Review articleThe effect of strategies, goals and stimulus material on the neural mechanisms of emotion regulation: A meta-analysis of fMRI studies
Introduction
Cognitive control of emotions by means of flexibly responding to affective events is of great importance in our daily social life and instrumental for our mental and physical well-being (Berking and Wupperman, 2012, Eftekhari et al., 2009, Gross et al., 2006, Gross and Muñoz, 1995). Emotion regulation comprises all extrinsic and intrinsic control processes whereby people monitor, evaluate and modify the occurrence, intensity and duration of emotional reactions (Thompson, 1994). In the last 10 years, more than 500 neuroimaging studies have investigated the neural basis of emotion regulation, and substantial progress has been made toward building neurally plausible models of emotion regulation that consider multiple cognitive control processes (e.g., Etkin et al., 2015, Gross, 2002, Kalisch, 2009, Koole, 2009, Ochsner et al., 2012, Phillips et al., 2008, Smith and Lane, 2015). However, single studies usually provide limited insight into the function of specific brain regions (Sarter et al., 1996, Yarkoni et al., 2010). Several reviews and meta-analyses have permitted some synthesis of imaging results on emotion regulation, specifically focusing on reappraisal as a regulation strategy (Buhle et al., 2014, Diekhof et al., 2011, Frank et al., 2014, Kohn et al., 2014, Messina et al., 2015). However, any comprehensive meta-analysis on emotion regulation requires consideration not only of the neural basis underlying reappraisal of negative emotions, but also of other emotion regulation strategies as well as the direction of the regulation strategy and the conditions that influence emotion regulation such as the nature of emotion induction method. Previously, several moderators of strategy effectiveness including factors related to strategy type, purpose of the regulation strategy and study design have been identified (Webb et al., 2012). Thus, different regulation strategies might engage different neural networks, and these networks might be recruited to a different extent depending on the regulation goal and the emotion induction. Here we provide a comprehensive and systematic meta-analysis of functional magnetic resonance imaging (fMRI) studies, which investigated a variety of emotion regulation strategies in combination with different emotion induction methods and regulation goals. The current meta-analysis examined whether 1) different emotion regulation strategies are based on different or the same neural networks; 2) which brain regions in particular support the up-regulation and the down-regulation of emotions, respectively; and 3) to which degree the neural networks realising emotion regulation depend on the stimulus material used to elicit emotions. In the following section we will briefly review some of the key findings of previous neuroimaging studies on emotion regulation and explain the sorting of studies for the current meta-analysis.
Different frameworks for conceptualising emotion regulation have been proposed that distinguish between different emotion regulation processes and strategies (Gross, 1998a, Gross, 1998b, Koole, 2009, Larsen, 2000, Parkinson and Totterdell, 1999, Thayer et al., 1994, Webb et al., 2012). The most widely used model to date is the process model of emotion regulation that proposes five distinct emotion regulation processes and corresponding emotion regulation strategies (Gross, 1998a, Gross, 1998b): situation selection, situation modification, attentional deployment (distraction, concentration), cognitive change (reappraisal), and response modulation (suppression). An alternative, more recent classification of emotion regulation strategies considers the target and function of regulation (Koole, 2009). In this view, emotion regulation targets three emotion-generating systems (attention, knowledge, and the body), and emotion regulation strategies can be classified along their psychological function as need-oriented (e.g., when targeting the attention system in terms of thinking pleasurable or relaxing thoughts), goal-oriented (e.g., when targeting the knowledge system by cognitive reappraisal), or person-oriented (e.g., when targeting the body by controlling breathing). Importantly, both frameworks are in agreement that emotion regulation is based on three main processes (Gross, 1998a, Gross, 1998b) or emotion generating systems (attentional deployment/attention, cognitive change/knowledge and response modulation/body; Koole, 2009) and use highly similar categories for goal-oriented emotion regulation strategies targeting these processes/systems (distraction, cognitive reappraisal, suppression). Within this framework, distraction is defined as a process in which selective attention is used to limit the extent to which the emotionally evocative aspects of an event or stimulus are attended and appraised (Kalisch et al., 2006, Kanske et al., 2011, McRae et al., 2010). Emotion regulation via reappraisal is more diverse and comprises various tactics such as reinterpretation (changing the meaning of a stimulus), detachment or distancing (creating a sense of physical or psychological distance from the emotional event; e.g., through perspective taking), reality challenge (changing the authenticity of what is being depicted), or acceptance (invoking the justification that sometimes bad things happen) (McRae et al., 2012). Finally, suppression is used to modify the behavioural or physiological response to an emotional stimulus, e.g., facial expressions (Vrticka et al., 2011).
According to contemporary dual-process models of emotion regulation, these strategies have further been categorized as deliberate/explicit (also called effortful, conscious or controlled) and automatic/implicit (also called incidental or unconscious) (Berkman and Lieberman, 2009, Gyurak et al., 2011, Koole and Rothermund, 2011, Mauss et al., 2007). According to this taxonomy, reappraisal constitutes an explicit form of emotion regulation, as it includes the active, voluntary reinterpretation of the meaning, cause, consequence or personal significance of the emotion-inducing stimulus by means of processes such as covert or overt verbal labelling (Ochsner and Gross, 2008). In contrast, incidental (or automatic) emotion regulation can be achieved as a by-product of engaging in other, intentional tasks, for example by providing a distraction from the emotional aspects of the stimulus. Thus, individuals are unaware of the modulation of emotional control (Gyurak et al., 2011).
Despite there being several emotion regulation strategies, only few studies compared these directly (e.g., Dörfel et al., 2014, Goldin et al., 2008, Kanske et al., 2012, McRae et al., 2010, Vanderhasselt et al., 2013), as most studies to date focussed on one strategy alone, namely reappraisal. This makes it difficult to disentangle the neural systems implicated in emotion regulation in general from those specifically involved in a particular strategy. Hence, our study was explicitly aimed at providing this strategy-specific overview. Previous meta-analyses on reappraisal established a widespread network of regions including bilateral ventrolateral prefrontal cortex (VLPFC) and dorsolateral prefrontal cortex (DLPFC), parietal and temporal regions, supplementary motor area (SMA) as well as cingulate cortex (Buhle et al., 2014, Kohn et al., 2014). Recent findings suggest that reappraisal in particular is associated with activation of VLPFC and orbitofrontal cortex (OFC) while other strategies (e.g., detachment, suppression and distraction) are linked to increased activation in a distinct right prefrontal-parietal network including inferior parietal cortex and DLPFC (Dörfel et al., 2014). McRae et al. (2010) showed that distraction compared to reappraisal was associated with increased activity in middle frontal gyrus and superior parietal lobe, while Kanske et al. (2011) found increased activity for distraction in comparison to reappraisal in dorsomedial prefrontal cortex (DMPFC), superior frontal gyrus (SFG), parietal regions and the insula. Suppression in contrast to reappraisal has been found to increase responses in DLPFC, VLPFC, ventromedial prefrontal cortex (VMPFC), anterior cingulate cortex (ACC) and parietal cortex, insula and amygdala (Goldin et al., 2008). Furthermore, suppression in contrast to reappraisal has been associated with increased activity in SFG and temporal regions (Hayes et al., 2010) as well as lateral prefrontal cortex (Vrtička et al., 2011). Taken together, these studies suggest that different emotion regulation strategies differentially engage neural systems involved in attention, response inhibition and cognitive reframing. Our aim was to investigate in more detail to which extent neural networks are strategy-specific. For this we sorted the emotion regulation literature by emotion regulation processes with specific strategies and strategy subtypes (Webb et al., 2012). We differentiate between three categories of emotion regulation strategies based on the existing prominent frameworks (Gross, 1998a, Gross, 1998b, Gyurak et al., 2011, Koole, 2009): a) attentional deployment by using distraction, concentration, mindfulness, and affect labelling (implicit emotion regulation); b) cognitive change by using reappraisal, reinterpretation, distancing and detachment (explicit emotion regulation); and c) response modulation via suppression (explicit emotion regulation).
A second complication in previous work is that emotion regulation usually serves the goal of down-regulating (i.e., decreasing) emotions, and this was often the main focus of most studies. However, up-regulation (i.e., increase) is also an emotion regulation strategy by which a person increases the magnitude of the emotional response. Given that most studies investigated negative emotions, up-regulation of them is usually maladaptive, for instance in anger rumination when a person is upset by a provocation (Bushman, 2002, Denson et al., 2009). However, the understanding of the neural underpinnings of up-regulation of emotions might inform emotional dysregulation as observed e.g., in depression, which has been linked to rumination (Aldao et al., 2010, Yoon et al., 2012). Few studies to date investigated up-regulation and down-regulation of emotions (e.g., Kim and Hamann, 2007, Morawetz et al., 2016a, Urry et al., 2009, Urry et al., 2006) and did not converge on a coherent pattern of activity when directly contrasting up- versus down-regulating. Holland et al. (2013) reported increased activity in inferior frontal gyrus (IFG), middle frontal gyrus (MFG), superior frontal gyrus (SFG), middle temporal gyrus (MTG) and parietal cortex for the down-regulation of emotion compared to the up-regulation, while Ochsner et al. (2004) found enhanced responses in the DLPFC and lateral orbitofrontal cortex (OFC) for the same contrast. Another study reported increased responses in a number of prefrontal regions (including OFC, MFG, SFG, IFG) and parietal as well as temporal regions (Kim and Hamann, 2007). The reverse contrast, up-regulation versus down-regulation of emotion, has been associated with increased activity in medial prefrontal cortex and posterior cingulate cortex (Ochsner et al., 2004) as well as in DLPFC, ACC, OFC, and amygdala (Eippert et al., 2007). Another study reported increased responses in VLPFC, SFG, insula, temporal and parietal regions for the up-regulation compared to the down-regulation of emotion (Morawetz et al., 2016a), while Kim and Hamann (2007) reported increased activity in medial frontal gyrus, MFG, and OFC for the same contrast. Finally, Holland et al. (2013) only observed one cluster in the superior temporal gyrus to be more activated for up-regulating compared to down-regulating of emotions. Prior research, which was aimed at disentangling this mixed pattern of results suggested that both increasing and decreasing of emotional responses recruit mostly similar left prefrontal regions; however, the down-regulation of emotion engaged right prefrontal regions to a greater extent than the up-regulation (Ochsner et al., 2012). Furthermore, increasing emotions has been suggested to differentially involve anterior portions of dorsomedial prefrontal cortex (DMPFC). Both emotion regulation goals might modulate activity in the striatum (including caudate and putamen), but might differ in the modulation of amygdala responses (Ochsner et al., 2012). To further shed light on these issues our study directly compared emotion regulation goals (up- and down-regulation of emotion) with each other, searching for joint as well as goal-specific activations.
Another important aspect of emotion regulation concerns the emotion-inducing stimuli. The majority of neuroimaging studies on emotion regulation used complex emotionally evocative scenes from the International Affective Picture System (IAPS) (Bradley and Lang, 2007) to induce negative emotions (e.g., Denny et al., 2015, Eippert et al., 2007, Ochsner et al., 2004, Urry et al., 2006, Wager et al., 2008). However, several studies started to move beyond this classical operationalization of emotion regulation and used different stimulus material ranging from faces and film clips, reward and pain, to scripts and shapes (e.g., Kalisch et al., 2005, Lévesque et al., 2003, Morawetz et al., 2016a, Morawetz et al., 2016b, Nelson et al., 2015, Sokol-Hessner et al., 2013). Previous research suggested that the mode of emotion induction rather than the process of emotion regulation per se might determine frontal activation patterns. For example, Kalisch (2009) found differences in frontal activation when comparing a meta-analysis based on reappraisal studies using IAPS pictures to an extended meta-analysis on reappraisal studies using IAPS and other stimuli such as pain and film clips. The latter analysis revealed an additional cluster in the right anterior middle frontal gyrus. In line with this, it has been shown recently that although reappraisal seems to operate in an abstract fashion independent of stimulus material, reappraisal in response to film clips was associated with more widespread activity within the emotion regulation network than in response to IAPS pictures (Morawetz et al., 2016a). Nevertheless, it is still unclear whether the observed patterns of activation for emotion regulation are dependent on the nature of the stimuli. Our study also explicitly addressed this question.
In summary, the present study extends previous meta-analyses in several aspects: First, we compared for the first time different emotion regulation strategies on a meta-analytic level. Second, we considered both the up- and down-regulation of emotional responses. Third, stimulus-specific effects on emotion regulation were tested directly using a meta-analytic approach. Moreover, in contrast to previous meta-analyses on emotion regulation, we used contrast and conjunction analyses to test for common and distinct activation patterns directly. We used an activation-likelihood-estimation (ALE) meta-analysis approach, which is based on the distribution of reported activation peaks in the brain (Eickhoff et al., 2012).
Section snippets
Selection criteria for data used
Literature research was conducted using PubMed (www.pubmed.com) searching for combinations of keywords: “emotion regulation”; “affective regulation”; “implicit emotion regulation”; “explicit emotion regulation”; “interpersonal emotion regulation”; “extrinsic emotion regulation; ‘intrinsic emotion regulation’; “reappraisal”; “suppression”; “distraction”; “detachment”; “labelling”; “affective labelling”; “reinterpretation”; “rumination”; “fMRI”; “neuroimaging”; “functional magnetic resonance
Reappraisal vs. control condition
Areas demonstrating greater activations for reappraisal compared to a control condition are illustrated in Fig. 1A and reported in Table 2. Prefrontal areas including bilateral IFG/VLPFC, bilateral SFG/DLPFC, and DMPFC were activated during reappraisal compared to a control condition. Left temporal (STG, MTG) and bilateral parietal regions (inferior parietal lobule) as well as the supplementary motor area (SMA) and preSMA also showed convergent activity during reappraisal compared to a control
Discussion
In the current meta-analyses, we demonstrate concordance across a large number of studies and reveal the common and distinct patterns of brain activity by different aspects of emotion regulation. In a data-driven fashion, we pooled over all coordinates from different experiments of 93 studies on the cognitive control of emotions and summarize the results with respect to emotion regulation strategy, emotion regulation goals, and emotion inducing stimulus material.
Conclusion
In this study, we investigated the neural substrates of the cognitive control of emotions via coordinate-based ALE meta-analyses of brain activity reported for emotion regulation tasks using a variety of emotion regulation strategies, different emotion regulation goals and various emotion inducing stimuli. Our study provides evidence for a pivotal role of the left VLPFC, the anterior insula and the SMA in emotion regulatory processes. These were the only regions consistently activated in all
Conflict of interest
The authors declare no competing financial interests.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References (152)
- et al.
Emotion-regulation strategies across psychopathology: a meta-analytic review
Clin. Psychol. Rev.
(2010) - et al.
Inhibition and the right inferior frontal cortex
Trends Cogn. Sci.
(2004) - et al.
Three key regions for supervisory attentional control: evidence from neuroimaging meta-analyses
Neurosci. Biobehav. Rev.
(2015) - et al.
Common and differential neural networks of emotion regulation by Detachment, Reinterpretation, Distraction, and Expressive Suppression: a comparative fMRI investigation
Neuroimage
(2014) - et al.
Multidimensional assessment of empathic abilities: neural correlates and gender differences
Psychoneuroendocrinology
(2010) - et al.
Fear is only as deep as the mind allows: a coordinate-based meta-analysis of neuroimaging studies on the regulation of negative affect
Neuroimage
(2011) - et al.
The role of the human ventral striatum and the medial orbitofrontal cortex in the representation of reward magnitude – an activation likelihood estimation meta-analysis of neuroimaging studies of passive reward expectancy and outcome processing
Neuropsychologia
(2012) - et al.
A core system for the implementation of task sets
Neuron
(2006) - et al.
A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data
Neuroimage
(2005) - et al.
Testing anatomically specified hypotheses in functional imaging using cytoarchitectonic maps
Neuroimage
(2006)