Can Neuromuscular Fatigue Explain Running Strategies and Performance in Ultra-Marathons? The Flush Model

Abstract

While the industrialized world adopts a largely sedentary lifestyle, ultra-marathon running races have become increasingly popular in the last few years in many countries. The ability to run long distances is also considered to have played a role in human evolution. This makes the issue of ultra-long distance physiology important. In the ability to run multiples of 10 km (up to 1000 km in one stage), fatigue resistance is critical. Fatigue is generally defined as strength loss (i.e. a decrease in maximal voluntary contraction [MVC]), which is known to be dependent on the type of exercise. Critical task variables include the intensity and duration of the activity, both of which are very specific to ultra-endurance sports. They also include the muscle groups involved and the type of muscle contraction, two variables that depend on the sport under consideration. The first part of this article focuses on the central and peripheral causes of the alterations to neuromuscular function that occur in ultra-marathon running. Neuromuscular function evaluation requires measurements of MVCs and maximal electrical/magnetic stimulations; these provide an insight into the factors in the CNS and the muscles implicated in fatigue. However, such measurements do not necessarily predict how muscle function may influence ultra-endurance running and whether this has an effect on speed regulation during a real competition (i.e. when pacing strategies are involved). In other words, the nature of the relationship between fatigue as measured using maximal contractions/stimulation and submaximal performance limitation/regulation is questionable. To investigate this issue, we are suggesting a holistic model in the second part of this article. This model can be applied to all endurance activities, but is specifically adapted to ultra-endurance running: the flush model. This model has the following four components: (i) the ball-cock (or buoy), which can be compared with the rate of perceived exertion, and can increase or decrease based on (ii) the filling rate and (iii) the water evacuated through the waste pipe, and (iv) a security reserve that allows the subject to prevent physiological damage. We are suggesting that central regulation is not only based on afferent signals arising from the muscles and peripheral organs, but is also dependent on peripheral fatigue and spinal/supraspinal inhibition (or disfacilitation) since these alterations imply a higher central drive for a given power output. This holistic model also explains how environmental conditions, sleep deprivation/mental fatigue, pain-killers or psychostimulants, cognitive or nutritional strategies may affect ultra-running performance.

Full text

Terms and Conditions for Use of PDF
The provision of PDFs for authors' personal use is subject to the following Terms & Conditions:
The PDF provided is protected by copyright. All rights not specifically granted in these Terms & Conditions are expressly
reserved. Printing and storage is for scholarly research and educational and personal use. Any copyright or other notices
or disclaimers must not be removed, obscured or modified. The PDF may not be posted on an open-access website
(including personal and university sites).
The PDF may be used as follows:
• to make copies of the article for your own personal use, including for your own classroom teaching use (this includes
posting on a closed website for exclusive use by course students);
• to make copies and distribute copies (including through e-mail) of the article to research colleagues, for the personal use
by such colleagues (but not commercially or systematically, e.g. via an e-mail list or list serve);
• to present the article at a meeting or conference and to distribute copies of such paper or article to the delegates
attending the meeting;
• to include the article in full or in part in a thesis or disser tation (provided that this is not to be published commercially).
This material is the copyright of the original publisher.
Unauthorised copying and distribution is prohibited.

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
Can Neuromuscular Fatigue Explain
Running Strategies and Performance
in Ultra-Marathons?
The Flush Model
Guillaume Y. Millet
1,2
1 Universite
´de Lyon, F-42023, Saint-Etienne, France
2 Inserm U1042, Grenoble, F-38000, France
Contents
Abstract................................................................................. 489
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
2. Neuromuscular Alterations in Ultra-Marathon Running . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
2.1 Central Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
2.2 Alterations at the Muscle Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
3. The Flush Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
3.1 Power Output Change in Ultra-Marathon Runners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
3.2 Description of the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
3.3 Feed-Forward and Feedback Mechanisms Influence the Filling Rate. . . . . . . . . . . . . . . . . . . . . . 497
3.4 Apart from the Filling Rate, Which Factors Influence the Quantity of Water?. . . . . . . . . . . . . . . . 499
3.5 The Waste Pipe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
3.6 The Security Reserve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
4. Conclusion ........................................................................... 502
Abstract While the industrialized world adopts a largely sedentary lifestyle, ultra-
marathon running races have become increasingly popular in the last few
years in many countries. The ability to run long distances is also considered to
have played a role in human evolution. This makes the issue of ultra-long
distance physiology important. In the ability to run multiples of 10 km (up to
1000 km in one stage), fatigue resistance is critical. Fatigue is generally de-
fined as strength loss (i.e. a decrease in maximal voluntary contraction
[MVC]), which is known to be dependent on the type of exercise. Critical task
variables include the intensity and duration of the activity, both of which are
very specific to ultra-endurance sports. They also include the muscle groups
involved and the type of muscle contraction, two variables that depend on the
sport under consideration. The first part of this article focuses on the central
and peripheral causes of the alterations to neuromuscular function that occur
in ultra-marathon running. Neuromuscular function evaluation requires
measurements of MVCs and maximal electrical/magnetic stimulations; these
provide an insight into the factors in the CNS and the muscles implicated in
fatigue. However, such measurements do not necessarily predict how muscle
REVIEW ARTICLE Sports Med 2011; 41 (6): 489-506
0112-1642/11/0006-0489/$49.95/0
ª2011 Adis Data Information BV. All rights reserved.

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
function may influence ultra-endurance running and whether this has an ef-
fect on speed regulation during a real competition (i.e. when pacing strategies
are involved). In other words, the nature of the relationship between fatigue
as measured using maximal contractions/stimulation and submaximal per-
formance limitation/regulation is questionable. To investigate this issue, we are
suggesting a holistic model in the second part of this article. This model can be
applied to all endurance activities, but is specifically adapted to ultra-endurance
running: the flush model. This model has the following four components: (i) the
ball-cock (or buoy), which can be compared with the rate of perceived exertion,
and can increase or decrease based on (ii) the filling rate and (iii) the water
evacuated through the waste pipe, and (iv) a security reserve that allows the
subject to prevent physiological damage. We are suggesting that central reg-
ulation is not only based on afferent signals arising from the muscles and peri-
pheral organs, but is also dependent on peripheral fatigue and spinal/supraspinal
inhibition (or disfacilitation) since these alterations imply a higher central drive
for a given power output. This holistic model also explains how environmental
conditions, sleep deprivation/mental fatigue, pain-killers or psychostimulants,
cognitive or nutritional strategies may affect ultra-running performance.
1. Introduction
More than 30 000 articles have been publish-
ed about fatigue. Limiting keywords to ‘muscle’
and ‘fatigue’ still gave us more than 12 000 arti-
cles. It is known that the magnitude and aetiology
of fatigue depend on the exercise under con-
sideration.
[1]
Critical task variables include the
muscle activation pattern, the type of muscle
group involved and, the type of muscle con-
traction. However, the intensity and dura-
tion of activity are probably among the most
important factors. This article focuses on ultra-
endurance running exercises, the so-called ultra-
marathons.
Throughout the world (e.g. in the US, Europe,
Japan, Korea, South Africa), ultra-marathons have
become increasingly popular in the last few years.
For example, Hoffman et al.
[2]
recently analysed
the participation in 161 km ultra-marathons in
North America and showed that the number of
finishes increased exponentially over the period
1977–2008 through a combination of increases
in the number of participants, average annual
number of races completed by each individual,
and number of races organized every year. It is
considered that more than 30 000 runners took
part in at least one ultra-marathon in France in
2009. There is no consensus about the definition
of contemporary ultra-marathons; some authors
consider it to be any distance greater than a
marathon, while for others, it is any event that
exceeds 4 hours
[3]
or 6 hours
[4]
in duration. Ultra-
marathons can last for up to 40 hours or even
several days (e.g. 6-day races) and are basically of
two types: (i) ultra-marathons performed on a
mostly flat road (24 hours, 100 km); and (ii) those
run on varying terrains (e.g. 100 miles in the US).
Contrary to what is usually claimed, ultra-marathon
running is not new; the Six-Day Professional
Pedestrian Races in London and New York have
existed since the 1880s.
[5]
Importantly, the ability
to run, rather than only walk, over long distances
(i.e. without fatigue) may have played a role in
human evolution.
[6]
For example, it has been sug-
gested that endurance running may have helped
hominids to exploit protein-rich resources. Thus,
while endurance running is now primarily a form
of recreation, its roots may be as ancient as the
origin of the human genus.
[6]
This type of extreme
event can also be seen as a testbed for ideas on
how some people manage to perform physical
feats at which others can only marvel.
[7]
Several models of fatigue have been proposed in
the literature. For example, Abbiss and Laursen
[8]
reviewed the following eight different models that
490 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
may be applied to prolonged cycling: cardiovascular/
anaerobic, energy supply/energy depletion, neuro-
muscular fatigue, muscle trauma, biomechanical,
thermoregulatory, psychological/motivational and
the central governor model. All of these are
interrelated. For example, changes in biomecha-
nical patterns may be both the cause and con-
sequence of neuromuscular fatigue, which is also
influenced by modifications in cardiovascular/
anaerobic metabolism, muscle trauma and thermal
conditions, and all potentially associated with the
central governor. Similarly, depending on the
authors and the scientific field, central fatigue has
been presented as a decrease in percentage max-
imal voluntary activation (%VA),
[9]
neurobiolo-
gical modifications in the brain,
[10]
a modification
of motor control
[11]
or alterations in cognitive
function.
[12]
In exercise physiology, most published articles
have defined fatigue as strength loss (i.e. a de-
crease in maximal voluntary contraction [MVC]).
Strength loss in the fatigued state is multifactorial
and is generally divided into central (i.e. above
the neuromuscular junction) and peripheral
(muscular), these two origins being interdependent
on the mediation of peripheral afferences. The
central/peripheral distinction was already pro-
posed by Bainbridge in 1931.
[13]
In this context,
central fatigue is an altered ability of the CNS to
recruit motor units at a higher discharge rate than
the frequency of tetanic fusion. In other words, a
decrease in maximal voluntary activation (i.e.
central fatigue) might be due to a de-recruitment
of motor units and/or a decrease of the discharge
frequency beyond the frequency of tetanic fusion,
both factors leading to force decline. Central fa-
tigue is variably implicated in total fatigue. It has
been shown that prolonged exercise is associated
with a large decrease in %VA, especially with
running.
[14]
However, the role of central fatigue
and its supraspinal and spinal components in the
cessation of exercise (if the intensity is fixed) or
in performance (if the intensity is self-chosen) is
not clear. The problem is further complicated by
the fact that (i) environmental conditions such as
hypoxia and hyperthermia may exacerbate central
fatigue or perceived exertion,
[15,16]
two different
forms of central alterations; and (ii) that mental
fatigue –another type of central alteration –has
been demonstrated to play a role in performance
limitation.
[17]
Although central fatigue is of great
importance, this does not imply that peripheral
fatigue is unimportant. How peripheral changes,
central fatigue, environmental conditions and a
runner’s strategies (whether cognitive, nutritional
or tactical) affect ultra-marathon performance
and regulation of speed during a race has never
been considered. Thus, the first aim of this article
is to review the central and peripheral factors that
might influence strength loss during very pro-
longed running exercise.
Describing these central and peripheral fac-
tors is essential but it does not predict how they
affect submaximal muscle function during ultra-
endurance running or how they influence speed
regulation during a real competition (i.e. including
pacing strategies). This poses the question of the
relationship between fatigue evidenced by measures
taken during maximal contractions/stimulation
and performance limitation/regulation. The second
aim of this article is then to propose a model that
integrates these different parameters, including
central and peripheral fatigue, which may help us
to understand pacing strategies and performance
during ultra-marathons. While the model is partic-
ularly well adapted to ultra-marathons, it can be
applied to any type of endurance performance.
2. Neuromuscular Alterations in
Ultra-Marathon Running
The alterations in neuromuscular function
after prolonged running, cycling and skiing were
reviewed by Millet and Lepers in 2004.
[14]
They
focused on the origin of muscle fatigue after
prolonged exercises lasting from 30 minutes to
several hours. The authors showed that the knee
extensors isometric strength loss increased in a
non-linear way with exercise duration when run-
ning for longer than 2 hours. Since then, several
articles have been published on this topic.
[18-23]
As shown in figure 1, the tendency toward no
further decrease in knee extensor strength with
increasing in running duration is confirmed.
Less is known about the decrease in peak
power after prolonged running. Nevertheless, it
The Flush Model of Fatigue 491
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
has been reported that the decrease in counter-
movement jump performance was around 45–60%
of the knee extensors isometric MVC decrease after
prolonged running.
[19,24,25,31]
Similarly, Lepers
et al.
[24]
reported that isokinetic strength loss was
smaller when measured over a concentric contrac-
tion compared with the eccentric or isometric
mode.
2.1 Central Contribution
Methods such as the twitch interpolation
technique, the ratio of electromyogram (EMG)
signal during MVC normalized to the M-wave
response (Mmax) or the comparison of the forces
achieved with voluntary and electrically evoked
contractions, have systematically shown that cen-
tral fatigue largely contributes to muscle fatigue
during long distance running.
[20,22,25-27,32]
It is
known that the decrease in central activation oc-
curring during exercise can be caused by several
factors at the spinal (motoneurone properties,
afferent input) and/or supraspinal levels.
[9,33,34]
A
few studies have measured the changes in strength
of muscles not involved in the exercise to further
explore the origin of the lower central drive post-
exercise.
[20,27,32]
It was hypothesized that a loss of
grip strength after running would be a good in-
dicator of supraspinal fatigue but no consistent
changes were observed in grip strength following
running exercise. Thus, this measurement did not
allow any conclusion of the existence or the absence
of supraspinal fatigue after prolonged running
because selective supraspinal fatigue may have
occurred. Ohta et al.
[35]
investigated biochemical
modifications during a 24-hour run and from
indirect measurements such as serum serotonin
and free tryptophan levels, they suggested that
this type of exercise induces some supraspinal
fatigue. It has been suggested for years that the
accumulation of serotonin in several brain regions
contributes to the development of fatigue during
prolonged exercise.
[36]
This was thought to be due
to an increase in the concentration ratio of free
tryptophan to branched-chain amino acids because
(i) branched-chain amino acids are oxidized; and
(ii) higher plasma free fatty acids during pro-
longed exercise cause a parallel increase in free
tryptophan since the free fatty acids displace
tryptophan from their usual binding sites on al-
bumin.
[36]
This in turn increases the concentra-
tion of free tryptophan (the serotonin precursor)
in the brain. Meeusen et al.
[10]
went further sug-
gesting that other neurotransmitters such as
dopamine probably also play a significant role
in supraspinal fatigue. Nevertheless, to the best
of our knowledge, no study has clearly shown
any evidence of central fatigue (e.g. a depressed
%VA) with an increased serotonin/dopamine
ratio or other biochemical changes in the CNS.
Meeusen et al.
[10]
acknowledged that fatigue is
probably due to a complex interaction between
45
Knee extensors strength loss (%MVC)
40
35 *
30
25
20
15
0 500 1000
Duration (min)
1500 2000 2500
Fig. 1. Relationship between strength loss in the knee extensors expressed as a percentage of maximal voluntary contraction (%MVC) at rest
and duration of running exercise.
[18,19,23-30] *
Indicates the value is from unpublished observations.
492 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
central and peripheral mechanisms. It is worth
noting that we recently observed a significant cor-
relation between %VA changes during a 24-hour
treadmill run for plantar flexor and knee extensor
muscles,
[26]
which could be indicative of a com-
mon supraspinal mechanism regulating the neural
drive to the working muscle. Another possibility
is that hyperventilation lowers the arterial carbon
dioxide tension and blunts the increase in cere-
bral blood flow, which can lead to an inadequate
oxygen delivery to the brain and contribute to the
development of fatigue.
[37]
Nevertheless, this is
less likely to occur during ultra-marathons, be-
cause of the relatively low level of ventilation.
While some central activation deficit has been
observed for knee extensor muscles in cycling,
[38]
there is a lower level of central fatigue after ac-
tivities that result in less muscle damage than do
running.
[39]
When marathon skiing
[40]
and 30-km
running
[27]
in similar competitive conditions and
duration were compared, the decrease in %VA
was more pronounced for running than for skiing.
Millet and Lepers
[14]
suggested that this result
indicated spinal modulation rather than cortical
alteration after the running exercise. Data from
reflex measurements, such as the Hoffmann
reflex (Hmax), provide interesting insights into
the origin of central fatigue. The Hmax/Mmax
ratio as been used as an indicator of motoneu-
ronal excitability, and more generally to evidence
modulations of neural drive at the spinal level.
[41]
This index was found to decline during a 24-hour
treadmill run and was correlated with decreases
in MVC and %VA,especiallyattheendoftheex-
ercise (personal observation). This finding con-
curs with those of Racinais et al.
[20]
who reported
depressed H-reflexes after a 90-minute run. This
could be due to reduced motoneurone excitability
or pre-synaptic inhibition. In both cases, in-
hibitory mechanisms could be limiting muscle
force production. Such inhibitory action may re-
sult from thin afferent fibre (group III–IV) sig-
nalling, which may have been sensitized by the
production of pro-inflammatory mediators, pro-
duced during prolonged exercise.
[42-46]
So, while
supraspinal fatigue may play a role in reduced
neural drive after prolonged exercises, it can be
suggested that spinal adaptation, such as inhibi-
tion from type III and IV group afferents or dis-
facilitation from muscle spindles contributes to
this reduced central drive. Group III–IV afferent
fibres may also contribute to the submaximal out-
put from the motor cortex (see Taylor et al.
[47]
).
Taken together, this suggests that high central
fatigue due to prolonged running is not due solely
to CNS biochemical changes but that afferent fi-
bres are probably involved. The twitch inter-
polation technique at the peripheral nerve does
not allow discrimination between central fatigue
originating from a supraspinal site and/or from
the spinal level. Researchers at the Prince of
Wales Medical Research Institute in Sydney (e.g.
Todd et al.
[48]
) have measured supraspinal deficit
by superimposing magnetic stimulations of the
motor cortex to voluntary contractions. Recently,
this method has been used to demonstrate the
existence of supraspinal %VA alteration in the
quadriceps after cycling exercise and prolonged
MVCs.
[49,50]
Future studies should use this new
method to further investigate central fatigue after
ultra-distance running exercises.
2.2 Alterations at the Muscle Level
Central fatigue alone cannot explain the entire
strength loss after prolonged running exercises.
Alterations of neuromuscular propagation, failure
of excitation-contraction coupling and mod-
ifications in the intrinsic capability of force pro-
duction may also be involved. To the best of our
knowledge, there has been no measure of change
in action potential conduction velocity using
high-density EMG after prolonged running ex-
ercise. Information about the propagation of the
action potentials can then only be deduced from
changes in the M-wave characteristics. This is
problematic,
[51]
especially in the case of ultra-
marathons, since muscle oedema and sweat can
complicate interpretation of the Mmax. Limits also
exist for mechanical twitch responses after ultra-
marathons,
[14]
in particular the fact that fully
potentiated twitches were not always used in the
past (e.g. Millet et al.
[25]
). Tetanic responses are
slight or not influenced by potentiation. A sig-
nificant but moderate (~10%) decrease in high-
frequency force response has been found for knee
The Flush Model of Fatigue 493
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
extensors after prolonged
[27]
or ultra-long
[26]
running exercise.
Indirect indices of muscle damage (creatine
kinase, myoglobin, C-reactive protein, myosin
heavy chain fragments, lactate dehydrogenase,
aspartate aminotransferase, alanine aminotrans-
ferase, cytokines) also suggest the existence of
some peripheral alterations
[26,44-46,52-54]
but from
the few results available, it appears that subjects
show wide variability in the degree of muscle
damage.
[26]
Potential explanations for this vari-
ability in inter-individual response include differ-
ences in genes, training (particularly the repeated
bout effect,
[55]
that likely occurs, particularly
during downhill running training
[56]
) flexibility,
oxidative stress,
[18]
muscle fibre type
[57]
and running
technique. No direct evidence exists to preferentially
support any one factor and it is probably a com-
bination of them all that explains the large dif-
ference in muscle damage among subjects.
Nevertheless, for the last two potential factors
(i.e. typology and running technique), it is inter-
esting to report that (i) there was no significant
relationship between knee extensor peripheral
fatigue and percentage type I muscle fibres in the
vastus lateralis;
[26]
and that (ii) ultra-long running
(from Paris to Beijing [i.e. 8500 km in 161 days,
~53 km daily]) modified the running patterns to-
wards a ‘smoother’ style in one case study.
[58]
This
latter point was evidenced by (i) a higher stride
frequency and duty factor; (ii) a reduced aerial
time with no change in contact time; (iii) a lower
maximal vertical force and loading rate at im-
pact; and (iv) a decrease in both potential and
kinetic energy changes at each step.
[58]
We also
measured a reduced running economy after the
trip. Thus, even if it is possible that the running
pattern changes could be linked to the decrease in
maximal strength also observed, we suggested
that running pattern modification was a strategy
to reduce the potential deleterious effects of his
ultra-long distance run rather than to decrease
the energy cost of running. Further studies should
also examine the potential influence of running
technique on muscle damage during ultra-endurance
running, particularly when running on variable
terrain (trails). Anecdotal information from qua-
lified coaches suggests that technical ability in
downhill sections might be a real determinant of
fatigue and performance.
Low-frequency fatigue (LFF; also called pro-
longed low-frequency force depression
[59]
) [i.e.
the preferential loss of force at low frequencies of
electrical stimulation] is a prominent character-
istic of exercises involving lengthening contrac-
tions of the active muscles such as eccentric- and
stretch shortening cycle-type exercises,
[60]
and has
been associated with failure of the excitation-
contraction coupling.
[61]
Contrary to what is gen-
erally observed after downhill running,
[56,62]
most
studies did not show LFF after prolonged run-
ning exercise despite several experiments that have
measured this factor during ultra-long running
exercises including the recent 24-hour treadmill
study.
[26-28,32]
Only recently have we been able
to able to measure LFF after one of the most
extreme exercises realized by humans in race
conditions: a 166-km mountain ultra-marathon
with 9500 m of positive and negative elevation
change.
[23]
It can then be suggested that minimal
exercise intensity is required to induce mechanical
or metabolic disturbances that can result in devel-
oping LFF. However, there is a limitation in that
it is theoretically possible that axonal hyperpolar-
ization preferentially depresses the high-frequency
response during tetanic muscle stimulation. Thus,
an absence of modification to the low- to high-
frequency ratio could have resulted from the
combined effects of LFF, which preferentially
depresses low-frequency response, and hyperpo-
larization, which preferentially depresses high-
frequency response.
[26]
3. The Flush Model
Measuring central activation changes or force/
EMG responses during MVCs or electrically
evoked stimulation after prolonged running gives
some insight into the potential factors implicated
in fatigue at the CNS and/or muscle level. How-
ever, it does not predict how this affects sub-
maximal muscle function during ultra-endurance
running or how this influences speed regulation
during a real competition when pacing strategies
are allowed. This poses the question whether there
is any relationship between fatigue evidenced by
494 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
measures taken during maximal contractions/
stimulation and performance limitation/regulation.
Before answering this question, the change in
power output during ultra-marathons must first
be described.
3.1 Power Output Change in Ultra-Marathon
Runners
Only one study
[31]
has examined the changes in
efficiency/energy cost to determine whether
power output can be connected to speed. Since it
is unlikely that a large uncoupling exists between
power output and velocity, monitoring the speed
change over flat-course ultra-marathons (the ty-
pical event being the 24-hour race) gives an ex-
cellent idea of the mechanical power produced by
the runner. During a self-paced 24-hour treadmill
run where the subjects were asked to give their
best performance as they would in a normal race,
Martin et al.
[26]
showed a clear decrease in velo-
city during the first 16 hours before there was a
tendency for this to level off. Over a shorter dis-
tance (68 km) in a real competition, Utter et al.
[63]
showed that subjects reached a rating of per-
ceived exertion (RPE) similar to the subjects of
Martin et al.
[26]
(i.e. 15.4 –0.4) but there was
an increase of RPE up to the end of the race.
While several studies have investigated 24-hour
races,
[64-66]
we are not aware of any data report-
ing velocity changes during an official race. An-
ecdotal evidence and personal data suggest that
speed is reduced during ultra-marathons, even in
elite athletes. Also, unpublished results showed
that the speed of the top five runners in the 2007
French 24-hour championship had a tendency to
decrease but that this was less pronounced for the
winner. Over a shorter distance (100 km at the
1995 International Association of Ultra Runners
World Challenge), Lambert et al.
[67]
reported
that the best runners (i) reduced their initial speed
less and later in the race; and (ii) they showed less
variation in their speeds than did their lesser-
performing counterparts. The general tendency
was still that these 107 runners reduced their
speed over time.
[67]
In marathon running, it has
been reported
[68]
that some elite athletes are able
to maintain their pace throughout the race but
their slower counterparts (especially young men
[69]
)
slow down over the distance.
It is more difficult to document power output
and speed change during ultra-trails (i.e. off-road
ultra-marathons) since the terrain is usually hilly,
even mountainous. Future studies should mea-
sure the changes in speed at given slopes using
global positioning system tools. Data from Utter
et al.
[63]
nevertheless suggest that heart rate (HR)
decreases over a 68-km ultra-marathon. Personal
data also show that HR generally decreases over
an ultra-endurance race in mountains (see figure 2a
and the first 22 hours in figure 2b) and that a
correlation exists between change in HR and
change in elevation speed (in m/hour) similar to
the relationship seen between change in HR and
speed variation after the first 10 km of a mara-
thon.
[70]
When using HR to predict speed varia-
tions, one should also consider the HR drift with
fatigue (i.e. HR increases at a given speed). This is
visible in the first 1–2 hours in figure 2 but the
change in HR due to cardiac drift is less than the
speed variation. However, any decrease in HR
during competitions can also only be due to a
decreased power output.
3.2 Description of the Model
Despite the lack of systematic studies on speed
change over ultra-marathons, the data presented
strongly suggest that power output/speed decreases
with time in ultra-marathon running, even in elite
athletes. Following the initial question about the
relationship between fatigue as evidenced by
measures taken during maximal contractions/
stimulations and performance limitation/regulation,
one should then ask whether the decreased speed
is due to strength loss. Based on the fact that knee
extensor and plantar flexor muscle strength de-
creased by ~30–40%and since the force devel-
oped at each step is very low, it does not appear
that strength loss can directly explain this speed
reduction. Marcora et al.
[71,72]
suggested that the
locomotor muscles’ capacity for force production
is always well above the requirements of high-
intensity cycling, which is about 20%of MVC. In
other words, even if this point has been debated,
[73]
failure to produce the force/power required by
The Flush Model of Fatigue 495
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
the exercise despite maximal voluntary effort
does not seem to limit submaximal performance
as commonly assumed.
While it is not possible to predict to what extent
the fatigue mechanisms identified during max-
imal isometric contractions would affect perfor-
mance during an ultra-endurance running event,
it may nevertheless be hypothesized that there
is an indirect effect. As proposed by the tele-
oanticipatory system
[74]
or the central governor
model,
[75,76]
the level of muscle activation (and so
the speed) is thought to be progressively reduced
to keep the RPE during running below a max-
imum tolerated level
[26]
(i.e. to maintain the body
below a homeostatistically acceptable exercise
intensity).
[77,78]
In relation to the data presented
above regarding central and peripheral fatigue,
the last part of the present article will discuss
2500
a
b
2000
1500
Altitude (m)Altitude (m)
HR (beats/min)HR (beats/min)
1000
500
0
3000
2500
2000
1500
1000
500
0
612 18 22 24
80
100
120
140
160
2 h
612
180
170
160
150
140
130
120
110
100
Race duration (h)
Race duration (h)
18
Fig. 2. Typical heart rate (HR) changes during two mountain ultra-marathon races: one crossing the island of la Reunion (155 km [a]) and one
around the Mont-Blanc (165 km [b]). The global tendency of heart change is indicated by the solid line. The profile of the course is also given
(grey shading).
496 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
(i) how RPE is progressively increased with time
for a constant (or even lower) running speed; and
(ii) how in relation to changes in environmental
conditions, sleep deprivation/mental fatigue, drugs,
cognitive or nutritional strategies, this may reg-
ulate performance in ultra-marathon running. We
propose a conceptual model based on the flush
toilet (figure 3). The ‘flush model’ is based on the
central governor model proposed by Noakes
et al.
[76]
(i.e. agrees with the fact that exercise
performance is regulated by the CNS specifically
to prevent catastrophic physiological failure). How-
ever, the flush model emphasizes the importance
of peripheral fatigue that has been described in
detail in the first part of the present article. Also,
because it has been suggested that the central
governor integrate the input from various sys-
tems all related to exercise,
[79]
the flush model was
also built to take into account changes not asso-
ciated with exercise. The present model was mainly
designed to explain the role of fatigue on perfor-
mance in ultra-marathon running and consists of
four components: the ball [or buoy, (1) in figure 3]
represents RPE and can increase or decrease
based on the filling rate (2) and the water evac-
uated through the waste pipe (3). There is also a
security reserve (4), also called the emergency
reserve,
[80]
which allows the subject to prevent
physiological harm.
[81]
We believe that the flush
model can help us to understand running strate-
gies during an ultra-marathon, but a few ques-
tions must first be addressed.
3.3 Feed-Forward and Feedback
Mechanisms Influence the Filling Rate
First, what influences the filling rate? At the
beginning of an ultra-marathon, running pace
is based on a control system which estimates
the optimal power output.
[82]
Depending on the
runner, his or her goal may be either simply fin-
ishing the race, finishing the race in a certain time
Death
Security reserve
RPE
RPE
Filling rate
Waste pipe
↑ Feed-forward
Ultra-marathon
↑ Feedback
Exercise cessation
I feel good (±)
Nociceptive information
(muscle, joint, tendon, blister,
digestive problems, etc.)
Peripheral fatigue + spinal and/or
supraspinal inhibition
1
2
3
4
Fig. 3. The flush model. Rating of perceived exertion (RPE) is assimilated to the volume of water in the tank (i.e. an increase in volume of
water signifies a higher RPE and decreasing the level of water in the tank indicates decreasing RPE). The water can get in (filling rate, 2) and
out (via waste pipe, 3) and the level of water can be detected by the ball (1). The level of water depends on the filling rate, mainly determined by
peripheral changes and central inhibition/disfacilitation (feedback and feed-forward mechanisms), but other factors such as mental fatigue,
nutritional strategies, sleep deprivation, environmental conditions and exceptional events during races can affect the level. The size of
the security reserve (4) is mainly determined by motivation. Psychostimulants and pain killers can modify the sensitivity of the RPE sensor
(i.e. the ball).
The Flush Model of Fatigue 497
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
or being well placed at the finish (for instance
being in the top ten runners). The estimation is
based on several factors such as distance, eleva-
tion, environmental conditions, training status
and runabilty (i.e. technical difficulty) of the course.
This teleoanticipation means that the runner has
a template that contains existing data on exercise
performance (the so-called ‘experience’ or per-
sonal history of running). In fact, Ulmer
[82]
con-
siders that this template may be inborn but it is
generally accepted that optimal pacing strategy is
the result of a learning process.
[67]
The initial pace
gives an initial filling rate of the tank and the
higher the initial speed for a given running, the
faster the filling rate. This is mainly due to both
feed-forward and feedback mechanisms.
Feedback mechanisms have been widely docu-
mented in the literature. For instance, Amann
et al.
[83]
nicely showed that by attenuating the
ascending activity of nociceptive and metabo-
receptive Ad(group III) and C fibres (group IV),
somatosensory feedback from the locomotor mus-
cles influences central motor drive. Even if the
this experimental study has been criticized for
lack of proper placebo procedures,
[84]
the authors
concluded that locomotor muscle afferent feed-
back, which also facilitates performance through
optimizing muscle oxygen delivery,
[85]
exerted an
inhibitory influence on the determination of cen-
tral motor drive during high-intensity exercise.
While acidosis or inorganic phosphate accumu-
lation is unlikely to occur in ultra-marathon run-
ning, other biochemical mediators, such as the
accumulation of extracellular potassium
[86]
or cyto-
kines (especially interleukin-6 and its antagonist
IL-ra) due to structural muscle damage
[44-46]
(see
section 2.2 about peripheral fatigue), could trig-
ger group III/IV afferent fibres and mediate the
sensation of fatigue.
[74]
Nociceptive information
is much more complicated; there is a potential
role for pain arising not only from the muscles
but also from other sites such as joints and tendons.
Even blisters (cutaneous afferences
[87]
) or diges-
tive problems
[42]
could all potentially play a role.
For shorter distances (i.e. higher intensity and
ventilation rate), dyspnoea may also be involved
in nociceptive information
[88]
but probably not in
ultra-marathon running. It should also be noted
that a modest temporary reduction in pressure
pain perception was observed after a 100-mile
(161-km) trail run, only in the faster runners.
[89]
Although the subject is debated,
[85]
feedback from
the locomotor muscles probably plays a major
role in central motor drive regulation in ultra-
marathons. As stated in section 2.1, the down-
regulation of group III/IV afferents at the spinal
and supra-spinal levels
[90]
is a probable explana-
tion for why maximal %VA is lower after running
than after cycling or skiing for similar intensity/
distance.
[14]
While afferent feedback certainly has a key
function, the regulation of central motor command
is complex and also depends on the environment.
For example, altitude and elevated temperature
are two conditions frequently encountered by
ultra-marathon runners (e.g. races in Nepal or in
deserts). Regarding altitude, using a sub-maximal
test until exhaustion in hypoxia/normoxia while the
muscles were maintained in identical complete
ischaemic conditions, we showed that (i) inhibitory
mechanisms from working muscles play a major
role in the cessation of the exercise in hypoxia and
that (ii) a minor but significant direct effect of
inspired oxygen fraction on the CNS could po-
tentiate this limiting mechanism and explain why
performance was slightly reduced in hypoxia.
[15]
Similarly, Amann et al.
[91]
showed that peripheral
fatigue measured with femoral nerve magnetic
stimulation at task failure was substantially less
severe in hypoxia compared with normoxia or
moderate hypoxia. This was attributed to brain
hypoxic effects on effort perception, leading the
subjects to stop earlier. Similar conclusions can
be deducted from hypoglycaemia
[92]
or hyper-
thermia experiments. Regarding this latter factor,
it is worth noting that high temperature does not
alter performance over brief contractions but
does cause reductions during sustained contrac-
tions
[16]
or prolonged exercises.
[93]
Thus, sensory feedback contributes to central
fatigue and effort perception, presumably through
its indirect projection into the anterior cingulate
cortex.
[85]
Other authors have argued that sensory
signals from peripheral receptors do not contribute
to perception of effort but generate other sensa-
tions experienced during exercise (e.g. muscle
498 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
pain and thermal sensation
[84,94]
). In all cases, the
increase in RPE with fatigue is not abolished by
spinal blockade of somatosensory feedback from
the muscles; there must be other mechanisms.
Besides triggering inhibitory mechanisms, peri-
pheral fatigue implies that greater muscle activa-
tion is required for a given mechanical power to
be produced in the fatigued condition. Indeed, at
a given force or power output, the onset of fatigue
is usually concomitant with a rise in neuromus-
cular cost (EMG signal amplitude), which points
to the recruitment of additional motor units and/
or a higher discharge rate in order to compensate
for peripheral alterations. It is known that RPE
changes and the increase in muscle activity dur-
ing a constant-load exercise are correlated.
[95]
Marcora et al.
[71]
showed that the reduced loco-
motor muscle force after drop jumps resulted in
a higher RPE at a given power output and a re-
duced time to exhaustion during high-intensity
constant-power cycling. They suggested that the
effects were mediated by the increased central
motor command required to exercise with weaker
locomotor muscles,
[71]
which increased the percep-
tion of effort probably throughout its corollary
discharge to sensory areas of the brain.
[84]
Similarly,
Gagnon et al.
[96]
tested the effects of pre-induced
quadriceps fatigue (using electrostimulation) on
endurance performance of healthy individuals
and patients with chronic obstructive pulmonary
disease. These authors demonstrated that endurance
time significantly decreased by 20–25%in the
experimental condition in both groups. There
has been hot debate about the role of afferent
feedback from fatigued locomotor muscles as
an important determinant of endurance exercise
performance.
[84,85]
Gagnon et al.
[96]
suggested
that that the enhanced metaboreflex is not the
main mechanism through which exercise toler-
ance was reduced in the fatigued state in both
study populations.
Moreover, a change in muscle efficiency after
ultra-marathon running
[31]
can affect muscle recruit-
ment to maintain a given task. Spinal inhibition
and/or disfacilitation (see section 2.1) after ultra-
marathon running could also necessitate higher
central command (i.e. reinforcing feed-forward
mechanisms) from supraspinal sites. Finally, the
gain of motoneurones decreases in fatigued con-
ditions such that additional synaptic drive at a
premotoneuronal level is required to maintain a
constant firing rate
[9]
(i.e. a larger descending drive
is needed to continue exercise at the same power
output). So, as well as the nociceptive signal coming
from the peripheral receptors, these feed-forward
mechanisms (also called the ‘sense of effort’
[1,79]
)
could also partly explain the RPE drift.
[97]
Interestingly, RPE for the same exercise could
vary with environmental conditions and is not
necessarily associated with a decrease in MVC
(e.g. at altitude).
[98]
In summary, the initial pace
and the adjustments made in response to these
feed-forward and feedback mechanisms
[81]
directly
affect filling rate. RPE is probably affected by
both central command output and muscle affer-
ents.
[85]
In other words, as stated by Smirmaul,
[94]
an interaction between the sense of effort and the
sensations obtained from afferent sensory feedback
that is probably the ultimate regulator of exercise
performance. We suggest that for ultra-marathons
performed over hilly terrain, feed-forward and feed-
back mechanisms are mainly implicated in uphill/
flat sections and downhill sections, respectively.
3.4 Apart from the Filling Rate, Which Factors
Influence the Quantity of Water?
The second question is whether the volume of
water (absolute RPE) depends only on the filling
rate? The answer is clearly no. It has recently been
argued that it is an interaction between the sense
of effort and the sensations obtained from affer-
ent sensory feedback that is probably the ultimate
regulator of exercise performance.
[94]
However,
while these two factors certainly play a crucial
role, it is, for example, possible to start an ex-
ercise with more water in the tank than usual (i.e.
with a higher RPE at the beginning of exercise
than is normally the case). Indeed, it is possible to
feel some fatigue without any exhaustive physical
load, for example, after a stressful day. More
importantly, Marcora et al.
[17]
reported that the
time to exhaustion at 80%of peak power output
was significantly reduced by ~15%after 90 minutes
of a demanding cognitive task. This was asso-
ciated with a higher RPE at the beginning of the
The Flush Model of Fatigue 499
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
cycling exercise compared with the control con-
dition. Since RPE increased similarly over time in
both conditions, mentally fatigued subjects reached
their maximal tolerated RPE and disengaged from
the cycling exercise earlier than did the controls.
[17]
Similarly, although one night of sleep depri-
vation does not usually affect MVC or intense
exercise,
[99]
several studies have demonstrated a
deleterious effect on endurance performance.
[100,101]
Interestingly, a higher RPE for a given load has
been observed after sleep deprivation.
[99,100,102]
Some authors
[100,101]
showed that, after sleep de-
privation, subjects ran a shorter distance during a
30-minute self-paced treadmill exercise than did
their controls, yet their perception of effort was
similar. The authors suggested that altered per-
ception of effort may account for decreased en-
durance performance after sleep deprivation.
Other data show that RPE is not only dependant
on sensory information and cortical output. For
example, unknown or unexpected running exercise
duration may affect RPE,
[103]
suggesting that
RPE has an affective component. Also, changing
the tempo of music that cyclists were listening to
influenced their self-chosen power and cadence.
[104]
In summary, effort perception probably involves
the integration of multiple signals from a variety
of perceptual cues. Alternatively, as suggested in
section 3.6 for the amphetamines or pain killers,
it can be argued that perception is a complex
neurocognitive process that does not depend only
on the intensity of the sensory signal because
sensory signals are processed at brain level and
interpreted by the subject. It is then possible that
mental fatigue, music and sleep deprivation affect
the processing of sensory signals rather than
provide additional sensory signals. Nevertheless,
in that case, the rate of increase in RPE (filling
rate) would be changed rather than RPE at the
beginning of the exercise, as is the case, for ex-
ample, for mental fatigue.
[17]
3.5 The Waste Pipe
The third question is whether rest is the only
way to decrease the level of water in the tank. It is
the most obvious but probably not the only one.
For example, it may be possible to reduce the
water level while running using suitable psycho-
logical strategies.
[105]
Various psychological routines
can be used by runners to attenuate the discomfort
of intense physical exertion. These strategies, la-
belled ‘dissociative thoughts’ (i.e. the runner dis-
tracts him-/herself by thoughts of a more external
nature) are performed to diminish the sensations
of pain during a marathon.
[105]
Nevertheless, it
has been reported that the best marathon runners
adopt associative cognitive strategies (i.e. are
centered on their own sensations),
[106]
probably
with the goal of maintaining optimal running
technique/efficiency and so decreasing peripheral
fatigue (and the filling rate). Another way to de-
crease the level of water in the tank might be
nutritional strategies. Chambers et al.
[107]
recently
showed that rinsing the mouth with solutions
containing glucose and maltodextrin could im-
prove cycling performance. The authors suggested
that this could be due to activation of brain re-
gions involved in reward and motor control since
functional MRI measurements showed that these
regions believed to mediate emotional and be-
havioural responses to a rewarding sensory stimulus
were activated. Thus, in addition to its peripheral
action (i.e. slowing the tank filling rate), glucose
ingestion could have some central effects (i.e. de-
creasing the water level). Unidentified oral recep-
tors in the mouth could counteract the increase in
RPE, permitting higher central command and
power output.
Interestingly, the suggestion of downhill cycling
through hypnotic manipulation decreased RPE
without altering exercise HR or blood pressure
responses.
[108]
One could also suggest that redu-
cing maximal neural drive (i.e. central fatigue, a
decrease in %VA) may reduce the size of the tank.
In this context, Søgaard et al.
[109]
acknowledged
that central fatigue can only be demonstrated
during MVCs, but these authors suggested that
adecreasein%VA may have contributed to the
increase in RPE during sustained low-intensity
contractions (i.e. 15%MVC for 43 minutes). They
based this speculation on the fact that central fa-
tigue was among the factors that predicted RPE
changes. To our knowledge, no neurophysiogical
basis exists regarding the potential role of central
fatigue to explain at least in part the RPE changes.
500 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
3.6 The Security Reserve
The fourth question is why exercise stops (in
the case of a time to exhaustion) or why the runner
decides to adjust his or her speed/power output
(in the case of a time trial). Assuming a constant
level of motivation, exercise cessation appears
to occur at the same RPE whatever the rising
slope
[110-112]
and the starting level.
[17,71]
Marcora
et al.
[17]
have incorporated Brehm’s theory –a
general motivation theory that does not specifi-
cally refer to exercise –to propose a psychobio-
logical model of endurance performance. Marcora
et al.’s model postulates that subjects decide to
withdraw effort (i.e. disengage) when an exercise
is perceived to be either too difficult or the effort
demanded exceeds the upper limit of what people
are willing to invest. Alternatively, it has been
proposed that exercise terminates when the feelings
of discomfort overwhelm the potential rewards of
continuing to exercise.
[112]
This was reviewed in
2003 by Kayser.
[79]
It is interesting to report that
an opiate antagonist, naloxone, leads to signif-
icant reductions in exercise performance when
compared with control trial.
[113]
The authors of
the article concluded that working capacity was
limited by the individual RPE, which can be at-
tenuated by endogenous opioids rather than by
physiological fatigue. Thus, it is not task failure
but task disengagement that sets the exercise limit
before reaching the security reserve. Even in highly
motivated competitors, task disengagement always
occurs before there is a threat to life.
[79]
Humans
do not usually exceed their security reserve. There
are a few exceptions where dramatic loss in body
homeostasis was reached causing collapse such as
during marathons or triathlons (e.g. Gabriela
Andersen-Schiess in the 1984 Olympics or Julie
Moss in the 1982 Hawaii Ironman), especially
in (i) hot environments with subjects not always
familiar with these environmental conditions (i.e.
unadapted template/sensor efficiency); or (ii) under
the effect of psychostimulants (e.g. Tom Simpson
who died).
One could pose the question whether elite
athletes finish with a lower security reserve. While
the common belief is that better athletes can ‘dig
deeper’ and work relatively harder than their less
successful counterparts,
[70]
we are not aware of
any scientific study supporting this concept. One
indirect argument has been proposed by Esteve-
Lanao et al.
[70]
who found that the pattern of
percentage maximum heart rate (%HR
max
) res-
ponse during an event was very similar in athletes
with large differences in running performance.
These authors concluded that better runners are
faster due to their underlying physiological ca-
pacity rather than to their ability to put greater
relative effort into their competition. Studies in-
vestigating the percentage of maximal oxygen
uptake ( .
VO
2max
) sustained during competition in
function of the level of performance are contra-
dictory. It was found that the faster running speed
of the more trained runners over 10–90 km was
not due a higher %.
VO
2max
during competition
but was due to their superior running economy.
[114]
In contrast, performance was significantly related
to the specific endurance (i.e. the average speed
sustained over a 24-hour running exercise ex-
pressed in %.
VO
2max
).
[57]
Nevertheless, a higher
%.
VO
2max
or %HR
max
sustained during competi-
tion does not necessarily mean that the elite ath-
letes can ‘dig deeper’ since this may be due to
physiological differences in terms of endurance.
Further studies must examine this interesting
question.
A very important factor in the flush model
is that the sensor may be deregulated (i.e. the
interpretation of the incoming signal may be
affected).
[115]
In other words, the processing of
sensory signals is affected rather than additional
sensory signals provided. This is true in two op-
posing cases: amphetamines
[80]
(or more generally
when dopaminergic system is manipulated
[115]
)
and pain killers,
[83,116]
both of which induce higher
peripheral fatigue and/or metabolic disruptions.
For example, higher lactate/HR
[80,116]
or lower
peak doublet response to a magnetic stimulation
at the cessation of exercise has been reported.
[83]
Interestingly, while more anecdotal, Amann et al.
[83]
reported that all their subjects needed assistance
in disembarking the cycling ergometer after in-
jection of intrathecal fentanyl. In the evening (i.e.
several hours after exercise cessation), their sub-
jects reported continuing problems with ambu-
lation and muscle soreness, which had never been
The Flush Model of Fatigue 501
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
observed in any of the many other studies re-
questing exhaustion conducted in that laboratory.
[83]
It should be noted that psychostimulants (e.g.
amphetamines, cocaine) could act by providing a
pleasure sensation (i.e. water leak throughout the
waste pipe rather than sensor deregulated). Some
runners, even elite athletes, use local anaesthetics/
anti-inflammatory drugs (e.g. Tissugel

in France),
particularly applied to the knee joint. While this
is not prohibited by the World Anti-Doping
Agency policy, the flush model suggests that this
may be beneficial for improving performance in
ultra-marathon running.
According to the flush model, there is always
a reserve for muscle recruitment (the security re-
serve) that can be used for the so-called ‘end
spurt’
[117]
when the runner is at his or her highest
level of peripheral fatigue. In ultra-marathons,
this is clearly illustrated in figure 2b representing
the HR data of an ultra-marathon runner per-
forming a race around the Mont-Blanc. In this
example, very particular race conditions (from
11th to 6th place with opponents regularly an-
nounced as potential targets) led the runner to
accelerate at the end of the race when his muscle
fatigue was higher than at ~22 hours. This further
illustrates that central regulation is not totally
based on peripheral changes. Since RPE was not
recorded, it is not possible to say whether this was
related to the enjoyment of overtaking several
opponents near the finish line, which counter-
acted the sensation of fatigue (i.e. the same level
of water despite increasing the power output with
some water being evacuated through the waste
pipe) and/or a decrease in the security reserve due
to increased motivation.
While mental processes are important in all
sports, it is probably particularly true for ultra-
marathon runners. Weir et al.
[118]
suggested that
the central governor is most applicable to endurance
exercise since the decline in muscle performance
under intensely fatiguing exercise conditions can
be directly attributed to peripheral fatigue. It is
further possible that for shorter distances, the
organism is equipped with a system to protect its
integrity, but at a peripheral level.
[79]
For ex-
ample, glycogen depletion may alter excitation-
contraction coupling, which in turn limits muscle
contraction capacity and so restricts muscle da-
mage. Also, it has been suggested that for exercise
of several minutes duration, RPE is increased
by sleep deprivation but when it is as short as
30 seconds, sleep deprivation causes only a small
change in the perception of exercise intensity
[99]
and thus no reduction in performance. Marino
et al.
[119]
recently argued that events such as the
marathon and, for example, walking 1000 km are
both quantitatively and qualitatively different,
and not simply because the 1000 km race is longer.
They proposed that it could be that the limits to
performance in a 1000-km race are predominantly
mental, while the limits to performance in the
marathon are predominantly physiological. Thus,
it can be suggested that ultra-marathon running
is an interesting model to study central regulation
of exercise. We believe that all interventions
designed to manipulate RPE and the sensation
of discomfort are of particular interest in this
sport.
4. Conclusion
As Marino et al.
[119]
recently pointed out,
Mosso concluded in his book La fatica (fatigue)
published 120 years ago,
[120]
that two phenomena
categorize fatigue, the diminution of muscular
force and the sensation of fatigue: ‘‘That is to say,
we have a physical fact which can be measured
and compared and a psychic fact which eludes
measurement.’’ Thus, the study of fatigue should
address both the perception of effort and the
decline in force that occurs during sustained ex-
ercise.
[1]
The aim of the present article was to re-
view these two aspects of fatigue and to propose a
model that integrates the ‘neuromuscular’ and
‘physiological’ factors of fatigue (responsible for
maximal force reduction) in ultra-marathon
running to explain the regulation of performance.
It has been argued
[76,79,121]
that fatigue can be under-
stood as a highly regulated strategy conserving
cellular integrity, function and, indeed, survival.
The flush model, dedicated to integrating the fati-
gue mechanisms in ultra-marathon performance
(and more generally to any type of endurance
performance), using a holistic approach, is fully
compatible with this statement.
502 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
Acknowledgements
The author would like to thank Professor Ken Nosaka for
his valuable comments on the manuscript and Wanda Lipski
for English language correction.
No sources of funding were used to conduct this study or
prepare this manuscript. The author has no conflicts of in-
terest that are directly relevant to this article.
References
1. Enoka RM, Stuart DG. Neurobiology of muscle fatigue.
J Appl Physiol 1992 May; 72 (5): 1631-48
2. Hoffman MD, Ong JC, Wang G. Historical analysis of
participation in 161-km ultramarathons in North America.
Int J History Sport 2010; 27 (11): 1877-91
3. Knez WL, Coombes JS, Jenkins DG. Ultra-endurance ex-
ercise and oxidative damage: implications for cardiovascular
health. Sports Med 2006; 36 (5): 429-41
4. Zaryski C, Smith DJ. Training principles and issues for
ultra-endurance athletes. Cur Sports Med Rep 2005; 4 (3):
165-70
5. Noakes TD. The limits of endurance exercise. Basic Res
Cardiol 2006 Sep; 101 (5): 408-17
6. Bramble DM, Lieberman DE. Endurance running and the
evolution of Homo. Nature 2004 Nov 18; 432 (7015): 345-52
7. Pearson H. Physiology: freaks of nature? Nature 2006 Dec
21; 444 (7122): 1000-1
8. Abbiss CR, Laursen PB. Models to explain fatigue during
prolonged endurance cycling. Sports Med 2005; 35 (10):
865-98
9. Gandevia SC. Spinal and supraspinal factors in human
muscle fatigue. Physiol Rev 2001; 81 (4): 1725-89
10. Meeusen R, Watson P, Hasegawa H, et al. Central fatigue:
the serotonin hypothesis and beyond. Sports Med 2006;
36 (10): 881-909
11. Forestier N, Nougier V. The effects of muscular fatigue on
the coordination of a multijoint movement in human.
Neurosci Lett 1998 Aug 21; 252 (3): 187-90
12. Brisswalter J, Collardeau M, Rene A. Effects of acute
physical exercise characteristics on cognitive perfor-
mance. Sports Med 2002; 32 (9): 555-66
13. Bainbridge FA. The physiology of muscular exercise.
New York: Longmans, Green and Co., 1931
14. Millet GY, Lepers R. Alterations of neuromuscular func-
tion after prolonged running, cycling and skiing exercises.
Sports Med 2004; 34 (2): 105-16
15. Millet GY, Aubert D, Favier FB, et al. Effect of acute hy-
poxia on central fatigue during repeated isometric leg
contractions. Scand J Med Sci Sports 2009 Oct; 19 (5):
695-702
16. Nybo L, Nielsen B. Hyperthermia and central fatigue
during prolonged exercise in humans. J Appl Physiol 2001
Sep; 91 (3): 1055-60
17. Marcora SM, Staiano W, Manning V. Mental fatigue im-
pairs physical performance in humans. J Appl Physiol
2009 Mar; 106 (3): 857-64
18. Gauche E, Lepers R, Rabita G, et al. Vitamin and mineral
supplementation and neuromuscular recovery after a
running race. Med Sci Sports Exerc 2006 Dec; 38 (12):
2110-7
19. Petersen K, Hansen CB, Aagaard P, et al. Muscle me-
chanical characteristics in fatigue and recovery from a
marathon race in highly trained runners. Eur J Appl
Physiol 2007 Oct; 101 (3): 385-96
20. Racinais S, Girard O, Micallef JP, et al. Failed excitability
of spinal motoneurons induced by prolonged running
exercise. J Neurophysiol 2007 Jan; 97 (1): 596-603
21. Ross EZ, Middleton N, Shave R, et al. Corticomotor ex-
citability contributes to neuromuscular fatigue following
marathon running in man. Exp Physiol 2007 Mar; 92 (2):
417-26
22. Saldanha A, Nordlund Ekblom MM, Thorstensson A.
Central fatigue affects plantar flexor strength after pro-
longed running. Scand J Med Sci Sports 2008 Jun; 18 (3):
383-8
23. Millet GY, Tomazin K, Verges S, et al. Neuromuscular
consequences of an extreme mountain ultra-marathon.
PLoS ONE 2011; 6 (2): e17059
24. Lepers R, Pousson M, Maffiuletti NA, et al. The effects
of a prolonged running exercise on strength characterisics.
Int J Sports Med 1999; 21: 275-80
25. Millet GY, Lepers R, Maffiuletti NA, et al. Alterations of
neuromuscular function after an ultramarathon. J Appl
Physiol 2002 Feb; 92 (2): 486-92
26. Martin V, Kerherve
´H, Messonnier LA, et al. Central and
peripheral contributions to neuromuscular fatigue in-
duced by a 24-h treadmill run. J Appl Physiol 2010; 108:
1224-33
27. Millet GY, Martin V, Lattier G, et al. Mechanisms con-
tributing to knee extensor strength loss after prolonged
running exercise. J Appl Physiol 2003 Jan; 94 (1): 193-8
28. Davies CTM, Thompson MW. Physiological responses to
prolonged exercise in ultramarathon athletes. J Appl
Physiol 1986; 61 (2): 611-7
29. Nicol C, Komi PV, Marconnet P. Fatigue effects of mara-
thon running on neuromuscular performance. II: changes
in force, intergrated electromyographic activity and en-
durance capacity. Scand J Med Sci Sports 1991; 1: 18-24
30. Nicol C, Komi PV, Marconnet P. Fatigue effects of mara-
thon running on neuromuscular performance. I: changes
in muscle force and stiffness characteristics. Scand J Med
Sci Sports 1991; 1: 10-7
31. Millet GY, Lepers R, Lattier G, et al. Influence of ultra-
long-term fatigue on the oxygen cost of two types of lo-
comotion. Eur J Appl Physiol 2000 Nov; 83 (4-5): 376-80
32. Place N, Lepers R, Deley G, et al. Time course of neuro-
muscular alterations during a prolonged running exercise.
Med Sci Sports Exerc 2004 Aug; 36 (8): 1347-56
33. Rasmussen P, Nielsen J, Overgaard M, et al. Reduced
muscle activation during exercise related to brain oxyge-
nation and metabolism in humans. J Physiol 2010 Jun 1;
588 (Pt 11): 1985-95
34. Taylor JL, Gandevia SC. A comparison of central aspects
of fatigue in submaximal and maximal voluntary con-
tractions. J Appl Physiol 2008 Feb; 104 (2): 542-50
35. Ohta M, Hirai N, Ono Y, et al. Clinical biochemical eva-
luation of central fatigue with 24-hour continuous ex-
ercise. Rinsho Byori 2005 Sep; 53 (9): 802-9
The Flush Model of Fatigue 503
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
36. Davis JM, Bailey SP. Possible mechanisms of central ner-
vous system fatigue during exercise. Med Sci Sports Exerc
1997; 29 (1): 45-57
37. Nybo L, Rasmussen P. Inadequate cerebral oxygen deliv-
ery and central fatigue during strenuous exercise. Exercise
and sport sciences reviews 2007 Jul; 35 (3): 110-8
38. Lepers R, Maffiuletti NA, Rochette L, et al. Neuromus-
cular fatigue during a long-duration cycling exercise.
J Appl Physiol 2002 Apr; 92 (4): 1487-93
39. Millet GY, Millet GP, Lattier G, et al. Alteration of neuro-
muscular function after a prolonged road cycling race.
Int J Sports Med 2003 Apr; 24 (3): 190-4
40. Millet GY, Martin V, Maffiuletti NA, et al. Neuromus-
cular fatigue after a ski skating marathon. Can J Appl
Physiol 2003 Jun; 28 (3): 434-45
41. Zehr PE. Considerations for use of the Hoffmann reflex in
exercise studies. Eur J Appl Physiol 2002 Apr; 86 (6):
455-68
42. Jeukendrup AE, Vet-Joop K, Sturk A, et al. Relationship
between gastro-intestinal complaints and endotoxaemia,
cytokine release and the acute-phase reaction during and
after a long-distance triathlon in highly trained men. Clin
Sci (Lond) 2000 Jan; 98 (1): 47-55
43. Kim HJ, Lee YH, Kim CK. Biomarkers of muscle and
cartilage damage and inflammation during a 200 km run.
Eur J Appl Physiol 2007 Mar; 99 (4): 443-7
44. Ostrowski K, Hermann C, Bangash A, et al. A trauma-like
elevation of plasma cytokines in humans in response
to treadmill running. J Physiol 1998 Dec 15; 513 (Pt 3):
889-94
45. Ostrowski K, Rohde T, Zacho M, et al. Evidence that
interleukin-6 is produced in human skeletal muscle during
prolonged running. J Physiol 1998 May 1; 508 (Pt 3):
949-53
46. Papassotiriou I, Alexiou VG, Tsironi M, et al. Severe
aseptic inflammation caused by long distance running
(246 km) does not increase procalcitonin. Eur J Clin In-
vest 2008 Apr; 38 (4): 276-9
47. Taylor JL, Todd G, Gandevia SC. Evidence for a su-
praspinal contribution to human muscle fatigue. Clin Exp
Pharmacol Physiol 2006 Apr; 33 (4): 400-5
48. Todd G, Taylor JL, Gandevia SC. Measurement of vol-
untary activation of fresh and fatigued human muscles
using transcranial magnetic stimulation. J Physiol 2003
Sep 1; 551 (Pt 2): 661-71
49. Goodall S, Romer LM, Ross EZ. Voluntary activation of
human knee extensors measured using transcranial mag-
netic stimulation. Exp Physiol 2009 Sep; 94 (9): 995-1004
50. Sidhu SK, Bentley DJ, Carroll TJ. Locomotor exercise in-
duces long-lasting impairments in the capacity of the hu-
man motor cortex to voluntarily activate knee extensor
muscles. J Appl Physiol 2009 Feb; 106 (2): 556-65
51. Dimitrova NA, Dimitrov GV. Amplitude-related char-
acteristics of motor unit and M-wave potentials during
fatigue: a simulation study using literature data on intra-
cellular potential changes found in vitro. J Electrom Ki-
nesiol 2002; 12: 339-249
52. Koller A, Mair J, Schobersberger W, et al. Effects of pro-
longed strenuous endurance exercise on plasma myosin
heavy chain fragments and other muscular proteins.
J Sports Med Phys Fitness 1998; 38 (1): 10-7
53. Mastaloudis A, Traber MG, Carstensen K, et al. Anti-
oxidants did not prevent muscle damage in response to an
ultramarathon run. Med Sci Sports Exerc 2006 Jan; 38 (1):
72-80
54. Skenderi KP, Kavouras SA, Anastasiou CA, et al. Exer-
tional rhabdomyolysis during a 246-km continuous run-
ning race. Med Sci Sports Exerc 2006 Jun; 38 (6): 1054-7
55. Nosaka K, Sakamoto K, Newton M, et al. How long does
the protective effect on eccentric exercise-induced muscle
damage last? Med Sci Sports Exerc 2001 Sep; 33 (9):
1490-5
56. Martin V, Millet GY, Lattier G, et al. Effects of recovery
modes after knee extensor muscles eccentric contractions.
Med Sci Sports Exerc 2004 Nov; 36 (11): 1907-15
57. Millet GY, Banfi JC, Kerherve
´H, et al. Physiological and
biological factors associated with a 24 h treadmill ultra-
marathon performance. Scand J Med Sci Sports 2011;
21 (1): 54-61
58. Millet GY, Morin JB, Degache F, et al. Running from
Paris to Beijing: biomechanical and physiological con-
sequences. Eur J Appl Physiol 2009 Dec; 107 (6): 731-8
59. Allen DG, Lamb GD, Westerblad H. Skeletal muscle fati-
gue: cellular mechanisms. Physiol Rev 2008 Jan; 88 (1):
287-332
60. Martin V, Millet GY, Martin A, et al. Assessment of low-
frequency fatigue with two methods of electrical stimula-
tion. J Appl Physiol 2004 Nov; 97 (5): 1923-9
61. Hill CA, Thompson MW, Ruell PA, et al. Sarcoplasmic
reticulum function and muscle contractile character fol-
lowing fatiguing exercise in humans. J Physiol 2001; 531:
871-8
62. Martin V, Millet GY, Lattier G, et al. Why does knee ex-
tensor muscles torque decrease after eccentric-type ex-
ercise? J Sports Med Phys Fitness 2005 Jun; 45 (2): 143-51
63. Utter AC, Kang J, Nieman DC, et al. Ratings of perceived
exertion throughout an ultramarathon during carbohy-
drate ingestion. Percept Mot Skills 2003; 97: 175-84
64. Kao WF, Shyu CL, Yang XW, et al. Athletic performance
and serial weight changes during 12- and 24-hour ultra-
marathons. Clin J Sport Med 2008 Mar; 18 (2): 155-8
65. Knechtle B, Wirth A, Knechtle P, et al. Personal best
marathon performance is associated with performance
in a 24-h run and not anthropometry or training volume.
Br J Sports Med 2009 Nov; 43 (11): 836-9
66. Wu HJ, Chen KT, Shee BW, et al. Effects of 24 h ultra-
marathon on biochemical and hematological parameters.
World J Gastroenterol 2004 Sep 15; 10 (18): 2711-4
67. Lambert MI, Dugas JP, Kirkman MC, et al. Changes in
running speed in a 100 km ultramarathon race. J Sports
Sci Med 2004; 3: 167-73
68. Ely MR, Martin DE, Cheuvront SN, et al. Effect of am-
bient temperature on marathon pacing is dependent on
runner ability. Med Sci Sports Exerc 2008; 40 (9): 1675-80
69. March DS, Vanderburgh PM, Titlebaum PJ, et al. Age,
sex, and finish time as determinants of pacing in the
marathon. J Strength Cond Res 2011; 25 (2): 386-91
504 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
70. Esteve-Lanao J, Lucia A, deKoning JJ, et al. How do hu-
mans control physiological strain during strenuous en-
durance exercise? PloS One 2008; 3 (8): e2943
71. Marcora SM, Bosio A, de Morree HM. Locomotor muscle
fatigue increases cardiorespiratory responses and reduces
performance during intense cycling exercise independently
from metabolic stress. Am J Physiol Regul Integr Comp
Physiol 2008 Mar; 294 (3): R874-83
72. Marcora SM, Staiano W. The limit to exercise tolerance in
humans: mind over muscle? Eur J Appl Physiol 2010 Mar
11; 763-70
73. Marcora SM, Staiano W. Reply to: What limits exercise
during high-intensity aerobic exercise? Eur J Appl Physiol
2010 Jul 2; 110: 663-4
74. Ament W, Verkerke GJ. Exercise and fatigue. Sports Med
2009; 39 (5): 389-422
75. Hampson DB, St Clair Gibson A, Lambert MI, et al. The
influence of sensory cues on the perception of exertion
during exercise and central regulation of exercise perfor-
mance. Sports Med 2001; 31 (13): 935-52
76. Noakes TD, St Clair Gibson A, Lambert EV. From cata-
strophe to complexity: a novel model of integrative central
neural regulation of effort and fatigue during exercise in
humans. Br J Sports Med 2004 Aug; 38 (4): 511-4
77. St Clair Gibson A, Noakes TD. Evidence for complex
system integration and dynamic neural regulation of ske-
letal muscle recruitment during exercise in humans. Br
J Sports Med 2004 Dec; 38 (6): 797-806
78. St-Clair Gibson A, Lambert MI, Noakes TD. Neural
control of force output during maximal and submaximal
exercise. Sports Med 2001; 31 (9): 637-50
79. Kayser B. Exercise starts and ends in the brain. Eur J Appl
Physiol 2003; 90: 411-9
80. Swart J, Lamberts RP, Lambert MI, et al. Exercising with
reserve: evidence that the central nervous system regulates
prolonged exercise performance. Br J Sports Med 2009
Oct; 43 (10): 782-8
81. Tucker R. The anticipatory regulation of performance: the
physiological basis for pacing strategies and the develop-
ment of a perception-based model for exercise perfor-
mance. Br J Sports Med 2009 Jun; 43 (6): 392-400
82. Ulmer HV. Concept of an extracellular regulation of
muscular metabolic rate during heavy exercise in humans
by psychophysiological feedback. Experientia 1996 May
15; 52 (5): 416-20
83. Amann M, Proctor LT, Sebranek JJ, et al. Opioid-medi-
ated muscle afferents inhibit central motor drive and limit
peripheral muscle fatigue development in humans. J
Physiol 2009 Jan 15; 587 (Pt 1): 271-83
84. Marcora S. Perception of effort during exercise is in-
dependent of afferent feedback from skeletal muscles,
heart, and lungs. J Appl Physiol 2009 Jun; 106 (6): 2060-2
85. Amann M, Secher NH. Point: Afferent feedback from fa-
tigued locomotor muscles is an important determinant of
endurance exercise performance. J Appl Physiol 2010 Feb;
108 (2): 452-4
86. Overgaard K, Lindstrom T, Ingemann-Hansen T, et al.
Membrane leakage and increased content of Na+-K+
pumps and Ca2+in human muscle after a 100-km run.
J Appl Physiol 2002 May; 92 (5): 1891-8
87. Martin PG, Gandevia SC, Taylor JL. Muscle fatigue
changes cutaneous suppression of propriospinal drive to
human upper limb muscles. J Physiol 2007 Apr 1; 580
(Pt 1): 211-23
88. Kayser B, Sliwinski P, Yan S, et al. Respiratory effort
sensation during exercise with induced expiratory-flow
limitation in healthy humans. J Appl Physiol 1997 Sep;
83 (3): 936-47
89. Hoffman MD, Lee J, Zhao H, et al. Pain perception after
running a 100-mile ultramarathon. Arch Phys Med Re-
habil 2007 Aug; 88 (8): 1042-8
90. Gandevia SC, Allen GM, Butler JE, et al. Supraspinal
factors in human muscle fatigue: evidence for suboptimal
output from the motor cortex. J Physiol 1996 Jan 15;
490 (Pt 2): 529-36
91. Amann M, Romer LM, Subudhi AW, et al. Severity of
arterial hypoxaemia affects the relative contributions
of peripheral muscle fatigue to exercise performance in
healthy humans. J Physiol 2007 May 15; 581 (Pt 1):
389-403
92. Nybo L. CNS fatigue and prolonged exercise: effect of
glucose supplementation. Med Sci Sports Exerc 2003 Apr;
35 (4): 589-94
93. Tucker R, Marle T, Lambert EV, et al. The rate of heat
storage mediates an anticipatory reduction in exercise in-
tensity during cycling at a fixed rating of perceived exer-
tion. J Physiol 2006; 574 (Pt 3): 905-15
94. Smirmaul BD. Sense of effort and other unpleasant sensa-
tions during exercise: clarifying concepts and mechan-
isms. Br J Sports Med. In press
95. Fontes EB, Smirmaul BP, Nakamura FY, et al. The Re-
lationship between rating of perceived exertion and
muscle activity during exhaustive constant-load cycling.
Int J Sports Med Oct; 31 (10): 683-8
96. Gagnon P, Saey D, Vivodtzev I, et al. Impact of preinduced
quadriceps fatigue on exercise response in chronic ob-
structive pulmonary disease and healthy subjects. J Appl
Physiol 2009 Sep; 107 (3): 832-40
97. Millet GY. Central fatigue is not the source but can explain
performance decrement due to afferent feedback. J Appl
Physiol 2010 Feb; 108 (2): 464
98. Fulco CS, Lewis SF, Frykman PN, et al. Muscle fatigue
and exhaustion during dynamic leg exercise in normoxia
and hypobaric hypoxia. J Appl Physiol 1996 Nov; 81 (5):
1891-900
99. Myles WS. Sleep deprivation, physical fatigue, and the
perception of exercise intensity. Med Sci Sports Exerc
1985 Oct; 17 (5): 580-4
100. Martin BJ. Effect of sleep deprivation on tolerance of
prolonged exercise. Eur J Appl Physiol Occup Physiol
1981; 47 (4): 345-54
101. Oliver SJ, Costa RJ, Laing SJ, et al. One night of sleep
deprivation decreases treadmill endurance performance.
Eur J Appl Physiol 2009 Sep; 107 (2): 155-61
102. Bond V, Balkissoon B, Franks BD, et al. Effects of sleep
deprivation on performance during submaximal and
maximal exercise. J Sports Med Phys Fitness 1986 Jun;
26 (2): 169-74
103. Baden DA, McLean TL, Tucker R, et al. Effect of antici-
pation during unknown or unexpected exercise duration
The Flush Model of Fatigue 505
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)

This material is
the copyright of the
original publisher.
Unauthorised copying
and distribution
is prohibited.
on rating of perceived exertion, affect, and physiological
function. Br J Sports Med 2005 Oct; 39 (10): 742-6; dis-
cussion -6
104. Waterhouse J, Hudson P, Edwards B. Effects of music
tempo upon submaximal cycling performance. Scand
J Med Sci Sports 2010 Aug; 20 (4): 662-9
105. Raglin JS. The psychology of the marathoner: of one mind
and many. Sports Med 2007; 37 (4-5): 404-7
106. Morgan WP, Pollock ML. Psychologic characterization of
the elite distance runner. Ann N Y Acad Sci 1977; 301:
382-403
107. Chambers ES, Bridge MW, Jones DA. Carbohydrate sensing
in the human mouth: effects on exercise performance and
brain activity. J Physiol 2009 Apr 15; 587 (Pt 8): 1779-94
108. Williamson JW, McColl R, Mathews D, et al. Hypnotic
manipulation of effort sense during dynamic exercise:
cardiovascular responses and brain activation. J Appl
Physiol 2001 Apr; 90 (4): 1392-9
109. Søgaard K, Gandevia SC, Todd G, et al. The effect of
sustained low-intensity contractions on supraspinal fati-
gue in human elbow flexor muscles. J Physiol 2006 Jun 1;
573 (Pt 2): 511-23
110. Baldwin J, Snow RJ, Gibala MJ, et al. Glycogen avail-
ability does not affect the TCA cycle or TAN pools during
prolonged, fatiguing exercise. J Appl Physiol 2003 Jun; 94
(6): 2181-7
111. Faulkner J, Parfitt G, Eston R. The rating of perceived
exertion during competitive running scales with time.
Psychophysiology 2008 Nov; 45 (6): 977-85
112. Noakes TD. Linear relationship between the perception of
effort and the duration of constant load exercise that re-
mains. J Appl Physiol 2004 Apr; 96 (4): 1571-2
113. Sgherza AL, Axen K, Fain R, et al. Effect of naloxone on
perceived exertion and exercise capacity during maximal
cycle ergometry. J Appl Physiol 2002 Dec; 93 (6): 2023-8
114. Scrimgeour AG, Noakes TD, Adams B, et al. The influence of
weekly training distance on fractional utilization of maximum
aerobic capacity in marathon and ultramarathon runners.
Eur J Appl Physiol Occup Physiol 1986; 55 (2): 202-9
115. Meeusen R. Fatigue: from muscle to brain or vice versa?
J Appl Physiol 2010 Feb; 108 (2): 459-60
116. Mauger AR, Jones AM, Williams CA. Influence of aceta-
minophen on performance during time trial cycling.
J Appl Physiol 2010 Jan; 108 (1): 98-104
117. Noakes TD, Lambert MI, Hauman R. Which lap is the
slowest? An analysis of 32 world mile record perfor-
mances. Br J Sports Med 2009 Oct; 43 (10): 760-4
118. Weir JP, Beck TW, Cramer JT, et al. Is fatigue all in your
head? A critical review of the central governor model. Br
J Sports Med 2006 Jul; 40 (7): 573-86
119. Marino FE, Gard M, Drinkwater E. The limits to exercise
performance and the future of fatigue research. Br J Sports
Med 2011; 45: 65-7
120. Mosso A. La fatica. Milan: Treves, 1891
121. McKenna MJ, Hargreaves M. Resolving fatigue mechan-
isms determining exercise performance: integrative phy-
siology at its finest! J Appl Physiol 2008 Jan; 104 (1): 286-7
Correspondence: Professor Guillaume Millet, Laboratoire de
Physiologie de l’Exercice (EA 4338), Me
´decine du Sport-
Myologie, Ho
ˆpital Bellevue, 42055 Saint Etienne, Cedex 2,
France.
E-mail: guillaume.millet@univ-st-etienne.fr
506 Millet
ª2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (6)