Modeling and Recovering Social Reward Differences in Autism with Multi-Agent Reinforcement Learning

The clinical assessment of Autism Spectrum Disorder (ASD) relies heavily on observing dyadic interaction, yet existing computational models are overwhelmingly single-agent frameworks that fail to account for reciprocal social dynamics. Here, we operationalize the Social Motivation Theory of ASD in a multi-agent reinforcement learning (MARL) environment, providing a testbed to evaluate how alterations in social reward shape movement patterns. By varying a scalar social reward parameter, $\rho$, across eight conditions from baseline ($\rho =1.0$) to profound deficit ($\rho =0.01$), joint policy optimization with Proximal Policy Optimization (PPO) generates behavioral phenotypes consistent with ASD without prescribing top-down behavioral rules. Mean peer distance provides a sensitive metric that scales inversely with $\rho$ (Spearman $r=-0.952$, $p<0.001$). Control experiments confirm that this effect is specific to social reward attenuation rather than a general reduction in reward magnitude, consistent with the atypical interpersonal distance reported in clinical cohorts. To determine whether these kinematic signatures are informative, we train a bidirectional LSTM on trajectory sequences. The network recovers the latent parameter $\rho$ at $5.06\times$ chance accuracy across nine classes. This recovery remains well above chance after removing explicit relational distance features and improves with extended observation windows. Additionally, mixed-$\rho$ pairings show directionally consistent coordination degradation, offering a preliminary reinforcement-learning account of the Double Empathy Problem. These findings demonstrate that the social reward differences hypothesized in ASD can be modeled in simulation and recovered from dyadic kinematic data.