Introduction
Exercise performance during an acute exposure to hypoxia is impaired via a reduction in arterial oxygen content (Fulco et al.
1996,
1998; Calbet et al.
2003a; Amann et al.
2006b; Romer et al.
2006). In moderate hypoxia, where the oxygen fraction of inspired air is reduced to ~ 13–15% (F
IO
2 0.13–0.15), decrements in performance have been attributed to a rise in peripheral markers of muscle fatigue, which generate afferent feedback to down-regulate motor output from the central nervous system (CNS) (Amann et al.
2006b,
2007; Romer et al.
2007). A so-called ‘sensory limit’ (Gandevia
2001), therefore, restricts the manifestation of peripheral fatigue to prevent catastrophic failure of any one system. In severe hypoxia (F
IO
2 < 0.115), larger reductions in exercise capacity have been described despite relatively less evidence of peripheral fatigue (Amann et al.
2007). Here, a hypoxia-sensitive ‘central’ component of fatigue mediates a reduction in central motor output via brain hypoxia (Subudhi et al.
2009; Vogiatzis et al.
2011; Millet et al.
2012; Goodall et al.
2012), thus limiting maximal exercise capacity. Indeed, in experiments where F
IO
2 is increased at the point of task failure, exercise performance can be prolonged in severe and moderate hypoxia (Amann et al.
2007; Torres-Peralta et al.
2016).
Central processing of the perception of effort and its role in setting exercise intensity is heavily debated. The subjective rating of perceived exertion, termed RPE, is a psychophysiological concept (Borg
1982; Morgan
1994) that centrally integrates perceptual, peripheral, experiential, and environmental sensory cues (Hampson et al.
2001). Indeed, Borg (
1982) described the RPE as a conscious representation of multiple inputs. To further understand how perceived exertion modulates self-regulated exercise, a fixed-RPE protocol was developed, referred to as the RPE clamp (Tucker et al.
2006). Here, participants exercise at a pre-determined fixed level of perceived exertion on the RPE scale (typically 16, < ‘very hard’) and modulate their workload according to the perceived mismatches between the expected and actual RPE (Tucker
2009). In an eloquent design, a recent study controlled the rate of arterial hypoxemia (SpO
2 98 to 70%) through manipulations in F
IO
2, demonstrating that the rate of decline in power output was reliant on the rate change of arterial oxygenation during self-regulated exercise performance (Farra et al.
2017). Whilst this provides evidence of the relationship between SpO
2 and the perception of exercise intensity in contrived ambient conditions, it overlooks the responses under fixed reductions in F
IO
2, such as that commonly encountered at altitude. This is important, since exposure to steady-state F
IO
2 conditions provides an opportunity for physiological compensations, such as increased muscle oxygen delivery or extraction. Based on their findings, the rate of change in SpO
2 under steady-state F
IO
2 is likely to determine exercise perception and tolerance, yet this is not currently known. These collective organ-level changes could feasibly offset the deleterious effects of hypoxia or, more importantly, complicate the afferent feedback process. We hypothesized that, during exercise at a fixed RPE, power output would decrease in accordance with a reduction in F
IO
2. Therefore, we examined exercise performance using a fixed RPE protocol in severe and moderate hypoxia relative to normoxia, with the aim of determining the relationship between time to exhaustion and a combination of acute physiological responses.
Discussion
The purpose of the study was to investigate the effect of severe and moderate hypoxia on exercise performed at a fixed RPE in reference to normoxia. As anticipated, our findings demonstrate that performance time was diminished when exposed to decreasing FIO2, meaning that participants down-regulated their work load as a result of increasing levels of hypoxia. Increases in breathing frequency and blood oxygen desaturation during the early stages of exercise were correlated with reductions in task performance. Despite these changes, oxygen extraction at the muscle (as indicated by NIRS) appeared to be tightly regulated to match the metabolic demand, suggesting that muscle oxygenation is not involved in determining perception during the early stages of setting exercise intensity. Together, the early rate of change in ventilation and arterial hypoxemia appears to drive the selection of exercise intensity associated with a fixed RPE in hypoxia.
Reductions in exercise performance in hypoxia have been attributed to depleted arterial oxygen content (Fulco et al.
1996,
1998; Calbet et al.
2003a; Amann et al.
2006b; Romer et al.
2006). Breathing hypoxic gas leads to a decrease in the arterial partial pressure of oxygen, oxygen saturation of haemoglobin, and the amount of oxygen dissolved in the plasma. Consequently, arterial oxygen content is reduced. Here, a reduction in F
IO
2 decreased exercise time by ~ 72% in severe hypoxia (F
IO
2 < 0.115) and by ~ 33% in moderate hypoxia (F
IO
2 ~ 0.15), relative to normoxia. There was some evidence of inter-individual responses with two participants showing a reduced sensitivity to hypoxia. Whilst a number factors have been presented to explain responders and non-responders to hypoxia and altitude (Fulco et al.
1998), these individuals performed relatively poorly in normoxia and recorded low aerobic capacity, suggesting that this may be related to fitness status. Across the group, SpO
2 was maintained within 3% of resting levels during normoxia; however, upon acute exposure to moderate hypoxia, resting S
PO
2 was reduced by ~ 4% and decreased by a further 8% across the exercise trial. In severe hypoxia, these reductions in SpO
2 were much greater, decreasing by ~ 14% at rest, with a further decrease of 14% (S
PO
2 ~ 72%) observed at end-exercise. Reductions in exercise performance in moderate hypoxia have largely been attributed to peripheral mechanisms, where a decrease in arterial oxygen content and impaired oxygen delivery to the working muscle leads to a subsequent metabolic perturbation (Hogan et al.
1999). This work has been advanced by studies describing comparable levels of peripheral muscle fatigue via evoked maximal contractions following exhaustive exercise in normoxia and hypoxia despite a substantial reduction in exercise time (Amann et al.
2006b; Romer et al.
2006,
2007; Goodall et al.
2012). Greater reductions in performance described in severe hypoxia have been attributed to greater impairments in pulmonary gas exchange, reduced limb blood flow, and reductions in cardiac output (Calbet et al.
2003a). However, experiments demonstrate rapid improvements in exercise performance and cerebral oxygenation following a fast transition from breathing gas that is severely hypoxic to hyperoxic, at the point of exhaustion, supporting a central role (Amann et al.
2006a).
We examined how the initial exposure to a hypoxic environment would impact determination of the exercise intensity associated with 16 on the Borg RPE scale (Borg
1982). Changes in perceived exertion may be determined during the beginning stages of exercise via a range of perceptual, peripheral, experiential, and environmental sensory cues to enable task completion within the physiological limits of the body (Hampson et al.
2001; St Clair Gibson et al.
2006). In both normoxic and moderate hypoxic conditions, the selected power output achieved within several minutes was comparable; however, in severe hypoxia, power output was reduced by ~ 18%. Whilst the ability to generate maximal power is unaffected by severe levels of hypoxia (Calbet et al.
2003b), supporting the argument that motor drive is unaffected upon acute exposure, the heightened perception of effort observed here may reflect a detrimental effect on decision-making processes (Niedermeier et al.
2017), cognition (McMorris et al.
2017), or a teleoanticipatory reduction in power output to maintain homeostasis (St Clair Gibson et al.
2006). During exercise in hypoxia, cerebral vascular conductance is continually adjusted to maintain oxygen delivery when a reduction in arterial oxygen concentration occurs (Curtelin et al.
2018). However, reductions in cerebral oxygenation have been described at rest and during exercise in hypoxia (Subudhi et al.
2009). Whilst the brain can compensate by increasing oxygen extraction (Gonzalez-Alonso et al.
2004), neuronal function can also be inhibited (Neubauer et al.
1990) which may impact higher cognitive functions. Whilst we did not directly measure cerebral oxygenation or cerebral blood flow, we did observe a reduction end-tidal carbon dioxide (PETCO
2) in severe hypoxia relative to normoxia. A close relationship exists between PETCO
2 and cerebral blood flow (Ide et al.
2003), suggesting that cerebral blood flow may have been reduced during exercise in severe hypoxia. Hypocapnia reduces cerebral blood flow by as much as ~ 3% for every ~ 1 mmHg change in PETCO
2 (Ringelstein et al.
1992). Based on an observed ~ 8 mmHg difference during exercise in normoxia and severe hypoxia (Fig.
4a), this could equate to a ~ 24% reduction in cerebral blood flow. It should also be noted that PETCO
2 sensitivity is increased by acute exposure to hypoxia (Jensen et al.
1996; Poulin et al.
2002) and the effect on PETCO
2 sensitivity has shown differential effects (Fortune et al.
1992; Vovk et al.
2002). Therefore, these observations should be further explored.
At submaximal exercise intensities when arterial oxygenation is reduced, it is probable that oxygen delivery to exercising muscles is compromised. However, compensatory mechanisms increase oxygen extraction or blood flow to maintain muscle oxygen supply (Amann and Calbet
2008). Remarkably, vastus lateralis oxygen saturation assessed via NIRS, did not differ, irrespective of the level of arterial hypoxemia experienced, suggesting comparable oxygenation in the primary exercising muscles during cycling. This has been previously observed during submaximal exercise in both severe and moderate hypoxia (Millet et al.
2012). Here, it was reasoned that afferent signalling, emanating from the metabolic environment at the muscle, is unlikely to have changed between conditions and, therefore, could not explain the reductions in exercise performance. On this basis, it was postulated that tissue deoxygenation was unlikely to play a role in determining exercise intensity. However, the metabolic adaptations and changes in blood flow to maintain constant tissue oxygenation may have contributed to afferent signalling. In the current study, the rate of change in TSI% during the first 3 min of exercise did not differ according to F
IO
2, yet it transitioned to a lower steady state from the onset of exercise across all the conditions. Therefore, working muscles were able to match oxygen delivery and extraction to meet the metabolic demand during submaximal exercise, despite an increasing physiological perturbation.
Reductions in F
IO
2 also corresponded with increases in minute ventilation. End-exercise minute ventilation was augmented by 41% and 24% in severe and moderate hypoxia, respectively, compared to normoxia. Ventilation during exercise is controlled via a balance of centrally mediated feed-forward commands and peripheral feedback that increases the rate and depth of breathing according to the demands imposed by the exercise intensity (Kaufman and Forster
1996). Changes in the partial pressure of blood gases are sensed by both central and peripheral chemoreceptors. In the brain, chemoreceptors respond to changes in brain tissue CO
2/[H
+] (Nattie and Li
2009; Tipton et al.
2017). Peripherally, chemoreceptors located in the carotid artery respond to low arterial O
2 and high arterial CO
2. The increased ventilatory response during the majority of the exercise trial was achieved by an increased tidal volume, which is typically the most efficient mechanical way to increase minute ventilation commensurate with metabolic needs and decreasing arterial blood gas tensions (Tipton et al.
2017). During the early stages of exercise, the rate of increase in breathing frequency was inversely related to the reductions in ambient F
IO
2 and was the strongest predictor of exercise time in hypoxia. The sensations of ventilation and breathing discomfort are consciously monitored during exercise (Robertson
1982). Indeed, the relationship between RPE and ventilation is well established (Cafarelli and Noble
1976; Robertson
1982; Killian
1998; Nicolo et al.
2016). In hypoxia, it is, therefore, possible that the rapid change in ventilation during exercise may have potentiated a greater conscious awareness, contributing to both the initial setting of exercise intensity and the modulation of perceived exercise intensity thereafter.
In an attempt to further examine the acute effects of hypoxia, we determined the rate change of the physiological measures taken during the first 3 min of exercise when an approximate steady state was obtained. We found that rate changes in breathing frequency and blood oxygen desaturation showed moderate-to-strong correlations with performance time. A recent study by Farra et al. (
2017) elegantly demonstrated that, when the rate of SpO
2 was altered via F
IO
2, faster arterial deoxygenation resulted in a greater decline in perceptually controlled exercise performance (Farra et al.
2017). They suggested that RPE was sensitive to both the rate of change and absolute magnitude of arterial deoxygenation, which we can partially support based on a fixed hypoxic environment. Interestingly, in contrast to our findings, they reported no difference in breathing frequency between the experimental conditions (fast, medium and slow desaturations). This may reflect the gradual reduction in S
PO
2 controlled by Farra et al. (
2017), which may have masked the relative contributions of the other candidate sensory cues, such as breathing rate, that combine to inform higher brain centres of the homeostatic disturbances. Indeed, these early cues appear to influence the selection of the initial exercise intensities and the subsequent power output that is sustainable for the entire exercise trial. Future work should explore the relationship between breathing frequency and exercise time in hypoxia and practical solutions to reduce breathing frequency may facilitate improvements to performance.
The relationship between exercise regulation, pacing, and RPE is still heavily debated. A three-dimensional framework has recently been proposed as a multidimensional model of volitional self-regulatory control and perceived fatigability (Venhorst et al.
2018). The model combines a sensory-discriminatory dimension (peripheral and central sensations), an affective-motivational dimension (arousal and motivation), and a cognitive-evaluative dimension (exertion and task aversion) (Venhorst et al.
2018). Importantly, this model accounts for both external and internal mediating factors in the generation of RPE. These multiple inputs are, therefore, continually processed, integrated, and interpreted, consciously or otherwise, to alter pacing behaviour in anticipation of potential threats to homeostasis (Hampson et al.
2001; Noakes
2004; Tucker
2009; Venhorst et al.
2018). Such complex psychophysiological interactions, therefore, provide a construct for the observed behavioural differences in pacing in severe and moderate hypoxia when exercising according to a fixed RPE. As we reported, the rate of decrease in power output did not differ between conditions once peak power was achieved; hence, the early setting of an acceptable perceived exercise intensity appears crucial to exercise performance. Whilst it is likely that the exercising template is updated as exercise ensues (Brick et al.
2016), the interplay between such dimensions in generating RPE when challenged with reduced F
IO
2 will require further investigation. Therefore, we propose that the early setting of task intensity in a hypoxic environment is chiefly based upon two primary physiological cues of ventilation and SpO
2, thus determining performance.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.