Feelings of shame, embarrassment and guilt and their neural correlates: A systematic review

Accepted Manuscript Title: Feelings of shame, embarrassment and guilt and their neural correlates: A Systematic Review Author: Coralie Bastin Ben J. Harrison Christopher Davey Jorge Moll Dr. Sarah Whittle PII: DOI: Reference:

S0149-7634(15)30287-6 http://dx.doi.org/doi:10.1016/j.neubiorev.2016.09.019 NBR 2605

To appear in: Received date: Revised date: Accepted date:

15-11-2015 19-8-2016 23-9-2016

Please cite this article as: Bastin, Coralie, Harrison, Ben J., Davey, Christopher, Moll, Jorge, Whittle, Sarah, Feelings of shame, embarrassment and guilt and their neural correlates: A Systematic Review.Neuroscience and Biobehavioral Reviews http://dx.doi.org/10.1016/j.neubiorev.2016.09.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Feelings of shame, embarrassment and guilt and their neural correlates: A Systematic Review

Coralie Bastin1, Ben J. Harrison1, Christopher Davey1,2, Jorge Moll3, Sarah Whittle1,*

1

Melbourne Neuropsychiatry Centre, Department of Psychiatry, The University of Melbourne and

Melbourne Health, Melbourne, VIC, Australia
 2

Orygen, The National Centre of Excellence in Youth Mental Health, and Centre for Youth Mental Health,

The University of Melbourne, Melbourne, VIC, Australia 3

Cognitive and Behavioral Neuroscience Unit, D’Or Institute for Research and Education (IDOR), Rio de

Janeiro, Brazil

*

Corresponding author: Dr Sarah Whittle, Melbourne Neuropsychiatry Centre - Department of Psychiatry,

The University of Melbourne, Alan Guilbert Building (level 3), 161 Barry Street, 3053 VIC, Melbourne, Australia. Fax: +61 3 9342 3743.

Highlights 

We reviewed the literature on the neural correlates of shame, embarrassment and guilt.



Research points to both common and unique neural underpinnings of these emotions.



All three emotions appear to be associated with anterior insula activity.



Varying methodologies to induce emotion prevent strong conclusions to be made.

1

Abstract This systematic review aimed to provide a comprehensive summary of the current literature on the neurobiological underpinnings of the experience of the negative moral emotions: shame, embarrassment and guilt. PsycINFO, PubMed and MEDLINE were used to identify existing studies. Twenty-one functional and structural magnetic resonance imaging and positron emission tomography studies were reviewed. Although studies differed considerably in methodology, their findings highlight both shared and distinct patterns of brain structure/function associated with these emotions. Shame was more likely to be associated with activity in the dorsolateral prefrontal cortex, posterior cingulate cortex and sensorimotor cortex; embarrassment was more likely to be associated with activity in the ventrolateral prefrontal cortex and amygdala; guilt was more likely to be associated with activity in ventral anterior cingulate cortex, posterior temporal regions and the precuneus. Although results point to some common and some distinct neural underpinnings of these emotions, further research is required to replicate findings.

Keywords: Negative moral emotions, guilt, shame, embarrassment, neuroimaging, MRI, fMRI, PET

2

1. INTRODUCTION Humans experience moral emotions as early as two years of age (Barrett et al., 1993). The onset of this ability coincides with the development of self-evaluative processing, including being able to distinguish self from others. They can be either positive (e.g., pride and gratitude) or negative (e.g., shame, embarrassment and guilt) and most of these emotions can be considered ‘pro-social’ as they tend to promote adaptive social behaviors (Bowles and Gintis, 2005; Tangney et al., 2011). Negatively valenced ‘self-blaming’ moral emotions are particularly important for social functioning. These emotions are crucial for the development and maintenance of interpersonal relationships because they act as important social regulators by encouraging a balance between the individual’s urges and the rights and needs of others. Further, dysregulation in the experience of these emotions may lead to poor mental health. For example, it has been suggested that the excessive experience of self-blaming emotions like shame and guilt may have particularly adverse consequences in the realm of mood and anxiety disorders, such as depression (Andrews, 1995; Kim et al., 2011; O'Connor et al., 1999; Tangney et al., 1992a). While a great deal of research has been conducted investigating the neural correlates of basic emotions (e.g., anger, fear, sadness, happiness), far less research has investigated the neural basis of the experience of the negative moral emotions, shame, embarrassment and guilt. Such research might be particularly useful for refining current conceptualizations of common and distinct features of these emotions. Indeed, shame, embarrassment and guilt have a number of broad characteristics in common; all three emotions occur when the rules, norms or social agreements, defining what is right or wrong, are broken. Conversely, these emotions are suggested to have distinct features. A generally accepted differentiation between shame and guilt, proposed by Tangney and colleagues, is that shame is associated with internal attributions while guilt is associated with behavioral attributions (Tangney et al., 1996). Guilt is generally referred to as ‘behavior-focused negative self-conscious emotion’ (Tangney et al., 2011). With guilt, the focus is on the ‘do’ (e.g., “I did something wrong”) (Tangney, 1995). Guilt may be associated

3

with attributing a transgression, of social or inner moral norms, to external (rather than intrinsic) factors. Furthermore, the feeling of guilt often generates a sentiment of remorse, a desire to have behaved differently regarding the transgressed social norm, or a need to make up for a fault by confessing, taking reparative action (Tangney et al., 2007), or employing other methods for releasing guilty feelings (Lindsay-Hartz, 1984; Tangney et al., 2011). Guilt, more so than shame, may imply empathy toward others and a real concern about acting badly and hurting someone in the process (Tangney, 1992; Tangney et al., 2007). While a guilty person may feel emotional ‘pain’, it will unlikely become overwhelming (Tangney et al., 2011). Shame, on the other hand, is generally referred to as ‘self-focused negative emotion’ (Tangney et al., 2011), and is experienced when a person believes that their transgression of certain rules defines who they are (Wong and Tsai, 2007). Shame is related to the way we perceive ourselves and how we believe others see us, and our failure or inadequacy to fulfill the desire to be a good person (e.g., “I am a bad person for lying to my friend”). People who experience shame may be concerned with their own evaluation and what others might think of them (Tangney et al., 1992b), even when experienced alone. Since shame refers to the entire self’s malfunction, negatively judging or questioning oneself as a person will likely directly affect one’s core identity (Wong and Tsai, 2007). Thus, feeling ashamed induces the sentiment of worthlessness, inferiority and incompetence, and often leads to a want to escape and withdraw socially (Tangney et al., 2011). As such, by directly affecting the self, shame is thought to generate greater pain and be a more distressing experience than guilt. As alluded to, shame and guilt have a different relationship to empathy, while guilt is associated to “other-oriented empathy”, shame’s connection to empathy appears to be disrupted and focuses on “self-oriented distress” (Tangney et al., 2007). While Tangney’s operationalization of shame and guilt is popular, the similarities and differences between these emotions remains a topic of debate (Pulcu et al., 2013; Tangney et al., 1996; Wong and Tsai, 2007), especially between different fields (i.e., criminology, social and clinical psychology, philosophy, etc.) (Tangney et al., 2007; Tibbetts, 2003). In contrast to the

4

above definitions, O’Connor (O'Connor et al., 1999) for example defines several classes of guilt as involving characterological self-blame. Embarrassment has long been conceived as a dimension of shame (Kaufman, 1989; Lewis, 1971), generally assumed to vary on a range of factors including intensity, public exposure, and physical reaction (e.g., blushing). However, in more recent years, embarrassment has been suggested to be considered as a distinct emotional response (Tangney et al., 1996). Compared to shame and guilt, embarrassment appears to be associated with more sudden and accidental violation of social conventions with a motivational response directed towards the preservation of one’s social reputation, rather than a concern for others’ wellbeing and a need to make amends as in guilt or with a concern for oneself with a need to hide as in shame (Eisenberg, 2000; Tangney et al., 1996). Embarrassment is always directly related to the response of the presence of an “audience” (real or imagined), in which the person worries about their social image as a result of their behavior being directly witnessed. Embarrassment appears to be less negative, generating less emotional pain, only affecting one’s presented self, and can therefore be perceived as a less ‘damaging’ emotion by playing a more adaptive role than shame in social interactions. As such, although there is some debate about their differences (Tangney et al., 1996), shame, embarrassment and guilt are theoretically separable constructs that can be differentiated on the basis of factors such as interpersonal context of experience, appraisal and resulting action, behavioral or situational emotional trigger and coping mechanisms, etc. (Tangney et al., 1996). While shame, embarrassment and guilt are emotions often used interchangeably in colloquial terms, and empirically, measurement of these emotions as separate constructs is more difficult (Gibson, 2013), recent neuroimaging research has attempted to differentiate them in order to understand their distinct neurobiological correlates. There is now growing evidence that shame and guilt can be partly differentiated in terms of their primary brain structural and functional correlates. However, this literature has not been summarized to date. As such, the intention of this paper is to conduct a systematic review in order to reconcile the available brain imaging literature

5

on the subject to date. The aim of the review is to a) provide a descriptive overview of the current literature on the neural correlates of the experience of shame, guilt and embarrassment, b) provide a preliminary quantitative analysis of the common and unique neural correlates of these emotions, c) critically appraise study methods and their potential impact on results, and d) highlight gaps in the literature for future research.

2. METHODS This literature review of neuroimaging studies was conducted under the PRISMA guidelines. Figure 1 presents the PRISMA Flow Diagram detailing the different phases of the literature search (Liberati et al., 2009). Studies were included if they reported the direct association between brain structure or function and the experience of guilt, shame and/or embarrassment. Studies were included if the experience of these emotions was inferred by the authors (based on theoretical reasoning), based on participant self-report or observational techniques. Studies employing structural magnetic resonance imaging (sMRI), functional magnetic resonance imaging (fMRI) or positron emission tomography (PET) were included. Studies including either or both healthy and psychiatric participants were included. Only studies with adult population were included. Furthermore, only studies using human subjects and written in English were included. Document types were limited to peer-reviewed journal articles. Studies were excluded if they did not investigate (or purport to investigate) experiential aspects of shame, embarrassment or guilt (e.g., studies were excluded if they investigated decision-making or other cognitive processes associated with moral emotions). Studies referring more

broadly

to

‘self-blaming

emotions’

or

‘negative

moral

emotions’

without

specifying/measuring shame, embarrassment or guilt, specifically, were excluded. In addition, in order to facilitate comparison across studies, studies whose subjects had neurodegenerative disorders or focal brain lesions were also excluded. Exceptions were made, however, for studies including patients with neurodegenerative disorders if the findings were reported separately for a

6

healthy control group. Finally, and again to facilitate comparison across studies, studies were excluded if they only reported associations between moral emotions and functional connectivity. Studies that were published or in press as of October 2014 were retrieved in from three databases: PsycINFO (American Psychological Association), PubMed (US National Library of Medicine National Institute of Health) and MEDLINE (Wolters Kluwer Health OvidSP). The database searches were conducted by combining keywords through Thesaurus (PsycINFO), MeSH term (PubMed), Search Tools (MEDLINE), and free text. While some of the papers mentioned in their abstract investigating specific emotions such as guilt, shame and/or embarrassment, they did not always report specific results related to them and were therefore excluded during the full text stage of the review. Details of study screening and eligibility assessment are presented in more detail in Figure 1. -- Insert Figure 1 -As a result of discrepancies and inconsistencies in the way that similar brain regions were labeled across studies, peak voxel locations (for each cluster reported in each study) were initially checked and labeled using the Anatomical Automatic Labeling (AAL) atlas in SPM8, and in the case of any ambiguities, a consensus was reached by three of the authors (CB, SW, BH). Further, if necessary, regions were relabeled to present a more unified scheme. For instance, brain regions reported as medial orbitofrontal cortex (mOFC) were relabeled to ventromedial prefrontal cortex (vmPFC). A description of the anatomical relabeling employed in this review can be found in the legend of Figure 3.

3. RESULTS The first brain imaging study of negative moral emotions was published fifteen years ago, and since there has been a steadily growing interest in the neurobiological underpinning of these emotions. Twenty-one studies were included in this review spanning the years 2000-2014. Search results were screened independently by two co-authors (CB and SW) and any uncertainties about

7

inclusion/exclusion criteria were discussed until consensus was reached. Table 1 summarizes the main characteristics of each study in terms of methodology, sample and results. It is of note that while studies often investigated several emotions, for clarity, the results are discussed for each emotion under separate headings. -- Insert Table 1 -3.1 Methodology Studies differed substantially in terms of the methodologies and statistical contrasts/analyses used. Differences are summarized briefly here, but are also mentioned in sections 3.2 to 3.4, below. Only three fMRI studies reported using imaging sequences optimized for coverage of areas where signal loss is common (i.e., ventral temporal lobe and orbitofrontal cortex) (Moll et al., 2007; Morey et al., 2012; Zahn et al., 2009). Upon inspection, however, it does not appear that the three studies using optimized sequences reported more activity in these regions as compared to other studies (see Figure 2). All but one study (Zahn et al., 2014) presented findings from whole brain analyses, although many used a combination of whole brain and region of interest (ROI) analyses. The majority of fMRI/PET studies employed script driven tasks (8/19) with instructions for participants to imagine themselves in the situation described (or remembered, in the case of autobiographical recall paradigms) and including ratings of stimuli or own affect (6/19). Five other studies involved other types of participation such as photo-driven tasks involving active rating of stimuli (3/19) and two studies employed more active participation by the using a game or combining the use of words and faces. An assessment of results based on the different induction techniques used in studies was conducted and the induction method did not seem to affect the results (see Figure S1 in supplementary material). In terms of statistical analysis, eight studies reported brain activity associated with one of the target negative moral emotions compared to a neutral (non-emotional) condition (Fourie et al., 2014; Michl et al., 2014; Moll et al., 2007; Morey et al., 2012; Roth et al., 2014; Shin et al., 2000; Takahashi et al., 2004; Wagner et al., 2011). Four studies reported brain activity associated with one of the target negative moral emotions compared to at least one other of these emotions (Michl et al., 2014;

8

Pulcu et al., 2014; Takahashi et al., 2004; Wagner et al., 2011). Five studies reported brain activity associated with one of the target negative moral emotions compared to another negative moral emotion (e.g., guilt > indignation) (Basile et al., 2011; Green et al., 2012; Kédia et al., 2008; Morey et al., 2012; Wagner et al., 2011), three studies reported findings using a ‘basic’ negative emotion contrast condition (Basile et al., 2014; Basile et al., 2011; Wagner et al., 2011), and three studies reported contrasting a target negative moral emotion with either a positive moral emotion (e.g., pride) (Fourie et al., 2014; Roth et al., 2014) or a matched ‘social’ condition (Finger et al., 2006). Finally, ten studies reported correlations between individual differences in a target negative moral emotion and either brain function or structure (Fourie et al., 2014; Morey et al., 2012; Morita et al., 2008; Morita et al., 2012; Morita et al., 2013; Pulcu et al., 2014; Sturm et al., 2012; Yu et al., 2014; Zahn et al., 2014; Zahn et al., 2009). Arguably, the most robust approach to support specific shame-, embarrassment- and guilt-related brain activity is to compare these emotions with one another. However, only four studies reported such contrasts in healthy populations (Michl et al., 2014; Pulcu et al., 2014; Takahashi et al., 2004; Wagner et al., 2011). Findings from these studies are illustrated in Figure S2 in the supplementary material and discussed below in the sections describing shame-, embarrassment-, and guilt-related brain activity compared to a specific emotion control condition. Note that the results of these studies are included in Figures 2 and 3.

3.2. The Neurobiology of Shame Only four functional studies specifically reported shame related brain activation: three using fMRI in healthy populations (Michl et al., 2014; Roth et al., 2014; Wagner et al., 2011) and one looking at patients with remitted Major Depressive Disorder (MDD) (Pulcu et al., 2014). Michl et al., (2014) used a paradigm involving the presentation of short sentences designed to evoke shame and other emotions (adapted from Takahashi et al., 2004). Roth et al., (2013) and Wagner et al. (2011) both used an autobiographical recall task where participants were instructed to remember personal events that elicited shame among other emotions.

9

3.2.1 Shame Compared to a Neutral Control Condition In these studies, shame compared to a neutral condition, was associated with increased activation in prefrontal areas including the dorsolateral prefrontal cortex (dlPFC) (Michl et al., 2014; Roth et al., 2014), the ventrolateral prefrontal cortex (vlPFC) (Roth et al., 2014) and the dorsomedial prefrontal cortex (dmPFC) (Michl et al., 2014; Wagner et al., 2011). Increased activation was also found in the anterior inferior temporal (aITC), the anterior middle temporal (aMTC) (Wagner et al., 2011) anterior superior temporal cortices (aSTC) (Michl et al., 2014), and anterior insula cortex (AIC) (Michl et al., 2014; Roth et al., 2014; Wagner et al., 2011). In addition, Wagner et al., (2011) found shame to be associated with increased activation in parietal regions including the angular gyrus and the precuneus whereas Michl et al., (2014) and Roth et al., (2013) reported activation in the posterior cingulate cortex (PCC) and subcortical involvement, including the parahippocampal gyrus (Michl et al., 2014), basal ganglia (pallidum and caudate head; Roth et al., 2014). Increased brain activation was also reported in sensorimotor areas (Michl et al., 2014; Roth et al., 2014), the visual cortex (Michl et al., 2014; Wagner et al., 2011) and the cerebellum (Michl et al., 2014; Roth et al., 2014; Wagner et al., 2011). Further Roth et al., (2014) reported shame to be associated with decreased activity in the superior parietal cortex (SPC) and supramarginal gyrus. 3.2.2 Shame Compared to a (Basic or Moral) Negative Emotion Control Condition Three studies compared shame to another (basic or moral) negative emotion in healthy populations (Michl et al., 2014; Pulcu et al., 2014; Wagner et al., 2011), but only two studies reported significant shame-related activation (Michl et al., 2014; Wagner et al., 2011). Wagner et al., (2011) reported increased activation in the primary motor cortex (precentral gyrus) when comparing shame to sadness but no significant effect when comparing shame to guilt. Michl et al., (2014), on the other hand, found increased activation for shame compared to guilt in the dlPFC, the AIC, the dorsal anterior cingulate cortex (dACC), the PCC, and in subcortical areas including the parahippocampal cortex and the dorsal anterior midbrain. While the current neuroimaging on shame in healthy individuals is still extremely limited,

10

it is notable that among the four studies reported, one investigated the neural correlates of individual differences in shame and reported a positive correlation of unpleasantness ratings for shame trials with posterior insula cortex (PIC) signal for shame versus guilt in healthy individuals (Pulcu et al., 2014). 3.2.3 Psychiatric Population Findings With regard to results from studies of psychiatric populations, Pulcu et al., (2014) examined shame in remitted MDD participants using a script based paradigm consisting of statements describing actions counter to social/moral standards, and requiring participants to rate their level of unpleasantness experienced. When looking at the effects of shame (shame > guilt trials based on individual post-scan classification), only in the remitted-MDD group was shame associated with activation in the right amygdala and right posterior insula cortex, whereas in the control group there were no significant effects (Pulcu et al., 2014).

3.3. The Neurobiology of Embarrassment More research has examined the neurobiological underpinning of embarrassment than shame. Embarrassment is arguably a more readily grasped moral emotion: it is a self-conscious emotion directly related to the presence of a public audience, and has been investigated in six fMRI studies to date. Five studies used healthy subjects (Finger et al., 2006; Moll et al., 2007; Morita et al., 2008; Morita et al., 2013; Takahashi et al., 2004) and one investigated the neural correlates of embarrassment in high-functioning Autism Spectrum Disorder (ASD) (Morita et al., 2012). In addition, structural MRI has been used to investigate embarrassment in patients suffering from behavioral variant frontotemporal dementia (bvFTD) and controls (Morita et al., 2012). The majority of these studies used statements or short scenarios to induce embarrassment. Takahashi et al., (2004) used three types of short sentences as stimuli designed to trigger embarrassment and other feelings. Finger et al., (2006) used written short stories depicting social situations moderated by the presence vs. absence of an audience to trigger embarrassment, and

11

Moll et al., (2007) used written scripts in a moral sentiment task to elicit the emotional experience of embarrassment. Three studies from the same group used facial feedback images of the self and others to induce embarrassment (Morita et al., 2008; Morita et al., 2012; Morita et al., 2013). Finally, one study induced embarrassment using video recordings of a karaoke singing task (Sturm et al., 2012). 3.3.1 Embarrassment Compared to a Neutral Control Condition Out of the six fMRI studies, two looked at embarrassment in relation to a neutral condition and found embarrassment to be associated with increased brain activation in the dmPFC (Moll et al., 2007; Takahashi et al., 2004), the vlPFC (Takahashi et al., 2004) , the dACC (Moll et al., 2007), the AIC (Moll et al., 2007), the anterior temporal lobes (ATL: aITC and aMTC; Takahashi et al., 2004), the posterior superior temporal sulcus (STS) (Moll et al., 2007; Takahashi et al., 2004) and the temporoparietal junction (TPJ) (Moll et al., 2007). In addition, Takahashi et al., (2004) also found increased brain activation in the left hippocampus and visual cortex (calcarine and lingual gyrus). 3.3.2 Embarrassment Compared to an Emotion Control Condition Two studies investigated embarrassment-related brain activity when compared to another emotion (Finger et al., 2006; Takahashi et al., 2004). Comparing embarrassment to guilt, increased brain activation was reported in the dmPFC, aMTC, bilateral hippocampus, dorsal midbrain and visual cortex (calcarine, lingual gyrus and inferior occipital gyrus; Takahashi et al., 2004). Comparing an embarrassment condition to a matched non-embarrassment social condition, Finger et al., (2006) found embarrassment to be associated with activation in the dmPFC, vlPFC and the amygdala. 3.3.3 Individual Differences in Self-reported Embarrassment Two studies by Morita and colleagues looked at individual differences in self-reported embarrassment in healthy populations (Morita et al., 2008; Morita et al., 2013). While Morita et al., (2013) showed a positive correlation between self-reported embarrassment and activity in the

12

AIC, Morita et al., (2008) found the opposite pattern with a negative correlation between AIC activity and embarrassment. In addition, Morita et al., (2008) reported self-reported embarrassment to be negatively correlated with activity in the dlPFC. 3.3.4. Psychiatric Population Findings Only one study examined the emotional experience of embarrassment in psychiatric population. Using the same paradigm as in studies described above (Morita et al., 2008; Morita et al., 2013), Morita et al., (2012) found embarrassment to be positively correlated with activity in the vlPFC and AIC in individuals with Autism Spectrum Disorder (ASD). 3.3.5 Structural MRI findings One sMRI study to date has investigated the neuroanatomical correlates of embarrassment. Sturm et al., (2013) induced embarrassment using a karaoke singing task and found that across both healthy controls and patients with frontal-temporal dementia (FTD), as well as within each group, larger right pregenual anterior cingulate cortex (pACC) gray matter volume was associated with higher levels embarrassment observed during the task. 3.4. The Neurobiology of Guilt Fifteen studies have investigated the neurobiological correlates of guilt, with one using PET, thirteen using fMRI, and one combining fMRI and sMRI. The most common methodology for eliciting guilt in participants was to use emotion induction statements (Finger et al., 2006; Green et al., 2012; Kédia et al., 2008; Michl et al., 2014; Moll et al., 2007; Morey et al., 2012; Pulcu et al., 2014; Takahashi et al., 2004; Zahn et al., 2009). Other studies used alternate methods such as emotion induction statements paired with facial expressions (Basile et al., 2014; Basile et al., 2011), autobiographical recall (Shin et al., 2000; Wagner et al., 2011), an interpersonal game paradigm (Yu et al., 2014), and an implicit-association test (IAT) (Fourie et al., 2014). The latter was a dual categorization task involving feedback indicating that the participant was prejudiced based on their performance. While most of these studies looked at healthy controls, some looked

13

at mental disorders such as remitted MDD and Obsessive Compulsive Disorder (OCD). 3.4.1 Guilt compared to a Neutral Control Condition Studies have found the experience of guilt to be associated with function and structure of prefrontal, temporal, cingulate, and subcortical brain regions. Studies examining guilt in relation to a neutral condition found increased guilt-related activity in prefrontal areas of the brain (dmPFC; Fourie et al., 2014; Michl et al., 2014; Moll et al., 2007; Takahashi et al., 2004; Wagner et al., 2011) (vlPFC; Moll et al., 2007; Wagner et al., 2011) (dlPFC; Michl et al., 2014; Wagner et al., 2011). Activation in the temporal lobes was also commonly associated with guilty feelings, with studies reporting increased activation in anterior (Michl et al., 2014; Moll et al., 2007; Shin et al., 2000; Wagner et al., 2011) and posterior (Michl et al., 2014; Moll et al., 2007; Takahashi et al., 2004; Wagner et al., 2011) temporal regions. Increased guilt-related brain activation was also reported in a number of studies in both the anterior (AIC; Fourie et al., 2014; Shin et al., 2000) and posterior (PIC; Wagner et al., 2011) regions of the insula cortex. Increased guilt-related brain activation was reported in a number of cingulate regions including the dACC (Fourie et al., 2014; Shin et al., 2000), middle cingulate cortex (MCC; Shin et al., 2000), and dorsal PCC (Fourie et al., 2014) and parietal regions including the supramarginal gyrus (Wagner et al., 2011), angular gyrus and parietal operculum (Wagner et al., 2011), and precuneus (Fourie et al., 2014; Wagner et al., 2011). An increase in brain activation in relation to guilt was also reported in several subcortical regions including the thalamus (Fourie et al., 2014), parahippocampal gyrus (Wagner et al., 2011), dorsal midbrain (Moll et al., 2007), and basal ganglia (Wagner et al., 2011). Finally, activation in several areas of the sensorimotor cortex (Michl et al., 2014; Moll et al., 2007; Shin et al., 2000; Wagner et al., 2011), the visual cortex (Michl et al., 2014; Takahashi et al., 2004; Wagner et al., 2011) and the cerebellum (Michl et al., 2014; Shin et al., 2000; Wagner et al., 2011) were associated with guilt. In addition, guilt was associated with decreased activity in the PIC, precuneus, sensorimotor cortex (precentral gyrus) and visual cortex (Shin et al., 2000). 3.4.2 Guilt Compared to an Emotion Control Condition

14

Thirteen studies investigated the experience of guilt in relation to other emotions, or investigated the experience of specific types of guilt. Some studies explored guilt by contrasting it to the experience of shame (Michl et al., 2014; Pulcu et al., 2014; Wagner et al., 2011), embarrassment (Takahashi et al., 2004), ‘basic negative’ emotions (e.g., anger, sadness) (Basile et al., 2011; Kédia et al., 2008; Wagner et al., 2011), indignation (Green et al., 2012), and indignation and pride (Zahn et al., 2009). Less frequently, guilt conditions were contrasted with other conditions that were designed to eliminate variance associated with the experience of moral emotions more generally (Fourie et al., 2014; Yu et al., 2014), or specific types of guilt were contrasted, such as guilt related to others versus the self (Morey et al., 2012) and deontological guilt versus altruistic guilt (Basile et al., 2011). Findings from these studies showed guilt-specific activation in the dmPFC (Basile et al., 2011; Finger et al., 2006; Fourie et al., 2014; Morey et al., 2012; Takahashi et al., 2004; Wagner et al., 2011), as well as other areas of the prefrontal cortex including the vlPFC and dlPFC (Wagner et al., 2011). Activation in cingulate regions were also commonly associated with guiltspecific feelings, especially in the dACC (Basile et al., 2011; Green et al., 2012; Wagner et al., 2011) and ventral ACC (Basile et al., 2011), but also in the MCC (Yu et al., 2014) and PCC (Fourie et al., 2014). A number of studies also reported an increase in activation in the AIC (Basile et al., 2011; Fourie et al., 2014; Wagner et al., 2011; Yu et al., 2014) during the experience of guilt. In terms of temporal lobe, activation in the posterior MTC (Finger et al., 2006; Green et al., 2012; Kédia et al., 2008; Michl et al., 2014; Morey et al., 2012) appeared to be most frequently reported in relation to guilt, with some studies also reporting increased brain activation in the anterior MTC (Green et al., 2012; Wagner et al., 2011), anterior STC (Finger et al., 2006), posterior STS (Fourie et al., 2014) and TPJ (Kédia et al., 2008). Increased activation in the precuneus (Basile et al., 2011; Kédia et al., 2008; Morey et al., 2012) was also commonly associated with guilty feelings. In addition, increased guilt-related activation was reported in several subcortical regions including the thalamus (Michl et al., 2014), the amygdala (Wagner et al., 2011) and basal ganglia (Fourie et al., 2014; Wagner et al., 2011). Finally, guilt-related activation was also reported in areas of the sensorimotor cortex (Green et al., 2012; Kédia et al.,

15

2008) the visual cortex (Green et al., 2012; Michl et al., 2014; Morey et al., 2012) and the cerebellum (Wagner et al., 2011). 3.4.3 Individual Differences in Self-reported Guilt Three studies investigated individual differences in self-reported guilt in healthy individuals. The experience of guilt was found to be positively correlated with activity in the ventromedial PFC (Zahn et al., 2009), dmPFC and vlPFC (Morey et al., 2012), the vACC (Zahn et al., 2009) and the inferior parietal cortex (IPC; Morey et al., 2012). On the other hand, guilt ratings were also found to be negatively correlated with activity in the dlPFC and vlPFC (Morey et al., 2012), the MCC (Morey et al., 2012) and vACC (Fourie et al., 2014), the PIC, the posterior MTC, the SPC and precuneus, as well as the postcentral gyrus cortex and cuneus (Morey et al., 2012). 3.4.4. Psychiatric population findings With regard to results from studies of psychiatric populations, three studies examined guilt relationship to specific mental disorders (Basile et al., 2014; Green et al., 2012; Pulcu et al., 2014). Both Pulcu et al., (2014) and Green et al., (2012) examined guilt in remitted MDD patients using a script based paradigm consisting of statements describing actions counter to social/moral standards, and requiring participants to rate their level of unpleasantness experienced. The third study, by Basile et al., (2014), investigated the experience of guilty feelings in OCD patients using an emotion induction statements paired with facial expressions where they were instructed to imagine oneself in a situation and rate their level of guilt experienced. One of these studies reported significant results for guilt across patient and control groups (remitted MDD and HC: guilt > indignation) with increased activation in the ventral ACC and basal ganglia (Green et al., 2012). Reduced guilt-related activation was found for patients compared to controls in other regions. Basile et al. (2014) found reduced guilt-related activity in OCD compared to control participants in a range of regions (Basile et al., 2014) while Pulcu et al. (2014) appeared to find reduced guilt-related activity in the amygdala and PIC in remitted MDD patients compared to controls (Pulcu et al., 2014).

16

3.4.5 Structural MRI findings One sMRI study conducted to date reported a positive correlation between gray matter volume in the anterior MCC and sensitivity to guilty feelings (Yu et al., 2014). Figures 2 and 3 present a simplified summary of the results of the reviewed studies. -- Insert Figure 2 --- Insert Figure 3 --

4. DISCUSSION To date, the majority of brain imaging research investigating emotional processes has focused on basic emotions such as anger, sadness, fear and happiness. While this research has been useful in providing information about the neural substrates of these emotions in mental health and illness, neuroimaging research that aims to characterize the neural basis of more complex social emotions is paramount. The identified neuroimaging studies of self-blaming negative moral emotions differed in terms of methodologies and statistical analyses employed (discussed further below). Some studies reported brain function or structure that was associated with shame, embarrassment or guilt without testing for specificity (e.g., variance in brain function/structure associated with the experience of negative affect more generally was not controlled for). In order to provide a meaningful interpretation of the shared and unique neural correlates of shame, embarrassment and guilt, only results from analyses that attempt to test specificity (i.e., by controlling for the experience of another basic or moral emotion) are discussed below. Results in healthy controls are discussed in sections 4.1 and 4.2. Results in psychiatric populations are discussed in section 4.4.

4.1. Common and Distinct Neural Circuitry Underlying Self-blaming Negative Moral Emotions A number of brain regions were associated with more than one self-blaming emotion. Shame and guilt appeared to share an association with AIC function (see Figures 2b and 3). The insula cortex

17

is known to contribute to diverse mental functions such as motor control, interoceptive awareness and emotion. The role of the insula in emotion is broad, and the function of the insula have been linked to the experience of a number of positive and negative emotions. The insula is known to play a role in representing emotional states, especially in relation to processing subjective feelings, empathy and uncertainty (Singer et al., 2009). Furthermore, the AIC has been shown to be involved in the experience of interoceptive awareness (Craig, 2009). These functions are all likely to contribute to the reported role of the insula in negative moral emotions, generally. Additionally, there is evidence from lesion studies for the role of the AIC in disgust, and it has been suggested that disgust toward the self may be present in negative moral emotions (Calder et al., 2001). The association between AIC activity, shame and guilt (but not embarrassment) may echo previous association of the AIC’s sensitivity in relation to emotional and cognitive aspects of pain (Pavuluri and May, 2015; Wiech et al., 2014), and the definition of shame and guilt as more painful and damaging negative emotions compared to embarrassment (Tangney et al., 1996). Shame and guilt were also both associated with dACC function. Different ACC regions have previously been involved in a range of functions including the experience of negative affect (Mayberg et al., 2014), the experience of social pain (Eisenberger and Lieberman, 2004) and intereoceptive awareness (Khalsa et al., 2009). The known role of the dACC, in the experience of distress and particularly social pain (Masten et al., 2011) may explain its common role in the experience of these emotions. Shame and embarrassment were both associated with hippocampal and midbrain function. While the hippocampus has been associated with several brain functions such as memory, emotion processing and regulation of stress (Brody et al., 2001; Morgane et al., 2005), the report of hippocampus-related activation with both shame and embarrassment is interesting. The role of the hippocampus in responding to psychosocial stress (Dranovsky and Hen, 2006; McEwen, 2001) may be particularly important for the experience of shame and embarrassment given that both of these emotions (perhaps more so than guilt) are associated with threat from the external environment to the self (Tangney, 1996; Tangney et al., 1992b).

18

Embarrassment and guilt were both associated with dmPFC, vlPFC and anterior temporal lobe function. The dmPFC plays a role in self-referential processing (Northoff et al., 2006) and mentalizing, also commonly referred to as the Theory of Mind (ToM) and the ability to infer others’ feelings, thoughts and intentions (Isoda and Noritake, 2013; Kang et al., 2013). The aptitude to estimate others’ intentions may parallel the necessary skills to read social cues and emotional states of others in order to experience self-blaming emotions when social/moral rules are broken. ToM may play more of a role in embarrassment and guilt than shame, given that shame is less likely to be associated with perceiving others’ feelings and more likely to be associated with self-focused cognition (Wong and Tsai, 2007). The vlPFC has frequently been implicated in moral decision-making and disturbances (Harrison et al., 2012). The anterior temporal lobe is suggested to play a role in social conceptual knowledge, where it is required to understand social concepts and rules (Ross and Olson, 2010) and to be aware of situations that are likely to elicit moral emotions (Olson et al., 2007; Zahn et al., 2007). It is unclear why these regions may be specifically associated with guilt and embarrassment, but not shame. Given the few studies focusing on shame, further research is required to confirm any emotion specificity. Shame, embarrassment and guilt were also found to have some neurobiological specificity. From the small number of studies (i.e., four) directly comparing shame, guilt and embarrassment with at least one other of these emotions, there was only evidence for guiltspecific activity in the dmPFC, posterior MTC, amygdala, basal ganglia, thalamus, and cerebellum (see Figure S1 in supplementary material). However, the review of studies that attempted to assess specificity in some way (either by controlling for negative affect or moral affect), provide some evidence that these three emotions might have different underlying neural circuitry. Shame appeared to be uniquely associated with dlPFC, PCC and sensorimotor cortex function. Considering previous findings showing the role played by the lateral PFC in the engagement of both cognitive and emotions regulation processes (Gray et al., 2002), it might be hypothesized that shameful feelings may require “harder mental work” to regulate the associated

19

negative affect in comparison to guilt and embarrassment. This assumption is in line with the theoretical definition of shame describing it as a more damaging and distressful emotional experience. This explanation is consistent with the literature on depression (commonly associated with elevated shame, (Andrews, 1995; Andrews et al., 2002; Kim et al., 2011; Pulcu et al., 2014)) that shows greater lateral PFC activation during emotion regulation in depressed patients (Johnstone et al., 2007). Recent studies have demonstrated the importance of the PCC in selfrepresentation (Leech et al., 2011), a mental process related to moral reflection and again in line with the definition of shame as an emotion associated with strong concern with one’s own evaluation and a particular focus on the self (Tangney et al., 2011; Wong and Tsai, 2007).

A number of brain regions appeared to be ‘guilt-specific’ including the ventral (pregenual and subgenual) ACC, posterior temporal areas including the TPJ, the precuneus and premotor cortex. While activation of the vACC has long been associated with negative affect (Drevets et al., 2008), its role in emotion regulation has more recently been acknowledged. In particular, the vACC has been associated with the inhibition of emotions (particularly fear), and it has been suggested that its activation may be associated with emotion regulation by facilitating planning of adaptive response (Etkin et al., 2011). This is in line with theoretical definition of guilt describing it as an intropunitive emotion linked to the fear of punishment and the need for related-reparative actions. Further, guilt-specific association with the vACC is consistent with lesion studies showing that patients with vACC lesions experience reduced levels of guilt (Krajbich et al., 2009). Finally, posterior temporal regions, associated in previous studies with aspects of social cognitive and emotional processing (Decety and Lamm, 2007; Schilbach et al., 2008), were also reported to play a role in guilt. Guilt-specific activity in the TPJ and posterior MTC/STS is in line with the role of these regions in ToM, given that guilt has been particularly linked with otherprocessing including reading other’s state of mind/feelings/thoughts (Tangney, 1992; Tangney et al., 2007). The precuneus, in close proximity to the PCC, has been demonstrated in recent studies to play a role in episodic memory retrieval, self-awareness-processing (Cavanna and Trimble, 2006; Kjaer et al., 2002; Morey et al., 2012) and emotional judgments (Morey et al., 2012). Given

20

that shame was also found to be associated with precuneus activation (albeit only in studies that contrasted shame with a neutral condition), further research is required to understand whether and why precuneus function may be particularly linked to guilt experiences. Finally, embarrassment appeared to be uniquely associated with vlPFC, amygdala and occipital function. The function played by the vlPFC in emotional processing (Stuss and Levine, 2002) and behavioral self-regulation (Beer et al., 2006; Goldstein, 1944), as well as the ventral pathway connection of this region to the amygdala in the limbic system may explain the involvement of both these regions during the experience of embarrassment. The increased functional coupling between the ventral PFC and the amygdala during reappraisal of negative emotion showing to result in reduced reported negative emotion (Banks et al., 2007; Schardt et al., 2010), whereas successful emotion regulation, is in line with the theoretical definition of embarrassment perceived as a less damaging and more adaptive emotion than shame. Additionally, the motivational response when feeling embarrassed being directed to selfpreservation of one’s social reputation (Tangney et al., 1996) is consistent with the literature on lesion studies looking at the orbitofrontal cortex in the context of social behavior showing that damages to this region is associated with inappropriate social behavior (Beer et al., 2006). Shame on the other hand, as mentioned earlier, appeared to solicit a more dorsal pathway with shamespecific activation in the dlPFC and PCC. While both ventrolateral and dorsolateral PFC have previously been shown to play a role in processing negative emotions and reappraisal (Banks et al., 2007), the difference in PFC laterality may lie in the specific nature of the definition of shame and embarrassment, with shame being more a distressing experience by affecting the entire self and therefore potentially requiring more complex cognitive processing during reappraisal of negative affects.

4.2. Self-blaming Negative Moral Emotions and Brain Networks It may be of relevance to interpret the neural correlates of shame, embarrassment and guilt in terms of brain networks. Parallels can be drawn between the regions implicated in each emotion

21

and well-characterized brain networks including the default mode network (DMN), the ‘social brain’ network, and the mesolimbic network. The brain network of potentially most relevance is the DMN. The DMN demonstrates increased activity while an individual is at ‘rest’ compared to when performing demanding (and typically non-self-referential) cognitive tasks (Shulman et al., 1997). DMN areas include the PCC, dmPFC, inferior parietal cortices, precuneus (Buckner et al., 2008; Harrison et al., 2008) and the TPJ (Mars et al., 2012). While the DMN was first associated with passive or resting state tasks, it is now thought to play a role in self-referential mental processes, autobiographical memory, and thinking about the future. A number of studies have hypothesized about the importance of DMN function for moral decision-making (Greene et al., 2004; Greene et al., 2001; Harrison et al., 2008). Given these identified functions of the DMN, and the fact that many of the brain regions found to be associated with negative moral emotions overlap with DMN regions, it is reasonable to conclude that the DMN likely plays a role in the experience of guilt, shame, and/or embarrassment. The emotional experience of self-blaming emotion implies some sort of personal reflection and evaluation about the self, including reference to specific personal experiences; it takes into account social rules and moral expectations from others and society. Parallels can also be drawn between brain regions whose activation was found to be associated with self-blaming negative moral emotions and a network of brain regions identified to underlie social cognitive processes (i.e., the ‘social brain’). The expression “social brain” refers to a complex network of brain regions that are involved in social interaction and processing of information about others including mentalizing (Blakemore, 2008; Frith, 2007). The brain regions that have been found to play a role in social cognition include the medial PFC, the vmPFC, the ACC, the TPJ, the posterior STS, and the amygdala (Blakemore, 2008; Frith, 2007; Kennedy and Adolphs, 2012). The ability to be aware of one’s own emotional state and to differentiate it from others’ is an important skill for survival, allowing for improved interindividual interaction and communication, and has been referred to as the ToM or ‘mentalizing’ (Schilbach et al., 2008). In a majority of cases, negative moral emotions are experienced in response to a ‘wrong/amoral’

22

action that has affected others and/or witnessed by others, which therefore implies the agent’s ability to not only reflect on their own behavior and identity but to also be able to read and relate to others’ emotional states. The neural regions found to be involved in the experience of selfblaming emotions appear to overlap with several brain regions associated with social cognitive skills such as mentalizing (Zaki and Ochsner, 2012). Correspondences can also be drawn between brain regions whose activation was found to be associated with self-blaming negative moral emotions and the mesolimbic emotional network. This network comprises a set of subcortical regions that have been found to play a role in various functions such as motivation, memory and emotion processing and regulation, including mood and stress responsiveness (Brody et al., 2001; Morgane et al., 2005). The overlap between the brain regions reported in the literature on self-blaming emotions and those recognized as part of the mesolimbic network include the amygdala, the hippocampus, the hypothalamus, the cingulate gyrus, and areas of the basal ganglia, the insula and the parahippocampal gyrus. This parallel between the functions of some of the brain regions involved in the mesolimbic system and negative moral emotions may be due (as mentioned above) to the role played by the mesolimbic system in the processing and experience of negative emotional states (Kelley and Berridge, 2002). It is of note that some research (not included in the review) has investigated the neural basis of self-blaming negative emotions using network approaches (i.e., functional connectivity). Three studies by Green and colleagues investigated guilt-related functional connectivity when compared to indignation in healthy and remitted MDD populations (Green et al., 2013a; Green et al., 2012; Green et al., 2010). Green et al., (2010) reported selective functional integration between subgenual cingulate/septal regions (SCSR) and superior ATL in healthy control (HC) individuals. Green et al., (2012) reported between group (HC>remitted MDD) connectivity between right superior ATL seed region and SCSR, bilateral medial frontal cortex, lateral hypothalamus and hippocampus, as well as within group (HC) connectivity between the SCSR, hippocampus, hypothalamus, amygdala, dACC and occipital pole. Functional connectivity results specific to the MDD population did not show significant results (Green et al., 2013a; Green et al.,

23

2012). Based on these findings, future work should employ measures of connectivity to further elucidate the role of neural networks in the experience of shame, embarrassment and guilt.

4.3. Methodological considerations The above summary of the available research on the neural correlates of shame, embarrassment and guilt offers preliminary insight into the brain substrates underlying the experience of negative moral emotions that are crucial for adaptive social functioning and that are commonly compromised in mental ill-health. As a result of the limited number of studies and the wide variety of methodology used, it is difficult to compare studies and draw confident conclusions. It is arguably difficult to illicit such emotions in experimental situations. For example, the ability of participants to project themselves into the situation presented by script-driven scenarios is uncertain, and as such validity of findings must be questioned.. Although several studies collected post-scan (Basile et al., 2014; Basile et al., 2011; Finger et al., 2006; Green et al., 2012; Michl et al., 2014; Moll et al., 2007; Morita et al., 2008; Morita et al., 2012; Morita et al., 2013; Pulcu et al., 2014; Takahashi et al., 2004; Yu et al., 2014), or in-scanner (Basile et al., 2014; Basile et al., 2011) self-reported ratings of the emotions experienced during the experiment, others assumed that a specific emotion was elicited without validation with self-report measures. In some cases self-report of emotion was arguably redundant (e.g., autobiographical recall). However, in many cases, it could be argued that the reported brain activation was not necessarily associated with the emotion intended by the research. Much neuroimaging research has utilized moral dilemmas to investigate the neural correlates of moral decision-making processes (Casebeer and Churchland, 2003; Greene et al., 2004; Heekeren et al., 2005; Heekeren et al., 2003). Typically, such studies employ paradigms using moral dilemmas, also called ethical dilemmas, which are hypothetical situations that require a decision to be made between two options that are equally balanced in terms of negative and positive outcomes. The process of decision-making in these types of dilemmas generally evokes strong conflict regarding justice, ethicality and morality. Defining whether moral dilemmas have

24

a ‘right’ or ‘wrong’ answer can be a real philosophical debate depending on various factors, and where pros and cons can be found on both sides. Yet, it is this fragile balance between the two possible choices as well as the fact that there are always consequences (both positive and negative) of the decision made (Moll et al., 2005) that may suggest good grounds for using moral situations to build new innovative and controlled tasks to elicit self-blaming emotions such as guilt, shame and embarrassment. Whereas existing paradigms used to elicit moral emotions suffer from methodological limitations (as noted above), utilizing moral dilemmas in new paradigms could be useful as a tool to capture brain activity during the experience of moral emotions. Another challenge in terms of study comparison is the variety of statistical contrasts used by researchers to look at specific negative moral emotions. The majority of studies compared one single emotion to a neutral condition, which is problematic given that one cannot be certain that results do not reflect activation related to the experience of negative affect. Furthermore, while several studies contrasted one specific emotion to another emotion (e.g., guilt versus indignation) or group of emotions (e.g., guilt versus basic emotions), and these were the results that formed the basis for our discussion, the specific control emotions used in contrasts varied widely (e.g., indignation, self-anger, group of basic emotions). Finally, some studies used regression analyses to assess linear associations between emotional ratings and brain function/structure, adding again to the complexity of comparing results between studies. There is therefore a need for replication and greater consistency between study methods. Ultimately, the definitions of these moral emotions have long been debated, and different studies may have drawn on different theoretical descriptions when designing stimuli. For example, emotions labeled as embarrassment in some work (Takahashi et al., 2004) were labeled as shame in other work (Michl et al., 2014). Further, the definitions of emotions were sometimes unclear. Thus, future work should be clear about the definitions of the emotions that they aim to probe, in addition to being clear about the rationale for using a specific paradigm to tap into those emotions. 4.4. Implications for psychopathology

25

While understanding the neurobiological underpinnings of self-blaming emotions in healthy individuals is paramount and there is still an undeniable need for more studies in this area, there is also a need for more neuroimaging research that investigates the neural correlates of negative moral emotions in individuals with mental health disorders, such as depression, anxiety and psychopathy. To date, the majority of brain imaging research examining shame, embarrassment and guilt has focused on healthy populations with only 4 studies including psychiatric populations. Among these, two studies examined shame and/or guilt in remitted MDD; one study reported a group by emotion (guilt, shame) effect for activation in the amygdala and posterior insula cortex (Pulcu et al., 2014). The second study found reduced guilt-related activity in patients in the PIC and parietal-occipital junction (Green et al., 2012). Another study examined individual differences in self-reported embarrassment in individuals with ASD and found a positive correlation with activity in the vlPFC and AIC (Morita et al., 2012). Finally, one study investigated guilt-related activity in patients with OCD (Basile et al., 2014), and found reduced activity in patients across a number of regions. The small number of studies and different methodologies/statistical contrasts used makes these findings difficult to integrate, but given that these emotions are known to be experienced ‘abnormally’ in a number of mental disorders, further research is clearly warranted. Guilt has for a long time been recognized to play an important role in depression, more recently, researchers have found that both shame and guilt are strongly associated with depressive disorders (Pulcu et al., 2014; Pulcu et al., 2013). It however remains uncertain whether and how shame and guilt are uniquely associated with depression or whether one has stronger links to the onset and maintenance of this disorder. One of the characteristics of depressed patients is their tendency to blame themselves for negative events (internal attribution), to generalize the situations (global attribution) and to perceive consistency in the cause of the event (stable attribution) (Abramson et al., 1989). It has been shown that both shame and guilt involve the tendency to make internal attributions for negative situations (Pineles et al., 2006). These characteristics resonate with the dysfunctional experience of self-blaming emotions: to blame the entire self for what happened

26

(shame) and to blame one’s own behavior for something bad happening (guilt). It has been shown in recent researches that people suffering from depression are more like to experience selfblaming emotions than emotions associated with blaming others, suggesting a more specific negative emotion tendency directed towards the self (Green et al., 2013b; Zahn et al., 2015). Furthermore, this relationship between the experience of self-blaming emotions such as shame/guilt and depression (Andrews, 1995; Andrews et al., 2002; Kim et al., 2011) is not specific and extends to other mental disorders, including anxiety disorders such as OCD (Basile et al., 2014; D’Olimpio and Mancini, 2014; Fergus et al., 2010; Hennig-Fast et al., 2015; Magee and Teachman, 2007; Mancini and Gangemi, 2011) and post-traumatic stress disorder (PTSD) (Amstadter and Vernon, 2008), eating disorders (ED) (Doran and Lewis, 2012; Goss and Allan, 2009; Troop et al., 2008) and personality disorders such as psychopathy (Blair et al., 1995; Holmqvist, 2008) and borderline personality disorder (BPD) (Lansky, 2003; Rüsch et al., 2007). Studying the neural correlates of self-blaming emotions across a range of patient populations is likely to provide a better understanding of the development of mental disorders. Neuroimaging studies of self-blaming emotions may lead to the discovery of biomarkers that could help us identify or predict individuals who may be more likely to experience pathological guilt and/or shame and therefore be a stepping stone to providing those affected with better and more adapted treatments.

5. Limitations Limitations of this systematic review include first, the focus on ‘negative’ moral emotions. While including positive moral emotions may have allowed for comparison more comprehensive review of the neural biology of moral emotion processing, to date, even fewer studies have investigated the neural correlates of positive moral emotions, precluding the worthwhile conduct of a systematic review. Second, the review did not include studies on sympathy, a widely investigated emotion that shares many relationships with moral emotions. However, considering the broader difficulties when comparing sympathy to guilt, shame and embarrassment, including sympathy

27

would have necessitated a much lengthier review, and it would have taken away the important focus on shame, embarrassment and guilt. Third, because functional deactivations and research on psychiatric populations were only reported across a handful of studies, these findings were reported within the result section but were not reviewed or interpreted in the discussion. Fourth, given that only three studies utilized fMRI sequences that were optimized for ventral temporal and frontal signal coverage, it is possible that our conclusions about guilt, shame and embarrassment-related neural activity are biased in terms of the regions implicated. Fifth, the variability in methodologies used by the reviewed studies (in terms of the paradigms used [e.g., scripts versus autobiographical recall], statistical contrasts employed [e.g., guilt>neutral versus guilt>“other emotion”], and group versus individual-based analyses) makes comparison of results difficult. While we based our discussion and interpretation of results primarily on studies employing ‘non-neutral’ contrast conditions (in order to avoid confounding from general negative emotional states or other social/moral processes), given the diversity of non-neutral contrast conditions (employed across and within studies), our interpretation of results must be taken with caution. Further, while a more in-depth critical appraisal of methodological aspects of studies such as sample size, sampling biases, data processing, etc., would be useful, the limited number of studies and varied methodologies used to date makes this difficult. Such in-depth appraisal should, however, be made in future when there are more studies available. Finally, we did not perform a meta-analysis of research findings, primarily because of the relatively small number of studies performed to date, in addition to the wide variation of methodologies employed. A reliable meta-analysis should be performed in future with the availability of further research.

6. Conclusions Negative moral emotions, particularly shame, embarrassment and guilt, are important emotions for adaptive functioning in society; they play a particularly important role in the maintenance and development of social relationships. This systematic review summarized the current knowledge about neural correlates of these emotions, and highlighted the similarities and differences between

28

them. All three emotions were suggested to be associated with function of neural systems underlying emotional processing, self-referential processing and social cognition. Further research is needed to comprehensively understand the neural correlates of these emotions (particularly shame and embarrassment, which have received the least study), and future research should be mindful of the methodological difficulties in inducing these kinds of moral emotions. Finally, the fact that self-blaming emotions can become dysfunctional when inappropriately experienced, and therefore contribute to the development and progression of mental disorders, suggests the need for more research in this area.

29

Acknowledgments This work was supported by the National Health and Medical Research Council of Australia (NHMRC; Career Development Fellowship to SW [ID: 1007716]), NHMRC Career Development Fellowship to CGD [ID: 1061757], and The University of Melbourne (Melbourne International Research Scholarship to CB).

30

References Abramson, L.Y., Metalsky, G.I., Alloy, L.B., 1989. Hopelessness depression: A theory-based subtype of depression. Psychological Review 96, 358. Amstadter, A.B., Vernon, L.L., 2008. Emotional reactions during and after trauma: A comparison of trauma types. Journal of Aggression, Maltreatment & Trauma 16, 391-408. Andrews, B., 1995. Bodily shame as a mediator between abusive experiences and depression. J Abnorm Psychol 104, 277-285. Andrews, B., Qian, M., Valentine, J.D., 2002. Predicting depressive symptoms with a new measure of shame: The Experience of Shame Scale. British Journal of Clinical Psychology 41, 29-42. Banks, S.J., Eddy, K.T., Angstadt, M., Nathan, P.J., Phan, K.L., 2007. Amygdala–frontal connectivity during emotion regulation. Social cognitive and affective neuroscience 2, 303-312. Barrett, K.C., Zahn-Waxler, C., Cole, P.M., 1993. Avoiders vs. amenders: Implications for the investigation of guilt and shame during toddlerhood? Cognition & Emotion 7, 481-505. Basile, B., Mancini, F., Macaluso, E., Caltagirone, C., Bozzali, M., 2014. Abnormal processing of deontological guilt in obsessive-compulsive disorder. Brain Struct Funct 219, 1321-1331. Basile, B., Mancini, F., Macaluso, E., Caltagirone, C., Frackowiak, R.S., Bozzali, M., 2011. Deontological and altruistic guilt: Evidence for distinct neurobiological substrates. Human brain mapping 32, 229-239. Beer, J.S., John, O.P., Scabini, D., Knight, R.T., 2006. Orbitofrontal cortex and social behavior: integrating self-monitoring and emotion-cognition interactions. Cognitive Neuroscience, Journal of 18, 871-879. Blair, R., Sellars, C., Strickland, I., Clark, F., Williams, A., Smith, M., Jones, L., 1995. Emotion attributions in the psychopath. Personality and Individual Differences 19, 431-437. Blakemore, S.-J., 2008. The social brain in adolescence. Nature Reviews Neuroscience 9, 267277.

31

Bowles, S., Gintis, H., 2005. Prosocial emotions. The Economy As an Evolving Complex System III. Santa Fe Institute, Santa Fe, NM, 339-366. Brody, A.L., Saxena, S., Mandelkern, M.A., Fairbanks, L.A., Ho, M.L., Baxter, L.R., 2001. Brain metabolic changes associated with symptom factor improvement in major depressive disorder. Biological psychiatry 50, 171-178. Buckner, R.L., Andrews‐Hanna, J.R., Schacter, D.L., 2008. The brain's default network. Annals of the New York Academy of Sciences 1124, 1-38. Calder, A.J., Lawrence, A.D., Young, A.W., 2001. Neuropsychology of fear and loathing. Nature Reviews Neuroscience 2, 352-363. Casebeer, W.D., Churchland, P.S., 2003. The neural mechanisms of moral cognition: A multipleaspect approach to moral judgment and decision-making. Biology and philosophy 18, 169-194. Cavanna, A.E., Trimble, M.R., 2006. The precuneus: a review of its functional anatomy and behavioural correlates. Brain 129, 564-583. Craig, A.D., 2009. How do you feel—now? The anterior insula and human awareness. Nature Reviews Neuroscience 10, 59-70. D’Olimpio, F., Mancini, F., 2014. Role of Deontological Guilt in Obsessive-Compulsive Disorder–Like Checking and Washing Behaviors. Clinical Psychological Science 2, 727-739. Decety, J., Lamm, C., 2007. The role of the right temporoparietal junction in social interaction: how low-level computational processes contribute to meta-cognition. The Neuroscientist 13, 580593. Doran, J., Lewis, C.A., 2012. Components of shame and eating disturbance among clinical and non‐clinical populations. European Eating Disorders Review 20, 265-270. Dranovsky, A., Hen, R., 2006. Hippocampal neurogenesis: regulation by stress and antidepressants. Biological psychiatry 59, 1136-1143. Drevets, W.C., Savitz, J., Trimble, M., 2008. The subgenual anterior cingulate cortex in mood disorders. CNS spectrums 13, 663. Eisenberg, N., 2000. Emotion, regulation, and moral development. Annual review of psychology 51, 665-697.

32

Eisenberger, N.I., Lieberman, M.D., 2004. Why rejection hurts: a common neural alarm system for physical and social pain. Trends in cognitive sciences 8, 294-300. Etkin, A., Egner, T., Kalisch, R., 2011. Emotional processing in anterior cingulate and medial prefrontal cortex. Trends in cognitive sciences 15, 85-93. Fergus, T.A., Valentiner, D.P., McGrath, P.B., Jencius, S., 2010. Shame-and guilt-proneness: Relationships with anxiety disorder symptoms in a clinical sample. Journal of anxiety disorders 24, 811-815. Finger, E.C., Marsh, A.A., Kamel, N., Mitchell, D.G., Blair, J.R., 2006. Caught in the act: the impact of audience on the neural response to morally and socially inappropriate behavior. Neuroimage 33, 414-421. Fourie, M.M., Thomas, K.G., Amodio, D.M., Warton, C.M., Meintjes, E.M., 2014. Neural correlates of experienced moral emotion: an fMRI investigation of emotion in response to prejudice feedback. Social neuroscience 9, 203-218. Frith, C.D., 2007. The social brain? Philosophical Transactions of the Royal Society of London B: Biological Sciences 362, 671-678. Gibson, M., 2013. Shame and guilt in child protection social work: new interpretations and opportunities for practice. Child & Family Social Work. Goldstein, K., 1944. The mental changes due to frontal lobe damage. The Journal of Psychology 17, 187-208. Goss, K., Allan, S., 2009. Shame, pride and eating disorders. Clinical psychology & psychotherapy 16, 303-316. Gray, J.R., Braver, T.S., Raichle, M.E., 2002. Integration of emotion and cognition in the lateral prefrontal cortex. Proceedings of the National Academy of Sciences 99, 4115-4120. Green, S., Lambon Ralph, M.A., Moll, J., Zakrzewski, J., Deakin, J.F.W., Grafman, J., Zahn, R., 2013a. The neural basis of conceptual–emotional integration and its role in major depressive disorder. Social neuroscience 8, 417-433.

33

Green, S., Moll, J., Deakin, J.F.W., Hulleman, J., Zahn, R., 2013b. Proneness to decreased negative emotions in major depressive disorder when blaming others rather than oneself. Psychopathology 46, 34-44. Green, S., Ralph, M.A.L., Moll, J., Deakin, J.F., Zahn, R., 2012. Guilt-Selective Functional Disconnection of Anterior Temporal and Subgenual Cortices in Major Depressive DisorderGuiltSelective Functional Disconnection. Archives of general psychiatry 69, 1014-1021. Green, S., Ralph, M.A.L., Moll, J., Stamatakis, E.A., Grafman, J., Zahn, R., 2010. Selective functional integration between anterior temporal and distinct fronto-mesolimbic regions during guilt and indignation. Neuroimage 52, 1720-1726. Greene, J.D., Nystrom, L.E., Engell, A.D., Darley, J.M., Cohen, J.D., 2004. The neural bases of cognitive conflict and control in moral judgment. Neuron 44, 389-400. Greene, J.D., Sommerville, R.B., Nystrom, L.E., Darley, J.M., Cohen, J.D., 2001. An fMRI investigation of emotional engagement in moral judgment. Science 293, 2105-2108. Harrison, B.J., Pujol, J., López-Solà, M., Hernández-Ribas, R., Deus, J., Ortiz, H., Soriano-Mas, C., Yücel, M., Pantelis, C., Cardoner, N, 2008. Consistency and functional specialization in the default mode brain network. Proceedings of the National Academy of Sciences 105, 9781-9786. Harrison, B.J., Pujol, J., Soriano-Mas, C., Hernández-Ribas, R., López-Solà, M., Ortiz, H., Alonso, P., Deus, J., Menchon, J.M., Real, E., 2012. Neural Correlates of Moral Sensitivity in Obsessive-Compulsive Disorder Archives of general psychiatry 69, 741-749. Heekeren, H.R., Wartenburger, I., Schmidt, H., Prehn, K., Schwintowski, H.-P., Villringer, A., 2005. Influence of bodily harm on neural correlates of semantic and moral decision-making. Neuroimage 24, 887-897. Heekeren, H.R., Wartenburger, I., Schmidt, H., Schwintowski, H.-P., Villringer, A., 2003. An fMRI study of simple ethical decision-making. Neuroreport 14, 1215-1219. Hennig-Fast, K., Michl, P., Müller, J., Niedermeier, N., Coates, U., Müller, N., Engel, R.R., Möller, H.-J., Reiser, M., Meindl, T., 2015. Obsessive-compulsive disorder–A question of conscience? An fMRI study of behavioural and neurofunctional correlates of shame and guilt. Journal of psychiatric research 68, 354-362.

34

Holmqvist, R., 2008. Psychopathy and affect consciousness in young criminal offenders. Journal of Interpersonal Violence 23, 209-224. Isoda, M., Noritake, A., 2013. What makes the dorsomedial frontal cortex active during reading the mental states of others? Frontiers in neuroscience 7. Johnstone, T., van Reekum, C.M., Urry, H.L., Kalin, N.H., Davidson, R.J., 2007. Failure to regulate: counterproductive recruitment of top-down prefrontal-subcortical circuitry in major depression. The Journal of neuroscience 27, 8877-8884. Kang, P., Lee, J., Sul, S., Kim, H., 2013. Dorsomedial prefrontal cortex activity predicts the accuracy in estimating others' preferences. Frontiers in human neuroscience 7. Kaufman, G., 1989. The Psychology of Shame. Theory and Treatment of Shame-Based Syndromes. New York (Springer) 1989. Kédia, G., Berthoz, S., Wessa, M., Hilton, D., Martinot, J.-L., 2008. An agent harms a victim: a functional magnetic resonance imaging study on specific moral emotions. Journal of cognitive neuroscience 20, 1788-1798. Kelley, A.E., Berridge, K.C., 2002. The neuroscience of natural rewards: relevance to addictive drugs. The Journal of neuroscience 22, 3306-3311. Kennedy, D.P., Adolphs, R., 2012. The social brain in psychiatric and neurological disorders. Trends in cognitive sciences 16, 559-572. Khalsa, S.S., Rudrauf, D., Feinstein, J.S., Tranel, D., 2009. The pathways of interoceptive awareness. Nature neuroscience 12, 1494-1496. Kim, S., Thibodeau, R., Jorgensen, R.S., 2011. Shame, guilt, and depressive symptoms: a metaanalytic review. Psychological bulletin 137, 68-96. Kjaer, T.W., Nowak, M., Lou, H.C., 2002. Reflective self-awareness and conscious states: PET evidence for a common midline parietofrontal core. Neuroimage 17, 1080-1086. Krajbich, I., Adolphs, R., Tranel, D., Denburg, N.L., Camerer, C.F., 2009. Economic games quantify diminished sense of guilt in patients with damage to the prefrontal cortex. The Journal of Neuroscience 29, 2188-2192.

35

Lansky, M.R., 2003. Discussion of Peter Fonagy et al.'s “The Developmental Roots of Borderline Personality Disorder in Early Attachment Relationships: A Theory and Some Evidence”. Psychoanalytic Inquiry 23, 460-472. Leech, R., Kamourieh, S., Beckmann, C.F., Sharp, D.J., 2011. Fractionating the default mode network: distinct contributions of the ventral and dorsal posterior cingulate cortex to cognitive control. The Journal of Neuroscience 31, 3217-3224. Lewis, H., 1971. Shame and Guilt in Neurosis. New York (International Universities Press) 1971. Liberati, A., Altman, D.G., Tetzlaff, J., Mulrow, C., Gøtzsche, P.C., Ioannidis, J.P., Clarke, M., Devereaux, P., Kleijnen, J., Moher, D., 2009. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Annals of internal medicine 151, W-65-W-94. Lindsay-Hartz, J., 1984. Contrasting experiences of shame and guilt. American Behavioral Scientist 27, 689-704. Magee, J.C., Teachman, B.A., 2007. Why did the white bear return? Obsessive–compulsive symptoms and attributions for unsuccessful thought suppression. Behaviour research and therapy 45, 2884-2898. Mancini, F., Gangemi, A., 2011. Fear of deontological guilt and fear of contamination in obsessive-compulsive disorder. Psicoterapia Cognitiva e Comportamentale 17, 395-404. Mars, R.B., Neubert, F.-X., Noonan, M.P., Sallet, J., Toni, I., Rushworth, M.F., 2012. On the relationship between the “default mode network” and the “social brain”. Frontiers in human neuroscience 6. Masten, C.L., Morelli, S.A., Eisenberger, N.I., 2011. An fMRI investigation of empathy for ‘social pain’and subsequent prosocial behavior. Neuroimage 55, 381-388. Mayberg, H.S., Liotti, M., Brannan, S.K., McGinnis, S., Mahurin, R.K., Jerabek, P.A., Silva, J.A., Tekell, J.L., Martin, C.C., Lancaster, J.L., 2014. Reciprocal limbic-cortical function and negative mood: converging PET findings in depression and normal sadness. American Journal of Psychiatry 165, 675-682.

36

McEwen, B.S., 2001. Plasticity of the hippocampus: adaptation to chronic stress and allostatic load. Annals of the New York Academy of Sciences 933, 265-277. Michl, P., Meindl, T., Meister, F., Born, C., Engel, R.R., Reiser, M., Hennig-Fast, K., 2014. Neurobiological underpinnings of shame and guilt: a pilot fMRI study. Social cognitive and affective neuroscience 9, 150-157. Moll, J., Oliveira-Souza, R.d., Garrido, G.J., Bramati, I.E., Caparelli-Daquer, E.M., Paiva, M. LMF, Zahn, R., Grafman, J., 2007. The self as a moral agent: linking the neural bases of social agency and moral sensitivity. Social Neuroscience 2, 336-352. Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., Grafman, J., 2005. The neural basis of human moral cognition. Nature Reviews Neuroscience 6, 799-809. Morey, R.A., McCarthy, G., Selgrade, E.S., Seth, S., Nasser, J.D., LaBar, K.S., 2012. Neural systems for guilt from actions affecting self versus others. Neuroimage 60, 683-692. Morgane, P.J., Galler, J.R., Mokler, D.J., 2005. A review of systems and networks of the limbic forebrain/limbic midbrain. Progress in neurobiology 75, 143-160. Morita, T., Itakura, S., Saito, D.N., Nakashita, S., Harada, T., Kochiyama, T., Sadato, N., 2008. The role of the right prefrontal cortex in self-evaluation of the face: a functional magnetic resonance imaging study. Journal of Cognitive Neuroscience 20, 342-355. Morita, T., Kosaka, H., Saito, D.N., Ishitobi, M., Munesue, T., Itakura, S., Omori, M., Okazawa, H., Wada, Y., Sadato, N., 2012. Emotional responses associated with self-face processing in individuals with autism spectrum disorders: An fMRI study. Social Neuroscience 7, 223-239. Morita, T., Tanabe, H.C., Sasaki, A.T., Shimada, K., Kakigi, R., Sadato, N., 2013. The anterior insular and anterior cingulate cortices in emotional processing for self-face recognition. Social cognitive and affective neuroscience, nst011. Northoff, G., Heinzel, A., De Greck, M., Bermpohl, F., Dobrowolny, H., Panksepp, J., 2006. Selfreferential processing in our brain—a meta-analysis of imaging studies on the self. Neuroimage 31, 440-457. O'Connor, L.E., Berry, J.W., Weiss, J., 1999. Interpersonal guilt, shame, and psychological problems. Journal of Social and Clinical Psychology 18, 181-203.

37

Olson, I.R., Plotzker, A., Ezzyat, Y., 2007. The enigmatic temporal pole: a review of findings on social and emotional processing. Brain 130, 1718-1731. Pavuluri, M., May, A., 2015. I feel, therefore, I am: the insula and its role in human emotion, cognition and the sensory-motor system. AIMS Neurosci 2, 18-27. Pineles, S.L., Street, A.E., Koenen, K.C., 2006. The Differential Relationships of Shame– Proneness and Guilt–Proneness to Psychological and Somatization Symptoms. Journal of Social and Clinical Psychology 25, 688-704. Pulcu, E., Lythe, K., Elliott, R., Green, S., Moll, J., Deakin, J.F., Zahn, R., 2014. Increased Amygdala Response to Shame in Remitted Major Depressive Disorder. PloS one 9, e86900. Pulcu, E., Zahn, R., Elliott, R., 2013. The role of self-blaming moral emotions in major depression and their impact on social-economical decision making. Front Psychol 4, 310. Ross, L.A., Olson, I.R., 2010. Social cognition and the anterior temporal lobes. Neuroimage 49, 3452-3462. Roth, L., Kaffenberger, T., Herwig, U., Brühl, A.B., 2014. Brain Activation Associated with Pride and Shame. Neuropsychobiology 69, 95-106. Rüsch, N., Lieb, K., Göttler, I., Hermann, C., Schramm, E., Richter, H., Jacob, G.A., Corrigan, P.W., Bohus, M., 2007. Shame and implicit self-concept in women with borderline personality disorder. American Journal of Psychiatry. Schardt, D.M., Erk, S., Nüsser, C., Nöthen, M.M., Cichon, S., Rietschel, M., Treutlein, J., Goschke, T., Walter, H., 2010. Volition diminishes genetically mediated amygdala hyperreactivity. Neuroimage 53, 943-951. Schilbach, L., Eickhoff, S.B., Rotarska-Jagiela, A., Fink, G.R., Vogeley, K., 2008. Minds at rest? Social cognition as the default mode of cognizing and its putative relationship to the “default system” of the brain. Consciousness and cognition 17, 457-467. Shin, L.M., Dougherty, D.D., Orr, S.P., Pitman, R.K., Lasko, M., Macklin, M.L., Alpert, N. M, Fischman, A. J, Rauch, S.L., 2000. Activation of anterior paralimbic structures during guiltrelated script-driven imagery. Biological psychiatry 48, 43-50.

38

Shulman, G.L., Corbetta, M., Buckner, R.L., Fiez, J.A., Miezin, F.M., Raichle, M.E., Petersen, S.E., 1997. Common blood flow changes across visual tasks: I. Increases in subcortical structures and cerebellum but not in nonvisual cortex. Journal of cognitive neuroscience 9, 624-647. Singer, T., Critchley, H.D., Preuschoff, K., 2009. A common role of insula in feelings, empathy and uncertainty. Trends in cognitive sciences 13, 334-340. Sturm, V.E., Sollberger, M., Seeley, W.W., Rankin, K.P., Ascher, E.A., Rosen, H.J., Miller, B.L., Levenson, R.W., 2012. Role of right pregenual anterior cingulate cortex in self-conscious emotional reactivity. Social cognitive and affective neuroscience, nss023. Stuss, D.T., Levine, B., 2002. Adult clinical neuropsychology: lessons from studies of the frontal lobes. Annual review of psychology 53, 401-433. Takahashi, H., Yahata, N., Koeda, M., Matsuda, T., Asai, K., Okubo, Y., 2004. Brain activation associated with evaluative processes of guilt and embarrassment: an fMRI study. Neuroimage 23, 967-974. Tangney, J.P., 1992. Situational detenninants of shame and guilt in young adulthood. Personality and Social Psychology Bulletin 18, 199-206. Tangney, J.P., 1995. Recent advances in the empirical study of shame and guilt. American Behavioral Scientist 38, 1132-1145. Tangney, J.P., 1996. Conceptual and methodological issues in the assessment of shame and guilt. Behaviour research and therapy 34, 741-754. Tangney, J.P., Miller, R.S., Flicker, L., Barrow, D., 1996. Are shame, guilt, and embarrassment distinct emotions? Journal of personality and social psychology 70, 1256-1269. Tangney, J.P., Stuewig, J., Mashek, D., Hastings, M., 2011. Assessing jail inmates’ proneness to shame and guilt: Feeling bad about the behavior or the self? Criminal justice and behavior, 0093854811405762. Tangney, J.P., Stuewig, J., Mashek, D.J., 2007. Moral emotions and moral behavior. Annual review of psychology 58, 345.

39

Tangney, J.P., Wagner, P., Fletcher, C., Gramzow, R., 1992a. Shamed into anger? The relation of shame and guilt to anger and self-reported aggression. Journal of personality and social psychology 62, 669. Tangney, J.P., Wagner, P., Gramzow, R., 1992b. Proneness to shame, proneness to guilt, and psychopathology. Journal of Abnormal Psychology 101, 469-478. Tibbetts, S.G., 2003. Self-conscious emotions and criminal offending. Psychological Reports 93, 101-126. Troop, N.A., Allan, S., Serpell, L., Treasure, J.L., 2008. Shame in women with a history of eating disorders. European Eating Disorders Review 16, 480-488. Wagner, U., N'Diaye, K., Ethofer, T., Vuilleumier, P., 2011. Guilt-specific processing in the prefrontal cortex. Cerebral cortex, bhr016. Wiech, K., Jbabdi, S., Lin, C., Andersson, J., Tracey, I., 2014. Differential structural and resting state connectivity between insular subdivisions and other pain-related brain regions. PAIN® 155, 2047-2055. Wong, Y., Tsai, J., 2007. Cultural models of shame and guilt. The self-conscious emotions: Theory and research, 209-223. Yu, H., Hu, J., Hu, L., Zhou, X., 2014. The voice of conscience: neural bases of interpersonal guilt and compensation. Social cognitive and affective neuroscience 9, 1150-1158. Zahn, R., Garrido, G., Moll, J., Grafman, J., 2014. Individual differences in posterior cortical volume correlate with proneness to pride and gratitude. Social cognitive and affective neuroscience 9, 1676-1683. Zahn, R., Lythe, K., Gethin, J., Green, S., Deakin, J., Workman, C., Moll, J., 2015. Negative emotions towards others are diminished in remitted major depression. European Psychiatry 30, 448-453. Zahn, R., Moll, J., Krueger, F., Huey, E.D., Garrido, G., Grafman, J., 2007. Social concepts are represented in the superior anterior temporal cortex. Proceedings of the National Academy of Sciences 104, 6430-6435.

40

Zahn, R., Moll, J., Paiva, M., Garrido, G., Krueger, F., Huey, E.D., Grafman, J., 2009. The neural basis of human social values: evidence from functional MRI. Cerebral Cortex 19, 276-283. Zaki, J., Ochsner, K.N., 2012. The neuroscience of empathy: progress, pitfalls and promise. Nature neuroscience 15, 675-680.

41

Database: PubMed

Figure captions:

Database: PsycInfo

Search terms: ((("Guilt"[Mesh]) OR "Shame"[Mesh])) AND (("Magne c Resonance Imaging"[Mesh]) OR ( "Neuroimaging"[Mesh] OR "Func onal Neuroimaging" [Mesh] ))

Search terms: ((exp guilt/ or exp shame/) OR ((guilt* or shame* or embarrassment* or moral emo on*).tw.)) AND ((exp magne c resonance imaging/ or exp neuroimaging/) OR ((fMRI or MRI or PET or "Magne c Resonance Imaging" or "Positron Emission Tomography" or neural).tw.))

Search terms: (((DE "Guilt") OR (DE "Shame")) OR (DE "Embarrassment") OR (TX guilt* OR TX shame* OR TX embarrassment* OR TX "moral emo on*”)) AND ((DE "Magne c Resonance Imaging" OR DE "Func onal Magne c Resonance Imaging" OR DE "Func onal Magne c Resonance Imaging") OR (DE "Neuroimaging" OR DE "Encephalography" OR DE "Magne c Resonance Imaging” OR DE "Roentgenography" OR DE "Tomography”)) OR (TX fMRI OR TX MRI ORTXPETORTX "Magne c Resonance Imaging" OR TX "Positron Emission Tomography" OR TX neural)

Filters: Language: English

Search limits: Language: English

Limiters: Language: English

Iden fica on

Figure 1: PRISMA flow diagram.

Database: MEDLINE

# of addi onal records iden fied through other sources: ((n=1)

Screening

# of records iden fied through database searching: PubMed: (n=28); MEDLINE: (n=178); PsycInfo: (n=241)

Eligibility

# of records a er duplicates removed: (n=340)

# of full-text ar cles assessed for eligibility: (n=41)

# of records excluded (not relevant from abstract): (n=299)

# of full-text ar cles excluded: (n=17)

42

Included

*Non Journal (n=3) # of studies included in qualita ve synthesis:(n=21)

Figure 2: a) Brain regions reported to be involved in the experience of shame, embarrassment and guilt (compared to a neutral condition)

b) Brain regions reported to be involved in the experience of shame, embarrassment and guilt (compared to an ‘other emotion’ control condition)

Note: Laterality of brain regions was not taken into account to illustrate results. Peak activation reported by studies has been shifted in some cases to enable illustration of all results on the five sagittal slices. The size of the dots is representative of cumulative frequency percentages as per Figure 3. A percentage of 100% referred to a dot diameter of 10 mm. Results reporting 10% or less are represented by a cross. Shame (in pink), embarrassment (in blue), and guilt (in green).

43

Figure 3: Cumulative frequency percentages of shame-, embarrassment- and guilt-related brain activations in healthy populations. 100.00 90.00 80.00

SHAME > emo on control condi on

Cumula ve frequency percentages

70.00

SHAME > neutral control condi on

60.00

EMBARRASSMENT > emo on control condi on

50.00

EMBARRASSEMENT > neutral control condi on

40.00

GUILT > emo on control condi on

30.00

GUILT > neutral control condi on

20.00 10.00

PR

PR EF RO

NT AL EF :d RO lP NT FC PR AL EF :d RO m NT PF C AL PR :* EF v m RO P NT FC AL :* CI vlP NG FC UL AT CI E: NG sA CC UL AT E: CI NG pA CC UL AT E CI : d NG AC UL C AT E CI :M NG SU CC BC UL OR AT E: TI CA PC SU L: C BC th OR SU al am TI BC CA us OR L: TI SU am CA BC yg L : OR da hi la pp TI CA oc SU L: am BC do p us rs OR al TI m CA id L: br ba ai n sa lg an gli IN a SU LA :A TE IC M IN PO SU RA LA TE L: :P M *a IC PO nt RA er io L: TE r *a IT M nt C PO er RA io r L: M TE *a TC M nt PO er RA io TE rS L: M TC po PO st RA er io L: TE r po IT M C st PO er io RA rM L: po TC st er i o TE rS M TC PO RA SE L T : NS EM ST OR S PO IM RA OT L: OR TP J :P SE re NS m OR ot or IM OT SE NS OR OR :M IM 1 OT OR : PA S1 RI ET AL :S PA PC RI PA ET RI AL ET :I AL PC :P re cu ne us OC CI PI TA CE L RE BE LL UM

0.00

Dorsolateral prefrontal cortex (dlPFC), dorsomedial prefrontal cortex (dmPFC), ventromedial prefrontal cortex (vmPFC), ventrolateral prefrontal cortex (vlPFC), subgenual anterior cingulate cortex (sACC), pregenual anterior cingulate cortex (pACC), dorsal anterior cingulate cortex (dACC), mid-cingulate cortex (MCC), posterior cingulate cortex (PCC), anterior insula cortex, posterior insula cortex (PIC), inferior temporal cortex (ITC), middle temporal cortex (MTC), superior temporal cortex (STC), superior temporal sulcus (STS), temporoparietal junction (TPJ), primary motor cortex (M1/PMC), first somato-sensori cortex (S1), superior parietal cortex (SPC), inferior parietal cortex (IPC). Described regions: PREFRONTAL AREAS: dlPFC, dmPFC, *vmPFC (including: medial orbifrontal cortex), *vlPFC (including: lateral orbifrontal cortex); CINGULATE AREAS: sACC, pACC, dACC, MCC, PCC; SUBCORTICAL AREAS: thalamus (including: septal/hippothalamic area), amygdala, hippocampus (including: parahippocampal gyrus), dorsal midbrain, basal ganglia (including: putamen, pallidum, caudate); INSULA CORTEX: AIC, PIC; TEMPORAL AREAS: *anterior ITC, *anterior MTC, *anterior STC, posterior ITC, posterior MTC, posterior STC, STS, TPJ; SENSORIMOTOR AREAS: premotor (including: supplementary motor cortex, pre-supplementary motor cortex), M1/PMC (including: pre-central gyrus, paracentral lobule, frontal operculum) , S1 (including: post-central gyrus); PARIETAL AREAS: superior parietal cortex (including: angular gyrus, parietal oppercullum), inferior parietal cortex (including: supramarginal gyrus), precuneus; OCCIPITAL AREAS (including: calcarine fissure, occipital inferior gyrus, fusiform gyrus, cuneus, supracalcarine, lingual gyrus). Note 1: The cumulative frequency percentage represents the number of times a specific brain region was implicated in a specific emotion (shame, embarrassment or guilt) across studies divided by the number of contrasts reporting activation for that particular emotion. This was done separately for neutral and emotion control conditions.

44 Note 2: Three studies claimed to have used optimized fMRI sequences to increase signal coverage in *orbitofrontal and *ventral temporal brain areas

Tables Table 1. Summary of studies on shame, embarrassment and guilt Study

Population

Imaging Modality

Type of Paradigm

Moral Emotions

Summary of Results

(Shin et al., 2000)

8 healthy individuals

PET

Auto-biographical recall

Guilt

Guilt > Neutral: dACC, MCC, AIC/vlPFC, anterior STC, M1: primary motor cortex (precentral gyrus), lateral cerebellum

(All males) Mean (SD) age: 25.0 (±4.4)

(Takahashi 2004)

et

al.,

19 healthy individuals

(Script-driven, instructed to imagine oneself in situation) fMRI

(10 M) Mean (SD) age: 30.8 (±6.2) (9 F) Mean (SD) age: 25.1 (±3.2)

(Finger et al., 2006)

16 healthy individuals

Emotion induction statements

Neutral > Guilt: PIC, M1: precentral gyrus, precuneus, visual cortex (cuneus, fusiform gyrus)

Guilt Embarrassment

Guilt > Neutral: dmPFC, posterior MTC, visual cortex (calcarine)

(script-driven, instructed to read the sentences, post scan guilt and embarrassment ratings)

fMRI

Emotion induction statements

(Gender/age not specified)

(script-driven, instructed to imagine oneself in situation)

Guilt > Embarrassment: dmPFC

Embarrassment > Guilt: hippocampus, anterior MTC, dorsal midbrain, visual cortex (calcarine, lingual gyrus, inferior occipital) Embarrassment > Neutral: dmPFC, vlPFC, posterior STS, anterior MTC, anterior ITC, hippocampus, visual cortex (calcarine, lingual gyrus)

Guilt Shame Embarrassment

Guilt (moral > social=neutral): dmPFC, anterior STC, posterior MTC Embarrassment/Shame (moral=social with audience) > (social without audience=neutral): dmPFC, vlPFC Embarrassment/Shame (audience>no audience): amygdala

45

(Moll et al., 2007)

12 healthy individuals

fMRI*

(6 M, 6 F) Mean (SD) age: 28.59 (±9.6)

(Morita et al., 2008)

19 healthy individuals

29 healthy individuals

fMRI

fMRI

(15 M, 14 F) Mean (SD) age: 21 (±1.98)

(Zahn et al., 2009b)

29 healthy individuals (15 M, 9 F) Mean (SD) age: 27.9 (± 7.3)

46

Guilt Embarrassment

(script-driven, instructed to imagine oneself in situation)

(9 M and 10 F) Mean (SD) age: 26.0 (±4.0)

(Kédia et al., 2008)

Emotion induction statements

Facial emotion induction (Photogenic ratings of self- and otherfaces, post scan embarrassment ratings) Emotion induction statements (script-driven, instructed to imagine oneself in situation)

fMRI*

Emotion induction statements (script-driven, instructed to judge pleasantness of own feeling, post-scan emotion ratings)

Guilt > Neutral: dmPFC, vlPFC, dorsal midbrain, anterior STC, posterior STS, M1: primary motor cortex (frontal operculum/precentral gyrus) Embarrassment > neutral: dmPFC, dACC, AIC, posterior STS, TPJ/posterior STC

Embarrassment

Self-reported embarrassment as parametric modulator of self-face condition: (negative correlation) dlPFC, AIC

Guilt Self-anger Other-anger Compassion

Guilt > Self-anger: posterior MTC, TPJ, preSMA, precuneus

Guilt Pride Indignation Gratitude

Negative self-agency (guilt condition) > fixation: no significant effect

*Note: similar regions were more active for other-anger and compassion > selfanger

Individual differences in the percentage of trials where guilt was experienced was positively correlated with activation in the vmPFC and vACC during negative self-agency trials, and the correlation was significantly greater for guilt compared to pride, indignation and gratitude

(Basile et al., 2011)

22 healthy individuals (9 M, 13 F) Mean (SD]) age: 26.8 (±3.0)

fMRI

Emotion induction statements associated with facial expressions (script-driven, instructed to imagine oneself in situation + forced choice guilt rating)

Deontological (DG) & Altruistic guilt (AG) Basic control emotions (i.e., anger and sadness)

Deontological+altruistic guilt > Basic emotions (anger & sadness): vACC, dACC, precuneus Deontological guilt > Sadness: dACC Altruistic guilt > Sadness: dmPFC (Deontological > anger) > (Altruistic guilt > sadness): AIC (Altruistic guilt > sadness) > (Deontological > anger): no significant effect

47

(Wagner et al., 2011)

15 healthy individuals

fMRI

(All females) age range: 25-30

Autobiographical recall

Guilt Shame Sadness

(script-driven, instructed to imagine oneself in situation)

Guilt > Neutral: dlPFC, vlPFC, dmPFC, PIC, anterior ITC, anterior MTC, anterior STC, posterior STS, preSMA, parahippocampal gyrus, IPC (supramarginal gyrus; angular gyrus), SPC (parietal operculum), basal ganglia (putamen), precuneus, visual cortex (calcarine), cerebellum (vermis) Guilt > Shame+Sadness: dmPFC, vlPFC Guilt > Shame: dlPFC, dmPFC, vlPFC, dACC, amygdala, basal ganglia (putamen, caudate, pallidum), AIC, anterior MTC, cerebellum Guilt > Sadness: vlPFC Shame > Neutral: dmPFC, AIC, anterior ITC, anterior MTC, IPC (angular gyrus), precuneus, visual cortex (lingual gyrus), cerebellum (vermis) Shame > Guilt: no significant effect Shame > Sadness: M1: primary motor cortex (precentral gyrus)

(Morita et al., 2012)

15 high-functioning ASD (14 M, 1 F) Mean (SD) age: 23.7 (± 4.3)

15 healthy controls

48

fMRI

Facial emotion induction (Photogenic ratings of self- and otherfaces, post scan embarrassment ratings)

Embarrassment

Within ASD group Self-reported embarrassment as parametric modulator of self-face condition: (positive correlation) vlPFC (negative correlation) AIC

(Morey et al., 2012)

16 healthy individuals

fMRI*

(All males) Mean (SD) age: 22.41 (±2.69)

Emotion induction statements

Guilt-self Guilt-other

Guilt-other > Guilt-self: posterior MTC, dmPFC, precuneus, visual cortex (calcarine) Guilt-self > Guilt-other: no significant effect

(script-driven, instructed to imagine oneself in situation, in-scan guilt ratings)

Main effect of guilt: (positive correlation with guilt ratings) dmPFC, vlPFC, IPC (supramarginal gyrus) Main effect of guilt: (negative correlations with guilt ratings) dlPFC, vlPFC, MCC, PIC, posterior MTC, S1: first somatosensoricortex (postcentral gyrus), SPC, precuneus, visual cortex (cuneus)

(Green et al., 2012)

25 remitted MDD participants 22 healthy controls (HC)

fMRI

Emotion induction statements (script-driven, unpleasantness ratings, post-scan individual classification of trials)

Guilt Indignation

Across groups (HC & remitted MDD) Guilt > Indignation: vACC (pregenual & subgenual), basal ganglia (caudate) HC > remitted MDD Guilt > Indignation PIC, parietal-occipital junction Remitted MDD > HC Guilt > Indignation no significant effects Within HC group Guilt > Indignation: dACC, anterior MTC, posterior MTC, preSMA, S1: first somatosensoricortex (postcentral), visual cortex (cuneus) Within remitted MDD group Guilt > Indignation: no significant effect

49

(Sturm et al., 2012)

27 behavioral variant frontotemporal dementia (bvFTD) patients

sMRI

Video recordings of a karaoke singing task

Embarrassment Both across groups (bvFTD & HC) and within each group Embarrassment (physiological reactivity and observed self-conscious behavior) associated with larger right vACC gray matter (after controlling for sadness associated reactivity/behavior)

(Passive viewing)

33 healthy controls

(Morita et al., 2013)

32 healthy individuals

fMRI

(16 M and 16 F) Mean (SD) age: 21.3 (±2.4)

(Michl et al., 2014)

14 healthy individuals (7 M, 7 F) Mean (SD) age: 29 (±2.95)

Facial emotion induction

Embarrassment

Self-reported embarrassment as a parametric modulator of self-face condition: (positive correlation) AIC

Guilt Shame

Guilt > Neutral: dlPFC, dmPFC, anterior MTC, anterior STC, posterior MTC, M1: primary motor cortex (frontal operculum, precentral gyrus), visual cortex (fusiform gyrus, calcarine, lingual gyrus), cerebellum

(Photogenic ratings of self- and otherfaces, post scan embarrassment ratings)

fMRI

Emotion induction statements (script-driven, instructed to imagine oneself in situation)

Guilt > Shame: thalamus, posterior MTC, visual cortex (fusiform gyrus) Shame > Neutral: dlPFC, dmPFC, hippocampus (parahippocampal gyrus), AIC, anterior STC, preSMA, visual cortex (middle occipital gyrus, calcarine), cerebellum Shame > Guilt: dlPFC, dACC, PCC, hippocampus (parahippocampal gyrus), dorsal (anterior) midbrain, AIC

50

(Yu et al., 2014)

(Basile et al., 2014)

27 healthy graduate & undergraduate students

13 OCD patients

fMRI sMRI using voxelbased morphometry (VBM)

fMRI

(10 M, 3 F) Mean (SD) age: 37.0 (11.1)

Interpersonal game paradigm

Interpersonal guilt

Interpersonal guilt: Positive correlation between gray matter (GM) volume in an anterior MCC cluster and compensation behavior during interpersonal game paradigm.

(dot estimation task, pain administered to other if self-incorrect, post-scan guilt ratings for each condition) Emotion induction statements associated with facial expressions

Deontological (DG) & Altruistic guilt (AG) Basic emotions (i.e., anger and sadness)

19 healthy controls (8 F & 11 M) Mean (SD) age: 26.2 (±2.1)

(Fourie et al., 2014)

22 healthy lowprejudice individuals Mean (SD) age: 19.32 (±1.11)

Implicit Association Test (written and facial stimilu) (categorization task followed by prejudice, egalitarian or neutral written feedback, post-scan emotion ratings)

Across groups (HC & OCD) Guilt (DG+AG) > Basic emotions No significant effect HC > OCD Guilt (DG+AG) > Basic emotions (anger + sadness): dmPFC, dACC, AIC, preSMA, precuneus

(script-driven, instructed to imagine oneself in situation + forced choice guilt rating)

fMRI

Interpersonal guilt (self-incorrect > both-incorrect contrasts values): anterior MCC, AIC

HC > OCD Guilt (DG only) > anger: AIC, preSMA, precuneus

Guilt Pride

HC > OCD Guilt (AG only) > sadness: no significant effect Guilt (prejudice feedback) > Neutral (neutral feedback): dmPFC, dACC, dorsal PCC, thalamus, AIC, precuneus Guilt (Prejudice Feedback) > Pride (Egalitarian Feedback): dmPFC, PCC, basal ganglia (caudate nucleus), AIC, posterior STS Negative correlation between individual guilt ratings and vACC activity during prejudice condition (controlling for other negative emotions)

51

(Pulcu et al., 2014)

21 remitted MDD

fMRI

(15 with no antidepressant medication)

Emotion induction statements

Shame Guilt

(script-driven, unpleasantness ratings, post-scan individual classification of trials)

18 healthy individuals (no personal/ family history of MDD)

MDD>HC Shame > Guilt: amygdala, PIC HC>MDD Shame > Guilt: no significant effect MDD Shame>Guilt: amygdala, PIC HC Shame>Guilt: No significant effect Guilt>Shame: No significant effect Across groups (HC & MDD) Shame > Guilt: no significant effect Guilt > Shame: no significant effect HC: positive correlation of unpleasantness of shame trials with PIC signal for shame vs. guilt.

(Roth et al., 2014)

25 healthy individuals (9 M, 16 F) Mean (SD) age: 31.9 (±10.2)

fMRI

Autobiographical recall (cue-driven, instructed to imagine oneself in situation)

Negative (shame/ guilt) Positive (pride) Neutral

Shame >Neutral: dlPFC, vlPFC, restrosplenial PCC, basal ganglia (pallidum, caudate head), AIC, SMA, cerebellum Neutral > Shame: SPC, IPC (supramarginal gyurs) Shame > Pride: no significant effect

52

(Zahn et al., 2014)

63 healthy individuals (33 M, 30 F) Mean (SD) age: 28.1 (±7.7)

sMRI

Emotion induction statements (script-driven, instructed to judge pleasantness of own feeling, post-scan emotion ratings)

Guilt Pride Indignation/anger Gratitude

Investigated structural variations and individual differences in proneness to specific moral emotions: Guilt-proneness associated with reduced volume in dlPFCa

* Study reported using optimized fMRI sequence for full brain coverage. a Result only found using a ROI analysis Structural magnetic resonance imaging (sMRI), functional magnetic resonance imaging (fMRI) Dorsolateral prefrontal cortex (dlPFC), dorsomedial prefrontal cortex (dmPFC), ventromedial prefrontal cortex (vmPFC), ventrolateral prefrontal cortex (vlPFC), subgenual anterior cingulate cortex (sACC), pregenual anterior cingulate cortex (pACC), dorsal anterior cingulate cortex (dACC), mid-cingulate cortex (MCC), posterior cingulate cortex (PCC), insula cortex (IC), anterior insula cortex (Roth et al., 2014), posterior insula cortex (PIC), anterior temporal lobe (ATL) inferior temporal cortex (ITC), middle temporal cortex (MTC), superior temporal cortex (STC), superior temporal sulcus (STS), temporoparietal junction (TPJ), supplementary motor cortex (SMA), pre-supplementary motor cortex (preSMA), primary motor cortex (M1/PMC), first somato-sensori cortex (S1), superior parietal cortex (SPC), inferior parietal cortex (IPC) Healthy control (HC), Major Depression Disorder (MDD), Obsessive Compulsive Disorder (OCD), Autism Spectrum Disorder (ASD), behavioral variant frontotemporal dementia (bvFTD)

53