1 Introduction

Major depressive disorder (MDD) and anxiety disorders (AD), one of its most common comorbid groups of disorders (Fava et al., 2000), are each among the most common psychiatric disorders worldwide, (Malhi & Mann, 2018; van Tol et al., 2021) with devastating effects on virtually all areas of a patient´s life as well as contributing significantly to the overall burden of disease (James et al., 2018; Kessler & Greenberg, 2002; Silverstone & Studnitz, 2003). Comorbid MDD and AD are shown to have significantly less favourable treatment outcome, more severe symptom presentation and higher impact on quality of life (van Tol et al., 2021).
Despite extensive research, both disorders are still lacking reliable biomarkers for diagnosis, leading to uncertainty both in diagnosis and treatment (Mulders et al., 2015) and high rates of patients failing to achieve remission (Mulders et al., 2015; Trivedi et al., 2008). In particular, linking differences in the engagement of individual brain areas to psychiatric disorders has been shown to be insufficient in explaining the full symptomatic complexity of these disorders (Menon, 2019; Menon, 2011). One of the most consistent findings has been a hyperactivation of the amygdala as a common core mechanism of anxiety disorders (Shin & Liberzon, 2010). However, this finding could not account for the dynamics and individual heterogeneity of symptoms associated with anxiety disorders (Menon, 2011; Sylvester et al., 2012). Therefore, in recent years, the focus has shifted to alterations in neural networks. One important tool for quantifying the activity of neural circuits has been resting-state functional connectivity (RSFC), which measures the correlation of spontaneous fluctuations of the blood oxygen level-dependent (BOLD) signal between brain regions at rest (Cullen et al., 2014; Greicius et al., 2007; Kelly et al., 2012; Ogawa et al., 1990; van den Heuvel & Hulshoff Pol, 2010). RSFC has proven to be an advantageous and reliable method to examine psychiatric disorders, with disruptions of brain networks being consistently reported in both MDD (Frodl et al., 2010; Javaheripour et al., 2021; Kaiser et al., 2015; Schirmer et al., 2023; Sharpley & Bitsika, 2013) and AD (Northoff, 2020; Peterson et al., 2014; Sylvester et al., 2012; Xu et al., 2019).
Of particular interest in this area has been the so-called triple network, encompassing the default mode network (DMN), the ventral attention network (VAN) and the executive control network (ECN). According to Menon´s (2011) model, the VAN, with key nodes in the insula and dorsal anterior cingulate cortex (Sridharan et al., 2008), associated with filtering and integrating information (Fan et al., 2017) and assigning saliency to stimuli, operates as a switch between the task-negative DMN and the task-positive ECN (Menon & Uddin, 2010). The DMN, with key nodes located in the ventromedial prefrontal cortex (VMPFC) and posterior cingulate cortex (PCC) (Buckner et al., 2008; Manoliu et al., 2013; Raichle et al., 2001), is activated during times of cognitive rest and associated with self-referential thoughts (Buckner et al., 2008; Raichle, 2015) and internal attention (Li et al., 2021). The ECN on the other hand includes areas of the dorsolateral prefrontal cortex (dlPFC) and posterior parietal cortex (PPC) (Fan et al., 2017; Jiang et al., 2017) and is associated with working memory, decision-making, and cognitive control (Ernst et al., 2019; Lerman et al., 2014).
A failure to adequately switch between the DMN and the ECN could lead to “deficits in engagement and disengagement” of these networks (Menon, 2011). Aberrant functional connectivity between these networks consistently found for both MDD (Kaiser et al., 2015; Mulders et al., 2015; Zheng et al., 2015) and AD (Etkin & Wager, 2007; Massullo et al., 2020; Sylvester et al., 2012; Williams, 2016) could therefore be related to a set of various core symptoms in these disorders, such as rumination, cognitive deficits or emotional dysregulation (Hamilton et al., 2013; Menon, 2011).
While MDD and AD have both separately been extensively studied in regards to alterations in RSFC, research examining comorbid MDD and AD has been scarce, with only three studies covering the topic to the authors knowledge at this time: Pannekoek et al. (2015) examined patients with comorbid MDD and one or more anxiety disorder compared to those with only MDD or anxiety disorders alone and healthy controls. For the comorbid group, they found significantly increased RSFC between the limbic network and a cluster containing the bilateral precuneus, intracalcarine cortex, lingual gyrus, and posterior cingulate, and with a cluster including the right precentral gyrus, inferior frontal gyrus, and middle frontal gyrus. No difference was found for the other groups. From the same cohort study, Nawjin et al. (2022) on the other hand, found no significant differences when comparing patients with comorbid MDD and anxiety disorders with patients with MDD and healthy controls. Oathes et al. (2015) did not find any significant results within the triple network.
While the posterior cingulate region found to be significant in Pannekoek et al. (2015) is considered part of the DMN, no study has yet examined the triple network specifically in the context of comorbid MDD and AD. For this reason, we investigated the RSFC between the three core networks belonging to the triple network, DMN, VAN and ECN, comparing patients with MDD, patients with comorbid MDD and AD and healthy controls, using a seed-based approach utilizing commonly used seed-regions of all three networks. The RSFC was compared both within and between all three networks studied and its relationship to symptom severity and medication status was explored.