Content uploaded by Jan Theeuwes
Author content
All content in this area was uploaded by Jan Theeuwes on Sep 18, 2020
Content may be subject to copyright.
Content uploaded by Benchi Wang
Author content
All content in this area was uploaded by Benchi Wang on Aug 11, 2020
Content may be subject to copyright.
Journal of Experimental Psychology:
Human Perception and Performance
Salience Determines Attentional Orienting in Visual
Selection
Benchi Wang and Jan Theeuwes
Online First Publication, August 6, 2020. http://dx.doi.org/10.1037/xhp0000796
CITATION
Wang, B., & Theeuwes, J. (2020, August 6). Salience Determines Attentional Orienting in Visual
Selection. Journal of Experimental Psychology: Human Perception and Performance. Advance
online publication. http://dx.doi.org/10.1037/xhp0000796
Salience Determines Attentional Orienting in Visual Selection
Benchi Wang
Ministry of Education, Guangzhou, China, and South China
Normal University
Jan Theeuwes
Vrije Universiteit Amsterdam
Recently the signal-suppression account was proposed, positing that salient stimuli automatically produce
a bottom-up salience signal that can be suppressed via top-down control processes. Evidence for this
hybrid account came from a capture-probe paradigm that showed that while searching for a specific
shape, observers suppressed the location of the irrelevant color singleton. Here we replicate these
findings but also show that this occurs only for search arrays with 4 elements. For larger array sizes when
both target and distractor singleton are salient, there is no evidence for suppression; instead and consistent
with the stimulus-driven account, there is clear evidence that the salient distractor captured attention. The
current study shows that the relative salience of items in the display is a crucial factor in attentional
control. In displays with a few heterogeneous items, top-down suppression is possible. However, in larger
displays in which both target and distractor singletons are salient, no top-down suppression is observed.
We conclude that the signal-suppression account cannot resolve the long-standing debate regarding
stimulus-driven and goal-driven attentional capture.
Public Significance Statement
This study replicated the critical findings (i.e., suppression effect) reported by Gaspelin, Leonard, and
Luck, 2015, when only four elements were presented on the display. Yet with larger search arrays
(six and 10 items), the target and distractor singleton become more salient, and then there was no sign
of any suppression; instead and consistent with the stimulus-driven account, there is clear evidence
that the salient distractor captured attention. We argued that signal suppression (inhibitory processes)
occurs only in displays with a limited number of nonsalient elements allowing for a peculiar (most
likely serial) search strategy. The current experiment provides the boundary conditions of when
top-down suppression is effective.
Keywords: visual selection, attentional capture, signal suppression, salience
It is well known that salient distractors can distract us and
interfere with our ongoing task (Theeuwes, 1991, 1992). Accord-
ing to the stimulus-driven account, physically salient objects cap-
ture attention, regardless of the intentions of the observers (Theeu-
wes, 2010). This effect was demonstrated in the so-called
additional singleton paradigm (Theeuwes, 1991, 1992), in which
observers searched for a shape singleton (e.g., a diamond shape
between circles) while an irrelevant color singleton was present.
The results showed that the irrelevant color singleton captured
attention, even though it was never relevant for the search goal.
The stimulus-driven account was challenged by research dem-
onstrating that if observers search for a specific shape (e.g., a
diamond shape among circle, square, and triangle distractors), the
capture by the irrelevant singleton can be avoided (Bacon & Egeth,
1994; Leber & Egeth, 2006). It was argued that instead of search-
ing for any salient singleton (called the singleton detection mode),
a top-down set for searching for a specific shape (called the feature
search mode) could prevent attentional capture (Bacon et al.,
1994). Clearly these studies demonstrate that under some circum-
stances capture can be prevented.
In the capture debate, recently a new and appealing view was
suggested that implicated a major step forward in resolving the
capture debate (Gaspelin & Luck, 2018b). Rather than simply
observing less capture when observers adopt a feature search
mode, the new account provides an explanation why capture is
smaller or even absent when the feature search mode is engaged.
According to the signal suppression account (Gaspelin et al., 2015;
Benchi Wang, Key Laboratory of Brain, Cognition and Education Sci-
ences (South China Normal University), Ministry of Education, Guang-
zhou, China, and Institute for Brain Research and Rehabilitation, Center for
Studies of Psychological Application, and Guangdong Key Laboratory of
Mental Health and Cognitive Science, South China Normal University; Jan
Theeuwes, Department of Experimental and Applied Psychology and In-
stitute of Brain and Behavior Amsterdam, Vrije Universiteit Amsterdam.
Benchi Wang and Jan Theeuwes contributed equally to the present
study.
Data and procedure can be accessed through https://github.com/
wangbenchi/Shared_data. None of the experiments was preregistered.
Correspondence concerning this article should be addressed to Benchi
Wang, Institute for Brain Research and Rehabilitation, South China Nor-
mal University, Zhongshan Road West 55, Guangzhou, China 510000.
E-mail: wangbenchi.swift@gmail.com
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Experimental Psychology:
Human Perception and Performance
© 2020 American Psychological Association 2020, Vol. 2, No. 999, 000
ISSN: 0096-1523 http://dx.doi.org/10.1037/xhp0000796
1
Sawaki & Luck, 2010), when observers engage the top-down
feature search mode, salient distractors still generate a large
bottom-up signal, but the signal can be proactively suppressed in
a top-down way, thereby preventing a shift of attention toward the
salient distractor. In other words, consistent with stimulus-driven
accounts (Theeuwes, 1992, 2010), it is argued that physically
salient stimuli have the ability to attract attention; yet, consistent
with goal-driven theories (Folk & Remington, 2010; Folk, Rem-
ington, & Johnston, 1992), inhibitory processes can suppress these
salient signals if participants exert top-down control. As such, the
signal suppression account seems to be able to resolve a long-
standing conflict between stimulus-driven and goal-driven theories
of attentional capture.
Gaspelin et al. (2015) provided compelling evidence for this
notion in an innovative study in which the additional singleton task
was combined with a letter probe task. In 70% of trials, partici-
pants searched for a target shape while ignoring a color singleton.
In 30% of trials, letter probes were briefly presented inside the
search elements, and participants were required to report as many
letters as possible. In conditions in which participants used the
singleton detection mode, Gaspelin et al. (2015) showed that
participants reported more letters when these were presented inside
the singleton distractor than presented inside the nonsingleton
distractors. This was considered as a clear evidence for attentional
capture by the salient distractor (see also Kim & Cave, 1999 for
similar results using a probe task). Critically, however, when
observers engaged the feature search mode, the accuracy for the
letter inside the irrelevant singleton distractor was reduced below
the accuracy observed for letters inside the nonsingleton distrac-
tors. This was considered as decisive evidence for active suppres-
sion of the singleton-distractor location. This result of subbaseline
suppression was considered to be the essential finding for resolv-
ing the 25-year-old debate regarding the role of top-down and
bottom-up control of attentional selection (Gaspelin & Luck,
2018b, 2018c).
Unlike the signal-suppression account of Gaspelin et al. (2015)
that assumes the capture is prevented because the salience signal of
the singleton-distractor is suppressed, an alternative account
known as rapid disengagement (Theeuwes, 2010) claims that spa-
tial attention is initially captured by the singleton distractor and
then rapidly disengaged from it. According to this account, top-
down suppression occurs following the initial shift of spatial
attention to the singleton distractor location. Gaspelin et al. (2015)
recognized that the rapid disengagement account was a viable
explanation for their results (Experiments 2 and 3) because the
probe was presented 200 ms after the search display, providing
plenty of time for disengagement to occur. Yet they addressed
these issues in their final Experiment 4 in which the letters were
presented simultaneously with the search display and found basi-
cally the same results. It rules out the rapid disengagement account
and firmly establishes their signal-suppression account.
Even though the final experiment of Gaspelin et al. (2015)
seems to be decisive in the debate, there is one major concern. In
their experiment (and also in a recent study by Chang and Egeth,
2019, and a replication with electroencephalographic recording by
Gaspelin & Luck, 2018a), the display consisted only of four
elements (a target, a singleton distractor, and two nonsingleton
distractors), rendering both the target and the distractor nonsalient.
In a study published in 2004, Theeuwes (2004) had already shown
that the number of elements in the display is critical for the
salience signal of both the target and distractor. This is especially
the case in displays in which distractor shapes are heterogeneous
(e.g., squares, hexagons and circles), which is typically done to
induce the feature search mode preventing the target from popping
out. The critical message of Theeuwes (2004) was that with a few
heterogeneous elements in the display, nothing in the display is
salient enough to capture attention. Not surprising that in displays
like these (i.e., displays used by Gaspelin et al., 2015 having only
a few heterogeneous items), the distractor no longer captures
attention, not because of a top-down search mode labeled as
feature search but simply because the singleton distractor is no
longer salient.
Previous research has shown that the saliency of items depends
on two factors: local feature contrast (Nothdurft, 1993) and
distractor-distractor similarity (Duncan & Humphreys, 1989). Lo-
cal feature contrast refers to how different an item is from nearby
items (Nothdurft, 1993). In small display sizes, when items are
equally spaced around the fixation point, they are relatively far
apart, reducing the local feature contrast. In addition, to force
participants to engage in feature search, the display needs to be
heterogeneous (e.g., squares, hexagons, and circles), which nega-
tively affects the distractor-distractor similarity (Duncan et al.,
1989). It was shown that distractor heterogeneity reduces search
efficiency, resulting in serial search (Duncan et al., 1989). It is
clear that both factors play an important role when engaging in
feature search among only a few items. Even though a distractor in
these displays with only a few items may be unique and as such
should be considered as a singleton, it does not necessarily mean
that the distractor is also salient. If none of the display items are
salient, it is not surprising that no capture is observed.
Given these concerns, we aimed at replicating the critical Ex-
periment 4 of Gaspelin et al. (2015) but now with various search
array sizes (four, six, and ten elements). When the number of
elements is relatively large, it is expected that both the target and
the singleton distractor are salient enough because they will stand
out from the background (see Theeuwes, 2004). As in Gaspelin et
al. (2015), we also used heterogeneous displays to induced the
feature search mode, and we presented the probes simultaneously
with the search display. According to the signal-suppression ac-
count, there should be active suppression of the salient singleton
distractor not only when the singleton distractor is relatively non-
salient (with search array size four, as in Gaspelin et al., 2015) but
also when the distractor is relatively salient (with search array
sizes six and 10). According to the stimulus-driven account
(Theeuwes, 1992, 2004, 2010), if a distractor is salient enough, it
should always capture attention, regardless of the search mode
used.
Method
The study was approved by both the Ethical Review Committee
of the Vrije Universiteit Amsterdam and the Ethical Review Com-
mittee of Zhejiang Normal University.
Participants
Seventy-two undergraduates (eight men and 64 women: with a
mean age of 19.2 ⫾1.1 years) were recruited from Zhejiang
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2WANG AND THEEUWES
Normal University in China for monetary compensation. All par-
ticipants provided written informed consent before the study and
reported normal color vision and normal or corrected-to-normal
visual acuity. Participants were equally and randomly allocated to
the different search array size conditions. Sample size was prede-
termined based on previous studies (Gaspelin et al., 2015; Gaspe-
lin & Luck, 2019). According to the power analysis described by
Gaspelin and Luck (2019; p. 8), “We estimated the population
effect size and standard deviation by pooling the probe suppression
effects (the difference between singleton distractor location and
nonsingleton distractor location in probe task) across participants
from three previous experiments that were similar in methodology
to the current experiment (Gaspelin et al., 2015, Experiments 2– 4),
yielding an Nof 72 participants. The probe suppression effects in
the pooled data were quite robust, with an effect size of dz ⫽.97.
Thus, to achieve a power of 95% and an alpha of 5% with this
effect size, a sample size of 16 participants would be needed.”
Because our experiment involved a between-subjects design, we
adopted a sample size of 24 subjects. If there is active suppression,
this large sample size should be enough to detect it.
Apparatus and Stimuli
Participants were seated in a dimly lit laboratory, 63 cm away
from the liquid crystal display color monitor with their chin on a
chinrest. The background was gray (red-green-blue [RGB] ⫽128,
128, 128). As illustrated in Figure 1A, the primary search display
for search array size four contained one circle with a radius of 0.7°,
one diamond (subtended by 1.6° ⫻1.6°), one square (subtended
by 1.6° ⫻1.6°), and one hexagon (subtended by 1.6° ⫻1.6°); for
search array size six, there was one circle, one diamond, two
squares, and two hexagons; for search array size 10, there was one
circle, one diamond, four squares, and four hexagons. Those
display elements were colored in red (RGB ⫽255, 0, 0) or green
(RGB ⫽0, 255, 0), and were centered 3.0° from the fixation (a
black cross, 0.5° ⫻0.5°, RGB ⫽0, 0, 0), containing a 0.2° black
dot located 0.2° left or right from the center of the element. On
search-probe trials, a white (RGB ⫽255, 255, 255) uppercase
letter (0.75° tall) was presented in Simhei typeface at the center of
each search element. All possible letters from the English alphabet
were presented subsequently as a response display. Stimulus pre-
sentation and response registration were controlled by custom
scripts written in Python 2.7.
Procedure and Design
The procedure and task were virtually identical to that of Ex-
periment 4 of Gaspelin et al. (2015) except that we varied search
array sizes (four, six, and 10 elements) between participants.
On each trial, a fixation cross was presented for 500 ms, fol-
lowed by a primary search display in which participants were
asked to search for a specific shape (for half of the participants a
circle shape; for the other half, a diamond shape). Participants
were required to keep fixation at the cross throughout a trial. The
search target was presented in each trial and appeared equally
often at each location. A uniquely colored singleton was used as
the salient distractor in half of the trials, with a different color as
the target and other display elements (red or green balanced
between participants).
On search-only trials (two thirds of the trials), the search array
was presented for 3,000 ms or until participants responded (as
shown in Figure 1B). Participants were required to indicate
Figure 1. (A) Example of search displays for different search array sizes. (B) The procedure in search-only
task, in which participants were required to search for a specific shape (for different participants either a circle
or a diamond shape) and to indicate the position (i.e., left or right) of the black dot inside. (C) The procedure
in search-probe task, in which the search display and the probe letters were presented simultaneously for 100 ms.
Participants were required to memorize the letters and to recall them by pressing the corresponding key on the
keyboard as accurately as possible. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
SALIENCE DETERMINED SELECTION
whether the dot was on the left or right side of the target shape (i.e.,
the specific shape, circle or diamond) by pressing the left or right
key on the keyboard using left hand as fast as possible, respec-
tively. Responses were speeded and feedback, “You did not re-
spond; please respond as fast as possible” or “Incorrect response;
please focus on the task,” was given when participants did not
respond or responded incorrectly, respectively.
On search-probe trials (one third of the trials), as shown in
Figure 1C, the search array and the probe display (containing
to-be-memorized letters) appeared simultaneously for 100 ms. For
those trials, participants did not have to respond to the search array
but had to attend and memorize as many letters as possible. For
each trial, the letters were selected randomly, without replacement,
from the English alphabet. Then a response display was presented
until that participants responded. Participants had to recall as many
letters as possible by pressing the corresponding letter keys on the
keyboard without time pressure, and only accuracy was empha-
sized. Once the letter was selected, it turned into red. When they
finished their response, they pressed the space key to continue.
Participants were first trained for a small number of trials to
make sure they understood the task before testing started. Different
search array sizes were tested between participants. For search
array size four, participants completed four blocks with each
containing 96 trials (a total of 384 trials). For search array size six,
they completed three blocks with each containing 144 trials (a total
of 432 trials). For search array size 10, they completed four blocks
with each containing 120 trials (a total of 480 trials). The search-
only and search-probe trials were mixed within blocks.
Results
Search Array Size Four
Search-only condition. Trials (1.7%) on which the response
times (RTs) were slower than 2,000 ms were removed from
analysis. Mean RTs are presented in Figure 2A, left panel. Mean
RTs were the same for distractor present (831 ms) and absent
conditions (834 ms), t(23) ⫽0.63, p⫽.53, Cohen’s d⫽0.21,
BF
01
⫽3.89. The same pattern of results was observed for error
rates: distractor present (1.1%) versus distractor absent (1.4%),
t(23) ⫽0.93, p⫽.36, Cohen’s d⫽0.02, BF
01
⫽3.15. These
findings are consistent with those of Gaspelin et al., 2015, showing
no attentional capture for search array size four when a feature
search mode is used.
Search-probe condition. We first calculated the proportion of
probes that were recalled when they were presented at the target
location, at the singleton distractor location, and at each nonsingle-
ton distractor location, and then averaged across the nonsingleton
distractor locations, to provide the probe recall accuracy for each
location. As illustrated in Figure 2B, left panel, the probe recall
accuracy was significantly lower when the probe was presented at
the singleton distractor location (19.4%) than at the nonsingleton
distractor location (22.8%), t(23) ⫽2.53, p⫽.02, Cohen’s d⫽
0.94, providing an exact replication of the critical finding of
Gaspelin et al., 2015.
For the remaining conditions, a repeated-measures ANOVA on
probe recall accuracy with the factors of distractor condition
Figure 2. The results for different search array sizes. (A) The mean response times in the singleton distractor
present and absent conditions. (B) The recall accuracy for probes presented at the target location, at the
nonsingleton distractor location, and at the singleton distractor location in both singleton distractor present and
absent conditions. Error bars denote 95% confidence intervals (CIs). Red stars indicate a significant difference
(p⬍.05). n.s. ⫽nonsignificant. See the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4WANG AND THEEUWES
(singleton distractor present and absent) and probe type (target and
nonsingleton distractor) was conducted. No significant main ef-
fects nor significant interactions were observed, all Fs⬍1.96, all
ps⬎.18.
Search Array Size Six
Search-only condition. Trials (1.2%) on which the RTs were
slower than 2,000 ms were removed from analysis. Mean RTs are
presented in Figure 2A, middle panel. Unlike the results of array size
four, with increasing the array size, singleton distractor started cap-
turing attention because mean RTs were higher when the distractor
was present (886 ms) than when it was absent (869 ms), t(23) ⫽2.39,
p⫽.03, Cohen’s d⫽0.97. There was no effect on mean error rates,
t(23) ⫽0.89, p⫽.38, Cohen’s d⫽0.02, BF
01
⫽3.26.
Search-probe condition. As shown in Figure 2B, middle
panel, the probe recall accuracy was basically identical for probes
presented at the singleton distractor location (19.3%) compared
with nonsingleton distractor location (20.4%), t(23) ⫽1.05, p⫽
.31, Cohen’s d⫽0.28, BF
01
⫽2.86.
For the remaining conditions, a repeated-measures ANOVA on
probe recall accuracy with the factors of distractor condition (single-
ton distractor present and absent) and probe type (target and nons-
ingleton distractor) was conducted. No significant main effects or
significant interactions were observed, all Fs⬍2.73, all ps⬎.11.
Search Array Size 10
Search-only condition. Trials (3.3%) on which the RTs were
slower than 2,000 ms were removed from analysis. Mean RTs are
presented in Figure 2A, right panel. For search array size 10, there
was a strong attentional capture because RTs were slower when
the distractor was present (988 ms) than when it was absent (945
ms), t(23) ⫽5.07, p⬍.01, Cohen’s d⫽2.21. There was no effect
on mean error rates, t(23) ⫽0.95, p⫽.35, Cohen’s d⫽0.01,
BF
01
⫽3.11.
Search-probe condition. Unlike what was found for array
size four, now there is a reversed pattern because probe recall
accuracy was higher when the probe was presented at the singleton
distractor location (18.1%) than at the nonsingleton distractor
location (16.2%), t(23) ⫽2.76, p⫽.01, Cohen’s d⫽0.51. Note
that recall accuracy for probes presented at the target location was
higher in the distractor absent condition than in the distractor
present condition, suggesting that when attention is allocated to the
singleton distractor, fewer resources are available for processing
the target (see similar results in Experiment 1 in Gaspelin et al.,
2015 when participants used the singleton detection mode to find
the target).
For the remaining conditions, a repeated-measures ANOVA on
probe recall accuracy with the factors of distractor condition
(singleton distractor present and absent) and probe type (target and
nonsingleton distractor) was conducted. The recall accuracy was
higher for probes presented at the target location than at the
nonsingleton distractor location, F(1, 23) ⫽14.87, p⬍.01, p
2⫽
.39. Moreover, the recall accuracy was also higher on singleton
distractor present versus absent trials, F(1, 23) ⫽7.89, p⫽.01,
p
2⫽.26. There was no significant interaction, F(1, 23) ⬍0.01,
p⫽.99, p
2⬍.01.
Analysis Across Different Search Array Sizes
For each subject, we first calculated the attentional capture
effect in the search-only task and the distractor suppression effect
in the search-probe task by using the mean RTs on distractor
present trials minus that on distractor absent trials and using the
recall accuracy for probes presented at nonsingleton distractor
location minus that for probes presented at singleton distractor
location, respectively. Then we compared them across different
search array sizes.
As illustrated in Figure 3A, there was a suppression effect for
search array size four (3.4%), but no such an effect for search array
size 6 (1.1%), and a reversed pattern for search array size 10
Figure 3. (A) The suppression effect (reflected by the recall accuracy for probes presented at the nonsingleton
distractor location minus that at the singleton distractor location) in the search-probe task and the attentional
capture (reflected by the mean response times on distractor present trials minus that on distractor absent trials)
in the search-only task for different search array sizes. (B) The correlation between the suppression effect in the
search-probe task and the attentional capture in the search-only task across different participants.
ⴱⴱ
p⬍.01. See
the online article for the color version of this figure.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
SALIENCE DETERMINED SELECTION
(⫺1.9%). Attentional capture was not found for search array size
four (⫺3 ms) but was present for search array size six (17 ms) and
for search array size 10 (43 ms). We compared search array sizes
four and 10 and found that the capture effect was larger for search
array size 10 than that for size four, t(46) ⫽4.63, p⬍.001,
Cohen’s d⫽1.34; and the suppression effect was smaller for
search array size 10 than that for size four, t(46) ⫽3.49, p⫽.001,
Cohen’s d⫽1.01. We also calculated the correlation between the
suppression effect and the attentional capture across 72 different
participants, with a significant Spearman correlation, r⫽⫺0.31,
p⫽.01, BF
01
⫽0.25 (same for Pearson correlation, r⫽⫺0.29,
p⫽.01, BF
01
⫽0.35; see Figure 3B). Overall, it indicates that the
suppression effect observed in the probe task is negatively related
to the capture effect in the search task.
Discussion
For search array size four, we perfectly replicated the results of
Gaspelin et al. (2015) showing suppression of the singleton dis-
tractor location. Yet for array size six, there was no evidence for
the suppression of the singleton distractor location. Critically, for
search array size 10, the effect was reversed, now showing en-
hanced processing of probes that were presented at the singleton
distractor location. This latter finding suggests that for larger
search array sizes, singleton distractor captured attention, even
though participants had to use the feature-search mode to find the
target. Indeed, in this experiment, the target shape was kept con-
stant and was presented among a heterogeneous set of distractor
shapes, resulting in feature-search mode (i.e., shape-based search)
and eliminating the possibility of searching for something unique
(i.e., discouraging the singleton detection mode).
The current findings suggest that the signal-suppression account
(Gaspelin et al., 2015), which assumes that salient-but-irrelevant
singletons can be suppressed when participants engage the top-
down feature search mode, is limited in scope because it seems to
work only for arrays with a few nonsalient items. As such, this
account alone cannot resolve the conflict between stimulus-driven
and goal-driven theories of attentional capture (Folk & Remington,
2010; Folk et al., 1992; Theeuwes, 1992, 2010). Note that the
current findings also showed that when both target and singleton
distractors were sufficiently salient, there was attentional capture,
regardless of what search mode was used (see also Theeuwes,
2004 for a similar argument).
The correlation that we reported between the suppression effect
and the capture effect is compelling. Overall, it indicates that at an
individual level, the suppression effect that is observed in the
probe task is significantly correlated with the capture effect in the
search task. This implies that, consistent with the stimulus-driven
account, if the singleton distractor is salient enough, attentional
resources are initially allocated to its location, resulting in a higher
recall accuracy for probes presented at the singleton distractor than
at nonsingleton distractor locations.
It is possible that there is suppression of the singleton distractor
location after spatial attention has been initially captured by the
singleton distractor, as has been suggested by the rapid disengage-
ment account of Theeuwes (2010). This is consistent with Kim et
al. (1999), who also used the additional singleton paradigm in
combination with a probe-detection task. They showed that 60 ms
after display onset, probe RTs at the singleton distractor location
were about 20 ms faster than at the target location, indicating
initial capture. Yet at an stimulus onset asymmetry of 150 ms, this
pattern was reversed: The probe RTs at the target location were
about 15 ms faster than at the distractor location, indicating that
attention was rapidly disengaged from the singleton-distractor
location and redirected to the target location.
It should be noted that recently Gaspelin and Luck (2018a)
provided neural evidence for the signal-suppression account. Spe-
cifically, they focused on inhibition-related component of the
event-related potential (ERP) signal called distractor positivity
(P
D
; Hickey, Di Lollo, & McDonald, 2009). It has been shown that
stimuli that fail to capture attention elicit a P
D
component (e.g.,
Burra & Kerzel, 2014; Eimer & Kiss, 2008; Feldmann-Wüstefeld,
Uengoer, & Schubö, 2015; Gaspar & McDonald, 2014). Gaspelin
and Luck (2018a) showed that across participants, the magnitude
of the suppression in a task as we used here was related to the
magnitude of the P
D
. This elegantly connected the behavioral
suppression to the neural measures of the suppression. Even
though these results are convincing, it should be noted that also in
their ERP study, Gaspelin and Luck (2018a) used only heteroge-
neous search arrays of four elements, which as shown here, renders
none of the elements salient enough to say anything about the
suppression of salient distractors.
Recently Barras and Kerzel (2016) also investigated this issue
and used eight instead of four items. Important for the present
discussion, when using eight elements, there was no sign of a P
D
,
even when participants engaged in the feature search mode. It
suggests that if the P
D
is a marker of suppression, this can be found
only in sparse displays in which none of the elements are salient
enough to generate a saliency signal. In another more recent ERP
study, Kerzel and Burra (2020) revisited this issue but now ex-
plicitly focused on small search displays. In fact, the displays were
identical to those of Gaspelin et al. (2015) and Gaspelin and Luck
(2018a). This study shows that the P
D
does not represent distractor
suppression but instead is the result of a paradoxical flip of the
contralateral voltage difference because of a peculiar search strat-
egy that is used in small displays (Kerzel et al., 2020).
The current study indicates that the relative salience of items in
the display is crucial in the capture-suppression debate. With a few
heterogeneous items on display, it is possible to obtain top-down
suppression (Chang et al., 2019; Gaspelin et al., 2015; Gaspelin &
Luck, 2018a). However, in displays in which there is enough local
feature contrast rendering both target and distractor singletons
truly salient, no top-down suppression is observed. Instead, and
consistent with the stimulus-driven account, the most salient item
in the display captures attention.
In summary, consistent with Gaspelin et al. (2015; Gaspelin &
Luck, 2018a), when engaged in feature search and the display
consists of only four elements, there is below-baseline suppression
of the singleton distractor. However, adding elements to the dis-
play renders both target and distractor singleton salient, resulting
in attentional capture by the distractor, even when participants
engage in the feature search mode. We conclude that signal-
suppression account is limited in scope and cannot resolve the
debate between stimulus-driven and goal-driven theories. If any-
thing, the below-baseline suppression observed by Gaspelin et al.
(2015) is the result of some idiosyncratic (most likely serial)
search strategy that can operate only in displays containing a
limited number of nonsalient elements.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6WANG AND THEEUWES
References
Bacon, W. F., & Egeth, H. E. (1994). Overriding stimulus-driven atten-
tional capture. Perception & Psychophysics, 55, 485– 496. http://dx.doi
.org/10.3758/BF03205306
Barras, C., & Kerzel, D. (2016). Active suppression of salient-but-
irrelevant stimuli does not underlie resistance to visual interference.
Biological Psychology, 121, 74 – 83. http://dx.doi.org/10.1016/j
.biopsycho.2016.10.004
Burra, N., & Kerzel, D. (2014). The distractor positivity (Pd) signals
lowering of attentional priority: Evidence from event-related potentials
and individual differences. Psychophysiology, 51, 685– 696. http://dx
.doi.org/10.1111/psyp.12215
Chang, S., & Egeth, H. E. (2019). Enhancement and suppression flexibly
guide attention. Psychological Science, 30, 1724 –1732. http://dx.doi
.org/10.1177/0956797619878813
Duncan, J., & Humphreys, G. W. (1989). Visual search and stimulus
similarity. Psychological Review, 96, 433– 458. http://dx.doi.org/10
.1037/0033-295X.96.3.433
Eimer, M., & Kiss, M. (2008). Involuntary attentional capture is deter-
mined by task set: Evidence from event-related brain potentials. Journal
of Cognitive Neuroscience, 20, 1423–1433. http://dx.doi.org/10.1162/
jocn.2008.20099
Feldmann-Wüstefeld, T., Uengoer, M., & Schubö, A. (2015). You see what
you have learned. Evidence for an interrelation of associative learning
and visual selective attention. Psychophysiology, 52, 1483–1497. http://
dx.doi.org/10.1111/psyp.12514
Folk, C. L., & Remington, R. (2010). A critical evaluation of the disen-
gagement hypothesis. Acta Psychologica, 135, 103–105. http://dx.doi
.org/10.1016/j.actpsy.2010.04.012
Folk, C. L., Remington, R. W., & Johnston, J. C. (1992). Involuntary
covert orienting is contingent on attentional control settings. Journal of
Experimental Psychology: Human Perception and Performance, 18,
1030 –1044. http://dx.doi.org/10.1037/0096-1523.18.4.1030
Gaspar, J. M., & McDonald, J. J. (2014). Suppression of salient objects
prevents distraction in visual search. Journal of Neuroscience: The
Official Journal of the Society for Neuroscience, 34, 5658 –5666. http://
dx.doi.org/10.1523/JNEUROSCI.4161-13.2014
Gaspelin, N., Leonard, C. J., & Luck, S. J. (2015). Direct evidence for
active suppression of salient-but-irrelevant sensory inputs. Psycho-
logical Science, 26, 1740 –1750. http://dx.doi.org/10.1177/
0956797615597913
Gaspelin, N., & Luck, S. J. (2018a). Combined electrophysiological and
behavioral evidence for the suppression of salient distractors. Journal of
Cognitive Neuroscience, 30, 1265–1280. http://dx.doi.org/10.1162/
jocn_a_01279
Gaspelin, N., & Luck, S. J. (2018b). Distinguishing among potential
mechanisms of singleton suppression. Journal of Experimental Psychol-
ogy: Human Perception and Performance, 44, 626 – 644. http://dx.doi
.org/10.1037/xhp0000484
Gaspelin, N., & Luck, S. J. (2018c). The role of inhibition in avoiding
distraction by salient stimuli. Trends in Cognitive Sciences, 22, 79 –92.
http://dx.doi.org/10.1016/j.tics.2017.11.001
Gaspelin, N., & Luck, S. J. (2019). Inhibition as a potential resolution to
the attentional capture debate. Current Opinion in Psychology, 29,
12–18. http://dx.doi.org/10.1016/j.copsyc.2018.10.013
Hickey, C., Di Lollo, V., & McDonald, J. J. (2009). Electrophysiological
indices of target and distractor processing in visual search. Journal of
Cognitive Neuroscience, 21, 760 –775. http://dx.doi.org/10.1162/jocn
.2009.21039
Kerzel, D., & Burra, N. (2020). Capture by context elements, not atten-
tional suppression of distractors, explains the PD with small search
displays. Journal of Cognitive Neuroscience, 32, 1170 –1183. http://dx
.doi.org/10.1162/jocn_a_01535
Kim, M.-S., & Cave, K. R. (1999). Top-down and bottom-up attentional
control: On the nature of interference from a salient distractor. Percep-
tion & Psychophysics, 61, 1009 –1023. http://dx.doi.org/10.3758/
BF03207609
Leber, A. B., & Egeth, H. E. (2006). It’s under control: Top-down search
strategies can override attentional capture. Psychonomic Bulletin &
Review, 13, 132–138. http://dx.doi.org/10.3758/BF03193824
Nothdurft, H.-C. (1993). The role of features in preattentive vision: Com-
parison of orientation, motion and color cues. Vision Research, 33,
1937–1958. http://dx.doi.org/10.1016/0042-6989(93)90020-W
Sawaki, R., & Luck, S. J. (2010). Capture versus suppression of attention
by salient singletons: Electrophysiological evidence for an automatic
attend-to-me signal. Attention, Perception & Psychophysics, 72, 1455–
1470. http://dx.doi.org/10.3758/APP.72.6.1455
Theeuwes, J. (1991). Cross-dimensional perceptual selectivity. Perception
& Psychophysics, 50, 184 –193. http://dx.doi.org/10.3758/BF03212219
Theeuwes, J. (1992). Perceptual selectivity for color and form. Perception
& Psychophysics, 51, 599 – 606. http://dx.doi.org/10.3758/BF03211656
Theeuwes, J. (2004). Top-down search strategies cannot override atten-
tional capture. Psychonomic Bulletin & Review, 11, 65–70. http://dx.doi
.org/10.3758/BF03206462
Theeuwes, J. (2010). Top-down and bottom-up control of visual selection.
Acta Psychologica, 135, 77–99. http://dx.doi.org/10.1016/j.actpsy.2010
.02.006
Received March 7, 2020
Revision received May 1, 2020
Accepted May 2, 2020 䡲
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
SALIENCE DETERMINED SELECTION
