Attention Makes Moving Objects Be Perceived To Move Faster

Attention makes moving objects be perceived to move faster
Massimo Turatto a,b,*, Massimo Vescovi a
, Matteo Valsecchi a
a
Department of Cognitive Sciences and Education, Rovereto, Italy
b
Center for Mind–Brain Sciences University of Trento, Rovereto, Italy
Received 24 February 2006; received in revised form 29 September 2006
Abstract
Although it is well established that attention affects visual performance in many ways, by using a novel paradigm [Carrasco, M., Ling,
S., & Read. S. (2004). Attention alters appearance. Nature Neuroscience, 7, 308–313.] it has recently been shown that attention can alter
the perception of different properties of stationary stimuli (e.g., contrast, spatial frequency, gap size). However, it is not clear whether
attention can also change the phenomenological appearance of moving stimuli, as to date psychophysical and neuro-imaging studies
have specifically shown that attention affects the adaptability of the visual motion system. Here, in five experiments we demonstrated
that attention effectively alters the perceived speed of moving stimuli, so that attended stimuli were judged as moving faster than less
attended stimuli. However, our results suggest that this change in visual performance was not accompanied by a corresponding change
in the phenomenological appearance of the speed of the moving stimulus.
2006 Published by Elsevier Ltd.
Keywords: Attention; Motion perception; Phenomenology
1. Introduction
What we are aware of seeing is what we attend to, since
attention profoundly restricts the contents of our limited
conscious perception (e.g., Mack Rock, 1998). Besides
determining what we see, does visual attention also affect
the way that objects are perceived? Although this has
recently become a timely and important question in visual
neuroscience (e.g., Carrasco, Ling, Read, 2004), the
question of whether the appearance of objects is modulated
by attention was a major issue that characterized the first
steps of modern experimental psychology almost a century
ago. Indeed, on the basis of independent informal observa-
tions, some prominent figures in the this field like William
James, Wilhelm Wundt and Hermann Ebbinghaus firmly
believed that attention intensifies the sensation of a stimu-
lus (e.g., Ebbinghaus, 1908; James, 1980).
Since then, a few studies have been conducted in the last
30 years that have thoroughly addressed this issue, some of
which agreed with the abovementioned view that attention
can boost sensation (e.g., Festinger, Coren, Rivers,
1970); other studies, quite surprisingly, provided evidence
to the contrary by showing that, for example, attention
reduces the perceived brightness contrast (Tsal, Shalev,
Zakay, Lubow, 1994) and perceived length (Tsal Sha-
lev, 1996). More recently, the relation between objects’
appearance and attention has been reconsidered and inves-
tigated by Prinzmetal, Nwachuku, Bodanski, Blumenfeld,
and Shimizu (1997). In a series of experiments the authors
addressed the effects of attention on perceived contrast and
brightness. To overcome possible methodological flaws in
previous studies and to obtain a more controlled manipu-
lation of attention allocation, Prinzmetal and colleagues
devised a novel experimental procedure based on a dual-
task paradigm. On each trial, participants were briefly pre-
sented with an array of letters at the centre of the screen,
plus a small achromatic disk either to the left or right of
the letters. In the (crucial) simultaneous onset condition
0042-6989/$ - see front matter 2006 Published by Elsevier Ltd.
doi:10.1016/j.visres.2006.10.002
*
Corresponding author.
E-mail address: massimo.turatto@unitn.it (M. Turatto).
www.elsevier.com/locate/visres
Vision Research 47 (2007) 166–178

the letters and the disk appeared concurrently, and partic-
ipants first had to report which of two possible target let-
ters was present, and then they had to indicate the
brightness (or contrast) of the disk by adjusting a cursor
using one of two vertical brightness response palettes.
The main result of Prinzmetal et al. (1997) can be summa-
rized as follows: attending to the stimuli did not make them
appear brighter or at higher contrast – namely attention
did not alter stimulus’ appearance.
In the last 2 years, however, the view that attention does
not change the phenomenological appearance of stimuli
has been challenged by two studies which reported evi-
dence to the contrary (Carrasco et al., 2004; Gobell
Carrasco, 2005). To study the effect of attention on objects’
appearance, Carrasco and her collaborators devised a nov-
el paradigm, which basically stems from that used in other
studies in which the effects of covert attention on contrast
sensitivity were evaluated (e.g., Cameron, Tai, Carrasco,
2002; Pestilli Carrasco, 2005). Shortly, the modified par-
adigm, which will be used in the present study as well, con-
sists in the brief (about 60 ms) presentation of a peripheral
visual onset (the cue) to transiently capture attention to a
given location (Nakayama Mackeben, 1989). About
50 ms later, two stimuli consisting of Gabor patches (sinu-
soidal achromatic gratings enveloped by a Gaussian filter)
of different contrasts and/or spatial frequencies are briefly
presented, one on each side of fixation. The task of the
observer is to report, for example, the orientation (left vs.
right) of the Gabors that appears higher in contrast (Carr-
asco et al., 2004). If contrast is the feature under investiga-
tion, one of the Gabor (referred to as the ‘‘standard’’) has a
fixed contrast value, while the other (referred to as the
‘‘test’’) is presented at the same, lower or higher contrast
values. To manipulate the spatial allocation of attention
in a controlled manner, the cue is randomly delivered either
to the standard, the test or at fixation (control condition).
By means of this paradigm, which has also been adopted to
address the phenomenological effect of attention on the
appearance of spatial frequency and gap size (Gobell
Carrasco, 2005), Carrasco et al. (2004) documented that
attention can alter the phenomenology of our visual per-
ception. In their study human observers perceived the
attended stationary Gabor of a pair as higher in contrast
than the less attended one. This is quite a remarkable
result, as it shows that visual attention not only plays a cru-
cial role in determining what we are aware of Mack and
Rock (1998), but that, to some extent and at least for static
stimulus features, attention also affects the way that we see.
One of the most pervasive characteristics common to
many stimuli in our visual field is motion. Indeed, in every-
day life moving objects are continuously part of our visual
experience, such as people moving around us or cars in the
street. Moreover, in many circumstances, and for different
reasons, we are often required to pay attention to these
moving objects, as when we look at the train approaching
on the track or a player running in a football game. Hence,
the important question we addressed in the present study is
whether attention, besides altering the perceived contrast,
can also alter the perception of a more complex stimulus
property such as motion speed.
Interactions between attention and motion perception
have already been documented by using brain imaging
and behavioral techniques. Single-cell recording (Marti-
nez-Trujillo Treue, 2002; Treue Maunsell, 1996) as
well as fMRI studies (Rees, Frith, Lavie, 1997) have
shown that attention modulates neural responses in the
motion-related brain area V5/MT. In addition, psycho-
physical studies have documented that the illusory motion
after-effect (MAE) is affected by visual attention (e.g.,
Chaudhuri, 1990). Specifically, in a MAE task participants
are first exposed to an adapting moving stimulus, and then
to a physically stationary one. Because of the prolonged
exposure to visual motion in one direction, the subsequent
stationary stimulus is perceived as moving in the opposite
direction, an illusion that fades over time. Interestingly,
the time course of the MAE has been shown to be sensitive
to attention, as less attended adapting stimuli reduce the
duration of the corresponding after-effect (Shulman,
1991, 1993). In addition, evidence exists showing that
attention not only modulates the duration of the MAE
but also its direction. In an elegant experiment Lankheet
and Verstraten (1995) presented their participants with
two superimposed random dot patterns moving in opposite
directions. In the control condition, when attention was
not biased to one of the two patterns, no MAE was
observed, as the two vectors of movement cancelled each
other out in the MAE. However, when attention was
deployed to the rightward pattern, for example, the MAE
shifted towards the left, thus showing a selective adaptation
of the visual system to the attended motion signal. For the
attended and non-attended motion signals shared the same
location (also see, Alais Blake, 1999), these results
extended those of Chaudhuri (1990), who reported a decre-
ment of the MAE for stimuli at non-attended locations, by
showing that the effect of attention on the motion system
can take place at feature analysis level rather than being
spatially specific.
Although these previous studies nicely documented that
attention affects motion perception, so far what has been
shown is that attention modulates the adaptability of the
visual motion system (e.g., Rezec, Krekelberg, Dobkins,
2004). Yet, none of them provide evidence for a direct and
phenomenological online modulation of an attended moving
object. Hence, whether or not attention (besides modulat-
ing the strength of the MAE) could effectively change the
actual appearance of motion perception, and specifically
the perceived motion speed, still remains an open issue.
In the present study we present five experiments aimed at
addressing this issue directly.
2. Experiment 1
As previously anticipated, to address whether attention
modulates perceived speed of moving stimuli we adapted
M. Turatto et al. / Vision Research 47 (2007) 166–178 167

the Carrasco et al. (2004) paradigm to the presentation of
moving objects. Recently, this paradigm has successfully
been used to demonstrate that attention can alter not only
the perceived contrast, but also spatial frequency and gap
size (Gobell Carrasco, 2005). In principle, the fact that
attention alters the perception of static stimulus properties
such as contrast, spatial frequency and gap size, does not
logically imply that attention can also change the phenom-
enology of motion perception. Indeed motion perception is
ultimately processed in different neurological brain areas
(V5/MT, MST) as compared to those involved in the
analysis of the static stimulus properties mentioned above
(V1, V2).
In the present study, spatial attention was exogenously
attracted to the left or right of fixation via a task-irrelevant
cue consisting of a peripheral onset (Jonides, 1981; Nakay-
ama Mackeben, 1989). After the cue onset, two vertical
achromatic sinusoidal gratings briefly appeared, one on
each side of the fixation point. The gratings moved either
leftward or rightward, and observers were asked to report
the direction of movement of the one that appeared to
move faster.
2.1. Method
2.1.1. Participants
Eighteen undergraduate students from the University of
Trento served as participants. They were compensated
either with course credits or 8€ for their participation. All
were naı̈ve as to the purpose of the experiment, reported
normal or corrected-to normal vision, and gave informed
written consent to participate in the study.
2.1.2. Apparatus
Participants sat approximately 60 cm in front of an
iiyama CRT 1900
(1024 · 768, 150 Hz) monitor. Generation
and presentation of the stimuli was controlled by a custom-
made program written using Matlab and the Psychophysics
Toolbox (Pelli, 1997), and running under Windows 2000 on
a Pentium IV Dell PC. In addition, to ensure that fixation
was maintained throughout the trial, eye movements were
monitored and recorded by an Eyelink II system (SR
Research, Ont., Canada). Head movements were restrained
by using a cheekbone and chin rest device.
2.1.3. Stimuli and procedure
The stimuli were presented over a grey background
(27.7 cd/m2
). As shown in Fig. 1, each trial started with
the presentation of a black fixation point (a disk subtend-
ing 0.2 of visual angle) in the centre of the screen, and
participants had to maintain their gaze at fixation through-
out the trial. 1000 ms later a visual onset (the cue) consist-
ing of a black disk (0.6) was transiently presented for
59 ms. The cue could appear either at fixation (control con-
dition), or just above (2 vertically) one of the two periph-
eral positions (4 left or right of fixation) where the target
stimuli appeared 53 ms later (see Fig. 1). The target stimuli
consisted of two Gabor patches (vertical sinusoidal achro-
matic gratings enveloped by a Gaussian filter; 3 · 3) pre-
sented for 200 ms 4 to the left and right of fixation. The
Gabors’ motion was achieved by independently shifting
the two gratings leftward or rightward. One of the two
Gabors (the standard) moved at a fixed speed (4.29/s),
whilst the other (the test) moved at seven different possible
speeds (1.88, 3.13, 3.75, 4.29, 5.02, 6.82, or 15/s). The
speeds were preliminarily determined in a pilot study,
and were chosen as to create a sigmoid-like function in
the control condition.
The contrast of the Gabors was 60%. Once the Gabors
disappeared a question mark replaced the fixation point,
signalling participants to report (see below) the direction
of movement of the Gabor that they judged to have moved
faster. The next trial began 1000 ms after the participants’
response.
Participants performed a 2 · 2 alternative forced choice
task: if the left Gabor was perceived as being faster, they
responded with their left hand by pressing the ‘‘Q’’ key
(middle finger) or the ‘‘W’’ key (index finger) depending
on whether motion was leftward or rightward, respectively.
By contrast, if the right Gabor was perceived as being fast-
er, they responded with their right hand by pressing the
‘‘O’’ key (index finger) or the ‘‘P’’ key (middle finger)
depending on whether motion was leftward or rightward,
respectively. Hence, though we asked participants to report
the direction of motion, we were actually interested in their
perception of motion speed. As pointed out by Carrasco
et al. (2004), this method has the advantage of reducing
any possible response bias in the dimension of interest
(here speed of motion).
2.1.4. Design
A 3 · 7 factorial design was used, with cue condition
and test speed as factors. The cue condition had three levels
(test, control, standard), whilst the test speed condition
Fig. 1. Example of the events occurring on each trial in the attentional
(peripheral cue) and control (cue at fixation) conditions, respectively. The
two horizontal white arrows (which were not present in the display)
indicate the direction of motion of the two Gabors. In this example both
Gabors moved outward but in opposite directions. The stimuli are not
drawn to scale.
168 M. Turatto et al. / Vision Research 47 (2007) 166–178

had seven levels (see above). In total each participant
performed 672 experimental trials (32 trials per condition)
divided in four blocks of 182 trials each. In addition, before
the first block began observers performed 40 practice trials;
10 warm-up trials were completed before each new block.
Practice and warm-up trials were not included in the
data analysis. The experiment took about 75 min to be
completed.
2.2. Results
We calculated the proportion of trials in which the test
Gabor appeared to move faster than the standard Gabor
as a function of whether attention was cued to the test,
the standard, or remained at fixation (control condition).
Results depicted in Fig. 2a (where each curve was generated
by a Gaussian fit of the corresponding data. The horizontal
line intersecting the curves indicates the Point of Subjective
Equivalence, PSE) show that visual attention increased the
perceived speed of motion over a wide range of speeds. This
effect was confirmed by a statistical analysis of variance
(ANOVA) with cue and test speed as factors. The factor test
speed was significant F(6,102) = 256.693, p .0001, indicat-
ing that the participants’ ability to appreciate a difference
in speed between the test and the standard was maximal
when the test was clearly slower or faster as compared
to the standard, while it was minimal when the two
Gabors had approximately the same velocity. Crucially,
we also observed a significant main effect of cue condition
F(2,34) = 15.233, p .0001, indicating a change in partici-
pants’ performance as a function of attention. Accordingly,
as compared to the control condition (PSE = 4.35/s,
SD = 1.65), the test-cued curve shifted to the left
(PSE = 3.85/s, SD = 1.81), whilst the standard-cued curve
shifted to the right (PSE = 4.84/s, SD = 1.67). In other
words, on the one hand when attention was allocated to
the test, it was judged to have the same speed as the stan-
dard even though it actually moved at a slower speed. On
the other hand, when attention was directed to the stan-
dard, the test needed to move at a higher speed to be seen
as moving at the same speed as the standard. In both cases
the attended Gabor was perceived as moving faster than the
less attended one. However, as one may note the difference
in speed produced by attention was, on average, 0.50/s, a
value that was smaller than the average just noticeable dif-
ference (JND) in the control condition (JND = 1.1/s, cal-
culated as half the difference between the speeds at the
25% and 75% cutoffs). These results may suggest that
although participants consistently and reliably reported
the attended Gabor as moving faster, it is likely that they
were not able to consciously appreciate the difference in
speed produced by attention. In Experiments 5a and 5b
we will address this issue directly.
The interaction between the two factors was also signif-
icant F(12,204) = 5.703, p .0001, which indicates that the
effect of attention on the perceived speed of motion tended
to be reduced for the extreme values of motion speeds. This
suggests that under the present experimental conditions
(i.e., levels of speed) the increment in the perceived speed
of the attended Gabor was not large enough for observers
to perceive the test Gabor as moving faster than the stan-
dard Gabor when the difference in the actual speed between
the two stimuli was very large (e.g., standard = 4.29/s vs.
test = 15/s). Finally, the same main effects and the interac-
tion were significant even when test vs. control and stan-
dard vs. control conditions were considered separately
(all ps .005).
To sum up, since participants judged the attended mov-
ing stimulus to move faster than the unattended moving
stimulus, the results of Experiment 1 indicate that attention
can alter the perception of visual stimuli, not only by
Fig. 2. (a) Results of Experiment 1 (53-ms ISI). Proportion of responses in which observers reported the speed of the test Gabor as faster than the
standard Gabor, plotted as a function of the test Gabor’s physical speed for the three cueing conditions. The standard Gabor moved at a fixed speed:
4.29/s. The three downward-pointing arrows indicate the test speed necessary for both stimuli to be perceived as moving at the same speed (PSE). When
either the test or the standard were cued, the PSE shifted to the left or right, respectively, indicating that the attended stimulus was judged as moving
faster. ( b) Results of Experiment 2 (500-ms ISI). When attention was no longer at the cued location because of the long cue-Gabor interval, the three plots
and the corresponding PSEs coincided. The curves in (a) and (b) are Gaussian fits of the data.

increasing their perceived contrast (Carrasco et al., 2004),
spatial frequency or gap size (Gobell Carrasco, 2005),
but also by altering the perceived speed of motion.
3. Experiment 2
The pattern of results from Experiment 1 points to a
modulation of the perceived speed of motion by attention.
In order to confirm that this result was a genuine attention-
al modulation of perceived motion speed we needed to rule
out an alternative explanation in terms of a decision bias
toward the stimulus sharing the same position as the cue.
In other words, it may be that when observers were uncer-
tain as to which was the faster Gabor, they could have
adopted a decision bias that tended to favor the selection
of the Gabor sharing the same spatial location as the cue.
To exclude this possibility, in the present experiment we
extended the inter stimulus interval (ISI) between the cue
and the Gabors to 500 ms. This manipulation was motivat-
ed by the fact that it is well established that the exogenous
orienting of attention is a very short-lived phenomenon
vanishing within 250 ms from cue onset (Müller Rabbitt,
1989; Nakayama Mackeben, 1989). So, if the results
observed in Experiment 1 were due to attention and not
to a decisional bias, by extending the ISI up to 500 ms
the prediction is that the effect of the cue on motion percep-
tion should vanish. On the contrary, if the results of Exper-
iment 1 were generated by a decision bias, the same pattern
of results are expected also with an ISI of 500 ms (also see,
Carrasco et al., 2004).
3.1. Method
3.1.1. Participants
Ten undergraduate students from the University of
written consent to participate in the study. None had taken
part in Experiment 1.
3.1.2. Apparatus, stimuli, design, and procedure
Everything was the same as in Experiment 1, with the
exception that the ISI was increased up to 500 ms.
3.2. Results
Fig. 2b depicts the participants’ performance in the test-
cued, standard-cued and control conditions (each curve
was generated by a Gaussian fit of the corresponding data)
Results differed substantially from those in Fig. 2a (53-ms
ISI), and showed that with a 500-ms ISI, participants’
responses no longer varied as a function of the cue condi-
tion, and that any effect of the cue disappeared. This was
confirmed by a further ANOVA which revealed only a sig-
nificant main effect of test speed F(6,54) = 236.418,
p .0001. However, the crucial main effect of cue, as well
as the interaction with test speed were not significant (with
the lowest p = .371). The lack of any cue effect on visual
(motion) perception with an ISI of 500 ms is what one
would have expected if attention, and not a decision bias,
accounted for the results of Experiment 1. Indeed, previous
studies on the temporal dynamics of attentional capture
have clearly documented that sudden visual onsets, such
as that used here, summon attention to the corresponding
location for no longer than 200 ms. After such an interval
the facilitatory effects of attentional capture decay very
rapidly (Jonides, 1981; Nakayama Mackeben, 1989).
In addition, the present findings are in agreement with
what was reported by Carrasco et al. (2004) for the percep-
tion of contrast when an ISI of 500 ms was used.
4. Experiment 3
The results of Experiment 2 suggested that the decision-
bias account can probably be dismissed. However, one may
note that in Experiment 2 we changed the ISI between the
cue and the Gabors, and that this could have affected,
beside attentional allocation, other low-level visual interac-
tions between the cue and the Gabor sharing the same spa-
tial location. As already proposed by Carrasco et al. (2004),
to circumvent this problem in the present experiment we
kept the same 53-ms ISI used in Experiment 1, but instead
of asking participants to report the direction of motion of
the faster Gabor, we required them to report the direction
of motion of the slower Gabor. In these conditions, any
low-level visual interactions between the stimuli in the dis-
play are preserved as in Experiment 1, and crucially, the
attentional- and decision-bias accounts make opposite
predictions. If the results of Experiment 1 were merely
due to a bias in selecting the cued Gabor, then we should
expect participants to select the cued Gabor also when
the task requires them to indicate the slower instead of
the faster one. By contrast, if attention is responsible for
the effect reported in Experiment 1, then one should expect
participants to perceive the uncued Gabor as the slower
stimulus.
4.1. Method
4.1.1. Participants
Twelve undergraduate students from the University of
written consent to participate in the study. None had taken
part in Experiments 1 or 2.
4.1.2. Apparatus, stimuli, design, and procedure
Everything was the same as Experiment 1, with the fol-
lowing exceptions: First, participants had to report the
direction of motion of the slower Gabor; second, in order

to reduce the overall duration of the experiment we used
only the control and standard-cued condition.
4.2. Results
Gabor appeared to move slower than the standard Gabor,
as a function of whether attention was cued to either the
standard or remained at fixation (control condition).
Results depicted in Fig. 3a (where each curve was generated
by a Gaussian fit of the corresponding data) confirmed that
visual attention increased the perceived speed of motion
when participants had to report the direction of motion of
the slower Gabor. An ANOVA with cue and test speed as
factors revealed that the factor test speed F(6,66) =
66.137, p .0001, and crucially the factor cue F(1,11) =
10.868, p .007, were significant. This latter result
documented a change in participants’ performance as a
function of attention that was consistent with the result of
Experiment 1, in which the task required to judge the
direction of motion of the faster Gabor. Accordingly, as
compared to the control condition (PSE = 4.25/s;
SD = 2.91), the curve in the standard-cued condition shift-
ed to the right (PSE = 4.72/s; SD = 2.77), indicating that
when the standard was cued, the test was more often select-
ed as the slower stimulus. The interaction between the two
factors only approached significance F(6,66) = 2.105,
p = .064.
On the basis of the results of the present experiment we
can safely dismiss the decision-bias account. Indeed, if a
decision bias, namely a tendency to select the cued Gabor,
was responsible for the findings of Experiment 1, then the
same bias in favor of the cued Gabor should have emerged
in the present experiment as well; this was clearly not the
case.
5. Experiment 4
Although the findings from Experiments 2 and 3
allowed us to rule out a decision bias as an explanation
of the results of Experiment 1, one additional alternative
interpretation must be dismissed before we can conclude
that visual attention modifies the perceived speed of mov-
ing objects.
Recently it has been shown that visual attention alters
the perceived contrast of the stimulus, so that the attended
stimulus of a pair is judged as being higher in contrast than
the less attended one (Carrasco et al., 2004). Furthermore,
previous psychophysical studies have established that the
perceived speed of moving gratings is affected by contrast,
so that gratings at higher contrast are perceived as moving
faster than gratings at lower contrast (e.g., Thompson,
1982; Stone Thompson, 1992; but see McKee, Silverman,
Nakayama, 1986). Hence, one might reasonably argue
that what we documented in Experiment 1 was not a direct
modulation of perceived motion speed by visual attention,
but rather the consequence of increasing the perceived con-
trast of the attended Gabor, which, in turn, caused the per-
ceived increment in motion speed. Therefore, at this point
we are left with the following two possibilities: either atten-
tion directly alters the perceived speed of moving gratings,
or alternatively, such an effect is due to a change in the per-
ceived contrast of the attended moving Gabor. Distinguish-
ing between these two accounts may be relevant to
understanding at what stage of information processing,
and possibly in which neural site(s) in the visual system
Fig. 3. (a) Results of Experiment 3. Proportion of responses in which observers reported the speed of the test Gabor as slower than the standard Gabor,
plotted as a function of the test Gabor’s physical speed for the two cueing conditions. The standard Gabor moved at a fixed speed: 4.29/s. The two
downward-pointing arrows indicate the PSE in the control and standard conditions. As compared to the PSE in the control condition, when the standard
was cued, the PSE shifted to the right, indicating that the attended stimulus was less likely to be judged as moving slower, or put differently that the
unattended stimulus (the test in the present experiment) was more likely to be perceived as moving slower. (b) Results of Experiment 4. Proportion of
responses in which observers reported the contrast of the test moving Gabor as higher than the standard moving Gabor, plotted as a function of the test
Gabor’s physical contrast for the two cueing conditions. The two Gabors moved at a fixed speed (4.29/s), while the standard Gabor had a fixed contrast
(60%). The two downward-pointing arrows indicate the test contrast necessary for both stimuli to be perceived as having the same contrast (PSE). As
compared to the PSE in the control condition, when the test was cued, the PSE shifted to the right, indicating that the attended stimulus was less likely to
be judged as higher in contrast. The curves in (a) and (b) are Gaussian fits of the data.

attention can affect the perception of motion speed. One
should note, however, that the view according to which
the boost up in perceived motion speed is the consequence
of a change in the perceived contrast of the attended Gabor
rests on the assumption that the contrast of moving grat-
ings, as that of static ones (Carrasco et al., 2004), is
increased by visual attention. Since, to the best of our
knowledge, there is no empirical evidence supporting this
(plausible) assumption, we devised an experimental proce-
dure aimed at directly addressing whether the contrast of
moving sinusoidal gratings is altered by visual attention.
If this were the case, our next step would be to distinguish
whether the increment in perceived motion speed was total-
ly attributable to a change in contrast, or whether attention
can, at least partially, alter the perception of motion speed
directly. On the other hand, however, if the contrast of a
moving Gabor is not increased by visual attention, then
there is no point in assuming that the boost in the percep-
tion of motion speed we reported was caused by a change
in the perceived contrast of the attended moving Gabor.
5.1. Method
5.1.1. Participants
Six undergraduate students from the University of Tren-
to served as participants. They were compensated either
with course credits or 8€ for their participation. All were
naı̈ve as to the purpose of the experiment, reported normal
or corrected-to normal vision, and gave informed written
consent to participate in the study. None had taken part
in previous experiments.
5.1.2. Apparatus, stimuli, and procedure
The same paradigm used in Experiment 1 was adapted
to study whether attention alters the apparent contrast of
moving Gabors. In other words, we used the basic para-
digm proposed by Carrasco et al. (2004), except that our
stimuli were moving Gabors (4.29/s) instead of stationary.
We used only the control and test-cued condition, so as to
maximize the number of data points per cell in the design,
while keeping the overall duration of the experiment below
1 h. We kept the contrast of the standard Gabor fixed
(60%), and varied that of the test Gabor from trial to trial
using seven levels of contrast (48%, 52%, 56%, 60%, 62%,
68%, and 72%). The contrast levels were preliminarily
determined in a pilot study as to create a sigmoid-like func-
tion in the control condition. Participants were instructed
to report the direction of movement (left vs. right) of the
Gabor that looked higher in contrast.
5.1.3. Design
A 2 · 7 factorial design with cue and test contrast as fac-
tors. The factor cue had two levels (test and control), whilst
the factor test contrast had seven levels (see above). In total
each participant performed 560 experimental trials (40 tri-
als per condition) divided into five blocks of 112 trials each.
In addition, before the first block began observers per-
formed 20 practice trials, and five warm-up trials were
completed before each new block. Practice and warm-up
trials were not included in data analysis.
5.2. Results
Gabor appeared to be higher in contrast than the standard
Gabor, as a function of whether attention was cued to the
test or remained at fixation (control condition). Results,
depicted in Fig. 3b (where each curve was generated by a
Gaussian fit of the corresponding data), showed that the
cued Gabor was not seen as being higher in contrast as
compared to the uncued one. Instead, since the curve in
the test cued condition shifted to the right of the curve in
the control condition, Fig. 3b suggests that, quite surpris-
ingly, the attended Gabor was perceived as lower in con-
trast. This impression was confirmed by an ANOVA in
which both the main effect of test contrast F(6,30) =
290.000, p .0001, and cue F(1,5) = 10.224, p .024 were
significant. In other words, participants judged the attend-
ed moving Gabor as being lower in contrast (control
condition, PSE = 60.63, SD = 8.3; test cued condition,
PSE = 63.33, SD = 8.6), a result that, as discussed in detail
later, is the opposite of what Carrasco et al. (2004) found
when the Gabors were stationary. The interaction between
the two factors only approached significance F(6,30) =
2.279, p = .062, suggesting that the effect of attention in
biasing the participants’ response was limited or absent
when the physical contrast difference between the cued
and uncued Gabor was very large.
The present results definitely undermine the possibility
that the boost in perceive speed observed in Experiment 1
can be attributed to an increment in the perceived contrast
of the attended moving Gabor. Indeed, this view would
have predicted a pattern of results opposite to that which
emerged in the present experiment; namely observers
should have reported the attended Gabor as being higher
in contrast. Therefore, we can safely conclude that attention
can directly alter the perceived speed of the selected stimu-
lus. The fact that participants reliably selected the unattend-
ed Gabor as the highest in contrast could be due to the fact
that attention, by increasing the perceived speed of the cued
Gabor, increased also the blurring of the corresponding
image, which in turns might have reduced its contrast.
6. Experiment 5a
The experiments presented so far provide converging
evidence that attention increases perceived motion speed.
This finding is in agreement with those of recent studies
showing that, when participants are asked to report which
stimulus of a pair is higher in contrast, or spatial frequency,
observers exhibited a significant bias in favor of the cued
stimulus as compared to the uncued one (Carrasco et al.,
2004; Gobell Carrasco, 2005). Crucially, in these studies,
and particularly in the study by Carrasco et al. (2004), it

has been claimed that this result can be accounted for by a
change in the phenomenological appearance of the attend-
ed stimulus, so that participants actually saw the attended
Gabor as being, for example, higher in contrast. Therefore,
one may wonder whether the effect of attention on motion
speed perception was also due to a change in the phenom-
enological appearance of the stimuli. Alternatively, the dif-
ference may be due only to a change in visual performance.
Put differently, did participants consciously and genuinely
see the attended Gabor as moving faster than the less
attended one, or was their performance based on implicit
speed perception? The fact that in Experiment 1 the
estimated effect of attention was smaller than the JND is
consistent with the latter possibility. However, to directly
address this issue participants were presented with two
Gabors moving either at the same or different speeds,
and were asked to report whether they noted a speed differ-
ence between the two Gabors. So, instead of forcing partic-
ipants to report which was the faster Gabor of a pair, as we
did in previous experiments, here we asked participants to
respond on the basis of their phenomenological awareness,
if any, of a difference in speed between the two stimuli.
6.1. Method
6.1.1. Participants
Four undergraduate students from the University of
Trento, plus two of the authors (M. T. and M. Ve.) served
as participants. The four students were compensated either
with course credits or 8€ for their participation, and were
naı̈ve as to the purpose of the experiment. All participants
reported normal or corrected-to normal vision, and gave
informed written consent to participate in the study. None
had taken part in previous experiments.
The paradigm used in Experiment 1 was modified as
follows. First, no cue was presented; second, in the same-
speed condition both Gabors moved at 4.29/s (the
standard speed used in previous experiments), whereas in
the different-speed condition one Gabor moved at 4.29/
s, whilst the other moved at one of four different speeds
(two slower and two faster than the standard speed):
3.75, 3.96, 4.70, or 5.02/s. The slowest (3.75/s) and the
fastest (5.02/s) speeds were already used in Experiment 1
as those closest to the standard speed. The remaining two
intermediate speeds (3.96 and 4.70/s) were chosen as the
best matching PSE-speed in the test-cued (3.85/s) and
standard-cued (4.85/s) condition, as observed in Experi-
ment 1. In other words, as compared to the control condi-
tion, they approximately corresponded to the increment
and decrement of the standard speed caused by visual
attention.
Participants were instructed to press the ‘‘B’’ key on the
computer keyboard if they saw a difference in speed
between the two Gabors, and the ‘‘N’’ key if they saw
the two Gabors moving at the same speed.
6.1.3. Design
Participants were submitted to 128 trials in the different-
speed condition (32 trials for each speed) and to 128 trial in
the same-speed condition, divided into two blocks. In addi-
tion, before the first block of trials observers performed 20
practice trials, which were not included in the data analysis.
6.2. Results
Overall participants reported that it was very difficult, if
not impossible, to consciously appreciate any speed differ-
ence between the two Gabors. However, to obtain an
objective measure of the participants’ capacity to detect
speed differences, data were analyzed using a measure of
signal sensitivity (d0
; Green Swets, 1974). The proportion
of trials in which participants correctly detected a speed
difference between the Gabors (Hits), and the proportion
of trials in which they reported a difference while the two
Gabors had the same speed (False alarms) were used to cal-
culate d0
and b for each of the four speed levels. In order to
gain statistical power, before comparing d0
mean scores
with 0 (lack of sensitivity) and b mean scores with 1 (no
bias in the response criterion), we averaged together data
from the two speeds (3.96 and 4.70/s) corresponding to
the PSE values into ‘‘speed-difference 1’’ condition, and
data from the slowest and fastest speeds (3.75 and 5.02/
s) into ‘‘speed-difference 2’’ condition. Results showed that
in both conditions d0
was not significantly different from 0
(speed-difference 1, d0
= 0.07, p = .184; speed-difference 2,
d0
= 0.13, p = .128); b was not significantly different from
1 in both conditions.
Although the results of Experiment 1 documented a reli-
able boost up of the speed of motion of the attended
Gabor, on the basis of the findings of the present experi-
ment it seems hard to conclude that attention altered the
phenomenological appearance of the speed of the cued
Gabor. However, the participants’ inability to consciously
perceive a speed difference in the present experiment is not
inconsistent with the fact that, for the same speed differ-
ence, they correctly selected the faster Gabor when forced
to do so in Experiment 1. These apparent discrepant out-
comes can probably be reconciled by noticing that the pro-
cedure adopted in Experiment 1 always forced participants
to guess which was the faster Gabor, and this even when in
the control condition the difference in speed between the
standard and the test was minimal (as in speed-difference
2 condition of the present experiment), or when one of
the two Gabors having similar speeds was cued. In these
conditions it is likely that participants were not aware of
having seen a difference in speed between the two Gabors.
Therefore, what we likely measured when we cued a Gabor
was the effect of attention on visual performance regardless
of any visual awareness. By contrast, in the present exper-
iment by asking participants to report whether they noted
any difference in speed between the two Gabors, we directly
addressed their phenomenological awareness of a difference
in speed. Put differently, it is likely that while Experiment 1

relied more on perception without awareness, Experiment 5a
was more concerned with conscious perception. The dissoci-
ation between perception with and without awareness is
well documented in cognitive psychology, and as discussed
later there are several cases in which human beings exhibit
visual performance above chance in the absence of aware-
ness (Merikle, Smilek, Eastwood, 2001).
In the next experiment (5b) we tried to provide further
support for this hypothesis. To this aim we presented par-
ticipants with the same speed values used in Experiment 5a,
but instead of asking participants to report any difference
in speed between the two Gabors, we asked them to decide
which one was faster. Our prediction was that, regardless
of any awareness of speed difference, under this condition
participants’ performance should vary from chance, or, in
other words, their sensitivity (d0
) to difference in speed
should be significantly higher than zero.
7. Experiment 5b
7.1. Method
7.1.1. Participants
Five undergraduate students from the University of
Trento, plus two of the authors (M. T. and M. Ve.) served
as participants. The five students were compensated either
with course credits or 8€ for their participation, and were
naı̈ve as to the purpose of the experiment. All participants
reported normal or corrected-to normal vision, and gave
informed written consent to participate in the study. With
the exception of the two authors none of the participants
had taken part in previous experiments.
As in Experiment 5a, except that participants were
instructed to report the direction of movement (left vs.
right) of the fastest Gabor (see Experiment 1 for details).
7.1.3. Design
Participants were submitted to 48 trials in each of the
five speed levels (3.75, 3.96, 4.29, 4.70, and 5.02/s), for a
total of 240 trials divided into 3 blocks of 80 trials each.
In addition, before the first block of trials participants per-
formed 20 practice trials, which were not included in the
data analysis.
7.2. Results
When informally probed at the end of the experiment,
participants reported that they had guessed which Gabor
moved faster, as it was very hard if not impossible to con-
sciously perceive a speed difference between the two
Gabors. We calculated d0
(using the formula for the
2AFC discrimination tasks, d0
¼ ½zðHitÞ zðFAÞ=
ffiffiffi
2
p
) for
each of the four speed levels in which the two Gabors
had different speeds (3.75, 3.96, 4.70, and 5.02/s). To this
aim we treated the condition in which the faster Gabor was
on the left as ‘‘signal’’, and that in which the faster Gabor
was on the right as ‘‘noise’’: Hits were considered those tri-
als in which participants correctly reported the left Gabor
as being faster; False alarms were those trials in which par-
ticipants selected the left Gabor when in fact the faster was
on the right. As in the previous experiment, to gain statis-
tical power we averaged together data from the two speeds
(3.96 and 4.70/s) corresponding to the PSE values into the
‘‘speed-difference 1’’ condition, and data from the slowest
and fastest speeds (3.75 and 5.02/s) into the ‘‘speed-
difference 2’’ condition. d0
values were 0.45 and 0.50 in
the speed-difference 1 and 2 conditions respectively, and
in both cases d0
was significantly different from 0 (all
ps .011).
Hence, despite participants claiming that they were not
aware of having seen any speed difference, as expected,
when they were forced to decide which Gabor moved faster
their performance was not at chance, as d0
results showed
that the visual system implicitly recognized the faster
Gabor. Note that this finding was obtained for the same
speed values for which, when required to report whether
or not a difference in speed was present (Experiment 5a),
participants’ sensitivity was not different from zero.
8. General discussion
After more than a century of research on visual atten-
tion it is well established that attention affects different
aspects of visual analysis. However, to date the scientific
enterprise committed to the understanding of attentional
phenomenon has mainly focused on how attention affects
visual performance (e.g., in terms of speed or accuracy)
rather than addressing whether the phenomenological
appearance of visual objects can be shaped by attention.
Indeed, the effects of attention on visual performance
have been documented on response times (e.g., Posner,
1980), accuracy (e.g., Bashinski Bacharach, 1980; Lyon,
1990), spatial resolution (e.g., Yeshurun Carrasco, 1999),
rate of information accrual (Carrasco McElree, 2001),
and contrast sensitivity (e.g., Cameron et al., 2002; Lu
Dosher, 1998; Treue, 2004). Although there is no doubt
that attention improves the efficiency of visual analysis
(but see Yeshurun Carrasco, 1998), it remains an open
issue as to whether attention also alters the phenomenolog-
ical appearance of visual stimuli. In other words, does an
attended stimulus look different compared to a less
attended one?
This intriguing question was considered, more than a
century ago, by famous pioneers of modern experimental
psychology like Wundt, James and Ebbingaus, who
believed and suggested that attended stimuli should be
perceived more vividly compared to less attended stimuli
(e.g., Ebbinghaus, 1908; James, 1980). In the last decade,
this important issue has been tackled by Prinzmetal and
colleagues (Prinzmetal et al., 1997), whose work, however,
seemed to provide evidence that attention does not alter the
phenomenological appearance of stimuli. Yet, as noted by

Carrasco et al. (2004) some methodological aspects of the
experimental procedure used in the Prinzmetal et al.
(1997) study may have precluded the possibility of demon-
strating the effect of attention on stimuli’s appearance. Per-
haps the most important concerns associated with
Prinzmetal et al.’s study regard the lack of a precise control
of the allocation of attention in space, the unclear contribu-
tion of eye movements, and an unlimited response time. To
assess the role of attention on the appearance of the stimuli
by overcoming these potential methodological flaws, Carr-
asco et al. (2004) devised a different procedure, in which
spatial attention was cued to a given location via a periph-
eral visual onset prior to the targets’ occurrence. The inter-
val of time between the cue and the targets (70 ms) was
chosen so as to magnify the effect of transient-automatic
attention (Jonides, 1981; Nakayama Mackeben, 1989),
and to preclude, given the brief ISI (50 ms) and target
presentation time (40 ms), any effect of eye movements.
In addition, the procedure proposed by Carrasco and col-
leagues required a direct online comparison between the
perceived contrast of two concurrently presented targets,
thus directly evaluating a possible phenomenological differ-
ence in appearance between the attended and less-attended
target stimuli. In these experimental conditions Carrasco
et al. (2004) showed that attention altered the perceived
contrast of the cued Gabor, which was judged to be higher
in contrast. In a following study (Gobell Carrasco, 2005)
the same procedure that proved to be successful in showing
that attention alters the perceived contrast was adapted to
demonstrate that other static features, such as spatial fre-
quency and gap size, can also be altered by visual attention.
Specifically, the authors provided evidence that transient
attention increased the apparent spatial frequency of the
selected Gabor stimulus, and increased apparent gap size
in a Landolt-square acuity task.
These findings, which testified that attention can change
the appearance of static visual stimuli, encouraged us to
address whether attention can also modify the perceived
speed of moving stimuli. Indeed, besides contrast, spatial
frequency and gap size, motion is another important visual
property of many objects or events that we usually encoun-
ter in our everyday visual experience.
With regard to possible interactions between attention
and visual motion perception, neuroimaging studies have
already documented a role of attention in shaping the neu-
ral responses in motion-sensitive cortical areas (e.g., Marti-
nez-Trujillo Treue, 2002; Treue Martinez-Trujillo,
1999; Treue Maunsell, 1996). For example, it has been
shown that the neural activity in V5/MT associated with
the MAE produced by exposure to distracting moving dots
in the background was strongly attenuated, if not com-
pletely abolished, when attention was diverted from
motion processing by engaging participants’ attention in
a highly demanding linguistic task at the center of the
screen (Rees et al., 1997). In addition, psychophysical stud-
ies have shown that attending to a pattern of moving stim-
uli increases the duration (Chaudhuri, 1990) and direction
(Alais Blake, 1999; Lankheet Verstraten, 1995) of the
corresponding MAE, and enhances the effect of adaptation
to motion processing (Rezec et al., 2004).
However, since no previous study has addressed whether
attention can directly alter the perception of motion speed,
we decided to investigate this issue directly. To this pur-
pose, we used the basic paradigm proposed by Carrasco
et al. (2004), which was adapted to study motion speed per-
ception. Instead of using static Gabors, we presented two
moving Gabors, where motion was achieved by indepen-
dently shifting the two sinusoidal gratings leftward or
rightward. The impact of attention on motion speed per-
ception, if any, was assessed by cueing one Gabor and by
asking participants the direction of motion (left vs. right)
of the faster Gabor. The result of Experiment 1 showed
that when the test and the standard Gabors moved at the
same physical speed, participants overestimated the motion
speed of the attended Gabor by approximately 10%.
Indeed, as compared to a PSE of 4.35/s in the control con-
dition, when the test was attended, the corresponding PSE
decreased to 3.85/s (see Fig. 2a). By contrast, when the
standard was cued the PSE of the test increased to 4.84/
s. So, in line with previous findings on perceived contrast,
these results support the hypothesis that attention
increased the perceived speed of a moving stimulus.
We then devised two control experiments to rule out the
possibility that such findings could be accounted for by a
decision bias rather than a change in perception (also see
Carrasco et al., 2004). Specifically, Experiment 2 showed
that the effect disappeared when the ISI between the cue
and the Gabors was increased to 500 ms (see Fig. 2b), a result
consistent with the temporal dynamic of transient attention
(Nakayama Mackeben, 1989). In addition, and most cru-
cially, when participants were asked to indicate the direction
of motion of the slower Gabor (Experiment 3), they showed
a significant perceptual bias in favor of the uncued Gabor
(see Fig. 3a). This result was the opposite than that predicted
by the decision-bias hypothesis, which would have predicted
that the cued Gabor always has higher chances of being
selected regardless of whether observers were required to
report the faster or the slower Gabor. However, this result
was not sufficient to completely rule out the decision-bias
account. One might argue, for instance, that participants
were more prone to select a stimulus as the one that has more
rather than less of a given feature (here speed) when it coin-
cides with attention. By contrast, they might have had the
tendency to associate the stimulus that had less rather than
more of a feature with the uncued location. This, in principle,
could explain why, in Experiment 3, participants tended to
report the Gabor at the uncued location as slower, i.e., the
one that was ‘‘not faster’’ or that possessed less speed. How-
ever, this further alternative decision-bias interpretation was
ruled out by the results of Experiment 4.
The main purpose of Experiment 4 was to verify
whether the increment in perceived motion speed for the
cued Gabor reported in Experiment 1 was a direct and
genuine effect of attention on speed perception. Indeed,

because stimuli at higher contrast are seen as moving faster
(e.g., Stone Thompson, 1992; Thompson, 1982; Thomp-
son, Stone, Swash, 1996), and given that attention
increases perceived contrast (Carrasco et al., 2004), one
could argue that the increment in motion speed we report-
ed might be a secondary effect of attention boosting con-
trast, and therefore the perceived speed. Even though this
argument might be plausible, it rests on the assumption
that attention, besides increasing the perceived contrast
of a static Gabor, would also increase the perceived con-
trast of a moving Gabor. Hence, our goal was to directly
verify whether this was the case. To this aim, in Experiment
4 we implemented essentially the same procedure used to
study perceived contrast with static Gabors, except that
our stimuli were two Gabors drifting at the same constant
speed. We asked participants to report which Gabor
looked higher in contrast, and quite surprisingly we found
that the attended Gabor was judged to appear lower in
contrast. This result has at least two important implica-
tions. First, it excludes a version of the decision-bias
account that assumes a tendency to select the Gabor at
the uncued location if it can be associated with a negative
dimension. Indeed, if on the one hand in Experiment 3
observers judged the uncued Gabor as the one that was
slower (i.e., not faster), on the other hand in Experiment
4 they still selected the uncued Gabor when asked to report
the one that was higher in contrast. Second, the data from
Experiment 4 provide evidence that attention reduces the
perceived contrast of a moving Gabor, a result that is
opposite to what has been reported by Carrasco et al.
(2004) with static Gabors. Why would attention decrease
the perceived contrast of a moving Gabor? Although we
do not have a straightforward explanation, one could spec-
ulate that the increased perceived speed of the cued Gabor
might have caused a larger degree of blurring in the corre-
sponding image, thus likely reducing its contrast (see how-
ever, Burr Morgan, 1997). However, while further
experimental work should be aimed at addressing this issue
more directly, this finding rules out the possibility that the
increase in speed reported in Experiment 1 was a secondary
effect of a change in apparent contrast.
The perceived speed of a moving sinusoidal grating is
affected by its contrast and spatial frequency. However,
while increasing contrast of a grating increases its perceived
speed (Stone Thompson, 1992), increasing spatial fre-
quency decreases the perceived speed (Smith Edgar,
1990; also see Priebe Lisberger, 2004). Interestingly,
Gobell and Carrasco (2005) have recently shown that
attention increases the perceived spatial frequency of static
Gabors. Hence, if in principle attention might increase the
contrast of the cued Gabor, which in turn may result mov-
ing at higher speed (but see the results of Experiment 4),
the results of Gobell and Carrasco (2005) would predict
that in Experiment 1 the attended Gabor should have been
perceived at higher spatial frequency, which should have
caused a reduction of the corresponding perceived speed.
The results of Experiment 1 clearly showed that this was
not the case; those data suggest that speed perception is
not secondary to effect of attention on the spatial frequency
of the moving Gabor, probably because it may be difficult
to estimate (consciously or unconsciously) small changes in
the spatial frequency of a moving grating.
So, the influence of static stimulus properties (contrast
and spatial frequency) on motion speed perception, and
the interactions between these factors and attention can
be summarized as follows: (a) attention increases perceived
contrast (Carrasco et al., 2004); (b) increasing contrast
increases perceived speed (Stone Thompson, 1992); (c)
attention increases perceived spatial frequency (Gobell
Carrasco, 2005); (d) increasing spatial frequency decreases
perceived speed (Smith Edgar, 1990); but attention
increases perceived speed (Experiment 1), and attention
decreases perceived contrast of moving stimuli (Experiment
4). Since with moving Gabors the attended one was judged
as being lower in contrast, and given that the perceived
speed is positively correlated with the degree of contrast
(at least with static Gabors), one may hypothesize that par-
ticipants might have underestimated the perceived speed of
the attended Gabor, thus reducing, if any, the observed
effect of attention on motion speed perception reported in
Experiment 1. In any case, this result suggests that the
modulation of speed we documented was a direct effect
of attention on speed perception.
A final issue that emerged from the results of the present
study regards the change in the phenomenological appear-
ance of the attended stimuli. In the case of the effect of atten-
tion on perceived contrast, Carrasco et al. (2004) stressed
the fact that their findings demonstrated that attention effec-
tively alters the way participants consciously perceive the
selected stimulus, namely that attention changes the phe-
nomenological appearance of the percept. As a straightfor-
ward example of how attention would change the way we
‘‘see’’ the stimuli, in their study the authors depicted (see
Fig. 4 of the Carrasco et al. study) three Gabors at different
levels of contrast, in which the visible difference between the
lowest (e.g., 16%) and the highest (e.g., 28%) levels corre-
sponds to the experimentally measured change in contrast
as a function of attention. Does attention also alter the phe-
nomenological experience of motion speed? As we pointed
out, in Experiment 1 the difference in speed caused by atten-
tion was smaller than the JND, a result that seemed to indi-
cate that participants perceived the attended Gabor as
moving faster without actually being aware of such a differ-
ence in speed. To directly test whether the increase in per-
ceived speed caused by attention corresponded to a real
phenomenological change, or was only due to a change in
visual performance without awareness (Merikle et al.,
2001), we conducted two experiments in which participants
were asked to detect a difference in speed that was equal to
(or a bit larger than) the magnitude of the difference caused
by attention in Experiment 1. Specifically, in Experiment 5a
participants’ task was to decide whether the two Gabors
moved at the same or different speed, but they were not able
to detect any speed difference (d0
was not different from

zero). However, when in Experiment 5b participants were
forced to decide which Gabor moved faster, despite report-
ing that they denied having seen any difference in speed,
their visual performance was not at chance, as d0
was signif-
icantly different from zero. This pattern of results is consis-
tent with the possibility that the boost in the perceived speed
documented in Experiment 1 occurred despite a lack of
change in visual awareness. In other words, attention chan-
ged the performance of the visual system with regards to
speed perception without altering the phenomenological
appearance of the attended Gabor’s speed. This conclusion
is in agreement with the view according to which stimuli
(and their properties) can be perceived even when there is
no awareness of perceiving (Merikle et al., 2001). Also, the
present findings can be considered as an example of a disso-
ciation between visual performance and visual awareness.
There are several cases in which the human being exhibits
a dissociation (Merikle et al., 2001), but perhaps one of
the most striking example is provided by the phenomenon
of Blindsight, a neuropsychological disorder in which
robust (albeit limited) visual performances remain despite
a lack of awareness (e.g., Stoerig Cowey, 1997). Indeed,
despite a lesion to visual cortex, which causes ‘‘blindness’’
of the contralateral visual field, blindsight patients exhibit
good visual performance (i.e. above chance) when they have
to judge certain properties (e.g., direction of motion) of visu-
al events in their blind field. In a sense, according to the
results of Experiments 5a and 5b, participants might have
behaved liked blindsight subjects, as they denied being
aware of any difference in speed, but they guessed above
chance which Gabor moved faster when forced to do so.
In the same vein, perhaps they were not aware of the change
in motion speed caused by attention (Experiment 1).
Because our findings have shown that attention alters
the (implicit) perception of motion speed without changing
its phenomenological appearance, one may wonder why
attention would alter actual appearance in the case of per-
ceived contrast (Carrasco et al., 2004) and spatial frequen-
cy (Gobell Carrasco, 2005). To this aim we would like to
point out that although the picture depicted in the Carras-
co et al. study showing the effect of attention on contrast
seems to provide compelling evidence of a change in visual
appearance, there are reasons to suspect that the difference
participants might have experienced was not as large as is
apparent in the figure. First, one should note that appreci-
ating the difference in contrast between stimuli on a printed
figure without any time constraints might be quite different
from the situation in which the same stimuli are presented
peripherally on a computer screen for 40 ms. Second, if one
tries to coarsely estimate from the figures the JND in the
Carrasco et al. (2004) and Gobell and Carrasco (2005)
experiments, it seems to be larger than the differences
between the attention conditions.1
So, although the results
reported by Carrasco and her colleagues are of high inter-
est, perhaps a more straightforward test of any possible
change in the phenomenological appearance of the stimuli
would require an experiment in which participants are
directly asked to report if they see any difference between
the stimuli when the difference in contrast or spatial fre-
quency is equal to the estimated effect produced by atten-
tion (see Experiment 5a of the present study).
One may wonder why attention increases the perceived
speed of moving objects rather than slow it down. As far
as static visual properties are concerned, Gobell and Carr-
asco (2005) have proposed that attention increases, for
example, the sensitivity of those spatial filters involved in
the stimulus analysis, thus altering the perceived spatial fre-
quency of the attended object as compared to the less
attended one. Motion itself can be conceived as a structure
of events that change in space and time with a certain fre-
quency (Palmer, 1999), and indeed Adelson and Berger
(1985) have proposed the existence of spatial-temporal fre-
quency based filters for motion analysis. Interestingly, such
motion filters can be tuned to different speeds of motion, so
that one can hypothesize that attention might increase the
perceived speed of a moving stimulus by shifting the sensi-
tivity of the filters to a higher speed (a mechanism similar
to the one proposed by Gobell and Carrasco, 2005, for spa-
tial frequency).
In conclusion, the present study presents the first evi-
dence that attention directly modulates the perceived
motion speed of moving stimuli, and specifically, attention
makes the brain believe that attended moving stimuli move
faster.
Acknowledgments
The authors thank David Burr, two anonymous review-
ers and the members of the Cognitive Neuropsychology
Laboratory at Harvard for their comments on earlier drafts
of this paper. We also thank Christian Valt for his help in
data collection.
References
Adelson, E. H., Berger, J. K. (1985). Spatio-temporal energy models for
the perception of motion. Journal of the Optic Society of America, 2,
285–299.
Alais, D., Blake, R. (1999). Neural strength of visual attention gauged
by motion adaptation. Nature Neuroscience, 2, 1015–1018.
Bashinski, H. S., Bacharach, V. R. (1980). Enhancement of perceptual
sensitivity as the result of selectively attending to spatial locations.
Perception Psychophysics, 28, 241–248.
Burr, D. C., Morgan, J. M. (1997). Motion deblurring in human vision.
Proceeding of the Royal Society, London B, 264, 431–436.
Cameron, E. L., Tai, J. C., Carrasco, M. (2002). Covert attention affects
the psychometric function of contrast sensitivity. Vision Research, 42,
949–967.
Carrasco, M., Ling, S., Read, S. (2004). Attention alters appearance.
Nature Neuroscience, 7, 308–313.
Carrasco, M., McElree, B. (2001). Covert attention accelerates the rate
of visual information processing. Proceeding of the National Academy
of Science, USA, 98, 5363–5367.
1
We thank an anonymous reviewer for having drawn our attention to
this relevant issue about the results of Experiment 1 of the present study.

Chaudhuri, A. (1990). Modulation of the motion after-effect by selective
attention. Nature, 344, 60–62.
Ebbinghaus, H. (1908). Psychology: An elementary text-book. Boston:
Heath.
Festinger, L., Coren, S., Rivers, G. (1970). The effect of attention
components in perception of size-at-a-distance. Perception Psycho-
physics, 83, 189–207.
Gobell, J., Carrasco, M. (2005). Attention alters the appearance of
spatial frequency and gap effect. Psychological Science, 16, 644–651.
Green, D. M., Swets, J. A. (1974). Signal detection theory and
psychophysics. Huntington, NY: Krieger.
James, W. (1980). The principle of psychology. New York: Henry Holt.
Jonides, J. (1981). Voluntary versus automatic control over the mind’s
eye’s movement. In J. B. Long A. D. Baddeley (Eds.), Attention and
performance IX (pp. 187–203). Hillsdale, NJ: Erlbaum.
Lankheet, M. J., Verstraten, F. A. (1995). Attentional modulation of
adaptation to two-component transparent motion. Vision Research,
35, 1401–1412.
Lu, Z., Dosher, B. A. (1998). External noise distinguishes attention
mechanisms. Vision Research, 38, 1183–1198.
Lyon, D. R. (1990). Large and rapid improvement in form discrimination
accuracy following a location precue. Acta Psychologica, 73, 69–82.
Mack, A., Rock, I. (1998). Inattentional blindness. Cambridge, MA:
MIT Press.
Martinez-Trujillo, J. C., Treue, S. (2002). Attentional modulation
strength in cortical area MT depends on stimulus contrast. Neuron, 35,
365–370.
McKee, S. P., Silverman, G., Nakayama, K. (1986). Precise velocity
discrimination despite random variations in temporal frequency and
contrast. Vision Research, 26, 609–619.
Merikle, P. M., Smilek, D., Eastwood, J. D. (2001). Perception without
awareness: perspective from cognitive psychology. Cognition, 79,
115–134.
Müller, H. J., Rabbitt, P. M. A. (1989). Reflexive and voluntary
orienting of visual attention: Time course of activation and resistance
to interruption. Journal of Experimental Psychology: Human Percep-
tion Performance, 15, 315–330.
Nakayama, K., Mackeben, M. (1989). Sustained and transient
components of focal visual attention. Vision Research, 29, 1631–1646.
Palmer, S. E. (1999). Vision Science – photons to phenomenology.
Cambridge, MA: MIT Press.
Pelli, D. G. (1997). The Video Toolbox software for visual psychophysics:
Transforming numbers into movies. Spatial Vision, 10, 437–442.
Pestilli, F., Carrasco, M. (2005). Attention enhances contrast sensitivity
at cued and impairs it at uncued locations. Vision Research, 45,
1867–1875.
Posner, M. I. (1980). Orienting of attention. Quarterly Journal of
Experimental Psychology, Part A, 32, 3–25.
Priebe, N. J., Lisberger, S. (2004). Estimating target speed from
population response in visual area MT. The Journal of Neuroscience,
24, 1907–1916.
Prinzmetal, W., Nwachuku, I., Bodanski, L., Blumenfeld, L., Shimizu,
N. (1997). The phenomenology of attention: 2. Brightness and
contrast. Consciousness and Cognition, 6, 372–412.
Rees, G., Frith, C. D., Lavie, N. (1997). Modulating irrelevant motion
perception by varying attentional load in an unrelated task. Science,
278, 1616–1619.
Rezec, A., Krekelberg, B., Dobkins, K. R. (2004). Attention enhances
adaptability: Evidence from motion adaptation experiments. Vision
Research, 44, 3035–3044.
Shulman, G. L. (1991). Attentional modulation of mechanisms that
analyse rotation in depth. Journal of Experimental Psychology: Human
Perception Performance, 17, 726–737.
Shulman, G. L. (1993). Attentional effects of adaptation of rotary motion
in the plane. Perception, 22, 947–961.
Smith, A. T., Edgar, G. K. (1990). The influence of spatial frequency on
perceived temporal frequency and perceived speed. Vision Research,
30, 1467–1474.
Stoerig, P., Cowey, A. (1997). Blindsight in man and monkey. Brain,
120, 535–559.
Stone, L. S., Thompson, P. (1992). Human speed perception is contrast
dependent. Vision Research, 32, 1535–1549.
Thompson, P. (1982). Perceived rate of movement depends on contrast.
Vision Research, 22, 377–380.
Thompson, P., Stone, L. S., Swash, S. (1996). Speed estimates from
gratings patches are not contrast-normalized. Vision Research, 36,
667–674.
Treue, S. (2004). Perceptual enhancement of contrast by attention. Trends
in Cognitive Scieneces, 8, 435–437.
Treue, S., Martinez-Trujillo, J. C. (1999). Feature-based attention
influences motion processing gain in macaque visual cortex. Nature,
399, 575–579.
Treue, S., Maunsell, J. H. R. (1996). Attentional modulation of visual
motion processing in cortical areas MT and MST. Nature, 382,
539–541.
Tsal, Y., Shalev, L. (1996). Inattention magnifies perceived length: The
attention receptive field hypothesis. Journal of Experimental Psychol-
ogy: Human Perception and Performance, 22, 233–243.
Tsal, Y., Shalev, L., Zakay, D., Lubow, R. E. (1994). Attention reduces
perceived brightness contrast. Quarterly Journal of Experimental
Psychology, Part A, 47, 865–893.
Yeshurun, Y., Carrasco, M. (1998). Attention improves or impairs
visual performance by enhancing spatial resolution. Nature, 396,
72–75.
Yeshurun, Y., Carrasco, M. (1999). Spatial attention improves
performance in spatial resolution tasks. Vision Research, 39, 293–306.

Attention Makes Moving Objects Be Perceived To Move Faster

More Related Content

Similar to Attention Makes Moving Objects Be Perceived To Move Faster (20)

More from Amy Roman (20)

Recently uploaded (20)

Attention Makes Moving Objects Be Perceived To Move Faster