Wanting It Too Much: The Unexpected Effect of Social Motivation on Emotion Recognition

Background: Deficits in facial emotion recognition (FER) have sometimes been found to be key challenges in autism spectrum disorders (ASD; Harms et al., 2010), yet the source and nature of these deficits remain unknown. The social motivation theory posits that deficits in social ‘wanting’ may be one predictor of FER ability in individuals with ASD (Dawson et al., 2005; Kohls et al., 2012), though this relation has not been examined. Additionally, individuals with ASD show categorical delays in early electrophysiological processing of faces (specifically latency of N170 ERPs; McPartland et al., 2011). Moreover, these delays correlate with impairments in FER, as evidenced by slower latencies in N170, especially with subtle (Wong et al., 2012) and adult faces (Lerner et al., 2013). Consequently, this ERP may provide insight into the neurocognitive processes facilitating a link between FER and social motivation.

Objectives: The aim of this study is to determine the relationships between social motivation, FER and the N170. We hypothesize that social motivation will be positively correlated with better FER performance and this correlation will be driven by low intensity and adult faces. Furthermore, social motivation will be positively correlated with faster N170 latencies and N170 latencies will moderate the relationship between social motivation and FER performance.

Methods: Participants included 34 well-characterized adolescents (26 male; M_age = 13.07, SD = 2.07) with ASD with intact cognitive ability. Parents completed the Dimensions of Mastery Questionnaire (DMQ), measuring social motivation. While, participants completed the Diagnostic Analysis of Nonverbal Accuracy-2 (DANVA-2), a well-established FER measure, ERP data, including the N170, were recorded.

Results: DMQ and DANVA-2 were significantly correlated (r = -.40, p = .016), indicating that greater social motivation predicted poorer emotion recognition. Post-hoc probing demonstrated that these effects were driven by the relationship between DMQ and low intensity (r = -.46, p = .003) and child faces (r = -.39, p = .011), and these results maintained after controlling for ASD severity (ADOS score) and IQ. N170 latency did not correlate with DMQ or overall DANVA-2, though it correlated with the DANVA-2 child faces subscale (r = .29, p = .05). To examine whether there is shared variance in the relations between DMQ, N170 latency, and DANVA-2 child faces relationships, a multiple regression was run predicting DANVA-2 child scores; this analysis suggested no reduction in either effect after control of the other, indicating two parallel pathways to emotion perception in child faces.

Conclusions: Contrary to our hypothesis, social motivation evinced a negative relationship with FER. These findings indicate that some individuals with ASD may be so highly socially motivated that they fail to attend to important (especially subtle and child peer-initiated) emotional cues in others. Thus, rather than being essential to adaptive outcomes in ASD (Kohls et al., 2012), social motivation may have important unanticipated consequences. Notably, while N170 latency revealed a modest relationship with FER, this effect appeared to be independent of the relationship with social motivation, suggesting that FER ability may be an equifinal and multifaceted construct in this population.