Secondary Logo

Journal Logo


Inadequate Cerebral Oxygen Delivery and Central Fatigue during Strenuous Exercise

Nybo, Lars; Rasmussen, Peter

Author Information
Exercise and Sport Sciences Reviews: July 2007 - Volume 35 - Issue 3 - p 110-118
doi: 10.1097/jes.0b013e3180a031ec
  • Free


Fatigue emerges during exercise as increased difficulty in retaining a required power or as impairment in the ability to produce force and power. The development of fatigue is complex and determined by an intricate interplay between psychological and physiological factors. Even in well-motivated subjects, the relative importance of muscular/peripheral and central factors seems to vary markedly depending on the mode, intensity, and duration of the exercise, besides the nutritional status of the subjects and the environmental setting. Some physiological factors may relate directly or indirectly to homeostatic disturbances in the skeletal muscles, whereas other factors may directly affect the central nervous system (CNS) and its ability to activate the skeletal muscles via the alpha motor neurons.

Previously, the "central fatigue hypothesis" has focused mainly on changes in extracellular neurotransmitter levels or exercise-induced alterations in the activity of different neurotransmitter systems, with the serotonin, dopamine, and ammonia-glutamate-glutamine "hypotheses" as the most dominant (3,20,24). The present review discusses the possibility that, during some exercise conditions, fatigue can be provoked or modulated by inadequate oxygen delivery to the brain and subsequently low cerebral capillary and mitochondrial oxygen tension (PO2). In turn, this could influence the function of neurons and astrocytes and thereby the ability to maintain motor activation. Although it is recognized that this impairment can occur during exercise at high altitude, it may also arise as a consequence of exercise-induced arterial hypoxemia (EIAH) or exercise conditions where hyperventilation-induced hypocapnia lowers cerebral blood flow (CBF). Hypocapnia may be provoked by exercise at intensities above the ventilatory threshold and, especially, exercise with hyperthermia, which may lower the cerebral perfusion by 20%-30% (23,24).

At rest, the brain is protected against hypoxia-induced reductions in arterial oxygen delivery because CBF increases when the arterial oxygen tension becomes low. When exercise and hypoxemia are combined, however, hyperventilation-induced reductions of the arterial carbon dioxide tension (PaCO2) may become so pronounced that it blunts the increase in CBF, and therefore, increased perfusion fails to compensate for the lower arterial oxygen content. Anecdotes from athletes experiencing blackout or fainting immediately after completing maximal exercise support the idea that oxygen delivery to the brain may become critically low, and studies of maximal rowing in subjects with EIAH show that the cerebral oxygenation falls during such exercise. Although athletes rarely faint during exercise studies in the laboratory, studies with different combinations of strenuous exercise and impaired oxygen delivery to the brain indicate that the CNS function may be compromised and that mitochondrial oxygen tension (Pmito) in the brain may become critically low. At rest and during moderate exercise, the brain is well perfused, and moderate reductions in CBF/cerebral oxygen delivery may be compensated for by increased extraction. However, a relatively high capillary oxygen tension is required to maintain the mitochondrial PO2 at an adequate level. The main questions raised in the present review are the following: 1) How large reduction in cerebral oxygen delivery and subsequent lowering of the mitochondrial PO2 can the brain tolerate before it begins to influence motor performance and contribute to the development of so-called central fatigue? 2) Which exercise conditions cause such reductions, and how can we monitor and evaluate whether cerebral oxygenation levels are sufficient?


Reductions in the arterial oxygen content (CaO2) lowers maximal work rate and impairs exercise endurance during whole-body exercise. Because changes in CaO2 influences oxygen delivery to all tissues, including the skeletal muscles and the brain, it may be difficult to determine whether performance deteriorates as direct effect of the low PO2 in some of the most active brain regions or whether central motor output becomes reduced secondary to feedback from the muscles and from the increased cardiorespiratory stress that accompanies exercise with a low atmospheric oxygen tension. At moderate hypoxia, the reduced work rate during maximal whole-body exercise is influenced by feedback from the working muscles. Reduced oxygen delivery to exercising muscles accelerates the accumulation of muscle metabolites, and these fatigue-related and sensory fiber-stimulating substances may activate group 3 and 4 muscle afferents and contribute to inhibitory feedback to the CNS during fatiguing exercise. In contrast, during severe hypoxia, afferent feedback seems to be of minor importance because epidural anesthesia (blocking or reducing feedback from group 3 and 4 muscle afferents) has no effect on performance or perceived exertion during maximal exercise, when the inspired oxygen fraction was lowered to approximately 10% (17).

Further support for a direct effect of reduced cerebral oxygen delivery on motor performance is provided by the study presented in Figure 1. In that study, cerebral oxygen delivery was varied by providing low, normal, or high PO2 in the inspired air either with normal, high, or reduced arterial CO2 tension, and motor performance was evaluated as maximal handgrip strength. When oxygen delivery was reduced by more than 15% below control levels, as a separate effect of either hyperventilation-induced reduction in CBF or inhalation of air with a low PO2, the maximal handgrip strength decreased concurrently with an increase in lactate spillover from the brain, which indicated that cerebral oxygen levels became inadequate to support optimal aerobic metabolism (Fig. 1 and "Cerebral Metabolism and Mitochondrial PO2" section). Although low arterial PO2 or hyperventilation-induced alkalosis may have influenced the ability of the muscles to produce a maximal force, the energy used in the skeletal muscles during a brief maximal contraction is delivered mainly by anaerobic metabolism (net adenosine triphosphate (ATP) and creatine phosphate degradation). It is likely, therefore, that the impaired motor performance observed in the study by Rasmussen et al. (29) relates to central fatigue arising subsequently from inadequate oxygen delivery to the brain.

Figure 1
Figure 1:
Cerebral lactate release (top) and maximal voluntary handgrip force (bottom) as a function of cerebral oxygen delivery (percent control). Alterations in central blood flow (CBF) (induced by CO2 breathing, hyperventilation, or normal ventilation) were combined with different levels of CaO2 and PaO2 (10%, 21%, or 100% oxygen in the inspired air) to vary the cerebral oxygen delivery from approximately 60% to approximately 140% of control (supine rest; see 29 for details). Data are mean ± SE for 12 subjects.

Central fatigue may not always appear during a single brief maximal voluntary contraction (MVC), and it may be necessary to use a protocol with repeated or sustained maximal contractions. Both for hypoglycemia- and hyperthermia-induced central fatigue (see 24 for review), well-motivated subjects are capable to establish a maximal neural drive to the muscle for a brief period, despite pronounced fatigue, whereas the ability to sustain a high firing rate of the alpha motor neurons for more than a few seconds becomes markedly impaired. Accordingly, MVC of a small and rested muscle group may be unaffected by exposure to moderate altitude (6), whereas others have reported (4,7) on the effect of extreme simulated altitude: "in some subjects, the responses to stimuli interpolated during repeated MVC provided evidence of central fatigue at altitude." If a period of rest is allowed between or before brief maximal efforts, there may be enough time for replenishment of brain glycogen and restoration of high-energy phosphate compounds in the relevant brain areas. In contrast, low mitochondrial oxygen tension during sustained motor activation may not allow for adequate maintenance of neuronal and astrocytic homeostasis, and eventually, this will impair the ability to maintain a high level of neuronal firing.


Because sustained static and dynamic exercise requires continuous or frequent repetitive neuronal firing in several motor areas, the brain seems to be more vulnerable to hypoxia during exercise. Furthermore, an elevated CBF does not compensate for a low arterial PO2 during strenuous exercise because pronounced hyperventilation will lower arterial carbon dioxide tension (PCO2) and cause constriction (less vasodilation) of the arterioles in the brain. This seems irrational because the primary function of the cerebral circulation is to ensure homeostasis in the brain by providing oxygen for mitochondrial respiration and removal of waste products from the cerebral metabolism. Cerebral autoregulation, metabolic regulation, and CO2 (hydrogen ion; H+)-mediated vasodilatation are the most important mechanisms to ensure that CBF remains adequate. However, CO2 reactivity seems to dominate, and during exercise associated with hyperventilation-induced reductions of the arterial CO2 tension, CBF may decline despite increased metabolic activity in motor areas of the brain (24). Both at rest and during exercise, the cerebral perfusion is influenced strongly by PaCO2, and the cerebral CO2 reactivity (percentage change in CBF per millimeters of mercury change in PaCO2) may even increase from approximately 3%-4% at rest to 4%-5% during exercise (13,23,28). It is not unusual to see reductions of the arterial PCO2 by 6-10 mm Hg during maximal exercise and exercise with hyperthermia, which, in turn, may reduce global CBF by 30% compared with moderate intense exercise at a normal temperature response (23,28).

Cerebral blood flow is distributed heterogeneously and largely depends on the neuronal activity in the different regions of the brain. Dynamic exercise is associated with activation of several areas of the brain, and when the intensity does not exceed the ventilatory threshold (and arterial PCO2 remains fairly stable), flow to these activated areas will increase linearly with the exercise intensity (13,30). When the intensity exceeds the ventilatory threshold, both global and regional CBF decline as PaCO2 decreases (Fig. 2). Although neuronal activity and metabolic needs keep increasing in the motor areas, regional CBF will decline as exercise intensity increases toward maximum, and regional CBF (e.g., to the motor cortex) may return to similar levels as at rest. According to equation 1 in "Cerebral Metabolism and Mitochondrial PO2" section, this uncoupling of metabolism and flow will cause a reduction in the mitochondrial PO2 in the activated motor areas.

Figure 2
Figure 2:
Middle cerebral artery blood flow velocity during incremental exercise, with (open symbols) and without (solid symbols) "adjustments" for arterial carbon dioxide tension (Rasmussen, P., H.H. Stie, B. Nielsen, et al., unpublished manuscript/observations, 2005). Data are mean ± SE for 7 subjects.

However, exercise causes hemoconcentration (secondary to increased filtering of plasma in the muscle capillaries) and, based on the assumption that the arterial saturation remains unchanged, CaO2 may increase by 5%-10% during exercise. Therefore, with a normal atmospheric (and arterial) PO2, the exercise-induced hemoconcentration causes the oxygen delivery to motor areas of the brain to be slightly higher during maximal exercise compared with that during rest but lower than that during submaximal exercise and lower than would be expected on the basis of the neuronal activity in the activated motor areas (14). Nonetheless, oxygen supply seems to be sufficient to support the aerobic metabolism in the brain during maximal exercise (Table). However, in subjects with severe EIAH, which may lower the arterial PO2 to less than 70 mm Hg and saturation to less than 88%, the cerebral oxygen delivery may be disturbed to an extent where it begins to influence aerobic metabolism in the brain as Pmito may be reduced by approximately 10 mm Hg. Similarly, low atmospheric and arterial PO2 when exercising at a high altitude may reduce the cerebral capillary oxygenation levels and limit mitochondrial PO2 to support the cerebral metabolism. At rest, lactate spillover increases when the cerebral oxygen delivery is reduced by more than approximately 15% and the mitochondrial PO2 decreases by more than 6-7 mm Hg (Fig. 1; see 29). It is more difficult to use increased lactate spillover as an indicator of inadequate oxygen delivery during strenuous exercise because the brain begins to metabolize lactate during high intensity exercise (2). Although endogenous lactate production in some brain areas may increase with strenuous exercise, it is likely that such an increase will be hidden by the high arterial lactate levels and the marked increase in the cerebral uptake of carbohydrates (including both blood glucose and lactate) in response to brain activation (2).

Comparisons of cerebral mitochondrial oxygen tension, cerebral metabolism, oxygen uptake, delivery, capillary saturation, and tension at rest during submaximal exercise and in response to strenuous exercise conditions.

Although exercise with superimposed hypoxia may compromise cerebral autoregulation and the ability to increase and to maintain an adequate cardiac output seems to influence the CBF response to exercise, the influence of changes in CBF arising secondary to changes in mean arterial pressure are of minor importance for maintenance of oxygen delivery to the brain. Thus, compared with the relatively large changes in CBF and arterial oxygen content with changes in arterial CO2 and O2 tension, the influence of arterial blood pressure is much smaller, and because mean arterial pressure increases during exercise, this would support rather than compromise oxygen delivery to the brain.


Because the brain relies on aerobic metabolism, it is clear that the oxygen tension within the mitochondria is critical for the cerebral function. However, it is not clear what the critical PO2 level is. Thus, impairment of the mitochondrial oxidation has been observed in the range of approximately 0.5-20 mm Hg (depending on the methodology used for evaluation of Pmito), and the physiological level in a resting human is estimated to range from close to zero and up to 40 mm Hg (5,27). Obviously, it is not possible to measure Pmito directly in vivo, and the wide range of the reported values for Pmito may relate to the methodological differences and difficulties; however, it may also relate to large differences within the brain. Although, the regional CBF is matched largely to the metabolic activity, flow and energy turnover are distributed heterogeneously in the brain, and there may be relatively large differences across brain regions. Pmito depends on the balance between oxygen supply and use and the oxygen conductance from the capillary to the mitochondria (L). Thus, the mitochondrial oxygen tension is determined by the capillary PO2, the oxygen use (cerebral metabolic rate (CMR); CMRO2), and L as expressed by equation 1 (8,29):

As the brain lacks capillary recruitment, O2 diffusibility (L) remains stable at any measured blood flow level and is determined at the lowest capillary oxygen tension, where Pmito is assumed to approach zero. PCap can be estimated from measurement of arterial and cerebral venous blood,


is the PO2 when hemoglobin is half saturated, and ha is Hill coefficient for arterial blood.

The average oxygen consumption may be calculated on the basis of the cerebral a-v difference for oxygen and CBF. However, it may not be simple to estimate flow during strenuous whole-body exercise, and interpretation and comparison of results from different studies require care, including consideration of whether a regional or a global CBF has been estimated. Mixed cerebral venous blood may be obtained from the internal jugular vein, but this blood may not be representative of all brain regions, and it provides only a measure of the average oxygen tension and saturation of the blood leaving the brain. Venous blood from active brain areas may have different oxygen content. Nevertheless, this method may provide the best functional estimate of changes in Pmito in exercising subjects. The calculation does not allow for an absolute determination of the cerebral Pmito, but it provides a useful indication of changes in the average cerebral Pmito relative to a basal level and thus the global oxygen state of the brain. As illustrated in Figure 1, it seems that a reduction in the average Pmito by more than 6-7 mm Hg - either induced by arterial hypoxia or reduced CBF - is associated with impaired cerebral aerobic metabolism (as increased lactate spillover indicates that aerobic pyruvate metabolism cannot match the required glycolytic rate). In addition, there is a reduced ability to activate the motor neurons at this Pmito level, implying that a drop by more than 6-7 mm Hg in Pmito or a reduction in cerebral oxygen delivery by more than 15% needs to be prevented for optimal motor function of the brain. Brain cells can survive for a relatively long time on anaerobic energy turnover; glycolytic metabolism by the astrocytes may protect against cell damage, and for a limited period, it may postpone fatigue. However, anaerobic metabolism of a glucose molecule only produces energy for resynthesis of 2 ATP, whereas aerobic metabolism of glucose is much more efficient and may provide up to 36 ATP. Thus, even if the brain cells could get rid of the produced lactate and maintain the lactate dehydrogenase-catalyzed conversion of pyruvate to lactate and concomitant interconversion of the reduced form of nicotinamide adenine dinucleotide to the oxidized form, NAD+ (as required to allow for further breakdown of glucose to pyruvate via the Embden-Meyerhof pathway), the brain would have to increase its glucose uptake substantially. In accordance, the cerebral glucose uptake increases during hypoxia and maximal exercise (2,29). However, during hypoxia, the increase is modest (2), and the enhanced cerebral uptake and anaerobic metabolism of glucose are not sufficient to maintain optimal function of the brain as judged from the impaired motor activation after 10-min exposure to hypoxia (29).

These observations have been made in resting subjects, and the tolerable reduction in Pmito may be even lower when the brain is activated by strenuous exercise, which raises the metabolism in the involved motor areas of the brain. Thus, the Pmito depends on the integration of supply and demand, and increased regional CMRO2 will lower the cerebral Pmito (equation 1) unless flow in the active brain area increases to compensate for the increased use. To maintain Pmito, or at least avoid critically large reductions of Pmito in the active brain areas, the capillary oxygen tension and saturation must be kept at a relatively high level, and this requires that the increase in flow is proportionally larger than the increase in metabolism. There is practically no capillary recruitment in the brain; the oxygen extraction rises linearly with distance as the blood traverses the capillary network from arterial to venous ends, and the capillary geometry is such that all segments of the capillary bed supply equal amounts of brain tissue (15). Therefore, compared with muscle tissue, the maximal extraction fraction is much lower for the brain.

Capillary Oxygen Tension and Saturation

As described by equation 2, the oxygen saturation and tension of the capillaries may be estimated on the basis of arterial and cerebral venous blood sampling. With this methodology, the lowest capillary saturation level consistent with unimpaired metabolic function of the brain is estimated to approximately 70%, corresponding to a capillary PO2 between 35 and 40 mm Hg (depending on temperature, pH, and CO2 level). In resting subjects with a normal PaO2, the highest extraction fraction (a-v difference for oxygen divided by arterial oxygen content) seems to be 55%-60%. Presyncope symptoms emerge during head-up tilt when the jugular venous saturation becomes reduced by approximately 15% (19), corresponding to a oxygen extraction fraction of approximately 55%. However, the global cerebral oxygen extraction rarely exceeds 50% during exercise (Table). Because physical or mental activation of the brain involves increased energy turnover in the active regions, whereas other regions may have unchanged or a decreased metabolic rate, the mixed venous blood may not reflect oxygen extraction in the activated brain regions. Furthermore, the highest obtainable global extraction fraction during hypoxia may become even lower because the PO2 gradient between the capillary and mitochondria will be too low to allow for sufficient oxygen flux/diffusion from blood to the mitochondria. Accordingly, the CMR for oxygen declines during exercise with severe hypoxia, and oxygen delivery to the brain becomes inadequate, although the oxygen extraction fraction is below 50% (Table).

Cerebral metabolism and oxygenation can be estimated noninvasively by determining capillary-oxygenation-level-dependent (COLD) monitoring with near-infrared spectroscopy (NIRS). Near-infrared spectroscopy works by emitting light at a defined wavelength into the tissue; after absorbance by hemoglobin that depends on its conformal state, some of the emitted light reaches to an optode, and oxygenation can be calculated according to a modified Lambert-Beer law. Near-infrared spectroscopy can be applied over any region of the skull; however, for subject's comfort, the forehead is usually preferred. The penetration depth of NIRS remains unknown in vivo but is likely restricted to a few centimeters. Although NIRS is a relative measure, it faithfully tracks the capillary oxygenation level as estimated from arterial and jugular venous blood saturation (29). The COLD signal, nonetheless, still depends on the chosen algorithm, and the signal may vary some 5% from "true" blood oxygenation, depending on the apparatus (29). Therefore, NIRS provides a rough estimate of capillary oxygen saturation, and the COLD signal is usually converted to hemoglobin oxygen saturation to provide a useful measure of the relation between oxygen availability and demand.

Cerebral Oxygen Uptake and Metabolism

Comparison of the CMR for oxygen during different exercise conditions could indicate whether oxygen delivery has been adequate to support the aerobic metabolism. Although global and regional CMRO2 may be determined with various methods during strenuous exercise, the estimate is associated with a host of problems. The rigorous movements associated with strenuous exercise preclude the use of advanced imaging techniques such as functional magnetic resonance imaging and positron emission tomography.

Therefore, evaluations of cerebral metabolism during whole-body exercise have concentrated on measurements of CBF and the arteriovenous differences across the brain. However, the cerebral circulation receives blood from both the carotid and spinal arteries, making evaluation of the inflow troublesome. Furthermore, the venous drainage goes to both the spinal venous plexus and the internal jugular veins. The internal jugular vein is relatively easily catheterized and, therefore, used for sampling of cerebral venous blood but may not represent venous blood from all regions of the brain. To further complicate the evaluation, contribution to the drainage of each vein depends on body position; it is asymmetrical and may vary between subjects (29). Despite these problems, the standard has been considered the Kety-Schmidt method, which derives CBF based on Fick principle - that is, wash out of a freely diffusible tracer determined from arterial and internal jugular venous blood samples (25). Although the evaluation of the CBF with the Kety-Schmidt technique provides a measure of jugular venous flow, the method can provide useful values of global CBF at rest and during submaximal exercise (18). However, a major difficulty with this method for determination of CBF and calculation of the CMRO2 during maximal exercise is the time resolution of the method. The Kety-Schmidt method requires approximately 30 min of tracer infusion, followed by at least 10 min of blood sampling, which is not compatible with maximal whole-body exercise.

Alternatively, CBF can be evaluated continuously by ultrasound Doppler sonography in the carotid artery or the basal cranial arteries. The exercise response in the basal arteries is similar to that in the carotid arteries (9), thereby indicating some relation between changes in regional and global CBF. However, determination of flow or flow velocities in the basal cerebral arteries estimates regional rather than global CBF, and data from studies using different methods for evaluation of CBF must be interpreted with caution (18). A crucial assumption in the interpretation of Doppler-derived flow velocity data is an unchanged caliber of the insonated vessel; however, it seems that vascular tone (vessel diameter) is regulated distally to the basal cerebral vessels, and in general, CBF velocities obtained by transcranial ultrasound Doppler seems to reflect regional CBF (13).

When evaluated with the Kety-Schmidt technique, whole-brain blood flow and CMRO2 seem to be unchanged during exercise compared with those during rest, although the 10%-25% increase in flow derived by ultrasound Doppler assessment of the arterial inflow to the brain or in the basal cerebral vessels indicates that flow to large parts of the brain increases in response to motor activation (10,11,13). However, even if global or regional CBF and CMRO2 may be determined, an unchanged or even elevated CMRO2 does not necessarily imply that oxygen delivery has been adequate. It may be argued that oxygen delivery is not sufficient to allow for a further increase in the CMRO2, which may be required during periods with intense neuronal firing (e.g., mentally demanding exercise).


During submaximal exercise, oxygen delivery to the brain seems to be more than adequate because Pmito increases. There is an unchanged lactate release and global metabolic rate during such exercise. However, Pmito and capillary oxygen saturation during strenuous exercise (Table) declines when hyperventilation-induced hypocapnia lowers or blunts an increase in CBF. During the three conditions where PaO2 and arterial oxygen saturation (SaO2) remain fairly unchanged (i.e., maximal exercise and submaximal and maximal exercise with hyperthermia), the drop in Pmito is limited to 5-6 mm Hg; therefore, it is below the level where Pmito begins to restrict the cerebral metabolic function and motor performance. The 5- to 6-mm Hg drop in Pmito is close to the reduction that may be tolerated without impairment in cerebral function, as confirmed by that lack of a change in lactate spillover from the brain during submaximal exercise with hyperthermia and during maximal exercise with or without hyperthermia, and the marked lactate uptake by the brain when the arterial lactate concentration rises. When cerebral oxygen delivery becomes inadequate to support aerobic metabolism during exercise with severe hypoxia, the brain does not take up lactate, although the arterial concentration increases to similar levels as during maximal exercise. The reduction in Pmito during submaximal exercise with progressive hyperthermia is similar to that of maximal exercise, but arterial lactate remains low (~2 mM), and there is no change in lactate balance across the brain. Taken together, these observations indicate that a reduction in cerebral Pmito of 5-6 mm Hg may be tolerated without impaired aerobic energy turnover and, consequently, without changes in anaerobic metabolism of glucose in the brain.

The observation of a 7% increase in the CMRO2 during exercise with hyperthermia, despite an approximately 20% reduction in CBF, provides support of an adequate oxygen delivery to the brain. On the other hand, the cerebral temperature at exhaustion, which is higher by 1.5°C-2°C if we compare with normothermic exercise and increased by approximately 3°C compared with rest, would predict an even greater increase in the CMRO2. Assuming a Q10 of 2, (which is the normal Q10 temperature coefficient for metabolic processes including the cerebral metabolism (24)), CMRO2 should increase by at least 13% compared with that in control exercise. Therefore, it cannot be excluded that low oxygen delivery during submaximal exercise with hyperthermia restricts an even larger increase in the CMRO2, and the blackout or fainting occasionally experienced by hyperthermic athletes may relate to inadequate oxygen delivery to the brain. However, the central fatigue arising with hyperthermia during submaximal exercise is probably largely related to a direct inhibitory effect of high brain temperature on motor outflow (24). Likewise, factors other than cerebral oxygen delivery seems to be of greater importance for the development of fatigue during maximal exercise. When maximal exercise is performed during hyperthermic conditions, cerebral oxygenation may decline at a faster rate than that during normothermic conditions (9), and that could be involved with the reduced time to exhaustion; however, it seems more likely that the marked reduction in cardiac output and subsequent impairment of oxygen delivery to the exercising muscles is the factor involved with reduced performance during maximal exercise with hyperthermia.

In contrast, when arterial hypoxia is superimposed either during exposure to high altitude or as consequence of EIAH, the reduction in Pmito is so pronounced that inadequate oxygen delivery to the brain may become a significant factor influencing the development of fatigue. As calculated from the data by Imray et al. (12) during exercise on Chacaltaya at 5260 m above sea level and by Rasmussen et al. (unpublished manuscript/observation, 2006) during simulated altitude (~6000 m), the reductions in cerebral oxygen delivery, cerebral oxygenation, and Pmito are likely to impair both the cerebral metabolism and motor performance. During exercise with severe arterial hypoxia, the brain does not display a net uptake of lactate, despite increases in arterial lactate to similar levels as during maximal exercise with normoxia, indicating that the cerebral oxygen status does not allow for an uptake and metabolism of lactate, and it is likely that endogenous lactate production by the brain becomes so large that the lactate gradient between the blood and the brain disappears. Furthermore, during acute hypoxia, the exercise CMRO2 is lower than that during exercise with normal PaO2 when both compared with the same absolute exercise intensity or matched to exercise, resulting in similar arterial lactate levels, cardiovascular stress, and perceived exertion (Rasmussen, Nybo, Peterson, et al., unpublished manuscript/observation, 2006). In contrast, acclimatized subjects exhibited no difference in CBF, oxidative metabolism, or lactate release when exercise performed on Chacaltaya was compared with rest or submaximal exercise at sea level (21). The differences between this study (21) and those by Imray et al. (12) and Rasmussen et al. (unpublished manuscript/observation, 2006) may relate to the differences between acute exposure and responses in acclimatized subjects, but it may also relate to the differences in exercise intensities. The workload in the study by Moller et al. (21) was rather low, and it seems that cerebral oxygen delivery in acclimatized subjects may be adequate at rest and low exercise intensities but becomes challenged as maximal workload is approached (Fig. 3).

Figure 3
Figure 3:
Changes in cerebral oxygen delivery at different altitudes (▪ 150 m; ♦, 3,610 m; •, 4,750 m; ▴, 5,260 m). (Reprinted from Imray, C.H., S.D. Myers, K.T. Pattinson, A.R. Bradwell, C.W. Chan, S. Harris, P. Collins, and A.D. Wright. Effect of exercise on cerebral perfusion in humans at high altitude. J. Appl. Physiol. 99:699-706, 2005. Copyright © 2005 The American Physiological Society. Used with permission.)

When subjects with EIAH perform maximal exercise, the cerebral oxygen delivery may decrease by approximately 25% because of the combined effect of reduced arterial saturation and lowered CBF, and this may cause reductions in Pmito from 6 and up to 13 mm Hg, depending on the level of arterial hypoxemia and hyperventilation-induced reduction of CBF. In endurance-trained subjects, SaO2 may decrease to below 88% and hyperventilation-induced reductions in PaCO2 by 8-10 mm Hg during maximal exercise (22), and Pmito is reduced by more than 10 mm Hg and to such an extent that it impairs motor performance. Similarly, Nielsen et al. (22) reported that well-trained rowers experiencing a drop in SaO2 from 98% to 92% during maximal exercise also displayed reduced NIRS-determined COLD from 80 to 65 mm Hg; however, cerebral desaturation was prevented, and performance improved when oxygen supplementation was provided and the arterial oxygen saturation maintained. Although the improvement in performance may relate to enhanced oxygen delivery to the exercising muscles and prevention of muscle fatigue (1), the reduction in cerebral oxygen saturation and Pmito is of such magnitude that it may affect motor performance during maximal exercise in subjects with EIAH. Therefore, oxygen supplementation may also contribute to improved performance by preventing central fatigue arising secondary to reductions of Pmito.


In summary, strenuous exercise conditions that cause hyperventilation-induced restrictions of CBF may compromise oxygen delivery to active brain areas and lower the cerebral Pmito, but reductions in cerebral oxygenation become relevant for the development of central fatigue primarily when arterial hypoxemia is superimposed. Under such conditions, the mitochondrial oxygen tension may decline by more than 10 mm Hg as compared with rest, and it is no longer adequate to support the cerebral metabolism and maintain motor function.


1. Amann, M., M.W. Eldridge, A.T. Lovering, M.K. Stickland, D.F. Pegelow, and J.A. Dempsey. Arterial oxygenation influences central motor output and exercise performance via effects on peripheral locomotor muscle fatigue in humans. J. Physiol. 575:937-952, 2006.
2. Dalsgaard, M.K. Fuelling cerebral activity in exercising man. J. Cereb. Blood Flow Metab. 26:731-750, 2006.
3. Davis, J.M., and S.P. Bailey. Possible mechanisms of central nervous system fatigue during exercise. Med. Sci. Sports Exerc. 29:45-57, 1997.
4. Dousset, E., J.G. Steinberg, N. Balon, and Y. Jammes. Effects of acute hypoxemia on force and surface EMG during sustained handgrip. Muscle Nerve 24:364-371, 2001.
5. Duong, T.Q., C. Iadecola, and S.G. Kim. Effect of hyperoxia, hypercapnia, and hypoxia on cerebral interstitial oxygen tension and cerebral blood flow. Magn. Reson. Med. 45:61-70, 2001.
6. Fulco, C.S., A. Cymerman, S.R. Muza, P.B. Rock, K.B. Pandolf, and S.F. Lewis. Adductor pollicis muscle fatigue during acute and chronic altitude exposure and return to sea level. J. Appl. Physiol. 77:179-183, 1994.
7. Garner, S.H., J.R. Sutton, R.L. Burse, A.J. McComas, A. Cymerman, and C.S. Houston. Operation Everest II: neuromuscular performance under conditions of extreme simulated altitude. J. Appl. Physiol. 68:1167-1172, 1990.
8. Gjedde, A., P. Johannsen, G.E. Cold, and L. Ostergaard. Cerebral metabolic response to low blood flow: possible role of cytochrome oxidase inhibition. J. Cereb. Blood Flow Metab. 25:1183-1196, 2005.
9. Gonzalez-Alonso, J., M.K. Dalsgaard, T. Osada, S. Volianitis, E.A. Dawson, C.C. Yoshiga, and N.H. Secher. Brain and central haemodynamics and oxygenation during maximal exercise in humans. J. Physiol. 557:331-342, 2004.
10. Hellstrom, G., W. Fischer-Colbrie, N.G. Wahlgren, and T. Jogestrand. Carotid artery blood flow and middle cerebral artery blood flow velocity during physical exercise. J. Appl. Physiol. 81:413-418, 1996.
11. Ide, K., and N.H. Secher. Cerebral blood flow and metabolism during exercise. Prog. Neurobiol. 61:397-414, 2000.
12. Imray, C.H., S.D. Myers, K.T. Pattinson, A.R. Bradwell, C.W. Chan, S. Harris, P. Collins, and A.D. Wright. Effect of exercise on cerebral perfusion in humans at high altitude. J. Appl. Physiol. 99:699-706, 2005.
13. Jorgensen, L.G. Transcranial Doppler ultrasound for cerebral perfusion. Acta Physiol. Scand. Suppl. 625:1-44, 1995.
14. Kastrup, A., G. Kruger, T. Neumann-Haefelin, G.H. Glover, and M.E. Moseley. Changes of cerebral blood flow, oxygenation, and oxidative metabolism during graded motor activation. Neuroimage 15:74-82, 2002.
15. Kety, S.S. Determinants of tissue oxygen tension. Fed. Proc. 16:666-671, 1957.
16. Kety, S.S., and C. Schmidt. The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am. J. Physiol. 143:53-66, 1945.
    17. Kjaer, M., B. Hanel, L. Worm, G. Perko, S.F. Lewis, K. Sahlin, H. Galbo, and N.H. Secher. Cardiovascular and neuroendocrine responses to exercise in hypoxia during impaired neural feedback from muscle. Am. J. Physiol. 277:R76-R85, 1999.
    18. Madsen, P.L., S. Holm, M. Herning, and N.A. Lassen. Average blood flow and oxygen uptake in the human brain during resting wakefulness: a critical appraisal of the Kety-Schmidt technique. J. Cereb. Blood Flow Metab. 13:646-655, 1993.
    19. Madsen, P., F. Pott, S.B. Olsen, H.B. Nielsen, I. Burcev, and N.H. Secher. Near-infrared spectrophotometry determined brain oxygenation during fainting. Acta Physiol. Scand. 162:501-507, 1998.
    20. Meeusen, R., P. Watson, H. Hasegawa, B. Roelands, and M.F. Piacentini. Central fatigue: the serotonin hypothesis and beyond. Sports Med. 36:881-909, 2006.
    21. Moller, K., O.B. Paulson, T.F. Hornbein, W.N. Colier, A.S. Paulson, R.C. Roach, S. Holm, and G.M. Knudsen. Unchanged cerebral blood flow and oxidative metabolism after acclimatization to high altitude. J. Cereb. Blood Flow Metab. 22:118-126, 2002.
    22. Nielsen, H.B., R. Boushel, P. Madsen, and N.H. Secher. Cerebral desaturation during exercise reversed by O2 supplementation. Am. J. Physiol. 277:H1045-H1052, 1999.
    23. Nybo, L., and B. Nielsen. Middle cerebral artery blood velocity is reduced with hyperthermia during prolonged exercise in humans. J. Physiol. 534:279-286, 2001.
    24. Nybo, L., and N.H. Secher. Cerebral perturbations provoked by prolonged exercise. Prog. Neurobiol. 72:223-261, 2004.
    25. Nybo, L., K. Moller, S. Volianitis, B. Nielsen, and N.H. Secher. Effects of hyperthermia on cerebral blood flow and metabolism during prolonged exercise in humans. J. Appl. Physiol. 93:58-64, 2002.
    26. Pott, F., C.A. Ray, H.L. Olesen, K. Ide, and N.H. Secher. Middle cerebral artery blood velocity, arterial diameter and muscle sympathetic nerve activity during post-exercise muscle ischaemia. Acta Physiol. Scand. 160:43-47, 1997.
      27. Rasio, E.A., and C.A. Goresky. Capillary limitation of oxygen distribution in the isolated rete mirabile of the eel (Anguilla anguilla). Circ. Res. 44:498-503, 1979.
      28. Rasmussen, P., H. Stie, B. Nielsen, and L. Nybo. Enhanced cerebral CO2 reactivity during strenuous exercise in man. Eur. J. Appl. Physiol. 96:299-304, 2006.
      29. Rasmussen, P., E.A. Dawson, L. Nybo, J.J. Van Lieshout, N.H. Secher, and A. Gjedde. Capillary-oxygenation-level-dependent near-infrared spectrometry in frontal lobe of humans. J. Cereb. Blood Flow Metab. 27:1082-1093, 2007.
      30. Thomas, S.N., T. Schroeder, N.H. Secher, and J.H. Mitchell. Cerebral blood flow during submaximal and maximal dynamic exercise in humans. J. Appl. Physiol. 67:744-748, 1989.

      central nervous system; cerebral blood flow; cerebral metabolism; hypoxia; mitochondria

      ©2007 The American College of Sports Medicine