Social anhedonia is associated with neural abnormalities during face emotion processing

Abstract

Human beings are social organisms with an intrinsic desire to seek and participate in social interactions. Social anhedonia is a personality trait characterized by a reduced desire for social affiliation and reduced pleasure derived from interpersonal interactions. Abnormally high levels of social anhedonia prospectively predict the development of schizophrenia and contribute to poorer outcomes for schizophrenia patients. Despite the strong association between social anhedonia and schizophrenia, the neural mechanisms that underlie individual differences in social anhedonia have not been studied and are thus poorly understood. Deficits in face emotion recognition are related to poorer social outcomes in schizophrenia, and it has been suggested that face emotion recognition deficits may be a behavioral marker for schizophrenia liability. In the current study, we used functional magnetic resonance imaging (fMRI) to see whether there are differences in the brain networks underlying basic face emotion processing in a community sample of individuals low vs. high in social anhedonia. We isolated the neural mechanisms related to face emotion processing by comparing face emotion discrimination with four other baseline conditions (identity discrimination of emotional faces, identity discrimination of neutral faces, object discrimination, and pattern discrimination). Results showed a group (high/low social anhedonia) × condition (emotion discrimination/control condition) interaction in the anterior portion of the rostral medial prefrontal cortex, right superior temporal gyrus, and left somatosensory cortex. As predicted, high (relative to low) social anhedonia participants showed less neural activity in face emotion processing regions during emotion discrimination as compared to each control condition. The findings suggest that social anhedonia is associated with abnormalities in networks responsible for basic processes associated with social cognition, and provide a starting point for understanding the neural basis of social motivation and our drive to seek social affiliation.

Introduction

As fundamentally social creatures, humans are driven by the desire for meaningful and frequent social interaction (Baumeister and Leary, 1995). There are individual differences in the strength of this desire, however, and some individuals exhibit a significantly reduced drive for social affiliation known as social anhedonia (Brown et al., 2007, Kwapil, 1998, Kwapil et al., 2009). Social anhedonia (SA) has been characterized as a deficiency in the need to belong to a social group and is distinct from other constructs that might also predict abnormalities in social interaction such as social anxiety (Brown et al., 2007, Kwapil et al., 2009). Individuals high in SA exhibit a genuine preference for solitude, disengagement during social interactions (Brown et al., 2007), and reduced negative affect when alone (Kwapil et al., 2009). Higher levels of SA are related to lower levels of social support and social functioning (Blanchard et al., 2011). Reduced social support and smaller social networks are associated with differences in immune functioning and other clinically significant health outcomes (Miller et al., 2009). Furthermore, high SA has been identified as one of the single most predictive traits for future onset of schizophrenia spectrum disorders (Kwapil, 1998) and has long been recognized as a core attribute of psychosis vulnerability (Bleuler, 1950, Horan et al., 2007, Kraepelin and Gosline, 1919, Meehl, 1962, Rado, 1953, Stone et al., 2005). Altogether, existing evidence indicates that SA is a deviation in a psychologically and clinically important social and emotional process that has broad implications for our understanding of normal and abnormal functioning.

Despite evidence for serious physical and mental health difficulties associated with reduced desire for social affiliation, no research to our knowledge has been done exploring the neural basis of SA in nonclinical populations. In schizophrenia, SA is considered a negative symptom that is stable (Blanchard et al., 2001) and can be reliably assessed (Horan et al., 2006). Studies of SA in schizophrenia have indicated that a number of neural systems may be involved in reduced desire for social affiliation, including the amygdala (Becerril and Barch, 2010), caudate nucleus (Dowd and Barch, 2010), dorsolateral prefrontal cortex (Becerril and Barch, 2010), and somatosensory areas (Arnfred and Chen, 2004). However, disease-related confounds and secondary characteristics of schizophrenia illness, such as psychosocial stress and neurodegenerative processes, make it difficult to generalize these findings to SA among healthy individuals or to identify whether neural abnormalities associated with SA are associated with the psychosis vulnerability (Lenzenweger, 2006).

Differences in the neural processing of face emotion provide a potential starting point for identifying abnormalities associated with high SA. Accurate face emotion recognition is critical for recognizing and responding to other's mental states and is a building block to more complex social behaviors (Adolphs, 2002). Importantly, face emotion recognition ability (but not face identity processing ability) predicts social functioning in schizophrenia participants (Hooker and Park, 2002) and varies with psychometric psychosis-proneness in nonclinical populations (Germine and Hooker, 2011). Previous work has also shown that face emotion perception is abnormal in individuals high in social anhedonia (Luh and Gooding, 1999). Thus, individual differences in social anhedonia may be related to deficits in the neural mechanisms supporting face emotion recognition.

The neural substrates of face emotion recognition are well characterized in healthy and clinical populations. Previous work indicates that effective emotion recognition involves the recruitment of a network of regions, including the amygdala (Adolphs, 2002, Adolphs et al., 1994, Anderson and Phelps, 2001), superior temporal sulcus (Allison et al., 2000, Haxby et al., 2000), medial prefrontal cortex (Amodio and Frith, 2006, Blair et al., 1999, Dolan et al., 1996, Gorno-Tempini et al., 2001, Gur et al., 2002a, Phillips et al., 1998, Sprengelmeyer et al., 1998, Wright et al., 2002), and somatosensory-related cortices (including insula, S1, S2, and anterior supramarginal gyrus) (Adolphs, 2002, Adolphs et al., 2000). Using functional neuroimaging, researchers have consistently found abnormalities in these regions during emotion recognition in individuals with schizophrenia (Das et al., 2007, Farrer et al., 2004, Gur et al., 2002b, Gur et al., 2007, Hall et al., 2004, Hempel et al., 2003, Holt et al., 2006, Kosaka et al., 2002, Phillips et al., 1999, Pinkham et al., 2008, Schneider et al., 1998, Spence et al., 1997, Taylor et al., 2002, Waberski et al., 2004, Williams et al., 2004).

Deficits in emotion recognition have been associated with lesions to the amygdala (Adolphs et al., 1994), somatosensory and related cortices (Adolphs et al., 2000) and medial prefrontal cortex (Heberlein et al., 2008). The medial prefrontal cortex, in particular, likely plays a broad role in many social-cognitive processes and has been implicated in lower-level emotion perception as well as higher-level processes including theory of mind attributions (Gallagher et al., 2000), self-referential processing (Mitchell et al., 2005), and distinguishing between self and other (Heatherton et al., 2006, Ochsner et al., 2004). In terms of functional divisions, the anterior portion of rostral medial prefrontal cortex (arMFC) has been consistently identified in measures of social cognition and emotion processing (Amodio and Frith, 2006) and in social cognition in schizophrenia (Brunet-Gouet and Decety, 2006).

In the present study, we used functional magnetic resonance imaging (fMRI) to examine differences in the neural circuitry underlying face emotion discrimination in otherwise normal individuals who were high versus low in SA. As our face emotion recognition task, we used the Queen Square Face Discrimination Test (QFDT; Garrido et al., 2009). Our primary hypothesis was that high SA would have specific deficits in face emotion processing even when controlling for broader, but equally complex aspects of face perception. The QFDT was chosen because it can dissociate face emotion processing and face identity processing (Banissy et al., 2011, Garrido et al., 2009, Germine and Hooker, 2011, Pitcher et al., 2008). In the QFDT, participants view sequentially presented emotional faces; in one condition they judge whether the two faces are expressing the same emotion, and in another condition they judge whether the two faces have the same identity. Importantly, the two conditions have identical stimuli and are equally difficult for healthy participants. As a result, any differences found between emotion discrimination and identity discrimination can be attributed to differences in specific cognitive processes related to emotion perception and cannot be attributed to differences in the stimuli, number of response options, or difficulty level of the two conditions. This feature of the task is an improvement over face processing studies where the experimental and control tasks differ along these dimensions (e.g. labeling emotional faces using four options vs. same/different identity of paired neutral faces). In the QFDT, the emotion recognition and identity recognition conditions use the same task structure (both are a forced choice same/different judgment) and the same stimuli. Therefore the comparison of emotion recognition and identity recognition of emotional faces isolates the specific cognitive processes for attending to, processing, and judging face emotions. Using a behavioral version of this task, we found that higher levels of psychosis risk (based on self-report of cognitive-perceptual, interpersonal, and disorganized psychosis-prone characteristics) are associated with reduced emotion discrimination performance, but normal identity discrimination performance (Germine and Hooker, 2011). The QFDT was also used by Pitcher et al. (2008), who found that applying transcranial magnetic stimulation (TMS) to the face area of somatosensory cortex impaired performance in the emotion discrimination condition, but not the identity discrimination condition. Thus, we have good reason to believe that the QFDT emotion discrimination condition depends on one or more processes specific to emotion processing that also vary with psychosis vulnerability. The current fMRI study included three additional control conditions designed to reveal potential group differences in the broader face emotion processing neural network. These conditions included identity discrimination of neutral faces, visual discrimination of objects, and visual discrimination of patterns. Given the putative relationship between SA and vulnerability for psychosis (Kwapil, 1998), we predicted that individuals high in SA would exhibit reduced recruitment of cortical regions involved in face emotion recognition, particularly superior temporal sulcus/gyrus, medial prefrontal cortex and somatosensory-related parietal regions, as well as reduced responses in the amygdala. Between group differences in one or more of these regions would indicate that higher levels of SA are associated with neural abnormalities during emotion perception, and help us better understand the neural basis of differences in the desire for social affiliation as well as psychosis vulnerability.