🤖 AI Summary
Existing cognitive modeling largely relies on idealized environmental assumptions, neglecting how dynamic environmental perturbations affect stimulus–response mapping. To address this, we propose the first unified framework integrating the Drift Diffusion Model (DDM) with Deep Reinforcement Learning (DRL), explicitly capturing the temporal dynamics of environmental stimuli and enabling both subject- and stimulus-specific modeling. Our approach incorporates dynamic environment representation and fine-grained behavioral temporal analysis. Evaluated across multi-task, multi-interference scenarios, the framework significantly improves prediction accuracy of cognitive responses: quantitatively outperforming state-of-the-art baseline models; qualitatively reproducing key human cognitive phenomena—including reaction time distributions and decision biases—with high fidelity. This work establishes a novel paradigm for interpretable and robust cognitive modeling in complex, real-world settings.
📝 Abstract
Using deep neural networks as computational models to simulate cognitive process can provide key insights into human behavioral dynamics. Challenges arise when environments are highly dynamic, obscuring stimulus-behavior relationships. However, the majority of current research focuses on simulating human cognitive behaviors under ideal conditions, neglecting the influence of environmental disturbances. We propose CogReact, integrating drift-diffusion with deep reinforcement learning to simulate granular effects of dynamic environmental stimuli on human cognitive process. Quantitatively, it improves cognition modelling by considering temporal effect of environmental stimuli on cognitive process and captures both subject-specific and stimuli-specific behavioural differences. Qualitatively, it captures general trends in human cognitive process under stimuli, better than baselines. Our approach is examined in diverse environmental influences on various cognitive tasks. Overall, it demonstrates a powerful, data-driven methodology to simulate, align with, and understand the vagaries of human cognitive response in dynamic contexts.
