Statistically defined visual chunks engage object-based attention

Gábor Lengyel,József Fiser,Márton Nagy

doi:10.1038/s41467-020-20589-z

Gábor Lengyel, József Fiser + Show 1 more

Open Access

https://doi.org/10.1038/s41467-020-20589-z

Copy DOI

Abstract

Although objects are the fundamental units of our representation interpreting the environment around us, it is still not clear how we handle and organize the incoming sensory information to form object representations. By utilizing previously well-documented advantages of within-object over across-object information processing, here we test whether learning involuntarily consistent visual statistical properties of stimuli that are free of any traditional segmentation cues might be sufficient to create object-like behavioral effects. Using a visual statistical learning paradigm and measuring efficiency of 3-AFC search and object-based attention, we find that statistically defined and implicitly learned visual chunks bias observers’ behavior in subsequent search tasks the same way as objects defined by visual boundaries do. These results suggest that learning consistent statistical contingencies based on the sensory input contributes to the emergence of object representations.

Highlights

Objects are the fundamental units of our representation interpreting the environment around us, it is still not clear how we handle and organize the incoming sensory information to form object representations
We started with an implicit learning paradigm called visual statistical learning (VSL), which uses a set of artificial shape stimuli to create novel scenes (Fig. 1a, VSL - Block 1)
If the shape-pairs that could only be learned from the co-occurrence probabilities of the shapes during VSL blocks behave to objects, the letter search should be facilitated in this setup by the chunks the same way as it would be by contour-based objects

Summary

Introduction

Objects are the fundamental units of our representation interpreting the environment around us, it is still not clear how we handle and organize the incoming sensory information to form object representations. If consistent statistical properties acquired by learning are fundamental in forming object representations, a set of newly learned arbitrary statistical contingencies, even if they are not connected to traditional cues and even if they are learned implicitly, should manifest the same kind of object-based behavioral-cognitive effects as true objects do To test this hypothesis, we started with an implicit learning paradigm called visual statistical learning (VSL), which uses a set of artificial shape stimuli to create novel scenes (Fig. 1a, VSL - Block 1). The low-level contrast edges, texture transitions, or Gestalt structures that can be important in forming classical object boundaries[12,17,26] cannot reveal the statistical structure of the chunks in these scenes Since these chunks are defined by stable statistical contingencies, according to our hypothesis, they qualify as newly learned objects, and they should induce object-based perceptual effects. Both experiments provided clear evidence that recently and implicitly learned statistical chunks without any visual boundary defined by luminance or other traditional cues elicited the same object-based effects as objects with explicit boundaries did

Methods

Results

Conclusion