Using NIRS, the results of this study are the first to our knowledge to demonstrate the novel findings that cerebral oxygenation and blood volume changes in female CFS subjects were significantly different from control subjects during incremental maximal exercise. These data support previous research that cerebral oxygenation is reduced in CFS under other experimental conditions (i.e. orthostatic intolerance test) (Tanaka et al., 2002), and support our hypothesis that cerebral differences exist between CFS and sedentary control subjects. Collectively, these data suggest that there is a link between impaired cerebral oxygenation and chronic fatigue during a maximal exercise challenge.
Cerebral oxygenation and haemodynamics
The interpretation of our NIRS data is as follows: first, during incremental exercise, there was an activation of the brain as reflected by the increased HbO2, tHb, HHb and decreased TOI% (reflecting increased O2 extraction) in both groups (Figs 1–4). Previous research has shown that changes in cerebral oxygenation is a reflection of neuronal activation (Ferrari et al., 2004; Bhambhani et al., 2007). Second, there was a significant difference (i.e. less change) in cerebral HbO2, tHb and HHb in the CFS versus the CON subjects. In addition, TOI decreased to a greater extent in the CON group than the CFS group (by 64·5%). Taken together, these results indicate that blood flow was likely compromised during incremental exercise in the CFS subjects, as reflected by less change in tHb (NIRS-generated tHb has been used as an indirect measure of blood flow (Nioka et al., 2006), and less oxygen transport and utilization (extraction) by the brain (as reflected by the lesser change in HbO2 and HHb in the CFS group).
It is well accepted that oxygen uptake is increased in the brain during exercise (Ide & Secher, 2000; Ferrari et al., 2004; Bhambhani et al., 2007; Wolf et al., 2007) when monitored using NIRS. Of the NIRS variables calculated, TOI is one of the most reliable parameters as it reflects the dynamic balance between O2 consumption and utilization, and is independent of the pathlength of near-infrared photons in cerebral tissue (Ferrari et al., 2004). In this study, there was a 3·8% decrease in TOI in the CON group, in comparison with only 1·3% decrease in the CFS group during the exercise period (Fig. 4). This equates to a 64·5% difference in the amount of O2 extraction between the two groups. In support of our observations, and under conditions of orthostatic intolerance, Tanaka et al. (2002) also used non-invasive NIRS to show that the majority of the CFS subjects in their study had decreased oxy-Hb concentration ([oxy-Hb]) in the brain during upright posture. They hypothesized that a reduced perfusion pressure and cerebral vasoconstriction may partly explain the reduction in [oxy-Hb]. This would support recent research by Rasmussen et al. (2007), which showed that a reduced cerebral oxygen delivery had a direct effect on motor performance. Thus, it is likely that the impaired motor performance demonstrated in that study (Rasmussen et al., 2007) as a result of the inadequate oxygen delivery to the brain resulted in the observed early onset of central fatigue that we observed in our CFS subjects in the current study, as demonstrated by the reduction in HbO2, tHb and HHb in comparison with the CON subjects.
Research has shown that cerebral blood flow is reduced in subjects with CFS when using transcranial Doppler sonography (Ichise et al., 1992; Yoshiuchi et al., 2006). In particular, Ichise et al. (1992) showed a significant blood-flow reduction in multiple regions of the brain of CFS subjects, including the prefrontal cortex. Some have suggested that this reduction in blood flow is related to an autonomic (cerebral autoregulation) dysregulation also noted in subjects exhibiting neurally medicated syncope under a number of different experimental conditions (Stewart et al., 1998; Yamamoto et al., 2003). Because cerebral autoregulation, metabolic regulation of O2 and CO2-mediated vasodilation are the most important mechanisms to ensure cerebral blood flow (Nybo & Rasmussen, 2007), our results would support the premise that the central nervous system of CFS subjects is somehow altered, and support previous research that suggests a CNS mechanism(s) is implicated in the pathogenesis of CFS (Georgiades et al., 2003; Chaudhuri & Behan, 2004; Siemionow et al., 2004b).
Third, it is noteworthy to mention here that cerebral HbO2 and tHb plateau at approximate 90% TTE (Figs 1 and 2). This supports previous research by others using healthy and active individuals (Bhambhani et al., 2007). This plateau in HbO2 and tHb before the termination of exercise is the result of a decline in end-tidal CO2 (PETCO2) and arterial CO2 content (PaCO2) that occurs above the respiratory compensation threshold (RCT). When exercise intensity exceeds the RCT, the decreased PaCO2 results in a reduction in the local cerebral blood flow (Bhambhani et al., 2007; Nybo & Rasmussen, 2007). Thus, we have demonstrated for the first time that this also occurred in the CFS subjects in our study but at a lower threshold level than CON subjects. Therefore, although our CFS subjects demonstrated a similar response above the RCT as did the CON subjects, i.e. a decline in cerebral oxygenation and total blood volume at maximal exercise, this would suggest that their cerebrovascular reactivity to the changing PaCO2 levels must still be functional. However, the significant differences in cerebral metabolism (i.e. HbO2 and tHb) between groups during submaximal and maximal levels of effort still suggests that cerebral blood flow regulation must be compromised in CFS sufferers. Unfortunately, we do not have blood pressure or direct blood flow measurements to confirm whether cerebral autoregulation was compromised.
Cardiovascular and performance variables
Although our groups were matched for sex, body size and general activity level, significant differences were found in their aerobic ability. This supports previous research that CFS subjects have a lower aerobic capacity than normal healthy untrained subjects of the same age and sex, whether measured using a submaximal predictive exercise test (Fulcher & White, 2000;Nijs et al., 2007) or a direct VO2max test (Inbar et al., 2001; Nijs & De Meirleir, 2004a). We used the predictive equation reported by Nijs and De Meirleir (Nijs & De Meirleir, 2004a), which has been shown to reliably predict VO2peak in CFS subjects. The CON subject’ VO2peak was predicted using the ACSM formulae (ACSM 2000) from the last steady-state (2-min) stage the subject was able to achieve. We acknowledge that we may have possibly overestimated VO2peak by using this predictive equation (ACSM). However, we did control for this by using the same predictive equations (ACSM 2000; Nijs & De Meirleir, 2004a) on each data set and found no significant differences between compared equations. Thus, we believe that we have provided a fair comparison between groups and this did not affect the cerebral oxygenation results as presented here. Therefore, we are confident that our calculated aerobic capacity for the subjects in this study is both valid and reliable. The reason for the reduced aerobic capacity in CFS subjects is speculative, but it has been suggested that both central (Pagani & Lucini, 1999) and peripheral (McCully & Natelson, 1999) factors contribute. Therefore, it is possible that both deconditioning and physiological factors associated with the reduction in cerebral oxygenation and blood flow limit exercise in CFS subjects. Further research is needed to confirm whether peripheral factors, such as muscle oxygenation and blood volume changes are altered in CFS during maximal incremental exercise, and postexercise recovery.
Heart rate was also significantly lower in our CFS (154 ± 13 beats min−1) versus CON (186 ± 11 beats min−1) subjects at maximal power output. The performance data showed that the peak PO and TTE were significantly lower in the CFS, reflecting their inability to perform incremental exercise for an extended period of time. The mean peak PO was 100 ± 39 W and 163 ± 34 W for the CFS and CON groups, respectively. These data are similar to other maximal values reported in the literature (Inbar et al., 2001; Nijs et al., 2004b; Wallman et al., 2004a).
We also used the HR:RPE ratio as a method to compare the actual and perceived exertion of the subjects during the incremental exercise test, and thus a way to control for self-reported fatigue (Cook et al., 2003b). The lower HR:RPE ratio for the CFS subjects indicates that at the same HR (or absolute workload), RPE was significantly higher, and this was a consistent finding for each of the three common submaximal workloads that were monitored (Fig. 5). Furthermore, HR was not different between groups for these workloads (35, 60, 85 W). This suggests that CFS individuals perceive their level of effort as being more difficult, and these results are in agreement with most previous research (Fulcher & White, 2000; Cook et al., 2003b; Georgiades et al., 2003; Wallman et al., 2004b), but in disagreement with others (Cook et al., 2003a). Whether these differences confirm that CFS individuals have an altered central dysregulation of perception of effort cannot be fully answered by our data. However, our documented changes in cerebral HbO2 and tHb provide evidence that physiological differences do exist between CFS and CON subjects. It is possible that the reduced oxygenation (and blood volume) delivery to and utilization by the brain alter neural function and perception of effort. Reduced neural activation to the working muscles was demonstrated in CFS subjects by Kent-Braun et al. (1993) using twitch interpolation methods during maximal voluntary contractions. Furthermore, Wallman et al. (2004a) have suggested that an increase in the effort sensation may also occur as a result of a reduced neural drive to the working muscles in CFS sufferers owing to psychological factors, such as fear of pain and fear of relapse (Inbar et al., 2001). This can result in a lack of motivation (Inbar et al., 2001), which may consequently give rise to the habitual inhibition or reduced facilitation of motor unit recruitment as demonstrated by Kent-Braun et al. (1993). Siemionow et al. (2004b) also showed that brain (EEG) signals were altered and significantly different in CFS versus CON when performing motor tasks, and furthermore, their results indicated that stronger voluntary efforts were needed to perform the same motor tasks as CON subjects. Collectively, these results clearly showed that CFS involves altered central nervous system signals in controlling voluntary muscle activities, especially when the activities induce fatigue. Thus, the dysregulation of RPE may be related to the reduced cerebral oxygenation and blood volume changes that we observed, but requires further experimentation.
Potential limitations of the study
Ideally, it would be better to have a larger sample size of CFS subjects, especially as CFS is an unknown disorder with many reported complications. This potentially could have increased the variance of our NIRS data (owing to potential heterogeneity of the disorder). However, when we examined the standard deviation of both the CFS and CON groups, this was not significantly different. Therefore, our small sample size did not contribute to the significant differences between the groups. In future studies, it is recommended that a large sample size be tested to confirm our results. Furthermore, we believe that our cerebral oxygenation data clearly differentiates CFS from CON subjects, even though the data was recorded from the left frontal lobe only. Future research should include ‘imaging’ the brain with NIRS probes on more regions of the brain to provide a global reflection of brain oxygenation, as it has been shown that blood flow is altered in different regions of the brain in CFS subjects (Ichise et al., 1992). A direct measurement of blood velocity (flow) using transcranial Doppler, simultaneous with NIRS and blood pressure, would also assist to verify changes in cerebral blood flow and a direct measure of cerebral autoregulation under conditions of rest, exercise and recovery. Finally, although we matched our groups for physical activity level [and did confirm that the current physical activity level of our groups was similar; Canadian Society for Exercise Physiology (CSEP), 2004)], it is possible that the CFS subjects were more de-conditioned because of their inability to carry out additional activities of normal living. However, all subjects stated that they exercised to their maximal ability when performing the maximal incremental exercise protocol.