Secondary Logo

Journal Logo


Caffeine: A Valuable Tool to Study Central Fatigue in Humans?

Kalmar, Jayne M.1; Cafarelli, E1,2

Author Information
Exercise and Sport Sciences Reviews: October 2004 - Volume 32 - Issue 4 - p 143-147
  • Free


The ability to perform any physical exercise is dependent on the activation of skeletal muscle. As exercise progresses, the ability to generate maximal voluntary force declines, and the central drive required to maintain a given intensity of exercise increases. Sources of failure within skeletal muscle have been well studied; however, the ability of the central nervous system (CNS) to maintain maximal activation of muscle during prolonged effort is not as clear. Central output to muscle may be diminished by 1) afferent input causing inhibition or disfacilitation at either a spinal or supraspinal level, 2) a decline in descending drive resulting from alterations in premotor output or the reduced excitability of corticospinal motor neurons, or 3) changes in intrinsic motor neuron properties that hyperpolarize the resting membrane potential (Fig. 1). In human subjects, it currently is not possible to measure changes at these sites directly. However, a variety of indirect protocols using voluntary contractions as well as evoked responses can be used to assess central drive. The purpose of this review is to describe how caffeine, a well-tolerated CNS stimulant with minimal side effects, may be used to study failure at various sites along the motor pathway. The ergogenic effects of caffeine, observed in many types of neuromuscular and sport performance, are likely the result of more than one biological mechanism. However, there is considerable evidence to suggest that caffeine exerts a measurable effect on the CNS, and that the drug’s ergogenic effect is at least in part the result of enhanced central drive. Therefore, the ability to manipulate various central mechanisms using caffeine could prove to be advantageous in the study of central fatigue.

Figure 1.
Figure 1.:
Sites of central failure and caffeine’s central effects. Failure may occur at many points along the motor pathway from the cortex to the contractile apparatus. Spinal as well as supraspinal mechanisms may contribute to a decline in central drive to the muscle and may be modulated at either level by sensory input from the muscle. The central effects of caffeine observed in several human neuromuscular studies are indicated at three sites (white circles). Experimental evidence and references are in the text at the bottom of the figure. It should be noted that the studies demonstrating a decline in central drive ② do not distinguish between spinal and supraspinal mechanisms. Furthermore, although the changes in force sensation and muscle pain ③ are likely the result of a central mechanism, the effect may be the result of a reduction in the peripheral activation of the sensory receptor.


Caffeine is known to enhance many types of exercise, including endurance sport performance, strength, speed, and power (5) as well as laboratory tests of muscle endurance (8,11). Several mechanisms have been proposed to explain its ergogenic effects, such as increased myofibrillar calcium availability, enhanced exercise metabolism and substrate availability, and stimulation of the CNS.

Direct Effects on Skeletal Muscle

It has been known, for many years, that caffeine induces skeletal muscle force in the absence of membrane depolarization and potentiates twitch and tetanic force production. These effects later were attributed to the direct interaction of caffeine with calcium-activated calcium channels of the sarcoplasmic reticulum (6). These findings led to the notion that caffeine increases calcium availability and improves muscle performance in vivo. However, the effects of caffeine on intracellular calcium and force production observed in vitro require millimolar concentrations of caffeine, which would be toxic in humans (3). Oral doses of caffeine used in human studies result in a plasma concentration of less than 70 μM, a level too low to increase the development of muscle tension even when applied directly to isolated, fatigued muscle (7). Nonetheless, there are a number of human studies that report delayed failure and enhanced recovery of contractile properties after fatiguing electrical stimulation (e.g., (13)), and the possibility that caffeine may directly affect skeletal muscle in some instances cannot be discounted.

Enhanced Exercise Metabolism and Substrate Availability

The increased duration of aerobic exercise after the caffeine ingestion was once attributed to enhanced fat oxidation and glycogen sparing as a consequence of increased plasma catecholamine concentration. However, this explanation lacks empirical evidence and recently was challenged (5). Moreover, these changes in exercise metabolism would not account for improved endurance reported in shorter duration muscle fatigue protocols.

Effects on the CNS

A third hypothesis is that caffeine increases time to fatigue through its effects on the CNS. This central effects hypothesis is attractive because caffeine has been well established as a CNS stimulant via its actions as an adenosine receptor antagonist at concentrations in the micromolar range (Fig. 2) (1,3). This hypothesis is particularly compelling because the effective concentration corresponds well with caffeine concentrations found in human plasma and cerebral spinal fluid after oral caffeine administration at dosages that elicit an ergogenic response.

Figure 2.
Figure 2.:
The effects of an adenosine receptor agonist and antagonist on voluntary run time in rats. Treadmill running time was reduced by 68% after an intracerebroventricular injection of 5′-N-ethylcarboxamidoadenosine (0.1 μg), and increased by 60% after the injection of caffeine (200 μg) when compared with the vehicle-only trial. The effects of the adenosine agonist were offset when caffeine was administered in combination with 5′-N-ethylcarboxamidoadenosine. Intraperitoneal injections of these drugs had no effect on treadmill running time. These data suggest that caffeine increases endurance time by blocking central adenosine receptors. (Reprinted from Davis J. M., Z. Zhao, H. S. Stock, K. A. Mehl, J. Buggy, and G. A. Hand. Central nervous system effects of caffeine and adenosine on fatigue. Am. J. Physiol. Integr. Comp. Physiol. 284:R399–R404, 2003. Copyright © 2003 American Physiological Society. Used with permission.)

The effects of adenosine and its antagonists, including caffeine, are diverse and complex. The idea that caffeine’s ergogenic effects may be the result of stimulation of the CNS has been dismissed as oversimplified. Furthermore, the drug may not enhance performance at all in athletes who are highly motivated, highly aroused, and removed from the controlled environment of the laboratory. Nonetheless, there is a growing body of evidence to suggest that caffeine’s central effects, despite their complexity, may contribute to the ergogenicity observed under experimental conditions. Thus, it may be possible to use caffeine as a perturbation of central excitability to study exercise-induced central failure after its effects on the human CNS have been better characterized. For this reason alone, the effects of caffeine on the CNS during fatiguing exercise warrants further study.


At the dosages used in human studies, the primary mechanism of caffeine’s actions is adenosine receptor antagonism (3). Adenosine is an endogenous neuromodulator with inhibitory effects on central excitability. It preferentially inhibits the release of excitatory neurotransmitters and decreases the firing rate of central neurons. Many in vitro studies have reported a reversal in the inhibitory effects of adenosine after the administration of caffeine and have reported increased excitatory neurotransmitter release and lower the thresholds for neuronal activation. The effects of adenosine as well as caffeine on the release and turnover of specific neurotransmitters have been reviewed elsewhere (3).

Many of the neurotransmitters affected by caffeine in vitro have been implicated in central fatigue (4), notably dopamine. Supraspinal dopaminergic transmission is associated with increased arousal and motivation as well as spontaneous motor activity and prolonged exercise time (4). Caffeine enhances dopaminergic transmission via both presynaptic as well as postsynaptic mechanisms (Fig. 3A). This is not to suggest that all the effects of caffeine on the CNS would tend to oppose central failure. The literature clearly indicates that the actions of adenosine and caffeine vary by target receptor and location in the CNS. The relationship between caffeine, serotonin, and fatigue is a good example of this complexity. Increased cortical serotonin after fatigue has been proposed as a cause of fatigue (4), which would suggest that a caffeine-induced increase in serotonergic transmission in the brain would promote, rather than offset, fatigue. However, serotonergic input at a spinal level (Fig. 3B) is associated with excitation of the α motor neuron pool and antinocioception (4), factors that may offset central failure. The fact that caffeine has numerous sites of action within the CNS may afford the opportunity to use the drug to study central fatigue. Studies that indicate these potential sites of action are discussed below.

Figure 3.
Figure 3.:
Presynaptic and postsynaptic effects of caffeine on two neurotransmitters that have been implicated in central fatigue. On presynaptic membranes (PRE), adenosine prevents the release of neurotransmitters such as (A) dopamine and (B) serotonin by binding to A1 receptors. This may be the result of a decrease in calcium influx or the direct effects neurotransmitter release machinery, or it may be secondary to decreased cyclic adenosine monophosphate (cAMP) production after activation of an inhibitory G protein. Caffeine blocks the inhibition of adenosine and increases the release of these neurotransmitters. Postsynaptically (POST), adenosine also binds to A2a receptors on postsynaptic membranes and prevents dopamine from binding to its receptor. Caffeine blocks the A2a receptor thereby promoting dopaminergic transmission (A). The postsynaptic effects of increased serotonergic input may increase spinal excitability via a decrease in K+ influx and the generation of a persistent inward calcium current that hyperpolarizes the resting membrane potential of the α motor neuron (B). Although the effects of adenosine and caffeine have been oversimplified for the sake of clarity in this schematic, the complexity of these mechanisms is detailed in a review by Fredholm et al. (3).

The behavioral effects of caffeine observed in vivo, particularly increased arousal and vigilance, are well known and likely contribute to the widespread use of the drug. Animal and human studies also have demonstrated an increase in spontaneous motor activity after the administration of caffeine (3). Although these observations are consistent with a central effects hypothesis, few of these experiments actually involve fatiguing exercise. More recently, however, a number of fatigue studies conducted on humans have provided some insight into the role that caffeine may play in offsetting failure through various supraspinal and spinal mechanisms.

Maximal voluntary activation of the nonfatigued quadriceps femoris, assessed using twitch interpolation, and associated maximal knee extension torque increases slightly but significantly after caffeine administration (8,11). In the same two investigations, the rate at which muscle activation increased to maintain the submaximal target force during the fatigue task and the ability to activate muscle voluntarily and to generate peak force after fatigue were unaffected by caffeine. The contractile properties and mass action potential amplitude and area were similarly unaffected by caffeine before, during, and after the fatigue protocol. However, both studies report an increase in endurance time. Voluntary activation increased at the same rate and to the same extent in the caffeine and placebo trials, but the fact that that more work was carried out in the caffeine trial would suggest that subjects were either more willing or motivated to tolerate discomfort or felt less discomfort. This latter possibility is supported by data demonstrating reductions in force sensation (Fig. 4) (11) and muscle pain (10).

Figure 4.
Figure 4.:
Caffeine decreases force sensation at the onset of sustained submaximal contractions. Subjects in this experiment were asked to meet a target force of 50% of their maximal voluntary contraction. As soon as the target was met, visual feedback was removed and a constant force sensation was maintained for 100 s. Although subjects sustained the same sensation of force, their force output declined over the course of the contraction, indicating that maintaining the 50% target force would have felt more difficult. Force declined more rapidly in the first 10 s (inset) of the placebo trial, indicating that caffeine diminished force sensation. This technique has the advantage of being extremely reproducible and less subjective than rating scales of pain and perceived exertion. (Reprinted from Plaskett, C. J., and E. Cafarelli. Caffeine increases endurance and attenuates force sensation during submaximal isometric contractions. J. Appl. Physiol. 91:1535–1544, 2001. Copyright © 2003 American Physiological Society. Used with permission.)

Under many conditions, the termination of exercise or reduction in optimal exercise intensity is a voluntary act that occurs because of discomfort. Nociception also seems to alter involuntary aspects of motor unit activation, such as motor unit firing rates, when force production is held constant (2). Although adenosine itself has an analgesic effect through its actions on peripheral A1 receptors, antagonism of central A2a receptors results in increased pain transmission (12). Nevertheless, the antinociceptive properties of caffeine are well known, and caffeine is included as an analgesic in a number of pharmacologic preparations. In human neuromuscular studies, caffeine diminishes the perception of muscle pain induced by exercise (10) and decreases force sensation during sustained isometric contractions (11). Although caffeine does not increase motor unit firing rates during nonfatigued submaximal contractions (8), its effects on pain-induced reductions in motor-unit firing rates have not yet been explored. It is also possible that caffeine’s effects on force and pain sensation are modulated peripherally; however, there are at least three observations that oppose this notion. The first is that the reduction in force sensation during sustained contractions is greatest at the onset of exercise and is no longer evident at the point of failure when the accumulation of metabolites and input to sensory receptors would be the greatest (11). Second, adenosine acts as a vasodilator in skeletal muscle such that caffeine would be expected to impair rather than enhance the removal of metabolites. Finally, caffeine does not affect the rate of decline in twitch amplitude, time to peak tension, or half relaxation time during sustained or intermittent fatigue protocols (8,11), suggesting that the reduction in force and pain sensation is not the result of greater efficiency of the contractile apparatus. Future studies may determine if caffeine disinhibits motor unit activation at a spinal or supraspinal level, modulates nociception at a spinal level, or acts at a cortical level to diminish the perception of pain.

In the past, attempts to demonstrate the effects of caffeine on human spinal excitability using maximal H reflex amplitude (Hmax) have been inconclusive. In the context of this review, “spinal excitability” refers to the net effect of membrane properties and the summation of all synaptic input to the α motor neuron pool. This term is broad by necessity, because the techniques used to assess spinal excitability in humans do not identify a specific mechanism. Recently, the slope of the H reflex recruitment curve normalized to the slope of the M wave recruitment curve was used to assess the excitability of the soleus α motor neuron pool before and after a 6-mg·kg−1 dose of caffeine (14). Using this technique, Walton et al. (14) demonstrated that caffeine does indeed increase excitability of the nonfatigued soleus motor neuron pool when compared with placebo. It therefore may be possible to use caffeine as a perturbation to determine whether changes in spinal excitability contribute to, or simply accompany, the decline in force-producing capacity. If inhibition or disfacilitation of the α motor neuron pool contributes to fatigue, then a caffeine-induced increase in excitability would be expected to offset failure.

During fatigue, changes in spinal excitability may be the result of alterations in descending drive, recurrent inhibition, presynaptic inhibition, as well as mechanicoceptive, chemicoceptive, or nociceptive afferent input. However, changes in intrinsic membrane properties of α motor neurons also have the capacity to alter spinal excitability. The development of plateau potentials cannot be measured directly in human studies; however, there are neurophysiological events, such as self-sustained firing, that generally are agreed to provide indirect evidence of these transient shifts in the resting membrane potential. Caffeine increases the incidence of self-sustained firing of human motor units (15), perhaps through caffeine’s actions on descending serotonergic or noradrenergic inputs that facilitate the production of plateau potentials. One role of plateau potentials may be to increase motor neuron excitability to reduce the level of descending drive required to maintain a contraction. Accordingly, these potentials may help offset fatigue-induced spinal inhibition, disfacilitation, or a decline in descending drive and could contribute to the increase in spinal excitability observed using H reflex recruitment curves.

Biochemical and pharmacologic investigations in animals also have provided insight into potential supraspinal sources of central failure. A number of indirect neurophysiological techniques, such as surface electromyography, twitch interpolation, transcranial electrical stimulation, transcranial magnetic stimulation (TMS), and transmastoid stimulation, have been used for this purpose. Such experiments have suggested that during exhaustive exercise: 1) the ability to generate maximal voluntary activation, generally present in the nonfatigued state, becomes compromised; 2) cortical excitability is diminished; and 3) afferent feedback to the corticospinal motor neuron may contribute to these findings (4). It is important to recognize that changes in voluntary surface electromyography or cortically evoked potentials can be considered evidence of supraspinal fatigue only if spinal and peripheral sources of failure are taken into account.

Although an increase in maximal voluntary activation after caffeine administration has been demonstrated at rest (8,11), there is currently no conclusive evidence that caffeine offsets the effects of fatigue on motor unit activation via a supraspinal mechanism. We recently used single-pulse TMS to demonstrate central failure in the first dorsal interosseous muscle, but found that most of the decline in the motor-evoked potential and surface EMG largely were attributable to peripheral transmission failure (9). In this study, caffeine increased postactivation potentiation of the cortically evoked potentials, but this does not necessarily reflect a supraspinal mechanism. Paired-pulse TMS maybe be of benefit in future studies to study the effects of caffeine on intracortical facilitation and inhibition before and after fatiguing exercise.


The biochemical actions of caffeine observed in vitro suggest that caffeine may offset central failure in vivo. Up to this point, human experiments have revealed an increase in central excitability (9), maximal voluntary activation (8,11), maximal voluntary force production (8,11), endurance (5,8,11), spinal excitability (14), and self-sustained firing (15), as well as decreases in force sensation (11) and muscle pain (10) after oral caffeine administration. Supraspinal mechanisms also are likely to contribute to caffeine’s effects on muscle activation, but at present, this hypothesis has only been supported by in vitro investigations. In the future, techniques such as paired-pulse TMS and brain imaging may provide further insight into the drug’s supraspinal effects. As its effects on the human motor system are revealed, caffeine will continue to be a useful tool in studying central fatigue.


Supported by NSERC (grant A-6655; EC), a Reebok Research Grant on Human Performance and Injury Prevention from the American College of Sports Medicine Foundation Grant, and NSERC PGS-B and OGS scholarships to JMK. Limitations in the numbers of references permitted prevent us from acknowledging many authors who have contributed to the understanding of central fatigue and caffeine’s effects on the CNS. Therefore, we have used review articles in some instances to direct the reader to more comprehensive summaries of the literature.


1. Davis, J. M., Z. Zhao, H. S. Stock, K. A. Mehl, J. Buggy, and G. A. Hand. Central nervous system effects of caffeine and adenosine on fatigue. Am. J. Physiol. Regul. Integr. Comp. Physiol. 284:R399–R404, 2003.
2. Farina. D., L. Arendt-Nielsen, R. Merletti, and T. Graven-Nielsen. Effect of experimental muscle pain on motor unit firing rate and conduction velocity. J. Neurophysiol/ 91:1250–1259, 2004.
3. Fredholm, B. B., K. Battig, J. Holmen, A. Nehlig, and E. E. Zvartau. Actions of caffeine in the brain with special reference to factors that contribute to its widespread use. Pharmacol. Rev. 51:83–133, 1999.
4. Gandevia, S. C. Spinal and supraspinal factors in human muscle fatigue. Physiol. Rev. 81:1725–1789, 2001.
5. Graham, T. E. Caffeine and exercise: metabolism, endurance and performance. Sports Med. 31:785–807, 2001.
6. Herrmann-Frank, A., H. C. Luttgau, and D. G. Stephenson. Caffeine and excitation-contraction coupling in skeletal muscle: a stimulating story. J. Muscle Res. Cell Motil. 20:223–237, 1999.
7. James, R. S., R. S. Wilson, and G. N. Askew. Effects of caffeine on mouse skeletal muscle power output during recovery from fatigue. J. Appl. Physiol. 96:545–552, 2004.
8. Kalmar, J. M., and E. Cafarelli. Effects of caffeine on neuromuscular function. J. Appl. Physiol. 87:801–808, 1999.
9. Kalmar, J. M., and E. Cafarelli. Central fatigue and transcranial magnetic stimulation: Effect of fatigue and the confound of peripheral transmission failure. J. Neurosci. Meth. 138:15–26, 2004.
10. Motl, R. W., P. J. O’Connor, and R. K. Dishman. Effect of caffeine on perceptions of leg muscle pain during moderate intensity cycling exercise. J. Pain 4:316–321, 2003.
11. Plaskett, C. J., and E. Cafarelli. Caffeine increases endurance and attenuates force sensation during submaximal isometric contractions. J. Appl. Physiol. 91:1535–1544, 2001.
12. Sawynok, J. Adenosine receptor activation and nociception. Eur. J. Pharmacol. 347:1–11, 1998.
13. Tarnopolsky, M., and C. Cupido. Caffeine potentiates low frequency skeletal muscle force in habitual and nonhabitual caffeine consumers. J. Appl. Physiol. 89:1719–1724, 2000.
14. Walton, C., J. Kalmar, and E. Cafarelli. Caffeine increases spinal excitability in humans. Muscle Nerve 28:359–364, 2003.
15. Walton, C., J. M. Kalmar, and E. Cafarelli. Effect of caffeine on self-sustained firing in human motor units. J. Physiol. 545:671–679, 2002.

CNS; muscle activation; supraspinal; spinal; TMS; EMG; adenosine

©2004 The American College of Sports Medicine