Social Discounting in Autism Spectrum Disorder

Background: In social discounting tasks, participants choose between a smaller monetary reward for themselves versus a larger reward for partners of varied social distance (e.g., a sibling versus an acquaintance; Jones & Rachlin, 2006). Participants’ choices thus implicitly measure the subjective value of rewards given to others, capturing constructs such as social closeness or perspective taking. Indeed, variability in social discounting has been linked to externalizing disorders (Sharp et al., 2012), episodic thinking (Yi et al., 2016), and empathy (Olson et al., 2016). Quantifying social discounting rates in ASD may help explain and predict relationships with close and distant others, but, to date, no research has examined social discounting in autism.

Objectives: To assess social discounting in adolescents and adults with ASD.

Methods: Forty-four typical individuals (16 males, average age=20.23 years) and sixteen individuals with ASD (13 males, average age=20.08 years) participated. We also analyzed data from a subset of 16 typical participants matched on age, sex, and verbal IQ. Participants were first told to list four individuals that they knew personally: Partner 1 was the person they were closest to; Partner 2 was someone else they knew well; Partner 3 was a person they knew a little; and Partner 4 was a person who they would recognize, but did not know well. Participants then completed a computerized social discounting task for each partner, understanding the task to be hypothetical. On each of six trials, participants chose between $100 for their partner versus a lower variable amount (titrated between $1.57 and $98.44) for themselves (Figure 1). This algorithm revealed the value at which the participant considered a reward to him or herself to be equal to giving $100 to the partner. Lower values corresponded to increased discounting (e.g., valuing others’ rewards less). As a control, participants completed a delay discounting task, in which they chose between receiving $100 at a future time point versus receiving a lower amount today. This ensured that effects were not driven by task comprehension, general discounting, or understanding of money.

Results: A repeated-measures ANOVA examining social discounting found a main effect of group membership, with ASD participants showing increased discounting. The interaction between group membership and social partner level was also significant (p=.041; Figure 2); ASD participants showed no difference in discounting for Partner 1, but increased discounting for more distant partners (i.e., showed less value for rewards given to these social partners; Partner 2, p=.002; Partner 4, p=.062). Delay discounting also revealed a main effect of group (p=.008), but no interaction between group and time point. Across both tasks, trends were similar but not significant for the matched typical group.

Conclusions: Overall, individuals with ASD showed no difference in social discounting for their closest social partner, but did show quicker drop-offs in the value of rewards given to individuals outside of this most intimate relationship. Additional data are being collected to examine the relation between social discounting and social cognitive abilities, to better assess the causes and consequences of social discounting.