Brain activation associated with evaluative processes of guilt and embarrassment: an fMRI study

We aimed to investigate the neural substrates associated with evaluative process of moral emotions. Using functional magnetic resonance imaging (fMRI), we examined the similarities and differences between evaluative process of guilt and that of embarrassment at the neural basis level. Study of the neural basis of judgments of moral emotions might contribute to a better understanding of the amoral behavior observed in neurological and psychiatric disorders. Nineteen healthy volunteers were studied. The participants read sentences carrying neutral, guilty, or embarrassing contents during the scans. Both guilt and embarrassment conditions commonly activated the medial prefrontal cortex (MPFC), left posterior superior temporal sulcus (STS), and visual cortex. Compared to guilt condition, embarrassment condition produced greater activation in the right temporal cortex (anterior), bilateral hippocampus, and visual cortex. Most of these regions have been implicated in the neural substrate of social cognition or Theory of Mind (ToM). Our results support the idea that both are self-conscious emotions, which are social emotions requiring the ability to represent the mental states of others. At the same time, our functional fMRI data are in favor of the notion that evaluative process of embarrassment might be a more complex process than that of guilt.

Introduction

Although there have been numerous neuroimaging studies on primary emotions (fear, disgust, happiness, and sadness) that have led to a better understanding of the neuroanatomical correlates of emotions (Phan et al., 2002), only a few studies on complex social emotions such as guilt, shame, and embarrassment have been reported. These social emotions have been viewed as moral emotions because they occur in response to moral violation and promote moral behavior, interpersonal etiquette, and personal hygiene (Eisenberg, 2000, Haidt, 2003). At the same time, these emotions inhibit transgression of social standards and motivate reparative action such as apology, confession, and atonement.

Impairment of possessing the mental states of these moral emotions could lead to amoral, inappropriate behaviors observed in neurological and psychiatric disorders such as brain injuries (Anderson et al., 1999, Beer et al., 2003), frontotemporal dementia (Miller et al., 2003, Snowden et al., 2002), autism (Capps et al., 1992, Frith, 2001, Hillier and Allinson, 2002), and antisocial personality (Brower and Price, 2001, Moll et al., 2003). Studying the neural substrates of judgments of moral emotions should add to the understanding of the neural basis of amoral behaviors observed in neurological and psychiatric disorders.

From a psychological point of view, guilt, shame, embarrassment, and pride are categorized into the same emotion family, “self-conscious emotions”. “Self-conscious emotions” are emotions founded in social relationship and arise from concerns about others' opinions of self or the behavior of self (Eisenberg, 2000, Haidt, 2003, Tangney and Dearing, 2002). Negative evaluation of self or the behavior of self is fundamental to guilt, shame, and embarrassment, while positive evaluation of self leads to pride. In other words, one needs the ability to represent the mental states of others (intention/emotion), that is, Theory of Mind (ToM), to recognize self-conscious emotions. The recognition of negative self-conscious emotions involves understanding of the violation of social norms and the negative evaluation of self, both important aspects of ToM. Children with autism demonstrating impaired ToM showed impaired recognition of self-conscious emotions (Heerey et al., 2003). In line with this notion, a recent functional magnetic resonance imaging (fMRI) study demonstrated activation in the medial prefrontal cortex (MPFC), temporal regions, and orbitofrontal cortex (OFC) during the emotional judgments of embarrassment (Berthoz et al., 2002). These areas have been implicated in ToM, social cognition, and moral judgment (Adolphs, 2001, Allison et al., 2000, Frith, 2001, Frith and Frith, 1999, Greene and Haidt, 2002, Greene et al., 2001, Moll et al., 2003, Pinkham et al., 2003). However, a previous positron emission tomography (PET) study using a guilt-related script reported a slightly different activation pattern in anterior paralimbic regions during the experience of guilt (Shin et al., 2000). In the former fMRI study, an emotional judgment task was used. Participants read various kinds of stories depicting embarrassing situations and were instructed to imagine what the story protagonist (participant himself or a third person in the story) would feel. In the latter PET study, an emotion induction method was used. Participants listened to audio-taped personal events involving the most guilt they had actually experienced. They recalled and imagined the event as if they were actually participating in it again. These two studies differed in the emotional tasks and measurement methods, making it difficult to compare the results directly. To our knowledge, no neuroimaging study has as yet investigated the different types of self-conscious emotions and compared the neural activation patterns directly in one session.

Although the distinctions among guilt, shame, and embarrassment are not clear-cut, psychologists have challenged this issue. Although embarrassment has traditionally been considered to be a variant of shame (Lewis, 1993), recent psychological data support the notion that embarrassment is an emotion distinct from other self-conscious emotions (Keltner and Buswell, 1997). Embarrassment has higher affinity to violation of social conventions, while guilt and shame have higher affinity to violation of a moral norm (Eisenberg, 2000, Haidt, 2003, Tangney et al., 1996). In this sense, among the negative self-conscious emotions, distinction between guilt and embarrassment is considered to be relatively clear-cut. Therefore, we focused on these two emotions.

We used block-design fMRI to measure regional activation associated with judgments of guilt and embarrassment during an emotional judgment task presenting short sentences. Emotion processing is composed of evaluative, experiential, and expressive components. We did not intend to induce emotional states because we thought it would be difficult to induce emotional states of guilt or embarrassment by merely having the subjects read short sentences. Moreover, it would be difficult to control the situation so as not to induce emotions other than guilt and embarrassment (e.g., anger, shame, sadness) as reported in the previous induction study (Shin et al., 2000). We aimed to elucidate the similarities and differences between the evaluative process of guilt and that of embarrassment at the neural basis level by measuring neural activation during judgments of both emotions in a session using fMRI.

We hypothesized that both emotional conditions would commonly activate the components of the neural substrates (MPFC, superior temporal sulcus (STS)) that have been implicated in social cognition (Adolphs, 2001, Allison et al., 2000) or ToM (Frith, 2001, Frith and Frith, 1999), and at the same time, would show differences in the extent of activation of the components.