The identification of pathophysiological mechanisms common to psychiatric disorders as well as mechanisms unique to specific psychiatric disorders are critical for elucidating the biological mechanisms underlying psychiatric symptoms and developing targeted treatment protocols (Insel et al., 2010; Nusslock et al., 2015). The purpose of this study was to identify differences in brain activation patterns between patients with psychiatric disorders and HCs during cognitive regulation of negative emotions. More specifically, we focused on the comparison of patients with MDD and PTSD. In this study, we employed a classical emotion regulation paradigm to identify emotion regulation networks common and specific to HCs as well as patients with MDD and PTSD by contrasting reappraisal of negative pictures with natural viewing of negative pictures. Both clinical groups showed higher scores of anxiety and depression compared to HCs, and with a higher self-reported severity of depression in MDD compared to PTSD patients. On the neurobiological level, exploratory analyses showed that MDD and PTSD patients had less right-lateralized activation of the inferior, middle and superior frontal gyrus during cognitive reappraisal (reappraise vs view) compared to HCs. Importantly, this cluster in the right IFG was negatively correlated with the scores within the hospital anxiety and depression scale (HADS), suggesting that less recruitment of right IFG may be associated with greater severity of anxiety as well as depressive symptomatology. Furthermore, our task-based and IC analyses confirmed group differences in connectivity during cognitive reappraisal. To our knowledge this is the first study directly comparing focal neural activation and FC patterns in HCs, MDD and PTSD patients.
Comparison with the existent literature on neural correlates of cognitive reappraisal in clinical populations.
Our whole brain fMRI analysis revealed significant activation of the bilateral inferior frontal gyrus (pars triangularis), bilateral temporal, middle frontal, and superior frontal gyrus during cognitive reappraisal. Previous studies consistently report that cognitive reappraisal engages a prefronto-parietal network which exerts top-down control on limbic areas (Costafreda et al., 2008; Kohn et al., 2014; Picó-Pérez et al., 2017). Accordingly, our results are largely consistent with the literature.
In terms of task-based fMRI connectivity, HCs exhibited increased FC between the rIFG seed and clusters distributed along posterior and frontal regions. Consistent with previous studies, our findings suggested the existence of several large-scale networks that are co-activated during cognitive reappraisal. Morawetz and colleagues (2020) identified four clusters of co-activation patterns during emotion-generative and emotion-regulatory processes to brain-networks underlying prominent psychological functions and evaluated their specificity in relation to emotion regulation. Increased connectivity between rIFG and superior and middle frontal gyri as well as with the precuneus and the left lingual gyrus overlaps with a network which plays an intermediary role in reappraisal by integrating information from prefrontal and subcortical areas to generate and regulate emotional responses.
Patients with depression and PTSD displayed reduced activation of the vIPFC (a crucial node of the emotion regulatory network) during downregulation of negative emotion (Zilverstand et al., 2017b). Wang et al., (2014) examined the neural mechanisms of self-related reappraisal in Chinese MDD outpatients and found that depressed individuals exhibited diminished activation in left IFG when detachment strategies were adopted (subjects should view the situation as fake or unreal and detach themselves from the situation). Contrary to some previous reports, we found that patients showed reduced activation, especially in the right-lateralized IFG, but not in the left-homologous region. However, this functional hemispheric asymmetry has also been found by other investigators reporting a greater role for the right hemisphere in the processing of negative affect in HCs (Davidson et al., 2002; Kim & Hamann, 2007) and impairments in clinical conditions. Furthermore, Wager and colleagues (2008) showed that right vlPFC activity predicted drops in self-reported negative emotion, and that this relationship was independently mediated by separate pathways through the amygdala and the ventral striatum, thus stressing the importance of the right vlPFC for emotion regulation. Interestingly, neuroimaging studies report that activation in the vlPFC is strongly lateralized based on the type of stimuli processed (Levy and Wagner, 2011). Studies addressing the differences between upregulation and downregulation of negative emotions report that upregulation engaged primarily left-lateralized prefrontal regions, whereas downregulation engaged bilateral prefrontal regions (Kim & Hamann, 2007).
In our study, we used strategies aimed at downregulating negative stimuli. Contrasting reappraisal versus view showed bilateral vlPFC activation in the control group. This is consistent with the pattern of right-lateralized activation for downregulation of negative emotions described by Ochsner et al. (2004). The altered right vlPFC activation in both patient groups implies an impaired cognitive control capacity of bottom-up systems, such as the amygdala that appraise the affective properties of stimuli. In general, the vlPFC - especially right vlPFC - is thought to play an important role in response inhibition as a particular form of executive control (Hampshire et al., 2010). Studies have highlighted the importance of the right hemisphere during cognitive reappraisal, engaging greater activation when the stimulus is more difficult and more cognitively demanding to reappraise (Ligeza et al., 2016). The correlation of right vlPFC activation with depressive and anxious symptomatology further supported the findings of the vlPFC as a key region involved in cognitive reappraisal across mood and anxiety disorders.
In addition to impairments across psychopathologies, (Zilverstand et al., 2016) reported deficits specific for mood and anxiety disorders. For patients with MDD, they found hyperactivity of the amygdala during downregulation of negative emotions, a finding in line with enhanced bottom-up responses and reduced top-down modulatory ability. Our ROI analysis indicated disorder-specific deficits for patients with MDD during cognitive reappraisal in the dmPFC, SMA, left middle temporal cortex and left precentral gyrus. These areas have been related to the execution of reappraisal, suggesting that a deficit in signaling the need for regulation in the vlPFC leads to less effective reappraisal (Kohn et al., 2014), a pattern specific to patients with MDD in our sample.
Regarding the neural correlates of emotion regulation deficits in PTSD, in addition to the reduced activation in the right IFG, the group of PTSD patients showed hypoactivation in the right dmPFC. Previous studies have focused on the structure, neurochemistry, and function of the amygdala, medial prefrontal cortex, and hippocampus in PTSD. Contrary to the hyperresponsiveness of the amygdala during emotion regulation, the medial prefrontal cortex appears to be volumetrically smaller and is hyporesponsive during symptomatic states and the performance of emotional cognitive tasks in PTSD (L. M. Shin et al., 2006). This finding is consistent with our results.
Task-based and intrinsic functional connectivity
FC analyses further investigated the importance of the group difference in the right IFG for cognitive reappraisal. In a task-based fMRI connectivity analysis using the right IFG as seed region, PTSD patients exhibited altered FC compared to HCs and MDD patients during reappraisal. Compared to HCs, PTSD patients showed increased FC of the rIFG with the left frontal orbital cortex, the left temporal pole, the left inferior frontal gyrus (pars opercularis), and the left frontal operculum. Interestingly, our results indicated altered connectivity predominantly in the PTSD group, suggesting higher FC in a fronto-temporo-parietal network. Higher connectivity in this network may result from a compensatory mechanism, taking into account that regions are important for emotion regulation and more specifically cognitive control mechanisms (Mathiak & Weber, 2006). Although it may be speculative, this fronto-temporo-parietal hyperconnectivity may also reflect the use of suppression strategies that rely on engagement of this network. Furthermore, abnormalities on fronto-limbic connectivity have been the most discriminant feature for classification of PTSD based on the resting-state connectivity between the prefrontal and limbic regions (Liu et al., 2015).
In addition, PTSD patients showed greater FC than MDD patients in regions including the precuneus, the anterior and posterior cingulate gyrus and the bilateral paracingulate gyrus. This finding is consistent with reports of PTSD patients showing increased FC in the anterior cingulate gyrus as a core component of the salience network (Chen et al., 2019). This may be related to the role played by the ACC in monitoring and appraisal of the external environment, reciprocally connecting brain regions to regulate stressor-related autonomic nervous system activity (Shin et al., 2013). The PCC is also implicated in stress neural circuit (Bremner, 2007), and researchers have found anatomical and functional change of this region in patients with PTSD (King et al., 2016; Zhang et al., 2012).
To investigate more general connectivity group differences, we employed an IC analysis as a whole-brain voxel-based connectivity measure that represents how well connected any given voxel is to the rest of the gray matter voxels in the brain. Compared to the conventional seed-based connectivity analyses, IC does not require a priori knowledge for the selection of the regions of interest (Martuzzi et al., 2011). In the present work, the comparison between HCs vs PTSD and PTSD vs MDD during cognitive reappraisal indicated IC abnormalities in PTSD patients. Previous studies have demonstrated impaired within- and between network FC in PTSD. For example, Akiki and colleagues (2017) report hypoactivity of the default mode network (DMN) and central executive network (CEN) that are putatively destabilized by an overactive and hyperconnected salience network (SN), which appears to be associated with an inefficient DMN-CEN modulation (Akiki et al., 2017). For the DMN network, we found that the group of PTSD patients differed from HCs in the activation of core regions such as the posterior cingulate cortex and precuneus. It has been suggested that DMN hypoactivity is associated with a deficit in the processing of autobiographical memory and self-references which may be a core underlying mechanism regarding PTSD symptomatology (Hayes et al., 2012).
Limitations
There are some limitations of the current study. First, the smaller sample size of PTSD compared to HCs and patients with MDD may have impeded the detection of significant differences between these groups. In addition, PTSD patients presented with relatively mild symptoms. And whereas most individuals in the MDD group were treated on the ward, a high percentage of PTSD patients was treated in an outpatient clinic. In addition, different subtypes of emotion regulation in PTSD may complicate interpretation of the data. A previous review described a model that includes these two types of emotion dysregulation in PTSD (Lanius et al., 2010). In this model, reexperiencing/hyperarousal reactivity is viewed as a form of emotion dysregulation that involves abnormally low activation in medial anterior brain regions, mediated by failure of prefrontal inhibition of limbic regions. In contrast, the dissociative subtype of PTSD is described as a form of emotion dysregulation that involves emotional overmodulation mediated by midline prefrontal inhibition of the same limbic regions. Both types of modulation are involved in a dynamic interplay and lead to alternating symptom profiles in PTSD (Lanius et al., 2010).