Attentional Inhibition Has
for Unfamiliar Faces
Mark J. Fenske,1Jane E. Raymond,2Klaus Kessler,3Nikki Westoby,2and Steven P. Tipper2
1Harvard Medical School;2University of Wales Bangor, Bangor, Wales, United Kingdom; and
3Heinrich Heine University D¨ usseldorf, D¨ usseldorf, Germany
ABSTRACT—Visual attention studies often rely on response
time measures to show the impact of attentional facilita-
tion andinhibition. Here weextendtheinvestigationofthe
effects of attention on behavior and show that prior at-
tentional states associated with unfamiliar faces can in-
faces. Participants were shown pairs of face images and
were asked to withhold a response if a transparent stop-
signal cue appeared over one of the faces. This served to
associate the cued face with an inhibitory state. Later,
when asked to make social-emotional choices about these
face pairs, participants chose uncued faces more often
than cued faces as ‘‘more trustworthy’’ and chose cued
faces more often than uncued faces as ‘‘less trustworthy.’’
For perceptual choices, there was no effect of how the
question was framed (which face is ‘‘on a lighter back-
ground’’ vs. ‘‘on a darker background’’). These results
suggest that attentional inhibition can be associated with
socially relevant stimuli, such as faces, and can have
Anumberof visual attention studies have demonstrated that the
time needed to identify a stimulus (or discriminate its location,
color, etc.) can depend on the task relevance of prior events.
Some events, such as the presentation of a cue indicating a
forthcoming target’s location, speed responding (Posner, 1980).
Other events slow responding. For example, when attention is
drawn to the location of an uninformative cue (and then with-
drawn), responses to subsequent stimuli are slower if they ap-
pear at that location than if they appear elsewhere (Posner &
Cohen, 1984). Neural mechanisms that speed or otherwise en-
hance processing of task-relevant information and those that
suppress processing of task-irrelevant information (for review,
see Kastner & Ungerleider, 2000) are thought to subserve at-
tentional facilitation and inhibition processes, respectively.
Although contributing greatly to the empirical basis for the
construct of attention, response time (RT) effects are relatively
small (often measured in tens of milliseconds). They seem un-
likely to be the only, or even the main, behavioral consequence
of a system that involves a large parallel network of brain areas
(e.g., Posner & Petersen, 1990). Indeed, other work has shown
that attention can alter visual sensitivity (Raymond, O’Donnell,
& Tipper, 1998), spatial resolution (Yeshurun & Carrasco,
1998), and even memory for visual events (Kessler & Tipper,
2004). Here we expand the search for the effects of attention by
specifically, we report a study in which we investigated whether
attentional inhibition activated at one point in time can influ-
ence social-emotional appraisal of faces seen later on.
There are two reasons why we thought this influence might
occur. First, neuroanatomical and neuroimaging evidence sug-
gests that the brain systems subserving emotion and attention
1984; Armony & Dolan, 2002; Bush, Luu, & Posner, 2000;
Vuilleumier, Armony, Driver, & Dolan, 2001; Yamasaki, LaBar,
& McCarthy, 2002). These links provide biological plausibility
for our expectation that attention can influence later emotional
behavior. Second, two previous studies have shown that prior
attention can modulate subjective emotional appraisal of ab-
stract stimuli (Fenske, Raymond, & Kunar, 2004; Raymond,
Fenske, & Tavassoli, 2003). In the study by Raymond et al., for
Address correspondence to Mark J. Fenske, MGH Martinos Center
Room 2301, Charlestown, MA 02129; e-mail: firstname.lastname@example.org.
Volume 16—Number 10
Copyright r 2005 American Psychological Society
example, participants evaluated the cheerfulness or dreariness
of colorful abstract images they had just viewed in a simple
distractors in the search task were rated more negatively than
novel images or images seen as targets. Ratings of previous
targets were no different from ratings of novel stimuli. This
pattern of results specifically suggests that inhibitory, though
perhaps not facilitatory, attentional processes are able to exert a
persistent influence on affective responses. Using a more com-
plex visual search paradigm, Fenske et al. obtained additional
evidence that attentional inhibition influences emotional ap-
praisal of abstract stimuli.
Central tothe inhibitoryexplanation for the effectofattention
on emotion is the idea that an attentional state, such as inhibi-
tion, can be associated with a stimulus, stored in memory, and
then later reinstated in a subsequent encounter (Treisman &
DeSchepper, 1996). Recently, Tipper, Grison, and Kessler
(2003) provided evidence that an inhibitory attentional state
produced through suppression of a response to an irrelevant
sudden-onset cue can be stored in memory for a relatively long
period of time. In their task, novel faces were first associated
13 min), when participants were required to detect simple tar-
gets, responses were slower to targets superimposed on previ-
ously cued faces than to targets appearing on previously viewed
but uncued faces. In the present study, we asked if retention of
an inhibitory state, like that evidenced by Tipper et al., could
have consequences for social-emotional appraisal of associated
Our study is based closely on that of Tipper et al. (2003), but
instead of measuring RT for targets superimposed on previously
seen faces, we assessed emotional appraisal of the perceived
attributes of the faces, specifically, their apparent trustworthi-
ness. We contrasted participants’ social-emotional judgments
with their perceptual judgments to assess whether the hypo-
thetical inhibition elicited by an abrupt-onset no-go signal has
a selective impact on emotional response. Participants were
more trustworthy, which was less trustworthy, which was on a
lighter background, and which was on a darker background. We
predicted that evaluative choices involving face pairs seen
previously with a no-go cue over one of the faces would spe-
cifically reflect the negative emotional impact of prior inhibi-
trustworthy and judgments of which face is less trustworthy
should show opposite effects; uncued (noninhibited) faces
more trustworthy, and cued faces should be chosen more often
than uncued faces as being less trustworthy. Moreover, if inhi-
bition specifically affects emotional response, then judgments
of which background is lighter and judgments of which back-
ground is darker should not show such opposite effects.
Twenty-four University of Wales Bangor students (18 females;
mean age 5 24.3 years) with normal or corrected-to-normal
vision participated in exchange for course credit or money. In-
formed consent was obtained.
A Pentium-4 computer, running E-Prime-1.0 (Schneider, Esch-
man, & Zuccolotto, 2002), recorded data and presented stimuli
on a 55.9-cm monitor (100 Hz, 1024 ? 768 resolution). The
viewing distance was 70 cm.
Stimuli were digital color photographs (Art-Explosion 250,000,
1995). Face images (height: 7.51) were frontal views, with
neutral or smiling expressions and visible hair, neck, and eyes.
Half were female and half were male. Pairs of face images,
matched for sex, race, expression, age, gaze direction, and hair
length and color, appeared on a black field (see Fig. 1). Filler
stimuli were neutral scenes of interiors and exteriors measuring
19.81 ? 14.61.
Two types of cues were used: a transparent red or green oval
forexperimental trialsandasingleletter(OorX)forfiller trials.
pair; these cues were sufficiently transparent to allow the un-
were positioned 6.01 above or below the center of the display.
The specific oval color used for each type of signal was coun-
terbalanced across participants: Half saw red as a no-go signal
the left face in half of the displays and over the right face in the
remaining displays. For filler-scene displays, the letter X was
the no-go signal, and the letter O was the go signal.
Design and Procedure
The experiment, approximately 45 min long, included 12
practice and 96 experimental trials with unique face-pair dis-
plays, interleaved with displays from filler trials with unique
scenes. Each trial consisted of a cue sequence of events and a
subsequent evaluation sequence of events. Experimental and
filler trials partially overlapped such that each experimental-
trial cue sequence was immediately followed by a filler-trial cue
sequence, then an experimental-trial evaluation sequence, and
finally a filler-trial evaluation sequence (see Fig. 1). To begin a
a letter for scenes) was then superimposed on the display. The
Volume 16—Number 10
Social-Emotional Consequences of Inhibition
original display remained visible for an additional 1,000 ms,
followed by a 500-ms blank interval.
When the cue was a no-go signal, participants were required
to refrain from pressing any key. For a go signal, participants
were required to quickly press a number-pad key with the index
finger ofeach hand(‘‘4’’ and ‘‘6’’forexperimental trials, ‘‘8’’and
‘‘2’’ for filler trials). Auditory feedback was provided for correct
and incorrect responses. Two thirds of trials were no-go trials,
and one third were go trials. Each cue type appeared equally
often on the left and right face in experimental trials and above
and subsequent evaluation question were randomly determined
for each participant.
Evaluation sequences for experimental trials began with a
display of one of four different questions, asking which person
was ‘‘more trustworthy,’’ ‘‘less trustworthy,’’ ‘‘on a lighter back-
ground,’’ and ‘‘on a darker background.’’ Evaluation sequences
for filler trials used four different questions, asking whether the
following scene was beautiful, was indoors, might be local, or
was bright. Pressing a ‘‘ready’’ key initiated a 500-ms blank
interval and 500-ms fixation cross, followed by a 150-ms pre-
scene (filler-trial evaluation) just viewed in the corresponding
cue sequence (approximately 9 s earlier). A central ‘‘?’’ prompt
then appeared for 1,500 ms. For experimental-trial evaluations,
choices were made during this time by pressing the ‘‘4’’ (‘‘left
face’’) or ‘‘6’’ (‘‘right face’’) key. For filler-trial evaluations,
participants indicated their response using the ‘‘8’’ (‘‘yes’’) or
‘‘2’’ (‘‘no’’) key. If no response was made within 1,650 ms, visual
feedback was given.
analysis. Experimental-trial evaluations were excluded from
Fig. 1. Examples of stimulus displays and the order of events in the cue and evaluation sequences of experimental trials. All stimuli were
experimental-trial evaluation sequence, and finally a filler-trial evaluation sequence. During experimental-trial cue sequences, a red or
green oval cue (depicted here in gray) appeared over one face in the pair; this cue was sufficiently transparent to allow the underlying face
image to be seen clearly. The question asked in each experimental-trial evaluation sequence was selected from the set of questions shown at
the bottom of the figure.
Volume 16—Number 10
M.J. Fenske et al.
analysis if the participant had responded incorrectly during the
corresponding cue sequence or if the evaluation response was
not made within the response interval. These criteria excluded
only 2% of trials.
Choice scores were calculated from the evaluation data by
subtracting the proportion of chosen faces previously superim-
posed with a cue from the proportion not associated with a cue.
Positive scores on this index indicate a tendency to select the
cued face; negative scores indicate a tendency to select the
other, uncued, face. A separate choice score was calculated for
each participant and each question following no-go trials and
following go trials. Averaged choice scores were submitted to
one-group t tests. Differences among choice scores for social-
emotional and perceptual questions were assessed using one-
way repeated measures analyses of variance.
The novel and important finding of our experiment is that on no-
go trials, the tendency to select a previously uncued face, rather
than a cued face, depended significantly on the type of question
asked, F(3, 69) 5 4.32, p < .01, Z25 .158. As can be seen in
Figure 2, following a no-go cue sequence, there was an overall
bias against choosing the previously cued face for three of the
questions, t(23) 5 ?3.71, p < .01. This suggests that the pre-
sentation of a no-go cue on a task-irrelevant object generally
produces avoidance of that object, a result that nicely mirrors
the slowed RTs Tipper et al. (2003) reported for detection of
targets presented on faces previously associated with a no-go
cue. In contrast, the average choice score for go trials did not
differ from 0, t(23) 5 0.551, p > .1, and was not affected by the
question asked, F(3, 69) < 1, Z25 .022, indicating no effect
of go cues on choice. These results suggest that it was not the
presentation of a cue per se that produced the bias against cued
faces on no-go trials.
A remarkable feature of our results is that on no-go trials, the
effect of question valence was different for perceptual versus
social-emotional questions. Figure 2 shows that the two per-
ground is darker?’’) produced the same marginal bias, t(23) 5
?2.03, p < .06, toward the uncued face. This appears to be a
general response bias, because a consideration of the two
questions’ meanings shows that there is no evidence of a ‘‘true’’
perceptual bias. That is, the percentage of uncued faces that
as the percentage of uncued faces (53%) that were chosen as
having the darker background, suggesting that the no-go cues
produced no systematic effect on the perception of lightness in
the face pictures, but did produce a modest response bias (away
from the cued image).
This pattern is in stark contrast to that observed for the op-
positely phrased emotional questions. When participants were
asked to choose the more trustworthy face, the uncued face was
chosen 57% of the time. This effect marginally exceeded the
response bias estimated from the choice-score data for the two
perceptual questions combined, F(1, 23) 5 3.30, p < .09, Z25
participants chose the cued face 55% of the time, clearly
overcoming any general response-related avoidance of the cued
face. The qualitative difference in the choice scores for the two
difference, F(1, 23) 5 12.06, p < .01, Z25 .344, clearly shows
that the no-go cues produced a social-emotional avoidance of
thepreviouslycued face.Thisavoidance cannotbeattributedto
simple color preferences or to learned associations between the
overlay color and an affective response (e.g., red 5 ‘‘avoid’’ or
‘‘do not trust’’), because the different colors of no-go cues were
counterbalanced across participants. A mixed-factors analysis
of variance confirmed that cue color (red vs. green no-go cues)
faces in response to any question (all Fs < 1).
The principal finding of this study is that the inhibition of a
simple action evoked by a sudden-onset cue can later influence
an evaluative judgment concerning an irrelevant stimulus that
Fig. 2. Meanchoicescores(theproportionofcuedfaceschosenminusthe
proportion of uncued faces chosen) for each question in the experimental
below zero indicate a tendency to select the uncued face. Results for no-go
trials are shown in the top panel, and results for go trials are shown in the
bottom panel. Error bars represent standard errors of the means.
Volume 16—Number 10
Social-Emotional Consequences of Inhibition
no-go cues produce a modest response bias to choose an alter-
native to the cue-associated stimulus, a result that is consistent
(Tipper et al., 2003). Kessler and Tipper (2004) observed a
similar bias in their study of memory effects of prior inhibition;
presented over the face that had, in fact, not been cued than to
been cued. However, the particularly remarkable finding in the
present study is that no-go cues have a substantial negative
impact on social-emotional evaluation. They have no impact on
These findings are important because they show that atten-
tional mechanisms can actively influence emotional response
mechanisms. Our results converge with previous demonstra-
patterns (Fenske et al., 2004; Raymond et al., 2003) and show
that attention has important consequences for human behavior
beyond the speeding or slowing of responses. Extending the
demonstration of the relation between attention and emotion to
situations involving the social-emotional appraisal of human
faces is particularly important. Although there is substantial
evidence that emotionally salient stimuli (e.g., angry faces) can
attract (e.g.,Eastwood, Smilek, &Merikle,2001; Vuilleumier &
and attention systems for prioritizing response to stimuli, es-
pecially social responses to other humans, a truly reciprocal
relation should exist between these two systems. The present
data, in combination with the results of previous studies, evi-
dence this clearly.
The bodily state of tensed inaction associated with an inhib-
ited response may also contribute to the negative affect attrib-
uted to faces on which a no-go cue is superimposed, just as
arm extension during stimulus presentation promotes negative
evaluative responses (e.g., Cacioppo, Priester, & Berntson,
1993; Chen & Bargh, 1999). Our findings suggest that such
influences canaffectsocial-emotional response to anunfamiliar
face even after a brief initial experience with it. Although future
research is needed to examine the specific social ramifications
of prior attentional states on human interaction, our findings
suggest that such states can be associated with socially relevant
stimuli and have specific, persistent effects on social-emotional
Acknowledgments—This work was supported by the Eco-
nomic and Social Research Council (ROPA-R022250212 to
Council (5/S1355 to S.P.T.).
Amaral, D.G., & Price, J.L. (1984). Amygdalo-cortical projections in
the monkey (Macaca fascicularis). Journal of Comparative Neu-
rology, 230, 465–496.
Armony, J.L., & Dolan, R.J. (2002). Modulation of spatial attention
by fear-conditioned stimuli: An event-related fMRI study. Neu-
ropsychologia, 40, 817–826.
Art-Explosion 250,000 [CD-ROM]. (1995). Calabasas, CA: Nova De-
Bush, G., Luu, P., & Posner, M.I. (2000). Cognitive and emotional in-
fluences in anterior cingulate cortex. Trends in Cognitive Sciences,
Cacioppo, J.T., Priester, J.R., & Berntson, G.G. (1993). Rudimentary
determinants of attitudes II: Arm flexion and extension have dif-
ferential effects on attitudes. Journal of Personality and Social
Psychology, 65, 5–17.
Chen, M., & Bargh, J.A. (1999). Consequences of automatic evalua-
tions: Immediate behavioral predispositions to approach or avoid
the stimulus. Personality and Social Psychology Bulletin, 25,
Eastwood, J.D., Smilek, D., & Merikle, P.M. (2001). Differential at-
tentional guidance by unattended faces expressing positive and
negative emotion. Perception & Psychophysics, 63, 1004–1013.
by faces expressing emotion: Evidence from flanker tasks. Emo-
tion, 3, 327–343.
Fenske, M.J., Raymond, J.E., & Kunar, M.A. (2004). The affective
consequences of visual attention in preview search. Psychonomic
Bulletin & Review, 11, 1034–1040.
Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening
stimuli draw or hold visual attention in subclinical anxiety?
Journal of Experimental Psychology: General, 130, 681–700.
Kastner, S., & Ungerleider, L.G. (2000). Mechanisms of visual atten-
tion in the human cortex. Annual Review of Neuroscience, 23,
Kessler, K., & Tipper, S.P. (2004). Retrieval of implicit inhibitory
processes: The impact of visual field, object-identity and memory
dynamics. Visual Cognition, 11, 965–995.
Posner, M.I. (1980). Orienting of attention. Quarterly Journal of Ex-
perimental Psychology, 32, 3–25.
Bouma & D. Bouwhuis (Eds.), Attention and performance X:
Control of language processes (pp. 531–556). Hillsdale, NJ: Erl-
Posner, M.I., & Petersen, S.E. (1990). The attention system of the hu-
man brain. Annual Review of Neuroscience, 13, 25–42.
Raymond, J.E., Fenske, M.J., & Tavassoli, N.T. (2003). Selective at-
tention determines emotional responses to novel visual stimuli.
Psychological Science, 14, 537–542.
Raymond, J.E.,O’Donnell, H.L.,&Tipper,S.P.(1998).Priming reveals
attentional modulation of human motion sensitivity. Vision Re-
search, 38, 2863–2867.
Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime user’s
guide. Pittsburgh, PA: Psychology Software Tools.
Tipper, S.P., Grison, S., & Kessler, K. (2003). Long-term inhibition of
return of attention. Psychological Science, 14, 19–25.
Treisman, A., & DeSchepper, B. (1996). Object tokens, attention, and
visual memory. In T. Inui & J. McClelland (Eds.), Attention and
performance XVI: Information integration in perception and com-
munication (pp. 15–46). Cambridge, MA: MIT Press.
Volume 16—Number 10
M.J. Fenske et al.
Vuilleumier,P.,Armony,J.L.,Driver,J.,&Dolan,R.J.(2001).Effectsof Download full-text
attention and emotion on face processing in the human brain: An
event-related fMRI study. Neuron, 30, 829–841.
Vuilleumier, P., & Schwartz, S. (2001). Emotional facial expressions
capture attention. Neurology, 56, 153–158.
Yamasaki, H., LaBar, K.S., & McCarthy, G. (2002). Dissociable pre-
frontal brain systems for attention and emotion. Proceedings of the
National Academy of Sciences, USA, 99, 11447–11451.
Yeshurun, Y., & Carrasco, M. (1998). Attention improves or impairs
visual performance by enhancing spatial resolution. Nature, 396,
(RECEIVED 4/5/04; REVISION ACCEPTED 3/17/05;
FINAL MATERIALS RECEIVED 3/24/05)
Volume 16—Number 10
Social-Emotional Consequences of Inhibition