The central governor model of fatigue

Fatigue is recognised as an important factor in the development of different pacing strategies.6 Because both physiological and cognitive factors could be of importance in regulating the pacing strategy, models that explain fatigue purely on the basis of peripheral changes in skeletal muscle cannot evaluate the possible role of cognitive factors. Recently, some novel physiological models have challenged the currently popular paradigm of “peripheral fatigue” by emphasising the crucial role played by the brain in the regulation of exercise performance. Ulmer14 has proposed a control system for the optimisation of “performance” during exercise, in which muscle power output is continuously modified as a result of an integrative control by the brain. In this model, referred to as teleoanticipation, physiological afferent signals are compared in a central “black box” which is pre-programmed as a result of antecedent experiences and where output is influenced by the projected exercise end point.

Noakes and colleagues15,–,18 have proposed a model referred to as the central governor model (CGM) that integrates information processing into the regulation of exercise performance. This is compatible with experimental observations during prolonged exercise such as marathon running,19 1-h cycling exercise20 or middle-distance cycling.6

This model proposes that physical activity is controlled by a central governor in the brain so that the human body functions as a complex system during exercise. Biological changes in peripheral systems would act as afferent signals to modulate control processes in the central nervous system (CNS) in a dynamic, non-linear, integrative manner to optimise performance and to prevent catastrophic failure during or after exercise.21 22

According to St Clair Gibson and colleagues,4 the brain area controlling pace integrates knowledge of the end point into an algorithm that incorporates an interpretation of knowledge of environmental and metabolic conditions to set a specific pacing strategy for a particular exercise bout. Hence, unconscious calculations continuously control emotion arousal and induce the choice of optimal intensity in order that the optimum exercise intensity can be maintained until the end point. Of course, the assessment of afferent feedback information by the brain does not occur only once during an exercise bout but must occur continuously.4 If the algorithm indicates that the pace is too fast such that the athlete will reach exhaustion before the end point of the race, efferent neural commands to the exercising muscle would be modified to reduce the power output. Conversely, if the algorithm indicates that the pace is too slow, the efferent neural commands will be increased to raise the power output. Therefore, using this integrative teleoanticipatory control algorithm,4 muscle power output would be continuously modified throughout the exercise bout by altering the number of skeletal muscle motor units recruited during exercise, thereby continuously varying the work rate and metabolic demand.15 These mechanisms correspond to the unconscious part of the process.15 17 18 23

Conscious control is also identified in the mechanisms of regulation of the intensity.15 17 18 Indeed, the sensation of fatigue is recognised as the conscious awareness originating from a subconscious control24 of the homeostatic, central governor control mechanisms18 and is related to the intensity of interoceptive processes.25 26

It has been proposed that during exercise until fatigue, the rising perception of exertion (RPE) or discomfort progressively reduces the conscious desire to override this control mechanism, so that the exercise intensity is regulated by these symptoms in order to protect the organism from catastrophic failure.18

This invites the question of how unconscious control mechanisms integrate the physiological responses directed by the CNS with the conscious awareness of fatigue and other sensations arising during exercise.

CNS activation mediating physiological homeostatic reflexes and emotional arousal

It is well known that the CNS is activated during exercise, inducing physiological homeostatic reflexes in order that bodily responses are closely matched to the rate of energy expenditure. For many homeostatic reflexes, the principal pathway begins with cranial visceral afferents entering the CNS via the nucleus of the solitary tract (NTS).27 This afferent information is projected to diverse sites, including efferent nuclei within the brainstem, spinal cord and hypothalamus.28 29 These produce a mix of reflexes from the paraventricular nucleus of the hypothalamus that integrates the neuroendocrine, cardiovascular, metabolic and homeostatic responses to exercise.29,–,33

Sympathetic neural outputs constricts vessels and sphincters and decrease gastrointestinal motility whereas parasympathetic nervous outputs produce the opposite responses.34 Peripheral and central chemoreceptors modulate sympathetic and ventilatory outputs.35

Hence, the hypothalamus is known to modulate physiological adaptation in response to physical, environmental, mental and many other stresses,34 integrating the function of the brain and the body.36 During exercise, hypothalamic output increases sympathetic tone and parasympathetic withdrawal in order to match the physiological responses to the rate of energy expenditure.37,–,40

These established physiological facts, among many others, suggest that a crucial role of the CNS during exercise is to match the physiological responses to energy expenditure. However, the question that these studies have not addressed is how these brain processes control the intensity of exercise, particularly through the production of the symptoms of fatigue.

The mechanisms by which afferent sensory feedback from the periphery interact with the so-called higher mental functions have been studied extensively. William James was one of the first to postulate that viscero-afferent feedback is closely linked to emotional experience, stating that “bodily changes follow directly the perception of the exciting fact and that our feelings of the same changes as they occur IS the emotion”.41

The first level of emotional experience phenomenology includes an integration of arousal, including the physiological responses. At the second-level, however, there is a conscious appraisal of these processes: Information from first-level physiological, expressive and experiential reactions is integrated with the behavioural context. Earlier “peripheral” models such as that of James and Lange argue that emotions arise from the interoception of physiological changes within the body.41

Recently, Gray et al42 have suggested that the processing of visceral information is closely related to emotional and motivational aspects of behaviour. Damasio et al43 specified neural networks underlying emotion, feeling and the consciousness of feelings, stressing that the body is the main source of emotions, either directly or via its representations in somatosensory areas of the brain. Moreover, Pollatos et al44 have shown that there is a strong relationship between the cortical processing of emotional stimuli and the perception of physiological state.

The perception of the heartbeat is the most extensively studied interoceptive process of the brain since it can be studied relatively easily. It has been argued that signals from the cardiac mechanoreceptors enter the brain primarily via afferent vagal fibres, projecting viscerotopically to the NTS of the brain stem.45 Axons from cardiovascular structures terminate within the dorsomedial portion of the NTS. Most of the fibres from the NTS project to the parabrachial nucleus, which provides projections to multiple higher centers such as the hypothalamus, thalamus and cerebral cortex. Here, the insular cortex plays a major role as the cortical projection area for viscerosensory input.46,–,50 Correspondingly, the most probable cardioafferent pathway giving rise to conscious visceral perception appears to be the NTS–parabrachial–thalamus–insula.45 Other studies indicated that emotional arousal activates the amygdala51 and that this activation, specifically that of the basolateral amygdala, results in modulation of memory-related processes in the hippocampus.52,–,57 These experiments are of great interest in describing the neurophysiologic mechanisms of the interoceptive processes.

Unfortunately, the mechanisms that determine the perception of effort during exercise are not as well understood as are those involved in the perception of the heartbeat. Nevertheless, it is accepted that there is a precise system of afferent feedback interpretation to monitor the exercise intensity at a subconscious level and that this interpretation mediates the conscious perception of effort.58 Indeed, RPE is well known to be related to chest and active mass muscle parameters,18 58 59 as well as heart rate,60,–,64 oxygen consumption,63 respiratory rate and minute ventilation,65 blood lactate concentrations66 and muscular strain.63 67,–,69 Importantly, no single physiological parameter predicts the RPE during exercise indicating that this is a complex system phenomenon.58 Indeed, the complexity of this system is confirmed by the finding that the RPE rises as a linear function of the exercise duration.70

The role of affective responses and motivation in determining the pacing strategy

According to the CGM, when exercise is performed until fatigue, the rising perception of exertion or discomfort progressively reduces the conscious desire to override this control mechanism, insuring that the exercise intensity is appropriately regulated.18 The evolution of negative affective responses is, therefore, an important component of the mechanisms regulating pacing strategies.

Other studies consider also the possible role of the valence (pleasure–displeasure) of emotion.41 Hence, according to the dual-mode theory,71 if a negative affective valence is observed during exercise, then a positive valence must also exist.

This is supported by numerous studies which show that affective responses are dependent on the intensity of exercise. Affective valence remains positive (pleasant) in low-intensity conditions whereas higher intensities are associated with less favourable affective responses (unpleasant).71,–,73 Hence, as the exercise intensity increases, the RPE rises progressively whereas positive affective responses decrease. Likewise, RPE increases when a constant exercise intensity is maintained until fatigue.20 74 Thus, logically, the relative proportion of positive and negative affective responses must decrease when exercise is performed until exhaustion.

Hence, not only negative but also positive affective responses should be considered when the mechanisms regulating pacing strategies are considered. We propose an affective loading (AL) scale as the difference between negative and positive affective responses during exercise. Hence, using Borg's category-ratio scale (CR-10)75 to quantify these two valences of affective responses, AL could be rated between −10 and +10. Indeed, at the beginning of low-intensity exercise, negative affective responses would be near 0 whereas the positive responses would be near maximal value so that the AL would be near −10. However, when exercise is performed until exhaustion, AL would increase to values nearer +10. Other scales such as the RPE76 scale could be also used; the method for the determination of AL remains the same, only the numbers would be different.

We propose that during exercise, the athlete must monitor not only the physiological reserves but also the AL in order that catastrophic failure in any physiological and emotional system does not occur before the finish line. Alternatively, the AL may represent part of the brain control that contributes to the regulation of the pacing strategy.

The CGM theory proposes that the discomfort that develops during exercise progressively reduces the conscious desire to override this control mechanism.18 This theory invites consideration of the role of motivation in the development of pacing strategies. The popular belief is that this factor is of great importance but is seldom directly studied. Hence, psychologists and exercise physiologists who study human motivation in contexts other than exercise recognise that motivation is goal directed and occurs within the context of self-regulation.77 78 Individuals develop strategies that allow them to reach their goals and mobilise and monitor their behaviours in order to attain their goals.77 Emotional phenomena represent central mechanisms of self-regulation that help humans deal effectively with their environments.77 79 80

Hulleman et al have previously shown that extrinsic motivation provided before a simulated competition did not improve performance in a 1500-m cycle time trial in well-trained cyclists, suggesting that pacing strategies are stable whatever the level of motivation.5 Nevertheless, only extrinsic motivation in the form of a monetary reward was provided in their study. However, this procedure could effect intrinsic motivation negatively.5 81 Moreover, other forms of motivation, such as the presence of high-level competitors, are known to influence pacing strategies. In this case, an all-out strategy is often used.4 To achieve this, athletes must accept more severe discomfort caused by the earlier onset of disturbances in physiological homeostasis and energy storage.5 Thus, motivation may improve unconscious control of physiological homeostasis as described in the CGM.15 18 23 An all-out strategy also means that the athlete must accept a high level of unpleasant sensations, involving the conscious control of the exercise intensity. It means that the level of motivation must be sufficient to overcome these negative sensations.

We propose that, for a fixed level of motivation, AL influences the pacing strategies to reach the preset goal. Hence, we suggest that the desire to sustain high levels of effort depends on the difference between the highest level of acceptance of AL for the expected duration of exercise and the level of AL present at any moment during the exercise bout (figure 1). The more positive are the affective responses during exercise, the greater will be the desire to maintain or to increase the exercise intensity. In contrast, the more negative are the affective responses, the less will be the desire to sustain the exercise intensity; as a result, the exercise intensity will likely decrease.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

The mechanisms of inter-relation between the nature of the task, the physiological responses, the central nervous system and the high mental processes on pacing strategies. The level of acceptance is defined as the highest tolerated affective loading for the expected duration. This parameter could be modulated by psychological factor before and during exercise.

Hence, when exercise is maintained until fatigue, the induced physiological responses would increase the AL, reducing the conscious desire to maintain the exercise intensity, thereby reducing the neural drive to the exercising muscles and thus reducing the exercise intensity. In contrast, when exercise is performed at low intensities, which minimised the physiological responses, the AL value is low, improving the desire to select higher intensities or reinforcing the positive affective responses if the intensity is maintained at that low level. Likewise, when the motivation is lower than the optimal level, the chosen exercise intensity would be lower for the same AL.

Hence, when the brain continuously performs subconscious calculations, by comparing the physiological demands with the athelete's physiological capacity, the level of motivation to support a high level of AL is also compared with the AL imposed by the selected exercise intensity in order to ensure that the exercise can be maintained for the expected duration without any catastrophic failure in physiological and emotional homeostasis. It seems probable that emotional arousal likely takes longer to develop than do the processes involved in physiological control. If correct, this might suggest that unconscious physiological processes should be considered as the most probable processes regulating rapid adjustments in power output. These would be the most important processes active during events of intermediate duration.5 6 8 Conscious processes may be active to reinforce subconscious controls during more prolonged exercise.

Hence, unconscious/physiological and conscious/emotional processes may interact with each other in order to regulate the pacing strategy. Since either process could regulate the performance under different circumstances, it follows that one aim of training should be to ensure that both processes are optimally adapted at the exact moment of the competition.

The role of memorised emotions on pacing strategies

Optimum pacing strategies occur in athletes who have developed a stable template of the appropriate time/distance versus power output profile for exercise of different durations.6 Foster et al suggest that this ability is learnt early in an athlete's experience.13 Memorised emotions developed by previous training could be part of this ability.

Hence, when a subject has no prior experience of a particular exercise, the representations of the task are imprecise because no element of comparison from a prior event is available in the memory. Hence, the expected time to cover a distance cannot be precisely estimated. Likewise, physiological capabilities and the capacity to resist the sensations of fatigue are not yet known. As a result, the optimum pacing strategy cannot be determined either before or during the exercise bout. Herbert et al25 26 have shown that there is a strong association between interoceptive sensitivity and the intensity of emotional experience, which corresponds to the quantity of memorised emotions. In accordance, we argue that cognitive information developed by each training session would reinforce the interoceptive sensitivity and the physiological awareness.

In a recent study, Foster et al have shown that the pattern of learning the performance template is primarily related to increased confidence that the trial can be completed without unreasonable levels of exertion or injury.82 Hence, pacing strategies can be thought of as a complex process in which the athlete controls the intensity of exercise at any time by taking into consideration physiological and psychological reserves in relation to the estimated time until the finish83 at that exercise intensity.

It could be suggested that training sessions improve the athlete's pacing strategies by associating a level of AL with the capability to maintain that effort for a specific duration without a catastrophic loss of performance. Hence, each memorised emotion for each training session would improve the quantity and quality of stored data (figure 1). Using these memorised data, the athlete would be able to compare the antecedent AL with those induced by the current exercise bout. Thus, the athlete would become able to efficiently control the exercise intensity,4 by selecting the appropriate AL level to sustain the exercise intensity for the entire expected duration.

The pacing strategy is dependent on the pre-established template and considerable time and effort may be required to change a specific strategy that has been followed frequently.84 St Clair Gibson et al have suggested that the motor sequences representative of the performance template are almost certainly programmed into the motor cortex from prior athletic performance, potentially from childhood.85 Thus, an interesting possibility is that training methods could be designed, even during childhood, to improve the efficiency and stability of the pacing strategies.

Perspectives in training

The CGM proposes that the CNS is an important determinant of the exercise performance. It is also known that the adoption of positive emotions can improve performance. For this reason, elite athletes adopt sophisticated mental preparation techniques. We argue that the emotional characteristics of training could be of great importance in the quest to achieve maximal performance in competition. Indeed, if experience is recognised as an important factor in pacing strategies,14 the valence (pleasure–displeasure) of memorised emotions induced by previous training exposures should also be taken into account during training.

Foster86 has proposed a method for monitoring training that integrates the exercise session RPE and the duration of the training session in order to minimise undesirable training outcomes as the so-called overtraining87 or underperformance syndrome.88 Overtraining is primarily related to sustained high-load training, which increases training monotony, and underperformance is often associated with frequent infections and a depressed mood state.87

Underperformance and chronic fatigue may depend on impaired transmission of ergotropic signals to target organs and to represent a complex strategy against an overload-dependent cellular damage.89 This corresponds to the unconscious control in which an unexpected sense of effort is also observed.88 This could be envisaged as a component of the process that protects against organ disturbance.

Athletes and coaches often ignore the symptoms of overtraining until performance is affected.87 Even sophisticated measurements of specific physiological markers such as neuro-endocrine variables do not identify the presence of overtraining until the athlete's exercise performance is affected.90 The most sensitive parameter for identifying overtraining appears to be the RPE86 because physiological systems are more robust and probably become affected at a later stage in the overtraining process.90 The RPE should be used to detect overtraining in the early stage and to monitor training.86 90 Overtraining may also influence the evolution of positive affective and AL responses, which might also be used to adjust training before underperformance occurs.

If the object of training is to improve the physiological responses in order that a greater physiological stress can be sustained during exercise, in the same way, training could also be designed to ensure that a more demanding emotional loading could also be accepted by the athlete. Hence, it might be proposed that the athlete should be trained also to accept high levels of AL during training and competition. However, training at lower levels of AL must also be performed in order to restore and to maximise not only the physiological capacities but also the optimal AL level. If this is ignored, overload of both the physiological and cognitive mechanisms will occur, leading to underperformance. This analysis predicts that underperformance can be due not only to physiological but also to emotional disorders, as it is often suggested by elite athletes.

Thus, training programs must be designed so that they are both physiologically efficient and psychologically acceptable.

We propose that previously memorised emotions may act either positively or negatively on higher mental processes, notably on the level of acceptance of AL for an expected duration of exercise, rather than directly on the level of AL58 (figure 2). Indeed, Hampson et al58 have already shown that either deception or other psychological interventions that alter pre-exercise expectations have a minimal feedforward effect on the exercise outcome as also recently confirmed.91

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Influence of memorised emotions of exercise (E) on the level of acceptance of affective loading of next exercises (E+n).

When the level of acceptance of AL is increased by previous training sessions, the desire to select high intensities would be increased for the same AL, but in contrast, when the level of acceptance of AL is decreased by inappropriate training sessions, the desire to select high intensities would be decreased for the same AL.

In fact, we suggest that training should be viewed as a “mental education” in order that athletes can approach as closely as possible their maximal performance allowed by their own physiological and emotional processes that control the fatigue mechanisms.

We propose that the final sessions before the competition must be constructed to remove monotony and to restore the positive emotions and self-confidence of the athlete. This is compatible with the traditional precompetitive training period which emphasises the intensity but decreases the volume of training. In these conditions, the athlete would be brought to an optimal psychological state to approach their intrinsic physiological limits.