Analysis of Default Mode Network in Social Anxiety Disorder: EEG Resting-State Effective Connectivity Study

Several neuroimaging findings by using different modalities (e.g., fMRI and PET) have suggested that social anxiety disorder (SAD) is correlated with alterations in regional or network-level brain function. However, these modalities do not quantify the fast dynamic connectivity of causal information networks due to their poor temporal resolution. In this study, SAD-related changes in brain connections within the default mode network (DMN) was investigated using Electroencephalogram (EEG). Partial directed coherence (PDC) was used to assess the causal influences of DMN regions on each other and indicate the changes in the DMN effective network related to SAD severity. The EEG data were collected from 88 subjects (control, mild, moderate, severe) and used to estimate the effective connectivity between DMN regions at different frequency bands. Among the healthy control (HC) and the three considered levels of severity of SAD, the results indicated a higher level of causal interactions for the mild and moderate SAD groups than for the severe and HC groups. Between the control and the severe SAD groups, the results indicated a higher level of causal connections for the control throughout all the DMN regions. We found significant increases in the mean PDC in the delta and alpha bands between the SAD groups. Among the DMN regions, the precuneus exhibited a higher level of causal influence than other regions. Therefore, it was suggested to be a major source hub that contributes to the mental exploration and emotional content of SAD. In contrast to the severe group, the HC exhibited higher resting-state connectivity at the mesial prefrontal cortex (mPFC), providing evidence for mPFC dysfunction in the severe SAD group. Furthermore, the total Social Interaction Anxiety Scale (SIAS) was positively correlated with the mean values of the PDC of the severe SAD group and negatively correlated with those of the HC group. The reported results may facilitate greater comprehension of the underlying potential SAD neural biomarkers and can be used to characterize possible targets for further medication.