Psychology
2011. Vol.2, No.8, 824-833
Copyright © 2011 SciRes. DOI:10.4236/psych.2011.28126
How the Strength of a Strong Object Mask Varies in Space and
Time When It Is Used as an Uninformative Singleton in Visual
Search for Target Location*
Talis Bachmann1, Endel Põder2, Carolina Murd1
1Institute of Public Law, University of Tartu, Tallinn, Estonia;
2Department of Psychology, University of Tartu, Tartu, Estonia.
Email: talis.bachmann@ut.ee
Received August 26th, 2011; revised September 28th, 2011; accepted October 29th, 2011.
Strong visual masking originates from sensory perceptual interactions between target and mask and also from
attentional competition between target and mask even though mask does not correspond to attentional control
settings. The relative contributions of these different masking mechanisms are difficult to estimate. One strategy
to begin approach this problem is to use the same stimulus as a mask and as a non-informative singleton in a
selective attention task. The purpose of the present study was to find the spatial and temporal intervals where a
strong object mask interferes with target-object search when used as a non-informative singleton. In visual
search for target location, we found that a visual object that has a strong forward and backward masking power
on target-object correct perception when spatially superimposed on target can impair target perception from a
spatially separated location only when presented up to 100 ms after the target and only from a spatially close
location. These results are explained by a processing account where the initial analysis of stimuli features allows
to determine the best candidate location for the target, but as soon as this location is established, a nearby later
appearing object may intrude it, replacing the target in explicit perception. The higher-level mechanisms based
interpretation is strengthened by the finding that any local masking effects of the same adjacent singleton were
absent in the task of single-target identification.
Keywords: Masking, Visual Search, Target Localization, Attentional Capture, Consciousness
Introduction
Impairment of Perception by Masking and
Attentional Miscueing
Surprisingly perhaps, but there are many ways normal people
with normal vision cannot see (Breitmeyer, 2010). Part of the
reasons why objects cannot be consciously seen owe to the
limitations and interactions of early sensory-perceptual mecha-
nisms; another part of the cases of not seeing takes place
because objects may remain unattended or inaccessible for
attention. In one of these phenomena called visual masking an
otherwise well perceptible brief stimulus object, a target,
becomes barely visible or completely invisible if accompanied
in space and time by another object—a mask. Masking is a
complex phenomenon occurring in different varieties and emer-
ging as an outcome of the effects of early sensory-perceptual
mechanisms and/or attentional mechanisms (Bachmann, 1994;
Enns, 2004). Thus, strong visual masking originates not only
from sensory-perceptual interactions between target and mask,
but also from attentional competition between target and mask
even though subjects try to ignore the mask as much as they can
(Di Lollo, Enns, & Rensink, 2000; Michaels & Turvey, 1979;
Smith, & Wolfgang, 2004). The relative contributions of these
different masking mechanisms are difficult to estimate. A mas-
king stimulus may at the same time inhibit target’s perceptual
processing, deprive the target from access to consciousness or
draw attention away from target processing. Unfortunately, the
power of a stimulus object to act as a mask and act as a cue for
capturing attention has been studied in completely different sets
of research.
One strategy to begin approach this problem would be to
start with setting up experiments where the same stimulus is
used as a mask and as a distractor in a selective attention task.
In one of its roles, a stimulus different from any target would be
used as a visual object-mask covering target location, but
presented either before the target (in forward masking) or after
the target (in backward masking). In the other of its roles, the
same stimulus is presented either not far from the target
location (as a valid attentional cue) or far from the target
location (as an invalid attentional cue). The selective attention
task is that of visual search—a typical task where subjects
search for a target stimulus pre-specified before each trial. The
main interest would be to compare the relative effectiveness of
the same stimulus as a mask and as a bottom-up attentional cue.
The relative expression of the effect of the cue/mask object as a
mask or as a cue for involuntary attentional capture on target
perception can be studied.
It is widely believed that reliable and explicit visual recogni-
tion depends critically on whether selective attention can parti-
cipate in the processing of object information. This standpoint
has been repeatedly advocated in the many studies of visual
search and attentional cueing (e.g., Cheal, & Lyon, 1991;
Davoli, Suszko, & Abrams, 2007; Di Lollo, Enns, & Rensink,
2000; Gibson et al., 2008; Kahneman, & Treisman, 1984;
Kawahara, & Miyatani, 2001; Müller & Krummenacher, 2006;
Vierck, & Miller, 2008; Yantis, & Jonides, 1990). However, it
is not unanimously clear whether the perception-improving or
impairing selective attention can be attracted also automatically
by uninformative singleton objects appearing somewhere in the
visual field (Müller & Krummenacher, 2006; Yeh, & Liao,
2010). An uninformative singleton that appears spatially close
to the target object supposedly draws attention to that generic
T. BACHMANN ET AL.
825
location and improves target processing; the same singleton, if
presented in a spatially remote location supposedly acts as an
invalid cue and therefore impairs target processing by sending
attention to a wrong place. On the other hand, if the perceptual-
attentional system can concentrate processing on the target and
ignore the uninformative singleton, singleton presentation may
remain without effect, neither a facilitatory nor a perturbing one.
This can hold especially when instead of selective attentional
focusing by spatial cueing, selective attention is engaged in
tasks of visual target search where attentional processes are pre-
set by target identity and not by spatial location.
It is generally accepted that the likelihood that an uninforma-
tive object will effectively shift attention away from a target
depends most of all on two factors: 1) its relative salience and 2)
attentional control settings or task relevance (e.g., Ansorge, &
Horstmann, 2007; Folk, Remington, & Johnston, 1992;
Lichtenstein-Vidne, Henik, & Safadi, 2007; Müller, & Krum-
menacher, 2006; Yantis, & Jonides, 1990). Even though the
controversy over whether spatial attention can be automatically
captured when the uninformative singleton is free from being
associated with control settings, has not been conclusively
solved (for the answer “yes” see, e.g., Davoli, Suszko, &
Abrams, 2007; Forster, & Lavie, 2008; Gibson, Folk, Theeuwes,
& Kingstone, 2008; Neo, & Chua, 2006; Turatto, & Galfano,
2000; Yeh, & Liao, 2010; for the answer “no” see, e.g., Becker,
2007; Gibson et al., 2008; Yantis, & Jonides, 1990; Jingling, &
Yeh, 2007), there seems to be at least one agreement. It is
believed that if attention is not spatially focused before
presentation of a set of alternative objects (including the target),
sufficiently salient uninformative singletons typically capture
attention in a bottom-up, involuntary manner and may do so in
advance of target exposure (Müller, & Krummenacher, 2006;
Neo, & Chua, 2006). This leads to costs in target discrimination
performance. Therefore, in the present research we use visual
search where target’s spatial position is not known to subjects
at the outset of their search trials and they themselves have to
figure out that position. As a corollary, we expect involuntary
attentional capture.
The Present Approach and Its Aims
If we compare the many studies carried out on visual search
and distracting effects on it, we can easily notice the wide
variety of the experimental variables and of the values of the
variables used. The problem of distractability of processing in
visual covert search cannot be easily solved when the dispute is
based on varied sets of data drawn from experiments using
different parameters of stimulation—different number of
alternatives, types of features, durations of stimuli, stimuli
onset asynchronies (positive and/or negative, narrowly selected,
time values), sizes of stimuli, spatial distances between stimuli,
luminances and contrasts, using masks or not, target-distractor
similarity values, number of varying dimensions of features, etc.
Often the yes-no type of target detection is used readily
allowing non-sensory bias effects; often the dependent measure
is reaction time whereby pre-conscious automatic facilitation
effects and conscious-level, explicit perception effects cannot
be separated (response priming effects can be heavily pre-
conscious—Mulckhuyse, & Theeuwes, 2010; Van den Bussche
et al., 2010). In the present work, we use correct localization of
targets among the many alernative positions as the measure of
veridical target perception and use one and the same stimulus as
the mask and as the uninformative bottom-up cue for attentional
capture. In the present study we examine the possible effects of
object masks in the role of uninformative singleton objects on
visual covert search by systematically varying the spatial and
temporal distance between the target and the singleton object
over a wide range of spatial and temporal values. We aim to
map the target-to-singleton spatial- and temporal-distance
values that lead either to improvement or impairment of target
processing compared to spatial and temporal intervals that do
not help get clear effects. First of all, we aim to assess
comparatively the strength of an invariant singleton object used
as a mask and used as an uninformative cue. Depending on the
results, we will discuss the possible support or inconsistency of
the data with regard to known processing mechanisms and
theories of attentive perception in target search.
The following simple hypotheses were put forward: 1)
When a singleton object as a forward mask is presented
optimally (0 - 150 ms) in advance of the display that contains a
pre-specified search-target object and when it spatially overlaps
the target location, strong masking is expected. 2) When a
singleton object as a backward mask is presented optimally (0 -
150 ms) subsequent to the display that contains a pre-specified
search-target object and when it spatially overlaps the target
location, strong masking is expected. 3) When a singleton
object is presented optimally (50 - 150 ms) in advance of the
display that contains a pre-specified search-target among
distractors and when it is spatially close to the target location (a
condition of valid spatial-attentional pre-cueing), target correct
localization rate will be higher compared to the overall mean
localization performance. 4) When a singleton object is pre-
sented optimally (50 - 150 ms) in advance of the display that
contains a pre-specified search-target among distractors and
when it is spatially far from the target location (acting as an
invalid spatial pre-cue), target correct localization rate will be
lower compared to the overall mean localization performance. 5)
When a singleton object is presented optimally (50 - 150 ms)
after the display that contains a pre-specified search-target
among distractors and when it is spatially close to the target
location, target correct localization rate will be higher compared
to the overall mean localization performance—the singleton
acting as a valid after-cue aiding selection from sensory visual
memory. 6) When a singleton object is presented optimally (50
- 150 ms) after the display that contains a pre-specified search-
target among distractors and when it is spatially far from the
target location (acting as an invalid spatial after-cue), target
correct localization rate will be lower compared to the overall
mean localization performance—visual sensory memory will be
selectively mis-cued. In the experiment the singleton object as a
mask/cue will be consistently mapped onto different color and
pattern features compared to the targets and fillers/distractors
(that will be mutually a varied mapping set). If some of the
hypotheses will get support, interpretation in favor or against
particular attentional processing accounts will depend on the
exact combination of the hypotheses that will have received
empirical support.
Independent of any particular theory of processing in search,
the visual system should solve the following sub-tasks: 1)
activate target identity nodes in the visual representation system
when target identity is specified before presentation of alterna-
tives, 2) when alternatives consisting in target and distractors
have been presented, test alternative object-features from
different locations for the match between target features and the
perceptual features of an actual object presented from certain
spatial locations, 3) when a fitting match is found, register the
corresponding spatial location, 4) respond by indicating this
location explicitly.
T. BACHMANN ET AL.
826
The singleton we used was chosen so that it remained fully
uninformative—it had to have no predictive value with regard
to target identity, location or presentation time neither due to
sharing its unique features nor because of being presented
predictably either before, simultaneously with, or after the
target and distractors.
Experiment 1
The aim of Experiment 1 was to find the spatial and temporal
intervals between singleton-object and target object where
singleton helps facilitate target processing and where it leads to
impairment of target processing in order to test the above
described hypotheses. In the within subjects design singleton-
to-target spatial separation and temporal separation between
singleton and target are the main independent variables and rate
of correct target localization is the principal dependent variable.
Method
Participants
A group of 6 subjects (mean age 23; 3 females, 3 males)
participated in the main experiment. They had normal or
corrected-to-normal vision; they participated as paid volunteers
(180 Estonian Kroon per experiment per subject). In piloting
the experiment with fewer trials per subject and using 8
subjects we found qualitatively highly similar data to the data
by the main group, therefore the results obtained with 6
observers can be considered as representative.
Stimuli and Apparatus
Both target and distracting items were selected from the
same set of 8 stimuli; thus a varied mapping format was used.
The stimuli were Gabor patches with varying colour (purple or
green), spatial frequency (2 or 4 cycles/deg) and orientation
(vertical or horizontal). Each one of these features was used an
equal number of times. As targets were defined by conjunction
of several features, pre-attentive filtering could not be used in
effective search. Stimuli subtended .62 degrees of visual angle.
Eight Gabors (the target and 7 fillers/distractors) were presented
in locations forming an imaginary circle (radius 4 deg, with a
small jitter) around the fixation dot. The uninformative feature-
singleton used as a mask and/or uninformative cue was a
yellow ring equal in size to the other 8 stimuli. It was presented
on the same imaginary ring, located either as superimposed
with possible target locations or placed in a midpoint between
two other stimuli. The stimuli were presented on a computer
monitor (Eizo Flex-scan T550), refresh rate 85 Hz, on a grey
background (57 cd/m2) with a duration of 24 ms. An example
of a stimulus display is given in Figure 1.
Procedure
Each trial started with presentation of a fixation cross (.2
deg). Participants initiated presentation of the stimuli by pres-
sing the Enter key on the keyboard. An example of one of the 8
Gabor-type stimuli then appeared for 500 ms in the center of
the display indicating a target stimulus for that trial. After 1200
ms, one of the three temporal types of trials ocurred: 1) yellow
uninformative feature-singleton appearing first, followed by 8
stimuli (1 target, 7 fillers) with SOA varying between 120 ms,
96 ms, 72 ms, 48 ms, and 24 ms; 2) 8 stimuli were presented
first, followed by yellow feature-singleton with SOA varying
between 120 ms, 96 ms, 72 ms, 48 ms, and 24 ms; 3) target and
distractor fillers presented simultaneously with the yellow
feature-singleton, this condition limited to the majority of trials
Figure 1.
An example of search display where one of the Gabor-like stimuli is a
target prespecified by showing it before this type of display, the other
Gabors are fillers/distractors and the yellow disc is a non-informative
singleton object.
where singleton occupied a location different from target.
Feature-singletons appeared at randomly chosen locations
relative to the target on an imaginary ring. Thus spatial distance
between the singleton and target varied between 0 deg, 1.55 deg,
3.04 deg, 4.42 deg, 5.66 deg, 6.65 deg, 7.38 deg, 7.84 deg, and
8.00 deg, including the one with spatial separation 0, i.e., the
singleton and the target were presented from the identical
spatial position (except when SOA = 0 ms where the singleton
was always in a different location from that of the target). After
stimuli presentation, participants indicated by a mouse-click the
spatial position where they perceived the target was presented.
Then the next trial followed. Each subject ran 1000 trials, with
6,000 trials across six participants.
Results and Discussion
Repeated measures ANOVA shows that SOA had no
significant main effect (F(9,45) = 2.01, p < .06). A tendency for
lower level of target correct localization performance with
increasing SOA can be noticed. Spatial distance (separation)
between singleton and target was highly significant (F(8,40) =
23.55, p < .0001). As can be seen also from Figure 2, this owes
first of all to the fact that when mask/cue singleton spatially
overlaps with target (separation distance 0 deg), target correct
localization level dramatically drops. (Vertical bars denote .95
confidence.) There was no interaction between SOA and spatial
separation (F(73,360) = 1.284, p < .074). In the condition with
spatially superimposed singleton and target (distance 0) when
singletons acted as forward masks (i.e., negative SOA values
were used), the target correct localization level dropped
significantly.
Taken together, the basic results are: 1) with most singleton-
to-target spatial and temporal distances neither facilitative nor
interfering effects on target correct localization were found,
which does not support hypotheses 3 - 6; 2) a strong forward-
masking effect emerged with spatially superimposed singletons
T. BACHMANN ET AL.
827
Figure 2.
Proportion of correct target localization responses as a function of SOA between uninformative singleton and target and
spatial distance between singleton and target. Singletons spatially superimposed with targets have a strong interfering effect
both when presented before and after the targets, with the strongest effect taking place with singletons that followed targets in
time. (Data from Experiment 1.)
and targets, supporting hypothesis 1; 3) when singletons that
followed the targets were presented from the same location,
backward-masking effect was strongly evident, supporting
hypothesis 2; 4) with the closest spatial separation between the
non-overlapping singleton and a target there is some significant
singleton effect specifically with later-appearing singletons,
which goes against hypothesis 5 and extends backward masking
effect to the conditions with clearly spatially separated targets
and singleton-as-mask. Spatial pre-cueing from a close target
location does not help to enhance correct target processing rate,
nor does an invalid type of precueing from locations far from
the target produce pre-cueing costs. Interference and masking
effects definitely overweighed any putative spatial-attentional
valid cueing effects supposed to help target perception in case
of spatially close cueing and impair target peception in case of
spatially distant, invalid pre-cueing. From these results we can
conclude that there is no competition for unspecific general
attentional resources between singleton and target, no attention-
capturing effect of the uninformative singletons as pre-cues or
after-cues, but a strong forward and backward masking effect
when the same singletons are used as forward and backward
masks. A novel result shows that a stimulus that is indifferent
for search of targets in its role as a spatial-attentional cue acts
as a strong visual mask for searched target perception. An
uninformative singleton stimulus incapable of engaging effec-
tive involuntary bottom-up attention (i.e., a stimulus that can be
ignored in searching for the pre-designated feature-conjunction
targets) cannot be ignored when used as a spatially overlapping
object mask and vice versa. The only condition where the
spatially non-overlapping singleton has its effect on correct
perception of target location is when it comes after the target
and from the spatially closest non-overlapping position. From
the attentional-mechanisms and iconic-memory mechanisms
point of view, the after-coming singleton presented from the
spatial location separated by slightly more than one degree of
the visual angle from the target should aid selection from iconic
memory and thus facilitate correct perception of target by
acting as a valid after-cue. Why this expected effect did not
take place, thereby contradicting hypothesis 5, will be discussed
later where we suggest possible mechanisms involved.
There are several problems with Experiment 1. First, the
results on the adverse effect of the spatially close, but non-
overlapping after-cue are not very distinct and the highly
significant effect of spatial separation can be attributed mainly
to the condition of spatially overlapping targets and singletons
(see Figure 2). It is therefore advisable to run an additional
experiment where only spatially non-overlapping singletons are
used in order to see whether the adverse effect of spatially close
singletons as after-cues will stay. Second, when in many trials
targets are masked by superimposed singletons as strong object
masks, subjects may have learned that distinctly different
singleton objects are detrimental for target perception (in these
strong masking trials they rarely experience targets in their
awareness) and because of this they become biased against
trying to use singletons as attentional cues. They may have
developed inherent “negative” attentional control settings
“wary” of yellow singleton cues. At the same time this also
may have increased the probability of target localization errors
stemming from intrusion effects of distractors from non-target
locations (e.g., see Chastain, 1990, about mislocalization as a
T. BACHMANN ET AL.
828
common source of errors). For these reasons, Experiment 2 was
run where only spatially non-overlapping singletons were used.
We also increased the number of participants in order to test for
the robustness of the effects.
Experiment 2
The aims of Experiment 2 were to replicate the results of
Experiment 1 using only spatially separated singletons and
targets in the conditions where subjects do not expect overlap-
ping objects masking and with a larger number of subjects.
Otherwise, the design, variables and hypotheses remain the
same, except that we could not test the hypothesis about strong
masking with spatially overlapping singletons and targets.
Method
Participants
Ten subjects participated (mean age 25, 4 females, 6 males).
They were paid volunteers (180 Estonian Kroon per experiment
per subject) who had normal or corrected-to-normal vision.
Stimuli and Apparatus
All stimuli and the apparatus were the same as in Experiment
1, except that less singleton locations were used and singletons
were never presented from the spatial locations superimposed
with targets’ and filler/distractors’ locations. Both target and
distracting items were selected from the same set of 8 stimuli as
in Experiment 1; thus a varied mapping format was used again.
Procedure
The procedure was basically the same as in Experiment 1.
Feature-singletons appeared at randomly chosen locations re-
lative to the target on an imaginary ring. The spatial distance
between the singleton and target varyied between 1.55 deg,
4.42 deg, 6.65 deg, and 7.84 deg of visual angle. Each subject
ran 1000 trials, with 10,000 trials across ten participants.
Results and Discussion
ANOVA was used to analyze of the effects of the factors of 1)
SOA between target and masking singleton (11 levels: –120,
–96, –72, –48, –24, 0, 24, 48, 72, 96, 120), 2) spatial distance
between masking singleton and target (4 levels: 1.55 deg, 4.42
deg, 6.65 deg, and 7.84 deg), 3) target identity (8 levels). A
repeated measures ANOVA shows that SOA had a moderately
significant effect on correct target localization (F(10, 90) = 2.11,
p < .031). As can be seen from Figure 3, this effect basically
depends on the detrimental effect of the singleton that was
presented more than 50 ms later than the target and only if
presented from the closest spatial position from target.
Consistent with this, the interaction between SOA and spatial
distance was highly significant (F(30,270) = 2.26; p < .0003).
Spatial distance between the singleton and target also had a
highly significant effect that was also based on the conditions
of closest spatial separation between target and singleton with
positive SOA values (F(3,27) = 15.10, p < .00001) (see Figure
3). These results basically repeat the results of Experiment 1 in
the conditions where singleton and Gabor stimuli do not
spatially overlap and, particularly, substantiate that the only
effect of the singleton emerges when it appears as an after-cue
presented from the close spatial position to target. However,
instead of the supposedly facilitative effect of the singleton
Figure 3.
Proportion of correct target localization responses as a function of SOA between uninformative singleton and target
presentation and spatial distance between singleton and target. Only spatially close singletons presented 60 - 120 ms after
target had a distractive e ff e ct . Vertical bars denote .95 confidence. (Data from Experiment 2.)
T. BACHMANN ET AL.
829
expected to act as a selective attentional after-cue (to foster
target selection from the iconic representation) there is an
opposite effect - backward maski n g. There is also no facilitative
precueing, no costly invalid precueing, no detrimental forward
masking. (We would like also to note that the SOA values we
used corresponded well to the optimal attentional pre-cueing
SOAs in the range of about –50 to –150 ms as found in earlier
research—e.g., Cheal and Lyon, 1991.)
Introspective observations of subjects gathered after the
experiment showed that in the trials where a near by aftercoming
singleton was effective in reducing target localization perfor-
mance, often no explicit experience of the target was possible.
Subjectively it appeared that the singleton substituted the
nearby stimuli in awareness. Either mutual masking between
objects (Bachmann, & Allik, 1976; Hommuk & Bachmann,
2009; Michaels, & Turvey, 1979) or a new type of object
substitution masking (Di Lollo et al., 2000; Enns, 2004) took
place. The additional data about the percentages of errors of
localization showed that subjects erred with roughly equal
frequency whether their responded location was far from the
actual target position or close to that position (percentages
between 13.5 and 16.5, ns). This means that it is unlikely that
errors in target localization originate primarily from trials
where target is expli citly percei ved, but slightly mislocalized.
To sum up in interim: 1) a feat urally and temporally uninfor-
mative singleton can impair target perception as measured by
correct target localization, but only in limited experimental
conditions, 2) the singleton effect appears even though the same
singleton does not have any facilitatory or interfering (costly)
effects of involuntary attentional capture on target localization
(when presented as a pre-cue) and thus the masking may not
originate directly from attentional mis-cueing, but appears as a
complication in the process of building up explicit re-
presentation of the target, 3) the effect appears only when the
singleton is presented more than 50 ms after target and from the
nearest spatial position. No other significant main effects or
interactions were found in this experiment. This pattern seems
to be well accounted for by some unusual form of backward
masking. However, it cannot be pattern masking because target
and singleton do not overlap (and pattern masking is typically
strongest with shortest target-to-mask SOAs—Bachmann,
1994). It is not also standard metacontrast masking because the
spatial distance between target and singleton is too large (1.55
deg) and most of the target space is not surrounded by singleton
elements. It is also difficult to accept the effect as object
substitution masking (OSM) in its commonly accepted interpre-
tation because the masks in the OSM paradigm have been
metacontrast-like weak masks surrounding the target, with their
effect depending on attention involvement. In our results, the
backward object-masking effect was combined with insensiti-
vity to attentional manipulations. Therefore, an option for
interpretation would be either 1) to extend OSM effects to the
ones where spatially neighbouring stimuli from as far as 1.55
degrees could intrude expli cit attentive perception instead of the
target or 2) to abandon the attentional explanation (that includes
OSM as a theory critically dependent on whether selective
attention can be deployed) and use some mechanism capable of
depriving the target information from reaching conscious-level
representation without directly invoking attention mechanisms.
We come to theoretical discussion later on.
In experiments 1 and 2 the most compelling result besides
strong forward and backward object masking was an effect of
spatially remote lateral backward masking. But because we
always used many distractor/filler objects we cannot be sure
whether this effect originates primarily from c lose-range la teral
interaction between target and the subsequent nearby singleton
without any processing capacity or spatial-location uncertainty
related effects or whether this new form of object masking
critically depends on these higher level factors.
Experiment 3
In order to test directly whether a lateral-inhibitory masking
of target identity features could be the main reason for the
effects in Experiments 1 and 2, a simple identification
experiment (Experiment 3) without distractors was run. We
explored the effect of an uninformative feature-singleton (that
was presented from different spatial locations never used for
presentation of targets) on correct target identification as a
function of spatial distance between singleton and target and on
temporal separations between singleton and target. If local
lateral inhibitory interaction between singleton and target is the
principal or main cause of the effects found in the previous
experiments, the close spatial distance between a singleton and
a target should lead to a strong masking in this experiment as
well. The basic design remained the same, but as the dependent
measure rate of correct identification of targets was used. We
test the hypothesis that when distractor objects are absent and
attention must not be divided between alternative objects, the
spatially close singleton will interfere with target processing,
indicating a close-range lateral inhibition as the likely cause of
the effects of Experiments 1 and 2.
Method
Subjects
Altogether 4 participants (mean age 35, two females, two
males) participated. They had normal or corrected-to-normal
vision.
Stimuli and Apparatus
Each stimulus used as a target was drawn from a set of 8
stimuli alternatives identical to those used in the previous
experiments. In each trial one of the stimuli was presented for
identification, located in one of the 8 locations forming an
imaginary circle around the fixation dot (with a small radial
jitter). Thus a target-stimulus appeared 3.8 - 4.3 degrees from
fixation. The uninformative feature singleton for exerting lateral
effects and for competing for bottom-up spatial attention with
the target was set as a yellow ring singleton identical to the one
used in the previous experiments. All of the equipment used
was identical to that of the previous experiments.
Procedure
Each trial started with a presentation of the fixation cross (.2
deg). Participants initiated presentation of the stimuli by
pressing the Enter key on the keyboard. After 1200 ms, one of
the three temporal types of trials occurred: a singleton appearing
first, followed by the target appearing randomly in one of the 8
possible positions, with SOA varying between 120 ms, 96 ms,
72 ms, 48 ms, and 24 ms; a target presented first, appearing
randomly in one of the 8 possible positions, followed by a
singleton with SOA varying between 120 ms, 96 ms, 72 ms, 48
ms, and 24 ms; a target appearing randomly in one of the 8
possible locations and a singleton that was presented simulta-
neously with target. The singleton appeared in randomly chosen
locations with the spatial distance between singleton and target
varying randomly between the two distances of 1.55 deg or
7.84 deg of visual angle. After stimuli presentation, participants
T. BACHMANN ET AL.
830
indicated by a mouse-click which one of the 8 possible targets
was presented at that trial, guessing if necessary. Then next trial
followed. Each subject ran between 540 and 960 trials, within a
preset time of 60 minutes.
Results and Discussion
Data from each subject was subjected to an ANOVA for the
analysis of the effects of the factors of 1) SOA between target
and singleton (11 levels: –120, –96, –72, –48, –24, 0, 24, 48, 72,
96, 120), 2) spatial distance between singleton and target (2
levels: 1.55 deg, 7.84 deg). Repeated measures ANOVA shows
no effect of SOA (F(10,30) = 1.42, p < .218), there was an
effect of the distance between target and singleton (F(1,3) =
25.1, p < .015), but no interaction between SOA and distance
showing that (and as different from the previous results) longer
SOAs did not selectively lead to a stronger singleton effect
(F(10,30) = .88, p < .562). (See Figure 4 for the graph depic-
ting the principal results.)
The suggested hypothesis predicted a difference in the level
of identification between the conditions of close and far
singleton-to-target separation stemming from the idea of low-
level sensory lateral interaction between singleton and target.
However, and importantly so with longer positive SOA values,
increased interfering effect from distracting singleton when far
and adjacent target-to-singleton distance conditions were
compared, was absent. Local lateral interactions between the
singleton and target are ineffective for selectively stronger
impairment of target identification at specific SOAs. The
uninformative singleton used in our experiments and having
within-stimulus visual features different from the features of the
targets (different unique singleton colour and absence of
grating-features in the singleton contrary to the gratings with
varying spatial frequencies and orientations used in targets) had
neither selectively increased facilitating effects nor distracting
effects as a function of the spatial distance from target. The
correct identification rate at about 60% excludes both ceiling
and floor effects. We conclude that a singleton used in our
study can be consideread as a “weak” lateral backward mask
when the possibility of local lateral interactions potentially
effective on target identification were examined and when we
assume insensitivity of the local early-level masking to the
extent of attention distribution. (Notice the smaller number of
the stimuli competing for attention in Experiment 3 compared
to first two experiments.) At the same time this masking
singleton was capable of strong masking effects in target
location search in Experiments 1 and 2.
General Discussion
This study shows that a singleton object that is a powerful
spatially overlapping visual mask in impairing target search
among distractors does not have power to capture involuntary
attention. This salient object that does not have power to influ-
ence visual target search when used as a singleton cue to cap-
ture attention (both as a valid cue expected to cause improve-
ment and as an invalid cue expected to cause processing costs)
nevertheless acts as a strong forward and backward mask when
spatially overlapping with target. Importantly, the only condi-
tion where an uninformative salient singleton has a clear effect
on target perception besides typical pattern/object masking is
Figure 4.
Proportion of correct target identification responses as a function of SOA between uninformative singleton and target and
spatial distance between singleton and target, averaged over 4 subjects. Correct target identification rate is same whether
singleton was presented adjacent t o ta r ge t or fr om a fa r location with longer SOAs. (Data from Experiment 3.)
T. BACHMANN ET AL.
831
when it is presented not far from the target and after it in time.
Notably, this salient but irrelevant object can be easily ignored
in a simple target identification task when no distractor ele-
ments are accompanying the target, but it has a strong adverse
effect in the visual search task with the same target objects
when their correct location has to be specified and filler/dis-
tractor items are presented simultaneouslyt with target. There-
fore, the explanation for the spatially remote masking effect
from the singletons that appear 50 - 100 ms after the target
excludes early-level lateral inhibition.
For a spatially non-overlapping target and an irrelevant ob-
ject there is no mutual interference with SOA set at zero or with
negative SOAs. This means that an irrelevant singleton does
not necessarily impair finding the target when presented just
adjacent to it. Also, it does not interrupt search when presented
after the target, but far from its spatial position. (Nor does it
facilitate target processing when presented before it in time and
near to it in space, as would be the case when it would act like
an effective, valid, pre-cue.)
At first sight the results found here seem to suggest an inter-
pretation quite similar to what is used for explaining the stan-
dard substitution masking (DiLollo, Enns, & Rensink, 2000;
Enns, 2004) and some other backward masking phenomena
where masking strength is determined not only by the low-level
visual-spatial and temporal relations between target and mask,
but also by the dynamics of attention (e.g. Smith, 2000; Smith,
& Wolfgang, 2004). Similarly to standard substitution-masking
which is studied with selective spatial attention controlled and
manipulated, target loss from explicit perception as found in the
present study may appear to presuppose an unfinished job of
focusing spatial attention in obtaining the masking effects.
However, as the singleton did not have any attentional cueing
effects, its attention-capturing power can be doubted and there-
fore we may need to consider involvement of the mechanisms
other than attention. A mechanism responsible for upgrading
the pre-conscious target representation to explicit conscious-
level representation would be a good candidate. This is espe-
cially so because recent research has shown independence or
even opposite effects of attention and consciousness—e.g.,
Koch and Tsuchiya, 2007; Bachmann and Murd, 2010. Many
distractor elements set the stage for the singleton mask to have
its effect and do this by depriving the target from the service by
the consciousness mechanism, the target becoming masked by
the singleton when it is close in space and follows the target.
The results show that stimulus identification mechanisms and
spatial selection mechanisms that are needed to localize preat-
tentively discriminated targets are at least partly autonomous. If
a spatial selection operation that ultimately serves explicit
perception would be dependent on feature-specific analysis,
singletons should not get an advantage over targets because
their features are clearly discriminable and different. It is
possible to discuss the results from the point of view of possible
beneficial selective spatial attentional effects mediated by
spatially localized singletons and also from the perspective of
possible interfering/inhibiting (costly) effects of the same
singletons. Our results show that singletons that exert neither
icreased local lateral sensory effects (either disturbing or
facilitating ones) nor invalid pre-cueing costs on target identifi-
cation do exert disturbing effects in the conditions where a pre-
known target has to be found among distractor items and
correctly localized. This effect comes about only retroactively,
i.e., when targets and distractors have been already presented.
In this case a singleton that in principle can be ignored as a
competing item because of its consistently different featural
identity in the identification task cannot be ignored as a token
that competes with the target for their location to be specified.
Uninformative singletons that are presented close to or over-
lapping with subsequent targets did not facilitate target
processing in the search/localization task. (Neither did spatially
remote singletons produce any invalid cueing costs.) This
makes our experimental conditions compatible with the require-
ments needed to satisfy the principle that uninformative single-
tons are inefficient as attention-capturing stimuli (e.g., Becker,
2007; Gibson et al., 2008; Yantis & Jonides, 1990; Jingling &
Yeh, 2007). This state of affairs does not mean that singletons
can be absolutely ignored or made ineffective due to some
suitable control setting. Our results stress that in order to claim
absence of any singleton effects on target processing a
systematic change in spatial and temporal values of intervals
between targets and singletons has to be used. This is because
we have found a highly selective singleton effect in space and
time. When a post-target singleton appears close to a target that
is pre-attentively processed but, in terms of its features—
explicitly unidentified, the singleton disturbs target processing
by substituting it or by making an obstacle for explicit location-
bound identification of it in the conscious representation. Our
present experiments do not specify whether this interference
takes place at the parallel stage of processing or at some stage
of serial processing when alternative items are analysed for a
match with the target cues. However, it has to occur before the
target has been explicitly localised by being experienced in
consciousness.
In guided search models (Hoffman, 1979; Wolfe, Cave, &
Franzel, 1989) it is possible to envisage a scenario where the
distracting singleton interferes at the unfinished parallel pre-
attentive stage of processing by substituting the target (akin to
singleton pop-out) or at the stage of subsequent focal selection
of an item to be responded to where, again, the singleton can
substitute for the target and indicate to an observer that this
location of interest, “unfortunately”, contains an irrelevant
stimulus. (The priority of spatial location over other attributes
in top-down controlled visual search is supported, e.g., by
Grabbe and Pratt, 2004; Kim and Cave, 1999.)
The time course of the interference effect suggests that the
target’s appearance in explicit perception with its actual location
being experienced cannot be generated very fast: the after-
coming singleton’s interfering effect on target perception
extended to more than 100 ms. It is therefore not surprising that
single-cell studies of stimulus pop-out also indicate that at least
70 - 230 ms post-target time appears to be necessary for a
secondary, top-down effect of target enhancement (e.g., Smith,
Kelly, & Lee, 2007). Moreover, the recent work on localized
attentional interference between neighboring visual-object
representations also showed that spatial distance takes its effect
mostly with longer target-to-mask/cue SOAs (Steelman-Allen,
McCarley, & Mounts, 2009). However, in the present study the
objects that do not show any capability to invoke spatial-
attentional capture facilitating target perception or impairing
target perception by effective misdirection of selective attention
nevertheless influence explicit perception of targets in their
correct location. Yet another attentional mechanism to be
considered for explaining our results is related to center-
surround profile of the focus of attention (e.g., Boehler, Tsotsos,
Schoenfeld, Heinze, & Hopf, 2009). In this case, however,
there are two problems. First, the optimal delay for the top-
down controlled and temporally delayed surround attenuation
effects (i.e., formation of the inhibitory area around the single-
ton in our experiment) was shown to be more than 175 ms,
T. BACHMANN ET AL.
832
which is by far too slow an effect compared to our data. Second,
this inhibitory surround effect has been related to feature-
binding operations, but our singleton is uninformative and does
not share varied-mapping features with targets. Third and most
importantly, our adverse effect of singleton on target perception
was found when the singleton followed the target in time and
therefore target must have been inhibiting the later-coming
singleton and not vice versa.
Recently, Munneke, Van der Stigchel and Theeuwes (2008)
showed that an irrelevant distractor can be made less efficient
by a top-down inhibitory mechanism that helps to better ignore
the irrelevant onset-stimulus. But even then the object-substitu-
tion effects strongly interfered with the target’s explicit identifi-
cation, even though the relevant spatial location has been
already successfully established. This means that in addition to
identity processing, visual awareness also requires correct
localization of the target stimulus within the map of stimuli
locations. It is even likely that, in principle, there cannot be
distinct visual awareness at all if an identified object has not
been granted its stable spatial position within the microgeneti-
cally evolving subjective perceptual image (e.g., Bachmann,
1994).
Taken together, the present study demonstrates a new version
of masking in visual search where the need for features-based
guidance of attentional search among the varied-mapping alter-
natives makes the target vulnerable from an otherwise ineffec-
tive masker. The singleton can substitute the target also from a
spatially shifted position, leaving the target often out of cons-
cious experience.
Acknowledgements
Support provided by Estonian Ministry of Education and
Research through Scientific Competency Council (targeted
financing research theme SF0182717s06, “Mechanisms of
Visual Attention”) is very much appreciated.
References
Ansorge, U., & Horstmann, G. (2007). Preemptive control of atten-
tional capture by colour: Evidence from trial-by-trial analyses and
orderings of onsets of capture effects in reaction time distributions.
Quarterly Journal of Experimental Psychology, 60, 952-975.
doi:10.1080/17470210600822795
Bachmann, T. (1994). Psychophysiology of visual masking. The fine
structure of conscious experience. Commack, New York: Nova Sci-
ence Publishers.
Bachmann, T., & Allik, J. (1976). Integration and interruption in the
masking of form by form. Perception, 5, 79-97.
doi:10.1068/p050079
Bachmann, T., & Murd, C. (2010). Covert spatial attention in search for
the location of a color-afterimage patch speeds up its decay from
awareness: Introducing a method useful for the study of neural cor-
relates of visual awareness. Vision Research, 50, 1048-1053.
doi:10.1016/j.visres.2010.03.013
Bachmann, T., Mäger, K., Sarv, M., Kahusk, N., & Turner, J. (1999).
Time-course of spatial-attentional focusing in the case of high proc-
essing demand on the peripheral precue. European Journal of Cogni-
tive Psychology, 11, 167-198. doi:10.1080/713752312
Becker, S. I. (2007). Irrelevant singletons in pop-out search: attentional
capture or filtering costs? Journal of Experimental Psychology: Hu-
man Perception and Performance, 33, 764-787.
doi:10.1037/0096-1523.33.4.764
Boehler, C. N., Tsotsos, J. K., Schoenfeld, M. A., Heinze, H.-J., &
Hopf, J.-M. (2009). The center-surround profile of the focus of atten-
tion arises from recurrent processing in visual cortex. Cerebral Cor-
tex, 19, 982-991. doi:10.1093/cercor/bhn139
Breitmeyer, B. G. (2010). Blindspots. The many ways we cannot see.
Oxford, New York: Oxford Universitgy Press.
Chastain, G. (1990). Representation of letters when mislocation errors
occur. Journal of General Psychology, 117, 143-151.
doi:10.1080/00221309.1990.9921131
Cheal, M. L., & Lyon, D. R. (1991). Central and peripheral precuing of
forced-choice discrimination. Quarterly Journal of Experimental
Psychology, 43, 859-880. doi:10.1080/14640749108400960
Davoli, C. D., Suszko, J. W., & Abrams, R. A. (2007). New objects can
capture attention without a unique luminance transient. Psychonomic
Bulletin and Review, 14, 338-343. doi:10.3758/BF03194074
Di Lollo, V., Enns, J. T., & Rensink, R. A. (2000). Competition for
consciousness among visual events: the psychophysics of re-entrant
visual processes. Journal of Experimental Psychology: General, 129,
481-507. doi:10.1037/0096-3445.129.4.481
Enns, J. T. (2004). Object substitution and its relation to other forms of
visual masking. Vision Research, 44, 1321-1331.
doi:10.1016/j.visres.2003.10.024
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. doi:10.1037/0096-1523.18.4.1030
Forster, S., & Lavie, N. (2008). Attentional capture by entirely irrele-
vant distractors. Visual Cognition, 16, 200-214.
doi:10.1080/13506280701465049
Gibson, B. S., Folk, C., Theeuwes, J., & Kingstone, A. (Eds.) (2008).
Attentional capture. Hove: Psychology Press.
Grabbe, Y., & Pratt, J. (2004). Competing top-down processes in visual
selection: Evidence that selection by location is stronger than selec-
tion by color. Journal of G e n er a l Psychology, 131, 137-149.
doi:10.3200/GENP.131.2.137-150
Hoffman, J. E. (1979). A two-stage model of visual search. Perception
& Psychophysics, 25, 319-327. doi:10.3758/BF03198811
Hommuk, K., & Bachmann, T. (2009). Temporal limitations in the
effective binding of attended target attributes in the mutual masking
of visual objects. Journal of Experimental Psychology: Human Per-
ception and Performance, 35, 648-660. doi:10.1037/a0013481
Jingling, L., & Yeh, S.-L. (2007). New objects do not capture attention
without a top-down setting: evidence from an inattentional blindness
task. Visual Cognition, 15, 661-684.
doi:10.1080/13506280600926695
Kahneman, D., & Treisman, A. (1984). Changing views of attention
and automaticity. In R. Parasuraman and D. R. Davis (Eds.), Varie-
ties of attention (pp. 29-61). Orlando: Academic Press.
Kawahara, J., & Miyatani, M. (2001). The effect of informative and
uninformative cueing of attention on feature integration. Journal of
General Psychology, 128, 51-75. doi:10.1080/00221300109598898
Kim, M. S., & Cave, K. R. (1999). Grouping effects on spatial attention
in visual search. Journal of General Psychology, 126, 326-352.
doi:10.1080/00221309909595370
Koch, C., & Tsuchiya, N. (2007). Attention and consciousness: two
distinct brain processes. Trends Cognitive Sciences, 11, 16-22.
doi:10.1016/j.tics.2006.10.012
Lichtenstein-Vidne, L., Henik, A., & Safadi, Z. (2007). Task-relevance
modulates the effects of peripheral distractors. Quarterly Journal of
Experimental Psychology, 60, 1216-1226.
doi:10.1080/17470210600917892
Michaels, C. F., & Turvey, M. T. (1979). Central sources of visual
masking: Indexing structures supporting seeing at a single, brief
glance. Psychologic al Research, 41, 1-61. doi:10.1007/BF00309423
Mulckhuyse, M., & Theeuwes, J. (2010). Unconscious attentional ori-
enting to exogenous cues: A review of the literature. Acta Psy-
chologica, 134, 299-309. doi:10.1016/j.actpsy.2010.03.002
Munneke, J., Van der Stigchel, & Theeuwes, J. (2008). Cueing the
location of a distractor: An inhibitory mechanism of spatial attention?
Acta Psychologica, 129, 101-107. doi:10.1016/j.actpsy.2008.05.004
Müller, H. J., & Krummenacher, J. (Eds.) (2006). Visual search and
attention. Hove: Psycho l og y Press.
Neo, G., & Chua, F. K. (2006). Capturing focused attention. Perception
& Psychophysics, 68, 1286-1296. doi:10.3758/BF03193728
Smith, M. A., Kelly, R. C., & Lee, T. S. (2007). Dynamics of response
to perceptual pop-out stimuli in macaque V1. Journal of Neurophysi-
ology, 98, 3436-3449. doi:10.1152/jn.00441.2007
T. BACHMANN ET AL.
833
Smith, P. L. (2000). Attention and luminance detection: Effects of cues,
masks, and pedestals. Journal of Experimental Psychology: Human
Perception and Performance, 2 6, 1401-1420.
doi:10.1037/0096-1523.26.4.1401
Smith, P. L., & Wolfgang, B. J. (2004). The attentional dynamics of
masked detection. Journal of Experimental Psychology: Human Per-
ception and Performance, 30, 119-136.
doi:10.1037/0096-1523.30.1.119
Steelman-Allen, K., McCarley, J. S., & Mounts, J. R. W. (2009).
Localized attentional interference reflects competition for reentrant
processing. Psychonomic Bulletin & Review, 16, 110-115.
doi:10.3758/PBR.16.1.110
Turatto, M., & Galfano, G. (2000). Color, form and luminance capture
attention in visual search. Vision Research, 40, 1639-1643.
doi:10.1016/S0042-6989(00)00061-4
Van den Bussche, E., Hughes, G., Van Humbeeck, N., & Reynvoet, B.
(2010). The relation between consciousness and attention: An em-
pirical study using the priming paradigm. Consciousness and Cogni-
tion, 19, 86-97. doi:10.1016/j.concog.2009.12.019
Vierck, E., & Miller, J. (2008). Precuing benefits for color and location
in a visual search task. Perception & Psychophysics, 70, 365-373.
doi:10.3758/PP.70.2.365
Wolfe, J. M., Cave, K. R., & Franzel, S. L. (1989). Guided search: An
alternative to the feature integration model for visual search. Journal
of Experimental Psychology: Human Perception and Performance,
15, 419-433. doi:10.1037/0096-1523.15.3.419
Yantis, S., & Jonides, J. (1990). Abrupt visual onsets and selective
attention: voluntary versus automatic allocation. Journal of Experi-
mental Psychology: Human P erception & Performance, 16, 121-134.
doi:10.1037/0096-1523.16.1.121
Yeh, S.-L., & Liao, H.-I. (2010). On the generality of the displaywide
contingent orienting hypothesis: Can a visual onset capture attention
without top-down control settings for displaywide onset? Acta Psy-
chologica, 135, 159-167. doi:10.1016/j.actpsy.2010.05.013