当前位置：文档库 › Attentional bias to angry faces using the dot-probe task-It depends when you look for it

Attentional bias to angry faces using the dot-probe task-It depends when you look for it

Behaviour Research and Therapy44(2006)1321–1329

Attentional bias to angry faces using the dot-probe task?

It depends when you look for it

Robbie M.Cooper?,Stephen https://www.wendangku.net/doc/f38402994.html,ngton

Department of Psychology,University of Stirling,Stirling,FK94LA,UK

Received30March2005;received in revised form22September2005;accepted3October2005

Abstract

A number of studies using the dot-probe task now report the existence of an attentional bias to angry faces in participants who rate highly on scales of anxiety;however,no equivalent bias has been observed in non-anxious populations,despite evidence to the contrary from studies using other tasks.One reason for this discrepancy may be that researchers using the dot-probe task have rarely investigated any effects which might emerge earlier than500ms following presentation of the threat-related faces.Accordingly,in the current study we presented pairs of face stimuli with emotional and neutral expressions and probed the allocation of attention to these stimuli for presentation times of100and500ms. Results showed that at100ms there was an attentional bias towards the location of the relatively threatening stimulus(the angry face in angry/neutral pairs and the neutral face in neutral/happy pairs)and this pattern reversed by500ms. Comparisons of reaction time(RT)scores with an appropriate baseline suggested that the early bias toward threatening faces may actually arise through inhibition of the relatively least threatening member of a face pair rather than through facilitation of,or vigilance towards,the more threatening stimulus.However the mechanisms governing the observed biases are interpreted,these data provide evidence that probing for the location of spatial attention at500ms is not necessarily indicative of the initial allocation of attention between competing emotional facial stimuli.

Keywords:Attention;Attentional bias;Dot probe;Threat;Face;Emotion

Introduction

Recent studies using the dot-probe task(see below)have established that people who rate highly on scales of trait and state anxiety show an attentional bias towards threat-related stimuli and that control groups who have low scores on these measures do not show such a bias(e.g.Bradley,Mogg,Falla,&Hamilton,1998; Mogg&Bradley,1999).However,given the fact that an angry facial expression is an evolutionarily old and salient signal of threat,it is somewhat surprising that the dot-probe task has failed to reveal an attentional bias in non-anxious individuals to threat-related facial expressions(e.g.Bradley et al.,1997).It is even more surprising given the wealth of evidence using a variety of other paradigms that angry faces do indeed capture

https://www.wendangku.net/doc/f38402994.html,/locate/brat

doi:10.1016/j.brat.2005.10.004

?Corresponding author.Tel.:+441786466375;fax:+441786467641.

E-mail address:r.m.cooper@https://www.wendangku.net/doc/f38402994.html,(R.M.Cooper).

attention in normal,non-anxious populations (Eastwood,Smilek,&Merikle,2001;Fox et al.,2000;Ohman,Lundqvist,&Esteves,2001;White,1995).The main aim of this paper is to establish a reason for this discrepancy.

In its most commonly used form,the dot-probe task involves the presentation of a pair of stimuli for a ?xed period of time (usually 500ms),followed by the appearance of a visual probe in one of the two stimulus locations.Participants then have to perform a task involving the probe (i.e.identi?cation or localisation)and the distribution of spatial attention between the initially presented stimulus pair is inferred by comparing the speed of manual responses to the probe at each of the stimulus locations (following Navon &Margalit,1983;Posner,Snyder,&Davidson,1980).This procedure was initially developed to measure attentional biases to threat in clinically anxious populations (Macleod,Mathews,&Tata,1986);however,its use has extended to many areas where attentional biases have been observed,such as among individuals with eating disorders (Ehrhardt et al.,2003),chronic pain sufferers (Dehghani,Sharpe,&Nicholas,2003),smokers (Hogarth,Mogg,Bradley,Duka,&Dickinson,2003)and various other clinical and non-clinical conditions (Townshend &Duka,2001;Wenzel &Holt,1999).

One reason for the lack of evidence for an attentional bias to threat in non-anxious individuals using the dot-probe task could be because,unlike the other paradigms used (e.g.the visual search task),the stimuli in the dot-probe paradigm are entirely irrelevant to the task the participant has to perform.This makes the dot-probe paradigm a relatively conservative test of whether or not a given stimulus is actually capturing attention (Driver et al.,1999).On the other hand,the failure to ?nd effects in this paradigm could also be attributable to the fact that the dot-probe task,in its standard form,only measures a snapshot of attention,i.e.where attention is allocated when the probe appears.This allows no insight into where attention is allocated before or after the measured snapshot.However,by varying the time between the onset of the stimuli of interest and the appearance of the probe,one can assess the time-course of the deployment of attention over the stimulus pair.Studies using this approach have indicated different patterns of attentional allocation to valenced stimuli (emotional words and faces)at 500ms compared with longer latencies (Bradley et al.,1998;Mogg,Bradley,Miles,&Dixon,2004).In these papers it is assumed that the probe’s appearance at 500ms is a measure of the initial allocation of attention to the stimuli.This assumption is supported by evidence showing that the location of the ?rst shift of attention,as assessed by manual key-press responses in the dot-probe task,is the same location as the initial eye movement made to the stimuli (Bradley,Mogg,&Millar,2000).In line with this evidence most researchers probe for initial allocation of attention at 500ms (e.g.Chen,Ehlers,Clark,&Mansell,2002;Egloff &Hock,2003;Macleod,Mathews,&Tata,1986;Mogg &Bradley,1999).However,while sampling the allocation of attention at the 500ms time window may re?ect the initial orienting of overt attention (i.e.where the eyes are directed),500ms is suf?cient time to allow more than one shift in covert attention (i.e.the shift of attention that precedes an eye movement but can move independently of it)(Posner &Peterson,1990).Indeed,measuring the initial allocation of covert attention may be crucial in the dot-probe task as Bradley et al.(2000)reported that the attentional bias shown with RT data was not dependent on overt orienting;more than half of their participants made eye movements on less than 10%of all trials.

The current study attempts to establish whether the inability to ?nd evidence for attentional capture by angry faces within the general population using the dot-probe task is due to the standard use of a 500ms presentation time.As already mentioned,given the wealth of evidence showing the presence of attentional biases to angry expressions using other paradigms,it is important to examine whether the dot-probe task,as a relatively conservative test of attention,will reveal such a bias when the allocation of attention is measured at an earlier time-frame.It is predicted that if an angry face is paired with a neutral face and presented for 100ms,an attentional bias will be observed towards the location of the angry face.A 100ms presentation time was chosen because it is thought that stimulus-driven covert orienting effects (as demonstrated with abrupt onsets)peak at cue-target stimulus onset asynchronies of around 100–150ms (e.g.Mu ller &Rabbitt,1989).Furthermore 100ms is too short to allow shifts in gaze between the stimuli.During active visual search,the intersaccadic interval is approximately 200–300ms (Kowler,1995).Thus a 100ms presentation time is likely to re?ect automatic initial shifts in attention rather than controlled strategies involved in sustained attention.Of more general interest for researchers using the dot-probe task to study attention,we expect that the predicted bias at 100ms will be absent when attention is sampled again at 500ms.This follows from a number of studies that have found apparent avoidance of threat-related stimuli in non-anxious controls with a 500ms

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton /Behaviour Research and Therapy 44(2006)1321–1329

1322

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton/Behaviour Research and Therapy44(2006)1321–13291323

presentation time(Macleod et al.,1986;Bradley et al.,1997).This outcome would mean that sampling attention at500ms does not represent the initial allocation of attention to an item in a stimulus pair.

A further aim of this study was to look at the role,if any,gaze direction plays in the allocation of attention to facial expressions of emotion displayed in the periphery.Recent work has suggested that the perception of anger is facilitated when it is associated with direct compared with averted gaze(Adams&Kleck,2003).Given this and other?ndings from the visual search task showing that direct gaze draws attention more ef?ciently than averted gaze(von Gru nau&Anston,1995),we expect that the predicted attentional bias towards the location of angry faces will be greater when the angry faces have direct compared with averted gaze. Method

Participants

Sixty undergraduate psychology students1(49female,with a mean age of20.5years)from the University of Stirling took part for course credit.Participants’anxiety levels were not measured.

Materials

Thirty pictures were taken from the Ekman and Friesen(1976)set of emotional expressions.The photographs comprised ten individuals(6females)each posing neutral,happy,and angry expressions with gaze directed into the camera.Happy as well as angry expressions were used to ensure that if the predicted effects with angry faces were observed,they could be attributed to the angry face speci?cally rather than emotional faces more generally.The external features of each of the faces were removed and the internal features were presented in a black rectangular frame.The image size(including the frame)measured8.5cm?3cm on the screen.An averted gaze version of each of the faces was created in Adobe Photoshop by moving the pupils laterally until they were at the corners of the eyes.Half of the gaze-averted faces were made to look to the left and the other half were made to look to the right.This gave a total of60photographs.

Each one of the60photographs was paired in Photoshop with a neutral face of the same identity with averted gaze.It was assumed that a face with a neutral expression and averted gaze would be least likely to attract attention compared with the other emotion and gaze combinations.There is some evidence that direct gaze captures attention(von Gru nau&Anston,1995),and as such pairing each of the faces with one which has averted gaze should allow us to assess the role gaze plays in the allocation of attention to the facial stimuli. The photographs were placed on a white background,one above the other,separated by6cm and were viewed by participants from a distance of approximately60cm.Two versions of each face pair were created with the relative positions switched so that each face could appear in both locations(either top or bottom).Stimuli were presented on an iMac using the experimental software SuperLab.Responses were collected using a Macintosh Keyboard.

Design and procedure

Participants saw a?xation cross for750ms followed by a pair of faces for100or500ms(depending on the group to which they were allocated)which were followed in turn by two dots.Participants’task was to identify whether the dots were horizontally or vertically oriented and that,as such,the faces had nothing to do with the task and so should be ignored.Participants had to press one key(z)when the dots were horizontal and another key(m)when the dots were vertical.The dots remained on the screen for two seconds or until a response was made,whichever came sooner.Participants were given10practice trials to familiarise themselves with the procedure and then240test trials with a break in the middle.These comprised20trials in each of the 1Allocation to the between subjects component of PT was not random.Initially36participants took part in the500ms PT condition and one month later24participants took part in the100ms PT condition.While this does not represent the ideal scenario for looking at differences in the allocation of attention across the two PTs,there is no reason to suspect that the two groups differed from one another on any variable that might be in?uential to the results of the current experiment(e.g.anxiety).

eight conditions:two emotional expressions (angry or happy),two probe locations (the location of the emotional face or the neutral face),and two gaze directions (direct or averted).Of the remaining 80trials,40contained face pairs where both faces were neutral but where the gaze direction differed (e.g.a neutral face with direct gaze paired with another neutral face of the same identity with averted gaze).These were included to examine any attentional effects arising from faces with direct compared to averted gaze.The other 40trials comprised pairs of faces of the same identity,each with neutral expressions and averted gaze.These were included to act as a baseline in order to help determine which mechanisms (e.g.facilitation or inhibition)might be responsible for any observed attentional biases (Koster,Crombez,Verschuere,De Houwer,2004).Test trials were presented in a new random order for each participant.

Results

The data analysis for the dot-probe task was based on reaction times for correct responses.Data from trials with errors were discarded (5%of data in both the two PT conditions)and not analysed further.Data from participants who made more errors than 2.5standard deviations from the mean were discarded (two participants,one from each of the between-subjects conditions,had an overall error rate of 24%).A preliminary analysis revealed no main effect of gaze,nor did the gaze manipulation interact with any of the measures re?ecting attentional bias (all p s 4.2).As such the results presented here will focus on the data collapsed across the factor of gaze.Median RTs were calculated for each of the experimental conditions for the 58remaining participants.The interparticipant means of these medians are displayed in Table 1.

The data were entered into a 2?2?2repeated measures ANOVA with two within-subjects factors of relative probe position (probe appears in location of emotional or neutral face)and type of emotional face (angry,happy),and a between-subjects factor of presentation time (100ms,500ms).This revealed a signi?cant effect of presentation time,F e1;56T?4:110,p ?:047,re?ecting the fact that average response time in the 100ms condition (554ms)was 53ms faster than the average response time in the 500ms condition (607ms).This effect was quali?ed by a three-way interaction between relative probe position,emotional face and presentation time F e1;56T?8:979,p ?:004.To simplify the three-way interaction a bias score was calculated for each type of emotional face pair,modifying the procedure of Macleod and Mathews (1988).Bias score ?PN –PE where PN is the mean RT to probes which appeared in the location of the neutral face and PE is the mean RT to probes which appeared in the location of the emotional face.Positive values re?ect attention towards the emotional face (vigilance)and negative values re?ect attention away from the emotional face (avoidance).These bias scores are displayed in Fig.1.A 2?2repeated measures ANOVA of the bias scores with emotional expression as a within-subjects factor and presentation time as a between-subjects factor yielded a signi?cant emotional expression ?presentation time interaction,F e1;56T?8:979,p ?:004(this result is equivalent to the three-way interaction with RTs).Comparisons of the bias scores against zero (z ero ?no attentional bias)at 100ms showed signi?cant avoidance of the happy face (bias score ?à12,t e22T?2:636,p ?:015,two-tailed)but no signi?cant vigilance of the angry face (bias score ?7,p 4:2).At 500ms there was signi?cant avoidance of the angry face (bias score ?à11,t e34T?2:13,p ?:041,two-tailed)and a trend towards vigilance of the happy face (bias score ?16,t e34T?1:732,p ?:092,two-tailed).Table 1

Mean reaction times (ms)for identifying probes that appear in a location previously occupied by an emotional (happy or angry)or neutral face,with a 100or 500ms presentation time (mean 7standard error)

Face Pair

Angry/Neutral

Happy/Neutral Baseline

Angry Neutral Happy Neutral 100ms

544714552720559?720559?721546723500ms 610718609716598?716602716618719?Signi?cantly different from baseline p o .05.

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton /Behaviour Research and Therapy 44(2006)1321–1329

1324

To clarify the interaction term further,t -tests were carried out on the differences in attentional bias scores between the two PTs.This revealed a signi?cant decrease in bias score for the angry/neutral pair between the short and long PT (angry bias at 100ms PT ?7ms,angry bias at 500ms PT ?à11ms,t e58T?2:192,p ?:033,two-tailed)and a signi?cant increase in the bias score for the happy/neutral pairs between the two PTs (happy bias at 100ms PT ?à12ms,happy bias at 500ms PT ?16ms,t e47T?à2:697,p ?:01,two-tailed).This indicates that the biases for each face pair are different at 100ms compared to 500ms.

Although the above analysis does illustrate the direction of attentional biases,it is unclear whether the bias toward the happy face at 500ms,for example,is a result of facilitation of attention at its location,inhibition of the neutral face location,or both.In order to establish which mechanisms are responsible for the observed attentional biases at both the 100and 500ms PTs,we compared the mean RTs from the four experimental conditions to baseline RTs obtained from trials comprised of neutral–neutral face pairs.Responses faster than the baseline RT would indicate that facilitation (vigilance)was taking place at that location compared to baseline responding.Responses slower than the baseline would indicate inhibition (avoidance)at that location compared to baseline responding (Koster et al.,2004).In the 100ms condition the baseline was 544ms.In happy/neutral trials the mean RT to probes appearing in the location of the happy face (559ms)was signi?cantly slower than the baseline 544ms,t e22T?2:724,p ?:012(two-tailed),whilst probes appearing in the location of the neutral face (546ms)attracted equivalent RTs to baseline (p 4.8).For angry/neutral trials,the mean RT to probes appearing in the location of the angry face (552ms)was no different from baseline (p

4.15),but the mean RT to probes in the location of the neutral face (559ms)was signi?cantly slower than the baseline,t e22T?2:332,p ?:03(two-tailed).To summarise,at the 100ms PT,the use of this baseline measure suggests that with angry/neutral pairs,participants inhibit the neutral face location but do not show vigilance to the angry face.For happy/neutral pairs,however,there is evidence for inhibition of the happy face location but no vigilance for the neutral face.

In the 500ms condition baseline responding was 610ms.For angry/neutral pairs,responses to probes which appeared in the location of the neutral face (598ms)were signi?cantly faster than baseline (610ms),t e35T?3:018,p ?:007(two-tailed),but performance in all other conditions was equivalent to baseline(all p s 4.14).Therefore at the longer PT,there is evidence for facilitation of attention to the neutral face location in angry/neutral pairs when comparing with the baseline,but no other attentional biases.

-25

-20

-15

-10-5

0510152025

M e a n R T B i a s (m s )A v o i d a n c e V i g i l a n c e

Fig.1.Mean attentional bias scores and standard errors for the angry and happy facial expressions with 100and 500ms presentation times.R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton /Behaviour Research and Therapy 44(2006)1321–13291325

Discussion

In this study,we investigated whether previous attempts to observe an attentional bias towards angry faces in the general population using the dot-probe task had failed due to the standard use of a 500ms PT.Indeed,our data indicate that the initial deployment of attention occurs much earlier than 500ms following presentation of a face pair.In fact,at 100ms the pattern of deployment of attention is the opposite of that observed at 500ms.More speci?cally,for the angry/neutral face pair at 100ms,there was a non-signi?cant 7ms bias towards the angry face and by 500ms there was a signi?cant 11ms bias away from the angry face.Although the initial bias was not signi?cant,this re?ects the pattern of data that was predicted for attention to the angry face.For the happy/neutral face pair,at 100ms there was a signi?cant 12ms bias away from the happy face and by 500ms there was a non-signi?cant 16ms bias towards the happy face.Clearly a PT of 100ms is suf?cient for an analysis of the emotional valance of the competing face stimuli and a deployment of attention between the stimuli on the basis of this analysis;a presentation time of 500ms does not,then,re?ect the initial allocation of attention as has been assumed in previous research.

Traditionally,biases towards and away from stimuli in a pair in the dot-probe task have been termed,respectively,vigilance and avoidance (e.g.Bradley et al.,1998;Mogg et al.,2000).For the angry/neutral face pair,the pattern of results just described would be interpreted as initial vigilance to the angry face at 100ms (albeit only a weak trend)followed by avoidance of the stimulus at 500ms.Indeed,the current results showing apparent avoidance of the angry face at 500ms replicate the ?ndings of Bradley et al.(1997).For the happy/neutral face pair,the data are suggestive of initial avoidance of the happy face at 100ms followed by vigilance at 500ms.However,these data can be interpreted in another way if we assume that on each trial attention is initially allocated to the most threatening face on the screen.In angry/neutral trials this would be the angry face but on happy/neutral trials this would be the neutral face.An expressionless face with gaze directed straight at the viewer could be considered neutral in terms of the emotion that is displayed,but the signal that this conveys could be considered hostile or,at the very least,ambiguous.If this assumption is made,then the pattern of data is identical in both the angry/neutral and happy/neutral trials with attention being initially allocated to the relatively threatening face at 100ms and then shifting to the other face by 500ms.

Thus,there are at least three ways of interpreting the advantage in RTs to probes appearing in the location of the neutral face in a neutral/happy pair over RTs to probes appearing in the location of the happy face in the same type of pair:avoidance of the happy face,vigilance to the neutral face (perhaps because it is relatively more threatening),or both.In standard forms of the dot-probe task,there would be no-way of choosing between these interpretations.All other studies assume that the neutral face (or neutral scene or word)plays an entirely passive role and that attention is either allocated to,or away from,the valenced stimulus.As the current study illustrates,it is not clear if differences in attentional bias are as a result of vigilance to one particular stimulus,avoidance of another,or both.One method that has been recently employed to try and tease these issues apart is the introduction of baseline trials which are comprised of neutral/neutral stimulus pairs (Koster et al.,2004;Koster,Verschuere,Crombez,&Van Damme,2005).Response times in experimental conditions are then compared with the baseline response times.If,in a particular condition,responses to identify the target are faster than the baseline,this implies that these responses are facilitated by a mechanism allocating attention to the region.If,however,response times are slower than the baseline,this implies that attention serves to inhibit processing at the location of the probe.

Using the baseline to interpret the nature of the observed biases in the current study radically alters their interpretation.For example,with the angry/neutral stimuli in the 100ms condition,the traditional method of computing attentional bias reveals a bias towards the angry face location (and hence away from the neutral face location)that would normally be interpreted as vigilance to the angry face.However,if the RT scores from these conditions are compared against the neutral/neutral baseline,the data suggests that differences in responses to probes following angry/neutral trials are due to inhibition of the neutral face rather than facilitation of the angry face.This pattern is mirrored in the happy/neutral face trials with the apparent vigilance to the angrier face (neutral)being due to inhibition of the happy face location.Our data therefore suggest that an early attentional bias towards threat may actually arise through the inhibition of the relatively least threatening stimulus rather than through the facilitation of the more threatening member of a stimulus pair.This highlights the problem of interpreting data in this paradigm (see Koster et al.,2004for a more

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton /Behaviour Research and Therapy 44(2006)1321–1329

1326

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton/Behaviour Research and Therapy44(2006)1321–13291327 detailed exploration of these ideas).What is clear is that results that would traditionally have been taken as evidence of avoidance or vigilance have to be reinterpreted when a baseline condition is employed.As such,we echo the sentiments of Koster et al.(2004)when they state that without a baseline measure from which comparisons can be made,inferences concerning attentional mechanisms in the dot-probe task are problematic.

Independent of the mechanism responsible for the effects observed in the current study,the data show a pattern that is consistent with other studies looking at attention to emotional faces in‘normal’populations, namely,facilitated processing of threat-related stimuli(Ohman et al.,2001;Fox et al.,2000).This is in line with recent models of attention to threat,which have suggested attentional bias towards threat is a universal feature of human cognitive processing(Mathews&Mackintosh,1998;Mogg&Bradley,1998)and is not just present in anxious individuals,as was suggested previously(Eysenck,1992).Furthermore,the current results are consistent with evolutionary accounts of threat detection(e.g.Le Doux,1996)which suggest evolutionarily old signals of threat,such as angry facial expressions,should be prioritised for processing because of their relevance to associated actions that are of immediate importance(e.g.?ght or?ight).

It is interesting to note that the allocation of attention in the current study was unaffected by the direction of gaze adopted by the face stimuli.This is somewhat surprising given that a number of recent studies have found that the processing of emotional expressions is modulated by gaze direction(Adams,Gordon,Baird, Ambady,&Kleck,2003;Adams&Kleck,2003;Wicker,Perrett,Baron Cohen,&Decety,2003).However, the critical difference between the current study and the previous work is that the current study was assessing which attributes of a face attract attention,whereas the previous work was measuring how the processing of the face occurs once the face has been attended.These are two entirely different levels of analysis.It may be that the interaction of expression and gaze requires focal attention whereas the work reported here was examining pre-attentive processing of faces.Allied to this,the previous studies required participants to make an explicit judgment about the face,whereas in the current study the faces were irrelevant to the task.Either of these differences,or a combination of the two,could account for the discrepancy between the current results and these others reporting an interaction between gaze and emotion processing.Further research is needed to address this issue.

In more general terms,the current results clearly demonstrate different patterns of attentional deployment at100ms compared to500ms.Far from representing the initial allocation of attention the results at the500ms PT are signi?cantly different to those observed at the100ms PT.This difference cannot be accounted for by assuming that no bias is present at100ms and that one emerges by500ms.At the100ms PT,a signi?cant bias away from the location of the happy face and towards the location of the neutral face was recorded which did not remain at the500ms PT.This means that a different pattern of bias was observed at the two presentation times.These results undermine the widely held assumption in the dot-probe literature that sampling attention at500ms is a re?ection of the initial orienting of attention towards the stimulus(e.g.Macleod,Mathews,& Tata,1986;Mogg&Bradley,1999;Chen,Ehlers,Clark,&Mansell,2002;Egloff&Hock,2003).This, combined with recent work showing different patterns of attentional deployment at100ms compared to 500ms,and500ms compared to1250ms for threat-related pictures(Koster et al.,2005),suggests that researchers looking to investigate attentional biases in both clinical and non-clinical populations should use a number of presentation times,including one shorter than the standard500ms condition(e.g.100ms). Measuring the initial allocation of spatial attention earlier than500ms would be in line with much research using other paradigms to study spatial attention(e.g.Posner et al.,1980;Langton&Bruce,1999;Friesen& Kingstone,1998).

One possible limitation with the current study is that the participants were primarily female.There is recent evidence that females show greater visual cortical activity to negatively valenced pictures,suggesting that females may be particularly sensitive to such information(Sabatinelli,Flaisch,Bradley,Fitzsimmons,&Lang, 2004).While the link between this evidence and the allocation of visual attention to threat-related stimuli is not known,it seems appropriate to take this into account in the design of future studies.

In sum,interpretation of observed attentional bias is ambiguous in the standard version of the dot-probe task.The use of baseline trials in the current experiment suggests bias to initial threat arises from inhibition of the least threatening stimulus,rather than vigilance to the threat.Thus the inclusion of baseline trials is one way of elucidating the attentional mechanisms that are responsible for any observed biases.Furthermore,

sampling attention to facial expressions at 100ms in the dot-probe task can reveal a different pattern of attentional allocation than is observed at 500ms.More generally,this is good evidence that sampling the location of attention at 500ms in the dot-probe task does not necessarily represent its deployment at 100ms.As such,researchers wishing to sample the initial allocation of attention to competing stimuli using this paradigm should take this into account in the design of their experiments.

Acknowledgements

We are grateful to Alan Kingstone and to two anonymous reviewers for their helpful comments on earlier versions of this manuscript.Robbie Cooper is funded by a Ph.D.studentship from the ESRC.References

Adams,R.B.,Gordon,H.L.,Baird,A.A.,Ambady,N.,&Kleck,R.E.(2003).Effects of gaze on amygdala sensitivity to anger and fear faces.Science ,300,1536.

Adams,R.B.,&Kleck,R.E.(2003).Perceived gaze direction and the processing of facial displays of emotion.Psychological Science ,14(6),644–647.

Bradley,B.P.,Mogg,K.,Falla,S.J.,&Hamilton,L.R.(1998).Attentional bias for threatening facial expressions in anxiety:Manipulation of stimulus duration.Cognition &Emotion ,12(6),737–753.

Bradley,B.P.,Mogg,K.,&Millar,N.H.(2000).Covert and overt orienting of attention to emotional faces in anxiety.Cognition &Emotion ,14(6),789–808.

Bradley,B.P.,Mogg,K.,Millar,N.,Bonham Carter,C.,et al.(1997).Attentional biases for emotional faces.Cognition &Emotion ,11(1),25–42.

Chen,Y.P.,Ehlers,A.,Clark,D.M.,&Mansell,W.(2002).Patients with generalized social phobia direct their attention away from faces.Behaviour Research and Therapy ,40(6),677–687.

Dehghani,M.,Sharpe,L.,&Nicholas,M.K.(2003).Selective attention to pain-related information in chronic musculoskeletal pain patients.Pain ,105(1–2),37–46.

Driver,J.,Davis,G.,Ricciardelli,P.,Kidd,P.,Maxwell,E.,&Baron Cohen,S.(1999).Gaze perception triggers re?exive visuospatial orienting.Visual Cognition ,6(5),509–540.

Eastwood,J.D.,Smilek,D.,&Merikle,P.M.(2001).Differential attentional guidance by unattended faces expressing positive and negative emotion.Perception and Psychophysics ,63(6),1004–1013.

Egloff,B.,&Hock,M.(2003).Assessing attention allocation toward threat-related stimuli:A comparison of the emotional Stroop task and the attentional probe task.Personality-and-individual-differences ,35(2),475–483.

Ehrhardt,J.,Weiss,T.,Musial,F.,Zimmanyi,A.,Englert,E.,Blanz,B.,et al.(2003).Attentional bias of patients with anorexia nervosa using a dot-probe paradigm.Journal of Psychophysiology ,17(3),147.

Ekman,P.,&Friesen,W.V.(1976).Pictures of facial affect .Palo Alo,CA:Consulting Psychologists Press.

Eysenck,M.W.(1992).Anxiety:The cognitive perspective .Hove:Erlbaum.

Fox,E.,Lester,V.,Russo,R.,Bowles,R.J.,Pichler,A.,&Dutton,K.(2000).Facial expressions of emotion:Are angry faces detected more ef?ciently?Cognition and Emotion ,14(1),61–92.

Friesen,C.K.,&Kingstone,A.(1998).The eyes have it!Re?exive orienting is triggered by nonpredictive gaze.Psychonomic Bulletin and Review ,5(3),490–495.

Hogarth,L.C.,Mogg,K.,Bradley,B.P.,Duka,T.,&Dickinson,A.(2003).Attentional orienting towards smoking-related stimuli.Behavioural Pharmacology ,14(2),153–160.

Koster,E.H.W.,Crombez,G.,Verschuere,B.,&De Houwer,J.(2004).Selective attention to threat in the dot probe paradigm:Differentiating vigilance and dif?culty to disengage.Behaviour Research and Therapy ,42(10),1183–1192.

Koster,E.H.W.,Verschuere,B.,Crombez,G.,&Van Damme (2005).Time course of attention for threatening pictures in high and low trait anxiety.Behaviour Research and Therapy ,43(8),1087–1098.

Kowler,E.(1995).Eye movements.In S.M.Kosslyn,&D.N.Osheron (Eds.),Visual Cognition .Cambridge,MA:MIT.

Langton,S.R.H.,&Bruce,V.(1999).Re?exive visual orienting in response to the social attention of others.Visual Cognition ,6(5),541–567.

Le Doux,J.E.(1996).The emotional brain .New York:Simon &Schuster.

MacLeod,C.,&Mathews,A.(1988).Anxiety and the allocation of attention to threat.Quarterly Journal of Experimental Psychology ,40,653–670.

MacLeod,C.,Mathews,A.,&Tata,P.(1986).Attentional bias in emotional disorders.Journal of Abnormal Psychology ,95(1),15–20.Mathews,A.,&Mackintosh,B.(1998).A cognitive model of selective processing in anxiety.Cognitive Therapy and Research ,22,539–560.Mogg,K.,&Bradley,B.P.(1998).A cognitive-motivational analysis of anxiety.Behaviour Research and Therapy ,36(9),809–848.Mogg,K.,&Bradley,B.P.(1999).Orienting of attention to threatening facial expressions presented under conditions of restricted awareness.Cognition and Emotion ,13(6),713–740.

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton /Behaviour Research and Therapy 44(2006)1321–1329

1328

R.M.Cooper,https://www.wendangku.net/doc/f38402994.html,ngton/Behaviour Research and Therapy44(2006)1321–13291329 Mogg,K.,Bradley,B.P.,Dixon,C.,Fisher,S.,Twelftree,H.,&McWilliams,A.(2000).Trait anxiety,defensiveness and selective processing of threat:An investigation using two measures of attentional bias.Personality and Individual Differences,28(6),1063–1077. Mogg,K.,Bradley,B.P.,Miles,F.,&Dixon,R.(2004).Time course of attentional bias for threat scenes:Testing the vigilance–avoidance hypothesis.Cognition&Emotion,18(5),689–700.

Mu ller,H.J.,&Rabbitt,P.M.A.(1989).Re?exive and voluntary orienting of visual attention:Time course of activation and resistance to interruption.Journal of Experimental Psychology:Human Perception and Performance,15,315–330.

Navon,D.,&Margalit,B.(1983).Allocation of attention according to informativeness in visual recognition.Quarterly Journal of Experimental Psychology,35A,497–512.

Ohman,A.,Lundqvist,D.,&Esteves,F.(2001).The face in the crowd revisited:A threat advantage with schematic stimuli.Journal of Personality and Social Psychology,80(3),381–396.

Posner,M.I.,&Petersen,S.E.(1990).The attention system of the human brain.Annual Review of Neuroscience,13,25–42. Posner,M.I.,Snyder,C.R.,&Davidson,B.J.(1980).Attention and the detection of signals.Journal of Experimental Psychology: General,109,160–174.

Sabatinelli,D.,Flaisch,T.,Bradley,M.M.,Fitzsimmons,J.R.,&Lang,P.J.(2004).Affective picture perception:Gender differences in visual cortex?Neuroreport,15(7),1109–1112.

Townshend,J.M.,&Duka,T.(2001).Attentional bias associated with alcohol cues:Differences between heavy and occasional social drinkers.Psychopharmacology,157(1),67–74.

von Gru nau,M.,&Anston,C.(1995).The detection of gaze direction:A stare-in-the-crowd effect.Perception,24(11),1297–1313. Wenzel,A.,&Holt,C.S.(1999).Dot probe performance in two speci?c phobias.British Journal of Clinical Psychology,38,407–410. White,M.(1995).Preattentive analysis of facial expressions of emotion.Cognition and Emotion,9(5),439–460.

Wicker,B.,Perrett,D.I.,Baron Cohen,S.,&Decety,J.(2003).Being the target of another’s emotion:A PET study.Neuropsychologia, 41(2),139–146.