Higher-level effects in crowding falter when low-level attributes are controlled for

Josephine Reuther, Ramakrishna Chakravarthi

Research output: Contribution to conferenceAbstractpeer-review

Abstract

Visual crowding impairs recognition of flanked objects in the periphery. Further, it is known that low-level feature differences between such objects (e.g., colour) weakens crowding. Recently, higher-level differences have also been posited to similarly modulate crowding: crowding has been shown to be weaker if objects differ in category membership or if the objects are familiar (meaningful). However, these effects of category and familiarity can be deemed to be higher-level influences only if all low-level differences between stimuli are controlled for, which has not been the case in previous research. In a series of experiments, we investigated the effects of a range of low-level attributes on higher-level crowding interactions. We found that the ostensible category
effect was fully accounted for by differences in features and overlap between target and flanker stimulus sets. However, the effect of familiarity was persistent even when low-level attributes were controlled for. Thus, at least some of the so-called higher-level crowding interactions are a result of inadequate stimulus control, while others seem robust. Therefore, we strongly encourage caution and a thorough control of stimulus-level properties before higher-level effects are claimed to influence crowding.
Original languageEnglish
Pages182-183
Number of pages2
Publication statusPublished - 2015
Event38th European Conference on Visual Perception (ECVP) 2015 Liverpool - Liverpool, United Kingdom
Duration: 23 Aug 201523 Aug 2015

Conference

Conference38th European Conference on Visual Perception (ECVP) 2015 Liverpool
Country/TerritoryUnited Kingdom
CityLiverpool
Period23/08/1523/08/15

Fingerprint

Dive into the research topics of 'Higher-level effects in crowding falter when low-level attributes are controlled for'. Together they form a unique fingerprint.

Cite this