Carbohydrate sensing in the human mouth: effects on exercise performance and brain activity

Abstract

Exercise studies have suggested that the presence of carbohydrate in the human mouth activates regions of the brain that can enhance exercise performance but direct evidence of such a mechanism is limited. The first aim of the present study was to observe how rinsing the mouth with solutions containing glucose and maltodextrin, disguised with artificial sweetener, would affect exercise performance. The second aim was to use functional magnetic resonance imaging (fMRI) to identify the brain regions activated by these substances. In Study 1A, eight endurance-trained cyclists (VO2max 60.8 +/- 4.1 ml kg(-1) min(-1)) completed a cycle time trial (total work = 914 +/- 29 kJ) significantly faster when rinsing their mouths with a 6.4% glucose solution compared with a placebo containing saccharin (60.4 +/- 3.7 and 61.6 +/- 3.8 min, respectively, P = 0.007). The corresponding fMRI study (Study 1B) revealed that oral exposure to glucose activated reward-related brain regions, including the anterior cingulate cortex and striatum, which were unresponsive to saccharin. In Study 2A, eight endurance-trained cyclists (VO2max 57.8 +/- 3.2 ml kg(-1) min(-1)) tested the effect of rinsing with a 6.4% maltodextrin solution on exercise performance, showing it to significantly reduce the time to complete the cycle time trial (total work = 837 +/- 68 kJ) compared to an artificially sweetened placebo (62.6 +/- 4.7 and 64.6 +/- 4.9 min, respectively, P = 0.012). The second neuroimaging study (Study 2B) compared the cortical response to oral maltodextrin and glucose, revealing a similar pattern of brain activation in response to the two carbohydrate solutions, including areas of the insula/frontal operculum, orbitofrontal cortex and striatum. The results suggest that the improvement in exercise performance that is observed when carbohydrate is present in the mouth may be due to the activation of brain regions believed to be involved in reward and motor control. The findings also suggest that there may be a class of so far unidentified oral receptors that respond to carbohydrate independently of those for sweetness.

Figure 1. A single trial of stimulus and control delivery

The stimulus was delivered at time 10 s and swallowing (SW) cued after 10 s and completed within a 2 s period. The control solution was delivered at 40 s and, 10 s later, swallowing was cued (SW). Each full trial lasted 52 s.

Figure 3. Average power output calculated at intervals during the trials compared to PLA during A, the GLU trial and B, during the MALT trial

Data are mean ±s.e.m.

Figure 2. Individual (open columns) and mean ±s.e.m. (filled columns) percentage change of power output compared to PLA during A, the GLU trial and B, the MALT trial

*Significant difference from PLA (P= 0.007) and ** (P= 0.012).

Figure 4. Activations in the insula/frontal operculum, the dorsolateral prefrontal cortex (DLPFC), the striatum and the cingulate cortex by the contrasts A, [Glucose – Control] and B, [Saccharin – Control]

The blue circles indicate the region where the peak voxel of activation was found (see Table 1) in the group analysis. The colour bar indicates the Z-score significance level. Clusters were formed by thresholding the Z-stat image for each contrast at Z > 2.7 (equivalent to a one-tailed P of 0.003). The y values are with respect to the MNI co-ordinate system.

Figure 5. Areas of common activation from the glucose and saccharin contrasts ([Glucose – Control] AND [Saccharin – Control])

A, the insula/frontal operculum (e.g. MNI coordinates: 54, 12, 14) and B, the dorsolateral prefrontal cortex (DLPFC) (e.g. MNI coordinates: −38, 38, 12). C and D demonstrate the average change in BOLD response produced by glucose, saccharin and their respective control solutions within the cluster of common activation highlighted in the insula/frontal operculum and the DLPFC, respectively. *Significant difference between solutions (P < 0.05); **(P < 0.001). Values are mean ±s.e.m.

Figure 6. Activations in the insula/frontal operculum, the orbitofrontal cortex (OFC), the dorsolateral prefrontal cortex (DLPFC), the striatum, and the cingulate cortex by the contrasts A, [Glucose – Control] and B, [Maltodextrin – Control]

The blue circles indicate the region where the peak voxel of activation was found (see Table 2) in the group analysis. The colour bar indicates the Z-score significance level. Clusters were formed by thresholding the Z-stat image for each contrast at Z > 2.7 (equivalent to a one-tailed P of 0.003). The y values are with respect to the MNI co-ordinate system.

Figure 8. Areas of common activation from the glucose and maltodextrin contrasts ([Glucose – Control] AND [Maltodextrin – Control])

A, the dorsolateral prefrontal cortex (DLPFC) (e.g. MNI coordinates: −28, 54, 20), B, the striatum (e.g. MNI coordinates: 10, 20, −10) and C, the anterior cingulate cortex (e.g. MNI coordinates: 8, 44, −4). D, E and F, show the average change in BOLD response produced by glucose, maltodextrin and their respective control solutions within the cluster of common activation highlighted in the dorsolateral prefrontal cortex, the striatum and the anterior cingulate cortex, respectively. *Significant difference between solutions (P < 0.05); **(P < 0.001). Values are mean ±s.e.m.

Figure 7. Areas of common activation from the glucose and maltodextrin contrasts ([Glucose – Control] AND [Maltodextrin – Control])

A, the insula/frontal operculum (e.g. MNI coordinates: 58, 8, 20) and B the medial orbitofrontal cortex (OFC) (e.g. MNI coordinates: 18, 34, −16). C and D show the average change in BOLD response produced by glucose, maltodextrin and their respective control solutions within the cluster of common activation highlighted in the insula/frontal operculum and the medial OFC, respectively. *Significant difference between solutions (P < 0.05); **(P < 0.001). Values are mean ±s.e.m.