Abstract
Foraging is a type of visual search in which observers look for an undetermined number of targets in a noisy environment. It is a ubiquitous task in human and animal daily life that has received a growing researching interest in the last decade, specifically in human behavior. Foraging in the real world is usually based on non-exhaustive search where the observer is free to abandon the search at will (e.g. berry picking, lego-blocks searches, etc.). Diffusion models have been widely used for modeling many cognitive tasks involving one or two-choice decision processes, including visual search. Here we present a two-stage model based on diffusion processes in the foraging domain with humans: A first stage is based on modeling the decision whether the observer determines if a given item is either a target or a distractor. The second stage models the one-choice decision on when to leave a certain patch/display of stimuli to start a new search or just to quit the search. We present simulation results and applications to experimental data on a large sample of almost 300 participants of different ages (from 4 to 25 years old). Observers must look for different targets in feature and conjunction foraging tasks within a controlled video game-like task. The results show an adequate fit to parameter values consistent with previous diffusion models applications, as well as sensitivity to changes in experimental conditions. The present model provides an innovative way to improve our understanding of foraging research in humans under experimental-controlled conditions by the application of diffusion models within a more mathematical modeling approach.