Intrinsic connectomes are a predictive biomarker of remission in major depressive disorder

Abstract

Although major depressive disorder (MDD) is associated with altered functional coupling between disparate neural networks, the degree to which such measures are ameliorated by antidepressant treatment is unclear. It is also unclear whether functional connectivity can be used as a predictive biomarker of treatment response. Here, we used whole-brain functional connectivity analysis to identify neural signatures of remission following antidepressant treatment, and to identify connectomic predictors of treatment response. 163 MDD and 62 healthy individuals underwent functional MRI during pre-treatment baseline and 8-week follow-up sessions. Patients were randomized to escitalopram, sertraline or venlafaxine-XR antidepressants and assessed at follow-up for remission. Baseline measures of intrinsic functional connectivity between each pair of 333 regions were analyzed to identify pre-treatment connectomic features that distinguish remitters from non-remitters. We then interrogated these connectomic differences to determine if they changed post-treatment, distinguished patients from controls, and were modulated by medication type. Irrespective of medication type, remitters were distinguished from non-remitters by greater connectivity within the default mode network (DMN); specifically, between the DMN, fronto-parietal and somatomotor networks, the DMN and visual, limbic, auditory and ventral attention networks, and between the fronto-parietal and somatomotor networks with cingulo-opercular and dorsal attention networks. This baseline hypo-connectivity for non-remitters also distinguished them from controls and increased following treatment. In contrast, connectivity for remitters was higher than controls at baseline and also following remission, suggesting a trait-like connectomic characteristic. Increased functional connectivity within and between large-scale intrinsic brain networks may characterize acute recovery with antidepressants in depression.

Trial registration: ClinicalTrials.gov NCT00693849.

Conflict of interest statement

L.M.W. has received consultant fees from BlackThorn Therapeutics. She has also received scientific advisory board fees from Psyberguide of the One Mind Institute. M.S.K., A.G.P., A.F. have no financial disclosures related to the work.

Figures

Fig. 1

Pre-treatment connectome networks differentiating remitters from non-remitters. (top row) The connectomic feature identified from the NBS analysis. Node colors indicate intrinsic resting state brain network membership as defined by the Gordon et al. [33] parcellation. (bottom row) The patterns of different intra- (loops) and inter-network connections comprising this feature (listed in Table 2) are shown. The thickness of the lines correspond to the number of significant connections between networks of interests relative to the total number of possible connections i.e. thicker lines implies more number of significant connections between the networks. Bar plots (means and SD) showing average connectivity estimates at baseline and post-treatment for each group. There was a significant time*group interaction for average connectivity in this network feature. Asterisks indicate significant post hoc findings (p < 0.05) for this interaction. DMN default mode network, FPN fronto-parietal network, SM somatomotor, VAN ventral attention network, DAN dorsal attention network, CON cingulo-opercular network, L left, R right, MDD-R Remitters, MDD-NR Non-remitters, Ctrl Healthy individuals

Fig. 2

Intra- and inter-network connections that differentiate Remitters from Non-Remitters. Bar plots show connectivity estimates at baseline and post-treatment (means and SD). There was a significant time*group interaction (FDRp p < 0.05) for this interaction. DMN default mode network, FPN fronto-parietal network, SM somatomotor, VAN ventral attention network, DAN dorsal attention network, CON cingulo-opercular network, L left, R right, MDD-R Remitters, MDD-NR Non-Remitters, Ctrl Healthy individuals