Hyperconnectivity in an introspective socio-affective network model in major depression

Overview: Alterations of affective and introspective processing are thought to play an important role in the etiology of major depression (MDD). The associated neurobiology, however, remains elusive. To test for pathological changes of such processes in MDD, we applied a model-based analysis approach to investigate functional connectivity (FC) within an introspective socio-affective (ISA) network. In line with recent findings suggestive of a 'hyperconnectivity hypothesis' of depression [1], we hypothesized to find a pathologically increased neurofunctional coupling within the ISA network in depressed patients.

Methods: The ISA network model was based on two previous meta-analyses [2], which identified a consistent network for affective and introspective processing, consisting of left amygdala (AmyL), dorsomedial prefrontal cortex (dmPFC), precuneus (PrC) as well as anterior (ACC) and subgenual cingulate cortex (SGC). A control analysis was performed for a network associated with decoding degraded speech [3]. FC was investigated using resting state fMRI of 57 patients with MDD and 57 age-matched healthy controls from two sites. After spatial preprocessing, temporal filtering was conducted including the removal of various confounds [4]. To quantify FC, correlations between time courses of all network nodes were computed for each subject and Fisher's Z transformed for group analysis. Influence of the factor 'site' and its interaction with the factor 'diagnosis' was removed before further analysis. Group differences in FC were assessed using non-parametric permutation statistics (P > 0.95). Rank-correlations were computed to test for relationships between FC and subjective symptom severity (BDI).

Results: We found significant differences between groups with MDD patients featuring significant hyperconnectivity between every node of the socio-affective network (all P > 0.984), except the dmPFC (Fig. 1). No significant group difference were found for any connection between regions of the control (language) network (all P

Conclucion: Using a model-based approach, we demonstrate neurofunctional alterations in a meta-analytically derived network for introspective socio-affective processing in depressed individuals. The negative result for the control network confirms the functional specificity of these FC changes associated with MDD. The hyperconnectivity within the ISA network is consistent with the idea that alterations of self-referential cognition and dysfunctional interpersonal expectations are an important feature of depressive symptomatology. Finally, the relationship between subjective symptom severity and FC of medial prefrontal brain areas associated with mood regulation [5] suggests that these neurofunctional alterations might be also present in other affective disorders.

Fig. 1

References:

[1] Perrinet al. (2012), Proc Natl Acad Sci U S A 109(14):5464 – 68.

[2] Schilbach et al. (2012), PLoS One 7(2):e30920.

[3] Clos et al. (2012), Hum Brain Mapp.

[4] Sommer et al. (2012), PLoS One 7(9):e43516.

[5] Davey et al. (2012), Psychol Med 42(10):2071 – 81.