4 Discussion

The present study examined RSFC network differences between patients with a diagnosis of MDD, patients with comorbid MDD and at least one anxiety disorder and healthy controls within the context of the triple network model. Comparing MDD+AD to HC, we found significantly reduced RSFC for the VAN with both the ECN and the DMN, whereas there were no alterations between the ECN and the DMN. This finding confirms the triple network model by Menon et al. (2010). In addition, the ECN showed significantly reduced within-network-connectivity. These effects were independent of severity of symptoms and medication status. No effects were found when comparing MDD with HC and MDD with MDD+AD.
Although we did not find any differences between patients with MDD and HC, aberrations of RSFC within the triple network are well documented (Kaiser et al., 2015; Mulders et al., 2015; Zheng et al., 2015). However, results in the RSFC-literature are generally heterogonous (Kaiser et al., 2016; Lydon-Staley et al., 2019) and with the relatively small sample size of this study and rigorous correction for multiple testing (Bonferroni correction is regarded as a comparatively conservative correction method (Chen et al., 2017)), smaller effects could have gone undetected. Therefore, we cannot exclude, that the MDD group does show aberrant RSFC compared to HC, in particular since this group does not differ to MDD+AD.
In research concerning MDD and the triple network model, the introspective qualities of the DMN and the more outwards oriented qualities associated with the ECN have been central to the discussion of the symptomatic correlates of the commonly found aberrations (Jiang et al., 2017; Li et al., 2021). Here, alterations in DMN connectivity have been found to be associated with higher levels of rumination (Brakowski et al., 2017; Kühn et al., 2012) and negative self-referential thoughts (Cullen et al., 2014). The ECN, on the other hand, shows a more outward-directed and stimulus-driven set of tasks, being involved in decision making and cognitive control (Li et al., 2021; Manoliu et al., 2013). Within the triple network model, alterations in the RSFC between these two networks and the VAN as a switch between them are thought to impair the engagement of the ECN and the disengagement of the DMN (Menon & Uddin, 2010), leading to maladaptive rumination and impaired cognitive abilities (Schimmelpfennig et al., 2023).
Similar to the MDD literature, in the much more sparse research regarding anxiety, the VAN is thought of as an important intermediary between the DMN and the ECN as well (Nawijn et al., 2022; Pannekoek et al., 2015). Here, however, other functions of the DMN and ECN are considered central to the symptomology, namely their role in the regulation of emotion and fear response (Sylvester et al., 2012). The DMN has been shown to be linked to emotion regulation (Macêdo et al., 2022), with a focus on emotion perception (Kim & Yoon, 2018), reinforcement expectancy (Blair, 2007) and fear extinction (Sylvester et al., 2012). The ECN on the other hand has been associated with more conscious and control oriented emotional regulation strategies, such as redirection of attention towards non-emotional stimuli and suppression of amygdala responses when attention is engaged with a non-emotional stimulus (Bishop et al., 2004; Sylvester et al., 2012). Impaired switching between the DMN and the ECN by the VAN is thought to be associated with less adequate control over fear responses and emotional regulation (Sylvester et al., 2012).
In summary, alterations within the triple network in the context of MDD are interpreted in terms of an inadequate switch between internally (DMN) and externally (ECN) oriented attention and stimulus processing. In contrast, in the context of AD, alterations in the triple network are interpreted as an imbalance between a more automated and unconscious processing of emotions (DMN) and conscious, cognitively controlled emotion regulation (ECN).
In this context, studying patients with comorbid MDD and AD serves to identify those neurological alterations within the triple network, that distinguish this important group of patients from those with MDD alone. We found reduced connectivity between the DMN and the VAN with clusters in the right middle and anterior cingulate and paracingulate cortex and the left cingulate cortex, between the ECN and the VAN, with clusters in the left and right middle cingulate and paracingulate gyrus and the right superior frontal gyrus, and within the ECN with a cluster in the left middle frontal gyrus. In the context of the interpretation of alterations discussed above, the alterations we found could lead to additional difficulties switching between the DMN and the ECN and in turn contribute to the specific symptomology of ADs by impairing the switch from unconscious, automated modes of emotional regulation to more conscious, cognitively controlled strategies (Fan et al., 2017). This could manifest in the high degree of internally oriented attention characteristic for AD (Fan et al., 2017). Additionally, the hypoconnectivity found within the ECN may be associated with a higher difficulty in using and applying these strategies effectively (Sylvester et al., 2012).
Furthermore, while anxiety symptoms accompanying depression are often associated with higher depression severity (Gaspersz et al., 2017), this was not the case in the current study as both groups showed similar severity of depression symptoms. Additionally, the severity of anxiety symptoms measured with the BAI was also not significantly different between groups. For this reason, our results may provide important indications of specific network signatures of comorbid MDD and anxiety disorders as opposed to greater overall symptom severity in the patient groups. Another result supporting this notion was our finding that the strength of the RSFC was not associated with symptom severity within the patient groups. This is consistent with previous findings (Pannekoek et al., 2015). Pannekoek et al. (2015) suggested that this may be due to the changes in RSFC being a result of trait rather than state characteristics or alternatively, the relatively mild overall symptom severity of their patient groups. However, as our patients were recruited from an inpatient setting, typically shortly after admission, our findings may further reinforce the qualitative nature of these alterations.