The intent to exercise influences the cerebral O2/carbohydrate uptake ratio in humans



During and after maximal exercise there is a 15–30 % decrease in the metabolic uptake ratio (O2/[glucose +1/2lactate]) and a net lactate uptake by the human brain. This study evaluated if this cerebral metabolic uptake ratio is influenced by the intent to exercise, and whether a change could be explained by substrates other than glucose and lactate. The arterial-internal jugular venous differences (a-v difference) for O2, glucose and lactate as well as for glutamate, glutamine, alanine, glycerol and free fatty acids were evaluated in 10 healthy human subjects in response to cycling. However, the a-v difference for the amino acids and glycerol did not change significantly, and there was only a minimal increase in the a-v difference for free fatty acids after maximal exercise. After maximal exercise the metabolic uptake ratio of the brain decreased from 6.1 ± 0.5 (mean ±s.e.m.) at rest to 3.7 ± 0.2 in the first minutes of the recovery (P < 0.01). Submaximal exercise did not change the uptake ratio significantly. Yet, in a second experiment, when submaximal exercise required a maximal effort due to partial neuromuscular blockade, the ratio decreased and remained low (4.9 ± 0.2) in the early recovery (n= 10; P < 0.05). The results indicate that glucose and lactate uptake by the brain are increased out of proportion to O2 when the brain is activated by exhaustive exercise, and that such metabolic changes are influenced by the will to exercise. We speculate that the uptake ratio for the brain may serve as a metabolic indicator of ‘central fatigue’.

The brain takes up O2 and glucose in a ratio close to 6:1, but the ratio becomes reduced during physiological activation as demonstrated with visual (Fox et al. 1988) and mental stimulation in man (Madsen et al. 1995b) and by sensory stimulation in the rat (Madsen et al. 1995a, 1999). Furthermore, this ratio between the brain uptake of O2 and carbohydrate decreases in response to exhaustive exercise and reaches its lowest value in the first minutes of the recovery (Ide et al. 2000). During activation the brain also takes up significant amounts of lactate provided that plasma lactate is also increased as seen during intense exercise (Ide et al. 2000). Thus, the disproportionate higher glucose and lactate to O2 uptake - reflected by a reduced O2/(glucose + 1/2lactate) ratio - becomes even more pronounced, especially in the early recovery period. Such a persisting higher uptake of carbohydrates, is also observed for several minutes following mental activity in man (Madsen et al. 1995b). As cerebral activation is associated with a reduced glycogen content in the brain of the rat (Swanson et al. 1992; Madsen et al. 1995a, 1999), a reduction in the metabolic ratio supports the idea that, especially in the recovery, some of the glucose taken up by the brain is used to replenish brain glycogen stores.

Williamson et al. (1997, 1999) and Nowak (2001) provide evidence that physiological activation of the brain increases with exercise intensity and intense neuronal activity can make energy demand exceed energy production (Sappey-Marinier et al. 1992). Such transient imbalance between energy demand and production may occur in brain regions engaged during exhaustive exercise and, in turn, cause glycogen depletion and termination of exercise. In this way, the metabolic ratio may be a unique metabolic equivalent to ‘central fatigue’ which so far has been difficult to explain.

Although Ide et al. (2000) reported a drop in the metabolic ratio following exhaustive exercise, it is unknown if non-fatiguing exercise also induces an excess carbohydrate uptake in the immediate recovery. The hypothesis of this study was that the O2/(glucose + 1/2lactate) ratio for the brain would be reduced after exhaustive exercise rather than following non-fatiguing exercise. We further considered that in animals, neuronal tissue has the capacity to oxidise amino acids (Larrabee, 1984; Sonnewald et al. 1997) and that such metabolism during exercise could affect the metabolic ratio. Hence, in addition to glucose and lactate, we determined the arterio-internal jugular venous differences (a-v difference) for glutamate, glutamine, alanine and also for glycerol and free fatty acids. Cerebral perfusion was evaluated by the transcranial Doppler determined middle cerebral artery (MCA) mean flow velocity (Vmean).

Exhaustive exercise requires cerebral activation to produce the ‘command’ to exercise and supposedly to perform the integration of the sensory input from working muscle. Yet, their individual contribution to the increased metabolism of the brain is unknown. Central command and sensory input from exercising skeletal muscle may be partially separated by administering a non-depolarising neuromuscular blocking agent that, in a dose-dependent manner, reduces muscle strength. Thus, under the influence of such agents, maintaining a given work rate requires increased discharge frequency or recruitment of additional motor units and consequently a larger central drive (Asmussen et al. 1964; Leonard et al. 1985; Galbo et al. 1987). Such effect is evaluated in an additional experiment.


The experiments were performed according to the Declaration of Helsinki as approved by the Ethics Committee of Copenhagen (KF 01–369/97). Every participant volunteered and informed written consent was obtained. In two studies separate subjects exercised in a semi-supine position on a modified Krogh cycle ergometer at 60 r.p.m. (Galbo et al. 1987), while the a-v difference across the brain for O2 and substrates were determined. Except for incremental exercise, each bout lasted 10 min and was followed by ∼1 h of recovery.

In the first experiment we included twelve subjects of both sexes. However, two females were excluded from the study due to difficulties of drawing from the catheters, leaving one female and nine male subjects to be studied (age 25 ± 4 years; height 179 ± 8 cm; weight 75 ± 9 kg; means ±s.e.m.). The protocol included exercise at light and moderate intensity (heart rate (HR) 90 and 120 beats min−1, respectively) and exercise to exhaustion that was reached by increasing the work rate by 30 W every second minute.

In the second study with partial neuromuscular blockade a further four females and seven males (age 25 ± 1 years; height 174 ± 3 cm; weight 71 ± 3 kg) attended the laboratory twice. First, they performed incremental exercise until exhaustion to determine the maximal O2 uptake (VO2,max, 3.2 ± 0.2 l min−1). On the second day they initially exercised at a work rate of about 40 % of VO2,max (control) followed by a resting period and then exercised at the same work rate but with partial neuromuscular blockade. Subjects were curarised by cisatracurium (Nimbex, Glaxo-Wellcome, Denmark) administered intravenously to a total dose of 1.2 (0.4–2.2) mg. Data for one subject was excluded because of apparatus failure (blood analysis). When muscular relaxation was induced an Ambu-E resuscitator apparatus, neostigmine and atropine were available, but never needed.

The a-v difference across the brain was obtained by means of a 2.2 mm (14 gauge) catheter introduced percutaneously into the right internal jugular vein and advanced to the venous bulb, and a 1.1 mm (20 gauge) catheter placed in the brachial artery of the non-dominant arm. Mean arterial pressure (MAP) was obtained from the artery and integrated by a monitor (Dialogue 2000, Danica Electronic, Copenhagen, Denmark) that also calculated HR from a three lead electrocardiogram. Catherisation was well tolerated by the subjects and was not reported to affect their exercise performance.

Arterial and venous blood samples were drawn simultaneously three times at rest, twice during submaximal exercise, every second minute during maximal exercise, and several times in the recovery period. Samples were drawn anaerobically for determination of glucose, lactate and blood gas variables and were placed immediately on ice for later analysis (ABL 625, Radiometer, Denmark). Samples drawn for the determination of alanine (Graßl & Supp, 1985), glutamate and glutamine (Lund, 1985), glycerol (Eggstein & Kuhlmann, 1970) and free fatty acids (FFA; NEFA C, ACS-ACOD Method, Wako Chemicals, Richmond, VA, USA) were placed on ice and centrifuged within 5 min. Plasma was frozen in liquid nitrogen and kept at −80°C until analysed. The assays applied to analyse for amino acids, glycerol and FFA have an accuracy that is better than 1%. Measurement by way of the ABL at the relevant concentrations has an inaccuracy (mean difference between the measured value on a group of test instruments and the estimated true value as assayed by the reference method) of: glucose and lactate 0.1 mm, PO2 0.08 kPa, haemoglobin 0.2 g dl−1 and O2 saturation 0.2 %.

The cerebral blood flow was evaluated by transcranial ultrasound Doppler. The proximal segment of the MCA was insonated (Multidop X, DWL, Sipplingen, Germany) through the right temporal window. After determination of the optimal signal-to-noise ratio, the probe was secured by adhesive ultrasonic gel (Tensive, Parker Laboratories, Fairfield, NJ, USA) and a headband. The MCA Vmean was calculated as the one minute average of continuously sampled maximal frequency shifts for each heart beat. Both at rest and during maximal cycling, determination of flow velocity in the MCA has a coefficient of variation of ∼5 % (Pott et al. 1996).

During the study with patial curarisation, the VO2 was measured breath by breath (MedGraphics 2001, St. Paul, MN, USA). The level of partial neuromuscular blockade was assessed both before and during exercise as the handgrip strength for the dominant arm by means of a strain gauge dynamometer connected to a measuring bridge (Caspersen & Nielsen, Copenhagen, Denmark). Two different questions were posed to assess the exercise intensity using a scale with units from ‘6′ to ‘20′ (Borg, 1970). The first question determined how hard the exercise felt by rating the perceived exertion (RPE), which is the regular application of the Borg scale. The second question quantified the ‘will to exercise’ by asking how hard the subject tried to exercise. Such differentiation is important when the neuromuscular transmission is impaired as the subjects may be trying mentally to exercise as hard as they can while at the same time experiencing only moderate intensity.

Variables are expressed as means ±s.e.m. or as median with range. The Friedman test was used to determine whether significant changes occurred between rest, exercise and the recovery which was divided into the first 5 min and the subsequent 25 min as well as, for maximal exercise, for an additional 30 min (Ide et al. 2000). Also, incremental exercise was divided into low and high intensity periods by an arbitrary ‘lactate threshold’, which in this context means the time point where the increase in plasma lactate becomes significant. Changes were identified with the Wilcoxon matched pairs test by rank. P < 0.05 was considered to be statistically significant.


Submaximal vs. maximal exercise

The resting HR was 67 ± 2 beats min−1, MAP, 89 ± 3 mmHg and MCA Vmean, 67 ± 6 cm s−1 with all values increasing during exercise. In response to light exercise (HR, 91 ± 5 beats min−1; MAP, 90 ± 3 mmHg and MCA Vmean, 71 ± 6 cm s−1) and in the subsequent recovery period there were no or only marginal changes in the a-v difference for O2, glucose, lactate, glutamate, glutamine, alanine, glycerol and FFA (Table 1). Thus, the O2/(glucose + 1/2lactate) uptake ratio for the brain did not change significantly (Fig. 1). With moderate exercise (HR was 122 ± 3 beats min−1; MAP, 96 ± 4 mmHg and MCA Vmean, 73 ± 5 cm s−1) the pattern was the same, except that blood lactate increased from 1 to ∼3 mm, resulting in an increase in the a-v difference for lactate to 0.1 mm (P < 0.01, Table 1).

Table 1. Arterial variables and arterial–venous differences over the brain at rest, and during light and moderate exercise and in the recovery (n= 10)
 RestLight exerciseRecoveryModerate exerciseRecovery
   0−5 min5−25 min 0−5 min5−25 min
  1. Values are means ±s.e.m. a−v difference, arterial–internal jugular venous difference; Pa,CO2, arterial carbon dioxide tension; pHa, arterial pH; FFA, free fatty acids; gluc., glucose; lac., lactate.*P < 0.05 and †P < 0.01 compared to rest.

P a,CO2 (kPa)5.71 ± 0.075.53 ± 0.29 *5.60 ± 0.065.74 ± 0.055.76 ± 0.115.48 ± 0.05 †5.65 ± 0.03
pHa7.41 ± 0.017.41 ± 0.017.43 ± 0.017.41 ± 0.00 *7.38 ± 0.01 †7.39 ± 0.01 †7.40 ± 0.00 †
Arterial blood contents (mm)
    O28.69 ± 0.228.93 ± 0.258.89 ± 0.278.68 ± 0.259.17 ± 0.27 †9.01 ± 0.28 *8.61 ± 0.22
    Glucose5.78 ± 0.265.09 ± 0.11 †5.25 ± 0.075.59 ± 0.184.96 ± 0.09 †4.93 ± 0.10 †5.32 ± 0.13
    Lactate0.94 ± 0.090.99 ± 0.080.91 ± 0.070.84 ± 0.063.16 ± 0.55 †2.61 ± 0.42 †1.34 ± 0.15
    Alanine0.36 ± 0.040.38 ± 0.030.34 ± 0.030.31 ± 0.030.40 ± 0.030.37 ± 0.030.32 ± 0.03
    Glutamate0.07 ± 0.010.05 ± 0.00 †0.06 ± 0.01 †0.07 ± 0.010.05 ± 0.01 †0.06 ± 0.01 *0.07 ± 0.00
    Glutamine0.59 ± 0.020.60 ± 0.020.57 ± 0.020.57 ± 0.020.61 ± 0.020.57 ± 0.010.58 ± 0.01
    Glycerole0.02 ± 0.000.03 ± 0.010.03 ± 0.010.04 ± 0.010.05 ± 0.010.06 ± 0.020.05 ± 0.02
    FFA0.25 ± 0.110.24 ± 0.100.35 ± 0.12 *0.28 ± 0.090.55 ± 0.19 *0.54 ± 0.15 *
a—v difference (mm)
    O23.31 ± 0.212.85 ± 0.13 *3.22 ± 0.233.12 ± 0.163.18 ± 0.173.36 ± 0.333.07 ± 0.16
    Glucose0.56 ± 0.040.50 ± 0.03 *0.63 ± 0.04 *0.59 ± 0.040.53 ± 0.030.60 ± 0.040.58 ± 0.03
    Lactate−0.04 ± 0.02−0.05 ± 0.01−0.08 ± 0.01−0.08 ± 0.020.12 ± 0.04 †−0.01 ± 0.02−0.05 ± 0.01
    Alanine−0.02 ± 0.01−0.01 ± 0.01−0.02 ± 0.01−0.01 ± 0.010.00 ± 0.01−0.01 ± 0.01−0.01 ± 0.01
    Glutamate0.00 ± 0.000.00 ± 0.000.00 ± 0.00−0.01 ± 0.000.00 ± 0.000.00 ± 0.000.00 ± 0.00
    Glutamine−0.03 ± 0.02−0.02 ± 0.02−0.03 ± 0.02−0.03 ± 0.02−0.01 ± 0.02−0.02 ± 0.0−0.01 ± 0.01
    Glycerol0.00 ± 0.000.00 ± 0.000.00 ± 0.000.01 ± 0.000.00 ± 0.000.00 ± 0.000.00 ± 0.00
    FFA−0.01 ± 0.000.00 ± 0.01−0.01 ± 0.00−0.01 ± 0.000.00 ± 0.01−0.02 ± 0.01
    Glucose +½lac.0.54 ± 0.040.47 ± 0.03 *0.59 ± 0.040.56 ± 0.040.58 ± 0.040.58 ± 0.040.55 ± 0.04
Ratios of a−v diffference       
    O2/glucose5.80 ± 0.355.95 ± 0.305.20 ± 0.235.36 ± 0.206.03 ± 0.285.46 ± 0.45.43 ± 0.30
    O2/(gluc. +½lac.)6.11 ± 0.456.41 ± 0.445.58 ± 0.275.76 ± 0.255.56 ± 0.355.58 ± 0.445.71 ± 0.38
Figure 1.

The metabolic uptake ratio of arterial-internal jugular venous differences in response to submaximal and maximal exercise

A, light, B, moderate and C, maximal exercise. Data are means ±s.e.m. for ten subjects. †P < 0.01, compared to rest.

Exhaustive exercise (peak values: HR, 170 ± 8 beats min−1; MAP, 122 ± 11 mmHg and MCA Vmean, 73 ± 11 cm s−1) did not change the a-v difference for O2 and glucose while there was a large uptake of lactate amounting to ∼80 % of the glucose uptake. This resulted in a decrease in the O2/(glucose + 1/2lactate) from 6.1 at rest to 4.4 at the point of exhaustion (Fig. 1, Table 2). In the initial 5 min of the recovery phase, glucose uptake increased by 30 % and lactate uptake by an additional 40 %. Since the a-v difference for O2 was not significantly increased, the O2/(glucose + 1/2lactate) fell to a value of 3.7, which normalised within the subsequent 25 min of the recovery period. There was an increase of the O2/glucose ratio in the 5–30 min recovery interval to a value of 7.2. Finally, there were no changes in the a-v difference of glutamate, glutamine, alanine, glycerol and FFA, except for a slight but statistically significant uptake of FFA in the recovery interval from 5 to 30 min.

Table 2. Arterial variables and arterial–venous differences over the brain during maximal exercise and in the subsequent recovery (n= 10)
 RestMaximal exerciseRecovery
  Below lactate thresholdAbove lactate threshold0–5 min5−30 min30−60 min
  1. Values are means ±s.e.m. a−v difference, arterial–internal jugular venous difference; FFA, free fatty acids; Pa,CO2, arterial carbon dioxide tension; pHa, arterial pH, gluc., glucose; lac., lactate; values at ‘Rest’ are identical to Table 1; *P < 0.05 and †P < 0.01 compared to rest.

P a,CO2 (kPa)5.71 ± 0.075.82 ± 0.085.12 ± 0.16 †4.38 ± 0.17 †4.75 ± 0.12 †5.55 ± 0.06 *
Alterial blood contents (mm)
O28.69 ± 0.228.78 ± 0.269.24 ± 0.25 †9.37 ± 0.26 †8.82 ± 0.308.43 ± 0.18
    Glucose5.78 ± 0.265.25 ± 0.15 *4.89 ± 0.16 *6.17 ± 0.235.74 ± 0.205.22 ± 0.19
    Lactate0.94 ± 0.091.44 ± 0.10 †6.95 ± 0.58 †14.88 ± 1.43 †9.00 ± 1.27 †2.73 ± 0.48 †
    Alanine0.36 ± 0.040.33 ± 0.030.42 ± 0.030.49 ± 0.03 *0.47 ± 0.040.37 ± 0.03
    Glutamate0.07 ± 0.010.05 ± 0.00 †0.04 ± 0.00 †0.05 ± 0.00 *0.07 ± 0.000.07 ± 0.01
    Glutamine0.59 ± 0.020.62 ± 0.010.62 ± 0.010.60 ± 0.020.58 ± 0.010.59 ± 0.02
    Glycerol0.02 ± 0.000.05 ± 0.010.07 ± 0.020.14 ± 0.030.13 ± 0.030.05 ± 0.01
    FFA0.25 ± 0.110.35 ± 0.090.27 ± 0.070.44 ± 0.080.52 ± 0.12 *0.48 ± 0.12 †
a—v difference (mm)
    O23.31 ± 0.212.96 ± 0.193.45 ± 0.274.04 ± 0.293.84 ± 0.243.07 ± 0.10
    Glucose0.56 ± 0.040.55 ± 0.040.61 ± 0.050.81 ± 0.07 †0.55 ± 0.040.57 ± 0.03
    Lactate−0.04 ± 0.02−0.02 ± 0.020.50 ± 0.08 †0.71 ± 0.13 †0.17 ± 0.09 *0.01 ± 0.05
    Alanine−0.02 ± 0.01−0.01 ± 0.010.00 ± 0.010.01 ± 0.010.00 ± 0.01−0.01 ± 0.01
    Glutamate0.00 ± 0.000.00 ± 0.000.01 ± 0.000.00 ± 0.000.01 ± 0.000.00 ± 0.00
    Glutamine−0.03 ± 0.020.00 ± 0.01−0.02 ± 0.01−0.01 ± 0.01−0.01 ± 0.01−0.01 ± 0.02
    Glycerol0.00 ± 0.000.00 ± 0.000.01 ± 0.010.02 ± 0.02−0.01 ± 0.000.00 ± 0.00
    FFA−0.01 ± 0.00−0.01 ± 0.01−0.01 ± 0.01−0.02 ± 0.020.01 ± 0.00 *0.00 ± 0.01
    Glucose +½lac.0.54 ± 0.040.55 ± 0.050.86 ± 0.09 *1.16 ± 0.12 †0.67 ± 0.07 *0.57 ± 0.04
Ratios of a−v diffference
    O2/glucose5.80 ± 0.355.55 ± 0.305.89 ± 0.285.13 ± 0.197.17 ± 0.30 †5.69 ± 0.37
    O2/(gluc. +½lac.)6.11 ± 0.455.68 ± 0.374.35 ± 0.22 †3.72 ± 0.16 †6.11 ± 0.435.76 ± 0.36

Exercise with partial neuromuscular blockade

Since exhaustive exercise activated the brain as described above, we conducted this second experiment to test if such activation was induced by the mental effort associated with exercise. This was achieved by comparing submaximal exercise with, and without, partial neuromuscular blockade.

In confirmation of the results from the first experiment, submaximal exercise alone (HR, 100 ± 5 beats min−1; MAP, 95 ± 5 mmHg and rating of perceived exertion (RPE) 11 (range, 6–12)) did not significantly affect the a-v difference for O2, glucose or lactate or the metabolic ratio (Table 3, Fig. 2).

Table 3. Arterial variables and arterial–venous differences at rest, and during exercise with and without neuromuscular blockade and in the recovery (n= 10)
 RestControlNeuromuscular blockade
   0−5 min5−30 min 0−5 min5−30 min
  1. Values are means ±s.e.m. a−v difference, arterial–internal jugular venous difference; Pa,CO2, arterial carbon dioxide tension; pHa, arterial pH; gluc., glucose; lac., lactate. *P < 0.05 and †P < 0.01 compared to rest.

P a,CO2 (kPa)5.43 ± 0.065.55 ± 0.095.41 ± 0.075.51 ± 0.06*5.20 ± 0.08*5.27 ± 0.07*5.47 ± 0.06*
pHa7.41 ± 0.007.41 ± 0.01*7.41 ± 0.007.40 ± 0.00†7.40 ± 0.01*7.40 ± 0.01*7.40 ± 0.00†
Arterial blood contents (mm)
    O28.31 ± 0.278.60 ± 0.30*8.33 ± 0.33*8.13 ± 0.28*8.34 ± 0.298.18 ± 0.328.15 ± 0.28
    Glucose5.56 ± 0.234.98 ± 0.12†5.02 ± 0.11*5.21 ± 0.12*5.12 ± 0.135.28 ± 0.135.13 ± 0.12
    Lactate0.96 ± 0.081.35 ± 0.141.19 ± 0.100.98 ± 0.072.30 ± 0.25†2.16 ± 0.25†1.39 ± 0.08†
a—v difference (mm)
    O22.84 ± 0.122.86 ± 0.202.93 ± 0.182.86 ± 0.172.66 ± 0.182.90 ± 0.172.84 ± 0.13
    Glucose0.55 ± 0.030.54 ± 0.040.60 ± 0.030.57 ± 0.020.50 ± 0.050.60 ± 0.030.55 ± 0.03
    Lactate−0.05 ± 0.02−0.02 ± 0.02−0.05 ± 0.02−0.05 ± 0.010.04 ± 0.03*0.01 ± 0.02*−0.04 ± 0.01
    Glucose +½lac.0.53 ± 0.030.53 ± 0.040.57 ± 0.030.54 ± 0.020.56 ± 0.050.60 ± 0.030.53 ± 0.03
Ratios of a−v difference
    O2/glucose5.25 ± 0.135.40 ± 0.224.99 ± 0.245.08 ± 0.195.27 ± 0.424.94 ± 0.255.32 ± 0.17
    O2/(gluc. +½lac.)5.40 ± 0.165.49 ± 0.235.25 ± 0.295.38 ± 0.244.60 ± 0.19*4.89 ± 0.21*5.64 ± 0.21
Figure 2.

The metabolic uptake ratio of arterial-internal jugular venous differences in response to exercise with and without neuromuscular blockade

Data are means ±s.e.m. for ten subjects. •, control; ○, partial neuromuscular blockade. *P < 0.05, compared to rest.

Exercise with neuromuscular blockade (handgrip strength 27 ± 4 % of the control value) increased cardiovascular variables more than control exercise (HR, 116 ± 6 beats min−1; MAP, 117 ± 5 mmHg) although work rate was the same or even slightly less for three subjects. The subjects rated their mental effort to a value of 20 (12–20). Hence, for all subjects except one who were minimally affected by curarisation, the mental effort to exercise was higher than the perceived exertion 16 (12–20).

In response to exercise with partial neuromuscular blockade the a-v difference for O2 and glucose remained stable, hence the O2/glucose ratio did not change significantly (Table 3). However, the a-v difference of lactate across the brain increased during exercise in the blockade condition and remained slightly positive during the early recovery. Thus, during exercise with neuromuscular blockade the O2/(glucose + 1/2lactate) uptake ratio decreased by 15 % from the resting value and it stayed low (4.9 ± 0.2) in the first 5 min of recovery (Fig. 2).


Following human brain activation glucose uptake increases out of proportion to the uptake of O2 (Madsen et al. 1995b). Furthermore, the brain activation elicited by intense exercise also provokes a significant lactate uptake, causing a large drop in the O2/carbohydrate ratio (Ide et al. 2000). The key finding of the present study explains these previous results, in that such changes in the metabolic uptake ratio of the brain took place in response to the enhanced mental effort associated with maximal exercise, but not in response to submaximal exercise. The changes in the metabolic uptake ratio for the brain could not be explained by contribution from glutamate, glutamine, alanine, glycerol or free fatty acids to brain energy metabolism.

Effect of exercise-induced brain activition on uptake of substrates

In addition to glucose, neurons can oxidise alanine (Larrabee, 1984) and glutamate (Sonnewald et al. 1997). Furthermore amino acids taken up by neuronal tissue may play a role, e.g. in the resynthesis of neurotransmitters (Sonnewald et al. 1997). Also, both nonesterified fatty acids and glycerol can penetrate the mammalian blood brain barrier and be metabolised (Evans et al. 1998). However, the fact that the uptake of amino acids and glycerol did not change in response to exercise and that there was only a marginal increase in the uptake of free fatty acids following maximal exercise suggest that these energy sources are not important to the changes in the global cerebral metabolism during exercise.

Lactate can serve as a neuronal energy source (Larrabee, 1995, 1996; Schurr et al. 1988). Yet, during moderate and maximal exercise the brain uptake of lactate could reflect loading of its distribution volume. If so, we would expect lactate to clear from the brain when the arterial concentration decreases after exercise. However, following maximal exercise the lactate a-v difference across the brain remained positive for 30 min, suggesting that the fate of lactate uptake is oxidation. In support, infusion of lactate in the immobilised rat caused no lactate uptake by the brain (Ide et al. 2000).

There is evidence that cerebral activity increases with exercise intensity. Williamson et al. (1999) reported that light intensity bicycle exercise caused an increase in regional cerebral blood flow (rCBF) to the leg motor region, and that high intensity exercise elicited an even greater increase. Furthermore, a submaximal handgrip increased rCBF with time to the contralateral hand- and forearm-motor regions (Williamson et al. 1999). Accordingly, light handgrip increased rCBF in the contralateral sensory-motor area from rest and prolonging the contraction until fatigue caused further elevation to close to 40 % above the resting value (Nowak, 2001). A similar increase is demonstrated with transcranial Doppler in this study and also noted previously (Jørgensen et al. 1992). More recently the representation of ventilation-induced increases during exercise in man was reported (Thornton et al. 2001), as well as an exercise intensity-dependent increase in blood flow to areas in the cerebellum of miniature swine that are likely to be important for integrating sensory input and motor output (Delp et al. 2001). The assessment of a-v difference over the brain in the present study, must be taken to reflect the integrated responses from such activated areas in response to exercise.

During exercise, cerebral perfusion influences the a-v differences. However, during exhaustive exercise both MCA Vmean and the cerebral uptake of O2 and carbohydrates increased simultaneously, which supports the idea that brain metabolism is increased. The metabolic ratio for brain is independent of cerebral blood flow (CBF) as this variable is present in both the numerator and the denominator:

display math


display math

Physiological neuronal activation is associated with an increase in energy requirement due to transport of neurotransmitters and ions. For example, visual stimulation increases the cerebral metabolic rate for glucose in the human occipital area (Fox et al. 1988) and evidence supports the notion of a coupling between astrocyte glucose metabolism and neuronal activity in the brain (Pellerin & Magistretti, 1994; Magistretti & Pellerin, 1999). Besides glucose metabolism, the glycogen level in the brain may also be affected by the level of neuronal activity (Swanson et al. 1992; Madsen et al. 1995a, 1999). For instance, neurotransmitters (vasoactive intestinal peptide and noradrenaline) and other substances (potassium and adenosine) released during synaptic activity can induce glycogen breakdown in cortical brain slices (Magistretti, 1988). Furthermore, astrocyte glycogen is reduced in response to a low glucose level of the surrounding medium in a rat brain culture (Dringen et al. 1992). In the retina of both the honeybee drone (Tsacopoulos et al. 1988; Evequoz-Mercier & Tsacopoulos, 1991) and in mammals (Poitry-Yamate et al. 1995) following light stimulation, glucose is taken up by glia, incorporated as glycogen and subsequently released to fuel the neurons during activation. It is therefore reasonable that astroglial glycogen may be depleted during intense cerebral stimulation to account for the increased energy demand from neuronal activity, and that the surplus carbohydrate taken up in the recovery period is used to rebuild glycogen levels.

The value obtained for the O2/glucose uptake ratio at rest is in agreement with data from studies considering the a-v difference across the brain (Ahlborg & Warren, 1972; Madsen et al. 1995b, 1999; Ide et al. 2000) but is larger than the regional value of 4.1 reported by Fox et al. (1988). In response to two different levels of non-fatiguing exercise the O2/glucose and O2/(glucose + 1/2lactate) uptake ratios did not change significantly. Yet during exhaustive exercise glucose tended to be taken up in excess of O2 and this disproportionate carbohydrate uptake was further aggravated by a lactate uptake. This reduction in the metabolic ratio is likely to reflect anaerobic glycolysis and lactate production as observed in response to a variety of cerebral activity (Ames, 2000; I. C. Schmalbruch, P. L. Madsen, O. B. Paulson & B. Quistorff, unpublished work). As discussed earlier this ‘anaerobic’ lactate production probably relates to a metabolic coupling between astrocytes and neurons involving a preference for lactate by the latter and not to a condition of low O2 tension of cerebral tissue. During exercise part of the decrease in the O2/(glucose + 1/2lactate) can presumably be explained by increased concentration of intermediates in response to enhanced metabolic flux (e.g. glucose-6-phosphate, fructose-6-phosphate and citrate) while during recovery, the process of glycogen resynthesis may cause a surplus glucose uptake. In fact, glycogen synthesis may not only be initiatied after but even during cerebral activation, with the consequence of a simultaneous glycogen breakdown and build up as suggested for skeletal muscle energetics, thus contributing to the reduction of the metabolic uptake ratio by the human brain before exercise is terminated (Shulman & Rothman, 2001).

After exhaustive exercise, the O2/glucose uptake ratio showed an overshoot following the nadir reached during the first 5 min of recovery and a peak at 7.2 after 20 min. A high metabolic flux during exercise and in the early recovery could increase several intermediates of the glycolytic and tricarboxylic acid pathways, and in turn cause a correspondingly lower glucose uptake by the brain. However, the O2/(glucose + 1/2lactate) ratio did not demonstrate an overshoot, suggesting that lactate taken up in the recovery is replacing glucose as a substrate for aerobic metabolism in the brain. Similarly, in the visual cortex the O2/glucose ratio increases above the resting value after stimulus cessation, although Fox et al. (1988) did not measure lactate uptake or release. Madsen et al. (1999) reported that the O2/glucose ratio reached a high value in the recovery period from sensory stimulation in the rat, while the brain was releasing lactate rather than taking it up. This situation differs from the present study where there was an uptake of lactate in the recovery period. However, it is not known if a similar global increase in brain lactate as observed in the rat (Madsen et al. 1999) occurs in man during brain activation. The fact that the uptake of lactate persisted during recovery argues against this notion. It is our hypothesis that the availability of lactate in the blood allows the activated brain to use this substrate for oxidation as an alternative to glucose (Rennie, 2000).

Exercise induced brain activation

Only exhaustive exercise reduced the metabolic ratio for the brain. Therefore, in a subsequent experiment we made the subjects work at a submaximal level but simultaneously we made this difficult by a partial neuromuscular blockade. The intention of the study was to force the subjects to mobilise all their mental effort to exercise, even though the actual work was identical to the control situation. Therefore, the subjects were encouraged to try to exercise as intensively as possible. We assume that the intent to exercise increased to a maximal level when the subjects were weakened by the neuromuscular blockade or even transiently failed to maintain the work rate. Apparently this was the case as the subjects gave near-maximal ratings when reporting how hard they were trying to exercise. Only one subject with a small reduction in handgrip muscle strength reported no increase in effort to exercise compared with control. Interestingly, during exercise with partial curarisation the subjects did not perceive a maximal exertion although they were unable to increase the work rate, and this may be because the subjects did not experience the muscle pain usually associated with intense exercise (Galbo et al. 1987). This indicates that central command alone cannot induce the maximal perception of fatigue, and probably explains why the maximal metabolic response of the brain in terms of the drop in O2/(glucose + 1/2lactate) of ∼3.7 observed in the first experiment and in the study by Ide et al. (2000) was not induced during exercise in the curarised condition (ratio 4.9). As stated earlier, a-v difference over the brain evaluates cerebral metabolism as a whole, i.e. the sum of metabolic activity in the regions drained by the internal jugular vein. We assume the more pronounced metabolic response to maximal exercise is mainly caused by cerebral integration of afferent input from fatiguing skeletal muscle, as opposed to exercise with partial neuromuscular blockade where the load on the limb muscle - and thereby also sensory input to the brain - is ‘submaximal’.

Exercise with neuromuscular blockade increased plasma lactate levels more than control exercise. Such an effect is probably due to the neuromuscular blocking agent's preference for slow twitch fibres (Mizuno et al. 1994), leading to a greater engagement of the predominately glycolytic fast twitch fibres (Galbo et al. 1987). As discussed earlier, it is unlikely that the increased plasma lactate concentration per se is responsible for the increase in lactate uptake by the brain. However, as exercise with partial neuromuscular blockade has been shown to enhance catecholamine levels in the blood (Galbo et al. 1987; Kjaer et al. 1987) we cannot rule out an adrenergic influence on cerebral metabolism during the curarised condition.

We included both females and males in the study, and whether gender differences exist in cerebral metabolism is a matter of controversy. Some data on the resting state support such difference (Yoshii et al. 1988) and some show no differences (Miura et al. 1990) or only regional differences (Gur et al. 1995). However, we consider the individual cerebral metabolic response to workload unlikely to be influenced by gender.

‘Central fatigue’ was defined by Mosso in the 19th century when a colleague of his, after giving a lecture, was unable to perform the same muscle work as he could normally accomplish (Mosso, 1904). It describes fatigue with no apparent muscular origin (Secher, 1992) and it may be related to an increased level in the brain of the amino acid tryptophan, the precursor for the neurotransmitter 5-hydroxytryptamine (Newsholme & Castell, 1996) or to a reduction in the amount of the neurotransmitters glutamate and γ-amino-butyric acid as a result of ammonia detoxication by glutamine synthesis (Guezennec et al. 1998). Yet, these mechanisms require marked elevation in blood of, e.g. free fatty acids (Davis et al. 1992) or ammonia (Guezennec et al. 1998), and are thereby unlikely explanations for the central fatigue experienced when only a small muscle mass is engaged in exercise, or for the phenomenon of an apparent increase in muscle strength at the point of exhaustion by opening the eyes (Asmussen & Mazin, 1978). Although the present experiments provide no direct measure of ‘central fatigue’ we consider such a central component to be present, and that the noted metabolic changes of the central nervous system, with the plausible relation to glycogen depletion, are either indicative of, or perhaps even a possible explanation of central fatigue.


We found that the metabolic ratio by the human brain only decreases when the will to exercise is intense as opposed to exercise requiring less mental effort. Moreover, the reduction in the metabolic ratio during exercise is not explained by glutamate, glutamine, alanine, glycerol or free fatty acids. Our results suggest that during maximal exercise ‘central command’ increases brain metabolism. We speculate that intense activity in cerebral regions causes energy demand to exceed production, in turn draining energy reserves. Such depletion of brain glycogen could play a role in central fatigue.


This study was supported by The Danish National Research Foundation Grant 504–4, Danish Medical Research Council, Grant 52–00-0098 and The Danish Medical Research Council Grant 9502885. Kojiro Ide was supported by Erik Kaj-Jensen and Mads K. Dalsgaard by a H:S Scholarship.