Reward Learning Influences Social Reciprocity: The Impact of Autistic and Behavioural Inhibition Traits

Background: Autism Spectrum Conditions are characterized by social deficits that often co-occur with symptoms of behavioural inhibition and social anxiety (White et al, 2009). Behavioural interventions based on reward learning (e.g. ABA therapy) have been found to be effective on improving these deficits (Zachor et al, 2007). However, the efficacy of such interventions shows considerable inter-individual variability (Howlin et al, 2009) and is lower for social avoidant individuals (Ingersoll et al, 2001). It is not clear whether autistic or behavioural inhibition traits modulate the sensitivity to a) reward learning or b) the ability to translate the learned associations to social behaviours.  

Objectives: In this study we tested: a) if autistic or behavioural inhibition traits could modulate the ability to implicitly associate a reward value to a social stimulus (reward learning/conditioning); b) if the learned association could modulate participants’ social behavior (i.e. social reciprocity); c) if the strength of this modulation was mediated by traits of autism or behavioural inhibition. 

Methods: 18 neurotypical adults (8 Males) were administered with an evaluative conditioning paradigm used to associate high or low reward values with neutral target faces. Subsequently, participants performed a ball-tossing game with the high-reward (PosFace) and low-reward (NegFace) faces. PosFace and NegFace reciprocated participant’s ball-tossing with the same probability (50%). To test the impact of reward conditioning on social reciprocity, we divided the number of tosses directed to the PosFace by the number of tosses directed to the NegFace. Subsequently we used the Implicit Association Task (IAT) to assess the strength of conditioning (i.e. the strength of the association between PosFace and win-pictures and NegFace and loss-pictures). All participants completed the Autism Spectrum Quotient (AQ) and Behavioural Inhibition System (BIS) questionnaires online. 

Results: 1) Overall, we found an effect of conditioning on IAT scores. Incongruent blocks (PosFaces paired with loss-items; NegFace paired with win-items) required longer reaction times then congruent blocks (PosFace paired with win-items; NegFace paired with lose-items).  This suggests that there was a significant effect of reward learning. This effect was not correlated to AQ or BIS.

2) We found that the number of ball tosses to the PosFace compared to the NegFace in the ball-tossing game was positively correlated with strength of conditioning (r=.524) and negatively correlated with the BIS scores (r=-.714). This suggests that reward learning has an effect on the extent of social reciprocity. No effect of AQ was found.

3) BIS scores were found to moderate the effect of conditioning on social reciprocity (β=2.4). This suggests that more inhibited the participants were, the less did the learned reward value of the face have an impact on social reciprocity.  

Conclusions: Our data show that neither AQ nor behavioural inhibition traits impact the ability of associate a reward value to a social stimulus. Nevertheless, high inhibition traits seem to impair the ability to translate this reward value to social reciprocity. This has potential implications for autism therapy, in that behavioural therapies relying on conditioning techniques may be less effective for autistic individuals who show high behavioural inhibition.