Sex differences in the influence of body mass index on anatomical architecture of brain networks

Abstract

Background/objectives: The brain has a central role in regulating ingestive behavior in obesity. Analogous to addiction behaviors, an imbalance in the processing of rewarding and salient stimuli results in maladaptive eating behaviors that override homeostatic needs. We performed network analysis based on graph theory to examine the association between body mass index (BMI) and network measures of integrity, information flow and global communication (centrality) in reward, salience and sensorimotor regions and to identify sex-related differences in these parameters.

Subjects/methods: Structural and diffusion tensor imaging were obtained in a sample of 124 individuals (61 males and 63 females). Graph theory was applied to calculate anatomical network properties (centrality) for regions of the reward, salience and sensorimotor networks. General linear models with linear contrasts were performed to test for BMI and sex-related differences in measures of centrality, while controlling for age.

Results: In both males and females, individuals with high BMI (obese and overweight) had greater anatomical centrality (greater connectivity) of reward (putamen) and salience (anterior insula) network regions. Sex differences were observed both in individuals with normal and elevated BMI. In individuals with high BMI, females compared to males showed greater centrality in reward (amygdala, hippocampus and nucleus accumbens) and salience (anterior mid-cingulate cortex) regions, while males compared to females had greater centrality in reward (putamen) and sensorimotor (posterior insula) regions.

Conclusions: In individuals with increased BMI, reward, salience and sensorimotor network regions are susceptible to topological restructuring in a sex-related manner. These findings highlight the influence of these regions on integrative processing of food-related stimuli and increased ingestive behavior in obesity, or in the influence of hedonic ingestion on brain topological restructuring. The observed sex differences emphasize the importance of considering sex differences in obesity pathophysiology.

Figure 1. Regions of Interest

Regions of interest used in the analyses are displayed separated by network Reward Network Regions of Interest: caudate nucleus, putamen, globus pallidus, nucleus accumbens (NAcc), amygdala, hippocampus, medial orbital frontal gyrus (mOFG) and includes parcellations from the medial orbital gyrus (mOG) and medial orbital sulcus (mOS) Salience Network Regions of Interest: anterior insula (aINS) and includes parcellations from the horizontal ramus of the anterior segment of the lateral sulcus (ALSHorp), anterior segment of the circular sulcus of the insula (ACirINS), vertical ramus of the anterior segment of the lateral sulcus (ALSVerp) and the short insular gyri (ShoInG); anterior mid cingulate cortex (aMCC) Sensorimotor Network Regions of Interest: thalamus, primary somatosensory cortex/S1 [which includes the postcentral gyrus (PosCG), postcentral sulcus (PosCS), and central sulcus (CS)], secondary somatosensory cortex/S2 [which includes the subcentral gyrus and sulcus (SbCGS)], primary motor cortex/M1 [which includes the precentral gyrus (PreCG), inferior part of the precentral sulcus (InfPreCS), and superior part of the precentral sulcus (SupPreCS)], mid insula (mINS) (superior segment of circular sulcus of the insula [SupCirINS]), posterior insula (pINS) [which includes the long insular gyrus and sulcus (LongINSGS), inferior segment of the circular sulcus of the insula (InfCirINS), and the posterior ramus of the lateral sulcus (PosLS)]

Figure 2

A: BMI-related differences in anatomical network metric measures of centrality Abbreviations: L, left; R, right; aINS, anterior insula (short insular gyri [ShoInG]) Node Strength is represented by both the number of edges emanating from a given node as well as the thickness of the edge (sized by weight). Betweenness Centrality and Eigenvector Centrality are depicted by colored nodes with no edges Significant differences in measures of centrality depicted as Nodes were colored according to the network they belong to (shown below). Black edges that were common to both groups were colored black and edges that were specific to one group were colored according to the network they belonged to (see below). B: Sex-related differences in anatomical measures of centrality in subjects with high BMI Abbreviations: L, left; R, right; AMYG, amygdala; HIPP, hippocampus; NACC, nucleus accumbens; aMCC; anterior mid cingulate cortex; pINS, posterior insula (long insular gyrus and sulcus [LongINSGS]) Node Strength is represented by both the number of edges emanating from a given node as well as the thickness of the edge (sized by weight). Betweenness Centrality and Eigenvector Centrality are depicted by colored nodes with no edges Significant differences in measures of centrality depicted as Nodes were colored according to the network they belong to (shown below). Black edges that were common to both groups were colored black and edges that were specific to one group were colored according to the network they belonged to (see below). C: BMI and sex-related differences in anatomical measures of centrality in female subjects with high BMI compared to female subjects with normal BMI Abbreviations: L, left; R, right; AMYG, amygdala; PMC/M1 primary motor cortex (specifically the inferior part of the precentral sulcus [InfPreCS]) Node Strength is represented by both the number of edges emanating from a given node as well as the thickness of the edge (sized by weight). Betweenness Centrality and Eigenvector Centrality are depicted by colored nodes with no edges Significant differences in measures of centrality depicted as Nodes were colored according to the network they belong to (shown below). Black edges that were common to both groups were colored black and edges that were specific to one group were colored according to the network they belonged to (see below). D: BMI and sex-related differences in anatomical measures of centrality in male subjects with high BMI compared to male subjects with normal BMI Abbreviations: L, left; R, right; HIPP, hippocampus; mOFG, medial orbital frontal gyrus; SSC/S2, secondary somatosensory cortex [which includes the subcentral gyrus and sulcus [SbCGS]); mINS, mid insula (superior segment of circular sulcus of the insula [SupCirINS]); pINS, posterior insula (which includes the long insular gyrus and sulcus [LongINSGS], inferior segment of the circular sulcus of the insula [InfCirINS], and the posterior ramus of the lateral sulcus [PosLS]) Node Strength is represented by both the number of edges emanating from a given node as well as the thickness of the edge (sized by weight). Betweenness Centrality and Eigenvector Centrality are depicted by colored nodes with no edges Significant differences in measures of centrality depicted as Nodes were colored according to the network they belong to (shown below). Black edges that were common to both groups were colored black and edges that were specific to one group were colored according to the network they belonged to (see below). Key

1. Reward

2. Salience

3. Sensorimotor