Neural Mechanisms of Reward Processing In ASD

Background: Performance monitoring refers to on-going cognitive processes evaluating the success or failure of an individual’s actions. It occurs during feedback appraisal and reward processing and entails error detection to optimize future behavior. The temporal resolution of event-related potential (ERP) research reveals distinct components of this process, presumed to originate from anterior cingulate cortex. The error-related negativity (ERN) occurs ~50-120 ms after an error commission and reflects self-perceived inconsistency between intended and actual outcomes. The feedback related negativity (FRN) occurs at ~250 ms in response to external feedback indicating an outcome worse than expected. Both the ERN and FRN have been applied to study reward processing in ASD. Most published work investigates the ERN, revealing a tendency toward attenuated intrinsic performance monitoring that correlates with degree of social impairment. A single study has investigated neural response to extrinsic feedback in ASD; Larson and colleagues (in press) compared FRN evoked by wins versus losses in a guessing game and found comparable FRN amplitudes among children with ASD and typically developing counterparts.

Objectives: The current study examined extrinsic feedback, indexed by the FRN, under conditions that more closely approximate the subtle feedback that characterizes social interactions (reward versus non-reward). Toward this end, children with ASD and typical counterparts completed a guessing game in which choices resulted in monetary gain or no reward. We predicted that in response to less concrete extrinsic feedback, children with ASD would display attenuated FRN.

Methods: ERPs were recorded from 15 high-functioning children with ASD and 15 age- and IQ-matched typically developing controls using a 128 electrode Geodesic Hydrocel Net. Participants performed a computerized feedback-reward task in which they selected one of four balloons. Outcomes were determined randomly and entailed monetary gain or no reward. FRN was extracted over frontocentral leads over a time window extending from 200 ms to 400 ms. Behavioral measures assessed social anxiety (SASC-R) and social personality traits (EPQ-Jr).

Results: Typically developing children showed an expected difference in FRN amplitude in response to feedback; neural responses of those with ASD did not discriminate between positive versus neutral outcome. Correlational analyses revealed relationships between FRN and social behavior in children with ASD.

Conclusions: Prior work demonstrates normative responses to extrinsic feedback regarding positive versus negative outcomes in children with ASD. Current findings show reduced sensitivity to more subtle feedback indicating positive versus neutral outcomes. We demonstrate, for the first time, that reward processing in ASD is contingent upon the nature of extrinsic feedback, offering insight into vulnerabilities in social learning in ASD. Experiential learning in social contexts may be negatively impacted in ASD by performance monitoring mechanisms reliant upon highly valenced outcomes rarely evident in social interactions. Results inform understanding of social learning processes in ASD and suggest potential utility of intervention strategies designed to tailor feedback on social performance to functional strengths in performance monitoring.