Semantically incongruent objects attract eye gaze when viewing scenes for change

Abstract

Past research has shown that change detection performance is often more efficient for target objects that are semantically incongruent with a surrounding scene context than for target objects that are semantically congruent with the scene context. One account of these findings is that attention is attracted to objects for which the identity of the object conflicts with the meaning of the scene, perhaps as a violation of expectancies created by earlier recruitment of scene gist information. An alternative account of the performance benefit for incongruent objects is that attention is more apt to linger on incongruent objects, as perhaps identifying these objects is more difficult due to conflicting information from the scene context. In the current experiment, we present natural scenes in a change detection task while monitoring eye movements. We find that eye gaze is attracted to these objects relatively early during scene processing. ARTICLE HISTORY

Extra-foveal Processing of Object Semantics Guides Early Overt Attention During Visual Search

“…Evidence of semantic relatedness on early overt attention also directly speaks about the degree of semantic processing that may happen outside the fovea. Conventionally, the visual field is characterised by three regions, going from the centre to the periphery of the retina: (1) the fovea, which subtends a visual angle of 1°eccentricity and is responsible for high resolution vision; (2) the parafovea, which stretches out to 4-5°; and (3) the periphery, which extends beyond the parafovea and cover the rest of the visual field (see Larson & Loschky, 2009 for provising a brief summary in the context of scene gist recognition). Although the visual acuity strongly decreases in the parafovea and in the periphery, i.e., in extra-foveal vision (e.g., Strasburger, Rentschler, & Jüttner, 2011), the area of the visual field from which observers can accrue useful information is quite large (see Rayner, 2014;Rosenholtz, 2016;B.…”

The Effect of Consistency on Short-Term Memory for Scenes

“…In early 1970s, Biederman and colleagues conducted a series of studies showing that coherent context facilitated object identification ( Biederman, 1972 ; Biederman et al, 1974 ) and visual search ( Biederman et al, 1973 ) compared to scrambled context. The conclusion was confirmed by many later studies (for review, see Bar, 2004 ; Oliva and Torralba, 2007 ; Zimmermann et al, 2010 ; LaPointe and Milliken, 2016 but see Hollingworth and Henderson, 1998 , 1999 ). Other studies directly compared the processing of objects in consistent versus inconsistent scenes and found that objects are recognized faster and more accurately when foreground objects are consistent with background contexts than when they are inconsistent ( Biederman et al, 1982 ; Boyce et al, 1989 ; Davenport and Potter, 2004 ; Davenport, 2007 ; Joubert et al, 2008 ; Mudrik et al, 2010 ), a phenomenon called “scene consistency effect.” For example, Davenport and Potter (2004) adopted color pictures of scenes to examine the effect of consistency on a naming task.…”

“…Intriguingly, most studies demonstrated that objects in consistent scenes, though having an advantage in identification, showed a disadvantage in capturing attention compared to objects in inconsistent scenes. First, it is well-established by eye movement studies that scene consistency influences attention orienting, with inconsistent scenes being fixated earlier and for a longer time than consistent scenes ( Loftus and Mackworth, 1978 ; De Graef et al, 1990 ; Underwood and Foulsham, 2006 ; Brockmole and Henderson, 2008 ; Underwood et al, 2008 ; Spotorno et al, 2015 ; LaPointe and Milliken, 2016 ). For example, Brockmole and Henderson (2008) added new objects to real-world scenes during a fixation or during a saccade.…”

Fixation-related brain activity during semantic integration of object-scene information

“…processing of object-scene semantics comes from studies in which inconsistent objects were selected for fixation earlier than consistent ones (e.g., LaPointe & Milliken, 2016;Mackworth & Morandi, 1967;Underwood et al, 2008). However, other studies have not found evidence for earlier selection of inconsistent objects (e.g., De Graef et al, 1990;Henderson et al, 1999;Võ & Henderson, 2009.…”

“…However, results from eye-movement studies on this issue have been mixed. A number of studies have indeed reported evidence for an inconsistent object advantage (e.g., Bonitz & Gordon, 2008;LaPointe & Milliken, 2016;Loftus & Mackworth, 1978;Underwood, Templeman, Lamming, & Foulsham, 2008). Among these studies, only Loftus & Mackworth, 1978 have reported evidence for an immediate extrafoveal attentional capture (i.e.…”