Analysis of Plasma PYY_3-36 and GLP-1_{7-36 amide} Levels and Mean Ad Libitum Energy Intake
(A–D) On each study visit, the hormone infusion was administered between t = 0 and t = 90 min. Shown are plasma PYY_3-36 levels (A) and plasma GLP-1_{7-36 amide} levels (B). Data are shown as mean ± SEM for 15 subjects. An ad libitum buffet meal was served immediately after the infusion in order to measure energy intake on all study days. Shown are the energy intake during the buffet meal following each infusion (C) and the reduction in energy intake during the buffet meal for each infusion versus the fasted saline infusion and also the sum of the individual effects of PYY_3-36 and GLP-1_{7-36 amide} in reducing energy intake (sum of PYY and GLP-1) (D). Data are shown as mean ± SEM, grouped for 15 subjects. ^∗∗p < 0.01 versus fasted saline. ^∗∗∗p = 0.0001 versus fasted saline. Abbreviations: EI, energy intake; kJ, kilojoules; PYY, PYY_3-36; GLP-1, GLP-1_{7-36 amide}. See also Figure S1.

Modulation of BOLD Signal across Brain ROIs by Feeding or Either Individual or Combined Gut Hormone Infusions
(A–D) Subjects underwent a 90 min infusion of saline (fasted saline) as a control. They also had four further infusions: saline after a standard breakfast (A), PYY_3-36 after an overnight fast (B), GLP-1_{7-36 amide} after an overnight fast (C), and combined PYY_3-36 + GLP-1_{7-36 amide} after an overnight fast (D). During each infusion, a BOLD fMRI scan was performed, incorporating a picture processing task where images of food and nonfood were shown. The mean percent BOLD signal change when subjects viewed images of food compared with when they viewed images of nonfood is shown for each of the infusions administered as a comparison with the fasted saline infusion: ^∗p = 0.015 for fed saline < fasted saline and p = 0.012 for PYY + GLP-1 < fasted saline. ^∗∗p = 0.005 for PYY < fasted saline. Data are shown for individual ROIs (amygdala, insula, caudate, nucleus accumbens [N Acc], OFC, and putamen), combined for left and right hemispheres and grouped for 15 subjects, shown as mean ± SEM. Abbreviations: PYY, PYY_3-36; GLP-1, GLP-1_{7-36 amide}. See also Figure S2.

Study Protocol
Following an overnight fast, 16 healthy, normal-weight subjects each received the following interventions, in random order, over 5 separate study days in a single-blinded fashion: (1) 90 min saline infusion (fasted saline, control visit); (2) standard breakfast, then 90 min saline infusion (fed saline); (3) 90 min PYY_3-36 infusion at 0.3 pmol/kg/min; (4) 90 min GLP-1_{7-36 amide} infusion at 0.8 pmol/kg/min; (5) 90 min combined PYY_3-36 and GLP-1_{7-36 amide} infusion, at 0.3 pmol/kg/min and 0.8 pmol/kg/min, respectively. On each visit, subjects underwent a 60 min fMRI scan, which commenced 20 min after the start of the infusion. During the fMRI scan, a picture processing task was performed where images of food and nonfood were shown. The mean percent change in BOLD signal in prespecified brain ROIs when viewing images of food compared to nonfood were determined for each study day. An ad libitum buffet meal was served immediately after the infusion on all study days in order to measure energy intake. Blood sampling (for PYY_3-36 and GLP-1_{7-36 amide}) and assessments of appetite (based on VAS after direct questioning) was performed. One subject was excluded from analysis due to excessive head movement during the scans, leaving data from 15 subjects for analysis.

Modulation of BOLD Signal across Brain ROIs
(A and B) During each infusion, a BOLD fMRI scan was performed, incorporating a picture processing task where images of food and nonfood were shown. For the fasted saline and fed saline infusions, a whole-brain map of brain regions activated (z-statistics) by food images (compared with nonfood images) is shown (A). Z indicates distance (mm) superior or inferior to the intercommissural plane in standard stereotactic space. Clusters were thresholded by means of parametric testing at the level of spatially contiguous suprathresholded clusters, simultaneously controlling the family-wise probability of type 1 error at p < 0.05, corrected. The mean percent BOLD signal change (food images minus nonfood images) across all six ROIs studied (amygdala, insula, caudate, nucleus accumbens, OFC, and putamen), is shown for each of the interventions (B). Data are shown as mean ± SEM, grouped for 15 subjects. Abbreviations: PYY, PYY_3-36; GLP-1, GLP-1_{7-36 amide}; ^∗∗p < 0.01; ^∗∗∗p < 0.001.