Habit Formation And The ACT-R: A Model Based on Perceptual Dynamics and Attentional Control

This paper presents a computational model to explain habit formation based on perceptual and attentional dynamics. The proposed model shows that habitual behavior arises when perceptually or attentionally salient cues are available in an environment, but only if they can be relied upon to elicit desired responses from participants. When no such cue is present, people engage in more controlled processing, even though the outcome may not turn out as desired due to interference between different tasks or goals for which there exist multiple associations with various cues. We discuss various Multibuffered extensions of the ACT-R Architecture and present a dynamic way of modelling cognitive processes with an emphasis on Attention and Fluency and how some of these patterns get repeated in the presence of a nonlinear environment. The goal of the paper is to argue that these phenomena are essential features that should be considered in models of cognitive systems and that they can provide a good explanation for a wide range of observations of how habitual behaviours arise.