The adult autism phenotype reflects compounded and compensatory processes as much as the original core biological atypicalities (Johnson et al. Reference Johnson, Jones and Gliga2015). Therefore, one cannot ask whether autism is characterised by a lack of social motivation without taking a lifelong approach and accounting for the high degree of heterogeneity in developmental pathways to the adult phenotype. To identify causal mechanisms that may allow for effective early interventions, one has to investigate early development.

Prospective studies of younger siblings of children with autism spectrum disorder (ASD) – henceforth, infant sibs – have been doing just that (Jones et al. Reference Jones, Gliga, Bedford, Charman and Johnson2014). About 20% of younger siblings develop ASD themselves (Ozonoff et al. Reference Ozonoff, Young, Carter, Messinger, Yirmiya, Zwaigenbaum, Bryson, Carver, Constantino, Dobkins, Hutman, Iverson, Landa, Rogers, Sigman and Stone2011). Because investigating social motivation in infancy cannot rely on self-reports, we design experimental paradigms that measure valid constructs of what human adults experience as rewarding interaction with others. Similar measurement challenges faced research using animal models to elucidate the biological bases of motivated behaviour (Berridge et al. Reference Berridge, Robinson and Aldridge2009). The critical contribution of animal research was to identify objective signatures of reward systems activation. These signatures may vary in manifestation but converge across species in terms of the function they serve and the neural substrates. For example, “wanting” signatures are behaviours that allow an individual to seek and consume or engage with the reward and may take the form of pressing a lever for food or pointing imperatively to food. “Liking” signatures indicate the hedonic value associated to the anticipation or the consumption of reward and may manifest as facial expressions or physiological reactions.

We (and others) designed paradigms that offer infant sibs the opportunity to seek social stimulation through the simplest behaviour available to them, eye movements. This research showed that in their first year of life, infant sibs who later developed ASD prefer to look at faces over other objects (Elsabbagh et al. Reference Elsabbagh, Gliga, Pickles, Hudry, Charman and Johnson2013). When scanning a face, they distribute their looking to the eyes and mouth in a similar manner as control participants (Elsabbagh et al. Reference Elsabbagh, Bedford, Senju, Charman, Pickles and Johnson2014; Jones & Klin Reference Jones and Klin2013). Just as typically developing infants do, they gradually decrease their looking to faces and eyes over the second year of life to explore other environmental cues that are critical for learning (Jones & Klin Reference Jones and Klin2013; Ozonoff et al. Reference Ozonoff, Iosif, Baguio, Cook, Hill, Hutman, Rogers, Rozga, Sangha, Sigman, Steinfeld and Young2010). Further, in an experimental situation which we directly tested social motivation, infant sibs who later developed ASD were again indistinguishable from controls (Vernetti et al. Reference Vernetti, Senju, Charman, Johnson and Gliga2018). In this study 24-month-olds received either social or non-social stimulation, contingent on whether they looked at an image of a person (who turned towards them and smiled) or of a mechanical toy (which spinned). Sibs who later developed ASD showed a range of typical signatures of socially motivated behaviour: They oriented to and maintained more attention to faces than to toys (the wanting component) and smiled more in response to social interaction (the liking component).

Taken together, these findings support the claims made by Jaswal & Akhtar (J&A), by showing that early in development ASD is not characterised by a lack of motivation to engage with social stimuli (Elsabbagh & Johnson Reference Elsabbagh and Johnson2016). However, more refinement in these claims is needed because these same studies show that social behaviour is both context-dependent and variable among individuals.

In Vernetti et al. (Reference Vernetti, Senju, Charman, Johnson and Gliga2018) all participating toddlers preferred social interaction when they could control it (the contingent condition). However, typically developing but not autistic toddlers preferred a more naturalistic social stimulus: a person who sometimes responded non-contingently and addressed the child in various ways, over a highly predictable stimulus (Vernetti et al. Reference Vernetti, Senju, Charman, Johnson and Gliga2018).

Prominent interindividual variability was also observed in early trajectories in most infant sibs studies. In one study, paradoxically, some infants in the group who later developed autism showed greater fixation to eyes than control subjects (Jones & Klin Reference Jones and Klin2013). This variability is important as it predicts later developmental outcomes. When exploring complex scenes requiring more endogenous control of attention, increased scanning of the mouth relative to the eyes in the first year of life predicted a subgroup with a more advanced language profile later in toddlerhood (Elsabbagh et al. Reference Elsabbagh, Bedford, Senju, Charman, Pickles and Johnson2014).

Is social motivation typical early in development in ASD? We believe the answer has to be yes and no. Whether a type of stimulus is rewarding or not is highly dependent on the context and on individual differences. For example, many rewards lose their value if they are delayed or unavailable (Blechert et al. Reference Blechert, Klackl, Miedl and Wilhelm2016). We suggest this is what may happen over development, in ASD, as the demands of social interaction become more complex and the rewards associated with social exchanges fail to materialize, the motivation to take part in those interactions decreases. However, when finding themselves in an optimal social environment, social motivation may surge again.

Thus, we agree with J&A that future research needs to characterise the nature of these optimal environments – that is, those in which autistic individuals will truly be, not just appear, socially motivated. However, we disagree with the suggestion that this requires revisiting the classical way of measuring motivation. We observed social motivation in ASD using the same criteria suggested to index the activation of reward systems – that is, seeking, maintaining, and enjoying the reward. We achieved this by making use of a simple behaviour, with minimal requirement in terms of motor abilities: eye movements. The nature of the behaviour that brings about the reward was never key to theories of motivated behaviour. Whatever other behaviours are employed to satisfy an existing social drive in ASD (e.g., echolalic speech or instrumental gestures) will have to fulfil the same criteria; research will have to demonstrate they are employed to bring a social stimulus closer and maintain engagement with it.