Simultaneous Gut-Brain Electrophysiology Shows Cognition and Satiety Specific Coupling

Figures

• 

  Figure 1. Procedure schematic. (A) Recording setup and electrode layout used for our 24-channel electroencephalography (EEG), highlighted with the 4 different regions (each 2 electrodes, left fronto-central (median of signals from electrodes F3, C3), right fronto-central (F4, C4), left parieto-occipital (P3, O1), right parieto-occipital areas (P4, O2) used for study analysis), and 8-channel electrogastrography (EGG, red: reference, blue: ground) is shown. (B) Task design for the working memory task is shown.

• 

  Figure 2. Phase Amplitude Coupling (PAC). (A) Illustration for computing PAC with instantaneous amplitude of the fast frequency EEG signal and instantaneous phase of the slow frequency EGG signal, and polar plot representation of the resultant coupling mean vector magnitude (x) at phase (y) (B) Illustrative examples of high coupling at phase ~100 degrees (top), and dispersed coupling spread over various phases (bottom).

• 

  Figure 3. Procedure schematic. (A) Global power spectral density (PSD) across all participant data showed a dominant EEG rhythm peaking in the alpha band (9–14 Hz) and dominant EGG rhythm peaking in the 0.03–0.07 Hz range. (B) Phase and normalized magnitude of the EEG-EGG analytical signal shown as polar plots (see Methods Figure 2) for two exemplar participants in the left fronto-central brain region (F3, C3), these participants data suggest diversity in brain-gut coupling across conditions (levels of satiety, cognition) at the level of individual participants. (C) The top plot (i) represents the normalized alpha band PSD in uV2/Hz. The bottom plot (ii) represents normalized and log-transformed PAC showing midline areas with elevated PAC. All signals survived FDR statistical corrections for multiple comparisons across electrodes. PAC: phase amplitude coupling.

• 

  Figure 4. PAC during cognitive and gut states. PAC values against permutation control distributions were FDR corrected, and are shown for eyes closed resting state and working memory cognitive state during hunger, satiety, and satiety-hunger contrast. Last, we present the spectral power density averaged for hunger and satiety states which showed increased density in the parieto-occipital area. There was not any significant difference between hunger and satiety in alpha PSD, in contrast to the significant PAC difference observed in frontal and central areas.

• 

  Figure 5. Phase amplitude gut-brain coupling in four brain regions differs by cognitive state and gut satiety state. Significant cognitive state (resting versus working memory) based differences in PAC (marked with ^) are observed in the left fronto-central (C3, F3), right fronto-central (C4, F4), and left parieto-occipital (P3, O1) regions. Modulations in the right parieto-occipital (P4, O2) region were not significant. Specifically, the left fronto-central region also showed significant (marked with #) hunger vs. satiety gut state effects, and significant cognitive state differences that vary with the levels of hunger (refer to Table 1 for statistical results).

• 

  Figure 6. Spearman correlations between PAC and behavior. A significant relationship was found between PAC during the working memory task and performance speed, with increased coupling corresponding to enhanced performance speed (r = 0.6, p = 0.002).