Ensemble Perception: A Review of the Field

Abstract

Ensemble perception refers to the ability of an observer to precisely estimate summary statistics of multiple objects (average, range, numerosity, etc.) at a glance. This article reviews the properties and research methodology of ensemble perception. Further, we consider the theoretical debate around mechanisms of information sampling and summary statistics calculation. One theory suggests a coarse, parallel and exhaustive mechanism, whereas another theory assumes high-precision processing of a small subsample of items to accomplish proxy statistics for the entire ensemble. We describe the evolving view of the internal ensemble representation that initially was viewed as a single magnitude (e.g., average) but later thought of as the entire feature distribution of all items. We also discuss the role of ensemble representations in various perceptual tasks. Finally, we describe potential neural correlates and neurally plausible models of ensemble perception.