Reduced activation in the mirror neuron system during a virtual social cognition task in euthymic bipolar disorder

Abstract

Social cognition entails both cognitive and affective processing, and impairments in both have accounted for residual symptoms of bipolar disorder (BD). However, there has been a lack of studies identifying neural substrates responsible for social cognitive difficulties in BD patients. Fourteen euthymic BD patients and 14 healthy normal controls underwent functional MRI while performing a virtual reality social cognition task, which incorporated both cognitive and emotional dimensions, simulating real-world social situations. During the scanning, subjects tried to guess (attribute) possible reasons for expressed emotion of virtual humans (avatars) while viewing their facial expressions, just after observing their verbal and nonverbal (facial) expressions which were emotionally valenced (happy, angry and neutral). BD patients compared to normal controls showed delayed reaction times in emotional conditions, with comparable response accuracy. Healthy normal controls activated the right anterior cingulate cortex, inferior frontal, and insular cortex in emotional conditions contrasted with neutral control conditions, that is, the regions that have been related to empathic processes during viewing others' emotional expression. Relative to normal controls, BD patients showed reduced activations in the ‘mirror neuron system’, including the right inferior frontal cortex, premotor cortex, and insula, mainly in angry or happy condition. These results may suggest that, even during euthymic state, BD patients have difficulties in recruiting brain regions for the utilization of emotional cues as a means for understanding others. Clinical attention should be paid to emotion-related residual symptoms to help improve social outcomes in these patients.

Introduction

Recent studies have recognized that social cognition and emotion are closely related with each other; brain structures important for emotional functioning are also critical for social cognition (Adolphs, 1999). Furthermore, even in terms of social cognition itself, there have been numerous inquiries regarding the differential contribution between emotional and cognitive aspects of social cognition (Olsson and Ochsner, Olsson and Ochsner, 2008, Shamay-Tsoory et al., 2009a, Shamay-Tsoory et al., 2009b).

Recent studies have suggested that people use overlapping brain regions when they feel their own emotions and observe the emotion of others (Jabbi et al., 2007, Singer et al., 2004). This overlapping pattern has been thought to reflect an ‘automatic resonance’ to others' emotion just as a response to one's own emotion (Gallese, 2007). The ‘simulation theory of social cognition’ attributes this implicit emotion sharing to the function of the ‘mirror neuron system (MNS)’ in mental imitation of others' nonverbal expression (Gallese, 2007). Thus, some studies suggested that MNS is involved in emotional aspect of social cognition (Shamay-Tsoory et al., 2009a, Singer et al., 2004). On the other hand, the ventromedial or dorsolateral prefrontal cortex is suggested to subserve higher-level cognitive processes, which is referred to as perspective-taking, a more reflective process in social cognition (Olsson and Ochsner, Olsson and Ochsner, 2008, Shamay-Tsoory et al., 2009a, Shamay-Tsoory et al., 2009b).

Bipolar disorder (BD) is a major psychiatric illness characterized by emotion regulation dysfunction. As well as this emotional dysfunction (Bozikas et al., 2006, Wessa et al., 2007), BD patients were reported to have subtle cognitive dysfunctions in their remitted state (Green et al., 2007), that were sometimes comparable to those of schizophrenia patients in the realm of cognitive and social functioning (Dickerson et al., 2001). Likewise, social cognitive impairment has also been reported in BD patients. In the prior literature, largely two aspects of social cognition have been examined in these patients. One is more relevant to cognitively-based social cognition such as the ‘theory of mind’ (ToM) (Bora et al., 2005, Kerr et al., 2003, Olley et al., 2005), while the other is more relevant to affectively-based social cognition such as facial affect processing (Harmer et al., 2002, Lembke and Ketter, 2002, Yurgelun-Todd et al., 2000). However, most of these previous studies examined BD patients' ability in either ToM or pictorial, facial affect processing per se (Harmer et al., 2002, Lembke and Ketter, 2002, Yurgelun-Todd et al., 2000) rather than in utilizing such emotional information as a means for understanding others' perspectives as people do in real social interactions.

Actually, it is difficult to quantify social cognitive ability that involves both affective and cognitive aspects, as well as to approach these two aspects separately (Olsson and Ochsner, 2008). Nevertheless, a task simultaneously entailing these two aspects might be more useful to identify the neural regions people may activate in their real social lives than a task requiring either of the two inseparable dimensions of social cognition.

Therefore, we developed a virtual reality (VR) social cognition task that incorporates both cognitive (attribution) and emotional (facial affect) dimensions, simulating real-world face-to-face situation, as follows: subjects tried to guess possible reasons for expressed emotion of virtual humans (avatars) while viewing their facial expressions just after observing their verbal and nonverbal (facial) expressions. With this task, intending to activate brain regions responsible for the processing of emotional stimuli to attribute such emotional statuses of others, recently we reported abnormal brain activation in MNS in schizophrenia patients (Park et al., 2009).

Interestingly, the regions corresponding to MNS, including the inferior frontal cortex (IFC) or insula, have not only been related to emotional aspects such as simple observation or imitation of facial emotion expression (Carr et al., 2003), implicit affect induction (Malhi et al., 2007), emotion regulation (Dolcos and McCarthy, 2006) and the selection of memory material for interpreting others' emotion (Olsson and Ochsner, 2008), but it also has been associated with working memory (Jonides et al., 1998), and language processing (Mason and Just, 2007). Collectively, these findings support an idea that MNS may be involved in both affective and cognitive processing of social stimuli during human interactions. Notably, prior studies have been related altered activation in these regions to affective or cognitive impairments in BD patients during manic or euthymic state. For example, BD patients showed abnormal activation in the IFC and precentral gyrus during a language processing task (Curtis et al., 2007), in inferior frontal and insular cortex in a sustained attention task (Strakowski et al., 2004), and reduced ventrolateral prefrontal activation during a facial emotion labeling task (Foland et al., 2008).

Based on these findings, we hypothesized that abnormal activation in MNS would be seen in euthymic BD patients during performing our ‘face-to-face’ social cognitive task. In relation to this hypothesis, we also predicted that social cognitive performance would be impaired in BD patients relative to normal controls when the processing of emotional versus neutral social stimuli is required.