The progression of visual search in multiple item displays: First relational, then feature-based

TL;DR

This study demonstrates that visual attention initially guides to relationally maximal items in a display, then sequentially to next-best items, supporting a rapid, context-independent relational search process in visual attention.

Contribution

It provides evidence that relational search in visual attention occurs spontaneously and rapidly, even without prior context learning or extensive training.

Findings

Relational search occurs on first fixations without prior context.

Attention sequentially shifts from relationally maximal to next-maximal items.

Relational tuning is strengthened by context learning but is not dependent on it.

Abstract

It is well-known that visual attention can be tuned in a context-dependent manner to elementary features, such as searching for all redder items or the reddest item, supporting a relational theory of visual attention. However, in previous studies, the conditions were often conducive for relational search, allowing successfully selecting the target relationally on 50% of trials or more. Moreover, the search displays were often only sparsely populated and presented repeatedly, rendering it possible that relational search was based on context learning and not spontaneous. The present study tested the shape of the attentional tuning function in 36-item search displays, when the target never had a maximal feature value (e.g., was never the reddest or yellowest item), and when only the target colour but not the context colour was known. The first fixations on a trial showed that these displays still reliably evoked relational search, even when participants had no advance information about the context and no on-task training. Context learning further strengthened relational tuning on subsequent trials, but was not necessary for relational search. Analysing the progression of visual search within a singe trial showed that attention is first guided to the relationally maximal item (e.g., reddest), then the next-maximal (e.g., next-reddest) item, and so forth, before attention can hone in on target-matching features. In sum, the results support two tenets of the relational account, that information about the dominant feature in a display can be rapidly extracted and used to guide attention to the relatively best-matching features.