Resting-state connectivity predictors of response to psychotherapy in major depressive disorder

Abstract

Despite the heterogeneous symptom presentation and complex etiology of major depressive disorder (MDD), functional neuroimaging studies have shown with remarkable consistency that dysfunction in mesocorticolimbic brain systems are central to the disorder. Relatively less research has focused on the identification of biological markers of response to antidepressant treatment that would serve to improve the personalized delivery of empirically supported antidepressant interventions. In the present study, we investigated whether resting-state functional brain connectivity (rs-fcMRI) predicted response to Behavioral Activation Treatment for Depression, an empirically validated psychotherapy modality designed to increase engagement with rewarding stimuli and reduce avoidance behaviors. Twenty-three unmedicated outpatients with MDD and 20 matched nondepressed controls completed rs-fcMRI scans after which the MDD group received an average of 12 sessions of psychotherapy. The mean change in Beck Depression Inventory-II scores after psychotherapy was 12.04 points, a clinically meaningful response. Resting-state neuroimaging data were analyzed with a seed-based approach to investigate functional connectivity with four canonical resting-state networks: the default mode network, the dorsal attention network, the executive control network, and the salience network. At baseline, the MDD group was characterized by relative hyperconnectivity of multiple regions with precuneus, anterior insula, dorsal anterior cingulate cortex (dACC), and left dorsolateral prefrontal cortex seeds and by relative hypoconnectivity with intraparietal sulcus, anterior insula, and dACC seeds. Additionally, connectivity of the precuneus with the left middle temporal gyrus and connectivity of the dACC with the parahippocampal gyrus predicted the magnitude of pretreatment MDD symptoms. Hierarchical linear modeling revealed that response to psychotherapy in the MDD group was predicted by pretreatment connectivity of the right insula with the right middle temporal gyrus and the left intraparietal sulcus with the orbital frontal cortex. These results add to the nascent body of literature investigating pretreatment rs-fcMRI predictors of antidepressant treatment response and is the first study to examine rs-fcMRI predictors of response to psychotherapy.

Figures

Figure 1

Group differences in resting-state connectivity in (a) the dorsal attention network (DAN); (b) the default mode network (DMN); (c) the executive control network (ECN); and (d) the salience network (SN). All results are cluster-corrected, p<0.05.

Figure 2

Overlap in the left middle temporal gyrus clusters that demonstrated differential connectivity with the right insula (in green), the left insula (in red), and the precuneus (in blue).

Figure 3

Left: Relations between pretreatment precuneus—left middle temporal gyrus connectivity and pretreatment BDI scores. At pretreatment, the MDD group was characterized by greater precuneus—left middle temporal gyrus connectivity relative to controls. Right: Relations between pretreatment dACC–parahippocampal connectivity and pretreatment BDI scores. At pretreatment, the MDD group was characterized by decreased dACC–parahippocampal connectivity relative to controls.

Figure 4

Biweekly BDI scores from individual MDD participants (thin lines) and average biweekly BDI scores from all MDD participants (thick line). Baseline assessments occurred before the start of treatment. Fourteen data points were not available.

Figure 5

Graphical illustration of the significant interaction between baseline connectivity and time predicting change in BDI scores from the HLM models. The lines represent the range of variability on connectivity: ‘low' represents the expectation for change in an individual who is a SD below the mean, ‘mean' for someone at the average, and ‘high' for someone a SD above. Note that the lines are model-based estimates and do not represent averages but rather ranges of brain connectivity variability.