We employed an ERP version of the Cyberball paradigm to measure brain activity in children with ASD during simulated social interactions. These are the first data to explore the temporal dynamics of brain activity associated with social exclusion in ASD, revealing a dissociation between reported distress and neural response in children with ASD and a difference in the temporal course of brain responses between children with ASD and typical peers. Both groups reported similar levels of distress in response to social exclusion. These results are generally consistent with those obtained in a prior study of ASD that examined self-reported experience alone (Sebastian et al., 2009
); however, we did not observe the slight decrease in mood evident solely in typical individuals in the prior study. Because we did not administer questionnaires prior to the Cyberball game, we cannot determine whether the comparable mood results between groups in the present study reflect a difference at baseline. Alternatively, modifications to the paradigm to enhance realism (i.e., matching players to participants’ ethnicity, gender, and age, using digitized photos and names, choosing a personal glove for play) may have increased investment on the part of participants with ASD, influencing self-reported mood. To examine this possibility, we created a standardized metric to compare scores by dividing mood scores for each group in each study (Current TD = 25.44, Current ASD = 26.11; Prior TD 4.8, Prior ASD 5.4) by the maximum possible score on the mood scale (Current study = 40; Prior study = 7), essentially creating a score representing proportion of the possible maximum. According to this metric, the typical group mood score was similar in the current study (.64) to the prior study (.69), while the ASD group’s mood score was lower in the current study (.65) than in the prior study (.77). The differential impact on mood in the present study suggests that our paradigm’s visual realism and “in-group” matching may be critical factors in the study of mood modulation by social exclusion. Future research on social cognition in ASD would benefit from specific investigation of such factors.
Despite similarity in self-reported experience, ERPs revealed distinct patterns of neural activity underlying reactions to rejection events during social exclusion for each of the groups. Consistent with earlier work from our group (Crowley et al., 2010
), the neural response of typically developing children was characterized by enhanced negativity, evident in a late slow wave over medial-frontal scalp electrodes, to rejection events versus “not my turn” events. Also in keeping with prior findings, differential amplitude associated with rejection events correlated with self-reported effects on mood and overall ostracism distress. Children with ASD showed a fundamentally distinct pattern of results. The late slow wave did not differentiate rejection from “not my turn” in the ASD group, nor were correlations observed among neural response and the reported experience of distress. In contrast to their typical peers, the children with ASD showed differentiation of rejection at an earlier frontal P2 component, although this response also did not associate with their reported experience of rejection. The absence of differential response to exclusion versus “not my turn” at the LSW suggests that children with ASD might be failing to make critical distinctions based on social context (i.e., the same outcome has a different connotation depending on the context in which it occurs, “not my turn” versus exclusion) at these late processing stages.
Results suggest that children with ASD are processing the experience of rejection differently than typical children. The temporal course of the early positivity (P2) indicates a role in more basic cognitive processing, such as visual attention. The frontal P2 usually appears in visual tasks and is related to selective visual attention (Key, Dove & Maguire, 2005
). We posit that reduced amplitude at the P2 for rejection events suggests attenuated engagement of attentional resources during the experience of exclusion in individuals with ASD. This may reflects the more dramatic impact of social exclusion on an already weakened social motivational system (Dawson, Webb & McPartland, 2005
), i.e. an accelerated acquisition of a sense of learned helplessness. It is understood that children with ASD display reduced orienting to social cues (Dawson, Toth, Abbott, Osterling, Munson, Estes & Liaw, 2004
); these results suggest this failure to discriminate meaningful social information is also evident when social meaning is conveyed by context alone. Furthermore, the selective impact in the context of exclusion implies this vulnerability is exacerbated by negative social experiences; the social attentional mechanisms of children with ASD may be most likely to dysfunction in the very contexts in which they are most vital. In contrast, typically developing children demonstrated preserved attentional engagement during rejection. Given the positive relationship between P2 amplitude during rejection and mood, this appeared to represent a protective factor. We propose that, in this paradigm, P2 represents adaptive engagement of attentional mechanisms to decipher cues relevant to determining social context and initiating the emotional processing indexed by subsequent components. The current results emphasize the importance of temporal dynamics in revealing processing strategies in typical and atypical development. In this instance, our measures of brain function revealed important group differences undetectable with behavioral methods alone. The current work highlights the need to understand the interplay of the forces of social drive and discouragement in ASD. Although it is often presumed that children with ASD possess reduced drive for social interaction (Dawson et al., 2005
), it has long been acknowledged that some children with ASD posses preserved social drive despite insufficient social agility to successfully navigate social interactions (Wing & Gould, 1979
). Current results emphasize that, rather than being invulnerable to social exclusion, children with ASD are also at risk for emotional and psychological consequences of ostracism. Our future work aims to examine the influence of social motivation on the experience of exclusion and to understand development of social motivation in the context of learning from aversive social experiences.
These findings are relevant to understanding emotion regulation in cognitive-behavioral therapy for children with ASD. Our results suggest under-responsiveness associated with brain mechanisms subserving regulation of emotional responses. In our study, children with ASD failed to display differentiation of rejection in the portion of their brain response hypothesized to reflect emotion regulation mechanisms, potentially reflecting ACC activity. Such late components have been posited to reflect facilitated attention to emotional stimuli (Cuthbert, Schupp, Bradley, Birbaumer & Lang, 2000
; Schupp, Cuthbert, Bradley, Cacioppo, Ito & Lang, 2000
) and emotional arousal (Schupp et al., 2000
; Hajcak & Dennis, 2009
) and have been shown to be reduced under conditions of voluntary reappraisal of negative emotion (Moser, Hajcak, Bukay & Simons, 2006
). In considering therapeutic approaches, these findings suggest the potential applicability of interventions addressing both cognitive reappraisal (corresponding to the functions indexed by the P2) and physiological down-regulation (corresponding to the LSW). For example, children with ASD might be taught strategies to reframe exclusion events (e.g., instead of thinking “They will never throw to me”, thinking “I am sure they will throw to me soon” or “I have other friends if they won’t play”). Emotional arousal might be directly addressed through strategies aimed at physiological responses, such as progressive relaxation or taking deep breaths. Work in progress is applying the current experimental paradigm as a tool to apply such targeted treatment approaches and provide an objective indicator of response to treatment. We predict that cognitive reappraisal techniques would differentially target early activity and that more physiologically-oriented methods would be reflected in reduced emotional arousal as indexed by the later component.
A limitation of the current study is its reliance on self-report for the measurement of emotional experience during social exclusion. A fundamental weakness of this methodology is that it inextricably relies on personal estimation of the magnitude of experience and presumes comparability of this metric between individuals. A problem with this assumption is that one individuals’ rating on a Likert scale does not necessarily reflect the same experience in another person; in other words, the ends of the scale may fall at different points in an experiential continuum for different individuals (Bartoshuk, Fast & Snyder, 2005
). This difficulty may be circumvented by assessing direct, objective measures of emotional arousal, such as electrodermal response or startle, as has been used in other social rejection studies (Downey, Mougios, Ayduk, London & Shoda, 2004
). The use of self-report measures is especially problematic in ASD, given possible difficulties with introspection. A second limitation of this study is that, though we speculate that children with ASD experience distress for different reasons, our behavioral questionnaire solely focused on distress associated with the social elements of the task. Given the absence of correlations between ERP measures and self-reported distress in the ASD group, it is possible that a distinct constellation of psychological contributors to distress may be at play in children with ASD. Although there are a number of factors other than social exclusion that could have influenced our results, such as differential rates of learning about probabilistic experiences, we speculate that the different response in ASD may reflect, in part, response to perceived violation of an implicit rule (Bolling et al., 2011
). These possibilities speak not only to experience during the course of the experiment but to broader social experience in ASD. The comments offered by participants during the exclusion portion of the experiment (e.g., “You’re annoying guys”, “I wish I could talk to them”, “I hate being left out”) suggest that irrespective of the cognitive source of ostracism-related distress (e.g., social exclusion, per se, deviation from equivalent probability of outcome, or rule violation), the encounter was subjectively experienced
in social terms. A third limitation of the current study is that because our sample included a limited number of individuals prescribed psychoactive medication, we lacked adequate statistical power to determine differential response to social exclusion in children receiving pharmacotherapy. This is a critical area for future research, as many children on the spectrum are prescribed medications specifically to address mood symptoms that might be affected by and might influence the experience of social exclusion. Finally, as is the case for most prior studies using the Cyberball paradigm, we cannot rule out the possibility that order of block administration may have influenced our results (discussed by Sebastian, Viding, Williams & Blakemore, 2010
). Future research could avert this confound by administering multiple blocks of exclusion, including a control group receiving two blocks of the same type, or by adding a throw type unrelated to exclusion that occurs in both blocks.
We see these findings as valuable, proximally, in terms of specifying the temporal dynamics involved in the neural mechanisms of perceiving and regulating emotional responses to social exclusion in ASD. In the longer term, an understanding of these mechanisms will be necessary for the development and implementation of therapies targeted to specific processes. For example, in the context of exclusion, cognitive-behavioral strategies might differentially address cognitive reappraisal of the perceived slight versus down-regulation of one’s emotional response to the experience; understanding brain function at each stage of this process is a needed step in advancing such interventions for children with ASD. More broadly, from a social neuroscience perspective, because brain regions activated during Cyberball (e.g., ACC, VLPFC, insula) are also implicated in the neuropathology of ASD (Barnea-Goraly, Kwon, Menon, Eliez, Lotspeich & Reiss, 2004
; Di Martino et al., 2009
), examination of their functional integrity is of direct relevance to understanding the neuropathology of ASD and in meaningfully defining subgroups within a heterogeneous phenotype.