Disrupted functional connectivity patterns of the insula subregions in drug-free major depressive disorder

Highlights

• The insular subregions show differential change patterns in MDD.

• The connectivity of anterior and middle insula are higher and positively correlated to HDRS in MDD.

• The connectivity of posterior insula is lower and negatively correlated to HDRS in MDD.

Abstract

Background

Major depressive disorder (MDD) is characterized by impairments in emotional and cognitive functions. Emerging studies have shown that cognition and emotion interact by reaching identical brain regions, and the insula is one such region with functional and structural heterogeneity. Although previous literatures have shown the role of insula in MDD，it remains unclear whether the insular subregions show differential change patterns in MDD.

Methods

Using the resting-state fMRI data in a group of 23 drug-free MDD patients and 34 healthy controls (HCs), we investigated whether the abnormal connectivity patterns of insular sub-regions or any behavioural correlates can be detected in MDD. Further hierarchical cluster analysis was used to identify the functional connectivity-clustering patterns of insular sub-regions.

Results

Compared with HCs, the MDD exhibited higher connectivities between dorsal agranular insula and inferior parietal lobule and between ventral dysgranular and granular insula and thalamus/habehula, and lower connectivity of hypergranular insula to subgenual anterior cingulate cortex. Moreover, the three subregions with significant group differences were in three separate functional systems along anterior-to-posteior gradient. The anterior and middle insula showed positive correlation with depressive severity, while the posterior insular was to the contrary.

Limitations

The small and unbalanced sample size, only included moderate and severe depression and the possible inter-individual differences may limit the interpretability.

Conclusions

These findings provided evidences for the MDD-related effects in functional connectivity patterns of insular subregions, and revealed that the subregions might be involved in different neural circuits associated with the contrary impacts on the depressive symptoms.

Introduction

Major depressive disorder (MDD) is a common psychiatric disorder characterized by impairments in emotional and cognitive functions including sustained negative affect or loss of pleasure, negative automatic thoughts or biases in attention, memory and interpretation, and concentration difficulties (Gotlib and Joormann, 2010). Previous neuroimaging studies have found that patients with depression showed functional (Diener et al., 2012, Mulders et al., 2015) and structural (Wang et al., 2017b) abnormalities in brain regions implicated in the dysregulation of emotional and cognitive processing. Recently, emerging studies have focused on the interplay between cognition and emotion (Cromheeke and Mueller, 2014), and have shown that cognition and emotion strongly interact by reaching identical brain regions in largely overlapping brain networks to modulate behavior (Diener et al., 2012, Pessoa, 2010). Therefore, exploring the underlying neural mechanism of this interaction may contribute to explain the pathology of MDD.

The insula cortex is a core brain area, serving as an interface between internal and extra-personal stimuli, for integrating information across neural systems underlying diverse cognitive, emotional and homeostatic processes (Craig, 2011, Simmons et al., 2013, Wang et al., 2014). Anatomically, the insula receives input from thalamic nuclei, and also receives reciprocal projections from amygdala, limbic system, and cortical structures (Craig, 2002, Zhuo, 2016). Given its diverse connectivity and functional heterogeneity, it is unsurprising that the insula, as a multi-scale organization, is formed from separate but highly interacted sub-regions in meta-analysis based on multi-modal neuroimaging studies (Kelly et al., 2012, Kurth et al., 2010b). The symptoms expressed in MDD might result either from changes within the insula, or from abnormal afferent and efferent projections into this region (Avery et al., 2014, Paulus and Stein, 2010). Structural and functional alterations in the insula constantly found in neuroimaging findings regard depression (Chen et al., 2015, Fitzgerald et al., 2008, Garrett et al., 2011, Guo et al., 2015, Lai and Wu, 2014, Zhang et al., 2016). More specifically, these changes have been associated with MDD symptom severity and illness duration (Belden et al., 2015, Sliz and Hayley, 2012).

The broad structural and functional connectivity between the insula and other brain regions in fronto-limbic circuit and associated cortical networks also provides evidence for its role in MDD. Decreased functional connectivity (FC) between insula and medial prefrontal cortex regions was evident when MDD patients prompted to maintain their reactions to negative stimuli (Perlman et al., 2012). Furthermore, the decreased fronto-insular cortical network activation during the resting state in MDD patients has been associated with greater maladaptive rumination (Hamilton et al., 2011). Besides, the resting-state functional connectivity (rs-FC) between the insula and fronto-limbic regions, including orbitofrontal cortex, anterior cingulate, amygdala, thalamus, and hippocampus, were obviously changed in MDD (Avery et al., 2014, Ho et al., 2014, Lui et al., 2011, Manoliu et al., 2013, Mulders et al., 2015, Zeng et al., 2014a). Moreover, severity of MDD was consistently related to the FC between the insula and amygdala (Avery et al., 2014), anterior cingulate (Horn et al., 2010). In addition, increased nodal local efficiency in insula and higher structural connectivity among cortical-limbic networks have been observed in MDD (Chen et al., 2017, Long et al., 2015).

Although previous literatures have consistently shown alterations in structural and functional connectivity with insula in MDD, it remains unclear whether the insular subregions show differentially functional change patterns in MDD. The insula showed both increased and decreased grey matter volume (Schmaal et al., 2017, Wang et al., 2017b), and hypo- and hyperactivity (Diener et al., 2012) in MDD. Differences in sample size, medication status, and methodological approaches may account for observed inconsistencies in MDD. It is also noteworthy that the insula is comprised of separate subregions, only assessing it as a whole region may obscure individual differences of functional connectivity with insula in MDD.

In addition, with the advance of network-based approach, accumulating evidence suggesting that depressive symptoms are associated with “dysfunction and disorganization of brain network” rather than with a single “brain lesions” (Chen et al., 2017, Ho et al., 2017, Zeng et al., 2012). Using seed-based methods, resting-state functional connectivity (RSFC) provides a promising imaging technique for measuring correlations in spontaneous neural activity between different brain regions (Cole et al., 2014, Power et al., 2011, Yeo et al., 2011). Thus, RSFC has advantages to detect functional abnormalities of brain circuit or network (Fox and Greicius, 2010), and to minimize confounds caused by specific tasks or patients’ cooperation.

Thus, we attempted to investigate the alterations in connectivity with separate insular subregions in MDD patients. According to the insular parcellation of the Brainnetome atlas (Fan et al., 2016), it can be subdivided into six sub-regions through an anatomical connectivity-based parcellation approach, which relates well to other functional and histological maps of the insular cortex (Chang et al., 2013, Kelly et al., 2012, Kurth et al., 2010a, Kurth et al., 2010b, Morel et al., 2013). Using the resting-state functional magnetic resonance imaging (fMRI) data in a group of 23 MDD patients and 34 healthy controls (HCs), we investigated the potentially abnormal connectivity patterns of insular sub-regions and any behavioural correlates in MDD patients. To further validate which insular functional subsystem the changes in these subregion-based functional connectivity pattern belonged to, we performed hierarchical cluster analysis to identify the functional connectivity-clustering patterns of insular sub-regions.