Pflügers Arch - Eur J Physiol (2001) 442:41–48

DOI 10.1007/s004240100515

O R I G I N A L A RT I C L E

B. Nielsen · T. Hyldig · F. Bidstrup

J. González-Alonso · G.R.J. Christoffersen

Brain activity and fatigue during prolonged exercise in the heat

Received: 14 April 2000 / Received after revision: 3 November 2000 / Accepted: 18 December 2000 / Published online: 1 March 2001
© Springer-Verlag 2001

Abstract We hypothesized that fatigue due to hyper- earlier fatigue in the heat are generally associated with
thermia during prolonged exercise in the heat is in part high body temperature, cardiovascular instability and al-
related to alterations in frontal cortical brain activity. The tered metabolism [9, 12, 15, 16]. We have recently dem-
electroencephalographic activity (EEG) of the frontal onstrated that combined dehydration and hyperthermia
cortex of the brain. was measured in seven cyclists [maxi- during exercise in the heat results in marked reductions
mal O.2 uptake (VO2max) 4.8±0.1 (SE) l min–1] cycling at in cardiac output, exercising muscle blood flow and con-
60% VO2max in a hot (H, 42°C) and a cool (C, 19°C) envi- comitant elevations in muscle glycogen utilization and
ronment. Fast Fourier transformation of the EEG was lactate production [15, 16]. However, despite diminished
used to obtain power spectrum areas in the α (8–13 Hz) blood flow and altered muscle metabolism, fatigue ap-
and β (13–30 Hz) frequencies. The ratio α/β was calcu- pears to be largely associated with high body tempera-
lated as an index of arousal level; an elevated α/β index ture [5, 15, 16, 17, 21, 23, 25, 26].
reflects suppressed arousal. In H, subjects fatigued after Support for this idea is also found in our recent obser-
34.4±1.4 min coinciding with an oesophageal tempera- vation that cyclists
. exercising at 60% maximal oxygen
ture (Toes) of 39.8±0.1°C, an almost maximal heart rate consumption (VO2max) fatigued at remarkably similar core
(HR 192±3 beats·min–1), a rating of perceived exertion and muscle temperatures (≅40 and 40.7°C, respectively),
(RPE) of 19.0±0.8 and significantly elevated α/β index regardless of the wide differences in the initial value and
(188±71% of the value after 2 min of exercise; P<0.05). the rate of rise of body temperature [17]. In the same
In C, subjects cycled for a similar period while Toes was light, recent studies show that rats exercising in hot envi-
below 38°C, HR and RPE were low, and the α/β index ronments fatigue at the same core and brain temperature
was not significantly elevated (59±27% of 2 min value; when the pre-exercise core temperature is manipulated
P=NS). Increases in the α/β index were strongly corre- [13, 29]. Therefore, it is clear that fatigue during exercise
lated to increases in Toes (r2=0.98; P=0.0001). in hot environments in trained subjects is closely linked
to high body temperature (i.e. high core, muscle and brain
Keywords EEG · Hyperthermia · Oesophageal temperature), and not to the duration of the work, albeit
temperature the underlying mechanisms have not been elucidated.
The possibility exists that hyperthermia causes fatigue
by a direct action of temperature on the brain and/or an
Introduction indirect effect on the brain of afferent signals originating
from skeletal muscle, cardiac muscle or internal organs
Endurance performance in hot environments can be in response to rising local temperature. Such effects may
markedly reduced compared to that in cool environ- result in altered central command and/or electrical poten-
ments. In trained subjects, impaired performance and tials of some brain regions [5, 6, 17]. Indeed, previous
studies have shown that high temperature is associated
B. Nielsen (✉) · T. Hyldig · F. Bidstrup · J. González-Alonso with altered brain activity in humans and animals, as re-
Institute of Exercise and Sport Sciences, flected by changes in electroencephalogram (EEG) activ-
Department of Human Physiology, University of Copenhagen, ity and sensory evoked potentials [10, 11, 22]. There is
Universitetsparken 13, 2100, Copenhagen Ø, Denmark
e-mail: bnielsen@aki.ku.dk evidence from studies of resting humans that the power
Tel.: +45-35-321620, Fax: +45-353-21600 spectrum of the EEG changes in proportion to the rise in
G.R.J. Christoffersen core temperature up to ≅41.8°C [10]. At present there is
August Krogh Institute, University of Copenhagen, no evidence for such an effect of hyperthermia on the
Copenhagen Ø, Denmark EEG power spectrum in exercising humans.
42

Therefore, the purpose of the present study was to de- chamber subjects were seated on the ergometer in a forward bend
termine whether progressive hyperthermia and ultimate- position (approx. 30° angle from the horizontal axis) while resting
their arms on the triathlon handlebars. The positions of the saddle
ly fatigue during exercise in hot environments are in part and the triathlon handlebars were adjusted for each subject in the
related to altered brain activity in the frontal cortex. We same way during both conditions. The subjects wore earplugs and
measured EEG activity of the frontal area of the brain in snow glasses without glass all the time in order to reduce the pe-
cyclists under two conditions: cycling at the same inten- ripheral field of vision and decrease external disturbances of brain
activity Furthermore, they faced a white sheet. During exercise
sity in (1) a hot environment until exhaustion occurred at each subject maintained the same position, with his eyes fixed on
a core temperature of ≅40°C, and (2) a cool environment an imaginary point in front of himself.
for a similar period of time. In the latter condition At rest, the EEG signals were measured for 30 s three different
core temperature was maintained below 38°C. We specu- times (see Methods below). Resting Toes, skin temperature (Tsk)
lated that a shift in EEG from the higher frequencies and HR were recorded before exercise started. At 2-, 5- and then
at 5-min intervals during exercise, a slight touch indicated that the
(β, 13–30 Hz) to the slower wave bands (α, 8–13 Hz) in subjects should shut their eyes and relax the jaw for a sequence of
the frontal area of the brain reflects a reduced state of EEG measurements. The final EEG was taken when a sign from
arousal. subjects indicated that the point of exhaustion was close. At the
abovementioned time periods after the start of exercise, EEG was
recorded twice: the first 30 s during exercise and then during the
subsequent 30-s pause while maintaining the same body position.
Subjects and methods Exercise was resumed upon completion of the EEG measurement.
During exercise, Toes, Tsk and HR were recorded continuously. A
Subjects rating of perceived exertion (RPE) was recorded after 13 min of
exercise and at exhaustion after the final EEG measurement, using
The subjects were seven endurance-trained male cyclists, with a Borg scale [4]. After exercise subjects were weighed again. The
the following characteristics (mean ±SE): age 26.3±1.9 years, degree of dehydration was estimated from the difference in body
height 186.1±1.7 cm, mass 78.7±2.3 kg, maximal heart rate mass before and after exercise.
.
198±3 beats·min–1, and VO2max 4.83±0.12 l·min–1. The subjects
were fully informed of any risks and discomfort associated with
the experiment before giving their written consent to participate. Measurements
This study was approved by the Ethics Committee of Copenhagen
and Frederiksberg
. communities. Prior to the experimental trials Temperatures
their VO2max was assessed on a cycle ergometer using an incremen-
tal protocol,
. consisting of 4 min of exercise at 60% of the predict- Temperature measurements were recorded using thermocouples
ed VO2max, followed by four to five 1-min work stages, in which connected to electrical thermometers with an accuracy of ±0.1°C
the work intensity was increased each minute by 15%, 10%, 5% (Ellab, Copenhagen: CTF 9008, and CTD 85). Toes was measured
and then by 5% until exhaustion was reached after 5–7 min. Fur- in the deep oesophagus at a depth, measured from the nostril, cor-
thermore, subjects were familiarized
. with the experimental proto- responding to 1/4 of the subject’s standing height (Ellab probe
col by cycling for 90 min at 60% VO2max in a 20°C environment. MOV-05005-A). Skin temperature was measured at the centre of
the forehead (Tfsk) in all subjects. Furthermore, in four subjects
mean Tsk was calculated from the skin temperatures (Ellab probe
Procedure MHC-40050-A) measured at six sites (i.e. chest, back, upper arm,
forearm, thigh, lower leg).
.
After a preliminary test cycling at 20°C, 60% VO2max for 90 min,
subjects
. cycled on two occasions separated by 1–6 days, at
60% VO2max (power output 213±4 W, 80 rpm, counterbalanced or- EEG recordings
der) in a climatic chamber while wearing a water-perfused jacket.
On one trial (hyperthermic trial, H), they exercised until they Beckmann Ag/AgCl electrodes (11 mm diameter) were affixed in
reached exhaustion (34.4±1.4 min) in a hot environment (i.e. two positions on the scalp (i.e. positions F3 and F4 according to
42°C, 18% relative humidity) while the jacket was perfused with the international “ten-twenty system” [19]) with conductive
water at 42°C. During the other trial (control trial, C), in a cool en- electrode paste (Bentonite paste, Dantec). Paired mastoid refer-
vironment (i.e. 19°C, 40% relative humidity) while 18°C water ences were used. In addition, the electrodes were secured with
was circulated through the jacket, they were stopped after OMNIFIX stretch tape and a specially made elastic “helmet”. The
36.4±1.4 min. When the control experiment was performed first inter-electrode resistance was below 5 kΩ. Signal conditioners
the subjects exercised for 5–10 min longer than the predicted time (CED1902, Cambridge Electronic Design) set at a gain of 1000
to fatigue in H. Measurements taken during the 5-min period clos- and –3 dB bandwidth between 0.15 Hz and 40 Hz (roll-off
est to the time of exhaustion in the following hot trial were used 12 dB/octave) were used. Scalp potentials were sampled at
for comparison. Exercise was performed on a cycle ergometer 1000 Hz for each electrode, using a CED 1401-plus A/D con-
(Monark Ergomed 818) mounted with toe clips and triathlon han- verter. An oscilloscope was used for continuous visual inspection
dlebars in order to secure a steady working posture. On the day of the EEG signals. Signals from only F4 are presented, as they
prior to the experimental trial, subjects consumed a carbohydrate- were not different from those from F3.
rich diet with copious amounts of fluid and refrained from hard
training. On both occasions subjects arrived at the laboratory at
the same time of the day after a light meal. Subjects were weighed Data analysis
on a digital scale (Ohaus 1–10) to ±20 g wearing short bicycle
tights. Thereafter, an oesophageal probe was inserted through the Fast Fourier transformation of recorded potentials of EEG was
nose, and the skin thermocouples were attached to the skin. Heart made with CED Spike2 software using steps of 0.2 Hz for the
rate (HR) was recorded with a Polar Sports Tester (Polar Electro, power spectrum. In the power spectrum the composite EEG is bro-
Finland). The EEG electrodes were then fixed to the scalp. Subject ken down into component waves of narrow frequency bands along
rested for 30 min in a supine position in a thermally comfortable the entire EEG frequency spectrum, and the square of the ampli-
condition. Subjects then walked quietly to the climatic chamber, tudes of these components calculated. As a representative re-
where they were fitted with the water-perfused jacket. Inside the sponse, Figure 1A shows an EEG recording from a subject exer-
 43

Fig. 1 A Electroencephalographic activity (EEG) recording from Statistical analysis

F4 in an exercising subject. B Power spectrum calculated from
EEG recording after 5 min of work, and C after 25 min of work Data were analysed using a two-way (trial by time) analysis of
variance with repeated measures. Following a significant F-test,
pair-wise differences were identified using Tukey’s honestly sig-
cising with closed eyes. Panels B and C show the power spectrum nificant difference (HSD). Significance was set at the 5% level.
of a 30-s EEG recording after 5 and 25 min, respectively, of cy- Results are presented as means ±SE. Furthermore, simple linear
cling in the hot environment. regression analysis was used to describe the relationship between
The focus of the present analysis was primarily on possible the rise in core temperature and the alterations in brain activity oc-
changes in relative amplitude between the α-bands and β-bands, curring during exercise.
rather than on changes in absolute EEG amplitude within each fre-
quency. Thus, data are presented as percent changes to correct for
day-to-day and between-subject variability. The changes in α-
bands and β-bands were calculated: (1) as a percentage of the rest- Results
ing measurement obtained while sitting on the ergometer on the
particular experimental day to describe differences between rest Body temperature, heart rate and rating
and exercise, and (2) as a percentage of the first exercise measure- of perceived exertion
ment (2 min) to describe the effect of hyperthermia during exer-
cise. In preliminary experiments the effects of isolated motor ac-
tivities, e.g. eye movements, swallowing and chewing, were stud- Measurements of Toes, mean Tsk and HR during the hy-
ied. All had considerable and characteristic effects on the EEG re- perthermic and control trials are presented in Fig. 2A–C.
cording, and sections including such noise were excluded from the In the hyperthermic trial, Toes was 36.8±0.1°C at the on-
analysis. During sitting rest immediately before the onset of exer-
cise, three 30-s recordings of the EEG signal were sampled. The set of exercise and rose steadily throughout exercise, un-
highest and lowest power spectra were discarded to exclude ex- til exhaustion occurred after 34.4±1.4 min at Toes
tremes, and the remainder was selected as the best representation 39.8±0.1 °C. From the onset of exercise, mean Tsk rose
of the brain activity at rest. At rest and during exercise, the area progressively from 36.0±0.1°C to 38.4±0.4°C, whereas
under the power spectrum between 8 Hz and 13 Hz (α) and be-
tween 13 Hz and 30 Hz (β) was calculated, as was the ratio α/β forehead temperature (Tfsk) was 36.7±0.1°C and in-
(α/β index). creased to 37.8±0.2°C. HR at 2 min was 150±4 beats
min–1 and thereafter rose gradually to a nearly maximal
value of 192±3 beats min–1 (97%HRmax). RPE was
44

Fig. 2 Mean oesophageal temperature, mean skin temperature

(n=4) and heart rate during exercise in the hyperthermic and con-
trol trials. Mean ±SE for 7 (or 4) subjects. *Significantly higher
than the initial rest value (P<0.05). †Significantly different from
control (P<0.05)

13±0.4 units after 13 min of exercise and 19.0±0.8 units

at exhaustion. The degree of dehydration due to sweat
loss represented 1.7±0.1% of body mass.
In the control trial, Toes increased from 36.8±0.0°C at
Fig. 3 Mean (±SE) percentage change from resting values in the
rest to 37.8±0.1°C after 15 min of exercise reaching a EEG power spectra in the α frequency (A) and β frequency range
plateau thereafter. Mean Tsk was 34.8±0.2°C at the start (B), when exercising and paused for 30 s during 35 min of exer-
and 35.6±0.2°C after 35 min, whereas Tfsk increased cise in a hot (42°C) environment. C α power spectrum and D β
from 33.9±0.2°C at rest to 35.1±0.1 at 10 min when it power spectrum in the control (18°C) environment. *Exercise and
pause significantly higher than the initial rest value (P<0.05). †Ex-
levelled off. Therefore, at the end of the hyperthermic ercise significantly different from pause (P<0.05)
compared to the control trial, Toes was ≅2°C higher and
mean Tsk and Tfsk were ≅3°C greater. In the control trial,
HR rose to 140±3 beats·min–1 after 2 min, increasing Electroencephalographic activity
thereafter up to 159±6 beats·min–1 at the end of the C
trial. RPE was 10.9±0.8 units at 13 min of exercise and Figure 3 depicts the EEG activity during exercise and the
13.7±0.7 units at the end of exercise. The degree of de- subsequent 30-s rest intervals in both the hyperthermic
hydration in the C trial represented 1.0±0.1% of body (Fig. 3A, B) and control (Fig. 3C, D) trials. In the transi-
mass. tion from rest to exercise, the α power spectrum area, α,
 45

Fig. 4 α power spectrum areas (A), β power spectrum areas (B) Fig. 5 α/β index from the exercise measurements versus oesopha-
and exercise α/β index (C), all as a percentage of the first measure- geal temperature (A) and α/β index versus heart rate (B) during
ment made after 2 min of exercise in the hot and control conditions. the hot and control trials. Correlation for α/β index versus heart
Mean ±SE for 7 subjects. *Significantly different from 2-min value rate: Hot y=4.2–645x, r2=0.94; Control y=3.1–421x, r2=0.89.
(P<0.05). †Significantly different from control (P<0.05) Mean ±SE for 7 subjects

increased significantly, both in H and C. After 2 min of 34.4±1.4 min of exercise, respectively, compared to the
exercise α rose 131±73% above the resting value in the 2-min value (Fig. 4B). In the control trial, however, β
hot condition (see Fig. 3A) and 36±22% in the control did not decline significantly over time (Fig. 4B).
trial (Fig. 3C). The difference between trials was not sta- The calculated α/β index rose steadily throughout the
tistically significant. After 2 min of exercise the β power hyperthermic trial until exhaustion, largely due to the de-
spectrum area, β, increased 109±30% and 115±30% cline in β activity, becoming significantly elevated after
(P<0.05) at the onset of exercise in H and C, respective- 25 min with a final value of 188±71% (Fig. 4C). In the
ly (Fig. 3B, D). During the initial 20 min, β was higher control trial, the α/β index tended to increase up to
during exercise than during the subsequent 30-s pauses 59±27% at 20 min when it reached a plateau of 58±28%.
(Fig. 3B, D). For α (Fig. 3A, C) the difference between The difference between H and C was significant after
exercise and rest was not statistically different. 25 min of exercise (Fig. 4C). The same pattern of re-
Figure 4 depicts power spectrum data during exercise sponse was observed in the α/β index estimated during
expressed as a percent change from the value obtained the pauses between exercise bouts (not shown).
after 2 min of exercise. α levelled off after the first 5 min A close to linear correlation between Toes and the α/β
at 15–25% in both H and C (Fig. 4A). In contrast, β de- index measured during exercise was observed up to Toes
clined significantly throughout the hyperthermic trial, of 38°C for both C and H, continuing to 39.8°C in H
being 50±5% and 59±5% lower (P<0.05) after 25 and (r2=0.98, P=0.0001; Fig. 5A). Such tight, common corre-
46

lation for both C and H was not found between HR and Dubois et al. [11] also observed that the late components
the α/β index (Fig. 5B). of somatosensory-evoked potentials to electrical stimula-
tion during hyperthermia treatment to Toes 41–42°C dis-
appeared at the high plateau temperature, but returned to
Discussion baseline with cooling. In another study, Dubois et al.
[10] found a progressive decrease in the higher frequen-
The main finding of this study was that during prolonged cies of the EEG spectra with increasing Toes with passive
exercise in the heat the electrical activity of the prefron- heating in cancer patients, to the point that they disap-
tal cortex initially increased in the transition from rest to peared completely at ≅39°C and reappeared when cool-
exercise, and thereafter showed a progressive decline in ing was applied. They demonstrated a direct inverse rela-
the high frequency β band, in association with the pro- tionship between spectral power and temperature, in
gressive rise in body temperature until exhaustion oc- agreement with the present findings [10]. In these stud-
curred at ≅40°C. Conversely, the electrical activity of the ies, however, the patients were sedated and at various
prefrontal area of the brain reached a plateau when body levels of consciousness, which in itself could affect cor-
temperature levelled off at ≅38°C during the control tical activity [10, 11, 22]. Together, the previous and
trial. These results indicate that these alterations in EEG present evidence suggests that a high body temperature
activity in the frontal cortex reflect activity changes in has the potential to alter the function of the central ner-
parts of the brain involved in the hyperthermia-associat- vous system at rest and during exercise.
ed reduced ability to exercise. EEG frequencies normally shift from high frequen-
The present study is, to our knowledge, the first to re- cies, β-bands, towards slower wave α-bands with
port EEG recordings during ongoing exercise in humans. changes in arousal from fully alert to drowsiness or sleep
The present observation that EEG activity during ongo- [20]. In our study the calculated α/β index increased, but
ing cycling in essence displayed the same pattern of re- only β changed significantly as hyperthermia developed
sponse as during the subsequent 30-s pauses supports the and exhaustion was reached. The reason for this may be
validity of our exercise EEG measurements. Further- that the measurements were made during exercise in
more, the major finding that EEG activity during exer- contrast to earlier rest studies. In this context changes in
cise in the heat differs significantly from that in a cool the cat EEG with fatigue have been reported during fa-
environment indicates that EEG is sensitive enough to tiguing treadmill exercise [1]. High-amplitude slow
record alterations in brain activity with large differences waves occurred at the time that the animals reduced their
in body temperature. The small number of electrodes and speed because of fatigue. However, no temperature mea-
the positions were chosen merely to investigate a possi- surements were made, and thus a plausible influence of
ble shift of general frontal cortical arousal. It is impor- temperature could not be assessed. In contrast, in our hy-
tant to note that great care was taken to reduce the exter- perthermic condition no abrupt changes in the EEG oc-
nal disturbances of brain activity by having subjects curred at the point of exhaustion. Fatigue, defined as the
wear earplugs and snow glasses. In addition, subjects inability to maintain the same work output, could reflect
maintained “relaxed” body and face positions while fac- a state of reduced alertness. In the present study the cal-
ing a white sheet. culated α/β index was used as an indicator of “alert-
The previous observation that the maximal isometric ness”. A significantly higher α/β index developed in the
force of the exercised leg and the rested arm was un- hyperthermic compared to the control trial during both
changed at the point of exhaustion in hyperthermic sub- exercise and the subsequent pauses. This may be as-
jects suggests that the reduced capacity for dynamic ex- cribed to the different core temperatures between the two
ercise with hyperthermia might not be mediated by pro- conditions (see Fig. 5A) and not to fatigue due to exer-
cesses located in the primary motor cortex or the pyrami- cise as such, since the exercise conditions and duration
dal tract [25]. The abrupt increases in α and β found in were the same.
this study at the onset of cycling in both environments Furthermore, the α/β index did not correlate as well
suggest that the frontal area is activated at the start of ex- with heart rate, indicated by a clear separation between the
ercise (Fig. 3). However, the activity of the prefrontal ar- calculated regression lines between the two conditions
ea further on during exercise and subsequent pauses de- (Fig. 5B). The RPE in the hyperthermic and control trials
pended on whether the subject was hyperthermic, as also appeared to be more closely correlated than heart rate
judged by the significant decline in β frequencies only to core temperature. A tight correlation between RPE and
with high body temperature. core temperature is not surprising, and agrees with previ-
The present results from exercising humans are con- ous results obtained during prolonged exercise with com-
sistent with earlier reports of animals and human patients bined dehydration and hyperthermia or hyperthermia
indicating that hyperthermia alters brain activity as re- alone [15, 16, 17, 18, 25, 26]. The differences in heart rate
flected by EEG recordings and sensory-evoked poten- and RPE between the hyperthermic and control trials are
tials [10, 11, 22, 27]. In a study of heat stroke patients, probably associated with the different skin temperature
Mustafa et al. [22] showed that brain stem auditory- [18]. Taken as a whole, the present reduction in β activity
evoked potentials displayed characteristic changes in and increase in α/β index appears to be most closely asso-
their mid-latency response at critical high temperatures. ciated with increases in body temperature.
 47

The mechanism by which hyperthermia might act on References

brain activity appears to be complex, possibly involving
local signals as well as afferent reflexes originating in 1. Angyán L, Czopf J (1998) Exercise-induced slow waves in the
the skeletal muscle, cardiac muscle and internal organs. EEG of cats. Physiol Behav 64:267–272
2. Bligh J (1973) Temperature regulation in mammals and other
We measured Toes, which is an index of the arterial blood vertebrates. North-Holland, Amsterdam
temperature that supplies the brain [7]. Therefore, it is 3. Bligh J (1998) Mammalian homeothermy: an integrative the-
reasonable to expect that the temperature of the hypo- sis. J Therm Biol 23:143–258
thalamus and the rest of the brain of exercising humans 4. Borg G (1975) Simple rating methods for estimation of per-
increases in parallel to Toes and reaches the same or a ceived exertion. In: Borg G (ed) Physical work and effort.
Pergamon, New York, pp 39–46
higher value than core temperature, since humans do not 5. Brück K, Olschewski H (1987) Body temperature related fac-
have a functional, selective brain-cooling mechanism tors diminishing the drive to exercise. Can J Physiol Pharma-
[24, 28]. The observations that rats exercising until ex- col 65:1274–1280
haustion on a treadmill fatigued at the same core and hy- 6. Caputa M, Feistkorn G, Jessen C (1986) Effects of brain and
trunk temperatures on exercise performance in goats. Pflügers
pothalamic temperature support this notion [13, 29]. Arch 406:184–189
Caputa and co-workers [6] cleverly demonstrated that 7. Cooper KE, Kenyon JR (1957) A comparison of temperatures
brain temperature has an independent effect on behavior- measured in the rectum, oesophagus and on the surface of the
al responses in a goat exercising on a treadmill. When aorta during hypothermia in man. Br J Surg 44:616–619
8. Davis JM, Bailey SP (1997) Possible mechanisms of central
independently increasing hypothalamic temperature to nervous system fatigue during exercise. Med Sci Sports Exerc
≅43°C, they showed that the goat decreased its running 29:45–57
speed despite the fact that the temperature of the rest of 9. Dill DB, Edwards HT, Bauer PS, Levenson EJ (1931) Physical
its body remained clamped at 40°C [6]. This compelling performance in relation to external temperature. Ar-
evidence clearly points to hypothalamic temperature be- beitsphysiologie 4:508–518
10. Dubois M, Sato S, Lees DE, Bull JM, Smith R, White BG,
ing a source of signals mediating the lowering of power Moore H, MacNamare TE (1980) Electroencephalographic
output and motor activity in the goat. The temperature changes during whole body hyperthermia in humans. Electro-
centres in the hypothalamus integrate efferent and affer- encephalogr Clin Neurophysiol 50:486–495
ent signals for the control of both the autonomic and the 11. Dubois M, Coppola R, Buschsbaum MS, Lees DE (1981)
Somatosensory evoked potentials during whole body hyper-
behavioral thermoregulatory responses in animals and thermia in humans. Electroencephalogr Clin Neurophysiol 52:
humans [2, 3]. Efferent signals could involve inhibitory 157–162
reflexes in response to local high temperatures, metabo- 12. Febbraio MA, Snow RJ, Stathis CG, Hargreaves M, Carey MF
lite accumulation, substrate depletion or other chemical (1994) Effect of heat stress on muscle metabolism during ex-
ercise. J Appl Physiol 77:2827–2894
changes (neurohormones, neuromodulators, cytokines, 13. Fuller A, Carter RN, Mitchell D (1998) Brain and abdominal
heat shock protein) [8, 14]. In the present study, signals temperatures at fatigue in rats exercising in the heat. J Appl
arising in the hypothalamus could have led to a change Physiol 84:877–883
in frontal cortex activity, reflected in the decline in β ac- 14. Gandevia SC (1998) Neural control in human muscle fatigue:
tivity and increased α/β index, which could possibly un- changes in muscle afferents, moto neurones and moto cortical
drive. Acta Physiol Scand 162:275–283
derlie an important physiological mechanism of fatigue 15. González-Alonso J, Calbet JAL, Nielsen B (1998) Muscle
in hyperthermic subjects. blood flow is reduced with dehydration during prolonged ex-
We cannot rule out the possibility that other mecha- ercise in humans. J Physiol (Lond) 513:895–905
nisms sensing cardiovascular, metabolic, chemical and 16. González-Alonso J, Calbet JAL, Nielsen B (1999) Metabolic
and thermodynamic responses to dehydration-induced reduc-
thermal reflexes from the exercising muscles or other in- tions in muscle blood flow in exercising humans. J Physiol
ternal organs could also be involved in the earlier fatigue (Lond) 520:577–589
with hyperthermia [8, 12, 14, 15, 16]. Based on previous 17. González-Alonso J, Teller C, Andersen SL, Jensen FB, Hyldig
studies [17, 18], marked differences in circulatory as T, Nielsen B (1999) Influence of body temperature on the de-
velopment of fatigue during prolonged exercise in the heat.
well as metabolic parameters between the hyperthermic J Appl Physiol 86:1032–1039
and control trials could be expected after 20–30 min of 18. González-Alonso J, Mora-Rodríguez R, Coyle EF (2000)
exercise. This points to the possibility of an interaction Stroke volume during exercise: interaction of environment and
between peripheral and local brain factors in the devel- hydration. Am J Physiol 278:H321–H330
opment of the generally observed fatigue at high body 19. Jasper H (1958) The ten twenty electrode system of the Inter-
national Federation. Electroencephalogr Clin Neurophysiol
temperatures. 10:371–375
In conclusion, the present findings indicate that dur- 20. Kandel ER, Schwartz JH (1985). Principles of neural science,
ing exercise in the heat, the core temperature rises and 2nd edn. Elsevier, Amsterdam
that there is a concomitant shift in the EEG power distri- 21. MacDougal JD, Reddan WG, Layton CR, Dempsey JA
(1974). Effects of metabolic hyperthermia on performance
bution, a decrease in β and a steady increase in the α/β during heavy prolonged exercise. J Appl Phyisol 36:538–544
index. These observations suggest that alterations in the 22. Mustafa MK, Kogali YM, Gumaa KA (1987). Central nervous
electrical activity of the brain’s frontal area indicate hy- system, blood clotting and respiratory function associated with
perthermia-associated fatigue. heat stroke. In: Hales JRS, Richards DAB (eds) Heat stress:
physical exertion and environment. Elsevier, Amsterdam,
pp 277–288
Acknowledgement This study belongs to a series of studies that 23. Nielsen B (1992) Heat stress causes fatigue. Med Sport Sci
were supported by grants from Team Denmark. 34:207–217
48
24. Nielsen B, Jessen C (1992) Evidence against brain stem cool- 27. Reeves DL, Justesen DR, Levinson DM, Riffle DW (1985)
ing by face fanning in severely hyperthermic humans. Pflügers Endogenous hyperthermia in normal human subjects: I. Exper-
Arch 422:168–172 imental study of evoked potentials and reaction time. Physiol
25. Nielsen B, Hales JRS, Strange S, Juel Christensen N, Warberg Psychol 13:258–267
J, Saltin B (1993) Human circulatory and thermoregulatory 28. Shiraki K, Sagawa S, Tajima F, Yokota A, Hashimoto M,
adaptations with heat acclimation and exercise in a hot, dry Brengelmann GL (1988) Independence of brain and tympanic
environment. J Physiol (Lond) 460:467–485 temperatures in an unanesthetized human. J Appl Physiol
26. Nielsen B, Strange S, Juel Christensen N, Warberg J, 65:482–486
Saltin B (1997) Acute and adaptive responses in humans to 29. Walters TJ, Ryan KJ, Tate LM, Mason PA (2000) Exercise in
exercise in warm, humid environment. Pflügers Arch 434:49– the heat is limited by a critical internal temperature. J Appl
56 Physiol 89:799–806