al. 2011) and, therefore, pro-
vide an integrative view of the F–V mechanical profile of
an athlete in his specific sprint running task. In particular,
although power output (yet quantified during other move-
ments than sprint running: vertical jump, sprint cycling)
was expected as highly correlated to sprint running per-
formance (e.g. Cronin and Hansen 2005; Cronin and
Sleivert 2005; Harris et al. 2008; Sleivert and Taingahue
2004), the relative importance of its force and velocity
components was unknown.
Recently, continuous GRF measurements in three
dimensions during a sprint acceleration were made possible
with the use of an instrumented sprint treadmill (Morin
et al. 2010). When comparing data of horizontal, vertical
and resultant GRF, these authors showed that during the
acceleration phase, the orientation of the resultant GRF
vector, related to athletes’ technical abilities, was a stron-
ger determinant of field sprint performance than the mag-
nitudes of vertical or resultant force vectors. Indeed, Morin
et al. (2011a) showed that the magnitude of the horizontal
component of the GRF per unit BW measured on the
treadmill over an accelerated run was highly correlated to
100-m performance (mean and top running speeds),
whereas the magnitude of the resultant GRF was not. They
also defined an index of force application technique (D
RF
),
which quantifies a runner’s ability to maintain a forward
Eur J Appl Physiol
123
orientation of the resultant GRF vector despite increasing
speed over the entire acceleration phase (see ‘‘Methods’’).
In two recent studies, the authors proposed and experi-
mentally supported the idea that the D
RF
index was sig-
nificantly related to field 100-m performance (Morin et al.
2011a) and significantly altered with fatigue over a repe-
ated sprint series (Morin et al. 2011b). They concluded that
the orientation of the resultant force vector applied against
the supporting ground during sprint acceleration was more
important to 100-m performance than its magnitude.
However, their conclusions were limited because these
results were obtained in subjects of rather low level of
sprint performance (ranging from non-specialists to regional-
level sprinters).
Last, in parallel with these functional abilities of force
production, some aspects of human body design have been
suggested to be requirements for high sprinting speed:
specifically a high BMI (Watts et al. 2011; Weyand and
Davis 2005) and long limbs (van Ingen Schenau et al.
1994). The present study also allowed us to further discuss
these anthropometric results.
The aim of this study was to investigate the detailed
mechanical variables associated with field 100-m perfor-
mance and discuss recent hypotheses about the mechanical
determinants of sprint performance (Morin et al. 2011a;
Weyand et al. 2000). To better elucidate the mechanical
correlates of 100-m sprint performance, we used instru-
mented sprint treadmill measurements, performed in a
group of subjects including a world-class and three
national-level sprinters. This was thought useful to poten-
tially moderate or strengthen the aforementioned results,
especially the recent conclusions of Morin et al. (2011a)
stating that the orientation of the resultant GRF vector onto
the ground was more important to 100-m performance than
its magnitude.
Methods
Subjects
Thirteen male subjects participated in the study. They had
different sprint performance levels: nine of them were
physical education students [age (mean ±SD) 26.5 ±1.8
years; body mass 72.6 ±8.4 kg; height 1.75 ±0.08 m]
who were all physically active and had all practiced
physical activities including sprints (e.g. soccer, basketball)
in the 6 months preceding the study, but were not sprint
specialists. Three were French national-level sprinters [age
(mean ±SD) 26.3 ±2.1 years; body mass 77.5 ±4.5 kg;
height 1.83 ±0.05 m]. Their personal best times on 100-m
relay (last update September 5, 2011) ranged from 10.31
to 10.61 s. And one subject was a world-class sprinter
(age 21 years; body mass 81.0 kg; height 1.91 m). His
official best performances were (last update September 5,
2011): 9.92 s on the 100-m and 19.80 s on the 200-m.
Among his official titles, he is French National Champion
and record holder on 100- and 200-m, he has won the
World Junior Championships on 200-m in 2008, and has
been European Champion in 2010 on 100-, 200- and
49100-m relay. More recently, he finished at the 4th and
3rd place on 100- and 200-m, respectively, at the 2011
World Championships in athletics. All subjects gave their
written informed consent to participate in this study after
being informed about the procedures approved by the local
ethical committee and in agreement with the Declaration of
Helsinki.
Experimental protocol
For each subject, two sets of measurements were per-
formed: (1) a laboratory test consisting in performing a 6-s
maximal sprint after full familiarization with the sprint
treadmill and an appropriate standardized warm up, and (2)
a field 100-m test performed on a standard synthetic track,
after an appropriate standardized warm up. The non-spe-
cialist subjects performed these treadmill and field tests
within a unique testing session, with treadmill and field
sprints performed in a randomized counterbalanced order
among subjects, and with about 30–45 min of passive rest
between tests (for full details, see Morin et al. 2011a). The
world-class and national-level sprinters were tested on two
distinct occasions: in mid-March and mid-April 2011
(treadmill and field performance measurements, respec-
tively). This corresponded to the training period just pre-
ceding the beginning of their official outdoor competitive
season.
Instrumented treadmill
The motorized instrumented treadmill (ADAL3D-WR,
Medical Development—HEF Tecmachine, Andre
´zieux-
Bouthe
´on, France) used has recently been validated for
sprint use (for details, see Morin et al. 2010). It is mounted
on a highly rigid metal frame fixed to the ground through
four piezoelectric force transducers (KI 9077b, Kistler,
Winterthur, Switzerland), and installed on a specially
engineered concrete slab to ensure maximal rigidity of the
supporting ground. The constant motor torque was set to
160 % of the default torque, i.e. the motor torque necessary
to overcome the friction on the belt due to subject’s body
weight. The default torque was measured by requiring
subjects to stand still and then increasing the driving torque
until observing a movement of the belt greater than 2 cm
over 5 s. This default torque setting as a function of belt
friction is in line with previous motorized-treadmill studies
Eur J Appl Physiol
123
(e.g. Chelly and Denis 2001; Jaskolska et al. 1999; Morin
et al. 2011a,b). Motor torque of 160 % of the default value
was selected after several preliminary measurements
comparing various torques. 160 % allowed subjects to
sprint in a comfortable manner and produce maximal effort
without risking loss of balance. Subjects were tethered by
means of a leather weightlifting belt and a thin stiff rope
(0.6 cm in diameter) rigidly anchored to the wall behind
the subjects by a 0.4-m vertical metal rail. When correctly
attached, subjects were able to lean forward in a typical
crouched sprint-start position with their preferred foot
forward. This starting position was standardized for all
trials. After a 3-s countdown, the treadmill was started and
the treadmill belt began to accelerate as subjects applied a
positive horizontal force.
Sprint mechanics
Mechanical data were sampled at 1,000 Hz continuously
over the sprints, the beginning of the sprint being deter-
mined with a velocity threshold of 0.2 m s
-1
. After
appropriate filtering (Butterworth-type 30 Hz low-pass
filter), instantaneous data of vertical, horizontal and resul-
tant GRF were averaged for each support phase (vertical
force above 30 N) over the 6-s sprints (F
V
,F
H
and F
Tot
,
respectively), and expressed in N and BW. For each 6-s
sprint, performance was described through mean and
maximal running speeds (Vand V-max, respectively).
These data were completed by measurements of the main
step kinematic variables: contact time (t
c
in s), aerial time
(t
a
in s), step frequency (SF in Hz), step length (SL in m)
and swing time (t
swing
), i.e. the time to reposition the limb,
from take-off to touch-down of the same foot.
For each step, the net power output in the horizontal
direction was computed according to Morin et al. (2010)as
P=F
H
V, and expressed in W kg
-1
. As for velocity, mean
and maximal mechanical power outputs were calculated
over the 6-s sprints (Pand P-max, respectively). For each
sprint, the linear F–V relationship was plotted from F
H
(expressed in N kg
-1
) and Vvalues of steps ranging from
the step at maximal F
H
(typically one of the three first
steps) to the step at V
max
, as for Morin et al. 2010). These
individual relationships were summarized through the
theoretical maximal horizontal force that the legs could
produce over one contact phase at null velocity (F
H
0in
Nkg
-1
), and the theoretical maximal velocity of the
treadmill belt that the legs could produce during the same
phase under zero load (V0inms
-1
).
To quantify this F
H
production compared to the F
Tot
production, a ratio of forces (RF in %) was calculated as
the ratio of F
H
to F
Tot
for one contact period (Morin et al.
2011a). This ratio basically represents, for a given support
phase, the percent of the resultant GRF that is applied in
the forward direction. As recently presented by Morin et al.
(2011a), an index of force application/orientation technique
(D
RF
) representing the decrement in RF with increasing
running velocity was computed for each subject as the
slope of the linear RF–velocity relationship calculated from
step-averaged values between the second step and the step
at top speed. A high value of D
RF
(i.e. a flat RF–velocity
relationship), indicates that the systematic linear decrease
in RF with increasing velocity is rather limited, and vice
versa (see for instance the typical comparison of two
individuals in Fig. 2b of Morin et al. 2011a).
Field 100-m performance
The four athletes used spiked shoes and starting blocks
during the field tests, which was not the case of the non-
specialists. The latter subjects used a standard crouched-
position start, similar to that used for the treadmill sprints.
The 100-m sprints were performed individually, and per-
formance was measured with a radar system (Stalker ATS
System, Radar Sales, Minneapolis MN, USA). This device
has been validated and used in previous human sprint
running experiments (e.g. Chelly and Denis 2001;Di
Prampero et al. 2005; Morin and Se
`ve 2011) and measures
the forward running speed of the subject at a sampling rate
of 35 Hz. It was placed on a tripod 10 m behind the sub-
jects at a height of 1 m (corresponding approximately to
the height of subjects’ centre of mass).
To better analyze the 100-m performance, radar speed–
time curves were fitted by a bi-exponential function (Morin
and Se
`ve 2011; Volkov and Lapin 1979):
SðtÞ¼Smax etþtSmaxÞ=s2ðÞðÞ
et=s1ðÞ
hi
ð1Þ
s1 and s2 being, respectively, the time constant for
acceleration and deceleration of this relationship,
determined by iterative computerized solving. Speed–
distance curves were then obtained from these modeled
speed–time curves by simple time-integration of modeled
speed data. For more clarity, and given the high quality of
the bi-exponential fitting of instantaneous radar data (see
for instance Morin and Se
`ve 2011), only the modeled speed
data were analyzed. From these data, maximal running
speed (S-max in m s
-1
) was obtained, as well as the 100-m
time and the corresponding mean 100-m speed (S
100
in
ms
-1
) for each subject. For the four sprinters, 100-m times
were also measured with a pair of photo-cells and a
chronometer triggered by a standard audio signal similar to
those of typical competitions. Last, to describe the
acceleration performance in relation to sports other than
track and field, and with a practical and simple index, the
4-s distance (d
4
in m) was measured as the distance
covered during the first 4 s of the 100-m.
Eur J Appl Physiol
123
Anthropometric measurements
Sub-ischial length (L, cm), referred to as leg length was mea-
sured as the great-trochanter-to-ground distance in a standing
position, measured with 0.5 cm accuracy. To facilitate com-
parison between subjects, the leg length to standing height ratio
(L/H) and body mass index (BMI, kg m
-2
) were used.
Statistical analyses
All data are presented as mean ±SD. After normality
checking by the Shapiro–Wilk test, and in case of normal
distribution, correlations between mechanical and perfor-
mance parameters were tested by means of Pearson’s corre-
lation coefficients. In case of absence of normal distribution,
the Spearman rank test was used to test these correlations. A
Pvalue of 0.05 was accepted as level of significance.
Results
The main field sprint performance (100-m time) recorded
during the experiment was 12.73 ±1.48 s, ranging from
10.35 to 15.03 s. For the world-class and national-level
sprinters, these performances corresponded to 95.8 ±1.6 %
of their personal best times. Table 1shows the main
performance and mechanical variables studied.
The results showed a significant correlation between
mean and maximal power output measured on the treadmill
and the three main 100-m performance variables: S-max,
S
100
and d
4
(Table 2).
All linear F–V regressions were significant (mean r
2
of
0.909, range 0.804–0.982; all P\0.001). We also tested
second degree polynomial regressions to model the F–V
relationship (data not shown), and the mean r
2
value for the
group only slightly increased (to 0.930 ±0.052). Although
this increase was statistically significant, the fact that (1)
linear regressions also gave significant and very high r
2
values, and that (2) our analysis and interpretation of the
mechanical F–V qualities was based on previous works
using a linear approach; we maintained this approach in our
study.
The theoretical maximal horizontal GRF (F
H
0) was not
significantly correlated to any of the performance variables
considered, whereas the maximal theoretical running
velocity V0 was (Table 2). The typical F–V relationships of
Table 1 Main 100-m
performance and mechanical
variables averaged over the
acceleration phase of the sprint
on the instrumented treadmill
(from the second step until top
speed)
Variable Mean (SD) Range
Field 100-m sprint performance
Average 100-m speed (m s
-1
) 7.96 (0.98) 6.65–9.66
Maximal 100-m speed (m s
-1
) 9.32 (1.15) 7.80–11.2
4-s distance (m) 25.1 (3.19) 20.9–31.0
Treadmill sprint kinematics
Contact time (s) 0.147 (0.019) 0.121–0.181
Aerial time (s) 0.094 (0.011) 0.077–0.121
Step frequency (Hz) 4.17 (0.27) 3.80–4.64
Step length (m) 1.41 (0.15) 1.03–1.56
Swing time (s) 0.330 (0.025) 0.297–0.371
Treadmill sprint kinetics
Index of force application technique -0.074 (0.015) -0.093 to -0.042
Horizontal GRF (N) 240 (37) 201–314
Horizontal GRF (BW) 0.322 (0.048) 0.224–0.398
Vertical GRF (N) 1,235 (183) 981–1,515
Vertical GRF (BW) 1.66 (0.15) 1.48–1.85
Resultant GRF (N) 1,263 (184) 1,009–1,549
Resultant GRF (BW) 1.7 (0.15) 1.52–1.90
Vertical GRF at maximal velocity (N) 1,371 (178) 1,109–1,657
Vertical GRF at maximal velocity (BW) 1.85 (0.14) 1.63–2.07
Treadmill force–velocity characteristics and power output
Maximal velocity (m s
-1
) 7.05 (0.91) 5.75–8.66
Maximal power output (W kg
-1
) 22.7 (4.88) 16.0–31.1
Mean power output (W kg
-1
) 18.1 (4.26) 11.1–25.5
Theoretical maximal horizontal force (N kg
-1
) 8.61 (1.09) 6.23–10.7
Theoretical maximal velocity (m s
-1
) 9.85 (1.70) 7.71–14.0
Eur J Appl Physiol
123
the fastest and slowest subjects presented in Fig. 1show
their relatively higher difference in V0 than in F
H
0 values.
These two individuals also strongly differed in terms of
peak power and optimal velocity, as shown by the com-
parison of their power–velocity relationships (Fig. 1).
Concerning GRF production and orientation onto the
ground, Table 3shows that D
RF
index was significantly
correlated to all the performance variables considered,
contrary to F
Tot
, which was only significantly correlated to
S-max (P=0.034). For the components of this resultant
GRF, F
H
was significantly correlated to 100-m perfor-
mance (P\0.05), whereas F
V
was only correlated to
S-max (P=0.039), and not to S
100
or d
4
.
These correlations between sprint performance (mean
100-m speed) and F
Tot
and D
RF
are shown in Fig. 2.
The ability to orient the resultant GRF vector effectively
(i.e. forward) during the acceleration phase on the treadmill
(analyzed through the D
RF
value) strongly differed between
the fastest and slowest individuals tested (Fig. 3). In
addition to being the individuals presenting the extreme
values of 100-m time, they had the highest (-0.042) and
second lowest (-0.091) values of D
RF
of the group.
Contact time and step frequency showed significant and
high correlations (P\0.01) with 100-m performance
(Table 4), which was also the case of the swing ti me
(P\0.05). However, neither aerial time (P[0.88) nor
step length (P[0.21) was related to sprint performance.
Last, BMI (23.2 ±2.2 kg m
-2
) and L/H ratio
(0.522 ±0.014) were not correlated to any of the perfor-
mance variables studied (all P[0.29).
Discussion
To our knowledge, since the work of Weyand et al. (2000),
this is the only study to specifically report experimental
data obtained in a group of subjects ranging from non-
specialists to national-level sprinters, and to a sub-10-s
athlete. Since pioneering works about human sprint per-
formance published in the late 1920s (Best and Partridge
1928; Furusawa et al. 1927) involving very fast runners
(estimated 100-m time of *10.8 s for subject H.A.R.,
probably 1928 Olympian sprinter Henry Argue Russel, in
the study of Furusawa et al. (1927), many studies involved
high-level athletes (e.g. Karamanidis et al. 2011; Mero and
Table 2 Correlations between mechanical variables (rows) of the force–velocity relationship and power output and 100-m performance
variables (columns)
Maximal speed (m s
-1
) Mean 100-m speed (m s
-1
) 4-s distance (m)
Maximal power output 0.863 (<0.01)0.850 (<0.01)0.892 (<0.001)
Average power output 0.810 (<0.01)0.839 (<0.01)0.903 (<0.001)
Theoretical maximal horizontal force F
H
0 0.560 (0.052) 0.447 (0.128) 0.432 (0.14)
Theoretical maximal horizontal velocity V00.819 (<0.01)0.735 (0.011)0.841 (<0.01)
Significant correlations are reported in bold. Values are presented as Pearson’s correlation coefficient (Pvalues in italics)
Fig. 1 Typical linear force–velocity and 2nd degree polynomial
power–velocity relationships obtained from instrumented treadmill
sprint data for the fastest (100-m best time: 9.92 s, 100-m time of
10.35 s during the study: black and dark grey circles) and slowest
(100-m time of 15.03 s during the study: white and light grey circles)
subjects of this study. All linear and 2nd degree polynomial
regressions were significant (r
2
[0.878; all P\0.001)
Table 3 Correlations between mechanical variables of sprint kinetics
measured during treadmill sprints (rows) and 100-m performance
variables (columns)
Maximal
speed (m s
-1
)
Mean 100-m
speed (m s
-1
)
4-s distance
(m)
Index of force
application
technique D
RF
0.875 (\0.01)0.729 (\0.05)0.683 (\0.05)
Horizontal GRF 0.773 (\0.01)0.834 (\0.01)0.773 (\0.05)
Vertical GRF 0.593 (\0.05) 0.385 (0.18) 0.404 (0.16)
Resultant GRF 0.611 (\0.05) 0.402 (0.16) 0.408 (0.16)
Significant correlations are reported in bold. Horizontal, vertical and
resultant GRF data are averaged values for the entire acceleration
phase. Values are presented as Pearson’s correlation coefficient
(Pvalues in italics)
Eur J Appl Physiol
123
Komi 1986) but not truly world-class individuals. The main
results of this study showed a higher importance of the
variables associated with velocity rather than force capa-
bilities (see below). As subjects’ level of 100-m increased,
this was particularly characterized at high running speeds
by the increasing ability to orient the resultant GRF gen-
erated by the lower limbs with a forward incline, i.e. to
produce higher amounts of horizontal net force at each
step, and not by increasing the amount of resultant force
produced.
During the treadmill sprint tests, we found a significant
and clear correlation between 100-m performance and
average or maximal mechanical power normalized to body
mass in the horizontal direction (r
s
[0.810; P\0.01;
Table 2). This was expected from previous findings (e.g.
Cronin and Hansen 2005; Cronin and Sleivert 2005; Harris
et al. 2008; Sleivert and Taingahue 2004), but the present
study had the novelty of reporting mechanical power data
measured during specific running exercise (Morin et al.
2010), contrary to the previously cited protocols in which
power output was assessed during vertical, horizontal or
incline push-offs, or cycling sprints (e.g. Morin et al. 2002).
When focusing on the two mechanical entities composing
power output (i.e. force and velocity) analyzed through the
linear F–V relationships, Table 2and Fig. 1clearly show
that with the increase in overall sprinting ability (i.e. from
non-specialists to national-level sprinters to the world-class
individual tested), the orientation of the F–V relationship
differs more on the velocity axis than on the force axis. The
theoretical maximal horizontal GRF (F
H
0) calculated from
linear F–V relationship was not significantly correlated to
any of the sprint performance variables considered (only a
tendency at P=0.052 with maximal running speed),
Fig. 2 Correlations between sprint performance parameter of mean
100-m running speed and mechanical variables of index of force
application (left panel), and resultant ground reaction force (right
panel) as measured during the 6-s treadmill sprints. r
s
Spearman
correlation coefficient
Fig. 3 Typical RF–velocity linear relationships during the acceler-
ation phase of the treadmill sprint for the fastest (100-m best time:
9.92 s, 100-m time of 10.35 s during the study: black circles) and
slowest (100-m time of 15.03 s during the study: white circles)
subjects of this study. These linear regressions were significant
(r
2
[0.936; P\0.001). Each point represents average values of ratio
of forces and velocity for one contact phase
Table 4 Correlations between mechanical variables of sprint step
kinematics measured during treadmill sprints (rows) and 100-m per-
formance variables (columns)
Maximal speed
(m s
-1
)
Mean 100-m
speed (m s
-1
)
4-s distance
(m)
Contact time -0.852 (\0.01) -0.751 (\0.01)-0.775 (\0.01)
Aerial time -0.018 (0.95) 0.773 (0.88) 0.002 (0.99)
Swing time -0.654 (\0.05)-0.630 (\0.05)-0.670 (\0.05)
Step
frequency
0.897 (\0.01)0.893 (\0.01)0.935 (\0.001)
Step length 0.363 (0.21) 0.337 (0.24) 0.212 (0.46)
Significant correlations are reported in bold. All mechanical data are
averaged values for the entire acceleration phase. Values are pre-
sented as Pearson’s correlation coefficient (Pvalues in italics)
Eur J Appl Physiol
123
whereas the theoretical maximal velocity (V0) was (all
P\0.011). When comparing the two extreme individuals
in terms of 100-m time (10.35 and 15.03 s during the field
test of the present study, Fig. 1), their V0 values (14.0 vs.
8.28 m s
-1
, respectively) differed much more than their
F
H
0 values (8.47 vs. 6.82 N kg
-1
, respectively). When
expressed as force normalized per kg body mass, F
H
dimensionally and conceptually corresponds to a forward
acceleration that would be the acceleration of the runner
should no resistive force applied on him. Indeed, the max-
imal values reported in Fig. 1for the fastest sprinter, i.e.
*6ms
-2
, are close to those reported by di Prampero et al.
(2005) using radar data obtained during field sprints. This
result could be interpreted as a higher importance of the
relative capability of the neuromuscular system to keep on
producing relatively high levels of horizontal force at high
and very high velocities, rather than to produce very high
levels of maximal force. Since sprint running acceleration
depends on mechanical power, and in turn on both force and
velocity outputs, further studies should investigate whether
and how sprint running performance could be maximized
through an optimal combination of force and velocity
capabilities, i.e. whether an ‘‘individually optimal F–V
profile’’ exists, as recently shown for jumping exercises by
Samozino et al. (2012).
This F–V profile characterized by substantially greater
horizontal force production at faster velocities can be
explained by the ability to produce a greater resultant force
magnitude at rapid movement velocities, which may be
partly related to favorable intrinsic muscular properties and
muscle fiber type (i.e. high proportion of fast-twitch fibers
(Baguet et al. 2011; Gollnick and Matoba 1984), but also
by the ability to orient the resultant force vector horizon-
tally during sprint acceleration. Indeed, we observed a high
and significant correlation between sprint performance and
the ability to produce net horizontal force per unit BW F
H
(Table 3). Given the much poorer correlation obtained with
resultant force production F
Tot
(only correlated to S-max,
and not to S
100
or d
4
; Table 3), the better ability to produce
and apply high F
H
onto the ground in skilled sprinters
comes mostly from a greater ability to orient the resultant
force vector forward during the entire acceleration phase,
despite increasing velocity. This is illustrated by the index
of force application technique D
RF
, which was significantly
correlated to the three main performance parameters tested
(all P\0.012; Table 3; Fig. 2).
As shown in the typical comparison between the fastest
and slowest individuals of this study (Fig. 3), the RF–
velocity linear relationship is overall less steep as level of
performance increases from non-specialist to world-class
levels. This means that sprint performance was related to
the ability to maintain a high RF with increasing speed
during the acceleration, which is illustrated by a high D
RF
index, as proposed by Morin et al. (2011a). The correlation
between D
RF
and 100-m performance (S
100
) recently
shown by Morin et al. (2011a) in a group of sportsmen
including three regional-level sprinters finds here a clear
confirmation with a more heterogeneous population,
including top-level sprinters and a sub-10-s individual.
In their recent study, these authors found that contrary to
D
RF
(r=0.779; P\0.01, Table 2of Morin et al. 2011a),
the resultant force magnitude while sprinting on the
treadmill (F
Tot
) was not related to S
100
(r=0.411;
P=0.19, Table 2of Morin et al. 2011a). The present
results almost exactly match those previously reported:
F
Tot
was not significantly related to S
100
when pooling the
data of all the subjects tested (r
s
=0.402; P=0.16,
Table 3), whereas D
RF
was (r
s
=0.729; P=0.012,
Table 3). Furthermore, the only performance parameter
significantly related to the vertical or resultant force pro-
duction was top speed (Table 3). The significant correla-
tion found here between F
V
or F
Tot
and field 100-m S-max
(r
s
=0.593 and 0.611, respectively; P\0.05) is consis-
tent with the results of Morin et al. (2011a), and the
hypothesis initially put forward by Weyand et al. (2000)
that top speed reached in the field is related to the ability to
produce high amounts of vertical GRF per unit BW (which
these authors measured at top-running velocity on the
treadmill).
Consistently with the importance of velocity production
capabilities earlier discussed, step kinematics showed a
significant positive correlation between step frequency and
sprint performance (Table 4), which was not the case of
step length. The higher SF measured over the 6-s sprint in
subjects with high sprinting skills resulted from a lower t
c
and t
swing
with similar t
a
(significant negative correlations
with all P\0.05 for the two former variables, no signifi-
cant correlations for the latter). The significant correlation
between t
swing
and sprint performance (acceleration, mean
and maximal 100-m speed) is contradicting the hypothesis
of Weyand et al. (2000) that the time to reposition the
limbs from one foot contact to another is not a key factor of
sprint performance. This may be due to the fact that t
swing
was considered here over the entire acceleration phase on
the treadmill, whereas Weyand et al. only considered the
steps at top speed. This fundamental difference in the
approaches (only top velocity phase vs. entire acceleration)
likely explains why the present study is different and
complementary with Weyand et al.’s one.
These step kinematics results also overall support the
idea that high running speeds are achieved through reduced
contact times (e.g. McMahon and Green 1979; Weyand
et al. 2000,2010). However, it must be kept in mind that
(1) the data presented here were obtained on a sprint
treadmill, on which lower top speeds are reached compared
to field conditions (Morin and Se
`ve 2011), and (2) they are
Eur J Appl Physiol
123
averaged data for the entire 6-s sprints, and not only data
averaged for the steps around top velocity. The same
applied to values of SL, which were much lower than what
is typically measured at top speed on the field (Table 1).
That said, we wanted our analysis to focus on the entire
acceleration phase of a sprint, and not only to the very
specific top-velocity phase hitherto studied by colleagues
(e.g. Bundle et al. 2003; Weyand et al. 2000). Although the
reliance on step kinematics (and the debated relative
importance of SL and SF) has recently been shown as
highly individual among elite athletes (Salo et al. 2011),
the present study shows in a heterogeneous group a clear
tendency (Table 4) toward the importance of t
c
,t
swing
and
SF variables.
Last, no correlation was found between the anthropo-
metric variables studied and 100-m performance (all
P[0.29). By comparing body mass and stature values for
the world’s fastest performers at track racing distances
from 100- to 10,000-m, Weyand and Davis (2005)
observed that BMI increased as running distance
decreased, which would allow sprinters to reach the
required support force earlier discussed (Watts et al. 2011;
Weyand and Davis 2005). We did not find such a tendency
in the group studied, and furthermore, it is worth noting
that the BMI of the world-class individual tested
(22.3 kg m
-2
) was consistently lower than the value (close
to 24.3 ±0.3 kg m
-2
) reported by Weyand and Davis
(2005). It will be of interest to follow whether, in this
young sprinter, an increase in BMI related to strength
training is to be associated with improved performances.
Concerning the L/H ratio, the present data do not seem to
indicate that long limbs could be a key factor by allowing
extended stride length and providing greater forward pro-
pulsion (van Ingen Schenau et al. 1994). That said,
according to the recent paper of Bejan et al. (2010), the
higher ratio observed in populations living in or originating
from Western Africa accounts for the domination of these
populations on sprint events. This was also proposed by
Rahmani et al. (2004), who compared Italian and Sene-
galese high-level sprinters. The ratio of the world-class
individual tested was the second highest of the group
(0.538 vs. 0.520 ±0.014 for the rest of the group). In
association with the functional abilities above discussed, it
is likely that his long limbs would provide him a further
advantage in sprinting.
One limitation of the present study is that sprint running
mechanics were investigated during sprints performed on an
instrumented treadmill, and not overground. The literature
is not clear as to the fundamental differences between these
two conditions (e.g. Frishberg 1983; Kivi et al. 2002).
However, the treadmill measurements performed here
aimed at quantifying subjects’ ability to apply/orient force
onto the ground while sprinting, as opposed to reproducing
exact field sprint conditions. Consequently, despite a lower
performance on the treadmill, but given the significant
correlations observed between field and treadmill sprint
performances (Morin and Se
`ve 2011), we can reasonably
assume that the inter-individual differences observed in
physical and technical capabilities did not fundamentally
differ between treadmill and track conditions. Finally, we
think that the advantage and novelty of being able to con-
tinuously measure GRF and RF and compute D
RF
over the
entire acceleration phase of a maximal sprint in such a
population outweighs the issue of lower sprint performance.
To conclude, this study including national- and world-
class level athletes as well as non-specialists provided
qualitative information toward a better understanding of
the biomechanical correlates of sprint running perfor-
mance, and confirmed recent hypotheses of the literature.
The main result of the present study is that a higher level of
acceleration and overall 100-m performance is mainly
associated with (1) a ‘‘velocity-oriented’’ force–velocity
profile, likely explained by (2) a higher ability to apply the
resultant GRF vector with a forward orientation over the
acceleration, and finally (3) a higher step frequency caused
by a shorter contact time. Contrastingly, resultant GRF
magnitude was not related to acceleration and overall
100-m performance, but only to top running speed. Further
studies should focus on the necessity, effectiveness and
practical feasibility of training programs/exercises that
could develop the key variables of sprint performance put
forward, and on the neuromuscular origin of the macro-
scopic results obtained here about the integrative variables
of lower limbs force and velocity outputs.
Acknowledgments We are very grateful to Pierre Carraz and the
athletes of the AS Aix-les-Bains Track and Field club for their
involvement in the protocol. We also thank Johan Cassirame
(Matsport, France), Thibault Lussiana and Nicolas Tordi (Centre
d’Optimisation de la Performance Sportive COPS, Universite
´de
Franche-Comte
´, France) and Mathieu Lacome and Olivier Rambaud
for their precious help in field and laboratory data collection. We also
gratefully thank the two anonymous reviewers for their supportive
and constructive comments.
Conflict of interest None.
References
Arsac LM, Locatelli E (2002) Modeling the energetics of 100-m
running by using speed curves of world champions. J Appl
Physiol 92:1781–1788
Baguet A, Everaert I, Hespel P, Petrovic M, Achten E, Derave W
(2011) A new method for non-invasive estimation of human
muscle fiber type composition. PLoS ONE 6:e21956
Bejan A, Jones EC, Charles JD (2010) The evolution of speed in
athletics: why the fastest runners are black and swimmers white.
Int J Des Nat 5:1–13
Eur J Appl Physiol
123
Beneke R, Taylor MJD (2010) What gives Bolt the edge—A.V. Hill
knew it already! J Biomech 43:2241–2243
Best CH, Partridge RC (1928) The equation of motion of a runner,
exerting a maximal effort. Proc R Soc B 103:218–225
Bret C, Messonnier L, Nouck Nouck JM, Freund H, Dufour AB,
Lacour J-R (2003) Differences in lactate exchange and removal
abilities in athletes specialised in different track running events
(100 to 1500 m). Int J Sports Med 24:108–113
Bundle MW, Hoyt RW, Weyand PG (2003) High-speed running
performance: a new approach to assessment and prediction.
J Appl Physiol 95:1955–1962
Chelly SM, Denis C (2001) Leg power and hopping stiffness:
relationship with sprint running performance. Med Sci Sports
Exerc 33:326–333
Cormie P, McGuigan MR, Newton RU (2011) Developing maximal
neuromuscular power: part 2—training considerations for
improving maximal power production. Sports Med 41:125–146
Coyle EF (2005) Improved muscular efficiency displayed as Tour de
France champion matures. J Appl Physiol 98:2191–2196
Cronin J, Hansen KT (2005) Strength and power predictors of sports
speed. J Strength Cond Res 19:349–357
Cronin J, Sleivert G (2005) Challenges in understanding the influence
of maximal power training on improving athletic performance.
Sports Med 35:213–234
Di Prampero PE, Fusi S, Sepulcri L, Morin JB, Belli A, Antonutto G
(2005) Sprint running: a new energetic approach. J Exp Biol
208:2809–2816
Dorel S, Couturier A, Lacour J-R, Vandewalle H, Hautier C, Hug F
(2010) Force–velocity relationship in cycling revisited: benefit of
two-dimensional pedal forces analysis. Med Sci Sports Exerc
42:1174–1183
Frishberg BA (1983) An analysis of overground and treadmill
sprinting. Med Sci Sports Exerc 15:478–485
Furusawa K, Hill AV, Parkinson JL (1927) The energy used in
‘‘sprint’’ running. Proc R Soc B 102(713):43–49
Gollnick PD, Matoba H (1984) The muscle fiber composition of
skeletal muscle as a predictor of athletic success. An overview.
Am J Sports Med 12:212–217
Harris NK, Cronin J, Hopkins WG, Hansen KT (2008) Relationship
between sprint times and the strength/power outputs of a
machine squat jump. J Strength Cond Res 22:691–698
Hirvonen J, Rehunen S, Rusko H, Harkonen M (1987) Breakdown of
high-energy phosphate compounds and lactate accumulation
during short supramaximal exercise. Eur J Appl Physiol Occup
Physiol 56:253–259
Hunter JP, Marshall RN, McNair PJ (2005) Relationships between
ground reaction force impulse and kinematics of sprint-running
acceleration. J Appl Biomech 21:31–43
Jaskolska A, Goossens P, Veenstra B, Jaskolski A, Skinner JS (1999)
Treadmill measurement of the force–velocity relationship and
power output in subjects with different maximal running
velocities. Sports Med Train Rehab 8:347–358
Jones AM (2006) The physiology of the world record holder for the
women’s marathon. Int J Sports Sci Coach 1:101–116
Karamanidis K, Albracht K, Braunstein B, Catala MM, Goldmann JP,
Bruggemann GP (2011) Lower leg musculoskeletal geometry
and sprint performance. Gait Posture 34:138–141
Kivi DM, Maraj BK, Gervais P (2002) A kinematic analysis of
highspeed treadmill sprinting over a range of velocities. Med Sci
Sports Exerc 34:662–666
Lacour J-R, Messonnier L, Bourdin M (2009) Physiological correlates
of performance. Case–study of a world-class rower. Eur J Appl
Physiol 106:407–413
Lucia A, Olivan J, Bravo J, Gonzalez-Freire M, Foster C (2008) The
key to top-level endurance running performance: a unique
example. Br J Sports Med 42:172–174
McMahon TA, Green PR (1979) The influence of track compliance
on running. J Biomech 12:893–904
Mero A, Komi PV (1986) Force-, EMG-, and elasticity–velocity
relationships at submaximal, maximal and supramaximal running
speeds in sprinters. Eur J Appl Physiol Occup Physiol 55:553–561
Moravec P, Ruzicka J, Susanka P, Dostal E, Kodejs M, Nosek M (1988)
The 1987 International Amateur Athletic Federation/IAAF Sci-
entific Project Report: time analysis of the 100 meters events at the
II World Championships in Athletics. New Stud Athl 3:61–96
Morin JB, Se
`ve P (2011) Sprint running performance: comparison
between treadmill and field conditions. Eur J Appl Physiol
111:1695–1703
Morin JB, Hintzy F, Belli A, Grappe F (2002) Force–velocity
relationships and sprint running performances in trained athletes.
Sci Sport 17:78–85 (French)
Morin JB, Samozino P, Bonnefoy R, Edouard P, Belli A (2010) Direct
measurement of power during one single sprint on treadmill.
J Biomech 43:1970–1975
Morin JB, Edouard P, Samozino P (2011a) Technical ability of force
application as a determinant factor of sprint performance. Med
Sci Sports Exerc 43:1680–1688
Morin JB, Samozino P, Edouard P, Tomazin K (2011b) Effect of
fatigue on force production and force application technique
during repeated sprints. J Biomech 44:2719–2723
Rahmani A, Locatelli E, Lacour JR (2004) Differences in morphology
and force/velocity relationship between Senegalese and Italian
sprinters. Eur J Appl Physiol 91:399–405
Roberts TJ, Scales JA (2002) Mechanical power output during
running accelerations in wild turkeys. J Exp Biol 205:1485–1494
Salo AI, Bezodis IN, Batterham AM, Kerwin DG (2011) Elite
sprinting: are athletes individually step frequency or step length
reliant? Med Sci Sports Exerc 43:1055–1062
Samozino P, Rejc E, Di Prampero PE, Belli A, Morin JB (2012)
Optimal force–velocity profile in ballistic movements. Altius:
citius or fortius? Med Sci Sports Exerc 44:313–322
Sleivert G, Taingahue M (2004) The relationship between maximal
jump-squat power and sprint acceleration in athletes. Eur J Appl
Physiol 91:46–52
van Ingen Schenau GJ, de Koning JJ, de Groot G (1994) Optimisation
of sprinting performance in running, cycling and speed skating.
Sports Med 17:259–275
Volkov NI, Lapin VI (1979) Analysis of the velocity curve in sprint
running. Med Sci Sports 11:332–337
Ward-Smith AJ, Radford PF (2000) Investigation of the kinetics of
anaerobic metabolism by analysis of the performance of elite
sprinters. J Biomech 33:997–1004
Watts AS, Coleman I, Nevill A (2011) The changing shape
characteristics associated with success in world-class sprinters.
J Sports Sci. doi:10.1080/02640414.2011.588957
Weyand PG, Davis JA (2005) Running performance has a structural
basis. J Exp Biol 208:2625–2631
Weyand PG, Sternlight DB, Bellizzi MJ, Wright S (2000) Faster top
running speeds are achieved with greater ground forces not more
rapid leg movements. J Appl Physiol 89:1991–1999
Weyand PG, Bundle MW, McGowan CP, Grabowski A, Brown MB,
Kram R, Herr H (2009) The fastest runner on artificial legs:
different limbs, similar function? J Appl Physiol 107:903–911
Weyand PG, Sandell RF, Prime DNL, Bundle MW (2010) The
biological limits to running speed are imposed from the ground
up. J Appl Physiol 108:950–961
Eur J Appl Physiol
123
