The desire for frequent and meaningful social interactions is a fundamental human motivation (Baumeister and Leary, 1995). Social anhedonia (SA), a trait-like disinterest in social contact and diminished capacity to experience pleasure from social interactions, is an example of when this need to belong goes awry (Silvia and Kwapil, 2011). Although socially anhedonic individuals report a genuine preference for solitude and reduced negative affect when alone (Brown et al., 2007; Kwapil et al., 2009), their asocial solitude negatively impacts their psychological well-being. High SA individuals report fewer social supports and less satisfaction with their existing social supports (Blanchard et al., 2011), avoidant attachment (Troisi et al., 2010), decreased social competence, and overall poor social functioning (Llerena et al., 2012) – all factors that are known to adversely impact important physical and mental health outcomes, possibly due to the lack of protective effects conveyed by social contact (Miller et al., 2008; Silvia and Kwapil, 2011). Indeed, high SA is consistently identified as a risk factor for psychiatric disorders (Watson and Naragon-Gainey, 2010), particularly schizophrenia-spectrum disorders. Meehl (1962) conceptualized SA as a feature of schizotypy – personality characteristics indicative of genetic liability for schizophrenia. Consistent with Meehl’s theory, SA is one of the strongest predictive traits of conversion to schizophrenia-spectrum disorders (Kwapil, 1998), and, similar to findings in non-clinical populations, is a key factor contributing to the characteristic social deficits in schizophrenia (Meehl, 1962; Blanchard et al., 1998, 2000). Collectively, existing evidence consistently associates high SA with poor social functioning in both healthy and clinical populations.

Although logically it follows that a reduced desire for social contact would lead to fewer friends and social engagements (i.e., poor social functioning), the underlying reason for why high SA individuals have poor social functioning is not entirely known. One proposal is that the consequences of SA may be related to deficits in attentional control, also termed effortful control, executive control, or cognitive control, and operationalized as the capacity to engage the inhibitory functions necessary to maintain task-relevant processing and goal oriented behavior (Derryberry and Reed, 2002; Lesh et al., 2011; Tully and Niendam, 2014). A core aspect of attentional control is the ability to inhibit prepotent responses in favor of subdominant ones – a self-regulatory mechanism that is likely a key component of successful social functioning (Heatherton and Wagner, 2011). Social interactions require the ability to filter out distracting/irrelevant information in order to attend to the relevant (e.g., in the “cocktail party” environment). This may be especially important in the context of emotional information: attentional control capabilities predict negative affect (Posner et al., 2002), response to conflict with a partner (Hooker et al., 2010), and response to social rejection (Gyurak et al., 2012), indicating that deficits in the ability to use attentional control to manage social/emotional information could harm social relationships over time, thereby negatively impacting social functioning. In this context, attentional control can be conceptualized as a domain-general system that assigns control resources to affect social behavior through multiple higher-level processes, such as the control of emotional information (Tully and Niendam, 2014), and other socially relevant processes such as theory of mind (ToM) and metacognition. This is consistent with the idea that neurocognitive impairments effect functioning via social cognitive impairments (Green et al., 2000; Schmidt et al., 2011), in which neurocognitive processes are the “building blocks” that precede socially relevant processes. In this study we examine attentional control as one of these neurocognitive processes.

Socially anhedonic but otherwise healthy individuals are impaired on tasks requiring attentional control, such as the Stroop paradigm (Giraldez et al., 2000) and the Wisconsin Card Sorting Task (Barrantes-Vidal et al., 2003). However, despite evidence demonstrating attentional control deficits in SA, the impact of these deficits on social functioning is rarely considered. In our previous work, we found high SA individuals demonstrated deficits in the attentional control of emotion information on an experimental task specifically designed to assess the ability to inhibit task-incongruent irrelevant negative faces, but these deficits did not relate to social impairments (Tully et al., 2012). However, highly specific experimental tasks may be too narrow to capture the effect of attentional control on social functioning. Here we sought to extend these findings by examining individual differences in attentional control as it naturally varies along a continuous dimension so as to better capture the multiple inhibitory demands of the social environment.

To test whether attentional control underlies the relationship between SA and social impairment, the present study investigates the relationship between individual differences in SA, attentional control, and social impairment in a large, representative, community sample. Specifically, we investigated whether attentional control mediates the relationship between SA and social impairment. Mediation analysis provides a meaningful statistical method for describing the mechanisms through which one variable exerts an effect on another (Hayes, 2009). We assessed SA using the Revised Social Anhedonia Scale (RSAS; Eckblad et al., 1982), attentional control using the Attentional Control Scale (ACS; Derryberry and Reed, 2002), and social impairment using the Social Adjustment Scale-Self-report (SAS-SR; Weissman et al., 1978). We hypothesized that: (1) high SA is associated with low attentional control and high social impairment, (2) low attentional control is associated with high social impairment, and (3) attentional control mediates the relationship between SA and social impairment.

Table 1 presents demographics and sample characteristics. There were no gender differences on any demographic or self-report measures. SA spanned the full range of possible scores (0–40), and had acceptable spread (interquartile range = 19). 35 participants (32%) – 20 female (34% female participants) and 15 male (32% male participants) – were high deviant scorers per the cut offs reported by Kwapil (1998). The distribution of SA scores is presented in Figure 1.

Table 3 presents bivariate Pearson correlation coefficients between IQ and all self-report variables. SA, attentional control, and social impairment were all significantly intercorrelated in the predicted directions; higher SA related to lower attentional control and greater social impairments, and lower attentional control related to higher social impairment. IQ did not significantly relate to any variables, indicating that these relationships are not due to IQ.

This study examined the relationship between individual differences in SA, attentional control, and social impairment in a large community sample of healthy individuals. Two main findings emerged: first, we replicated the association between high SA and high social impairment found in previous studies (Katsanis et al., 1992; Cohen et al., 2006; Blanchard et al., 2011) providing further evidence for the presence of social impairments in socially anhedonic but otherwise healthy individuals. Additionally, our results are consistent with prior evidence demonstrating attentional control deficits in high SA and schizophrenia samples, providing further support for the proposal that impaired attentional processes are characteristic of schizophrenia liability (Erlenmeyer-Kimling and Cornblatt, 1992). Second, attentional control partially mediated the relationship between SA and social impairment, accounting for 19% of the variance. Specifically, individuals with higher SA reported lower attentional control and lower social functioning. Although both attentional deficits and social impairments have been separately noted in SA, the relationship between SA, attentional control, and social impairments in this large community sample is a novel contribution to the literature. These findings suggest attentional control is one of the mechanisms underlying aberrations in the fundamental human need for social contact. Specifically, results suggest that the ability to engage attentional control processes, particularly focusing attention in the presence of irrelevant and distracting stimuli, is a cognitive feature of SA that contributes to social impairments.

Our results have implications for understanding how SA, a dimension of schizotypy that conveys risk for developing schizophrenia-spectrum disorders and a core negative symptom of schizophrenia, relates to the characteristic social impairments of the disorder and those at risk for developing it. Here, mediation analysis demonstrated attentional control is a mediator that accounts for only 19% of the variance in social impairment. This suggests that the effect of attentional control on social impairment may operate through additional variables that reflect the multiple ways that attentional control is used in social contexts. Examination of additional variables would likely account for more variance in social impairment. This is consistent with the conceptualization of attentional control mechanisms as a domain-general system that assigns control resources to facilitate multiple higher-order processes (Lesh et al., 2011), as well as literature demonstrating neurocognitive impairments affect functioning via social cognition and metacognition in schizophrenia (Schmidt et al., 2011). Socially relevant processes such as regulation of emotional information, reward processing, and metacognition are three such higher-order processes through which attentional control could indirectly influence social functioning. We discuss each one briefly below and highlight avenues for future research examining the underlying mechanisms of social impairment in SA.

The role of attentional control in the regulation of emotional information may be particularly important for understanding the pathway between attentional control deficits and social impairment. Social interactions by nature involve affectively salient information, thus deficits in the regulation of emotional information could adversely affect response to interpersonal stressors and consequently social functioning (Tully and Niendam, 2014). Our prior research is consistent with this proposal. Impaired attentional control in schizophrenia contributes to negative affective information exerting inappropriate influence on social judgments (Hooker et al., 2011), and failure to recruit neural mechanisms of attentional control predicts maladaptive responses to interpersonal conflict in healthy (Hooker et al., 2010), schizophrenia (Tully et al., 2014), and high SA samples (Hooker et al., 2014). These findings support a model in which attentional control deficits impact social functioning – particularly response to interpersonal conflict – via impaired attentional control of emotional information.

A second pathway through which attentional control could affect social functioning could be via reward processing mechanisms. For example, impaired engagement of attentional control mechanisms to down-regulate negative affective information could be accompanied by a complimentary deficit in the up-regulation of positive affective information, which is thought to underlie the anticipatory pleasure deficit in anhedonia (Pizzagalli, 2010) and could contribute to the associated reward/motivational impairments seen in high SA (Horan et al., 2008). Preliminary evidence is consistent with this proposed role for attentional control in the management of both positive and negative affect (e.g., Vasey et al., 2013). Relatedly, anticipatory pleasure deficits could also be due to an impaired ability to generate representations of the reward value of future pleasurable activities (e.g., socializing with a friend), a process dependent on attentional control functions (Burbridge and Barch, 2007). Thus, the inability to co-opt attentional control mechanisms to facilitate the generation and use of reward representations and positive affect could result in a lack of motivation to engage in pleasurable activity (Germans and Kring, 2000) and consequently negatively impact social functioning.

Finally, there is a role for metacognition – processes involved in the capacity to think about thinking, including forming representations of one’s own mental state (self-reflection) and others’ mental states (ToM; Lysaker et al., 2010a), processes that are important for successful navigation of the social world and may be closely related to attentional control processes (Fernandez-Duque et al., 2000). Impaired metacognition is considered to be characteristic of schizophrenia (Pickup and Frith, 2001) and is observable in high-risk populations (Morrison et al., 2007; Bora and Pantelis, 2013). In schizophrenia, metacognition predicts occupational (Lysaker et al., 2010a), and social functioning (Brüne et al., 2007; Lysaker et al., 2011), and mediates the relationship between neurocognition and social functioning (Lysaker et al., 2010b). Moreover, a recent study found neural activation during ToM mediates between SA and social functioning in schizophrenia participants suggesting that, at least in patient populations, SA partially impacts social functioning through ToM (Dodell-Feder et al., 2014). This may be because difficulty understanding the minds of others causes social interactions to be stressful/unpleasant, resulting in social withdrawal and/or social rejection and consequently poor social functioning (Salvatore et al., 2008). Further research examining putative relationships between SA, ToM, and social impairment are needed.

Clearly, social functioning is complex and multiply determined; a comprehensive model of how SA impacts social functioning would need to incorporate multiple neurocognitive and social cognitive factors (Schmidt et al., 2011). Here we identified attentional control as one proximal mediator of the relationship between SA and social impairment, but there is still much variance to be explained. Future research could include measures of attentional control of emotional information, reward processing, and metacognition in order to examine additional mediators of the relationship between SA and social impairment.

Limitations must be acknowledged. The sole use of self-report assessments in the current study design could have resulted in inflated relationships due to common measurement bias, thus replication samples and follow-up studies using alternative measures are necessary. In particular, the assessment of attentional control using a self-report questionnaire, although partially validated here in a sub-sample of participants and in prior literature, is less than ideal. Future studies employing behavioral assessments of attentional control are necessary. Similarly, the ability of high SA individuals to self-report their level of SA could be confounded by accompanying deficits in self-reflection/metacognition. As discussed above, future research should include measures of metacognitive processes in order to delineate the unique contribution of SA and metacognition to social impairment. Finally, our sample – although representative of the Greater Boston Community – was comprised solely of healthy individuals, warranting caution when generalizing the results to patient populations. Parallel investigations of the relationship between SA and social impairment in patient samples are needed.

It is important to note that the current study is unable to establish causal directions between variables because of its cross-sectional design; a statistically significant mediated effect does not determine the causal direction of a relationship (Preacher and Hayes, 2004). We conducted the current study on a strong theoretical foundation based on prior research demonstrating that SA predicts social impairment (Katsanis et al., 1992; Cohen et al., 2006; Blanchard et al., 2011) and that SA can be partially characterized by attentional control deficits (Giraldez et al., 2000; Barrantes-Vidal et al., 2003), leading us to hypothesize that attentional control is an underlying characteristic of high SA that explains the relationship between SA and social impairment. It is possible that the opposite is true – that social impairment impacts SA via attentional control. However, that this reverse mediation model was not significant lends further support to our theory. Although, it is possible that once social impairments are present there exists a vicious cycle whereby social impairment exacerbates/potentiates SA through some other process (e.g., reward processing, as discussed above), which in turn potentiates social impairment and so on ad infinitum. The question of causality remains, however, and only a longitudinal study design in which attentional control and SA are tracked across the life span in relation to social functioning can truly determine causal priority of the model. The New York High Risk Project partially investigated this in relation to physical anhedonia and attention in a longitudinal follow-up of individuals at genetic risk for schizophrenia (Erlenmeyer-Kimling et al., 1993). Results indicated that poor attentional capabilities as a child predicted physical anhedonia as an adolescent, which in turn predicted social functioning as an adult. It is possible that a similar relationship exists between the development of attentional control, SA, and social functioning. Future research should conduct longitudinal examination of these causal pathways and potential maintenance processes.

The current study demonstrates that the relationship between SA and social impairment is partially mediated by attentional control. This has implications for our understanding of a fundamental human desire, the need to belong, and suggests attentional control is one of the mechanisms necessary for successful social interactions. This finding also illuminates one of the mechanisms underlying the relationship between a well-established negative symptom of schizophrenia and social impairment.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.