Our main goal was to identify associations (linear and quadratic) of BMI and characteristics of eating behavior (CR, DIS) with BOLD activation during volitional regulation of food craving. We tested separate regression models to individually assess the relationship of BMI, CR, DIS or regulation success and the respective regulation contrasts (REGULATE_TASTY>ADMIT_TASTY, REGULATE_TASTY>REGULATE_NOT_TASTY) including age (analyses of BMI, CR, DIS, regulation success) or age and BMI (analysis of BMI 2 ) as covariates. To assess the relationship of craving intensity and appetitive brain activity, separate regression models were tested on the respective craving contrasts (ADMIT_TASTY>REGULATE_TASTY, ADMIT_TASTY>ADMIT_NOT_TASTY). Please see Supplementary Table III for a summary of performed regression analyses. Second-level maps were thresholded voxelwise at P<0.001 and corrected for multiple comparisons at a cluster threshold of P<0.05 (family-wise error) for the whole brain.
Useful connectivity analysis
Functional connectivity was assessed by means of psychophysiological interaction (PPI) analysis. 28 Source regions were based on the above-mentioned regression analysis of BOLD activation and BMI, our primary research focus. Individual BOLD signal time series within 4-mm spheres surrounding detected peak coordinates were extracted (based on the inverted U-shaped relationship of BMI and REGULATE_TASTY>ADMIT_TASTY, please see ‘Performance’ section and Table 2 for details). General linear models were estimated separately for every source region including the following regressors: Time course of the respective source region (physiological vector), a vector coding for the main effect (psychological vector; REGULATE_TASTY>ADMIT_TASTY; with the former term weighted as +1 and the latter one weighted as ?1), and the PPI term (element-by-element product between the time course of the source region and the vector coding the main effect). The models also included realignment parameters as nuisance regressors. Single-subject contrasts for the PPI regressors were calculated. In the second-level analysis, we aimed to identify regions whose functional connectivity was related to BMI (linear and quadratic) or characteristics of eating behavior (CR, DIS). Therefore, the PPI terms were regressed on these measures in separate multiple regression analyses. Second-level models also included the regressors of no interest mentioned under subsection ‘Analysis of BOLD response’. Second-level maps were thresholded voxelwise at P<0.001 and corrected for multiple comparisons at a cluster threshold of P<0.05 (family-wise error) for the whole brain. Clusters were considered to be significant at P<0.017 (Bonferroni adjustment to account for the number of investigated seeds). Please see Supplementary Table III for a summary of performed regression analyses.
We seen a powerful confident correlation away from Body mass index and you can DIS (Roentgen dos =0.285, P>0.001, Pearson correlation, Additional Figure Ia). Multiple regression analysis shown a poor association out of Bmi dos with CR (R dos =0.151, P=0.038, covariate Body mass index; Secondary Figure Ib), proving how to find a sugar daddy in Columbus Oh OH an upside down U-shaped matchmaking. Urge intensity don’t correlate with Bmi (R=?0.206, P=0.185, Pearson relationship). We discover a pattern of a bad relationship between controls achievements and you can Bmi (R=?0.295, P=0.055, Pearson correlation). See Desk step 1 getting descriptive analytics.
To manage its desire, the players (especially overweight volunteers) thought new bad enough time-term effects out-of dining the fresh new illustrated palatable food. Very professionals switched anywhere between additional regulation steps for the duration of this new try (pick Additional Desk IV to own all about means use). Whenever coached so you’re able to admit, all players imagined taste or structure of your own presented foodstuff.
Dating ranging from Challenging hobby and Body mass index, eating choices, need strength or subjective control achievements
Activity in a cluster comprising left putamen, amygdala and insula was nonlinearly (inverted U-shaped) related to BMI during volitional regulation devoid of craving influences (REGULATE_TASTY>ADMIT_TASTY; Table 2, Figure 2). Activation during regulation specific to hedonic food (REGULATE_TASTY>REGULATE_NOT_TASTY) was unrelated to BMI. We found no linear relationships with BMI. Craving intensity correlated positively with activity in the right hippocampus/amygdala during craving devoid of volitional regulatory influences (ADMIT_TASTY>REGULATE_TASTY; Table 2, Supplementary Figure X), but did not correlate with activation during craving specific to hedonic food (ADMIT_TASTY>ADMIT_NOT_TASTY). Neither subjective regulation success nor measures of eating behavior were significantly related to task-related BOLD activity. The above-mentioned results indicate some lateralization of the findings. However, when a less strict threshold was applied, bilateral BOLD activation of all mentioned regions associated with BMI and craving intensity was observed (relationship of BOLD and BMI: t-values thresholded at P<0.05, uncorrected; relationship of BOLD and craving intensity: t-values thresholded at P<0.001, uncorrected).