The dynamics of idealized attention in complex learning environments

Abstract

Effective allocation of attention is crucial for many cognitive functions, and attentional disorders (e.g., ADHD) negatively impact learning. Despite the importance of the attentional system, the origins of inattentional behavior remain hazy. Here we present a model of an ideal learner that maximizes information gain in an environment containing multiple objects, each containing a set amount of information to be learned. When constraints on the speed of information decay and ease of shifting attention between objects are added to the system, patterns of attentional switching behavior emerge. These predictions can account for results reported from multiple object tracking tasks. Further, they highlight multiple possible causes underlying the atypical behaviors associated with attentional disorders.