Resting-state connectivity within the brain’s reward system predicts weight loss and correlates with leptin

Abstract Weight gain is often associated with the pleasure of eating food rich in calories. This idea is based on the findings that people with obesity showed increased neural activity in the reward and motivation systems of the brain in response to food cues. Such correlations, however, overlook the possibility that obesity may be associated with a metabolic state that impacts the functioning of reward and motivation systems, which in turn could be linked to reactivity to food and eating behaviour and weight gain. In a study involving 44 female participants [14 patients with obesity, aged 20–63 years (mean: 42, SEM: 3.2 years), and 30 matched lean controls, aged 22–60 years (mean: 37, SEM: 1.8 years)], we investigated how ventromedial prefrontal cortex seed-to-voxel resting-state connectivity distinguished between lean and obese participants at baseline. We used the results of this first step of our analyses to examine whether changes in ventromedial prefrontal cortex resting-state connectivity over 8 months could formally predict weight gain or loss. It is important to note that participants with obesity underwent bariatric surgery at the beginning of our investigation period. We found that ventromedial prefrontal cortex–ventral striatum resting-state connectivity and ventromedial–dorsolateral prefrontal cortex resting-state connectivity were sensitive to obesity at baseline. However, only the ventromedial prefrontal cortex–ventral striatum resting-state connectivity predicted weight changes over time using cross-validation, out-of-sample prediction analysis. Such an out-of-sample prediction analysis uses the data of all participants of a training set to predict the actually observed data in one independent participant in the hold-out validation sample and is then repeated for all participants. In seeking to explain the reason why ventromedial pre-frontal cortex–ventral striatum resting-state connectivity as the central hub of the brain’s reward and motivational system may predict weight change over time, we linked weight loss surgery-induced changes in ventromedial prefrontal cortex–ventral striatum resting-state connectivity to surgery-induced changes in homeostatic hormone regulation. More specifically, we focussed on changes in fasting state systemic leptin, a homeostatic hormone signalling satiety, and inhibiting reward-related dopamine signalling. We found that the surgery-induced increase in ventromedial prefrontal cortex–ventral striatum resting-state connectivity was correlated with a decrease in fasting-state systemic leptin. These findings establish the first link between individual differences in brain connectivity in reward circuits in a more tonic state at rest, weight change over time and homeostatic hormone regulation.


Additional clinical assessments of morbidly obese patients before surgery
Before bariatric surgery, obese participants were assessed on the following: depression, using the Beck Depression Inventory (BDI); alcohol abuse, using the Alcohol Use Disorders Identification Test (AUDIT); nicotine consumption, using the Fagerstrom test; dietary restraint, disinhibition, and hunger, using the Three-Factor Eating Questionnaire (TFEQ); and diabetes (clinical assessment). Glycemia was assessed by measuring blood glucose levels before (fasting conditions) and after a standardized meal test. The test results (Supplementary  We investigated differences in RSC in the brain's valuation system with the vmPFC as seed between the obese and lean participants. In other words, we looked at the main effect of participant group irrespective of time and found that participants with obesity presented stronger vmPFC RSC to a set of frontal brain regions including the dorsolateral prefrontal cortex (dlPFC), the ventrolateral prefrontal cortex (vlPFC), and the medial prefrontal cortex (mPFC) (cluster-corrected pFDR < 0.05). We also tested for a main effect of time, but found no differences between T0 and T8 across the whole participant sample, even at a more lenient uncorrected threshold of p<0.001.
We investigated whether RYGB surgery affected the RSC of the vmPFC, and, if so, whether it would affect its RSC to other brain regions involved in reward and motivation processing and control. In more detail, we compared the difference in the RSC of the vmPFC in the participants with obesity after versus before RYGB surgery to the change over time in the RSC of the vmPFC in the lean participants (i.e., the obese group > lean group by time T8 > T0 interaction). We found stronger RSC between the vmPFC and the vStr RSC for this interaction (MNI coordinates [-10 6 -2], punc < 0.001, extent threshold k = 50 voxels; Supplemental Figure 2).

Additional statistical analysis and results: Residual leptin and vmPFC-vStr connectivity
As a robustness check we also calculated residual leptin values by regressing out any variance of leptin explained by kg body fat. We then correlated the difference of before minus after surgery in residual leptin values to the difference in before minus after surgery in vmPFC-vStr RSC, respectively. It revealed a significant covariance (r = 0.41, p = 0.08, 95% CI due to chance: -0.45-0.46; for % body fat: r = 0.52, p = 0.05, 95% CI due to chance: -0.45-0.46).

Additional statistical analysis and results: Insulin sensitivity and vmPFC-vStr connectivity
Other metabolic measures such as insulin also have also been shown to be indirectly linked to where GPF corresponds to the fasting-state glycemia, and IPF to fasting-state insulin.
HOMA-IR scores were smaller after bariatric surgery than before, indicating a decrease in insulin resistance after surgery, which was significant (t (13)=3.3, p=0.005). The magnitude of this decrease (HOMA-IRT8 -HOMA-IRT0) was positively correlated with the change in vmPFC-vStr RSC after surgery (r = 0.30, p = 0.2), indicating the same trend as leptin, albeit not significant.

Supplementary Figure 1
Comparisons of vmPFC to brain resting-state connectivity in lean participants compared to those with obesity after bariatric surgery (T8). Statistical parametric maps (SPMs) of the seed-to-voxel restingthe vmPFC seed ROI and the rest of the brain at 8 months post-surgery (T8) (n = 44) Significant voxels are displayed for visualization purposes in orange at p < 0.001 uncorrected, k corresponding to a false discovery rate (FDR) corrected threshold of pFDR < 0.05 on the average structural image obtained from the lean participants. The [x, y, z] coordinates correspond to MNI coordinates and are taken at maxima of interest. The line graphs on the right side depict average correlation coefficients between resting state activity of the seed region, the vmPFC, and the right vlPFC in lean (dark grey) and obese (light grey) participants.

Supplementary Figure 2
Activity in the vmPFC seed correlated significantly more to resting-state activity in the striatum in obese participants after surgery compared to before surgery and to lean participants for the time between baseline (T0) and eight months later (T8) assessments (N = 44, p < 0.001 uncorrected). on the left panel display all voxels activated on the axial slice taken at the global maximum indicated arrow.
Statistical parametric maps are superimposed on the average structural image obtained from the lean participants. AUDIT (alcohol use disorders): in women, a score of ≥7 indicates alcohol abuse, and ≥11 indicates alcohol dependence. Fagerstrom score (nicotine dependence): 0 to 2 = none, 3 to 4 = weak, 5 to 6 = moderate, 7 to 10 = strong. TFEQ score (severity of dietary restraint, disinhibition, and hunger): l = low, 2 = moderate, 3 = high. Glycemia reflects the number of participants with glycemia and hyperglycemia.