Escitalopram Decreases Cross-Regional Functional Connectivity within the Default-Mode Network

Abstract

The default-mode network (DMN), which comprises medial frontal, temporal and parietal regions, is part of the brain's intrinsic organization. The serotonergic (5-HT) neurotransmitter system projects to DMN regions from midbrain efferents, and manipulation of this system could thus reveal insights into the neurobiological mechanisms of DMN functioning. Here, we investigate intrinsic functional connectivity of the DMN as a function of activity of the serotonergic system, through the administration of the selective serotonin reuptake inhibitor (SSRI) escitalopram. We quantified DMN functional connectivity using an approach based on dual-regression. Specifically, we decomposed group data of a subset of the functional time series using spatial independent component analysis, and projected the group spatial modes to the same and an independent resting state time series of individual participants. We found no effects of escitalopram on global functional connectivity of the DMN at the map-level; that is, escitalopram did not alter the global functional architecture of the DMN. However, we found that escitalopram decreased DMN regional pairwise connectivity, which included anterior and posterior cingulate cortex, hippocampal complex and lateral parietal regions. Further, regional DMN connectivity covaried with alertness ratings across participants. Our findings show that escitalopram altered intrinsic regional DMN connectivity, which suggests that the serotonergic system plays an important role in DMN connectivity and its contribution to cognition. Pharmacological challenge designs may be a useful addition to resting-state functional MRI to investigate intrinsic brain functional organization.

Figure 1. Default-mode networks after escitalopram (upper row) and placebo administration (bottom row).

The figure shows three transverse slices (Talairach-Z coordinates 22, 2 and −14) of ICA-derived functional connectivity maps of the DMN superimposed on the anatomical average of the participants. The left hemisphere is depicted on the left side of each image.

Figure 2. ROI-based RMANOVA results.

Each panel depicts the color-coded RMANOVA F-values of the pairwise ROI functional connections for factors Drug (A), Time (B) and Drug x Time (C). Repetitive values in the ROI connections matrix are replaced with black. Degrees of freedom for plotted F-values are 1 and 9 in all cases. Asterisks mark ROI connections with significant results at q(FDR) = 0.05.

Figure 3. Escitalopram-induced changes in DMN functional connectivity.

A: Bars show means of functional connectivity values (Fisher’s Z normalized correlation coefficients) as a function of drug condition, and for each resting-state segment, for each of the six connections that were found significant after RMANOVA testing. Grey bars represent connectivity after Escitalopram administration, black bars after Placebo. B-D: Schematic 3D display of the DMN ROIs and their pairwise functional connections. ROIs are displayed as blue orbs (radius = 10 mm), significant connections are displayed as red beams connecting the orbs, non-significant connections are shown as grey beams. B: Lateral view (sagittal; left side is anterior). C: Top view (transverse; left side is left). D: Tilted side view.

Figure 4. Scatterplots.

Shown are the scatterplots of functional connectivity between PCC and RIPC (A), and between LIPC and RPHC (B) as a function of alertness ratings of the Bond and Lader VAS scale. Each data point represents one participant. A regression line (least squares fit) is drawn through the data points in each plot. Functional connectivity and VAS rating values are the average of the Escitalopram and Placebo conditions.