Decreased modular segregation of the frontal-parietal network in major depressive disorder

Abstract

Major depressive disorder (MDD) is a common psychiatric condition associated with aberrant large-scale distributed brain networks. However, it is unclear how the network dysfunction in MDD patients is characterized by imbalance or derangement of network modular segregation. Fifty-one MDD patients and forty-three matched healthy controls (HC) were recruited in the present study. We analyzed intrinsic brain activity derived from resting-state functional magnetic resonance imaging (R-fMRI) and then examined brain network segregation by computing the participation coefficient (PC). Further intra- and inter-modular connections analysis were preformed to explain atypical PC. Besides, we explored the potential relationship between the above graph theory measures and symptom severity in MDD. Lower modular segregation of the frontal-parietal network (FPN) was found in MDD compared with the HC group. The MDD group exhibited increased inter-module connections between the FPN and cingulo-opercular network (CON), between the FPN and cerebellum (Cere), between the CON and Cere. At the nodal level, the PC of the anterior prefrontal cortex, anterior cingulate cortex, inferior parietal lobule (IPL), and intraparietal sulcus showed larger in MDD. Additionally, the inter-module connections between the FPN and CON and the PC values of the IPL were negatively correlated with depression symptom in the MDD group. These findings might give evidence about abnormal FPN in MDD from the perspective of modular segregation in brain networks.

Keywords: fMRI; frontal–parietal network; major depressive disorder; modular segregation; participation coefficient.

FIGURE 1

The schematic illustration of processing steps. This figure showed the data analysis pipeline of the present study. (A) Resting-state fMRI (R-fMRI) images were acquired from all subjects. (B) The Dosenbach 160 regions were used as network nodes of the whole brain and the 160 nodes were parcellated into six functional modules. (C) We conducted Pearson’s correlation analysis between each pair of nodes to construct a 160 × 160 correlation matrix for each participant. (D) The mean participation coefficient (PC) for each module was calculated to quantify the modular segregation. (E) The number of intra- and inter-module connections were calculated to explore the reasons for PC abnormalities among the two groups.

FIGURE 2

Between-group differences of mean participant coefficient (PC). Major depressive disorder (MDD) patients showed significantly higher PC on the fronto-parietal network (FPN) and cerebellum (Cere) than the healthy controls (HC). *Indicates p < 0.05.

FIGURE 3

Between-group comparison of inter-module connections and their correlation with HAMD scores. (A) Compared with the healthy controls (HC), the major depressive disorder (MDD) patients exhibited significantly increased inter-module connections between the frontal–parietal network (FPN) and cingulo-opercular network (CON), FPN and cerebellum (Cere), CON and Cere. (B) The inter-module connections between the FPN and CON are significantly negatively correlated with HAMD scores in the MDD group. *Indicates p < 0.05, **indicates p < 0.01.

FIGURE 4

Between-group differences of participation coefficient (PC) of the nodes in frontal–parietal network (FPN) and their correlation with HAMD scores. (A) The major depressive disorder (MDD) patients exhibited significantly increased PC in the bilateral ventral anterior prefrontal cortex, left anterior cingulate cortex, right inferior parietal lobule, and left intraparietal sulcus relative to the healthy controls (HC). (B) The PC values of the right inferior parietal lobule were significantly negatively correlated with HAMD scores in the MDD group. *Indicates p < 0.05, **indicates p < 0.01.