Emotion regulation in mood and anxiety disorders: A meta-analysis of fMRI cognitive reappraisal studies

Highlights

• Patient groups exhibit both abnormal decreases and increases of brain activity.

• The fronto-parietal network is hypoactivated during reappraisal in patient groups.

• Hyperactivations may relate to both emotion experience and compensatory mechanisms.

• Hypoactivations are observed both with disorder-specific and unspecific stimuli.

• The cognitive reappraisal strategy employed has a significant effect on findings.

Abstract

Emotion regulation by means of cognitive reappraisal has been widely studied with functional magnetic resonance imaging (fMRI). To date, several meta-analyses of studies using cognitive reappraisal tasks in healthy volunteers have been carried out, but no meta-analyses have yet been performed on the fMRI data of clinical populations with identified alterations in emotion regulation capacity.

We provide a comprehensive meta-analysis of cognitive reappraisal fMRI studies in populations of patients with mood or anxiety disorders, yielding a pooled sample of 247 patients and 262 controls from thirteen independent studies. As a distinguishing feature of this meta-analysis, original statistical brain maps were obtained from six of these studies.

Our primary results demonstrated that patients with mood and anxiety disorders recruited the regulatory fronto-parietal network involved in cognitive reappraisal to a lesser extent in comparison to healthy controls. Conversely, they presented increased activation in regions that may be associated with the emotional experience (i.e., insula, cerebellum, precentral and inferior occipital gyri) and in regions whose activation may be the consequence of compensatory mechanisms (i.e., supramarginal gyri and superior parietal lobule). Moreover, activations in the left ventrolateral prefrontal cortex and the left superior temporal gyrus were associated with reinterpretation emotion regulation strategies, whereas medial frontal and parietal activations were associated with the deployment of distancing strategies.

The regions revealed by this meta-analysis conform to a pattern of dysfunctional brain activation during cognitive reappraisal common to mood and anxiety disorders. As such, this neural pattern may reflect a transdiagnostic feature of these disorders.

Introduction

Throughout our day-to-day lives we are confronted with situations that trigger negative emotions. The unpleasantness of the emotion or contextual factors can lead us to try to change the way we feel. To this end, we draw upon an inherent human capacity: emotion regulation. Emotion regulation has been defined as the processes by which individuals influence their emotions, when they have them, and how they experience and express these emotions (Gross, 1998a). There are several ways to carry out emotion regulation, but not all are equally adaptive. An emotion regulation strategy is considered maladaptive if it is unsuccessful in reducing emotional response or if it is associated with costs that potentially outweigh the short-term benefits brought about by diminishing acute emotions. Conversely, an emotion regulation strategy is considered adaptive if it decreases subjective distress and/or physiological arousal while maintaining one's ability to pursue meaningful short- and long-term goals (Campbell-Sills et al., 2014).

Emotion regulation strategies are classified according to when their primary effect appears during the emotion-generative process (Gross, 1998a). Thus, antecedent-focused strategies are those acting before emotional responses have been completely generated, while response-focused strategies are those put into practice after the full development of the emotional response. Overall, antecedent-focused strategies are considered more adaptive than response-focused strategies (Gross, 1998b). An example of an antecedent-focus strategy which has been widely studied is cognitive reappraisal. This strategy has been associated with decreased sympathetic nervous system activity and enhanced cognitive control of emotions, leading to decreased levels of negative affect and higher levels of positive emotions. Successful employment of this strategy subsequently brings about better interpersonal functioning along with physical and psychological well-being (Gross, 1998b, Gross and John, 2003, Webb et al., 2012, Gross, 2014, Hu et al., 2014). Importantly, several studies have shown that many patients with psychiatric disorders have difficulties in using cognitive reappraisal, and it has been suggested that ineffective emotion regulation may represent a transdiagnostic feature of mood and anxiety disorders (Campbell-Sills et al., 2006). Specifically, behavioral studies show a wide range of emotion regulation deficits in these patients, featuring the more frequent use of maladaptive strategies such as expressive suppression or less awareness and acceptance of emotions (Aldao et al., 2010, Cutuli, 2014, Ehring et al., 2008, Görlach et al., 2016). Likewise, neuroimaging studies show structural and functional abnormalities in the prefrontal cortex circuits related with top-down inhibitory control, which is necessary to deploy cognitive reappraisal strategies. This coincides with hyperreactivity in limbic structures implicated in emotion generation (Etkin and Wager, 2007, Phillips et al., 2008, Rive et al., 2013). Moreover, such inefficient use of cognitive reappraisal strategies may have relevant consequences not only for the development and maintenance of mental health alterations, but also for treatment response. Thus, Cognitive Behavioral Therapy (CBT), the most frequently used psychotherapy technique for these disorders (Beck, 2005), is tightly linked to improving cognitive reappraisal abilities (Taylor and Liberzon, 2007). For instance, it has been shown that, after treatment with CBT, patients with social anxiety disorder improve their performance on a cognitive reappraisal task, both at the behavioral and neurobiological level (Goldin et al., 2013).

The neurofunctional correlates of cognitive reappraisal have been widely studied with functional magnetic resonance imaging (fMRI). Emotion regulation paradigms typically expose subjects to stimuli of negative emotional content (e.g., images or videos), and although experiments differ in trial timelines, they generally alternate between task-blocks or trials in which participants are instructed to experience the negative emotions evoked by the images (i.e., maintain condition) with others in which participants are instructed to reduce the intensity of evoked negative emotions via cognitive reappraisal (i.e., reappraise condition) (Ochsner et al., 2002, Phan et al., 2005). Maintain is the most common control condition, and subjects are characteristically instructed to not down-regulate the evoked emotion, similarly to what was reported in Ochsner et al. (2002). Likewise, regarding the reappraise condition, in most studies participants are instructed to use distancing or reinterpretation as reappraisal strategies. The former refers to rationalizing the content of a situation by adopting the perspective of an uninvolved observer (e.g., when viewing a scene depicting a wounded person, presuming that the person is actually an actor). The latter refers to changing the meaning of stimuli in order to view the outcome of a situation in a more positive light (e.g., deciding that an image of weeping people outside a church is actually of a wedding instead of a funeral) (Ochsner et al., 2012, Dörfel et al., 2014).

The results of reviews and meta-analyses on the neurofunctional correlates of cognitive reappraisal in healthy volunteers have been rather homogenous. Regulating negative affective states involves activation of the prefronto-parietal network, and at times, the middle temporal gyrus. These prefronto-parietal activations are accompanied by significant deactivations of the limbic subcortical network (Kalisch, 2009, Diekhof et al., 2011, Ochsner et al., 2012, Buhle et al., 2013, Kohn et al., 2014, Etkin et al., 2015). More specifically, the regions consistently recruited as part of the prefronto-parietal network are the dorsolateral, medial and ventrolateral prefrontal cortices, the dorsal anterior cingulate cortex and the inferior parietal lobule. Notably, these regions have been traditionally associated with cognitive processes such as conflict monitoring, selective attention, working memory, mental state attribution, response selection and inhibition and semantic processing (Pessoa et al., 2003, Wager and Smith, 2003, Botvinick et al., 2004, Thompson-Schill et al., 2005, Aron et al., 2014). All of these processes are believed to be relevant for implementing successful cognitive reappraisal (Ochsner and Gross, 2014). Likewise, downregulated regions in the limbic network commonly include the amygdala, the ventral striatum and the insula, regions associated with the detection of arousing and potentially threatening stimuli, reward processing and the integration of information about body states, respectively (Craig, 2003, Haber and Knutson, 2010, Morrison and Salzman, 2010).

Although the cognitive reappraisal paradigm has also been extensively used in clinical populations to study the neurobiological correlates of altered emotion regulation capacity, to our knowledge, only one review (Zilverstand et al., 2016) has been conducted to summarize this information and no meta-analysis has been performed to provide a comprehensive description of the neurobiological commonalities underlying emotion regulation deficits across different mental health conditions. The aim of the present study is to identify, by means of a meta-analysis of fMRI studies assessing cognitive reappraisal in samples of patients with mood or anxiety disorders, the neural correlates of impaired emotion regulation. We specifically focused on mood and anxiety as disorders where concurring emotion regulation alterations have been consistently described (Campbell-Sills et al., 2006, Aldao et al., 2010). Our analyses were centered on comparing reappraise and maintain blocks during the presentation of images of negative emotional content in order to identify regions presenting both increased and decreased activation during cognitive reappraisal. Moreover, we explored the differences between reinterpretation and distancing strategies.

We hypothesized that activations in healthy controls (vs. patient group) during reappraise blocks would substantially overlap with previously reported regions in the prefronto-parietal network (Buhle et al., 2013, Diekhof et al., 2011, Kalisch, 2009, Kohn et al., 2014). In the patient group, we expected to find decreased activation of the prefronto-parietal network in combination with an ineffective downregulation of emotion generation regions (i.e., limbic regions). Likewise, from the sparse literature comparing different reappraisal strategies, we expected distancing to specifically activate parietal regions related to perspective taking and spatial attention, while reinterpretation would be linked to ventral prefrontal regions implicated in response selection and inhibition, along with temporal regions related to linguistic and semantic processing (Ochsner et al., 2012, Dörfel et al., 2014).