基于证据积累的认知决策神经网络模型

Reaction time (RT) is a window into understanding human decision-making processes. The Evidence Accumulation Model (EAM) is the dominant theory of computational framework for modeling RT. However, EAMs, such as the Drift Diffusion Model (DDM), offer statistical descriptions of decision outcomes but without detailed algorithms for stimulus encoding or neural mechanisms, thereby omitting the algorithmic and hardware levels in David Marrs three-level framework (computation, algorithm, and hardware). We suggest that these limitations can be addressed by combining Artificial Neural Networks (ANNs) and evidence accumulation model to simulate the entire decision-making processfrom stimulus encoding to evidence accumulation and decision output. These new models, termed Cognitive Decision Neural Networks, enable comprehensive modeling of human decision-making on non-biological hardware (in silico), providing a novel approach to understanding cognitive processes. Cognitive Decision Neural Networks have demonstrated preliminary potential in multi-option decision-making, temporal stimulus processing, and neural activation simulation. Such models provide a novel approach to simulating the full spectrum of human decision-making processes. In the future, integration with digital twin brain models could extend their applicability to more complex decision-making scenarios, thereby advancing a deeper understanding of human cognition.