Jaswal & Akhtar (J&A) provide a novel and compelling argument to rethink presumptions about social motivation in youth with autism spectrum disorder (ASD) – that behaviors presumed to be indicative of reduced social motivation in ASD may reflect causes unrelated to motivational factors. For example, they suggest that self-regulation, rather than reduced social motivation, may influence averted eye contact. However, an alternative explanation bears consideration: that both in the target article and in the preexisting literature, social motivation is an imprecisely specified construct (Bottini Reference Bottini2018; Clements et al. Reference Clements, Zoltowski, Yankowitz, Yerys, Schultz and Herrington2018). That is, while social motivation models have proliferated, they have done so in the absence of clear definitions of constructs and specific (and falsifiable) measurement approaches.

J&A focus on social interest and use this interchangeably with the term social motivation. However, drawing from the broad motivation literature, it is clear that “social interest” is but one part of a larger construct of social motivation. Social motivation may be parsed into four components that have largely distinct neurobiological elements, namely, reward motivation (or “wanting”), which is comparable to the social interest term used by J&A, reward processing (or “liking”), reward learning, and habit formation (Berridge & Kringelbach Reference Berridge and Kringelbach2008; Reference Berridge and Kringelbach2011; Graybiel Reference Graybiel2008).

Reward motivation (wanting) is defined as a state of incentive motivation salience that regulates approach behaviors toward biologically relevant goals (Table 1; Berridge & Robinson Reference Berridge and Robinson2003; Berridge et al. Reference Berridge, Robinson and Aldridge2009). Reward motivation (wanting) also encompasses the willingness to expend effort to obtain rewards, and it is referred to as maintenance behavior in the social motivation literature (Chevallier et al. Reference Chevallier, Kohls, Troiani, Brodkin and Schultz2012b). Social wanting, therefore, may be observed in prosocial behaviors (social initiations, i.e., effort expended to maintain social relationships) and in the brain as heightened activity in reward-related mesocorticolimbic regions during the anticipation of social feedback. Reward processing/responsiveness (liking) is associated with the hedonic impact associated with reward acquisition (Tindell et al. Reference Tindell, Smith, Peciña, Berridge and Aldridge2006). Social liking includes subjective, behavioral, and biological components of responses to rewards such as self-reported social pleasure and heightened electrophysiological response to social feedback (e.g., reward positivity or late positive potential event-related potential [ERP]; Stavropoulos & Carver Reference Stavropoulos and Carver2014). Reward learning includes the processes by which behavior is modified when a reward is presented or omitted (Dayan & Berridge Reference Dayan and Berridge2014). Over time, social reward learning is observable behaviorally, for example, when an individual differentially attends to social partners who are friendly and avoids those who treat them poorly. Implicit learning tasks using socially relevant stimuli can elicit social reward learning in a laboratory setting (Scott-Van Zeeland et al. Reference Scott-Van Zeeland, Dapretto, Ghahremani, Poldrack and Bookheimer2010). Habit formation refers to over-learned, largely automatic responses that were once associated with positive hedonic responses but that persist even in the absence of reward presentation (Graybiel Reference Graybiel2008; Smith & Graybiel Reference Smith and Graybiel2014). There are many culturally specific social habits, such as waving or shaking hands when greeting someone, that require effort to learn, but once learned, are largely automatic overtures. The transition from effortful social learning to social habit formation is adaptive in that it frees cognitive resources for other tasks.

Table 1

Summary of the four putative components of social motivation across levels of analysis

J&A suggest the predominant view is that social motivation differences are driven by a lack of social interest (wanting) in individuals with ASD. However, most ASD research does not specify any component as decisively representing social motivation. For instance, the seminal paper on social motivation in ASD (Dawson et al. Reference Dawson, Webb, Wijsman, Schellenberg, Estes, Munson and Faja2005) posits that an absence of preferential processing of social information may be indicative of a social motivation deficit, implicating both social wanting and liking. Other well-cited papers point to various aspects of reward processing and maintenance (e.g., Chevallier et al. Reference Chevallier, Kohls, Troiani, Brodkin and Schultz2012b; Kohls et al. Reference Kohls, Chevallier, Troiani and Schultz2012). Without consensus on the scope of the social motivation construct, J&A's claim of an extant (but differently expressed) desire for social connection among most individuals with ASD may be appropriate – but it may not impact the plausibility of the social motivation model.

To move the field forward, a number of theoretical and practical issues related to the operationalization and measurement of social motivation should be precisely specified. Presently, researchers often use broad self-report measures (such as the Dimensions of Mastery Questionnaire; Morgan et al. Reference Morgan, Wang, Barrett, Liao, Wang, Huang and Jozsa2015) or the Social Responsiveness Scale (2nd ed.; Constantino Reference Constantino2012) to measure social motivation. Though these measures are easy to administer, and they do not differentiate components of social motivation. Task-based paradigms that can be paired with EEG and/or fMRI are helpful in elucidating the neural correlates of social motivation. Although these measures may be more effective in differentiating components of social motivation, they have limited ecological validity. Thus, we suggest three improvements: First, the assessment of social motivation should be multimethod to investigate each of the four components, combining survey, behavioral, and neural measures. Second, assessment of social motivation should also be multi-informant, incorporating the insight of the individual with ASD as well as outside reporters (e.g., parents, teachers, and clinicians) to provide information about social motivation in different contexts. Third, when reporting results, researchers should specify which of the four components are being investigated to promote clarity and consistency.

We appreciate the novel approach J&A took to the reconceptualization of behaviors frequently assumed to stem from a lack of “social interest” in individuals with ASD. We agree that differences in the expression of social interest should not necessarily indicate a complete absence of social motivation. Their focus on social interest as a proxy for social motivation is indicative of the inconsistency in the field over how to define social motivation. Drawing from the broader motivation literature, we set out four components that may contribute to a larger social motivation construct. In doing so, we clarify that heightened social interest may coexist with attenuated levels of the other three social motivation components. Social motivation as a construct is complex and difficult to operationalize. To better understand the ways in which social motivation is expressed – and may or may not be impacted – in individuals with and without ASD, it is essential for the field to converge on specific definitions of components of this construct.