A Hierarchical Model of Attention over Time

A key feature of attention is that it moves over time, guided both exogenously by changing external circumstances and endogenously by internal cognitive states. However, the endogenous mechanisms guiding the movement of attention in the absence of external cues remain poorly understood. This paper develops and validates a computational model of how internal attentional states, motivated by the Adaptive Gain Theory of the locus coeruleus-norepinephrine (LC-NE) system, can guide the movement of visual attention over time. By fitting our model to young children's gaze data as they perform a visual object tracking task, we investigate developmental changes in higher-order patterns of attending behavior between 3.5-6 years of age, and we hypothesize about how the LC-NE system might mediate these changes.