The initial stage of visual selection is controlled by top-down task set: new ERP evidence
Ulrich Ansorge
0
Monika Kiss
0
Franziska Worschech
0
Martin Eimer
0
0
Ulrich Ansorge, Fakultt fr Psychologie,
Universitt Wien
,
Austria
;
Institut fr Kognitionswissenschaft, Universitt Osnabrck
,
Germany
;
Birkbeck College
,
London, UK
; Franziska Worschech,
Institut fr Kognitionswissenschaft, Universitt Osnabrck
,
Germany
;
Birkbeck College
,
London, UK
; Monika Kiss,
Birkbeck College
,
London, UK
; Martin Eimer,
Birkbeck College
,
London, UK
1
) Faculty of Psychology, Universitt Wien
, Liebiggasse 5, 1010 Wien,
Austria
Salient visual singleton stimuli produce spatial cueing effects indicative of attentional capture only when they match current task sets, suggesting that capture is subject to top-down control. However, such task-set contingent capture effects could be associated with the top-down controlled disengagement of attention from nonmatching stimuli that follows their initial bottom-up salience-driven selection. Using the N2pc component as an event-related potential marker of attentional capture, we demonstrate that top-down task set already controls the initial rapid selection of salient visual singleton stimuli prior to any subsequent attentional disengagement. These findings provide new evidence for the primacy of top-down control over bottom-up salience in attentional capture.
-
Our visual system has evolved to detect and identify
currently relevant objects. Attentional mechanisms play a
critical role in visual cognition because they determine
which of the many objects that are simultaneously present
in the visual world are selected as targets for eye
movements, in-depth analysis, and identification. A critical
question is whether the attentional selection of visual
objects is fully under endogenous top-down control or
whether properties of the visual world dictate which objects
are selected. Some have claimed that top-down control
drives the allocation of attention to different parts of the
visual input from the very start (cf. Bichot, Rossi, &
Desimone, 2005). Others have argued that endogenous
control of attentional selection can only be exerted after an
initial phase during which visual stimuli attract attention in
an exogenous bottom-up fashion that is determined by
salience alone (cf. Bergen & Julesz, 1983; Itti & Koch,
2001).
Results from visual search experiments have suggested
that perceptually salient but task-irrelevant visual events
can capture attention independently of current top-down
task sets. When participants search for a shape-defined
singleton target among non-target shapes (such as a
diamond target among circle distractors), reaction times
(RTs) are delayed when a salient but task-irrelevant color
singleton is present relative to trials without an additional
color distractor (e.g., Theeuwes, 1992). This indicates that
color singletons capture attention due to their bottom-up
salience and irrespective of current search intentions.
However, other studies (e.g., Folk & Remington, 1998;
Folk, Remington, & Johnston, 1992; Folk, Remington, &
Wright, 1994) have demonstrated that the ability of salient
visual events to capture attention is determined by
topdown task sets. When spatially uninformative singleton
cues precede visual search displays, faster RTs for visual
search targets at cued locations indicate attentional capture
by these cues. Critically, such spatial cueing effects are only
present when cues share features with currently
taskrelevant stimuli (e.g., red color singleton cues in blocks
where targets are also red), but not when cue features are
task-irrelevant. Based on such findings, Folk et al. (1992)
proposed their contingent involuntary orienting hypothesis,
which postulates that salient visual feature singletons
capture attention only when their features match an active
top-down attentional setting. Once such a control setting is
established (via experimental instructions that specify
target-defining visual features), stimuli that share critical
properties will attract attention even when they are known
to be task-irrelevant (hence involuntary orienting).
The hypothesis that attention is captured in a bottom-up
fashion by salient visual singletons regardless of their task
relevance (Theeuwes, 1992) appears at odds with the
observation that spatial cueing effects indicative of
attentional capture are absent when singleton cues do not share
attributes with targets (Folk et al., 1992). However,
Theeuwes and colleagues (Belopolsky, Schreij, & Theeuwes,
2010; Theeuwes, 2010; Theeuwes, Atchley, & Kramer,
2000) have claimed that bottom-up attentional capture is
perfectly consistent with results such as those observed by
Folk et al. (1992). They argue that visually salient singleton
stimuli will always capture attention, regardless of whether
they match a currently active task set. Contingent capture
effects are the result of top-down control mechanisms that
affect selective attentional processing after the initial
salience-driven attentional capture. According t (...truncated)